Review

Received: 10 March 2010 Revised: 6 August 2010 Accepted: 6 August 2010 Published online in Wiley Online Library: 2 September 2010

(wileyonlinelibrary.com) DOI 10.1002/jsfa.4158

Nutrition facts and functional potential of quinoa (Chenopodium quinoa willd.), an ancient Andean grain: a review
Antonio Vega-Galvez,a Margarita Miranda,a Judith Vergara,a Elsa Uribe,a b and Enrique A Martnezc Luis Puente
Abstract
Quinoa, Chenopodium quinoa Willd., is an Amaranthacean, stress-tolerant plant cultivated along the Andes for the last 7000 years, challenging highly different environmental conditions ranging from Bolivia, up to 4.500 m of altitude, to sea level, in Chile. Its grains have higher nutritive value than traditional cereals and it is a promising worldwide cultivar for human consumption and nutrition. The quinoa has been called a pseudo-cereal for botanical reasons but also because of its unusual composition and exceptional balance between oil, protein and fat. The quinoa is an excellent example of functional food that aims at lowering the risk of various diseases. Functional properties are given also by minerals, vitamins, fatty acids and antioxidants that can make a strong contribution to human nutrition, particularly to protect cell membranes, with proven good results in brain neuronal functions. Its minerals work as cofactors in antioxidant enzymes, adding higher value to its rich proteins. Quinoa also contains phytohormones, which offer an advantage over other plant foods for human nutrition. c 2010 Society of Chemical Industry Keywords: Andean crops; functional foods; human nutrition; physiologically active compounds; quinoa; stress tolerance

INTRODUCTION
A strong earthquake and tsunami recently impacted our Chilean territory.1 This strike left two million people without shelter and food supplies.2 This situation is imposing a large-scale challenge for good-quality food to be readily available. One alternative source of staple food is Chenopodium quinoa Willd., a crop present in Chile, although poorly known, and it emerges as a good food candidate due to its exceptional nutritive value but also due to the strong tolerance to stressing abiotic conditions. The genus Chenopodium is distributed worldwide and includes 250 species. This review focuses on the nutritional and functional properties of Chenopodium quinoa, which is a tetraploid species, a close relative of beets and amaranth that originated in the Andean region of Bolivia and Peru.3,4 It has been cultivated in this area for the last 50007000 years5,6 and from there it was transmitted by livestock migrations and traded to other ancient cultures to the northern (Venezuela) and southern extremes of South America, namely Argentina and Chile.7 This is why it is known with different local names, according to voices of different cultures such as tupapa supha in Aymar , suba in Chibcha, ayara in Quechua, dawe a in Mapudungun (southern Chile) or just quinoa or quinua. This plant was called by the Incas the mother grain and it was given a sacred status, a gift from their gods. After the Spanish conquest, it remained only where Europeans could not arrive and introduce grains such as wheat, rye and oat (Altiplano in the High Andes above 3500 m above sea level (a.s.l.)) or in isolated regions where roads are cut off in winter or where the ancient cultures still remain strong and attached to their agricultural practices and to their traditional food consumption habits (aymaras in the northern Chilean Altiplano, isolated farmers of the coast of central Chile

and within the mapuches people in southern Chile). Concerning the more accessible arable lands, the European-introduced crops replaced quinoa.5,7,8 This separation and subsequent isolation determined a strong pattern of genetic differentiation: the high Andes ecotypes are genetically different from the southern ones, as detected by microsatellites, a kind of highly polymorphic molecular marker.9 About 3000 varieties are conserved in South American germplasm banks assuring conservation and characterization, and opening possibilities for informed interchange of seed materials. Quinoa has been authorized to be sown in Europe, North America, Asia and Africa.10 Particularly in Europe a project was approved in 1993 entitled Quinoa: a multipurpose crop for the European Community, for agriculture diversication.11,12 While quinoa is an ancient crop, available technical information regarding the properties of chemical composition and functional properties is limited. Therefore, this work is an updated review of the chemical composition, physiologically active compounds and some functional properties of Chenopodium quinoa that gives this grain outstanding potential in human nutrition. Its varied

Correspondence to: Antonio Vega-G lvez, Department of Food Engineering, a Universidad de La Serena, Av. Raul Bitr n s/n, Box 599, La Serena, Chile. a E-mail: avegag@userena.cl

a Department of Food Engineering, Universidad de La Serena, La Serena, Chile b Department of Food Science and Chemical Technology, Universidad de Chile, 7500906, Santiago, Chile.

2541

c Center for Advanced Studies in Arid Zones, CEAZA, Universidad de La Serena, La Serena, Chile

J Sci Food Agric 2010; 90: 25412547

www.soci.org

c 2010 Society of Chemical Industry

www.soci.org

A Vega-G lvez et al. a

Table 1. Main uses of quinoa Main uses Foods and drinks Animal food Medicine Component implied Vitamins Proteins Vitamins Proteins Immune system Skin applications Circulatory applications Insects Plant organ Seeds and leaves Seeds and leaves Whole plant Whole plant Leaves and seeds Leaves and seeds Leaves and seeds Leaves and seed coat

Repellent FAO.26

Figure 1. Quinoa panicules (Chenopodium quinoa Willd).

nutritional properties are better understood when the botanical and environmental diversity of quinoa is also known. We start our review with a brief description on its great adaptation to abiotic stress, a particularly relevant aspect in a world with strong trends towards increased soil degradation, desertication and critical climatic change.

BOTANICAL AND AGRONOMICAL DIVERSITY OF QUINOA IN VARIED AND STRESSING ENVIRONMENTAL CONDITIONS
Quinoa is a plant that produces grains even if cultivated up to 4500 m a.s.l. and with higher nutritive value than traditional cereals,8 as for instance Amarilla de Marangani and Blanca de Junn, two commercial varieties grown in greater proportion in the Andes of southern Peru.13 Most varieties of quinoa commonly differ in the morphology, phenology and the chemical composition of the tissues.14 The quinoa (a dicot plant) is not a true grain, like typical cereal (monocot) grains, it is rather a fruit, so that it has been called a pseudo-cereal and even a pseudo-oilseed. This is also because of its unusual composition and exceptional balance between oil, protein and fat.15 Quinoa, according to sowing density, can grow from 1 to 3 m high. The seeds can germinate very fast, i.e. in a few hours after having been exposed to moisture. The roots can reach a depth of up to 30 cm if sown deep in the soil. The stem is cylindrical, 3.5 cm in diameter; it can be either as a straight stem or branched and its color is variable. Depending on the variety, it changes from white, yellow or light brown to red. Leaves are shaped like a goose foot. The owers are incomplete and do not have petals. Quinoa has both hermaphrodite owers, located at the distal end of a group, and female owers, located at the proximal end(Fig. 1).16 Quinoa seeds are round and attened, and they may measure from about 1.5 mm in diameter to 4 mm; about 350 seeds weigh 1 g.17 Seed size and color are variable.18 Seed colors go from white to grey and black, potentially having tones of yellow, rose, red and purple and violet, often with very colorful mixes in the same panicule, where black is dominant over red and yellow, which in turn are dominant to white seed color (as seen in Fig. 1).19 The classication of quinoa was rst made from the color of the plant and fruits. Subsequently, it was based on the morphological

types of the plant. Despite the wide variation observed, quinoa is considered to be one single species.16 The cultivation of quinoa is related to crop rotation with potatoes, also a crop of Andean origin. This is a common practice, which improves quinoa yield and preserves soil fertility. Moreover, the biological cycle of several pathogenic microorganisms is broken down.16 The cultivation cycle lasts 8 months in the high Andes but it can be as short as 4 months in arid central Chile.20 It is sown in November in the Altiplano, close to the Equator (close to 12 h daylight) and from September to August in the lowlands of more southern latitudes (longer spring and summer days). Maturation and harvest, according to daylength, is done in May in the Altiplano and from February to March in the centersouth of Chile. Here, some ecotypes could attain maturity and seed production under irrigation equivalent to only 50 mm of rainfall per season, which is an extremely low irrigation for any crop species.21 It also seems to have exceptional physiological adaptations for high water use efciency under stomatal closure21 besides efcient roots for water capture, as earlier pointed out by Wood.22 In arid regions the addition of organic matter also increases water use efciency and grain yields.21 Strong tolerance has been also demonstrated for other stressing conditions such as salty soils and cold climate.12,23 Quinoa can be grown on various types of soils, including marginal soils,8 under a wide range of acid/alkaline conditions (from pH 6.0 to 8.5). The plant is not affected from around 1 C. However, it tolerates high temperatures up to 35 C. Quinoa is resistant to freezing temperatures if the frost occurs before owering. However, if the frost occurs after owering, signicant damage may affect the plant. As mentioned above, quinoa is a drought-tolerant crop having low water requirements, although yield is signicantly affected by irrigation.24 It is able to develop even in regions where the annual rainfall is in the range of 200400 mm,16 but it has been proven that it can be grown in southern Chile with an annual precipitation as high as 3000 mm.20 Although having good a response in poor soils, quinoa does respond well to nitrogen fertilization. Thus nitrogen signicantly increases seed production and protein content.25

BIOCHEMICAL AND NUTRITIONAL COMPOSITION OF QUINOA

Consumption of seeds is the most common use of quinoa (Table 1) and the review will be focused on its composition (Table 2). However, consumption of sprouts is becoming increasingly

2542

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 25412547

Nutrition and functional potential of quinoa

www.soci.org

Table 2. Proximate analysis of quinoa (g 100 g1 fresh weight) References Component Protein Fat Ash Carbohydrate Crude ber Kozio27 16.5 6.3 3.8 69.0 3.8 Wright et al.28 16.7 5.5 3.2 74.7 10.5 De Bruin29 15.6 7.4 3.0 69.7 2.9 Dini et al.30 12.5 8.5 3.7 60.0 1.92

Table 3. Essential amino acid prole (g 100 g1 protein) References Amino acid His Ile Leu Met + Cys Phe + Tyr Thr Val Lys Trp Kozio27 3.2 4.4 6.6 4.8 7.3 3.8 4.5 6.1 1.2 Dini et al.30 2.0 7.4 7.5 4.5 7.5 3.5 6.0 4.6 ND Repo-Carrasco et al.33 2.7 3.4 6.1 4.8 6.2 3.4 4.2 5.6 1.1 Wright et al.28 3.1 3.3 5.8 2.0a 6.2 2.5 4.0 6.1 ND Gonz lez a et al.34 ND ND ND 2.4a ND ND ND 6.6 1.1

popular among people interested in improving and maintaining their health status by changing dietary habits. The seeds and sprouts are both excellent examples of functional food, dened as lowering the risk of various diseases and/or exerting healthpromoting effects, in addition to its nutritive value.31 Most of the recently published papers are focused mainly on studies of typical sprouts such as buckwheat, broccoli, mung bean, and soybean, which are already readily available on the market. The sprouts of amaranth and quinoa are new vegetables, which can be used in vegetarian nutrition and as a common diet too.31 Todays health-conscious consumers are illustrating a preference towards value-added products. The opportunity to supplement or completely replace common cereal grains (corn, rice and wheat) with a cereal of higher nutritional value (such as quinoa) is inherently benecial to the public interest.32 We will review the known composition and nutritional facts reported for quinoa, before describing the potential for functional properties and for human health, particularly for certain consumers (the elderly, children, high-performance athletes, diabetics, celiacs, people who are gluten or lactose intolerant). Proteins Protein nutritional quality is determined by the proportions of essential amino acids, which cannot be synthesized by animals and hence must be provided in the diet. If only one of these amino acids is limiting, the others will be broken down and excreted, resulting in poor growth of livestock and humans and loss of nitrogen in the diet. Ten amino acids are strictly essential: lysine, isoleucine, leucine, phenylalanine, tyrosine, threonine, tryptophan, valine, histidine and methionine, all of which are present in quinoa (Table 3), providing it with a similar value to casein, the protein of milk.33,35 Koziol27 showed that protein content in quinoa grain ranges from 13.8% to 16.5%, with an average 15%. Wright et al.28 reported a protein content of 14.8% and 15.7% for sweet and bitter quinoa, respectively, from Bolivia. De Bruin29 studied the protein content of four genotypes of quinoa, reporting a range of 12.915.1%. According to values indicated by FAO/WHO/UNU,36,37 quinoa protein can supply around 180% of the histidine, 274% of the isoleucine, 338% of the lysine, 212% of the methionine + cysteine, 320% of the phenylalanine + tyrosine, 331% of the threonine, 228% of the tryptophan and 323% of the valine recommended in protein sources for adult nutrition. In addition, the sulfur-containing amino acids cystine, and methionine are found in concentrations that are unusually high compared to other plants,38 probably due to the type of land (volcanic) where this plant originated. The content of essential amino acids in quinoa (Table 3) is higher than in common cereals.28,39 Mahoney et al.,40 working with Bolivian quinoa, concluded that protein contained high amounts of lysine

His, histidine; Ile, isoleucine; Leu, leucine; Met + Cys, methionine + cystine; Phe + Tyr, phenylalanine + tyrosine; Thr, threonine; Val, valine; Lys, lysine; Trp, tryptophan. ND, not detected. a Only methionine reported.

and methionine even though there is considerable variation between these varieties in the contents of such amino acids. Dini et al.,30 using decorticated quinoa, found that the composition of quinoa is nutritionally comparable or superior to other commonly consumed cereals. When extracted, quinoa proteins solubility could be improved by enzymatic hydrolisis.41 Quinoa is also considered as one of the best leaf protein concentrate sources and so has potential as a protein substitute for food and fodder and in the pharmaceutical industry.42 Carbohydrates Starch, the major biopolymeric constituent of plants (grains, seeds and tubers) occurs typically as granular forms of various shapes and sizes.43 Starch provides the major source of physiological energy in the human diet and accordingly it is classied, in general, as available carbohydrate.44 In quinoa, starch is the most important carbohydrate in all grains, making up approximately 58.164.2% of the dry matter, according to studies of Repo-Carrasco et al.,33 of which 11% is amylose.45,46 Granules of quinoa starch have a polygonal form with a diameter of 2 m, being smaller than starch of the common grains. The extremely small size of the starch granule can be benecially exploited by using it as a biodegradable ller in polymer packaging. Its excellent freezethaw stability makes it an ideal thickener in frozen foods and other applications where resistance to retrogradation is desired.47 In addition, quinoa our contains high percentages of D-xylose and maltose, and low contents of glucose and fructose, which allows its use in malted drink formulations.48 Also, its content of D-ribose and D-galactose and maltose would result in a low fructose glycemic index. RepoCarrasco et al.33 reported for quinoa 1.70 mg 100 g1 of glucose, 0.20 mg 100 g1 of fructose, 2.90 mg 100 g1 of saccharose and 1.40 mg 100 g1 of maltose. Minerals Quinoa has a high content of calcium, magnesium, iron, copper and zinc.27,29,33,49 Many minerals in quinoa are found at concentrations greater than that reported for most grain crops. Providing they are found in bioavailable forms, calcium, magnesium and potassium are found in sufcient quantities for a balanced human diet38

2543

J Sci Food Agric 2010; 90: 25412547

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org

A Vega-G lvez et al. a

Table 4. Mineral composition (mg kg1 dry weight) Minerals References Ca P Mg Fe Zn K Cu

Table 6. Unsaturated fatty acid (g 100 g1 of oil extract) Fatty acid Reference Kozio27 Repo-Carrasco et al.33 Ruales and Nair17 Oleic 23.3 26.0 24.8 Linoleic 53.1 50.2 52.3 Linolenic 6.2 4.8 3.9

Kozio27 1487 3837 2496 132 940 1400 2700 168 Repo-Carrasco et al.33 Ruales and Nair17 874 5300 260 81 Bhargava et al.10 1274 3869 ND 20 Konishi et al.50 863 4110 5020 150 Dini et al.30 275 4244 ND 26 565 4689 1760 14 Sanders52 Gonz lez et al.34 a 1020 1400 ND 105 ND, not detected.

44 9267 51 48 ND 37 36 12000 10 48 6967 ND 40 7320 ND 27.5 75 ND 28 11930 2 ND 8225 ND

Table 5. Vitamin concentration (mg 100 g1 dry weight) References Vitamin Ascorbic acid (C) -Tocoferol (E) Thiamin (B1 ) Riboavin (B2 ) Niacin (B3 ) ND,not detected. Kozio27 4.0 5.37 0.38 0.39 1.06 Ruales and Nair17 16.4 2.6 0.4 ND ND

appreciable amounts of thiamin (0.4 mg 100 g1 ), folic acid (78.1 mg 100 g1 ) and vitamin C (16.4 mg 100 g1 ). Kozio27 compared the vitamin contents of quinoa with some cereals (rice, barley and wheat) and reported that quinoa contains substantially more riboavin (B2 ), -tocopherol (vitamin E) and carotene than those cereals. In terms of a 100 g edible portion, quinoa supplies 0.20 mg vitamin B6 , 0.61 mg pantothenic acid, 23.5 g folic acid and 7.1 g biotin.10 Repo-Carrasco et al.33 also reported that quinoa is rich in vitamin A, B2 and E. The content of vitamin E in quinoa is important since this vitamin acts as an antioxidant at the cell membrane level, protecting the fatty acids of the cell membranes against damage caused by free radicals.33 Lipids Oil content in quinoa ranges from 1.8% to 9.5%, with an average of 5.07.2%, which is higher than that of maize (34%).27 Numerous fatty acids are synthesized by the human body, and these are known as non-essential fatty acids because they are not essentially needed in the diet.54 However, because the body cannot produce all the types of fatty acids it requires, some must come from the diet; these fatty acids are called essential fatty acids or EFAs (Table 6). The EFAs are metabolized to longer-chain fatty acids of 20 and 22 carbon atoms.37 There are two known families of EFAs: omega-3 (-3) and omega-6 (-6).55 Linoleic acid is metabolized to arachidonic acid and linolenic acid to eicosapentaenoic acid (EPA) and docosahexaenoic acid (DHA). Linoleic acid is one of the most abundant polyunsaturated fatty acids identied in quinoa; polyunsaturated fatty acids have several positive effects on cardiovascular disease and improved insulin sensitivity.37 The reported total lipid content in quinoa is 14.5% with an unsaturated level of about 70%, having linoleic and oleic acids in percentages of 38.9% and 27.7% respectively,30 while Ahamed et al.47 reported in another study that quinoa fat had a high content of oleic acid (24%) and linoleic acid (52%). All fatty acids present in quinoa are well protected by the presence of vitamin E, which acts as a natural antioxidant.56 Repo-Carrasco et al.33 reported from Peruvian quinoa the highest percentage of fatty acids being 50.2% for linoleic acid (-6), 4.8% of linolenic acid (-3).

(Table 4). For instance, Ruales and Nair39 and Ahamed et al.47 reported that iron, calcium and phosphorus levels are higher than those of maize and barley: iron was 81 mg kg1 and calcium was 874 mg kg1 . It also has about 0.26% of magnesium in comparison to 0.16% of wheat and 0.14% of corn. Schlick and Bubenheim,38 comparing quinoa from different sources (USA, Peru, Bolivia and Chile), reported that the mineral concentrations for quinoa seem to vary dramatically. This may occur due to the soil type and mineral composition of the region and/or fertilizer application. With respect to the distribution of minerals within the grain, Konishi et al.50 used scanning electron microscopy with energydispersive X-ray detection on seed with and without epicarp, nding that minerals like P, K and Mg were located in the embryo, while Ca and P in the pericarp were associated with pectic compounds of the cell wall. Thus abrasive procedures to remove saponins might cause losses of 40% and 10%, respectively. Sulfur is found uniformly distributed within the embryo. The iron has been reported as highly soluble and thus could be easily available to anemic populations.51 Vitamins Vitamins are compounds essential for the health of humans and animals; according to their solubility they are divided into two groups: hydro- and lipo-soluble. Traditionally, vitamins have been among the most widely applied chemical agents to enhance the nutritional values of food products. Some vitamins may also help to lower the levels of toxic compounds formed in chemical reactions such as the Maillard reaction.53 Table 5 shows the main vitamins found in quinoa. The quinoa is found to be rich in -carotene and niacin. Ruales and Nair39 have reported

ANTIOXIDANT ACTIVITY
Recently, much attention has been given to naturally occurring antioxidants, which may play an important role in inhibiting both free radicals and oxidative chain reactions within tissues and membranes.57 Therefore, the evaluation of antioxidant activities of extracts and fractions is considered an important step prior to the isolation of antioxidant phytochemicals they contain. Antioxidants are compounds that can delay or inhibit the oxidation of lipids or other molecules by inhibiting the initiation or propagation

2544

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 25412547

Nutrition and functional potential of quinoa of oxidizing chain reactions. When added to foods, antioxidants minimize rancidity, delay the formation of toxic oxidation products, maintain nutritional quality and increase shelf life.52 Pa ko et al.31 showed that pseudocereal seeds and sprouts s show relatively high antioxidant activity. Nsimba et al.57 evaluated the antioxidant activity of various extracts from quinoa (Japan) and from its relative Amaranthus, nding different values among the samples. In addition, Pa ko et al.31 reported that quinoa s presents higher antioxidant activity than amaranth using different methodologies: ferric reducing/antioxidant potential (FRAP), 2,2 -azinobis(3-ethylbenzothiazoline 6-sulfonate) (ABTS) and 2, 2-diphenyl-2-picryl-hydrazyl (DPPH). Antioxidant activity of quinoa might be of particular interest to medical researchers and needs further attention regarding its utilization as a natural potent antioxidant.10

www.soci.org

SUMMARY OF QUINOA FUNCTIONAL POTENTIAL FOR HUMAN DIET

Some Leguminosae in combination with some cereals might improve proteic proles of high-quality foods due to amino acid compensation, a good strategy also used with quinoa food for children in the Andean region. Cerezal et al.74 designed a food for 3- to 5-year-old children, with high amino acid content (3540% of daily requirements). Nsimba et al.57 used quinoa and amaranth in products such as bread, pastas and baby foods. Lorenz and Coulter45 evaluated quinoa our extrusion mixed with maize grits to develop snacks with moderate acceptance. Moreover, there is evidence concerning other physiologically active compounds present in quinoa seeds such as tannins17 and betaines.75 Tannins, which are polyphenolic compounds, form complexes with dietary proteins and also with digestive enzymes.76 In addition to proteins, humans require minerals for their normal life processes, particularly essential minerals, those necessary to support adequate growth, reproduction and health throughout the life cycle. Because they cannot be synthesized, minerals are necessarily obtained from the diet, and thus animals require a mineral intake for a long-term maintenance of body mineral reserves.77 Minerals are involved in many important functions in the body, e.g. cofactors of hundreds of enzymatic reactions, bone mineralization, as well as protection of cells and lipids in biological membranes (antioxidant properties). Low intake or reduced bioavailability of minerals may lead to deciencies, which causes serious impairment of body functions.78 Quinoa content is rich in minerals such as calcium, iron, zinc, magnesium and manganese, which give the grains high value for different target populations: for instance, adults and children benet from calcium for bones and from iron for blood functions.27,33 Antioxidant properties conferred by vitamin E and -3 fatty acids, plus the neuronal activity of tryptophan amino acid and vitamin B complex, can be powerful aids in brain function. Strong effects on protection of stressed neurons given by quinoa consumption in lab rats has recently been demonstrated, with evident effects on neuronal gene expression under stressing conditions, and also on improving spatial memory and promotion of low anxiety in the same animals.79 All these effects should be important in adults as well as in child populations. Besides, zinc helps the immunological system and magnesium is also important during the formation of neuromessengers and neuron modulators. Quinoa also improves some insulin-like forms which are active as growth hormones.80 The low glycemic index makes quinoa good for diabetic patients (low fructose and glucose), as mentioned by Oshodi et al.81 Celiac and lactose-intolerant subjects should also be quinoa consumers because of its gluten-free nature and its rich protein levels, similar to milk casein quality.82

ANTINUTRITIONAL FACTORS
Several antinutritional substances have been found in quinoa, such as, saponins, phytic acid, tannins and protease inhibitors,34,58 which can have a negative effect on performance and survival of monogastric animals when it is used as the primary dietary energy source.58 Saponins were found to be the primary anti-quality factors associated with quinoa,58 but they have also some interesting biological properties.59 Saponins are natural detergents made of glycosylated secondary metabolites, distributed throughout the plant kingdom; they include a diverse group of compounds characterized by their structure containing a steroidal or triterpenoid aglycone and one or more sugar chains.60 The quinoa is surrounded by an epicarp that contains saponins showing a characteristic bitter or astringent taste.61 Quinoa can be classied in accordance with the saponin concentration and their content depends on the quinoa variety: sweet (free from or containing <0.11% of free saponins) or bitter (containing >0.11% of free saponins).21,27,62 64 Stuardo and San Martin63 reported that the content of saponins varies in quinoa between 0.1% and 5%. From the nutritional or pharmacological point of view saponins could also have some value. They can increase membrane permeability, thus enabling use for increasing food intake at the intestinal level or even for drug assimilation.64,65 Other applications include raw materials for production of hormones66 and immunological adjuvants,67 and there are also reported to be active ingredients in various natural health prod ucts, such as herbal extracts.68 Stuardo and San Martn,63 Keukens et al.69 and Armah et al.70 reported antifungal activity of quinoa saponins due to its capacity to associate with steroids of fungal membranes, causing damage to its integrity and pore formation, probably the basis of the novel molluscicide derived from the husks of quinoa, discovered and developed by San Martn et al.71 Phytic acid is not only present in the outer layers of quinoa, as in the case of rye and wheat,47 but is also evenly distributed in the endosperm. Phytic acid binds minerals, thereby rendering them unavailable for metabolism.47,72 Ranges of 10.513.5 mg g1 of phytic acid for ve different varieties of quinoa were reported by Kozio.27 Protease inhibitors, broadly distributed in nature, are proteins that form very stable complexes with proteolytic enzymes.73 The concentration of protease inhibitors in quinoa seeds is <50 p.p.m.47 Ahamed et al.47 and Improta and Kellems58 reported that quinoa contains small amounts of trypsin inhibitors which are much lower than those in commonly consumed grains and hence do not pose any serious concern.

ISOFLAVONES
Finally, a recent yet unpublished thesis83 showed that quinoa seeds from different origins, including long-distance regions of Chile, show different isoavone concentrations, particularly daidzein and genistein. These hormones are implicated in plant physiology (protection from pathogens, from UV light and nitrogen-limited soils) and could be recognized by alpha and beta receptors of estrogens in humans. These endoplasmic reticulum-linked receptors are implicated as inhibitors of tyrosine kinase enzymes, and as antagonists of vessel contraction. They also reduce arterial resistance, benet bone density and stimulate osteoprogerin secretion by osteoblasts, in addition to its antioxidant properties.83

2545

J Sci Food Agric 2010; 90: 25412547

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org

A Vega-G lvez et al. a

10 Bhargava A, Shukla S and Ohri D, Chenopodium quinoa: an Indian perspective. Ind Crops Prod 23:7387 (2006). 11 Jacobsen SE, Adaptation of quinoa (Chenopodium quinoa) to northern European agriculture: studies on developmental pattern. Euphytica 96:4148 (1997). 12 Jacobsen SE, Developmental stability of quinoa under European conditions. Ind Crops Prod 7:169174 (1998). 13 Ormachea E and Quispe D, Evaluacion de parasitoides de la polilla de la quinua Eurysacca melanocampta, en el Cusco, in XXXVConvenci n o Nacional de Entomologa, Arequipa, Peru, 1114 November (1993). 14 Bertero HD, De La Vega AJ, Correa G, Jacobsen SE and Mujica A, Genotype and genotype-by-environment interaction effects for grain yield and grain size of quinoa (Chenopodium quinoa Willd.) as revealed by pattern analysis of international multi-environment trials. Field Crops Res 89:299318 (2004). 15 Cusack DF, Quinua: grain of the Incas. Ecologist 14:2131 (1984). 16 Valencia-Chamorro SA, Quinoa, in Encyclopedia of Food Science and Nutrition, ed. by Caballero B. Academic Press, Amsterdam, pp. 48954902 (2003). 17 Ruales J and Nair BM, Saponins, phytic acid, tannins and protease inhibitors in quinoa (Chenopodium quinoa Willd.). J Sci Food Agric 54:211219 (1993). 18 Mujica A, Andean grains and legumes, in Neglected Crops: 1492 from a Different Perspective, ed. by Hernando Bermujo JE, Leon J. FAO, Rome, pp. 131148 (1994). 19 Risic JC and Galwey NW, The Chenopodium grains of the Andes: Inca crops for modern agriculture, in Advances in Applied Microbiology, ed. by Coaker TH. Academic Press, London, pp. 145217 (1984). 20 Martnez EA, Delatorre J and Von Baer I, Qunoa: las potencialidades de un cultivo sub-utilizado en Chile. Tierra Adentro (INIA) 75:2427 (2007). 21 Martnez EA, Veas E, Jorquera C, San Martn R and Jara P, Re introduction of Chenopodium quinoa Willd. into arid Chile: cultivation of two lowland races under extremely low irrigation. J Agron Crop Sci 195:110 (2009). 22 Wood TR, Tale of a food survivor: Quinoa. East West J 4:6368 (1985). 23 Mujica A, Jacobsen SE and Izquierdo J, Resistencia a factores adversos de la quinua, in Quinua (Chenopodium quinoa Willd.): Ancestral Cultivo Andino, Alimento del Presente y Futuro, ed. by Mujica A, Jacobsen S-E, Izquierdo J, Marathee JP. FAO, UNA-Puno, CIP, Santiago, pp. 162183 (2001). 24 Oelke EA, Putnam DH, Teynor TM and Oplinger ES, Alternative eld crops manual. University of Wisconsin Cooperative Extension Service, University of Minnesota Extension Service, Centre for Alternative Plant and Animal Products (1992). 25 Thanapornpoonpong SN, Vearaslip S, Pawelzik E and Gorinstein S, Inuence of various nitrogen applications on protein and amino acid proles of amaranth and quinoa. JAgricFoodChem 56:1146411470 (2008). 26 FAO, Food and Agriculture Organization, Ecocrop Database (2007). [Online]. Available: http://ecocrop.fao.org/ecocrop/srv/en/ dataSheet?id=2509 [29 May 2009]. 27 Kozio M, Chemical composition and nutritional evaluation of Quinoa. (Chenopodium quinoa Willd). J Food Comp Anal 5:3568 (1992). 28 Wright KH, Pike OA, Fairbanks DJ and Huber SC, Composition of Atriplex hortensis, sweet and bitter Chenopodium quinoa seeds. Food Chem Toxicol 67:13831385 (2002). 29 De Bruin A, Investigation of the food value of quinoa and canihua seed. J Food Sci 29:872876 (1963). 30 Dini A, Rastrelli L, Saturnino P and Schettino O, A compositional study of Chenopodium quinoa seeds. Nahrung 36:400404 (1992). 31 Pa ko P, Barton H, Zagrodzki P, Gorinstein S, Folta M and Zachwieja Z, s Anthocyanins, total polyphenols and antioxidant activity in amaranth and quinoa seeds and sprouts during their growth. Food Chem 115:994998 (2009). 32 Brady K, Ho C, Rosen R, Sang S and Karwe M, Effects of processing on the nutraceutical prole of quinoa. Food Chem 100:12091216 (2007). 33 Repo-Carrasco R, Espinoza C and Jacobsen SE, Nutritional value and use of the Andean crops quinoa (Chenopodium quinoa) and kaniwa (Chenopodium pallidicaule). Food Rev Int 19:179189 (2003). 34 Gonz lez JA, Rold n A, Gallardo M, Escudero T and Prado FE, a a Quantitative determinations of chemical compounds with nutritional value from inca crops: Chenopodium quinoa (quinoa). Plant Foods Hum Nutr 39:331337 (1989).

CONCLUSIONS
The outstanding physicochemical, nutritional and functional properties of quinoa have been reviewed. From ancient and historical data to current laboratory scientic evidence, quinoa was cultivated for its nutritional value, and after being abandoned in favor of old-world crops it is now starting to be rediscovered by modern scientic approaches. Bitter seed coat saponins, while probably giving slower speed to quinoa recovery, might now be helpful for its take-off among farmers for a broader range of consumers, as such saponins also have important agropharmacological and cosmetic industrial uses. From the point of view of vegetarian consumers, quinoa in combination with other cereals might easily replace meat, with a great future in modern, conscient and more ecological food habits. Functional properties given by strongly active compounds like minerals, vitamins, fatty acids and antioxidants make of this small and noble grain a strong contribution to human nutrition, particularly for all cell processes requiring antioxidant protection of membranes, like neuronal activity, with minerals and amino acid contents with potential implications for aiding memory and lowering anxiety under stressful conditions.

ACKNOWLEDGEMENTS
The authors gratefully acknowledge the Research Department at Universidad de La Serena (DIULS), Chile, for providing nancial support to the project (DIULS PI07302). In addition, we wish to thank Project Fondecyt 1060281, 1100638, and funding by ANR-IMAS, TWAS-ICGEB and IRSES agencies.

REFERENCES
1 USGS, United States Geological Survey. [Online]. Available: http:// earthquake.usgs.gov/earthquakes/eqinthenews/2010/us2010tfan/ [2 March 2010]. 2 Legrand C, [Online]. Available: <http://www.lemonde.fr/ameriques/ article/2010/03/01/des-millions-de-refugies-au-chili-apresun-fort-seisme 1312672 3222.html#xtor=EPR-32280229[NL Titresdujour]-20100301-[zoneb]&ens id=1312017> [2 March 2010]. 3 Maughan PJ, Kolano BA, Maluszynska ND, Coles ND, Bonifacio A, Rojas J, et al, Molecular and cytological characterization of ribosomal RNA genes in Chenopodium quinoa and Chenopodium berlandieri. Genome 49:825839 (2006). 4 Christensen SA, Pratt DB, Nelson PT, Stevens MR, Jellen EN, Coleman CE, et al, Assessment of genetic diversity in the USDA and CIP-FAO international nursery collections of quinoa (Chenopodium quinoa Willd.) using microsatellite markers. Plant Genet Resour 5:8295 (2007). 5 National Research Council (NRC), Lost Crops of the Incas: Little Known Plants of the Andes with Promise for Worldwide Cultivation. National Academy Press, Washington, DC, pp. 148161 (1989). 6 Mujica A, Izquierdo J and Marathee JP, Origen y descripcion de la quinua, in Quinua (Chenopodium quinoa Willd.): Ancestral cultivo andino, alimento del presente y futuro, ed. by Mujica A, Jacobsen S-E, Izquierdo J, Marathee J. FAO, UNA-Puno, CIP, Santiago, Chile, pp. 929 (2001). 7 Tagle MB and Planella MT, La Quinoa en la zona central de Chile: Superviviencia de una tradici n prehispana. Editorial IKU, Santiago, o Chile (2002). 8 Tapia M, Cultivos andinos subexplotados y su aporte a la alimentaci n o (2nd edn). FAO, Ocina Regional para Am rica Latina y el Caribe, e Santiago, Chile (1997). 9 Fuentes FF, Martnez EA, Hinrichsen PV, Jellen EN and Maughan PJ, Assessment of genetic diversity patterns in Chilean quinoa (Chenopodium quinoa Willd.) germplasm using multiplex uorescente microsatellite markers. Conserv Genet 10:369377 (2008).

2546

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 25412547

Nutrition and functional potential of quinoa

35 Ridout CL, Price KP, Dupont MS, Parker ML and Fenwiclk GR, Quinoa saponins: analysis and preliminary investigation into the effects of reduction by processing. J Sci Food Agric 54:165176 (1991). 36 FAO/WHO/UNU, Food and Agriculture Organization of the United States/World Health Organization/United Nations University, Energy and protein requirements. Report of a joint FAO/WHO/UNU meeting. World Health Organization, Geneva (1985). 37 Abugoch LE, Quinoa (Chenopodium quinoa Willd.): composition, chemistry, nutritional, and functional properties. Adv Food Nutr Res 58:131 (2009). 38 Schlick G and Bubenheim DL, Quinoa: candidate crop for NASAs controlled ecological life support systems, in Progress in New Crops, ed. by Janick J. ASHS Press, Arlington, VA, pp. 632640 (1996). 39 Ruales J and Nair BM, Properties of starch and dietary bre in raw and processed quinoa (Chenopodium quinoa, Willd.) seeds. Plant Foods Hum Nutr 45:223246 (1994). 40 Mahoney AW, Lopez JG and Hendricks DG, An evaluation of the protein quality of quinoa. J Agric Food Chem 23:190193 (1975). 41 Aluko RE and Monu E, Functional and bioactive properties of quinoa seed protein hydrolysates. J Food Sci 68:12541258 (2003). 42 Bhargava A, Rana TS, Shukla S and Ohri D, Seed protein electrophoresis of some cultivated and wild species of Chenopodium. Biol Plant 49:505511 (2005). 43 Tharanathan RN, Starch: the polysaccharide of high abundance and usefulness. J Sci Ind Res 54:452458 (1995). 44 Tharanathan RN and Mahadevamma S, Grain legumes: a boom to human nutrition. Trends Food Sci Technol 14:507518 (2003). 45 Lorenz K and Coulter L, Quinoa our in baked products. Plant Foods Hum Nutr 41:213223 (1991). 46 Jian YQ and Kuhn M, Characterization of Amaranthus cruentus and Chenopodium quinoa starch. StarchSt rke 51:116120 (1999). a 47 Ahamed NT, Singhal RS, Kulkarni PR and Mohinder P, A lesser-known grain, Chenopodium quinoa: review of the chemical composition of its edible parts. Food Nutr Bull 19:6170 (1998). 48 Ogunbengle HN, Nutritional evaluation and functional properties of quinoa (Chenopodium quinoa) our. Int J Food Sci Nutr 54:153158 (2003). 49 Dini I, Tenore GD and Dini A, Nutritional and antinutritional composition of Kancolla seeds: an interesting and underexploited andine food plant. Food Chem 92:125132 (2005). 50 Konishi Y, Hirano S, Tsuboi H and Wada M, Distribution of minerals in quinoa (Chenopodium quinoa Willd.) seeds. Biosci Biotechnol Biochem 68:231234 (2004). 51 Valencia S, Svanberg U, Sandberg AS and Ruales J, Processing of quinoa (Chenopodium quinoa Willd.): effects on in vitro iron availability and phytate hydrolysis. Int J Food Sci Nutr 50:203211 (1999). 52 Sanders M, Estudio del secado industrial de la quinoa (Chenopodium quinoa Willd.) cultivada en Chile: efecto de la temperatura sobre su composici n. Tesis de Pregrado. Departament of Food Engineering, o Universidad de La Serena, Chile (2009). 53 Zeng X, Cheng K-W, Jiang Y, Lin Z-X, Shi J-J, Ou S-Y, et al, Inhibition of acrylamide formation by vitamins in model reactions and fried potato strips. Food Chem 116:3439 (2009). 54 Insel P, Turner RE and Ross D, Discovering Nutrition. American Dietetic Association, Jones & Bartlett, Boston, MA (2003). 55 Moyad MA, An introduction to dietary/supplemental omega-3 fatty acids for general health and prevention. Part I. Urol Oncol 23:2835 (2005). 56 Ng SC, Anderson A, Coker J and Ondrus M, Characterization of lipid oxidation products in quinoa (Chenopodium quinoa). Food Chem 101:185192 (2007). 57 Nsimba RY, Kikuzaki H and Konishi Y, Antioxidant activity of various extracts fractions of Chenopodium quinoa and Amaranthus spp. seeds. Food Chem 106:760766 (2008). 58 Improta F and Kellems RO, Comparison of raw, washed and polished quinoa (Chenopodium quinoa Willd.) to wheat, sorghum or maize based diets on growth and survival of broiler chicks. [Online]. Livest Res Rural Dev 13 (2001). Available: http://www.cipav.org.co/lrrd/lrrd13/1/impr131.htm [29 May 2009]. 59 Sparg SG, Light ME and Van Staden J, Biological activities and distribution of plant saponins. J Ethnopharmacol 94:219243 (2004). 60 Guclu-Ustundag O and Mazza G, Saponins: properties, applications and processing. Crit Rev Food Sci Nutr 47:231258 (2007).

www.soci.org
61 Tarade KM, Singhal RS, Jayram RV and Pandit AB, Kinetics of degradation of saponins in soybean our (Glycine max) during food processing. J Food Eng 76:440445 (2006). 62 Soliz-Guerrero JB, Jasso de Rodrguez D, Rodrguez-Garca R, Angulo S nchez JL and M ndez-Padilla G, Quinoa saponins: concentration a e and composition analysis, in Trends in New Crops and New Uses, ed. by Janick J and Whipkey A. ASHS Press, Alexandria, VA, pp. 110114 (2002). 63 Stuardo M and San Martn R, Antifungal properties of quinoa (Chenopodium quinoa Willd.) alkali treated saponins against Botrytis cinerea. Ind Crops Prod 27:296302 (2008). 64 Gee JM, Price KR, Ridout CL, Wortley GM, Hurrel RF and Johnson IT, Saponins of quinoa (Chenopodium quinoa): effect of processing on their abundance in quinoa products and their biological effects on intestinal mucosal tissue. J Sci Food Agric 63:201209 (1993). 65 Oakenfull D and Sidhu G, Could saponins be a useful treatment for hypercholesterolaemia? Eur J Clin Nutr 44:7988 (1990). 66 Blunden G, Culling C and Jewers K, Steroidal sapogenins: a review of actual and potential plant sources. Trop Sci 17:139154 (1975). 67 Kensil CR, Mo AX and Truneh A, Current vaccine adjuvants: an overview of a diverse class. Frontiers Biosci 9:29722988 (2004). 68 Balandrin MF, Commercial utilization of plant-derived saponins: an overview of medicinal, pharmaceutical, and industrial applications, in Saponins Used in Traditional Medicine, ed. by Waller GR and Yamasaki K. Plenum Press, New York, pp. 114 (1996). 69 Keukens AJ, De Vrije T, Van den Boom C, De Waard P, Plasman HH, Thiel F, et al, Molecular basis of glycoalkaloid induced membrane disruption. Biochem Biophys Acta 1240:216228 (1995). 70 Armah CN, Mackie AR, Roy C, Price K, Osbourn AE, Bowyer P, et al, The membrane permeabilizing effect of avenacin A-1 involves the reorganization of bilayer cholesterol. Biophys J 76:281290 (1999). 71 San Martn R, Ndjoko K and Hostettmann K, Novel molluscicide against Pomacea canaliculata based on quinoa (Chenopodium quinoa) saponins. Crop Prot 27:310319 (2008). 72 Khattak AB, Zeb A, Bibi N, Khalil SA and Khattak MS, Inuence of germination techniques on phytic acid and polyphenols content of chickpea (Cicer arietinum L.) sprouts. Food Chem 104:10741079 (2007). 73 Aguirre C, V ldez-Rodrguez S, Mendoza-Hern ndez G, Rojoa a Domnguez A and Blanco-Labra A, A novel 8.7 kDA protease inhibitor from chan seeds (Hyptis suaveolens L.) inhibits proteases from the larger grain borer Prostephanus truncatus (Coleoptera: Bostrichidae). Comp Biochem Physiol Part B 138:8189 (2004). 74 Cerezal P, Carrasco A, Pinto K, Romero N and Arcos R, Suplemento alimenticio de alto contenido prot ico para ninos de 25 e anos, desarrollo de la formulacion y aceptabilidad. Interciencia 32:857864 (2007). 75 Dini I, Tenore GC, Trimarco E and Dini A, Two novel betaine derivatives from Kancolla seeds (Chenopodiaceae). Food Chem 98:209213 (2006). 76 Singh U and Eggum BO, Factors, affecting the quality of pigeonpea (Cajanus cajan L.). Plant Foods Hum Nutr 34:273283 (1984). 77 McDowell LR, Minerals in Animal and Human Nutrition (2nd edn). Elsevier, Amsterdam (2003). 78 Schlenker ED and Williams SR, Essentials of Nutrition and Diet Therapy. Elsevier, Amsterdam (2003). 79 Munoz-Llancao P, Martnez EA, Wyneken U and Dagnino-Subiabre A, Pseudo cereal quinoa improves memory and decreases anxiety of stressed rats: behavioral and molecular approaches, in I IBRO/LARC Congress of Neurosciences of Latin America, Caribbean and Iberian Peninsula, Buzios, Brazil, 14 September (2008). 80 Ruales J, de Grijalva Y, Lopez-Jaramillo P and Nair BM, The nutritional quality of an infant food from quinoa and its effect on the plasma level of insulin-like growth factor-1 (IGF-1) in undernourished children. Int J Food Sci Nutr 53:143154 (2002). 81 Oshodi AA, Ogungbenle HN and Oladimeji MO, Chemical composition, nutritionally valuable minerals and functional properties of beniseed (Sesamum radiatum), pearl millet (Pennisetum typhoides) and quinoa (Chenopodium quinoa) ours. Int J Food Sci Nutr 50:325331 (1999). 82 Herencia LI, Ala MJ, Gonz lez A and Urbano P, Cultivo de la quinoa a (Chenopodium Quinoa Willd.) en la region Centro. [The culture of quinoa in the central region]. Vida Rural VI:2833 (1999). 83 Martnez AM, Contenidodedaidzeinaygenisteinaenecotiposdesemillas de qunoa (Chenopodium quinoa Willd.). Thesis, Chemistry and Pharmacy, Universidad de Valparaso, Chile (2008).

2547

J Sci Food Agric 2010; 90: 25412547

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

Research Article
Received: 26 October 2009 Revised: 1 July 2010 Accepted: 2 July 2010 Published online in Wiley Online Library: 17 August 2010

(wileyonlinelibrary.com) DOI 10.1002/jsfa.4116

Cultivar choice provides options for local production of organic and conventionally produced tomatoes with higher quality and antioxidant content
Heather Troxell Aldrich,a Karen Salandanan,a Patricia Kendall,b Marisa Bunning,b Frank Stonaker,a Oktay Kulena and Cecil Stushnoffa
Abstract
BACKGROUND: Tomatoes (Solanum lycopersicum L.) are widely consumed and well known for their health benets, many of which have been associated with the high levels of antioxidants present in tomatoes. With a growing interest in local and organic foods, it would be helpful to determine whether farmers could naturally improve the quality and antioxidant content of tomatoes for sale in local markets. This study evaluated antioxidant properties, quality attributes, and yield for 10 tomato cultivars grown for 2 years using certied organic and conventional practices. RESULTS: Cultivar and year effects impacted (P < 0.05) all tests conducted, while growing method inuenced (P < 0.05) yield, soluble solids content, ascorbic acid, and antioxidant radical scavenging capacity. Even when accounting for year-to-year variability, cultivars in the highest groups had 1.35- to 1.67-fold higher antioxidant levels than cultivars in the lowest groups. New Girl, Jet Star, Fantastic, and First Lady were always in the highest groups, while Roma and Early Girl consistently had the lowest antioxidant content. CONCLUSION: Compared to production practices and environmental effects of years that are generally beyond the control of small-scale producers, choice of cultivar provides the simplest and most effective means of increasing antioxidant properties. Knowledge of tomato cultivars with naturally higher antioxidant levels could assist smaller-scale producers to grow fruit that may provide a competitive advantage and the opportunity to capitalize on the increasing popularity of locally grown, high-quality fresh produce. c 2010 Society of Chemical Industry Keywords: antioxidants; total phenolics; ascorbic acid; cultivar; organic production

INTRODUCTION
Tomatoes (Solanum lycopersicum L.) are widely consumed, ranking second to potatoes among vegetable and melon per capita use in the USA.1 Tomatoes are good sources of vitamin C, folate, and potassium, as well as many phytochemicals.2 5 Awareness for the important role fruits and vegetables have in a healthy diet is increasing,6,7 as is interest in organic foods. Organic foods have become one of the fastest-growing food categories, with sales increasing nearly 20% each year since 1990.8 Consumer studies have shown that organic produce is perceived to be safer, more nutritious, and better tasting than conventionally grown produce.9 12 However, research comparing such attributes has produced inconsistent or inconclusive results, most likely due to uncontrolled variables such as differences in growing conditions and cultivars studied.13,14 Additional well-controlled studies are needed to better understand the impact of organic and conventional growing methods on produce quality and nutritional attributes.15 Farmers markets are becoming a popular place to purchase produce in the USA. According to the United States Department

of Agriculture (USDA), the number of US farmers markets more than doubled from 1755 in 1994 to 4685 in 2008.16 Also, based on the results of a national survey conducted in 2006, three out of four respondents had shopped at a farmers market within the last year.17 Consumers indicated greater willingness to pay a premium for produce that had a higher nutritional value (i.e. melons with 25% more vitamin C), was grown locally, and/or was grown using organic production methods. Support for local farmers was found to be more important than organic production and documented nutritional attributes also seemed to be favored over production certications.18

Correspondence to: Cecil Stushnoff, Department of Horticulture and Landscape Architecture, Colorado State University, Fort Collins, CO 80523-1173, USA. E-mail: Cecil.Stushnoff@colostate.edu

a Department of Horticulture and Landscape Architecture, Colorado State University, Fort Collins, CO 80523, USA b Department of Food Science and Human Nutrition, Colorado State University, Fort Collins, CO 80523, USA

2548

J Sci Food Agric 2010; 90: 25482555

www.soci.org

c 2010 Society of Chemical Industry

Cultivar choice provides nutritious tomato options This increased popularity and demand for high-quality, fresh produce, along with the growing interest in locally produced food and organic production,17,18 provides small-scale growers with a unique opportunity. The opportunity, however, can best be realized if local small-scale growers can differentiate their produce from that available from other sources. The objectives of this study were to compare the antioxidant levels, quality attributes, and yield among 10 tomato cultivars grown using organic and conventional production methods, and to determine whether production method and/or cultivar choice would assist small- and medium-scale farmers to market nutritionally superior tomatoes.

www.soci.org Pest management was needed in 2005, as there was a combination of pressure from potato psyllid (Paratrioza cockerelli) and beet leafhopper (Circulifer tenellus). Potato psyllids were controlled in the organic plots using an approved botanically derived pyrethrum (Pyganic, MGK Co., Golden Valley, MN, USA) insecticide. Provado (Imidicloprid, Bayer CropScience, Research Triangle Park, NC, USA) was applied to control psyllids and leafhoppers in the conventional plots. There was no psyllid infestation during the 2006 growing season. Temperature and solar radiation data collection Data on temperature and solar radiation during the 2005 and 2006 growing seasons were obtained from a Northern Colorado Water Conservancy District weather station, located within 100 m of the research plots. To examine possible effects of temperature, daily growing degree days (GDD) was computed by subtracting the base temperature (10 C) from the average of the daily maximum and minimum temperatures (daily GDD = [(Tmax + Tmin )/2] base temperature, where Tmax and Tmin are maximum and minimum daily air temperatures). Each daily GDD was summed over the growing season and 30 days prior to harvest. Solar radiation data were recorded with an Eppley pyranometer (Newport, RI, USA) and expressed as langleys. Post-harvest analyses Once the tomatoes reached the vine ripe stage, indicated by full red color and the onset of softening, they were harvested manually early each morning. Note that while the latest harvest date in 2005 was 20 days longer than in 2006 due to the difference in heat unit accumulation, fruit was harvested when it was judged to be at essentially the same stage of ripeness for each cultivar. All tomatoes were harvested and weighed from each replicate, and yield was expressed in kg per plant. On one day during the height of the growing season, three uniformly sized and colored tomato fruits per replicate plot were randomly selected from the bulked harvest. The tomato samples (n = 180 in 2005 and 240 in 2006) were transported at ambient temperature to the laboratory within 30 min for evaluation of soluble solids content and pH, as well as preparation for antioxidant analyses. Soluble solids content and pH Percent soluble solids of each tomato sample was measured using a handheld Reichert temperature-compensated refractometer (Reichert Analytical Instruments, Depew, NY, USA). The pH of the tomatoes was tested using a Beckman pH meter (Fullerton, CA, USA). Antioxidant analyses Sample preparation Tomatoes were washed well to remove any debris on their outer surface, and then cut in half vertically. Thin radial slices were cut from the tomato halves. Thin slices from each tomato (3540 g) were freeze-dried using a Genesis freeze drier (Virtis, Inc., Gardiner, NY, USA). Lyophilized samples were weighed to determine percent dry matter content and ground in preparation for extraction. The dried samples were ground into a ne powder using a mortar and pestle and sieved with a No. 20 Tyler sieve (WS Tyler Inc., Mentor, OH, USA). Samples were prepared by extracting 200 mg lyophilized powder in 5 mL of 80% acetone (Fisher Scientic, Fair Lawn, NJ,

MATERIALS AND METHODS

Plant material Tomatoes were grown at Colorado State Universitys (CSU) Horticulture Field Research Center (HFRC) in Fort Collins, CO, during the summers of 2005 and 2006. Ten cultivars of tomatoes were grown simultaneously on organic and conventional plots. The cultivars grown included Big Beef, Early Girl, Celebrity, Fantastic, First Lady, Husky Red, Jet Star, Red Sun, New Girl, and Roma. Seeds were obtained from Johnnys Selected Seeds, Winslow, ME, USA, and Harris Seeds, Rochester, NY, USA. All cultivars are beefsteak type except Roma (a plum type). Plants were started in the CSU Plant Environmental Research Center greenhouses in Landmark T1204 cell packs (Landmark Plastic Corp., Akron, OH, USA) using Sunshine Organic Basic planting media (Sun Gro Horticulture, Bellvue, WA, USA) with 20% vermicompost (local source). After 6 weeks, the tomatoes were transplanted to the eld, spaced evenly in black plastic mulched beds (plants approximately 45 cm apart within a row, and rows 150 cm apart). Soil at the HFRC is classied as Nunn clay with a pH of 7.8. The organic plots have been USDA certied organic since 2001. The tomatoes were planted in a split plot design with production system as the whole plot factor. Tomato cultivars were randomized as sub-plots within each of three replicated plots (15 plants per plot) in 2005 and within four replicated plots (eight plants per plot) in 2006. The two production systems have identical soil textures and were located approximately 50 m from each other, separated by an appropriate buffer zone with vegetation.19 Prior to planting, soil tests were conducted on the organic and conventional plots to determine pH, electroconductivity (EC), lime, organic matter, N, P, K, Zn, Fe, Mn, and Cu levels. Soil fertility of the two plots was equilibrated at the beginning of the growing season using either organic or conventional fertilizers. Based on the soil tests, 22 679 kg ha1 of Evergreen poultry compost (A1 Organics, Eaton, CO, USA) was applied pre-planting to the organic plots. The compost was applied with a spreader (Millcreek Mfg, Leola, PA, USA) and disked into the soil immediately following the application. To match the amount of nutrients in the organic block, 389 kg ha1 of urea (45N-0P-0K) and 882 kg ha1 of triple superphosphate (0N-20.1P-0K) were applied to the conventional block using a broadcast spreader. Crops were irrigated using drip irrigation with low-salt-content, high-quality municipal water. Soil moisture levels were determined using Watermark granular matrix sensors (Irrometer Co., Riverside, CA, USA). Soil moisture was monitored to ensure the tomatoes were watered adequately in order to prevent water stress by not allowing water tension to drop below 100 kPa throughout the growing season.

2549

J Sci Food Agric 2010; 90: 25482555

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org USA) with vortexing and 15 min rotary mixing in 15 mL centrifuge tubes in the dark at 4 C. The samples were centrifuged (4 C; 1771 g relative centrifugal force) for 15 min. One-milliliter aliquots of supernatant were vacufuged at 45 C to dryness (approximately 23 h). Samples were stored at 20 C until analytical tests were completed. Total phenolic content Total phenolic content was measured using a microplate-based FolinCiocalteu assay adapted from Singleton and Rossi,20 Spanos and Wrolstad,21 and Rivera et al.22 Briey, vacufuged extractions were reconstituted with 1.0 mL of 80% acetone, diluted 1/9 with nanopure water and reacted with 150 L of 0.2 mol L1 FolinCiocalteu reagent (Sigma-Aldrich, Inc., St Louis, MO, USA). After 5 min at room temperature, microplate samples were reacted with 115 L of 7.5% (w/v) Na2 CO3 (Fisher Scientic), incubated at 45 C for 30 min and cooled to room temperature for 1 h, before reading absorbance at 765 nm in a Spectra Max Plus (Molecular Devices, Sunnyvale, CA, USA) spectrophotometer using Softmax Pro software (Molecular Devices). Total phenolic content was calculated by regression from a gallic acid (Sigma Chemical Co., St Louis, MO, USA) standard curve and expressed as milligrams of gallic acid equivalents per kilogram of tomato fresh weight (mg GAE kg1 FW). Ascorbic acid content Ascorbic acid (vitamin C) content was determined using highperformance liquid chromatography (HPLC) as described by Rivera et al.22 and modied from Dale et al.23 Freeze-dried samples were extracted for 15 min at 4 C in the dark with a 5% w/v aqueous solution of metaphosphoric acid containing 1% w/v dithiothreitol (DTT) (Promega Corp., Madison, WI, USA). The samples were centrifuged for 5 min at 1771 g at 4 C and the supernatant was ltered through a 0.45 m nylon syringe lter. The extraction process was repeated and the supernatant from both extractions was placed in an amber HPLC vial. Ascorbic acid standards were prepared with 100 mg DTT (Promega Corp.), 10 mg ascorbic acid (Sigma-Aldrich), and 10 mL of 100% methanol before diluting to ve concentrations for the standard curve. All analyses were run in duplicate and were analyzed by HPLC (Hewlett Packard Model 1050 Series, Palo Alto, CA, USA) using Chem Station for LC Rev A 09.01 software (Agilent Technologies, Palo Alto, CA, USA). Samples were injected into an Inertsil C4 column (Agilent Technologies) run with a phosphoric acid/methanol gradient and absorbance read at 254 nm. ABTS Trolox equivalent antioxidant capacity The 2,2 -azino-bis(3-ethylbenzothiazoline-6-sulfonic acid) diammonium salt (ABTS) assay was used to estimate antioxidant hydroxyl radical scavenging capacity. The protocol used was based on the microplate method described by Rivera et al.,22 as modied from Miller and Rice-Evans.24 The ABTS solution was prepared by mixing 40 mg ABTS (Calbiochem, EMD Biosciences, La Jolla, CA, USA), 15 mL distilled water, and 2.0 0.5 g MnO2 (Sigma-Aldrich). After 20 min, the MnO2 was removed using double ltration, rst with a vacuum ltration and second with a 0.2 m syringe lter. The absorbance value of the ABTS solution was adjusted to 0.70 AU at 734 nm in the Spectra Max Plus spectrophotometer using Softmax Pro software with 5.0 mmol L1 phosphate buffer solution. The ABTS solution was held at 30 C and used within 4 h.

HT Aldrich et al.

Vacufuged samples were reconstituted with 1 mL of 80% acetone (Fisher Scientic). Twenty-ve microliters of sample were reacted with 250 L of the ABTS solution, and the absorbance value was read after exactly 60 s at 30 C in a temperature-controlled microplate reader. ABTS antioxidant capacity was reported as Trolox equivalent antioxidant capacity (TEAC) per gram of fresh sample (TEAC g1 FW) and was calculated by regression of a Trolox (Calbiochem) standard curve. Analyses were run in triplicate at three dilutions for a total of nine assays per sample. DPPH Trolox equivalent antioxidant capacity The 2,2-diphenyl-1-picrylhydrazyl (DPPH) assay was also used to estimate antioxidant capacity according to the method of Lu and Foo25 with some modications. Briey, vacufuged samples were reconstituted with 1.0 mL of 5.0 mmol L1 phosphate buffer solution and reacted with 0.1 mmol L1 DPPH solution in 100% methanol. Absorbance was read in the Spectra Max Plus spectrophotometer using Softmax Pro software at 515 nm and adjusted to 0.95 AU. Fifteen microliters of the reconstituted samples were mixed with 285 L of the DPPH solution, held for exactly 3 min at 25 C, and read at 515 nm. The results were determined by regression from a Trolox (Calbiochem) standard curve and expressed as mol TEAC kg1 FW. Data and statistical analysis Results were analyzed using SAS statistical software (Version 9.1, Cary, NC, USA). A factorial analysis of variance was performed (SAS Proc Mixed) with differences between means assessed using a signicance of P < 0.05 with the TukeyKramer adjustment for multiple comparisons. Fixed effects included cultivar, growing method, and year; replication was included as a random effect. Correlation analysis was done using Pearson distribution (SAS Proc Corr).

RESULTS
Temperature and solar radiation From eld planting to harvest, GDD was higher in 2006 than in 2005; thus tomatoes grown in 2006 were exposed to higher cumulative heat units (Fig. 1(A)). During the 2006 growing season, there were also more days with maximum temperatures above 30 C (Fig. 1(B)). Solar radiation received by tomato plants from planting to harvest was nearly the same for both 2005 and 2006 (Fig. 1(C)). Yield Year, growing method, and cultivar all impacted (P < 0.0001) tomato yields (Table 1). Overall, 2006 yields per plant were higher than 2005 yields (average of 4.17 kg per plant versus 2.99 kg per plant, respectively, data not shown); and conventional yields were higher than organic yields (average of 4.03 kg per plant versus 3.14 kg per plant, respectively; Table 2). While average cultivar yields varied by growing method, Early Girl consistently produced the highest yields and Roma the lowest yields. Percent dry matter The percent dry matter (% DM) was signicantly affected by year (Table 1; average of 5.3% in 2005 versus 7.4% in 2006), while growing method did not impact % DM, with both methods averaging 6.3%. Early Girl had the lowest % DM (P < 0.05), while cultivars First Lady, New Girl, and Jet Star had the highest (P < 0.05) over both years (Table 2).

2550

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 25482555

Cultivar choice provides nutritious tomato options

A Heat Accumulation 1,200 1,000 800 600 400 200 0 1 11 21 31 41 51 61 71 81 91 101 Days After Transplanting 2005 B Maximum temperatures (degrees C) 40 35 30 25 20 15 10 5 0 1 11 21 31 41 51 61 71 81 91 101 Days After Transplanting 2005 C Solar radiation (Langleys) 800 700 600 500 400 300 200 100 0 1 11 21 31 41 51 61 71 81 91 101 111 Days After Transplanting 2005 2006 2006 2006

www.soci.org (P < 0.001) by cultivar, with Early Girl having the lowest SSC (Table 2). Antioxidant analyses Overall, year had a signicant impact on all antioxidant tests (Table 1). Total phenolic content was higher in 2005, the cooler year, than in 2006 (average of 0.78 versus 0.74 g GAE kg1 FW, respectively), while the 2006 fruit had higher ABTS values (average of 1533 versus 1400 mol TEAC kg1 FW), higher DPPH (average of 2573 versus 1705 mol TEAC kg1 FW), and higher ascorbic acid (average of 0.23 versus 0.15 g kg1 FW; data not shown). Organically grown tomatoes had higher overall (P < 0.0001) ABTS (average of 1546 versus 1387 mol TEAC kg1 FW), DPPH (average of 2194 versus 2084 mol TEAC kg1 FW), and ascorbic acid (average of 0.198 versus 0.185 g kg1 FW) compared to the conventionally grown fruit (Table 3). Cultivar differences were very signicant (P < 0.0001) for all antioxidant analyses (Table 1), though they varied by test (Table 3). To compare cultivars by growing method, we placed cultivars into lowest, midrange, and highest groups, with values for cultivars in lowest and highest groups signicantly different (P < 0.05) based on TukeyKramer mean separation tests, using the method developed by Salandanan et al.26 This was done for both organic and conventional results averaged for the 2 years, since year differences are impossible to control (Table 4). Based on ratios comparing highest and lowest cultivar groups, cultivars falling in the highest category were found to have 1.35- to 1.67-fold higher values for all the antioxidant tests compared to cultivars in the lowest groups. The spread between lowest and highest groups was similar for organic and conventionally grown tomatoes. Ratios used to compare organic versus conventional results for the lowest, midrange, and highest cultivar groups tended to hover around 1.00, while organic results were slightly higher for all but two groups (Table 4). A simple additive antioxidant index was used to compare cultivars over both years regardless of growing method. This index was calculated by determining the average of the total phenolic content, ascorbic acid content, and antioxidant capacity (using an average of the TEAC values obtained from the ABTS and DPPH assays; Fig. 2). The antioxidant index rank of the cultivars varied slightly between the 2 years; however, despite the environmental effects of each growing season, the same cultivars appear in the top and bottom tiers. Roma and Early Girl consistently had the lowest antioxidant index scores, while New Girl, Jet Star, Fantastic, and First Lady consistently had the highest index scores (Fig. 2 and Table 3). Correlation analysis When all 10 cultivars were considered over both years and production methods, fruit with the highest % DM also had the highest pH and SSC, as well as TEAC radical scavenging capacity and ascorbic acid content (Table 5). Total phenolic levels were not strongly related to fruit quality parameters nor to ascorbic acid or TEAC values. While yield was signicantly correlated with a few parameters, none, except possibly DPPH, had high r values and were therefore considered not to contribute substantially to these correlations (Table 5).

Figure 1. (A) Heat accumulation (in growing degree-days or GDD) from planting to harvest, GDD = [(Tmax + Tmin )/2] 10 C, where Tmax and Tmin are maximum and minimum daily air temperatures; 10 C is the base temperature for warm season crops. (B) Number of days with temperature greater than 30 C as indicator of heat stress. (C) Daily net solar radiation (in langleys) from planting to harvest.

pH and soluble solids content Cooler conditions in 2005 signicantly impacted pH, with 2005 tomatoes being more acidic than in 2006 (overall average of 3.84 versus 4.14, respectively), while growing method did not signicantly affect pH (Tables 1 and 2). The pH also varied by cultivar, with Celebrity and Early Girl ranking among the lowest in pH and Jet Star the highest across growing method and year. Cultivar pH tended to rank in a similar order, regardless of growing method and year (Table 2). Soluble solids content (SSC) was impacted by year (P < 0.0001; Table 1), with 2006 tomatoes having higher SSC than in 2005 (average of 4.9 versus 4.3 Brix, respectively). Overall, conventional tomatoes had a slight, though signicantly higher SSC than organic (means 4.6 versus 4.4 Brix, respectively). SSC also varied

DISCUSSION
In this study, year-to-year variability was signicant, with tomatoes grown in the warmest year, 2006, having higher (P < 0.05) ascorbic

2551

J Sci Food Agric 2010; 90: 25482555

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org

HT Aldrich et al.

Table 1. Analysis of variance for the main effects year (Y), cultivar (C), growing method (M), and their interactions on yield, dry matter (% DM), pH, soluble solids content (SSC), total phenolic content (TPC), ascorbic acid (AA), and antioxidant activity (ABTS and DPPH) Yield Year Cultivar YC Method YM CM YCM <0.0001 <0.0001 0.0437 <0.0001 0.8109 0.5094 0.3712 % DM <0.0001 <0.0001 0.3034 0.7891 <0.0001 0.1612 0.4037 pH <0.0001 <0.0001 0.0210 0.1126 0.0022 0.1269 0.2641 SSC <0.0001 <0.0001 0.0079 0.0034 0.0473 0.0020 0.0907 TPC 0.0034 <0.0001 <0.0001 0.4944 <0.0001 <0.0001 <0.0001 AA <0.0001 <0.0001 0.0356 0.0001 0.5746 0.0083 0.0119 ABTS 0.0008 <0.0001 <0.0001 <0.0001 0.4440 <0.0001 <0.0001 DPPH <0.0001 <0.0001 <0.0001 <0.0001 0.0027 <0.0001 <0.0001

Table 2. Yield, % DM, pH, and soluble solids content resultsa for 10 tomato cultivars grown organically and conventionally in 2005 and 2006 Yield (kg per plant) Cultivar Big Beef Celebrity Early Girl Fantastic First Lady Husky Red Jet Star New Girl Red Sun Roma Mean SEb
a b

% DM Organic 6.5bc 6.2bc 5.1a 6.5bc 7.1c 6.3bc 6.8c 6.7c 6.2bc 5.7ab 6.3 0.18 Conv. 6.4bcde 5.6ab 5.2a 6.6cde 7.2e 6.3bcde 6.8de 7.2e 5.9abc 6.0abcd 6.3 Organic 3.97abc 3.83a 3.85ab 4.01cd 3.91ab 4.02cde 4.18e 3.95abc 3.98bcd 4.13de 3.98

pH Conv. 4.02abc 3.94ab 3.89a 3.97ab 3.95ab 4.03abc 4.14c 3.97abc 4.04bc 4.08bc 4.00 0.03

Soluble solids content ( Brix) Organic 4.62c 4.33bc 3.52a 4.60c 4.82c 4.50bc 4.66c 4.39bc 4.26abc 3.79ab 4.35 0.15 Conv. 4.96cde 4.01ab 3.46a 5.22de 5.22de 4.28bc 4.58bcde 5.31e 4.48bcd 4.03ab 4.56

Organic 2.74bc 2.89bc 4.51a 2.96abc 3.66ab 3.40ab 2.66bc 3.57ab 3.35ab 1.62c 3.14 0.32

Conv. 3.32bc 3.89abc 5.21a 4.44ab 4.33ab 4.18ab 4.20ab 3.74abc 4.59ab 2.39c 4.03

Results represent the mean (SE) of seven replications (three replications in 2005 and four replications in 2006). Standard error of means listed in column. Cultivars for the same year and growing method with different letters are signicantly different (P < 0.05).

Table 3. Total phenolic content (TPC), ascorbic acid, and antioxidant activity (ABTS and DPPH) resultsa for 10 tomato cultivars grown organically and conventionally in 2005 and 2006 TPC (g GAE kg1 FW) Cultivar Big Beef Celebrity Early Girl Fantastic First Lady Husky Red Jet Star New Girl Red Sun Roma Mean SEb Organic 0.78abc 0.80abc 0.69ab 0.87c 0.77abc 0.64a 0.89c 0.81bc 0.69ab 0.68ab 0.76 0.032 Conv. 0.66ab 0.70ab 0.56a 0.94d 0.67ab 0.75bc 0.90cd 0.93d 0.67ab 0.73b 0.74 Ascorbic Acid (g kg1 FW) Organic 0.17a 0.20abc 0.17a 0.22bc 0.22bc 0.17a 0.22bcA 0.23c 0.19ab 0.19ab 0.198 0.007 Conv. 0.18abc 0.21bc 0.17ab 0.19abc 0.20bc 0.18abc 0.18abcB 0.21c 0.18abc 0.15a 0.185 ABTS (mol TEAC kg1 FW) Organic 1318a 1641abcA 1222a 1967cd 2117dA 1241a 1501ab 1775bcA 1405ab 1274a 1546 91 Conv. 1474ab 1106aB 1278ab 1568b 1534bB 1284ab 1527b 1340abB 1403ab 1354ab 1387 DPPH (mol TEAC kg1 FW) Organic 1859ab 1874ab 1956abc 2379d 2711fA 2083bc 2585defA 2605ef 2166cd 1722a 2194 55 Conv. 1939ab 1877a 1751a 2436c 2292cB 2217bc 2219bcB 2466c 1958ab 1688a 2084

a Results represent the mean (SE) of seven replications for TPC and ABTS+ (three replications in 2005 and four replications in 2006) and six replications for DPPH+ and ascorbic acid (three replications both years). b Standard error of means listed in column. Cultivars for the same year and growing method with different lower-case letters are signicantly different (P < 0.05). Growing methods for the same year and cultivar with different upper-case letters are signicantly different (P < 0.05).

2552

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 25482555

Cultivar choice provides nutritious tomato options

140 Antioxidant Index 120 100 80 60 40 20 0

www.soci.org

Table 4. Comparison of means for total phenolics, antioxidant activity (ABTS and DPPH), and ascorbic acid for high, midrange, and low cultivar groups in organic and conventional production systems averaged over 2 years Ratio (organic/ Organic z Conventional z conventional)

Group

Total phenolic content (g GAE kg1 FW) Lowest 0.64 1 0.56 Midrange 0.75 7 0.73 Highest 0.88 2 0.94 Ratio (highest/lowest) 1.38 1.67 Ascorbic acid (g kg1 FW) Lowest 0.17 3 0.15 Midrange 0.21 6 0.19 Highest 0.23 1 0.21 Ratio (Highest/lowest) 1.35 1.38 ABTS antioxidant activity (mol TEAC kg1 FW) Lowest 1264 4 1106 Midrange 1658 5 1356 Highest 2117 1 1544 Ratio (Highest/lowest) 1.67 1.40 DPPH antioxidant activity (mol TEAC kg1 FW) Lowest 1722 1 1772 Midrange 2188 8 2083 Highest 2711 1 2398 Ratio (Highest/lowest) 1.57 1.35

us

Fi

1 8 1

1.13 1.11 1.10

Cultivar 2005 2006

1 6 3

1.14 1.22 1.37

Figure 2. Antioxidant Index (SE) of 10 tomato cultivars for 2005 and 2006 growing seasons (n = 6 for each bar, three replications of two growing methods). Antioxidant index was calculated as [( vit. C + TPC + antioxidant capacity)/3], with antioxidant capacity being an average of the ABTS and DPPH antioxidant capacity assay results. Antioxidant indices were calculated per 100 g to allocate equal weight to each factor.

3 4 3

0.97 1.05 1.13

z, number of cultivars in lowest, midrange, and highest groups with values for cultivars in lowest and highest groups signicantly different (P < 0.05) based on TukeyKramer mean separation tests.

acid, antioxidant activity (ABTS and DPPH), soluble solids content, pH, yield per plant and percent dry matter, while tomatoes grown in 2005 had higher (P < 0.05) total phenolic content. Higher light intensity during production has been associated with higher ascorbic acid content in fresh produce,27 which may explain why the ascorbic acid levels were higher in 2006 than in 2005. Also, environmental temperatures during tomato growth have been shown to affect antioxidant levels, though mechanisms are not well understood and optimal temperatures may differ for different groups of phytochemicals.28,29 Although precipitation amounts were quite different in the 3 months preceding harvest in 2005 (10.5 cm) and 2006 (3.5 cm), soil moisture was monitored and plots were

irrigated to prevent water stress. Clearly, environmental factors such as temperature, precipitation, and light intensity have important implications in the development of quality and nutritional characteristics of crops, but they are difcult to control and more research is needed to better understand their specic impacts.30 Some evidence indicates there may be a trade-off between yield and nutritional quality of produce.31 34 However, such studies have primarily focused on comparing historical variation of fruits and vegetables, which entail averaging many foods over time and include a great deal of confounding variables such as production methods, sampling procedures, and test methodologies.33 As a result, Davis recommended the use of side-by-side comparisons as a more accurate, controlled approach in which specic foods and nutrients can be evaluated.33 Few such side-by-side comparisons have been conducted, though studies looking at potato,35 wheat,36,37 and broccoli38 have found that cultivars with higher yields (or larger heads for broccoli) often experience low to moderate declines in many nutrients compared to lower-yielding (or smaller-head) cultivars. Our side-by-side comparative study does not support the argument that high-yielding cultivars compensate by reducing nutritional quality. Low Pearson r values and non-signicance of some correlations suggest only a small proportion of the variation

Table 5. Correlation analysis (Pearson r) between percent dry matter (% DM), pH, soluble solids content (SSC), total phenolics content (TPC), antioxidant activity (ABTS and DPPH tests), ascorbic acid (AA), and yield. Correlation results include means from 10 cultivars, 2 years, two growing methods, and three replications (n = 120) % DM % DM pH SSC TPC ABTS+ DPPH+ AA Yield 0.65 0.79 0.13 n.s. 0.40 0.86 0.81 0.29 pH SSC TPC ABTS DPPH AA

0.46 0.05 n.s. 0.16 n.s. 0.61 0.66 0.23

0.25 0.31 0.69 0.57 0.18

0.33 0.23 0.10 n.s. 0.06 n.s.

0.50 0.42 0.09 n.s.

0.83 0.48

Ea

Fa

1 7 2

1.14 1.02 0.94

tic

ed

ef om a R

irl

rit y

ad

Su

tS

ew

tL

eb

Be

as

nt

ed

ky

Je

rs

el

Bi

rly

n.s., not signicant; asterisks indicate signicance at P < 0.05; P < 0.01; P < 0.0001.

G irl

ta

0.35

2553
wileyonlinelibrary.com/jsfa

J Sci Food Agric 2010; 90: 25482555

c 2010 Society of Chemical Industry

www.soci.org in quality and antioxidant parameters may be directly related to yield on a per plant basis. We chose to normalize yield on a per plant basis to avoid confounding effects that may occur from density and planting conguration. Organic versus conventional growing method differences have been a topic of much research and debate. A literature review of nutritional differences by Magkos and colleagues39 found a slight trend toward higher ascorbic acid content in organic leafy vegetables and potatoes and a trend toward slightly lower (but higher-quality) protein levels in some organic crops compared to conventional counterparts. In another review article, organically grown produce was found to have signicantly more vitamin C, iron, magnesium, and phosphorus, as well as signicantly lower nitrate levels compared to conventionally grown produce,40 while others have indicated organic produce may be higher in antioxidants than conventional.41,42 However, these authors and others acknowledge that most organic and conventional research studies did not control cultivars, maturity at harvest, soil, and other environmental conditions that may impact results as much as, if not more than, growing method.13 15,26 Recently, more controlled studies have been conducted comparing various attributes of tomatoes grown on paired organic and conventional farms. In two studies done in California, researchers found the soluble solids content to be higher in the organic tomatoes than in the conventionally grown tomatoes.43,44 However, another project with a similar design conducted in Taiwan found no consistent differences between soluble solids content for organic and conventional tomatoes.45 All three of these studies also found no signicant yield differences or clear trends that one method produces higher antioxidant levels in tomatoes.43 45 Another study comparing tomatoes grown on organic and conventional plots on the same farm for 3 years found organic tomatoes to be higher in ascorbic acid and soluble solids content, but year and cultivar variability made it difcult to draw conclusions about other nutrient and quality characteristics studied.46 In a study by Caris-Veyrat et al.,47 organically grown tomatoes had signicantly higher levels of antioxidants compared to those grown conventionally when reporting on a fresh weight basis, but the results were not signicant on a dry weight basis. Our results indicate there may be a difference in yield potential, with conventional production yielding more tomatoes than organic production. This trend was also seen in a study done in Tunisia, which found organically grown tomato yields to be 63% of conventionally grown tomato yields.48 Riahi and colleagues found no consistent difference in quality and antioxidant results between the two growing systems. The results of our project found method overall to impact (P < 0.05) antioxidant capacity (ABTS and DPPH), ascorbic acid content, and soluble solids content, with organic tomatoes being higher in antioxidant capacity and ascorbic acid content and conventional tomatoes being higher in soluble solids content. With some results favoring each method, it is difcult to argue that one method produces higher quality tomatoes than the other; continued research in well-controlled farm settings is encouraged to further explore this question. Cultivar differences greatly inuenced (P < 0.05) all of the parameters studied in this project, and this is an area where farmers can easily make choices that will increase the quality and nutritional factors of the crops they grow. Chassy et al.46 also found cultivar to have the greatest impact on phytochemical results in a comparison of two cultivars grown organically and conventionally. High genetic diversity with regard to antioxidant

HT Aldrich et al.

potential was found in a screening of 50 tomato cultivars.49 Even though the fruit of many tomato cultivars look visually the same, slight changes in size and pigmentation may inuence antioxidant levels.50 The importance of genotype on antioxidant composition has also been documented in other fruit, such as strawberries,51 stone fruit,52 54 and blueberries.55 Growers and seed companies may wish to consider incorporating additional antioxidant and quality parameters into their cultivar trials in addition to yield, pest resistance, appearance, and other standard criteria. With research showing consumers may be willing to pay more for produce with higher nutritional levels, as well as produce that is produced locally or organically,12,18,56 this appears to be a promising area where farmers, especially those selling at farmers markets and other local venues, could differentiate their produce from others.

CONCLUSIONS
A better understanding of how the complex interaction of environmental effects, growing method, and cultivar choices inuence the quality and antioxidant properties of tomatoes and other fresh produce is fundamental to producing high-quality food. In this study, year-to-year variability and the production method used to grow tomatoes affected quality and nutritional characteristics, though the trends were not consistent. Genotype differences were shown to dramatically impact quality and antioxidant characteristics, and based on the cultivars used in this study, by choosing cultivars in the highest groups, a 1.35- to 1.67-fold gain in antioxidant levels could be achieved. Further research on the importance of cultivar choices that support production of nutritiously superior tomatoes could benet local producers who target consumers seeking nutritious produce.

ACKNOWLEDGEMENTS
This project was supported by the National Research Initiative of the USDA Cooperative State Research, Education and Extension Service, grant number 2005-55618-15634, and Colorado State University Agricultural Experiment Station project number 690.

REFERENCES
1 Lucier G and Dettman R, Vegetables and Melons Situation and Outlook Yearbook. [Online]. Economic Research Service, United States Department of Agriculture (2008). Available: http://www.ers.usda. gov/Publications/VGS/2008/05May/VGS2008.pdf [12 August 2008]. 2 Beecher GR, Nutrient content of tomatoes and tomato products. Proc Soc Exp Biol Med 218:98100 (1998). 3 Leonardi C, Ambrosino P, Esposito F and Fogliano V, Antioxidative activity and carotenoid and tomatine contents in different typologies of fresh consumption tomatoes. J Agric Food Chem 48:47234727 (2000). 4 Djuric Z and Powell LC, Antioxidant capacity of lycopene-containing foods. Int J Food Sci Technol 52:143149 (2001). 5 Willcox JK, Catignani GL and Lazarus S, Tomatoes and cardiovascular health. Crit Rev Food Sci 43:118 (2003). 6 Goldman IL, Recognition of fruit and vegetables as healthful: vitamins and phytonutrients. HortTechnology 13:252258 (2003). 7 Bazzano LA, The high cost of not consuming fruits and vegetables. J Am Diet Assoc 106:13641368 (2006). 8 Winter CK and Davis SF, Organic foods. J Food Sci 71:R117R124 (2006). 9 Shepherd R, Magnusson M and Sjoden PO, Determinants of consumer behavior related to organic foods. Ambio 34:352359 (2005). 10 Torjusen H, Lieblein G, Wandel M and Francis CA, Food system orientation and quality perception among consumers and producers of organic food in Hedmark County, Norway. Food Qual Pref 12:207216 (2001).

2554

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 25482555

Cultivar choice provides nutritious tomato options

11 Magnusson MK, Arvola A, Hursti UKK, Aberg L and Sjoden PO, Choice of organic foods is related to perceived consequences for human health and to environmentally friendly behaviour. Appetite 40:109117 (2003). 12 Yiridoe EK, Bonti-Ankomah S and Martin RC, Comparison of consumer perceptions and preference toward organic versus conventionally produced foods: a review and update of the literature. Renew Agric Food Syst 20:193205 (2005). 13 Harker FR, Organic food claims cannot be substantiated through testing of samples intercepted in the marketplace: a horticulturalists opinion. Food Qual Pref 15:9195 (2004). 14 Lester GE, Organic versus conventionally grown produce: quality differences, and guidelines for comparison studies. HortScience 41:296300 (2006). 15 Bourn D and Prescott J, A comparison of the nutritional value, sensory qualities, and food safety of organically and conventionally produced foods. Crit Rev Food Sci Nutr 42:134 (2002). 16 USDA-AMS, Farmers Market Growth. [Online]. United States Department of Agriculture Agricultural Marketing Service (2008). Available: http://www.ams.usda.gov/farmersmarkets/FarmersMarketGrowth. htm [25 May 2009]. 17 Keeling-Bond J, Thimany D and Bond C, Direct marketing of fresh produce: understanding consumer purchasing decisions. Choices 21:229235 (2006). 18 Bond CA, Thilmany D and Keeling Bond J, Understanding consumer interest in product and process-based attributes for fresh produce. Agribusiness 24:231252 (2008). 19 USDA-NOP, Code of Federal Regulations Title 7, Part 205: National Organic Program. (2010). [Online]. Available: http://www.access.gpo.gov/nara/cfr/waisidx 10/7cfr205 10.html [6 May 2010]. 20 Singleton VL and Rossi JA, Colorimetry of total phenolics with phosphomolybdicphosphotungstic acid reagents. Am J Enol Viticult 16:144158 (1965). 21 Spanos GA and Wrolstad RE, Inuence of processing and storage on the phenolic composition of Thompson seedless grape juice. J Agric Food Chem 38:15651571 (1990). 22 Rivera JRE, Stone MB, Stushnoff C, Pilon-Smits E and Kendall PA, Effects of ascorbic acid applied by two hydrocooling methods on physical and chemical properties of green leaf lettuce stored at 5 C. J Food Sci 71:S270S276 (2006). 23 Dale MFB, Grifths DW and Todd DT, Effects of genotype, environment, and postharvest storage on the total ascorbate content of potato (Solanum tuberosum) tubers. J Agric Food Chem 51:244248 (2003). 24 Miller NJ and Rice-Evans CA, Factors inuencing the antioxidant activity determined by the ABTS radical cation assay. Free Radic Res 26:195199 (1997). 25 Lu YR and Foo LY, Antioxidant and radical scavenging activities of polyphenols from apple pomace. Food Chem 68:8185 (2000). 26 Salandanan K, Bunning M, Stonaker F, Kulen O, Kendall P and Stushnoff C, Comparative analysis of antioxidant properties and fruit quality attributes of organically and conventionally grown melons (Cucumis melo L.). HortScience 44:18 (2009). 27 Lee SK and Kader AA, Preharvest and postharvest factors inuencing vitamin C content of horticultural crops. Postharvest Biol Technol 20:207220 (2000). 28 Dumas Y, Dadomo M, Di Lucca G and Grolier P, Effects of environmental factors and agricultural techniques on antioxidant content of tomatoes. J Sci Food Agric 83:369382 (2003). 29 Gautier H, Diakou-Verdin V, Benard C, Reich M, Buret M, Bourgaud F, et al, How does tomato quality (sugar, acid, and nutritional quality) vary with ripening stage, temperature, and irradiance? J Agric Food Chem 56:12411250 (2008). 30 Weston LA and Barth MM, Preharvest factors affecting postharvest quality of vegetables. HortScience 32:812816 (1997). 31 Mayer A-M, Historical changes in the mineral content of fruits and vegetables. Br Food J 99:207211 (1997). 32 Davis DR, Epp MD and Riordan HD, Changes in USDA food composition data for 43 garden crops, 1950 to 1999. J Am Coll Nutr 23:669682 (2004). 33 Davis DR, Declining fruit and vegetable nutrient composition: what is the evidence? HortScience 44:1519 (2009). 34 White PJ and Broadley MR, Historical variation in the mineral composition of edible horticultural products. J Hortic Sci Biotechnol 80:660667 (2005).

www.soci.org
35 White PJ, Bradshaw JE, Dale MFB, Ramsay G, Hammond JP and Broadley MR, Relationships between yield and mineral concentrations in potato tubers. HortScience 44:611 (2009). 36 Fan MS, Zhao FJ, Fairweather-Tait SJ, Poulton PR, Dunham SJ and McGrath SP, Evidence of decreasing mineral density in wheat grain over the last 160 years. J Trace Elem Med Biol 22:315324 (2008). 37 Garvin DF, Welch RM and Finley JW, Historical shifts in the seed mineral micronutrient concentration of US hard red winter wheat germplasm. J Sci Food Agric 86:22132220 (2006). 38 Farnham MW, Grusak MA and Wang M, Calcium and magnesium concentration of inbred and hybrid broccoli heads. J Am Soc Hortic Sci 125:344349 (2000). 39 Magkos F, Arvaniti F and Zampelas A, Organic food: nutritious food or food for thought? A review of the evidence. Int J Food Sci Technol 54:357371 (2003). 40 Worthington V, Nutritional quality of organic versus conventional fruits, vegetables, and grains. J Altern Complement Med 7:161173 (2001). 41 Zhao X, Carey EE, Wang WQ and Rajashekar CB, Does organic production enhance phytochemical content of fruit and vegetables? Current knowledge and prospects for research. HortTechnology 16:449456 (2006). 42 Brandt S, Pek Z, Barna E, Lugasi A and Helyes L, Lycopene content and colour of ripening tomatoes as affected by environmental conditions. J Sci Food Agric 86:568572 (2006). 43 Barrett DM, Weakley C, Diaz JV and Watnik M, Qualitative and nutritional differences in processing tomatoes grown under commercial organic and conventional production systems. J Food Sci 72:C441C451 (2007). 44 Pieper JR and Barrett DM, Effects of organic and conventional production systems on quality and nutritional parameters of processing tomatoes. J Sci Food Agric 89:177194 (2009). 45 Juroszek P, Lumkin HM, Yang RY, Ledesma DR and Ma CH, Fruit quality and bioactive compounds with antioxidant activity of tomatoes grown on-farm: comparison of organic and conventional management systems. J Agric Food Chem 57:11881194 (2009). 46 Chassy AW, Bui L, Renaud ENC, Van Horn M and Mitchell AE, Threeyear comparison of the content of antioxidant microconstituents and several quality characteristics in organic and conventionally managed tomatoes and bell peppers. J Agric Food Chem 54:82448252 (2006). 47 Caris-Veyrat C, Amiot MJ, Tyssandier V, Grasselly D, Buret M, Mikolajczak M, et al, Inuence of organic versus conventional agricultural practice on the antioxidant microconstituent content of tomatoes and derived purees: consequences on antioxidant plasma status in humans. J Agric Food Chem 52:65036509 (2004). 48 Riahi A, Hdider C, Sanaa M, Tarchoun N, Ben Kheder M and Guezal I, Effect of conventional and organic production systems on the yield and quality of eld tomato cultivars grown in Tunisia. J Sci Food Agric 89:22752282 (2009). 49 Hanson PM, Yang RY, Wu J, Chen JT, Ledesma D, Tsou SCS, et al, Variation for antioxidant activity and antioxidants in tomato. J Am Soc Hortic Sci 129:704711 (2004). 50 Toor RK, Lister CE and Savage GP, Antioxidant activities of New Zealand-grown tomatoes. Int J Food Sci Nutr 56:597605 (2005). 51 Scalzo J, Politi A, Pellegrini N, Mezzetti B and Battino M, Plant genotype affects total antioxidant capacity and phenolic contents in fruit. Nutrition 21:207213 (2005). 52 Gil MI, Tomas-Barberan FA, Hess-Pierce B and Kader AA, Antioxidant capacities, phenolic compounds, carotenoids, and vitamin C contents of nectarine, peach, and plum cultivars from California. J Agric Food Chem 50:49764982 (2002). 53 Drogoudi PD, Vemmos S, Pantelidis G, Petri E, Tzoutzoukou C and Karayiannis I, Physical characters and antioxidant, sugar, and mineral nutrient contents in fruit from 29 apricot (Prunus armeniaca L.) cultivars and hybrids. J Agric Food Chem 56:1075410760 (2008). 54 Tavarini S, DeglInnocenti E, Remorini D, Massai R and Guidi L, Preliminary characterisation of peach cultivars for their antioxidant capacity. Int J Food Sci Technol 43:810815 (2008). 55 Howard LR, Clark JR and Brownmiller C, Antioxidant capacity and phenolic content in blueberries as affected by genotype and growing season. J Sci Food Agric 83:12381247 (2003). 56 Lin BH, Smith TA and Huang CL, Organic premiums of US fresh produce. Renew Agric Food Syst 23:208216 (2008).

2555

J Sci Food Agric 2010; 90: 25482555

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

Research Article
Received: 5 April 2010 Revised: 15 June 2010 Accepted: 5 July 2010 Published online in Wiley Online Library: 5 August 2010

(wileyonlinelibrary.com) DOI 10.1002/jsfa.4118

Inhibition of enzymatic browning in actual food systems by the Maillard reaction products
Burce Atac Mogol, Asl Yldrm and Vural Gokmen
Abstract
BACKGROUND: The Maillard reaction occurring between amino acids and sugars produces neo-formed compounds having certain levels of antioxidant activity depending on the reaction conditions and the type of reactants. The objective of this study was to investigate enzymatic browning inhibition capacity of Maillard reaction products (MRPs) formed from different amino acids including arginine (Arg), histidine (His), lysine (Lys) and proline (Pro). RESULTS: The inhibitory effects of the MRPs on polyphenol oxidase (PPO) were determined. The total antioxidant capacity (TAC) of MRPs derived from different amino acids were in the order Arg > His > Lys > Pro. The TAC and PPO inhibition of MRPs were evaluated as a function of temperature (80120 C), time (16 h) and pH (212). Arg-Glc and His-Glc MRPs exhibited strong TAC and PPO inhibition. Increasing temperature (up to 100 C) and time also increased TAC and PPO inhibition. Kinetics analysis indicated a mixed type inhibition of PPO by MRPs. CONCLUSION: The results indicate that the MRPs derived from Arg and His under certain reaction conditions signicantly prevent enzymatic browning in actual food systems. The intermediate compounds capable of preventing enzymatic browning are reductones and dehydroreductones, as conrmed by liquid chromatographicmass spectrometric analyses. c 2010 Society of Chemical Industry Keywords: Maillard reaction products; enzymatic browning inhibition; polyphenol oxidase; reductones; dehydroreductones

INTRODUCTION
Enzymatic browning is catalyzed by polyphenol oxidase (PPO; EC 1.10.3.1), which oxidizes o-quinones to polymerized dark-colored pigments. This enzymatic reaction is considered to be the main contributor to browning in fruits and vegetables and greatly affects nutritional, functional and organoleptic properties of the product.1 There are many chemical compounds with different mechanisms avoiding or minimizing this discoloration. Sulting agents are the most effective and the cheapest antibrowning agents. They act both on o-quinones and enzymes. However, they have adverse health effects on asthmatic people and their use tends to be limited or even prohibited by food regulations.2,3 Thus the search for alternative inhibitors in the prevention of such oxidative reactions has great importance. The Maillard reaction is a chemical reaction between amino groups and reducing sugars and forms colored and/or colorless products. The functional properties of the Maillard reaction products depend on the reaction conditions,4 pH,5 and type of reactants.6 The MRPs have been found to have antioxidant properties.7 MRPs formed from glycine and glucose, cysteinerelated compounds and various monosaccharides have been reported as natural antibrowning agents.2,8 10 The objective of this study was to investigate the antioxidant capacity of MRPs formed from different amino acids, and their inhibitory effect on enzymatic browning. The Maillard reaction intermediates responsible for the inhibition of antioxidant activity were identied by liquid chromatography coupled to mass spectrometry (LC-MS). The inhibition of apple PPO was kinetically characterized in the presence of MRPs. The antibrowning potential

of MRPs was also tested for apple pur homogenized with MRP, ee and for potato cubes dipped in MRP solution.

EXPERIMENTAL
Chemicals and consumables Apples of the variety Golden Delicious and potato tubers of the variety Agria were obtained from a local supermarket in Ankara, Turkey. Glucose, arginine, histidine, proline, disodium hydrogen phosphate anhydrous and potassium dihydrogen phosphate were purchased from Merck (Darmstadt, Germany). Lysine, catechol, citric acid, polyvinylpolypyrrolidone (PVPP), methanol, ABTS (2,2 azino-bis(3-ethylbenz-thioline-6-sulfonic acid)) and DPPH (1,1diphenyl-2-picrylhydrazyl) were purchased from Sigma (St Louis, MO, USA). Trolox (6-hydroxy-2,5,7,8-tetra-methylchromone-2carboxylic acid) and potassium persulfate (di-potassium peroxdisulfate) were purchased from Fluka (Buchs, Switzerland). Preparation of Maillard reaction products Equimolar amounts of different amino acids (Arg, His, Lys and Pro) and glucose (Glc) were mixed in 1.5 mL glass vials to a nal concentration of 0.125 mol L1 . The mixtures were heated in an oven (Memmert UNE 400, Braunschweig, Germany) at different

Correspondenceto:Vural G kmen,DepartmentofFoodEngineering,Hacettepe o University, 06800 Beytepe, Ankara, Turkey. E-mail: vgokmen@hacettepe.edu.tr Department of Food Engineering, Hacettepe University, 06800 Beytepe, Ankara, Turkey

2556

J Sci Food Agric 2010; 90: 25562562

www.soci.org

c 2010 Society of Chemical Industry

Maillard reaction products inhibit enzymatic browning temperatures (80120 C) for different times (16 h). In order to determine the effect of pH, initial pH values of the reaction mediums were adjusted between 2.0 and 12.0 with H3 PO4 (85%) and NaOH (4 mol L1 ). The mixtures were heated at 100 C for 3 h. The color of MRPs was measured as absorbance at 420 nm using a variable-wavelength UV-visible spectrometer (model 2101 PC, Shimadzu, Kyoto, Japan).

www.soci.org PPO inhibition kinetics To determine the kinetics of PPO inhibition by MRPs, catechol concentration was varied from 3.33 to 166.67 mmol L1 in the reaction volume of 3 mL. MRPs (100 L) prepared from Arg-Glc or His-Glc was added as inhibitors for each substrate concentration. The reaction rates were determined from the initial slopes of absorbance versus time plots (Abs. min1 ). Inhibition constants (Ki and Ki ) were calculated according to LineweaverBurk plots.14 The velocity equation for mixed type inhibition is V= Vmax [S] Km + [S] (2)

Measurement of the total antioxidant capacity (TAC) The TAC of MRPs was measured by using DPPH and ABTS radical scavenging methods. The results were expressed as mmol L1 of Trolox equivalent antioxidant capacity (TEAC).

and the LineweaverBurk equation for this type inhibition is DPPH radical scavenging method A working solution of DPPH radicals was prepared in methanol to a nal absorbance of 1.00 at 517 nm.11 The reaction mixture containing 1.9 mL DPPH working solution and 0.1 mL MRPs was kept at room temperature. Absorbance was measured at 517 nm exactly after 30 min using a Shimadzu 2101 PC model UV-visible spectrophotometer. The control was prepared in the same manner, except that distilled water was added instead of MRPs. 1 1 Km 1 = + V Vmax [S] Vmax where =1+ and =1+ [I] Ki [I] Ki (3)

(4)

(5)

ABTS radical scavenging method The blue-green ABTS stock solution was produced by reacting 7 mmol L1 ABTS with 2.45 mmol L1 potassium persulfate and allowing the mixture to stand in the dark at room temperature for 1216 h before use.12 The ABTS working solution was obtained by diluting the stock solution with water to an absorbance of 0.70 (0.02) at 734 nm. The reaction mixture containing 1.95 mL ABTS and 50 L MRPs was kept at room temperature. The absorbance was measured at 734 nm exactly after 6 min using a Shimadzu 2101 PC model UV-visible spectrophotometer. The control was prepared with distilled water as described previously.

Antibrowning activity of MRPs Antibrowning activity of MRPs was tested in apple pur and ee potato cubes as the actual food systems. Apple (2 g) was homogenized with 2 mL Arg-Glc or His-Glc MRPs and the volume was made up to 12 mL with distilled water. The control was prepared with 2 mL distilled water instead of MRPs. The potato tuber was washed, peeled and sliced into 1 cm cubes. Three cubes of potato were dipped into 20 mL MRPs for 1 min. Change of color was monitored using a color spectrometer (CM-3600d, Minolta,) up to 2 h and 6 h for apple puree and potato cubes, respectively. The total color difference E was calculated for each time interval using the following equation: E= (L L )2 + (a a )2 + (b b )2 t t 0 0 0 0 (6)

Inhibition of PPO PPO extraction Washed, cored and peeled apples were used to extract PPO. Apple (100 g) was blended at 4 C with 160 mL of pH 6.5 phosphate buffer. PVPP (2 g) was added to prevent excessive browning during blending. The homogenate was squeezed through layers of cloth and centrifuged at 10 000 g for 15 min at 4 C. The supernatant was used as crude PPO extract.

The browning inhibition was calculated as follows: Browning Inhibition % = 1 AUC (inhibitor) AUC (control) 100 (7)

Measurement of PPO activity The PPO activity was assayed spectrophotometrically.13 Inhibition studies were carried out under buffered conditions. The sample cuvette contained 1.6 mL of pH 4.5 citratephosphate buffer (Mc Ilvaines), 200 L MRPs, 200 L of 1 mol L1 catechol as substrate and 1 mL PPO extract (3 mL total volume). PPO was added last to initiate the enzymatic reaction. The control was prepared with distilled water instead of MRPs. The increase in absorbance at 400 nm was recorded up to 3 min. The reaction rate was calculated from the initial slope of the progress curve and percentage PPO inhibition values were calculated as follows: PPO Inhibition (PPO) = 1 Initial rate (sample) Initial rate (control) 100 (1)

where AUC stands for the area under the curve. The trapezoid rule was used to calculate AUC from the plots of E versus time up to 2 h and 6 h for apple puree and potato cubes, respectively. Analysis of MRPs by LC-MS LC-MS analyses of MRPs were performed using an Agilent 1200 HPLC system (Waldbronn, Germany) consisting of a binary pump, autosampler and temperature-controlled column oven, coupled to an Agilent 6130 MS detector equipped with an electrospray ionization (ESI) interface operated using the following interface parameters: drying gas (N2 , 30 psig) ow 13 L min1 , nebulizer pressure 30 psig, drying gas temperature 300 C, capillary voltage 4 kV, and fragmenter voltage 80 eV. Analytical separation was performed on an Atlantis T3 column (2504.6 mm, i.d. 5 m) using the isocratic mixture of 10 mmol L1 formic acid and methanol (90 : 10, v/v) at a ow rate of 0.5 mL min1

2557

J Sci Food Agric 2010; 90: 25562562

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org at 25 C. The LC eluent was directed to the MS system for the detection of Maillard reaction intermediates formed during heating in Arg-Glc and His-Glc mixtures. Data acquisition was performed in scan mode with positive ionization in an m/z range of 501000. The presence of intermediate compounds such as N-glycosylamine, reductone, dehydroreductone, Schiff base and decarboxylated Schiff base of Arg and His were investigated in MRPs. Total ion chromatograms were extracted to conrm the presence of corresponding intermediate structures. Statistical analysis The least signicance difference test was used to determine signicant difference between means at a signicance level of P < 0.05.

BA Mogol, A Yldrm, V Gokmen the TAC of resulting MRPs for all amino acids. The TAC of MRPs derived from different amino acids were in the order Arg > His > Lys > Pro. It should be noted that Lys-Glc and Pro-Glc MRPs had very low levels of TAC compared to Arg-Glc and His-Glc MRPs. For example, the MRPs derived from Arg at 100 C for 3 h had approximately 50 times greater TAC than the MRPs prepared from Lys and Pro under the same conditions (P < 0.05). It has been previously reported that His-Glc MRPs possess peroxyl radical scavenging activity assayed by the ORAC method.15 The TAC values determined by ABTS and DPPH methods were found to be comparable (data not shown). There was a positive correlation (r = 0.93) between the TAC measures obtained by ABTS and DPPH methods for His-Glc MRPs. In this study, inhibition potency of MRPs derived from different amino acids was also characterized on apple PPO. Figure 2 shows the PPO inhibition percentages of MRPs together with their TAC values. The reaction mixtures prepared from Arg-Glc and His-Glc had a PPO inhibition potency of 15.01% and 13.68%, respectively. In general, MRPs having higher TAC also showed higher percentage inhibition of PPO regardless of the type of amino acid. There was a positive correlation between the TAC and PPO inhibition percentages of MRPs. The correlation coefcient was determined to be 0.81 and 0.83 for Arg-Glc and His-Glc MRPs, respectively. Similar to TAC values, MRPs derived from Arg were found to be the best inhibitor of PPO, followed by MRPs derived from His, Lys and Pro (Fig. 2(a, b)). At 100 C, increasing the reaction time also increased PPO inhibition potency of MRPs derived from Arg and His, while it was relatively stable for MRPs derived from Lys (Fig. 2(c)). MRPs formed at temperatures above 100 C showed different PPO inhibition potency. The PPO inhibition capacity of MRPs rst increased to an apparent maximum and then decreased gradually regardless of the type of amino acid (data not shown). This fact has been explained as the progressive formation of
(b)

RESULTS AND DISCUSSION

Effect of heating time and temperature on TAC and PPO inhibition of MRPs Some MRPs, especially melanoidins, have been known to have antioxidant activity through scavenging oxygen radicals or chelating metals.15,16 In this study, effect of heating time and temperature on TAC and PPO inhibition capability of the MRPs obtained from different amino acids (Arg, His, Lys and Pro) and sugars (Glc and Fru) were investigated. Preliminary studies revealed that MRPs prepared by glucose showed greater TAC and PPO inhibition than those obtained from fructose (P < 0.05) (data not shown). Other researchers have previously reported similar results.2,10 Figure 1 shows the TAC of MRPs prepared from different amino acids as inuenced by temperature and heating time. The reaction mixtures were found to have TAC levels less than 0.05 mmol L1 TEAC initially. In general, increasing the temperature also increased
(a)

(c)

(d)

Figure 1. The TAC of MRPs determined using DPPH radical scavenging method: (a) Arg-Glc; (b) His-Glc; (c) Lys-Glc; (d) Pro-Glc (vertical bars indicate standard deviations).

2558

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 25562562

Maillard reaction products inhibit enzymatic browning

(a) (b)

www.soci.org

(c)

Figure 2. TAC and PPO inhibition properties of MRPs synthesized from different amino acids under different conditions: (a) 100 C for 3 h; (b) 120 C for 3 h; (c) 100 C for 1, 2, 3 and 6 h (vertical bars indicate standard deviations).

insoluble melanoidins during heating.17 However, our results did not conrm this explanation, because the MRPs obtained were optically clear. No insoluble melanoidins formed under the reaction conditions applied in this study. Therefore, intermediates formed during the Maillard reaction rather than melanoidins are thought to be responsible for the inhibition of PPO. Color was measured as absorbance at 420 nm in MRPs. In general, Arg generated darker MRPs during heating, followed by His, Lys and Pro (data not shown). This order was valid for the TAC, which meant there was a positive correlation between browning intensity and TAC. Other researchers have also found positive correlation between color and antioxidant properties in processed foods.18 Even though Lys was known to be a very reactive amino acid in the Maillard reaction,19 it was surprising that the MRPs derived from Lys showed very low levels of TAC and PPO inhibition capability. Therefore, this study was continued with Arg-Glc and His-Glc MRPs for further characterization. Effect of initial pH of MRPs on enzyme inhibition and antioxidant activity Initial pH of the reaction mixture (Arg-Glc or His-Glc) was adjusted between 2.0 and 12.0 in order to see the effect of pH on TAC and PPO inhibition of MRPs formed at 100 C for 3 h. Initial pH of the reaction mixture had a signicant effect on the characteristics of MRPs. As shown in Table 1, the MRPs derived from Arg under acidic conditions had very low levels of TAC and PPO inhibition capacity. For Arg, reaction conditions having pH 10.02 and 11.97 were found more favorable to form MRPs having both higher TAC and PPO inhibition. The PPO inhibition percentage was highest at a pH of 11.97 for Arg-Glc MRPs. For His, MRPs formed at pH values of 6.00 and 8.02 showed remarkable PPO inhibition capacity. The initial pH values of unheated Arg-Glc and His-Glc solutions are

Table 1. Inuence of initial pH on TAC and PPO inhibition capacity of MRPs prepared from Arg-Glc and His-Glc mixtures by heating at 100 C for 3 h pH MRP Arg-Glc Initial 2.02 2.98 3.98 5.08 6.06 8.06 10.02 11.97 2.02 2.99 4.04 5.05 6.00 8.02 10.00 11.98 Final 2.28 2.86 4.00 4.79 5.47 6.59 8.95 9.32 2.05 3.00 3.99 4.87 5.72 7.14 8.74 9.27 TAC (mmol L1 TEAC) 0.39 0.02f 0.35 0.02f 0.22 0.01f 1.32 0.07e 6.46 0.32d 11.50 0.57a 8.36 0.42c 9.89 0.49b 0.68 0.03f 0.73 0.04f 0.75 0.04f 1.57 0.08e 3.21 0.16c 2.13 0.11d 8.32 0.42b 12.74 0.64a PPO inhibition (%) 12.12 0.61o 7.48 0.37p 12.52 0.36o 16.94 0.85n 24.08 1.20m 10.07 0.50of 29.86 1.49l 83.11 4.16k 0.00n 0.00n 0.00n 19.46 0.97l 54.50 2.72k 55.14 2.76k 12.54 0.63m 10.63 0.53m

His-Glc

Different letters and letters with prime symbols indicate statistically different groups for Arg-Glc and His-Glc MRPs, respectively.

10.76 and 7.59, respectively. These pH values make MRPs from both amino acids practically applicable without changing their initial pH values. The absorbance of MRPs measured at 420 nm changed as the initial pH of reaction medium changed. Absorbance increased

2559

J Sci Food Agric 2010; 90: 25562562

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org gradually with pH for both Arg and His (data not shown). The MRPs derived from His-Glc had lighter color intensity, whereas Arg-Glc MRPs were darker. Inhibition kinetics To characterize the inhibition kinetics of MRPs on PPO, Arg-Glc and His-Glc MRPs were chosen from among others because they showed higher TAC and PPO inhibition percentages. A mixed type inhibition activity of Arg-Glc and His-Glc MRPs on PPO was determined from the LineweaverBurk double reciprocal plots (Fig. 3). The Km and Vmax values of enzyme without inhibitor were calculated as 7.10 mmol L1 and 0.24 Abs. min1 , respectively. The Km increased while Vmax reduced in the presence of inhibitors. In mixed type inhibition, inhibitor can interact not only with the free enzyme but also with enzymesubstrate complex at a site other than the active site. The Ki and Ki values were calculated as 22.79 L and 82.90 L for Arg-Glc MRPs by means of the LineweaverBurk plots and related equations (Eqns (2)(5)). The Ki and Ki values were signicantly higher for His-Glc MRPs than Arg-Glc MRPs, and determined as 49.98 L and 159.27 L, respectively. The higher constants calculated for the MRPs derived from His conrmed that they were stronger PPO inhibitors (P < 0.05). For both Arg-Glc and His-Glc MRPs, Ki values were lower than Ki values, which indicated that the afnity of MRPs towards the free enzyme was higher than the enzymesubstrate complex. Brun-M ee et al. have reported that MRPs prepared using an erim aqueous model system containing equimolar glucose or fructose with glutathione (GSH) were mixed-type inhibitors, glucose/GSH being the most efcient model system.20 Potential use of MRPs as antibrowning agents in foods To test the potential use of Arg-Glc and His-Glc MRPs for the inhibition of PPO in actual food systems, apple puree and potato cubes were used as the model systems. Apple pur was prepared ee by homogenizing 10 g apple slices with 1 mL MRPs or 1 mL water for control. Measuring the color for up to 2 h monitored browning tendency of the puree. Total color difference ( E) was calculated to evaluate the antibrowning activity of MRPs in apple puree. Figure 4(a) shows the change of E with time in apple puree in the presence of His-Glc MRPs. It is clear from the results that E rapidly increased, reaching a plateau value within 1 h at room temperature. Compared to the control, addition of His-Glc MRPs signicantly decreased the plateau value of E attained after 1 h

BA Mogol, A Yldrm, V Gokmen in the apple puree model system (P < 0.05). From a practical point of view, a E value that exceeded 3.0 was undesirable because the color of the puree became apparently brownish after this point. The color turned to brownish ( E > 3.0) within 10 min in control, while it was relatively stable for 2 h in pur treated with His-Glc ee MRPs. Change in the surface color of potato cubes dipped in MRPs or water as control was monitored for 6 h. Similar to the results obtained in the apple puree system, enzymatic browning could be signicantly prevented in potato cubes dipped in MRPs (Fig. 4(b)). Using Eqn (7), percentage inhibition of browning was calculated by comparing the area under the curves of E versus time plots for apple puree and potato cubes. In apple puree kept at room temperature for 2 h, browning inhibition rates were 75% 4% and 52% 3% for Arg-Glc and His-Glc MRPs, respectively. Browning rates could be effectively decreased in potato cubes dipped in MRPs derived from Arg and His. In potato cubes, browning rates decreased by 51% 3% and 53% 3% for Arg-Glc and His-Glc MRPs, respectively, within 6 h of storage at room temperature. These results revealed that the MRPs derived from Arg and His can act as potential antibrowning agents in actual food systems. It has been reported that Arg-Glc MRPs are limited in practical use because of their intense color.21 Since they have a lighter color, His-Glc MRPs are considered more desirable for food applications.

(a)

(b)

Figure 3. LineweaverBurk double-reciprocal plots showing inhibition of PPO by MRPs synthesized at 100 C for 3 h.

Figure 4. Change of total color difference ( E) with time in actual food systems as inuenced by MRPs: (a) apple pur e homogenized with water e (control) and His-Glc MRPs prepared at 100 C for 3 and 6 h; (b) potato cubes dipped in water (control), His-Glc MRPs prepared at 100 C for 6 h, and Arg-Glc MRPs prepared at 100 C for 3 h.

2560

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 25562562

Maillard reaction products inhibit enzymatic browning Analysis of MRPs by LC-MS The MRPs derived from Arg and His were analyzed by LCMS in scan mode with positive ionization in an m/z range of 501000 to conrm the formation of reductones and dehydroreductones as potential antioxidants and antibrowning agents. Total ion chromatograms were extracted to conrm the presence of corresponding intermediate structures using the parent and compound-specic ions having certain m/z values for Arg-Glc and His-Glc. In positive electrospray ionization mode, the parent ions used to conrm N-glycosylamines, reductones, dehydroreductones and Schiff bases were 337, 301, 299, and 283, respectively, for Arg-Glc MRPs. They were 318, 282, 280, and 264 for His-Glc MRPs. The extracted ion chromatograms conrmed the presence of reductones and dehydroreductones, as well as N-glycosylamines and Schiff bases in the MRPs. Among the intermediates, N-glycosylamines were the most abundant forms in the MRPs prepared by heating mixtures of Arg-Glc and His-Glc at 100 C. This indicated that the Maillard reaction was at early and intermediate stages under the reaction conditions applied in this study. The rate of change of reductones in the MRPs of Arg-Glc and His-Glc was comparable during heating at 100 C for up to 6 h (Fig. 5(a)), but levels of dehydroreductones were signicantly higher in the Arg-Glc mixture than in the His-Glc mixture (Fig. 5(b)). It is thought that the initial pH of 10.47 accelerated the formation of dehydroreductones in the Arg-Glc mixture during heating. Since the reaction products were mainly characterized by the intermediate compounds formed during the Maillard reaction at 100 C, high TAC and PPO inhibition capacity should be due to these intermediates rather than high-molecularweight melanoidins. Among these intermediates, reductones and dehydroreductones are capable of active involvement in

www.soci.org oxidationreduction reactions, possibly by a hydrogen transfer mechanism.22 Due to the fact that we determined signicant levels of reductones and dehydroreductones in the MRPs derived from Arg and His, these reductones formed during the Maillard reaction might be responsible for reversing the oxidation of polyphenols to quinones catalyzed by PPO.

CONCLUSION
Functional characteristics of MRPs were investigated in terms of TAC, PPO inhibition and antibrowning capacity in this study. The type of amino acids, initial pH of the reaction medium, temperature and time were found to inuence these characteristics in the resulting MRPs. Among others, the MRPs derived from Arg and His under certain reaction conditions had higher TAC and signicantly prevented enzymatic browning in actual food systems. Owing to its lighter color intensity, the MRPs from His are considered more desirable for food applications. In-depth mass spectrometric analyses conrmed the presence of reductones and dehydroreductones in MRPs derived from Arg and His. These compounds, especially reductones, are thought to be potentially active inhibitors of enzymatic browning. Since similar MRPs are produced under food-processing conditions such as baking and roasting, they are considered to be usual constituents of foods as PPO inhibitors that may facilitate their approval as antibrowning agents. However, the absence of any harmful compound in the mixture of MRPs that may be formed during the Maillard reaction should be conrmed by detailed investigation. If desired, removal of aroma compounds in the mixture of MRPs should also be considered to prevent any adverse effect on the sensorial properties of foods.

(a)

ACKNOWLEDGEMENTS
We thank the Scientic and Technical Research Council of Turkey for nancial support (Project TOVAG COST 928).

REFERENCES
1 Mayer AM and Harel E, Polyphenol oxidases in plants. Phytochemistry 18:193215 (1979). 2 Billaud C, Maraschin C and Nicolas J, Inhibition of polyphenoloxidase from apple by Maillard reaction products prepared from glucose or fructose with L-cysteine under various conditions of pH and temperature. LWT Food Sci Technol 37:6978 (2004). 3 Nicolas J, Richard-Forget F, Goupy PM, Amiot M-J and Aubert SY, Enzymatic browning reactions in apple and apple products. Crit Rev Food Sci Nutr 34:109157 (1994). 4 Renn PT and Sathe SK, Effects of pH, temperature, and reactant molar ratio on L-leucine and D-glucose Maillard browning reaction in an aqueous system. J Agric Food Chem 45:37823787 (1997). 5 Cioroi M, Leonte M, Munari M, Mastracola D and Lerici CR, Chemicalphysical changes in sugaramino acid model systems due to Maillard reaction. Note 1. The evolution of the pH and the colour of the glucoselysine model system heat treated. Rev Roum Chimie 45:153157 (2000). 6 Ajandouz EH and Puigserver A, Nonenzymatic browning reaction of essential amino acids: effect of pH on caramelization and Maillard reaction kinetics. J Agric Food Chem 47:17861793 (1999). 7 Franzke C and Iwansky H, Zur antioxydativen Wirksamveit der Melanoidine. Dtsch Lebens Rundsch 50:251254 (1954). 8 Lee M-K and Park I, Inhibition of potato polyphenol oxidase by Maillard reaction products. Food Chem 91:5761 (2005). 9 Nicoli MC, Elizalde BE, Pitotti A and Lerici CR, Effects of sugars and Maillard reaction products on polyphenol oxidase and peroxidase activity in food. J Food Biochem 15:169184 (1991).

(b)

Figure 5. Changes in peak areas for reductones and dehydroreductones in MRPs prepared at 100 C: (a) Arg-Glc; (b) His-Glc.

2561

J Sci Food Agric 2010; 90: 25562562

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org
10 Billaud C, Maraschin C, Chow Y-N, Cheriot S, Peyrat-Maillard M-N and Nicolas J, Maillard reaction products as natural antibrowning agents in fruit and vegetable technology. Mol Nutr Food Res 49:656662 (2005). 11 Brand-Williams W, Cuvelier ME and Berset C, Use of a free radical method to evaluate antioxidant activity. LWT Food Sci Technol 28:2530 (1995). 12 Re R, Pellegrini N, Proteggente A, Pannala A, Yang M and Rice-Evans C, Antioxidant activity applying an improved ABTS radical cation decolorization assay. Free Radic Biol Med 26:12311237 (1999). 13 Billaud C, Brun-M rim e S, Louarme L and Nicolas J, Effect of e e glutathione and Maillard reaction products prepared from glucose or fructose with glutathione on polyphenoloxidase from apple. I. Enzymatic browning and enzyme activity inhibition. Food Chem 84:223233 (2004). 14 Lineweaver H and Burk D, The determination of enzyme dissociation constants. J Am Chem Soc 56:658666 (1934). 15 Yilmaz Y and Toledo R, Antioxidant activity of water-soluble Maillard reaction products. Food Chem 93:273278 (2005). 16 Rendleman JA, Complexation of calcium by melanoidin and its role in determining bioavailability. J Food Sci 52:16991705 (1987).

BA Mogol, A Yldrm, V Gokmen

17 Maillard M-N, Billaud C, Chow C-N, Ordonaud C and Nicolas J, Free radical scavenging, inhibition of polyphenoloxidase activity and copper chelating properties of model Maillard systems. LWT Food Sci Technol 40:14341444 (2007). 18 Manzocco L, Calligaris S, Mastracola D, Nicoli MC and Lerici CR, Review of nonenzymatic browning and antioxidant capacity in processed foods. Trends Food Sci Technol 11:340346 (2001). 19 Morales FJ and Jim nez-Perez S, Free radical scavenging capacity of e Maillard reaction products as related to colour and uorescence. Food Chem 72:119125 (2001). 20 Brun-M rim e S, Billaud C, Louarme L and Nicolas J, Effect of e e glutathione and Maillard reaction products prepared from glucose or fructose with glutathione on polyphenoloxidase from apple. II. Kinetic study and mechanism of inhibition. Food Chem 84:235241 (2004). 21 Tan BK and Harris ND, Maillard reaction products inhibit apple polyphenoloxidase. Food Chem 53:267273 (1995). 22 Belitz H-D, Grosch W and Schieberle P, Food Chemistry (4th edn). Springer, Berlin, pp. 248337 (2009).

2562
wileyonlinelibrary.com/jsfa
c 2010 Society of Chemical Industry J Sci Food Agric 2010; 90: 25562562

Research Article
Received: 30 April 2010 Revised: 18 June 2010 Accepted: 7 July 2010 Published online in Wiley Online Library: 24 August 2010

(wileyonlinelibrary.com) DOI 10.1002/jsfa.4120

Development of efcient enzymatic production of theanine by -glutamyltranspeptidase from a newly isolated strain of Bacillus subtilis, SK11.004
Yuying Shuai,a,b Tao Zhang,a Bo Jianga, and Wanmeng Mua
Abstract
BACKGROUND: Theanine, a unique amino acid found almost exclusively in tea plants, has various favourable physiological and pharmacological functions in humans. Gamma-glutamyltranspeptidase (GGT, EC 2.3.2.2) is considered to be the most effective enzyme for the production of theanine. In fact, GGT can catalyse the transfer of -glutamyl moieties from -glutamyl compounds to water (hydrolysis) or to amino acids and peptides (transpeptidation). RESULTS: A novel strain, SK11.004, which produces GGT with high theanine-forming ability was isolated from fermented shrimp paste and identied as Bacillus subtilis through its physiological and biochemical properties as well as its 16S rDNA sequence analysis. Theanine (18.9 mmol L1 ) was synthesised by GGT (0.06 U mL1 ) through transfer reaction in the presence of glutamine (20 mmol L1 ) as a donor and ethylamine HCl (50 mmol L1 ) as an acceptor at pH 10 and 37 C for 4 h, the conversion rate being up to 94%. CONCLUSION: The enzymatic synthesis of theanine using GGT from a newly isolated strain Bacillus subtilis SK11.004 was found to be an efcient method. Moreover, compared with others, the GGT from B. subtilis SK11.004 exhibited the highest ratio of transferring activity to hydrolytic activity using glutamine, suggesting a high potential application in the production of theanine and other functional -glutamyl compounds. c 2010 Society of Chemical Industry Keywords: theanine; gamma-glutamyltranspeptidase; enzymatic synthesis; Bacillus subtilis

INTRODUCTION
Theanine ( -glutamylethylamide), a unique amino acid found almost exclusively in tea plants, elicits umami (avour potential) and a sweet taste, which determine the quality of green tea. Moreover, theanine offers various favourable physiological and pharmacological functions: it promotes relaxation,1 helps to reduce blood pressure,2 inhibits the negative effects of caffeine,3 enhances anti-tumour activity,4 provides neuroprotection,5 acts as an anti-obesity compound,6 and improves learning ability.7 The US Food and Drug Administration (FDA) admitted and accepted theanine early in 1985, and conrmed that theanine is Generally Recognised As Safe (GRAS) with no limitation on the quantity during application. For instance, Sun-theanine (Taiyo Kagaku, Co. Ltd, Yokkaichi, Mie, Japan) has been used to facilitate relaxation and in the treatment of anxiety and depression, which are prominent psychiatric illnesses in todays society. Therefore, many attempts have been made to produce theanine commercially, including tea callus cultivation, chemical synthesis, extraction from fresh leaves of the tea plant and enzymatic synthesis. The latter has been intensively investigated, involving theanine synthetase (EC 6.3.1.6), glutamine synthetase (EC 6.3.1.2),8 glutaminase (EC 3.5.1.2) and -glutamyltranspeptidase (gamma-glutamyltransferase, GGT, EC 2.3.2.2).9 However, most of

these methods appear to use complicated processes with a low yield and are expensive. GGT catalyses the transfer of the -glutamyl moieties from glutamyl compounds to water (hydrolysis) or to amino acids and peptides (transpeptidation). Its transpeptidation reaction could be exploited for the production of theanine, and seemed to be one of the most effective methods due to the availability of enzyme source, short reaction time and non-requirement of ATP. However, theanine was produced at a relatively low yield with a high quantity of the by-product -glutamylglutamine using GGT from Escherichia coli K-12.9 In fact, various microorganisms such as Aspergillus oryzae,10 E. coli K-12,11 Helicobacter pylori,12 Proteus mirabilis,13 Saccharomyces cerevisiae,14 Treponema denticola15 and many Bacillus species, including Bacillus sp. KUN-17,16 Bacillus

Correspondence to: Bo Jiang, State Key Laboratory of Food Science and Technology, Jiangnan University, 1800 Lihu Avenue, 214122 Wuxi, Jiangsu, China. E-mail: bjiang2001@hotmail.com

a State Key Laboratory of Food Science and Technology, Jiangnan University, 1800 Lihu Avenue, 214122 Wuxi, Jiangsu, China

2563

b School of Food Science and Technology, Jiangnan University, 1800 Lihu Avenue, 214122 Wuxi, Jiangsu, China

J Sci Food Agric 2010; 90: 25632567

www.soci.org

c 2010 Society of Chemical Industry

www.soci.org pumilus,17 Bacillus subtilis 168,18 Bacillus subtilis NAFM5,19 Bacillus subtilis natto20 and Bacillus subtilis NX-2,21 have been found to produce GGT. Fermented shrimp paste, a traditional and popular seasoning in Asian countries, was shown to contain many strains of Bacillus species.22 The main objective of this study was to isolate strains producing GGT with high theanine-forming activity from a traditional fermented shrimp paste and develop a more efcient method to produce theanine.

Y Shuai et al.

was calculated from the difference in absorbance at 410 nm between reaction mixtures with and without glycylglycine. One unit of GGT was dened as the amount of enzyme that released 1 mol of p-nitroaniline per minute from -GpNA through the transpeptidation reaction. Enzymatic synthesis of theanine with -glutamyltranspeptidase The reaction system (10 mL in a ask) constituted of 0.04 U mL1 GGT, 20 mmol L1 glutamine (Gln) and 100 mmol L1 ethylamine HCl. The reaction was carried out in a shaking water bath at 37 C and 150 rpm for 5 h. Reaction mixtures without the addition of GGT served as controls and were incubated similarly. The reaction was terminated by the addition of 100 g L1 trichloroacetic acid. Analysis of theanine, glutamine and glutamic acid Theanine, Gln and glutamic acid (Glu) were determined using HPLC equipped with a C18 ODS HYPERSIL column (Agilent 1200; Agilent Technologies, Palo Alto, CA, USA), with a gradient elution at 40 C and a ow rate of 1 mL min1 . The gradient of mobile phase was formed with buffer A (8 g L1 CH3 COONa3H2 O, containing 0.16 mg L1 triethylamine, and 4.4 mg L1 tetrahydrofuran, pH 7.2), and buffer B (20 g L1 CH3 COONa3H2 O, pH 7.2/acetonitrile/methanol (1 : 2 : 2, by volume). A/B ratios were 92 : 8, 62 : 38, 0 : 100, 0 : 100 and 92 : 8, at run times of 0, 20, 24, 25.5 and 28.5 min, respectively. o-Phthaldialdehyde (Sigma) was used as the pre-column derivation reagent and detection was performed by a UV detector (model LC-9A; Shimadzu, Kyoto, Japan) at 338 nm, with excitation at 262 nm. Typically, Glu, Gln and theanine standards (Sigma) were eluted after 3.2, 7.5 and 13.6 min, respectively. Analysis of puried theanine The puried theanine was identied by a Waters Synapt QTOF mass spectrometer (Waters Corporation, Manchester, UK). The instrument was operated in positive ion mode with the electro spray ionisation (ESI) capillary set to 3.5 kV. The nitrogen desolvation gas was set to 600 L h1 , the source block temperature to 100 C, the desolvation temperature to 250 C, and the cone voltage to 30 V.

MATERIALS AND METHODS

Strain isolation and identication Shrimp paste (1 g; Lee Kum Kee ) from a local market was transferred into a 250 mL Erlenmeyer ask containing 30 mL medium A containing 20 g glucose, 15 g yeast extract, 10 g corn steep liquor, 0.5 g MgSO4 7H2 O and 1 g K2 HPO4 3H2 O L1 at pH 7.2. The inoculated medium was incubated at 37 C in a rotary shaker at 200 rpm. After 30 h, a loop of the culture was spread over separate plates of solid medium B containing 10 g glucose, 5 g yeast extract, 5 g glutamate and 0.5 g KH2 PO4 L1 at pH 7.2, and incubated at 37 C. After 48 h, fast-growing colonies were picked and incubated in medium A again. Then the culture broth was centrifuged at 10 000 g and 4 C for 30 min to remove insoluble materials. The supernatant was tested for GGT activity. The strains that produced GGT with high activity were taken for a second screening to determine the theanine-forming capability of their GGTs according to the enzymatic synthesis described below. The strain producing GGT with high theanine-forming capability was selected for further research and then identied. The morphological and biochemical tests of the isolated strain were conducted using Biolog Microstation System (Biolog Inc., Hayward, CA, USA). Genomic DNA was extracted with a Bacteria Genomic DNA Isolation Kit (Shanghai Chaoshi Bio Technologies Co. Ltd, Shanghai, China). Liquor (13 L) of the supernatant was used as the template DNA for PCR amplication. The 16S rDNA gene was amplied with bacterial universal primers 27F (5 -GAGTTTGATCCTGGCTCAG3 ) and 1527R (5 -AGAAAGGAGGTGATCCAGCC-3 ), which were used for sequencing (China Center for Type Culture Collection, Wuhan, China). Homology searches were performed against the sequences of database using the BLAST program (NCBI, Bethesda, MD, USA). Enzyme production For the production of GGT with a high activity, a loop of the isolated strain was inoculated into 30 mL optimised medium containing 25 g sucrose, 5 g tryptone, 15 g corn steep liquor, 0.5 g MgSO4 7H2 O and 1 g K2 HPO4 3H2 O L1 at pH 7.2 in 250 mL ask and incubated at 37 C in a rotary shaker at 200 rpm for 20 h. The culture broth was centrifuged at 10 000 g and 4 C for 30 min to remove insoluble materials. The supernatant was used as the enzyme. Assay of -glutamyltranspeptidase activity The enzyme assay was performed using the method of Orlowski and Meister with slight modications.23 The assay mixture (1 mL) was composed of 50 mmol L1 borateNaOH (pH 10), 5 mmol L1 L- -glutamyl-p-nitroanilide ( -GpNA; Sigma, St Louis, MO, USA), 20 mmol L1 glycylglycine and the enzyme solution. After incubation at 37 C for 30 min, the reaction was terminated by the addition of 0.1 mol L1 HCl. The transferase activity

RESULTS AND DISCUSSION

Screening and identication of strains that produce -glutamyltranspeptidase About 20 GGT-producing strains were isolated from fermented shrimp paste, among which four strains produced GGT with transferring activity (above 2.0 U mL1 ) relatively close to the highest GGT activity (3.2 U mL1 ) exhibited by B. subtilis NX-2.21 Some characteristics of GGT produced by the four strains were determined and compared (Table 1). Results showed that different GGTs have dissimilar specicity toward the same substrate Gln. Obviously, using GGT from strain SK11.004 as a catalyst, the highest production of theanine was 16 mmol L1 with a conversion yield of 80%. The amount of theanine formed was much greater than that produced by GGT from E. coli K-12,9 which has been reported to be efcient for theanine production. It is worth noting that the ratio of transferring activity to hydrolytic activity of Gln was found to be 16.8 and this ratio is much higher than those observed for glutamyl transferring enzymes from other microorganisms.11,20,24

2564

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 25632567

Enzymatic production of theanine by B. subtilis GGT

20

www.soci.org
5 4 3 10 2 5 1 0 0.02 0.04 0.06 0.08 0.10 GGT concentration (U mL-1) Glu (mmol L-1) Glu (mmol L-1)

Table 1. Characteristics of the GGTs produced by four isolated strains of B. subtilis Strain SK11.001 SK11.002 SK11.003 SK11.004
a

Theanine:Glub 5.29 0.16 8.00 0.21 6.42 0.33 16.8 0.26

2.33 0.05 2.01 0.04 2.15 0.03 2.30 0.04

56.3 0.14 62.5 0.16 64.2 0.1 80.0 0.2

The conversion rate was calculated taking into account the amount of substrate (glutamine) converted into product (theanine). b Ratio of theanine to glutamic acid (Glu). Data were expressed as mean SD from three independent experiments.

Theanine (mmol L-1)

GGT activity (U mL1 )

Conversion ratioa (%)

15

20

15 Theanine (mM)

Figure 2. Effect of enzyme concentration on theanine synthesis. The reaction was carried out with 20 mmol L1 Gln and 50 mmol L1 ethylamine HCl, at pH 10 and 37 C for 5 h, the concentration of GGT being varied as indicated in the gure. Theanine (shaded columns) and Glu (unlled columns) were analysed by HPLC. Data were expressed as mean SD from three independent experiments.

10 2 5 1 0 0.01 0.02 0.03 0.04 0.05 0.06 0.07 Ethylamine HCI concentration (mol L-1) 0

Glu (mM)

20

5 4 3

Theanine (mmol L-1)

15

10 2 5 1 0 9.0 9.5 10.0 10.5 pH values 11.0

Figure 1. Effect of ethylamine HCl concentration on theanine synthesis. The reaction was carried out with 20 mmol L1 Gln and 0.04 U mL1 GGT at pH 10 and 37 C for 5 h, the concentration of ethylamine HCl being varied as indicated in the gure. Theanine (shaded columns) and Glu (unlled columns) were analysed by HPLC. Data were expressed as mean SD from three independent experiments.

These ndings suggested a potential application of strain SK11.004 in theanine production. Strain SK11.004 was a rod-shaped, endospore-forming and Gram-positive bacterium, which grows well in aerobic culture. Its 16S rDNA sequence (1432 bp) showed 99% of similarity with Bacillus subtilis 79-9 16S rDNA gene (EU624 322.1) and has been deposited in GenBank under the accession number FJ437210. It was identied as B. subtilis based on its physiological and biochemical properties as well as its 16S rDNA sequence analysis. Enzymatic synthesis of theanine with -glutamyltranspeptidase To increase the yield of theanine, the reaction conditions for its synthesis were investigated. When the concentration of donor (Gln) was xed at 20 mmol L1 , the optimal concentration of acceptor (ethylamine HCl) was found to be 50 mmol L1 (Fig. 1). Moreover, Gln was not detected in the reaction product, indicating that Gln has been used up as a -glutamyl donor. Similarly, the optimum concentration of GGT was 0.06 U mL1 (Fig. 2). When higher concentrations of GGT were used, the amounts of theanine and Glu decreased after 5 h. This may be due to the fact that GGT catalyses the auto-transpeptidation reaction in which Gln itself serves as an acceptor of the -glutamyl group when the

Figure 3. Effect of pH on theanine synthesis. The reaction was carried out with 20 mmol L1 Gln, 50 mmol L1 ethylamine HCl, and 0.06 U mL1 GGT at 37 C for 5 h, the pH of reaction mixture being varied as indicated in the gure. Theanine (shaded columns) and Glu (unlled columns) were analysed by HPLC. Data were expressed as mean SD from three independent experiments.

concentration of GGT is increased. The effect of pH was also studied (Fig. 3). After 5 h of incubation, the best conversion yield was obtained at pH 10. The content of Glu increased with decreasing pH from 10 to 7. Results revealed that at pH values lower than 10, the hydrolytic reaction is preferred. However, when the pH is above 10, the auto-transpeptidation reaction occurs. These facts indicated that if the pH of reaction mixture is well adjusted, GGT can selectively catalyse the transpeptidation reaction. The optimum incubation time was 4 h (Fig. 4). The optimum conditions for the synthesis of theanine were determined to be 20 mmol L1 Gln, 50 mmol L1 ethylamine HCl, 0.06 U mL1 GGT, pH 10, and incubation at 37 C for 4 h. The yield was 18.9 mmol L1 and the conversion rate of Gln to theanine reached 94% (Fig. 5). Theanine production was achieved in a shorter time with a small amount of enzyme and by-product. However, the yield was higher than those obtained using GGT from Pseudomonas taetrolens Y-30,8 Methylovorus mays No. 925 and Pseudomonas citronellosis GEA.26

2565

J Sci Food Agric 2010; 90: 25632567

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org
20 5 4 3 10 2 5 1 0 1 2 3 4 5 Incubation time (h) 6 Glu (mmol L-1)
200 Response value (mAu) 100 0 0 B Glu Theanine 5 10 15 400 300 A

Y Shuai et al.

Theanine (mmol L-1)

15

20

100 80 60 40 20

Gln

Figure 4. Effect of incubation time on theanine synthesis. The reaction was carried out with 20 mmol L1 Gln, 50 mmol L1 ethylamine HCl, and 0.06 U mL1 GGT at 37 C and pH 10. Theanine (shaded columns) and Glu (unlled columns) were analysed by HPLC. Data were expressed as mean SD from three independent experiments.

10 Time (min)

15

20

Figure 5. HPLC of the enzymatic synthesis of theanine by GGT under optimised conditions: (A) product of enzymatic synthesis; (B) Glu, Gln and theanine standards.

Purication and identication of theanine After theanine had been synthesised, the enzymatic reaction mixture (20 mL) was applied to a column (20 mL) of 201 7 strong basic resin (Cl form) to absorb Glu, theanine being eluted with Milli Q water. The fractions containing theanine were then applied to a column (20 mL) of 001 7 strong acidic resin (H+ form). Theanine was eluted with ammonia solution, and the fractions containing only theanine were collected and lyophilised after removing ammonia under vacuum at 60 C. The puried theanine was subjected to positive-mode ESI-MS analysis. The molecular mass of theanine was determined to be 174 Da (Fig. 6). As shown in Fig. 6, a strong signal with an m/z value of 175, which corresponds to the ionised theanine in positive mode, was observed. The protonated molecule at m/z 175 loses the ammonia residue to yield a fragment at m/z 158, which in turn

loses the ethylamine group to generate an ion at m/z 130. These results indicated that theanine had been successfully synthesised and puried.

CONCLUSIONS
In this study, Bacillus subtilis SK11.004, which produces GGT with high theanine-forming activity, was isolated from fermented shrimp paste. The goal of developing an effective method for the production of theanine was achieved using the GGT from B. subtilis SK11.004. These ndings are of considerable importance in the feasibility of industrial production of this functional food additive and pharmaceutical intermediate to meet

2566

Figure 6. Q-TOF mass spectra of the [M+H]+ ion of the puried theanine.

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 25632567

Enzymatic production of theanine by B. subtilis GGT the increasing needs of humans. Moreover, compared with others, the GGT from B. subtilis SK11.004 exhibited the highest ratio of transferring activity to hydrolytic activity of Gln (the donor of -glutamyl moiety), suggesting a potential application in the production of other functional -glutamyl compounds. However, further investigation is required to extensively characterise the GGT produced by the newly isolated strain.

www.soci.org
10 Tomita K, Ito M, Yano T, Kumagai H and Tochikura T, Glutamyltranspeptidase activity and the properties of the extracellular glutaminase from Aspergillus oryzae. Agric Biol Chem 52:11591163 (1988). 11 Suzuki H, Kumagai H and Tochikura T, -Glutamyltranspeptidase from Escherichia coli K-12: purication and properties. J Bacteriol 168:13251331 (1986). 12 Boanca G, Sand A and Barycki JJ, Uncoupling the enzymatic and autoprocessing activities of Helicobacter pylori glutamyltranspeptidase. J Biol Chem 281:1902919037 (2006). 13 Nakayama R, Kumagai H and Tochikura T, Purication and properties of -glutamyltranspeptidase from Proteus mirabilis. J Bacteriol 160:341346 (1984). 14 Mehdi K, Thierie J and Penninckx MJ, -Glutamyltranspeptidase in the yeast Saccharomyces cerevisiae and its role in the vacuolar transport and metabolism of glutathione. Biochem J 359:631637 (2001). 15 Chu L, Xu X, Dong Z, Cappelli D and Ebersole JL, Role for recombinant -glutamyltransferase from Treponema denticola in glutathione metabolism. Infect Immun 71:335342 (2003). 16 Hwang SY, Ryang JH, Lim WJ, Yoo ID and Oishi K, Purication and properties of -glutamyl transpeptidase from Bacillus sp. KUN-17. J Microb Biotechnol 6:238244 (1996). 17 Moallic C, Dabonne S, Colas B and Sine JP, Identication and characterization of a -glutamyltranspeptidase from a thermoalcalophile strain of Bacillus pumilus. Protein J 25:391397 (2006). 18 Minami H, Suzuki H and Kumagai H, Salt-tolerant glutamyltranspeptidase from Bacillus subtilis 168 with glutaminase activity. Enzyme Microb Technol 32:431438 (2003). 19 Kimura K, Tran LS, Uchida I and Itoh Y, Characterization of Bacillus subtilis -glutamyltransferase and its involvement in the degradation of capsule poly-gamma-glutamate. Microbiology 150:41154123 (2004). 20 Ogawa Y, Hosoyama H, Hamano M and Motai H, Purication and properties of -glutamyltranspeptidase from Bacillussubtilis (natto). Agric Biol Chem 55:29712977 (1991). 21 Wu Q, Xu H, Zhang L, Yao J and Ouyang P, Production, purication and properties of -glutamyltranspeptidase from a newly isolated Bacillus subtilis NX-2. J Mol Catal B: Enzym 43:113117 (2006). 22 Phromraksa P, Nagano H, Kanamaru Y, Izumi H, Yamada C and Khamboonruang C, Characterization of Bacillussubtilis isolated from Asian fermented foods, Food Sci Technol Res 15:659666 (2009). 23 Orlowski M and Meister A, -Glutamyl-p-nitroanilide: a new convenient substrate for determination and study of Land D- -glutamyltranspeptidase activities. Biochim Biophys Acta 73:679681 (1963). 24 Nakayama R, Kumagai H and Tochikura T, Purication and properties of -glutamyltranspeptidase from Proteus mirabilis. J Bacteriol 160:341346 (1984). 25 Sachiko Y, Mamoru W and Takashi T, Characterization of theanineforming enzyme from Methylovorus mays No. 9 in respect to utilization of theanine production. Biosci Biotechnol Biochem 71:545552 (2007). 26 Tachiki T, Okada Y, Ozeki M, Okubo T, Juneja L and Yamazaki N, Process for producing theanine. JP Patent 2002229026 (2002).

ACKNOWLEDGEMENTS
This research was nancially supported by the National Natural Science Foundation of China (No. 20376029), and the Research Program of State Key Laboratory of Food Science and Technology, Jiangnan University (No. SKLF-MB-200804 and SKLF-TS-200805).

REFERENCES
1 Lu K, Gray MA, Oliver C, Liley DT, Harrison BJ, Bartholomeusz CF, et al., The acute effects of L-theanine in comparison with alprazolam on anticipatory anxiety in humans. Hum Psychopharmacol Clin Exp 19:457465 (2004). 2 Yokogoshi H, Kato Y, Sagesaka YM, Takihara-Matsuura T, Kakuda T and Takeuchi N, Reduction effect of theanine on blood pressure and brain 5-hydroxyindoles in spontaneously hypertensive rats. Biosci Biotechnol Biochem 59:615618 (1995). 3 Kakuda T, Nozawa A, Unno T, Okamura N and Okai O, Inhibiting effects of theanine on caffeine stimulation evaluated by EEG in the rat. Biosci Biotechnol Biochem 64:287293 (2000). 4 Sadzuka Y, Sugiyama T and Sonobe T, Improvement of idarubicin induced antitumor activity and bone marrow suppression by theanine, a component of tea. Cancer Lett 158:119124 (2000). 5 Egashira N, Hayakawa K, Mishima K, Kimura H, Iwasaki K and Fujiwara M, Neuroprotective effect of -glutamylethylamide (theanine) on cerebral infarction in mice. Neurosci Lett 363:5861 (2004). 6 Zheng G, Sayama K, Okubo T, Juneja LR and Oguni I, Anti-obesity effects of three major components of green tea, catechins, caffeine and theanine, in mice. In Vivo 18:5562 (2004). 7 Juneja LR, Chu DC, Okubo T, Nagato Y and Yokogoshi H, Ltheanine a unique amino acid of green tea and its relaxation effect in humans. Trends Food Sci Technol 10:199204 (1999). 8 Yamamoto S, Wakayama M and Tachiki T, Cloning and expression of Pseudomonas taetrolens Y-30 gene encoding glutamine synthetase: an enzyme available for theanine production by coupled fermentation with energy transfer. Biosci Biotechnol Biochem 70:500507 (2006). 9 Suzuki H, Izuka S, Miyakawa N and Kumagai H, Enzymatic production of theanine, an umami component of tea, from glutamine and ethylamine with bacterial -glutamyltranspeptidase. Enzyme Microb Technol 31:884889 (2002).

2567
wileyonlinelibrary.com/jsfa

J Sci Food Agric 2010; 90: 25632567

c 2010 Society of Chemical Industry

Research Article
Received: 4 February 2010 Revised: 7 July 2010 Accepted: 9 July 2010 Published online in Wiley Online Library: 17 August 2010

(wileyonlinelibrary.com) DOI 10.1002/jsfa.4122

Freshness characterisation of whiting (Merlangius merlangus) using an SPME/GC/MS method and a statistical multivariate approach
Guillaume Duos,a Francois Leduc,a Assi NGuessan,b Frederic Krzewinski,c c Ossarath Kol and Pierre Mallea
Abstract
BACKGROUND: The freshness of whiting was studied at ve stages of ice storage by comparing the analysis of volatile compounds obtained through solid phase microextraction/gas chromatography/mass spectrometry (SPME/GC/MS) with two sensory methods. RESULTS: Of the volatile compounds identied, 38 were analysed using a statistical multivariate approach and classied according to their role in the estimation of freshness during storage as markers of freshness or spoilage. Regarding the evolution of the presence or absence of individual compounds, three categories were dened. For example, the volatile compounds propanal, hexanal, 1-penten-3-ol, pentanal, 2,3-pentanedione, 1-penten-3-one, heptanal, (E)-2-pentenal, 2,3-octanedione, (Z)-2-penten-1-ol, 1-pentanol, butanal, octanal, 3,5,5-trimethyl-2-hexene, 1-hexanol and 4,4-dimethyl-1,3-dioxane appeared highly relevant, because they are found throughout storage and can be divided into several categories that are directly related to the quality of sh. CONCLUSION: SPME/GC/MS combined with a statistical multivariate approach may be a useful method to identify volatile compounds and characterise sh freshness during storage. c 2010 Society of Chemical Industry Keywords: sh freshness; SPME; GC/MS; principal component analysis; volatile compounds

INTRODUCTION
Fish is a fragile product. Its quality is a health and commercial issue that concerns industrialists, ofcial inspection authorities and consumers. Fish freshness is therefore a key attribute of its quality. Odour is one of the main indicators that consumers use to assess the freshness of sh. The smell of sh changes rapidly according to the products degree of freshness, and this is why sensory analyses are used by consumers and industrialists to assess sh quality. Since odour is one of the main parameters used to determine the sensory quality of products, the key volatile compounds that contribute to this characteristic odour can be measured and used as quality indicators.1 3 These volatile aromatic compounds that characterise smell are generated by the action of bacterial and tissue enzymes and lipid autoxidation. The microbial production of volatile compounds is caused by specic spoilage bacteria.4 These include mainly Gram-negative psychrotrophs such as Pseudomonas spp. and Alteromonas spp.5 and more specically Shewanella putrefaciens and Photobacterium phosphoreum.6,7 These bacteria generate odours by producing alcohols, carbonyls, esters and sulfur compounds8 10 corresponding to different kinds of odour such as at and neutral odours, and then rancid, putrid odours or odours of ammonia or sulfur changing the avour and taste of sh during storage.11 13 On the other hand, odours of fresh sh are mainly associated with compounds

that contain carbonyls and alcohols and are characterised by light, green, delicate, melon, marine plant and iodised notes.11,13 In view of the role that volatile compounds play in the assessment of sh freshness, studies have been undertaken to identify and quantify these compounds in various sh (mackerel, herring, cod, sardine and sea bream) using either static headspace analysis (SHA), dynamic headspace analysis (DHA) or an electronic nose in various storage phases.3,4,14 18 Another technique is solid phase microextraction (SPME). SPME is a relatively new method that can be used to extract volatiles from foods. It is a rapid and simple extraction technique that does not require the use of solvents. SPME was rst used to assess sh

Correspondence to: Guillaume Duos, ANSES, Laboratoire des Produits de la P che, Bld Bassin Napolon, F-62200 Boulogne-sur-Mer, France. e e E-mail: guillaume.duos@anses.fr

a ANSES, Laboratoire des Produits de la P che, Bld Bassin Napolon, F-62200 e e Boulogne-sur-Mer, France b Laboratoire Paul Painlev, UMR CNRS 8524 & Ecole Polytechnique Universitaire e de Lille, Universit de Lille 1, Avenue Paul Langevin, F-59655 Villeneuve dAscq e Cedex, France c Unite de Glycobiologie Structurale et Fonctionnelle, UMR CNRS 8576, Universit e de Lille 1, F-59655 Villeneuve dAscq Cedex, France

2568

J Sci Food Agric 2010; 90: 25682575

www.soci.org

c 2010 Society of Chemical Industry

Freshness characterisation of whiting quality by Bene et al.19,20 when studying the volatile substances in salmon and whiting. Other researchers have used the technique on craysh, rainbow trout, salmon, mackerel, sardine, tuna, sea bream, cod and shrimp to study their volatile compounds during storage.16,21 26 SHA, DHA and SPME are often combined with gas chromatography/mass spectrometry (GC/MS) to determine the quality and quantity of various volatile aromatic compounds that can be relevant quality indicators. The aims of this study were to examine the various volatile compounds in whiting during storage using the SPME technique combined with GC/MS, to ascertain if there is a relationship between these compounds and sh quality as assessed through a sensory analysis method and to determine which of these compounds may be markers of sh freshness or spoilage using a statistical multivariate approach.

www.soci.org The second method is the quality index method (QIM) evaluation system adapted to whiting. It is based on changes in the sensory characteristics of raw sh during spoilage. Scores of 01, 02 or 03 demerit (or index) points are attributed according to changes observed in the smell, texture, eye appearance, skin and gills. The points are added up to obtain an overall sensory score or quality index (QI).

EXPERIMENTAL
Chemicals Carboxen (CAR)/polydimethylsiloxane (PDMS) StableFlex bre (65 m) came from Supelco (Bellefonte, PA, USA). Before rst use, each SPME bre was conditioned as recommended by the manufacturer. NaCl came from Oxoid Ltd (Basingstoke, UK), MilliQ water (high-performance liquid chromatographic water) from Fisher Scientic Labosi (Elancourt, France) and 3-methyl-3-buten1-ol from Sigma-Aldrich (Saint Quentin Fallavier, France). Sample preparation Whiting (Merlangius merlangus), caught the night before the start of the study, was acquired from Cooperative Maritime Etaploise (Boulogne-sur-Mer, France). Two different catches (20 and 15 sh respectively) were analysed. The sh were stored in crushed ice at 4 C in self-draining polystyrene boxes for 7 days. Fresh crushed ice was added daily. Sensory evaluation and volatile analysis were performed on days 1, 2, 3, 4 and 7 (on seven different sh each day). Sensory evaluation was performed by two panellists familiar with the sensory evaluation of sh. Following sensory evaluation, each sh was lleted. The next step was according to previous work.27 The llets were cut into 1 cm cubes, then 50 g of esh was introduced into a stomacher bag with 100 mL of ultrapure water saturated with NaCl. The contents were mixed for 2 min in a Stomacher Lab-Blender 400 (Seward, Thetford, UK). The aqueous phase was removed and centrifuged at 12 000 g for 10 min at 4 C (Multifuge 3 S-R Heraeus, Kendro Laboratory Products, Courtaboeuf, France). Sensory evaluation Two methods were used for the sensory evaluation of sh. The rst method is the European Unions grading system presented in European Directive 2406/96. This system distinguishes between three freshness categories, E, A and B, corresponding to various levels of spoilage. Category E corresponds to the highest quality level, followed by categories A and B, while sh graded below B is considered to be non-edible. Another category corresponds to the products discard level. In order to rate this evaluation, these letters have been replaced with numbers: 0 = E, 1 = A, 2 = B and 3 = unacceptable. The lower the number, the fresher the sh; conversely, the higher the number, the greater the spoilage (below 18 the sh is acceptable).

SPME procedure According to previous studies,27 11 mL of aqueous phase from the supernatant of sample preparation was introduced into a hermetically sealed 20 mL vial. The vial was placed in the sample tray of a Combi PAL (CTC Analytics, Zwingen, Switzerland) and then transferred to the mixer, where it was heated at 50 C and mixed at 500 rpm for 10 min. After this equilibrium time the CAR/PDMS SPME bre was inserted into the headspace of the sample and held there for 40 min at 50 C. The bre was then removed from the headspace and inserted into the Merlin Microseal injector (250 C) of a GC-17A gas chromatograph (GC) equipped with an MS-QP5000 mass spectrometer (MS) (Shimadzu, Kyoto, Japan) for desorption. The bre was maintained 10 seconds.

GC/MS procedure The GC was equipped with a BPX5 capillary column (60 m 0.25 mm 0.25 m) (SGE, Courtaboeuf, France). The GC conditions were as follows: oven temperature set initially at 35 C (5 min hold), increased to 100 C at 10 C min1 , then increased to 280 C at 20 C min1 and maintained at 280 C for 5 min; the splitless mode was used for injection, with a purge time of 2 min. The bre was maintained in the injection port for 10 s. The electron impact MS conditions were as follows: temperature of interface, 260 C; ionisation voltage, 70 eV; mass range, m/z 33200; scan speed, 250 per 0.5 s. After each injection the bre was heated to 300 C for 10 min in the SPME bre conditioner. Volatile compounds were identied by matching their mass spectra with Mass Spectral Libraries 21 and 107 of the National Institute of Standards and Technology (NIST) (developed for Shimadzu by NIST, July 2002). Semi-quantication of the components was based on arbitrary units of total current ion peak area counts.

Statistical analysis The statistical analysis of the 35 samples was divided into two parts. First we described and analysed the freshness and spoilage indices on the various sampling days using PASW Statistics 18 (SPSS, Paris, France). Two graphic approaches were used to that end. One approach, using box plots, shows a statistical summary of the indices by sampling day. Another graphic approach based on control charts of the individual index values was used for this rst part. These control charts were used to analyse the daily progression of freshness and/or spoilage indices. The second part of our statistical analysis proposes a grouping of analysed sh samples and identies the volatile compounds that characterise their categories. To do so, we used SPAD 7 (SPAD, Paris, France) to select a set of relevant volatile compounds via the rst two trends in a principal component analysis (PCA). We then performed an ascending hierarchical classication (AHC) of the rst two factors. The categories were then analysed in relation to the selected volatile compounds and the freshness and/or spoilage indices.

2569

J Sci Food Agric 2010; 90: 25682575

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org

G Duos et al.

Table 1. Volatile compounds identied during spoilage analysis of whiting Variable name a b c d e f g h i j k l m o p q r s t v w x y z aa ab ac ae Compound name Acetaldehyde Methanethiol Trimethylamine Ethanol Pentane Propanal Dimethyl sulde Methylene chloride Carbon disulde 2,3-Butanedione Butanal 2-Butanone 2-Butanol Ethyl acetate Acetic acid, ethyl ester 2-Methyl-1-propanol 3-Methyl butanal 1-Butanol 2-Methyl butanal 1-Penten-3-ol 1-Penten-3-one Heptane 2-Ethyl furan 3-Pentanone 2,3-Pentanedione Pentanal 3-Pentanol 3-Hydroxy-2-butanone Rention time (min) 3.099 3.400 3.426 3.527 3.699 3.795 4.019 4.195 4.253 5.009 5.075 5.110 5.370 5.442 5.639 5.793 6.525 6.695 6.725 7.197 7.298 7.453 7.577 7.598 7.610 7.643 7.892 8.503 Variable name ag ah aj ak al am an ao ap aq ar as at au av aw ax ay az ba bb bc bd be bf bg bh Compound name 3-Methyl-1-butanol 2-Methyl-1-butanol (E)-2-Pentenal 1-Pentanol (Z)-2-Penten-1-ol 2-Hexanone Hexanal 4,4-Dimethyl-1,3-dioxane (E)-2-Hexenal 1-Hexanol 3-Heptanone (Z)-4-Heptenal Heptanal Cyclohexanone 1-Heptanol 3,5,5-Trimethyl-2-hexene Benzaldehyde 1-Octen-3-ol 2,3-Octanedione 2,4-Heptadienal Octanal (E,E)-2,4-Heptadienal 2-Ethyl-1-hexanol Limonene 1-Octanol 3,5-Octadien-2-one Nonanal Rention time (min) 8.790 8.801 9.331 9.565 9.640 10.154 10.484 11.484 11.882 12.122 12.497 12.817 12.858 12.968 14.018 14.106 14.142 14.192 14.248 14.534 14.567 14.769 14.818 14.856 15.052 15.459 15.852

Figure 1. Comparison of freshness indices according to sampling day.

Figure 2. Comparison of QIM scores according to sampling day.

RESULTS AND DISCUSSION

Analysis of freshness or spoilage indices by sampling day The study consists of 35 measurements taken during ve periods, D1 (day 1), D2 (day 2), D3 (day 3), D4 (day 4) and D7 (day 7), at a rate of seven measurements per period and some 60 volatile compounds observed per sample (Table 1). Figures 1 and 2 show the respective freshness indices and QIM scores associated with freshness and spoilage of the analysed whiting catches.

They illustrate both the average trend and the variability of the two indices (freshness and QIM) by sampling day. We note that these two freshness and/or spoilage indicators tend to rise over sampling time. The daily variability of the two indicators is illustrated by the height of each box plot. This variation is greatest on D3 and D4 for the freshness index and on D3 and D7 for the QIM score. Values outside the boxes (outliers) testify to the atypical character of some indicator values (abnormally high or low). Two sh have

2570

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 25682575

Freshness characterisation of whiting

www.soci.org

Figure 3. Chart of individual freshness indices: upper and lower control limits.

, mean; - - - - ,

Figure 4. Chart of individual QIM scores: lower control limits.

, mean; - - - - , upper and

freshness indices that are relatively far from the average on D1 and D3, while one sh has a QIM score that is relatively far from the average on D1. We note for example that the mean value of the freshness index increases from 1.393 on D1 to 14.000 on D7 with high variability on D3 (standard deviation of 1.1603). As for the QIM score, it varies on average from 2.976 on D1 to 18.333 on D7 with fairly high variability on D3 (standard deviation of 2.0189). Concerning the freshness index, the estimate of means and standard deviations clearly shows a relatively signicant difference between D2 and D3 on the one hand and between D4 and D7 on the other hand. It thus appears that D3 represent a breaking point between high and low quality of the analysed sh samples freshness. Freshness therefore decreases between D2 and D3. Then there is a transition phase from D3 to D4, leading to a conrmed spoilage zone starting on D4. This grouped analysis of freshness indices is combined with a descriptive analysis of each of the two indicators individual values in Figs 3 and 4. These two graphs explain the upward trend observed above and specify the quite linear nature of this trend in relation to the sampling days. They also highlight a transition zone delimited by the line located between the two broken lines (upper and lower control limits) and marked by a rst change in values of the two indicators between D2 and D3 and a second change between D4 and D7. These jumps appear clearly evident for the QIM indicator. Statistical tests for small samples were applied and gave a signicance of around 103 (Fig. 4): the mean value on D2 is signicantly different from that on D3 with an error level of 5% for freshness index and QIM score. D4 presents the same difference as D7 for the two indicators. Classication of sh using PCA The results of the PCA show that the rst two components, a linear combination of volatile compounds, account for about 65% (Tables 2 and 3) of the total information contained in the data

table. The rst principal component with an eigenvalue of 28.477 (Table 2, column VP) accounts for 54.75% (projected inertia) of the 65%, while the second principal component with an eigenvalue of 5.535 (Table 3, column VP) accounts for only 10.64%. The circle of correlation (Fig. 5: projection of variables on the plane formed by the rst two principal components) clearly illustrates two major volatile compound trends. The rst/second trend involves a group of volatile compounds that are positively projected onto the rst/second principal component. From these two trends we extracted the most characteristic volatile compounds (Tables 2 and 3) in order of their importance in each of the two principal components. The COOR column corresponds to the coordinate of the projection of the volatile compound on each factor, while the columns QLT and CRT respectively represent the quality of the representation and the relative contribution of this compound to the formation of the principal trend. We note therefore that 30 volatile compounds are largely responsible for the formation of axis 1 (Fig. 5). These 30 volatile compounds account for more than 80% of the 54.75% of inertia projected on the rst component and range from 1-pentanol (ak) to (E)-2-hexenal (ap) and methylene chloride (h) and from 2-ethyl1-hexanol (bd) to octanal (bb), heptanal (at) and 1-octen-3-ol (ay). As for the second principal component, it largely comprises eight volatile compounds, ethanol (d), 3-methyl-1-butanol (ag), 2,3-butanedione (j), 2-methyl-1-butanol (ah), 3-methyl butanal (r), 2-methyl-1-propanol (q), limonene (be) and ethyl acetate (o), which make up nearly 63% of the 10.64% of inertia projected on this second component. Classication description and interpretation After the PCA an AHC was conducted using SPAD 7 and the rst two principal components to dene categories. Table 4 presents a distribution into three categories and gives the composition of each of them as well as the sh with proles that are closest to the categorys centre of gravity (distance to the centre of the categorys column).

2571

J Sci Food Agric 2010; 90: 25682575

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org

G Duos et al.

Table 2. Variables of interpretation from rst principal component Variable name ak ap ba bg av aj al as w aw bb at an bc ab ao az ay v am ax aa k bh aq b f x bd h % total contribution COOR 0.981 0.976 0.969 0.969 0.964 0.964 0.963 0.963 0.956 0.949 0.946 0.945 0.940 0.937 0.935 0.933 0.931 0.928 0.927 0.927 0.924 0.905 0.901 0.895 0.891 0.879 0.870 0.854 0.839 0.827 QLT 0.963 0.952 0.939 0.938 0.930 0.929 0.928 0.928 0.914 0.900 0.894 0.893 0.883 0.878 0.874 0.870 0.866 0.861 0.859 0.859 0.854 0.819 0.811 0.802 0.794 0.772 0.757 0.729 0.703 0.684 VP 28.477 28.477 28.477 28.477 28.477 28.477 28.477 28.477 28.477 28.477 28.477 28.477 28.477 28.477 28.477 28.477 28.477 28.477 28.477 28.477 28.477 28.477 28.477 28.477 28.477 28.477 28.477 28.477 28.477 28.477 CRT 0.034 0.033 0.033 0.033 0.033 0.033 0.033 0.033 0.032 0.032 0.031 0.031 0.031 0.031 0.031 0.031 0.030 0.030 0.030 0.030 0.030 0.029 0.028 0.028 0.028 0.027 0.027 0.026 0.025 0.024 0.905

Figure 5. Circle of correlations in principal plane.

Table 3. Variables of interpretation from second principal component Variable name d ag j ah r q be o % total contribution COOR 0.779 0.769 0.747 0.741 0.695 0.644 0.635 0.634 QLT 0.607 0.591 0.559 0.549 0.483 0.415 0.403 0.402 VP 5.535 5.535 5.535 5.535 5.535 5.535 5.535 5.535 CRT 0.110 0.107 0.101 0.099 0.087 0.075 0.073 0.073 0.724

Category 1/3 consists of 17 sh from the two catches of whiting with samples taken on D1, D2 and D3. Category 2/3 consists of 7 sh from the two catches with samples taken on D3 and D4. Category 3/3 consists of 11 sh from the two catches. It includes some of the samples taken on D4 and all samples taken on D7. Figure 6 illustrates the simultaneous projection of the centres of the three categories and the sh. The plan of projection comprising

the rst two principal components represents more than 75% of the total information. The diameter of each category (each class) allows us to assess the quality of representation and importance of the category in the projection plan. The large the diameter, the more the category is well represented: better quality for category 1/3 than for categories 3/3 and 2/3. The analysis shows that the centre of category 1/3 projects negatively onto the rst two principal components. The squared cosine of the angle formed by the category 1/3 centre of gravity and the rst principal component is 0.794 with a signicance of about 104 . As a result, if we use an error level of 5% (ve out of 100 chances that the conclusion will be false), the correlation between this category and the rst principal component is statistically signicant. The results also show that category 1/3 is statistically signicantly correlated with the second principal component with a correlation of 0.191. Similarly, category 2/3 is signicantly positively correlated with the second principal component with a correlation of 0.517 and a virtually null signicance. Category 3/3 is signicantly positively correlated with the rst principal component with a correlation of 0.991 and a virtually null signicance. The diagrams in Figs 7 and 8 make it possible to quantify the degree of freshness and/or spoilage of each of the three sh categories. Overall, category 1/3 has the lowest values of freshness index and QIM score. The freshness quality of the sh in this category is excellent. In category 2/3 the freshness indices and QIM scores are higher than in category 1/3 and overlap category 3/3, with signicant dispersion within this category only for the freshness index. It appears to be an intermediate category between freshness and spoilage. Category 3/3 shows higher freshness index values with less dispersion. This category is synonymous with spoilage (poor freshness quality of sh). To characterise each of the three sh categories with volatile compounds, we compared the compounds mean category value with its mean value calculated from all analysed sh. The selected compounds were then compared with those from the PCA. We therefore propose three typologies of volatile compounds: compounds present in the category on average and highly signicant, those present and not highly signicant and those completely and signicantly absent on average (Table 4). In category 3/3, which consists of sh at an advanced stage of spoilage (Figs 7 and 8), the volatile compounds whose average presence is signicant are those that contribute strongly to the

2572

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 25682575

Freshness characterisation of whiting

www.soci.org

Table 4. Characterisation of categories by volatile compounds Distance to the centre of the category 5.932 Highly signicant average presence f, an, v, ab, aa, w, h, at, aj, az, al, ak, ay, aq, b, aw, as, bb, k, av, ap, ao, bc, ax, bd, bh, ba, x, bg, am Slightly signicant average presence r Highly signicant average absence

Category 3/3

Name L01P03D07

6.780 10.452 13.765 18.874 22.325 23.574 39.567 39.972 43.094 56.905 2/3 9.989 12.390 20.353 38.469 49.812 52.347 57.513 0.680

L01P02D07 L02P01D04 L01P01D07 L02P01D07 L02P02D04 L01P04D04 L02P02D07 L02P03D04 L02P03D07 L01P04D07 L02P03D03 L01P03D04 L01P02D04 L01P02D03 L01P01D03 L02P02D03 L01P01D04 L02P01D01 d, ag, r, ah, q, j, o, be w, ao am

1/3

f, an, v, ab, aa, w, at, aj, ay, ak, az, al, k, bb, aw, aq, ao, b, av, ap, bd, bh, bc, ax, bg d, q, o, j

h, as, x, ba, am,

1.532 2.169 2.366 3.336 3.376 3.427 3.547 4.516 4.550 6.350 7.754 10.033 10.781 12.540 14.050 20.665

L01P03D02 L01P04D01 L02P01D02 L01P02D01 L01P04D02 L01P02D02 L02P01D03 L01P03D01 L02P03D01 L02P02D01 L01P04D03 L02P02D02 L01P01D02 L01P01D01 L01P01D03 L02P03D02

formation of the rst trend in the PCA (Table 2). The majority of these compounds have a category mean that is signicantly higher than their general mean (mean calculated using all samples of analysed sh). The most characteristic compounds on average are propanal (f), hexanal (an), 1-penten-3-ol (v), pentanal (ab), 2,3-pentanedione (aa), 1-penten-3-one (w), methylene chloride (h), heptanal (at), (E)-2-pentenal (aj), 2,3-octanedione (az), (Z)-2penten-1-ol (al), 1-pentanol (ak), 1-octen-3-ol (ay), 1-hexanol (aq), methanethiol (b), 3,5,5-trimethyl-2-hexene (aw), (Z)-4-heptenal (as), octanal (bb), butanal (k), 1-heptanol (av), (E)-2-hexenal (ap), 4,4-dimethyl-1,3-dioxane (ao), (E,E)-2,4-heptadienal (bc), benzaldehyde (ax), 2-ethyl-1-hexanol (bd), nonanal (bh), 2,4heptadienal (ba), heptane (x), 3,5-octadien-2-one (bg) and 2hexanone (am). We also noted the average presence of a volatile

compound that contributes to the second principal component: 3-methyl butanal (r). This compound has a mean in the category that is lower than its general mean, but the difference between the two means is not statistically signicant. Category 2/3, comprising seven sh, can be qualied as an intermediate category because it contains neither sh with an excellent state of freshness nor sh with signicant spoilage (sh with a poor state of freshness). This transitional category between freshness and spoilage is mainly characterised by the compounds ethanol (d), 3-methyl-1-butanol (ag), 2,3-butanedione (j), 2-methyl-1-butanol (ah), 3-methyl butanal (r), 2-methyl-1propanol (q), limonene (be) and ethyl acetate (o) whose mean in the category is signicantly higher than their general mean. These volatile compounds are also those involved in the second

2573

J Sci Food Agric 2010; 90: 25682575

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org

G Duos et al.

Figure 6. Projections of categories and sh on rst two factors.

Figure 7. Characterisation of categories by freshness indices.

compounds that characterise this second category, these three volatile compounds have a mean value in category 2/3 that is slightly lower or even null compared with their general mean. Category 1/3 consists of sh with an excellent state of freshness. The volatile compounds that characterise it generally have a category mean that is signicantly smaller than their general mean. These compounds (d, q, o, j) are linked to both the rst and second components in the PCA. The most prominent volatile compounds in category 1/3 can be divided into two groups: a group of compounds (f, an, v, ab, aa, w, at, aj, ay, ak, az, al, k, bb, aw, aq, ao, d, q, o, j, b, av, ap, bd, bh, bc, ax, bg) whose mean in the category is non-null but signicantly smaller than the mean in the sample (very signicant average presence) and a group of compounds (h, as, x, ba, am) whose mean in the category is totally null (very signicant average absence). This latter group of compounds can also be found in category 3/3 but with the completely opposite behaviour. The sh L02P01D01 has the prole that is closest to the centre of the category. The compounds that characterise these two axes (Fig. 6) may be compounds that represent sh freshness and spoilage. To sum up the three categories, the rst category 1/3 corresponds to a category with an excellent freshness index, which characterises excellent product quality. For category 2/3, this quality decreases and the whiting starts to deteriorate. For category 3/3, which is the worst category from a freshness/quality viewpoint, spoilage of whiting continues but the product remains acceptable for consumption (freshness index <18, Fig. 3). We may conclude that axis 1 represents the freshness of whiting and axis 2 characterises spoilage.

CONCLUSIONS
Figure 8. Characterisation of categories by QIM scores.

main trend determined by the PCA. This category 2/3 also contains volatile compounds related to the formation of the PCAs rst principal component: 1-penten-3-one (w), 4,4-dimethyl-1,3dioxane (ao) and 2-hexanone (am). These latter three volatile compounds contribute to the formation of the PCAs rst principal component and the characterisation of category 3/3. Unlike the

The volatile compounds propanal (f), hexanal (an), 1-penten3-ol (v), pentanal (ab), 2,3-pentanedione (aa), 1-penten-3-one (w), heptanal (at), (E)-2-pentenal (aj), 2,3-octanedione (az), (Z)2-penten-1-ol (al), 1-octen-3-ol (ay), 1-pentanol (ak), butanal (k), octanal (bb), 3,5,5-trimethyl-2-hexene (aw), 1-hexanol (aq), 4,4-dimethyl-1,3-dioxane (ao), methanethiol (b), 1-heptanol (av), (E)-2-hexenal (ap), 2-ethyl-1-hexanol (bd), nonanal (bh), (E,E)-2,4heptadienal (bc), benzaldehyde (ax) and 3,5-octadien-2-one (bg) are found in both category 1/3 (category with excellent freshness quality; negative test values) and category 3/3 (considered a

2574

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 25682575

Freshness characterisation of whiting spoilage category; positive test values). Thus they are relevant compounds because they are found throughout the spoilage process but with a difference in quantity, which shows an alteration in the whitings quality. These volatile compounds may be considered representative markers of whiting and markers of future freshness. The compounds methylene chloride (h), (Z)-4-heptenal (as), heptane (x), 2,4-heptadienal (ba) and 2-hexanone (am) are signicantly absent from category 1/3 and highly present in category 3/3. They are therefore relevant in the sense that they can be considered spoilage markers. A very fresh product will not contain these compounds; conversely, they will be found in a deteriorated product. Other compounds present in category 2/3 could be representative of spoilage in whiting and can be considered to be spoilage markers: ethanol (d), 3-methyl-1-butanol (ag), 3-methyl butanal (r), 2-methyl-1-butanol (ah), 2-methyl-1-propanol (q), 2,3butanedione (j), ethyl acetate (o) and limonene (be). 2-Ethyl furan (y) is worth exploring for use as a freshness marker, which does not characterise axes 1 and 2 but is only signicant for category 1/3. This method is well adapted for recovery of the volatile compounds, because our results present a large number of compounds described in previous studies:3,4,15,17,23,24,26 compounds characterising freshness, i.e. hexanal (an), heptanal (at), 1-penten-3-ol (v), (Z)-2-penten-1-ol (al), 1-octen-3-ol (ay), octanal (bb), (E)-2-pentenal (aj) and (E)-2-hexenal, or spoilage, i.e. 3-methyl-1-butanol (ag), ethanol (d), 2-methyl-1-propanol (q), 3-methyl butanal (r), ethyl acetate (o), 2,3-butanedione (j) and 3-hydroxy-2-butanone (ae). SPME/GC/MS may be a useful method at the same time to identify the volatile compounds then to characterise freshness. Complementary studies could allow applications to other species.

www.soci.org
redsh (Sebastes marinus) stored in ice and modied atmosphere bulk storage. J Aquat Food Prod Technol 11:229249 (2002). Miller III A, Scanlan RA, Lee JS and Libbey LM, Identication of the volatile compounds produced in sterile sh muscle (Sebastes melanops) by Pseudomonas fragi. J Appl Microbiol 25:952955 (1973). Miller III A, Scanlan RA, Lee JS, Libbey LM and Morgan ME, Volatile compounds produced in sterile sh muscle (Sebastes melanops) by Pseudomonas perolens. J Appl Microbiol 25:257261 (1973). Lindsay RC, Josephson DB and Olafsdottir G, Chemical and biochemical indices for assessing the quality of sh packaged in controlled atmospheres, in Seafood Quality Determination, Proceedings of an International Symposium (University of Alaska Sea Grant Program, Anchorage, Alaska), ed. by Kramer DE and Liston J. Elsevier Science, Amsterdam, pp. 221234 (1986). Josephson DB and Lindsay RC, Enzymic generation of volatile aroma compounds from fresh sh. ACS Symp Ser 317:201219 (1986). Durnford E and Shahidi F, Flavour of sh meat, in Flavor of Meat, Meat Products and Seafoods, ed. by Shahidi F. Springer - Verlag, Berlin, pp. 131158 (1998). Kawai T, Fish avor. Crit Rev Food Sci 36:257298 (1996). Alasalvar C, Quantick PC and Grigor JM, Aroma compounds of fresh and stored mackerel (Scomber scombrus). ACS Symp Ser 674:3954 (1997). Alasalvar C, Taylor KDA and Shahidi F, Comparison of volatiles of cultured and wild sea bream (Sparus aurata) during storage in ice by dynamic headspace analysis/gas chromatographymass spectrometry. J Agric Food Chem 53:26162622 (2005). Jonsdottir R, Bragadottir M and Olafsdottir G, The role of volatile compounds in odor development during hemoglobin-mediated oxidation of cod muscle membrane lipids. J Aquat Food Prod Technol 16:6786 (2007). Prost C, Hallier A, Cardinal M, Serot T and Courcoux P, Effect of storage time on raw sardine (Sardina pilchardus) avor and aroma quality. J Food Sci 69:198204 (2004). Duos G, Cornu M, Coin VM, Antinelli JF and Malle P, Determination of volatile compounds to characterize sh spoilage using HS/MS and SPMEGC/MS analysis. J Agric Food Chem 86:600611 (2006). Bene A, Fornage A, Luisier J, Pichler P and Villettaz J, A new method for the rapid determination of volatile substances: the SPME-direct method. Part I: Apparatus and working conditions. Sensors Actuators B 72:184187 (2001). Bene A, Hayman A, Reynard E, Luisier J and Villettaz J, A new method for the rapid determination of volatile substances: the SPME-direct method. Part II: Determination of the freshness of sh. Sensors Actuators B 72:204207 (2001). Baek HH and Cadwallader KR, Volatile compounds in avor concentrates produced from craysh-processing by-products with and without protease treatement. J Agric Food Chem 44:32623267 (1996). Guill n MD and Errecalde MC, Volatile components of raw and e smoked black bream (Brama raii) and rainbow trout (Oncorhynchus mykiss) studied by means of solid phase microextraction and gas chromatography/mass spectrometry. J Sci Food Agric 82:945952 (2002). Mansur MA, Bhadra A, Takamura H and Matoba T, Volatile avor compounds of some sea sh and prawn species. Fish Sci 69:864866 (2003). Wierda RL, Fletcher G, Xu L and Dufour J-P, Analysis of volatile compounds as spoilage indicators in fresh king salmon (Oncorhynchus tshawytscha) during storage using SPME-GCMS. J Agric Food Chem 54:84808490 (2006). Edirisinghe RKB, Graffham AJ and Taylor SJ, Characterisation of the volatiles of yellown tuna (Thunnus albacares) during storage by solid phase microextraction and GCMS and their relationship to sh quality parameters. Int J Food Sci Technol 42:11391147 (2007). Iglesias J and Medina I, Solid-phase microextraction method for the determination of volatile compounds associated to oxidation of sh muscle. J Chromatogr A 1192(1):916 (2008). Duos G, Coin VM, Moine F and Malle P, Determination of volatile compounds in whiting using SPME GCMS. J Chromatogr Sci 43:304312 (2005).

9 10

11 12 13 14 15

16

17 18 19

ACKNOWLEDGEMENTS
We thank the Nord-Pas de Calais region (France) and France Agrimer for nancial assistance.
20

REFERENCES
1 Olafsdottir G and Fleurence J, Evaluation of sh freshness using volatile compounds classication of volatile compounds in sh, in Methods to Determine the Freshness of Fish in Research and Industry, Proceedings of the Final Meeting of the Concerted Action Evaluation of Fish Freshness AIR3CT94 2283, ed by Olafsdottir G. International institute of refrigeration, Paris, France, pp. 5169 (1997). 2 Alasalvar C, Aishima T and Quantick PC, Dynamic headspace analysis of volatile aroma products in fresh and deteriorated mackerel (Scomber scombrus). Food Sci Technol Int 1:125127 (1995). 3 Aro T, Tahovonen R, Koskinen L and Kallio H, Volatile compounds of Baltic herring analysed by dynamic headspace samplinggas chromatographymass spectrometry. Eur Food Res Technol 216:483488 (2003). 4 Olafsdottir G, Jonsdottir R, Lauzon HL, Luten J and Kristbergsson K, Characterization of volatile compounds in chilled cod (Gadus morhua) llets by gas chromatography and detection of quality indicators by an electronic nose. J Agric Food Chem 53:1014010147 (2005). 5 Shewan JM, The bacteriology of fresh and spoiling sh and some related chemical changes. Recent Adv Food Sci 1:167193 (1962). 6 Van Spreekens KJA, Characterization of some sh and shrimp spoiling bacteria. Antonie van Leeuwenhoek: Int J Gen Mol Microbiol 43:283303 (1977). 7 Olafsdottir G, Li X, Lauzon HL and Jonsdottir R, Precision and application of electronic nose for freshness monitoring of whole

21

22

23 24

25

26 27

2575
wileyonlinelibrary.com/jsfa

J Sci Food Agric 2010; 90: 25682575

c 2010 Society of Chemical Industry

Research Article
Received: 23 July 2009 Revised: 8 July 2010 Accepted: 12 July 2010 Published online in Wiley Online Library: 5 August 2010

(wileyonlinelibrary.com) DOI 10.1002/jsfa.4123

Release of phenolic acids from defatted rice bran by subcritical water treatment
Cynthia Fabian, Ngoc Yen Tran- Thi, Novy S Kasim and Yi-Hsu Ju
Abstract
BACKGROUND: Oil production from rice bran, an undervalued by-product of rice milling, produces defatted rice bran (DRB) as a waste material. Although it is considered a less valuable product, DRB still contains useful substances such as phenolic compounds with antioxidant, UV-B-protecting and anti-tumour activities. In this study the phenolic acids in DRB were extracted with subcritical water at temperatures of 125, 150, 175 and 200 C. RESULTS: Analysis of total phenolics using FolinCiocalteu reagent showed about 220 g gallic acid equivalent kg1 bran in the extracts. High-performance liquid chromatography analysis showed low contents of phenolic acids (about 0.42 g kg1 bran). Ferulic, p-coumaric, gallic and caffeic acids were the major phenolic acids identied in the extracts. Thermal analysis of the phenolic acids was also done. The thermogravimetric curves showed that p-coumaric, caffeic and ferulic acids started to decompose at about 170 C, while gallic acid did not start to decompose until about 200 C. CONCLUSION: Subcritical water can be used to hydrolyse rice bran and release phenolic compounds, but the high temperatures used in the extraction can also cause the decomposition of phenolic acids. c 2010 Society of Chemical Industry Supporting information may be found in the online version of this article. Keywords: antioxidant; phenolic acids; rice bran; subcritical water; thermogravimetry

INTRODUCTION
Lipid oxidation decreases the quality and nutritional value of foods.1,2 To remedy this, antioxidants that can prevent or retard the oxidation of fats and oils are introduced. Increasing evidence has shown that intake of antioxidants can also lower the risk of degenerative diseases such as cancer, cardiovascular diseases, diabetes, arthritis, cataract formation and aging.3 5 Interest in antioxidants is being directed towards the identication and extraction of natural compounds because of the growing concern about the safety of synthetic ingredients contained in various foods.6 Cereals, including rice (Oryza sativa L.), contain a wide range of phenolic compounds and are claimed to be a good source of natural antioxidants.4,7 10 Some of these compounds are predominantly found in grains and are not present signicantly in fruits and vegetables.11 These phenolic acids are known to contribute to the antioxidant potential of cereal grains.12 15 Zhou et al.16 found high levels of phenolic acids (ferulic and p-coumaric acids) in brown rice but lower levels in milled rice. This indicates that phenolic acids are concentrated in rice bran. However, there are few studies reporting the proles of phenolic acids in rice bran and defatted rice bran (DRB). Rice bran is a by-product of the rice-milling industry and is generally used as an animal feed.7,17 19 In some countries, oil is extracted from rice bran for food application, which produces DRB as a waste material. Renuka Devi and Arumughan4 studied the antioxidant efcacy of DRB extract and its phytochemical constituents, which include ferulic acid. Since phenolic acids are polar compounds, most are retained in the bran after extraction

of lipids. They can be utilised if extracted and separated from the DRB. Extraction of phenolic compounds from DRB would facilitate the production of value-added products. Also, the extraction of bioactive components from biodiesel residue provides options to lower the production cost of biodiesel.20 The utilisation of DRB, a by-product of biodiesel production if rice bran is used as the raw material, has the potential to reduce the cost of biodiesel production. In this study, subcritical water was employed to extract phenolic compounds from DRB. Subcritical water extraction, a new, environmentally friendly technique, has been considered as an alternative for the isolation of antioxidant constituents.2,21,22 Since only water is employed for extraction, the process avoids the use and disposal of large volumes of ammable or toxic solvents. Subcritical water is achieved at temperatures between 100 and 374 C (the critical point of water is at 374 C and 22 MPa) and at a pressure high enough to keep the water in a liquid state.23 At subcritical condition the ionic product of water increases such that the water becomes a rich source of H+ and OH ions suitable

Correspondence to: Yi-Hsu Ju, National Taiwan University of Science and Technology, 43 Keelung Road Sec. 4, Taipei 106-07, Taiwan. E-mail: yhju@mail.ntust.edu.tw Part of this paper was presented at the Conference of Food Engineering, Columbus, Ohio, USA, 58 April 2009. National Taiwan University of Science and Technology, 43 Keelung Road Sec. 4, Taipei 106-07, Taiwan

2576

J Sci Food Agric 2010; 90: 25762581

www.soci.org

c 2010 Society of Chemical Industry

Phenolic acids from defatted rice bran for hydrolysis reactions.24 Also, the dielectric constant of water decreases as it approaches the critical point, so a wide range of organic compounds can be solubilised and extracted.25 Previous studies by Wiboonsirikul et al.,18,19 Hata et al.26 and Sereewatthanawut et al.27 on subcritical extraction from rice bran focused on proteins, carbohydrates and antioxidant capacity of the subcritical extracts. The identication of antioxidants such as phenolic acids in the rice bran extracts was not performed. In the present study, high-performance liquid chromatography (HPLC) was utilised to identify phenolic acids in the extracts. Since high temperatures were employed during extraction, thermogravimetric analyses of the identied phenolic acids were also carried out to determine their decomposition temperatures.

www.soci.org and separated to obtain the methanolic extract. The extract was then passed through a 0.45 m lter and analysed by HPLC. The HPLC analysis of phenolic acids employed in this study was based on the method of Zhou et al.16 with modication. A 20 L aliquot of the methanolic extract was separated using a Jasco HPLC system. Peaks were detected with a Jasco multiwavelength UVvisible detector operated at 280 and 325 nm. The absorbance at 325 nm is optimal for all hydroxycinnamic acids, while 280 nm is the most accepted compromise for all phenols, including hydroxybenzoic and hydroxycinnamic acids. Separations were achieved on a 5 m Phenomenex Luna C18 column (240 mm 4.6 mm; Phenomenex, Torrance, CA, USA). Gradient elution was performed using solvent A (water/acetic acid, 100 : 1 v/v) and solvent B (methanol/acetonitrile/acetic acid, 95 : 5:1 v/v/v) in the following sequence: 02 min, 5% B; 210 min, 525% B; 1020 min, 2540% B; 2030 min, 4050% B; 3040 min, 50100% B; 4045 min, 100% B; 4555 min, 1005% B. The solvent ow rate was xed at 1 mL min1 . Peak identities were conrmed from retention data and by spiking extracts with authentic standards. HPLC data were processed using ChromPass Version 1.8.6.1 (Jasco, Tokyo, Japan). Quantication was accomplished via external standard calibration curves plotted using pure standards of the four phenolic acids (Supporting information). Regression analysis of each calibration curve was also carried out to obtain a correlation coefcient between peak area and standard concentration. These correlation coefcients were later used in the quantication of phenolic acid content in samples. Thermogravimetric analysis A Perkin Elmer Diamond TG/DTA instrument (PerkinElmer, Shelton, CT, USA) was used for thermal stability studies. Approximately 6 mg of phenolic acid was placed on a platinum pan. The sample was then heated from 30 to 950 C at 10 C min1 to determine the temperature at which decomposition occurred. During the entire run, air at atmospheric pressure was allowed to ow at 20 mL min1 through the system containing the sample.

MATERIALS AND METHODS

Materials Rice bran fresh from milling was purchased from a local rice mill in Taoyuan County (Taiwan). The bran is not specically from one variety of rice but is a mixture of rice harvested in northern Taiwan. Bran collected from the mill was stored at 60 C before use. Defatting of rice bran was done using hexane in a Soxhlet extractor at 60 C for 4 h. Phenolic acid standards (gallic (monohydrate), caffeic, pcoumaric and ferulic acids) and the FolinCiocalteu (FC) reagent used for total phenolic analysis were purchased from Sigma Aldrich (St Louis, MO, USA). HPLC-grade methanol and acetonitrile were used in the analyses. Other chemicals used were of analytical grade. Extraction with subcritical water DRB was mixed with distilled water to a content of about 400 g L1 solids. About 70 mL of this mixture was charged into a 90 mL stainless steel batch reactor (Jia Chan Company, Tainan, Taiwan) equipped with a thermocouple and a pressure gauge. The pressure was increased to 20 bar using nitrogen gas. The reactor was then heated to the desired temperature (125, 150, 175 or 200 C), which took about 30 min, and held at that temperature for 5 min. The reactor was then cooled to room temperature and its content was centrifuged. The supernatant was collected, freeze-dried and analysed for phenolic acid content. All extraction experiments and subsequent analyses were carried out at least twice. Analysis of phenolic acids The supernatant was rst analysed for total phenolic compounds extracted using the methods of Wiboonsirikul et al.18 About 100 L of the diluted subcritical water extract (10 dilution) was combined with 400 L of FC reagent and 1 mL of 75 g L1 sodium carbonate solution. Distilled water was added to the mixture to a total volume of 5 mL and the solution was kept in the dark at ambient temperature for 2 h to complete the reaction. Total phenolic content was determined by measuring the absorbance at 765 nm using a Jasco UV-V 550 UVvisible spectrophotometer (Jasco, Easton, MD, USA). Gallic acid was used as a standard, and results were calculated as g gallic acid equivalent kg1 bran. Individual phenolic acids in the DRB extract were identied using an HPLC UVvisible detector. Since many compounds were present in the subcritical extract, phenolic acids were separated by solvent extraction from the freeze-dried supernatant. About 0.1 g of the freeze-dried supernatant was mixed with 4 mL of 800 mL L1 methanol at 30 C for 2 h. The mixture was centrifuged

RESULTS AND DISCUSSION

Total phenolics in subcritical extracts from DRB The phenolic acids in DRB were extracted by subcritical water and analysed using two different methods, i.e. FC reagent and HPLC. As can be seen in Table 1, the FC method showed an increase in the amount of phenolic compounds in the subcritical extract with increasing temperature. At 125 C the amount of phenolics was about 2 g kg1 bran, but this increased signicantly to about 20 g kg1 bran at 200 C. This trend is similar to the ndings of Wiboonsirikul et al.,18 with differences in specic values being attributable to the different sources of rice bran. The increase in phenolic content of the extract at high temperatures is probably due to the release of phenolic substances, interfering substances or both by subcritical water treatment. At subcritical condition, hydrolysis of rice bran is favoured, because the ionic product of water increases such that there are more H+ and OH ions available to catalyse the hydrolysis.25 Hydrolysis of the bran resulted in an increase in released phenolic acids, because hydroxycinnamic acids such as ferulic, sinapic, caffeic and p-coumaric acids are found not only as soluble forms in the cytoplasm but also covalently attached to the plant cell wall.28 Quantication of the four major phenolic acids found in the rice bran extract (gallic, caffeic, ferulic and p-coumaric acids) by

2577

J Sci Food Agric 2010; 90: 25762581

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org

C Fabian et al.

Table 1. Phenolic content of subcritical water extracts of defatted rice bran Temperature ( C) 125 150 175 200 FolinCiocalteu method (g gallic acid equivalent kg1 bran) 2.011 0.056 3.552 0.046 7.833 0.178 19.480 0.148 HPLC analysis (g total phenolic acids kg1 bran) 0.528 0.006 1.794 0.045 2.125 0.077 0.380 0.029

the HPLC method (based on the study by Zhou showed a different trend of phenolic acid content in rice bran extract from that determined by the FC method. The total amount of phenolic acids found was about 0.5 g kg1 bran at 125 C, increased to about 2 g kg1 bran at 175 C and then fell sharply to about 0.4 g kg1 bran at 200 C. Zhou et al.16 extracted about 0.5 and 0.105 g kg1 phenolic acids from brown and milled rice respectively. Comparing these three sets of data, there is a noticeable tendency that the phenolic acid content (about 0.42 g kg1 bran) in rice bran matrix is higher than those in brown and milled rice matrices. Thus further study needs to be carried out to investigate the likelihood of the phenolic acid distribution being in the following order of magnitude: phenolic acid content in rice bran > phenolic acid content in brown rice > phenolic acid content in milled rice. The comparative analyses of all major phenolic compounds using FC reagent and HPLC were also studied, since the two methods gave different trends of phenolic acid content in rice bran extract. Emphasis was placed on the applicability and reliability of FC reagent to analyse the phenolic acid content in rice bran. Data on rice phenolic content reported by Zhou et al.,16 Bunzel et al.,29 Hudson et al.30 and Kuroda et al.31 suggest that most of the phenolic compounds in rice bran are phenolic acids. A signicant difference in the quantication results of the two methods was observed, which may be attributed to interfering compounds during the analysis using FC reagent. According to Yu,32 the use of FC reagent for the spectrophotometric analysis of phenolic compounds has the disadvantages of low specicity for phenols and interference of other reducing agents that can react with the reagent. Wiboonsirikul et al.18 also mentioned that the subcritical extract contains carbohydrates, proteins and amino acids, which are considered as interfering substances for FC analysis. These compounds can reduce the reagent. FC analysis will give poor reliability in terms of accuracy and data reproduction in analysing the phenolic acid content in rice bran; therefore this analysis may be inappropriate for the determination of phenolic acid content in rice bran matrix. It was also noted that at 175 and 200 C there was a pronounced brown discolouration of the extract, and phenolic acids were not the major components in the methanolic extract. A high-intensity peak with a retention time of about 10 min was observed at these temperatures (Fig. 1). The brown pigment in the extract is a polymeric material resulting from oxidative polymerisation of phenols as well as possible formation of Maillard products. According to Somoza,33 the Maillard reaction is a series of reactions between proteins and carbohydrates during heating. The results of early-stage Maillard reactions are Amadori products. When higher temperatures are applied for longer times, advanced brown-pigmented Maillard reaction products termed melanoidins are formed. Adams et al.34 stated that other food constituents were incorporated in the

et al.16 )

3 4 5

c b a 2 4 6 8 10 12 14 16 18 20 22 24 26 28 30 Retention time (min)

Figure 1. HPLC chromatograms of methanolic extract from freeze-dried subcritical water extract of defatted rice bran at (a) 125, (b) 150, (c) 175 and (d) 200 C recorded by UV detection at 280 nm. Peak numbers: 1, gallic acid; 2, unknown compound; 3, caffeic acid; 4, p-coumaric acid; 5, ferulic acid.

melanoidin structure as well, such as lipid oxidation products in tomato melanoidins and phenolic compounds in coffee melanoidins. Wiboonsirikul et al.18 conrmed the occurrence of an advanced Maillard reaction when rice bran was subjected to subcritical water at a temperature above 150 C. From Fig. 1 it can be seen that, when the temperature was raised to 175 C, the intensities of peak 4 (p-coumaric acid) and peak 5 (ferulic acid) decreased while the intensity of peak 2 increased sharply. It is highly unlikely that the latter peak is due to a melanoidin, since melanoidins are heterogeneous polymers and will appear in an HPLC chromatogram as a broad region of intensity. From the fact that the intensities of peaks 4 and 5 decreased while the intensity of peak 2 increased, it is likely that peak 2 represents a product derived from p-coumaric and ferulic acids. Further investigations are needed to identify peak 2. The chromatogram of the subcritical extract at 175 C was compared with that of the subcritical extract at 175 C treated with 2 mol L1 NaOH for 4 h at 25 C.35 It was observed that the area of peak 2 decreased by 94% while the area of the ferulic acid peak increased by about 700% (data not shown). This may be a result of the breakdown of ester linkages of ferulic acid by alkaline hydrolysis.35 Phenolic acid proles From the HPLC method based on that of Zhou et al.,16 four phenolic acids, i.e. gallic, caffeic, p-coumaric and ferulic acids, were detected in all extracts obtained at different subcritical extraction temperatures. Gallic acid was observed in the rice bran extracts. During milling, some rice hull ends up in the rice bran fraction. Rice hull is abundant in gallic acid because of the presence of tannin,

2578

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 25762581

Phenolic acids from defatted rice bran which upon hydrolysis yields gallic acid.36 Hydroxycinnamic acids such as ferulic, p-coumaric and caffeic acids were also found in the extracts, since these phenolic acids are constituents of the cell wall and cytoplasm of plant cells.28 Figure 2 presents the relation between the amount of each phenolic acid extracted and the temperature of extraction. Figure 2(a) shows that, when the temperature was increased from 150 to 175 C, gallic acid was the rice bran phenolic acid to increase most rapidly in the extract. When the temperature was further increased to 200 C, gallic acid was the one that decreased most rapidly (see also Table 1). This is because gallic acid is the most abundant of all phenolic acids in rice bran extract. Moreover, thermogravimetric analysis of the four phenolic acids (Figs 35) indicated that ferulic acid was the most heat-labile phenolic acid in rice bran, and it lost its weight because of its thermal decomposition at a lower temperature. During subcritical extraction, phenolic acids are released from the bran. This is probably due to several reasons, such as hydrolysis of the rice bran cell wall, a change in solvent viscosity allowing permeability through the bran, and the behaviour of water at subcritical condition being similar to that of alcohol in terms of polarity. Since a mixture of phenolic acids is present in the bran, the decreasing polarity of water allows the extraction of more phenolic acids as the temperature is increased. At 175 C the extracted caffeic and p-coumaric acids have already decreased, while the extracted ferulic and gallic acids decrease only at 200 C. The decrease in extracted phenolic acids with increasing temperature of subcritical treatment is attributable to two possible phenomena, i.e. thermal degradation of these phenolic acids and their involvement in the Maillard reaction. Thermal stability studies of the four phenolic acids were carried out. For the thermogravimetric analysis the weight loss of each phenolic acid as

www.soci.org a function of temperature is presented in Figs 35. The derivative of the weight loss curve is also shown in the gures to indicate the temperature at which weight loss is apparent. Figure 3 presents the thermogravimetric curve for gallic acid monohydrate. The rst abrupt decrease in weight of the sample was observed at 91.7 C, which corresponds to the release of water attached to the acid. The second (207308 C), third (308365 C) and fourth (376576 C) events in the thermogravimetric curve correspond to the thermal degradation of gallic acid. Major weight loss was observed at 207308 C. In the thermal analysis of gallic acid, Oridorga et al.37 noted that the peak at about 260 C corresponds to the formation of decarboxylation products from gallic acid. Heating of gallic acid at this temperature caused the evolution of gaseous products and the formation of pyrogallol. At about 308 C, which corresponds to the boiling point of pyrogallol, further weight loss of the sample occurred. Lastly, heating above 308 C resulted in the subsequent decomposition of pyrogallol. From Fig. 3 it can be seen that the decomposition of gallic acid started at about 207 C. The subcritical extraction at 200 C also showed a decrease in the amount of gallic acid in the extract (Fig. 2). The decrease in gallic acid in the extract at 200 C may be attributed to the conversion of gallic acid into its decarboxylation products such as pyrogallol and other gaseous products. Similar thermogravimetric curves for caffeic and p-coumaric acids were obtained (Fig. 4). Four events of thermal decomposition occurred, with peaks at 221, 300, 332 and 539 C for caffeic acid and at 229, 360, 394 and 547 C for p-coumaric acid. Thermal decomposition started at about 172 C for these two phenolic acids, and the amount extracted decreased at 175 C (Fig. 2). In the thermal stability analysis of ferulic acid it was observed that two decomposition steps occurred (Fig. 5). The major decomposition of ferulic acid started at 173 C and peaked at

Figure 2. Concentrations of phenolic acids in subcritical water extract of defatted rice bran at different temperatures: (a) gallic acid; (b) caffeic acid; (c) p-coumaric acid; (d) ferulic acid.

2579

J Sci Food Agric 2010; 90: 25762581

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org

C Fabian et al.

Figure 3. Thermogravimetric curves of gallic acid: - - - - , derivative of weight loss.

, weight loss;

Figure 5. Thermogravimetric curves of ferulic acid: - - - - , derivative of weight loss.

, weight loss;

acid identied in rice bran cell wall, where it may be bound to arabinoxylans or crosslinked to cell wall components.28 At 175 C, ferulic acid was extracted by subcritical water, but the free ferulic acid started to decompose, resulting in a lower yield. The amount of ferulic acid extracted decreased at 200 C.

CONCLUSIONS
Subcritical water can be used to hydrolyse rice bran and release phenolic compounds. Phenolic acids at a level of about 2 g kg1 rice bran can be obtained at 175 C. At higher temperatures, phenolic acids may be damaged and/or Maillard reaction products may be formed that may incorporate the phenolic acids. The temperature used in the extraction must be carefully chosen so as to release the maximum amount of phenolics while simultaneously minimising their destruction.

ACKNOWLEDGEMENTS
Financial support from Taiwans Ministry of Education, Amia Enterprise Co., Ltd and National Taiwan University of Science and Technology through project 94DI062 is greatly appreciated. The authors thank Dr Fred Quarnstrom for his assistance during the preparation of this manuscript. Supporting information Supporting information may be found in the online version of this article.

Figure 4. Thermogravimetric curves of (a) caffeic acid and (b) p-coumaric , weight loss; - - - - , derivative of weight loss. acid:

245 C. This result agrees with those of Taniguchi et al.,38 who stated that ferulic acid starts to decompose at about 176 C, and Fiddler et al.,39 who reported that the peak decomposition of ferulic acid occurs at 245 C. Fiddler et al.39 stated that the rst step in the decomposition of ferulic acid corresponds to its decarboxylation, while the second step comprises different reactions occurring simultaneously. It is impossible to propose a single mechanism. Although the decomposition of ferulic acid started at 173 C, in the subcritical extract, ferulic acid kept increasing up to 175 C (Fig. 2). This can be expected, since ferulic acid is the major phenolic

REFERENCES
1 Cosgrove J, Church D and Pryor W, The kinetics of the autoxidation of polyunsaturated fatty acids. Lipids 22:299304 (1987). 2 Ibanez E, Kubatova A, Senorans FJ, Cavero S, Reglero G and Hawthorne SB, Subcritical water extraction of antioxidant compounds from rosemary plants. J Agric Food Chem 51:375382 (2002). 3 Landrault N, Poucheret P, Ravel P, Gasc F, Cros G and Teissedre P-L, Antioxidant capacities and phenolics levels of French wines from different varieties and vintages. J Agric Food Chem 49:33413348 (2001). 4 Renuka Devi R and Arumughan C, Phytochemical characterization of defatted rice bran and optimization of a process for their extraction and enrichment. Bioresour Technol 98:30373043 (2007).

2580

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 25762581

Phenolic acids from defatted rice bran

5 Wu X, Gu L, Holden J, Haytowitz DB, Gebhardt SE, Beecher G, et al, Development of a database for total antioxidant capacity in foods: a preliminary study. J Food Compos Anal 17:407422 (2004). 6 Prior RL, Phytochemicals: mechanism of action, in Absorption and Metabolism of Anthocyanins: Potential Health Effects, ed. by Meskin M, Bidlack WR, Devies AJ, Lewis DS and Randolph RK. CRC Press, Boca Raton, FL, pp. 19 (2004). 7 Iqbal S, Bhanger MI and Anwar F, Antioxidant properties and components of some commercially available varieties of rice bran in Pakistan. Food Chem 93:265272 (2005). 8 Krings U, El-Saharty YS, El-Zeany BA, Pabel B and Berger RG, Antioxidant activity of extracts from roasted wheat germ. Food Chem 71:9195 (2000). 9 Renuka Devi R, Jayalekshmy A and Arumughan C, Antioxidant efcacy of phytochemical extracts from defatted rice bran in in-vitro model emulsions. Int J Food Sci Technol 43:878885 (2008). 10 Shin ZI, Chang YS, Kang WS and Jung SU, Antioxidant extracted from the defatted rice bran. US Patent 5175 (1992). 11 Wojdyo A and Oszmainski J, Comparison of the content of phenolic acid, -tocopherol and their antioxidant activity in oat and naked and weeded. ElectronJEnvironAgricFoodChem 6:19801988 (2007). 12 Adom KK, Sorrells ME and Liu RH, Phytochemicals and antioxidant activity of milled fractions of different wheat varieties. J Agric Food Chem 53:22972306 (2005). 13 Maillard M-N and Berset C, Evolution of antioxidant activity during kilning: role of insoluble bound phenolic acids of barley and malt. J Agric Food Chem 43:17891793 (2002). 14 Goupy P, Hugues M, Boivin P and Amiot MJ, Antioxidant composition and activity of barley (Hordeum vulgare) and malt extracts and of isolated phenolic compounds. J Sci Food Agric 79:16251634 (1999). 15 Pussayanawin V, Wetzel DL and Fulcher RG, Fluorescence detection and measurement of ferulic acid in wheat milling fractions by microscopy and HPLC. J Agric Food Chem 36:515520 (2002). 16 Zhou Z, Robards K, Helliwell S and Blanchard C, The distribution of phenolic acids in rice. Food Chem 87:401406 (2004). 17 Juliano BO, Rice: Chemistry and Technology. American Association of Cereal Chemists, St Paul, MN (1985). 18 Wiboonsirikul J, Kimura Y, Kadota M, Morita H, Tsuno T and Adachi S, Properties of extracts from defatted rice bran by its subcritical water treatment. J Agric Food Chem 55:87598765 (2007). 19 Wiboonsirikul J, Kimura Y, Kanaya Y, Tsuno T and Adachi S, Production and characterization of functional substances from a by-product of rice bran oil and protein production by a compressed hot water treatment. Biosci Biotechnol Biochem 72:384392 (2008). 20 Zullaikah S, Lai C-C, Vali SR and Ju Y-H, A two-step acid-catalyzed process for the production of biodiesel from rice bran oil. Bioresour Technol 96:18891896 (2005). 21 Rodrguez-Meizoso I, Marin FR, Herrero M, Senorans FJ, Reglero G, Cifuentes A, et al, Subcritical water extraction of nutraceuticals

www.soci.org
with antioxidant activity from oregano. Chemical and functional characterization. J Pharmaceut Biomed Anal 41:15601565 (2006). Ju Z and Howard LR, Subcritical water and sulfured water extraction of anthocyanins and other phenolics from dried red grape skin. J Food Sci 70:S270S276 (2005). Ramos L, Kristenson EM and Brinkman UAT, Current use of pressurised liquid extraction and subcritical water extraction in environmental analysis. J Chromatogr 975:329 (2002). Patrick HR, Grifth K, Liotta CL, Eckert CA and Glaser R, Near-critical water: a benign medium for catalytic reactions. Ind Eng Chem Res 40:60636067 (2001). Galkin AA and Lunin VV, Subcritical and supercritical water: a universal medium for chemical reactions. Russ Chem Rev 74:2135 (2005). Hata S, Wiboonsirikul J, Maeda A, Kimura Y and Adachi S, Extraction of defatted rice bran by subcritical water treatment. Biochem Eng J 40:4453 (2008). Sereewatthanawut I, Prapintip S, Watchiraruji K, Goto M, Sasaki M and Shotipruk A, Extraction of protein and amino acids from deoiled rice bran by subcritical water hydrolysis. Bioresour Technol 99:555561 (2008). Faulds CB and Williamson G, The role of hydroxycinnamates in the plant cell wall. J Sci Food Agric 79:393395 (1999). Bunzel M, Allerdings E, Sinwell V, Ralph J and Steinhart H, Cell wall hydroxycinnamates in wild rice (Zizania aquatica L.) insoluble dietary bre. Eur Food Res Technol 214:482488 (2002). Hudson EA, Dinh PA, Kokubun T, Simmonds MSJ and Gescher A, Characterization of potentially chemopreventive phenols in extracts of brown rice that inhibit the growth of human breast and colon cancer cells. Cancer Epidemiol Biomarkers Prev 9:11631170 (2000). Kuroda K-I, Suzuki A, Kato M and Imai K, Analysis of rice (Oryza sativa L.) lignin by pyrolysisgas chromatography. J Anal Appl Pyrol 34:112 (1995). Yu L, Wheat Antioxidants. Wiley-Interscience, Hoboken, NJ (2008). Somoza V, Five years of research on health risks and benets of Maillard reaction products: an update. Mol Nutr Food Res 49:663672 (2005). Adams A, Borrelli RC, Fogliano V and De Kimpe N, Thermal degradation studies of food melanoidins. J Agric Food Chem 53:41364142 (2005). Ribereau-Gayon P, Plant Phenolics. Oliver and Boyd, Edinburgh (1972). Whiting D, Natural phenolic compounds 19002000: a birds eye view of a centurys chemistry. Nat Prod Rep 18:583606 (2001). Oridorga VA, Korobeinikova II and Yasnitski BG, Thermographic study of gallic acid. Khimiya Prirodnykh Soedinenii 4:527528 (1978). Taniguchi H, Nomura E, Hosoda A, Tsuno T and Maruta Y, Thermally stable ferulic acid derivatives. US Patent 20040152912 (2005). Fiddler W, Parker WE, Wasserman AE and Doerr RC, Thermal decomposition of ferulic acid. J Agric Food Chem 15:757761 (1967).

22 23 24 25 26 27

28 29 30

31 32 33 34 35 36 37 38 39

2581
wileyonlinelibrary.com/jsfa

J Sci Food Agric 2010; 90: 25762581

c 2010 Society of Chemical Industry

Research Article
Received: 12 May 2009 Revised: 15 February 2010 Accepted: 12 July 2010 Published online in Wiley Online Library: 17 August 2010

(wileyonlinelibrary.com) DOI 10.1002/jsfa.4124

Nutritional composition and condensed tannin concentration changes as browse leaves become litter
Amanda Acero,a James P Muirb and Richard M Wolfeb
Abstract
BACKGROUND: The role of condensed tannins (CT) in ruminant nutrition and health makes changes in leaf litter (LL) after abscission of interest. This study compared the effect of different drying methods of green leaves (GL) with that of natural drying of LL on CT, bre, crude protein (CP) and phosphorus (P) concentrations in nine Texas browse species. Leaves harvested before autumn shedding were oven-dried (OD) or freeze-dried (FD). RESULTS: Where different (P < 0.05), extractable CT concentrations were higher while protein- and bre-bound CT concentrations were lower in GL-FD than in LL. Drying method changed total CT concentration in three species. Where different, bre fraction concentrations were greater in LL than in GL, regardless of drying method. In some species, CP and P concentrations were lower in LL than in GL, but in ve species they did not change (P > 0.05) from GL to LL, with CP concentrations ranging from 63 to 151 g kg1 in the latter. CONCLUSION: Browse LL had high nutritive value and CT concentrations, explaining why browsing ruminants utilise this feed resource. However, changes in nutrient and CT concentrations as leaves become litter in some species mean that information on one is not necessarily applicable to the other. c 2010 Society of Chemical Industry Keywords: browse; leaf litter; nutritive value; tree forage quality; drying method

INTRODUCTION
Goats use tree, brush and vine leaf litter (LL) as a feed source when green browse is not available. This has been documented in warm climates with prolonged annual droughts1,2 as well as in cooler climates where leaf loss is associated with shorter day length and cool temperatures.3 Although studies have indicated that the nutritive value of LL is generally lower than that of green leaves (GL),3,4 the dynamics of condensed tannin (CT) concentrations and CT fractions as leaves change to litter are not known. Greater knowledge of CT and other nutritive values such as crude protein (CP), bre fractions and phosphorus (P) in LL will help to understand the role they play in environments where this feed resource is important to browsing ruminants. Tannins are polyphenolic compounds with the ability to precipitate proteins and play a role in ecological processes such as LL decomposition, nutrient cycling, nitrogen sequestration and microbial activity.5,6 They can be divided into two major classes, namely condensed and hydrolysable.7 CT have both positive and negative effects in ruminants, depending on their concentration in forages. According to Barry and McNabb,8 the ideal concentration of CT in forage legumes fed to ruminants is between 20 and 40 g kg1 . In higher concentrations, CT can reduce feed intake, protein and carbohydrate digestibility and animal performance.9 Moderate levels of CT increase the efciency of CP utilisation through greater ow of protein to the duodenum10 by inhibiting ruminal degradation of plant proteins and enhancing ruminal protein

escape (bypass).11 Research has also shown that plants containing CT may have anthelmintic properties.12 14 Feeding ruminants these plants can suppress gastrointestinal parasite faecal egg counts.15 Previous research on herbaceous species indicates that concentrations of CT, neutral detergent bre (NDF), acid detergent bre (ADF) and acid detergent lignin (ADL)16,17 change with drying method, i.e. oven-drying (OD) or freeze-drying (FD), owing to the formation of complexes between tannins and cell wall compounds, increasing CT bound to protein and bre.18 However, the effect of natural drying of LL on CT fractions present before abscission has not been thoroughly studied. To better understand the CT and nutritive value changes that occur as leaves become LL, browse species known to have CT in greater concentrations than 20 g kg1 on a dry matter (DM) basis were selected for this study. These included two vine species, Smilax rotundifolia (SR) and Smilax bona-nox (SB), and seven tree or shrub species, Celtis occidentalis (CO), Quercus stellata (QS), Ulmus

Correspondence to: James P Muir, Texas AgriLife Research, 1229 North U.S. Highway 281, Stephenville, TX 76401, USA. E-mail: j-muir@tamu.edu

a Universidad de Cundinamarca, Facultad de Ciencias Agropecuarias, Diagonal 18 No. 20-29, Fusagasug , Colombia a b Texas AgriLife Research, 1229 North U.S. Highway 281, Stephenville, TX 76401, USA

2582

J Sci Food Agric 2010; 90: 25822585

www.soci.org

c 2010 Society of Chemical Industry

Compositional changes as browse leaves become litter crassifolia (UC), Bumelia lanuginosa (BL), Quercus virginiana (QV), Carya illinoinensis (CI) and Gleditsia triacanthos (GT). The objective of our study was to determine the composition of key compounds and nutrients in pre-abscission leaves and LL in order to better predict their potential nutritive value and health contributions to ruminants consuming LL during winter or dry seasons. We also wanted to compare the effect of OD or FD of leaves with that of the natural drying that occurs when leaves become LL.

www.soci.org The dependent variables were ECT, PBCT, FBCT, TCT, CP, P, NDFom, ADFom and ADLom.

RESULTS
Nutritional composition There were no differences in CP concentration among GL-OD, GL-FD and LL in all species except QS, UC and CI, where CP concentration was lower in LL (Table 1). P concentration decreased in LL relative to GL-FD of SB, QS, UC and BL as well as relative to GL-OD of SR, SB, CO, QS, UC and BL. NDFom concentration was affected by drying method in SR, CI and GT, where higher values were found in LL (Table 2). The vines SR and SB had higher ADFom concentrations in LL than in GL-FD, but there were no differences with GL-OD. ADFom concentrations in QS, CI, UC and GT were higher in LL than in GL-OD and GL-FD. ADLom concentration was likewise higher in LL of QS and CI, while SR, SB and CO had higher ADLom concentrations in LL and GL-OD than in GL-FD. Condensed tannins ECT concentration was higher in GL-FD than in GL-OD and LL of SB, CO and QS, while CI, UC and GT had higher ECT concentration in GL-FD than in LL, but no different from GL-OD (Table 3). PBCT concentration was greater in GL-OD than in GL-FD and LL of UC and CI. In contrast, there were no differences in PBCT concentration between GL-OD and LL of SB and BL, but PBCT concentration in GL-FD was lower in these species. FBCT concentration was higher in GL-OD than in GL-FD and LL of SR, SB, QS and UC, while QV had higher FBCT concentration in LL than in GL-OD and GL-FD. Drying method changed TCT only in GT, UC and CI.

MATERIALS AND METHODS

Leaves from plants of each species were harvested at three locations in October 2007 at the Texas AgriLife Research Center, Stephenville, TX, USA (32 15 N, 98 12 W, altitude 395 m) and immediately frozen with dry ice. Subsequently, half were ovendried (GL-OD) at 55 C for 48 h and the other half were freeze-dried (GL-FD) at 40 C for 72 h (FTS System Dur-Top MP, SP Industries, Warminster, PA, USA). Shade cloth nets were placed under the plants at the same time to collect LL throughout January 2008. Nutrient and CT fraction compositions were determined for GLOD, GL-FD and LL. Samples were digested by the Dumas method19 using an Elementar Vario Macro C : N analyser (Elementar Americas, Inc., Mt Laurel, NJ, USA) to determine nitrogen (N) concentration. CP was estimated as N 6.25.20 An ANKOM200 bre analyser (Ankom Technologies, Macedon, NY, USA) was used to determine neutral detergent bre (NDFom), acid detergent bre (ADFom) and acid detergent lignin (ADLom) in a sequential analysis with the addition of sodium sulte and heat-stable -amylase. These values were corrected to an organic matter (om) basis. Samples of 0.5 g were digested in sulfuric acid and extracts were used to determine P concentration by a modied ascorbic acid method.21 Extractable (ECT), protein-bound (PBCT), bre-bound (FBCT) and total (TCT) CT in GL-OD, GL-FD and LL were determined by the butanol/HCl procedure.22 CT were extracted from each species to develop a specic standard curve for each species.17 Plant harvested (three for each species) was considered as replicate. Data were analysed for variance by species using a least signicant difference test (P 0.05) to determine mean treatment differences. Differences were considered signicant at P 0.05 unless noted otherwise in the text. The general lineal model procedure of SAS/STAT 9.123 was used based on the statistical model Yij = + j + Eij , where j is the effect of drying method.

DISCUSSION
CP concentrations measured in the LL samples indicate that some species may be good sources of CP for concentrate-selective ruminants such as goats. The lowest CP concentration in LL was 63 g kg1 for CI, while the highest was 151 g kg1 for SB. These concentrations are appropriate for satisfying protein requirements in goats, whose reported requirements vary between 2.03 and 3.07 g kg1 (body weight)0.75 for maintenance.24,25 LL from most

Table 1. Crude protein (CP) and phosphorus (P) concentrations in green leaves oven-dried (GL-OD) or freeze-dried (GL-FD) and leaf litter (LL) of Texas trees and vines CP (g kg1 ) Species Bumelia lanuginosa Carya illinoinensis Celtis occidentalis Gleditsia triacanthos Quercus stellata Quercus virginiana Smilax rotundifolia Smilax bona-nox Ulmus crassifolia GL-FD 112a 120a 123a 134a 121a 91a 139a 159a 121a GL-OD 112a 125a 135a 143a 124a 91a 155a 164a 120a LL 95a 63b 87a 107a 67b 102a 137a 151a 87b P 0.34 0.01 0.05 0.10 0.01 0.42 0.10 0.33 0.01 GL-FD 1.1a 1.2a 1.2b 2.3a 1.1b 2.0a 1.4ab 1.7b 1.2b P (g kg1 ) GL-OD 1.2a 1.4a 1.7a 2.7a 1.6a 1.7a 2.0a 2.2a 1.5a LL 0.6b 0.8a 1.1b 1.8a 0.2c 0.2a 0.9b 1.0c 0.4c P 0.04 0.06 0.01 0.66 0.01 0.34 0.02 0.01 0.01

Means within rows under the same subheading followed by different letters differ according to a least signicant difference multiple mean separation (P 0.05).

2583

J Sci Food Agric 2010; 90: 25822585

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org

A Acero, JP Muir, RM Wolfe

Table 2. Neutral detergent bre (NDFom), acid detergent bre (ADFom) and acid detergent lignin (ADLom) concentrations in green leaves oven-dried (GL-OD) or freeze-dried (GL-FD) and leaf litter (LL) of Texas trees and vines NDFom (g kg1 ) Species Bumelia lanuginosa Carya illinoinensis Celtis occidentalis Gleditsia triacanthos Quercus stellata Quercus virginiana Smilax bona-nox Smilax rotundifolia Ulmus crassifolia GL-FD 250a 248b 344a 322b 396a 460a 344a 336c 501a GL-OD 352a 282b 323a 325a 399a 450a 388a 378b 482a LL 341a 349a 330a 420a 398a 482a 409a 411a 421a P 0.57 0.01 0.87 0.01 0.98 0.32 0.13 0.01 0.22 GL-FD 166a 148b 166a 238b 225b 316a 215b 220b 168b ADFom (g kg1 ) GL-OD 173a 160b 153a 244b 225b 310a 250a 242ab 193b LL 233a 211a 209a 319a 263a 333a 258a 261a 265a P 0.46 0.01 0.09 0.01 0.01 0.44 0.01 0.04 0.01 GL-FD 74a 63b 44b 144b 102b 121a 60b 60b 58b ADLom (g kg1 ) GL-OD 87a 73b 59a 168ab 99b 127a 95a 87a 85ab LL 117a 100a 80a 211a 125a 136a 86a 95a 123a P 0.29 0.01 0.01 0.03 0.02 0.37 0.01 0.01 0.01

Means within rows under the same subheading followed by different letters differ according to a least signicant difference multiple mean separation (P 0.05).

Table 3. Extractable (ECT), protein-bound (PBCT), bre-bound (FBCT) and total (TCT) condensed tannin concentrations in green leaves oven-dried (GL-OD) or freeze dried (GL-FD) and leaf litter (LL) of Texas trees and vines ECT (g kg1 ) Species Bumelia lanuginosa Carya illinoinensis Celtis occidentalis Gleditsia triacanthos Quercus stellata Quercus virginiana Smilax bona-nox Smilax rotundifolia Ulmus crassifolia GLFD 62.8a 25.3a 19.0a 36.9a 22.2a 40.1a 44.0a 16.9a 37.2a GLOD 32.1a 15.3ab 1.1b 24.6ab 6.1b 34.3a 17.3c 8.1a 36.4a LL 37.0a 5.8b 0.0b 8.2b 10.4b 30.3a 25.0b 23.2a 5.3b P 0.10 0.02 0.01 0.03 0.01 0.27 0.01 0.08 0.01 GLFD 21.7b 20.6b 3.7a 27.6a 7.8a 9.4a 13.8b 16.4a 9.8b PBCT (g kg1 ) GLOD 35.7a 40.5a 2.8a 27.7a 12.5a 16.4a 27.5a 21.2a 41.6a LL 35.2a 24.0b 4.7a 25.7a 6.9a 13.2a 25.5a 22.1a 15.6b P 0.04 0.01 0.09 0.95 0.07 0.08 0.01 0.46 0.03 GLFD 2.7a 7.7a 5.0a 8.5a 2.5b 2.2b 3.5b 3.9b 4.3c FBCT (g kg1 ) GLOD 7.0a 8.6a 12.4a 6.9a 7.4a 2.8b 14.8a 23.3a 23.7a LL 4.1a 5.6a 5.6a 5.6a 2.9b 5.8a 2.6b 2.6b 10.2b P 0.05 0.27 0.06 0.27 0.01 0.01 0.01 0.003 0.01 GLFD 76.2a 53.6a 27.6a 73.0a 32.5a 51.7a 61.3a 37.2a 51.3b TCT (g kg1 ) GLOD 87.2a 64.4a 16.3a 45.6b 26.0a 53.5a 59.6a 52.6a 91.7a LL 74.9a 35.5b 10.2a 39.6b 30.2a 49.3a 53.1a 48.0a 31.1c P 0.63 0.05 0.08 0.01 0.49 0.85 0.15 0.23 0.01

Means within rows under the same subheading followed by different letters differ according to a least signicant difference multiple mean separation (P 0.05).

of these species could contribute to meeting these requirements when other feed alternatives are limited. However, digestibility and intake of this material, with resulting nutrient bioavailability to ruminants, must be determined before usefulness of this feed can be dened. The lower P concentration in LL compared with GL may be due to the normal process of translocating nutrients in the plant before LL fall.26 This phenomenon may not be as developed for N in most of the species studied, since only QS, UC and CI had lower N concentration in LL compared with GL. The effect of drying method on the bre fraction became progressively more evident with each assay step. The NDFom fraction was greater in the LL of only three species, ADFom was greater in six species, whereas ADLom was more concentrated in seven species compared with GL. This indicates that ADLom is more stable and degrades less in LL compared with hanging leaves than ADFom, which in turn is more stable than NDFom. Schoeld et al.,27 in a study of willow tree LL, found that extractable phenolics and tannins were rapidly lost from the leaves and had half-lives of approximately 2.4 weeks. Lower-molecular-weight tannins were lost more rapidly than

higher-molecular-weight tannins, suggesting that the primary route for loss of tannins is leaching. These results parallel those of three species in our study, but not the six for which there was no decline in TCT from GL to LL. Our results would indicate that molecular weights in the latter six species may be greater than those in the rst three species. TCT concentrations in LL ranged between 10 and 75 g kg1 DM. Four species in our study had LL TCT concentrations greater than 45 g kg1 , levels that might adversely affect the palatability to or performance of ruminants.28 Further research is needed to determine if the dynamics of the CTruminant interactions in LL follow the same general trends as those of fresh browse or forage, especially since PBCT and FBCT increase in LL compared with GL as a proportion of TCT. Further research is also recommended on those species with high variability in CT concentrations among collection sites (replications) as indicated by lack of statistical differences among treatments with large numerical differences, for example S. rotundifolia (Table 3). Hattenschwiler et al.29 reported differences in tree population CT concentrations as inuenced by soil properties in tropical forests, so this or other factors may be in play and should be identied.

2584

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 25822585

Compositional changes as browse leaves become litter Effect of drying method on PBCT and FBCT fractions differed among species. Three had greater PBCT concentrations in GL-OD, with no differences between GL-FD and LL. The same pattern was observed in FBCT concentrations of three species, where GL-OD had greater concentrations compared with GL-FD and LL. This could be partly a result of OD causing the formation of protein and bre tannin complexes as reported in previous studies,16,17 but why this pattern was not uniform across all species needs to be further studied.

www.soci.org
9 Reed JD, Soller H and Woodward A, Fodder tree and straw diets for sheep: intake, growth, digestibility, and the effects of phenolics on nitrogen utilization. Anim Feed Sci Technol 30:3950 (1990). 10 McNabb WC, Waghorn GC, Barry TN and Shelton ID, The effect of condensed tannins of Lotus pedunculatus on the digestion and metabolism of methionine, cystine and inorganic sulphur in sheep. Br J Nutr 70:647661 (1993). 11 Makkar HPS, Effects and fate of tannins in ruminant animals, adaptation to tannins, and strategies to overcome detrimental effects of feeding tannin-rich feeds. Small Rumin Res 49:241256 (2003). 12 Molan AL, Waghorn GC, Min BR and McNabb WC, The effect of condensed tannins from seven herbages on Trichostrongylus colubriformis larval migration in vitro. Folia Parasitol 47:3944 (2000). 13 Paolini V, Bergeaud JP, Grized C, Prevot F, Dorchies P and Hoste H, Effects of condensed tannins on goats experimentally infected with Haemonchus contortus. Vet Parasitol 113:253261 (2003). 14 Lange KC, Olcott DD, Miller JE, Mosjidis JA, Terrill TH, Burke JM, et al, Effect of Sericea lespedeza (Lespedeza cuneata) fed as hay, on natural and experimental Haemonchus contortus infections in lambs. Vet Parasitol 141:273278 (2006). 15 Min BR and Hart SP, Tannins for suppression of internal parasites. J Anim Sci 81:E102E109 (2003). 16 Steward JL, Mould F and Mueller-Harvey I, The effect of drying treatment on the fodder quality and tannin content of two provenances of Calliandra calothyrsus Meissner. J Sci Food Agric 80:14611468 (2000). 17 Wolfe RM, Terrill TH and Muir JP, Drying method and origin of standard affect condensed tannin (CT) concentrations in perennial herbaceous legumes using simplied butanol-HCl CT analysis. J Sci Food Agric 88:10601067 (2008). 18 Terrill HT, Windham WR, Evans JJ and Hoveland C, Effect of drying method and condensed tannin on detergent ber analysis of Sericea lespedeza. J Sci Food Agric 66:337343 (1994). 19 AOAC, Ofcial Methods of Analysis (15th edn), Vol. 1. Association of Ofcial Analytical Chemists. Arlington, VA (1990). 20 Van Soest PJ, Nutritional Ecology of the Ruminant (2nd edn). Cornell University Press, Ithaca, NY (1994). 21 APHA, Phosphorus, in Standard Methods for the Examination of Water and Wastewater, ed. by Eaton AD, Clesceri LS and Greenberg AE. American Public Health Association, Bethesda, MD, pp. 106115 (1995). 22 Terrill TH, Rowan AM, Douglas GD and Barry TN, Determination of extractable and bound condensed tannin concentrations in forage plants, protein concentrate meals and cereal grains. J Sci Food Agric 58:321329 (1992). 23 SAS, SAS/STAT 9.1 Users Guide. SAS Institute, Cary, NC (2004). 24 NRC, Nutrient Requirements of Small Ruminants: Sheep, Goats, Cervids, and New World Camelids. National Research Council, Washington, DC (2007). 25 Luo J, Goetsch AL, Nsahlai IV, Sahlu T, Ferrell CL, Owens FN, et al, Metabolizable protein requirements for maintenance and gain of growing goats. Small Rumin Res 53:309326 (2004). 26 Mafongoya PL, Guiller KE and Palm CA, Decomposition and nitrogen release patterns of tree prunings and litter. Agroforest Syst 38:7797 (1998). 27 Schoeld JA, Hagerman AE and Harold A, Loss of tannins and other phenolics from willow leaf litter. J Chem Ecol 24:14091421 (1998). 28 Min BR, Barry TN, Attwood GT and McNabb WC, The effect of condensed tannins on the nutrition and health of ruminants fed fresh temperate forages: a review. Anim Feed Sci Technol 106:319 (2003). 29 Hattenschwiler S, Hagermand AE and Vitousek PM, Polyphenols in litter from tropical montane forests across a wide range in soil fertility. Biogeochemistry 64:129148 (2003).

CONCLUSIONS
The transformation from GL to LL reduces nutritive value in some species but not others, and these characteristics may explain ruminant selectivity and performance where LL is an important ruminant feed resource. LL of some species may be a useful source of CP for browsers in situations, such as winter, where green material is not available. Fibre and CT concentrations are not excessive in the species studied, so these should not compromise animal performance. However, the implications of decreased proportions of ECT and increased PBCT and FBCT in some species as leaves abscise need to be further studied, especially in regard to ruminal protein bypass (escape) or other secondary effects on ruminant health. The effect of drying method on the variables evaluated did not show a uniform pattern among the species studied; however, OD of GL tended to closer emulate LL compared with FD of GL, although patterns were not consistent across species. More studies are necessary to better understand the interaction between drying, CT and cell wall bre components. For example, a better understanding of the CT chemical characteristics in these species could help to explain the variation in response to leaf drying and abscission.

ACKNOWLEDGEMENT
This research was funded in part by grant 2005-51300-02392 from the USDA-CSREES Integrated Organic Program.

REFERENCES
1 Pster JA and Malechek JC, Dietary selection by goats and sheep in a deciduous woodland of northeastern Brazil. J Range Manag 39:2428 (1984). 2 Scholte PT, Leaf litter and Acacia pods as feed for livestock during the dry season in Acacia-Commiphora bushland, Kenya. J Arid Environ 22:271276 (1992). 3 Packard CE, Muir JP and Wittie RD, Effects of groundnut stover or bermudagrass hay supplementation to doe kids on winter hardwood range. Small Rumin Res 67:16 (2007). 4 Kronberg SL and Malechek JC, Relationships between nutrition and foraging behavior of free-ranging sheep and goats. J Anim Sci 75:17561763 (1997). 5 Kraus TEC, Dahlgren RA and Zasoski RJ, Tannins in nutrient dynamics of forest ecosystems. Plant Soil 256:4166 (2003). 6 Haase K and Wantzen KM, Analysis and decomposition of condensed tannins in tree leaves. Environ Chem Lett 6:7175 (2008). 7 Reed JD, Nutritional toxicology of tannins and related polyphenols in forage legumes. J Anim Sci 73:15161528 (1995). 8 Barry TN and McNabb WC, The implications of condensed tannins on the nutritive value of temperate forages fed to ruminants. Br J Nutr 81:263272 (1999).

2585
wileyonlinelibrary.com/jsfa

J Sci Food Agric 2010; 90: 25822585

c 2010 Society of Chemical Industry

Research Article
Received: 14 April 2010 Revised: 7 June 2010 Accepted: 13 July 2010 Published online in Wiley Online Library: 17 August 2010

(wileyonlinelibrary.com) DOI 10.1002/jsfa.4125

Fruit dry matter concentration: a new quality metric for apples

John W Palmer,a F Roger Harker,b D Stuart Tustinc and Jason Johnstonb
Abstract
BACKGROUND: In the fresh apple market fruit must be crisp and juicy to attract buyers to purchase again. However, recent studies have shown that consumer acceptability could be further enhanced by improving taste. This study evaluates the use of fruit dry matter concentration (DMC) as a new fruit quality metric for apple. RESULTS: Fruit samples collected at harvest, in the two main fruit growing regions of New Zealand, showed a variation in mean fruit DMC from 130 to 156 g kg1 with Royal Gala and with Scifresh from 152 to 176 g kg1 . Individual fruit DMC showed a larger range, from 108 to 189 g kg1 with Royal Gala and from 125 to 201 g kg1 with Scifresh. Fruit DMC proved a more reliable predictor of total soluble solids after 12 weeks of air storage at 0.5 C than TSS at harvest for both Royal Gala and Scifresh. Fruit DMC was also positively related to esh rmness, although this relationship was not as strong as that seen with soluble solids and was more dependent on cultivar. Consumer studies showed that consumer preference was positively related to fruit DMC of Royal Gala apples. CONCLUSION: Fruit DMC can therefore be measured before or at harvest, and be used to predict the sensory potential for the fruit after storage. c 2010 Society of Chemical Industry Keywords: Malus domestica Borkh.; esh rmness; fruit maturity; soluble solids; consumer panel; dry matter concentration

INTRODUCTION
Eating quality in the broadest sense (texture, taste, odour) is one of the primary reasons consumers purchase fruit1 3 and consequently the development of knowledge and indices to predict eating quality has become a signicant focus for postharvest biologists and technologists.4 6 Research has generally sought to establish key objective measurements that can be used to predict consumer liking and willingness to purchase fruit4,7 or specic sensory attributes as measured by trained panellists.8 10 This need for objective measurements to predict consumer responses has evolved in two ways: (1) a move from destructive to nondestructive instruments so that each individual fruit can be sorted into the appropriate quality categories;11,12 and (2) the desire to apply quality indices based on the fundamental biology of the crop, which can be monitored through fruit growth and development and applied at harvest to represent the quality of an entire crop.13 The advantage of the latter approach is that it provides early insights for logistics and marketing and can help to identify physiological processes and crop management practices that can be modied in order to provide solutions that improve quality across an entire fruit sector. Traditionally, the best examples of the biologically based quality standards involve the monitoring of fruit maturity to support decisions on the timing of harvest.14 16 In the fresh apple market, fruit with good textural properties are paramount; the fruit must be crisp and juicy to attract buyers to purchase again. A recent consumer study of fruit from ve apple cultivars in the USA showed that esh rmness is the primary edible quality factor determining consumer acceptance.4 Much of postharvest science over the last 30 years has been devoted to

the control and maintenance of esh rmness in apples. However, once fruit are above a minimum threshold for esh rmness, consumer acceptability could be further enhanced by improving taste (total soluble solids (TSS) and titratable acidity).4 In other words, although texture is paramount, the customer is also looking for high taste intensity in the fresh product. While volatiles provide the signature orthonasal and retronasal stimuli responsible for characteristic avour of apple,17 our ability to establish which individual or combinations of aroma compounds contribute to consumers preferences for whole apples has proven difcult until recently,18 probably due to biological variability associated with both apples and human ability to perceive odour. Thus there may be opportunities for a more holistic predictor of eating quality based on knowledge of physiological and metabolic processes. In recent times there has been an increasing focus on fruit dry matter concentration (DMC), either as it relates to maturity (e.g. avocado) or to consumer preference in itself.13,19 For example, fruit DMC of kiwifruit at harvest has been shown to be a good predictor of ripe soluble solids content after storage.13 Consumer

Correspondence to: John W Palmer, New Zealand Institute for Plant and Food Research Limited (PFR), 55 Old Mill Road, RD 3, Motueka 7198, New Zealand. E-mail: John.Palmer@plantandfood.co.nz

a New Zealand Institute for Plant and Food Research Limited (PFR), Motueka 7198, New Zealand b PFR, Mt Albert Research Centre, Auckland 1142, New Zealand c PFR,HawkesBayResearch Centre,HavelockNorth,Hastings4157,NewZealand

2586

J Sci Food Agric 2010; 90: 25862594

www.soci.org

c 2010 Society of Chemical Industry

Apple fruit dry matter concentration studies have also shown that kiwifruit harvested with high DMC are preferred, after ripening, to fruit harvested with low DMC levels.13 As these authors point out, the improvement in consumer liking is not just associated with the higher TSS, because, for example, titratable acidity was also positively correlated with fruit DMC. In similar studies with avocado, consumers showed a progressive increase in both liking and intent to buy as DMC increased from 200 g kg1 (minimally mature) to nearly 400 g kg1 (very mature).19 It is surprising that there has not been an earlier attempt to use fruit DMC as a quality index for apples, given its success with other fruits. It has been shown that fruit DMC of Royal Gala apple at harvest can be used to predict TSS after cool storage.20 Few people seem to have followed up on this linkage between apple fruit DMC and TSS after storage. This current work set out rstly to quantify the variation in apple fruit DMC, within and between fruit samples, and secondly to extend this earlier work across several apple cultivars but, in addition, to examine possible links between fruit DMC and other key quality attributes such as esh rmness, in the light of the importance of fruit texture for apple. However, in attempting to develop a new quality index, one of the critical early steps is to establish that the measurement can predict consumer liking, and so the third part of this work addresses this issue. We believe this is the rst extensive study of its kind with apple.

www.soci.org For sensory assessment experiments, Royal Gala apples (100110 count) were harvested from nine orchard blocks in Hawkes Bay and nine orchard blocks in Nelson during February 2008. Within each region there were three DMC categories (low, moderate and high), and three orchard blocks per DMC. Fruit were representative of those harvested during the rst main commercial pick (excluding an initial skim pick), with fruit either sampled directly from the picked bins or directly from the tree 12 days before commercial harvest. All fruit met commercial requirements for blush coverage and background colour. A 30-fruit sample from each orchard block was then assessed for at-harvest characteristics (within 24 h of harvest), and the remainder (170200 fruit) were placed in air storage at 0.5 C for 1012 weeks. At-harvest assessments included skin background colour, internal ethylene concentration, fruit DMC, rmness, TSS and SPI (using the methodology already described). Internal ethylene concentration was determined by injecting a 1 mL core gas sample from each fruit into a gas chromatograph equipped with a ame ionisation detector. Post storage, each apple was assessed for rmness, DMC, TSS and titratable acidity before using the remaining tissue for sensory assessment. A wedge adjacent to the puncture test was used for titratable acidity determination. A subsample of 5 g of tissue was homogenised in 25 mL water using a Polytron homogeniser (Kinematica, Luzern, Switzerland), and titrated to an endpoint of pH 8.1 with 0.1 mol L1 sodium hydroxide (NaOH) using an automatic titrator (Model 702 SM Titrino, Metrohm, Herisau, Switzerland). Results were expressed as malic acid equivalent (g kg1 ). One hundred and four consumers from the general population of New Zealand living in Auckland were recruited on the basis of being between the ages of 18 and 65 years and being regular apple eaters, i.e. eating apples more than once a week. Fifty-two percent were male, 29% aged 1830, 38% aged 3145 and 33% aged 4660 years. The vast majority (92%) consumed Royal Gala apples. Each consumer tasted four sets of three fruit. Two sets were from Nelson and two from Hawkes Bay, and each set included a fruit from each DMC category. The order in which fruit were tasted was according to a randomised complete block statistical design that accounted for order and carryover effects. Care was taken to ensure all orchards were equally represented within each DMC category, although each consumer did not taste fruit from all orchards. Consumers liking of apples was assessed on a 9-point category scale using a modied rank-rating approach, in which consumers were asked to consider the relative as well as absolute scores.13,19 Following this they indicated taste acceptability and their willingness to purchase fruit if available at an average price ($NZ 2.60 kg1 ) using a 6-category scale.19 Tastings were undertaken under red-coloured lighting in sensory facilities described previously.3 Statistical analysis of variation in fruit DMC across orchards and relationships between fruit DMC and other quality variates was completed using linear regression with Genstat (version 10.1.0.71, VSN International Ltd, Hemel Hempstead, UK) or Origin (version 7.5, OriginLab Corp., Northampton, MA, USA). Mixed-effects models relating instrumental measurements to consumer responses were tted using PROC Mixed in the SAS 9.1 statistical package. The region and DMC categories were treated as xed effects, with the effect related to orchard being modelled as random. This allowed the estimates to take into account the structure of the data, in that fruit from within the same orchard may be more similar than fruit from different orchards. Additionally the

MATERIALS AND METHODS

Fifty-fruit samples of Royal Gala and Scifresh fruit were collected just before the start of commercial harvest in 2005 from 12 orchards in the Nelson (latitude 41 S, longitude 173 E) and 17 in the Hawkes Bay (latitude 40 S, longitude 177 E) regions of New Zealand. Fruit DMC was measured by drying two wedges per fruit at 65 C in a forced draught oven for at least 48 h to a constant weight. The fruit wedges were taken from opposite sides of each fruit, using two longitudinal cuts. This sampling procedure was adopted to allow for the variation in fruit DMC within individual fruit.21 The sampled tissue fresh weight was 1525 g before drying. Further fruit samples were taken from the rst and the third select picks of Royal Gala and Scifresh from four orchards in each of Nelson and Hawkes Bay. From each pick, 10 fruit were used to record fruit DMC and a further 10 fruit used to assess at-harvest quality: TSS using an Atago Pocket PAL-1 digital refractometer, esh rmness using a Guss Fruit Texture Analyser tted with an 11.1 mm probe, starch pattern index (SPI) using a scale from 0 to 6, no clearance to full starch clearance, and skin background colour using ENZA swatches from 1 to 10, green to yellow. Additional 40-fruit samples were air-stored for either 6 or 12 weeks at 0.5 C. After storage, fruit DMC, fruit rmness and TSS were recorded. This was repeated for Royal Gala in the 2005/6 growing season. A further dataset was obtained in the 2005/6 growing season from eight commercial apple cultivars (Coxs Orange Pippin clone Greenmeadows, Fuji, Granny Smith, Scifresh, Sciearly, Sciros, Pink Lady and Royal Gala). The fruit were sourced from Hawkes Bay during the commercial harvesting period, couriered to PFR, Auckland, and stored at 0.5 C, or 3 C for Cox. After storage, two adjacent wedges of fruit were removed from each of 22 fruit of each cultivar and one used for measurement of TSS and one for fruit DMC. In this case, TSS was determined on juice obtained by passing the fruit wedge through a domestic juicer, rather than using the juice on the tip of the penetrometer probe used in the other trials.

2587

J Sci Food Agric 2010; 90: 25862594

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org

JW Palmer et al.

Table 1. Fruit DMC (g kg1 ) of orchard samples (50 fruit) of Royal Gala and Scifresh apples harvested just prior to or on the day of the rst commercial harvest. Samples were collected from 12 orchards in the Nelson and 17 orchards in the Hawkes Bay (HB) regions of New Zealand Cultivar: Region: Mean Range of samples Range of individual fruit Royal Gala Nelson 138 130150 108175 HB 147 137156 113189 Scifresh Nelson 159 152165 132198 HB 164 156176 125201

panellist may have a similar effect on liking scores and purchase intent. For these analyses, the panellist effect was also treated as random. Since the acceptability data were binary, with acceptable or not acceptable the only possible answers, a generalised linear mixed effects model was used to account for the orchard and panellist effects. This was estimated in Genstat, and used the approach of Breslow and Clayton.22 The tted curves in the plots of physical measurements against liking scores were estimated using a LOESS smoother routine using R version 2.90 (R Development Core Team http://www.R-project.org). This had the advantage of not having to constrain the data to a preconceived theoretical model.

Similar trends were evident with Scifresh orchard samples to those observed with Royal Gala orchard samples. The relationship between fruit DMC at harvest and TSS improved with storage time (Fig. 3). In this case, the relationship was strongest after 6 weeks in storage. Fruit DMC at harvest again proved a better predictor of TSS after storage than TSS at harvest. After 6 weeks of storage, variance accounted for by the regression between harvest fruit DMC and TSS was 78%, compared with 57% when harvest TSS was used. After 12 weeks, the two values were 54% and 32% respectively. On an individual fruit basis at harvest, TSS of Scifresh was poorly related to fruit DMC (r2 = 0.17). This relationship improved with storage after 6 weeks (r2 = 0.61) and after 12 weeks (r2 = 0.68). Considering orchard samples of both cultivars together, the relationship between TSS and fruit DMC, where both variates were measured after 12 weeks, was highly statistically signicant (r2 = 0.97). In the absence of a signicant effect of cultivar, both cultivars could be represented by the same linear relationship (Fig. 4). A similar result was found in the study of individual fruit from eight cultivars (Fig. 5) where TSS after storage was highly correlated with fruit DMC recorded after storage (r2 = 0.77). Relationship between fruit DMC and esh rmness There was no statistically signicant relationship between fruit DMC at harvest and rmness for Royal Gala orchard samples at harvest or rmness after 6 or 12 weeks of storage. Flesh rmness after storage was primarily related to esh rmness at harvest, with 52% and 55% of the variation at 6 and 12 weeks explained by esh rmness at harvest. However, when the data for individual fruit were analysed, there was a signicant positive relationship between fruit DMC and rmness, although the proportion of the variation accounted for by the regression was 6%, 10% and 12% for samples at harvest, 6 weeks and 12 weeks respectively. However, with Scifresh orchard samples, fruit DMC at harvest was signicantly related to esh rmness both at harvest and after storage, with the regressions accounting for 22%, 52% and 44% of the variation in esh rmness at harvest and after 6 weeks and 12 weeks of storage, respectively. Flesh rmness at harvest, however, was a better predictor of esh rmness after storage than fruit DMC at harvest, with regressions accounting for 70% and 73% of the variation in esh rmness after 6 weeks and 12 weeks of storage respectively. Post storage, esh rmness in the orchard samples was signicantly related to fruit DMC, where fruit DMC was recorded after storage, with the regressions accounting for 70% and 67% of the variance in esh rmness at 6 weeks and 12 weeks respectively. Relationship between fruit DMC and consumer response All apples were allocated into treatments on the basis of whether they were harvested from low, medium or high DM orchards, which is fundamentally different from previous studies, which have allocated fruit to treatments on the basis of instrumental measurement made just before consumer assessments.4,13,19 Given the potential high risk that individual apples allocated to treatments on the basis of orchard might not have met required DMC criteria, it was important to evaluate the effectiveness of the sorting process. The process of sorting the apples into different DMC categories was successful (Table 2), with differences between categories pronounced. All DMC categories were signicantly different from each other (P < 0.001 and P < 0.001 for all three

RESULTS
Variation across orchard blocks Considerable variation in fruit DMC was evident among orchard samples of apples (Table 1), with a 1320 g kg1 difference among orchard samples and a 70 g kg1 difference between individual fruit for any region/cultivar combination. Within one region, Scifresh orchard samples had a consistently higher fruit DMC than samples of Royal Gala. Relationship between fruit DMC and total soluble solids Although the relationship between fruit DMC at harvest and TSS at harvest of Royal Gala orchard samples was statistically signicant (r2 = 0.32), fruit DMC at harvest was more closely related to TSS after storage, with an improvement in the relationship the longer the fruit remained in storage. After 12 weeks of storage the relationship accounted for 82% of the variation (Fig. 1). The results from the 2005/6 season were similar to those of the 2004/5 season and the combined relationship after 12 weeks of storage is given in Fig. 2. The effect of year was not statistically signicant. When comparisons were made using TSS at harvest or fruit DMC at harvest as the predictive variate for TSS of Royal Gala after 6 weeks of storage, the relationships were similar; both relationships accounted for 49% of the variation. However, after 12 weeks of storage, fruit DMC accounted for 81% of the variation in TSS compared with 37% of the variation where TSS at harvest was used as the predictive variate. It was obviously not possible to use fruit DMC at harvest as a predictor of TSS after storage on the same fruit, as fruit DMC is a destructive test. However, TSS could be related to fruit DMC when the fruit DMC was determined. On an individual fruit basis, TSS of Royal Gala at harvest was poorly related to fruit DMC at harvest (r2 = 0.10). This relationship improved with storage after 6 weeks (r2 = 0.52) and after 12 weeks (r2 = 0.63).

2588

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 25862594

Apple fruit dry matter concentration

A Hawke's Bay Nelson r2 = 0.32 Soluble solids after 6 weeks storage (Brix) 14 Soluble solids at harvest (Brix)

www.soci.org
B 14 Hawke's Bay Nelson

r2 = 0.53

13

13

12

12

11

11

10 120 125 130 135 140 145 150 155 160 Dry matter concentration at harvest (g kg-1) C Soluble solids after 12 weeks storage (Brix) 14 Hawke's Bay Nelson 13 r2 = 0.82

10 120 125 130 135 140 145 150 155 160 Dry matter concentration at harvest (g kg-1)

12

11

10 120 125 130 135 140 145 150 155 160 Dry matter concentration at harvest (g kg-1)

Figure 1. Relationship between harvest dry matter concentration and (A) total soluble solids at harvest, (B) after six weeks of cool storage and (C) after 12 weeks of cool storage of Royal Gala apple fruit from the rst and third select picks of four orchards in Hawkes Bay and four orchards in Nelson, New Zealand, 2005.

14 Soluble solids after 12 weeks storage (Brix)

r = 0.78

13

12

11 2004/5 2005/6 10 120 125 130 135 140 145 150 155 160

Fruit dry matter concentration at harvest (g kg-1)

Figure 2. Relationship between harvest DMC and TSS after 12 weeks of cool storage of Royal Gala apple fruit over two seasons (200406) from the rst and third select picks of four orchards in Hawkes Bay and four orchards in Nelson, New Zealand.

pairwise comparisons after Tukeys HSD adjustment for multiple testing differences). The process of allocating apples to individual consumers ensured minimal overlap in DMC categories for any set of three apples tasted by an individual consumer. Most pairs of fruit had positive differences in DMC values (i.e. very few fruit from a higher DMC category that were actually lower in DMC) and few fruit that had no (zero) difference between DMC categories. Rather, the average DMC differences going from low to moderate, from moderate to high and from low to high were 11, 14 and 25 g kg1 respectively. Consumers were asked to taste apples of differing DMC to indicate their absolute level of liking of each fruit, and to pay attention to the relative pattern of liking between these different apples. There was a signicant increase in consumer liking for apples from the high DMC category (P = 0.001; pairwise comparisons after Tukeys HSD adjustment for multiple testing differences, high versus moderate P = 0.004, high versus low P < 0.001, moderate versus low P = 0.318). Average scores for consumer liking for the apples increased from 5.2 for the low DMC category to 6.3 for the highest DMC category (Fig. 6(A)). For the supplementary question (Was the taste of the fruit acceptable or unacceptable?), the acceptability of the DMC categories increased from 69% to 83% of apples as the DMC increased from low to high (Fig. 6(B)). DMC categories had a signicant inuence on acceptability of apples (P = 0.028). There was a signicant increase in consumer acceptance for

2589

J Sci Food Agric 2010; 90: 25862594

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org
A
16 Soluble solidsat harvest (Brix) Soluble solids after 6 weeks storage (Brix) Hawke's Bay Nelson r2 = 0.42 16 Hawke's Bay Nelson

JW Palmer et al.

B
r2 = 0.79

15

15

14

14

13

13

12 140 145 150 155 160 165 170 175 180 Dry matter concentration at harvest (g kg-1) 16 Hawke's Bay Nelson 15

12 140 145 150 155 160 165 170 175 180 Drymatter concentration at harvest (g kg-1)

Soluble solids after 12 weeks storage (Brix)

r2 = 0.57

14

13

12 140 145 150 155 160 165 170 175 180 Dry matter concentration at harvest (g kg-1)

Figure 3. Relationship between harvest DMC and (A) TSS at harvest, (B) after 6 weeks of cool storage and (C) after 12 weeks of cool storage of Scifresh apple fruit from the rst and third select picks of four orchards in Hawkes Bay and ve orchards in Nelson, New Zealand, 2005.

16 Royal Gala Scifresh 15 Soluble solids (Brix) r2 = 0.97

14

13

12

11 130

140

150

160

170

180

Fruit dry matter concentration (g kg-1)

Figure 4. Relationship between DMC and TSS after 12 weeks of cool storage of Royal Gala and Scifresh apple fruit from the rst and third select picks of four orchards in Hawkes Bay and four orchards in Nelson, New Zealand, 2005.

pairwise comparisons between low and high DMC (P < 0.011), while comparisons between high and moderate (P = 0.109) and moderate and low (P = 0.221) were not signicant. There were signicant increases in consumers intention to purchase apples at the stated price of $2.60 kg1 for each stepwise increase in DMC band (P = 0.001). Tukeys pairwise comparisons showed signicant differences between high to low DMC (P = 0.001) and high to moderate DMC (P = 0.022), but no signicant differences between moderate and low DMC (P = 0.209) (Fig. 6(C)). An alternative approach to investigating the relation between DMC and consumer responses is to group the results by actual DMC rather by an arbitrary DMC category (e.g. low, moderate and high) allocated at harvest. This type of analysis is only possible in retrospect, since actual DMC were not available until some days after the consumer tasting sessions. Analysis of these more detailed categories shown in Fig. 7(A) suggests a curvilinear relationship between increasing DMC of apples and increasing consumer liking for fruit, although there are limited sample sizes at very high and very low fruit DMC. The curve (LOESS smoother) was generated using all individual consumers responses, and the points represent consumer liking as was calculated for discrete bands of DMC. It was also possible to examine the inuence of other instrumental attributes (TSS, fruit rmness and titratable acidity) besides that of fruit DMC. Fruit from categories with higher DMC

2590

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 25862594

Apple fruit dry matter concentration

17 16 Soluble solids concentration (Brix) 15 14 13 12 11 10 9 100

www.soci.org

Cox Royal Gala Pink Lady Pacific Rose Pacific Beauty Scifresh Granny Smith Fuji

r2 = 0.77

110

120

130

140

150

160 kg-1)

170

180

190

Fruit dry matter concentration (g

Figure 5. Relationship between DMC and TSS after 12 weeks of cool storage of 22 individual fruit from eight apple cultivars sourced from Hawkes Bay, New Zealand, 2006.

Table 2. DMC, TSS, fruit rmness and titratable acidity (malic acid equivalent g kg1 ) of Royal Gala apple fruit after 1012 weeks at 0.5 C and 1 day at 20 C. Means are based on six orchard blocks (two regions, three orchard blocks per region, 8691 fruit per block), analysis of variance P-values (12 d.f.e.), and 5% least signicant differences (LSD) are displayed Dry matter category High Moderate Low P-value 5% LSD DMC (g kg1 ) 157 143 132 <0.001 4.5 Soluble solids ( Brix) 13.6 12.7 11.8 <0.001 0.40 Firmness (kgf) 6.62 6.21 6.19 0.1133 0.45 Titratable acidity (g kg1 ) 2.9 2.6 2.4 0.02 0.30

also had higher TSS, rmness and titratable acidity (Table 2). Plots of consumer liking against each of these physical attributes indicate that all these traits were also positively associated with increasing consumer preference (Fig. 7(B)(D)). There were no signicant differences in consumer liking for Nelson compared with Hawkes Bay apples (data not shown; P = 0.6), or for apples harvested in different seasons (data not shown). This is in keeping with the instrumental measurements, where no regional differences were detected in rmness, TSS and titratable acidity. The relationship of increased consumer liking with increasing DMC was similar for both regions. The lack of any regional differences suggests that DMC is a robust predictor of consumer preference. An important component of this study was to evaluate which at-harvest characteristic of Royal Gala best predicted consumer responses after storage. For the orchard blocks in this study, atharvest values ranged from 0.1 to 3.0 L L1 for internal ethylene, 7.38.8 kgf for rmness, 10.712.6% for TSS, 126164 g kg1 for fruit DMC and 1.64.7 for SPI. Background colour did not vary greatly, but this is to be expected given this index is used

commercially to pick fruit selectively. For DMC to be a successful quality index in a commercial context, it has to be sufciently robust to contend with this background variation. It is therefore possible to compare the effectiveness of DMC as an index for predicting eating quality before storage with other more traditional indices such as rmness, TSS and SPI. However, this should be done with care, since the experiment was not designed with this in mind. Correlations of at-harvest characteristics and sensory responses after storage showed that for mean sample liking scores DMC (r2 = 0.58) and SPI (r2 = 0.53) were the two best predictors, with rmness having a lower predictive association (r2 = 0.45). The percentage of consumers that rated the apples as acceptable was best predicted by DMC and SPI, although both r2 values were low to moderate (r2 = 0.41). The willingness to buy was best predicted by DMC (r2 = 0.55), followed by SPI (r2 = 0.46). The co-correlation amongst the different at-harvest variables also means that there is limited value in performing multivariate procedures, such as multiple linear regressions, to determine if combinations of two or more variables explain a higher degree of variation in consumer responses. Overall these results suggest that DMC was the best at-harvest predictor of sensory responses, followed closely by SPI, and that the relationships were linear.

DISCUSSION
Variation in fruit DMC The variation in apple fruit DMC observed in this study was similar to that reported elsewhere. Belgian work showed that fruit DMC of individual Jonagold apples varied within a single year from 134 to 196 g kg1 .23 The range of just over 60 g kg1 is very similar to the ranges recorded for both Royal Gala and Scifresh. Although the mean fruit DMC of Scifresh was higher than that of Royal Gala, the range of fruit DMC across individual apples was similar. This variation in DMC in fruit within commercially picked samples represents a considerable variation in possible consumer acceptability. It must be noted that these fruit samples were from the rst commercial pick. This spread in fruit DMC might be wider

2591

J Sci Food Agric 2010; 90: 25862594

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org

JW Palmer et al.

A 8

100

B a

100

80 a 6 Liking Score b Acceptability % b b

ab Likelihood of Purchase %

80

a 60 b b 40

60

40

20

20

0
Low Moderate High

0
Low Moderate High

0
Low Moderate High

DMC Category

DMC Category

DMC Category

Figure 6. Consumers scores for (A) liking, (B) acceptability and (C) purchase intent for Royal Gala apples from different DMC categories after 1012 weeks of cool storage. Letters show pairwise comparisons after Tukeys HSD adjustment for multiple testing differences.

7 Liking score Liking score

3 100 110 120 130 140 150 160 170 180 190 Fruit dry matter concentration (g kg-1)

3 4 5 6 7 8 9 Flesh firmness (kgf)

C 8

7 Liking score 9 10 11 12 13 14 15 16 Soluble solids (Brix)

Liking score

3 0.0 0.5 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 Titratable acidity (malic acid equivalents g kg-1)

Figure 7. Relationship between consumer liking scores (scale 1 dislike extremely; to 9 like extremely) and fruit DMC (A), fruit rmness (B), TSS (C) and titratable acidity (D) for Royal Gala apples after 1012 weeks of cool storage. The range of each of the four physical measures was divided into 20 equally spaced bins. The mean liking scores of each of these bins are shown plotted against the respective bin mid-points. The points denoted by the star symbol indicate mean liking scores based on ve or fewer fruit. The curves were calculated from a LOESS smoother routine with a span of 0.5 and a degree of 1.

2592

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 25862594

Apple fruit dry matter concentration if later-picked fruit included those from the more shaded regions of older, larger trees, which have a reduced fruit DMC.24 The variation in fruit DMC between samples from different regions appeared to be similar. The tendency for higher fruit DMC to be recorded in Hawkes Bay than in Nelson, particularly for Royal Gala, may suggest a regional difference related to differences in seasonal conditions, or to the small number of sampled orchard blocks. In general, fruit DMC was higher for Scifresh than for Royal Gala, suggesting cultivar differences. In the comparison of eight cultivars, Pink Lady apples were notable for their high fruit DMC (mean of 172 g kg1 ) and Royal Gala apples for their low fruit DMC (mean of 130 g kg1 ). Similar consistent differences in fruit DMC between cultivars have been reported by several authors25,26 and recently a survey of 106 apple cultivars showed a range in esh DMC from 80 to 186 g kg1 .27 Fruit growth modelling of apples normally involves using carbon as the currency, with an unwritten assumption that fruit DMC is constant. The results presented here clearly show that fruit DMC can vary from fruit to fruit and from orchard sample to orchard sample. Additionally, mean fruit DMC of Coxs Orange Pippin apples from the same trees varied from 159 to 206 g kg1 over a 9-year period.28 Similarly fruit DMC of Golden Delicious varied from 139 to 176 g kg1 in two consecutive seasons.29 Relationship between fruit DMC and TSS The results presented here show a clear relationship between fruit DMC recorded at harvest and TSS after storage demonstrated across several cultivars. This therefore conrms and extends earlier work,20 which was done solely with the cultivar Royal Gala. The strong relationship between fruit DMC and TSS, particularly when measured on the same fruit, follows from the high proportion of fruit dry matter contributed by soluble compounds, particularly sugars, in apple fruit. The proportion of dry matter contributed by sugars, starch and organic acids in fresh apples of seven cultivars was reported to range from 71% to 82%, with a mean of 78%.30 A nine-apple cultivar study showed a similar range.31 Most apples are harvested before starch solubilisation is complete, and therefore TSS at harvest may not be a good predictor of TSS after storage, as has been clearly shown by this work. Relationship between fruit DMC and esh rmness The linkage between esh rmness and fruit DMC on these data appears be very cultivar specic. We might hypothesise that fruit with DMC would have more cell wall material and/or a higher osmotic potential and therefore greater esh rmness. Fruit maturity, however, is likely to be a major determinant of esh rmness, so that with both cultivars the relationship between esh rmness at harvest and esh rmness after storage accounted for a greater proportion of the variation in esh rmness after storage than fruit DMC at harvest. Nevertheless, with Scifresh, fruit with high DMC at harvest were associated with fruit with greater esh rmness. This connection between esh rmness and fruit DMC in Royal Gala needs further attention, particularly considering the stronger relationship between these two variates seen in the orchard samples taken for the sensory study in 2008 than in the orchard samples taken in 2005. Consumer studies This study provides evidence of a good relationship between DMC and consumer preference for Royal Gala apples, with the

www.soci.org response being similar for fruit grown in two regions and in two seasons. Given the robust relationship between DMC and consumer preference, it is perhaps useful to speculate what DMC represents in terms of sensory quality. Accumulation of DMC is a fundamental measure of complex physiological and metabolic processes that contribute to the ow of carbon, nitrogen and minerals into fruit.13 This ux of material into the fruit will include (1) compounds that may contribute directly to avour, (2) compounds that contribute to avour as a result of metabolic processes that occur within the fruit and (3) compounds that will be incorporated in the structural elements of the cell associated with texture. Thus a holistic measure such as DMC offers a complex array of potential physicalchemical interactions that individually and in combination contribute to sensory outcomes. For apples, rmness and TSS are sometimes thought of as proxies for texture attributes such as crispness and taste attributes such as sweetness and avour, respectively. However, caution is needed before making these assumptions here, as the avour or texture characteristics in these DMC samples can only be determined by a trained sensory panel, which was not used in the current study. Trained panels are used to describe the individual textural and avour characteristics of fruit, whereas (untrained) consumer panels (as employed in this study) are used to determine preference. At this stage, we suspect that the consumer responses to DMC are not specically driven by TSS, as industry might be tempted to conclude. Rather, we speculate that DMC is a more holistic measure of quality. Dry matter concentration specically represents the biological processes responsible for setting up the textural characteristics, carbohydrate status and avour potential of the fruit. Application of DMC as a new quality predictor It must be emphasised that we are proposing fruit DMC as a new quality metric for apples. It is not to be confused with or thought of as a maturity metric. For both avocado and kiwifruit, fruit DMC has been proposed as both a quality metric and a harvest maturity metric. Just before and during harvest of apple, maturity variates such as background colour, SPI and internal ethylene concentration can change rapidly, but fruit DMC shows only minor change over this period of time.32 Fruit DMC can therefore be measured before or at harvest, and used to predict the sensory potential for the fruit after many months of storage. In contrast, measurements such as rmness and titratable acidity are more dynamic and change appreciably during maturation and storage. The dynamic nature of these latter indices makes it difcult to predict storage responses from their measurement at harvest. However, this does not mean that the traditional harvest indices are redundant, as they are indicators of harvest maturity; instead, DMC can be viewed as a complementary quality index. Fruit DMC can be used to compare and contrast the sensory potential for different orchard blocks before or at harvest, while rmness and other quality indices can be used to monitor the progression of the fruit during the harvest window and in storage, to determine if the sensory potential is realised in the marketplace. For example, a high DMC fruit will only achieve its high sensory potential if it is harvested at the correct stage of maturity and then stored in a manner in which rmness and acidity are optimally conserved. It is very unlikely that high DMC will compensate for poor texture.

CONCLUSIONS
This work has highlighted the variation in fruit DMC that can occur within individual cultivars and also between cultivars,

2593

J Sci Food Agric 2010; 90: 25862594

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org but we believe that fruit DMC will nd its prime usefulness in distinguishing between fruit from the same cultivar, rather than distinguishing between different cultivars. The innate genetic control within individual cultivars of characters such as size, shape, colour, sugaracid balance, avour and crispness will determine the potential for that cultivar to appeal to individual customers, but we believe that high dry matter fruit of any cultivar will prove to be more acceptable to the customer than low dry matter fruit.

JW Palmer et al.

ACKNOWLEDGEMENTS
The authors would like to gratefully acknowledge the major input from our colleagues Murray Oliver, Robert Diack, Daya Dayatilake, Shona Seymour, Robert Henriod, Ken Breen, Amy Paisley, Mark Wohlers, Michelle Beresford and Judith Bowen; help from New Zealand apple growers, particularly Heartland Fruit in Nelson and Fruitpackers Cooperative (Hawkes Bay) Ltd; and funding from Pipfruit New Zealand Inc. and the New Zealand Foundation for Science Research and Technology (Contract No. CO6X0705).

REFERENCES
1 Anon, State of the plate 2005: Study on Americas consumption of fruits and vegetables. Produce for Better Heath Foundation, Wilmington, DE (2005). 2 Harker FR, Gunson FA and Jaeger SR, The case for fruit quality: an interpretive review of consumer attitudes, and preferences for apples. Postharvest Biol Technol 28:333347 (2003). 3 Jaeger SR, Rossiter KL, Wismer WV and Harker FR, Consumer-driven product development in the kiwifruit industry. Food Qual Pref 14:187198 (2003). 4 Harker FR, Kupferman EM, Marin AB, Gunson FA and Triggs CM, Eating quality standards for apples based on consumer preferences. Postharvest Biol Techmol 50:7078 (2008). 5 Herregods M, Postharvest market quality preferences for fruit and vegetables. Acta Hortic 518:207212 (2000). 6 Crisosto CH, Garner D, Crisosto GM and Bowerman E, Increasing Blackamber plum (Prunus salicina Lindell) consumer acceptance. Postharvest Biol Technol 34:237244 (2004). 7 McCluskey JJ, Mittelhammer RC, Marin AB and Wright KS, Effect of quality characteristics on consumers willingness to pay for Gala apples. Can J Agric Econ 55:217231 (2007). 8 Harker FR, Maindonald J, Murray SH, Gunson FA, Hallett IC and Walker SB, Sensory interpretation of instrumental measurements. 1. Texture of apple fruit. Postharvest Biol Technol 24:225239 (2002). 9 Harker FR, Marsh KB, Young H, Murray SH, Gunson FA and Walker SB, Sensory interpretation of instrumental measurements. 2. Sweet and acid taste of apple fruit. Postharvest Biol Technol 24:241250 (2002). 10 Mehinagic E, Royer G, Symoneaux R, Bertrand D and Jourjon F, Prediction of the sensory quality of apples by physical measurements. Postharvest Biol Technol 34:257269 (2004). 11 Molina-Delgado D, Alegrec S, Barreirod P, Valerod C, Ruiz-Altisent M and Recasensa I, Addressing potential sources of variation in several non-destructive techniques for measuring rmness in apples. Biosyst Eng 104:3346 (2009). 12 Hertog MLATM, Nicolai BM, De Ketelaere B, Lammertyn J and De Baerdemaeker J, Non-destructive techniques and quality models for the supply chain: a review. Acta Hortic 768:375384 (2008).

13 Harker FR, Carr BT, Lenjo M, MacRae EA, Wismer WV, Marsh KB, et al, Consumer liking for kiwifruit avour: a meta-analysis of ve studies on fruit quality. Food Qual Pref 20:3041 (2009). 14 Blankenship SM, Parker M and Unrath CR, Use of maturity indices for predicting poststorage rmness of Fuji apples. HortScience 32:909910 (1997). 15 Knee M, Hateld SGS and Smith SM, Evaluation of various indicators of maturity for harvest of apple fruit intended for. J Hortic Sci 64:403411 (1989). 16 Kingston CM, Maturity indices for apple and pear. Hortic Rev 13:407432 (1992). 17 Rowan DD, Hunt MB, Alspach PA, Whitworth CJ and Oraguzie NC, Heritability and genetic and phenotypic correlations of apple (Malus domestica) fruit volatiles in a genetically diverse breeding population. J Agric Food Chem 57:79447952 (2009). 18 Altisent R, Echeverria G, Lara I, Lopez ML and Graell J, Shelf-life of Golden Reinders apples after ultra low oxygen storage: effect on aroma volatile compounds, standard quality parameters, sensory attributes and acceptability. Food Sci Technol Int 15:481493 (2009). 19 Gamble J, Harker FR, Jaeger SR, White A, Bava C, Beresford M, et al, The impact of dry matter, ripeness and internal defects on consumer perceptions of avocado quality and intentions to purchase. Postharvest Biol Technol 57:3543 (2010). 20 McGlone VA, Jordan RB, Seelye R and Clark CJ, Dry matter a better predictor of the post-storage soluble solids in apples? Postharvest Biol Technol 28:431435 (2003). 21 Perring MA, Changes in dry matter concentration gradients in stored apples. J Sci Food Agric 46:439449 (1989). 22 Breslow NE and Clayton DG, Approximate inference in generalized linear mixed models. J Am Statist Assoc 88:925 (1993). 23 Moons E, Sinnaeve G and Dardenne P, Non-destructive visible and NIR spectroscopy measurement for the determination of apple internal quality. Acta Hortic 517:441448 (2000). 24 Haynes RJ and Goh KM, Variation in the nutrient content of leaves and fruit with season and crown position for two apple varieties. Aust J Agric Res 31:739748 (1980). 25 Planchon V, Lateur M, Dupont P and Lognay G, Ascorbic acid levels of Belgian apple genetic resources. Sci Hortic 100:5161 (2004). 26 Lata B, Relationship between apple peel and the whole fruit antioxidant content: year and cultivar variation. J Agric Food Chem 55:663671 (2007). 27 Roen D, Ekholm A and Rumpunen K, Estimating useful diversity in the Norwegian core collection of apples. Acta Hortic 814:131136 (2009). 28 Perring MA, The technique and application of fruit analysis, in Proceedings of the 8th International Colloquium on Plant Analysis and Fertilizer Problems, Auckland, New Zealand NZ DSIR Information Series No 134, ed. by Ferguson AR, Bieleski RL and Ferguson IB. Government Printer, Wellington, pp. 375382 (1978). 29 Marguery P and Sangwan BS, Sources of variation between apple fruits within a season, and between seasons. J Hortic Sci 68:309315 (1993). 30 Suni M, Nyman M, Eriksson NA, Bjork L and Bjorck I, Carbohydrate composition and content of organic acids in fresh and stored apples. J Sci Food Agric 80:15381544 (2000). 31 Salo ML and Korhonen I, Carbohydrate and acid composition of some apple varieties. J Sci Agric Soc Finl 44:6367 (1972). 32 Bowen JH and Watkins CB, Fruit maturity, carbohydrate and mineral content relationships with watercore in Fuji apples. Postharvest Biol Technol 11:3138 (1997).

2594
wileyonlinelibrary.com/jsfa
c 2010 Society of Chemical Industry J Sci Food Agric 2010; 90: 25862594

Research Article
Received: 13 April 2010 Revised: 13 July 2010 Accepted: 15 July 2010 Published online in Wiley Online Library: 17 August 2010

(wileyonlinelibrary.com) DOI 10.1002/jsfa.4126

Characterization and discrimination of premium-quality, waxy, and black-pigmented rice based on odor-active compounds
Dong Sik Yang,a,b Kyu-Seong Leec and Stanley J. Kaysa
Abstract
BACKGROUND: Odor-active compounds have been studied in cooked aromatic rice, but not in specialty rice types that have distinctly different avors. We analyzed the odor-active compounds emanating from three different types of specialty rice (premium-quality, waxy and black-pigmented) and identied the differences in odor-active compounds among them. RESULTS: Twenty-one, 21 and 23 odorants were detected using GC-O for cooked samples of premium-quality, waxy and blackpigmented rice cultivars, respectively. Hexanal was the main odorant in premium-quality and waxy cultivars; however, waxy cultivars had 16 times higher hexanal odor activity values (OAVs) than premium-quality cultivars, indicating premium-quality rice had a less pronounced overall aroma. 2-Acetyl-1-pyrroline was the main contributor to overall aroma in black-pigmented rice, followed by guaiacol. The three types of specialty rice were clearly discriminated based on the OAVs of their odor-active compounds using multivariate analyses. Six odor-active compounds [2-acetyl-1-pyrroline, guaiacol, hexanal, (E)-2-nonenal, octanal and heptanal] contributed the most in discriminating the three types of specialty rice. Six very similar superior cultivars of premium rice could likewise be readily separated using aroma chemistry. CONCLUSION: The ability to discriminate the aroma among rice types using the OAVs of the principal odor-active compounds facilitates our understanding of the aroma chemistry of specialty rice. c 2010 Society of Chemical Industry Keywords: specialty rice; hierarchical cluster analysis (HCA); principal component analysis (PCA); GC-O; plant breeding; selection criteria

INTRODUCTION
Specialty rice is a term used to distinguish cultivars of rice that have unique properties (e.g. avor, color, nutrition, chemical composition). The demand has been increasing in recent years for various types of specialty rice, which are sold for as much as 50% more than traditional rice cultivars.1 They are widely grown in India, Pakistan and Thailand, and are popular in Asia, the Middle East, Europe, and the United States.2 Examples of types of specialty rice are: aromatic (unique aromas); premium (glossiness, stickiness, and smooth texture); black (unique color and avor); and waxy (very low amylose content and superior processing quality).1,3,4 New rice cultivars are developed in breeding programs where the progeny are screened for a wide range of quantitative and qualitative traits. The criteria used in evaluating rice varies among programs, reecting differences in preferences among consumer populations. Many traits can be fairly readily and accurately quantied. For example, traits such as milling quality, grain size, shape, color and appearance, and cooking characteristics (e.g. gelatinization temperature, amylose content, grain elongation, aroma), are commonly used in making progeny selection decisions.5,6 Assessment of other traits (e.g. avor, insect resistance) tend to be much more difcult and are often more subjective. Flavor, for example, requires the use of sensory panels

that typically require 1415 individuals and generally only ve to six samples can be accurately tested at one setting. As a consequence, the time, cost, and subjectivity of traditional sensory analysis tremendously limit the number of progeny that can be assessed. The net effect is that, out of necessity, avor is generally relegated to one of the last traits in the selection hierarchy, greatly decreasing the rate at which improvement can be made. If sensory analysis could be replaced in initial progeny screening with a rapid, accurate analytical technique for the measurement of avor, large numbers of progeny could be screened, greatly facilitating the development of superior new cultivars.7

Correspondence to: Stanley J. Kays, 1111 Plant Science Building, Department of Horticulture, University of Georgia, Athens, GA 30602, USA. E-mail: kaysstan@uga.edu

a 1111 Plant Science Building, Department of Horticulture, University of Georgia, Athens, GA 30602, USA b Plant Biology Division, Samuel Roberts Noble Foundation, Ardmore, OK 73401, USA

2595

c Green Growth & Future Strategy Team, RDA, Suwon 441-707, Republic of Korea

J Sci Food Agric 2010; 90: 25952601

www.soci.org

c 2010 Society of Chemical Industry

www.soci.org In rice, aroma is the primary contributor to the overall avor. As a consequence, considerable interest has been focused on the identication and quantication of the volatile compounds emanating from cooked rice.8 Over 300 volatiles have been reported using a cross-section of isolation techniques.9 13 Dynamic headspace trapping onto Tenax TA (2,6-diphenylene oxide-based porous polymer), in particular, is widely used for studying organic volatiles in air, food and plants due to its ability to adsorb and desorb a wide range of compounds.14 Tenax TA, which has a low afnity for water, is used for collecting volatiles from high moisture samples such as cooked foods. A dynamic headspace system using a Tenax trap in conjunction with short path thermal desorption can detect volatile organic compounds in the ppb range.15 Among the volatiles identied in cooked rice, approximately 20 odor-active compounds (e.g. 2-acetyl-1-pyrroline, hexanal, (E)2-nonenal, 2-amino acetophenone) are known to contribute to the unique avor of a cross-section of rice cultivars.10,13,16 For example, the popcorn-like odor in aromatic rice is conferred by 2-acetyl-1-pyrroline, while hexanal is thought to be associated with the development of a rancid odor in aged rice.9 2Amino acetophenone, conferring medicinal and phenolic odors, contributes to the overall aroma of cooked brown rice.16 Recently the odor-active compounds in black-pigmented rice were determined and 2-acetyl-1-pyrroline and guaiacol were found to be the primary components of its unique aroma.4 In addition, differences in the unique aromas of basmati, jasmine, blackpigmented, and other rice avor types were explained using 13 critical odor-active compounds.17 Characterization of odor-active compounds, however, has been mainly limited to aromatic rice, even though other specialty rices have distinctly different avors. A better understanding of the aroma chemistry of speciality rice will facilitate the development of superior new cultivars in rice breeding programs focusing on avor. As an initial step in developing an analytical method for screening rice progeny for avor, we characterized the odor-active compounds emanating from three different types of specialty rice (premium-quality, waxy and black-pigmented). Differences in odor-active compounds affecting avor quality among the main specialty rice types were assessed using hierarchical cluster analysis (HCA) and principal component analysis (PCA).

DS Yang, K-S Lee, SJ Kays the rice was sealed in glass containers and held at 20 C until analysis.

Chemicals Analytical standards utilized for identication and quantication were: benzaldehyde, decane, decanal, (E)-2-decenal, guaiacol, heptanal, heptane, 2-heptanone, nonane, nonanal, octane, pentadecane, 1-pentanol, tridecane, tetradecane, undecane (SigmaAldrich Inc., St Louis, MO, USA); (E,E)-2,4-decadienal, hexanal, (E)-2-hexenal, 2-nonanone, (E)-2-octenal, 1-octen-3-ol (Aldrich Chem. Co., Milwaukee, USA); p-menthan-3-one, naphthalene, octanal, p-xylene (Fluka Chem. Co., St. Louis, MO, USA); dodecane, (E,E)-2,4-nonadienal, (E)-2-nonenal and 2-pentylfuran (TCI America, Portland, OR, USA); 3-octen-2-one (Alfa Aesar, Ward Hill, MA, USA). 2,4,6-Trimethylpyridine (TMP) was purchased from Aldrich Chemical Co. and is routinely used as a internal standard to quantify the concentration of 2-acetyl-1-pyrroline.

Preparation of cooked rice Rice samples (100 g) were cooked in distilled water (150 mL) for 30 min at 100 C in a specially constructed 1 L glass beaker with entry and exit ports. The entry and exit ports were tightly sealed with aluminum foil during cooking. Due to differences in water absorption, 100 mL of distilled water was added to black rice samples (100 g) and similarly cooked for 30 min as described previously by Yang et al.18 Three replications of each cultivar were analyzed.

MATERIALS AND METHODS

Materials Ten cultivars including six premium-quality [Oryza sativa L. japonica cv. Hwaseongbyeo (P1), Ilpumbyeo (P2), Gopumbyeo (P3), Taebongbyeo (P4), Chucheongbyeo (P5), and Samkwangbyeo (P6)], two waxy cultivars [O. sativa L. japonica cv. Hwaseonchalbyeo (W1), and O. sativa L. indica cv. Hangangchalbyeo (W2)], and two black-pigmented [O. sativa L. japonica cv. Heugjinjubyeo (B1), and Heugkwangbyeo (B2)] were grown at the National Institute of Crop Science, Suwon, South Korea in 2006. Rice transplants of each cultivar were planted on 24 May. Heugjinjubyeo and Taebongbyeo were harvested on 22 September; Heugkwangbyeo, Hangangchalbyeo, Hwaseonchalbyeo and Hwaseongbyeo on 3 October; Gopumbyeo, Ilpumbyeo, Chucheongbyeo and Samkwangbyeo on 22 October. N, P, and K fertilizers were applied at the rate of 9.0 105 , 4.5 105 , and 5.7 105 kg m2 , respectively. During growth, the average temperature was 21.7 3.6 C and there was 1150 mm of precipitation. After harvest and milling,

Dynamic headspace using Tenax trap Isolation of volatile compounds emanating from cooked rice was performed using an all-glass dynamic headspace system described previously.4,17 The glass beaker with freshly cooked rice was immediately placed in a hot water bath and maintained at 70 C during sampling. Headspace volatiles emanating from cooked rice samples were collected using a Tenax trap and a vacuum sampling pump (Aircheck Sampler, Model 224-44XR; SKC Inc., Eighty Four, PA, USA) that swept volatiles from the sample headspace. Incoming air was puried by passing it through a charcoal lter (Alltech Assoc. Inc., Deereld, IL, USA; 1 cm i.d., 10 cm long Pyrex glass tube with a 7 cm bed of charcoal) connected to the entry port of the glass beaker at a rate of 150 mL min1 for 60 min. A 50 mL Erlenmeyer ask was placed between the exit port and the trap to collect any condensation. One milliliter of 18.34 mg L1 TMP solution in 0.1 mol L1 HCl was injected into the Erlenmeyer ask at the beginning of volatile collection. TMP was chosen as an internal standard for quantication of 2-AP due to similar properties with 2-AP (e.g. basic, water solubility, volatility, stability, retention time).19 After sampling, the Tenax trap was connected to an automated short path thermal desorption system (Model TD-5; Scientic Instrument Services, Ringoes, NJ, USA) on the injector port of the gas chromatograph/mass spectrometer (GC/MS, Model 6890N/5973; Agilent, Wilmington, DE, USA). The volatiles were desorbed at 250 C for 5 min with He at a ow rate of 10 mL min1 and the analytes collected on the rst 4 cm of the GC column using a CO2 cooled cryofocus trap (40 C) (SIS 2 Cryo-Trap; Scientic Instrument Services). After desorption, the cryofocus trap was rapidly heated to 200 C and condensed analytes were separated utilizing GC temperature programming.

2596

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 25952601

Aroma of premium-quality, waxy and black-pigmented rice GC/MS and GC-Olfactometry Desorbed odorants were split between the mass spectrometer for identication and quantication and an Olfactory Detector Outlet (ODO II; SGE Intl., Austin, TX, USA) used to describe the individual odorants and to assess their intensity. The GC was equipped with a 30 m length, 0.25 mm i.d., 0.25 m lm thickness (5% phenyl methyl silicone) fused silica capillary column (HP-5 MS; Agilent). The injection port temperature was 225 C with a split ratio of 0.5 : 1. Helium was used as the carrier gas with a ow rate of 2.0 mL min1 . The column temperature was programmed to hold at 40 C for 1 min and then increase 1.5 C min1 to 65 C for 1 min, then 2 C min1 to 120 C for 1 min, and nally 15 C min1 to 280 C for 5 min. MS conditions were: ion source 230 C; electron energy 70 eV; multiplier voltage 1247 V; a transfer line 280 C; and a scan range of 35350 atomic mass units. GC-O analysis was made by three assessors who had considerable previous experience in sensory tests. Acceptance of an odorant required substantiation by at least two of the three assessors with each individual assessing replicated samples of the cultivars.17 The assessors were trained by describing 15 materials with different odors: popcorn-like (popcorn), starchy (rice starch), woody (toothpicks), cooked grain (cream of wheat), corn (cream style corn), nutty (roasted peanut), oral (jasmine scent), dairy (2% milk), hay (hay), barn (white pepper), buttery (butter), green (alfalfa), rancid (vegetable oil), waxy (candle) and earthy (mushroom) odors. An aroma extract of the sample was characterized by describing the aroma of the individual components and assessing their odor intensity on a scale of 1 to 5, where 1 = very weak; 2 = weak; 3 = intermediate; 4 = strong; 5 = very strong. Odorants perceived by all three assessors were considered odor-active compounds (odor description and intensity data are not shown). Identication and quantication of odorants Odorants were identied based on their mass spectra using NIST 02 and Wiley 7 libraries. Identication was conrmed using Kovats retention indices (RI) and odor descriptors with authentic standards. RIs were determined using a non-polar HP-5 MS column and a series of n-hydrocarbons (C7 C15 ) and compared with those reported previously and at http://webbook.nist.gov/chemistry/name-ser.html/. In the absence of an authentic standard, the identication of 2-acetyl1-pyrroline (2-AP) was conrmed by mass spectra, RI and its distinctive popcorn odor. The concentration of each odorant was quantied using standard curves for each compound in hexane (5, 10, 20, 50, 100, 200, and 500 L1 ) and three replications of 1 L of each standard solution was injected directly into the GC-MS using a microsyringe. Linearity, sensitivity, and precision of each compound were shown in Yang et al.4 2-AP was expressed as TMP equivalents. Odor thresholds in air and odor activity values Odor thresholds in air were determined using a modied Ulrich and Grosch GC-O method by the assessors to ascertain the impact of individual odorants.20 A solution of each odorant (10 mg) was prepared in 10 mL of hexane and 0.5 L of the solution was injected into the GC and sniffed using the olfactory detector outlet. The solution was diluted stepwise (1 : 1, v/v) until the odorant was not detected (minimum detectable concentration). The odor threshold in air for each odorant was calculated relative to minimum detectable concentration of (E)-2-decenal in air (2.7 ng L1 ). The odor threshold for 2-AP was reported

www.soci.org previously.21 Odor activity values (OAVs) were calculated by dividing the emission concentration of each odorant in a sample over a constant collection period by its odor threshold in air. Statistical data analysis Hierarchical cluster analysis (HCA) and principal component analysis (PCA) were carried out using the SAS system for Windows v8 with three replications of each cultivar (six premium quality, two waxy and two black-pigmented cultivars). HCA was performed using Ward minimum variance method to determine if the OAVs of the odor-active compounds in each specialty rice cultivars were able to segregate the rice samples. PCA was performed using the covariance matrix obtained from the data matrix (i.e. OAVs of the primary odor-active compounds) to determine which odor-active compounds had the highest loadings and contributed the most to differences in aroma among the samples.

RESULTS AND DISCUSSION

Odor-active compounds and their odor activity values Among the premium-quality cultivars (Hwaseongbyeo, Ilpumbyeo, Gopumbyeo, Taebongbyeo, Chucheongbyeo and Samkwangbyeo) there were 21 odorants identied that included 12 aldehydes, two alcohols, three ketones, and four aromatic compounds (Table 1). Based on the OAV, hexanal was the dominant odorant, followed by (E)-2-nonenal, octanal, heptanal, (E,E)-2,4nonadienal and nonanal in all six cultivars. Collectively, these compounds accounted for over 95% of the total OAV and were key aroma compounds in premium rice. However, these cultivars had lower OAVs than waxy and black-pigmented rice cultivars, indicating that the aroma of the premium cultivars was less pronounced. The six premium-quality cultivars were selected in breeding programs based on six primary quality traits: (1) grain length/width ratio prior to milling (1.72.0); (2) appearance (size 172 g per 1000 milled grains), translucency, gloss and very low chalkiness; (3) physicochemical properties (amylose content below 20%, protein content of brown rice 79%, high alkali digestion value); (4) appearance and palatability when cooked (glossy appearance, intact grains, weakly aromatic fragrance, sticky, good taste and texture); (5) slow retrogradation of cooked rice; and (6) milling recovery from rough to polished rice of above 75% and a head rice ratio of above 90%.3 With regard to the palatability of cooked Korean premium-quality rice, our results indicated that the six premium-qualtiy cultivars have a less pronounced overall aroma that corresponds to the preference of Korean consumers.3 Twenty-one odorants were identied in the waxy rice cultivar Hwaseonchalbyeo (W1) and 20 in Hangangchalbyeo (W2) that included 12 aldehydes, one alcohol, three ketones, and four aromatic compounds. 1-Pentanol was present only in Hwaseonchalbyeo (Table 1). Hexanal in Hwaseonchalbyeo (316), a Japonica type and Hangangchalbyeo (448), an Indica type, had the highest OAVs, 8.8 and 12.4 times higher than that in Samkwangbyeo (P6) which had the highest OAV in the premium-quality cultivars (Table 1). The very high hexanal concentration appears to be due to the higher lipid content in waxy rice and its subsequent degradation resulting in the formation of hexanal.22 Hexanal, a linoleic acid oxidation product, had a higher concentration in broken rice than in head rice due to greater surface lipids and free fatty acids. The concentration of hexanal in broken rice increased signicantly during storage, resulting in the development of off-avor.23 The odor of hexanal is described as green or green tomato

2597

J Sci Food Agric 2010; 90: 25952601

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

2598
Premium-quality riceb P2 ND 447.53 9.39 0.30 0.00 0.03 0.00 2.29 0.01 7.35 0.03 ND 0.74 0.03 ND 2.73 0.38 0.53 0.05 6.80 0.13 0.27 0.03 1.24 0.71 ND 0.01 0.00 2.26 0.05 0.33 0.04 10.42 1.72 0.02 0.01 tr 0.33 0.06 2.38 0.84 0.13 0.02 0.40 0.01 0.22 0.06 0.34 0.11 13.38 1.04 0.03 0.00 tr 0.22 0.04 1.86 0.50 0.11 0.02 0.35 0.02 ND 0.01 0.01 0.51 0.06 0.15 0.00 3.87 0.12 0.05 0.03 0.60 0.34 1.12 0.63 0.01 0.00 2.37 0.26 0.03 0.00 5.46 1.08 0.04 0.01 ND 0.27 0.03 1.64 0.07 0.07 0.01 0.14 0.03 ND 0.05 0.02 0.24 0.19 0.10 0.02 5.02 1.18 0.08 0.04 0.56 0.12 1.07 0.20 ND 2.94 0.68 0.10 0.09 7.29 1.82 0.04 0.01 ND 0.38 0.07 1.79 0.04 0.09 0.04 0.23 0.08 P3 P4 P5 P6 W1 W2 B1 B2 Waxy ricec Black-pigmented rice d OTg 0.01 0.00 0.01 0.00 153i 88.52 3.88 36.61 8.42 1.1i 0.10 0.01 0.11 0.03 3.1i 0.02 0.00 0.06 0.01 46.9 0.68 0.00 0.93 0.00 3.5i 2.73 0.07 4.00 0.78 0.9i 569.24 94.96 285.60 114.74 0.02j 13.4 85i 2.7i 19i 0.4i 6.7i 2.7i 1.5i 31i 2.6i 4.7i 0.09i 33 202 2.6i 0.2i 2.7i 2.3i 0.07 0.06 ND 0.13 0.07 0.05 0.01 2.22 0.81 0.01 0.00 0.24 0.09 ND tr 1.14 0.41 0.02 0.02 3.58 1.24 0.03 0.01 ND 0.15 0.07 0.80 0.10 0.07 0.01 0.16 0.06 0.09 0.03 0.25 0.10 0.10 0.01 0.36 0.04 0.09 0.02 0.30 0.02 4.51 1.47 0.03 0.01 ND 0.16 0.06 1.27 0.19 5.77 0.83 0.03 0.01 ND 0.23 0.08 1.23 0.11 4.66 0.53 0.03 0.00 ND 0.16 0.02 1.04 0.06 6.37 1.25 0.03 0.01 ND 0.26 0.06 1.22 0.05 0.39 0.05 ND tr 1.04 0.35 0.04 0.03 0.42 0.08 ND tr 1.68 0.40 0.01 0.01 0.33 0.01 ND tr 1.02 0.10 0.04 0.03 0.42 0.08 ND tr 1.47 0.50 0.04 0.03 1.55 0.34 ND 0.01 0.00 3.48 0.91 0.03 0.01 ND 0.21 0.05 0.08 0.01 2.16 0.43 0.04 0.02 ND 0.27 0.07 0.09 0.03 3.30 0.82 0.05 0.01 ND 0.20 0.00 0.08 0.00 2.18 0.03 0.01 0.00 ND 0.23 0.03 0.10 0.02 3.13 0.97 0.02 0.01 ND 1.19 0.28 0.30 0.08 11.24 2.67 0.14 0.04 0.15 0.04 0.17 0.04 0.14 0.01 0.19 0.03 0.52 0.10

Table 1. Odor activity value (OAV)a of odor-active compounds in six premium-quality, two waxy, and two black-pigmented rice cultivars

wileyonlinelibrary.com/jsfa

No. Odorant

RIe IDf

P1

1 2 3 4 5 6 7

766 803 857 859 895 903 918

A 0.01 0.00 trh 0.01 0.00 0.01 0.00 0.01 0.00 tr 0.02 0.00 A 15.53 4.79 13.43 5.55 22.49 2.08 31.70 8.30 26.67 0.30 35.26 12.15 315.52 4.37 A 0.05 0.01 0.03 0.02 0.05 0.01 0.05 0.01 0.06 0.01 0.04 0.03 0.15 0.08 A 0.01 0.00 0.01 0.00 0.01 0.00 0.01 0.00 0.01 0.00 0.01 0.00 0.01 0.00 A 0.16 0.03 0.10 0.10 0.20 0.03 0.25 0.05 0.23 0.03 0.24 0.05 1.14 0.26 A 1.08 0.23 1.37 0.49 1.50 0.20 1.82 0.44 1.43 0.02 2.14 0.48 5.22 1.22 B ND ND ND ND ND ND ND

952 A

0.09 0.05

9 10 11 12 13

954 984 992 1005 1036

A A A A A

ND 0.14 0.05 0.07 0.01 1.52 0.23 0.02 0.01

www.soci.org

c 2010 Society of Chemical Industry

14 15 16 17 18

1058 1086 1093 1106 1152

A A A A A

0.31 0.10 ND tr 0.65 0.18 0.02 0.01

19 20 21 22 23

1160 1172 1200 1206 1212

A A A A A

3.32 0.58 0.02 0.01 ND 0.12 0.04 0.89 0.12

24 25

1-Pentanol Hexanal (E)-2-Hexenal p-Xylene 2-Heptanone Heptanal 2-Acetyl-1pyrroline (E)-2-Heptanal Benzaldehyde 1-Octen-3-ol 2-Pentylfuran Octanal 3-Octen-2one (E)-2-Octenal Guaiacol 2-Nonanone Nonanal p-Menthan-3one (E)-2-Nonenal Naphthalene Dodecane Decanal (E,E)-2,4Nonadienal (E)-2-Decenal (E,E)-2,4Decadienal

1262 A 1315 A

0.07 0.02 0.24 0.07

J Sci Food Agric 2010; 90: 25952601

OAVs are given as the mean standard deviation of three replications of each cultivar. P1, Hwaseongbyeo; P2, Ilpumbyeo; P3, Gopumbyeo; P4, Taebongbyeo; P5, Chucheongbyeo; P6, Samkwangbyeo. c B1, Heugjinjubyeo; B2, Heugkwangbyeo. d W1, Hwaseonchalbyeo; W2, Hangangchalbyeo. e Retention index based on HP-5 MS column using a series of n-hydrocarbons. f Method of identication: A, compounds were identied by comparison of mass spectrum and RI of an perceived compound with those of an authentic compound.; B, compounds were tentatively identied by comparison of mass spectrum, odor description, and RI with those from the literature when reference compounds were not available. g OT, odor threshold in air (ng L1 ). h tr, trace value (<0.01); ND, not detected. i Data from Yang et al.17 j Data from Schieberle.21

DS Yang, K-S Lee, SJ Kays

Aroma of premium-quality, waxy and black-pigmented rice and has been associated with rancidity and consumer rejection of rice.9 Monsoor and Proctor23 suggested that water washing of rice grains can reduce the off-avor products including hexanal. In addition to hexanal, (E)-2-nonenal, octanal, heptanal, nonanal, (E,E)-2,4-nonadienal, (E)-2-octenal, 1-octen-3-ol, and 2heptanone had OAVs greater than 1 (Table 1). OAVs of octanal in Hwaseonchalbyeo and heptanal in Hanganchalbyeo were 4.7 and 4.6 times higher than the average OAVs of octanal and heptanal in the six premium-quality cultivars, respectively. Twenty-three odorants were identied in the black-pigmented cultivars Heugjinjubyeo (B1) and 22 in Heugkwangbyeo (B2) (Table 1). Eleven aldehydes, two alcohols, two ketones, six aromatics and one nitrogen-containing compound were found in all cultivars. 2-Nonanone was present in Heugjinjubyeo but not Heugkwangbyeo. In Heugjinjubyeo and Heugkwangbyeo, 2acetyl-1-pyrroline (2-AP) had the highest OAV (e.g. 569 and 286, respectively) followed by hexanal, (E)-2-nonenal, octanal, heptanal, nonanal, (E,E)-2,4-nonadienal and guaiacol. 2-AP, which confers a popcorn-like odor, was a signicant component in the aroma of the cultivar. The concentration of 2-AP, used as an indicator of aromatic rice during selection in some breeding programs, varied substantially between the two black rice cultivars from a low of 5.71 ng g1 in Heugnambyeo to a high of 11.38 ng g1 in Heugjinjubyeo (data not shown). According to a previous report, there was a wide variation in 2-AP concentration among aromatic cultivars.24 Guaiacol was detected only in black rice where it is a key aroma compound that confers a smoked odor and has a low odor threshold.4 It had an OAV of 1.12 and 1.07 in Heugjinjubyeo, and Heugkwangbyeo, respectively. Even though hexanal had high OAVs, guaiacol and 2-AP contributed to the unique aroma of black rice due to their unique odors and low odor threshold values. Rice aroma is affected by genetic, preharvest and postharvest (e.g. drainage and harvest dates, harvest moisture content, degree of milling, storage conditions, washing, soaking, cooking method) factors.25 The different aroma patterns of cultivars among each rice type (premium-quality, waxy, and black-pigmented rice types) appear to be predominantly due to genetic differences since all cultivars were cultivated, harvested, and stored the same. Genetic differences in existing compounds (e.g. 2-AP) and those synthesized during cooking from the basic chemical components of rice (e.g. lipids, proteins, starch) largely dictated the differences among cultivars in the volatiles released from cooked rice.18 HCA and PCA of the odor-active compounds in specialty rice The dendogram obtained by HCA demonstrated distinct separation of three different classications in avor among samples that coincided with the three types of specialty rice (Fig. 1): cluster 1 (premium-quality rice; P1 to P6), cluster 2 (waxy rice; W1 and W2), and cluster 3 (black-pigmneted rice; B1 and B2). Premium-quality (P1 to P6), waxy (W1 and W2), and black-pigmented (B1 and B3) types were clearly separated with the rst (PC 1) and second (PC 2) principal components accounting for 62.8% and 37.2% of the total variance, respectively (Fig. 2), conrming the clustering obtained from HCA. The black-pigmented cultivars positioned on the positive side of PC 1, were clearly separated from the premiumquality and waxy cultivars positioned on the negative side of PC 1. 2-Acetyl-1-pyrroline (7) was the most important contributor inuencing the separation of black-pigmented cultivars from premium-quality and waxy cultivars, followed by guaiacol (15). The premium-quality cultivars positioned on the negative side of PC 2 were clearly separated from the waxy cultivars positioned on the positive side of PC 2. The main compounds inuencing

www.soci.org the differences on PC 2 were, in decreasing order of contribution, hexanal (2), (E)-2-nonenal (19), octanal (12) and heptanal ( 6). Potential for using aroma chemistry to make progeny selection decisions in breeding programs The potential for using aroma chemistry for making progeny selection decisions in rice breeding programs is predicated on being able to screen large numbers of progeny and accurately identify those with desirable characteristics. The six cultivars of premium rice in the test that were very uniform in their overall avor as indicated by their very tight clustering in Fig. 2, were readily separated using HCA (Fig. 3). HCA indicated three different avor groups within the cultivars: cluster 1 (P1 and P2), cluster 2 (P3 and P5), and cluster 3 (P4 and P6) (Fig. 3) with cluster 1 being more distant than clusters 2 and 3. Figure 4 illustrates the pattern of odor-active compounds for the six cultivars with PC 1, accounting for 99.6% of the total variance. The separations were matched to clusters obtained from HCA. The main compound inuencing PC 1 was hexanal (2). Based on PC 2, the groups were separated by heptanal (6), octanal (12), nonanal (17), and (E)-2-nonenal (19); however, PC 2 explained only 0.3% of total variance, indicating their contribution is very weak. The method allowed precise separation of the premium-quality rice cultivars into three groups, which was conrmed by HCA and PCA. It should be noted that in a breeding program, one of the primary advantages of using an analytical method for making selection decisions is to allow screening large numbers of progeny for avor, in contrast to the very restricted number that can be currently assessed using sensory analysis. The cultivars tested represent superior lines with very desirable avors. In a breeding program, the range in avor chemistry among progeny will be far greater than the cultivars tested herein, and therefore, less favorable lines will be separated much more distinctly from the existing cultivars. Screening a far greater number of progeny to identify desirable lines facilitates the development of new cultivars with superior avor. Selection decisions are also made using other analytical assessments methods (e.g. texture, stickiness, amylose to amylopectin ratio) that are currently widely used. An analytical method for assessing avor does not totally eliminate the use of sensory analysis; however, it allows testing only a small number of advanced lines prior to nal acceptance of potential new cultivars.

CONCLUSIONS
The OAV of odor-active compounds was useful in classifying the ten cultivars into three groups (premium-quality, waxy, and black-pigmented) (Figs. 1 and 2). Six odor-active compounds [2acetyl-1-pyrroline, guaiacol, hexanal, (E)-2-nonenal, octanal, and heptanal] contributed the most in distinguishing differences in aroma among the rice types. The ability to discriminate the aroma among rice types using the OAVs of odor-active compounds will facilitate dening the avor chemistry contributing to differences in consumer avor preference. The methods employed also demonstrated distinct differences within rice types (i.e. premium). If sufciently precise separation can be made among progeny within a type of specialty rice, the analytical method described for assessing the aroma component of avor could potentially be used for making progeny selection decisions in rice breeding programs, characterizing the chemistry of preference of targeted consumer populations and greatly accelerating the development of superior new cultivars.

2599

J Sci Food Agric 2010; 90: 25952601

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org

DS Yang, K-S Lee, SJ Kays

Figure 1. HCA dendrogram of ten cooked specialty rice cultivars based on their average odor active values (OAVs) (n = 3): premium-quality rice (P1-P6), waxy rice (W1 and W2), and black-pigmented rice (B1 and B2).

Figure 2. PCA biplot of odor-active compounds in cooked specialty rices based on their average odor active values (OAVs) (n = 3). Numbers correspond to compounds listed in Table 1.

Figure 3. HCA dendrogram of six cooked premium-quality rice cultivars based on their average odor active values (OAVs) (n = 3): Hwaseongbyeo (P1), Ilpumbyeo (P2), Gopumbyeo (P3), Taebongbyeo (P4), Chucheongbyeo (P5), Samkwangbyeo (P6).

2600

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 25952601

Aroma of premium-quality, waxy and black-pigmented rice

www.soci.org

Figure 4. PCA biplot of odor-active compounds in cooked premium-quality rices based on their average odor active values (OAVs) (n = 3). Numbers correspond to compounds listed in Table 1.

ACKNOWLEDGEMENTS
The authors wish to express their appreciation to Drs H.P. Moon and S.K. Hahn for their contributions to our rice avor chemistry research program.

REFERENCES
1 Chaudhary RC, Speciality rices of the world: effect of WTO and IPR on its production trend and marketing. J Food Agric Environ 1:3441 (2003). 2 Singh RK, Singh US and Khush GS, Prologue, in Aromatic Rices, ed. by Singh RK, Singh US and Khush GS. Science Publishers, Eneld, NH, pp. 13 (2000). 3 Choi HC, Current status and perspectives in varietal improvement of rice cultivars for high-quality and value-added products. Kor J Crop Sci 47:1532 (2000). 4 Yang DS, Lee KS, Jeong OY, Kim KJ and Kays SJ, Characterization of volatile aroma compounds in cooked black rice. J Agric Food Chem 56:235240 (2008). 5 Abdel-Aal ESM, Young JC and Rabalski I, Anthocyanin composition in black, blue, pink, purple, and red cereal grains. J Agric Food Chem 54:46964704 (2006). 6 Dela Cruz N and Khush GS, Rice grain quality evaluation procedures, in Aromatic Rices, ed. by Singh RK, Singh US and Khush GS. Science Publishers, Eneld, NH, pp. 1528 (2000). 7 Yang DS, Lee KS, Jeong OY, Kim KJ and Kays SJ, Fingerprinting rice avor, in Improving Human Health Through Biofortied Rice. International Symposium on Rice Fortication. National Institute of Crop Science, Suwon, S. Korea, pp. 7195 (2006). 8 Champagne E and Bett-Garber K, Challenges of measuring rice aroma and avor, in Proceedings of the United StatesJapan UJNR CooperativePrograminNaturalResources,FoodandAgriculturePanel, 36th Annual Meeting, National Food Research Institute, Tsukuba, Japan, pp. 217222 (2007). 9 Bergman CJ, Delgado JT, Bryant R, Grimm C, Cadwallader KR and Webb BD, Rapid gas chromatographic technique for quantifying 2-acetyl-1-pyrroline and hexanal in rice (Oryza sativa L.). Cereal Chem 77:454458 (2000). 10 Widjaja R, Craske JD and Wootton M, Comparative studies on volatile components of non-fragrant and fragrant rices. J Sci Food Agric 70:151161 (1996).

11 Tava A and Bocchi S, Aroma of cooked rice (Oryza sativa): comparison between commercial Basmati and Italian line B5-3. Cereal Chem 76:526529 (1999). 12 Grimm CC, Bergman C, Delgado JT and Bryant R, Screening for 2acetyl-1-pyrroline in the headspace of rice using SPME/GC-MS. J Agric Food Chem 49:245249 (2001). 13 Buttery RG, Turnbaugh JG and Ling LC, Contribution of volatiles to rice aroma. J Agric Food Chem 36:10061009 (1988). 14 Reineccius G, Flavor Chemistry and Technology, 2nd edition. Taylor and Francis, Boca Raton, FL (2006). 15 Peng CY and Batterman S, Performance evaluation of a sorbent tube sampling method using short path thermal desorption for volatile organic compounds. J Environ Monit 2:313324 (2000). 16 Jezussek M, Juliano BO and Schieberle P, Comparison of key aroma compounds in cooked brown rice varieties based on aroma extract dilution analysis. J Agric Food Chem 50:11011105 (2002). 17 Yang DS, Shewfelt RL, Lee KS and Kays SJ, Comparison of odor-active compounds from six distinctly different rice avor types. J Agric Food Chem 56:27802787 (2008). 18 Yang DS, Lee KS, Kim KJ and Kays SJ, Site of origin of volatile compounds in cooked rice. Cereal Chem 85:591598 (2008). 19 Buttery RG, Ling LC and Thomas RM, Quantitative analysis of 2-acetyl1-pyrroline in rice. J Agric Food Chem 34:112114 (1986). 20 Ullrich F and Grosch W, Identication of the most intense volatile avour compounds formed during autoxidation of linoleic acid. Zeitschrift f r Lebensmitteluntersuchung und -Forschung A u 184:277282 (1987). 21 Schieberle P, Primary odorants in popcorn. J Agric Food Chem 39:11411144 (1991). 22 Shin HS and Rhee JY, Comparative studies on the lipid content and neutral lipid composition in nonglutinous and glutinous rice. Kor J Food Sci Technol 18:1337142 (1986). 23 Monsoor MA and Proctor A, Volatile component analysis of commercially milled head and broken rice. J Food Sci 69:C632C636 (2004). 24 Buttery RG, Juliano BO and Ling LC, Cooked rice aroma and 2-acetyl1-pyrroline. J Agric Food Chem 31:823826 (1983). 25 Champagne E, Rice aroma and avor: a literature review. Cereal Chem 85:445454 (2008).

2601
wileyonlinelibrary.com/jsfa

J Sci Food Agric 2010; 90: 25952601

c 2010 Society of Chemical Industry

Research Article
Received: 11 May 2010 Revised: 14 July 2010 Accepted: 19 July 2010 Published online in Wiley Online Library: 5 August 2010

(wileyonlinelibrary.com) DOI 10.1002/jsfa.4128

Relationships between disease control, green leaf duration, grain quality and the production of alcohol from winter wheat
Andrew M Watson,a Martin C Hare,a Peter S Kettlewell,a James M Brosnanb and Reginald C Agub
Abstract
BACKGROUND: Since demand for distilling wheat is expected to increase rapidly as a result of the development of the bioethanol industry, efcient production will become of increasing importance. Achieving this will require an understanding of the agronomic factors that inuence both grain yield and alcohol yield. Therefore ve eld experiments using the winter distilling wheat variety Glasgow were conducted over three seasons (20062007, 20072008 and 20082009) to study the relationships between foliar disease and alcohol yield. RESULTS: There was a signicant relationship between alcohol yield and the severity of the disease septoria leaf blotch (Septoria tritici), which was present in the experiments from natural infection. Retention of green ag leaf area as affected by disease control following fungicide application was also shown to be important for achieving high alcohol yields. Measurements of grain quality showed that high thousand-grain weight and low grain protein concentration were signicantly related to increased alcohol yield. CONCLUSION: The experiments showed the importance of disease management to protect alcohol yields in the distilling wheat crop. Fungicides that provide greater disease control and improved green leaf retention are likely to be benecial to alcohol yield. c 2010 Society of Chemical Industry Keywords: grain quality; alcohol; bioethanol; wheat; disease; green leaf duration; fungicide

INTRODUCTION
Traditionally, distilling wheat in the UK has been grown for the Scotch whisky industry. This market requires an annual production of 7 105 t of wheat to produce approximately 3 108 L of alcohol (LA) each year.1 There is growing interest in the production of distilling wheat for the emerging fuel alcohol market. Fuel alcohol, commonly known as bioethanol, is produced using similar techniques to those used by the potable alcohol industry. Growing interest in biofuels has led to concerns over the production of large areas of non-food crops threatening the availability and affordability of food.2 There have also been concerns over the amount of CO2 emitted during production and processing of the fuel.3 If biofuels are to be produced, it is important to address these concerns by maximising efciency both in the eld and during processing. Achieving a high level of alcohol production per hectare will help minimise the cultivated area of non-food crops, easing pressures on both food production and the environment. Maximising the efciency of crop production can also help limit the amount of CO2 emitted during production. Achieving a high grain alcohol yield (LA t1 ) will also help maximise processing efciency and reduce energy requirements. To achieve these objectives, it is necessary to have a thorough understanding of the agronomic factors that inuence alcohol production. Previous investigations have studied relationships between disease control, green ag

leaf area duration and grain quality,4,5 but none has looked at the inuence of these factors on alcohol yield. This study is therefore designed to expand on what is known to determine how these factors inuence the production of alcohol.

MATERIALS AND METHODS

Experimental design Five eld experiments were conducted in Shropshire, UK at Harper Adams University College. During the study, experiment 1 was established for the 20062007 growing season, experiments 2 and 3 were conducted in the 20072008 growing season and experiments 4 and 5 were conducted in the 20082009 growing season. Each experiment used the soft winter wheat variety Glasgow recommended for distilling. All experiments were conducted using a randomised block design with seven replicates.

Correspondence to: Andrew M Watson, Crop and Environment Research Centre, Harper Adams University College, Newport, Shropshire TF10 8NB, UK. E-mail: awatson@harper-adams.ac.uk

a Crop and Environment Research Centre, Harper Adams University College, Newport, Shropshire TF10 8NB, UK b Scotch Whisky Research Institute (SWRI), Riccarton, Edinburgh EH14 4AP, UK

2602

J Sci Food Agric 2010; 90: 26022607

www.soci.org

c 2010 Society of Chemical Industry

Relationships affecting alcohol production from winter wheat

www.soci.org treatments at GS 59 were applied using a hand-held eld trial sprayer propelled by CO2 . Spraying was performed using a spray pressure of 2 bar to provide medium spray quality, with Lurmark F110 02 nozzles. Treatments were applied using a water volume of 200 L ha1 and a forward speed of 1 m s1 . Weather data Weather data were collected from a meteorological station within 1 km of the eld experiments. Data for monthly rainfall and average minimum and maximum temperatures were compared with data collected over a period of 30 years (19611990) from the same meteorological station. Preharvest assessments Septoria leaf blotch (Septoria tritici) was the only foliar disease that caused signicant symptoms in these experiments throughout the growing season. This disease was assessed on the ag leaf 60 days after leaf emergence on ten leaves per plot using visual assessment keys,7 which provided a pictorial comparison of % disease symptoms at given levels (1, 2 and 5%). Despite being a rst wheat, parts of experiment 2 became infected with take-all (Gaeumannomyces graminis). This was visually assessed using the method of Tilston et al.,8 whereby the number of white heads within 1 m2 were counted at three points in each plot using a systematic sampling method. These data were then used as a covariate when analysing the results for this experiment. Towards the end of the growing season the decline in green ag leaf area was monitored by performing visual assessments of % green leaf area every 3 days throughout the period of senescence on ten randomly selected shoots per plot. Harvest and postharvest assessments Yield and grain weight Experiments were harvested using a plot combine, a 1 kg sample of grain being taken from each plot; all further measurements conducted derived from this sample. Thousand-grain weight was measured using an automated seed counter with a known weight of grain. From this the average weight of a grain was determined and then multiplied by 1000. Grain protein concentration Grain nitrogen concentration was measured by near-infrared (NIR) using an Infratec 1241 grain analyser (Foss, Warrington, UK). Grain protein concentration was measured by multiplying the nitrogen concentration by 5.7.9 Predicted alcohol yield The Scotch Whisky Research Institute (SWRI) has developed a calibration for the NIR grain analyser to predict alcohol yields, using grain samples from across the UK collected over 7 years. This technique is now commercially used within Scottish grain distilleries and was employed in this study to predict the alcohol yields on the plot grain samples. In each of the three years, ten samples were selected for their high and low alcohol yields. These selected samples were subjected to a wheat cook method10 to conrm the accuracy of the predicted alcohol yield data. This method closely simulates the distillation process conducted in grain distilleries.

Table 1. Fungicide treatments applied at GS 59 Treatment 1 2 3 4 5 6 Fungicide treatment at GS 59 Untreateda Prothioconazole + tebuconazole 62.5 : 62.5 g ha1 Prothioconazole + tebuconazole 62.5 : 62.5 g ha1 + azoxystrobin 62.5 g ha1b Prothioconazole + tebuconazole 62.5 : 62.5 g ha1 + azoxystrobin 125 g ha1 Prothioconazole + tebuconazole 62.5 : 62.5 g ha1 + azoxystrobin 187.5 g ha1 Prothioconazole + tebuconazole 62.5 : 62.5 g ha1 + azoxystrobin 250 g ha1

a Received no fungicide at GS 59 (i.e. only received fungicides at GS 30, 32 and 39). b Azoxystrobin 62.5 g ha1 treatment was not present in experiment 1.

Each plot measured 12 m 1.8 m with a 1.5 m buffer zone between replicates and a 0.2 m buffer zone between plots within each replicate. All plots in all ve experiments received a fungicide application of 300 g L1 metrafenone (Attenzo, BASF, Cheadle Hulme, UK) at 0.2 L ha1 applied with 375 g L1 chlorothalonil + 62.5 g L1 propiconazole + 50 g L1 cyproconazole (Cherokee, Syngenta Crop Protection UK Ltd, Fulbourn, UK) at 0.75 L ha1 at growth stage (GS) 30 (pseudostem erect).6 A further 125 g L1 epoxiconazole (Opus, BASF) at 0.6 L ha1 + 500 g L1 chlorothalonil (Bravo 500, Syngenta Crop Protection UK Ltd) at 1 L ha1 was applied at both GS 32 (second node detectable) and GS 39 (ag leaf blade all visible). Experimental treatments were 125 g L1 prothioconazole + 125 g L1 tebuconazole (Prosaro, Bayer CropScience, Milton, UK) at 0.5 L ha1 with varying rates of 250 g L1 azoxystrobin (Amistar, Syngenta Crop Protection UK Ltd) applied at GS 59 (ear fully emerged) (Table 1). The base fungicide programme used for this study was designed to provide a high level of disease control to enable any potential physiological effects of the strobilurin fungicide azoxystrobin on green leaf duration, grain quality and alcohol to be detected. Crop establishment and management All ve experiments were established as the rst wheat crop after oilseed rape into a plough-based seedbed consisting of a free-draining sandy loam soil. The experiments were all drilled in early October using a seed rate of 300 seeds m2 . In the autumn of each year a broad-spectrum herbicide along with an insecticide was applied to control weeds and aphids in each of the experiments. Soil nitrogen was measured prior to fertiliser applications, enabling nitrogen inputs to be adjusted accordingly. This ensured that crops grown in each experiment would have approximately 180 kg available nitrogen ha1 . In experiment 1 (2007), nitrogen was applied on 8 March (GS 25) at 40 kg ha1 , on 14 April (GS 32) at 60 kg ha1 and on 27 April (GS 37) at 60 kg ha1 . In experiments 2 and 3 (2008), applications were made on 6 March (GS 25) at 42.5 kg ha1 , on 15 April (GS 32) at 42.5 kg ha1 and on 3 May (GS 37) at 40 kg ha1 . In experiments 4 and 5 (2009), nitrogen was applied on 13 March (GS 29) at 40.5 kg ha1 , on 9 April (GS 31) at 45 kg ha1 and on 15 May (GS 39) at 41.5 kg ha1 . A growth regulator of chlormequat (1.29 kg ha1 ) was applied at GS 31. Fungicide treatments at GS 30, 32 and 39 were applied to all plots using a 12 m tractor-mounted sprayer. Experimental

2603

J Sci Food Agric 2010; 90: 26022607

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org

AM Watson et al.

Table 2. Monthly rainfall, average minimum temperature and average maximum temperature for May, June, July and August 2007, 2008 and 2009 in comparison with 30 year mean (19611990) of recordings taken at Harper Adams University College, Shropshire, UK Year Rainfall (mm) 2007 2008 2009 19611990 Average minimum temperature ( C) 2007 2008 2009 19611990 Average maximum temperature ( C) 2007 2008 2009 19611990 May June July August

Table 3. Correlation matrix showing relationships between septoria leaf blotch on ag leaf 45 days after leaf emergence (SLB, %), green ag leaf area duration (GFLAD, days to Gompertz M), grain yield at 85% dry matter (GY, t ha1 ) and grain quality parameters thousand-grain weight (TGW, g), grain protein concentration (GPC, %) and predicted alcohol yield (PAY, LA t1 ) SLB GFLAD 0.8117 0.3313 0.2085 0.3953 GY TGW GPC

107.2 47.5 50.2 57.2 7.2 9.1 7.2 6.0 16.8 18.2 17.5 15.6

236.8 35.2 92.2 54.2 11.4 9.0 9.7 8.8 19.7 19.2 20.4 18.7

125.8 94.4 110.6 49.1 11.4 12.0 11.5 10.6 19.5 21.4 21.0 20.5

20.4 83.6 37.8 60.4 11.1 12.8 11.8 10.5 20.8 20.6 21.7 20.2

GFLAD GY TGW GPC PAY

0.4660 0.3047 0.1307NS 0.1807 0.4778

0.3108 0.2660 0.4188

0.1318NS 0.2382 0.3349

Data presented are from all ve experiments. Signicance: P < 0.05; P < 0.01; P < 0.001; NS, not signicant (P > 0.05).

Statistical analysis Data analysis was performed using the statistical program Genstat Version 12.0. Green ag leaf area duration was assessed by using regression analysis with standard curves and tting the Gompertz model.11 This enabled data to be combined from assessments made throughout the period of green leaf decline to give the time (days) to 37% green ag leaf area (Gompertz M) for each plot. A large proportion of the plots had low levels of disease, which skewed the data. Loge transformation was performed. The results from experiment 2 were analysed using take-all as a covariate. All data from the ve experiments were combined in a spreadsheet, and one-way analysis of variance (ANOVA) in randomised blocks was performed using experiment as a factor. The residual data from the ANOVA table were then used for further analysis. The residuals were used to remove the systematic effect of experiment, enabling overall regressions to be performed on the data. The general mean was then added to the residuals for presentation of the results. Correlation and linear regression were used to study the relationships between the various parameters.

(P = 0.003) to an increase in the predicted alcohol yield observed in the experiments (Table 3 and Fig. 1). The data can be linked together through the physiological mechanisms involved. This shows how disease infection in the eld results in a decline in green leaf area, which in turn is important for grain quality, which ultimately inuences predicted alcohol yield (Fig. 1). Where levels of septoria leaf blotch symptoms were less, green ag leaf area duration was signicantly (P < 0.001) greater. Increased green ag leaf area duration showed a relationship with increased grain lling, resulting in a higher thousand-grain weight, which in turn resulted in an increased predicted alcohol yield. Septoria leaf blotch and green ag leaf area duration can have a large impact on grain quality and were therefore directly linked to the predicted alcohol yield of the grain (Figs 2 and 3). Septoria leaf blotch and green ag leaf area duration Septoria leaf blotch was the only foliar disease present at high enough levels for assessments to be conducted. The average severity of symptoms in plots that received no fungicide at GS 59 across the ve experiments was 17.0% on the ag leaf in mid-July (60 days after ag leaf emergence). Greater symptoms of the disease inevitably resulted in a decline in green leaf area (P < 0.001). Septoria leaf blotch and grain quality The disease appeared to decrease thousand-grain weight, but the relationship was not signicant (P = 0.157). This is likely to be the result of the relatively low disease levels present in these experiments. A signicant (P = 0.050) relationship between septoria leaf blotch and grain protein concentration was detected, however. This showed that, where there was a 1% increase in the level of septoria leaf blotch on the ag leaf, grain protein concentrations increased by 0.04%. In addition, it was shown that a 1% increase in the disease reduced alcohol yields by 1.12 LA t1 (P < 0.001). Green ag leaf area duration, grain quality and yield Increased green ag leaf area duration had a signicant (P < 0.001) relationship with thousand-grain weight, which showed an increase of 0.18 g for each day taken to reach Gompertz M. A signicant (P < 0.001) increase in grain yield of 0.20 t ha1 was

RESULTS
The weather data collected showed that there was a high level of rainfall in May, June and July 2007 (experiment 1), July 2008 and June and July 2009 (Table 2). Temperature measurements showed no major variations from the 30 year mean. Therefore water does not appear to be a limiting factor to grain development within these experiments. The experiments were designed to study how the alcohol yield of the grain was inuenced by factors occurring in the eld, such as disease infection, green leaf retention and the relationships these had with grain quality. The overall relationships from the ve experiments are presented in Table 3. Increasing levels of septoria leaf blotch disease symptoms resulted in a signicant (P < 0.001) reduction in green ag leaf area duration, which showed a signicant (P < 0.001) relationship with increasing grain weight. Increases in thousand-grain weight were related

2604

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 26022607

Relationships affecting alcohol production from winter wheat

www.soci.org
y = 0.3546x + 0.0233 R 2 = 0.05 (P = 0.006)

y = 0.1816x 0.0406 R 2 = 0.11 (P < 0.001)

TGW

PAY

y = 1.3907x 0.0006 R 2= 0.22 (P < 0.001)

GFLAD GFLAD

TGW

SLB
Figure 1. Effect of septoria leaf blotch (SLB) on green ag leaf area duration (GFLAD) and how this inuences thousand-grain weight (TGW) to ultimately inuence predicted alcohol yield (PAY) of grain.

4650 4625 4600 4575 4550 4525 4500 4475 4450 3 4 5 6 SLB (%) 7

y = 1.12253x + 464.15 R2 = 0.23 (P < 0.001)

4650 4625 4600 4575 4550 4525 4500 4475 4450

y = 0.3359x + 437.26 R 2 = 0.15 (P < 0.001)

PAY (LA t1)

10

PAY (LA t1)

50

55

60

65

70

GFLAD (days to Gompertz M)

Figure 2. Overall relationship between septoria leaf blotch (SLB) on ag leaf (60 days after leaf emergence) and predicted alcohol yield (PAY) from all ve experiments (data presented are residuals added to general mean).

Figure 3. Overall relationship between green ag leaf area duration (GFLAD) and predicted alcohol yield (PAY) from all ve experiments (data presented are residuals added to general mean).

also observed. Grain protein concentration increased (P = 0.006) by 0.02% for each day taken to reach Gompertz M. Predicted alcohol yield increased by 0.34 LA t1 for each day taken to reach Gompertz M (P < 0.001). Grain quality and predicted alcohol yield Thousand-grain weight Increasing thousand-grain weight appeared to result in a higher grain protein concentration in these experiments, but the relationship was not signicant (P = 0.096). A signicant (P = 0.003) relationship was seen with predicted alcohol yield, which increased by 0.36 LA t1 for each 1 g increase in thousandgrain weight.

Protein and predicted alcohol yield An inverse relationship (P < 0.001) between grain protein concentration and predicted alcohol yield was shown in these experiments. For every 1% increase in the concentration of protein within the grain, the predicted alcohol yield fell by 3.23 LA t1 . Predicted alcohol yield and actual alcohol yield Distillations were conducted to determine the accuracy of the predicted alcohol yields. The results showed that there was a signicant (P < 0.001) relationship between the two methods. For every 1 LA t1 increase in predicted alcohol yield the actual alcohol yield increased by 0.71 LA t1 .

2605

J Sci Food Agric 2010; 90: 26022607

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org

AM Watson et al.

DISCUSSION
The results show that factors affecting the crop later in the season, such as disease and green ag leaf area duration, can have substantial effects on alcohol yield. Maintenance of green ag leaf area is of key importance in achieving high crop yields (t ha1 ) as well as helping to maximise the alcohol yield of the grain (LA t1 ). Delaying senescence helps maximise the crops ability to photosynthesise. In these experiments, green ag leaf area duration was manipulated by fungicide applications. Prolonged green ag leaf area duration is largely the result of disease control when nutrition and water are not limiting factors. In addition, fungicides have been shown to prolong green ag leaf area duration through physiological effects such as promoting the growth hormone cytokinin and delaying the inhibitor ethylene.12 Other studies have shown increases in green ag leaf area duration through the control of phylloplane microora such as Cladosporium herbarum.13,14 Prolonged green ag leaf area duration increases the potential grain-lling period and hence increases grain weight. In this study an average thousand-grain weight increase of 0.18 g occurred for each additional day the ag leaf retained above 37% green leaf area. However, studies by Dimmock and Gooding15 showed that the ability of the crop to prolong grain lling with increased green ag leaf area duration was dependent on the variety grown. They found that early-maturing varieties beneted less from extending green ag leaf area duration, as these varieties had a limited grainlling period compared with later-maturing varieties. The variety Glasgow has an average maturity date which is identical (0 days ) to the standard solstice.16 Therefore it is likely that the application of fungicides that promote green ag leaf area duration would be more benecial in increasing alcohol yield in late-maturing varieties such as Invicta (+3 days). However, it should be noted that green ag leaf area duration and grain lling will also be strongly inuenced by the growing conditions of the crop, such as temperature and the availability of water and nitrogen.17 19 Therefore this must be taken into account when making decisions on whether to apply additional fungicides to distilling wheat crops. Where increases in thousand-grain weight occur, it could be assumed that grain protein concentration would be diluted as a result of greater accumulation of carbohydrates later in the season. However, despite increases in thousand-grain weight in this study, no dilution of grain protein was observed, suggesting a late season accumulation of nitrogen within the grain. Therefore grain protein increased at a comparable rate to starch. As a result, there was a relatively high level of protein at a given starch concentration. Therefore the relationship between protein and alcohol yield was less (3.23 LA t1 per 1% increase in grain protein) than in previous studies (7.50 LA t1 ).1 Weather data collected over the experiments showed a high level of rainfall in the later period of the growing season in each of the years. According Monaghan et al.,20 this can result in an increased proportion of nitrogen accumulated post-anthesis. In addition, Gooding et al.21 and Ford et al.22 observed that late season fungicide applications had the potential to delay senescence in the root system, thus providing crops with the potential to translocate a greater amount of nitrogen to the grain. This may therefore offer a potential explanation for the results in the current study. Studies of the growing crop also showed that the control of septoria leaf blotch had a signicant relationship with alcohol yield. For every 1% increase in septoria leaf blotch severity on the ag leaf (60 days after ag leaf emergence) there was a reduction in alcohol yield of 1.12 LA t1 . Septoria leaf blotch is a facultative

pathogen. Facultative pathogens are highly destructive, secreting enzymes that destroy the plant cell wall to enable them to feed on the glucose within the cell.23 This reduces the plants grain-lling potential, thus reducing the thousand-grain weight. According to Gooding et al.,24 facultative pathogens are more detrimental to carbohydrate accumulation than to nitrogen. This was seen in the present study with an increase in grain protein concentration of 0.04% for every 1% increase in septoria leaf blotch severity. Previous studies25 27 have also observed increases in grain protein concentration as a result of septoria leaf blotch. The ability of this disease to reduce carbohydrate accumulation, thereby increasing grain protein concentration, can be directly related to the reductions in predicted alcohol yield observed in the present study. These ndings suggest that other facultative pathogens such as eyespot (Oculimacula spp.) and take-all (G. graminis var. tritici) may also have a similar relationship, resulting in reduced alcohol yields. The occurrence of obligate pathogens such as rust (Puccinia spp.) and mildew (Blumeria graminis), however, is likely to have a lesser effect. These pathogens are less destructive, developing a haustorial feeding complex to continuously absorb nutrients from the living plant cell,28 which results in less of an effect upon starch translocation. Studies by Dimmock and Gooding29 suggest that mildew retains nitrogen in the leaves, which is likely to reduce the grain protein concentration further. This is supported in studies by Johnson et al.30 and Puppala et al.,26 whose results appear to show a reduction in grain protein concentration with increased levels of mildew. Therefore infection with these diseases is likely to have little effect on grain alcohol yield per tonne but will reduce the overall crop yield and therefore alcohol yield per hectare. In conclusion, this study has shown that fungicide applications have the potential to substantially increase the alcohol yield of wheat grain as a result of disease control and extended green ag leaf area duration when the crop is not sink-limited. In addition to increasing the alcohol yield per tonne of grain, the concomitant effect of increased crop yield will result in a substantial increase in alcohol yield per hectare. In the present study, disease levels remained relatively low in the later stages of crop development. This was the result of the robust fungicide programme applied; in situations where disease levels are greater, the inuence of applying a late season fungicide is likely to be of increased importance. It should be noted that at the time of writing this paper there are no known price incentives in place for growers to produce crops with high alcohol yields in the UK. However, the cost of fungicide inputs in this study (12.50/ha + 10.80/ha for application) was more than covered by the additional yield benet observed (+0.85 t ha1 at 100/t = 85/ha).31

ACKNOWLEDGEMENTS
The authors thank Syngenta Crop Protection UK Ltd for sponsorship of the project, David Ranner of Syngenta Crop Protection UK Ltd for advice, Tom Bringhurst of the Scotch Whisky Research Institute for advice and the technical staff of the Crop and Environment Research Centre, Harper Adams University College for establishment and support of the eld experiments.

REFERENCES
1 Smith TC, Kindred DR, Brosnan JM, Weightman RM, Shepherd M and Sylvester-Bradley R, Wheat as a Feedstock for Alcohol Production (HGCA Research Review No. 61). HGCA, London (2006).

2606

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 26022607

Relationships affecting alcohol production from winter wheat

2 Gurgel A, Reilly JM and Paltsev S, Potential land use implications for global biofuels industry. J Agric Food Ind Organisat 5:134 (2007). 3 Reijnders L and Huijbregts MAJ, Life cycle greenhouse gas emissions, fossil fuel demand and solar energy conversion efciency in European bioethanol production for automotive purposes. JCleaner Prod 15:18061812 (2007). 4 Ruske RE, Gooding MJ and Jones SA, The effects of adding picoxystrobin, azoxystrobin and nitrogen to a triazole programme on disease control, ag leaf senescence, yield and grain quality of winter wheat. Crop Protect 22:975987 (2003). 5 Paveley ND, Sylvester-Bradley R, Scott RK, Craigon J and Day W, Steps in predicting the relationship of yield on fungicide dose. Phytopathology 91:708716 (2001). 6 Zadocks JC, Chang TT and Konzak CF, A decimal code for the growth stages of cereals. Weed Res 114:415421 (1974). 7 Cereal Disease Guide. AgrEvo UK Ltd, Kings Lynn (1997). 8 Tilston EL, Pitt D, Fuller MP and Groenhof AC, Compost increases yield and decreases take-all severity in winter wheat. Field Crops Res 94:176188 (2005). 9 Teller GL, Non-protein nitrogen compounds in cereals and their relation to the nitrogen factor for protein in cereals and bread. Cereal Chem 9:261274 (1932). 10 Agu RC, Bringhurst TA and Brosnan JM, Production of grain whisky and ethanol from wheat, maize and other cereals. J Inst Brew 112:314332 (2006). 11 Gooding MJ, Dimmock JPRE, France J and Jones SA, Green leaf area decline of wheat ag leaves: the inuence of fungicides and relationships with mean grain weight and grain quality. Ann Appl Biol 136:7784 (2000). 12 Grossmann K and Retzlaff G, Bioregulatory effects of the fungicidal strobulurin kresoxim-methyl in wheat (Triticum aestivum). Pestic Sci 50:1120 (1997). 13 Bertelsen JR, Neergaard E and Smedegaard-Petersen V, Fungicidal effects of azoxystrobin and epoxiconazole on phyllosphere fungi, senescence and yield of winter wheat. Plant Pathol 50:190205 (2001). 14 Hussain Z and Leith MH, The effect of applied sulphur on the growth, grain yield and control of powdery mildew in spring wheat. Ann Appl Biol 147:4956 (2005). 15 Dimmock JPRE and Gooding MJ, The effects of fungicides on rate and duration of grain lling in winter wheat in relation to maintenance of ag leaf green area. J Agric Sci 138:116 (2002). 16 Home Grown Cereals Authority, HGCARecommendedList,WinterWheat 2010/11. HGCA, London (2010). 17 Pan J, Yan J and WeiXing C, Modelling plant carbon ow and grain starch accumulation in wheat. Field Crops Res 101:276284 (2007).

www.soci.org
18 Gooding MJ, Ellis RH, Shewry PR and Schoeld JD, Effects of restricted water availability and increased temperature on the grain lling, drying and quality of winter wheat. J Cereal Sci 37:295309 (2003). 19 Kindred DR, Verhoeven TMO, Weightman RM, Swanston JS, Agu RC, Brosnan JM, et al., Effects of variety and fertiliser nitrogen on alcohol yield, grain yield, starch and protein composition of winter wheat. J Cereal Sci 48:4657 (2008). 20 Monaghan JM, Snape JW, Chojecki JS and Kettlewell PS, The use of grain protein deviation for identifying wheat cultivars with high protein concentrations and yield. Euphytica 122:309317 (2001). 21 Gooding MJ, Gregory PJ, Ford KE and Pepler S, Fungicide and cultivar affect post-anthesis patterns of nitrogen uptake, remobilization and utilization efciency in wheat. J Agric Sci 143:503518 (2005). 22 Ford KE, Gregory PJ, Gooding MJ and Pepler S, Genoptype and fungicide effects on late-season root growth of winter wheat. Plant Soil 284:3344 (2006). 23 Farrar JF and Lewis DH, Nutrient relations in biotrophic interactions, in Fungal Infection of Plants, ed. by Pegg GF and Ayres AG. Cambridge University Press, Cambridge, pp. 92132 (1987). 24 Gooding MJ, Smith SP, Davies WP and Kettlewell PS, Effects of lateseason applications of propiconazole and tridemorph on disease, senescence, grain development and the breadmaking quality of winter wheat. Crop Protect 13:362369 (1994). 25 Pepler S, Gooding MJ, Ford KE, Ellis RH and Jones SA, Delaying senescence of wheat with fungicides has interacting effects with cultivar on grain sulphur concentration but not with sulphur yield or nitrogen : sulphur ratios. Eur J Agron 22:405416 (2005). 26 Puppala V, Herrman TJ, Bockus WW and Loughin T, Quality response of twelve hard red winter wheat cultivars to foliar disease across four locations in central Kansas. Cereal Chem 75:9499 (1998). 27 Ruske RE, Gooding MJ and Dobraszczyk BJ, Effects of triazole and strobulurin fungicide programmes with and without late season nitrogen fertiliser, on baking quality of Malacca winter wheat. J Cereal Sci 40:18 (2004). 28 Sutton PN, Gilbert MJ, Williams LE and Hall JL, Powdery mildew infection of wheat leaves changes host solute transport and invertase activity. Physiol Plant 129:787795 (2007). 29 Dimmock JPRE and Gooding MJ, The effects of fungicides on wheat grain protein concentration. Aspects Appl Biol 64:217218 (2001). 30 Johnson JW, Baenziger PS, Yamazaki WT and Smith RT, Effects of powdery mildew on yield and quality of isogenic lines of Chancellor wheat. Crop Sci 19:349352 (1979). 31 Nix J, The John Nix Farm Management Pocketbook (40th edn). Andersons Centre, Melton Mowbray (2009).

2607
wileyonlinelibrary.com/jsfa

J Sci Food Agric 2010; 90: 26022607

c 2010 Society of Chemical Industry

Research Article
Received: 30 April 2010 Revised: 22 June 2010 Accepted: 19 July 2010 Published online in Wiley Online Library: 17 August 2010

(wileyonlinelibrary.com) DOI 10.1002/jsfa.4129

Modulation of salt (NaCl)-induced effects on oil composition and fatty acid prole of sunower (Helianthus annuus L.) by exogenous application of salicylic acid
Sibgha Noreena and Muhammad Ashrafa,b
Abstract
BACKGROUND: Salicylic acid (SA) is a potential endogenous plant hormone that plays an important role in plant growth and development. Since sunower yield and its seed oil yield are adversely affected by salinity, in this study the role of SA in modulating salt (NaCl)-induced effects on various yield and oil characteristics of sunower was investigated. For this purpose a greenhouse experiment comprising two sunower hybrid lines (Hysun-33 and SF-187), two NaCl levels (0 and 120 mmol L1 ) and four SA levels (0, 100, 200 and 300 mg L1 ) was conducted. RESULTS: Salt stress markedly reduced yield, oil content, linoleic acid and -tocopherol in both sunower lines, while it increased linolenic acid, palmitic acid, stearic acid and - and -tocopherols. However, increasing levels of foliar-applied SA resulted in improved achene yield and hundred-achene weight in both lines. Foliar-applied SA caused a signicant decrease in oil stearic acid and - and -tocopherols in both lines under non-saline and saline conditions. CONCLUSION: Salt-induced harmful effects on achene yield and oil characteristics of sunower could be alleviated by exogenous application of SA. High doses of SA caused a marked increase in sunower achene oil content as well as some key fatty acids. c 2010 Society of Chemical Industry Keywords: sunower; salt stress; salicylic acid; oil composition; tocopherols; antioxidant activity

INTRODUCTION
Plant growth and development are directly regulated by plant hormones. These inuence plants in multifarious ways, affecting a number of physiological/biochemical processes in plants subjected to abiotic stresses.1 4 Salicylic acid (SA) and related compounds belong to a diverse group of plant phenolics. This acid acts as an endogenous signal molecule and is responsible for alleviating adverse effects of abiotic stresses on plants.5,6 SA has also shown promise in mitigating oxidative and osmotic stresses due to salt stress by enhancing proteins, total free amino acids and soluble sugars in different plant parts of several crops.3,7 9 Apart from enhancement of growth and development in various crop plants, foliar-applied SA has been reported to improve grain yield of mung bean (Vigna radiata),10 Phaseolus vulgaris11 and wheat.3,12 A signicant increase in number of pods per plant and grain yield of soybean by exogenously applied SA was also observed.10 Sunower (Helianthus annuus L.) is one of a number of potential oilseed crops and is gaining considerable popularity because it is a short-duration crop. The oil content of sunower varies from 40 to 45% depending on the quality of seed. It contains 8595% polyunsaturated fatty acids. Oleic and linoleic acids are abundant, while it is low in cholesterol.13 Sunower is rated as a moderately salt-tolerant crop.14 16 However, as the salinity of irrigation water increases, plant height, stem diameter and head diameter of sunower decrease signicantly.17

Signicant decreases in achene oil concentration and oil yield may occur when the salt content of irrigation water exceeds l g L1 .18 Foliar injury from saline sprinkling water containing high Na+ or Cl concentration may occur when the salt level is greater than 20 mmol L1 . Francois13 and Maas16 reported that soil salinity up to ECe 4.8 dS m1 did not have an adverse effect on relative seed yield of four hybrids of sunower, but there was a reduction in yield by 5% for each unit increase in salinity above 4.8 dS m1 . However, it has been observed that fatty acid composition is greatly affected by salinity: under saline conditions, oleic acid increased and linoleic acid decreased progressively with increasing salinity level owing to salt-induced inhibition of oleate desaturase enzyme.19 Various researchers have reported the effects of salinity on oil content and fatty acid composition in different oilseed crops such as safower (Carthamus tinctorius L.),20,21 olive,22 rapeseed (Brassica napus L.), stock (Matthiola tricuspidata) and evening primose (Oenothera biennis).23 Oil yield decreased signicantly from 38.3 to 3.8 g per head and unsaturated fatty acids such as

Correspondence to: Muhammad Ashraf, Department of Botany, University of Agriculture, Faisalabad 38040, Pakistan. E-mail: ashrafbot@yahoo.com

a Department of Botany, University of Agriculture, Faisalabad, Pakistan b King Saud University, Riyadh, Saudi Arabia

2608

J Sci Food Agric 2010; 90: 26082616

www.soci.org

c 2010 Society of Chemical Industry

Modulation of salt-induced effects on sunower by salicylic acid oleic and linoleic acids, which account for 8595% of the total fatty acids in sunower seed, were greatly affected by salt stress.24 The oil content of sunower seed decreased from 524 to 508 mg oil g1 seed with increasing salt level, while oleic acid increased from 82.8 to 86.0%; conversely, linoleic acid decreased from 6.9 to 2.8% owing to inhibition of oleate desaturase occurring under salt stress.25 Francois and Kleiman25 also reported an increase in oleic acid with a concurrent decrease in linoleic acid in safower (C. tinctorius L.) under saline conditions (3.77.9 dS m1 ). The increase in oleic/linoleic acid ratio could be due to water stress, lipid peroxidation and/or accumulation of Na+ and Cl under salt stress.26 In view of the great economic importance of sunower as a potential oilseed crop, it was hypothesised that exogenously applied SA could improve salt tolerance in sunower for protable yield production as well as improved quality of oil. The objective of this study was to examine whether or not exogenously applied SA could alter the quantity and quality of achene oil from saltstressed sunower plants. In addition, relationships between yield attributes and various oil components of sunower supplied with SA were established.

www.soci.org volume of 9 mL of solution per three plants was sprayed in all cases with a manually operated sprayer. The nozzle of the hand-held sprayer was adjusted to deliver 1 mL of solution per spray, with each plant receiving 3 mL of the SA solution. Yield and yield components Capitulum weight per plant, hundred-achene weight and achene yield per plant were measured at maturity on day 113 after sowing. Harvested heads were sun dried and thrashed manually to clean the grains. Oil extraction Dried seeds (100 g) from each treatment were crushed. The oil was extracted with 0.5 L of n-hexane using a Soxhlet apparatus. Oil content was determined by evaporating the extractant in a rotary evaporator. Refractive index The refractive index of sunower oil was determined using an Abbe refractometer. The temperature of the prism was maintained at 40 C. The refractive index (n) was read after putting a clear drop of oil on the prism: n = sin(i)/ sin(r) = velocity of light in air/velocity of light in oil Specic gravity The specic gravity of sunower oil was determined as follows: specic gravity = density of oil at 20 C/density of water at 20 C Fatty acid composition First, fatty acid methyl esters (FAMEs) were prepared by the standard IUPAC method. Samples were derivatised into FAMEs of triglycerides by saponication of glycosides and esterication of fatty acids using methanol. The oil sample (0.2 g) was placed in a 100 mL round-necked and round-bottomed ask, then 30 mL of methanol and one pellet of KOH were added. The contents of the ask were reuxed for 25 min until the droplets of fat dissolved. On cooling, the reaction mixture was gently transferred to a separating funnel and a small amount of n-hexane was added. The funnel was shaken gently by rotating it several times and the upper layer of n-hexane was separated. The solution was dried over anhydrous sodium sulfate, ltered and stored in a sealed sample tube in a freezer until analysis. The fatty acid composition (stearic, palmatic, oleic and linoleic acids) was determined by gas chromatography using a Shimadzu 17-A gas chromatograph tted with an SP-2330 methyl lignoserate-coated polar capillary column (30 m 0.32 mm) and a ame ionisation detector (Shimadzu, Kyoto, Japan). Determination of tocopherols Tocopherols were analysed by the method of Lee et al.28 A highperformance liquid chromatograph (HPLC) tted with an S-3210 UV detector (Sykam GmbH, Kleinsthein, Germany) was used. The oil sample (1 g) was placed in a 10 mL volumetric ask and the volume was made up to the mark with acetonitrile. The ask was wrapped in foil to retard the oxidation process. The material was subsequently ltered and a 20 L aliquot was injected into

MATERIALS AND METHODS

The present investigation was conducted in a wire-house at the University of Agriculture, Faisalabad, Pakistan under natural light. During the experimental period, mean maximum temperatures ranged from 27.6 to 40.3 C, mean minimum temperatures from 13.0 to 27.3 C and mean relative humidity from 26.2 to 52.4%. The experimental design was a completely randomised three-factorial (2 2 4) design with four replications. Two sunower (H. annuus L.) hybrid lines (Hysun-33 and SF-187), two salinity levels (0 and 120 mmol L1 NaCl) and four SA levels (0, 100, 200 and 300 mg L1 ) were assessed. Before the start of the experiment the seeds of both sunower hybrids were treated with 5 g L1 sodium hypochlorite solution for 5 min to prevent the occurrence of seedling diseases. The seeds were then washed several times with distilled water to remove excess sodium hypochlorite solution. The sterilised seeds were air dried at room temperature. Plastic pots of 24.5 cm diameter and 28 cm depth were each lled with 10 kg of airdried river sand. The sunower achenes were selected on the basis of uniform size and maturity and freedom from physical damage or infection. Five achenes were dibbled at 5 mm depth in each pot, and irrigation with modied half-strength Hoaglands nutrient solution was carried out every 2 days after sowing.27 After complete emergence, i.e. on day 4 after sowing, thinning to three seedlings per pot was carried out. Neutral NaCl was used to impose salt stress. Two levels of salinity (0 and 120 mmol L1 NaCl) were prepared in full-strength Hoaglands nutrient solution. The salt stress was imposed on day 15 after sowing. The plants were salinised progressively at 1 day intervals at an increment of 40 mmol L1 NaCl. SA (2-hydroxybenzoic acid (C7 H6 O3 ), molecular weight 138.1; Sigma Aldrich, Tokyo, Japan) was initially dissolved in a few drops of dimethyl sulfoxide and then in distilled water and the pH was adjusted to 5.5 0.2 with 0.5 mol L1 KOH. Four levels of SA (0, 100, 200 and 300 mg L1 ) were prepared. The SA was applied exogenously in combination with 1 mL L1 Tween-20 (polyoxyethylene sorbitan laurate) as a surfactant to ensure spreading of the applied solution on the leaf surface and to maximise penetration of the exogenously applied SA into the leaf tissues. The rst foliar application of SA was done on day 20 after sowing and the second on day 3 after the rst application. A

2609

J Sci Food Agric 2010; 90: 26082616

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org

S Noreen, M Ashraf

Table 1. Mean squares from analysis of variance of data for yield attributes, oil physical properties and oil composition of sunower (Helianthus annuus L.) when varying levels of salicylic acid were applied as a foliar spray to 24-day-old plants subjected to normal or saline conditions Source of variation Salt (S) Hybrid line (HBL) Salicylic acid (SA) S HBL S SA HBL SA S HBL SA Error Source of variation Salt (S) Hybrid line (HBL) Salicylic acid (SA) S HBL S SA HBL SA S HBL SA Error Source of variation Salt (S) Hybrid line (HBL) Salicylic acid (SA) S HBL S SA HBL SA S HBL SA Error DF 1 1 3 1 3 3 3 48 DF 1 1 3 1 3 3 3 48 DF 1 1 3 1 3 3 3 48 Hundred-achene weight 2.080 0.151ns 0.235 3.850 0.077ns 0.257ns 0.489 0.163 Specic gravity 0.443 0.382 0.023ns 0.163 0.024ns 0.011ns 0.004ns 0.036

Achene yield per plant 12.60 64.20 0.85ns 39.20 4.25 1.40ns 2.72ns 1.12 -Tocopherol 508.50 20652.00 18166.00 2458.00 2477.00 3981.00 111.20 31.6

Capitulum weight per plant 1.32ns 138.40 0.90ns 61.60 12.80 5.25ns 5.85ns 2.65 -Tocopherol 3609.00 44.40ns 603.10 53.30 76.60 21.30ns 29.60ns 12.90 Linolenic acid 2.080 0.102 0.072 0.00001ns 0.007ns 0.012ns 0.003ns 0.017

Seed oil content 142072.4 15221.4 23167.5 858.4 1500.5 4531.7 5996.2 164.1 -Tocopherol 1160.00 1559.00 727.10 1.40 4.45 252.30 19.40 14.70 Palmitic acid 107.40 0.47ns 0.12ns 3.60ns 1.57ns 0.66ns 0.37ns 2.60

Refractive index 0.0010ns 0.0070 0.0010ns 0.0001ns 0.0004ns 0.0020 0.0001ns 0.004 Oleic acid 183.40 5114.00 14.53ns 79.81ns 3.95ns 44.10ns 64.40ns 34.40 Stearic acid 508.50 20652.00 18166.00 2458.00 2477.00 3981.00 111.20 31.60

Linoleic acid 371.60 1863.00 120.90 11.50ns 3.39ns 4.46ns 11.20ns 20.6

Signicance: P < 0.05; P < 0.01: P < 0.001; ns, not signicant.

a Hypersil ODS reverse phase (C-18) guard column for elution with a solvent comprising HPLC-grade methanol and acetonitrile (65 : 35 v/v). A ow rate of 1.3 mL min1 at 30 C was used to achieve separation. The absorbance of the separated solution was detected at 292 nm. The different forms of tocopherol were identied by comparison of retention times. Quantication was carried out by comparing the peak areas of the unknown samples with those of pure blank standards of -, - and -tocopherols (Sigma Chemical Co., St Louis, MO, USA). Data were calculated from the peak areas using PeakSimple chromatography data acquisition and integration software (SRI Instruments, Torrance, CA, USA). Statistical analysis Analysis of variance of all data was carried out using CoStat software (CoHort, Berkeley, CA, USA). Signicant differences between mean values were assessed using the least signicant difference test at P < 0.05.29

RESULTS
Salt stress (120 mmol L1 NaCl) was found to be very effective in reducing hundred-achene weight in both sunower hybrid lines. There were no signicant differences between the two hybrid lines in this parameter (Table 1, Fig. 1). However, exogenous

application of varying levels of SA caused a signicant change in grain weight of both sunower hybrid lines under salt-stressed and non-stressed conditions. Increasing levels of foliar-applied SA resulted in improved achene weight in both hybrid lines. Achene yield in both hybrid lines was signicantly reduced by increasing salt concentration in the substrate. However, the responses of the two lines to salt stress with respect to this yield attribute were signicantly different. Exogenous application of 100 and 300 mg L1 SA caused a substantial increase in grain yield of line SF-187 under salt-stressed and non-stressed conditions (Table 1, Fig. 1). Another potential yield component, capitulum weight, was not signicantly affected by salinity. However, the two hybrid lines differed signicantly in this attribute (Table 1, Fig. 1). Under nonstressed conditions, SF-187 attained more capitulum weight after application of 100 mg L1 SA, but no gain in capitulum weight of Hysun-33 was observed until application of 200 mg L1 SA (Fig. 1). Exogenous application of 300 mg L1 SA proved most effective in increasing capitulum weight in both hybrid lines under salt stressed and non-stressed conditions. Addition of NaCl to the growth medium caused a substantial reduction in oil content of sunower. The two hybrid lines differed signicantly in seed oil content. Hysun-33 produced more seed oil than SF-187 under both saline and normal growth conditions. Seed oil content was signicantly improved by foliar application of SA in both hybrid lines under salt-stressed and non-stressed

2610

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 26082616

Modulation of salt-induced effects on sunower by salicylic acid

200 mg L1

www.soci.org

0 mg L1 9 Achene yield (g per plant) 8 7 6 5 4 3 2 1 0 3.5 Hundred-achene weight (g) 3 2.5 2 1.5 1 0.5 0 16 14 12 10 8 6 4 2 0 Control SF-187

100 mg L1

300 mg L1 Salicylic acid

Capitulum weight (g per plant)

Saline

Control HYSUN-33

Saline

Figure 1. Yield attributes of two sunower (Helianthus annuus L.) hybrid lines when varying levels of salicylic acid were applied as a foliar spray to 24-day-old plants subjected to normal or saline conditions (mean standard error, n = 4).

conditions (Table 1, Fig. 2). Thus exogenously applied SA had a promotive effect on seed oil content and was more effective when 300 mg L1 SA was applied as a foliar spray. Furthermore, plants of Hysun-33 grown on saline medium and sprayed with SA produced more seed oil than those of SF-187, while the opposite was true under non-saline conditions (Fig. 2). The refractive index of seed oil of both lines was little affected by salinity. Foliar application of SA did not signicantly affect the refractive index of either line (Table 1, Fig. 2). The specic gravity of achene oil of both sunower lines was signicantly reduced by addition of salt to the growth medium. The two lines differed signicantly under non-saline conditions. Seed oil of SF-187 had higher specic gravity than that of Hysun-33 under non-saline conditions. The specic gravity of seed oil of both hybrid lines was little inuenced by foliar spraying of SA on plants subjected to saline conditions. Salt stress signicantly increased seed oil linolenic, palmitic and stearic acid contents but decreased seed oil linoleic acid content in both lines of sunower (Table 1). The two lines also differed signicantly in all these fatty acids except palmitic acid. Hysun-33 was higher in oil palmitic and linolenic acid contents but lower in oil stearic acid content than SF-187 under saline conditions. Foliar

application of 300 mg L1 SA resulted in increased oil stearic acid content in both lines under saline conditions. Similarly, foliar application of various levels of SA caused a marked improvement in oil linolenic acid content in both hybrid lines under saline and normal conditions (Fig. 3). Saline conditions caused a signicant reduction in oil oleic acid content, particularly in SF-187 (Table 1, Fig. 3). However, Hysun-33 had higher oil oleic acid content than SF-187 under both salt-stressed and non-stressed conditions. Exogenously applied SA improved oil oleic acid content only in salt-stressed plants of SF-187 (Fig. 3). Linoleic acid content in both sunower lines was signicantly reduced (Table 1). SF-187 was higher in oil linoleic acid content than Hysun-33 under both salt-stressed and non-stressed conditions (Fig. 3). Exogenously applied SA resulted in a signicant reduction in oil linoleic acid content in plants of both hybrid lines grown under saline and normal conditions. Linolenic acid content in both lines was higher under salt stress than under normal conditions. It generally increased owing to imposition of salt on the rooting medium (Fig. 3). Hysun-33 had higher oil linolenic acid content than SF-187 under both non-saline and saline conditions. Oil palmitic acid content in both sunower hybrids was higher under salt stress than under non-stressed conditions. Exogenous

2611

J Sci Food Agric 2010; 90: 26082616

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org
0 mg L1 Seed oil contents (g oil kg-1 seed) 450 400 350 300 250 200 150 100 50 0 4.5 4 3.5 Specific gravity 3 2.5 2 1.5 1 0.5 0 1.5 1.48 1.46 1.44 1.42 1.4 1.38 1.36 1.34 Control SF-187 Saline Control HYSUN-33 Saline 100 mg L1 200 mg L1 300 mg L1 Salicylic acid

S Noreen, M Ashraf

Figure 2. Physical properties of seed oil of two sunower (Helianthus annuus L.) hybrid lines when varying levels of salicylic acid were applied as a foliar spray to 24-day-old plants subjected to normal or saline conditions (mean standard error, n = 4).

Refractive index

application of SA had a positive effect in increasing palmitic acid content in plants under saline conditions. Hysun-33 maintained higher oil palmitic acid content than SF-187 under salt stress. In contrast, SF-187 maintained higher oil palmitic acid content than Hysun-33 under non-saline conditions (Fig. 3). Oil stearic acid content decreased with increasing concentration of foliar-applied SA under non-stressed conditions. On the other hand, it increased with exogenous SA application under salt stress. SF-187 sprayed with 300 mg L1 SA had maximum oil stearic acid content under salt stress. Foliar application of 200 mg L1 SA resulted in a reduction in oil stearic acid content compared with foliar application of 100 mg L1 SA in plants of both hybrid lines under salinity stress. Analysis of tocopherol data showed that salt stress signicantly increased - and -tocopherol contents but decreased -tocopherol content in the seed oil of both sunower lines (Table 1, Fig. 4). SF-187 was higher in oil - and -tocopherol contents than Hysun-33 under both salt-stressed and non-stressed conditions. However, the two lines showed little difference in oil -tocopherol content under salt-stressed and non-stressed conditions. Foliar-applied SA caused a signicant decrease in all three tocopherols in both lines under non-saline and saline conditions. Maximum reduction in all seed oil tocopherols in both lines

was observed at 300 mg L1 SA under salt-stressed and normal conditions. The analysis of correlation coefcients (r) among various yield and achene oil characteristics presented in Table 2 shows that only yield attributes such as hundred-achene weight/achene yield (r = 0.94 ), seed oil content/achene yield (r = 0.76 ) and specic gravity/achene yield (r = 0.64 ) are signicantly correlated; among the oil composition characteristics, -, - and tocopherols are signicantly correlated with oleic acid (r = 0.75 , 0.76 and 0.76 respectively).

DISCUSSION
Crop yield is a vital determinant for assessing crop productivity of a plant under stress conditions. In sunower, size of capitulum and number and size of achenes per capitulum determine the actual yield, since various physiological and biochemical processes that govern yield are adversely affected by salinity.30 34 Therefore it is not difcult to suggest that decline in achene yield due to salt stress with concurrent enhancement in achene yield due to SA application was principally reponsible for change in achene yield (achene yield vs hundred-achene weight, r = 0.94 ). These results can be discussed by reviewing earlier studies where

2612

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 26082616

Modulation of salt-induced effects on sunower by salicylic acid

0 mg L1 100 mg L1 200 mg L1

www.soci.org
300 mg L1 Salicylic acid

12 Palmitic acid (%) Oleic acid (%) Linoleic acid (%) Linolenic acid (%) 10 8 6 4 2 0 80 70 60 50 40 30 20 10 0 60 50 40 30 20 10 0 1 0.8 0.6 0.4 0.2 0

Control SF-187

Saline

Control HYSUN-33

Saline

Figure 3. Fatty acid prole of seed oil of two sunower (Helianthus annuus L.) hybrid lines when varying levels of salicylic acid were applied as a foliar spray to 24-day-old plants subjected to normal or saline conditions (mean standard error, n = 4).

plants treated with salicylates had higher yield when cultivated in either greenhouse or open conditions. For example, it was reported that SA caused production of larger-sized tubers of carrot (60%), beet (16%) and radish (200%).35 Increase in capitulum size of sunower lines through foliar-applied SA could have been due to either development of additional owers or rapid ower initiation with subsequent ower development.36 For example, in a study of Xanthium strumarium, Cleland and Ajami37 reported that exogenously applied SA stimulated owering. Similarly, owering in Pistia stratiotes, a member of the family Arecaceae, was also accelerated by incorporating SA in the culture medium.38 In another study, Kumar et al.39 compared the owerinducing effect of SA with that of gibberellic acid (GA), kinetin, 1-naphthaleneacetic acid, ethral and chlorocholine chloride. They found that SA and/or GA were more effective than any other combination of these hormones. However, the specic owerinducing mechanism that involves SA is yet to be elucidated. Based on information available in the literature, it is evident that owering is induced by chelating agents in Laminaceae40,41 through the activity of benzoic acid42,43 and other non-chelating phenolic

compounds.44 It was hypothesised that the free o-hydroxyl group of benzoic acid induces a metal-chelating characteristic that favours ower induction.45 47 Moreover, it was concluded that additional ower-inducing mechanisms may be heavily involved in this process.46,47 Thus the results from the present study also support this notion that SA-induced increase in achene yield in saltstressed plants of both sunower lines might have been due to the development of additional owers, thereby increasing capitulum size. In the present study, SA-induced increase in achene size of salt-stressed plants of both sunower lines might have been due to transport of an additional amount of photoassimilates from vegetative organs (source) to achenes (sink) during their development, thereby increasing achene size. Zhou et al.48 also reported a 9% higher grain weight of maize by injecting stems with SA compared with sucrose and/or distilled water treatments. Similarly, in a study of wheat, Arfan et al.12 observed that SAinduced improvement in grain yield occurred because of SAinduced enhancement in grain size. Achene oil content in both sunower lines was decreased by salinity. This result is in line with previous studies in which

2613

J Sci Food Agric 2010; 90: 26082616

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org

S Noreen, M Ashraf

250 a-Tocopherol (mg kg1 oil) 200 150 100 50 0 80 d-Tocopherol (mg kg1 oil) 70 60 50 40 30 20 10 0 g-Tocopherol (mg kg1 oil) 70 60 50 40 30 20 10 0

0 mg L1

100 mg L1

200 mg L1

300 mg L1 Salicylic acid

Control SF-187

Saline

Control HYSUN-33

Saline

Figure 4. Tocopherol contents of seed oil of two sunower (Helianthus annuus L.) hybrid lines when varying levels of salicylic acid were applied as a foliar spray to 24-day-old plants subjected to normal or saline conditions (mean standard error, n = 4).

it was shown that salt stress reduced seed oil content in ax and safower49 and sunower.19 However, exogenously applied SA improved achene oil content in both sunower lines. This improvement was greatest in salt-stressed plants of Hisun-33 treated with 300 mg L1 SA foliar spray. It is widely accepted that the quality of achene oil is associated with its fatty acid composition, mainly the proportion of oleic, linoleic and linolenic acids. In the present study, salinity stress increased palmitic, stearic and linolenic acid contents in both sunower lines, while linoleic acid content decreased. Thus salt stress reduced achene oil quality in both sunower lines, which is similar to what was observed previously in stock (M. tricuspidata), chia (Salvia hispanica) and evening primrose (O. biennis)23 and Matthiola incana.50 Of all fatty acids, the quantity of oleic and linoleic acids is most important from the viewpoint of oil quality. Exogenous application of SA did not change palmitic, stearic and oleic acid contents, while linolenic acid content increased considerably in both sunower lines. These results suggest that exogenous SA application counteracted the salt-induced harmful effects on achene oil quality in both sunower lines. Tocopherols (vitamin E) are non-enzymatic antioxidant compounds found to a varying extent in seed oils. They counteract

oxidative stress, thereby maintaining oil quality.51 Although different types of tocopherol occur in oils, -tocopherol is the most prominent, as it has considerable antioxidant activity.52 Salinity stress can generate reactive oxygen species (ROS) in plants to a great extent,31,53 55 which damage seed lipids.56 To combat ROS, plants generate various types of antioxidant, including tocopherols (vitamin E).57 In the present study, salt stress signicantly increased - and -tocopherol contents in the seed oil of both sunower hybrids. These results are in agreement with those of Anwar et al.,58 who reported that salt stress increased tocopherol content in the seed oil of Moringa oleifera plants. A considerable variation was observed in both sunower cultivars with regard to this attribute. However, when parallels were drawn between tocopherol content and yield attributes, a non-signicant association was observed. In conclusion, salt-induced harmful effects on achene yield and oil characteristics of sunower could be allayed by exogenous application of SA, but the effectiveness of SA in mitigating the negative effects of salt stress was dose-dependent. However, high doses of SA caused a marked increase in sunower achene oil content as well as some key fatty acids such as linolenic, oleic and palmitic acids.

2614

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 26082616

Modulation of salt-induced effects on sunower by salicylic acid

www.soci.org

Table 2. Correlation coefcients (r) among yield attributes, oil physical properties and oil composition of sunower (Helianthus annuus L) when varying levels of salicylic acid were applied as a foliar spray to 24-day-old plants subjected to normal or saline conditions Achene yield Achene yield Hundredachene weight Capitulum weight Seed oil content Refractive index Specic gravity Palmitic acid Stearic acid Oleic acid Linoleic acid Linolenic acid -Tocopherol -Tocopherol -Tocopherol 0.94 Hundredachene Capitulum Seed oil Refractive Specic Palmitic Stearic weight weight content index gravity acid acid Oleic acid Linoleic Linolenic acid acid

-Toc

-Toc

0.24ns 0.76 0.28ns 0.64 0.91 0.81 0.06ns 0.42ns 0.85 0.23ns 0.41ns 0.41ns

0.32ns 0.8 0.13ns 0.6 0.79 0.81 0.12ns 0.29ns 0.73 0.25ns 0.37ns 0.37ns 0.38ns 0.25ns 0.10ns 0.35ns 0.45ns 0.27ns 0.26ns 0.17ns 0.16ns 0.07ns 0.07ns 0.04ns 0.13ns 0.69 0.79 0.4ns 0.09ns 0.44ns 0.07ns 0.18ns 0.18ns 0.61 0.2ns 0.05ns 0.57ns 0.74 0.39ns 0.53ns 0.6 0.6 0.52ns 0.37ns 0.87 0.4ns 0.18ns 0.4ns 0.79 0.29ns 0.12ns 0.75 0.79 0.87 0.76 0.05ns 0.44ns 0.05ns 0.16ns 0.75 0.54ns 0.17ns 0.03ns 0.76 0.55ns 0.17ns 0.03ns 0.76

0.6 0.6 0.7 0.7

0.08ns 0.95 0.95

0.95 0.99

0.99

Signicance: P < 0.05; P < 0.01: P < 0.001; ns, not signicant.

REFERENCES
1 Reymond P and Farmer EE, Jasmonate and salicylate as global signals for defense gene expression. Curr Opin Plant Biol 1:404411 (1998). 2 Steduto P, Albrizio R, Giorio P and Sorrentino G, Gas exchange response and stomatal and non-stomatal limitations to carbon assimilation of sunower under salinity. Environ Exp Bot 44:243255 (2000). 3 Shakirova FM, Sakhabutdinova AR, Bezrukova MV, Fathutdinova RA and Fathutdinova DR, Changes in hormonal status of wheat seedlings induced by salicylic acid and salinity. Plant Sci 164:317322 (2003). 4 Ashraf M, Inducing drought tolerance in plants: recent advances. Biotechnol Adv 28:169183 (2010). 5 Senaratna T, Touchell D, Bunn T and Dixon K, Acetyl salicylic (aspirin) and salicylic acid induce multiple stress tolerance in bean and tomato plants. Plant Growth Regul 30:157161 (2000). 6 Khodary SEA, Effect of salicylic acid on growth, photosynthesis and carbohydrate metabolism in salt stressed maize plants. Int J Agric Biol 6:58 (2004). 7 Borsani O, Valpuesta V and Botella MA, Evidence for a role of salicylic acid in the oxidative damage generated by NaCl and osmotic stress in Arabidopsis seedlings. Plant Physiol 126:10241030 (2001). 8 El-Tayeb MA, Response of barley grains to the interactive effect of salinity and salicylic acid. Plant Growth Regul 45:215224 (2005). 9 Gunes A, Inal A, Alpaslan M, Eraslan F, Bagci EG and Cicek N, Salicylic acid induced changes on some physiological parameters symptomatic for oxidative stress and mineral nutrition in maize (Zea mays L.) grown under salinity. J Plant Physiol 164:728736 (2006). 10 Singh G and Kaur M, Effect of growth regulators on podding and yield of mung bean (Vigna radiata L.) Wilezck. Indian J Plant Physiol 23:366370 (1980). 11 Rendon SLA, Hormonal control of reproductive organs in Phaseolus vulgaris L. cv. Cacahuate-72. Master of Science Thesis, C.P. Chapingo, Mexico, (1983). 12 Arfan M, Athar HR and Ashraf M, Does exogenous application of salicylic acid through the rooting medium modulate growth and photosynthetic capacity in differently adapted spring wheat cultivated under salt stress? J Plant Physiol 6:685694 (2007). 13 Francois LE, Salinity effects on four sunower hybrids. Agron J 88:215219 (1996).

14 Richards RA, Increasing salinity tolerance of grain crops. Is it worthwhile? Plant Soil 146:8998 (1992). 15 Flowers TJ and L uchli A, Sodium versus potassium: substitution and a compartmentation, in Inorganic Plant Nutrition, Vol. 15B, ed. by L uchli A and Bieleski RL. Springer, Berlin, pp. 651681 (1983). a 16 Maas EV, Salt tolerance in plants. Appl Agric Res 1:1226 (1986). 17 Farah MA, Daoud AM, Barakat MA and Anta IM, The effect of the depth to water table and salinity of irrigation water on growth and various parameters of sunower. Agric Res Rev 59:199209 (1984). 18 Raju KV and Ranganayakulu C, Effect of saline water irrigation on salt distribution in prole and performance of sunower crop. Indian J Agric Res 12:183186 (1978). 19 Flagella Z, Giuliani MM, Rotunno T, Caterina RD and De Caro A, Effect of saline water on oil yield and quality of a high oleic sunower (Helianthus annuus L.) hybrid. Eur J Agron 21:267272 (2004). 20 Irving DW, Shannon MC, Breda VA and Mackey BE, Salinity effects on yield and oil quality of high-linoleate and high-oleate cultivars of safower (Carthamus tinctorius L.). J Agric Food Chem 36:3742 (1988). 21 Bassil ES and Kaffka SR, Response of safower (Carthamus tinctorius L.) to saline soils and irrigation: II. Crop response to salinity. Agric Water Manag 54:8192 (2002). 22 Zarrouk M, Marzouk B, Ben-Miled D and Cherif A, Oil accumulation in olives and effect of salt on their composition. Olivae 61:4145 (1996). 23 Heuer B, Yaniv Z and Ravina I, Effect of late salinization of chia (Salvia hispanica), stock (Matthiola tricuspidata) and evening primrose (Oenothera biennis) on their oil content and quality. Indian Crop Prod 15:163167 (2002). 24 Lagravere T, Kleiber D and Dayd J, Perfomance agronomique et e conduits culturales du tournesol ol ique. Realites et perspectives. e Ol agineux, Corps Gras, Lipides 5:477485 (1998). e 25 Francois LE and Kleiman R, Salinity effects on vegetative growth, seed yield, and fatty acid composition of crambe. Agron J 82:11101114 (1990). 26 Baldini M, Giovanardi R, Tahmasebi-Enferadi S and Vannozzi GP, Effects of water regime on fatty acid accumulation and nal fatty acid composition in the oil of standard and high oleic sunower hybrids. Ital J Agron 6:119126 (2002).

2615

J Sci Food Agric 2010; 90: 26082616

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org
27 Hoagland DR and Arnon DI, The water culture method for growing plants without soil. Calif Agric Exp Sta Circ. University of California, Berkeley, 347 p. 32 (1950). 28 Lee BL, New AL and Ong CN, Simultaneous determination of tocotrienols, tocopherols, retinol, and major carotenoids in human plasma. Clin Chem 49:20562066 (2003). 29 Snedecor GW and Cochran WG, Statistical Methods (7th edn). The Iowa State University Press, Ames, IA (1980). 30 Ashraf M, Some important physiological selection criteria for salt tolerance in plants. Flora 199:361376 (2004). 31 Munns R and Tester D, Mechanisms of salinity tolerance. Annu Rev Plant Biol 59:651681 (2008). 32 Ashraf M and Akram NA, Improving salinity tolerance of plants through conventional breeding and genetic engineering: an analytical comparison. Biotechnol Adv 27:744752 (2009). 33 Noreen S and Ashraf M, Alleviation of adverse effects of salt-stress on sunower (Helianthus annuus L.) by exogenous application of salicylic acid: growth and photosynthesis. Pak J Bot 40:16571663 (2008). 34 Noreen S, Ashraf M, Hussain M and Jamil A, Exogenous application of salicylic acid enhances antioxidative capacity in salt-stressed sunower (Helianthus annuus L.) plants. Pak J Bot 41:473479 (2009). 35 Aristeo-Cortes P, Reguladores de crecimiento. XIV: Efectos del acido saliclico ydimetilsulfoxido en el crecimiento de zanahoria, betabel y r bano. Tesis de Licenciatura, Facultad de Ciencias UNAM, M xico a e (1998). 36 Hayat S, Ali B and Ahmad A, Salicylic acid biosynthesis, metabolism and physiological role in plants, in Salicylic Acid. A Plant Hormone, ed. by Hayat S and Ahmad A. Springer, Berlin, pp. 114 (2007). 37 Cleland CF and Ajami A, Identication of the ower-inducing factor isolated from aphid honeydew as being salicylic acid. Plant Physiol 54:904906 (1974). 38 Piterse AH, A review of chemically induced owering in Lemma gibba G3 and Pistia stratiotes. Aquat Bot 13:2128 (1982). 39 Kumar P, Lakshmi NJ and Mani VP, Interactive effects of salicylic acid and phytohormones on photosynthesis and grain yield of soybean (Glycine max L. Merrill). Physiol Mol Biol Plants 6:179186 (2000). 40 Seth PN, Venkatarman R and Maheshwari SC, Studies on growth and owering of a short-day plant, Wolfa microscopica. III. Role of metal ions and chelates. Planta 90:349359 (1970). 41 Oota Y, The response of Lemma gibba G3 to a single long day in the presence of EDTA. Plant Cell Physiol 13:575580 (1972). 42 Watanabe K and Takimoto A, Flower inducing effect of benzoic acid and some related compounds in Lamma paucicostata. Plant Cell Physiol 22:14691479 (1979). 43 Fujioka S, Yamaguchi I, Murofushi N, Takahashi N, Kaihara S and Takimoto A, The role of plant hormones and benzoic acid in

S Noreen, M Ashraf
owering of Lamma paucicostata. Plant Cell Physiol 24:241246 (1983). Watanabe K, Fujita T and Takimoto A, Relationship between structure and ower inducing activity of benzoic acid derivatives in Lemma paucicostata. Plant Cell Physiol 20:847850 (1981). Oota Y, Short-day owering of Lamma gibba G3 induced by salicylic acid. Plant Cell Physiol 16:11311135 (1975). Raskin I, Role of salicylic acid in plants. Annu Rev Plant Physiol Plant Mol Biol 43:439463 (1992). Raskin I, Salicylate, a new plant hormone. Plant Physiol 99:799803 (1992). Zhou XM, Mackeuzie AF, Madramootoo CA and Smith DLJ, Effect of some injected plant growth regulators, with or without sucrose, on grain production, biomass and photosynthetic activity of eldgrown corn plants. J Agron Crop Sci 183:103110 (1999). Beke GJ and Volkmar KM, Mineral composition of ax (Linum usitatissimum L.) and safower (Carthamus tinctorius L.) on a saline soil high in sulphate salts. Can J Plant Sci 75:399404 (1995). Heuer B, Ravina I and Davidov S, Seed yield, oil content, and fatty acid composition of stock (Matthiola incana) under saline irrigation. Aust J Agric Res 56:4547 (2005). Kamal-Eldin A and Appelqvist LA, The chemistry and antioxidant properties of tocopherols and tocotrienols. Lipids 31:631701 (1996). Pongracz G, Weiser H and Matzinger D, Tocopherol. Antioxidant der Nat. Fat Sci Technol 97:90104 (1995). Wise RR and Naylor AW, Chilling-enhanced photo-oxidation: evidence for the role of singlet oxygen and endogenous antioxidants. Plant Physiol 83:278282 (1987). Mittova V, Guy M, Tal M and Volokita M, Response of the cultivated tomato and its wild salt-tolerant relative Lycopersicon pennellii to salt-dependent oxidative stress: increased activities of antioxidant enzymes in root plastids. Free Radic Res 36:195202 (2002). Mittova V, Volokita M, Guy M and Tal M, Activities of SOD and the ascorbategluthione cycle enzymes in subcellular compartments in leaves and roots of the cultivated tomato and its wild salt-tolerant relative Lycopersicon pennellii. Physiol Plant 110:4251 (2000). Dat JS, Vandenabeele E, Vranova M, Montagu V, Inze D and Breusegem FV, Dual action of the active oxygen species during plant stress responses. Cell Mol Life Sci 57:779795 (2000). Foyer CH and Noctor G, Redox sensing and signalling associated with reactive oxygen in chloroplasts, peroxisomes and mitochondria. Physiol Plant 119:355364 (2003). Anwar F, Hussain AI, Ashraf M, Jamil A and Iqbal S, Effect of salinity on yield and quality of Moringa oleifera seed oil. Grasas Aceites 57:394401 (2006).

44 45 46 47 48

49 50 51 52 53 54

55

56 57 58

2616
wileyonlinelibrary.com/jsfa
c 2010 Society of Chemical Industry J Sci Food Agric 2010; 90: 26082616

Research Article
Received: 15 April 2010 Revised: 15 July 2010 Accepted: 16 July 2010 Published online in Wiley Online Library: 17 August 2010

(wileyonlinelibrary.com) DOI 10.1002/jsfa.4130

Use of microfungi in the treatment of oak chips: possible effects on wine

Leonardo Petruzzi,a Antonio Bevilacqua,a,b Claudio Ciccarone,c Giuseppe Gambacorta,a,b Giuseppina Irlante,a Sandra Patia,b and Milena Sinigagliaa,b
Abstract
BACKGROUND: Oak barrels are commonly used in the aging of wines and spirits because of their positive effects on the product. In recent years the addition of oak chips has been used to introduce desirable wood aromas and avours into wines. In this study, oak chips in saline solution or laboratory medium were inoculated with Penicillium purpurogenum, Aureobasidium pullulans, Phialemonium obovatum, Phanerochaete chrysosporium and a combination of Ph. chrysosporium and A. pullulans. After 12 weeks of incubation, oak chips (2 g L1 ) were macerated in a red wine for 17 days. Gas chromatography/mass spectrometry and highperformance liquid chromatography were used to evaluate 14 compounds, namely furfural, furfuryl alcohol, guaiacol, syringol, cis--methyl- -octalactone, 2-phenylethanol, 4-vinylguaiacol, benzyl alcohol, 2,3-butanediol, -butyrolactone, benzaldehyde, 4-ethylguaiacol, gallic acid and ellagic acid. RESULTS: The microfungal treatments increased the concentration of some components. In particular, P. purpurogenum resulted in a signicant improvement in the levels of guaiacol, furfural, syringol, furfuryl alcohol and 2-phenylethanol. CONCLUSION: Penicillium purpurogenum and Ph. chrysosporium showed a constant trend (enrichment of furfural and benzaldehyde) independent to some extent of the medium used for chip treatment. c 2010 Society of Chemical Industry Keywords: wood; seasoning; oak chips; fungi; wine aging

INTRODUCTION
Oak wood has traditionally been used in barrel making because of both its mechanical properties and extractable compounds, which can induce positive changes in the composition and avour of aged wine.1 The use of oak barrels for wine aging involves a high cost outlay for wineries; therefore some suitable alternatives to this traditional method have been sought.2 In recent years the addition of oak chips has been used to introduce desirable oak aromas and avours into wines and accelerate the aging process.2 4 Oak chips, obtained from wood scrap wastes produced during barrel manufacturing, are prepared following traditional cooperage methods;5 their use is considered a valid alternative to traditional barrel aging or barrel fermentation.4 Perez-Coello et al.6 reported that the amount of compounds released into the wine depends on the type of wood employed, the seasoning and charring treatments applied, as well as other factors, such as how long the wood is in contact with the wine, the temperature and humidity, etc. The production of staves for barrels or chips includes an outdoor seasoning step that generally takes between 24 and 36 months; this is not a simple process of drying but a rening stage, since the wood undergoes slow chemical and biochemical transformations of biopolymers and extractable compounds by fungi and bacteria that contribute to its maturation.7 The importance of mycotic wood infection is a matter discussed as a main theme in several studies.8 11

Focusing on the effect of fungi, basidiomycetes are the most important organisms able to degrade lignocelluloses, mainly the white rot and related litter-decomposing fungi.12 In particular, Phanerochaete chrysosporium is the most studied among the wood-rotting fungi owing to its production of two peroxidases, lignin peroxidase (LP) and manganese peroxidase (MnP), known to play an important role in the degradation of lignin.13 Vivas et al.10 studied the evolution of the microora of wood staves throughout seasoning and found that 83% of the total fungal community was represented by Aureobasidium pullulans, followed by 15% of Trichoderma harzianum and Trichoderma konigii. These three fungi have an enzymatic activity able to

Correspondence to: Antonio Bevilacqua, Department of Food Science, Faculty of Agricultural Science, University of Foggia, Via Napoli 25, I-71122 Foggia, Italy. E-mail: a.bevilacqua@unifg.it; abevi@libero.it

Current address: Dipartimento di Biologia e Chimica Agro-forestale e Ambientale (DiBCA), University of Bari, Via Amendola 165, Bari, Italy. a Department of Food Science, Faculty of Agricultural Science, University of Foggia, Via Napoli 25, I-71122 Foggia, Italy b Food Quality and Health Research Center (BIOAGROMED), University of Foggia, Via Napoli 52, I-71122 Foggia, Italy

2617

c Department of Agro-Environmental, Chemistry and Crop-Protection, Faculty of Agricultural Science, University of Foggia, Via Napoli 25, I-71122 Foggia, Italy

J Sci Food Agric 2010; 90: 26172626

www.soci.org

c 2010 Society of Chemical Industry

www.soci.org hydrolyse many wood heterosides (ellagitannins, coumarins and polysaccharides), thus resulting in a decrease in astringency and bitterness.9,11 In particular, A. pullulans is a ubiquitous species producing cellulosolytic, pectinolytic and ligninolytic enzymes and is able to adapt to different environmental conditions. After 2436 months of stave seasoning, it is possible to note an evolution of the fungal community, with an increase in species of the genera Penicillium, Geomyces and Geotrichum; throughout this period, Penicillium purpurogenum is the secondary species most represented.10 This fungus is an active producer of a variety of xylan-degrading enzymes and several esterases.14 An ecological study carried out by Roulland et al.8 revealed that Candida sp., Paecilomyces variotii, Phialemonium sp. and the strain E were the moulds that colonised the internal layers of staves. The authors measured the ability of these species to degrade cellulose, xylan and aesculin: in a synthetic liquid medium the four fungal species degraded aesculin and xylan, but only Phialemonium sp. showed a measurable growth rate in a cellulose medium.8 Some authors hypothesised the possibility to inoculate staves with selected mould strains to control metabolic reactions of fungi; in particular, the use of A. pullulans was suggested.11 13 The use of selected fungal inocula on a large scale could lead to the creation of new seasoning places. Jourez et al.15 reported that a speeding-up of the aging of staves is possible by treatment with a solution of natural enzymes obtained from fungal species usually colonising wood staves; this enzymatic treatment could be considered a very interesting economical alternative, permitting a reduction in storage time of the shooks from 1236 months to only 1 month before assembly. Therefore, as an alternative to articial drying, this paper proposes the treatment of chips with four fungi (P. purpurogenum, A. pullulans, Phialemonium obovatum and Ph. chrysosporium) and a combination of two of them (Ph. chrysosporium + A. pullulans), representative of both the primary and secondary microora recovered in the natural seasoning of staves. This treatment was proposed to improve the impact of oak chips in the maceration of red wine and attain an effect comparable to that observed in barrel aging. The chips were inoculated with the aforementioned fungi both in saline solution and in laboratory medium and used after 12 weeks for the accelerated aging of a red wine in order to study the effect of the fungal inocula on the concentration of volatile compounds (furfural, furfuryl alcohol, guaiacol, syringol, cis--methyl- -octalactone, 2-phenylethanol, 4-vinylguaiacol, benzyl alcohol, 2,3-butanediol, -butyrolactone, benzaldehyde and 4-ethylguaiacol) and gallic and ellagic acids in the wine. Gas chromatography/mass spectrometry (GC/MS) and high-performance liquid chromatography (HPLC) were used to evaluate these 14 compounds.

L Petruzzi et al.

extract agar (Difco Laboratories) plates and kept at 70% relative humidity in complete darkness as follows: Ph. chrysosporium was incubated for 7 days at 37 C; P. purpurogenum was incubated for 7 days at 28 C; A. pullulans and Phi. obovatum were incubated for 14 days at 24 C. Afterwards, an agar plug (8 mm in diameter, cut along the edge of an actively growing colony) of each fungus was used for preparation of the inoculum for oak chips (reported in the following subsections and divided into different steps: preculture, fungal inoculum preparation and oak chip inoculation). Preculture Fungi were grown on Petri plates (90 mm in diameter) with different lignocellulosic supplements: wheat bran (Ph. chrysosporium and A. pullulans), wheat straw (P. purpurogenum) and oat bran (Phi. obovatum). The medium was prepared according to Moredo et al.16 and comprised 10 g L1 glucose (J.T. Baker, Milan, Italy), 15 g L1 agar technical No. 3 (Oxoid, Milan, Italy), 3.5 g L1 malt extract (Oxoid) and 7.5 g L1 triturated lignocellulosic material. Fungi were incubated at 30 C for 7 days (Ph. chrysosporium and P. purpurogenum), 10 days (A. pullulans) or 14 days (Phi. obovatum). Fungal inoculum preparation Each inoculum was prepared in several steps. First, the central part (5 cm in diameter) of the colony cultivated on a preculture plate was blended for 30 s with 50 mL of sterile distilled water in a laboratory blender (Stomacher 400, PBI International, Milan, Italy). Then 5 mL of homogenate was added to a 250 mL Erlenmeyer ask containing 50 mL of growth medium (sterilised in an autoclave at 121 C for 15 min) (see Appendix). All experiments were done in duplicate. Cultures were incubated statically (Ph. chrysosporium, 37 C for 7 days; P. purpurogenum, 37 C for 7 days; Phi. obovatum, 28 C for 6 days) or with agitation on a rotary shaker (A. pullulans, 28 C for 35 days, 200 rpm). The asks, capped with cellulose stoppers to permit passive aeration, were maintained in complete darkness at 90% relative humidity. Afterwards, fungus and medium were aseptically homogenised in a blender (3 20 s cycles with 1 min interval). The resulting homogenate (5 mL) was added to a 250 mL Erlenmeyer ask containing 50 mL of growth medium (sterilised in an autoclave at 121 C for 15 min). Cultures were incubated statically (Ph. chrysosporium, 30 C for 5 days) or with agitation (P. purpurogenum, 30 C for 57 days, 135 rpm; Phi. obovatum, 28 C for 7 days, 150 rpm; A. pullulans, 28 C for 7 days, 200 rpm). Afterwards, fungus and medium were aseptically homogenised in a blender (Ph. chrysosporium and P. purpurogenum, 3 20 s cycles with 1 min interval; A. pullulans and Phi. obovatum, 20 s). The nal homogenate (fungus + medium) of each fungal species was used to inoculate the oak chips. Oak chip inoculation French chips toasted at low degree and of medium size (F4L Bois elevage Lavoisier, AEB Group SpA, Bologna, Italy) were used. A 3 mL aliquot of homogenate was added to a 250 mL Erlenmeyer ask containing 4 g of oak chips and 12 mL of production medium (laboratory medium) (see Appendix) or 12 mL of saline solution; chips were sterilised with laboratory medium or saline solution at 121 C for 15 min. The saline solution contained 3 g L1 NH4 NO3 , 3 g L1 KH2 PO and 0.5 g L1 MgSO4 7H2 O;17 all components were purchased from J.T. Baker.

EXPERIMENTAL
Oak chip treatment Micro-organisms The following fungi purchased from the culture collection of Mycotheca Universitatis Taurinensis (MUT), University of Torino, Italy were used throughout this study: Ph. chrysosporium Burds. (MUT 2660), P. purpurogenum Stoll (MUT 3316), A. pullulans (de Bary) G. Arnaud (MUT 3237) and Phi. obovatum W. Gams & McGinnis (MUT 2702). All fungi were non-mycotoxinogenic. Fungi were maintained on potato dextrose agar (Difco Laboratories, Detroit, MI, USA) plates at 4 C, then revived on malt

2618

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 26172626

Oak chips inoculated with moulds Cultures were incubated for 12 weeks in darkness under static conditions at different temperatures: 2426 C (P. purpurogenum, A. pullulans and Phi. obovatum) or 30 C (Ph. chrysosporium). Each experiment was performed twice (on two independent batches). Aliquots of laboratory medium or saline solution containing oak chips but not inoculum were used as controls. In addition to the single inocula, a combination of A. pullulans and Ph. chrysosporium was investigated. Oak chips were rst inoculated with A. pullulans for 6 weeks, then sterilised, charged with laboratory medium or saline solution and inoculated with Ph. chrysosporium for 6 weeks. Wine aging After the 12 week incubation period, oak chips were removed from the asks, cleaned with a small brush to remove visible mycelia and used for aging red wine elaborated by the traditional red winemaking process from two grape varieties, Montepulciano dAbruzzo (70%) and Merlot (30%) (2006 vintage), from vineyards in southern Italy (San Severo, Apulia). Articial aging was performed by placing 1 g (dry weight) portions of oak chips in 500 mL glass bottles containing red wine and storing the bottles in a cool dark room (temperature of about 20 C at 82% relative humidity) for 17 days. Chemical analyses Identication and quantication of volatile compounds by GC and GC/MS The method of Di Stefano,18 appropriately modied, was used for the extraction and concentration of volatile compounds (furfural, furfuryl alcohol, guaiacol, syringol, cis--methyl- -octalactone, 2-phenylethanol, 4-vinylguaiacol, benzyl alcohol, 2,3-butanediol, -butyrolactone, benzaldehyde and 4-ethylguaiacol). Liquid/liquid extraction was carried out in a glass ask at room temperature under magnetic stirring, using 7.5 mL of dichloromethane (J.T. Baker) as extracting solvent, 0.25 mL of 200 mg L1 2-octanol solution (Fluka, Buchs, Switzerland) as internal standard and 50 mL of wine. The emulsion was recovered after 1 h of extraction and frozen for 2 h at 20 C. After freezing, the organic phase was passed through two separation funnels (100 and 25 mL respectively), dried over Na2 SO4 (J.T. Baker) and concentrated to 1.5 mL under a stream of pure N2 (0.5 L min1 ) for GC and GC/MS analyses. A 1 L aliquot of concentrated extract was injected in splitless mode into an Agilent 6890N gas chromatograph (Agilent Technologies, Palo Alto, CA, USA) equipped with a split/splitless injector, a DB-Wax fused silica capillary column (60 m length, 0.25 mm i.d., 0.25 m lm thickness; J&W Scientic, Folsom, CA, USA) and a ame ionisation detector. Helium was used as the carrier gas at a linear velocity of 37 cm s1 . The oven temperature was programmed as follows: 40 C for 3 min; from 40 to 200 C at a rate of 4 C min1 ; 200 C for 20 min. The detector and injector temperatures were 240 and 230 C respectively. Identication of compounds was performed using an Agilent 5975C quadrupole mass spectrometer coupled with an Agilent 6890N gas chromatograph. The same column and same temperature programme as for GC analysis were employed. Electron impact mass spectra were recorded with an ion source energy of 70 eV. Chromatographic peaks were identied by comparing their mass spectra with those of standards (Sigma-Aldrich, St Louis, MO, USA) and those reported in the NIST 2.0 commercial library.

www.soci.org Identication and quantication of gallic and ellagic acids by HPLC Acid hydrolysis was used to determine the content of gallic and ellagic acids,19 modied as follows: 7.5 mL of HCl was added to 10 mL of wine and heat-shocked for 3 h in a water bath held at 100 C. After ltration through a 13 mm polytetrauoroethylene lter (0.20 m pore size), a 10 L sample was directly injected into an Agilent 1200 HPLC system equipped with an autosampler, a binary pump, a diode array detector and a 4 m Synergi MAX-RP C12 column (150 mm 2 mm; Phenomenex, Milan, Italy). The eluent was fed at 0.2 mL min1 and consisted of two solvents: A, water; B, acetonitrile. Both solvents were acidied with 10 mL L1 formic acid for the determination of ellagic acid and with 20 mL L1 formic acid for the determination of gallic acid. Acetonitrile and water were of HPLC ultra gradient grade and were purchased from J.T. Baker. The elution gradient for the determination of ellagic acid was as follows: from 0 to 4 min, 18% B; from 4 to 14 min, 20% B; then an increase to 90% B (time 14.2 min). The elution gradient for the determination of gallic acid was as follows: from 0 to 10 min, 1% B; then an increase to 95% B (time 10.2 min). The spectra were obtained by scanning from 225 to 390 nm. Ellagic acid was detected at 250 nm and gallic acid at 280 nm. Peaks were identied via spectra and retention times of standards (Sigma-Aldrich). Chemstation Version 03.01 (Agilent Technologies) was used for data acquisition and analysis. Quantication was achieved by means of a calibration curve obtained by injecting solutions with variable and known amounts of standards in order to cover the desired concentration range. The main calibration characteristics are reported in Table 1. Statistical analysis All analyses were performed in duplicate over two different batches. Data were submitted to one-way analysis of variance (ANOVA) and Tukeys test using Statistica for Windows (StatSoft, Tulsa, OK, USA). In addition, differences among samples were determined by principal component analysis (PCA) using the Excel component XLSTAT 7 (Addinsoft, Paris, France).

RESULTS AND DISCUSSION

It is known that some compounds, namely furfural, furfuryl alcohol, guaiacol, syringol and cis--methyl- -octalactone, are present in wine largely as a result of their extraction from oak wood. Other compounds such as 2-phenylethanol and 4-vinylguaiacol are to some extent fermentation products, but oak aging has been shown to increase levels of these chemicals to a certain degree. Benzyl alcohol, 2,3-butanediol, -butyrolactone and benzaldehyde are grape or fermentation products and are not known to be affected by oak aging.20 4-Ethylguaiacol, produced by the contaminant yeasts Brettanomyces and Dekkera, is related to the Brett character of spoiled wines and has been associated with descriptive expressions such as bacon or smoked.21 Gallic and ellagic acids were studied in this research because they contribute to wine astringency and bitterness.22 It is interesting to note that cis--methyl- -octalactone, which is associated with oaky, coconut and vanilla notes, appeared in trace concentrations in all samples.4,23 The concentration of cis-methyl- -octalactone is strongly related to the origin of the

2619

J Sci Food Agric 2010; 90: 26172626

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org

L Petruzzi et al.

Table 1. Main calibration( characteristics related to ellagic and gallic acid determination Compound Ellagic acid Gallic acid Retention time (min) 8.15 6.12 Linear regression equation y = 268.06x 136.58 (R2 = 0.9987) y = 177.62x 2.91 (R2 = 0.9999) Detection limit (S/N = 3) mg L1 ) 0.6 0.07 Concentration range of calibration (mg L1 ) 110 210

S/N, signal/noise ratio; R2 , coefcient of determination.

Table 2. Chemicals recovered in wine aged with oak chips treated in (a) saline solution (S) or (b) laboratory medium (M) and inoculated with fungi (a) Compound Furfural (g Guaiacol (g L1 ) Syringol (g L1 ) Furfuryl alcohol (g L1 ) 2-Phenylethanol (mg L1 ) 4-Vinylguaiacol (g L1 ) Benzyl alcohol (mg L1 ) 2,3-Butanediol (mg L1 ) -Butyrolactone (mg L1 ) Benzaldehyde (g L1 ) 4-Ethylguaiacol (g L1 ) Ellagic acid (mg L1 ) Gallic acid (mg L1 ) cis--Methyl- -octalactone (g L1 ) (b) Compound Furfural (g L1 ) Guaiacol (g L1 ) Syringol (g L1 ) Furfuryl alcohol (g L1 ) 2-Phenylethanol (mg L1 ) 4-Vinylguaiacol (g L1 ) Benzyl alcohol (mg L1 ) 2,3-Butanediol (mg L1 ) -Butyrolactone (mg L1 ) Benzaldehyde (g L1 ) 4-Ethylguaiacol (g L1 ) Ellagic acid (mg L1 ) Gallic acid (mg L1 ) cis--Methyl- -octalactone (g L1 ) TM 16.5 2.5 2.85 0.77 35.5 3.53 11.5 2.12 0.88 0.11 26 2.82 0.16 0.007 10.39 0.74 0.49 0.02 4 1.41 48.5 9.19 8.21 1 9.34 0.25 Tr PCM 168.5 5.5 8.5 3.53 123.5 0.7 255 35.35 10.49 0.57 123.5 0.7 1.88 0.02 1.5 0.62 6.19 0.14 88 12.72 53 1.41 1.37 0.22 8.31 0.89 Tr PPM 185 5 50 11.31 580 84.85 230 28.28 47.99 0.14 230 55.15 1.93 0.07 2.68 0.72 6.26 1.15 45 8.48 63.5 7.77 6.65 1.83 9.25 1.76 Tr APM 14 3 119 15.55 725.5 92.63 120 28.28 10.06 2.04 228.5 45.96 2.26 0.35 3.31 1.15 5.36 0.007 55.5 12.02 56.5 6.36 1.59 0.02 9.23 0.14 Tr POM 23.5 4.5 17.5 3.53 139.5 57.27 106.5 9.19 11.28 0.21 210.5 38.89 4.21 0.65 5.34 1.22 7.96 0.29 9.5 2.12 65 7.07 1.36 0.43 3.95 1.24 Tr A+PM 21.5 2.5 15 7.07 225 63.63 105 21.21 8.58 1.52 145.5 44.54 3.5 0.68 7.44 1.04 6.11 1.07 79.5 16.26 83.5 2.12 1.97 0.38 3.96 0.1 Tr L1 ) TS 22.45 0.85 13.5 3.53 227.5 45.96 100 14.14 9.57 0.26 66 31.11 3.58 0.16 11.01 0.97 0.4 0.17 42.5 9.19 57.5 10.6 7.18 0.86 8.72 0.36 Tr PCS 113.5 11.5 11.5 3.53 232.5 38.89 43.5 10.6 9.51 1.52 213 5.65 1.59 0.02 3.8 0.39 6.65 0.78 74.5 9.19 75 7.07 2.81 0.8 7.39 0.64 Tr PPS 130 10 44.5 3.53 615 91.92 57.5 17.67 10.47 1.73 162 26.87 1.62 0.13 3.26 1.49 5.6 0.17 32.5 9.19 55 2.82 3.22 0.78 7.81 1.44 Tr APS 25.5 1.5 62.5 6.36 362 59.39 116 2.82 9.34 0.19 342.5 45.96 2.62 0.09 9.99 0.02 6.98 0.01 6 2.82 56.5 6.36 1.32 0.25 4.75 0.35 Tr POS 16.5 1.5 50.5 9.19 502.5 51.61 72 1.41 11.67 0.65 401 24.04 2.62 0.28 3.66 1.04 7.19 0.26 4 1.41 65.5 9.19 1.14 0.5 3.49 0.58 Tr A+PS 20 7 19 7.07 269.5 71.41 115 7.07 8.2 0.65 376.5 38.98 2.14 0.42 4.92 1.01 5.27 0.07 7.1 0.14 76 7.07 2.14 0.44 5.04 1.08 Tr

Data are the mean of two values standard deviation. TS/TM, control; PCS/PCM, Ph. chrysosporium; PPS/PPM, P. purpurogenum; APSS/APM, A. pullulans; POS/POM, Phi. obovatum; A+PS/A+PM, A. pullulans + Ph. chrysosporium; Tr, trace amount.

wood and the country of seasoning.23 Moreover, Guchu et al.4 reported that non-toasted oak chips released more oak lactones in wine than toasted oak chips did, probably owing to the thermal degradation of these heat-sensitive compounds or their loss by volatilisation when oak wood is treated at very high temperatures. Concerning volatile compounds and phenols, their concentration was inuenced by the kind of medium and the mould used for the chip treatment. Table 2 reports the concentrations of wine compounds. The chemicals for which an inuence of the fungal inoculum was found are discussed in the following sections.

Furfural The concentration of furfural was signicantly affected by Ph. chrysosporium (PC) and P. purpurogenum (PP) and increased from 22.45 to 113.5130 g L1 and from 16.5 to 168.5185 g L1 for saline solution- and laboratory medium-treated oak chips respectively, without attaining the perception threshold level (14 100 g L1 )24 (Fig. 1). In the other samples, no signicant changes were recorded. Furfural contributes to the character of dried fruits, particularly to that of burned almonds.3 Owing to the high threshold, Garde-Cerd n et al.25 reported that this a

2620

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 26172626

Oak chips inoculated with moulds

250 225 200 175 Amount (g L-1) 150 b 125 100 75 50 25 0 a a a a b

www.soci.org

Perception threshold (1/60)

b b

a a

TS

PCS PPS APS POS A+PS

TM PCM PPM APM POM A+PM

Figure 1. Amounts ( standard deviation) of furfural in wine aged with oak chips treated in saline solution (S) or laboratory medium (M) and inoculated with fungi: T, control; PC, Phanerochate chrysosporium; PP, Penicillium purpurogenum; AP, Aureobasidium pullulans; PO, Phialemonium obovatum; A+P, A. pullulans + Ph. chrysosporium. For each kind of chips (S or M), bars with different letters are signicantly different (one-way ANOVA and Tukeys test, P < 0.05). Perception threshold: the line represents a fraction of the real breakpoint (1/60).

compound does not play an important role in the aroma of wine, although it might strengthen the aroma of lactones. Furfural is formed by the degradation of hemicelluloses during wood toasting and is an indicator of the relative toast level of wood.3,24 Therefore Spillman et al.26 reported that furfural appears to be the most susceptible of the various oak wood volatile compounds to microbial transformations, being almost totally transformed during alcoholic and malolactic fermentation and during red wine maturation. Guaiacol The treatment of oak chips with fungi caused a signicant increase in the content of guaiacol. In particular in the case of saline solution (Fig. 2), the increase was observed for chips inoculated with P. purpurogenum (PPS), A. pullulans (APS) and Phi. obovatum (POS) (the average value of this compound was 44.562.5 g L1 , compared with 13.5 g L1 for non-treated chips). On the other hand, in the case of laboratory medium the differences were signicant only for P. purpurogenum (PPM) and A. pullulans (APM) (50 and 119 g L1 respectively). It is interesting to note that in the case of A. pullulans the content of guaiacol was above its perception threshold (75 g L1 ).27 Guaiacol is produced by the breakdown of lignin during wood toasting and is responsible for the burnt overtones of wine aroma.3 Syringol For saline solution-treated chips the increase in syringol was signicant only in the case of P. purpurogenum (PPS) (615 g L1 ) in comparison with the control (227.5 g L1 ) (Fig. 3). In laboratory medium the control showed a syringol content of 35.5 g L1 , while higher levels (above the perception threshold of 570 g L1 )28 were found in wine treated with oak chips inoculated with P. purpurogenum (PPM) (580 g L1 ) and A. pullulans (APM) (725.5 g L1 ).

Syringol is synthesised during wood toasting and is an indicator of the relative toast level of wood. In comparison with guaiacol, syringol has a weak odour and little impact on wine avour.29

Benzaldehyde In saline solution the fungal treatment decreased the concentration of benzaldehyde (ranging between 4 and 32.5 g L1 , while in the control its content was 42.5 g L1 ), with the sole exception of Ph. chrysosprorium (PCS) (74.5 g L1 ) (Fig. 4). In the case of oak chips aged in laboratory medium, a signicant increase in benzaldehyde was observed (4588 g L1 , compared with 4 g L1 in the control). A characteristic bitter almond odour is generally attributed to the presence of benzaldehyde, which has a perception threshold of 20 mg L1 .30 Benzaldehyde in wine is probably formed by the oxidation of benzyl alcohol,31 which is occasionally used as a plasticiser in epoxy resins or found as a contaminant in liquid gelatin.32 Benzaldehyde in wine is also formed by the action of micro-organisms on aromatic amino acids (e.g. phenylalanine), phenol compounds of the grape or some secondary compounds such as phenyl acetic acid and p-hydroxybenzoic acid.31 Several micro-organisms can oxidise benzyl alcohol to benzaldehyde, including Botrytis cinerea and the yeasts Schizosaccharomyces pombe and Zygosaccharomyces bailii.32

2,3-Butanediol Figure 5 shows the content of 2,3-butanediol. Saline solution and laboratory medium showed similar trends. In saline solution the level of this chemical was lower in wine samples aged with oak chips inoculated with fungi (3.264.92 mg L1 , compared with 11.01 mg L1 in the control), with the sole exception of A. pullulans (APS). Similar results were obtained for oak chips treated in laboratory medium.

2621

J Sci Food Agric 2010; 90: 26172626

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org
180

L Petruzzi et al.

150 c 120 Amount (g L-1)

90
Perception threshold

b b b

60

b a a a a a a

30

TS

PCS PPS APS POS A+PS

TM PCM PPM APM POM A+PM

Figure 2. Amounts ( standard deviation) of guaiacol in wine aged with oak chips treated in saline solution (S) or laboratory medium (M) and inoculated with fungi: T, control; PC, Phanerochate chrysosporium; PP, Penicillium purpurogenum; AP, Aureobasidium pullulans; PO, Phialemonium obovatum; A+P, A. pullulans + Ph. chrysosporium. For each kind of chips (S or M), bars with different letters are signicantly different (one-way ANOVA and Tukeys test, P < 0.05).
900 b 800 c 700
Perception threshold

b b, c

600 Amount (g L-1) 500 a, b 400 300 200 a a

a, b a a a

100 0

a TS PCS PPS APS POS A+PS TM PCM PPM APM POM A+PM

Figure 3. Amounts ( standard deviation) of syringol in wine aged with oak chips treated in saline solution (S) or laboratory medium (M) and inoculated with fungi: T, control; PC, Phanerochate chrysosporium; PP, Penicillium purpurogenum; AP, Aureobasidium pullulans; PO, Phialemonium obovatum; A+P, A. pullulans + Ph. chrysosporium. For each kind of chips (S or M), bars with different letters are signicantly different (one-way ANOVA and Tukeys test, P < 0.05).

Although 2,3-butanediol is odourless, as its perception threshold is very high (600 mg L1 ), it contributes to the sweet taste in wine.33 Bartowsky and Henschke34 reported that 2,3-butanediol is a stable compound derived from the reduction of acetoin. Inuence of chip fungal treatment on ellagic and gallic acids Figure 6 presents the results for ellagic acid. In saline solution this compound was found at mean levels ranging from 1.14 to 3.22 mg L1 (compared with 7.18 mg L1 for the control), while in laboratory medium it varied from from 1.36 to 1.97 mg L1 (compared with 8.21 mg L1 for the control), thus suggesting

that ellagic acid was probably metabolised by fungi, the only exception being represented by P. purpurogenum in laboratory medium (PPM) (6.65 mg L1 ). Similar results were found in the case of gallic acid (data not shown). A possible explanation can be found in the paper of Roulland et al.,8 who reported that some fungi (in particular Phialemonium sp.) were able to metabolise these two acids. Principal component analysis As a nal step in this paper we propose a multivariate approach to try to give a simple insight into a complex phenomenon, such as the

2622

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 26172626

Oak chips inoculated with moulds

150
Perception threshold (1/15)

www.soci.org

125

c
Amount (g L-1) 100

b, c

d b, c c b, c b

75

50

25

a
0 TS

PCS PPS APS POS A+PS

TM PCM PPM APM POM A+PM

Figure 4. Amounts ( standard deviation) of benzaldehyde in wine aged with oak chips treated in saline solution (S) or laboratory medium (M) and inoculated with fungi: T, control; PC, Phanerochate chrysosporium; PP, Penicillium purpurogenum; AP, Aureobasidium pullulans; PO, Phialemonium obovatum; A+P, A. pullulans + Ph. chrysosporium. For each kind of chips (S or M), bars with different letters are signicantly different (one-way ANOVA and Tukeys test, P < 0.05). Perception threshold: the line represents a fraction of the real breakpoint (1/15).
24 21 18 Amount (g L-1) 15 a 12 9 b 6 3 0 b b b a, b a a, b b, c a c, d d 1/60

Perception threshold (1/30)

TS

PCS PPS APS POS A+PS

TM PCM PPM APM POM A+PM

Figure 5. Amounts ( standard deviation) of 2,3-butanediol in wine aged with oak chips treated in saline solution (S) or laboratory medium (M) and inoculated with fungi: T, control; PC, Phanerochate chrysosporium; PP, Penicillium purpurogenum; AP, Aureobasidium pullulans; PO, Phialemonium obovatum; A+P, A. pullulans + Ph. chrysosporium. For each kind of chips (S or M), bars with different letters are signicantly different (one-way ANOVA and Tukeys test, P < 0.05). Perception threshold: the line represents a fraction of the real breakpoint (1/30).

pretreatment with some selected fungi and the chip effect on the chemical prole of phenols and volatile compounds of wine. Only those compounds showing signicant differences among samples, i.e. furfural, benzaldehyde, 2,3-butanediol, guaiacol, syringol, gallic acid and ellagic acid, were used as input data. The results are shown in Fig. 7. The different samples are represented as a function of components 1 and 2, which together accounted for 79% (saline solution, Fig. 7(a)) and 68% (laboratory medium, Fig. 7(b)) of the total variance. Table 3 lists the correlation coefcients of each compound with axis components. In the case of saline solution, benzaldehyde, guaicol, ellagic acid and gallic acid were associated with principal component

1 (PC1 ); on the other hand, principal component 2 showed a good correlation with 2,3-butanediol. Finally, syringol showed a partial correlation with both PC1 and PC2 (0.571 and 0.599 respectively). Figure 7(a) shows the distribution of the different samples in a 2D plot. As evidenced by this score plot, the fungal treatment of chips signicantly affected the chemical prole of wine, resulting in a distribution into two groups, i.e. group A represented by Phi. obovatum (POS), A. pullulans + Ph. chrysosporium (A+PS) and A. pullulans (APS) and group B represented by P. purpurogenum (PPS) and Ph. chrysosporium (PCS). Samples treated with oak chips of group A showed an increase in guaiacol and syringol levels,

2623

J Sci Food Agric 2010; 90: 26172626

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org
12

L Petruzzi et al.

10 b Amount (mg L-1) 8

b b

6 a a 2 a a a a a a

TS

PCS PPS APS POS A+PS

TM PCM PPM APM POM A+PM

Figure 6. Amounts ( standard deviation) of ellagic acid in wine aged with oak chips treated in saline solution (S) or laboratory medium (M) and inoculated with fungi: T, control; PC, Phanerochate chrysosporium; PP, Penicillium purpurogenum; AP, Aureobasidium pullulans; PO, Phialemonium obovatum; A+P, A. pullulans + Ph. chrysosporium. For each kind of chips (S or M), bars with different letters are signicantly different (one-way ANOVA and Tukeys test, P < 0.05).

whereas samples treated with oak chips of group B showed higher concentrations of furfural and benzaldehyde. The effect of the fungi appeared to be quite variable and was inuenced by the kind of medium used for the treatment of chips, as evidenced by the score plot in Fig. 7(b). In particular, in the case of oak chips treated with laboratory medium, component 1 was associated with 2,3-butanediol, guaiacol and syringol, while gallic acid and ellagic acid showed high correlation coefcients with component 2. Although the variance explained for laboratory medium-treated chips was lower (68%), the PCA showed some qualitative trend, conrming the data obtained in saline solution, but with some differences. The samples were again distributed into two groups, i.e. group C represented by Phi. obovatum (POM) and A. pullulans + P. purpurogenum (A+PM) and group D represented by P. purpurogenum (PPM), A. pullulans (APM) and Ph. chrysosporium (PCM). Group D was characterised by higher concentrations of furfural, benzaldehyde, guaiacol and syringol, thus conrming the effects of P. purpurogenum and Ph. chrysosporium.

aspect should be considered as a fundamental prerequisite for the selection of suitable strains. Further investigations are required to provide a more complete understanding of the biology and enzymatic prole of fungi. The present promising results suggest the application of this methodology on the one hand to determine the extent of the relationship between P. purpurogenum and oak chips as a function of their different sizes and degrees of toasting and on the other hand to study other fungi (e.g. Trichoderma sp.) that colonise oak wood more or less deeply during seasoning or other white rot fungi such as Bjerkandera adusta and Trametes versicolor.

APPENDIX: MEDIA FOR FUNGI

Growth media A liquid medium adapted from Keller et al.35 was used for Ph. chrysosporium. It consisted of 3 g L1 NaNO3 , 0.5 g L1 KCl, 0.5 g L1 MgSO4 7H2 O, 0.5 g L1 FeSO4 7H2 O, 1 g L1 KH2 PO4 , 20 g L1 glucose and 1 g L1 wheat bran. All components were purchased from J.T. Baker. The liquid medium described by Kheng and Omar36 was used for P. purpurogenum, with some modication. It was prepared with the following composition: 0.3 g L1 urea (Sigma-Aldrich), 0.7 g L1 bacteriological peptone (Oxoid), 0.2 g L1 yeast extract (Oxoid), 1 g L1 KH2 PO4 , 0.3 g L1 CaCl2 (J.T. Baker), 1.4 g L1 (NH4 )2 SO4 (J.T. Baker), 0.3 g L1 MgSO4 7H2 O, 10 g L1 glucose, 1 g L1 wheat straw, 5 mg L1 FeSO4 7H2 O, 1.6 mg L1 MnSO4 4H2 O, 1.4 mg L1 ZnSO4 7H2 O (J.T. Baker) and 20 mg L1 CoCl2 6H2 O (J.T. Baker). Aureobasidium pullulans was grown on the liquid medium reported by Lee et al.,37 modied as follows: 2.5 g L1 yeast extract, 0.6 g L1 (NH4 )2 SO4 , 5 g L1 KH2 PO4 , 1 g L1 NaCl, 0.2 g L1 MgSO4 7H2 O, 0.2 g L1 chloramphenicol (C. Erba, Milan, Italy), 50 g L1 glucose and 1 g L1 wheat bran. A liquid medium adapted from De Jong et al.38 was used for Phi. obovatum. It contained 5 g L1 bacteriological peptone, 3 g L1

CONCLUSIONS
Oak chips are a suitable alternative to barrels to obtain, in a much shorter period of time (17 days), wine with the peculiar characteristics given by oak wood. Based on the results reported in this paper, we could suggest that microfungal treatment of chips increased the concentration of some components during the aging period of wine. In particular, some interesting data were obtained in the case of P. purpurogenum and Ph. chrysosporium, as they showed a constant trend (enrichment of furfural and benzaldehyde) independent to some extent of the medium used for chip treatment. Objections to the proposed method could arise owing to the possible production of toxins by fungi; however, the fungi used throughout this study are non-mycotoxigenic. Moreover, this

2624

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 26172626

Oak chips inoculated with moulds

www.soci.org
B PPS

(a)

3 2.5 2

-- Component 2 (30%) -->

1.5 1 0.5 0 -0.5 A+PS -1 -1.5 -2 -2.5 -3 -2 A APS 3 POS 4 7 1 5 2 6

PCS

TS

-1

-- Component 1 (49%) --> (b) 2.5 2 -- Component 2 (27%) --> 1.5 1 0.5 0 -0.5 -1 -1.5 -2 -3 -2 -1 0 -- Component 1 (41%) -->
Figure 7. PCA of chemical prole of wine aged with oak chips treated in (a) saline solution (S) or (b) laboratory medium (M) and inoculated with fungi: T, control; PC, Phanerochate chrysosporium; PP, Penicillium purpurogenum; AP, Aureobasidium pullulans; PO, Phialemonium obovatum; A+P, A. pullulans + Ph. chrysosporium. Compounds: 1, furfural; 2, benzaldehyde; 3, 2,3-butanediol; 4, guaiacol; 5, syringol; 6, gallic acid; 7, ellagic acid.
2 PCM A+PM POM C APM 5 4 1 D PPM 6 7 3 TM

Table 3. Compounds associated with principal components 1 (PC1 ) and 2 (PC2 ) in PCA: correlation coefcients S Compound Furfural (1) Benzaldehyde (2) 2,3-Butanediol (3) Guaiacol (4) Syringol (5) Gallic acid (6) Ellagic acid (7) PC1 0.372 0.855 0.230 0.834 0.571 0.905 0.828 PC2 0.890 0.381 0.826 0.071 0.599 0.185 0.281 PC1 0.528 0.601 0.885 0.726 0.808 0.414 0.344 M PC2 0.082 0.535 0.225 0.271 0.306 0.792 0.858

yeast extract, 1 g L1 K2 HPO4 , 0.5 g L1 MgSO4 7H2 O, 0.2 g L1 chloramphenicol, 20 g L1 glucose and 1 g L1 oat bran. Production media The composition of the production media was almost the same as reported for the growth media; however, they did not contain lignocellulosic materials and there were some slight differences. The liquid medium for P. chrysosporium was supplemented with 0.5 g L1 asparagine (Sigma-Aldrich), 1 mg L1 vitamin B1 (Sigma-Aldrich) and 1 mL L1 Tween 80 (Biolife, Milan, Italy); the content of glucose was only 10 g L1 . The liquid medium for P. purpurogenum was supplemented with 0.2 mL L1 Tween 80. The liquid medium for A. pullulans was supplemented with 5 mL L1 Tween 80; the content of glucose was only 30 g L1 . In the liquid medium for Phi. obovatum the content of MgSO4 7H2 O was 4 g L1 and the content of KH2 PO4 was 5 g L1 ; the content of glucose was only 15 g L1 .

S, oak chips treated in saline solution; M, oak chips treated in laboratory medium. The numbers in parentheses correspond to compound numbers in Fig. 7.

2625

J Sci Food Agric 2010; 90: 26172626

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org

L Petruzzi et al.

REFERENCES
1 Citron G, Uso del legno in enologia: specie botaniche utilizzate, anatomia e classicazione. Inform Agrar 50:6972 (2003). 2 Garde-Cerd n T and Ancn-Azpilicueta C, Review of quality factors on a wine ageing in oak barrels. Trends Food Sci Technol 17:438447 (2006). 3 Arapitsas P, Antonopoulos A, Stefanou E and Dourtoglou VG, Articial ageing of wines using oak chips. Food Chem 86:563570 (2004). 4 Guchu E, Daz-Maroto MC, Perez-Coello MS, Gonz les-Vinas MA a and Cabezudo Ib nez MD, Volatile composition and sensory a characteristics of Chardonnay wines treated with American and Hungarian oak chips. Food Chem 99:350359 (2006). 5 Bozalongo R, Carrillo JD, Torroba MAF and Tena MT, Analysis of French and American oak chips with different toasting degrees by headspace solid-phase microextractiongas chromatographymass spectrometry. J Chromatogr A 1173:1017 (2007). 6 Perez-Coello MS, S nchez MA, Garca E, Gonz les-Vinas MA, Sanz J a a and Cabezudo MD, Fermentation of white wines in the presence of wood chips of America and French oak. J Agric Food Chem 48:885889 (2000). 7 Ribereau-Gayon P, Glories Y, Maujean A and Dubourdieu D, Handbook of Enology. The Chemistry of Wine, Stabilization and Treatments (2nd edn), Vol. 2. Wiley, Chichester (2006). 8 Roulland C, Snakkers G and Cantagrel R, Approche exp rimentale du e role des micro-organismes dans le processus de maturation des bois de tonnellerie. J Int Sci Vigne Vin 33:6778 (1999). 9 Vivas N, Glories Y, Doneche B and Gueho E, Observation sur la microore du bois de ch ne (Quercus sp.) au cours de son sechage e naturel. Ann Sci Nat Bot 11:149153 (1991). 10 Vivas N, Saint-Cricq De Gaulejac N, Doneche B and Glories Y, Incidence de la dur e du sechage naturel de Quercus petraea Liebl. et Quercus e robur L. sur la diversit de la ore fongique en place et sur quelques e aspects de son ecologie. J Sci Technol Tonnel 3:1725 (1997). 11 Vivas N and Glories Y, Etude de la ore fongique du ch ne (Quercus sp.) e caract ristique du s chage naturel des bois destines a la tonnellerie. e e Cryptogr Mycol 14:127148 (1993). 12 Dhouib A, Hamza M, Zouari H, Mechichi T, Hmidi R, Labat M, et al, Autochthonous fungal strains with high ligninolytic activities from Tunisian biotopes. Afr J Biotechnol 4:431436 (2005). 13 Orth AB, Royse DJ and Tien M, Ubiquity of lignin-degrading peroxidases among various wood-degrading fungi. Appl Environ Microbiol 59:40174023 (1993). 14 Ch vez R, Bull P and Eyzaguirre J, The xylanolytic enzyme system from a the genus Penicillium. J Biotechnol 123:413433 (2006). 15 Jourez B, Charron S and Quin GP, Propri t s des merrains afn s dans ee e une solution denzymes naturels et destin s a la tonnellerie. Ann e ` Forest Sci 60:123130 (2003). 16 Moredo N, Lorenzo M, Domnguez A, Moldes D, Cameselle C and Sanroman A, Enhanced ligninolytic enzyme production and degrading capability of Phanerochaete chrysosporium and Trametes versicolor. World J Microbiol Biotechnol 19:665669 (2003). 17 Nilsson T, Studies on wood degradation and cellulolytic activity of microfungi. Stud Forest Suecica 104:140 (1973). 18 Di Stefano R, Metodi chimici nella caratterizzazione varietale. Ann Ist Sper Enol 27:3353 (1996). 19 Puech JL, Rabier P, Bories-Azeau J, Sarni F and Moutounet M, Determination of ellagitannins in extracts of oak wood and in distilled beverages matured in oak barrels. J Assoc Off Anal Chem 73:498501 (1990).

20 Towey JP and Waterhouse AL, The extraction of volatile compounds from French and American oak barrels in Chardonnay during three successive vintages. Am J Enol Vitic 47:163172 (1996). 21 Su rez R, Su rez-Lepe JA, Morata A and Calderon F, The production of a a ethylphenols in wine by yeasts of the genera Brettanomyces and Dekkera: a review. Food Chem 102:1021 (2006). 22 Parodi G, Lafnamento del vino in legno. Vignevini 10:4959 (1997). 23 Spillman PJ, Sefton MA and Gawel R, The contribution of volatile compounds derived during oak barrel maturation to the aroma of a Chardonnay and Cabernet Sauvignon wine. Aust J Grape Wine Res 10:227235 (2004). 24 Ferreira V, Lopez R and Cacho JF, Quantitative determination of the odorants of young red wines from different grape varieties. J Sci Food Agric 80:16591667 (2000). 25 Garde-Cerd n T, Rodrguez-Mozaz A and Ancn-Azpilicueta C, Volatile a composition of aged wine in used barrels of French oak and of American oak. Food Res Int 35:603610 (2002). 26 Spillman PJ, Iland PG and Sefton MA, Accumulation of volatile oak compounds in a model wine stored in American and Limousin oak barrels. Aust J Grape Wine Res 4:6773 (1998). 27 Boidro JN, Chatonnet P and Pons M, Inuence du bois sur certaines substances odorantes des vins. Conn Vigne Vin 22:275294 (1988). 28 Lopez R, Aznar M, Cacho JF and Ferreira V, Determination of minor and trace volatile compounds in wine by solid-phase extraction and gas chromatography with mass spectrometric detection. J Chromatogr A 966:167177 (2002). 29 Sefton MA, How does oak barrel maturation contribute to wine avour? Aust NZ Wine Ind J 6:1720 (1991). 30 Peinado RA, Moreno J, Bueno JE, Moreno JA and Mauricio JC, Comparative study of aromatic compounds in two young white wines subjected to pre-fermentative cryomaceration. Food Chem 84:585590 (2004). 31 Genovese A, Gambuti A, Piombino P and Moio L, Sensory properties and aroma compounds of sweet Fiano wine. Food Chem 103:12281236 (2007). 32 Jackson RS, Wine Science: Principles and Applications (3rd edn). Academic Press, London (2008). 33 Gonz lez-Marco A, Jim nez-Moreno N and Ancn-Azpilicueta C, a e Concentration of volatile compounds in Chardonnay wine fermented in stainless steel tanks and oak barrels. Food Chem 108:213219 (2008). 34 Bartowsky EJ and Henschke PA, The buttery attribute of wine diacetyl desirability, spoilage and beyond. Int J Food Microbiol 96:235252 (2004). 35 Keller FA, Hamilton JE and Nguyen QA, Microbial pretreatment of biomass: potential for reducing severity of thermochemical biomass pretreatment. Appl Biochem Biotechnol 105/108:2741 (2003). 36 Kheng PP and Omar IC, Xylanase production by a local fungal isolate, Aspergillus niger USM AI 1 via solid state fermentation using palm kernel cake (PKC) as substrate. Songklanakarin J Sci Technol 27:325336 (2005). 37 Lee JH, Kim JH, Zhu IH, Zhan XB, Lee JW, Shin DH, et al, Optimization of conditions for the production of pullulan and high molecular weight pullulan by Aureobasidium pullulans. Biotechnol Lett 23:817820 (2001). 38 De Jong E, Chandra RP and Saddler JN, Effects of a fungal treatment on the brightness and strength properties of a mechanical pulp from Douglas-r. Bioresour Technol 61:6168 (1997).

2626
wileyonlinelibrary.com/jsfa
c 2010 Society of Chemical Industry J Sci Food Agric 2010; 90: 26172626

Research Article
Received: 31 March 2010 Revised: 20 June 2010 Accepted: 19 July 2010 Published online in Wiley Online Library: 17 August 2010

(wileyonlinelibrary.com) DOI 10.1002/jsfa.4131

Temperature effects on type I pepsin-solubilised collagen extraction from silver-line grunt skin and its in vitro bril self-assembly
Nuntaporn Aukkanit and Wunwiboon Garnjanagoonchorn
Abstract
BACKGROUND: Fish skin is a potential source of collagen. Increasing the extraction temperature increases the yield of collagen. However, it may also result in degradation of the peptide chains, thus damaging the 3D structure of collagen that is vital for its application as a biomaterial. This work investigated the effects of extraction temperature on the yield and characteristics, including bril self-assembly, of type I pepsin-solubilised sh skin collagen. RESULTS: Pepsin-solubilised collagens were extracted from fresh skin of silver-line grunt at 4, 10, 20 and 28 C for 6 h. Extraction at 10 C gave the highest yield of collagens (439.32 96.43 mg g1 fresh skin, dry basis), which were identied as type I and comprised , 1 and 2 subunits. Extraction at higher temperatures (20 and 28 C) resulted in the formation of low-molecularweight peptide fragments, thus reducing the yield of the resultant type I collagen. The denaturation temperatures of collagens extracted at 4 and 10 C, as determined by thermal analysis using differential scanning calorimetry, were 39.5 and 37.5 C respectively. In vitro bril self-assembly of 1 mg mL1 collagen solution (pH 6) incubated at 25 C was only observed with collagens extracted at 4 and 10 C. The 10 C collagen not only showed a higher rate of self-assembly, but its matrix also had a larger bril diameter of 0.50 0.07 m (compared with 0.41 0.07 m for the 4 C collagen) after 4 h of incubation. CONCLUSION: The results indicated strong effects of extraction temperature on the yield and characteristics of the collagen obtained. Extraction of pepsin-solubilised collagen from silver-line grunt skin at 410 C gave a high yield of type I collagen with molecular integrity suitable for tissue-engineering applications. c 2010 Society of Chemical Industry Keywords: type I collagen; extraction; temperature; bril self-assembly

INTRODUCTION
Collagen is widely used as a food and cosmetic ingredient and as a biomaterial. The collagen molecule must maintain its native 3D structure to function properly when used as a biomaterial, e.g. to form lms or porous sheets.1,2 It has also been shown that sh skins from many species are good sources of collagen.3 6 To obtain collagen from animal skin, the cleaned skin is hydrolysed with an acid such as acetic acid7 or treated with an acid and then digested with enzymes such as pepsin to remove telopeptides.8,9 After digestion, type I collagen is separated by salting out with 0.9 mol L1 NaCl in 0.5 mol L1 acetic acid. Most extraction studies have been conducted at low temperatures (usually 4 C) to avoid damaging collagen integrity. However, Lin and Liu10 reported the optimum temperature for collagen extraction from pepsintreated bird feet as 12 C, while the optimum temperature for digestion with commercial porcine pepsin was specied as 37 C. Moreover, collagen has been shown to denature at temperatures close to animal body temperature.11 For example, the acid-soluble collagen and pepsin-solubilised collagen extracted from the skin of silver-line grunt, a warm water sh, denature at 34 and 33.8 C respectively.8 It is therefore of interest to investigate the effects of different temperatures (4, 10, 20 and 28 C) on pepsin-solubilised

collagen extraction from silver-line grunt skin, one of the major by-products from the frozen sh industry, and to determine the characteristics of the extracted collagen, including its bril self-assembly.

MATERIALS AND METHODS

Collagen preparation from sh skin Silver-line grunt (Pomadasys kaakan) skins taken from sh of average length 37 cm were obtained as frozen blocks from Union Frozen Product Co., Ltd (Samutsakorn, Thailand). Fish skins were cleaned by scraping off scales and remaining meat and washed in distilled water at room temperature. The skins were then suspended in 0.05 mol L1 NaOH (pH 12) for 4 h at 4 C to remove non-collagenous proteins. Finally, the treated skins were cut into

Correspondence to: Wunwiboon Garnjanagoonchorn, Department of Food ScienceandTechnology,FacultyofAgro-Industry,KasetsartUniversity,Bangkok 10900, Thailand. E-mail: fagiwwg@ku.ac.th

2627

Department of Food Science and Technology, Faculty of Agro-Industry, Kasetsart University, Bangkok 10900, Thailand

J Sci Food Agric 2010; 90: 26272632

www.soci.org

c 2010 Society of Chemical Industry

www.soci.org 23 cm pieces and collected as a pooled sample for collagen extraction. To extract collagen, 20 g portions of the collected skins were blended with 200 mL of extracting solution (specied later) at a medium speed for 3 min. The collagen extraction was carried out at different temperatures (4, 10, 20 and 28 C) with 1 mg mL1 pepsin (EC 3.4.23.1, activity 936 units mg1 protein; Sigma Chemicals, St Louis, MO, USA) in 0.5 mol L1 acetic acid (at a skin/acetic acid ratio of 1 g/10 mL) by continuous shaking for 6 h in a shaking water bath (BS-11, Jeio Tech Co., Ltd, Daejeon, Korea) tted with a temperature controller. The extract was centrifuged at 20 000 g for 30 min at 4 C in a refrigerated centrifuge (Sorvall RC 5C, Dupont, Newtown, CT, USA) to remove undigested materials. The type I pepsin-solubilised collagen was then precipitated according to Noitup et al.5 by adjusting the concentration of NaCl in the supernatant to 0.9 mol L1 and allowed to stand for 30 min at 4 C. The resultant precipitate was collected by centrifugation at 20 000 g for 30 min, dissolved in four volumes of 0.5 mol L1 acetic acid and dialysed (Cellu Sep dialysis membrane, molecular weight cut-off 12 00014 000; Membrane Filtration Products Co., Ltd, Seguin, Texas, USA) at 4 C against distilled and deionised water until the medium maintained neutral pH. The dialysate was nally lyophilised to obtain the collagen sample. The yield of collagen (mg g1 fresh skin, dry basis) was calculated as follows: yield = [weight of lyophilised collagen (g)/ weight of fresh skin (g, dry basis)] 1000 Viscosity of extracting solution Viscosity was determined with a Brookeld viscometer (RVT, Brookeld Engineering Laboratories, Inc., Middleboro, Massachusetts, USA) using spindle No. 5 at a speed of 5 rpm for the solutions extracted at 4, 10 and 20 C and a speed of 100 rpm for the solution extracted at 28 C. The viscosity was read hourly over a 6 h period. Collagen subunits Collagen subunit patterns were determined by sodium dodecyl sulfate polyacrylamide gel electrophoresis (SDS-PAGE) (modied method of Laemmli12 ) on 75 g L1 separating gel with 40 g L1 stacking gel, using Tris-glycine containing 1 g L1 SDS (pH 8.3) as electrophoresis buffer. Lyophilised samples were solubilised with 50 g L1 SDS solution (containing 50 g L1 2-mercaptoethanol as reducing agent) to a concentration of 5 mg mL1 , then 5 L sample solutions were applied onto the gel. The molecular weights of the protein bands were determined by comparing their mobilities with those of high-molecular-weight markers ranging from 36 to 205 kDa (Sigma Chemicals), namely rabbit myosin (205 kDa), Escherichia coli -galactosidase (116 kDa), rabbit phosphorylase (97 kDa), rabbit fructose-6-phosphate kinase (84 kDa), bovine albumin (66 kDa), bovine glutamic dehydrogenase (55 kDa), ovalbumin (45 kDa) and rabbit glyceraldehyde-3-phosphate dehydrogenase (36 kDa). After electrophoresis was completed, the protein bands were stained with Coomassie brilliant blue R-250. The subunit patterns were compared with that of the standard acid-soluble type I collagen from calf skin (ICN 9007-34-5, ICN Biomedical, Livermore, CA, USA). The proportions of protein bands on three gel slabs (from triplicate experiments) were determined by densitometry. Thermal analysis Collagen samples were prepared by the modied method of Komsa-Penkova et al.13 and Rochdi et al.,14 whereby lyophilised

N Aukkanit, W Garnjanagoonchorn collagens were dissolved in deionised water (1 : 40 w/v) and allowed to stand at 4 C for 24 h. A 15 mg aliquot of sample solution was sealed in a differential scanning calorimetry (DSC) cell and kept at 4 C for 24 h. Thermal analysis was carried out by DSC (Peris 1, Perkin Elmer Inc., Waltham, MA, USA). Samples were scanned at 1 C min1 over the temperature range 2050 C. The characteristic onset (To ), peak (Tp ) and recovery (Tr ) temperatures were recorded from each DSC curve. Collagen self-assembly In vitro self-assembly was performed according to the method of Noitup et al.5 . Lyophilised collagen was dissolved in 0.5 mol L1 acetic acid at 4 C to give a concentration of 1 mg mL1 and dialysed against 67 mmol L1 phosphate buffer (pH 6). After centrifugation at 20 000 g for 30 min the supernatant was incubated at 25 C to reconstruct the matrix. The rate of collagen self-assembly was monitored every 15 min by measuring the turbidity changes as observed by the increase in absorbance at 310 nm using a UV spectrophotometer (Spectro 22, Labomed Inc., Culver city, CA, USA). Structure of collagen matrix reconstruct The collagen solution was prepared according to the method described in the previous section. The supernatant was incubated at 25 C for 4 or 5 h to reconstruct the matrix. At the end of the incubation period the collagen matrix reconstruct was separated by centrifugation at 5000 g for 10 min and the structure was resolved according to the method of Wong et al.15 The matrix was stained with uorescein isothiocyanate (FITC) (Invitrogen, Carlsbad, CA, USA) for 1 h and rinsed for 30 s with distilled water to remove excess FITC. The sample was then covered with a coverglass and its structure was examined using a confocal laser scanning microscope (Carl Zeiss LSM 5 PASCAL, Axio Imager MI, Carl Zeiss Pte Ltd, Gottingen, Germany) and an He/Ne laser with an excitation wavelength of 488 nm. Fibril diameters of the collagen matrix were measured in two areas of 0.6 cm2 and the mean of 100 brils was reported. Statistical analysis Data were subjected to analysis of variance and further analysed using Duncans multiple range test to determine differences between treatment means at a signicance level of 95%. Unless specied otherwise, the mean of at least two replicates was used in each study.

RESULTS AND DISCUSSION

Viscosity of extracting solution Different temperatures and times of pepsin digestion affect both the yield and the properties of the collagen obtained. Temperature has been shown to affect the protein digestion rate of pepsin.10 At the early stage of skin collagen extraction by acid containing 1 mg mL1 pepsin the collagen bres swelled, causing an increase in solution viscosity as shown in Fig. 1(a). Increasing extraction time resulted in loss of skin integrity, and the solution appeared more viscous. After 6 h the extraction at 10 C gave the highest viscosity, followed by those at 4, 20 and 28 C respectively. At high extraction temperature (28 C) the viscosity of the extracting solution decreased abruptly after 3 h owing to the rapid rate of acid and pepsin digestion of collagen molecules (Fig. 1(b)). The increased viscosity of the extracting solution was related to the

2628

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 26272632

Temperature effects on collagen extraction from silver-line grunt skin

www.soci.org

Figure 1. Viscosity of sh skin collagen solutions extracted at different temperatures: (a) 4, 10 and 20 C; (b) 28 C.

Figure 2. Yield of lyophilised sh skin type I collagen at 4, 10, 20 and 28 C after 6 h extraction time. Values with different letters are signicantly different (P < 0.05). Bars indicate standard deviation of the means.

increased yield of pepsin-solubilised collagen, as can be seen in the following section. However, the severe digestion conditions at 28 C resulted in the formation of low-molecular-weight fragments, which lowered the viscosity of the extracting solution (Fig. 1) and the yield of the extracted collagen as shown in Fig. 2. Yield of collagen The yield of type I collagen extract, isolated by precipitation with 0.9 mol L1 NaCl in 0.5 mol L1 acetic acid, was calculated from

the nal lyophilised product. Extraction temperature also had a signicant effect on the yield of lyophilised collagen obtained (P < 0.05) owing to the reduced activity of pepsin at lower temperature. The extraction at 10 C for 6 h gave the highest yield of 439.32 96.43 mg g1 fresh skin (dry basis) (Fig. 2), leaving skin residues of 544.60 mg g1 fresh skin (dry basis) that could go through the extraction process again. According to the viscosity measurement at low temperature (Fig. 1), the yield of collagen increased with the viscosity of the extracting solution. However,

2629

J Sci Food Agric 2010; 90: 26272632

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org

N Aukkanit, W Garnjanagoonchorn

Figure 3. SDS-PAGE patterns of sh skin collagen extracted at 4, 10, 20 and 28 C on 75 g L1 gel: M, standard marker proteins; C, calf skin collagen. Table 1. Composition of collagen subunits on SDS-PAGE gel slab of sh skin collagen extracted at 4, 10, 20 and 28 C and calf skin collagen Fish skin collagen Collagen subunita >250 kDa 1 + 2 <50 kDa Total
a

Calf skin collagen 0.096 0.015 0.512 0.031 0.392 0.045 1.000

4 C 0.045 0.017 0.566 0.018 0.389 0.023 1.000

10 C 0.039 0.016 0.561 0.020 0.400 0.034 1.000

20 C 0.132 0.013 0.254 0.065 0.614 0.072 1.000

28 C 0.033 0.010 0.177 0.055 0.790 0.059 1.000

Collagen subunits as shown in Fig. 3.

the low-molecular-weight peptides that resulted from the higher digestion temperatures (20 and 28 C) could be responsible for the lower yield, since small peptides do not precipitate in the 0.5 mol L1 acetic acid solution containing 0.9 mol L1 NaCl that was used to fractionate type I collagen.10

Table 2. Thermal analysis of sh skin collagen extracted at 4 and 10 C and calf skin collagen Denaturation temperaturea ( C) Sample Fish skin, 4 C Fish skin, 10 C Calf skinb To 38.5 1.48b 35.9 0.21c 45.0 0.03a Tp 39.5 0.63b 37.5 0.07c 45.6 0.00a Tr 40.1 0.68b 39.6 0.65b 46.9 0.02a

SDS-PAGE patterns The SDS-PAGE patterns of sh skin collagen obtained by precipitation with 0.9 mol L1 NaCl in 0.5 mol L1 acetic acid are shown in Fig. 3. The collagen samples, extracted by acid pepsin digestion at 4, 10, 20 and 28 C for 6 h, all contained type I collagen similar to that in calf skin. However, different proportions of each subunit were seen at different extraction temperatures (Table 1). The 4 and 10 C collagens gave very similar results. The 20 and 28 C extractions produced collagen with much reduced proportions of both and subunits but with a high proportion of low-molecularweight peptides (<50 kDa) owing to the excessive digestion at high temperatures. The temperature of the extraction process should therefore be kept between 4 and 10 C to ensure that the extraction product contains type I collagen with a minimum amount of low-molecular-weight peptide fragments.

Values with different letters within a column are signicantly different (P < 0.05). a T , onset temperature; T , peak temperature; T , recovery temperao p r ture. b Calf skin acid-soluble collagen type I.

Thermal analysis The thermal analysis of silver-line grunt skin collagen extracted at 4 and 10 C and calf skin acid-soluble type I collagen is shown in Table 2. The denaturation temperature (Td ) or peak temperature (Tp ) of pepsin-solubilised sh skin collagen extracted at 4 C (Tp 39.5 C) was signicantly higher (P < 0.05) than that

2630

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 26272632

Temperature effects on collagen extraction from silver-line grunt skin

www.soci.org

Figure 4. Rate of self-assembly of sh skin collagen extracted at 4, 10, 20 and 28 C.

of collagen extracted at 10 C (Tp 37.5 C), which could be due to the reason suggested by Noitup et al.8 and Wong,16 namely that collagen that retains a higher degree of inter- and intramolecular crosslinks is more heat-stable. In the case of calf skin collagen the denaturation temperature reported was higher than that of sh skin collagen owing to the higher imino acid content, which results in more stable helices.8,16 The higher Td of pepsinsolubilised collagen from silver-line grunt skin in this study as compared with that reported by Noitup et al.8 was mainly due to differences in the sample preparation methods used for thermal analysis. Denaturation is thought to involve the disruption of interand intramolecular crosslinking of collagen molecules.17 In the present study, lyophilised collagens were dissolved in deionised water, whereas Noitup et al.8 used acetate buffer (pH 3.5) in which collagen solubility is higher. The characteristic temperatures To , Tp and Tr of the collagens extracted at 20 and 28 C could not be detected under the present experimental conditions used for thermal analysis. Rate of collagen self-assembly The rate of collagen bril reconstruction is inuenced by several experimental variables, including the concentration of collagen and the pH and temperature of the solution.18 In this study, collagen self-assembly was performed according to Noitup et al.,5 who reported that 1 mg mL1 collagen solution (pH 6) incubated at 25 C gave the highest rate of in vitro collagen self-assembly. The progressive aggregation of collagen molecules was monitored by the increase in absorbance at 310 nm of the solution, and the changes in absorbance at 310 nm against incubation time indicated the rate of collagen matrix reconstruction. The results obtained (Fig. 4) show typical sigmoidal curves similar to those reported by Engel19 and Veis and George.20 The sh skin collagens extracted at 20 and 28 C did not show self-assembly, while that extracted at 10 C again exhibited the highest rate of collagen bril reconstruction (Fig. 4). Similarly, Lin and Liu10 reported that bird feet collagen extracted at 12 C had a higher rate of selfassembly than those extracted at 4, 18 and 24 C. The results from peptide SDS-PAGE patterns indicated that the sh skin collagens extracted at 4 and 10 C contained fewer low-molecular-weight

peptides with higher 1, 2 and subunits when compared with the samples treated at 20 and 28 C. These data support previous reports1,2 that molecular integrity of the collagen molecule is necessary for its application as a biomaterial for tissue-engineering purposes. Structure of collagen matrix reconstruct The matrices reconstructed from sh skin type I collagens extracted at 4 and 10 C with 4 and 5 h of assembly incubation time were examined under a confocal laser scanning microscope. Fibrils were observed as green uorescence against a dark background; however, all micrographs in Fig. 5 are formatted and shown in grey scale. The resultant micrographs revealed that after 4 h of incubation the diameter of the brils formed from acid/pepsindigested collagen extracted at 10 C (0.50 0.07 m, mean standard deviation) (Fig. 5(c)) was signicantly larger (P < 0.05) than that of the brils formed from collagen extracted at 4 C (0.41 0.07 m) (Fig. 5(a)). Incubation time also plays a role: a longer incubation time leads to an increase in bril diameter (Figs 5(a) and 5(b)). Likewise, the absorbance at 310 nm is related to the diameter of brils: the larger the bril diameter, the higher is the absorbance (Fig. 4). Different rates of collagen monomer assembly have also been shown to produce brils with different diameters.21 In this study we found that under the same assembly environment the bril diameter increases as the collagen self-assembly rate increases.

CONCLUSIONS
Different temperatures and times of pepsin digestion affect both the yield and the properties of the collagen obtained. Collagen extraction from fresh silver-line grunt skin with 0.5 mol L1 acetic acid containing 1 mg mL1 pepsin (at a skin/acid ratio of 1 g/10 mL) incubated at 10 C for 6 h gave the highest yield (439.32 96.43 mg g1 fresh skin, dry basis). Extraction at higher temperatures (20 and 28 C) gave lower yields of collagen, resulting from the formation of low-molecular-weight peptide fragments not precipitated by 0.9 mol L1 NaCl. The Td of collagen extracted at 4 C was higher than that of collagen

2631

J Sci Food Agric 2010; 90: 26272632

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org

N Aukkanit, W Garnjanagoonchorn

ACKNOWLEDGEMENTS
The authors thank The Graduate School, Kasetsart University for nancial support and Union Frozen Product Co., Ltd, Samutsakorn, Thailand for the generous supply of sh skins.

REFERENCES
1 Lee CH, Singla A and Lee Y, Biomedical applications of collagen. Int J Pharmaceut 221:122 (2001). 2 Ikada Y, Biological materials, in Integrated Biomaterials Science, ed. by Barbucci R. Kluwer Academic/Plenum, New York, NY, pp. 17 (2001). 3 Kimura S, Zhu X, Matsui R, Shijoh M and Takamizawa S, Characterization of sh muscle type I collagen, J Food Sci 53:13151318 (1988). 4 Sato K, Yoshinaka R, Itoh Y and Sato M, Molecular species of collagen in the intra-muscular connective tissue of sh. CompBiochemPhysiol B 92:8792 (1989). 5 Noitup P, Morrissey MT and Garnjanagoonchorn W, In vitro selfassembly of silver line grunt type I collagen: effect of collagen concentrations, pH and temperatures on collagen self-assembly. J Food Biochem 30:547555 (2006). 6 Zhang JJ, Duan R, Tian Y and Konno K, Characterisation of acid-soluble collagen from skin of silver carp (Hypophthalmichthys molitrix). Food Chem 116:318322 (2009). 7 Kittipattanabowon P, Benjakul S, Visessanguan W, Nagai T and Tanaka M, Characterisation of acid-soluble collagen from skin and bone of bigeye snapper (Priacanthus tayenus). Food Chem 89:363372 (2005). 8 Noitup P, Garnjanagoonchorn W and Morrissey MT, Fish skin type I collagen: characteristic comparison of albacore tuna (Thunnus alalunga) and silver-line grunt (Pomadasys kaaka). J Aquat Food Prod Technol 14:1728 (2005). 9 Senaratne LS, Park PJ and Kim SK, Isolation and characterization of collagen from brown backed toadsh (Lagocephalus gloveri) skin. Bioresour Technol 97:191197 (2006). 10 Lin YK and Liu DC, Effects of pepsin digestion at different temperatures and times on properties of telopeptide-poor collagen from bird feet. Food Chem 94:621625 (2006). 11 Bailey AJ and Light ND, Molecular and ber structure of collagen, in Connective Tissue in Meat and Meat Products, ed. by Bailey AJ and Light ND. Elsevier Science, London, pp. 3334 (1989). 12 Laemmli UK, Cleavage of structural proteins during assembly of the head of bacteriophage T4. Nature 227:680685 (1970). 13 Komsa-Penkova R, Koyonava R, Kostov G and Tenchov B, Discrete reduction of type I collagen thermal stability upon oxidation. Biophys Chem 83:185195 (1999). 14 Rochdi A, Foucat L and Renou J, NMR and DSC studies during thermal denaturation of collagen. Food Chem 69:295299 (2000). 15 Wong JW, Frank JF and Bailey S, Visualization of eggshell membranes and their interaction with Salmonella enteritidis using confocal laser scanning microscopy. J Food Protect 60:10221028 (1997). 16 Wong DWS, Proteins, in Mechanism and Theory in Food Chemistry, ed. by Wong DWS. Van Nostrand Reinhold, New York, NY, pp. 4897 (1989). 17 Purslow PP, The fracture properties and thermal analysis of collagenous tissues, in Advances in Meat Research, ed. by Pearson AM, Dutson TR and Bailey AJ. AVI Books, New York, NY, pp. 187208 (1987). 18 Hulmes DJS, The collagen superfamily diverse structures and assemblies. Essays Biochem 27:4967 (1992). 19 Engel J, Concepts of self-assembly in biological systems, in Extracellular Matrix Assembly and Structure, ed. by Yurchenco PD, Birk DE and Mecham RP. Academic Press, San Diego, CA, pp. 114 (1994). 20 Veis A and George A, Fundamentals of interstitial collagen selfassembly, in Extracellular Matrix Assembly and Structure, ed. by Yurchenco PD, Birk DE and Mecham RP. Academic Press, San Diego, CA, pp. 1545 (1994). 21 Miyahara M, Njieha FK and Prockop DJ, Formation of collagen brils in vitro by cleavage of procollagen with procollagen proteinase. J Biol Chem 257:84428448 (1982).

Figure 5. Confocal laser scanning micrographs of reconstructed sh skin collagen brils obtained using a 100 oil immersion objective lens: (a) sh skin collagen extracted at 4 C incubated for 4 h; (b) sh skin collagen extracted at 4 C incubated for 5 h; (c) sh skin collagen extracted at 10 C incubated for 4 h. Fibril diameters (mean standard deviation) determined from 100 brils are shown at the bottom of each micrograph; means with different superscripts are signicantly different (P < 0.05). The scale bar in (a) represents 5 m.

extracted at 10 C, but the collagen extracted at 10 C had a higher rate of collagen self-assembly with a larger diameter of reconstructed brils. In summary, the extraction of collagen from silver-line grunt skin by acid/pepsin digestion at 10 C (or lower) produced collagen in high yield and with promising molecular integrity for its application in tissue-engineering work.

2632

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 26272632

Research Article
Received: 24 May 2010 Revised: 19 July 2010 Accepted: 20 July 2010 Published online in Wiley Online Library: 24 August 2010

(wileyonlinelibrary.com) DOI 10.1002/jsfa.4132

Iron supply to soybean plants through the foliar application of IDHA/Fe3+: effect of plant nutritional status and adjuvants
Patricia Rodrguez-Lucena, Edgar Ropero, Lourdes Hernandez-Apaolaza and Juan J Lucena
Abstract
BACKGROUND: Synthetic Fe chelates are commonly used to overcome Fe deciencies in crops, but most of them are scarcely biodegradable. Iminodisuccinic acid (IDHA) is a biodegradable chelating agent that is currently being evaluated as an alternative to EDTA. In this work, the efcacy of the foliar application of IDHA/Fe3+ to soybean chlorotic plants under controlled conditions was studied, testing the inuence of the adjuvant used and of the plant nutritional status. RESULTS: When IDHA/Fe3+ was applied to soybean plants with severe Fe chlorosis and the foliar sprays were the sole source of Fe, this chelate behaved similarly to the EDTA/Fe3+ and the recovery of the plants was slight in both cases. The same chelates were tested when foliar sprays were an additional source of Fe for mildly chlorotic plants, which were also being supplied with low concentrations of Fe applied to the nutrient solution. Then, plant recovery was appreciable in all cases, and the IDHA/Fe3+ was as effective as EDTA/Fe3+ . Among the adjuvants studied, a urea-based product was the only one that did not damage the leaf surface and that could improve the efciency of IDHA/Fe3+ up tp the level of EDTA/Fe3+ . CONCLUSIONS: Thus, it was concluded the foliar application of IDHA/Fe3+ can be an environmentally friendly alternative to the non-biodegradable chelate EDTA/Fe3+ when the appropriate adjuvant is used. c 2010 Society of Chemical Industry Keywords: biodegradable; chelate; chlorosis; foliar spray; IDHA; soybean

INTRODUCTION
Iron chlorosis is a widespread agricultural problem, especially for crops grown in calcareous soils, where calcium carbonate buffers soil solution pH in the 7.58.5 range1 and there is a high bicarbonate concentration. The solubility of Fe in soil is controlled by Fe oxides2 and the most soluble Fe oxide limits total soluble Fe concentration to around 1010 mol L1 in calcareous soils, much lower than that required (108 mol L1 ) for optimal plant growth.3 Caused by a reduction in leaf photosynthetic pigment concentration, intervenial leaf yellowing is the most characteristic visual symptom in chlorotic plants, whose fruit quality, size and yield are severely reduced. Fe chelates applied to soils are the most efcient remedy to control Fe chlorosis. Most of these chelates degrade very slowly in the environment, and concern about the environmental risk of their application4 has grown in recent decades. Moreover, the risk of leaching out of the root zone in regimes of high water availability, which can be very high for the highly efcient chelate EDDHA/Fe3+ ,5 constitutes an important constraint when synthetic chelates are applied to soils.6 The biodegradable chelating agent N-(1,2-dicarboxyethyl)-D,L-aspartic acid, commonly known as iminodisuccinic acid or IDHA (Fig. 1), has recently been proposed for its use in agriculture.7,8 IDHA shares structural similarities with EDTA (Fig. 1), but contains only ve functional groups able to complex Fe. Due to this, IDHA/Fe3+ records a lower stability than EDTA/Fe3+

and a higher reactivity in agronomic conditions. However, the efciency of IDHA/Fe3+ has been comparable to EDTA/Fe3+ at providing Fe through nutrient solution to cucumber and soybean plants, cultivated in a growth chamber in hydroponics under calcareous soil conditions.8 When applied to tomato or green bean plants grown under eld conditions or in commercial hydroponics, IDHA/Fe3+ behaved similarly to EDTA/Fe3+ solving Fe chlorosis.9 As an alternative to a conventional supply of Fe to the soil or nutrient solution, the foliar application of Fe compounds is now under consideration, but controversial results have been obtained due to the numerous uncertainties involved in this type of fertilisation. The aerial parts of plants are covered by a cuticle that is the major barrier to be overcome when chemicals are sprayed onto leaves. Cuticles have both hydrophilic and lipophilic properties, and the routes and factors controlling penetration of ionic species are still poorly understood. The use of adjuvants enhances the retention of foliar sprays and can increase cuticular penetration, diffusion into the apoplast and uptake by leaf cells.10 Nevertheless,

Correspondence to: Juan J Lucena, Agricultural Chemistry Department, Aut noma University of Madrid, E-28049 Madrid, Spain. o E-mail: juanjose.lucena@uam.es

2633

Agricultural Chemistry Department, Universidad Aut noma de Madrid, o Francisco Tom s y Valiente N 7, 28049 Madrid, Spain a

J Sci Food Agric 2010; 90: 26332640

www.soci.org

c 2010 Society of Chemical Industry

www.soci.org
EDTA HOOC NH HOOC NH COOH COOH

P Rodrguez-Lucena et al. was studied when the foliar sprays were the sole source used to provide Fe to plants, and different adjuvants (non-ionic or glycine-based) were applied in combination with the chelate. In a second experiment, the efciency of the IDHA/Fe3+ applied with urea or with a non-ionic adjuvant to provide foliar Fe to plants was tested when applied to mildly chlorotic plants that were also being supplied with Fe at low concentrations (as EDTA/Fe3+ ) through the nutrient solution to simulate the low availability of Fe in calcareous soils. For comparative purposes, EDTA/Fe3+ , frequently used in foliar applications in the eld, was evaluated in both experiments.

IDHA COOH H N HOOC COOH

HOOC

EXPERIMENTAL
Fe-containing compounds IDHA/Fe3+ (8.7% (w/w) Fe as IDHA/Fe3+ ) was provided by ADOB PPC (Poznan, Poland) and prepared by dissolution of the solid and ltration, while EDTA/Fe3+ was prepared in the laboratory; to do so, the ligand (Na2 EDTA; Merck, Darmstadt, Germany) was rst dissolved and then an amount of Fe(NO3 )3 9H2 O (Merck), calculated to be 5% in excess of the molar amount of ligand, was added slowly. During chelation, the pH was maintained between 6.0 and 8.0 and adjusted to 5.5 at the end. The solution was left to stand overnight to allow any excess element to precipitate as oxides. Finally, the solution was ltered through a 0.45 m Millipore membrane and made up to volume with water. Exposure of all the chelate solutions to light was avoided during their preparation and storage because of the potential photodecomposition of chelates.15 Both chelates were prepared at a concentration of Fe 5 mmol L1 . Adjuvants Different adjuvants were tested: two non-ionic products with ethylene oxide groups as hydrophilic component (an alkylpolyglucoside (APG) and polyoxyethylene (20) sorbitan monooleate (PS)), an anionic glycine-based solution (GLY) and a urea-based adjuvant (U). Plant material, experimental design and treatments Soybean seeds (Glycine max L. cv Stine 0480, kindly provided by Professor R. Goos, North Dakota State University, Fargo, USA) were germinated using a standard seed-growing procedure in sterilised trays. The seeds were washed with water for 30 min and then placed in trays between two sheets of cellulose paper moistened with distilled water. The trays were kept in darkness at 28 C for 3 days in a thermostatted oven. After germination, the seedlings were transferred to a growth chamber, where they were grown under controlled climatic conditions: day/night photoperiod, 16/8 h; temperature (day/night) 30/25 C; relative humidity (RH) (day/night) 50/70%. Seedlings of similar development were placed on a holed plate, oating in containers with a continuously aerated 1/5 diluted nutrient solution for 6 days. The diluted nutrient solution was then replaced by a fullstrength solution with the following composition: (macronutrients in mmol L1 ) 1.0 Ca(NO3 )2 , 0.9 KNO3 , 0.3 MgSO4 , 0.1 KH2 PO4 ; (cationic micronutrients in mol L1 , as buffered micronutrient solution) 2.5 MnSO4 , 1.0 CuSO4 , 10.0 ZnSO4 , 1.0 NiCl2 , 1.0 CoCl2 , 5.0 Fe(NO3 )3 , 120.5 Na2 EDTA, 50.0 KOH; (anionic micronutrients in mol L1 ) 35.0 NaCl, 10.0 H3 BO3 , 0.05 Na2 MoO4 . The pH was buffered with HEPES 1.0 104 mol L1 and adjusted to 7.5 with KOH 1.0 mol L1 . The seedlings were kept in this solution for 6 days, until chlorotic symptoms were observed. After this time,

o,o-EDDHA OH O NH HN HO O

OH

HO

Figure 1. Fe chelates described in the text.

as many adjuvants can traverse the cuticle and damage cells and cell membranes, those applied with foliar fertilisers should either not penetrate or do so very slowly. Moreover, they should be biodegradable and non-phytotoxic products and perform efciently when applied at low concentration. The foliar application of IDHA/Fe3+ has been assayed previously under eld conditions with variable results. Although some studies showed that FeSO4 could re-green chorotic leaves more efciently than other Fe compounds when supplied through foliar sprays,11,12 the benecial effect seems related to the low pH used in the application of Fe sulfate (pH 4) in order to avoid natural oxidation of Fe. In addition, the Fe supplied by this salt may not remain available in leaves after a short period of time and leaf surface is damaged due to the low pH of the solution.13,14 On the other hand, recent studies have corroborated that synthetic chelates behave more efciently that sulfate when the appropriate adjuvant is included in the formulation.13,14 With regard to the type of chelate used, working with Fe-decient peach trees Fern ndez et al.13 observed a that EDTA/Fe3+ was more efcient than IDHA/Fe3+ when applied with an alkyl-glucoside adjuvant. In a different experiment testing the inuence of the type of adjuvant on the performance of these chelates when applied to chlorotic peach trees, the same authors concluded that their effectiveness was highly dependent on the type of adjuvant used. In general, EDTA/Fe3+ promoted higher leaf re-greening than IDHA/Fe3+ , while leaf Fe concentration was greater in the trees sprayed with IDHA/Fe3+ .14 Given the variability of the results obtained in eld experiments, the aim of this work was to asses the ability of IDHA/Fe3+ foliar sprays to overcome Fe chlorosis under controlled conditions when applied to soybean (Glycine max. cv Stine 0480), a non-efcient model plant for Fe nutrition. First, the effectiveness of the foliar application of IDHA/Fe3+ to soybean plants with severe chlorosis

2634

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 26332640

Foliar application of IDHA/Fe3+ to soybean

www.soci.org Experiment I: Foliar application of IDHA/Fe3+ and EDTA/Fe3+ to soybean plants affected by severe chlorosis without an additional source of Fe Each pair of plants was sprayed with 2 mL of the 5 mmol L1 IDHA/Fe3+ or EDTA/Fe3+ solutions at pH 5.5 (Table 1). Leaf sprays were applied both on the adaxial and the abaxial surface with a nebuliser system. The corresponding adjuvant (Table 1) was added to each solution just before leaf spraying at the rate indicated in Table 1. A control was included in which no Fe was supplied (control Fe). Experiment II: Foliar application of IDHA/Fe3+ and EDTA/Fe3+ to mildly chlorotic soybean plants, as a complement to Fe supply through the roots Application was performed as described in Experiment I, but only the rst four levels of leaves were treated in order to evaluate the redistribution of the Fe applied through foliar sprays. The corresponding adjuvant was added to each solution just before leaf spraying (Table 2), with the rate varying depending on the surface active agent. Two control treatments (control Fe and control FeEDTA) were established. For control Fe, no Fe was added to the nutrient solution or applied through foliar sprays. In the case of control FeEDTA (used to evaluate the inuence on plant development of the Fe supplied through the nutrient solution) 5.0 mol L1 EDTA/Fe3+ was applied to the nutrient solution. For control treatments, foliar applications were not performed. Measurements During the experiments, regular Soil and Plant Analyzer Development (SPAD) readings were taken with a chlorophyll meter (Minolta SPAD-502; Minolta, Osaka, Japan) for all the leaf stages (average of three readings per leaf). Whole plants were sampled 8 days (two pairs of plants) and 22 days (one pair of plants) after the rst application of the treatment. Sampled roots, stems, treated leaves and untreated leaves (only in Experiment II) were separated, weighed, washed with 0.1% HCl and 0.01% non-ionic detergent solution, and rinsed twice with ultrapure water. The samples were then dried in a forced air oven at 65 C for 3 days. Micronutrients were determined in the leaves and roots after dry digestion procedure by Atomic absorption spectroscopy (AAS, PerkinElmer AAnalyst 800; Perkin Elmer, Waltham, MA, USA). Statistical analysis Data were statistically evaluated using analysis of variance (ANOVA) with the program SPSS 15.0 to asses the signicance of the main factors and interactions. Means were also compared using Duncans test at P 0.05 in order to nd signicant differences between treatments.

Table 1. Treatments, adjuvants and doses tested in experiment I Foliar Fe treatments FeIDHA + APG FeIDHA + GLY FeIDHA + PS Treatment 5 mmol L1 IDHA/Fe3+ Adjuvant Nutrient solution

FeEDTA + APG

5 mmol L1 EDTA/Fe3+

Oligomeric No added Fe alkylpolyglucoside (APG) 0.1% (v/v) Glycine based (GLY) 0.3% (v/v) Polyoxyethylene (20) sorbitan monooleate (PS) 0.1% (v/v) Oligomeric alkylpolyglucoside (APG) 0.1% (v/v) No added Fe

Control Control Fe

No added Fe

Table 2. Treatments, adjuvants and doses tested in Experiment II Foliar Fe treatments FeIDHA + APG I FeIDHA + APG II FeIDHA + U FeEDTA + APG 5 mmol L1 EDTA/Fe3+ Treatment 5 mmol L1 IDHA/Fe3+ Adjuvant Nutrient solution

Oligomeric No added Fe alkylpolyglucoside (APG) 0.1% (v/v) 5 mol L1 EDTA/Fe3+ Urea (U) 0.15% (v/v) Oligomeric alkylpolyglucoside (APG) 0.1% (v/v) No added Fe 5 mol L1 EDTA/Fe3+

Control Control Fe Control FeEDTA

No added Fe No added Fe

the stems of two plants were wrapped together with foam, and placed in 2 L polyethylene vessels (three holes in the lid, six plants per pot) containing 2 L of a full strength nutrient solution with the same composition as the initial one, except in the content of micronutrients (unbuffered micronutrient solution, in mol L1 ): 1.0 MnSO4 , 0.5 M CuSO4 , 0.5 ZnSO4 , 0.1 NiCl2 , 0.1 CoCl2 . In Experiment I, Fe was not added to this nutrient solution, while in Experiment II, 5.0 mol L1 EDTA/Fe3+ was added to avoid severe chlorosis. The pH was adjusted to 7.5 with KOH 1.0 mol L1 and buffered with HEPES 1.0 104 mol L1 , and 0.4 g1 of solid CaCO3 per pot. Water was added every 2 days, and the solution was renewed every week. Three replicates were prepared for each treatment and for each control. At this point, Fe treatments were applied (Table 1 for Experiment I and Table 2 for Experiment II), and applications were repeated after 8 and 15 days.

RESULTS
Experiment I: foliar application of IDHA/Fe3+ and EDTA/Fe3+ to soybean plants affected by severe chlorosis without an additional source of Fe Micronutrient contents and biometric data In this experiment, the foliar sprays described in Table 1 were the sole source of Fe for plants, which showed severe chlorosis symptoms. The effect of the foliar application of the treatments on plant dry weight, root Fe and Mn concentrations, and Fe/Mn and Fe/(Mn + Zn + Cu) ratios was studied using one-way ANOVA. Statistical differences were recorded only for root Fe

2635

J Sci Food Agric 2010; 90: 26332640

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org

P Rodrguez-Lucena et al.

Table 3. SPAD values, dry weight, root Fe and Mn concentrations, and root Fe/Mn and Fe/(Mn + Zn + Cu) ratios in soybean plants sprayed with the Fe chelates tested, at the end of Experiment I SPAD FeIDHA + APG FeIDHA + GLY FeIDHA + PS FeEDTA + APG Control Fe 11.3 0.7a 8.9 0.6ab 9.6 0.7ab 11.6 1.3a 7.3 1.1b Root DW (g plant1 DW) 0.51 0.05ns 0.57 0.02 0.62 0.13 0.69 0.12 0.43 0.12 [Fe] (g g1 DW) 61.1 1.4ab 74.2 7.9a 50.6 10.7b 50.2 0.5b 41.5 3.3b [Mn] (g g1 DW) 166 21b 186 28ab 187 7ab 236 21ab 306 73a Fe/Mn 0.38 0.04a 0.36 0.07a 0.28 0.07ab 0.21 0.01ab 0.15 0.03b Fe/(Mn + Zn + Cu) 0.20 0.02a 0.20 0.02a 0.15 0.04ab 0.13 0.01ab 0.09 0.01b

Data are means standard error (SE) of three independent replicates, except for SPAD values (average of 18 measurements). Different letters in the same column denote signicant differences between the treatments (P 0.05). DW: Dry weight basis; ns, not signicant.

concentration (at P 0.05), indicating that the combination of FeIDHA + GLY was more effective than the other Fe-containing compounds and adjuvants tested. In order to evaluate the redistribution of the foliarly applied Fe from leaves to other plant organs, Fe and other micronutrient concentrations in roots were measured. Fe and Mn concentrations, as well as the Fe/Mn and Fe/(Mn + Zn + Cu) ratios in the roots, are presented in Table 3. The same measurements were also performed on treated leaves, but as it was not possible to ensure that the fraction of Fe applied through foliar sprays and not absorbed by leaves was completely removed after washing, these data were not taken into consideration. The highest root Fe concentrations corresponded to FeIDHA + GLY, which was the only treatment statistically different to control Fe, and to FeIDHA + APG. Mn concentrations also indicate that the uptake of Mn from the nutrient solution was lower in plants treated with Fe sprays, even though most of these treatments were not statistically different to control Fe. The low Fe/Mn and Fe/(Mn + Cu + Zn) ratios for control Fe with regard to treated plants indicate the increased capacity of roots to take up other cations under Fe deciency conditions. This favoured uptake is related to the involvement of the same transporter implied in the uptake of Fe by roots, the Iron regulated transporter 1 (IRT1), in the absorption of other cations such as Mn or Zn.16 In spite of the differences in micronutrient concentrations and ratios, none of the treatments presented statistical differences in plant dry weight with control Fe. SPAD measurements SPAD was measured periodically in all leaf levels. Two-way ANOVA analysis showed that the SPAD values measured for each treatment in the third level of leaves were strongly inuenced (at P 0.001) by the treatment applied and by the age of the plants (days since the rst application of the treatments), but there was no interaction between the two factors. Average SPAD values are reported in Table 3, and indicate that for all the treatments the average SPAD was higher than for control Fe. The best results corresponded to plants treated with Fe compounds combined with APG as adjuvant (FeIDHA + APG and FeEDTA + APG), while the other treatments were not statistically different to control Fe. Experiment II: foliar application of IDHA/Fe3+ and EDTA/Fe3+ to mildly chlorotic soybean plants, as a complement to Fe supply through roots In the second experiment, foliar sprays were a complement to Fe nutrition through the roots (Table 2), as usually occurs when Fe foliar sprays are applied under eld conditions. Consequently,

plants exhibited mild chlorosis. In order to test the redistribution from treated leaves to untreated leaves only the rst four levels of leaves were sprayed. Regarding the adjuvants, a urea-based solution (U) was compared with the APG used in Experiment I when applied in combination with IDHA/Fe3+ through foliar sprays. Micronutrient contents and biometric data The inuence of the treatments on plant dry weight, Fe and Mn concentrations, and Fe/Mn and Fe/(Mn + Zn + Cu) ratios in untreated leaves and roots was tested using one-way ANOVA. The treatments strongly affected Fe and Mn concentrations and the Fe/Mn and Fe/(Mn + Cu + Zn) ratios in roots (at P 0.001 in all the cases). The data corresponding to dry weight, Fe and Mn concentrations and the Fe/Mn and Fe/(Mn + Zn + Cu) ratios in untreated leaves and roots at the end of the experiment are reported in Table 4. In untreated leaves, no statistical differences due to the treatments were observed for any of the parameters studied. However, when micronutrient concentrations and ratios were analyzed in roots, IDHA treatments including the supply of Fe to the roots (FeIDHA + APG II and FeIDHA + U) were the most effective and presented statistical differences with control Fe. Nevertheless, in most cases these treatments were not statistically different to control FeEDTA (where foliar applications were not performed). The only treatment in which Fe was not added to the nutrient solution (FeIDHA + APG I) recorded no statistical differences with control Fe. With regard to root dry weight, the highest values were found for FeIDHA + U and FeEDTA + APG, while no statistical differences were observed in root dry weight between FeIDHA + APG I (with no Fe supply through the nutrient solution) and control Fe. SPAD measurements SPAD was measured periodically in all leaf levels. Two-way ANOVA analysis showed that SPAD values for each treatment in the third level of leaves were strongly inuenced (at P 0.001) by the treatment applied and by the age of the plants (days since the rst application of the treatments), but there was no interaction between the two factors. Average SPAD values are reported in Table 4, and indicate that all the treatments except FeIDHA + APG I (with no Fe added to the nutrient solution) were statistically different to control Fe. On the other hand, very low SPAD measurements were recorded when no foliar sprays were performed (control FeEDTA). The best treatments for improving SPAD values were FeIDHA + U and FeEDTA + APG.

2636

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 26332640

2637

J Sci Food Agric 2010; 90: 26332640

Biomass (g plant1 DW) Fe concentration (g g1 DW) Untreated leaf Root 47.9 1.9c 135.3 21.7a 105.3 9.1ab 89.4 10.4b 33.1 1.8c 140.4 17.9a 70.8 63.8 12.1 70.4 20.6 97.6 25.5 1.8ns 215.9 36.6a 23.3 2.1b 12.5 1.9b 8.6 0.2b 194.8 26.9a 15.8 0.4b Root 46.5 45.1 8.9 40.7 3.3 31.4 4.2 0.5ns Untreated leaf Mn concentration (g g1 DW) Root 0.56 0.10ab 0.93 0.24ab 1.64 0.47a 1.55 0.59ab 0.46 0.04b 1.07 0.22ab Fe/Mn Untreated leaf 0.7 0.8 0.3 0.7 0.2 0.4 0.1 0.01ns Root 0.2 0.1c 5.9 1.1b 8.7 0.9a 10.4 1.2a 0.2 0.1c 8.9 1.1a Fe/(Mn + Cu + Zn) Untreated leaf 0.4 0.5 0.2 0.4 0.1 0.2 0.1 0.1ns Root 0.20 0.02b 2.3 0.3a 2.3 0.1a 2.0 0.1a 0.10 0.04b 2.5 0.2a 0.11ab

Foliar application of IDHA/Fe3+ to soybean

Table 4. SPAD values, untreated leaves and root dry weight, untreated leaves and root Fe and Mn concentrations, and untreated leaves and roots Fe/Mn and Fe/(Mn + Zn + Cu) ratios in soybean plants sprayed with the Fe chelates tested, at the end of Experiment II

SPAD

Untreated leaf

c 2010 Society of Chemical Industry

FeIDHA + APG I FeIDHA + APG II FeIDHA + U FeEDTA + APG Control Fe Control FeEDTA

4.4 0.7d 10.1 1.4bc 14.8 1.4a 14.1 1.6ab 3.2 0.4d 6.5 0.7cd

1.20 1.50 0.37a 0.49 0.15b 0.58 0.10ab

www.soci.org

Data are means standard error (SE) of three independent replicates, except for SPAD values (average of 24 measurements). Different letters in the same column denote signicant differences between the treatments (P < 0.05, n = 3). DW: Dry weight basis. FeIDHA + APG I did not develop untreated leaves.

wileyonlinelibrary.com/jsfa

www.soci.org

P Rodrguez-Lucena et al. Comparison of adjuvant solutions Three biodegradable adjuvants were tested in this rst experiment: (1) two non-ionic products (polyoxyethylene (20) sorbitan monooleate (PS) and an oligomeric alkylpolyglucoside (APG) product); and (2) an anionic glycine-based (GLY) solution (Table 1). Non-ionic adjuvants are the most commonly used in foliar applications, since interaction with the active ingredient is minimised, albeit not completely avoided,18 in this type of compounds. At the same time, they increase the cuticular penetration of the active ingredient through complex interactions between the active ingredient, the adjuvants and leaf.21 PS adjuvants have not been commonly applied in foliar formulations. In the case of APG, these adjuvants favour cuticular penetration of solutions in general22 and Fe solutions in particular.13 For anionic adjuvants, it is important to bear in mind that these molecules may interact with the cations of the Fe compound solutions, forming large molecules that may block cuticular pores and interfere with the process of leaf Fe penetration.23 However, when applied through foliar sprays, GLY could penetrate leaf surfaces and translocate to other plant organs,24 so solutions based on this amino acid may improve Fe penetration. Previous studies13,14 have tested the foliar application of IDHA/Fe3+ and EDTA/Fe3+ in combination with an APG and a GLY-based adjuvant, concluding that both chelates could increase green surface and chlorophyll content in peach leaves. However, as far as we know, the foliar application of solutions containing only Fe compounds and GLY-based or PS adjuvants has not been assayed previously. Our results indicate that the application of IDHA/Fe3+ in combination with GLY (FeIDHA + GLY) or PS adjuvants (FeIDHA + PS) was slightly less efcient than the use of an APG adjuvant (FeIDHA + APG), in spite of the interaction between IDHA/Fe3+ and APG adjuvants described elsewhere.13,14 Regardless of the Fe compound or adjuvant used, leaf burn was always observed in treated leaves, even when the nonphytotoxic APG surface active agent was applied. This observation is consistent with the results obtained by Fern ndez et al.,18 who a detected the ionisation of a non-ionic adjuvant in the presence of EDTA. However, these symptoms were similar to the necrotic spots observed for control Fe, so it was not possible to conrm whether they were due to the foliar sprays (the damage observed on leaves might be related to the Fe compound applied, to the dissolution of the cuticle as a negative consequence of the application of adjuvants, to a detrimental interaction between the adjuvant and the Fe compound) or to the severe chlorosis of the plants. Experiment II: foliar application of IDHA/Fe3+ and EDTA/Fe3+ to mildly chlorotic soybean plants, as a complement to Fe supply through roots Comparison of Fe compounds The results of Experiment I indicated that foliarly applied Fe can re-green leaves and be distributed to other plant organs, although due to the limited absorption of Fe through leaf surfaces18 and to the low mobility of Fe in the phloem19,25 this type of application cannot fully satisfy plant Fe demand. Since Fe fertilisation through foliar sprays can be used only as a complementary technique to the direct application into soil or to a nutrient solution of Fe chelates,11,26 in the second experiment presented in this work foliar sprays were applied as a complement to the application of Fe chelates to the nutrient solution of soybean plants grown in hydroponics under calcareous soil conditions. The results of this assay conrmed that even though Fe foliar sprays cannot fully substitute the conventional supply of Fe

DISCUSSION
Experiment I: foliar application of IDHA/Fe3+ and EDTA/Fe3+ to soybean plants affected by severe chlorosis without an additional source of Fe Comparison of Fe compounds IDHA is a biodegradable chelating agent recently proposed for agricultural use.7,8 Consequently, there are currently few studies available that evaluate the effectiveness of IDHA/Fe3+ to overcome Fe chlorosis. The ability of this chelate to solve Fe deciencies when applied to roots under controlled8,9 or eld conditions9 has been tested successfully, but due to its relatively low stability its application through foliar sprays is also under consideration. Accordingly, several eld trials to assess its performance have been carried out in recent years,13 although to our knowledge the foliar application of IDHA/Fe3+ chelates under controlled conditions to correct Fe deciencies in plants has yet to be studied. One of the objectives of this work was to evaluate the behaviour of this chelate under controlled conditions. All possible sources of Fe would therefore be considered in the experimental design and any improvement in plant Fe nutrition could be attributed to the Fe supplied through the treatments. In the studies assessing the effectiveness of IDHA/Fe3+ foliar sprays when applied in the eld to peach trees,13,14 IDHA/Fe3+ performed as efciently as EDTA/Fe3+ when the appropriate adjuvant was employed, although the results were always very variable. Foliar sprays are not the only source of Fe, since in soil experiments a fraction of Fe is available for plants, even under calcareous soil conditions. Thus, the results will be inuenced by soil conditions and Fe availability in it. In the rst experiment described in this work, foliar sprays were the sole source of Fe for plants. In general, IDHA/Fe3+ behaved as efciently as EDTA/Fe3+ . Root Fe concentration and the other nutritional parameters reported in Table 3 suggest that Fe was redistributed from leaves to roots. Evidence of translocation from leaves has been observed in other experiments working with cucumber and tomato and using labelled Fe.17 For our results, differences in Fe and Mn concentrations and ratios in roots with regard to control Fe were observed, even though these differences were not signicant in all cases. This suggests that the foliarly applied Fe was redistributed, although at low rates due to the limited absorption of Fe through leaf surfaces18 and to the scarce mobility of this micronutrient through the phloem.19 Data corresponding to Fe/Mn and Fe/(Mn + Zn + Cu) ratios indicate that the afnity of roots for Mn, Cu and Zn decreased when Fe sprays were performed, supporting the idea of Fe redistribution to roots. Taken together, these data would to some extent reect the relative normalisation of the metabolic processes of the plants, as already observed in previous works with IDHA/Fe3+ applied through foliar sprays.13 Moreover, these results indicate that the foliar application of biodegradable IDHA/Fe3+ is a good alternative to EDTA/Fe3+ when applied under appropriate conditions, as expected due to the similarities between both chelates. The better performance of EDTA/Fe3+ with regard to other chelates when applied through foliar sprays has been attributed, at least partially, to the photoreduction of this chelate20 that would favour Fe uptake by leaf cells. However, our results suggest that this phenomenon cannot explain the different behaviour observed for EDTA/Fe3+ and IDHA/Fe3+ in previous assays, which can be attributed to different experimental conditions. The biodegradation of IDHA, which may be favoured more in crops grown in the eld than under the controlled conditions of a growth chamber, could explain the lower effectiveness of IDHA/Fe3+ with regard to EDTA/Fe3+ when used under eld conditions.

2638

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 26332640

Foliar application of IDHA/Fe3+ to soybean chelates to the roots, they help to overcome Fe chlorosis when applied as a complementary source of Fe. Under these conditions, the re-greening of leaves and the redistribution of Fe from treated leaves to other organs occurred. The foliar application of Fe chelates when EDTA/Fe3+ was added to the nutrient solution (FeIDHA + APG II, FeIDHA + U and FeEDTA + APG) increased root Fe concentration. This suggests that Fe was translocated to the roots, as already observed when labeled Fe was used,17,27,28 and/or that Fe uptake from the nutrient solution was enhanced due to the transmission of signals from the shoots29 that activate the root Fe uptake mechanism. None of the treatments tested was clearly more efcient than the rest at favouring Fe redistribution to untreated organs, but the highest SPAD values corresponded to FeIDHA + U. Comparison of adjuvant solutions Although the APG adjuvant was the most efcient product used in combination with IDHA/Fe3+ , the interaction between both compounds has been described in previous studies. Thus, the effectiveness of the APG was compared with urea when added in IDHA/Fe3+ formulations. The use of urea in foliar sprays favours Fe assimilation by leaves,30 given both its capacity to increase the permeability of leaf membranes and its surfactant properties. At the same time, these applications can play an additive role. It has already been demonstrated that under Fe deciency conditions the capacity of roots to acquire nitrate is limited and this anion accumulates at low rates in leaf tissues.31 Thus, the simultaneous application of Fe and N (as urea) through foliar sprays may have deal with this deciency and favoured leaf re-greening. This benecial effect of the application would explain the high SPAD values for the FeIDHA + U treatment. In any case, urea must be supplied at low rates to avoid the inhibition of photosynthesis.32 Regarding burn signs on leaf surfaces, they were less abundant in the mildly chlorotic plants used in this experiment. However, except for FeIDHA + U, treated leaves always presented dark spots that could be associated with leaf damage. This conrms that the use of APG negatively affected the leaf surface even though this type of adjuvant has been described as non-phytotoxic. Negative interactions between spray components and leaf surfaces did not seem to occur for urea.

www.soci.org Education FPI pre-doctoral contract co-nanced by the European Social Fund.

REFERENCES
1 Lindsay WL and Schwab AP, The chemistry of iron soils and its availability to plants. J Plant Nutr 5:821840 (1982). 2 Lindsay WL, Iron oxide solubilization by organic matter and its effect on iron availability, in Iron Nutrition and Interaction in Plants, ed. by Chen Y and Hadar Y. Kluwer Academic Publishers, Dordrecht, pp. 2936 (1991). 3 Romheld V and Marschner H, Mobilization of iron in the rhizosphere of different plant species, in Advances in Plant Nutrition, vol. 2, ed. by Tinker B and Lauchli A. Praeger, New York, pp. 155204 (1986). 4 Hyvonen H, Orama M, Saarinen H and Aksela R, Studies on biodegradable chelating agents: complexation of iminodisuccinic acid (ISA) with Cu(II), Zn(II), Mn(II) and Fe(III) ions in aqueous solution. Green Chem 5:410414 (2003). 5 Cesco S, Romheld V, Varanini Z and Pinton R, Solubilization of iron by water-extractable humic-substances. J Plant Nutr 163:285290 (2000). 6 Rombol` AD and Tagliavini M, Iron nutrition of fruit tree crops, in Iron a Nutrition in Plants and Rhizospheric Microorganisms, ed. by Barton LL and Abada J. Springer-Verlag Academic Publishers, Dordrecht, pp. 6183 (2006). 7 Mitschker A, Moritz RJ and Nawrocki A, Chelated plant micronutrients. Bayer Chemicals A.-G., Germany; Przedsiebiorstwo Produkcyjno Consultingowe Adob. European Patent EP1411037 A120040421Appl (2004). 8 Vill n M, Garca-Arsuaga A and Lucena JJ, Potential use e of biodegradable chelate N-(1,2-dicarboxyethyl)-D,L-aspartic acid/Fe3+ as an Fe fertilizer. J Agric Food Chem 55:402407 (2007). 9 Lucena JJ, Sents JA, Vill n M, Lao T and P rez-S ez M, IDHA chelates e e a as a micronutrient source for green bean and tomato in fertigation and hydroponics. Agron J 100:813818 (2008). 10 Schonherr J and Baur P, Effects of temperature, surfactants and other adjuvants on rates of uptake of organic compounds, in Plant Cuticles An Integrated Functional Approach, ed. by Kerstiens G. Bios Science Publisher Ltd, Oxford, pp. 135156 (1996). 11 Alvarez-Fern ndez A, Garca-Lavina P, Fidalgo C, Abada J and a Abada A, Foliar fertilization to control iron chlorosis in pear (Pyrus communis L.) trees. Plant Soil 263:515 (2004). 12 Rombol` AD, Bruggemann W, Tagliavini M, Marangoni B and a Moog PR, Iron source affects iron reduction and re-greening of kiwifruit (Actinidia deliciosa) leaves. J Plant Nutr 23:17511765 (2000). 13 Fern ndez V, Ro V, Abada J and Abada A, Foliar iron fertilization a of peach (Prunus persica (L.) Batsch): effects of iron compounds, surfactants and other adjuvants. Plant Soil 289:239252 (2006). 14 Fern ndez V, Del Ro V, Pumarino L, Igartua E, Abada J and Abada A, a Foliar fertilization of peach (Prunus persica (L.) Batsch) with different iron formulations: Effects on re-greening, iron concentration and mineral composition in treated and untreated leaf surfaces. Sci Hortic 117:241248 (2008). 15 Hill-Cottingham DG, Photosensitivity of iron chelates. Nature 175:347348 (1955). 16 Vert G, Grotz N, D dald champ F, Gaymard F, Guerinot ML, Briat JF, e e et al, IRT1, an arabidopsis transporter essential for iron uptake from the soil and for plant growth. Plant Cell 14:12231233 (2002). 17 Rodrguez-Lucena P, Tomasi T, Pinton R, Hern ndez-Apaolaza L, a Lucena JJ and Cesco S, Evaluation of 59 Fe-lignosulfonates complexes as Fe-sources for plants. Plant Soil 325:5363 (2009). 18 Fern ndez V, Orera I, Abada J and Abada A, Foliar iron-fertilisation of a fruit trees: present knowledge and future perspectives a review. J Hort Sci Biotechnol 84:16 (2009). 19 Kim AS and Guerinot ML, Mining iron: iron uptake and transport in plants. FEBS Lett 581:22732280 (2007). 20 Bityustkii NP, Effects of carboxylic and phosphonic Fe-chelates on root and foliar plant nutrition. Russ J Plant Physiol 42:444453 (1995). 21 Stock D and Holloway PJ, Possible mechanisms for surfactant-induced foliar uptake of agrochemicals. Pestic Sci 38:165177 (1993). 22 Schonherr J, Cuticular penetration of calcium salts: effects of humidity, anions, and adjuvants. J Plant Nutr Soil Sci 164:225231 (2001). 23 Fern ndez V and Ebert G, Foliar iron fertilization: a critical review. a J Plant Nutr 28:21132124 (2005).

CONCLUSIONS
IDHA/Fe3+ performed similarly to EDTA/Fe3+ and revealed a potential ability to provide Fe to soybean when applied through foliar sprays under controlled conditions and with the appropriate adjuvant. Moreover, the application of IDHA/Fe3+ was more effective when urea was also used. Thus, foliar application with IDHA/Fe3+ and urea formulations, both degradable compounds, may constitute an environmentally friendly alternative to the traditional use of EDTA/Fe3+ in foliar Fe fertiliser formulations, although further research is recommended to optimise formulations and application procedures.

ACKNOWLEDGEMENTS
This work was partly supported by ADOB PPC, and by the Spanish Ministry of Science and Education Project AGL200763756, co-nanced with ERDF and the European Commission. P. Rodrguez-Lucena was on a Spanish Ministry of Science and

2639

J Sci Food Agric 2010; 90: 26332640

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org
24 Espasa-Manresa R, La fertilizacion foliar con amino cidos. Horticultura a 2:3335 (1983). 25 Briat JF, Curie C and Gaymard F, Iron utilization and metabolism in plants. Plant Biol 10:276282 (2007). 26 Pestana M, Correia PJ, De Varennes A, Abada J and Araujo-Faria E, Effectiveness of different foliar iron applications to control iron chlorosis in orange trees grown on a calcareous soil. J Plant Nutr 24:613622 (2001). 27 Huve K, Remus R, Luttschwager D and Merbach W, Transport of foliar applied iron (59 Fe) in Vicia faba. J Plant Nutr 26:22312242 (2003). 28 Nikolic M, Cesco S, Romheld V, Varanini Z and Pinton R, Uptake of iron (59 Fe) complexed to water-extractable humic substances by sunower leaves. J Plant Nutr 26:22432252 (2003).

P Rodrguez-Lucena et al.
29 Romera FJ, Lucena C and Alc ntara E, Plant hormones inuencing a iron uptake in plants, in Iron Nutrition in Plants and Rhizospheric Microorganisms, ed. by Barton LL and Abada J. Springer-Verlag Academic Publishers, Dordrecht, pp. 255278 (2006). 30 Swietlik D and Faust M, Foliar nutrition of fruit crops. Hort Rev 6:287356 (1984). 31 Nikolic M, Cesco S, Romheld V, Varanini Z and Pinton R, Short-term interactions between nitrate and iron nutrition in cucumber. Funct Plant Biol 34:402408 (2007). 32 Orbovic V, Jifon JL and Syvertsen JP, Foliar-applied surfactants and urea temporally reduce carbon assimilation of grapefruit leaves. J Am Soc Hort Sci 126:486490 (2001).

2640
wileyonlinelibrary.com/jsfa
c 2010 Society of Chemical Industry J Sci Food Agric 2010; 90: 26332640

Research Article
Received: 5 May 2010 Revised: 8 July 2010 Accepted: 20 July 2010 Published online in Wiley Online Library: 24 August 2010

(wileyonlinelibrary.com) DOI 10.1002/jsfa.4133

Inuence of genetic matrix and crop year on chemical and sensory proles of Italian monovarietal extra-virgin olive oils
Annalisa Rotondi,a Barbara Alfei,b Massimiliano Maglia and Giorgio Pannellic
Abstract
BACKGROUND: Commercial virgin olive oils belonging to the cultivars (Bosana, Carolea, Coratina, Frantoio, Itrana, Leccino, Moraiolo, Peranzana, Piantone di Mogliano and Ravece) most represented at the Italian National Review of Monovarietal olive oils (Rassegna Nazionale Italiana degli oli Monovarietali) were considered. The evaluation of the inuence of the cultivar and of the crop year as well as their interaction on oil composition were statistically analysed by a complete factorial design by principal components analysis and by linear discriminant analysis. RESULTS: In fatty acids composition, the effect of the cultivar and crop year and their interaction were highly signicant. The statistical analysis showed that the sensory attributes (olive fruity, grassy, fresh almond, artichoke, tomato, aromatic herbs, bitter and pungent) were strongly inuenced by the cultivar. The prevalent effect of the cultivar on the sensory prole was also demonstrated by the low or absent level of signicance observed in the crop year. CONCLUSION: The construction of a databank based on a large number of samples, which is available at URL http://www. olimonovarietali.it, has contributed to the reduction of the variable effects involved in the oil production process. Knowledge of the chemical and sensory proles of the Italian monovarietal olive oils could start a certication process of these oils, thus giving greater guarantees about their origin. c 2010 Society of Chemical Industry Keywords: monovarietal in extra virgin olive oils; Olea europaea; cultivar; chemical and sensory identity; statistical analyses

INTRODUCTION
The safeguarding of Italian autochtonous olive cultivars represents an important aim which each region is pursuing. This is demonstrated by the continuous increase of monovarietal olive oils participating in the Italian National Review of Monovarietal olive oils organised in Marche Region. Between 2005 and 2008, 986 olive oil samples originating from 117 varieties and 18 different Italian regions, have reached the Review organised by Agenzia Servizi Settore Agroalimentare delle Marche (ASSAM), Istituto Centro di Ricerca per lOlivicoltura e lIndustria Olearia, sede distaccata di Spoleto (CRA- OLI) and Il Sole24ore-Business Media. The richness of Italian autochthonous olive germplasm, represented by 822 cultivars,1 which are still increasing, guarantees the production of high quality extra-virgin olive oils, contributing in such a way as to maintain a large part of the olive genetic ancient biodiversity. Since the chemical and sensory characters of extra-virgin olive oil (EVOO) are strongly affected by the genotype of origin, the safeguarding and characterisation of cultivars and clones play a key role in the marketing of high quality standard olive oils. Several typical Italian EVOOs have received a European Protected Origin Denomination (POD) trademark or a European Protected Geographical Indication (PGI) trademark. These typical oils are generally blends of different varieties or monovarietal

EVOOs. The studies of the quality of monovarietal oils increase the value of the product and, at the same time, inform the consumer of their nutritional and organoleptic value. In Italy the market of monovarietal and organic oils is increasing thanks to the presence of consumers who are paying greater attention to the pleasurable aspect as well as the health aspect.2,3 The knowledge and the increase in the characteristics of Italian monovarietal EVOOs will also improve the knowledge of areas where these oils are produced, with a consequent increase in tourism, a very important sector for the Italian economy. In Italy, a new regulation was recently introduced compelling virgin and extra-virgin olive oil producers to indicate the

Correspondence to: Annalisa Rotondi, Istituto di Biometeorologia, Consiglio Nazionale delle Ricerche, Via Gobetti 101, I-40129 Bologna, Italy. E-mail: a.rotondi@ibimet.cnr.it

a Istituto di Biometeorologia, Consiglio Nazionale delle Ricerche, Via Gobetti 101, I-40129 Bologna, Italy b Agenzia Servizi Settore Agroalimentare delle Marche, Via Alpi 21 I-60131 Ancona, Italy

2641

c CRA, Istituto Centro di Ricerca per lOlivicoltura e lIndustria Olearia, sede distaccata di Spoleto, Via Nursina 2, I-06049 Spoleto, Perugia, Italy

J Sci Food Agric 2010; 90: 26412648

www.soci.org

c 2010 Society of Chemical Industry

www.soci.org

A Rotondi et al.

Table 1. Cultivar, geographical region and crop year of 489 monovarietal olive oils Crop year (no. of samples) Cultivar Bosana Carolea Coratina Frantoio Itrana Leccino Moraiolo Peranzana Piantone Di Mogliano Ravece Total no. of samples Geographical region Sardegna Calabria Puglia Marche, Lazio, Toscana, Umbria Lazio Abruzzo, Marche, Toscana, Umbria Toscana, Umbria Puglia Marche Campania 2005/2006 11 8 8 14 13 8 7 7 10 4 90 2006/2007 20 5 11 19 19 13 14 7 6 10 124 2007/2008 28 5 16 21 9 21 17 7 7 14 145 2008/2009 23 3 14 13 14 17 15 6 11 14 130 Total no. of samples 82 21 49 67 55 59 53 27 34 42 489

600

500 Total phenols (mg gallic acid kg-1)

400

300

200

100

0
Bosana Carolea Coratina Frantoio Itrana Leccino Moraiolo Peranzana Piantone M. Ravece

Varietes

Figure 1. Mean values and standard error of the mean of the total phenols relative to the 489 monovarietal olive oil samples. Contents were expressed as mg kg1 acid gallic of oil.

geographical location of olive harvesting and oil production on the label.4 More recently, the European Community (EC) Council of Regulation established compulsory standards on olive oil production regarding labelling giving the origin for extra-virgin and virgin olive oils.5 Quality and typicality of EVOO are primarily determined by genetic, agronomical and environmental factors, even if the technological parameters of oil processing also play an important role.6,7 Nutritional and sensory properties of EVOOs are strongly affected by several agronomical factors such as cultivar, fruit ripeness, crop yield, and growing area.8 10 With regard to fatty acid composition, Uceda and Hermoso,11 in a preliminary olive germplasm bank evaluation, concluded that the cultivar was the main source of variability for the major fatty acids. The strong relationship between the variety and the composition of fatty acids is also shown by several studies.12,13 EVOO is also well known for its high content of phenolic substances, which are thought to have health-promoting

properties.14,15 Phenolic compounds are the main responsible agents for the resistance against autoxidation and photoxidation.16 They also inuence some sensorial properties of olive fruits and virgin olive oils.17,18 The large increase of the demand of EVOO is due not only to its health virtues, but also to its organoleptic properties. The richness of Italian olive cultivars allows the production of different monovarietal EVOOs characterised by a wide range of pleasant sensory avours. Also, the phenols, as well as the sensory properties of EVOO, are inuenced by the cultivar19 and environmental factors20 or their interaction. Some authors investigated the effect of cultivarenvironment interaction on the composition of EVOO.21,22 Environments where olive trees grow may inuence plant physiology as well as the olive ripening process. During the ripening process, the fruit changes its chemical composition, with the activation and inhibition of different enzymatic activities, also affecting oil composition. A smaller number of studies has considered the seasonal effect on the chemical and sensory prole of EVOOs. The seasonality which

2642

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 26412648

2643

J Sci Food Agric 2010; 90: 26412648

Carolea 0.40 0.01 0.33 0.03 0.07 0.01 0.08 0.00 7.10 0.16 0.78 0.02 78.37 0.24 10.77 0.16 0.45 0.03 1.65 0.06 0.36 0.01 0.29 0.01 0.06 0.01 0.10 0.00 7.01 0.12 0.69 0.01 76.90 0.23 12.13 0.14 0.82 0.03 1.63 0.04 0.31 0.01 0.22 0.02 0.06 0.00 0.09 0.00 6.38 0.14 0.77 0.02 78.12 0.25 11.62 0.15 0.83 0.03 1.59 0.05 0.32 0.01 0.22 0.01 0.06 0.01 0.10 0.00 6.69 0.17 0.74 0.02 75.56 0.26 13.65 0.13 1.28 0.17 1.54 0.04 0.36 0.01 0.27 0.02 0.05 0.00 0.09 0.00 7.14 0.12 0.76 0.02 76.23 0.23 12.54 0.17 0.76 0.02 1.53 0.05 Coratina Frantoio Itrana Leccino Moraiolo Peranzana 0.39 0.01 0.35 0.04 0.05 0.00 0.08 0.00 9.50 0.26 0.78 0.02 73.88 0.40 12.48 0.25 0.75 0.02 1.75 0.06 Piantone di Mogliano 0.42 0.01 0.32 0.03 0.12 0.01 0.23 0.01 6.67 0.19 0.79 0.02 77.44 0.35 11.43 0.22 0.69 0.03 1.86 0.05 Ravece 0.57 0.14 0.01 0.06 0.06 0.01 0.09 0.01 9.52 0.20 0.78 0.02 73.43 0.28 12.31 0.13 0.66 0.03 2.56 0.08

Table 2. Mean values and standard error of the mean of the fatty acids relative to the 489 monovarietal olive oil samples

Genetic matrix and crop year effects on Italian monovarietal olive oil quality

c 2010 Society of Chemical Industry

Fatty acid

Bosana

Eicosanoic Eicosenoic Heptadecanoic Heptadecenoic Linoleic Linolenic Oleic Palmitic Palmitoleic Stearic

0.41 0.01 0.26 0.01 0.07 0.01 0.10 0.01 10.24 0.14 0.72 0.02 73.05 0.20 12.48 0.11 0.76 0.02 1.93 0.05

0.39 0.02 0.28 0.02 0.14 0.02 0.24 0.04 7.13 0.26 0.61 0.02 74.81 0.60 13.31 0.34 1.16 0.11 1.92 0.08

www.soci.org

wileyonlinelibrary.com/jsfa

www.soci.org is deeply related to the different climate events of the crop year, may inuence the ripening process of olives, affecting in this way the oil composition.23,24 In order to create a database of chemical and sensory proles of Italian Monovarietal EVOOs, an elevated number of observations, available every year, could allow the continuous update of the database and give a more accurate data set. The aim of the study is to describe the nutritional properties, expressed as fatty acid and total phenols contents, and the sensory proles of Italian monovarietal EVOOs. In addition, the evaluation of the inuence of the cultivar and seasonality, as well as their interaction on oil composition, were analysed.

A Rotondi et al.

Table 3. Variability expressed as % of the total sum of the squares for fatty acid composition and total phenols Parameter Eicosanoic Eicosenoic Heptadecanoic Heptadecenoic Linoleic Linolenic Oleic Palmitic Palmitoleic Stearic Total phenols Cultivar 19.67 12.75 53.08 67.04 86.17 12.14 82.28 63.90 62.85 53.39 37.40 Crop year 10.56 78.53 20.69 7.92 6.07 81.20 9.06 28.27 12.10 36.45 49.02

Cultivar crop year 69.77 8.72 26.23 25.04 7.76 6.66 8.66 7.83ns 25.05ns 10.16 13.58 0.01 or

MATERIALS AND METHODS

Commercial virgin olive oil samples (n = 489) belonging to the ten cultivars most represented (n 20) at the Review and declared monovarietal by the producer, were obtained from the Italian National Review of Monovarietal Olive Oils during four successive editions (crops 2005/2006, 2006/2007, 2007/2008 and 2008/2009). Table 1 shows geographical region, cultivar, crop year and number of samples. Sensory and chemical analyses of oils were evaluated each year during the Review 24 months after their processing. Sensory analysis was determined by the ASSAM Marche Panel recognised by the IOOC (International Olive Oil Council) and the Italian Ministry for Agriculture, Food and Forestry Policy. The evaluation of the samples was performed under the conditions described in the EC Regulation 640/2008.25 Each taster smelled and tasted the oil under consideration, in order to analyse the olfactory, gustatory, tactile and kinaesthetic characteristics. Thirteen attributes were evaluated: nine during the olfactory phase (olive fruity, olive fresh leaf, grass, fresh almond, artichoke, tomato, apple, berries and aromatic herbs) and four during the gustatory phase (olive fruity, bitter, pungent and the uidity). Attributes were assessed on an oriented 10 cm line scale and quantied measuring the location of the mark from the origin. The data obtained for the 13 descriptors were used to dene the sensory prole of each sample using the median values. Fatty acids composition was determined by gas chromatography (GC) according to Regulation EC Reg.796/200226 methodology by Centro Agrochimico ASSAM, Jesi (AN), Italy. Total phenolic content, expressed in mg kg1 acid gallic of oil, was determined by a colorimetric method using a spectrophotometer (Perkin Elmer model Lambda 3b; Perkin Elmer, Waltham, MA, USA). Chemical and sensory data were processed using SAS software 9.1.3 (SAS Institute Inc., Cary, NC, USA). Explorative analysis and descriptive statistics were performed for each set of data in order to identify outliers, extreme observations and to obtain distributional properties of the data. Descriptive measures (moments, basic measures of location and variability, condence intervals for the mean, standard deviation, and variance) of chemical and sensory variables were calculated for each monovarietal oil. Subsequently the statistical procedure was based on the analysis of variance (ANOVA) by a complete factorial design in order to examine treatment interdependencies (variety and crop year). A principal components analysis (PCA) and a linear discriminant analysis (LDA) were also performed on chemical and sensory data separately, using mean values of each crop year of each cultivar collected. The fatty acid content of monovarietal olive oil samples was submitted to the ANOVA procedure by a complete factorial design.

, , Signicant F-values: the respectively; ns = not signicant.

0.05,

0.001 levels,

Fatty acids with the highest index of variability (heptadecenoic, linoleic, oleic, stearic and palmitic) were selected according to their P level and F values and submitted to PCA and LDA.

RESULTS AND DISCUSSION

Table 2 reports mean values and standard errors of the mean of the fatty acids belonging to the 489 monovarietal olive oil samples. Regarding the oleic acid, whose range contents reported to EC Ofcial Reg. EC Regulation 702/200727 varies from 55% to 83%, the monovarietal EVOOs considered in this study are generally characterised by high levels (superior to 73%), moreover Coratina and Itrana monovarietal EVOOs showed the highest amount of acid oleic (above 78%). The percentage of the other individual fatty acids (markers of genuineness), such as myristic, linolenic, arachidonic, eicosenoic, and behenic acid were within the limits set by the ofcial normal standard.27 As shown in Fig. 1, oil of the Coratina cultivar presented the highest value of total phenols (543.06 mg kg1 ), while Itrana oils showed the lowest level of 305.02 mg kg1 . However, the phenolic content of the other monovarietal EVOOs, having a range between 376.37 and 498.90, showed their high levels of nutritional quality. In order to evaluate the treatment interdependencies (cultivar and crop year) on the fatty acid and phenolic contents, an ANOVA procedure by complete factorial design was made. In Table 3 it is reported that in all fatty acids analysed, the effect of the cultivar and crop year is highly signicant. The effect of the interaction between the two factors is also highly signicant on the content of fatty acid, except palmitic and palmitoleic acids. Also, the total phenolic contents are deeply inuenced both by the cultivar and year, while the interaction between the two factors has showed a minor level of signicance. It is interesting to underline that both factors (cultivar and crop year) inuence at the same level of signicance the contents of the most important fatty acids as linoleic, linolenic, oleic, palmitic and palmitoleic; however for oleic and linoleic acid the main factor was the cultivar, in fact ANOVA procedure explains the 82.28% and 86.17% of its variation respectively. The cultivar did not represent a great source of variability for linoleic acid: only 12.14%, while the crop year shows a variation of 81.20%. Figure 2 shows the PCA, applied on the mean values of each monovarietal olive oil and fatty acids content. The gure describes

2644

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 26412648

Genetic matrix and crop year effects on Italian monovarietal olive oil quality

www.soci.org

Figure 2. Principal component analysis applied to the fatty acids (heptadecenoic, linoleic, oleic, stearic and palmitic) expressing the highest index of variability. The gure describes each monovarietal EVOO considering the four following crop years individually. B, Bosana; P, PeranzanA; R, Ravece; CA, Carolea; L, Leccino; PM, Piantone di Mogliano; M, Moraiolo; C, Coratina; F, Frantoio; I, Itrana.

each monovarietal EVOO considering the four following crop years individually. This analysis explains the 71.16% of variability representing the discrimination among some cultivars. At the top of the gure are localised Bosana (B), Peranzana (P) and Ravece (R) oils. These are characterised by a higher level of linoleic acid, while cv. Itrana (I) and Coratina (C) by a higher level of oleic acid. Cv. Carolea (CA) showed a higher level of heptadecenoic acid. Data plotted on the principal components are, however, spread inside some cultivars, conrming the signicant effect of the cultivar and crop year on the fatty acid composition, as previously shown by the ANOVA procedure. Linear discriminant analysis (LDA) on fatty acids conrmed the result of the PCA. The analysis of the misclassied observations showed that the LDA model presented a relatively low specicity (77.5%), thus, while being able to identify four varieties (Itrana, Peranzana, Piantone di Mogliano and Ravece), its ability to classify the rest of observations is fairly limited; the majority of these misclassied observations were Bosana (50%) and Frantoio (75%) (Table 4). Considering organoleptic quality for each of the 489 olive oil samples collected, the sensory prole was dened. Table 5 shows mean values and standard error of the mean of the sensory attributes of the monovarietal EVOOs. The monovarietal EVOOs produced by cultivars Bosana, Frantoio, Itrana, Peranzana and Ravece exhibited a higher intensity (mean, about 5) of olive fruity odour. The intensity of olive fruity perceived in the above mentioned oils, was also conrmed by the gustative route. All monovarietal EVOOs have presented signicant intensities of grass attribute with the highest levels of 3.01 and 2.90, respectively, perceived in Itrana and Ravece oils. Considering the peculiar attributes which are deeply cultivar-dependent, like fresh almond, artichoke, tomato, aromatic herbs and berries,28 30 oils produced by Coratina, Frantoio, Leccino, Moraiolo and Piantone di Mogliano

Table 4. Linear discriminant analysis (error count estimates for varieties) applied on fatty acids Cultivar Bosana Carolea Coratina Frantoio Itrana Leccino Moraiolo Peranzana Piantone di Mogliano Ravece 0.50 0.25 0.25 0.75 0.00 0.25 0.25 0.00 0.00 0.00

are distinguished for their high intensity of fresh almond, a typical pleasant avour which characterised these cultivars. All monovarietal oils considered in this study presented a signicant intensity of artichoke avours. Bosana, Peranzana and Ravece and oils exhibited the highest intensity, while in Piantone di Mogliano and Leccino oils, artichoke and tomato attributes were slightly perceived. With regard to this last attribute, oils of Ravece and Itrana are distinguished also for their high intensity of tomato avour. Itrana oil, previously described for its high intensity of olive fruity odour, is also characterised by a high intensity of tomato avours. Aromatic herbs avour is signicantly present only in Itrana and Ravece oils. All monovarietal EVOOs considered in this study were characterised by a signicant level of bitterness showing a range from 3.65 to 5.14. In particular, Piantone di Mogliano, Leccino and Itrana

2645

J Sci Food Agric 2010; 90: 26412648

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org

A Rotondi et al.

Table 5. Mean values and standard error of the median of the sensory attributes relative to the 489 monovarietal olive oil samples Sensory attribute Olive fruity (olf.) Olive fruity (gus) Grass Fresh almond Artichoke Tomato Aromatic herbs Bitter Pungent Piantone Di Mogliano

Bosana

Carolea

Coratina

Frantoio

Itrana

Leccino

Moraiolo

Peranzana

Ravece

5.08 0.07 4.81 0.24 4.80 0.11 5.05 0.09 5.39 0.15 4.63 0.13 4.91 0.09 5.06 0.09 4.31 0.13 5.69 0.10 5.00 0.06 4.63 0.19 4.83 0.08 4.92 0.08 5.08 0.12 4.38 0.11 4.81 0.10 4.88 0.12 4.28 0.14 5.39 0.09 2.61 0.13 2.14 0.33 1.56 0.19 2.39 0.17 3.01 0.20 1.61 0.19 1.84 0.19 2.72 0.28 1.50 0.20 2.90 0.20 1.64 0.12 1.61 0.24 2.20 0.18 2.61 0.13 1.22 0.16 2.34 0.13 2.15 0.15 1.51 0.20 2.09 0.15 1.08 0.15 2.16 0.12 1.73 0.26 1.69 0.17 1.51 0.16 1.83 0.17 0.99 0.14 1.80 0.16 2.28 0.23 0.98 0.17 2.15 0.15 0.76 0.11 0.41 0.14 0.43 0.11 0.39 0.10 2.14 0.18 0.19 0.06 0.53 0.12 1.11 0.23 0.41 0.14 2.47 0.22 0.09 0.04 0.06 0.06 0.06 0.03 0.02 0.02 0.29 0.09 0.03 0.02 0.02 0.02 0.14 0.08 0.09 0.06 0.30 0.12 4.84 0.08 4.30 0.33 5.14 0.14 4.61 0.13 4.01 0.13 3.99 0.16 4.58 0.15 4.13 0.18 3.65 0.15 4.70 0.13 4.52 0.07 3.78 0.24 4.79 0.11 4.39 0.10 3.85 0.13 3.99 0.13 4.39 0.12 4.08 0.12 4.03 0.16 4.75 0.12

Table 6. Variability expressed as % of the total sum of the squares for sensory attributes Parameter Olive fruity (olf.) Olive fruity (gus.) Grass Fresh almond Artichoke Tomato Aromatic herbs Bitter Pungent

Cultivar 72.65 68.07 65.34 71.53 56.63 83.58 42.22 73.20 57.34

Crop year 8.84 5.01ns 5.29ns 7.48 4.87ns 5.16ns 10.49ns 8.21 10.69

Cultivar crop year 18.51ns 26.92ns 29.37ns 20.99ns 38.50 11.26 47.29ns 18.59ns 31.97 0.01 or

clear grouping inside each cultivar (Fig. 3), explaining the 79% of variability. LDA analysis on sensory attributes conrmed the result of the PCA. The analysis of the misclassied observations showed that the LDA model presented a high specicity (85.0%). The model is able to identify correctly ve varieties (Coratina, Itrana, Leccino, Peranzana and Ravece); the majority of these misclassied observations were Bosana observations (50%) (Table 7).

CONCLUSION
The decision to carry out a study of an increased number of labelled commercial extra-virgin olive oils has had the precise aim to provide the consumer with information about the chemical and sensory properties of extra-virgin olive oils which are actually available on the Italian market. The authors are aware of the numerous variables: mills typology, olive ripening index and agronomic practices, which inuence the overall olive oil quality. These variables are not usually known for commercial oils. The construction of a databank based on a large number of samples which is available at URL http://www.olimonovarietali.it, has contributed to the reduction of the variable effects involved in the oil production process. Moreover the continuous increase of oil samples participating in the Italian National Review of monovarietal olive oils will allow the improvement of the chemical and sensory proles of each oil. The exact knowledge of sensory proles of the Italian monovarietal olive oils represents an important step, considering that the production of monovarietal EVOOs has greatly increased during the past few years. As reported by Chiavaro et al.,2 the characterisation of monovarietal oils is needed to identify specic properties that distinguish them and increase their value with respect to other niches of EVOOs such as PDO. In this study only ten of the 117 cultivars, available in the databank, which represent the Italian olive genetic resources are reported. Further studies will be carried out involving other cultivars. Considering the cultivars Frantoio and Leccino, which are widely diffused along the Italian peninsula, studies are in progress to estimate the effect of the environment (climate, altitude and latitude) on the chemical and sensory proles of olive oils participating in the Italian National Review.

, , Signicant F-values the respectively; ns = not signicant.

0.05,

0.001 levels,

oils presented the lowest intensity of bitterness. It is interesting to underline that the same oils were also characterised by the lowest contents of total phenols (see Fig. 1). On the contrary, the Coratina oil which exhibited the highest intensity of bitterness, also presents the highest phenolic content. These results show a clear positive correlation between bitterness and total phenols (r2 = 0.65) according to other authors.31 In the present study the monovarietal oils show values of pungency quite similar among them. In fact the range varied from 3.78 to 4.75, but correlation between bitterness and pungency is high (r2 = 0.75). The sensory proles of the 489 monovarietal olive oil samples were submitted to the ANOVA procedure by a complete factorial design. Table 6 shows that all sensory attributes: olive fruity (odour), olive fruity (taste), grass, fresh almond, artichoke, tomato, aromatic herbs, bitter and pungent were strongly inuenced by the cultivar. The prevalent effect of the cultivar on the sensory prole of monovarietal oils was demonstrated by the low or absent level of signicance observed in the crop year, relative interaction and by the variability, expressed as percentage of the total sum of the squares, of the factor cultivar characterised by a range from 42.22% to 83.58%. The prevalent effect of the cultivar with respect to the crop year is also evidenced by the PCA, where the data plotted showed a

2646

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 26412648

Genetic matrix and crop year effects on Italian monovarietal olive oil quality

www.soci.org

Figure 3. Principal component analysis applied to the sensory attributes (olfactive and gustative olive fruity, grass, fresh almond, artichoke, tomato, aromatic herbs, bitter and pungent). The gure describes each monovarietal EVOO considering the four following crop years individually. B, Bosana; P, PeranzanA; R, Ravece; CA, Carolea; L, Leccino; PM, Piantone di Mogliano; M, Moraiolo; C, Coratina; F, Frantoio; I, Itrana.

Table 7. Linear discriminant analysis (error count estimates for varieties) applied on sensory attributes Cultivar Bosana Carolea Coratina Frantoio Itrana Leccino Moraiolo Peranzana Piantone di Mogliano Ravece 0.50 0.25 0.00 0.25 0.00 0.00 0.25 0.00 0.25 0.00

REFERENCES
1 Bartolini G, Olive Germplasm (Olea europaea L.): cultivars, synonyms, cultivation area, collections (2008). Available: http://www.oleadb.eu/. [30 April 2010]. 2 Chiavaro E, Vittadini E, Rodriguez-Estrada MT, Cerretani L and Bendini A, Monovarietal extra virgin olive oils. Correlation between thermal properties and chemical composition: heating thermograms. J Agric Food Chem 56:496501 (2008). 3 Consolandi C, Palmieri L, Severgnini M, Maestri E, Marmiroli N, Agrimonti C, et al, A procedure for olive oil traceability and authenticity: DNA extraction, multiplex PCR and LDR-universal array analysis. Eur Food Res Technol 227:14291438 (2008). 4 Italian Government Legislative Decree of 18 October 2007 Regulation on compulsory indications which appear on the labelling of virgin and extra virgin olive oil. Ofcial Gazette Italian Republic 243:89 (2007). 5 European Commission Regulation 182/2009. Ofcial Journal of European Community, 6 March 2009, amending Regulation (EC) N. 1019/2002, L63:68 (2009). 6 Cerretani L, Bendini A, Rotondi A, Lercker G and Gallina Toschi T, Analytical comparison of monovarietal virgin olive oils obtained by both a continuous industrial plant and a low-scale mill. Eur J Lipid Sci Technol 107:93100 (2005). 7 Angerosa F, Mostallino R, Basti C and Vito R, Inuence of malaxation temperature and time on the quality of virgin olive oils. Food Chem 72:1928 (2001). 8 Lazzez A, Perri E, Caravita M, Khalifa M and Cossentini M, Inuence of olive maturity stage and geographical origin on some minor components in virgin olive oil of the Chemlali variety. J Agric Food Chem 56:982988 (2008). 9 Rotondi A, Bendini A, Cerretani L, Mari M, Lercker G and Gallina Toschi T, Effect of olive ripening degree on the oxidative stability and organoleptic properties of cv. Nostrana di Brisighella extra virgin olive oil. J Agric Food Chem 52:36493654 (2004). ` 10 Morello JR, Romero MP and Motilva MJ, Effect of the maturation process of the olive fruit on the phenolic fraction of drupes and

The knowledge and the information of the chemical and sensory proles of the Italian monovarietal olive oils could start a certication process of these oils, giving in such a way greater guarantees about their origin and consequently greater guarantees of quality for the consumer. The importance of monovarietal oils is becoming more signicant in restaurants and eating establishment, as a result of associating local oils with typical dishes.

ACKNOWLEDGEMENTS
We gratefully thank Il Sole24ore-Business Media, Srl, Via Patecchio 2, Milano, Italy.

2647

J Sci Food Agric 2010; 90: 26412648

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org
oils from Arbequina, Farga, and Morrut cultivars. J Agric Food Chem 52:60026009 (2004). Uceda M and Hermoso M, La Calidad del Aceite de Oliva, in El Cultivo del Olivo, ed. by Junta de A, Sevilla and Mundi-Prensa, Madrid, pp. 589614 (2001). DImperio M, Dugo G, Alfa M, Mannina L and Segre AL, Statistical analysis on Sicilian olive oils. Food Chem 102:956965 (2007). ` Allalout A, Krichnene D, Methenni K, Taamalli A, Oueslati I, Daoud D, et al, Characterization of virgin olive oil from super intensive Spanish and Greek varieties grown in northern Tunisia. Sci Hort 129:7783 (2009). Menendez JA, V squez-Martn A, Colomer R, Carrasco-Pancorbo A, a Garca-Villalba R, Fern ndez-Guti rrez A, et al, Olive oils bitter a e principle reverses acquired autoresistance to trastuzumab (Herceptin) in HER2-overexpressing breast cancer cells. BMC Cancer 7:780 (2007). Bendini A, Cerretani L, Carrasco-Pancorbo A, Gomez-Caravaca AM, Segura-Carretero A, Fern ndez-Guti rrez A, et al, Phenolic a e molecules in virgin olive oils. Survey of their sensory properties, health effects, antioxidant activity and analytical methods. An Overview of the last decade. Molecules 12:16781719 (2007). Baccouri O, Cerretani L, Bendini A, Lercker G, Zarrouk M and Ben Miled D, Phenol contents correlated to antioxidant activity and gustative characteristics of Tunisian monovarietal virgin olive oils. Riv Ital Sost Grasse 85:189195 (2008). Boti J, Ortuno A, Benavente-Garci O and Baidez AG, Modulation of a a the biosynthesis of some phenolic compounds in Olea europaea L. fruits: their inuence on olive quality. J Agric Food Chem 49:355358 (2001). Carrasco-Pancorbo A, Gomez-Caravaca AM, Segura-Carretero A, Cerretani L, Bendini A and Fern ndez-Guti rrez A, Use of capillary a e electrophoresis with UV detection to compare the phenolic proles of extra virgin olive oils belonging to Spanish and Italian PDOs and their relation to sensorial properties. J Sci Food Agric 89:21442155 (2009). Oliveras-Lopez MJ, Innocenti M, Giaccherini C, Ieri F, Romani A and Mulinacci N, Study of the phenolic composition of Spanish and Italian monocultivar extra virgin olive oils: distribution of lignans, secoiridoidic, simple phenols and avonoids. Talanta 73:726732 (2007).

A Rotondi et al.

11 12 13

14

15

16

17

18

19

20 Paz Aguilera M, Beltr n G, Ortega D, Fern ndez A, Jimenez A and a a Uceda M, Characterization of virgin olive oil of Italian olive cultivars: Frantoio and Leccino, grown in Andalusia. Food Chem 89:387391 (2005). 21 Kotti F, Chiavaro E, Cerretani L, Barnaba C, Gargouri M and Bendini A, Chemical and thermal characterization of Tunisian extra virgin olive oil from Chetoui and Chemlali cultivars and different geographical origin. Eur Food Res Technol 228:735742 (2009). 22 Rotondi A and Lapucci C. Nutritional properties of extra virgin olive oils from the Emilia-Romagna region: proles of phenols, vitamins and fatty acids, in Olives and Olive Oil in Health and Disease Prevention, ed. by Preedy VR and Watson RR. Academic Press, Oxford, pp. 725733 (2010). 23 Rotondi A and Magli M, Ripening of olives var. Correggiolo: modication of oxidative stability of oils during fruit ripening and oil storage. J Food Agric Environ 2:193199 (2004). 24 Beltr n G, Del Rio C, S nchez S and Martnez L, Inuence of harvest a a date and crop yield on the fatty acid composition of virgin olive oils from cv. Picual. J Agric Food Chem 52:34343440 (2004). 25 European Commission Regulation No. 640/2008. Off J Eur Commun L 178:1116 (2008). 26 European Commission Regulation No. 796/2002. Off J Eur Commun L 128:828 (2002). 27 European Commission Regulation No. 702/2007. Off J Eur Commun L 161:1127 (2007). 28 Angerosa F, Inuence of volatile compounds on virgin olive oil quality evaluate by analytical approaches and sensory panels. Eur J Lipid Sci Technol 104:639660 (2002). 29 Aparicio R, Morales MT and Alonso V, Authentication of European virgin olive oils by their chemical compounds, sensory attributes and consumer attitudes. J Agric Food Chem 45:10761083 (1997). 30 Tura D, Prenzler PD, Bedgood D R, Antolovich M and Robards K, Varietal and processing effects on the volatile prole of Australian olive oils. Food Chem 84:341349 (2004). 31 Inarejos-Garca AM, Androulaky A, Salvador D, Fregapane G and Tsimidou, MZ, Discussion on the objective evaluation of virgin olive oil bitterness. Food Res Int 42:279284 (2009).

2648
wileyonlinelibrary.com/jsfa
c 2010 Society of Chemical Industry J Sci Food Agric 2010; 90: 26412648

Research Article
Received: 27 February 2010 Revised: 23 July 2010 Accepted: 23 July 2010 Published online in Wiley Online Library: 25 August 2010

(wileyonlinelibrary.com) DOI 10.1002/jsfa.4135

Physiological optimality, allocation trade-offs and antioxidant protection linked to better leaf yield performance in drought exposed mulberry
Anirban Guha, Debashree Sengupta and Attipalli Ramachandra Reddy
Abstract
BACKGROUND: Mulberry (Morus spp. L.), usually linked to silkworm rearing, is now considered as a potential forage for livestock feeding and has great potential in world agriculture. Trait-based investigations for leaf yield stability in mulberry under water stress have not been studied extensively. The present study aims to identify candidate traits conferring leaf yield stability in mulberry under drought. RESULTS: Four popular, indigenous mulberry cultivars (Morus indica L. cvs AR-12, K-2, M. Local and V-1) were investigated. Low leaf temperature (Tl ), higher internal/ambient CO2 ratios (Ci /Ca ), greater stomatal conductance to CO2 (gs ) and stability in photosystem II efciency were associated with better net photosynthetic rates (Pn ) in V-1, generating maximum leaf yield when compared to other drought-exposed cultivars. Increased accumulation of foliar -tocopherol and ascorbic acidglutathione pool, associated with higher carotenoids, proline and glycine betaine, facilitated lower lipid peroxidation and better leaf yield in V-1 under drought. CONCLUSION: Minimal plasticity in photosynthetic gas exchange traits and better quantitative growth characteristics were attributed to leaf yield stability under drought. Lower photoinhibition, stabilized photochemistry, effective osmoregulation and enhanced activity of foliar antioxidants extensively contributed to drought tolerance and higher leaf yield in mulberry. c 2010 Society of Chemical Industry Keywords: antioxidants; chlorophyll a uorescence; drought; leaf yield stability; mulberry; photosynthesis

INTRODUCTION
Mulberry (Morus spp L.), a pioneer tree of secondary succession, was one of the rst commercialized foliage crops in the world and has been cultivated transcontinentally, covering 50 countries across the globe.1 In the past, mulberry cultivation was predominantly linked to the silkworm (Bombyx mori L.) and sericulture industry. However, research studies over the last two decades have revealed several other potential uses of mulberry. Recently, mulberry has been gaining popularity as forage for livestock feeding as the leaves are highly rich in protein, antioxidants and minerals, without any toxic elements. Several reports have emphasized the potential of mulberry foliage as a feed for ruminant and non-ruminant animals.2,3 Hence the purpose of mulberry research is largely aimed at producing more foliage with palatable, succulent leaves of high nutritive value. Recently evolved high-yielding mulberry genotypes and improved crop husbandry techniques have caused a massive increase in harvestable leaf yield in the moriculture sectors of India.4 However, mulberry is also cultivated under resourcelimited agro-ecosystems and many regions of the country are currently facing the prospect of producing yields with low water availability. Nearly 48% of the cultivated mulberry areas in India have been clustered under rainfed water-stress conditions.

Mulberry requires 500700 L water to produce 1 kg of fresh leaf.5 High-yielding mulberry cultivars (cvs) consume large quantities of water due to their faster growth rate, high biomassproducing foliage, large leaf area and canopy size; hence water deprivation can severely arrest mulberry growth, leading to yield loss. The relationship between yield loss and water stress severity can differ largely among crop genotypes. This has given rise to the concept of drought tolerance (DT), with some genotypes performing better under a given severity of water stress than others. Even though such variations have been recognized for many years, progress towards understanding and exploiting the mechanisms that confer tolerance has been slow in many crop species due to inconsistent use of the term DT and the difculties encountered while quantifying it. From an agro-economical viewpoint, the functional denition of DT

Correspondence to: Attipalli Ramachandra Reddy, Photosynthesis and Plant Stress Biology Laboratory, Department of Plant Sciences, School of Life Sciences, University of Hyderabad, Hyderabad 500046, India. E-mail: arrsl@uohyd.ernet.in Photosynthesis and Plant Stress Biology Laboratory, Department of Plant Sciences, School of Life Sciences, University of Hyderabad, Hyderabad 500046, India

2649

J Sci Food Agric 2010; 90: 26492659

www.soci.org

c 2010 Society of Chemical Industry

www.soci.org should be based on yield stability, which precisely indicates less uctuation in yield components in a drought-tolerant genotype compared to the susceptible when exposed to water decit. Stabilizing yield performance requires optimization of the physiological processes involved in the critical stages of plant response to soil dehydration. Several morphophysiological and biochemical mechanisms including water use efciency, photosynthetic capacity, radiation use efciency, rooting vigour, antioxidative protection, osmotic adjustment and staying green, are linked to enhanced performance and yield stability in agreement with stress tolerance syndrome.6,7 A priori knowledge of these candidate crop traits contributing to DT is essential before designing crop improvement programmes for any crop model. In spite of being an age-old crop, such trait-based investigations for yield stability in drought-exposed mulberry cvs have not been carried out extensively, except for a few reports.8 11 Hence the present investigation was focused on identifying candidate crop traits conferring DT in mulberry. Our major objectives were: (a) to identify the morphophysiological traits that would maximize biomass production as well as the use of available soil moisture under drought; (b) to investigate the drought-induced allocation trade-offs, partitioning to leaf biomass for enhanced crop water productivity (yield achieved per amount of water used); and (c) to link the oxidative stress and mechanisms of antioxidant protection to understand the overall physiological and growth performance of mulberry cvs under water decit conditions.

A Guha, D Sengupta, AR Reddy season I (FebruaryApril 2009) and season II (MayJuly 2009). Mean precipitation during the experimental period was only 49 mm and was considered insignicant to inuence the study; mean air temperature recorded during daytime ranged from 30.7 to 40.8 C; mean photosynthetically active radiation (PAR) measured between 09.00 and 11.00 solar hours ranged from 1230 to 2500 mol m2 s1 . The soil of the experimental site was sandy loam with a pH of 7.5. Experimental design, stress treatments and sampling The experimental design was a randomized block with a pit system of plantation and split plot arrangement of treatments,4,12 including four replications. The mulberry cvs were evaluated under two water regimes: well-watered (control) and water-stressed (drought). The control plot received two irrigations per week, whereas the stressed plot was exposed to natural drought and only received survival irrigation (two to three irrigations in an entire growing season) to avoid depletion of soil moisture content below 20% (at a soil depth of 3040 cm). To verify the degree of drought intensity, soil moisture content at 3040 cm depth of the control and stressed plots was determined gravimetrically on every sampling day. The soil moisture content of the wellwatered and water-stressed plots ranged from 55% to 65% and 22% to 28%, respectively, throughout the experimental period. The mulberry plantations were maintained as medium bush by pruning techniques and other farm practices were followed as recommended.4 All the samplings and measurements were performed in fully developed young leaves of the third or fourth position from the apex of top branches. The rst sampling and measurements were conducted in the rst week of March 2009. Thereafter, all estimations were periodically conducted at intervals of 2025 days in each growing season and the results reported are the mean of all periodic data obtained over two summer growing seasons (seasons I and II). Measurement of leaf gas exchange parameters The rate of leaf gas exchange was measured using a portable infrared CO2 /H2 O gas analyser (IRGA) (LCpro-32 070, ADC Bioscientic Ltd, Great Amwell, UK) equipped with a broad leaf chamber. The gas analyser was used to measure instantaneous net photosynthetic rate (Pn ), stomatal conductance to CO2 (gs ) and transpiration rate (E), periodically during each growing season between 10.00 and 11.00 a.m. on clear sunny days. The instant water

MATERIALS AND METHODS

Field experiments Plant materials, study site and climatic conditions The tested plant materials consisted of four 2-year-old indigenous Morus indica cvs AR-12, K-2, M. Local and V-1 grown at the experimental plots of the University of Hyderabad, India (17.3 10 N, 78 23 E; 542.6 m above sea level). The origin and some of the important genotypic characters of the tested cvs are listed in Table 1. Hyderabad, the region of the present study, is in the hottest part of the state of Andhra Pradesh in India and has a hot steppe climate characterized by high irradiance and air temperature during summer (FebruaryJuly), followed by southwest monsoon (AugustOctober). All samplings and estimations were performed during the dry summer months (FebruaryJuly) of 2009 over two consecutive growth seasons:

Table 1. Genotypic characteristics of four mulberry cvs (AR-12, K-2, M. Local and V-1) selected for drought tolerance experiment in this study Characteristics Genetic origin Branching nature Phyllotaxy Internodal distance (cm) Leaf nature Leaf lobation Leaf shape Leaf colour Sex expression Ploidy level Rooting (%)
a

AR-12 Cross-breed Erect 1/3 4.6 Homophyllous Unlobed Cordate Dark green Unisexual male 3n = 42 >80 OPHa

K-2 selection Erect Mixed 4.5 Homophyllous Unlobed Ovate Light green Unisexual female 2n = 30 >80

M. Local Land race Erect 1/3 3.7 Heterophyllous Medium lobed Ovate Dark green Unisexual male or female 2n = 28 >90

V-1 OPH selection Erect Mixed 4.5 Homophyllous Unlobed Narrow ovate Dark green Unisexual male 2n = 28 >90

2650

OPH, open pollinated hybrid.

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 26492659

Characteristics conferring leaf yield stability in drought-stressed mulberry use efciency (WUEi ) was calculated (WUEi = Pn /E). The plants were also analysed for internal CO2 concentration (Ci ) and internal to ambient CO2 ratios (Ci /Ca ). Leaf temperature (Tl ) was measured simultaneously by an integral leaf thermistor probe (M.PLC-011, LICOR, ADC) attached to the leaf chamber. All photosynthetic measurements were performed in situ on young, well-expanded and light-exposed leaves, randomly chosen from the upper half of the plant canopy of uniform plants in each replicate. Measurement of leaf yield (LY) and determination of drought susceptibility index (DSI) At the end of each growing season, the leaves were harvested and the weight recorded as fresh weight (FW) of leaves in grams. Drought susceptibility index (DSI) was calculated:13 DSI = (1 Yd /Yc )/DI where Yd and Yc are the leaf yields of the waterstressed and well-watered plants of one particular cv., respectively and DI is the drought intensity for the conditions of the given experimental time span. DI is dened as DI = 1 Xd /Xc , where, Xd and Xc are the mean leaf yields of all compared cvs under water-stressed and well-watered conditions, respectively Thus, DSI = (1 Yd /Yc )/(1 Xd /Xc ); if DSI <1 the cv. is drought tolerant, whereas if DSI >1 the cv. is drought susceptible (in terms of leaf yield harvest). Glasshouse experiments Experimental design, glasshouse conditions and sampling The second phase of study was executed under a glasshouse chamber to avoid difculties in eld observations due to the southwest monsoon, which usually arrives in our study zone immediately after the dry spell of 6 months (FebruaryJuly). Threemonth-old, healthy potted (pot size 35 L) saplings of all the tested mulberry cvs were arranged in a completely randomized block design (CRBD) with four replications. The plants were randomly submitted to two watering treatments: water-stressed pots were maintained at 2530% pot capacity (PC), whereas the well-watered plants were maintained at 100% PC. The study was undertaken for a time period of 75 days (20 August 2009 to 30 October 2009), calculated from the day after initiation of treatments. Chamber walls and ceiling were transparent to sunlight. Mean air temperature inside the glasshouse ranged from 22 1 C (early morning) to 34 4 C (early afternoon) and relative humidity from 20% 5% to 41% 2%. Samplings were conducted periodically at an interval of 1518 days and the results represented are the mean values of all periodic data. Measurement of leaf water status To monitor the effectiveness of the dry-down treatment, leaf relative water content (RWC) was measured and calculated:14 RWC (%) = [(FW DW)/(SW DW)] 100, where FW is fresh weight, SW is the turgid mass after rehydration obtained by storing leaf samples for 24 h in distilled water and DW is oven-dried weight (105 C for 24 h) of leaves. Leaf moisture content (LMC) was estimated by using the formula LMC (%) = [(FW DW)/FW] 100. Measurement of light-saturated net photosynthetic rates (LSPn ), photosynthetic radiation use efciency (PRUE) and chlorophyll a uorescence Non-detached, young and fully expanded leaves from each cv. under each treatment were used to measure LSPn during 09.00 to 11.00 a.m. by using a portable IRGA at a saturating PAR of

www.soci.org

1600 mol m2 s1 . The PAR was provided by an LED light source attached to a broad light unit (LCpro Lamp 32 070 Broad, ADC) of the leaf chamber. The air humidity in the leaf chamber was about 60% with a CO2 concentration of 350360 mol mol1 and ambient air temperature was 25 2 C. PRUE was determined:15 PRUE = light-saturated net photosynthetic rate (LSPn )/incident PAR. Chlorophyll a uorescence was measured using a portable plant efciency analyseruorometer (Handy PEA-2126, Hansatech, Kings Lynn, UK). Measurements were made on dark-adapted (30 min) leaves at pre-dawn (5.005.30 a.m.). The ground uorescence yield (Fo ) and maximal uorescence induction (Fm ) were obtained by illuminating the leaves with a beam of saturating light (3000 mol m2 s1 ) of 650 nm peak wavelength obtained from three light-emitting diodes, focused on a circle of 5 mm diameter of the leaf sample. The rst reliably measured point of the uorescence transient is at 20 s, which was taken as Fo . The variable uorescence (Fv = Fm Fo ) and the maximum quantum yield of PSII (Fv /Fm ) was then estimated. Measurements were done in 10 replications. A single leaf per plant constituted each replicate. Investigation on growth, biomass allocation and yield parameters Periodically, some of the potted mulberry plants were completely harvested (both shoots and roots), while remaining plants were harvested at the end of the experiment (75th day) to obtain periodic as well as nal growth and yield records. The numbers of branches were calculated and the length of all the branches of a plant was added to obtain the total shoot length (TSL). Leaf weight and stem weight of each plant were recorded as fresh weight in grams and added to obtain above-ground biomass. Leaf area (cm2 ) was recorded using a LICOR leaf area meter (LI 1600) and instant leaf area duration (LAD, cm2 d1 ) was computed. Root fresh weight (g) was recorded and root volume (cm3 ) was measured following the water displacement method.16 Total plant biomass, leaf mass ratio (LMR), stem mass ratio (SMR), root mass ratio (RMR) and root : shoot ratio were calculated using the data of leaf, stem and root weights. The periodically as well as nally harvested plant tissues were oven-dried at 70 C for 72 h and, based on these data, relative growth rate (RGR), net assimilation rate (NAR) and biomass duration (BMD) were calculated.17 Analysing antioxidative protection, osmoprotectants and lipid peroxidation In the third phase of study we narrowed down to only two mulberry cvs (one drought tolerant and one drought susceptible, screened out from the rst two phases of our experiment). Carotenoids were extracted from fresh leaves (0.5 g) with 10 mL of 80% (v/v) acetone followed by centrifugation at 10 000 g for 5 min. The absorbance of cleared extract was read at 663.2, 646.8 and 470 nm in a UV-visible 160A spectrophotometer (Shimadzu, Tokyo, Japan) and total carotenoid content was calculated.18 For foliar ascorbic acid(AA) determination, fresh leaf tissue (0.5 g) was homogenized with 5 mL of 10% trichloroacetic acid (TCA), centrifuged at 10 000 g for 20 min at room temperature, re-extracted twice, supernatants were pooled and 0.5 mL of the extract was used to estimate AA.19 For measuring total glutathione, the reaction mixture (1 mL) contained 100 L sample, 100 L distilled water, 700 L of 0.3 mmol L1 NADPH in potassium phosphate buffer (20 mmol L1 , pH 7.5) and 6 mmol L1 5 -dithiobis(2-nitrobenzoic acid) (DNTB). The rate of DNTB reduction was recorded by adding 10 L glutathione reductase (GR) at 412 nm for 3 min.20 For quantifying -tocopherol,21 fresh leaf tissue

2651

J Sci Food Agric 2010; 90: 26492659

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org

A Guha, D Sengupta, AR Reddy

Figure 1. Effect of water decit on photosynthetic gas exchange characteristics and leaf yield performance: (A) net photosynthetic rate (Pn ); (B) stomatal conductance to CO2 (gs ); (C) transpiration rate (E); (D) instant water use efciency (WUEi ); (E) internal CO2 concentration (Ci ); (F) minimal value of internal to ambient CO2 ratio (Ci /Ca )min ; (G) leaf temperature (Tl ); (H) leaf yield/plant; (I) drought susceptibility index (DSI, based on leaf yield) in four mulberry cvs grown under two water regimes (well-watered and water-stressed). All measurements and leaf harvests were conducted during two summer seasons of the study zone: season I (FebruaryApril 2009) and season II (MayJuly 2009). Data represented are the average over two summer seasons. Values are mean SD.

(1 g) was homogenized in liquid nitrogen with 10 mL of cold 80% methanol, centrifuged at 3000 g for 15 min at 4 C. The supernatant was ltered through a 25 mm Millipore lter and the ltrate was stored in ice under dark conditions. The -tocopherol content of the ltrate was determined by high-performance liquid chromatography (HPLC) with a C18 reverse-phase column (250 10.00 mm 5 m, Phenomenex, Maccleseld, UK) at a ow rate of 1 mL min1 , using an isocratic solvent system methanol and ethyl acetate (1 : 4, v/v) as an eluant. -Tocopherol level was quantied by a UV-visible detector system at 295 nm (LC-10 AT VP, Shimadzu, Japan). Peak identication was performed by comparing the retention times with pure -tocopherol standard (Sigma, St Louis, MO, USA). To determine the content of free proline, fresh leaf tissue (0.5 g) was homogenized in 10 mL of 3% sulfosalicyclic acid. The homogenate was centrifuged at 9000 g for 15 min at room temperature and the supernatant was used to determine free proline content.22 For glycine betaine (GB) quantication, fresh leaf tissue was nely ground, mechanically shaken with 20 mL deionized water for a period for 24 h at 25 C. The samples were then ltered and the collected ltrate was used for estimation of GB.23 The extent of lipid peroxidation in leaf tissues was determined by measuring the malondialdehyde (MDA) content. Tissue samples (0.5 g) were homogenized in liquid nitrogen with 5 mL of 0.1% TCA followed by centrifugation at 5000 g for 10 min at 4 C. Thereafter, 500 L of the extract was used for determination of MDA equivalents.24 Statistical analysis Results were represented as mean standard deviation (SD). The signicance of the differences between the treatments was determined using paired t-tests. Correlation coefcient (r) and coefcient of determinations (r2 ) of linear relationships between the investigated parameters were established by using linear regression. The linear regression slopes were analysed using bivariate correlation signicance tests. Data were also analysed by analysis of variance (ANOVA) to determine signicant differences

between individual cv. and water stress treatments. All statistical analyses were performed and graphs drawn using the statistical packages Excel and Sigma Plot 11.0.

RESULTS
Field experiments Photosynthetic leaf gas exchange and leaf yield performance Effect of drought stress on Pn , gs and E depicted contrasting variability among the cvs and between the treatments. Signicant genotypic variation was recorded for Pn among the four mulberry cvs in both water regimes (Fig. 1(A)). Under well-watered conditions, cv. AR-12 exhibited the highest Pn (13.2 mol m2 s1 ), followed by V-1 (12.58 mol m2 s1 ), K-2 (11.45 mol m2 s1 ) and M. Local (9.03 mol m2 s1 ). Water decit signicantly (P < 0.05) reduced Pn in all the mulberry cvs. Reduction in Pn was strong in cv. AR-12 (70.7%), followed by K-2 (64%) and M. Local (61.4%) under drought stress. However, higher Pn under water decit was recorded in V-1 (7.9 mol m2 s1 ) with a minimum reduction of 37.2% compared to well-watered counterparts. The gs ranged from 0.46 to 0.26 mol m2 s1 and from 0.19 to 0.05 mol m2 s1 in control and stress treatments, respectively (Fig. 1(B)). Under well-watered conditions, V-1 (0.46 mol m2 s1 ) exhibited maximum gs , followed by K-2 (0.40 mol m2 s1 ) and AR-12 (0.38 mol m2 s1 ), whereas gs was lower in M. Local (0.26 mol m2 s1 ). Soil water deprivation caused a dramatic downregulation in the rates of gs for all mulberry cvs. The lowest value for gs under low water regime was recorded in AR-12 (0.05 mol m2 s1 ), followed by K-2 (0.06 mol m2 s1 ) and M. Local (0.07 mol m2 s1 ). However, cv. V-1 exhibited maximum gs (0.19 mol m2 s1 ) under low soil moisture. Drought caused similar drastic reduction of E (70%) associated with lack of irrigation in all the cvs (Fig. 1(C)). V-1 maintained relatively higher rates for E under low water regimes. Relative to unstressed plants, drought led to a signicant (P < 0.01) increase in WUEi in mulberry cvs V-1 (26.6%) and M. Local (36.6%) (Fig. 1(D)). In K-2 there was a marginal increment of 9.7% in WUEi under water stress, whereas no signicant change in WUEi of drought-exposed AR-12 was

2652

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 26492659

Characteristics conferring leaf yield stability in drought-stressed mulberry

www.soci.org

Figure 2. Relationships between: (A) stomatal conductance to CO2 (gs ) versus net photosynthetic rate (Pn ); (B) transpiration rate (E) versus instant water use efciency (WUEi ); (C) internal CO2 concentration (Ci ) versus net photosynthetic rate (Pn ); (D) stomatal conductance to CO2 (gs ) versus internal CO2 concentration (Ci ). Linear regressions were tted to data for four mulberry cvs grown under well-watered and water-stressed conditions during February 2009 to July 2009. Each point represents the mean of ve independent measurements from an individual plant. The correlations were signicant at P < 0.001.

witnessed when compared to control plants. Ci of well-watered cvs ranged from 257 to 267 mol mol1 ; however, drought resulted in an apparent reduction of Ci in all the mulberry cvs, although to a different extent depending upon the genotype (Fig. 1(E)). The reduction in Ci was signicantly (P < 0.05) sharper in K-2 (39.4%) and M. Local (36.3%) compared to V-1 (11.4%) during the drought stress period. The lowest values recorded for Ci /Ca [(Ci /Ca )min ] under dry-land conditions were 0.23, 0.29, 0.33 and 0.47 for K-2, AR-12, M. Local and V-1, respectively (Fig. 1(F)). Water stress caused signicant (P < 0.01) elevation in Tl for the drought exposed mulberry cvs (Fig. 1(G)). The highest Tl of stressed mulberry leaves was observed in AR-12 (36.5 C) followed by M. Local (35.9 C) and K-2 (35.4 C), while V-1 maintained the lowest value for Tl (32.6 C) throughout the periods of water deprivation. The assessment of leaf yield production depicted wide genetic variability between both the cv. and treatment levels (Fig. 1(H)). Under well-irrigated conditions, the highest leaf yield per plant was recorded in V-1 (1190 g per plant) followed by AR-12 (698 g per plant) and M. Local (624 g per plant). A signicant (P < 0.01) decline in leaf yield was recorded in all the cvs under water-stressed treatments. However, maximum leaf yield per plant was obtained in V-1 (450 g per plant) under drought conditions exhibiting a DSI of 0.85, whereas minimum leaf yield was recorded in AR-12 (166 g per plant, DSI 1.12) followed by K-2 (168 g per plant, DSI 1.02) and M. Local (180 g per plant, DSI 1.05) (Fig. 1(I)). Relationships among photosynthetic gas exchange variables The drought-induced regulation of Pn by gs followed a linear function, resulting in a signicant positive correlation between Pn and gs in all the mulberry cvs (Fig. 2(A)). The regression slopes were steeper for the data from K-2 (r2 = 0.93, P < 0.001) and M. Local (r2 = 0.81, P < 0.001). For V-1, the correlation between Pn and gs was comparatively weak though signicant (r2 = 0.64, P < 0.001). Signicant negative correlations were evident when

Table 2. Leaf moisture content and leaf relative water content in four mulberry cvs grown in a glasshouse under two water regimes (well-watered and water-stressed). Values are mean SD Leaf moisture content (%) Cultivars Well-watered Water-stressed AR-12 K-2 M. Local V-1 74.3 1.2 72.5 1.1 70.2 1.4 76.1 0.8 66.3 0.7 67.5 0.9 66.5 0.7 72.6 0.8 Leaf relative water content (%) Wellwatered 82.5 2.0 82.1 3.8 80.2 1.9 86.8 2.7 Water-stressed 67.4 1.7 68.5 0.9 68.6 2.6 74.4 1.9

data for E and WUEi were plotted together for all the mulberry cvs (Fig. 2(B)). The correlations were stronger for V-1 (r2 = 0.97, P < 0.001) and M. Local (r2 = 0.82, P < 0.001); however, they were weak for K-2 (r2 = 0.26 P < 0.001) and AR-12 (r2 = 0.41, P < 0.001). The relationships between Pn versus Ci and Ci versus gs were linear and positively correlated (Fig. 2(C, D)). However, the stomatal closure induced a prompt and higher inhibition of Pn compared to Ci . In fact, for a reduction of 0.28 mol m2 s1 in gs , Pn decreased 58.3%, while Ci decreased 30%. The decline in Ci seemed to depend linearly on gs and was strongly correlated for all four mulberry cvs. The slope of regression for Ci versus gs curve was steeper for AR-12 (r2 = 0.94, P < 0.001), followed by K-2 (r2 = 0.89, P < 0.001), M. Local (r2 = 0.86, P < 0.001) and V-1 (r2 = 0.84, P < 0.001). Glasshouse experiments Leaf water status Effect of drought stress on LMC and leaf RWC of mulberry cvs are presented in Table 2. For all the cvs LMC decreased under water-

2653

J Sci Food Agric 2010; 90: 26492659

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org

A Guha, D Sengupta, AR Reddy

Table 3. Changes in growth and yield characteristics in four mulberry cvs subjected to drought stress under glasshouse conditions AR-12 Variables Plant height (cm) TSL (cm) No. branches per plant Total leaf area (cm2 ) Leaf weight (g) LAD (cm2 d1 ) Stem weight (g) Root weight (g) Root length (cm) Root volume (cm3 ) Root : shoot (g g1 ) Above ground biomass (g) Total plant biomass (g) RGR (g g1 d1 ) NAR (g m2 d1 ) BMD (g d1 ) Wellwatered 76.6 173.1 6.2 2894.2 156.2 41.5 118.1 142.7 87.4 55.5 0.52 274.3 417.4 0.31 1.6 6.1 Water-stressed 36.2 77.2 3.7 908.6 18.4 13.2 24.3 58.8 67.8 30.5 1.38 42.7 101.5 0.24 0.28 1.4

K-2 Wellwatered 84.2 211.3 7.2 2782.5 125.3 40.9 128.2 144.8 86.5 58.7 0.57 253.5 398.3 0.32 1.5 5.9 Water-stressed 31.1 76.2 3.2 924.4 21.6 13.4 34.9 66.2 65.6 35.5 1.17 56.5 122.5 0.22 0.33 1.8

M. Local Wellwatered 95.3 188.4 7.0 2112.3 148.4 31.5 136.6 166.4 108.1 67.2 0.58 285.3 451.4 0.30 1.7 6.6 Water-stressed 37.3 71.3 3.2 921.7 24.3 13.5 46.5 76.2 78.4 38.8 1.07 70.8 147.1 0.26 0.35 2.2

V-1 Wellwatered 97.4 196.5 7.5 3041.6 191.7 44.6 241.7 171.1 108.6 98.4 0.40 433.4 604.5 0.36 1.8 8.9 Water-stressed 46.3 113.2 4.5 1274.6 46.5 20.7 42.7 108.3 88.9 65.3 1.21 89.2 197.5 0.31 n 0.77 3.1

Effects of drought were tested by paired t-test. P < 0.001; P < 0.01; P < 0.05; n.s., not signicant.

stressed conditions, but for V-1 this decrease was minimal and the cv. maintained a comparatively higher LMC of 72.6%. Under low water regimes, the RWC was reduced signicantly (P < 0.05) in all four mulberry cvs compared to control (Table 2). However, V-1 showed tolerance to drought with the highest RWC of 74.4%, whereas the decrease in RWC was higher in AR-12 and dropped to 67% during periods of water deprivation. PRUE and chlorophyll a uorescence Signicant genotypic variation was encountered for LSPn and PRUE among the four mulberry cvs in both water regimes (Fig. 3(A)). LSPn showed essentially the same variations as PRUE

(hence the LSPn values were not shown), since the PAR was xed and due to the high correlation between LSPn and PRUE (r2 = 0.96, P < 0.001). Under well-irrigated conditions, PRUE ranged from 8.8 to 6.2 (mmol (CO2 ) mol1 (photons)). AR-12 had the highest PRUE, followed by V-1 (8.8 and 8.3 mmol (CO2 ) mol1 (photons), respectively) under normal physiological conditions, while PRUE values were comparatively low in the control plants of M. Local (6.2 mmol (CO2 ) mol1 (photons)). Drought imposed on the mulberry cvs decreased PRUE signicantly (P < 0.05) and the values ranged from 5.2 to 2.3 (mmol (CO2 ) mol1 (photons)) under low water regimes. The maximum drop in PRUE was evident in stressed AR-12 plants (71.5%), followed by K-2 (64.4%) and

Figure 3. Effect of water stress on photosynthetic radiation use efciency (PRUE) and chlorophyll a uorescence characteristics: (A) PRUE; (B) percentage reduction in PRUE; (C) ground uorescence (Fo ); (D) maximal uorescence (Fm ); (E) maximal quantum yield of PSII (Fv /Fm ) in four mulberry cvs grown in a glasshouse under two water regimes (well-watered and water-stressed). Values are mean SD.

2654

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 26492659

Characteristics conferring leaf yield stability in drought-stressed mulberry

www.soci.org

Table 4. Signicance of the inuence of treatment, cultivar and product of treatment cultivar on the variance of 10 growth parameters measured under glasshouse conditions according to multivariate analyses of variance Growth parameter TSL Total leaf area Leaf weight Stem weight Above-ground biomass Root length Root weight Root volume Root : shoot mass ratio Total plant biomass

Treatment

Cultivar

Treatment cultivar

P < 0.001; P < 0.01; P < 0.05.

Figure 4. Biomass allocation patterns in leaf (leaf mass ratio, LMR) stem (stem mass ratio, SMR) and root (root mass ratio, RMR) fractions of four mulberry cvs grown in a glasshouse under well-watered (left half of pie diagrams) and water-stressed (right half of the pie diagrams) conditions. Data on fresh weights of leaf, stem and root fractions from Table 3 have been taken to depict the pie diagrams.

M. Local (61.6%) (Fig. 3(B)). However, the highest PRUE value was observed in V-1 (5.2 mmol (CO2 ) mol1 (photons)) during water decit conditions. V-1 also executed a minimum reduction of 37.3% in PRUE compared to the well-watered counterparts. Chlorophyll a uorescence parameters were signicantly affected by drought stress. In all the tested mulberry cvs, drought stress induced an increase in Fo which was signicantly more in K-2, AR-12 and M. Local when compared to V-1 (Fig. 3(C)). Unlike Fo , a reduction in Fm was encountered in the stressed leaves of mulberry cvs compared to the well-watered counterparts (Fig. 3(D)). Depression in Fm was more discrete in AR-12 (12%), whereas the reduction was minimal in the stressed leaves of V-1 (0.8%). The Fv /Fm ratio in control leaves of four mulberry cvs remained above 0.8 throughout the experiment. The ratio was also practically not altered in V-1 (0.81) under low water regimes. However, in K-2 the ratio dropped signicantly (0.70, P < 0.05) in response to drought when compared to well-watered plants (Fig. 3(E)). Growth, allocation trade-offs and yield performance Substantial alterations in growth, allocation and productivity traits were detected in response to drought in four mulberry cvs (Table 3). The main effect of water stress was a signicant reduction in all shoot criteria (plant height, TSL, number of branches, stem weight, leaf weight, cumulative leaf area and above-ground biomass) as assessed at the end of experimental period. AR-12, K-2 and M. Local cvs suffered greater reductions in all shoot growth criteria than V-1. The interaction treatment cv. was signicant for all these parameters (Table 4). The four mulberry cvs also varied signicantly in respect to different root characteristics (Table 3). Discrete reduction in root FW was observed in mulberry cvs (15.5%) under water stress treatments; however, V-1 exhibited highest root FW compared to other cvs. Vertical proliferation of roots was retarded, in all the mulberry cvs under low water regimes,

which was reected in reduction of root length in the mulberry cvs of either group. However, the root length in V-1 was relatively larger irrespective of the watering regimes (108.6 cm and 88.9 cm in control and water stress treatments, respectively). The root volume was reduced to 45% and 42.2% in AR-12 and M. Local, respectively, in response to water stress; however, the reduction was minimum in drought-exposed V-1 (31%). The interaction treatment cv was signicant for the majority of root characteristics (Table 4). Drought stress also led to an apparent reduction in total plant biomass accumulation in all the mulberry cvs (Table 3). However, V-1 maintained comparatively higher plant biomass under waterlimited conditions when compared to others. The drought-treated plants showed a larger reduction in shoot growth compared to root, indicating a shift in biomass allocation towards below-ground organs. However, the four mulberry cvs differed in their response to biomass allocation. As indicated in Fig. 4, under well-watered conditions the biomass allocations to leaf and root tissues in the mulberry cvs were approximately 33.5%, and 33.7%, respectively. A substantial shift to root biomass allocation (RMR 0.56 g g1 ) was recorded in the water-stressed cvs compared to control counterparts (RMR 0.33 g g1 ), which concurrently increased the root : shoot mass ratios in all tested mulberry cvs (Table 3). A discrete depression in leaf biomass allocation was recorded in AR-12 (LMR, 0.18 g g1 ), K-2 (LMR, 0.17 g g1 ) and M. Local (LMR, 0.16 g g1 ) cvs, while droughtexposed V-1 exhibited signicantly more allocation to leaf biomass (LMR, 0.24 g g1 ) than the remaining cvs. V-1 had a greater allocation to stem (SMR, 0.39 g g1 ) compared to AR-12 (SMR, 0.28 g g1 ), M. Local (SMR, 0.31 g g1 ) and K-2 (SMR, 0.32 g g1 ) under well-irrigated conditions (Fig. 4(D)). However, allocation to stem was maximal in M. Local (SMR, 0.32 g g1 ) and minimal in V-1 (SMR, 0.22 g g1 ) under water-limited conditions (Fig. 4(B,D)). LAD ranged from 31.5 to 44.6 cm2 d1 and from 13.2 to 20.7 cm2 1 d under control and drought treatments, respectively (Table 3). Deprivation in soil moisture led to a severe reduction in LAD in all the mulberry cvs; however, V-1 exhibited the highest value for LAD during water stress periods. Under well-watered conditions, the cvs did not differ much in respect to RGR and NAR; however, drought stress led to a signicant drawdown in RGR and NAR in all the tested cvs (Table 3). V-1 exhibited relatively higher values for RGR (0.31 g g1 d1 ) and NAR (0.77 g m2 d1 ) under low water regimes, whereas RGR and NAR were found to be lowest in K-2 and

2655

J Sci Food Agric 2010; 90: 26492659

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org AR-12, respectively (Table 3). Regardless of the watering regimes, V-1 maintained higher values for BMD, while drought caused a dramatic reduction of BMD in other mulberry cvs with an average decline of 69.2% (Table 3). Signicant positive correlations were achieved when data for PRUE and leaf weight were plotted together for all the cvs (Fig. 5(A)). The regression slopes were steeper and the degree of signicance was high for all the mulberry cvs (AR-12, r2 = 0.96, P < 0.001; K-2, r2 = 0.98 P < 0.001; M. Local, r2 = 0.97, P < 0.001; V-1, r2 = 0.98, P < 0.001). The response of PRUE to total plant biomass accumulation was also strongly signicant for all mulberry cvs (M. Local (r2 = 0.98, P < 0.001), V-1 (r2 = 0.95, P < 0.001), AR-12 (r2 = 0.93, P < 0.001) and K-2 (r2 = 0.92, P < 0.001)] (Fig. 5(B)). Antioxidants and osmoprotectants Total carotenoid content was not signicantly different in the leaves of well-watered cvs. However, an increasing trend in total carotenoid accumulation was recorded in both the cvs under drought stress (Fig. 6(A)). V-1 showed a maximum level of total carotenoids (0.69 mg g1 FW) compared to K-2 (0.55 mg g1 FW) under drought conditions. The foliar AA concentration of wellirrigated plants was not signicantly different between the two tested cvs (AA 1.05 mg g1 FW). However, an increase in AA level was witnessed in the leaves of stressed V-1 plants exhibiting signicant elevation (about 36% increase, P < 0.05) compared to the minor increment observed in K-2 (only 7.5% increase) (Fig. 6(B)). -Tocopherol content showed contrasting variation

A Guha, D Sengupta, AR Reddy

Figure 5. Relationships between photosynthetic radiation use efciency (PRUE) versus (A) leaf weight and (B) total plant biomass. Linear regressions were tted to data for four mulberry cvs grown in a glasshouse under well-watered and water-stressed conditions. The solid lines are the linear regressions tted to all points for each cv. and are signicant for P 0.001.

Figure 6. Effect of water stress on the contents of (A) total carotenoids, (B) ascorbic acid (AA), (C) -tocopherol, (D) glutathione, (E) proline, (F) glycine betaine and (G) foliar malondialdehyde (MDA) in two mulberry cvs (V-1 and K-2) grown under glasshouse conditions. Values are mean SD.

2656

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 26492659

Characteristics conferring leaf yield stability in drought-stressed mulberry among the two mulberry cvs (Fig. 6(C)). V-1 exhibited a signicantly (P < 0.05) elevated level of -tocopherol under drought stress with respect to control, whereas the endogenous -tocopherol level of K-2 was affected severely and a net -tocopherol loss was visible in stressed leaves of K-2. V-1 had endogenously higher concentration of glutathione compared to K-2 (5.9 and 3.7 mol g1 FW, respectively). Under water decit conditions, V-1 exhibited more glutathione accumulation (7 mol g1 FW) than K-2 (4 mol g1 FW) (Fig. 6(D)). Differential levels of endogenous proline content were monitored in the two mulberry cvs in both well-watered and waterstressed plants (Fig. 6(E)). V-1 showed inherently higher proline content compared to K-2 during unstressed conditions and also accumulated signicantly (threefold increase, P < 0.05) under low water regimes. Similar to proline, concentration of GB increased in the water-stressed leaves of both mulberry cvs (Fig. 6(F)). However, increment in the level of GB was substantial in V-1 (52%, P < 0.05) when compared to K-2 (31%, P < 0.05). Water deciency augmented the MDA content in leaf tissues of the two tested mulberry cvs. An apparent increase in leaf MDA content occurred in stressed K-2 (46.2% increase, P < 0.05) compared to control, while MDA accumulation was considerably less in the stressed leaf tissues of V-1. An increase of 29% (P < 0.05) in MDA content was evident in drought-exposed V-1 leaves when compared to control (Fig. 6(G)).

www.soci.org

DISCUSSION
In the present study, the most important plastic response of drought stress in mulberry was the reduction in leaf yield per plant and associated traits, which included cumulative leaf area, leaf biomass, number of branches and LAD. However, such hypersensitive, phenotypic plasticity causing diminution of foliage cannot be linked to enhanced performance under drought as leaf yield is equally important besides crop sustenance in mulberry. Anomalous to drought-susceptible mulberry cvs (AR-12, M. Local and K-2) that underwent severe loss in leaf yield (DSI >1), V-1 showed drought tolerance (DSI <1), maintaining relatively higher and stable leaf yield during drought-exposed conditions. In agreement with the stress tolerator syndrome, not only higher biomass allocation to yield organ (LMR) but also a minimal plasticity in foliar gas exchange physiology conferred enhanced genotypic performance to V-1 under water-limited conditions. Focusing on leaf yield harvest per plant, we analysed the evidence of yield losses in the susceptible mulberry cvs compared to tolerant V-1. Limitation in leaf yield due to drought stress can be explained as a consequence of signicant decline in the rates of photosynthesis (Pn ), stomatal conductance to CO2 (gs ) and transpiration rates (E) in the tested mulberry cvs. As Pn decreased in parallel with gs , stomatal inhibition to CO2 conductance seem to be the principal cause of photosynthetic downregulation in mulberry leaves under water stress. The reductions in Pn and gs were accompanied by acute depressions in Ci during drought stress, supporting strong stomatal inhibition. Stomatal closure is a drought avoidance strategy that allows leaf moisture restoration, but a concomitant drawdown in Ci can dramatically downregulate Pn.25 27 There may be uncertainties in calculating Ci in water-stressed, heteroboric mulberry leaves due to stomatal patchiness.28,29 As we did not measure patchy stomatal conductance,30 32 we cannot be sure that our Ci values were not subjected to the patchiness problems. However, it should be noted that: (i) patchiness should not be an important problem under progressively drought-exposed

conditions in the eld; and (ii) gs values measured during the water-stressed periods for two consecutive summer months were always higher than 0.03 mol m2 s1 , below which the patchiness phenomenon can be important.32,33 The explicit corelationships between the photosynthetic gas exchange traits provided a vivid understanding of the functional interrelations amongst those traits. The positive linear relationships between Pn gs , Pn Ci and Ci gs , as evident within the tested mulberry cvs, prominently indicate the regulatory functions of gs and Ci on Pn in the dehydrated leaves of mulberry. Elevated Tl (>35 C), as recorded in control and stressed leaves of drought-susceptible mulberry cvs might also adversely affect mesophyll conductance to CO2 and cause damage to the photosynthetic machinery.34 Interestingly, the drought-tolerant cv. V-1 was able to maintain low Tl and therefore heat-induced damage on the photosynthetic apparatus was also presumably less critical for the same cv. The negative relationship between E and WUEi of the mulberry cvs was in agreement with ndings in other tree species such as European chestnut35 and western red cedar.36 However, in spite of sufcient reduction in E, WUEi was not found to be signicantly improved in the susceptible mulberry cvs, owing to their inability to maintain higher Pn under water stress conditions, whereas in the tolerant V-1 Pn was found to be higher and less sensitive to E; hence WUEi was consequently much improved in the same cv. The leaf-level efciency of carboxylation, commonly termed photosynthetic efciency, is the central process for plant performance and can be determined in terms of instantaneous PRUE, which characterizes net CO2 xation per incoming PAR on the leaf surface.37 Moreover, the efciency with which absorbed photons are nally used for photosynthetic electron transport and carbon xation can be determined using chlorophyll a uorescence techniques and denoted as maximum PSII photochemical efciency (Fv /Fm ).38 All three processes (photochemical efciency of PSII, leaf-level PRUE and Pn ) are ultimately light-driven and determine the carbon gain and biomass acquisition in plants. In the present study, water constraint directly affected Fv /Fm , which presumably contributed to jeopardizing Pn and PRUE in the susceptible mulberry cvs compared to tolerant V-1. The light-harvesting complexes reduced their efciency due to drought-induced damage to the antennae, as evidenced by an increase in Fo in droughtexposed leaves of susceptible cvs; however, in the stressed leaves of drought-tolerant V-1 the increase in Fo was relatively less and was balanced by the maintenance of Fm , which contributed to keep similar Fv /Fm . The decrease in Fm in the susceptible mulberry cvs may be related to the decrease in activity of the water-splitting enzyme complex and perhaps a concomitant cyclic electron transport with or around PSII.39 The decrease in Fv /Fm in the susceptible cvs suggests a chronic photoinhibition due to photoinactivation of PSII centres, possibly attributable to D1 protein damage.40 The strong, positive inuence of PRUE on leaf weight and total plant biomass (Fig. 5) suggests the importance of unimpaired photochemical efciency for photosynthetic yield acquisition in mulberry cvs under well-watered as well as water decit situations. The susceptible mulberry cvs revealed signicant differences in their responses to drought with respect to leaf and root morphology, whole-plant productivity and allocation when compared to tolerant V-1. Mulberry cv. V-1 not only showed the maximum LMR and cumulative leaf area in both the drought and control treatments, but also excelled in growth rates (NAR, RGR) and BMD (Table 3). A comparatively well-developed root system as evident in V-1 might be helpful in better hydraulic

2657

J Sci Food Agric 2010; 90: 26492659

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org conductance of the root system during drought and subsequently could maintain higher rates of E and gs , displaying highly efcient water use strategy.41 It is a widespread opinion that plants adapted to dry habitats reduce their foliage and thus their water loss as a conservative water use strategy. However, stable optimization strategies can confer the ability to consume large amounts of resources due to high explorative ability and thus allow stabilization in leaf morphophysiological characteristics, as witnessed in drought-tolerant cv. V-1.42 Better physiological functioning and growth performance in crops under water stress conditions have been connected to the enhanced activity of antioxidants for scavenging reactive oxygen species (ROS).43 Our results indicate a paramount increase in the concentrations of non-enzymatic antioxidants in V-1 compared to the susceptible K-2, exhibiting elevated antioxidative protection under water-deprived situations. AA was the major water-soluble antioxidant in V-1 leaves and the enhanced levels of AA, glutathione and -tocopherol in water-stressed V-1 elucidate the well-maintained redox buffering capacity of the acorbateglutathione cycle.44 An apparent increase in total carotenoids and -tocopherol content in V-1 indicates efcient scavenging of highly destructive singlet oxygen (1 O2 ) and lipid peroxidation products, thereby stabilizing the photosynthetic membranes of chloroplasts.45 It is interesting to note that the drought-susceptible K-2 showed endogenous loss in -tocopherol when exposed to drought which might occur due to irreversible degradation of -tocopherol as a limited AA availability in chloroplast of the same cv.46 The efcient antioxidative protection in stressed V-1 facilitated low MDA accumulation, whereas in K-2 increased levels of MDA accumulated as an aftermath of poor antioxidative defence response, making the cv succumb more to oxidative damage. A positive relationship between accumulation of compatible solutes (proline, GB) and drought tolerance was also observed in V-1, which could provide better osmotic equilibrium and cell membrane stability during water stress regime.

A Guha, D Sengupta, AR Reddy

REFERENCES
1 Vijayan K and Chatterjee SN, ISSR proling of Indian cultivars of mulberry (Morus spp.) and its relevance to breeding programs. Euphytica 131:5363 (2003). 2 Papanastasis VP, Yiakoulaki MD, Decandia M and Dini-Papanastasi O, Integrating woody species into livestock feeding in the Mediterranean areas of Europe. Animal Feed Sci Technol 140:117 (2008). 3 Machii H, Koyama A and Yamanouchi H, Mulberry breeding, cultivation and utilization in Japan, in MulberryforAnimalProduction: FAO Electronic Conference, Feed Resources Group (AGAP), FAO, Rome, 1 May31 June (2000). 4 Dandin SB, Jayaswal J and Giridhar K, Mulberry cultivation, in Handbook of Sericulture Technologies, ed. by Dandin SB, Jayaswal J and Giridhar K. Central Silk Board, Bangalore, pp. 3145 (2003). 5 Karaba NN, Sheshshayee MS and Udaykumar M, Biotech News: Options for Improvement Mulberry, Vol. III(5) (2008). 6 Chaves MM, Flexas J and Pinheiro C, Photosynthesis under drought and salt stress: regulation mechanisms from whole plant to cell. Ann Bot 103:551560 (2009). 7 Murchie EH, Pinto M and Horton P, Agriculture and the new challenges for photosynthesis research. New Phytol 181:532552 (2009). 8 Susheelamma BN, Jolly MS, Giridhar K and Sengupta K, Evaluation of germplasm genotypes for the drought resistance in mulberry. Sericologia 30:327340 (1990). 9 Ramanjulu S, Giridara NKS and Sudhakar C, Photosynthetic characteristics in mulberry during water stress and rewatering. Photosynthetica 35:259263 (1998). 10 Chaitanya KV, Jutur PP, Sundar D and Reddy AR, Water stress effects on photosynthesis in different mulberry cultivars. Plant Growth Regul 40:7580 (2003). 11 Karatassiou M, Parissi ZM, Abraham EM and Kyriazopoulos AP, Growth of Morus alba L. under water decit conditions. Options M diterran ennes Series A 70:315318 (2008). e e 12 Chaturvedi HK and Sarkar A, Optimum size and shape of the plot for mulberry experiments. Indian J Seric 39:6669 (2000). 13 Fisher RA and Maurer R, Drought resistance in spring wheat cultivars. I. Grain yield response. Aust J Agric Res 29:897912 (1978). 14 Castillo FJ, Antioxidative protection in the inducible CAM plant Sedum album L. following the imposition of severe water stress and recovery. Oecologia 107:469477 (1996). 15 Ma CC, Gao YB, Guo HY and Wang JL, Photosynthesis, transpiration, and water use efciency of Caragana microphylla, C. intermedia, and C. korshinskii. Photosynthetica 42:6570 (2004). 16 Burdett AN, A nondestructive method for measuring the volume of intact plant parts. Can J For Res 9:120122 (1979). 17 Lamers John PA, Khamzina A and Worbes M, The analyses of physiological and morphological attributes of 10 tree species for early determination of their suitability to afforest degraded landscapes in the Aral Sea Basin of Uzbekistan. For Ecol Manage 221:249259 (2006). 18 Lichtenthaler HK, Chlorophylls and carotenoids: pigments of photosynthetic biomembranes. Methods Enzymol 148:350382 (1987). 19 Omaye ST, Turnbull JD and Sauberilich HE, Selected methods for the determination of ascorbic acid in animal cells, tissues and uids. Methods Enzymol 62:311 (1979). 20 Grifth OW and Meister A, Potent and specic inhibition of glutathione synthesis by buthionine sulfoximine (s-n-butylhomocysteine sulfoximine). J Biol Chem 254:75587560 (1979). 21 Yen GC, Wu SC and Duh PD, Extraction and identication of antioxidant components from the leaves of mulberry (Morus alba L.). J Agric Food Chem 44:16871690 (1996). 22 Bates LS, Walderen RP and Teare ID, Rapid determination of free proline for water stress studies. Plant Soil 39:205207 (1973). 23 Grieve CM and Grattam SR, Rapid assay for determination of water soluble quaternary ammonium compounds. Plant Soil 70:303307 (1983). 24 Fu J and Huang B, Involvement of antioxidants and lipid peroxidation in the adaptation of two cool-season grasses to localized drought stress. Environ Exp Bot 45:105114 (2001). 25 Cornic G, Drought stress inhibits photosynthesis by decreasing stomatal aperture not by affecting ATP synthesis. Trends Plant Sci 5:187188 (2000).

CONCLUSION
From our present investigation, the following conclusions can be ascertained: (i) in agreement with stress tolerance syndrome, not only an increased biomass allocation to yield organ but also a minimal plasticity in foliar gas exchange characteristics conferred enhanced leaf yield performance to cv. V-1 under water-limited conditions; (ii) less photoinhibition and stabilized photochemistry facilitated higher PRUE in tolerant cv. V-1, showing enhanced photosynthetic yield under drought; (iii) the effective osmoregulation and increased activities of antioxidants also complemented drought tolerance in cv. V-1. Net photoassimilation rates and carbon allocation are the culmination of all these biological responses and chemical processes, which were relatively less affected in the drought-tolerant mulberry cv. V-1, resulting in better yield performance under a water stress environment.

ACKNOWLEDGEMENTS
We are grateful to Regional Sericultural Research Station, Salem, India, for providing the mulberry germplasm. We acknowledge the Department of Science and Technology (DST), Government of India, for nancial assistance (Grant SR/SO/PS-27/05). AG gratefully acknowledges DST for a research fellowship.

2658

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 26492659

Characteristics conferring leaf yield stability in drought-stressed mulberry

26 Flexas J and Medrano H, Drought inhibition of photosynthesis in C3 plants: stomatal and non stomatal limitations revisited. Ann Bot 89:183189 (2002). 27 Medrano H, Escalona JM, Bota J, Gulas J and Flexas J, Regulation of photosynthesis of C3 plants in response to progressive drought: stomatal conductance as a reference parameter. Ann Bot 89:895905 (2002). 28 Nikolopoulos D, Liakopoulos G, Drossopoulos I and Karabourniotis G, The relationship between anatomy and photosynthetic performance of heteroboric leaves. Plant Physiol 129:235243 (2002). 29 Gomes FP, Oliva MA, Mielke MS, Almeida A-AF de, Leite HG and Aquino LA, Photosynthetic limitations in leaves of young Brazilian Green Dwarf coconut (Cocos nucifera L. nana) palm under wellwatered conditions or recovering from drought stress. Environ Exp Bot 62:195204 (2008). 30 Terashima I, Wong SC, Osmond CB and Farquhar GD, Characterisation of non-uniform photosynthesis induced by abscisic acid in leaves having different mesophyll anatomies. PlantCellPhysiol 29:385394 (1988). 31 Mott KA and Buckley TN, Patchy stomatal conductance: emergent collective behaviour of stomata. Trends Plant Sci 5:258262 (2000). 32 Flexas J, Bota J, Escalona JM, Sampol B and Medrano H, Effects of drought on photosynthesis in grapevines under eld conditions: an evaluation of stomatal and mesophyll limitations. Funct Plant Biol 29:461471 (2002). 33 Grassi G and Magnani F, Stomatal, mesophyll conductance and biochemical limitations to photosynthesis as affected by drought and leaf ontogeny in ash and oak trees. Plant Cell Environ 28:834849 (2005). 34 Bernacchi CJ, Portis AR, Caemmerer HS von and Long SP, Temperature response of mesophyll conductance: implications for the determination of Rubisco enzyme kinetics and for limitations to photosynthesis in vivo. Plant Physiol 130:19921998 (2002). 35 Lauteri M, Pliura A, Monteverdi MC, Brugnoli E, Villani F and Eriksson G, Genetic variation in carbon isotope discrimination in six European

www.soci.org

36 37

38 39 40

41 42 43 44 45 46

populations of Castanea sativa Mill. originating from contrasting localities. J Evol Biol 17:12861296 (2004). Fan S, Grossnickle SC and Russell JH, Morphological and physiological variation in western redcedar (Thuja plicata) populations under contrasting soil water conditions. Trees 22:671683 (2008). Winkel T, Methy M and Thenot F, Radiation use efciency, chlorophyll uorescence, and reectance indices associated with ontogenic changes in water-limited Chenopodium quinoa leaves. Photosynthetica 40:227232 (2002). Baker NR, Chlorophyll uorescence: a probe of photosynthesis in vivo. Annu Rev Plant Biol 59:89113 (2008). Zlatev ZS and Yordanov IT, Effects of soil drought on photosynthesis and chlorophyll uorescence in bean plants. Bul J Plant Physiol 30:318 (2004). Ohnishi N, Allakhverdiev SI, Takahashi S, Higashi S, Watanabe M, Nishiyama Y, et al, Two-step mechanism of photo damage to photosystem II: step 1 occurs at the oxygen-evolving complex and step 2 occurs at the photochemical reaction centre. Biochemistry 44:84948499 (2005). Susiluoto S and Berninger F, Interactions between morphological and physiological drought responses in Eucalyptus microtheca. Silva Fenn 41:221233 (2007). Aspelmeier S and Leuschner C, Genotypic variation in drought response of silver birch (Betula pendula Roth): leaf and root morphology and carbon partitioning. Trees 20:4252 (2006). Foyer CH and Noctor G, Oxygen processing in photosynthesis: regulation and signalling. New Phytol 146:359388 (2000). Reddy AR, Chaitanya KV and Vivekanandan M, Drought-induced responses of photosynthesis and antioxidant metabolism in higher plants. J Plant Physiol 161:11891202 (2004). Munn -Bosch S, The role of -tocopherol in plant stress tolerance. e J Plant Physiol 162:743748 (2005). Munn -Bosch S and Jon F, New insights into the function of e tocopherols in plants. Planta 218:323326 (2004).

2659
wileyonlinelibrary.com/jsfa

J Sci Food Agric 2010; 90: 26492659

c 2010 Society of Chemical Industry

Research Article
Received: 12 November 2009 Revised: 24 June 2010 Accepted: 23 July 2010 Published online in Wiley Online Library: 18 August 2010

(wileyonlinelibrary.com) DOI 10.1002/jsfa.4136

Effects of dietary taurine on egg production, egg quality and cholesterol levels in Japanese quail
Fu-Rong Wang,a,b Xiao-Fang Dong,b Xiao-Ming Zhang,a Jian-Ming Tong,a,b Zhong-Guo Xiea and Qi Zhangb
Abstract
BACKGROUND: Taurine is a semi-essential amino acid and has many biological properties. The objective of this study was to determine the effect of dietary supplementation with taurine on egg production, egg quality, and cholesterol level in serum and egg yolk of quails. A total of 108 quails aged 6 weeks were randomly allocated to three dietary treatments. Each treatment consisted of four replicates of nine quails. The diets were supplemented with 0, 100, and 500 mg kg1 of taurine for 8 weeks. RESULTS: Dietary 500 mg kg1 taurine signicantly affected egg production rate and feed conversion ratio, but had no signicant effects on body weight gain, feed consumption, or egg weight. Dietary taurine had no signicant effect on egg quality parameters studied. Serum triglyceride concentration was reduced signicantly with supplementation of taurine at 100 and 500 mg kg1 . Egg yolk cholesterol content was reduced signicantly, and the contents of serum taurine and egg yolk taurine were increased signicantly with taurine supplementation at 500 mg kg1 . CONCLUSION: Results of the present study indicated that adding 500 mg kg1 taurine reduced yolk cholesterol concentration and increased yolk taurine content without adverse effects on performance and egg quality of laying quails. c 2010 Society of Chemical Industry Keywords: taurine; quail; performance; egg quality; cholesterol

INTRODUCTION
Taurine (2-aminoethanesulfonic acid) is a free, semi-essential, sulfur-containing -amino acid.1 It is not incorporated into proteins and is in fact the most abundant free amino acid in many tissues.2 Taurine has been reported to have many biological properties, including osmoregulation,3 membrane stabilization,4 antioxidant,5 calcium homeostasis,6 modulation of immunity,7 and growth modulation.8 Taurine plays an important role in conjugation of bile acids that are formed from cholesterol in the liver, suggesting that there is a close relationship between taurine and cholesterol metabolism. Taurine also plays an important role in lipid metabolism. Recent studies have shown the hypolipidemic effect of taurine in various species, including rats,9 hamsters,10 mice,11 and rabbits,12 with hypocholesterolemia induced by feeding a high-fat diet. A previous study found that the serum cholesterol concentration was signicantly reduced in quails receiving taurine when compared to a control treatment without taurine supplementation.13 However, the literature concerned with the effects of taurine on egg production, egg quality, egg yolk cholesterol and egg taurine content is limited. The specic objectives of this experiment were to determine the effect of dietary taurine supplementation on performance, egg quality, cholesterol in serum and egg yolk, and taurine content in serum and egg yolk in Japanese quails (Coturnix coturnix japonica).

MATERIALS AND METHODS

Birds, diets, and management A total of 108 quails (Coturnix coturnix japonica) aged 6 weeks were randomly allocated to three dietary treatments. The birds were fed either a basal diet or the basal diet with taurine supplementation at a rate of 100 or 500 mg kg1 feed. Each treatment consisted of four replicates of nine quails. They were housed in cages in a windowed poultry house with a light regime of 16L:8D. Feed and water were provided for ad libitum consumption. The experiment was conducted for 8 weeks. The ingredient and chemical composition of the diets are given in Table 1. The basal diet contained 2810 kcal kg1 metabolizable energy (ME) and 201.3 g kg1 crude protein, and was calculated to

Correspondence to: Jian-Ming Tong, State Key Laboratory of Food Science and Technology, and School of Food Science and Technology, Jiangnan University, Jiangsu Wuxi 214122, China. E-mail: tjm606caas@yahoo.com.cn

a State Key Laboratory of Food Science and Technology, and School of Food Science and Technology, Jiangnan University, Jiangsu Wuxi 214122, China b State Key Laboratory of Animal Nutrition, Institute of Animal Science, Chinese Academy of Agricultural Science, Beijing 100193, Chinasupported by the National Scientic and Technological Supporting Project of Peoples Republic of China (2006BAD12B05), Agricultural S&T Achievement Transfer Fund Program (2007 GB23260400), and Basic Science Research Program

2660

J Sci Food Agric 2010; 90: 26602663

www.soci.org

c 2010 Society of Chemical Industry

Effects of taurine on laying quails

www.soci.org

Table 1. Composition of the basal level(%) diet and its nutrient levels Ingredient Corn Soybean meal Corn protein Soybean oil CaHPO4 Limestone Methionine Vitaminmineral premixa Salt Nutrient composition MEb (kcal kg1 ) Crude protein Calcium Phosphorus Lysine Methionine Composition (g kg1 ) 511.7 32.50 3.50 30.0 14.0 55.0 1.3 10.0 3.0 Level 2810 201.3 25.1 4.6 10.3 4.5

Table 2. The effects of taurine on performance of laying quails Taurine (mg kg1 ) 0 Body weight gain (g) Egg weight (g) Laying rate (%) Daily feed intake (g) Feed conversion rate 29.46 6.99a 10.86 0.26a 82.72 0.93a 23.00 0.68a 2.59 0.12b 100 28.64 4.42a 10.88 0.15a 83.42 1.46a 22.54 0.78a 2.53 0.07b 500 27.44 2.99a 10.80 0.22a 86.19 1.26b 21.89 0.60a 2.36 0.10a

Means in the same row with no common letters differ signicantly (P < 0.05).

a Supplied per kg diet: 3300 IU vitamin A, 900 IU cholecalciferol, 25 IU vitamin E, 1.04 mg vitamin K3 , 5 mg riboavin, 2.04 mg thiamine mononitrate, 7.5 mg D-biotin, 3450 mg choline chloride, 1.03 mg folic acid, 20.10 mg niacin, 16.64 mg calcium pantothenate, 3.06 mg pyridoxine hydrochloride, 0.3 mg vitamin B12 (cyanocobalamin), 142.83 mg ZnSO4 .H2 O, 189.67 mg FeSO4 .H2 O, 13.08 mg CuSO4 , 0.46 mg Na2 SeO3 , 192.53 mg MnSO4 .H2 O, 0.41 mg KI. 2 ME: Metabolizable energy.

meet or slightly exceed the nutrient requirements recommended by the National Research Council (1994).14

Cholesterol analyses Feed was removed 12 h before collecting blood. Blood samples were collected by cardiac puncture from eight randomly selected quails per treatment group (two from each replicate) at the end of the experiment and centrifuged at 1500 g for 10 min. The serum was then collected from the blood and stored at 20 C for determination of serum parameters. The levels of total cholesterol and triglyceride in the serum were determined enzymatically with commercially available reagent kits according to the manufacturers instructions (Bio Sino Bio-technology and Science Inc., Beijing, China). Yolk cholesterol was determined by using a spectrophotometer during the last week of the trial by the method of Rudel and Morris.15

Laying performance Quails were weighed individually at the beginning and at the end of the experiment. Eggs were collected daily and egg production was calculated on a birdday basis. Egg weight was recorded daily for each replicate. Feed consumption was recorded weekly. The feed conversion rate (FCR) was calculated as kilograms of feed per kilogram of egg. Mortality was recorded when it occurred and was expressed as a percentage.

Taurine analyses Plasma serum was deproteinized by 6% sulfosalicylic acid and centrifugation at 3000 g for 15 min. The extraction of taurine from egg yolk was carried out by homogenization in 6% sulfosalicylic acid and centrifugation at 3000 g for 15 min. The supernatants of serum and egg yolk were collected separately, and taurine content was then determined by high-performance liquid chromatography (HPLC) as described by Chen et al.16

Egg quality measurements Eight eggs from each treatment (two eggs from each replicate) were randomly collected on the 2nd, 4th, 6th and 8th week of the experiment to study the egg quality traits: egg weight (g), egg length and width (cm), eggshell strength (kg cm2 ) and thickness (mm), albumen height (mm), width and height (mm) of yolk. Eggs were weighed by electronic scale; the eggshell strength was measured in kilograms per centimeters squared using an eggshell force gauge (model-II, Robotmation Co. Ltd, Tokyo, Japan). Egg albumen height and yolk height were measured using an egg quality gauge, and yolk width was measured by vernier caliper. The eggshell thickness was measured using the vernier caliper at three points: the two narrow ends and in the middle of the egg. Egg shape index was measured by vernier caliper that considers width : length ratio as a percentage. The yolk index was calculated as height of yolk (mm)/average width of yolk (mm). Interior egg quality was measured by Haugh unit. Haugh unit was calculated as 100 log (H 1.7 W0.37 + 7.57), where H is albumen height (mm) and W is egg weight in (g).

Statistical analyses All the data were analyzed by SPSS 13.0 software for Windows (SPSS Inc., Chicago, IL, USA). The differences among treatments in each group were determined by one-way analysis of variance (ANOVA). The signicance of mean differences between groups was tested by Duncans test. The means with P-values 0.05 were considered signicantly different. Data were expressed as mean SD.

RESULTS
Laying performance The effects of taurine supplementation on performance are shown in Table 2. There were no signicant differences (P > 0.05) in body weight gain, egg weight or daily feed intake among the treatments. Furthermore, 100 mg kg1 taurine had no signicant effect on egg production rate or FCR (P > 0.05). However, the addition of 500 mg kg1 taurine in the diet signicantly increased egg production rate and decreased FCR (P < 0.05).

2661

J Sci Food Agric 2010; 90: 26602663

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org

F-R Wang et al.

Table 3. The effects of taurine on egg quality of laying quails Taurine (mg kg1 ) 0 Eggshell thickness (mm) Eggshell strength (kg cm2 ) Egg shape index (%) Egg yolk index (%) Egg Haugh unit 0.185 0.017a 1.02 0.26a 80.48 3.48a 41.16 2.67a 78.28 6.97a 100 0.193 0.016a 1.09 0.16a 80.16 3.46a 41.92 3.43a 78.76 4.07a 500 0.193 0.013a 1.18 0.22a 80.25 2.48a 41.49 2.16a 78.23 4.77a

Means in the same row with no common letters differ signicantly (P < 0.05).

Table 4. The effects of taurine on serum biochemical, egg yolk cholesterol and taurine concentrations of laying quails Taurine (mg kg1 ) 0 Blood serum cholesterol (mg dL1 ) Blood serum triglyceride (mmol L1 ) Blood serum taurine (g 100 mL1 ) Egg yolk cholesterol (g kg1 yolk) Egg yolk taurine (mg kg1 ) 144.51 10.58a 15.29 0.45b 38.32 4.33a 17.94 0.44b 44.68 3.58a 100 142.81 10.61a 14.81 1.51a 42.67 7.94ab 17.28 0.66ab 48.57 5.04ab 500 140.41 22.58a 13.15 0.72a 49.12 4.07b 16.81 0.97a 50.91 2.39b

Means in the same row with no common letters differ signicantly (P < 0.05).

Egg quality The effect of dietary supplementation of taurine on egg quality is presented in Table 3. There were no signicant differences in the eggshell thickness, egg shape index, eggshell strength, egg yolk index or egg Haugh unit among the three groups (P > 0.05). Serum cholesterol and triglyceride concentration and egg cholesterol There was no signicant reduction in serum cholesterol concentration when laying quails were fed the diets with 100 or 500 mg kg1 taurine (Table 4). The serum triglyceride concentration decreased signicantly due to supplementation of taurine at 100 and 500 mg kg1 . The addition of 500 mg kg1 taurine signicantly decreased yolk cholesterol content (P < 0.05). Serum and egg yolk taurine The serum concentration of taurine and egg yolk taurine content from the quails fed 500 mg kg1 taurine were signicantly higher than those of the control group (Table 4).

DISCUSSION
The results indicated that the supplementation of taurine at 500 mg kg1 improved the egg production and feed conversion ratio in laying quails, while egg weight and body weight gain were not inuenced. Yamazaki and Takemasa17 found that dietary 2.5 g kg1 taurine in laying hens enhanced egg production and feed conversion but the egg weight decreased. They also observed that egg weight was decreased without affecting egg production, feed conversion, or body weight by dietary 5 g kg1 taurine in laying hens. No information is available with regard to the effect of dietary taurine on serum triglyceride concentration of the

laying quails. Our present study demonstrated that taurine supplementation decreased serum triglyceride concentration. This result is consistent with similar ndings in other animals and humans.18 20 Though the addition of taurine to the diet did not reduce serum level of cholesterol in normal rats,21,22 many studies showed that it can diminish the degree of increase in the serum cholesterol concentration induced by feeding a diet containing a large amount of cholesterol.9 13 Our present study demonstrated that the taurine supplementation at 500 mg kg1 level decreased the cholesterol level signicantly in the egg yolk. The results of these studies tend to vary depending on the species of animal used, and the composition of taurine supplementation and of the experimental diet. Taurine is the most abundant free amino acid in several tissues and in the cellular components of blood.23 Dietary taurine supplementation at 500 mg kg1 increased serum taurine level signicantly. Yokogoshi et al.22 found that the serum taurine concentration in rat was decreased by the intake of the highcholesterol diet, but the decrease was gradually restored by the administration of taurine in a dose-dependent manner. Similarly, serum taurine concentration in cats increased with increased dietary supplementation of taurine.24 These results imply that dietary taurine could increase serum taurine in a dose-dependent manner. Moreover, earlier observations in rats,25,26 cats27 and humans28 obtained similar results. Taurine mostly existed only in the yolk, but not in the albumen. The present study showed that the egg yolk taurine level was signicantly increased with the dietary taurine supplementation at 500 mg kg1 . Hu et al.29 found that adding taurine to the diet similarly increased milk taurine content. Taurine has many biological functions; taurine supplements have been shown to be benecial for infants and some groups of elderly subjects, and it is especially essential to the fetus and newborn for their

2662

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 26602663

Effects of taurine on laying quails development.24,30 Thus the results from the present study, along with ndings from the previous studies mentioned above, suggest that increased levels of taurine in egg yolk through dietary supplementation may aid in improving the nutritional value of eggs.

www.soci.org
11 Murakami S, Kondo-Ohta Y and Tomisawa K, Improvement in cholesterol metabolism in mice given chronic treatment of taurine and fed a high-fat diet. Life Sci 64:8391 (1999). 12 Petty MA, Kintz J and DiFrancesco GF, The effects of taurine on atherosclerosis development in cholesterol-fed rabbits. Eur J Pharmacol 180:119127 (1990). 13 Jackson JA and Burns MJ, Effects of cystine, niacin and taurine on cholesterol concentration in the Japanese quail with comments on bile acid metabolism. Comp Biochem Physiol 48:6168 (1974). 14 NRC, Nutrient Requirements of Poultry (9th rev. edn). National Academies Press, Washington, DC (1994). 15 Rudel LL and Morris MD, Determination of cholesterol using o-phthaladehyde. J Lipid Res 14:364366 (1973). 16 Chen ZL, Xu G, Specht K, Yang RJ and She SW, Determination of taurine in biological samples by reversed-phase liquid chromatography with precolumn derivatization with dinitrouorobenzene. Anal Chim Acta 296:249253 (1994). 17 Yamazaki M and Takemasa M, Effects of dietary taurine on egg weight. Poult Sci 77:10241026 (1998). 18 Nakaya Y, Minami A, Harada N, Sakamoto S, Niwa Y and Ohnaka M, Taurine improves insulin sensitivity in the Otsuka Long-Evans Tokushima Fatty rat, a model of spontaneous type 2 diabetes. Am J Clin Nutr 71:5458 (2000). 19 Balkan J, Kanbagli O, Hatipoglu A, Kucuk M, Cevikbas U, AykacToker G, et al, Improving effect of dietary taurine supplementation on the oxidative stress and lipid levels in the plasma, liver and aorta of rabbits fed on a high-cholesterol diet. Biosci Biotechnol Biochem 66:17551758 (2002). 20 Zhang M, Bi LF, Fang JH, Su XL, Da GL, Kuwamori T, et al, Benecial effects of taurine on serum lipids in overweight or obese nondiabetic subjects. Amino Acids 26:267271 (2004). 21 Mochizuki H, Oda H and Yokogoshi H, Increasing effect of dietary taurine on the serum HDL-cholesterol concentration in rats. Biosci Biotechnol Biochem 62:578579 (1998). 22 Yokogoshi H, Mochizuki H, Nanami K, Hida Y, Miyachi F and Oda H, Dietary taurine enhances cholesterol degradation and reduces serum and liver cholesterol concentrations in rats fed a highcholesterol diet. J Nutr 129:17051711 (1999). 23 Schuller-Levis G and Park E, Taurine: new implications for an old amino acid. Fems Microbiol Lett 226:195202 (2003). 24 Sturman JA, Taurine in development. Physiol Rev 73:119147 (1993). 25 Lombardini JB and Medina EV, Effects of dietary inorganic sulfate, taurine, and methionine on tissue levels of taurine in the growing rat. J Nutr 108:428433 (1978). 26 Dawson R, Liu S, Eppler B and Patterson T, Effects of dietary taurine supplementation or deprivation in aged male Fischer 344 rats. Mech Ageing Dev 107:7391 (1999). 27 Laidlaw SA, Sturman JA and Kopple JD, Effect of dietary taurine on plasma and blood cell taurine concentrations in cats. J Nutr 117:19451949 (1987). 28 Vinton NE, Laidlaw SA, Ament ME and Kopple JD, Taurine concentrations in plasma and blood cells of patients undergoing long-term parenteral nutrition. Am J Clin Nutr 44:398404 (1986). 29 Hu JM, Rho JY, Suzuki M, Nishihara M and Takahashi M, Effect of taurine in rat milk on the growth of offspring. J Vet Med Sci 62:693698 (2000). 30 Cho KH, Kim ES, Chen JD, Zhang S, Kim H, Kim SY, et al, Serum and urine taurine levels in elderly patients undergoing long-term enteral nutrition are reduced overtime. Nutr Res 22:10171025 (2002).

CONCLUSIONS
On the basis of the above experiment, it is concluded that dietary supplementation of taurine at 500 mg kg1 in laying quails diets could enhance egg production, improve the egg yolk taurine content, and reduce the egg yolk cholesterol concentration and serum triglyceride level without adversely affecting on egg quality.

ACKNOWLEDGEMENTS
This research was supported, in part, by the National Scientic and Technological Supporting Project of the Peoples Republic of China (2006BAD12B05), Agricultural Science and Technology Achievements Transformation Fund Programs (2007 GB23260400), and the Basic Research Operating Expenses of the Central-level Non-prot Research Institutes (ywf-td-4). The authors thank all the people who offered help in this study.

REFERENCES
1 Gaull GE, Taurine in human milk: growth modulator or conditionally essential amino acid. J Pediatr Gastroenterol Nutr 21:52665271 (1983). 2 Huxtable RJ, Physiological actions of taurine. Physiol Rev 72:101163 (1992). 3 Pasantes-Morales H and Schousboe A, Role of taurine in osmoregulation in brain cells: mechanisms and functional implications. Amino Acids 12:281292 (1997). 4 Pasantes-Morales H, Wright CE and Gaull GE, Taurine protection of lymphoblastoid cells from iron-ascorbate induced damage. Biochem Pharmacol 34:22052207 (1985). 5 Gurer H, Ozgunes H, Saygin E and Ercal N, Antioxidant effect of taurine against lead-induced oxidative stress. Arch Environ Contam Toxicol 41:397402 (2001). 6 Foos TM and Wu JY, The role of taurine in the central nervous system and the modulation of intracellular calcium homeostasis. Neurochem Res 27:2126 (2002). 7 Redmond HP, Stapleton PP, Neary P and Bouchier-Hayes D, Immunonutrition: the role of taurine. Nutrition 14:599604 (1998). 8 Hayes KC, Stephen ZF and Sturman JA, Growth depression in taurinedepleted infant monkeys. J Nutr 110:20582064 (1980). 9 Park T and Lee K, Dietary taurine supplementation reduces plasma and liver cholesterol and triglyceride levels in rats fed a high-cholesterol or a cholesterol-free diet. Adv Exp Med Biol 442:319325 (1998). 10 Murakami S, Kondo Y, Toda Y, Kitajima H, Kameo K, Sakono M, et al, Effect of taurine on cholesterol metabolism in hamsters: upregulation of low density lipoprotein (LDL) receptor by taurine. Life Sci 70:23552366 (2002).

2663
wileyonlinelibrary.com/jsfa

J Sci Food Agric 2010; 90: 26602663

c 2010 Society of Chemical Industry

Research Article
Received: 13 October 2009 Revised: 24 June 2010 Accepted: 23 July 2010 Published online in Wiley Online Library: 24 August 2010

(wileyonlinelibrary.com) DOI 10.1002/jsfa.4137

Postmortem degradation of desmin and calpain in breast and leg and thigh muscles from Taiwan black-feathered country chickens
Ya-Shiou Chang and Rong-Ghi R. Chou
Abstract
BACKGROUND: Several studies have reported that the postmortem changes are more rapid in breast muscles (BM) than in leg and thigh muscles (LM) of chickens. However, the reasons for the differences in postmortem proteolysis of BM and LM are still uncertain. The purpose of this study was therefore to compare the postmortem degradation of desmin and calpains in BM and LM from Taiwan black-feathered country chickens at 5 C. RESULTS: The pH was lower (P < 0.05) in BM than in LM. Western blot indicated that postmortem desmin degradation was more rapid in BM than in LM. Casein zymograms showed that at-death -calpain activity was higher in BM than in LM. As postmortem time proceeded, -calpain was activated and autolyzed more extensively in BM than in LM. However, the /m-calpain activity remained stable during postmortem storage in both BM and LM. CONCLUSION: Our results suggest that the more rapid postmortem proteolysis found in BM than in LM at 5 C similar with the previous study could be mainly explained by both greater amounts and faster activation and autolysis of -calpain in BM. c 2010 Society of Chemical Industry Keywords: chicken muscle; postmortem changes; calpain; desmin

INTRODUCTION
It is generally accepted that postmortem degradation of cytoskeletal proteins such as desmin by calcium-dependent proteases (calpains) may improve meat tenderness.1 3 There are two ubiquitous calpains (- and m-calpain) and one tissue-specic calpain, p94, that are present in mammalian muscle cells.4 Instead of m-calpain, /m-calpain, which contains sequence homology and Ca2+ sensitivity between - and m-calpain, is present in chicken muscle.5 Free calcium concentration required for chicken - and /m-calpain activation is lower than for mammalian - and mcalpain.5 Among those calpains, -calpain is thought to play an essential role in the postmortem proteolysis of bovine muscle at 5 C,6 although m-calpain may also be involved.7,8 Several studies9,10 have indicated that postmortem proteolysis is more rapid in avian breast muscles (BM) than in leg and thigh muscles (LM). Early studies9,11 reported that the difference in ber type composition between BM (white) and LM (red) may explain the difference in the rate of postmortem proteolysis in these two muscle types. However, studies show that because red muscles (m. vastus intermedius and m. soleus) exhibit the same rate of desmin and troponin-T postmortem degradation as white muscle (m. semitendinosus) in pigs, the difference in rate of postmortem proteolysis cannot be solely explained by the difference in ber type composition.12 Therefore, more studies are needed to examine the differences in postmortem proteolysis of BM and LM. Taiwan black-feathered country chickens were very popular in local markets, and its production value in 2008 was close to $300 million in US currency.13 Little information, however, was available regarding postmortem changes in the muscles of this

strain. Additionally, because previous studies10 found that desmin degraded very rapidly in 1-day postmortem chicken BM, it would be interesting to examine the relationship between calpain autolysis and desmin degradation within 24 h post mortem. Thus the purpose of this study was to compare the degradation of desmin and calpains within 24 h post mortem in BM and LM from Taiwan black-feathered country chickens at 5 C.

MATERIALS AND METHODS

Sample preparation Taiwan black-feathered country chickens (female, 100 days old with an average live weight of 1.5 kg) were slaughtered in a local abattoir by using standard commercial practices. The carcasses (3040 min post mortem) were vacuum-packed and stored at 5 C. Breast and leg and thigh muscles were sampled at 3, 6, 12 and 24 h of storage. The 0 h samples were taken at the time of killing (within 510 min post mortem). Twenty chickens were randomly assigned to each of the ve sampling times. Four chickens were sampled at each time in each of three replications. Muscle samples (7080 g) from the four chickens at each sampling time were combined and ground through a 3 mm plate. The ground samples were

Correspondence to: Rong-Ghi R. Chou, Department of Animal Science, National Chiayi University, 300 University Road, Chiayi City, 60083 Taiwan. E-mail: chourg@mail.ncyu.edu.tw Department of Animal Science, National Chiayi University, Chiayi City, Taiwan

2664

J Sci Food Agric 2010; 90: 26642668

www.soci.org

c 2010 Society of Chemical Industry

Postmortem changes in breast and leg muscles in chickens immediately frozen in liquid nitrogen and stored at 80 C before pH measurement, western blot analysis, and casein zymography. The pH was determined by the method of Farouk and Swan.14

www.soci.org Image analysis Two to three representative gels or blots from each replication were used for image analysis. The bands in blots and casein gels incubated with 4 mmol L1 Ca2+ to activate all calpain isoforms were digitized using a scanner (model J131B, Epson Taiwan Technology and Trading LTD, Taiwan) using Photoshop software. The resulting signals were quantied using Image Gauge (version 3.46, Science Lab 99 for Windows, Fuji Film, Tokyo, Japan). Statistical analysis Split-plot design was used in this study. Whole unit was the type of muscle, and subunit was the muscle samples taken at each sampling time. This experiment was done with three replications, and triplicate samples were taken at each sampling time in each of three replications. All data were analyzed by the procedure of the SAS-GLM program (SAS Institute Inc., Cary NC, USA), and mean separation was determined using the least squares means procedure.22

Western blot analysis Breast and leg and thigh myobrils from the pooled samples were isolated via the method of Huff-Lonergan et al.15 The protein concentration of myobrils was determined by using a modied biuret method.16 Myobril samples for sodium dodecyl sulfatepolyacrylamide gel electrophoresis (SDSPAGE) were prepared by the method of Wang et al.17 The SDS-PAGE was done in a 120 g kg1 tris-glycine slab gel (acrylamide : methylenebisacrylamide = 37.5 : 1, w/w) by the method of Laemmli.18 The same amount of protein (150 g) from each sample was loaded on each well of the gels. Proteins were transferred from a 120 g kg1 slab gel to a nitrocellulose membrane by the method of Towbin et al.19 After the transfer, the membrane was incubated in a 50 g kg1 bovine serum albuminphosphate buffer solution (BSA-PBS) for 30 min at 37 C and then washed three times (5 min each) in a 1 g kg1 BSA-PBS solution at room temperature. A desmin monoclonal antibody (clone DE-U-10; D-1033, Sigma-Aldrich Co., St Louis, MO, USA) was used as a primary antibody. The membrane was incubated with the primary antibody for 2 h at room temperature, washed three times (5 min each) in 1 g kg1 BSA-PBS, incubated with a secondary antibody goat anti-mouse-HRP (A-9044, Sigma-Aldrich) for 2 h at room temperature, and washed twice (5 min each) in 1 g kg1 BSAPBS solution and twice (1 min each) in deionized water. The color was developed by SIGMAFAST 3,3 -diaminobenzidine tablets (D-4418, Sigma-Aldrich).

RESULTS AND DISCUSSION

The average at-death pH was 6.11 0.05 and 6.45 0.08 in breast (BM) and leg and thigh (LM) muscles, respectively, and the pH decreased signicantly (P < 0.05) to 5.70 0.04 and 6.28 0.06, respectively, by 6 h post mortem (Fig. 1). However, the pH changed insignicantly (P > 0.05) from 6 to 24 h post mortem in the BM (5.68 0.07) and LM (6.27 0.06) samples (Fig. 1). These results were consistent with the ndings of Lee et al.,23 who showed that the ultimate pH in chicken muscle was reached by 6 h post mortem. Figure 1 also shows that the pH was signicantly (P < 0.05) lower in the BM than in the LM samples after 24 h postmortem storage at 5 C, consistent with our previous report.24 This difference might be attributed to the higher content25 and to the faster utilization23 of muscle glycogen in chicken breast muscle. Desmin, an intermediate lament protein, is a calpain-sensitive protein.26 Western blots labeled with a desmin monoclonal antibody indicate that BM desmin (Fig. 2) decreased more rapidly than LM desmin (Fig. 2), conrming previous studies.24 In BM samples, most of the desmin remained intact in 3 h postmortem samples, which contained 85% of the desmin present in BM sampled at death. However, the most extensive desmin degradation (P < 0.05) was observed in 6 h postmortem samples, which lost 75% of the at-death desmin. Between 6 and 12 h post mortem, there was a small decrease (<5%) in BM desmin and little additional desmin decrease was seen in 24 h BM samples. On the other hand, 87% of the at-death LM desmin was left in 3 h postmortem samples and remained after 624 h post mortem. Casein gel zymography is a sensitive method to detect calpain proteases.20 In agreement with previous studies on chicken muscle,27 casein zymograms showed that the bands of - and /m-calpain in the at-death BM and LM samples began to appear faintly in the presence of 10 (Fig. 3(A)) and 30 mol L1 Ca2+ (Fig. 3(B)), respectively. The bands on casein zymograms became more apparent as Ca2+ concentration increased. However, no extra bands were found in the presence of 100 mol L1 (Fig. 3(C)) or 4 mmol L1 Ca2+ (Fig. 3(D)) in both BM and LM at-death samples. Results also indicated that -calpain, instead of /mcalpain, migrated as a doublet on gels, similar to the studies of Veiseth et al.20 in postmortem ovine muscle, which showed that native -calpain (lower band) and its autolyzed form (upper band) migrated very closely on casein gels as seen on ours. Therefore it

Casein zymography The procedure used for protein extraction was based on the method of Veiseth et al.20 Briey, a 5 g sample was homogenized in three volumes of extraction buffer (100 mmol L1 Tris base, 10 mmol L1 ethylenediaminetetraacetic acid (EDTA), 0.5 g kg1 2-mercaptoethanol, pH 8.3) at 5 C. Homogenates were centrifuged at 22 000 g for 25 min at 5 C. The protein concentration of the supernatant was determined by using a modied biuret method.16 Zymograms were routinely run in 120 g kg1 gels (acrylamide : methylenebisacrylamide = 37.5 : 1, w/w) containing 2.1 g kg1 casein (w/v) by the method of Raser et al.21 The sample buffer (150 mmol L1 Tris-HCl, pH 6.8, 200 g kg1 glycerol, 7.5 g kg1 2-mercaptoethanol (MCE), 0.2 g kg1 (w/v) bromophenol blue) was added to the protein extract at a ratio of two parts of the buffer to three parts of protein extract (v/v). The same amount of protein (250 g) from each sample was loaded on to each well of the casein gels. The casein minigels (0.75 mm, Bio-Rad Laboratories, Hercules, CA, USA) were pre-run at 100 V for 15 min (5 C), with a running buffer containing 25 mmol L1 Tris-HCl, 0.5 g kg1 MCE, 192 mmol L1 glycine, and 1 mmol L1 EDTA (pH 8.3) before samples were loaded on to the wells. The gels were run at 100 V for 2 h (5 C) and then incubated at room temperature in a 50 mmol L1 Tris-HCl (pH 7.5) buffer containing 0.5 g kg1 MCE and CaCl2 (0.01, 0.03, 0.1 or 4 mmol L1 ) with slow shaking for 1 h (three changes of the buffer). This was followed by a 16 h incubation in the same buffer at 37 C. The gels were then stained for 2 h with Coomassie blue (R-250) and destained with 200 g kg1 methanol and 70 g kg1 acetic acid.

2665

J Sci Food Agric 2010; 90: 26642668

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org
BM LM

Y-S Chang, R-GR Chou

150 125 Relative activity, % 100 75 b 50 25 0 -calpain /m-calpain a a a

7.0 6.5 pH 6.0 5.5 5.0

9 12 15 18 Time post mortem, h

21

24

Figure 1. Changes in pH of breast (BM) and leg (LM) muscles from Taiwan black-feathered country chickens during postmortem storage at 5 C. , BM samples; , LM samples.

Figure 4. Activity of - and /m-calpain in at-death BM and LM samples from Taiwan black-feathered country chickens. Activities measured in the at-death BM samples are taken as 100%. Vertical bars show the standard deviation of the means. Different letters within each calpain isoform indicate signicant differences. , BM; , LM.

Figure 2. Western blotting showing changes in desmin of BM (A) and LM (B) samples from the Taiwan black-feathered country chickens during postmortem storage at 5 C. These blots are representative of three replications of combined samples. D, Desmin; lane 1, at-death; lane 2, 3 h; lane 3, 6 h; lane 4, 12 h; lane 5, 24 h.

Figure 3. Zymograms showing - and /m-calpains in the presence of 10 mol L1 (A), 30 mol L1 (B), 100 mol L1 (C) and 4 mmol L1 (D) calcium in at-death BM and LM samples from Taiwan black-feathered country chickens. These gels are representative of three replications of combined samples. Lane 1, BM; lane 2, LM. - = -calpain; /m- = /mcalpain.

was reasonable to predict that the doublet found in BM samples might consist of -calpain and its autolyzed form. Because Ji and Takahashi28 reported that the free calcium concentration in chicken muscle is 70 mol L1 20 min post mortem, it is possible that -calpain in the at-death samples, which were collected within 510 min post mortem, had been activated and autolyzed. Image analysis (Fig. 4) results show that the at-death activity of -calpain was approximately two times higher (P < 0.05) in BM than in LM samples. The at-death activity of /m-calpain

was approximately 10% lower in BM than in LM samples (Fig. 3), although the difference was insignicant (P > 0.05). Figure 5 shows casein zymograms of - and /m-calpain activity in BM and LM samples at different times of postmortem storage at 5 C. The results indicated that -calpain activity decreased as postmortem time proceeded (Fig. 5). Figure 6 shows that the greatest reduction (P < 0.05) of -calpain activity was found in the 6 h postmortem BM samples, which lost 60% of the at-death BM -calpain activity. This implies that -calpain autolysis was very extensive by 6 h post mortem and explains why BM desmin degraded extensively simultaneously (Fig. 2). By 6 h post mortem, the level of -calpain activity was similar in BM and LM samples because the reduction of -calpain activity was much slower in LM, if the at-death BM -calpain activity (approximately two times higher than the at-death LM samples) was taken as 100% (Fig. 6). After 24 h post mortem, -calpain activity in both BM and LM samples further decreased to 20% of the at-death BM activity (Fig. 6). Recent studies28 in postmortem bovine muscle showed that the 80 kDa subunit of -calpain was autolyzed rst to a 78 kDa intermediate product, from which NH2 -terminal 14 amino acids of the 80 kDa subunit were removed,4,29 and completely autolyzed to produce a 76 kDa form, from which an additional NH2 -terminal 12 amino acids were removed.4,29 However, the 28 kDa small subunit of -calpain was not autolyzed in postmortem bovine muscle.28 Camou et al.29 concluded that the 76/28 kDa -calpain was proteolytically inactive and that this accounted for the loss of -calpain activity during postmortem storage. It has been reported that postmortem proteolysis is more rapid in BM than in LM.9 11 As shown in Fig. 1, the LM samples had a higher ultimate pH (6.3) than the BM samples (5.7). It was also reported that little calpain activity could be found at pH 5.8 in bovine muscle.7 Presumably -calpain, which is essential for postmortem proteolysis at 5 C,6,30 was more active in the LM samples than in the BM samples and might favor more extensive postmortem desmin degradation. However, results (Fig. 4) not only showed that BM contained approximately two times greater calpain activity than LM in the at-death samples, but also indicated that most of the -calpain (60% of the at-death activity) was autolyzed in BM samples (Fig. 6) before the pH reached 5.7 by 6 h postmortem (Fig. 1). It was recently proposed that calpain activation might explain the variation in postmortem desmin degradation in pig muscle.3 Therefore the differences in the extent

2666

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 26642668

Postmortem changes in breast and leg muscles in chickens

www.soci.org

Figure 5. Zymograms showing postmortem changes in - and /m-calpains in the presence of 4 mmol L1 calcium in BM (A) and LM (B) samples from Taiwan black-feathered country chickens at 5 C. These gels are representative of three replications of combined samples. , -calpain; /m, /m-calpain. Lane 1, at-death; lane 2, 3 h; lane 3, 6 h; lane 4, 12 h; lane 5, 24 h.

120 Relative activity, % 100 80 60 40 20 0 0 3 6 9 12 15 18 21 24

Postmortem time, h
Figure 6. Postmortem changes in -calpain activity of BM and LM samples from Taiwan black-feathered country chickens at 5 C. Activity is expressed as a percentage of the at-death BM -calpain activity, which is taken as 100%. , BM; , LM.

200 Relative activity, % 150 100 50 0

that little BM or LM desmin degraded after 6 h post mortem. This implies that /m-calpain may play a relatively minor role in desmin degradation after 6 h post mortem. Therefore more research would be needed to understand the precise role of /m-calpain in the postmortem proteolysis of chicken muscle. In summary, we compared postmortem degradation of desmin and calpain in BM and LM from Taiwan black-feathered country chicken at 5 C. The results showed that the pH was lower (P < 0.05) in BM than in LM at death and at 3, 6, 12 and 24 h post mortem. Western blot indicated that postmortem desmin degradation was more rapid in BM than in LM. Casein zymograms showed that at-death -calpain activity was higher in BM than in LM. As postmortem time proceeded, -calpain was activated and autolyzed more extensively in BM than in LM. However, the /m-calpain activity remained stable during postmortem storage in both BM and LM. Therefore our results suggest that the more rapid postmortem proteolysis found in BM than in LM at 5 C could be mainly explained by both greater amounts and faster activation and autolysis of -calpain in BM.

REFERENCES
1 Hopkins DL and Thompson JM, Factors contributing to proteolysis and disruption of myobrillar proteins and the impact on tenderization in beef and sheep meat. Aust J Agric Res 53:149166 (2002). 2 Robson RM, Huff-Lonergan E, Parrish FC Jr, Ho C-Y, Stromer MH, Huiatt TW, et al, Postmortem changes in the myobrillar and other cytoskeletal proteins in muscle. Proc Recip Meat Confer 50:4352 (1997). 3 Huff-Lonergan E and Lonergan SM, Mechanism of water-holding capacity of meat: the role of postmortem biochemical and structural changes. Meat Sci 71:194204 (2005). 4 Goll DE, Thompson VF, Li H, Wei W and Cong J, Calpain system. Physiol Rev 83:731801 (2003). 5 Sorimachi H, Tsukahara T, Okada-Ban M, Sugita H, Ishiura S and Suzuki K, Identication of a third ubiquitous calpain species: chicken muscle expresses four distinct calpains. Biochim BIophys Acta 1261:381393 (1995). 6 Geesink GH, Kuchay S, Chishti AH and Koohmaraie M, -Calpain is essential for postmortem proteolysis of muscle proteins. J Anim Sci 84:28342840 (2006). 7 Camou JP, Marchello JA, Thompson VF, Mares SW and Goll DE, Effect of postmortem storage on activity of - and m-calpain in ve bovine muscles. J Anim Sci 85:26702681 (2007). 8 Pomponio L, Lametsch R, Karlsson AH, Costa LN, Grossi A and Ertbjerg P, Evidence for post-mortem m-calpain autolysis in porcine muscle. Meat Sci 80:761764 (2008). 9 Paxhia JM and Parrish Jr FC, Effect of postmortem storage on titin and nebulin in pork and poultry light and dark muscles. J Food Sci 53:15991601 (1988). 10 Chou R-GR, Tseng T-F, Lin K-J and Yang J-H, Post-mortem changes in myobrillar proteins of breast and leg muscles from broilers, spent hens and Taiwanese country chickens. J Sci Food Agric 65:297302 (1994). 11 Hay JD, Currie RW and Wolf FH, Effect of postmortem aging on chicken muscle brils. J Food Sci 38:981987 (1973).

12

15

18

21

24

Postmortem time, h
Figure 7. Postmortem changes in /m-calpain activity of BM and LM samples from Taiwan black-feathered country chickens at 5 C. Activity is expressed as percentage of the at-death BM /m-calpain activity, which is taken as 100%. , BM; , LM.

of postmortem proteolysis between BM and LM probably depend more on -calpain level and the rate of its activation and autolysis than on the difference in nal pH between the two muscle types. On the other hand, /m-calpain activity was very stable during postmortem storage in both BM and LM samples (Figs 5 and 7). The bands below the /m-calpain band in the 6, 12 and 24 h postmortem BM and LM samples were visible in the presence of 30 mol L1 Ca2+ (results not shown) and clearly seen in the presence of 4 mmol L1 Ca2+ (Fig. 5). It was suggested that these bands might be generated from /m-calpain23 because the free calcium concentration in 6 h postmortem chicken muscle exceeded 120 mol L1 ,28 which is enough for /m-calpain activation and autolysis. Additionally, these bands were more apparent in the LM than in BM samples (Fig. 5) because the higher ultimate pH would favor continued proteolytic activity (Fig. 1). Results indicated that the /m-calpain in BM and LM was active after 6 h post mortem. However, our western blot (Fig. 2) showed

2667

J Sci Food Agric 2010; 90: 26642668

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org
12 Christensen M, Henckel P and Purslow PP, Effect of muscle type on the rate of post-mortem proteolysis in pigs. Meat Sci 66:595601 (2004). 13 Council of Agriculture, Executive Yuan, ROC. Agriculture Statistics Yearbook (2008). [Online]. Available: http://www.coa.gov.tw/htmlarea le/web articles/coa/11499/ 010.xls [26 April 2010]. 14 Farouk MM and Swan JE, Acceptability and functional properties of restructured roast from frozen pre-rigor injected beef. Meat Sci 46:5766 (1997). 15 Huff-Lonergan E, Parrish FC Jr and Robson RM, Effects of postmortem aging time, animal age, and sex on degradation of titin and nebulin in bovine longissimus muscle. J Anim Sci 73:10641073 (1995). 16 Robson RM Goll DE and Temple MJ, Determination of protein in Tris buffer by the biuret reaction. Anal Biochem 24:339341 (1968). 17 Wang S-M, Greaser ML, Schultz E, Bulinski JC, Lin JJ-C and Lessard JL, Studies on cardiac myobrillogenesis with antibodies to titin, actin, tropomyosin, and myosin. J Cell Biol 107:10751083 (1988). 18 Laemmli UK, Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227:680685 (1970). 19 Towbin H, Staehelin T and Gordon J, Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheet: procedure and some application. Proc Natl Acad Sci USA 76:43504354 (1979). 20 Veiseth E, Shackelford SD, Wheeler TL and Koohmaraie M, Effect of postmortem storage on -calpain and m-calpain in ovine skeletal muscle. J Anim Sci 79:15021508 (2001). 21 Raser KJ, Posner A and Wang KW, Casein zymography: a method to study -calpain, m-calpain and their inhibitory agents. ArchBiochem Biophys 319:211216 (1995).

Y-S Chang, R-GR Chou

22 SAS Institute Inc., User Guide: Statistics, version 8.01. SAS Institute, Cary, NC (1986). 23 Lee HL, Sante-Lhoutellier V, Vigouroux S, Briand Y and Briand M, Role of calpains of postmortem proteolysis in chicken muscle. Poult Sci 87:21262132 (2008). 24 Tsai S-F, Lin C-Y, Lu J-J and Chou R-GR, Postmortem proteolysis of breast and leg muscles from Taiwan colored chickens and silkie bantam. Asia Aus J Anim Sci 19:739743 (2006). 25 Schreurs FJG, Post-mortem changes in chicken muscle. Worlds Poult Sci J 56:319346 (2000). 26 OShea JM, Robson RM, Hartzer MK, Huiatt TW, Rathbun WE and Stromer MH, Purication of desmin from adult mammalian skeletal muscle. Biochem J 195:345356 (1981). 27 Lee HL, Sante-Lhoutellier V, Vigouroux S, Briand Y and Briand M, Calpain specicity and expression in chicken tissue. Comp Biochem Physiol B 146:8893 (2007). 28 Ji JR and Takahashi K, Changes in concentration of sarcoplasmic free calcium during post mortem ageing of meat. Meat Sci 73:395403 (2006). 29 Camou JP, Marchello JA, Thompson VF, Mares SW and Goll DE, Effect of postmortem storage on activity of - and m-calpain in ve bovine muscles. J Anim Sci 85:26702681 (2007). 30 Zimmerman UJP and Schlaepfer WW, Two stage autolysis of the catalytic subunit initiates activation of calpain. Biochim Biophys Acta 1078:192198 (1991).

2668
wileyonlinelibrary.com/jsfa
c 2010 Society of Chemical Industry J Sci Food Agric 2010; 90: 26642668

Research Article
Received: 9 October 2009 Revised: 27 June 2010 Accepted: 23 July 2010 Published online in Wiley Online Library: 25 August 2010

(wileyonlinelibrary.com) DOI 10.1002/jsfa.4138

Biogenic amine formation in Turkish dry fermented sausage (sucuk) as affected by nisin and nitrite
Sukru Kurta and Omer Zorbab
Abstract
BACKGROUND: The effects of nitrite (0, 100, and 200 mg kg1 ) and nisin (0, 250, and 500 mg kg1 ) on biogenic amine formation in sucuk were investigated by utilising a central composite design of response surface methodology. RESULTS: The addition of nitrite led to decreased levels of tryptamine, 2-phenylethylamine, putrescine, cadaverine, tyramine, and histamine, whereas nisin decreased the tryptamine level and counts of lactic acid bacteria. However, nisin increased putrescine, cadaverine, and spermidine levels. Their interactive effect was also found to be signicant (P < 0.05) for putrescine values. CONCLUSION: The additional nitrite levels can be decreased by the addition of nisin, which will hinder biogenic amine formation. c 2010 Society of Chemical Industry Keywords: biogenic amines; dry fermented sausage; nitrite; nisin; sucuk

INTRODUCTION
Sucuk is a dry fermented sausage produced from beef or sheep meat. A variety of microorganisms are involved in sucuk fermentation. The production conditions of this meat product may result in the formation of biogenic amines, such as putrescine, cadaverine, tyramine, and histamine. Biogenic amines are organic bases with aliphatic, aromatic, or heterocyclic structures that can be found in several food products. They are mainly generated via decarboxylation of the corresponding precursors (i.e. amino acids), through substrate-specic microbial enzymes. Microbial growth, acidication, and proteolysis provide favourable conditions for the formation of biogenic amines during meat fermentation. Some of these microbial activities are of concern in relation to human health due to the potential for toxicological effects.1 Some biogenic amines, such as putrescine and cadaverine, form carcinogenic nitrosamines by reacting with nitrite. Nitrite is an indispensable additive that plays an important role in the generation of high-quality sucuk. However, nitrite may also cause haemoglobinaemia.2 In addition to the toxicological effect of biogenic amines, these metabolic products are also of concern in relation to food hygiene. High amounts of certain amines may be found in food as a consequence of the using poor-quality raw materials, contamination, and inappropriate conditions during food processing and storage.3,4 Therefore, biogenic amines and nitrite levels are factors that need to be considered in the production of dry fermented sausages. Decreasing nitrite levels and preventing the accumulation of biogenic amines are critical in the production of dry fermented sausages.3 To prevent biogenic amine accumulation, raw materials must be of higher microbial quality, through controlled production and ripening conditions.

New approaches, such as using bacteriocinogenic lactic acid bacteria (LAB) cultures and/or the bacteriocins of these cultures, to control pathogenic and spoilage microorganisms have been developed in order to increase food safety.5,6 Among bacteriocins, nisin is mainly used in food, and particularly in dairy products. However, less is known regarding the effects of nisin in meat products.5 Nisin exhibits activity against a range of Gram-positive bacteria. The spectrum of activity of this compound may also be extended to Gram-negative bacteria through utilising nisin in combination with other agents. Such a combinatory application may also enhance the activity of nisin against Gram-positive bacteria.7 Nisin is still the only bacteriocin approved as a food preservative (E234) in more than 50 countries worldwide, including the USA, the European Union, Brazil, and China.5 The effects of nitrite and nisin levels on biogenic amine formation were evaluated using response surface methodology in this study.

MATERIALS AND METHODS

Reagents and standards Dansyl chloride from Acros (Acros Organics, Geel, Belgium), and ammonia (25%), acetone, acetonitrile (high-performance liquid

Correspondence to: Sukr Kurt, Vocational School, University of Adyaman, u TR-02040 Adiyaman, Turkey. E-mail: sukrukurt@hotmail.com

a Department of Food Technology, Vocational School, University of Adyaman, TR-02040 Adiyaman, Turkey b Department of Food Engineering, University of Y z nc Yl, 65080 Van, Turkey u u u

2669

J Sci Food Agric 2010; 90: 26692674

www.soci.org

c 2010 Society of Chemical Industry

www.soci.org

S Kurt, O Zorba

Table 1. Central composite design of two independent variables Codied level Run order 1 2 3 4 5 6 7 8 9 10 X1 (nitrite) 1 1 1 0 0 0 0 1 1 1 X2 (nisin) 1 0 1 1 0 0 1 1 0 1 Actual level Nitrite (mg kg1 ) 0 0 0 100 100 100 100 200 200 200 Nisin (mg kg1 ) 0 250 500 0 250 250 500 0 250 500

Figure 1. Three-dimensional plots of the effects of nitrite and nisin on LAB. Table 2. Results of biogenic amine analysis of fresh meat Biogenic amine Tryptamine 2-Phenylethylamine Histamine Spermidine Cadaverine Tyramine Putrescine Spermine (mg kg1 ) 0.00 0.00 0.00 3.24 8.41 15.40 17.28 27.34

chromatography (HPLC) grade), ammonium acetate, and perchloric acid from Merck (Darmstadt, Germany) were used in HPLC analysis. Biogenic amine standards used were 1,7-diaminoheptane (internal standard), spermidine trihydrochloride, spermine tetrahydrochloride from Sigma (St Louis, MO, USA), and cadaverine dihydrochloride, histamine dihydrochloride, putrescine dihydrochloride, tryptamine hydrochloride, 2-phenylethylamine hydrochloride, and tyramine hydrochloride from Acros (Acros Organics). Nisin (Nisaplin MS-50) was obtained from Danisco (Niebull, Ger many). Nitrite was used in the form of sodium nitrite (Merck). Sucuk preparation Sucuk was prepared according to the following recipe:8 84.6% meat (beef), 9.4% lamb tail fat, 1.9% salt, 0.94% garlic, 0.66% red pepper, 0.47% black pepper, 0.85% cumin, 0.24% allspice, 0.47% sugar and 0.47% phosphate (K2 HPO4 ; Merck). The meat and fat pieces (4 cm3 in size), spices, garlic, salt, sugar, and phosphate were mixed and minced in a grinder (Cem, Turkey). Starter cultures (Lactobacillus sake, Pediococcus pentosaceus, Staphylococcus carnosus, and Staphylococcus xylosus; Bactoferm ; Chr. Hansen, Hrsholm, Denmark) were added to the sucuk batter and mixed in. Sucuk batter was divided into 10 equal parts and varying amounts of nisin and nitrite, which were dissolved in 20 mL distilled water, were added to each part as shown in Table 1. Each of the resulting batches of batter was rested for 12 h at 4 C and stuffed into collagen casings (Naturin Darm, Weinheim, Germany) of 35 mm diameter using a stufng machine (Cem, Istanbul, Turkey). Each sample was washed under running water, and then a 10% potassium sorbate solution was sprayed on to it.

Figure 2. Three-dimensional plots of the effects of nitrite and nisin on tryptamine.

Samples were ripened at 20 1 C for 13 days. For equilibration, relative humidity was adjusted to 60 3% during the rst 6 h of the ripening period and then increased to 87 3% and decreased every day by 1 unit. After the ripening period, samples were stored at 20 1 C until analysis. Microbiology analysis A 25 g sample was aseptically collected from each sucuk and homogenised in 225 mL of a sterile salt solution (0.85% NaCl; Merck) using a blender (Waring 80 011S, Torrington, CT, USA). The LAB number was determined on De Man Rogosa Sharpe Agar (MRS; Oxoid, Basingstoke, UK), which was incubated at 30 C for 72 h. Determination of biogenic amines The chromatographic method of Eerola et al.9 was used for the determination of tyramine, putrescine, cadaverine, histamine,

2670

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 26692674

Biogenic amine formation in Turkish dry fermented sausage

www.soci.org

Table 3. Analysis of variance of the effects of nitrite and nisin on biogenic amine values of sucuk Source of variation d.f. LAB R2 = 0.82 F-value 0.211 7.895 0.316 4.538 3.271 0.267 Tryptamine R2 = 0.95 F-value 51.468 19.410 3.866 1.706 5.450 33.799 2-Phenylethylamine R2 = 0.83 F-value 11.378 5.676 1.156 0.352 0.748 153.558 Tyramine R2 = 0.97 F-value 117.031 1.756 0.235 15.140 0.121 8.817 Putrescine R2 = 0.95 F-value 42.639 19.607 10.439 0.013 4.565 10.173 Cadaverine R2 = 0.95 F-value 16.331 44.081 4.222 2.354 10.352 22.013 Histamine R2 = 0.87 F-value 22.003 3.463 0.603 0.266 0.866 6.580 Spermidine R2 = 0.86 F-value 2.721 12.623 0.002 0.133 8.755 11.935 Spermine R2 = 0.67 F-value 5.753 1.292 0.185 0.021 0.943 11.765

X1 (nitrite) X2 (nisin) X1 X2 X1 X1 X2 X2 Lack of t Corrected total

1 1 1 1 1 3 9

P < 0.01 signicance level; P < 0.05 signicance level; X1 , nitrite (mg kg1 ); X3 , nisin (mg kg1 ); d.f., degrees of freedom.

Table 4. Predicted model equations for the effects of nitrite (X1 ) and nisin (X2 ) on biogenic amine values of sucuk Parameter LAB Tryptamine 2-Phenylethylamine Putrescine Predicted model equation Y = 8.512 + 0.013X1 0.082X2 0.099X1 2 0.084X2 2 + 0.020X1 X2 Y = 7.754 12.050X1 7.400X2 + 3.518X1 2 + 6.288X2 2 + 4.045X1 X2 Y = 5.820 3.138X1 2.217X2 + 0.885X1 2 + 1.290X2 2 + 1.225X1 X2 Y = 163.764 102.653X1 + 69.610X2 2.902X1 2 53.862X2 2 62.208X1 X2 Y = 283.13965.777X1 8.057X2 37.937X1 2 3.387X2 2 +3.610X1 X2 Y = 343.075 86.443X1 + 142.020X2 52.625X1 2 110.365X2 2 53.828X1 X2 Y = 2.397 1.727X1 + 0.685X2 0.304X1 2 0.549X2 2 0.350X1 X2 Y = 4.332 0.087X1 + 0.187X2 + 0.031X1 2 0.249X2 2 0.003X1 X2 Y = 37.650 + 1.502X1 + 0.712X2 + 0.145X1 2 + 0.975X2 2 0.330X1 X2

Tyramine Cadaverine

Histamine Spermidine Spermine

Figure 3. Three-dimensional plots of the effects of nitrite and nisin on 2-phenylethylamine.

2-phenylethylamine, tryptamine, spermidine, and spermine levels. Amines were separated using HPLC (Agilent 1100, Agilent Technologies, Boeblingen, Germany). The separation was carried out by gradient elution with 0.1 mol L1 ammonium acetate/acetonitrile on a reverse-phase column (Spherisorb ODS-2; 5 m, 125 4 mm; Waters Corporation, Milford, MA, USA) at a ow rate of 1 mL min1 using a diode array detector (G1315B DAD, Agilent Technologies) at 254 nm with 550 nm as a reference. A 4 g sample was weighed into test tube and added 250 L internal standard (1.7 diaminoheptane) and 10 mL of 0.4 mol L1 perchloric acid solution and homogenised with a homogeniser (Pro260, Pro, Oxford, CT, USA). The homogenised sample was centrifuged (Universal 32R, Hettich International, Tuttlingen, Germany) for 10 min at 2400 g and rinsed with supernatant into a 25 mL bottle through lter paper. The extraction was repeated with 10 mL of 0.4 mol L1 perchloric acid solution, mixed thoroughly with a Vortex mixer (Reax Top, Heidolph, Schwabach, Germany) and centrifuged as above. Supernatants were combined and adjusted to 25 mL with 0.4 mol L1 perchloric acid solution.

The alkalinity of a 500 L sample extract was adjusted using 200 L of 2 mol L1 sodium hydroxide solution. A 300 L saturated sodium bicarbonate was added as a buffer. A 1 mL dansyl chloride solution (10 mg mL1 acetone) was added and incubated at 40 C for 45 min. Residual dansyl chloride was removed by adding 100 L ammonia (25%). After 30 min, dansylated extract was adjusted to 5 mL with 0.1 mol L1 ammonium acetate/acetonitrile (1 : 1), and ltered through a 0.45 m syringe lter (Sartorius, Goettingen, Germany). Quantities of biogenic amines in the sample were calculated as Cu = 250 RF (HA /Hi ) Ci /WS where Cu is the unknown (mg kg1 sample), 250 is the dilution factor, RF is the response factor, HA is the peak height of unknown, Hi is the peak height of internal standard, Ci is the concentration of the internal standard, and WS is the weight of sample. Experimental design and statistical analysis The experimental design and statistical analysis were performed using JMP Software (SAS Institute Inc., Cary, NC, USA). The experiments, based on a central composite design with a total

2671

J Sci Food Agric 2010; 90: 26692674

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org

S Kurt, O Zorba

Figure 4. Three-dimensional plots of the effects of nitrite and nisin on tyramine.

Figure 5. Three-dimensional plots of the effects of nitrite and nisin on putrescine.

of 10 combinations, including two replicates of the central point, were carried out in random order. The codied and actual levels are given in Table 1. The variables were coded according to the following equation: Xi = (xi x i )/ xi where Xi is the coded value of an independent variable, xi is the real value of an independent variable, xi is the real value of an independent variable at the centre point, and xi is the step change. The variance for each factor assessed was partitioned into linear, quadratic, and interactive components and were represented using a second-order polynomial equation. The equation is
k k k 2 ii xii + i=1
i=1 j=1 i<j

Y = 0 +
i=1

i xi +

ij xi xj

where Y is the estimated response, 0 , i , ii , and ij are constant coefcients, k is the number of factor variables, and Xi , Xii , and Xij represent the linear, quadratic, and interactive effects of the independent variables, respectively. The analysis was performed using uncoded units.

Figure 6. Three-dimensional plots of the effects of nitrite and nisin on cadaverine.

RESULTS AND DISCUSSION

The results of a biogenic amine analysis of fresh meat are summarised in Table 2. The addition of nitrite and/or nisin affected some biogenic amine values, such as values for tryptamine, 2phenylethylamine, putrescine, cadaverine, histamine, tyramine, and spermidine. The linear effects of nisin were found to be signicant (P < 0.05) on LAB (Table 3). As shown in Fig. 1, the increasing level of nisin decreased counts of LAB, which may be the result of the inhibitory effect of nisin on Grampositive microorganisms.6,10 Dykes et al.11 reported that nisin had signicant effects on Lactobacillus plantarum, Lactobacillus brevis, and Lactococcus lactis. This result is important because these bacteria play an important role in dry fermented sausage production.

The linear effects of nitrite and nisin on tryptamine values were found to be signicant (P < 0.01, P < 0.05 respectively; Table 3). Tryptamine values decreased with increasing nitrite or nisin levels (Fig. 2). These results are important, since tryptamine is a vasoactive amine. The 2-phenylethylamine level decreased in sucuk with the addition of nitrite, which was statistically signicant (P < 0.05) (Table 3; Fig. 3). 2-Phenylethylamine generally occurs when a high level of tyramine is present.12 Nitrite was also found to have a similar effect on tyramine values (Fig. 4). The linear effect of nitrite on tyramine was found to be signicant (P < 0.01) (Table 3), and the quadratic effect was also found to be signicant (P < 0.05) for tyramine levels. The increasing rate of nitrite decreased tyramine values (Fig. 4), where this may result from the antimicrobial effects of nitrite on proteolytic organisms, which causes tyramine formation. Higher nitrite levels were more effective than lower concentrations at decreasing the rate of tyramine formation

2672

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 26692674

Biogenic amine formation in Turkish dry fermented sausage

www.soci.org

Figure 7. Three-dimensional plots of the effects of nitrite and nisin on histamine.

Figure 8. Three-dimensional plots of the effects of nitrite and nisin on spermidine.

(Fig. 4). The ndings of previous investigations8,13,14 support this result for sucuk production. The effect of nisin was not found to result in signicant effects (P > 0.05) on tyramine levels (Table 3). u Espinosa et al.15 similarly found that the addition of nisin to Cl a cheese did not reduce the amount of tyramine. The linear effect of nitrite was found to be signicant (P < 0.01) for putrescine (Table 3). The increasing rate of nitrite decreased putrescine values (Fig. 5). However, nisin increased putrescine values (P < 0.05) (Table 3 and Fig. 5). Their interactive effect was also found to be signicant (P < 0.05) for putrescine values. Some previous studies reported that the formation of putrescine is closely linked to the total aerobic viable count.16 18 In addition to aerobic bacteria, some LAB and Enterobacter can cause putrescine formation.4,12,19,20 Kurt,8 as well as Bozkurt and Erkmen,21,22 reported that nitrite signicantly decreased the counts of total aerobic mesophilic bacteria in sucuk production. Raju et al.23 reported that nisin at 50 p.p.m. was effective in reducing the total plate count and spore count in the sh sausage after 2 days of storage. The linear effect of nisin was found to be signicant (P < 0.01) on cadaverine levels (Table 3), and the quadratic effect was also found to be signicant (P < 0.05). The increased rate of nisin resulted in an augmentation of increased cadaverine values, and this increased rate was higher in the presence of low levels of nisin rather than in higher levels (Fig. 6). Espinosa et al.15 reported that the addition of nisin increased the amounts of cadaverine in Cl a cheese. However, nitrite decreased cadaverine u values signicantly (P < 0.05) (Fig. 6). The effects of nitrite on cadaverine were supported by previous investigations of sucuk with similar results.8,13,14 Moreover, nitrite was reported to decrease cadaverine concentrations in dry fermented Italian sausage.12 Members of Enterobacteriaceae and enterococci play an important role in the formation of cadaverine.12,19,24 Ercoskun et al.16 reported that cadaverine concentrations were directly related to Enterobacteriaceae counts. Histamine and spermidine were found at lower concentrations in sucuk (Figs 7 and 8). Nitrite signicantly (P < 0.01) decreased histamine levels. However, histamine levels were not signicantly affected by the addition of nisin (Table 3). The linear and quadratic

effects of nisin were found to have signicant effects (P < 0.05) on spermidine levels (Table 3). In particular, lower levels of nisin increased spermidine values (Fig. 8). The increasing amount of spermidine was less than 1 ppm. Spermidine and spermine were not considered an indicator of spoilage at certain levels, since these compounds can be natural components of muscles.25 The effects of nitrite and nisin on LAB count and biogenic amines are also expressed mathematically in Table 4. These predicted model equations are useful for understanding the signicance of amine levels and the interactions between studied factors.

CONCLUSION
Decreases in tryptamine, 2-phenylethylamine, putrescine, cadaverine, tyramine, and histamine levels demonstrated that nitrite is an indispensable component of sucuk, since the physical, chemical, and microbial quality of sucuk is partially related to nitrite presence. However, the additional nitrite levels can be decreased with the addition of nisin, which hinders biogenic amine formation. Some antimicrobial agents, which were effective on genera of Gram-negative bacteria, may be considered for utilisation in combination with nisin in dry fermented sausage to prevent the formation of biogenic amines. Thus nisin can play a more effective role in decreasing additional nitrite levels. However, the phenomenon of increasing levels of the some biogenic amines due to the addition of nisin needs further study to determine the nisinbiogenic amine interactions.

REFERENCES
1 Bover-Cid S, Izquierdo-Pulido M and Vidal-Carou MC, Inuence of hygienic quality of raw materials on biogenic amine production during ripening and storage of dry fermented sausages. J Food Prot 63:15441550 (2000). 2 Gokalp HY, N-Nitroso bile ikleri, kanserojenik etkileri, ce itli gdalarn N-nitrosamin icerikleri ve ce itli kaynaklardan bunyeye alnan N nitrosamin miktarlar (in Turkish). Gda 6:317324 (1984). 3 Hal sz A, Bar th A, Simon-Sarkadi L and Holzapfel W, Biogenic amines a a and their production by microorganisms in food. Trends Food Sci Technol 5:4249 (1994). 4 Bover-Cid S, Izquierdo-Pulido M and Vidal-Carou MC, Changes in biogenic amine and polyamine contents in slightly fermented

2673

J Sci Food Agric 2010; 90: 26692674

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org
sausages manufactured with and without sugar. Meat Sci 57:215221 (2001). Reunanen J and Saris PEJ, Bioassay for nisin in sausage; a shelf life study of nisin in cooked sausage. Meat Sci 66:515518 (2004). de Martinis ECP, Alves VF and Franco BDGM, Fundamentals and perspectives for the use of bacteriocins produced by lactic acid bacteria in meat products. Food Rev Int 18:191208 (2002). Singh B, Falahee MB and Adams MR, Synergistic inhibition of Listeria monocytogenes by nisin and garlic extract. Food Microbiol 18:133139 (2001). Kurt S, The effects of fermentation time, nitrite level and heat treatment on biogenic amine formation and some properties of sucuk. PhD thesis. Yuzuncu Yl University, Van, Turkey (2006). Eerola S, Hinkkanene R, Lindfors E and Hirvi T, Liquid chromatographic determination of biogenic amines in dry sausages. J AOAC Int 76:575578 (1993). Chen H and Hoover DG, Bacteriosins and their food applications. Comp Rev Food Sci Food Saf 2:82100 (2003). Dykes GA, Amarowicz R and Pegg RB, Enhancement of nisin antibacterial activity by a bearberry (Arctostaphylos uva-ursi) leaf extract. Food Microbiol 20:211216 (2003). Suzzi G and Gardini F, Biogenic amines in dry fermented sausages: a review. Int J Food Microbiol 2731:114 (2003). Kurt S and Zorba O, The effects of ripening period, nitrite level and heat treatment on biogenic amine formation of sucuk; a Turkish dry fermented sausage. Meat Sci 82:179184 (2009). Genccelep H, Kaban G and Kaya M, Effects of starter cultures and nitrite levels on formation of biogenic amines in sucuk. Meat Sci 77:424430 (2007). Espinosa D, de Lamo S, Hernandez M, Lopez T and Roig AX, Formation of biogenic amines in Cl a cheese treated by high hydrostatic u

S Kurt, O Zorba
pressure, in Health Implications of Dietary Amines: A Joint COST Action 922 and Biochemical Society Focused Meeting, University of Aberdeen, 1921 October (2006). Ercoskun H, Con AH and Gokalp HY, Biyojenik aminler ve gdalarda mikroorganizmalarca uretimi (in Turkish). Standard 5661 (2000). Ruiz-Capillas C and Jimenez-Colmenero F, Biogenic amines in meat and meat products. Crit Rev Food Sci Nutr 44:489499 (2004). Bozkurt H, Utilization of natural antioxidants: green tea extract and Thymbra spicata oil in Turkish dry-fermented sausage. Meat Sci 73:442450 (2006). Shalaby AR, Signicance of biogenic amines to food safety and human health. Food Res Int 29:675690 (1996). Bover-Cid S, Migu lez-Arrizado J and Vidal-Carou MC, Biogenic amine e accumulation in ripened sausages affected by the addition of sodium sulphite. Meat Sci 59:391396 (2001). Bozkurt H and Erkmen O, Effect of nitrate/nitrite on the quality of sausage (sucuk) during ripening and storage. J Sci Food Agric 84:279286 (2004). Bozkurt H and Erkmen O, Effects of some commercial additives on the quality of sucuk (Turkish dry-fermented sausage). Food Chem 101:14821490 (2007). Raju CV, Shamasundar BA and Udupa KS, The use of nisin as a preservative in sh sausage stored at ambient (28 2 C) and refrigerated (6 2 C) temperatures. Int J Food Sci Technol 38:171185 (2003). Montel MC, Masson F and Talon R, Bacterial role in avour development. Meat Sci 49:111113 (1998). Hern ndez-Jover T, Izquierdo-Pulido M, Veciana-Nogues MT, Marinea Font A and Vidal-Carou MC, Biogenic amine and polyamine contents in meat and meat products. J Agric Food Chem 45:20982102 (1997).

5 6 7 8 9 10 11 12 13 14 15

16 17 18 19 20 21 22 23

24 25

2674
wileyonlinelibrary.com/jsfa
c 2010 Society of Chemical Industry J Sci Food Agric 2010; 90: 26692674

Research Article
Received: 31 May 2010 Revised: 5 July 2010 Accepted: 26 July 2010 Published online in Wiley Online Library: 2 September 2010

(wileyonlinelibrary.com) DOI 10.1002/jsfa.4139

Relationship between changes in the total concentration of acetic acid bacteria and major volatile compounds during the acetic acid fermentation of white wine
Silvia Baena-Ruano,a Ines M Santos-Duenas,a Juan C Mauriciob a and Isidoro Garca-Garca
Abstract
BACKGROUND: In the scope of the wine vinegar production, this paper provides comprehensive information about the evolution of some volatile compounds during the biological acetication cycle. These data were compared with the acidity, cell concentration and ethanol concentration. Such information may allow a better understanding of the complex biological processes involved. RESULTS: The volatile compounds 2-phenylethanol, diethyl succinate (diethyl butanedioate), meso-2,3-butanediol (mesobutane-2,3-diol), levo-2,3-butanediol (levo-butane-2,3-diol), methanol and ethyl acetate exhibited no signicant changes between the starting wine and produced vinegar, whereas the rest [acetoin (3-hydroxybutan-2-one) excepted] ethyl lactate (ethyl 2-hydroxypropanoate), isoamyl alcohols (3-methylbutan-1-ol and 2-methylbutan-1-ol), isobutanol (2-methylpropan-1ol), 1-propanol (propan-1-ol), and acetaldehyde were consumed in substantial amounts during the process. Additionally, their specic evolution patterns alongside bacterial cell concentrations, acidity and ethanol concentration are shown. CONCLUSION: Concentrations of acetic acid bacteria at the end of the acetication cycle were found to vary because of cell lysis, a result of the high acidity and low ethanol concentration of the medium. Variations were similar to those in some volatile compounds, which suggests their involvement in the metabolism of acetic bacteria. The results testify to the usefulness of this pioneering study and suggest that there should be interest in similar, more detailed studies for a better knowledge of the presence of certain volatile compounds and metabolic activity in cells effecting the acetication of wine. c 2010 Society of Chemical Industry Keywords: wine; acetication; vinegar; volatile compounds; acetic acid bacteria

INTRODUCTION
Wine vinegar is an increasingly appreciated product by virtue of its sensory quality and richness. In fact, today, as much importance is attached to vinegar as to wine in winemaking areas which have traditionally obtained the vinegar as a by-product;1 this had led to the application of strict analytical and control methods to the acetication process with a view to improving the quality of the end product, and facilitating its characterisation and discrimination.2,3 Vinegar owes much of its sensory character to its aroma, which is a combination of the individual contributions of many volatile products.4 7 The nal composition of vinegar in such products depends on the particular raw material used, the way the alcoholic fermentation and subsequent biological oxidation are conducted, and the ageing procedure employed, if any. Many vinegars, some with a designation of origin included, are biologically oxidised in modern industrial fermentation tanks where the culture medium is subjected to substantial aeration. Despite the high efciency of the aeration process, there is always the risk of some volatile compounds being lost and the quality of the end product diminished as a result. However, using volatile condensers at

the gas output and optimising oxygen input to the reactor during the aeration process can help to substantially avoid volatile losses. Although, as noted earlier, the volatile composition of vinegar is widely variable, it usually includes higher alcohols, esters and some aldehydes and ketones such as acetoin, acetaldehyde, ethyl lactate, 2,3-butanediol, isoamyl alcohols, ethyl acetate, methanol and 2-phenylethanol as major components.8

Correspondence to: Isidoro Garca-Garca, Departamento de Ingeniera Qumica, Edicio Marie Curie, Facultad de Ciencias, Universidad de C rdoba, o Campus Universitario de Rabanales, 14071 C rdoba, Spain. o E-mail: isidoro.garcia@uco.es

a Departamento de Ingeniera Qumica, Edicio Marie Curie, Facultad de Ciencias, Universidad de C rdoba, Campus Universitario de Rabanales, o 14071 C rdoba, Spain o b Departamento de Microbiologa, Edicio Severo Ochoa, Facultad de Ciencias, Universidad de C rdoba, Campus Universitario de Rabanales, 14071 C rdoba, o o Spain

2675

J Sci Food Agric 2010; 90: 26752681

www.soci.org

c 2010 Society of Chemical Industry

www.soci.org Acetaldehyde present in vinegar is an intermediate in the catabolism of carbon-containing substrates used by acetic acid bacteria.9 This aldehyde is formed by oxidation of both ethanol and pyruvate, which in turn can be produced by the EntnerDoudoroff reaction of sugars, if available, or from lactate, especially at low ethanol levels.10 Acetoin is produced in substantial amounts during the acetication process. In fact, it is often used as a marker for the biological origin of vinegar since its synthesis is related to cell metabolism in acetic bacteria.11 Alcohols are also important to vinegar. Thus, isoamyl alcohols are usually consumed during the acetication of wine.12 On the other hand, levo- and meso-2,3-butanediol in wine are reportedly oxidised to acetoin,5 but were found to change little in content with respect to the starting wine here. Esters are also important as regards concentration and their potential inuence on vinegar aroma. Wine vinegar usually contains some, such as ethyl lactate, diethyl succinate, ethyl acetate, methyl acetate and isoamyl acetate.13 Also, isoamyl acetate and ethyl acetate are among the compounds with the highest odour activity value (OAV) in vinegar.14 Although the volatile composition of vinegar is widely documented, its changes during biological acetication of wine have seemingly been studied only once.13 Also, the study in question was quite comprehensive and interesting, the authors only analysed samples at the beginning, middle and end of the acetication cycle, which was inadequate to assess changes in volatile compounds throughout the acetication cycle with a view to identifying potential alterations in their assumed evolution patterns. Recently,15 changes in volatiles in balsamic and red wine vinegars stored in wooden casks and bottles were studied, but only between the start and end of the process. No study of the potential relationship between volatiles and microbial concentration changes during the acetication cycle appears to have been conducted to date. Elucidating such a relationship, if it does exist, would be helpful with a view to relating the synthesis and evolution of these compounds with chemical and biological activity in the acetication system. In this work, we studied the variations in major volatiles in wine vinegar and examined their potential relationship to the total concentration of cells throughout the acetication cycle and to various other important variables.

S Baena-Ruano et al. concentration of cells as measured with a method described elsewhere16 ranged from about 1 108 to 3.5 108 cells mL1 during the acetication cycle. Culture medium and fermentation conditions The raw material used was white wine from the MontillaMoriles region (Spain) (which is similar to sherry wine) containing 89.7 3.9 g ethanol L1 and having an initial acidity of 4.0 g acetic acid L1 . The bioreactor employed was operated in a semi-continuous mode (Fig. 1). Thus, once an ethanol concentration of 3.9 g ethanol L1 was reached, a portion of 75% of the total volume of culture medium was unloaded and the reactor replenished at a constant ow-rate of 0.01 L wine min1 . This cycle can be repeated an indenite number of times. The reactor was a Frings 8 L fermenter and operated at 31 C, using an aeration regime of 7.5 L air h1 L1 medium. The fermenter was loaded, unloaded and monitored in an automatic manner according to a programmed sequence. Proper operation in this situation entailed careful on-line measurement of the total ethanol concentration and volume of medium. An online probe Alcosens (Heinrich Frings GmbH & Co. KG, Bonn, Germany) and a differential pressure sensor (Yokogawa Iberia S.A., Madrid, Spain) were used for ethanol and volume determination respectively.17 Analysis of volatiles Major volatile compounds and polyols were quantied on a Model 6890 gas chromatograph from Agilent Technologies (Palo Alto, CA, USA), using the method described by Peinado et al.18 A CP-WAX 57 CB capillary column (60 m long 0.25 mm i.d., 0.4 m lm thickness) from Varian (Palo Alto, CA) was used, and 0.5 L aliquots from 10 mL samples previously supplied with 1 mL of 1 g L1 4-methyl-2-pentanol as internal standard were injected into the instrument. Tartaric acid in the wine was removed by precipitation with 0.2 g of calcium carbonate and centrifugation at 1380 g. Quantication was based on the response factors obtained for standard solutions of each compound. A split ratio of 30 : 1, an FID, and a temperature program involving an initial temperature of 50 C (15 min), a 4 C min1 ramp and a nal temperature of 190 C (35 min) were used. The injector and detector temperatures were 270 and 300 C, respectively. The ow rate of carrier gas (helium) was initially set at 0.7 mL min1 (16 min) and followed by a 0.2 mL min1 ramp to the nal value (1.1 mL min1 ), which was held for 52 min.

EXPERIMENTAL
Microorganism The inoculum used was a mixed culture of acetic bacteria from an industrial fermentation tank in full operation. The total

2676

Figure 1. Stages of a semi-continuous acetication process.

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 26752681

Relationship between acetic acid bacteria and volatile compounds in wine vinegar

www.soci.org

Figure 2. Variation of the concentrations of ethanol and oxygen, as well as the acidity and volume of the medium, in relation to the total number of cells during the acetication cycle. Bars represent standard deviations. The corresponding mean standard deviations for ethanol, oxygen and volume were about 3%, 2% and 2%, respectively.

Figure 3. Contents in volatiles of the starting wine and the resulting vinegar. Bars represent standard deviations.

RESULTS AND DISCUSSION

Figures 2 to 8 show the experimental data. The results shown in this work are the means for four cycles. Bars represent standard deviations. Figure 2 shows the variation of the main system variables during the acetication cycle. As can be seen, the volume and concentration of ethanol in the medium increased with time during the rst 10 h (loading stage); by contrast, the concentration of bacterial cells and the acidity of the medium decreased markedly by effect of dilution over the same period. The ethanol and acidity values at the end of the loading phase are the result of both the addition of wine as well as the bacteria activity. In fact, a global mass balance can shows that, approximately, 10 g L1 of acetic acid has been produced during this stage. Acetic bacteria

were subjected to a high stress as a result of abrupt changes in their environment during the loading stage, which was followed by an adaptation lag stage that lasted 8 h. Then, bacterial cells entered an exponential growth stage which lasted about 4 h. The last stage of the vinegar production cycle was especially interesting on account of the marked changes undergone by the bacterial cells. In a previous study19 on the variation of amino acids concentrations during the vinegar production cycle, such changes were found to be due to cell lysis phenomena. In fact, each anecdotal decrease in cell concentration was accompanied by a simultaneous anecdotal increase in the contents of many amino acids present that was ascribed to cellular autolysis. There was also evidence of the opposite changes (i.e. an increase in cell concentration concomitant with a decrease in amino acid levels).

2677

J Sci Food Agric 2010; 90: 26752681

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org

S Baena-Ruano et al.

Figure 4. Variation of the concentrations of acetoin and ethyl acetate in relation to the total number of cells during the acetication cycle. Bars represent standard deviations.

Figure 5. Variation of the concentrations of isoamyl alcohols, isobutanol and 1-propanol in relation to the total number of cells during the acetication cycle. Bars represent standard deviations.

These facts may be indicative or even provide evidence for a complex response of bacterial cells to stressing conditions in order to ensure survival of their population via programmed cell death, which allows cells to live on lysis products from dead cells. The previous results led us to examine changes in various compounds with a strong impact on the sensory properties of vinegar.5 Fig. 3 shows their concentrations in the starting wine and end product (vinegar), as well as the intervening changes. The compounds inside the box exhibited no signicant changes, whereas those outside it, acetoin excepted, were consumed in substantial amounts during the process. Acetoin is known to be involved in the biological oxidation of ethanol by acetic bacteria. Therefore, it is present in virtually all types of vinegar, albeit at widely variable concentrations. Thus, its content in pineapple

vinegar is typically in the region of 2 mg L1 ,20 whereas that in vinegar from sherry wines21 or cider22,23 can be as high as 1000 mg L1 or even higher. Acetoin can form in various ways5,11 including the condensation of two acetaldehyde molecules, the reaction between pyruvate and acetaldehyde, and the oxidation of 2,3-butanediol. Figure 4 shows the variation of the acetoin concentration alongside that of the bacterial cell concentration. As can be seen, both evolved virtually in parallel throughout the process. Because of its low level in the starting wine, the acetoin concentration decreased by effect of dilution during the reactor loading stage. Worthy of special note were the oscillations observed at the end of the process, consistent with changes in the total cell concentration which, as noted earlier, may be a result of cell lysis

2678

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 26752681

Relationship between acetic acid bacteria and volatile compounds in wine vinegar

www.soci.org

Figure 6. Variation of the concentrations of ethyl lactate and acetaldehyde in relation to the total number of cells during the acetication cycle. Bars represent standard deviations.

Figure 7. Variation of the concentrations of methanol and 2-phenylethanol in relation to the total number of cells during the acetication cycle. Bars represent standard deviations.

caused by an increased acidity and a reduced nutrient availability in the medium. The concomitance of these oscillations in the acetoin and cell concentrations provides further evidence for a relationship of the synthesis and changes in acetoin to biological activity in the system. Figure 4 also shows the variation of the ethyl acetate concentrations. Although the nal content of the vinegar in this compound was roughly the same as in the starting wine, this does not exclude potential changes in its concentration during the process. In fact, as can clearly be seen from the gure, this compound exhibited strong changes and an early evolution pattern differing markedly from that for acetoin. Judging by the high concentrations of acetic acid and ethanol present in the medium, the ester was most likely

formed by esterication outside bacterial cells. As fresh wine was added to the fermenter during the loading stage, the medium was supplied with additional ethanol that reacted with acetic acid accumulating in it to give the ester. Once loading was nished and ethanol started to be consumed by acetic bacteria, the reverse reaction (hydrolysis of ethyl acetate) gradually prevailed and led to a decrease in the ester concentration. Although the esterication reaction must occur outside cells, oscillations in cell concentrations had a marked effect on the ethyl acetate concentration (for example, the release of acetate to the medium by effect of cell disruption can displace equilibria to the ester formation). Despite the subsequent oscillations in cell concentrations, the increasing scarcity of ethanol in the medium led to a more limited esterica-

2679

J Sci Food Agric 2010; 90: 26752681

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org

S Baena-Ruano et al.

Figure 8. Variation of the concentration of 2,3-butanediol in relation to the total number of cells during the acetication cycle. Bars represent standard deviations.

tion reaction and prevented further ester concentration rises by effect of cell lysis. The four alcohols of Fig. 5 exhibited an increase in concentration at the start of the cycle as a result of their being supplied during the loading process. Subsequently, the alcohols started to decrease in parallel with the increase in acidity of the medium, which suggests that esterication with acetic acid was the main factor governing their evolution. However, the small, transient increase at the start of the oscillations in cell concentrations may also indicate that the three alcohols are involved in the metabolism of acetic bacteria. As can be seen from Fig. 6, there was a close relationship between cell, lactate and acetaldehyde concentrations. This is unsurprising if one considers the signicance of acetaldehyde as an intermediate in the oxidative metabolism of ethanol and various other compounds in acetic bacteria.9 Also, there is the well-known ability of acetic bacteria in using lactate to produce acetate via pyruvate rst and acetaldehyde then. The other volatile compounds (Figs 7 and 8) exhibited no signicant differences in concentration between the starting wine and the vinegar. Again, there were anecdotal changes in concentration at the end of the production cycle, which suggests the involvement of these compounds in the metabolism of acetic bacteria.

ACKNOWLEDGEMENTS
The authors are grateful to Spains Ministry of Education and Science, and to Grupo SOS, S.A. for funding this research in the framework of Projects AGL2002-01712, PET2006-0827, AGL20052494-E-ALI and AGL2009-08117-E-ALI. Co-funding by FEDER is also gratefully acknowledged.

REFERENCES
1 Dur n-Guerrero E, Control de los procesos de elaboraci n, calidad y a o trazabilidad del Vinagre de Jerez. PhD thesis. Puerto Real, C diz a (2008). 2 Callejon RM, Morales ML, Troncoso AM and Ferreira ACS, Targeting key aromatic substances on the typical aroma of sherry vinegar. J Agric Food Chem 56:66316639 (2008). 3 Tesfaye W, Morales ML, Garcia-Parrilla MC and Troncoso AM, Improvement of wine vinegar elaboration and quality analysis: instrumental and human sensory evaluation. Food Rev Int 25:142156 (2009). 4 Blanch GP, Tabera J, Sanz J, Herraiz M and Reglero G, Volatile composition of vinegars simultaneous distillation extraction and gas-chromatographic mass-spectrometric analysis. J Agric Food Chem 40:10461049 (1992). 5 Tesfaye W, Morales ML, Garca-Parrilla MC and Troncoso AM, Jerez Vinegar, in Vinegars of the World, ed. by Solieri L and Giudici P. Springer-Verlag Italia, Milan, pp. 180195 (2009). 6 Troncoso AM, Evaluacion sicoqumica y sensorial de vinagres, in Second Symposium on R+D+I for Vinegar Production, ed. by Garca Garca I. Servicio Publicaciones de la Universidad de Cordoba, Cordoba, pp. 195199 (2006). 7 Duran-Guerrero E, Marin RN, Mejias RC and Barroso CG, Stir bar sorptive extraction of volatile compounds in vinegar: validation study and comparison with solid phase microextraction. J Chromatogr A 1167:1826 (2007). 8 Callejon RM, Torija MJ, Mas A, Morales ML and Troncoso AM, Changes of volatile compounds in wine vinegars during their elaboration in barrels made from different woods. Food Chem 120:561571 (2010). 9 Bidan P, Divies C and Cachon R, Microorganismos de alteracion de los vinos, in Enologa: Fundamentos Cientcos y Tecnol gicos, ed. by o Flanzy C. A Madrid Vicente Ediciones and Ediciones Mundi-Prensa, Madrid, pp. 344356 (2000). 10 Morales ML, Gonzalez GA, Casas JA and Troncoso AM, Multivariate analysis of commercial and laboratory produced sherry wine

CONCLUSION
In summary, this paper provides for the rst time comprehensive information about the evolution of some volatile compounds during the biological acetication cycle. Such information may allow a better understanding of the complex biological processes involved. Thus, the oscillations in cell concentrations observed at the end of the cycle, which can be ascribed to cell adaptation and survival mechanisms, coincided with similar oscillations in the concentrations of the target compounds. Their apparent relationship may be of use to identify the specic compounds involved in the metabolism of acetic bacteria. To our knowledge, no study of this type has been conducted to date.

2680

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 26752681

Relationship between acetic acid bacteria and volatile compounds in wine vinegar
vinegars: inuence of acetication and aging. Eur Food Res Technol 212:676682 (2001). Polo MC and S nchez-Luengo AA, Las bacterias aceticas, in El Vinagre a de Vino, ed. by Llaguno C and Polo MC. Consejo Superior de Investigaciones Cientcas, Madrid, pp. 2547 (1991). Nieto J, Gonz lez-Vinas MA, Barba P, Martn-Alvarez PJ, Aldave L, a Garca-Romero E and Cabezudo MD, Recent progress in wine vinegar R&D and some indicators for the future, in Food Flavour, Ingredients and Composition, ed. by Charalombous G. Elsevier Science, New York, pp. 469499 (1993). Morales M, Tesfaye W, Garcia-Parrilla MC, Casas JA and Troncoso AM, Sherry wine vinegar: physicochemical changes during the acetication process. J Sci Food Agric 81:611619 (2001). Callejon RM, Morales ML, Ferreira ACS and Troncoso AM, Dening the typical aroma of sherry vinegar: sensory and chemical approach. J Agric Food Chem 56:80868095 (2008). Callejon RM, Tesfaye W, Torija MJ, Mas A, Troncoso AM and Morales ML, Volatile compounds in red wine vinegars obtained by submerged and surface acetication in different woods. Food Chem 113:12521259 (2009). Baena-Ruano S, Jimenez-Ot C, Santos-Duenas I, Cantero-Moreno D, Barja F and Garcia-Garcia I, Rapid method for total, viable and non-viable acetic acid bacteria determination during acetication process. Process Biochem 41:11601164 (2006). Garca-Garca I, Santos-Duenas I, Jimenez-Ot C, Jimenez-Hornero J and Bonilla-Venceslada J, Vinegar Engineering, in Vinegars of the

www.soci.org

11 12

18

19

13 14 15

20 21

22 23

16

World, ed. by Solieri L and Giudici P. Springer-Verlag Italia, Milan, pp. 97120 (2009). Peinado R, Moreno J, Munoz D, Medina M and Moreno J, Gas chromatographic quantication of major volatile compounds and polyols in wine by direct injection. J Agric Food Chem 52:63896393 (2004). Maestre O, Santos-Duenas I, Peinado R, Jimenez-Ot C, Garcia-Garcia I and Mauricio JC, Changes in amino acid composition during wine vinegar production in a fully automatic pilot acetator. Process Biochem 43:803807 (2008). Ou ASM and Chang RC, Taiwan fruit vinegar, in Vinegars of the World, ed. by Solieri L and Giudici P. Springer-Verlag Italia, Milan, pp. 223242 (2009). Morales ML, Tesfaye W, Garcia-Parrilla MC, Casas JA and Troncoso AM, Evolution of the aroma prole of sherry wine vinegars during an experimental aging in wood. J Agric Food Chem 50:31733178 (2002). Llaguno C, Denicion y tipos de vinagre, in El Vinagre de Vino, ed. by Llaguno C and Polo MC. Consejo Superior de Investigaciones Cientcas, Madrid, pp. 133145 (1991). Valero E, Berlanga T, Roldan P, Jimenez C, Garcia I and Mauricio JC, Free amino acids and volatile compounds in vinegars obtained from different types of substrate. J Sci Food Agric 85:603608 (2005).

17

2681
wileyonlinelibrary.com/jsfa

J Sci Food Agric 2010; 90: 26752681

c 2010 Society of Chemical Industry

Research Article
Received: 2 November 2009 Revised: 28 April 2010 Accepted: 26 July 2010 Published online in Wiley Online Library: 14 October 2010

(wileyonlinelibrary.com) DOI 10.1002/jsfa.4140

Chemical composition, antioxidant and antibacterial properties of the essential oils of Etlingera elatior and Cinnamomum pubescens Kochummen
Siddig Ibrahim Abdelwahab,a, Faridah Qamaruz Zaman,b,c Abdalbasit Adam Mariod,d Muhammad Yaacob,c Adil Hassan Ahmed Abdelmageedc and Shamsul Khamisc
Abstract
BACKGROUND: Plant essential oils are widely used as fragrances and avours. Therefore, the essential oils from the leaves of Cinnamomum pubescens Kochummen (CP) and the whole plant of Etlingera elatior (EE) were investigated for their antioxidant, antibacterial and phytochemical properties. RESULTS: CP and EE were found to contain appreciable levels of total phenolic contents (50.6 and 33.41 g kg1 as gallic acid equivalent) and total avonoid contents (205.6 and 244.8 g kg1 as rutin equivalent), respectively. DPPH free radical scavenging activity of CP is superior to EE (P < 0.05) showing IC50 of 77.2 and 995.1 g mL1 , respectively. Methicillin-resistant Staphylococcus aureus (MRSA), Bacillus subtilis, Pseudomonas aeruginosa and Salmonella choleraesuis were tested against CP and EE. Only MRSA was the most susceptible bacteria to CP. GC/MS studies resulted in the identication of 79 and 73 compounds in CP and EE, respectively. The most abundant components of EE included -pinene (24.92%) and 1-dodecene (24.31%). While the major compound in CP were 1,6-octadien-3-ol,3,7-dimethyl (11.55%), cinnamaldehyde (56.15%) and 1-phenyl-propane-2,2-diol diethanoate (11.38%). CONCLUSION: This study suggests that the essential oils from Cinnamomum pubescens Kochummen and Etlingera elatior could be potentially used as a new source of natural antioxidant and antibacterial in the food and pharmaceutical industries. c 2010 Society of Chemical Industry Keywords: antibacterial activities; antioxidant; chemical composition; Cinnamomum pubescens Kochummen; Etlingera elatior

INTRODUCTION
Essential oils from herbal sources are used in food avours, perfumes and pharmaceutical preparations for their functional properties.1 The commercial use of essential oils in aromatherapy constitutes little more than 2.0% of the total market.2 Moreover, the antibacterial properties of these herbal essential oils and their components are exploited in such diverse commercial products as dental root canal sealers, antiseptics and animal feed supplements.3 Besides the antibacterial properties, some of essential oils were proven to possess antioxidant properties.4,5 Williams and Harborne screened 39 species of ginger (Zingiberaceae) for their phytochemical constituents. Leaves of Alpinia and Zingiber were found to contain kaempferol and quercetin glycosides, and myricetin and quercetin glycosides, respectively.6 Flavonoids in the leaves of Etlingera elatior (Zingiberaceae) have been identied as kaempferol 3-glucuronide, quercetin 3glucuronide, quercetin 3-glucoside, and quercetin 3-rhamnoside. Members of the Etlingera genus have various remedial and commercial uses. Young shoots, owers and fruits are eaten either raw, cooked as a vegetable, or used as a condiment.7 Inorescences

of E. elatior are widely cultivated as spice for curry.8 Fruits are used to treat ear itching, while leaves are applied for healing wounds.7 E. elatior was also found to have anti-tumour promoting and cytotoxic activities.9,10

Correspondence to: Siddig Ibrahim Abdelwahab, Department of Pharmacy, Faculty of Medicine, University of Malaya, Kuala Lumpur, Malaysia. E-mail: siddigroa@yahoo.com

S.I. Abdelwahab and F.Q. Zaman contributed equally to this paper. a Department of Pharmacy, Faculty of Medicine, University of Malaya, Kuala Lumpur, Malaysia b Department of Biology, Faculty of Science, Universiti Putra Malaysia, Serdang, Malaysia c Biodiversity Unit, Laboratory of Natural Products, Institute of Bioscience, Universiti Putra Malaysia, Serdang, Malaysia d Food Science & Technology Department, College of Agricultural Studies, Sudan University of Science & Technology, Khartoum north, Sudan

2682

J Sci Food Agric 2010; 90: 26822668

www.soci.org

c 2010 Society of Chemical Industry

Composition and properties of oils from E. elatior and C. pubescens Cinnamomum pubescens Kochummen (family: Lauraceae) is a small tree up to 7 m tall. It is indigenous to Peninsular Malaysia; the leaves are opposite to alternate, with a hairy stalk 12 cm in length. The blade is leathery, drying greenish yellow, lanceolate, apex pointed, base cuneate, nely pubescent on the under surface, midrib and secondary nerves attened to sunken on the upper surface, raised on the lower surface.11,12 As with all cinnamomum species, the whole parts of the tree are aromatic and are used in traditional medicine as medang, but this species has not been used commercially.13 The leaf oil of C. pubescens showed signicant larvacidal and platelet-activating factor (PAF) receptor-binding antagonist activities.14 In this present study, the essential oils obtained from the whole plant of E. elatior and leaves of C. pubescens Kochummen (referred to as C. pubescens for the remainder of this paper) were analysed by GC/MS and their antioxidant activities using diphenylpicrylhydrazyl (DPPH) free-radical scavenging activity, antibacterial properties and their chemical composition were compared. To the best of our knowledge, this study is the rst of its kind to report the antioxidant activities and antibacterial effect of essential oils from E. elatior and C. pubescens.

www.soci.org Determination of total avonoid content Total avonoid content (TFC) was determined by the AlCl3 method, using rutin as a standard.16 The test samples were dissolved in dimethyl sulfoxide (DMSO). The sample solution (1.0 mL) was mixed with 1.0 mL of AlCl3 (0.15 mol L1 ). After 10 min of incubation at ambient temperature, the absorbance of the supernatant was measured at 435 nm using a Shimadzu UVvisible spectrophotometer (Mini 1240). Three replicates were made for each test sample. The total avonoid content was expressed as rutin equivalents (RE, mg g1 ). For the rutin, the curve was established by plotting concentration (mg mL1 ) versus absorbance (nm) (y = 5.6752x 0.0312; R2 = 0.994), where y is the absorbance and x is the concentration. DPPH radical scavenging antioxidant assay Radical scavenging activity of plant essential oils against stable DPPH (2, 2-diphenyl-2-picrylhydrazyl hydrate) (SigmaAldrich Chemie, Steinheim, Germany) was determined spectrophotometrically. When DPPH reacts with an antioxidant compound, which can donate hydrogen and is reduced, the changes in colour (from deep-violet to light-yellow) were measured at 517 nm wavelength.17 Radical scavenging activity of essential oils was measured by a slight modication of the method by Ao et al.18 Stock solutions were prepared in 10 mg mL1 in methanol. The working solution was prepared using methanol in a concentration of 2.0 mg mL1 (Labsystems iEMS Reader MF; Eichenweg, Aumuhle, Germany). The solution of DPPH in methanol (1 mmol L1 ) was freshly prepared, before UV measurements. Five microlitres of this solution were mixed with 100 L of serial dilutions of samples (15.6252000 g mL1 ) in a 96-well plate. The samples were kept in the dark for 30 min at ambient temperature and then the decrease in absorption was measured each 30 min for 2 h. Absorption of a blank sample containing the same amount of methanol and DPPH solution was prepared and measured daily. The experiment was carried out in triplicate. Radical scavenging activity was calculated by the following % inhibition = [(AB AA )/AB ] 100, where AB is the absorption of blank sample (t = 0 min); and AA is the absorption of tested samples (t = 30 min). The inhibitory concentration 50% was determined as well as the kinetics of DPPH scavenging reaction. Commercial standard antioxidant butylated hydroxytoluene (BHT) was also tested against DPPH and used as a reference. Antibacterial assay Microbial strains The antibacterial activity of essential oil samples was evaluated using two Gram-positive bacteria, methicillin resistant Staphylococcus aureus (MRSA) and Bacillus subtilis B29, and two Gram-negative bacteria, Pseudomonas aeruginosa 60 690 and Salmonella choleraesuis. All bacterial strains were obtained from the Laboratory of Molecular Biomedicine, Institute of Bioscience, Universiti Putra Malaysia, Serdang, Malaysia. Disc diffusion method Screening for the antibacterial effect of the essential oils was carried out by determining the zone of inhibition using paper disc (6 mm in diameter, Whatman No. 1) diffusion method.19 The microorganism strains obtained were inoculated in a Petri dish containing nutrient broth at 37 C for 24 h and were referred as seeded broth. The density of the bacterial suspension was standardised and the concentrations of the cultures were adjusted turbidometrically at

MATERIALS AND METHODS

All solvents used were of analytical grade. Methanol, ethyl acetate, hexane, chloroform, butylated hydroxytoluene (BHT) and FolinCiocalteu reagent were obtained from Merck (Darmstadt, Germany). Plant materials Fresh leaves of C. pubescens and whole plant of E. elatior were collected from Pahang and Selangor states, Malaysia, respectively, in 2009. Plants were identied by Assistant Professor Shamsul Khamis at the Unit of Biodiversity, Institute of Bioscience, Universiti Putra Malaysia, Malaysia. The voucher specimens under the plants names were deposited in the unit herbarium. Isolation procedure for the essential oils The fresh whole plant of E. elatior and leaves of C. pubescens, were steam distilled, separately, in a hydrodistillation apparatus (Clevenger-type) for 8 h. The essential oils were dried over anhydrous sodium sulfate and stored at 46 C before analysis. Antioxidant activity Determination of total phenolic content Total phenolic content (TPC) in essential oils of E. elatior and C. pubescens was determined with FolinCiocalteu reagent following the method of Kaur et al.15 Stock solutions of oils were prepared in a concentration of 10 mg mL1 , and a 50 L from this solution was transferred to a test tube (n = 3). To this tube, 0.4 mL of FolinCiocalteu reagent (1 : 10) was added and the tube was shaken thoroughly. After 1 min, 0.8 mL of sodium bicarbonate solution (0.9 mol L1 ) was added and the mixture allowed standing in dark room for 30 min with intermittent shaking. Absorbance was measured at 765 nm using a Shimadzu UVvisible spectrophotometer (Mini 1240; Shimadzu, Columbia, MD, USA). The total phenolic content (TPC) was expressed as gallic acid equivalent (GAE) in mg per g oil from the calibration curve of gallic acid standard solution. For the gallic acid, the curve was established by plotting concentration (mg mL1 ) versus absorbance (nm) (y = 5.145x + 0.014; R2 = 0.9975), where y is the absorbance and x is the concentration.

2683

J Sci Food Agric 2010; 90: 26822668

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org wavelength of 600 nm to 5 105 to 106 colony forming units per mL. The essential oils were dissolved in DMSO which was previously tested for antibacterial activity against all test bacteria and found to have no activity. Essential oils were diluted to a concentration of 100 mg mL1 and nally sterilised by ltration using 0.45 m Millipore lters. The sterile discs were impregnated with oil solution (0.05 mL from 100 mg mL1 ) to achieve desired concentration and placed in inoculated agar. Streptomycin (10 g mL1 ) susceptibility discs and methanol-impregnated discs were used as positive and negative controls, respectively. After incubation overnight at 37 C, inhibition zones were measured and recorded as mean diameter (mm). Antibacterial activity was also expressed as inhibition percentage of streptomycin. Minimum inhibitory concentration The least possible inhibitory concentrations of essential oils against MRSA were estimated using the agar disc method (ADM). Inoculation of 1.0 mL of MRSA was poured into each Petri dish and the agar was later dispensed and permitted to set. Wells were bored using a sterile 3.0 mm cork borer. Serial dilutions of the essential oil were added into the wells. The plates were incubated at 37 C for 24 h. The growth was observed to determine the sensitivity of MRSA using clear zones of no microbial growth. The least concentration of the essential oil that had inhibitory effect was taken as the minimum inhibitory concentration (MIC). Gas chromatography mass spectrometry The essential oils of E. elatior and C. pubescens were analysed by Shimadzu GC-MS (Model GC-17A). A FT-DB-5 capillary column (30 m 0.25 mm i.d. 0.25 m) was used for gas chromatographic separation of the analytes.20 The injection volume was 1.0 L with a split ratio of 13 : 1; the injector temperature was held constant at 230 C. Helium was used as the carrier gas with an inlet pressure of 21.0 kPa, corresponding to a ow rate of 1.0 mL min1 . The column oven temperature was set at 30 C (held for 3 min), raised at 8 C min1 to 230 C (held for 5 min), and nally held at 245 C for 10 min. The mass spectrometer was operated in the electron impact (EI) mode with ionisation energy of 70 eV. The transfer line was set at 290 C. The chemical constituents of the analytes were identied by comparing the MS fragmentation patterns with those of NIST/EPA/NIH mass special database library of the GC/MS system. Statistical analyses In order to determine whether there is a statistically signicant difference between the obtained results for the different assays the independent t-test was carried out using the SPSS 17.0 software package.

SI Abdulwahab et al.

Table 1. Total phenolic content, total avonoid content and DPPH IC50 (g mL1 ) of Etlingera elatior and Cinnamomum pubescens Kochummen Total avonoid content (RE g kg1 ) 244.83 15.5 205.65 30.4 Total phenolic content (GAE g kg1 ) 3341.2 92.1 5060.5 58.6 DPPH IC50 (g mL1 ) 995.1 123 77.2 8.5

Sample E. elatior C. pubescens

Results are expressed as average SD (n = 3). RE, rutin equivalent; GAE, gallic acid equivalent.

content (TFC) was determined by the AlCl3 method. Results were expressed as milligrams of rutin equivalent (RE) in one gram of oil (Table 1). The TFC for E. elatior was observed to be 244.83 15.5 g RE kg1 , which is not statistically different from the TFC for C. pubescens (205.65 30.4 g RE kg1 ) (Table 1). The current study shows that the essential oil from E. elatior has high TPC and TFC. However, it was reported earlier that a methanolic extract (not the essential oil) from fresh leaves of E. elatior has also shown high TPC.7 Antioxidant activity of essential oils from E. elatior and C. pubescens was also evaluated using the DPPH assay, which was conducted for 240 min (Fig. 1). Samples were able to reduce violet DPPH to the yellow DPPH-H, with an IC50 of 995.1 123 and 77.2 8.5 g mL1 for C. pubescens and E. elatior, respectively (Table 1). No earlier reports are available regarding the DPPH radical scavenging activity of the essential oils of E. elatior and C. pubescens with which to compare the results of our present analysis. However, Chan et al., reported the IC50 for a methanolic extract of fresh leaves of E. elatior to be 37.5 mg kg1 .7 The higher TPC of C. pubescens may explain its superiority compared to E. elatior. In addition, DPPH protection by C. pubescens may be due to the antioxidative action of eugenol (7.27%), which has been detected by GC-MS. This phenolic compound has been detected earlier in the essential oils of eight Cinnamomum species.21 Despite the fact that E. elatior essential oil has a higher avonoid content (Table 1), signicant contributors to the high antiradical effect, it did not show a scavenging effect compared to C. pubescens. These surprising results might be explained by the existence of -pinene (monoterpene hydrocarbons). Some isolated terpenes have been previously tested individually in order to determine the antioxidant nature of the oils, such as -pinene, but none has exhibited antioxidant activity.22 To the best of our knowledge, no data have been published on antioxidant activity, using the DPPH method, on E. elatior and C. pubescens essential oils. Antibacterial activities Among the tested bacteria, only MRSA was the most sensitive organism to the essential oils of C. pubescens. The mean diameter of the zone of inhibition of C. pubescens was 15 mm for MRSA which represents 75% inhibition compared to streptomycin MRSA (Table 2). The essential oil of C. pubescens was reported earlier to be antifungal.21 Cinnamaldehyde (the major constituent of C. pubescens, 56.15%) is known to be antibacterial against Escherichia coli and Salmonella typhimurium; it did not disintegrate the outer membrane or deplete the intracellular ATP pool.23 The carbonyl group of some kind of potential antibacterial agents is thought to bind to proteins, preventing the action

RESULTS AND DISCUSSION

Antioxidant capacity Hydro-distillation of fresh leaves of C. pubescens and the whole plant of E. elatior, afforded colourless pleasant-smelling essential oils. The total phenolic content (TPC) of these essential oils was determined using the FolinCiocalteu method and expressed in mg GAE g1 . Results presented in Table 1 showed that C. pubescens (50.6 0.58 g GAE kg1 ) had higher TPC when compared to E. elatior (33.41 0.92 g GAE kg1 ). Independent t-test statistical analysis showed the mean TPC of C. pubescens is signicantly different (n = 3) at the 0.05 level of signicance. Total avonoid

2684

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 26822668

Composition and properties of oils from E. elatior and C. pubescens

www.soci.org

Figure 1. Antioxidant activity of Etlingera elatior (upper photo), Cinnamomum pubescens (lower photo).

Table 2. Antibacterial activity of essential oils of Etlingera elatior and Cinnamomum pubescens Kochummen against bacteria using the disc-diffusion method and minimum inhibitory concentration Methicillin resistant Staphylococcus aureus Inhibition zone (mm) 15 (75) 20 MIC (mg mL1 ) 10 Pseudomonas aeruginosa Inhibition zone (mm) 20 MIC (mg mL1 ) Salmonella choleraesuis Inhibition zone (mm) 23 MIC (mg mL1 ) Bacillus subtilis Inhibition zone (mm) 23 MIC (mg mL1 )

Sample Etlingera elatior Cinnamomum pubescens Kochummen Control (streptomycin) Methanol

a The screening of the essential oils antibacterial effect was carried out by determining the zone of inhibition using paper disc (6 mm in diameter, Whatman No. 1) diffusion method (n = 2). Figures in parentheses are inhibition percentages compared to streptomycin. MIC: Minimum inhibitory concentration.

2685
wileyonlinelibrary.com/jsfa

J Sci Food Agric 2010; 90: 26822668

c 2010 Society of Chemical Industry

www.soci.org of amino acid decarboxylases in Enterobacter aerogenes.24 Grampositive Bacillus subtilis was observed to be resistant for both E. elatior and C. pubescens. Previous studies on the antimicrobial activity of essential oils obtained from ginger species also showed weak inhibition of bacteria.25 The mean diameter of the zone of inhibition of streptomycin was 20 mm for MRSA and Pseudomonas aeruginosa and 23 mm for Salmonella choleraesuis and B. subtilis (Table 2). The solvent used to prepare the reference and test samples showed no inhibitory effect on the bacteria used.

SI Abdulwahab et al. Essential oils of E. elatior and C. pubescens failed to inhibit Gram-negative S. choleraesuis and P. aeruginosa (Table 2). Selective permeability of the outer membrane of Gram-negative bacteria makes it generally less susceptible to volatile oils than the Gram-positive bacteria. Gram-negative P. aeruginosa is known to have a high level of intrinsic resistance to virtually almost all known antimicrobials and antibiotics, due to a very restricted outer membrane barrier, highly resistant even to synthetic drugs.26 To the best of our knowledge, this is the rst study reporting antibacterial activities of the essential oils of E. elatior and C. pubescens.

Table 3. Compounds tentatively identied in the essential oil of Cinnamomum pubescens Kochummen Compound no. 1 2 3 4 5 6 7 8 9 10 11 12 12 out of 79
a b

RTa 14.132 15.019 16.377 16.475 19.408 23.132 26.547 26.731 27.870 32.470 35.476 39.822 Total

RCb (%) 0.94 0.71 0.77 0.51 0.72 11.55 1.40 1.13 0.82 56.15 7.27 11.38 93.35

Compoundc Bicyclo[3.1.1]hept-2-ene, 2,6,6-trimethyl Camphene Benzaldehyde -Pinene Eucalyptol 1,6-Octadien-3-ol,3,7-dimethyl Benzene propanal Borneol ()--Terpineol (p-menth-1-en-8-ol) Cinnamaldehyde Phenol, 2-methoxy-4-(2-propenyl)-, acetate (eugenol) 1-Phenyl-propane-2,2-diol diethanoate

Molecular weight 136 136 106 136 154 154 134 154 154 132 206 236

Similarity (%) 95 94 94 88 85 95 95 94 94 98 76 77

RT, Retention time (min). Relative area percentage (peak area relative to the total peak area percentage). c Compounds are listed in order of their relative area percentage.

Table 4. Compounds tentatively identied in the essential oil of Etlingera elatior Compound no. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 20 out of 73
a b

RTa 14.151 16.528 17.206 26.919 27.908 27.917 28.325 32.448 32.639 37.937 38.074 39.554 39.744 40.935 41.675 45.447 46.405 46.663 48.837 49.334 Total

RCb (%) 11.59 24.92 0.60 0.86 0.66 0.74 0.69 0.74 1.38 8.15 2.49 2.41 1.99 24.31 0.90 2.56 3.49 1.22 0.63 2.27 92.6%

Compoundc Bicyclo[3.1.1]hept-2-ene, 2,6,6-trimethyl -Pinene -Myrcene Bicyclo[3.1.1]heptan-3-one, 2,6,6-trimethyl ()--Terpineol (p-menth-1-en-8-ol) 7-Methylene-9-oxabicyclo[6.1.0]non-2-ene Decanal 2-Undecanone 3-Bromo-7-methyl-1-adamantane carboxylic acid Dodecanal -Farnesene 1,6,10-Dodecatriene, 7,11-dimethyl-3-methylene -Caryophyllene 1-Dodecene 2-Tridecanone trans-(Z)--Bisabolene epoxide Acetic acid 2-Pentadecyn-1-ol (E)-10-Pentadecenol 1,3-Propanediol, 2-dodecyl

Molecular weight 136 136 136 152 154 136 156 170 272 184 204 204 204 168 198 220 228 224 226 244

Similarity (%) 97 96 91 92 86 78 94 95 78 97 87 91 94 96 96 86 87 83 92 93

RT, Retention time (min). Relative area percentage (peak area relative to the total peak area percentage). c Compounds are listed in order of their relative area percentage.

2686

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 26822668

Composition and properties of oils from E. elatior and C. pubescens

www.soci.org

Figure 2. Total ionic chromatogram (GC-MS) of essential oils of Etlingera elatior (A) and Cinnamomum pubescens Kochummen (B) obtained with 70 eV using a A FT-DB-5 capillary column (30 m 0.25 mm i.d. 0.25 m). Helium was used as the carrier gas with an inlet pressure of 21.0 kPa, corresponding to a ow rate of 1.0 mL min1 .

Chemical composition The chemical composition of E. elatior and C. pubescens essential oils studied by GC-MS is presented in Tables 3 and 4. Total ionic chromatograms of both E. elatior and C. pubescens are presented in Fig. 2. A total of 79 compounds were characterised in C. pubescens while the essential oil of E. elatior consists of 73 compounds. The most abundant components in the leaf essential oil of E. elatior included bicyclo[3.1.1] hept-2-ene, 2,6,6-trimethyl (11.59%), -pinene (24.92%), 3-bromo-7-methyl1-adamantanecarboxylic acid (1.38%), dodecanal (8.15%), farnesene (2.49%), 1,6,10,-dodecatriene, 7,11-dimethyl-3-methylene (2.41%), -caryophyllene (1.99%), 1-dodecene (24.31%), trans-(Z)--bisabolene epoxide (2.56%), acetic acid (3.49%) and 1,3-propanediol, 2-dodecyl (2.27%) (Table 3). 1,6-Octadien3-ol,3,7-dimethyl (11.55%), borneol (1.13%), cinnamaldehyde (56.15%), phenol, 2-methoxy-4-(2-propenyl)-, acetate (7.27%) and 1-phenyl-propane-2,2-diol diethanoate (11.38%) were the major compounds in C. pubescens (Table 3). Bicyclo [3.1.1] hept-2-ene, 2,6,6-trimethyl was found to be more abundant in the essential oil of E. elatior (11.59%) compared to C. pubescens (0.94%), while -pinene is abundant in the latter

(24.92%) compared to C. pubescens (0.51%) (Tables 3 and 4). Caryophyllene was reported previously in E. elatior and C. pubescens; however, the current study revealed the absence of this sesquiterpene hydrocarbon in C. pubescens.25 Previously, Jaafar et al. analysed the essential oils isolated from different parts (leaves, stems, owers and rhizomes) of Malaysian E. elatior using GC-MS.9 The leaf essential oil was found to contain -pinene (19.7%), -caryophyllene (15.4%) and trans--farnesene (27.1%) as the major compounds whereas the stem essential oil was largely dominated by 1,1-dodecanediol diacetate (34.3%) and trans-5-dodecene (27.0%). The essential oils of the owers and rhizomes contained 1,1-dodecanediol diacetate (24.4% and 40.4%, respectively) and cyclododecane (47.3% and 34.5%, respectively) as the major compounds. The current study demonstrated a higher percentage of -pinene and lower for -caryophyllene and trans--farnesene.

CONCLUSIONS
Essential oils of C. pubescens and E. elatior have signicant differences in their chemical composition and antibacterial

2687

J Sci Food Agric 2010; 90: 26822668

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org activities. C. pubesecens essential oil showed an interesting antibacterial effect against methicillin resistant Staphylococcus aureus, a bacterium responsible for difcult-to-treat infections in humans. In view of the antioxidant properties of these two essential oils they might be considered for inclusion as natural antioxidants in nutraceutical and pharmaceutical preparations.

SI Abdulwahab et al.
the Straits Settlements and Federated Malay States by the Crown Agents for the Colonies, London (1935). Jantan I, Ra IAA and Jalil J, Platelet-activating factor (PAF) receptorbinding antagonist activity of Malaysian medicinal plants. Phytomed 12:8892 (2005). Kaur R, Arora S and Singh B, Antioxidant activity of the phenol rich fractions of leaves of Chukrasia tabularis A. Juss. Bioresour Technol 99:76927698 (2008). Quettier-Deleu C, Gressier B, Vasseur J, Dine T, Brunet C and Luyckx M, Phenolic compounds and antioxidant activities of buckwheat (Fagopyrum esculentum Moench) hulls and our. J Ethnopharmacol 72:3542 (2000). Brand-Williams W, Cuvelier ME and Berset C, Use of a free radical method to evaluate antioxidant activity. LWT-Food Sci Technol 28:2530 (1995). Ao C, Li A, Elzaawely AA, Xuan TD and Tawata S, Evaluation of antioxidant and antibacterial activities of Ficus microcarpa L. l. extract. Food Cont 19:940948 (2008). Sahoo S, Kar DM, Mohapatra S, Rout SP and Dash SK, Antibacterial activity of Hybanthus enneaspermus against selected urinary tract pathogens. Indian J Pharm Sci 68:653 (2006). Ibrahim H, Aziz AN, Syamsir DR, Ali NAM, Mohtar M and Ali RM, Essential oils of Alpinia conchigera Griff. and their antimicrobial activities. Food Chem 113:575577 (2009). Jantan I, Karim Moharam BA, Santhanam J and Jamal JA, Correlation between chemical composition and antifungal activity of the essential oils of eight Cinnamomum species. Pharma Biol 46:406412 (2008). Marin R, Apel MA, Limberger RP, Raseira MCB, Pereira JFM and Zuanazzi JAS, Volatile components and antioxidant activity from some Myrtaceous fruits cultivated in Southern Brazil. Lat Am J Pharm 27:172177 (2008). Friedman M, Kozukue N and Harden LA, Cinnamaldehyde content in foods determined by gas chromatography-mass spectrometry. J Agric Food Chem 48:57025709 (2000). Wendakoon CN and Sakaguchi M, Inhibition of amino acid decarboxylase activity of Enterobacter aerogenes by active components in spices. J Food Protect 58:280283 (1995). Mackeen MM, Ali AM, El-Sharkawy SH, Manap MY, Salleh KM and Lajis NH, Antimicrobial and cytotoxic properties of some Malaysian traditional vegetables (ulam). Pharma Biol 35:174178 (1997). Mann CM, Cox SD and Markham JL, The outer membrane of Pseudomonas aeruginosa NCTC 6749 contributes to its tolerance to the essential oil of Melaleuca alternifolia (tea tree oil). Lett Appl Microbiol 30:294297 (2000).

14 15 16

REFERENCES
1 Surburg H, Panten J and Bauer K, Common Fragrance and Flavor Materials: Preparation, Properties and Uses. Vch Verlagsgesellschaft Mbh, Weinheim, West Germany (2006). 2 Burt S, Essential oils: their antibacterial properties and potential applications in foods a review. Int J Food Microbiol 94:223253 (2004). 3 Doel MA and Segrott J, Materializing complementary and alternative medicine: aromatherapy, chiropractic, and Chinese herbal medicine in the UK. Geoforum 35:727738 (2004). 4 Kelen M and Tepe B, Chemical composition, antioxidant and antimicrobial properties of the essential oils of three Salvia species from Turkish ora. Bioresour Technol 99:40964104 (2008). 5 Kivrak , Duru ME, Ozturk M, Mercan N, Harmandar M and Topcu G, Antioxidant, anticholinesterase and antimicrobial constituents from the essential oil and ethanol extract of Salvia potentillifolia. Food Chem 116:470479 (2009). 6 Williams CA and Harborne JB, The leaf avonoids of the Zingiberales. Biochem Syst Ecol 5:221229 (1977). 7 Chan EWC, Lim YY and Omar M, Antioxidant and antibacterial activity of leaves of Etlingera species (Zingiberaceae) in Peninsular Malaysia. Food Chem 104:15861593 (2007). 8 Lemmens R, Soerianegara I and Wong WC, Plant Resources of Southeast Asia No. 5 (2). Timber Trees: Minor Commercial Timbers. Backhuys Publishers, Leiden (1995). 9 Jaafar FM, Osman CP, Ismail NH and Awang K, Analysis of essential oils of leaves, stems, owers and rhizomes of Etlingera elatior (Jack) RM Smith. Malaysian J Anal Sci 11:269273 (2007). 10 Habsah M, Ali AM, Lajis NH, Sukari MA, Yap YH and Kikuzaki H, Antitumor promoting and cytotoxic constituents of Etlingera elatior. Malaysian J Med Sci 12:612 (2005). 11 Verheij EWM and Coronel RE, Plant Resources of South-East Asia. Pudoc, Jakarta, Kuala Lumpur (1991). 12 Kochummen KM, Family: Lauraceae, Tree Flora of Malaya, Vol. 4. Longmans, Kuala Lumpur (1989). 13 Burkill IH and Birtwistle W, A Dictionary of the Economic Products of the Malay Peninsula. Published on behalf of the Governments of 17 18 19 20 21

22

23 24 25 26

2688
wileyonlinelibrary.com/jsfa
c 2010 Society of Chemical Industry J Sci Food Agric 2010; 90: 26822668

Research Article
Received: 9 June 2010 Revised: 21 July 2010 Accepted: 28 July 2010 Published online in Wiley Online Library: 17 September 2010

(wileyonlinelibrary.com) DOI 10.1002/jsfa.4141

Morphological and qualitative characterisation of globe artichoke head from new seed-propagated cultivars
Anna Bonasia,a Giulia Conversa,a Corrado Lazzizera,a Giuseppe Gambacortab and Antonio Eliaa
Abstract
BACKGROUND: Three new artichoke seed-propagated hybrids (Tempo, Opal and Madrigal) were compared with two standard cultivated varietal types [Catanese and Violet du Provence (VP)] in terms of head morphology, processing performance, nutritional or technological qualitative traits, in order to dene their best use. RESULTS: Compared to the other genotypes, Opal and Madrigal had more rounded, heavier, larger heads, higher processing yield (>400 g of heart kg1 raw head) and lower total phenol (TP) content (2.4 g of gallic acid equivalents kg1 FW). VP gave a higher processing yield than Catanese and showed the highest TP content (6.5 g kg1 FW). Tempo hearts were more similar to those of VP in biometrical and chemical terms (P, Na, K, Ca); they had the highest dry matter content (163 g kg1 FW) and the waste left after peeling had the highest TP content. CONCLUSIONS: Hybrid artichokes, especially Opal and Madrigal, appear more suitable for the processing industry and also for fresh-cut production due to their highest processing yield and lowest total phenol content. Because of its high total phenol content, Tempo waste represents a possible source of natural antioxidant in the pharmaceutical eld and in the food industry (as a food additive). c 2010 Society of Chemical Industry Keywords: processing yield; polyphenols; antioxidant activity

INTRODUCTION
Globe artichoke [Cynara cardunculus L. subsp. scolymus (L.) Hayek] is a perennial rosette plant grown for its capitula, commonly referred to as heads or buds. The edible part (heart) consists of a receptacle and innermost tender bracts. Artichoke has a marked antioxidative and health protective potential due to its high levels of phenolic compounds, which are also important from a technological point of view. In fact, in stored artichoke enzymatic or chemical reactions are responsible for the appearance of browning phenomena,1 representing a signicant problem both for industrial processing and, above all, for fresh-cut preparation. Globe artichoke production is traditionally based on vegetatively propagated cultivated varietal types. Those widely grown in Apulia are Violet du Provence (VP) and Catanese, both with reowering characteristic, whose heads are harvested both during autumnwinter (early production) and spring (late production). Normally, early production is for the fresh market, whereas the last part of the late production is industrially processed (canned or frozen). In the last three decades artichoke breeding programmes have produced some seed (achens) propagated hybrids, which are starting to be considered in the main artichoke producing areas (even in Apulia) as an alternative to the traditional ones. Being propagated from seeds, these hybrids are free of the

main endemic diseases (e.g. viruses and fungi, like Verticillium daliae), infecting propagation material of cultivated. They have high agronomic performance2 4 and they are also suitable both for the fresh market and processing industry,5,6 even if they have a late or mediumlate harvest. In previous studies agronomic,2,6,7 morphological or qualitative8,9 traits were evaluated on different hybrids. However, there is little or no information on the head/heart biometrical and qualitative traits of the new seed-propagated hybrids Madrigal, Opal and Tempo. With the aim of dening the nutritional and technological quality of the heads of these latter hybrids, in the present work the morphological and the chemical characteristics of heads in Madrigal, Opal and Tempo were evaluated and compared with those of the standard cultivated varietal types (Violet du Provence

Correspondenceto:Anna Bonasia,DepartmentofAgro-Environmental Science, Chemistry and Plants Protection DiSACD University of Foggia, via Napoli 25, 71100 Foggia, Italy. E-mail: a.bonasia@unifg.it

a Department of Agro-Environmental Science, Chemistry and Plants Protection DiSACD University of Foggia, via Napoli 25, 71100 Foggia, Italy b Department of Engineering and Management of the Agricultural, Livestock and Forest Systems PROGESA University of Bari, Via Amendola 165/A, 70126 Bari, Italy

2689

J Sci Food Agric 2010; 90: 26892693

www.soci.org

c 2010 Society of Chemical Industry

www.soci.org and Catanese) for the Apulia region, one of the most important areas for artichoke production in Italy.

A Bonasia et al.

MATERIALS AND METHODS

Chemicals -Carotene, linoleic acid and Tween 40 were purchased from Sigma Chemical Co. (St Louis, MO, USA). Sodium bicarbonate, gallic acid, FolinCiocalteu reagent, hydrochloric acid, chloroform and methanol analytical grade from Carlo Erba (Rodano, Milan, Italy). Ultrapure water (Millipore, Billerica, MA, USA) was used. Collection of raw material On 18 April 2008 artichoke heads of the cultivated types (Catanese and Violet du Provence (VP)) and of the new hybrids (Tempo, Opal and Madrigal (Nunhems Netherlands BV, Haelen, The Netherlands) were simultaneously harvested from commercial elds in the Apulia region. The heads were picked at the optimal stage for fresh consumption and, from the whole mass of heads, four replications of 20 heads were randomly selected for each genotype. Biometric measurements and head peeling After completely removing the oral stem, the heads were processed to obtain the edible part (hearts) as follows: the 30 outermost tough bracts were rst removed and the apex of the head was cut across at 70% of head height, then starting from the 31st outermost bract, the cutting force of each single bract was measured and it was eliminated if the cutting force was higher than a threshold value. A digital pressure tester (model 53 205, TR; Turoni & C. s.n.c., Forl`, Italy), equipped with a prismatic body was used (width 3 mm and length 10 mm). The threshold value was dened as the level of tenderness acceptable for processing and was xed at 40 N by preliminary tests. The cutting force was measured 5 mm below the transversal point of the upper cut of the bract, parallel to this and in the middle position with respect to the breadth of the bract. The inedible apex and the eliminated bracts were collected as waste. The fresh weight (FW), diameter and length of raw head, FW and diameter of heart, number and FW of inner bracts, number of outer bracts and the FW of waste were all measured. In addition, the ratio between length and diameter of heads was calculated as a representative shape index (SI). Moreover, the processed yield was expressed as grams of produce obtained from 1 kg of raw head, and the number of heads needed to obtain 1 kg of produce was determined. Samples of hearts were dried at 65 C in a thermo-ventilated oven till constant weight to determine dry weight (DW). Preparation of plant material for chemical analyses Samples of hearts and waste were separately sliced into small pieces, treated by liquid nitrogen, lyophilised (model Lio5P; CinquePascal s.r.l., Trezzano, Milano, Italy), and subsequently nely ground using a mortar. The modied method described by Gil-Izquierdo et al.10 was used for preparation of plant extracts: lyophilised samples (1 g) were extracted with 2 20 mL of water/methanol (20 : 80; v/v) solution using a refrigerated centrifuge Beckman Coulter Allegra 25 (Fullerton, CA, USA) at 21 000 g for 15 min at 4 C. The supernatant was extracted twice, recovered and ltered through Whatman no. 1 lter paper. The ltrate was considered as the artichoke extract and kept in the dark at 20 C until use in the following assays. All samples were analysed in triplicate.

Content of inorganic cations and phosphorus Inorganic cations were extracted from 1 g of lyophilised heart samples, previously ashed in a mufe furnace at 550 C for 6 h, and digested with 20 mL of 1 mol L1 HCl in boiling water for 30 min. They were then determined by ion chromatography (Dionex ICS 3000; Dionex, Sunnivale, CA, USA) with a conductivity detector, using the pre-column IonPack CG12A and the column of separation IonPack CS12A (4 250 mm, 5 m), according to the method reported by Serio et al.11 Phosphorus was determined on lyophilised heart samples by spectrophotometry (Shimadzu UV-1800; Shimadzu Scientic Instruments, Columbia, MD, USA), following the method proposed by Miller.12 Total phenol content as determined by the FolinCiocalteu assay Total phenol (TP) content was determined on methanolic heart and waste extracts according to the method of Singleton and Rossi.13 The methanolic extracts (100 L) were mixed with 0.5 mL of FolinCiocalteu reagent and left to stand at room temperature for 5 min before 1.0 mL of sodium bicarbonate solution (Na2 CO3 , 20%) was added to the mixtures. After 45 min at 30 C, absorbance was read at 750 nm (Shimadzu UV-1800). Results were expressed as grams of gallic acid equivalents kg1 FW, using a calibration curve. Antioxidant activity by -carotene/linoleic acid assay The antioxidant activity (AA) of heart and waste methanolic extracts was assayed based on the -carotene bleaching method developed by Ismail et al.14 In this assay, the solution of -carotene/linoleic acid mixture was prepared as follows: in 3 mL -carotene solution (5 mg -carotene in 50 mL chloroform) were dissolved in Tween 40 (0.400 g) and linoleic acid (0.040 g). Chloroform was removed under vacuum at reduced pressure using a rotary evaporator Buchi R-200 (Buchi Labortechnik AG, Flawil, Switzerland). Following evaporation, 100 mL of distilled water saturated with oxygen was added to the mixture, with vigorous shaking to form an emulsion. The mixture was then added to artichoke extracts or methanol (as control). Then, the test tubes were incubated in a water bath at 50 C for 2 h. The absorbance was measured before and after the incubation phase at 470 nm (Shimadzu UV-1800). AA was expressed as a percentage, using the formula Ai,s Af,s AA = Ai,c Af,c where Ai,s is the initial absorbance of the sample; Af,s is the nal absorbance of the sample; Ai,c is the initial absorbance of the control; and Af,c is the nal absorbance of the control. Statistical analysis Data were analysed by ANOVA using the GLM procedure of the SAS Software (SAS, Cary, NC, USA) and mean separation was performed using the LSD test.

RESULTS
Morphological characteristics and processing features of heads Raw heads of Madrigal and Opal had the highest values in morphological parameters (on average, FW 277.8 g, length

2690

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 26892693

Quality in new seed-propagated artichokes

www.soci.org

Table 1. Morphological characteristics of raw artichoke heads and related processing performance in the different artichoke genotypes Raw head Genotype Catanese VP Madrigal Opal Tempo Signicance
ad

Heart FW (g) 117.8c 141.9bc 292.9a 262.7a 164.9b

Waste Bracts (n) 40c 56ab 50b 63a 49b

Diameter (mm) 65.4c 69.4c 94.9a 93.8a 76.6b

Length (mm) 88.3abc 81.7c 94.5a 94.1ab 87.3bc

Diameter (mm) 36.4c 46.8b 58.3a 59.2a 47.8b

FW (g) 35.4c 52.5b 119.5a 112.3a 61.5b

Bracts (g) 19.8d 30.8c 74.7a 65.6b 36.8c

Bracts (n) 33 34 38 37 37 NS

FW (g) 80.1c 89.2c 172.7a 149.5b 102.8c

Processed yield (g kg1 raw head) 300d 384bc 408ab 427a 374c

Means in columns not sharing the same letter are signicantly different according to LSD test (P < 0.05). NS, , not signicant or signicant at P 0.01 or P 0.001, respectively. It is included also the weight of head apex.

94.3 mm, diameter 94.3 mm), while raw heads of standard cultivated types showed the lowest values (on average, FW 129.8 g, length 85.0 mm, diameter 67.4 mm). The morphological parameters in Tempo raw head were intermediate between those of cultivated types and hybrids (diameter) or similar to VP (length and FW). The results for Tempo were similar to standard type VP even in shape (shape index (SI) = 1.1; data not shown). For the standard types, the raw head of Catanese was slightly more oblong (SI = 1.3) compared to VP (data not shown), even if the length and diameter of the raw heads were not signicantly different between the two cultivated varietal types (on average, 85.0 and 67.4 mm, respectively) (Table 1). Moreover, Madrigal and Opal raw heads resulted were completely round in shape, with SI = 1.0, on average (data not shown). Once processed, Madrigal and Opal also showed the highest heart FW and diameter (on average, 115.9 g FW, 58.7 mm diameter). Tempo showed hearts similar to those of VP (on average, 57.0 g FW, 47.3 mm diameter), while Catanese had the smallest hearts (35.4 g FW, 36.4 mm diameter). After peeling, Opal hearts presented the highest number of remaining inner bracts (63), followed by those of Madrigal (50) and Tempo (49). Catanese hearts showed the lowest number of inner bracts (40), while in VP it was similar to the hybrids. The FW of inner bracts was the highest in Madrigal hearts (74.7 g), followed by Opal hearts (65.6 g); in Tempo it was similar to VP (on average, 33.8 g), while the lowest value was observed in Catanese hearts (19.8 g).

The number of (outer) bracts eliminated during the peeling process was not affected by the genotype, being 36, on average. Waste FW (outer bracts plus head apex) was the highest in Madrigal (172.7 g), followed by Opal (149.5 g), while Tempo and the two cultivated varietal types showed the lowest value (on average, 90.7 g). The highest processing yield was in Opal and Madrigal (on average, 417 g kg1 raw head), while it was lowest in Catanese (300 g kg1 raw head). Tempo recorded a intermediate value to VP (on average, 379 g kg1 raw head). Quantitative characterisation of the hearts and the waste Results for the quantitative characterisation of hearts and waste are given in Table 2. The hearts of Tempo had the highest dry matter content (163 g kg1 FW), followed by Catanese (133 g kg1 FW), while Madrigal showed the lowest value (107 g kg1 FW). Intermediate values, not dissimilar from those of Madrigal and Catanese, were found in VP and Opal (on average, 122 g kg1 FW). The highest contents of mineral elements in the hearts were observed in Madrigal for P and K (6.7 and 29.9 g kg1 DW, respectively), in Madrigal and Catanese for Ca (on average, 3.5 g kg1 DW), and in Opal and Catanese for Na (on average, 4.9 g kg1 DW). The lowest values were observed in Tempo and VP for P and K content (on average, 4.2 and 14.5 g kg1 DW, respectively), in Tempo and Opal for Ca content (on average, 2.0 g kg1 DW) and in Tempo, VP and Madrigal for Na level (on average, 1.6 g kg1 DW). Intermediate values were found in

Table 2. Quantitative characteristics of the heart, and total phenol and antioxidant activity of the waste in the different artichoke genotypes Minerals in heart (g kg1 DW) Genotype Catanese VP Madrigal Opal Tempo Signicance
ac

Total phenol (g gallic acid equivalent kg1 FW) Ca 3.5a 2.5ab 3.5a 2.8b 1.3b

Antioxidant activity (%) Heart 72a 78a 74a 77a 53b

Dry matter (g kg1 FW) 133b 117bc 107c 127bc 163a

P 5.7ab 4.0b 6.7a 6.1ab 4.5b

Na 5.6a 1.8b 1.5b 4.2a 1.6b

K 24.2ab 16.4bc 29.9a 20.3abc 12.7c

Mg 3.2 2.6 2.3 2.5 1.3 NS

Heart 3.8b 6.5a 2.3c 2.6c 2.6c

Waste 6.9a 7.2a 2.0b 3.5b 8.0a

Waste 63 67 64 67 61 NS

NS, , , and , not signicant or signicant at P 0.05, P 0.01 or P 0.001, respectively. Means in columns not sharing the same letter are signicantly different according to LSD test (P < 0.05).

2691
wileyonlinelibrary.com/jsfa

J Sci Food Agric 2010; 90: 26892693

c 2010 Society of Chemical Industry

www.soci.org Catanese and Opal for P and K (on average, 5.9 and 22.0 g kg1 DW, respectively) and in VP for Ca (2.5 g kg1 DW). Mg heart content was not affected by the genotype (on average 2.4 g kg1 DW). VP hearts had the highest TP content (6.5 g of gallic acid equivalent kg1 FW), followed by Catanese (3.8 g kg1 FW), while the hybrids had the lowest content (on average, 2.5 g kg1 FW). Tempo, VP and Catanese showed the highest TP content in waste (on average, 7.4 g kg1 FW), Madrigal and Opal had the lowest (on average, 2.7 g kg1 FW). Tempo hearts had lower AA than the other genotypes (53% vs. 75%). The AA in waste parts was not affected by the genotype, being 64%, on average.

A Bonasia et al.

DISCUSSION
Though in the last decade a number of papers have conrmed the superiority of seed-propagated artichoke hybrids in terms of yield and quality compared to traditional cultivar types,2,4,6,7 the size of their heads when used for processing is still questionable from a technical point of view. Peeling machines in artichoke canning factories require correctly sized vegetables, and they are set to work with the small-sized heads of cultivated varietal types, especially at the end of the harvesting season. These new cultivars may be protably used for heart production, but only by setting the peeling machines to work with larger-sized artichokes. The new peeling machines are able to work with a head size ranging from 30 to 100 mm in diameter. Among the genotypes tested in this research, the seedpropagated hybrids, in particular Madrigal and Opal, stand out compared to the vegetatively propagated types; in fact, they had the largest heads in terms of both weight and size, especially due to their larger diameter. The number of bracts eliminated during the peeling process to obtain the edible hearts did not differ between the tested genotypes, but the weight of waste (outer bracts + apex) was considerably higher in Madrigal and Opal (Table 1). Since this weight consists almost exclusively of the external bracts, the latter observation conrms that in Madrigal and Opal external bracts are thicker than in the other genotypes, as also revealed by visual analysis. Despite the higher waste weight recorded in Madrigal and Opal, its incidence on the total head weight was lower than in the standard types, especially Catanese. Indeed, the lowest processing yield was found in this latter cultivated type (300 g kg1 raw head), while in Madrigal and Opal it was more than 400 g kg1 raw head (Table 1). Tempo and VP processing yield values were more similar to the latter (Table 1). Conrming the results of previous work on Madrigal,15,16 nine heads were necessary for 1 kg of hearts with Madrigal and Opal, while a considerably higher number was necessary in Tempo (17), VP (20), and especially in Catanese (29) (data not shown). Catanese showed a processing yield in agreement with that observed in other research15 and with the average processing yield recorded for the most cultivated European varieties;17 these values are largely lower compared to those gathered from the best two hybrids considered in this research. As the inner bract FW in Madrigal and Opal was almost double that of Tempo and VP and almost four-fold higher than in Catanese, and considering that their number was on average slightly higher only if compared to Catanese (Table 1), their greater thickness and eshiness is conrmed.

After head peeling, the incidence of the remaining bracts on the heart weight was about 59% for all the genotypes, underlining a substantial equilibrium between the two parts of the heart (receptacle and inner bracts). Tempo was similar to VP both from a biometric point of view (FW, diameter and shape of head, number and FW of inner bract, processed yield) (Table 1) and in chemical composition (P, Na, K, Ca) (Table 2). The TP content in the hearts of standard types, and particularly those of VP (more than 6.0 g of gallic acid equivalent kg1 FW), was higher than that of the hybrids (Table 2). This is in agreement with Lombardo et al.,9 who have recently found a lower TP content in the inner bract and receptacle of Madrigal and Tempo than in Violetto di Sicilia (a cultivated varietal type very close to Catanese) and VP. The high abundance of these compounds in the standard types gives them a greater nutritional value. However, the lower phenol content in seed-propagated hybrids brings them the advantage of lower browning processes during storage, thus minimising the need for antioxidant treatments in processing steps.18 This factor demonstrates the greater suitability of the hybrids compared to standard types for standard industrial processing and especially in fresh-cut produce. A recent study evaluating different artichoke genotypes for fresh-cut preparation has underlined that those with the lowest polyphenol content are more suitable for this transformation.19 Therefore, the polyphenol content of an artichoke cultivar plays a key role in dening how it can best be used.1 Although TP content was far lower in the hybrid hearts than in those of standard types, the antioxidant capacity did not differ between hybrid, especially in Madrigal and Opal, and cultivated varietal type hearts (Table 2). This indicates that the AA capacity in Madrigal and Opal hearts, as well as the phenol fraction, is also formed by other concurrent antioxidative compounds. When screening different artichoke cultivars, Cabezas-Serrano et al.19 found the lowest TP content in Catanese, but it had the highest AA and vitamin C, and these authors concluded that the latter is the concurrent antioxidant compound. The average values of AA indicate there is good control capacity of oxidative damage in the tested artichoke genotypes (Table 2), in agreement with Lattanzio et al.20 In addition to processing aspects, Madrigal and Opal also stand out for their high nutritional value in terms of mineral content and particularly that of P, K and Ca (Table 2). The cation prole has been determined for vegetatively propagated types such as Violetto di Toscana,21 but no reference data has been found for the new hybrid cultivars considered in this research. By comparing our ndings in mineral composition with the only available values, found on the INRAN website (National Research Institute for Food and Nutrition Italian Ministry of Agriculture), P content overlaps with INRAN data,22 while the content of the main cations is, on average, lower.

CONCLUSIONS
Tempo and Catanese stand out for their higher phenol content in the waste than in the heart. The chemical characterisation of the waste contributes to indicating artichoke by-products as a potential source of natural (and non-toxic) antioxidant phenols.23 The results of this work contribute to the characterisation of new cultivars of artichokes and dene the best way they can be used. Opal and Madrigal have high potential suitability in the processing

2692

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 26892693

Quality in new seed-propagated artichokes industry, especially in fresh-cut production, due to their very low TP content, a key factor for limiting browning phenomena in processed artichokes. Moreover, their high AA in hearts could be strategic in prolonging shelf life. In contrast, vegetatively propagated types can be assessed as best suited to fresh consumption, because they are characterised by a lower processing yield compared to seed-propagated types, and by a high nutritional value in terms of antioxidant compounds. Among the hybrids studied, Tempo was the most similar to the standard types, especially to VP, from a biometric point of view, in chemical composition and in processing performance, but stands out for the high dry-matter content of its hearts, which could be useful in some processing preparations (dehydratated or fried produce), and for the high TP content of the waste, which could represent a source of natural antioxidants.

www.soci.org
9 Lombardo S, Pandino G, Mauromicale G, Knodler M, Carle R and Schieber A, Inuence of genotype, harvest time and plant part on polyphenolic composition of globe artichoke [Cynara cardunculus L. var. scolymus (L.) Fiori]. Food Chem 119:11751181 (2009). 10 Gil-Izquierdo A, Gil MI, Conesa MA and Ferreres F, The effect of storage temperatures on vitamin C and phenolics content of artichoke (Cynara scolymus L.) heads. Innov Food Sci Emerg Technol 2:199202 (2001). 11 Serio F, De Gara L, Caretto S, Leo L and Santamaria P, Inuence of an increased NaCl concentration on yield and quality of cherry tomato grown in posidonia (Posidonia oceanica (L) Delile). J Sci Food Agric 84:18851890 (2004). 12 Miller RO, Extractable chloride, nitrate, orthophosphate, potassium, and sulphate-sulfur in plant tissue: 2% acetic acid extraction, in Handbook of Reference Methods for Plant Analysis, ed. by Klara YP. CRC Press, Boca Raton, FL, pp. 115118 (1998). 13 Singleton VL and Rossi Jr JA, Colorimetry of total phenolics with phosphomolybdicphosphotungstic acid reagents. Am J Enol Vitic 16:144158 (1965). 14 Ismail A, Marjan ZM and Foong CW, Total antioxidant activity and phenolic content in selected vegetables. Food Chem 87:581586 (2004). 15 Conversa G, Bonasia A, Lazzizera C and Elia A, Head processing suitability in Madrigal and Brindisino artichoke cultivar. Acta Hortic (2010), in press. 16 Del Nobile MA, Conte A, Scrocco C, Laverse J, Brescai I, Conversa G, et al, New packaging strategies to preserve fresh-cut artichoke quality during refrigerated storage. Innov Food Sci Emerg Technol 10:128133 (2009). 17 Lahoz L, Malumbres A, Macua JI, Bozal JM, Urmentea I and Arrondo MA, Agricultural and industrial study of the principal European varieties of artichoke in Navarra. Acta Hortic 660:531537 (2004). 18 Garcia EL and Barrett DM, Preservative treatments for fresh-cut fruits and vegetables, in Fresh-Cut Fruits and Vegetables. Science, Technology and Market, ed. by Lamikanra O. CRC Press, Boca Raton, FL, pp. 267303 (2002). 19 Cabezas-Serrano AB, Amodio ML, Cornacchia R, Rinaldi R and Colelli G, Screening quality and browning susceptibility off ve artichoke cultivars for fresh-cut processing. J Sci Food Agric 89:25882594 (2009). 20 Lattanzio V, Cicco N and Linsalata V, Antioxidant activities of artichoke phenolics. Acta Hortic 681:421427 (2005). 21 Romani A, Pinelli P, Cantini C, Cimato A and Heimler D, Characterization of Violetto di Toscana, a typical Italian variety of artichoke (Cynara scolymus L.). Food Chem 95:221225 (2006). 22 INRAN. Istituto Nazionale di ricerca per gli Alimenti e la Nutrizione. Ministero delle Politiche Agricole, Alimentari e Forestali, Tabelle di composizione degli alimenti. Available: http://www.inran.it/ 646/tabelle di composizione degli alimenti.html?idalimento= 005120&quant=100 [10 May 2010]. 23 Llorach R, Espn JC, Tom s-Barber n FA and Ferreres F, Artichoke a a (Cynara scolymus L.) by-products as a potential source of healthpromoting antioxidant phenolics. J Agric Food Chem 50:34583464 (2002).

ACKNOWLEDGEMENTS
We are grateful to Mrs M.A. Previtali for technical assistance with the laboratory analyses.

REFERENCES
1 Lattanzio V, Kroon PA, Linsalata V and Cardinali A, Globe artichoke: A functional food and source of nutraceutical ingredients. J Functional Foods 1:131144 (2009). 2 Miguel A, Baixauli C, Aguilar JM, Giner A, Maroto JV, Loperz S, et al, Cultivar trials of seed propagated artichoke. Acta Hortic 660:111114 (2004). 3 Calabrese N, De Palma E and Bianco VV, Yield and quality of new commercial seed grown artichoke hybrids. Acta Hortic 660:7782 (2004). 4 Calabrese N, De Palma E and Bianco VV, Yield and quality of seed propagated artichoke hybrid cultivars grown for four years. Acta Hortic 681:135142 (2005). 5 Calabrese N, Cardinali A, Di Venere D, Linsalata V, Pieralice M, Sergio L, et al, Technological parameters and suitability to freezing of seed grown artichoke hybrids. Acta Hortic 681:495501 (2005). 6 Baixauli C, Giner A, Miguel S, Lopez B, Pascual B and Maroto JV, Agronomic behaviour of seed propagated artichoke cultivars in the spanish mediterranean area. Acta Hortic 730:143147 (2007). 7 Mauromicale G and Ierna A, Characteristics of head of seed-grown globe artichoke [(Cynara cardunculus L. var. scolymus (L.) Fiori] as affected by harvest period, sowing date and gibberellic acid. Agronomie 20:197204 (2000). 8 Di Venere D, Linsalata V, Calabrese N, Cardinali A, Sergio L and Pieralice M, Biochemical characterization of new seed propagated artichoke cultivars. Acta Hortic 681:517522 (2005).

2693
wileyonlinelibrary.com/jsfa

J Sci Food Agric 2010; 90: 26892693

c 2010 Society of Chemical Industry

Research Article
Received: 26 May 2010 Revised: 27 July 2010 Accepted: 28 July 2010 Published online in Wiley Online Library: 2 September 2010

(wileyonlinelibrary.com) DOI 10.1002/jsfa.4142

Effect of storage on chemical and sensory proles of peanut pastes prepared with high-oleic and normal peanuts
Cecilia G Riveros,a Marta G Mestrallet,a Maria F Gayol,b Patricia R Quiroga,a Valeria Nepoteb and Nelson R Grossoa
Abstract
BACKGROUND: Peanut paste and peanut butter have high oil contents and are thus susceptible to developing rancidity and off-avours through lipid oxidation. Preservation of the chemical and sensory quality of these products is one of the main problems in the peanut industry. The purpose of this study was to compare the chemical and sensory stability of peanut paste prepared with high-oleic peanuts (cv. Granoleico, GO-P) with that of peanut paste prepared with normal peanuts (cv. Tegua, T-P) from Argentina. RESULTS: Chemical (peroxide and p-anisidine values and conjugated dienes) and sensory (roasted peanutty, oxidised and cardboard avours) indicators of lipid oxidation were measured in peanut pastes stored at 4, 23 and 40 C. Chemical indicator values and oxidised and cardboard avours showed lower increments in GO-P than in T-P during storage. T-P had signicantly higher peroxide value than GO-P. Roasted peanutty avour showed a lower decrease in GO-P. Peanut paste prepared with high-oleic peanuts had four (at 4 C), two (at 23 C) and three (at 40 C) times longer shelf-life than peanut paste prepared with normal peanuts. CONCLUSION: These results indicate that high-oleic Granoleico kernels provide peanut paste with higher protection against lipid oxidation. c 2010 Society of Chemical Industry Keywords: peanuts; high oleic; sensory; stability

INTRODUCTION
A large proportion of peanut production worldwide is destined for domestic foods.1 Peanuts contain high levels of oil (450540 g kg1 ) and protein (250310 g kg1 ). Owing to their high oil content and elevated unsaturated fatty acid concentration (300350 g kg1 linoleic acid, 450500 g kg1 oleic acid), peanuts are susceptible to lipid oxidation.2 Peanut butter and peanut paste are important peanut products used for direct consumption or as ingredients in the preparation of other foods. Preservation of the chemical and sensory quality of these products is an ongoing concern in the peanut industry. Many factors inuence the shelf-life of peanut products, such as variety, kernel ripeness at harvest, seed size, processing and storage conditions (temperature, time, light and oxygen). Researchers have shown increased interest in high-oleic peanut cultivars owing to their low degree of lipid oxidation during storage, which signicantly improves the preservation of sensory and chemical quality parameters. This high-oleic trait was initially reported by Norden et al.3 Other researchers have reported peanut lines from the USA with about 800 g kg1 oleic acid and 2030 g kg1 linoleic acid.4,5 Peanut products made from high-oleic varieties are expected to have higher stability. This effect was detected in roasted peanuts6 and fried-salted peanuts.7 Furthermore, studies have shown that the consumption of

high-oleic oils has potential health benets, such as lowering blood cholesterol levels in hypercholesterolaemic women.2 The concentration of oleic acid in high-oleic peanut oil is higher than that in olive oil,8 which is known for its heart-healthy characteristics. Dry-roasted peanuts6 and fried-salted peanuts7 prepared with high-oleic peanuts showed longer stability during storage. However, the chemical and sensory stability of peanut paste prepared with high-oleic peanuts has not been studied in depth, especially that prepared with Argentinean peanut varieties. The use of high-oleic peanuts rather than normal peanuts would increase shelf-life and improve the oxidative stability of peanut paste, thus preventing loss of sensory and nutritional quality. The purpose of this study was to compare the chemical and sensory stability of peanut paste prepared with high-oleic peanuts

Correspondence to: Nelson R Grosso, Qumica Biol gica, Facultad de Ciencias o Agropecuarias (UNC), IMBIV-CONICET, CC 509, 5000 C rdoba, Argentina. o E-mail: nrgrosso@agro.unc.edu.ar

a Qumica Biol gica, Facultad de Ciencias Agropecuarias (UNC), IMBIV-CONICET, o CC 509, 5000 C rdoba, Argentina o b ICTA, Facultad de Ciencias Exactas, Fsicas y Naturales, IMBIV-CONICET, C rdoba, Argentina o

2694

J Sci Food Agric 2010; 90: 26942699

www.soci.org

c 2010 Society of Chemical Industry

Peanut pastes prepared with high-oleic peanuts with that of peanut paste prepared with normal peanuts from Argentina.

www.soci.org samples. All panellists were selected according to the following criteria: (a) people with natural dentition; (b) people without food allergies; (c) non-smokers; (d) people between the ages of 18 and 64; (e) people who consumed roasted peanuts and/or peanut products at least once a month; (f) people available for all sessions; (g) people interested in participating; (h) people able to verbally communicate their observations regarding the product.15 Before being qualied, all panellists showed a perfect score in a taste sensitivity test and the ability to identify ve of seven commonly found food avours. The panellists were trained and calibrated in six training sessions for evaluating peanut pastes. Each training session lasted 3 h. The descriptive analysis test procedures described by Meilgaard et al.,16 Grosso and Resurreccion10 and Nepote et al.17 were used to train the panellists. A hybrid descriptive analysis method combining the Quantitative Descriptive Analysis (Tragon Corp., Redwood City, CA, USA) and Spectrum Analysis (Sensory Spectrum, Inc., Chatham, NJ, USA) methods was used by the panellists for evaluating samples. A 150 mm unstructured linear scale was employed.15 A list of denitions and a sheet with warm-up and reference intensity ratings (Table 1) were developed during the training sessions.10 The attribute denitions were based on the peanut lexicon.9 All samples were evaluated in partitioned booths under uorescent light at room temperature. Product samples (10 g) were placed in plastic cups with lids coded with three-digit random numbers. The nal lists of warm-up and reference intensity ratings and denitions (Table 1) were posted in the booths for all test sessions.10 The panellists were instructed to rst familiarise themselves with the reference standard intensities (Table 1) and then evaluate the sensory attributes of the peanut paste samples. A completely randomised block design was used for testing samples. The data were registered on paper ballots. Statistical analysis The experiment was replicated three times. The data were analysed using InfoStat Version 1.1 (Facultad de Ciencias Agropecuarias, Universidad Nacional de Cordoba, Argentina). Means and standard deviations were calculated. Analysis of variance and Duncan tests ( = 0.05) were used to detect signicant differences in sensory attributes and chemical analyses between sampling days. Pearson coefcients were used to calculate correlations between dependent variables. Second-order polynomial equations were used in the regression analyses.

MATERIALS AND METHODS

Peanut samples High-oleic peanuts (cv. Granoleico) and normal peanuts (cv. Tegua) were provided by Lorenzati, Ruetsch & Cia (Ticino, Cordoba, Argentina). Sound and mature seeds of blanched peanuts, size 4050 kernels oz1 , were selected. To prepare Granoleico (GO-P) and Tegua (T-P) peanut pastes, the blanched peanuts were heated in an oven (Model 600, Memert, Schwabach, Germany) at 140 C for 30 min to a medium roast, measured as an average Hunter colour lightness (L) value of 50 1.9 The roasted peanuts were then ground in a colloid mill (COLMIL Mod. AD 50 VR, Munro, Buenos Aires, Argentina). Storage conditions and sampling After preparation, GO-P and T-P samples were packaged in 350 g plastic jars and stored at 4 C (refrigeration chamber), 23 C (conditioner room) and 40 C (oven) to reproduce refrigeration, room temperature and accelerated storage conditions respectively.10 Samples were removed from storage after 0, 35, 70, 105, 140 and 175 days for chemical and descriptive analyses. Chemical analysis Oil was obtained from the peanut paste by cold pressing using a 20 ton press (HE-DU, Hermes I. Dupraz SRL, Cordoba, Argentina). The peanut oil was used for chemical analyses: fatty acid, peroxide, p-anisidine and conjugate diene determinations. Fatty acid methyl esters were prepared from peanut paste (GO-P and T-P) oils by transmethylation with a 30 g L1 solution of sulfuric acid in methanol as described by Grosso et al.11 The fatty acid methyl esters of total lipids were analysed in a Clarus 500 gas/liquid chromatograph (Perkin Elmer, Waltham, MA, USA) equipped with a ame ionisation detector. A CP-Wax 52 CB capillary column (30 m 0.25 mm 0.25 m; Varian, Lake Forest, CA, USA) was used. The column temperature was increased from 180 C (held for 1 min) to 230 C at 4 C min1 . The carrier gas was nitrogen at a ow rate of 1 mL min1 . The separated fatty acid methyl esters were identied by comparing their retention times with those of authentic samples purchased from Sigma Chemical Co. (St Louis, MO, USA). Quantitative fatty acid analysis was performed using heptadecanoic acid methyl ester (Sigma Chemical Co.) as internal standard. Iodine value (IV) was calculated using the formula11 IV = (%C18 : 1 0.8601) + (%C18 : 2 1.7321) + (%C20 : 1 0.7854) Peroxide value (PV) was evaluated by the AOAC12 standard method and expressed as milliequivalents active oxygen (meqO2 ) kg1 oil. Conjugated dienes (CD) and p-anisidine value (AV) were evaluated in an HP 8452A UVvisible diode array spectrophotometer (Hewlett Packard, Palo Alto, CA, USA) according to the IUPAC13 and COI14 standard methods respectively. Descriptive analysis A total of ten trained panellists (six female and four male), each with at least 6 years of experience in evaluating peanut products, participated in the descriptive analysis of peanut paste

RESULTS AND DISCUSSION

Chemical analyses GO-P had higher oleic acid content but lower linoleic acid content (785 and 46 g kg1 respectively) than T-P (458 and 333 g kg1 respectively). Similar results for these fatty acid contents were also found in others peanut products6,7,17 prepared with high-oleic and normal peanut lines from Argentina and other countries.18,19 The oleic/linoleic ratio was 17.06 in GO-P and 1.38 in T-P. IV was lower in GO-P (77) than in T-P (98). In addition, GO-P had higher eicosenoic acid content but lower palmitic acid content (25.1 and 57.7 g kg1 respectively) than T-P (17.2 and 99.3 g kg1 respectively). Other fatty acids did not differ signicantly between GO-P and T-P. The changes in PV, CD and AV during storage at 4, 23 and 40 C are shown in Fig. 1. Except for GO-P samples stored at 4 C, PV and CD increased signicantly with storage time. In each treatment,

2695

J Sci Food Agric 2010; 90: 26942699

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org

CG Riveros et al.

Table 1. Denitions of attributes, standard references and warm-up intensity ratings used in descriptive analysis of peanut pastes Attributea Appearance 1. Brown colour 2. Uneven colour Denition Reference Reference intensityb Warm-up intensityb,c

3. Glossiness Aroma 4. Roasted peanutty 5. Oxidised 6. Cardboard 7. Burnt 8. Raw/beany Taste 9. Sweetness

Intensity or strength of brown colour from light to dark brown Amount of speckles in peanut paste coming from residual particles of peanut skins Appearance associated with amount of light reected by product surface Aroma associated with medium-roasted peanuts Aroma associated with rancid fats and oils Aroma associated with wet cardboard Aroma associated with over-roasted peanuts Aroma associated with uncooked or raw peanuts Taste on tongue associated with sucrose solutions

Hazelnut and cocoa spread butterd Hazelnut and cocoa spread butterd Hazelnut and cocoa spread butterd Dry-roasted peanutse Rancid peanuts Moist cardboard 8 g of coffeef in 250 mL of distilled water Raw peanutsg

120 75

44 15

100

115

44 75 45 80 65

55 1 5 20 30

20 g kg1 sucrose solution 50 g kg1 sucrose solution 100 g kg1 sucrose solution 2 g kg1 NaCl solution 3.5 g kg1 NaCl solution 5 g kg1 NaCl solution 0.5 g kg1 citric acid solution 0.8 g kg1 citric acid solution 1.5 g kg1 citric acid solution 0.5 g kg1 caffeine solution 0.8 g kg1 caffeine solution 1.5 g kg1 caffeine solution

20 50 100 25 50 85 20 50 100 20 50 100 60

25

10. Saltiness

Taste on tongue associated with sodium chloride solutions

10

11. Sourness

Taste on tongue associated with acid agents such as citric acid solutions

12. Bitterness

Taste on tongue associated with bitter solutions such as caffeine

15

Mouthfeel 13. Astringency Textureh 14. Oiliness 15. Adhesiveness

Puckering or drying sensation on mouth or tongue surface Degree to which free oil is perceived in mouth Force required for removing material that adheres to palate during normal eating process Degree to which grains or granules are perceived in mouth

8 g of coffeef in 250 mL of distilled water Hazelnut and cocoa spread butterd Hazelnut and cocoa spread butterd Hazelnut and cocoa spread butterd

35

40 45

55 80

16. Graininess
a

25

35

Attributes listed in same order as perceived by panellists. Intensity ratings based on 150 mm unstructured linear scale. c Standard peanut paste (L = 50 1) prepared with roasted peanuts (cv. Runner, blanched). d Hazelnut and cocoa spread butter: Nutella (Ferrero SpA, Alba, Italy). e Dry-roasted peanuts: cv. Runner (JL SA, Ticino, Cordoba, Argentina). f Coffee: Nescaf Cl sico (Nestl Argentina SA, Buenos Aires, Argentina). e a e g Raw peanuts: size 4050 kernels oz1 (Lorenzati, Ruetsch & Cia, Ticino, Cordoba, Argentina). h The texture descriptors and their corresponding denitions were adapted from Muego-Ganasekharan and Resurreccion.22
b

AV did not show any marked increase during the storage period. However, GO-P samples stored at 4 and 23 C had signicantly lower AV than the other treatments. T-P had higher PV and showed signicant differences during storage after day 0 at 23 and 40 C and after day 35 at 4 C with respect to GO-P. T-P had slightly higher AV than GO-P.

Abegaz et al.20 reported that storage time was a signicant factor inuencing PV in model peanut butter confections stored at 21 C. Their results indicated that PV increased markedly within 4 weeks. In roasted peanuts in shell,21 roasted peanuts6 and fried-salted peanuts,7 PV and CD obtained at the end of the storage period were higher than those found in peanut

2696

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 26942699

Peanut pastes prepared with high-oleic peanuts

(a) 16 Peroxide value (meqO2 Kg1) 14 11 9 7 5 2 0 0 35 70 105 Days 140 175

www.soci.org

(a) 16 14 Oxidized intensity rating 11 9 7 5 2 0 0 35 70 Days (b) 14 105 140 175

(b) 3.0 Conjugated Dienes (E1; %232nm) 2.5 2.0 1.5 1.0 0.5 0.0 0 35 70 105 Days 140 175

Cardboard intensity rating

12 9 7 5 2 0 0 35 70 Days 105 140 175

(c)

(c) 60 Roasted peanutty intensity rating

p-Anisidine value

58 56 54 53 51 49 47 0 35 70 Days 105 140 175

35

70 Days

105

140

175

Figure 1. Changes in (a) peroxide value, (b) conjugated dienes and (c) panisidine value in Granoleico (GO-P) and Tegua (T-P) peanut pastes during storage: , GO-P at 4 C; , GO-P at 23 C; , GO-P at 40 C; , T-P at 4 C; , T-P at 23 C; , T-P at 40 C.

pastes (T-P and GO-P) in the present study. These results indicate that peanut paste is more resistant to lipid oxidation processes in comparison with other peanut products. After the roasting process, roasted peanuts release moisture, forming small cells that trap air inside. It is likely that the oxygen in these cells accelerates the lipid oxidation process. When roasted peanuts are nely ground to produce peanut paste, these air cells are destroyed, so peanut paste traps less air than roasted peanut kernels.

Figure 2. Changes in intensity rating of (a) oxidised, (b) cardboard and (c) roasted peanutty avours in Granoleico (GO-P) and Tegua (T-P) peanut pastes during storage: , GO-P at 4 C; , GO-P at 23 C; , GO-P at 40 C; , T-P at 4 C; , T-P at 23 C; , T-P at 40 C.

Descriptive analyses Sixteen sensory attributes were evaluated, but only three (roasted peanutty, oxidised and cardboard avours) showed signicant changes in their intensity ratings during storage (Fig. 2). The intensity of oxidised avour was higher in T-P except for samples stored at 4 C, in which there were no signicant differences

2697

J Sci Food Agric 2010; 90: 26942699

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org

CG Riveros et al.

Table 2. Coefcients and adjusted R2 values from regression equations of peroxide value (PV), conjugated dienes (CD), p-anisidine value (AV) and sensory attributes (oxidised, cardboard and roasted peanutty avours) in Granoleico (GO-P) and Tegua (T-P) peanut pastes at 4, 23 and 40 C Regression coefcientsa Sample GO-P Dependent variable PV Temperature ( C) 4 23 40 4 23 40 4 23 40 4 23 40 4 23 40 4 23 40 4 23 40 4 23 40 4 23 40 4 23 40 4 23 40 4 23 40 0 0.006779 0.112643 0.155357 0.912810 0.926633 0.928497 0.030389 0.042354 0.109461 2.618214 2.726071 2.848929 4.473929 4.132857 4.553214 57.351071 57.217857 56.917857 0.493571 0.182893 0.597857 1.134650 1.193871 1.253916 0.454742 0.449254 0.437830 2.990357 2.771429 2.761071 4.850000 4.793214 4.280000 57.209643 57.210714 57.143929 1 0.000620 0.013597 0.015676 0.002166 0.000322 0.00019 0.000437 0.002560 0.008736 0.025170 0.024197 0.050405 0.021881 0.039245 0.043870 0.045866 0.037104 0.060618 0.064076 0.114488 0.142239 0.006293 0.014912 0.021565 0.000771 0.002440 0.000962 0.005997 0.040857 0.057497 0.019253 0.025691 0.049404 0.090605 0.094533 0.107528 11 0.000017 0.000155 0.000198 0.000011 0.000015 0.000017 0.000001 0.000009 0.000035 0.000028 0.000033 0.000101 0.000032 0.000097 0.000121 0.000152 0.000047 0.000182 0.000042 0.000273 0.000390 0.000003 0.000036 0.000072 0.000005 0.000005 0.000004 0.000164 0.000023 0.000043 0.000055 0.000055 0.000019 0.000302 0.000265 0.000352 R2 0.955723 0.984468 0.968748 0.700428 0.933126 0.991511 0.949463 0.946595 0.852145 0.924024 0.979966 0.994754 0.943499 0.992439 0.951792 0.930223 0.967110 0.984936 0.979634 0.996934 0.981749 0.913256 0.982638 0.980057 0.069723 0.642913 0.905300 0.984725 0.957019 0.974840 0.993361 0.989509 0.953164 0.992060 0.996901 0.994401

CD

AV

Oxidised

Cardboard

Roasted peanutty

T-P

PV

CD

AV

Oxidised

Cardboard

Roasted peanutty

a Regression coefcients for the general regression equation Y = 0 + 1 X + 11 X 2 , where Y is the dependent variable (PV, CD, AV, oxidised, cardboard or roasted peanutty) and X is the independent variable (days of storage).

between GO-P and T-P. Signicant differences ( = 0.05) in oxidised avour intensity between GO-P and T-P were observed from day 70 at 23 and 40 C. The intensity ratings of cardboard avour were higher in T-P. Signicant differences ( = 0.05) in cardboard avour intensity between GO-P and T-P were detected from day 70 at 4, 23 and 40 C. Other authors reported similar results for normal peanut (cv. Florunner) paste stored at 40 C for 161 days22 and model (ideal) peanut butter confection stored at 21 C20,23 with respect to the results obtained for T-P in the present study. Roasted peanutty avour is the attribute used to characterise typical roasted peanut avour in peanut products.9 Bett and Boylston24 and Nepote et al.7 reported that roasted peanutty avour intensity decreased in roasted peanuts during storage. In the present study the intensity ratings of roasted peanutty

avour in GO-P and T-P decreased during storage. The intensities of this sensory attribute were lower in T-P. Signicant differences ( = 0.05) in roasted peanutty intensity between GO-P and T-P were detected from day 35 at all temperatures tested. Similar results were observed for peanut butter prepared with normal peanuts and stored at 25 C for 29 days25 in comparison with the results observed for T-P in the present study. Correlation and regression analyses The variables that changed during storage (PV, CD, AV and oxidised, cardboard and roasted peanutty avours) were correlated using Pearson coefcients. Positive correlations (higher than 0.60) were detected between PV, CD, AV and oxidised and cardboard avours, the highest being between PV and CD (0.98) and between oxidised and cardboard avours (0.92). Negative correlations of

2698

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 26942699

Peanut pastes prepared with high-oleic peanuts

www.soci.org
2 Frankel EN, Lipid Oxidation. The Oily Press, Bridgewater (2005). 3 Norden AAJ, Gorbet DW, Knauff DDA and Young CT, Variability in oil quality among peanut genotypes in the Florida Breeding Program. Peanut Sci 14:711 (1987). 4 Braddock JC, Sims CA and OKeefe SF, Flavour and oxidative stability of roasted high oleic acid peanuts. J Food Sci 60:489493 (1995). 5 Mugendi JB, Sims CA, Gorbet DW and OKeefe SF, Flavor stability of high oleic peanuts stored at low humidity. J Am Oil Chem Soc 75:2125 (1998). 6 Nepote V, Mestrallet MG, Accietto RH, Galizzi M and Grosso NR, Chemical and sensory stability of roasted high-oleic peanuts from Argentina. J Sci Food Agric 86:944952 (2006). 7 Nepote V, Mestrallet MG and Grosso NR, Oxidative stability in friedsalted peanuts elaborated with high-oleic and regular peanut from Argentina. Int J Food Sci Technol 41:900909 (2006). 8 Cabrini L, Barzanti V, Cipollone M, Florenini D, Grossa G, Tolomelli B, et al, Antioxidants and total perosyl radical-trapping ability of olive and seed oils. J Agric Food Chem 49:60266032 (2001). 9 Johnsen PB, Civille GV, Vercellotti JR, Sanders TH and Dus CA, Development of a lexicon for the description of peanut avor. J Sensory Stud 3:917 (1988). 10 Grosso NR and Resurreccion AVA, Predicting consumer acceptance ratings of cracker-coated and roasted peanuts from descriptive analysis and hexanal measurements. J Food Sci 67:15301537 (2002). 11 Grosso NR, Nepote V and Guzm n CA, Chemical composition of some a wild peanut species (Arachis L.) seeds. JAgricFoodChem 48:806809 (2000). 12 AOAC, Ofcial Methods of Analysis (16th edn). Association of Ofcial Analytical Chemists, Washington, DC (1995). 13 IUPAC, Method number 2.504. Determination of the p-anisidine value (p-A.V.), in Standard Methods for the Analysis of Oils, Fats and Derivatives (7th edn), ed. by Paquot C and Hautfenne A. Blackwell, Oxford, pp. 1347 (1987). 14 COI, Metodo de an lisis, prueba espectrofotom trica en el ultravioleta. a e Document COI/T, 20/Doc. No. 19/Rev. 1, International Olive Oil Council, Madrid (2001). 15 Plemmons LE and Resurreccion AVA, A warm-up sample improves reliability of responses in descriptive analysis. J Sensory Stud 13:359376 (1998). 16 Meilgaard M, Civille GV and Carr BT, Sensory Evaluation Techniques (2nd edn). CRC Press, Boca Raton, FL, pp. 135235 (1991). 17 Nepote V, Olmedo RH, Mestrallet MG and Grosso NR, A study of the relationships among consumer acceptance, oxidation chemical indicators and sensory attributes in high-oleic and normal peanuts. J Food Sci 74:18 (2009). 18 Ozcan M and Seven S, Physical and chemical analysis and fatty acid composition of peanut, peanut oil and peanut butter from COM and NC-7 cultivars. Grasas Aceites 54:1218 (2003). 19 Isleib TG, Pattee HE, Sanders TH, Hendrix KW and Dean LO, Compositional and sensory comparisons between normal and higholeic peanuts. J Agric Food Chem 54:17591763 (2006). 20 Abegaz EG, Kerr WL and Koehler PE, The role of moisture in avor changes of model peanut confections during storage. Lebensm Wiss Technol 37:215225 (2004). 21 Mozingo RW, OKeefe SF, Sanders TH and Hendrix KW, Improving shelf life of roasted and salted in shell peanuts using high oleic fatty acid chemistry. Peanut Sci 31:4045 (2004). 22 Muego-Ganasekharan KF and Resurreccion AVA, Physicochemical and sensory characteristics of peanut paste stored at different temperatures. J Food Sci 57:13851389 (1992). 23 Abegaz EG, Kerr WL and Koehler PE, Descriptive sensory analysis of stored model peanut confections with different sugar, moisture and antioxidant levels. Peanut Sci 33:5359 (2006). 24 Bett KL and Boylston TD, Effect of storage on roasted peanut quality. ACS Symp Ser 500:322343 (1992). 25 Felland SL and Koehler PE, Sensory, chemical and physical changes in increased water activity peanut butter products. J Food Qual 20:145156 (1997). 26 CAA, Captulo VII, Alimentos Grasos, Artculo No 531 (Res. 2012, 19.10.84). C digo Alimentario Argentino, Ley 18.248 (18.7.1969), De la o Canal y Asociados, Buenos Aires, pp. 164165 (1996).

Table 3. Shelf-life (days to reach a peroxide value of 10 meqO2 kg1 ) of Granoleico (GO-P) and Tegua (T-P) peanut pastes at 4, 23 and 40 C estimated using regression equation Shelf-life (days) Sample T-P GO-P 4 C 187 786 23 C 128 300 40 C 87 266

roasted peanutty avour with PV, CD, AV and oxidised and cardboard avours were found, the highest being with PV (0.87), oxidised avour (0.87) and CD (0.83). These negative correlations indicated that roasted peanutty avour decreased as lipid oxidation indicators increased during storage time. The regression equations of dependent variables (PV, CV, AV and oxidised, cardboard and roasted peanutty avours) for GO-P and T-P are presented in Table 2. The dependent variables showed R2 > 0.60 except for AV in T-P stored at 4 C. These regressions equations could be used to estimate the effect of storage time at 4, 23 and 40 C on peanut paste prepared with normal and high-oleic peanuts. In the Food Code of Argentina, 10 meqO2 kg1 is the maximum PV allowed for peanut products.26 A PV of 10 meqO2 kg1 may be useful as a reference value for the endpoint of peanut paste shelflife. Using the prediction equation, the shelf-lives of peanut pastes prepared with high-oleic (GO-P) and normal (T-P) peanuts are presented in Table 3. PV levels above 10 meqO2 kg1 were reached in T-P before GO-P during storage at different temperatures. Using the prediction equation, it was estimated that 10 meqO2 kg1 is reached on day 128 in T-P and on day 300 in GO-P when the peanut paste is stored at 23 C. The shelf-life of GO-P is approximately four (at 4 C), two (at 23 C) and three (at 40 C) times longer than that of T-P. These results indicate that high-oleic Granoleico kernels provide peanut paste (or peanut butter) with higher protection against lipid oxidation and improve the shelf-life of this product considerably.

CONCLUSIONS
This study provides an equation to estimate the shelf-life of peanut paste using chemical measurements of PV, AV and CD and intensity ratings of sensory attributes (oxidised, cardboard and roasted peanutty avours) from descriptive analysis. The results obtained show that peanut paste prepared with high-oleic peanuts (GO-P) has higher storage stability than peanut paste prepared with normal peanuts (T-P). This is due to the increased resistance of the former to lipid oxidation.

ACKNOWLEDGEMENTS
This work was supported by CONICET and SECYT-UNC.

REFERENCES
1 Riveros CG, Mestrallet MG, Nepote V and Grosso NR, Chemical composition and sensory analysis of peanut pastes elaborated with high-oleic and regular peanut from Argentina. Grasas Aceites 60:388395 (2009).

2699
wileyonlinelibrary.com/jsfa

J Sci Food Agric 2010; 90: 26942699

c 2010 Society of Chemical Industry

Research Article
Received: 17 April 2010 Revised: 25 June 2010 Accepted: 29 July 2010 Published online in Wiley Online Library: 25 August 2010

(wileyonlinelibrary.com) DOI 10.1002/jsfa.4143

Meta-analyses of effects of phytochemicals on digestibility and rumen fermentation characteristics associated with methanogenesis
Amlan K Patra
Abstract
BACKGROUND: A meta-analysis study was conducted to investigate the changes in rumen fermentation characteristics when methane inhibition by phytochemicals is employed. The whole database containing 185 treatment means from 36 published studies was divided into four subsets according to the major phytochemicals used in the studies, i.e. saponins, tannins, essential oils (EO) and organosulfur compounds (OS). RESULTS: Changes in protozoal numbers showed linear relationships with changes in methane production by saponins (R2 = 0.48), tannins (R2 = 0.30) and EO (R2 = 0.20) but not OS. Concentrations of total volatile fatty acids (VFA) and acetate did not show any relationship (P > 0.1) with changes in methane due to saponins. However, propionate production increased linearly with increasing inhibition of methane (R2 = 0.31), which resulted in a linear (R2 = 0.26) decrease in acetate/propionate ratio (A/P) with decreasing methane production. Concentrations of total VFA, acetate and propionate did not change with changes in methane production by tannins. However, A/P showed a signicant linear relationship (R2 = 0.27) with decreasing methane formation. Concentrations of total VFA (R2 = 0.44) and propionate (R2 = 0.15) changed linearly and positively with changes in methane production by EO. However, acetate production (R2 = 0.22) and A/P (R2 = 0.17) increased linearly with increasing inhibition of methane by EO. Changes in concentrations of total VFA (R2 = 0.60) and acetate (R2 = 0.35) decreased linearly while those of propionate increased linearly (R2 = 0.23) with increasing inhibition of methane by OS. Consequently, A/P decreased linearly (R2 = 0.30) with decreasing methane production by OS. Digestibilities of organic matter (OM) and neutral detergent bre were not affected by inhibition of methane production by saponins, EO and OS, but digestibility of OM decreased with decreasing methane production by tannins. CONCLUSION: The inhibition of methane production by phytochemicals results in changes in rumen fermentation that differ depending on the types of phytochemicals. c 2010 Society of Chemical Industry Keywords: phytochemicals; methane production; rumen fermentation; meta-analysis

INTRODUCTION
Methane is the second most anthropogenic greenhouse gas after carbon dioxide and thus contributes to global warming and climate change.1 Agriculture is responsible for about 47% of total anthropogenic methane emissions, of which 32% comes from enteric fermentation in livestock.1 Enteric methane is produced during the fermentation of feeds by methanogenic archaea, mostly in the rumen. Hence, in addition to greenhouse effects, methane production in ruminants represents a loss of about 215% of feed energy,2 which decreases the metabolisable energy content of feeds. From the calculation of energy balances3,4 as cited by Beauchemin et al.,5 it has been suggested that a 25% decrease in methane production in ruminants might result in an increase in milk production of 1 L day1 in high-yielding dairy cows or growth of 75 g day1 in growing cattle. Therefore the development of feeding strategies to decrease methane emissions in ruminants merits research attention for long-term mitigation of greenhouse

gas emissions into the atmosphere as well as short-term economic benets.6 The use of phytochemicals to inhibit methane production in the rumen could provide benets over chemical feed additives in relation to the presence of chemical residues in animalderived foods. Many phytochemicals, broadly saponins, essential oils (EO), tannins and organosulfur compounds (OS), have been shown to lower methane production in vitro and in vivo.7 10 These phytochemicals also affect rumen metabolism, e.g. volatile fatty

Correspondence to: Amlan K Patra, Department of Animal Nutrition, Faculty of Veterinary and Animal Sciences, West Bengal University of Animal and Fishery Sciences, 37 K. B. Sarani, Belgachia, Kolkata 700037, India. E-mail: patra amlan@yahoo.com Department of Animal Nutrition, Faculty of Veterinary and Animal Sciences, WestBengal UniversityofAnimal andFisherySciences,37K.B.Sarani,Belgachia, Kolkata 700037, India

2700

J Sci Food Agric 2010; 90: 27002708

www.soci.org

c 2010 Society of Chemical Industry

Effects of phytochemicals on rumen fermentation associated with methanogenesis acid (VFA) production pattern and digestibility of nutrients,7,9 and ultimately animal performance. Hence it is important to consider rumen fermentation characteristics when methane mitigation strategies involving these phytochemicals are employed. A metaanalysis of data obtained from previously published studies can explain the responses of rumen fermentation pattern when methane inhibition is targeted by different phytochemicals under a variety of experimental conditions.11,12 Therefore a meta-analysis was conducted to study rumen fermentation and digestibility in relation to methane inhibition by phytochemicals.

www.soci.org

oil, ropadiar, cinnamon oil, juniper berry oil, eucalyptus oils, Syzygium aromaticum, Foeniculum vulgare, anethol and cymene. OS included allicin, diallyl disulde, garlic oil, allyl isothiocyanate and horseradish oil. Statistical analysis Although the statistical analysis procedure used for meta-analysis of the database has been described elsewhere,13 a brief account is presented here. All statistical computations were carried out using the PROC MIXED, PROC REG and PROC CORR procedures of SAS Version 8.2.14 Data were analysed according to St-Pierre,15 taking into account the random effect of study, using PROC MIXED14 with the following model:
2 Yij = B0 + B1 Xij + B2 Xij + si + bi Xij + eij

MATERIALS AND METHODS

Description of database A meta-analysis of the effects of phytochemicals on rumen fermentation characteristics and digestibilities in association with methanogenesis was conducted by pooling and analysing data from previously published studies. The studies that were included in the database reported the effects of phytochemicals or plants/plant extracts rich in phytochemicals on methanogenesis and rumen fermentation and/or digestibilities of nutrients. These studies were published in peer-reviewed English language journals. Although several other studies have reported the effects of phytochemicals on methanogenesis, they were not included in the database, since the main active components were generally not identied. Overall, 36 published studies of 185 different treatment groups were considered (Appendix). The database included data on composition of feeds, digestibilities of organic matter (OM), neutral detergent bre (NDF), acid detergent bre (ADF) and crude protein (CP), protozoal numbers and rumen fermentation characteristics. Changes in rumen fermentation, digestibilities and protozoal numbers compared with control values were calculated (i.e. change = (value in phytochemical-supplemented treatment - control value)/control value)) to investigate how changes in methanogenesis by phytochemicals alter digestibilities, VFA and protozoal counts. Changes in methane production were calculated after expressing methane per unit of dry matter (DM) or OM digestibility. All variables were not available across all observations in the database. Hence the numbers of observations used for regression analyses varied between dietary and response variables depending on the regressor variables available. Since many of the studies did not report concentrations of phytochemicals, it was not possible to consider this factor in the database. Many studies also reported additional outcomes, but only the outcomes of interest were used here. Data reported in different units of measurement were transformed to the same units. Some records were incomplete or not reported uniformly, which necessitated calculations from the reported data. When a study did not report all possible outcomes and it was not possible to calculate from the reported data, missing variables were considered as missing data. The whole database was divided into four subsets according to the major phytochemicals used in the studies, i.e. saponins (n = 53), tannins (n = 48), EO (n = 44) and OS (n = 40). Most of the data were from in vitro measurements. Saponins and saponincontaining plants or extracts included Yucca saponaria, Sapindus saponaria, tea saponins, alfalfa saponins, Acacia concinna, Quillaja saponaria, Enterolobium cyclocarpum, Trigonella foenum-graecum, Pithecellobium saman and Sesbania sesban. Tannins included quebracho, Castanea sativa, Acacia mearnsii, myrabolam, chestnut, Lespedeza striata, Lespedeza cuneata, Biophytum petersianum, Acacia mangium, Psidium guajava, Jatropa curcas, Phaleria papuana, Quercus incana and Persea americana. EO included peppermint

where Yij is the expected outcome for the dependent variable Y observed at level j of the continuous variable X in study i; B0 is the overall intercept across all studies (xed effect); B1 and B2 are the overall linear and quadratic regression coefcients of Y on X respectively across all studies (xed effect); Xij is the synthetic datum value j of the continuous variable X in study i; si is the random effect of study i; bi is the random effect of study i on the regression coefcient of Y on X in study i; and eij is the unexplained residual error. The variable study was declared in the CLASS statement. The slope and intercept by study were included as random effects, as suggested by St-Pierre.15 In addition, an unstructured variancecovariance matrix (type = un) was performed in the random part of the model, as suggested by St-Pierre,15 to avoid the positive correlation between intercept and slope. When random covariance of the slope and intercept or random slope and intercept was not signicant (P > 0.1), a variance component (type = vc) of variancecovariance structure was performed or slope and intercept were not included in the model respectively.15 When the square term of slope was not signicant (P > 0.1), it was not included in the model. Because the data on changes in rumen fermentation parameters, protozoal counts and digestibility involve relative rates, they were subjected to logarithmic transformation before analysis to meet the criterion of homogeneity of variance. Since the values of some of these parameters were negative, data were coded by adding 1 prior to logarithmic transformation (i.e. log(x + 1), where x is the variable). For proper graphical representation of statistical results from the multidimensional space of studies in two-dimensional space, the Y observation was adjusted to take into account the random effect of study.15 The regression coefcient (R2 ) calculations also used the adjusted Y observation.15

RESULTS
Description of data set Means and standard deviations for selected variables in the four subsets are reported in Table 1. In the saponin subset, NDF concentrations of feeds ranged from 344 to 682 g kg1 DM with a mean value of 468 g kg1 DM, while CP contents ranged from 73 to 287 g kg1 DM with a mean value of 148.5 g kg1 DM. The effect of saponins on methane production varied from stimulatory (19%) to inhibitory (42%) with an average of 11% inhibition, while the growth of protozoa ranged from an increase by 60% to a decrease by 79% with an average of 28% lower in the

2701

J Sci Food Agric 2010; 90: 27002708

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org

AK Patra

Table 1. Descriptive statistics of dietary composition and percentage changes in rumen fermentation and digestibility variables relative to control values (i.e. % change versus control) as affected by different phytochemicals Saponins Parameter OM content (g kg1 ) NDF content (g kg1 ) ADF content (g kg1 ) CP content (g kg1 ) Methane Protozoa TVFA Propionate Acetate A/P OM digestibility NDF digestibility ADF digestibility CP digestibility Mean 917.6 467.9 273.4 148.5 10.96 28.27 1.24 8.03 0.60 8.61 4.34 10.56 5.27 1.32 SD 21.03 97.87 58.56 40.75 12.96 34.52 8.11 11.03 5.02 11.62 6.52 17.57 9.15 4.23 Mean 927.9 490.8 254.5 132.7 19.13 14.61 5.75 4.47 4.01 0.13 2.63 2.95 0.62 3.69 Tannins SD 33.44 112.65 50.27 23.39 18.57 28.14 5.70 6.46 5.24 8.77 5.10 9.63 9.32 6.79 Essential oils Mean 907.2 609.9 335.6 119.5 35.35 34.92 1.06 3.29 1.30 3.38 8.46 23.16 ND ND SD 8.74 85.07 29.56 11.21 32.41 29.86 10.91 9.07 5.66 14.10 9.19 9.56 ND ND Organosulfurs Mean 914.4 434.3 292.3 154.9 23.09 14.58 3.00 6.64 1.84 7.67 2.01 3.77 ND ND SD 8.44 110.19 71.50 27.39 30.82 31.59 10.48 9.94 5.82 12.53 4.95 9.22 ND ND

OM, organic matter; NDF, neutral detergent bre; ADF, acid detergent bre; CP, crude protein; TVFA, total volatile fatty acids; A/P, acetate/propionate ratio; SD, standard deviation; ND, not determined owing to insufcient data.

saponin database. Acetate concentrations were lower by 12% to higher by 18% compared with controls. Changes in propionate concentration ranged from 8 to 39% compared with controls, which resulted in a mean decrease in acetate/propionate ratio (A/P) by 8.6%. Overall, digestibility of OM (4.3%), especially NDF (10.6%), was reduced by saponins, whereas digestibility of CP (1.3%) was hardly affected. In the tannin subset, NDF and CP contents ranged from 401 to 711 g kg1 DM and from 97 to 229 g kg1 DM respectively, suggesting that roughage-based medium-quality feeds were used in the studies. Like saponins, tannins affected protozoal numbers (69 to 53%) and methane production (50 to 32%) both positively and negatively, with overall decreases in protozoal counts and methane production. In this data set, total VFA (16 to 5%), acetate (14 to 6%), propionate (21 to 7%) and OM (16 to 11%) and NDF (20 to 15%) digestibilities were little affected in general. In the EO data set, there was substantial inhibition of methane production (90 to 21%) and protozoal growth (88 to 18%) compared with controls. Digestibilities of OM (28 to 5%) and NDF (33 to 10%) were also reduced considerably by EO. Although the effects of EO on rumen VFA varied noticeably, overall effects were minimal. In the OS data set, methane production decreased by 90 to 9%. Overall, protozoal growth was stimulated (15%) by OS. While total VFA (21 to 19%) and acetate (12 to 6%) concentrations were only slightly affected, propionate (8 to 32%) concentrations increased considerably, resulting in signicant reductions in A/P (33 to 10%). The minimum and maximum values of OM and NDF digestibility changes compared with control values were 17 and 3% and 27 and 5% respectively. The results presented here should be used with caution when input variables are outside the range of variables in the data set. Correlations Pearson correlation matrices for the data sets are shown in Table 2. In the saponin data set, changes in methane production correlated positively (P < 0.1) with changes in protozoal numbers and OM digestibility. Other parameters did not show signicant

Table 2. Simple correlation coefcients (P < 0.1) between changes in methane production and changes in rumen fermentation characteristics and digestibility as affected by different phytochemicals Parameter Protozoa TVFA Propionate Acetate OM digestibility NDF digestibility ADF digestibility CP digestibility Saponins Tannins Essential oils Organosulfurs 0.64 NS NS NS 0.29 NS NS NS 0.65 0.28 NS 0.55 0.39 NS NS NS 0.48 0.38 0.43 0.55 NS NS ND ND NS NS 0.32 0.61 0.58 0.67 ND ND

TVFA, total volatile fatty acids; OM, organic matter; NDF, neutral detergent bre; ADF, acid detergent bre; CP, crude protein; NS, not signicant; ND, not determined owing to insufcient data.

correlations with methane production. In the tannin data set, there were also positive correlations (P < 0.1) between changes in methane production and changes in protozoal counts, total VFA concentration, acetate concentration and OM digestibility. In the EO data set, changes in protozoal populations and concentrations of total VFA and propionate showed signicant (P < 0.1) positive correlations with changes in methane production. However, changes in acetate concentration correlated (P < 0.1) negatively with changes in methane production. In the OS data set, protozoal numbers and total VFA concentration did not show any correlation (P > 0.1) with changes in methane production. Changes in methane production correlated negatively with changes in propionate concentration but positively with changes in acetate concentration and digestibilities of OM and NDF. Effects of saponins on protozoa, VFA and digestibility Changes in protozoal numbers showed a linear relationship (R2 = 0.48) with changes in methane production caused by

2702

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 27002708

Effects of phytochemicals on rumen fermentation associated with methanogenesis

0.15 Changes in methane (log transformed) 0.1 0.05 0 -0.05 -0.1 -0.15 -0.4

www.soci.org

weak (R2 = 0.20) (Fig. 3): log(change in methane + 1) = 0.0628 (SE = 0.0374, P = 0.02) + log(change in protozoal numbers + 1) 0.513 (SE = 0.179, P = 0.08) Concentrations of total VFA (P = 0.06, R2 = 0.44) and propionate (P = 0.09, R2 = 0.15) changed linearly and positively with positive changes in methane production (Table 3). However, acetate concentrations decreased linearly (P = 0.01, R2 = 0.22) with positive in methane production. A/P increased linearly (P = 0.02) with increasing inhibition of methane by EO, though the relationship was very weak (R2 = 0.17). Changes in digestibilities of OM and NDF were not affected by inhibition of methane by EO. Effects of organosulfur compounds on protozoa, VFA and digestibility Methane inhibition by OS, unlike that by other phytochemicals, was not associated with protozoal populations. Changes in concentrations of total VFA (P = 0.07, R2 = 0.60) and acetate (P = 0.02, R2 = 0.35) decreased linearly while those of propionate increased linearly (P = 0.06, R2 = 0.23) with increasing inhibition of methane by OS (Table 3). Hence changes in A/P decreased linearly (P = 0.03, R2 = 0.30) with decreasing methane production by OS. Digestibilities of OM and NDF were not affected by inhibition of methane production by OS. Effects of dietary composition on methane inhibition by phytochemicals In the whole database, CP content did not affect methane inhibition by phytochemicals. However, suppression of methane by saponins decreased with increasing CP content in diets, but the degree of determination was low (R2 = 0.23): log(change in methane + 1) = 0.0875(SE = 0.0208, P < 0.01) + CP content (%) 0.00392 (SE = 0.00194, P = 0.06) For other types of phytochemicals, no signicant relationship was evident. The above relationship signies that methane inhibition by saponins could be higher for low-protein diets than for high-protein diets. In the whole database, methane production was inhibited by phytochemicals increasingly with greater concentrations of NDF in diets (R2 = 0.15): log(change in methane + 1) = 0.0173 (SE = 0.0431, P = 0.65) NDF content (%) 0.00143 (SE = 0.00301, P = 0.09) Data sets for individual types of phytochemicals did not reveal any relationship between NDF content in diets and methane depression by phytochemicals. This, in general, suggests that the effect of phytochemicals on methane inhibition could be greater for roughage-based diets than for concentrate-based diets.

-0.2 0 0.2 0.4 Changes in protozoal numbers (log transformed)

0.6

Figure 1. Effect of changes in protozoal numbers on changes in methane production by saponins. The relationship was as follows: log(change in methane +1) = 0.0253 (SE = 0.00875, P < 0.01) + log(change in protozoal numbers +1) 0.0.184 (SE = 0.0802, P = 0.04), R2 = 0.48.

saponins (Fig. 1): log(change in methane + 1) = 0.0253 (SE = 0.00875, P < 0.01) + log(change in protozoal numbers + 1) 0.184 (SE = 0.0802, P = 0.04) Concentrations of total VFA and acetate did not show any relationship with methane due to saponins (Table 3). However, propionate production increased linearly (P = 0.10) with increasing inhibition of methane, though the relationship was weak (R2 = 0.31). This probably resulted in the linear decrease in A/P (P = 0.07, R2 = 0.26) with decreasing methane production. Digestibilities of all nutrients did not change with decreasing methane production by saponins. Effects of tannins on protozoa, VFA and digestibility Inhibition of methane production by tannins, like that by saponins, changed linearly (R2 = 0.30) with decreasing numbers of protozoa (Fig. 2): log(change in methane + 1) = 0.0609 (SE = 0.0113, P < 0.01) + log(change in protozoal numbers + 1) 0.327 (SE = 0.155, P = 0.08) Concentrations of total VFA, acetate and propionate did not change with changes in methane production by tannins (Table 3). However, A/P showed a signicant linear relationship (P = 0.09, R2 = 0.27) with increasing inhibition of methane production. Digestibilities of OM (P = 0.08, R2 = 0.19) and NDF (P = 0.12, R2 = 0.23) decreased linearly with decreasing methane production by tannins, but digestibility of CP did not show any relationship with methane inhibition by tannins. Effects of essential oils on protozoa, VFA and digestibility For EO, changes in methane production were linearly associated with changes in protozoal numbers, though the relationship was

DISCUSSION
In order to take advantage of the effects of phytochemicals on methane inhibition as well as to design future methane mitigation technologies, it is necessary to understand the effects of phytochemicals on rumen metabolism. It appears that different

2703

J Sci Food Agric 2010; 90: 27002708

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org

AK Patra

Table 3. Effects (equations for linear regressionsa) of changes in methane production on changes in rumen fermentation parameters and digestibility as impacted by different phytochemicals Parameter Saponins TVFA Acetate Propionate A/P OM digestibility NDF digestibility Tannins TVFA Acetate Propionate A/P OM digestibility NDF digestibility Essential oils TVFA Acetate Propionate A/P OM digestibility NDF digestibility Organosulfurs TVFA Acetate Propionate A/P OM digestibility NDF digestibility Intercept SEIntercept P Slope SESlope P RMSE Adj. R2

0.0068 0.00285 0.0175 0.0178 0.00636 0.0321 0.0143 0.00592 0.0178 0.00337 0.00581 0.0157 0.00602 0.00656 0.00059 0.0740 0.0242 0.0529 0.0127 0.00499 0.00587 0.0149 0.00158 0.00319

0.0104 0.00422 0.00935 0.00761 0.00581 0.0141 0.00606 0.00461 0.00620 0.00976 0.00540 0.0167 0.0164 0.00258 0.00956 0.0135 0.00992 0.0186 0.0151 0.00835 0.0210 0.0176 0.00230 0.0151

0.52 0.51 0.08 0.03 0.29 0.04 0.04 0.23 0.02 0.73 0.31 0.38 0.72 0.04 0.95 0.58 0.02 0.04 0.43 0.57 0.72 0.42 0.50 0.75

0.141 0.0340 0.294 0.262 0.0857 0.0251 0.135 0.0629 0.00946 0.179 0.121 0.233 0.219 0.0526 0.125 0.265 0.0511 0.211 0.368 0.0771 0.193 0.168 0.0142 0.273

0.101 0.0843 0.164 0.127 0.0121 0.211 0.133 0.0857 0.0678 0.0909 0.0519 0.152 0.119 0.0193 0.0714 0.104 0.0672 0.146 0.175 0.0319 0.0981 0.0742 0.0793 0.197

0.17 0.69 0.10 0.07 0.50 0.90 0.36 0.50 0.89 0.09 0.08 0.12 0.06 0.01 0.09 0.02 0.45 0.22 0.07 0.02 0.06 0.03 0.65 0.28

0.0132 0.0125 0.0311 0.0222 0.0128 0.0443 0.0166 0.0122 0.0258 0.0211 0.0177 0.0305 0.0164 0.00940 0.0348 0.0501 0.0362 0.0424 0.00792 0.0103 0.0333 0.0247 0.00872 0.00828

0.14 0 0.31 0.26 0.08 0 0.24 0.08 0 0.27 0.19 0.23 0.44 0.22 0.15 0.17 0.0 0.10 0.60 0.35 0.23 0.30 0.10 0.32

TVFA, total volatile fatty acids; A/P, acetate/propionate ratio; OM, organic matter; NDF, neutral detergent bre; SE, standard error; RMSE, root mean square error. a No signicant quadratic relationship was noted between changes in methane and changes in volatile fatty acid concentrations and digestibility. Data on changes in methane production, rumen fermentation and digestibility are logarithmically transformed (log(x + 1), where x is the variable).

0.15 Changes in methane (log transformed) 0.1 0.05 0 -0.05 -0.1 -0.15 -0.2 -0.4 Changes in methane (log transformed)

0.1 0.05 0 -0.05 -0.1 -0.15 -0.2 -0.25 -0.3 -0.4 -0.2 -0.1 0 0.1 0.2

-0.2 0 0.2 Changes in protozoal numbers (log transformed)

0.4

Changes in protozoal numbers (log transformed)

Figure 3. Effect of changes in protozoal numbers on changes in methane production by essential oils. The relationship was as follows: log(change in methane +1) = 0.0628 (SE = 0.0374, P = 0.02) + log(change in protozoal numbers +1) 0.513 (SE = 0.179, P = 0.08), R2 = 0.20.

Figure 2. Effect of changes in protozoal numbers on changes in methane production by tannins. The relationship was as follows: log(change in methane +1) = 0.0609 (SE = 0.0113, P < 0.01) + log(change in protozoal numbers +1) 0.327 (SE = 0.155, P = 0.08), R2 = 0.30.

2704

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 27002708

Effects of phytochemicals on rumen fermentation associated with methanogenesis types of phytochemicals such as saponins, tannins, EO and OS have different effects on rumen fermentation and microorganisms. Inhibition of methane production by saponins was associated with an increase in propionate concentration and thus resulted in a decrease in A/P. For tannins, a decrease in methane emission did not relate to changes in acetate and propionate concentrations, although A/P decreased quadratically with inhibition of methane. For EO, acetate and propionate concentrations relative to controls decreased quadratically and linearly respectively with increasing inhibition of methane. Unlike saponins, tannins and OS, A/P increased with decreasing methane production by EO. OS-mediated methane inhibition resulted in a linear decrease in acetate concentration, a linear increase in propionate concentration and a linear decrease in A/P, which was very similar to decreases in methane by direct inhibition of methanogens in the rumen. According to the results of this meta-analysis, it appears that different types of phytochemicals inhibit methane by different modes, resulting in different patterns of rumen fermentation. A decrease in methane production by phytochemicals may be confounded with several factors such as suppression of protozoa, archaea and hydrogen-producing microbial populations and decreased bre digestion in the rumen.6,16 From the relationship between changes in methane production and protozoal counts for different phytochemicals, it is apparent that each 1% suppression of protozoal numbers accounted for about 0.17, 0.29 and 0.45% inhibition of methane by saponins, tannins and EO respectively. Thus methane inhibition by saponins perhaps results predominantly from decreased protozoal populations. Methanogens associated with protozoa account for decreased methane production of about 925%17 or as much as 37%.18 Sterol-binding capability of saponins19 has been implicated in the destruction of protozoal cell membranes.9,16 A decrease in the activities of methane-producing genes or rate of methane production per methanogenic cell by saponins has also been suggested,20,21 which might increase propionate production as a result of rechannelling of hydrogen from methane to propionate.22 Hence a decrease in protozoal numbers by saponins may result in an increase in propionate concentration and a decrease in A/P. Besides, saponins sometimes stimulate the growth of Selenomonas ruminantium,23 which predominantly produces propionate from succinate metabolism. Suppression of methane by tannins could entail reductions in numbers of protozoa and methanogens depending on the chemical structure of tannins and methanogenic species.6,24 Besides their direct effect on methanogens, tannins exert antimicrobial action on microbial growth, including cellulolytic bacteria and fungi,16 which could decrease hydrogen availability and thus lower methanogenesis25 to some extent. Therefore propionate and acetate production was not signicantly affected by tannin-induced suppression of methane, but A/P decreased with increasing methane inhibition. In contrast to saponins, EO-mediated inhibition of methane could primarily involve suppression of methanogens7,9,16 and hydrogen-producing micro-organisms such as Lachnospira multiparus,Ruminococcusalbus and Ruminococcusavifaciens, including inhibition of protozoal growth.7,16 Because of less formation of hydrogen, propionate concentrations decrease, resulting in an increase in A/P. Besides, EO such as thymol may decrease propionate formation by inhibiting the growth of S. ruminantium.26 These characteristic changes in rumen fermentation by EO are

www.soci.org

not nutritionally favourable for ruminant animals in the context of energy utilisation. In an in vitro study with mixed rumen microbial populations, Evans and Martin26 also reported that thymol (a component of essential oils) decreased concentrations of acetate, propionate, methane and hydrogen and increased A/P. OS might specically inhibit methanogenic archaea.7,27 Via analysis by real-time polymerase chain reaction, McAllister and Newbold22 observed that allicin at 20 mg L1 decreased methanogen numbers in Rusitec, but not at 2 mg L1 . It has been suggested that OS found in garlic oil perhaps directly inhibit rumen methanogenic archaea through inhibition of the enzyme 3-hydroxy-3-methyl-glutaryl coenzyme A (HMG-CoA) reductase.27 Methanogenic archaea have unique membrane lipids that contain glycerol joined by ether linkages to long-chain isoprenoid alcohols, which is not present in other rumen microorganisms.28 The synthesis of isoprenoid units in methanogenic archaea is catalysed by HMG-CoA reductase.27 Inhibition of methane production usually increases the concentration of hydrogen, which is rechannelled to other hydrogen sinks such as propionate, resulting in increased concentration of propionate.22 Conversely, production of acetate in the rumen results in large quantities of hydrogen and depends on the availability of reducing equivalents27 such as NAD+ . The high partial pressure of hydrogen and high NADH/NAD+ ratio in the rumen due to inhibition of methanogenesis may result in a decrease in acetate production.27,29 The characteristics of VFA patterns noted due to suppression of methane by OS suggest a direct effect on methanogens without much effect on other rumen microbial populations. A strategy for suppressing methane inhibition from ruminants would only be practical if it has no hostile effects on ruminal dynamics. Among the phytochemicals, OS hold promise to decrease enteric methane emissions, as they do not exert much adverse effect on ruminal fermentation and digestibility compared with other phytochemicals.

CONCLUSIONS
From the patterns of changes in protozoal numbers and VFA prole by different groups of phytochemicals, it appears that methane inhibition by saponins may predominantly involve suppression of protozoal growth, which is responsible for greater propionate production and lower A/P. However, the decrease in methane production by tannins may be due to inhibition of protozoa, methanogens and, to a lesser extent, hydrogen-producing microbial population. For EO, depression of methane production may primarily entail inhibition of methanogens, hydrogenproducing rumen micro-organisms and, to a lesser extent, protozoa, resulting in decreased concentrations of propionate and acetate and increased A/P. OS may have a direct antimethanogenic effect, which results in decreased acetate, increased propionate and decreased A/P. The different phytochemicals exert their characteristic effects on rumen fermentation in association with the inhibition of methanogenesis.

ACKNOWLEDGEMENT
The research grant provided by the Indian Council of Agricultural Research, New Delhi is gratefully acknowledged.

2705
wileyonlinelibrary.com/jsfa

J Sci Food Agric 2010; 90: 27002708

c 2010 Society of Chemical Industry

www.soci.org

AK Patra

APPENDIX
Table A1. List of previously published studies included in database for analysing changes in rumen fermentation characteristics associated with methanogenesis Phytochemicals Saponins Study Goel et Hess et al.2 Hess et al.3 Hess et al.4 Holtshausen et al.5 Holtshausen et al.5 Hu et al.6 Hu et al.7 Klita et al.8 Lila et al.9 Patra et al.10 Pen et al.11 Santoso et al.12 Sliwinski et al.13 Sliwinski et al.14 Wang et al.15 Animut et al.16 Beauchemin et al.17 Bhatta et al.18 Bodas et al.19 Carrula et al.20 Hariadi and Santoso21 Hess et al.4 Patra et al.10 Patra et al.22 Patra et al.23 Puchala et al.24 Sliwinski et al.13 Sliwinski et al.14 Agarwal et al.25 Beauchemin and McGinn26 Chaves et al.27 Kumar et al.28 Patra et al.29 Sallam et al.30 Tatsuoka et al.31 Wang et al.15 Busquet et al.32 Chaves et al.27 Garcia-Gonzalez et al.33 Kamel et al.34 Lila et al.35 Mohammed et al.36 Patra et al.23 Patra et al.29 Tatsuoka et al.31 al.1 Diet Hay; hay/concentrate (32 : 68) Forage/concentrate (2 : 1) Meadow grass/Arachis pintoi hay/barley straw (56 : 22:11) Brachiaria grass (66100%) Barley silage/concentrate (51 : 49) Barley silage/concentrate (51 : 49) Grass hay/corn (1 : 1) Grass hay/corn (1 : 1) Grass hay Corn starch; potato starch; hay/concentrate (3 : 2) Wheat straw/concentrate (1 : 1) Ryegrass hay/concentrate (3 : 2) Orchard grass silage/concentrate (7 : 3) Hay/barley-based concentrate (1 : 1) Grass silage and hay/barley (77 : 23) Mixed hay/concentrate (3 : 1) Lespedeza forage/sorghum-sudan grass (1 : 2, 2 : 1 and 3 : 0) Barley silage 69.5%, barley grain 21.5% Timothy hay/concentrate (65 : 35) Lucerne/grass hay/barley grain (5 : 4:1) Ryegrass/lucerne (1 : 1) Elephant grass Brachiaria grass (66100%) Wheat straw/concentrate (1 : 1) Wheat straw/concentrate (1 : 1) Wheat straw/concentrate (1 : 1) Lespedeza cuneata pasture Hay/barley-based concentrate (1 : 1) Grass silage and hay/barley (77 : 23) Wheat straw/concentrate (1 : 1) Barley silage/concentrate (3 : 1) Soluble starch Wheat straw/concentrate (1 : 1) Wheat straw/concentrate (1 : 1) Roughage/concentrate (1 : 1) Glucose/cellobiose (1 : 1) Mixed hay/concentrate (3 : 1) Lucerne hay/concentrate (1 : 1) Soluble starch Lucerne/grass hay/barley (5 : 2:3) Lucerne hay/concentrate (1 : 1) Corn starch; potato starch; hay/concentrate (3 : 2) Sudan grass/concentrate (3 : 2) Wheat straw/concentrate (1 : 1) Wheat straw/concentrate (1 : 1) Soluble sugar METH Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y PROT Y Y Y Y N Y Y Y N Y Y Y N Y Y N Y N Y N Y Y Y Y Y Y N Y Y Y N Y Y Y Y N N N Y N N Y Y Y Y N VFA N Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y N Y Y Y Y Y Y Y Y Y Y Y DIG N Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y Y

Tannins

Essential oils

Organosulfurs

METH, methane; PROT, protozoa; VFA, volatile fatty acids; DIG, digestibility; Y, reported; N, not reported. 1 Goel G, Makkar HPS and Becker K, Effect of Sesbania sesban and Carduus pycnocephalus leaves and fenugreek (Trigonella foenum-graecum L.) seeds and their extracts on partitioning of nutrient from roughage and concentrate based feeds to methane. Anim Feed Sci Technol 147:7289 (2008). 2 Hess HD, Beuret RA, Lotscher M, Hindrichsen IK, Machmuller A, Carulla JE, et al., Ruminal fermentation, methanogenesis and nitrogen utilization of sheep receiving tropical grass hayconcentrate diets offered with Sapindus saponaria fruits and Cratylia argentea foliage. Anim Sci 79:177189 (2004). 3 Hess HD, Kreuzer M, Diaz TE, Lascano CE, Carulla JE and Solvia CR, Saponin rich tropical fruits affect fermentation and methanogenesis in faunated and defaunated uid. Anim Feed Sci Technol 109:7994 (2003). 4 Hess HD, Monsalve LM, Lascano CE, Carulla JE, Diaz TE and Kreuzer M, Supplementation of a tropical grass diet with forage legumes and Sapindus saponaria fruits: effects on in vitro ruminal nitrogen turnover and methanogenesis. Aust J Agric Res 54:703713 (2003).

2706
wileyonlinelibrary.com/jsfa
c 2010 Society of Chemical Industry J Sci Food Agric 2010; 90: 27002708

Effects of phytochemicals on rumen fermentation associated with methanogenesis

www.soci.org

Table A1. (Continued) Holtshausen L, Chaves AV, Beauchemin KA, McGinn SM, McAllister TA, Odongo NE, et al., Feeding saponin-containing Yucca schidigera and Quillaja saponaria to decrease enteric methane production in dairy cows. J Dairy Sci 92:28092821 (2009). 6 Hu W, Wu Y, Liu J, Guo Y and Ye J, Tea saponins affect in vitro fermentation and methanogenesis in faunated and defaunated rumen uid. J Zhejiang Univ Sci B 6:787792 (2005). 7 Hu W, Liu J, Wu Y, Guo Y and Ye J, Effects of tea saponins on in vitro ruminal fermentation and growth performance in growing Boer goat. Arch Anim Nutr 60:8997 (2006). 8 Klita PT, Mathison GW, Fenton TW and Hardin RT, Effects of alfalfa root saponins on digestive function in sheep. J Anim Sci 74:11441156 (1996). 9 Lila ZA, Mohammed N, Kanda S, Kamada T and Itabashi H, Effect of sarsaponin on rumen fermentation with particular reference to methane production in vitro. J Dairy Sci 86:33303336 (2003). 10 Patra AK, Kamra DN and Agarwal N, Effect of plant extracts on in vitro methanogenesis, enzyme activities and fermentation of feed in rumen liquor of buffalo. Anim Feed Sci Technol 128:276291 (2006). 11 Pen B, Takaura K, Yamaguchia S, Asa R and Takahashi J, Effects of Yucca schidigera and Quillaja saponaria with or without -1,4 galactooligosaccharides on ruminal fermentation, methane production and nitrogen utilization in sheep. Anim Feed Sci Technol 138:7588 (2007). 12 Santoso B, Mwenya B, Sar C, Gamo Y, Kobayashi T, Morikawa R, et al., Effects of supplementing galacto-oligosaccharides, Yucca schidigera and nisin on rumen methanogenesis, nitrogen and energy metabolism in sheep. Livest Prod Sci 91:209217 (2004). 13 Sliwinski BJ, Kreuzer M, Wettstein HR and Machmuller A, Rumen fermentation and nitrogen balance of lambs fed diets containing plant extracts rich in tannins and saponins and associated emissions of nitrogen and methane. Arch Anim Nutr 56:379392 (2002). 14 Sliwinski BJ, Solvia CR, Machmuller A and Kreuzer M, Efcacy of plant extracts rich in secondary constituents to modify rumen fermentation. Anim Feed Sci Technol 101:101114 (2002). 15 Wang CJ, Wang SP and Zhou H, Inuences of avomycin, ropadiar, and saponin on nutrient digestibility, rumen fermentation, and methane emission from sheep. Anim Feed Sci Technol 148:157166 (2009). 16 Animut G, Goetsch AL, Puchala R, Patra AK, Sahlu T, Varel VH, et al., Methane emission by goats consuming diets with different levels of condensed tannins from lespedeza. Anim Feed Sci Technol 144:212227 (2008). 17 Beauchemin KA, McGinn SM, Martinez TF and McAllister TA, Use of condensed tannin extract from quebracho trees to reduce methane emissions. J Anim Sci 85:19901996 (2007). 18 Bhatta R, Uyeno Y, Tajima K, Takenaka A, Yabumoto Y, Nonaka I, et al., Difference in the nature of tannins on in vitro ruminal methane and volatile fatty acid production and on methanogenic archaea and protozoal populations. J Dairy Sci 92:55125522 (2009). 19 Bodas R, Lopez S, Fernandez M, Garcia-Gonzalez R, Rodrguez AB, Wallace RJ, et al., In vitro screening of the potential of numerous plant species as antimethanogenic feed additives for ruminants. Anim Feed Sci Technol 145:245258 (2008). 20 Carulla JE, Kreuzer M, Machmuller A and Hess HD, Supplementation of Acacia mearnsii tannins decreases methanogenesis and urinary nitrogen in forage-fed sheep. Aust J Agric Res 56:961970 (2005). 21 Hariadi BT and Santoso B, Evaluation of tropical plants containing tannin on in vitro methanogenesis and fermentation parameters using rumen uid. J Sci Food Agric 90:456461 (2010). 22 Patra AK, Kamra DN and Agarwal N, Effect of extracts of leaves on rumen methanogenesis, enzyme activities and fermentation in in vitro gas production test. Indian J Anim Sci 78:9196 (2008). 23 Patra AK, Kamra DN and Agarwal N, Effect of plants containing secondary metabolites on in vitro methanogenesis, enzyme prole and fermentation of feed with rumen liquor of buffalo. Anim Nutr Feed Technol 6:203213 (2006). 24 Puchala R, Min BR, Goetsch AL and Sahlu T, The effect of a condensed tannin-containing forage on methane emission by goats. J Anim Sci 83:182186 (2005). 25 Agarwal N, Shekhar C, Kumar R, Chaudhary LC and Kamra DN, Effect of peppermint (Mentha piperita) oil on in vitro methanogenesis and fermentation of feed with buffalo rumen liquor. Anim Feed Sci Technol 148:321327 (2009). 26 Beauchemin KA and McGinn SM, Methane emissions from beef cattle: effects of fumaric acid, essential oil, and canola oil. J Anim Sci 84:14891496 (2006). 27 Chaves AV, He ML, Yang WZ, Hristov AN, McAllister TA and Benchaar C, Effects of essential oils on proteolytic, deaminative and methanogenic activities of mixed ruminal bacteria. Can J Anim Sci 89:97104 (2008). 28 Kumar R, Kamra DN, Agrawal N and Chaudhary LC, Effect of eucalyptus (Eucalyptus globulus) oil on in vitro methanogenesis and fermentation of feed with buffalo rumen liquor. Anim Nutr Feed Technol 9:237243 (2009). 29 Patra AK, Kamra DN and Agarwal N, Effects of extracts of spices on rumen methanogenesis, enzyme activities and fermentation of feeds in vitro. J Sci Food Agric 90:511520 (2010). 30 Sallam SMA, Bueno ICS, Brigide P, Godoy PB, Vitti DMSS and Abdalla AL, Efcacy of eucalyptus oil on in vitro rumen fermentation and methane production. Options Mediterraneennes 85:267272 (2009). 31 Tatsuoka N, Hara K, Mikuni K, Hara K, Hashimoto H and Itabashi H, Effects of the essential oil cyclodextrin complexes on ruminal methane production in vitro. Anim Sci J 79:6875 (2008). 32 Busquet M, Calsamiglia S, Ferret A, Carro MD and Kamel C, Effect of garlic oil and four of its compounds on rumen microbial fermentation. J Dairy Sci 88:43934404 (2005). 33 Garcia-Gonzalez R, Lopez S, Fernandez M, Bodas R and Gonzalez JS, Screening the activity of plants and spices for decreasing ruminal methane production in vitro. Anim Feed Sci Technol 147:3652 (2008). 34 Kamel C, Greathead HMR, Tejido ML, Ranilla MJ and Carro MD, Effects of allicin and diallyl disulde on in vitro rumen fermentation of a mixed diet. Anim Feed Sci Technol 145:351363 (2008). 35 Lila ZA, Mohammed N, Kanda S, Kamada T and Itabashi H, Effect of -cyclodextrin allyl isothiocyanate on ruminal microbial methane production in vitro. Anim Sci J 74:321326 (2003). 36 Mohammed N, Ajisaka N, Lila ZA, Hara K, Mikuni K, Hara K, et al., Effect of Japanese horseradish oil on methane production and ruminal fermentation in vitro and in steers. J Anim Sci 82:18391846 (2004).
5

2707
wileyonlinelibrary.com/jsfa

J Sci Food Agric 2010; 90: 27002708

c 2010 Society of Chemical Industry

www.soci.org

AK Patra

REFERENCES
1 IPCC, Summary for policymakers, in Climate Change 2007: the Physical ScienceBasis. ContributionofWorkingGroupItotheFourthAssessment Report of the Intergovernmental Panel on Climate Change, ed. by Solomon S, Qin D, Manning M, Chen Z, Marquis M, Averyt KB, et al. Cambridge University Press, Cambridge, pp. 118 (2007). 2 Johnson KA and Johnson DE, Methane emission from cattle. J Anim Sci 73:24832492 (1995). 3 Bruinenberg MH, van der Honing Y, Agnew RE, Yan T, van Vuuren AM and Valk H, Energy metabolism of dairy cows fed on grass. Livest Prod Sci 75:117128 (2002). 4 Nkrumah JD, Okine EK, Mathison GW, Schmid K, Li C, Basarab JA, et al, Relationships of feedlot feed efciency, performance, and feeding behavior with metabolic rate, methane production, and energy partitioning in beef cattle. J Anim Sci 84:145153 (2006). 5 Beauchemin KA, McGinn SM, Martinez TF and McAllister TA, Use of condensed tannin extract from quebracho trees to reduce methane emissions from cattle. J Anim Sci 85:19901996 (2007). 6 Animut G, Goetsch AL, Puchala R, Patra AK, Sahlu T, Varel VH, et al, Methane emission by goats consuming diets with different levels of condensed tannins from lespedeza. Anim Feed Sci Technol 144:212227 (2008). 7 Hart KJ, Yanez-Ruiz DR, Duval SM, McEwan NR and Newbold CJ, Plant extracts to manipulate rumen fermentation. Anim Feed Sci Technol 147:835 (2008). 8 Kamra DN, Patra AK, Chatterjee PN, Kumar R, Agarwal N and Chaudhary LC, Effect of plant extract on methanogenesis and microbial prole of the rumen of buffalo: a brief overview. Aust J Exp Agric 48:175178 (2008). 9 Patra AK and Saxena J, A review of the effect and mode of action of saponins on microbial population and fermentation in the rumen and ruminant production. Nutr Res Rev 22:204219 (2009). 10 Patra AK, Kamra DN and Agarwal N, Effects of extracts of spices on rumen methanogenesis, enzyme activities and fermentation of feeds in vitro J Sci Food Agric 90:511520 (2010). 11 Patra AK, Responses of intake, digestibility and nitrogen utilization in goats fed low-quality roughages supplementing with tree foliages. J Sci Food Agric 89:14621472 (2009). 12 Sauvant D, Schmidely P, Daudin JJ and St-Pierre NR, Meta-analyses of experimental data in animal nutrition. Animal 2:12031214 (2008). 13 Patra AK, Aspects of nitrogen utilization in sheep fed mixed diets containing foliages from trees and browses. Br J Nutr 103:13191330 (2010). 14 SAS, SAS/STAT Users Guide, Version 8.2. SAS Institute, Cary, NC (2001). 15 St-Pierre NR, Integrating quantitative ndings from multiple studies using mixed model methodology. J Dairy Sci 84:741755 (2001).

16 Patra AK and Saxena J, Dietary phytochemicals as rumen modiers: a review of the effects on microbial populations. Antonie van Leeuwenhoek 96:363375 (2009). 17 Newbold CJ, Lassalas B and Jouany JP, The importance of methanogenesis associated with ciliate protozoa in ruminal methane production in vitro. Lett Appl Microbiol 21:230234 (1995). 18 Finlay BJ, Esteban G, Clarke KJ, Williams AG, Embley TM and Hirt RR, Some rumen ciliates have endosymbiotic methanogens. FEMS Microbiol Lett 117:157162 (1994). 19 Hostettmann K and Marston A, Saponins. Cambridge University Press, Cambridge (1995). 20 Guo YQ, Liu J-X, Lu Y, Zhu WY, Denman SE and McSweeney CS, Effect of tea saponin on methanogenesis, microbial community structure and expression of mcrA gene, in cultures of rumen micro-organisms. Lett Appl Microbiol 47:421426 (2008). 21 Hess HD, Monsalve LM, Lascano CE, Carulla JE, Diaz TE and Kreuzer M, Supplementation of a tropical grass diet with forage legumes and Sapindus saponaria fruits: effects on in vitro ruminal nitrogen turnover and methanogenesis. Aust J Agric Res 54:703713 (2003). 22 McAllister TA and Newbold CJ, Redirecting rumen fermentation to reduce methanogenesis. Aust J Exp Agric 48:713 (2008). 23 Wang Y, McAllister TA, Yanke LJ and Cheeke PR, Effect of steroidal saponin from Yucca schidigera extract on ruminal microbes. J Appl Microbiol 88:887896 (2000). 24 Tavendale MH, Meagher LP, Pacheco D, Walker N, Attwood GT and Sivakumaran S, Methane production from in vitro rumen incubations with Lotus pedunculatus and Medicago sativa, and effects of extractable condensed tannin fractions on methanogenesis. Anim Feed Sci Technol 123/124:403419 (2005). 25 Carulla JE, Kreuzer M, Machmuller A and Hess HD, Supplementation of Acacia mearnsii tannins decreases methanogenesis and urinary nitrogen in forage-fed sheep. Aust J Agric Res 56:961970 (2005). 26 Evans JD and Martin SA, Effects of thymol on ruminal micro-organisms. Curr Microbiol 41:336340 (2000). 27 Busquet M, Calsamiglia S, Ferret A, Carro MD and Kamel C, Effect of garlic oil and four of its compounds on rumen microbial fermentation. J Dairy Sci 88:43934404 (2005). 28 De Rosa M, Gambacorta A and Gliozzi A, Structure, biosynthesis, and physicochemical properties of archaebacterial lipids. Microbiol Rev 50:7080 (1986). 29 Miller TL, Ecology of methane production and hydrogen sinks in the rumen, in Ruminant Physiology: digestion, metabolism, growth and reproduction. ed. by Engelhardt W, Leonhard-Marek S, Breves G and Giesecke D. Ferdinand Enke Verlag, Stuttgart, Germany, pp. 317331 (1995).

2708
wileyonlinelibrary.com/jsfa
c 2010 Society of Chemical Industry J Sci Food Agric 2010; 90: 27002708

Research Article
Received: 15 April 2010 Revised: 22 June 2010 Accepted: 30 July 2010 Published online in Wiley Online Library: 2 September 2010

(wileyonlinelibrary.com) DOI 10.1002/jsfa.4144

Red mold dioscorea-induced G2/M arrest and apoptosis in human oral cancer cells
Wei-Hsuan Hsu, Bao-Hong Lee and Tzu-Ming Pan
Abstract
BACKGROUND: Monascus-fermented products are among the most commonly used traditional food supplements. Dioscorea is known to exhibit anticancer properties. In this study the effects of the ethanol extract of red mold dioscorea (RMDE) on cell proliferation, cell cycle and apoptosis in human oral cancer cells were investigated. RESULTS: RMDE exercised growth inhibition on squamous cell carcinoma-25 (SCC-25) cells. RMDE-mediated G2/M phase arrest was associated with the down-regulation of NF-B, resulting in the inhibition of cyclin B1 and CDK1 expression; this may be the mechanism by which RMDE inhibits cancer cells. Furthermore, the proapoptotic activity of RMDE was revealed by the Annexin V-FITC/PI double-staining assay. In addition, the proapoptotic effect of RMDE was evident by the inhibition of Bax expression in the mitochondria, resulting in the activation of caspase-9 and caspase-3 and subsequent triggering of the mitochondrial apoptotic pathway. RMDE also enhanced caspase-8 activity, indicating the involvement of the death receptor pathway in RMDE-mediated SCC-25 cell apoptosis. CONCLUSION: RMDE treatment inhibited the growth of SCC-25 cells by arresting cell cycle at the G2/M phase and induced apoptosis in a time- and dose-dependent manner. Therefore RMDE may be a good candidate for development as a dietary supplement against oral cancer. c 2010 Society of Chemical Industry Supporting information may be found in the online version of this article. Keywords: apoptosis; cell cycle; red mold dioscorea; Monascus-fermented products; oral cancer

INTRODUCTION
Oral cancer is the fth most common neoplasm worldwide, accounting for more than 500 000 cases annually.1 In Taiwan it has the fastest-rising incidence and mortality rate of any cancer and is the sixth most common cause of cancer death, being more prevalent in males than in females. Tobacco and alcohol consumption have been reported to be the major factors in the development of oral cancer.2 Diets low in carotenoids and vitamin A, poor oral hygiene and indoor air pollution are also recognised as factors in oral cancer.3,4 However, betel quid chewing is one of the most important causes of oral cancer in Taiwan, with high mortality and poor prognosis. Therefore, in an effort to improve patient survival and quality of life, new therapeutic approaches focusing on the molecular target and mechanism that mediate tumour cell growth or cell death have gained much attention. In recent years, natural food products have received increased attention because of their potential role in the prevention and/or intervention of carcinogenesis and neoplastic progression. Monascus species have been used as traditional food fungi in eastern Asia for several centuries. In previous studies, Monascusfermented rice, known as red mold rice (RMR), was found to show antioxidative ability5 as well as anti-Alzheimers disease and anticancer effects.6 RMR has many functional secondary metabolites. One of these secondary metabolites, monacolin K, acts by competitively inhibiting 3-hydroxy-3-methylglutaryl coenzyme A (HMG-CoA) reductase in the cholesterol biosynthetic pathway7 and also plays a role in antitumorigenic activity.8 In

addition, several yellow pigments such as monascin and ankaavin from Monascus have been reported to have anti-inammatory potential and cytotoxic/cytostatic activities.9,10 Monascus-fermented dioscorea, known as red mold dioscorea (RMD), comprises a dioscorea root substance as well as several Monascus metabolites. Dioscorea is regarded as a functional food or a valuable herb because of its content of many benecial ingredients for the prevention of various diseases.11 Dioscorin, polysaccharides, avones, vitamin C and polyphenols of dioscorea have been proven to exhibit high antioxidative ability.12 In addition, dioscorea shows antitumour activity.13,14 Thus Monascus fermentation of dioscorea should strengthen the anticancer effects of RMD. In a previous study the ethanol extract of RMR (RMRE) was found to effectively inhibit oral cancer carcinogenesis in a hamster buccal pouch model.15 Hence Monascus fermentation of dioscorea may lead to stronger anticancer effects. In the present study, to understand the effect of the ethanol extract of RMD (RMDE) fermented by Monascus purpureus NTU 568 on human oral cancer cells, we selected human tongue cancer squamous cell carcinoma-25 (SCC-25) cells for examining the effects on cell

Correspondence to: Tzu-Ming Pan, Institute of Microbiology and Biochemistry, College of Life Science, National Taiwan University, No. 1, Sec. 4, Roosevelt Road, Taipei 10617, Taiwan. E-mail: tmpan@ntu.edu.tw

2709

Institute of Microbiology and Biochemistry, College of Life Science, National Taiwan University, No. 1, Sec. 4, Roosevelt Road, Taipei 10617, Taiwan

J Sci Food Agric 2010; 90: 27092715

www.soci.org

c 2010 Society of Chemical Industry

www.soci.org viability, cell proliferation and the mechanisms underlying the induction of cell cycle arrest and apoptosis.

W-H Hsu, B-H Lee, T-M Pan Cell culture Human SCC-4, SCC-9 and SCC-25 tongue, FaDu and HEp-2 pharynx cell lines were obtained from the Bioresource Collection and Research Center (Hsinchu, Taiwan). SCC cells were maintained in DMEM/Hams F-12 (1 : 1 v/v) medium supplemented with 100 mL L1 FBS, 1.5 g L1 sodium bicarbonate, 400 ng mL1 hydrocortisone and 10 mL L1 antibiotic/antimitotic solution. FaDu and HEp-2 cells were maintained in DMEM medium containing 100 mL L1 FBS and 10 mL L1 antibiotic/antimitotic solution. Cells were incubated in 5% CO2 /95% humidied atmosphere at 37 C. Determination of cell viability The cell-killing effect of RMDE and RMRE against oral cancer cells was measured using the crystal violet staining assay. Cells were seeded on 24-well plates (3 104 cells per well) and treated with various concentrations of RMDE and RMRE for 24 and 48 h. The medium was then removed, washed with phosphatebuffered saline (PBS) and stained with 2 g L1 crystal violet in 100 mL L1 phosphate-buffered formaldehyde for 20 min before being washed with water. The crystal violet bound to the cells was dissolved in 20 g L1 SDS solution and its absorbance at 600 nm was measured. The 50% inhibitory concentration (IC50 ) of RMDE/RMRE was calculated from a sigmoidal doseresponse curve. Clonogenic survival assay Cells were seeded on six-well plates at 250 cells per well in a nal volume of 2 mL of medium containing either vehicle or an appropriate RMDE concentration. All cultures were incubated for an additional 14 days until colonies were large enough to be clearly discerned. At this point the medium was aspirated and the dishes were washed once with PBS and stained with a ltered solution of 5 g L1 crystal violet for 10 min. Colonies containing 50 cells were scored. Clonogenic survival was expressed as the percentage of colonies formed in RMDE-treated cells with respect to vehicle-treated cells. Cell cycle distribution After 12 and 24 h of exposure to RMDE the medium was aspirated and adherent cells were harvested and centrifuged at 300 g for 5 min. Cells were washed with PBS, xed with 700 mL L1 ice-cold ethanol at 20 C overnight and then stained with PI at room temperature for 30 min. The cell cycle distribution was analysed by ow cytometry using a FACScan-LSR ow cytometer equipped with CellQuest software (BD Biosciences, San Jose, CA, USA). Apoptosis analysis For apoptosis detection, oating cells in the medium and adherent cells were collected after 12 and 24 h of RMDE treatment. Cells were harvested, washed in ice-cold PBS and resuspended in 200 L of binding buffer before being incubated in 5 L of Annexin V-FITC (BD Biosciences) solution and 5 L of PI at room temperature for 15 min in the dark. Then 300 L of binding buffer was added. Cells were analysed by ow cytometry. Untreated cells were used as the control for double staining. Reverse transcription polymerase chain reaction (RT-PCR) Total RNA was isolated using Trizol (Life Technologies) according to the manufacturers instructions. A 5 g aliquot of puried

MATERIALS AND METHODS

Chemicals and reagents Crystal violet, propidium iodide (PI), sodium dodecyl sulfate (SDS), Triton X-100, trypsin, trypan blue, FolinCiocalteus reagent, gallic acid and quercetin were purchased from Sigma Chemical Co. (St Louis, MO, USA). Foetal bovine serum (FBS) was purchased from Life Technologies (Auckland, New Zealand). Dimethyl sulfoxide (DMSO) was purchased from Wako Pure Chemical Industries (Saitama, Japan). Ethanol (950 mL L1 ) was purchased from Taiwan Tobacco and Liquor Corp. (Taipei, Taiwan). Dulbeccos modied Eagles medium (DMEM), Hams F-12, sodium bicarbonate, hydrocortisone, penicillin and streptomycin were purchased from HyClone Laboratories (Logan, UT, USA). Aluminium nitrate (Al(NO3 )3 ) and potassium acetate (CH3 COOK) were purchased from J.T. Baker Co. (Phillipsburg, NJ, USA). Preparation of RMDE and RMRE The M. purpureus NTU 568 culture strain was maintained on potato dextrose agar (Difco Co., Detroit, MI, USA) slants at 10 C and transferred monthly. Dioscorea root (Dioscorea batatas Dence) purchased from a local supermarket in Taiwan was used to produce RMD by the method of solid state culture.14 The preparation of RMR was carried out with a substrate of long-grain rice (Oryza sativa) using the method of solid state culture. Briey, 500 g of substrate was soaked in deionised water for 8 h and excess water was removed with a sieve. The substrate was autoclaved (HL-341, Gemmy Corp., Taipei, Taiwan) at 121 C for 20 min in a wooden koji dish with dimensions of 30 cm 20 cm 5 cm. After the substrate had cooled, it was inoculated with 50 g L1 spore suspension and cultivated at 30 C for 10 days. During the culturing stage, 100 mL of water was added daily to the substrate from the second day to the fth day. At the end of cultivation the crushed and dried product with the mold was extracted with 950 mL L1 ethanol or water at 50 C for 3 days. Extracts were further freeze-dried to powder and stored at 20 C until use. Samples were dissolved in DMSO and the concentration was kept below 1 mL L1 in the experimental design. Determination of total phenolic compounds and avonoids For the determination of total phenolic compounds, freezedried dioscorea or RMD powder was dissolved in deionised water and the concentration of total phenolic compounds was measured spectrophotometrically using FolinCiocalteus reagent. Sample solution (100 L), FolinCiocalteus reagent (500 L), sodium carbonate (400 L, 75 g L1 ) and deionised water (5 mL) were mixed thoroughly and kept at room temperature for 30 min before the absorbance at 760 nm was measured. Total phenolic content was determined using gallic acid as standard. For the determination of avonoids, sample solution (500 L), ethanol (1.5 mL), Al(NO3 )3 (100 L, 100 g L1 ), CH3 COOK (100 L, 1 mol L1 ) and water (2.8 mL) were mixed thoroughly and kept at ambient temperature for 40 min before the absorbance at 415 nm was measured with a spectrophotometer (U-2000, Hitachi, Tokyo, Japan). Total avonoid content was calculated according to a standard curve established with quercetin. The results indicated that the levels of total phenols and avonoids were 119 and 191 mg kg1 dioscorea respectively and 179 and 249 mg kg1 RMD respectively.

2710

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 27092715

Effects of red mold dioscorea on SCC-25 cells total RNA was employed for reverse transcription using SuperScript III (Life Technologies). The reaction mixture was incubated at 42 C for 1 h and the reaction was terminated by heating at 70 C for 10 min. Amplication of the RT product by PCR was performed using Promega Taq DNA Polymerase (Promega Co., Madison, WI, USA). All reactions were carried out in a thermal cycler (Model 2400, Perkin-Elmer, Norwalk, CT, USA) with the following primers: GAPDH sense, 5 -CATCAAGAAGGTGGTGAAGCAG-3 , and antisense, 5 -CCACCACCCTGTTGCTGTAGCCA3 , GST-P sense, 5 -TCATCTACACCAACTATGAG-3 , and antisense, 5 -GCCACATAGGCAGAGAGCAG-3 , cyclin B1 sense, 5 -CTTATACTAAGCACCAAATC-3 , and antisense, 5 -CTTGGCTAAATCTTGAACT3 , and CDK1 sense, 5 -CTTATGCAGGATTCCAGGTT-3 , and antisense, 5 -GGTGCCTATACTCCAAATGTC-3 (PREMIER Biosoft Int., Palo Alto, CA, USA). The conditions for standard amplication were 95 C for 5 min, then 30 cycles of 95 C for 30 s (to amplify GAPDH, cyclin B1 and CDK1) or 95 C for 1 min (to amplify GST-P), 65 C for 30 s, 74 C for 30 s and 74 C for 10 min. Products of the reaction were separated on 20 g L1 agarose gel, stained with 1 g mL1 ethidium bromide and visualised using a UVPGDS-7900 digital imaging system (UVP AutoChemi System, Cambridge, UK). All amplications were conducted within the linear range of the assay, normalised to respective GAPDH levels using SPSS Version 17.0 (SPSS Inc., Chicago, IL, USA). Colorimetric estimation of caspase-3, caspase-8 and caspase-9 activities Caspase-3, caspase-8 and caspase-9 activities were determined using kits from Biovision (Mountain View, CA, USA). Cells (1 106 ) were treated with RMDE for 6, 12 and 24 h, washed with PBS, suspended in 50 L of cell lysis buffer and incubated on ice for 10 min. Following centrifugation at 10 000 g for 1 min, the supernatant (cytosolic extract) was transferred to a fresh tube and put on ice for immediate assay. Protein concentration was assayed using 50200 g of standard protein in 50 L of cell lysis buffer for each assay. The cytosolic extract was mixed with 50 L of 2 reaction buffer (containing 10 mmol L1 dithiothreitol) and 5 L of 4 mmol L1 substrate (200 mol L1 nal concentration), incubated at 37 C for 1 h and its absorbance at 405 nm was measured. Immunoblot analysis Proteins separated by SDS polyacrylamide gel electrophoresis were electrophoretically transferred to polyvinylidene diuoride membranes. Blots were rst incubated in PBS containing 50 g L1 non-fat dry milk for 2 h to block non-specic binding sites, then in a 1 : 1000 dilution of primary antibodies at 4 C overnight and nally, after washing, in a 1 : 20 000 dilution of horseradish peroxidaseconjugated secondary antibodies (GeneTex, Inc., San Antonio, TX, USA) at room temperature for 1 h. After washing, immunoreactive proteins were visualised using enhanced chemiluminescence detection reagents (Sigma Chemical Co.). Densitometry was quantitated with SPSS Version 17.0 (SPSS Institute, Inc.). Statistical analysis Data were expressed as mean standard deviation (SD). Statistical signicance was determined by one-way analysis of variance (ANOVA) using the general linear model procedure of SPSS Version 17.0 (SPSS Institute, Inc.), followed by ANOVA with Duncans test. All comparisons were made relative to controls, and signicant differences are indicated as P < 0.05, P < 0.01 or P < 0.001.

www.soci.org

Figure 1. Representative cell viability of different oral cancer cells: (A) effects of 24 h treatment with ethanol and water extracts of RMR and RMD on viability of SCC-25 cells; (B) clonogenic survival assay for RMDE on oral cancer cells. Results are expressed as mean SD (n = 3).

Table 1. Growth inhibition (IC50 , g mL1 ) of different oral cancer cell lines by RMRE and RMDE Cell line Extract RMRE 24 h 48 h RMDE 24 h 48 h SCC-4 SCC-9 SCC-25 FaDu HEp-2

>250 >250 >250 142.6

>250 237.3 >250 151.9

>250 188.6 78.1 36.4

>250 >250 244.8 117.8

>250 >250 162.6 155.5

RESULTS
Effect of RMD/RMR ethanol and water extracts on cell viability and proliferation Figure 1(A) shows the dose-dependent effects of ethanol and water extracts of RMD and RMR on the viability of oral cancer SCC-25 cells after a 24 h treatment. Water extracts of RMD (RMDW) and RMR (RMRW) exerted only weak cell toxicity, whereas RMDE and RMRE reduced cell viability signicantly. Thus we investigated the growth-inhibitory effects of RMDE and RMRE on different oral cancer cell lines. The results showed that IC50 of RMDE was less than that of RMRE, with a time-dependent decrease in growth inhibition. Among the ve cell lines tested, the maximum cytotoxic effect was observed on SCC-25 cells (Table 1). RMDE affected the long-term survival of oral cancer cells more severely (Fig. 1(B)) than their short-term viability (Table 1),

2711

J Sci Food Agric 2010; 90: 27092715

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org

W-H Hsu, B-H Lee, T-M Pan

Table 2. Cell cycle distribution (%) in SCC-25 cells treated with RMDE for 24 h Phase Treatment Untreated 100 g mL1 200 g mL1 G0/G1 74.0 5.02 55.1 2.20 50.6 3.04 S 6.1 0.15 10.0 0.19 9.8 0.12 G2/M 18.9 2.04 34.0 1.94 38.3 3.06

Results are expressed as mean SD (n = 3).

indicating that the treatment with RMDE resulted in long-term cell damage and subsequent inhibition of cell proliferation. Effect of RMDE on cell cycle distribution To elucidate the cytotoxic mechanism of RMDE, further experiments were performed using the SCC-25 cell line. To examine the mechanism underlying the inhibitory effects of RMDE on SCC-25 cell proliferation, the cell cycle distribution was evaluated using ow cytometry. A 24 h treatment with RMDE caused cell cycle arrest at the G2/M phase, and this effect was dose-dependent (Table 2). Thus the cytotoxic effect of RMDE on SCC-25 cells might be attributed to the induction of G2/M arrest. Effect of RMDE on GST-P, cyclin B1 and CDK1 expression We next examined the effect of RMDE on cell cycle-regulatory molecules. Figure 2 shows the results of the RT-PCR analysis of

Figure 2. Changes in level of mRNA associated with cell proliferation by RT-PCR: (A) expression of GST-P, cyclin B1 and CDK1 after exposure to 100 g mL1 RMDE for 3 and 6 h; (B) densitometric analysis of RT-PCR images.

cell survival markers and G2/M phase arrest-associated factors of SCC-25 cells. In comparison with the control group, the expression

Figure 3. Effects of RMDE-induced apoptosis on SCC-25 cells after 24 h (cells in lower right quadrant represent early apoptosis; cells in upper right quadrant represent late apoptosis): (A) untreated group; (B) 100 g mL1 RMDE treatment; (C) 200 g mL1 RMDE treatment; (D) analysis of cells in early and late stage apoptosis.

2712

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 27092715

Effects of red mold dioscorea on SCC-25 cells of GST-P, cyclin B1 and CDK1 was decreased in cells treated with 100 g mL1 RMDE for 3 and 6 h. Thus the inhibitory effect of RMDE on SCC-25 cell proliferation and cell cycle progression might be attributable to the down-regulation of GST-P, cyclin B1 and CDK1. Effect of RMDE on SCC-25 cell apoptotic induction To further elucidate the mechanism of action of RMDE on SCC-25 cells, we examined the effect of RMDE on apoptosis by performing Annexin V-FITC/PI double staining. This staining method along with ow cytometry enables the quantitative assessment of living (Annexin V-FITC negative/PI negative), early apoptotic (Annexin V-FITC positive/PI negative), late apoptotic/necrotic (Annexin V-FITC positive/PI positive) and dead (Annexin V-FITC negative/PI positive) cells. The effects of a 24 h RMDE treatment on SCC-25 cell apoptosis are shown in Fig. 3. The cells in the lower right quadrant represent early apoptosis, while those in the upper right quadrant represent late apoptosis. The results indicated that, after treatment with 100 g mL1 RMDE, most cells were in the early apoptotic stage (69.88%) and a few were in the late apoptotic stage (7.87%), i.e. 77.75% of cells underwent apoptosis. On the other hand, 95.93% of cells underwent apoptosis in the group treated with 200 g mL1 RMDE, 52.3% being in the early apoptotic stage and 43.63% in the late apoptotic stage. Effect of RMDE on caspase-3, caspase-8 and caspase-9 activities Apoptosis has been found to be regulated by two main pathways, the mitochondrial pathway and the death receptor pathway, both of which activate caspase-3. Caspase-8 and caspase-9 are activators/initiators of the death receptor pathway and the mitochondrial pathway respectively.16 We examined the effect of RMDE on caspase activities in order to understand the mechanism of RMDE cell proliferation inhibition. Figure 4 indicates that RMDE promoted the expression of caspase-3, caspase-8 and caspase-9 in a dose- and time-dependent manner. These results suggested that the RMDE-induced apoptosis was mediated by the induction of both the death receptor pathway and the mitochondrial pathway. Effect of RMDE on NF-B, I-B and Bax expression The NF-B transcription factor is reported to be overexpressed in highly proliferating tumours.17 Besides aiding tumour cell proliferation, NF-B has also been reported to enhance tumour development by inhibiting apoptosis.18 NF-B is sequestered in the cytoplasm by inhibitory proteins such as I-B, which mask the nuclear localisation signal of NF-B. An increase in Bax expression can block mitochondria-mediated apoptosis by preventing cytochrome c release from the mitochondria, thereby inhibiting the activation of caspase-3. Figure 5 shows a representative Western blot analysis of SCC-25 cells treated with RMDE. The expression of NF-B, I-B and Bax was detected as bands of molecular weight 65, 43 and 21 kDa respectively. In comparison with the control group, RMDE treatment decreased the expression of NF-B and increased the expression of I-B and Bax.

www.soci.org

Figure 4. Effects of RMDE on (A) caspase-3, (B) caspase-8 and (C) caspase-9 activities: , 50 g mL1 RMDE; , 100 g mL1 RMDE; , 200 g mL1 RMDE. Results are expressed as mean SD (n = 3). Signicantly different from control group at P < 0.05. Signicantly different from control group at P < 0.01.

DISCUSSION
Dioscorea is proven to have antioxidative as well as antitumour properties.13,14,19 The antiproliferative activity of polyphenols such as delphinidin, cyanidin, peonidin, petunidin and malvidin has also

been reported.20 Additionally, a polyphenol in black tea has been shown to effectively inhibit hamster buccal pouch carcinogenesis and induce apoptosis,21,22 and its antioxidative property prevents oral carcinogenesis.23 In a previous study we found that ethanol extracts of Monascus-fermented products had antioxidative properties, including reducing power and 1,1-diphenyl-2-pichrylhydrazyl radical-scavenging activity.24 Furthermore, Monascus-fermented products contain various antioxidants such as dimerumic acid, tannins and phenols.14 Thus Monascus-fermented products exhibit tumour-inhibitory properties, and Monascus fermentation of dioscorea may result in stronger anticancer effects. The levels of total phenols and avonoids were 119 and 191 mg kg1 dioscorea respectively and 179 and 249 mg kg1 RMD respectively (data not shown). Thus the total phenol and avonoid levels increased as a result of Monascus fermentation, and this may confer RMD with anti-oral cancer properties. On the other hand, Monascus species produce yellow pigments such as monascin and ankaavin, which have been reported to have cancer cell-cytotoxic activities.9,10 As a result, apart from analysing phenol and avonoid contents, we

2713

J Sci Food Agric 2010; 90: 27092715

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org

W-H Hsu, B-H Lee, T-M Pan

Figure 5. (A) Western blot analysis of protein extracts obtained from SCC-25 cells treated with different concentrations of RMDE for different times. Densitometric analysis of (B) NF-B, (C) I-B and (D) Bax.

also used high-performance liquid chromatography to identify the bioactive chemical constituents in RMDE. Monascin and ankaavin were present at concentrations of 137.2 and 34.14 g kg1 respectively in Monascus-fermented dioscorea,25 higher than the levels found in RMRE.15 Dioscorea has proven antitumour ability,13 and RMDE has stronger anticancer activity than RMRE, RMDW and RMRW used in this study. We suggest that the greater anticancer activity of RMDE compared with RMRE might be attributable to RMDE containing higher levels of anticancer substances such monascin and ankaavin. Monascus-fermented products have been reported to inhibit tumour progression and tumour metastasis-associated factors and to reduce angiogenesis.8 RMRE is known to exert inhibitory effects against oral carcinogenesis.15 However, in the present study, RMDE exerted stronger cell cytotoxicity than RMRE (Table 1) and was a more potent cell proliferation inhibitor than RMDW and RMRW (Fig. 1(A)). These ndings suggested that RMDE had the potential to repress oral cancer cell growth. Therefore we examined the effect of RMDE on the induction of apoptosis. Phytochemicals have been shown to induce cell cycle arrest, cause apoptosis and affect the differentiation and proliferation of cells mediated by the effect of intracellular reactive oxygen species on the signal transduction pathway.26 Further, two apoptosis signalling pathways converge during the activation of caspases in the inhibition of cell proliferation. Capsase-8 is activated by the death pathway, which is initiated by the death receptor that contains the intracellular death domain and is propagated through the Fas-associated protein with death domain (FADD) adaptor protein. The ligand bound to the death receptor forms a

signal complex, FADD, which increases caspase-8 and caspase-3 activities to induce apoptosis.16,27 The mitochondrial pathway is regulated by mitochondria-released cytochrome c, which leads to caspase-9 activation via apoptotic protease-activating factor-1. Treatment of cells with RMDE for 24 h resulted in cell cycle arrest at the G2/M phase (Table 2). This effect was also associated with the repression of CDK1 and cyclin B1 mRNA levels, resulting in cell proliferation inhibition (Fig. 2). The overexpression of GST-P and NF-B enhances cell proliferation and prevents apoptosis reaction. RMDE treatment decreased the mRNA levels of GST-P (Fig. 2), indicating that RMDE inhibited cell proliferation, possibly via lowering CDK1 and cyclin B1 expression and then arresting cells at the G2/M phase. In addition, RMDE effectively induced SCC-25 cell apoptosis in a dose-dependent manner (Figs 3 and 5(D)). Our results suggested that the 6, 12 and 24 h RMDE (200 g mL1 ) treatments induced the activity of caspase-3, caspase-8 and caspase-9 in SCC-25 cells (Fig. 4). The activity of caspase-8 was signicantly increased compared with that of caspase-9, indicating that RMDE induced SCC-25 cell apoptosis mainly via activating caspase-8. RMDE treatment also resulted in a decrease in NF-B levels and an increase in I-B (Fig. 5), thus indicating that RMDE inhibited cell proliferation by repressing NF-B transcription. In conclusion, RMDE treatment resulted in the following signicant changes in SCC-25 cells: (a) selective inhibition of SCC25 cells, (b) dose-dependent cell cycle arrest at the G2/M phase and (c) time- and dose-dependent induction of apoptosis (Fig. 6). Therefore RMDE has potential to be used as a functional food in adjuvant chemotherapy for treating human oral cancer.

2714

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 27092715

Effects of red mold dioscorea on SCC-25 cells

www.soci.org
10 Su NW, Lin YL, Lee MH and Ho CY, Ankaavin from Monascusfermented red rice exhibits selective cytotoxic effect and induces cell death on Hep G2 cells. J Agric Food Chem 53:19491954 (2005). 11 Chang WC, Yu YM, Wu CH, Tseng YH and Wu KY, Reduction of oxidative stress and atherosclerosis in hyperlipidemic rabbits by Dioscorea rhizome. Can J Physiol Pharmacol 83:423430 (2005). 12 Wang G, Chen H, Huang M, Wang N, Zhang J, Zhang Y, et al., Methyl protodioscin induces G2/M cell cycle arrest and apoptosis in HepG2 liver cancer cells. Cancer Lett 241:102109 (2006). 13 Park MK, Kwon HY, Ahn WS, Bae S, Rhyu MR and Lee Y, Estrogen activities and the cellular effects of natural progesterone from wild yam extract in mcf-7 human breast cancer cells. Am J Chin Med 37:159167 (2009). 14 Lee CL, Wang JJ, Kuo SL and Pan TM, Monascus fermentation of dioscorea for increasing the production of cholesterol-lowering agent-monacolin K and anti-inammation agent-monascin. Appl Microbiol Biotechnol 72:12541262 (2006). 15 Tsai RL, Ho BY and Pan TM, Red mold rice mitigates oral carcinogenesis in 7,12-dimethyl-1,2-benz[a]anthracene-induced oral carcinogenesis in hamster. Evid Based Compl Altern Med DOI: 10.1093/ecam/nep215 (2009). 16 Nagata S, Apoptosis by death factor. Cell 88:355365 (1997). 17 Dolcet X, Llobet D, Pallares J and Matias-Guiu X, NF-B in development and progression of human cancer. Virchows Arch 446:475482 (2005). 18 Tamatani M, Che YH, Matsuzaki H, Ogawa S, Okado H, Miyake S, et al., Tumor necrosis factor induces Bcl-2 and Bcl-x expression through NFkappaB activation in primary hippocampal neurons. J Biol Chem 274:85318538 (1999). 19 Lin YM and Lin KW, Antioxidative ability, dioscorin stability, and the quality of yam chips from various yam species as affected by processing method. J Food Sci 74:118125 (2009). 20 Jacob JK, Hakimuddin F, Paliyath G and Fisher H, Antioxidant and antiproliferative activity of polyphenols in novel high-polyphenol grape lines. Food Res Int 41:419428 (2008). 21 Chandra-Mohan KVP, Devaraj H, Prathiba D, Hara Y and Nagini S, Antiproliferative and apoptosis inducing effect of lactoferrin and black tea polyphenol combination on hamster buccal pouch carcinogenesis. Biochim Biophys Acta 1760:15361544 (2006). 22 Chandra-Mohan KVP, Vidjaya-Letchoumy P, Hara Y and Nagini S, Combination chemoprevention of hamster buccal pouch carcinogenesis by bovine milk lactoferrin and black tea polyphenols. Cancer Invest 26:193201 (2008). 23 Vidjaya-Letchoumy P, Chandra-Mohan KVP, Stegeman JJ, Gelboin HV, Hara Y and Nagini S, In vitro antioxidative potential of lactoferrin and black tea polyphenols and protective effects in vivo on carcinogen activation, DNA damage, proliferation, invasion, and angiogenesis during experimental oral carcinogenesis. Oncol Res 17:193203 (2008). 24 Lee CL, Hung HK, Wang JJ and Pan TM, Red mold dioscorea has greater hypolipidemic and antiatherosclerotic effect than traditional red mold rice and unfermented dioscorea in hamsters. J Agric Food Chem 55:71627169 (2007). 25 Hsu WH, Lee BH and Pan TM, Protection of Monascus-fermented dioscorea against DMBA-induced oral injury in hamster by antiinammatory and antioxidative potentials. J Agric Food Chem 58:67156720 (2010). 26 Hu R and Kong AHT, Activation of MAP kinases, apoptosis and nutrigenomics of gene expression elicited by dietary cancerprevention compounds. Nutrition 20:8388 (2004). 27 Pietenpol JA and Stewart ZA, Cell cycle checkpoint signaling: cell cycle arrest versus apoptosis. Toxicology 181/182:475481 (2002).

Figure 6. Proposed signal pathway of RMDE-induced G2/M arrest and apoptosis in human oral cancer SCC-25 cells.

ACKNOWLEDGEMENT
This research work and subsidiary spending were supported by Paolyta Co., Ltd (Taipei, Taiwan). Supporting information Supporting information may be found in the online version of this article.

REFERENCES
1 Parkin DM, Bray F, Ferlay J and Pisani P, Estimating the world cancer burden: Globocan 2000. Int J Cancer 94:153156 (2001). 2 Blot WJ, McLaughlin JK, Winn DM, Austin DF, Greenberg RS, PrestonMartin S, et al., Smoking and drinking in relation to oral and pharyngeal cancer. Cancer Res 48:32823287 (1988). 3 McLaughlin JK, Gridley G, Block G, Winn DM, Preston-Martin S, Schoenberg JB, et al., Dietary factors in oral and pharyngeal cancer. J Natl Cancer Inst 80:12371243 (1988). 4 Pintos J, Franco EL, Kowalski LP, Oliveira BV and Curado MP, Use of wood stoves and risk of cancers of the upper aero-degestive tract: a case-control study. Int J Epidemiol 27:936940 (1998). 5 Lee BH, Ho BY, Wang CT and Pan TM, Red mold rice promoted antioxidase activity against oxidative injury and improved the memory ability of zinc-decient rats. J Agric Food Chem 57:1060010607 (2009). 6 Lee CL, Kuo TF, Wang JJ and Pan TM, Red mold rice ameliorates impairment of memory and learning ability in intracerebroventricular amyloid beta-infused rat by repressing amyloid beta accumulation. J Neurosci Res 85:31713182 (2007). 7 Endo A and Monacolin K, a new hypocholesterolemic agent produced by a Monascus species. J Antibiot 32:852854 (1979). 8 Ho BY and Pan TM, The Monascus metabolite monacolin K reduces tumor progression and metastasis of Lewis lung carcinoma cells. J Agric Food Chem 57:82588265 (2009). 9 Akihisa T, Tokuda H, Ukiya M, Kiyota A, Yasukawa K, Sakamoto N, et al., Anti-tumor-initiating effects of monascin, an azaphilonoid pigment from the extract of Monascus pilosus fermented rice (red-mold rice). Chem Biodiversity 2:13051309 (2005).

2715
wileyonlinelibrary.com/jsfa

J Sci Food Agric 2010; 90: 27092715

c 2010 Society of Chemical Industry

Research Article
Received: 6 May 2010 Revised: 3 August 2010 Accepted: 5 August 2010 Published online in Wiley Online Library: 2 September 2010

(wileyonlinelibrary.com) DOI 10.1002/jsfa.4147

Effects of postharvest treatments on fruit quality of sweet pepper at low temperature

Paula Cuadra-Crespo and Francisco M del Amor
Abstract
BACKGROUND: Postharvest storage of sweet pepper fruits (Capsicum annuum L.) at low temperatures could impair their physical and chemical composition. Therefore, maintenance of essential nutrition support or altered gas exchange could preserve fruit quality, minimizing chilling injury. Thus our aim was to determine the response to postharvest application of a low concentration of nitrogen (urea) or antitranspirant (pinolene) during a period of 21 days at 5 C. RESULTS: The results indicate that storage at 5 C was effective with respect to maintaining rmness of sweet pepper fruits for 21 days, while application of antitranspirant increased rmness compared with non-sprayed fruits. Additionally, urea maintained color while increasing total phenolics and the activity of catalase and ascorbate peroxidase, lowering lipid peroxidation. Composition of free amino acids was affected to a minor extent. CONCLUSION: Maintaining quality is of paramount importance in the postharvest period. This study shows the effect of both temperature and spraying treatments with regard to maintaining fruit quality during this period, and provides new insights into the physiological role of enzymes of the antioxidant system during pepper storage at low temperature. c 2010 Society of Chemical Industry Keywords: Capsicum annuum L.; urea; antitranspirant; color; antioxidant enzymes; chilling injury

INTRODUCTION
With the increasing demand for fresh fruit and vegetables, postharvest technology for extending shelf-life of these perishable commodities has gained signicant importance in recent years.1 The principal physiological factors that negatively impact pepper fruit during shipment and storage and subsequent marketing are water loss and chilling injury.2 The skin of fruit and vegetables plays an important role in gas exchange between the product and the surrounding environment,3 and for this reason the protection of the pericarp against dehydration is particularly important after harvest, when fruits do not receive water or nutrients from the plant. Therefore, implementation of techniques to preserve the physicochemical properties of the pericarp could help to preserve fruit quality during the storage period. Storage temperature has a great inuence on the physiological response, since temperature and humidity are the environmental factors that have the strongest inuence on fruit quality.3 Many studies have pointed out an inuence of storage temperature on water loss rate, texture and overall quality of fresh pepper fruits. Thus temperature regulation is the most effective tool for extending the storage life of fresh commodities, including pepper.4 However, storage at low temperatures could produce unusual ripening, water loss, an increase of CO2 , and higher permeability in cellular membranes, inducing ion leakage; the damage is progressively more severe during long-term storage at low a temperatures.5 Gonz lez-Aguilar et al.6 indicated that symptoms are accompanied by biochemical and physiological changes, which are generated by the direct effect of low temperature on cellular constituents.

If free radicals are not neutralized in humans and plants, they damage cells and organs, causing many degenerative diseases and susceptibility to biotic and abiotic stresses.7 Reactive oxygen species (ROS) are strongly associated with lipid peroxidation and consequent deterioration of food materials, but have also been involved in the development of several diseases.8 Low temperatures can induce free radical production and ROS may contribute to the loss of cellular functions through lipid peroxidation.9 Oxidation and peroxidation of membrane lipids and proteins could be caused by ROS.10 Therefore, antioxidant enzymes are the most important components in the scavenging system of ROS11 and peppers contain many active substances that are important for protection against oxidative damage by free radicals.12 Plants are able to use several forms of N, nitrate and ammonium being the most important ones. Urea is one of the most widely used foliar-N fertilizers, characterized by high leaf penetration rate and low cost. Additionally, urea has been considered the most suitable form of foliar N because of its rapid absorption, low phytotoxicity and high solubility in both oil and water.13 The nutrient with the single greatest effect on fruit quality is N.14 Moreover, pepper fruit yield and quality are affected by different agricultural practices and

Correspondence to: Francisco M del Amor, Departamento de Citricultura y Calidad Alimentaria, Instituto Murciano de Investigaci n y Desarrollo Agrario o y Alimentario (IMIDA), C/Mayor s/n, 30150 Murcia, Spain. E-mail: franciscom.delamor@carm.es Departamento de Citricultura y Calidad Alimentaria, Instituto Murciano de Investigaci n y Desarrollo Agrario y Alimentario (IMIDA), C/Mayor s/n, 30150 o Murcia, Spain

2716

J Sci Food Agric 2010; 90: 27162722

www.soci.org

c 2010 Society of Chemical Industry

Postharvest quality of sweet pepper by the availability of N to the plant.15,16 No N source is provided to the fruit once detached from the plant, but many physiological process could still require N to avoid metabolic unbalance in the pericarp. Therefore, rapid absorption of N applied directly to the fruit could help to maintain its characteristics. Thus, urea sprays enhanced the green pigmentation in apple after a storage period,17 increased chlorophyll content in broccoli heads18 and had a benecial effect in pepper leaves, maintaining membrane permeability.19 Antitranspirants are compounds applied to plant leaves to reduce transpiration, since they reduce the stomatal opening and increase the leaf resistance to water vapor diffusion, and they have often been applied in an attempt to prevent water stress.20 Reducing rates of transpiration directly by antitranspirant sprays avoided the need for drastically altering environmental conditions in experiments designed to evaluate the effects of transpiration.21 Antitranspirants have been used successfully in agriculture to control leaf transpiration and improve quality of sweet pepper,22 tomatoes,23 onions24 and potatoes.25 The aim of this study was to determine the effect of foliar urea and antitranspirant applications on pepper fruits, regarding both physical and chemical postharvest characters (oxidative metabolism and amino acids). Therefore, we studied several quality parameters such as fruit rmness, color, lipid peroxidation, antioxidants such as catalase (CAT), ascorbate peroxidase (APOX) and phenolic compounds, and the free amino acids prole. Our result could help to dene an efcient postharvest strategy (from a physiological and agronomical perspective) to maintain fruit quality of sweet pepper fruits under low temperature.

www.soci.org a Konica-Minolta CR-300 colorimeter (Minolta, Osaka, Japan), light source D65, making three measurements for each pepper. After rmness and color determinations, destructive measurements were carried out for measuring the enzymatic response; therefore these parameters were measured in different fruits of the same treatment at each harvest time. Color data are provided as CIELAB (L a b ) coordinates, which dene the color in a three-dimensional space: L indicates lightness, and a and b are the chromaticity coordinates greenred and blueyellow, respectively. L is an approximate measurement of luminosity, which is the property according to which each color can be considered as equivalent to a member of the gray scale, between black and white, taking values within the range 0100; a takes positive values for reddish colors and negative values for greenish ones, whereas b takes positive values for yellowish colors and negative values for bluish ones.16 Total phenolic compounds Total phenolic compounds were extracted from 0.4 g of frozen pepper fruits (80 C) with 4 mL methanol and 0.1 mol L1 HCl. The homogenate was centrifuged at 15 000 g for 20 min at 4 C. For the determination, FolinCiocalteu reagent was used diluted with distilled water (1 : 10). The diluted reagent (2 mL) was mixed with 400 L supernatant and 1600 L sodium carbonate (7.5%) was added. The mixture was kept for 30 min in the dark and then centrifuged at 5000 g for 5 min. The supernatant was separated and absorbance was measured at 765 nm according to methodology of K hkonen et al.27 The total phenolic content was a expressed as gallic acid equivalents in mg mL1 fresh weight. Antioxidant enzymes Antioxidant enzymes were determined according to Del Amor et al.28 Briey, extracts for the determination of catalase (CAT) and ascorbate peroxidase (APOX) activities were homogenized and centrifuged and the supernatant was used for the assays. CAT activity was determined by measuring the decrease in absorption at 240 nm. The H2 O2 -dependent oxidation of ascorbate was followed as a decrease in absorbance at 290 nm (e, 2.8 mmol L1 cm1 ). Lipid peroxidation Lipid peroxidation was extracted using the method described by our group in 2009.28 Briey, fresh fruit were homogenized and centrifuged. The supernatant was separated and a mixture of trichloroacetic acid (TCA), thiobarbituric acid (TBA) and butylated hydroxytoluene (BHT) was added. The mixture was heated and then quickly cooled on ice. The contents were centrifuged and the absorbance was measured at 532 nm. The value for nonspecic absorption at 600 nm was subtracted. The concentration of TBARS was calculated using an extinction coefcient of 155 mmol L1 cm1 .29 Free amino acids Free amino acids were extracted from fruits frozen at 80 C: sap was extracted after vortexing at 5000 rpm (10 min, 4 C) and determined following the AccQTag-ultra ultra-performance liquid chromatography (UPLC) method.30 For derivatization, 70 L borate buffer was added to the hydrolyzed sample or to 10 L of the fruit sap. Next, 20 L reagent solution was added. The reaction mixture was mixed immediately and heated at 55 C

MATERIALS AND METHODS

Plant material and storage conditions Uniform sweet pepper plants (Capsicum annuum L.) cv. Herminio were obtained from a commercial nursery. Plants were grown in 1.2 m long coconut ber-lled bags in a greenhouse equipped with a computer-regulated drip irrigation system, under controlled environmental conditions. Each bag had three plants with three 4 L h1 drippers. Irrigation management was according to local commercial soilless cultivation and the drainage percentage was maintained at 30%.26 The pH of the nutrient solution was maintained between 5.6 and 6.0. Plants were irrigated with nutrient solution of the following composition (meq L1 ): NO3 , 12.5; H2 PO4 , 1.5; SO4 2 , 7.5; K+ , 7.5; Ca2+ , 9.5; Mg2+ , 4.5. One hundred and fty days after transplanting (DAT), fruits were harvested, weighed and stored at 5 C and 9095% relative humidity in a dark chamber for 21 days. Fruit harvesting was performed at the green stage of ripening. Treatments consisted of the application of antitranspirant (AT, pinolene 5%, as commercial preparation Vapor Gard = 96% pinolene in 4% inert ingredients), foliar urea (UR, 15 g L1 ) and non-sprayed fruits (control (CN)). At 0 (before storage), 7, 14 and 21 days after storage, 15 fruits per treatment and per day of storage were processed by measuring color, rmness, total phenolic compounds, lipid peroxidation (TBARS), amino acids, and the antioxidant enzymes catalase (CAT) and ascorbate peroxidase (APOX). Each fruit was considered a sample. Skin color and rmness Firmness was determined from the surface in the equatorial area, using a Bertuzzi FT011 penetrometer (Fruit tester, Alfonsine, Italy) tted with an 8 mm diameter probe. Color was determined with

2717

J Sci Food Agric 2010; 90: 27162722

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org for 10 min. After cooling, an aliquot of the reaction mixture was used for UPLC injection. UPLC was performed on an Acquity system (Waters, Milford, MA, USA), equipped with a uorescence detection (FLR) system. The column used was a BEH C18 100 mm 2.1 mm, 1.7 m (Waters). The ow rate was 0.7 mL min1 and the column temperature was kept at 55 C. The injection volume was 1 L. Wavelength excitation (ex ) and emission (em ) were set at 266 and 473 nm, respectively. The solvent system consisted of two eluents: (A) AccQTag-ultra eluent A concentrate (5%, v/v) and water (95%, v/v); (B) AccQTag ultra eluent B. The following gradient elution was used: 00.54 min, 99.9% A0.1% B; 5.74 min, 90.9% A9.1% B; 7.74 min, 78.8% A21.2% B; 8.04 min, 40.4% A59.6% B; 8.058.64 min. 10% A90% B; 8.7310 min, 99.9% A0.1% B. Empower 2 (Waters) software was used for system control and data acquisition. External standards (Thermo scientic) were used for quantication of (NH3 ) ammonia; (ala) alanine; (arg) arginine; (asp) aspartic acid; (cys) cysteine; (glu) glutamic acid; (gly) glycine; (his) histidine; (ile) isoleucine; (leu) leucine; (lys) lysine; (met) methionine; (phe) phenylalanine; (pro) proline; (ser) serine; (thr) threonine; (tyr) tyrosine; (val) valine. Statistical analysis Data were tested rst for homogeneity of variance and normality of distribution, and were analyzed by analysis of variance using the Duncan multiple range test to determine differences between means (P 0.05) for treatments at different days of storage. The statistical analyses were done using SPSS 12.0 (SPSS Science, Chicago, IL, USA).

P Cuadra-Crespo, FM del Amor

Figure 1. Effects of the application of antitranspirant (AT) or foliar urea (UR) on fruit rmness. CN indicates control non-sprayed fruit. Vertical bars indicate standard errors of means, only shown when larger than symbol size. Values with the same letter are not signicantly different at P < 0.05 (Duncans multiple range test).

RESULTS AND DISCUSSION

Firmness and color Fruit rmness measurement is a good way to monitor fruit softening and to predict bruising damage during harvest and postharvest handling;31 thus accelerated loss of texture is considered one of the main factors that limit the shelf-life of fresh tissue.32 In our study, storage at 5 C was effective for maintenance of fruit rmness of control fruits (non-sprayed) during 21 days (Fig. 1). Already after 7 days, a signicant increase in rmness was observed for those fruits sprayed with urea or AT compared with control fruits. After 14 days, rmness was reduced in those fruits sprayed with urea to values close to the control fruits, but AT increased fruit rmness by 29.1% compared with the non-sprayed fruits after 14 days of storage, and by 29.5% after 21 days. Lowering the temperature of non-climacteric fruits such as sweet pepper lowers their rate of ripening and deterioration.33 However, fresh peppers are highly sensitive to freezing injury and susceptible to chilling injury, and the characteristic symptoms of chilling injury in sweet pepper are softening, pitting and a predisposition to decay.34 Our study with cv. Herminio showed no variations in rmness at 5 C for control fruits, while a transient increase in rmness was observed for urea. However, a more stable and long-lasting difference compared with the control was maintained when AT was applied. Softening of fruits during ripening is characterized by the solubilization of pectins.35 Thus the observed effect might be due to an increase of the pectin viscosity of cell walls, implicating a rapid and short-lasting effect of urea due to a quick absorption, and a longer-lasting effect for AT, due to an effective covering of the skin of the fruit until 21 days with no absorption or alteration. Color change in the pepper surface takes place as a result of chlorophyll degradation and a considerable increase in carotenoid

content.16 Meir36 reported that pepper fruit harvested at the mature-green stage is sensitive to temperatures below 6 C and developed chilling injury. Color parameter L represents lightness, ranging between 0 (black) to 100 (white), and color change is observed as a decrease in L .37 Thus already after 7 days of storage at 5 C, both sprayed and non-sprayed fruits become darker but after 21 days no signicant differences were observed between the treatments (Fig. 2(A)). The parameter a showed an inverse pattern compared with the L , but after 21 days fruits treated with urea had signicantly reduced values of this parameter compared with the control fruits (Fig. 2(B)). This increase in a in control or antitranspirant treatments will result in reddish fruits and therefore an increase in ripening (these pepper fruits turn from green to red). Thus urea could delay senescence as it was able to maintain lower a values after 21 days. Additionally, the higher a values of control and antitranspirant fruits compared with urea could indicate a signicant degradation of chlorophyll pigments in these treatments. In a previous study38 we demonstrated that the chlorophyll pigment concentration was related to the supply of N to the roots, and urea could provide a source of N for maintaining these pigments. Similar results were found by del Amor et al.:28 applications of urea during pepper cultivation increased L and a and reduced b , compared with the limited N supply treatment. Some differences in the responses to urea between that study, under greenhouse conditions, and this one (postharvest) could be also due to differences in the intensity of the N stress, as our fruits were harvested from well-nourished plants and the additional effect in L could be minimized. Total phenolic compounds Phenolic compounds are secondary metabolites in plants. Their functions are not always known, but some are structural polymers, UV screens, antioxidants or attractants, and others are involved in non-specic defense mechanisms.39 The phenolic composition of fruits and hence their antioxidant properties may be modied by environmental and postharvest factors, including storage and processing.40 Also, phenolics are of great importance in determining some quality attributes and properties in fresh fruits and vegetables, like color, texture, taste and avour. Additionally,

2718

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 27162722

Postharvest quality of sweet pepper

www.soci.org a pattern similar to that of the control fruits. Shin et al.43 observed that phenolic compounds increased during storage but in our study this effect was only as a result of urea application. Padda and Picha44 showed that phenolic contents are inuenced by cultivar and other pre- and/or postharvest conditions and longterm exposure to low temperature. Cordenunsi et al.45 found that total phenolics contents remained constant or even decreased. In contrast, Robards et al.40 found that total phenol content increased signicantly in apples stored at 0 C, while AyalaZavala et al.46 found that total phenolics increased during cold storage of strawberry fruits. In our control fruits, the total phenolic compounds showed a slight tendency to increase after 7 days of storage.

Figure 2. Effects of the application of antitranspirant (AT) or foliar urea (UR) on fruit color. CN indicates control non-sprayed fruit. Vertical bars indicate standard errors of means, only shown when larger than symbol size. Values with the same letter are not signicantly different at P < 0.05 (Duncans multiple range test).

phenolic compounds may act as antioxidants in different plant reactions. Thus numerous studies have demonstrated that the accumulation of phenolic compounds such as avonoids and anthocyanins provides a defensive mechanism, and the concentration of avonoids in plants could be affected by NO3 supply.28,41 In our study, total phenolics of control fruits were not affected at 21 days of storage at 5 C (Fig. 3(A)), while a signicant increase was observed when urea was applied. Some researchers found that NO3 application was favorable to phenolic compound accumulation.42 Del Amor et al.28 showed the effect of foliar N fertilization on some phenolic compounds like anthocyanins, which increased with the application of urea. In the case of AT treatment, phenolic compounds did not change and they followed

Antioxidant enzymes Plants have an antioxidant defense system that can prevent the accumulation of ROS and repair oxidative damage. Chilling injury in plants results in elevated levels of ROS and antioxidant enzymes protect against these potentially dangerous molecules.47 Hydrogen peroxide (H2 O2 ) is a potentially toxic compound which is reduced to water by CAT and APOX.48 Thus catalase protects cells against the H2 O2 that is generated in these cells, catalyzing its conversion to H2 O and molecular O2 , and destroying toxic substances, which could enter the cells.49 However the response of CAT and APOX was not exactly the same when urea was applied to growing, N-decient fruits28 or to well-nourished fruits under storage at low temperatures. In this study, catalase was affected already after 7 days by urea or AT, while it was reduced in control fruits to a minor extent (Fig. 3(B)). Thus the catalase activity in AT fruits rst increased transiently during the rst 7 days and then decreased permanently over the next period. However, the effect of urea was the inverse of that observed for AT (while AT decreased, urea increased it) but at 21 days both treatments showed the same activity, signicantly higher than for control fruits. A similar pattern was observed for APOX activity (Fig. 3(C)). However, differences between AT and urea were still maintained at 21 days although with the opposite tendency. Imahori et al.50 found that fruit stored at 6 C showed a gradual decline in CAT and APOX activity at 15 days of storage and reported that changes in CAT activity during cold storage are also related to both the chilling resistance and the development of oxidative stress. Thus Baker51 found that CAT activity in pepper leaves declined before visible symptoms of senescence were observed, while Zilkah et al.52 also reported a benecial effect of foliar urea: increased freezing tolerance in avocado and peach. Consequently, our results indicate that the increased enzyme activity triggered by urea could confer increased protection during storage at low temperatures as, compared with the control, lower activity of defensive enzymes at chilling temperatures can impair the plants ability to break products of oxygen photoreduction.51 Additionally, Lim et al.53 found that chilling increases the level of active oxygen species (AOS) in chilling-sensitive plants and reduction of the chilling in resistant plants may be related to their ability to reduce and/or scavenge free radicals through increased enzyme activity. Thus lower activities of CAT were observed in chilling-sensitive fruit than in tolerant fruit after storage at 0 C.54 However, as previously pointed out, CAT and APOX tended to decrease with AT but increase for urea after 7 days. A different pattern during the rst 7 days of storage compared with the rest of the storage period was observed.

2719

J Sci Food Agric 2010; 90: 27162722

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org

P Cuadra-Crespo, FM del Amor

Figure 3. Effects of the application of antitranspirant (AT) or foliar urea (UR) on total phenolics compounds (A) catalase (B), ascorbate peroxidase (C) and lipid peroxidation (D). CN indicates control non-sprayed fruit. Vertical bars indicate standard errors of means, only shown when larger than symbol size. Values with the same letter are not signicantly different at P < 0.05 (Duncans multiple range test).

Lipid peroxidation Environmental stress induces active oxygen species, which could lead to oxidation of membrane lipids and disrupted membranes.55 Lipid peroxidation contributes to the development of chilling injury56 and chilling has been found to induce lipid degradation in cucumber fruit and tomato pericarp.57 The thiobarbituric acid (TBA) assay is widely used to measure thiobarbituric acid-reactive substances (TBARS) resulting from lipid oxidation, the TBARS reaction being an indicator of lipid peroxidation.58 Changes in lipid peroxidation levels in a tissue can be a good indicator of the structural integrity of the membranes of plants subjected to low temperature.59 Our data show that lipid peroxidation was reduced in all treatments during the rst 7 days of storage at low temperature. Afterwards, control fruits showed a tendency to increase TBARS, while a signicant increase was observed in fruits treated with AT after 14 days of storage (Fig. 3(D)). Fruits treated with urea showed signicantly lower values of TBARS than control or AT fruits. Oxidative stress may be dened as an increment of oxidant species and/or a depletion of antioxidant defenses. Thus this differential effect of AT could be attributed to a sharply decreasing activity of CAT and APOX observed after 7 days of storage. Thus it is envisaged that both activities in AT fruits will continue decreasing below the control levels. However, the dramatic increase in TBARS was produced before a more evident decrease of antioxidant enzymes in the AT treatment was observed. This could indicate a delayed induction of the defense system that counteracts the increase of TBARS at low temperature in this treatment. Wismer et al.60 reported that low temperatures can modify the biophysical properties of membranes via the composition of membrane lipids, which contributes to the visible symptoms of damage and induces oxidative stress in the cell. Therefore, maintenance of the membrane integrity at low temperature has been considered important in the resistance to low temperature. Thus the observed increase in TBARS could be associated with changes in fatty acid unsaturation and in the phospholipid composition of mitochondrial membranes that, in turn, cause changes in membrane uidity and the activity of respiratory

complexes.61 However, it is at least as likely that the TBARS increase resulted from peroxidation of plastid galactolipids, which are rich in linoleic and linolenics acids (source of malondialdehyde).

Amino acid composition Amino acid metabolism is one of the most important biochemical processes in plants. Free amino acids are involved in secondary plant metabolism and the biosynthesis of compounds, such as glucosinolates and phenolics, which directly or indirectly play an important role in plantenvironment interaction and human health; thus free amino acid prole determination is important.62 Additionally, amino acids are important for human nutrition and affect food quality, including taste, aroma and color.63 In this study arginine was the main free amino acid in sweet pepper and accounted for half of the total amino acid content, while values were moderate for Asp, Thr, Ala, Lys and Try and very low for Pro, Ile, Leu and Phe (Table 1). His, Gly, Glu, Cys were not detected in the sweet pepper extracts. In general, concentrations of amino acids were not affected when pepper was stored at 5 C for 21 days, and were maintained close to the initial concentrations before storage. Arginine, the main amino acid in sweet pepper, is considered a semi-essential amino acid for humans, being required to ensure that the liver, joints, muscles (including the heart muscle) and skin are kept healthy. Additionally, Oliveira et al.64 found that arginine is also of great importance as an intermediary product in urea synthesis, being (theoretically) the most efcient form of storage N because of its low C/N ratio.65 The preservation of the amino acid prole at the studied temperature is important for the nutritional characteristics of sweet pepper. Therefore, our results demonstrate that although the enzymatic metabolism related with antioxidant enzymes or color was altered by low temperature, only minor changes were observed for the amino acid composition of sweet pepper. Additionally, a positive differential effect was found in the urea treatment, which improved color and the mitigation of lipid peroxidation through increased catalase and ascorbate peroxidase.

2720

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 27162722

Postharvest quality of sweet pepper

www.soci.org

Table 1. Effect of foliar urea (UR) and antitranspirant (AT) on the free amino acid concentration of sweet pepper at 7, 14 and 21 days of storage at 5 C Free amino acids (mol L1 ) Days 0 7 CN UR AT CN UR AT CN UR AT Treatment Ser 5.08 4.31 4.25 5.59 4.60 4.33 5.63 5.29 5.98 5.70 Arg 16.13ab 15.89ab 16.42ab 18.86bc 16.00ab 16.44ab 22.67c 16.02ab 13.00a 15.37ab Asp 4.11a 5.22bc 5.87cd 6.18d 4.11a 5.66bcd 6.25d 4.91ab 5.12bc 5.92cd Thr 2.54 2.35 2.52 2.92 2.31 1.97 2.54 2.38 2.68 2.58 Ala 3.01 3.30 3.34 3.35 3.19 2.56 3.31 3.45 3.58 3.18 Pro 0.00 0.00 0.00 0.25 0.00 0.00 0.00 0.25 0.27 0.00 Lys 2.15 1.97 1.74 2.29 1.92 1.90 2.71 2.50 2.34 1.59 Tyr 0.42a 3.67d 1.97abcd 2.53bcd 1.41abc 1.65abc 3.19cd 0.95ab 0.96ab 1.45abc Val 2.02a 2.55b 1.87a 2.03a 2.52b 1.86a 1.74a 1.79a 1.78a 1.68a Ile 0.00a 0.00a 0.00a 0.23ab 0.00a 0.00a 0.00a 0.00a 0.48b 0.00a Leu 0.98 0.24 0.00 0.70 0.54 0.24 0.74 0.30 1.04 1.12 Phe 0.22 0.00 0.00 0.24 0.00 0.00 0.24 0.24 0.25 0.00

14

21

Values with the same letter within the same column are not signicantly different at P < 0.05 (Duncans multiple range test).

ACKNOWLEDGEMENTS
Paula Cuadra-Crespo is the recipient of a pre-doctoral fellowship from the IMIDA. The authors thank MC Pinero for his technical assistance. This work has been supported by the Instituto Nacional de Investigaciones Agrarias (INIA), through project RTA2008-00089 and POI 07-021. Part of this work was also funded by the European Social Fund.
13 14

15

16

REFERENCES
1 Banaras M, Bosland PW and Lownds NK, Effects of harvest time and growth conditions on storage and post-storage quality of fresh peppers (Capsicum annuum L.). Pak J Bot 37:337344 (2005). 2 Smith DL, Stommel JR, Fung RWM, Wang CY and Whitaker BD, Inuence of cultivar and harvest method on postharvest storage quality of pepper (Capsicumannuum L.) fruit. PostharvestBiolTechnol 42:243247 (2006). 3 Daz-Perez JC, Muy-Rangel MD and Mascorro AG, Fruit size and stage of ripeness affect postharvest water loss in bell pepper fruit (Capsicum annuum L.). J Sci Food Agric 87:6873 (2007). 4 Paull RE, Chilling injury of crops of tropical and subtropical origin, in Chilling Injury of Horticultural Crops, ed. by Wang CY. CRC Press, Boca Raton, FL, pp. 1736 (1990). 5 Kehr E, Susceptibilidad a dano por enfriamiento en poscosecha de pimiento y tratamientos para disminuir su efecto. Agric T c e 62:509518 (2002). 6 Gonz lez-Aguilar GA, Gayoso L, Cruz R, Fortiz J, B ez R and Wang CY, a a Polyamines induced by hot water treatments reduce chilling injury and decay in pepper fruit. Postharvest Biol Technol 18:1926 (2000). 7 Singh BK, Sharma SR and Singh B, Combining ability for superoxide dismutase, peroxidase and catalase enzymes in cabbage head (Brassica oleracea var. capitata L.). Sci Hortic 122:195199 (2009). 8 Murcia MA, Jim nez AM and Martnez-Tom M, Vegetables e e antioxidant losses during industrial processing and refrigerated storage. Food Res Int 42:10461052 (2009). 9 Boonsiri K, Ketsa S and Van Doorn WG, Seed browning of hot peppers during low temperature storage. Postharvest Biol Technol 45:358365 (2007). 10 Koc A, Gasch AP, Rutherford JC, Kim HY and Gladyshev VN, Methionine sulfoxide reductase regulation of yeast lifespan reveals reactive oxygen species-dependent and -independent components of aging. Proc Natl Acad Sci USA 101:79998004 (2004). 11 Nunez M, Mazzafera P, Mazorra LM, Siquira WJ and Zullo MAT, Inuence of a brassinosteroid analogue on antioxidant enzymes in rice grown in culture medium with NaCl. Biol Plant 47:6770 (2003). 12 Ogiso Y, Hosoda-Yabe R, Kawamoto Y, Kawamoto T, Kato K and Yabe T, An antioxidant of dried chili pepper maintained its activity 17

18

19

20

21

22

23

24

25 26

27

through postharvest ripening for 18 months. Biosci Biotechnol Biochem 72:32973300 (2008). Knoche M, Petracek PD and Bukovac MJ, Urea penetration of isolated tomato fruit cuticles. J Am Soc Hortic Sci 119:761764 (1994). Crisosto CH and Mitchell JP, Preharvest factors affecting fruit and vegetable quality, in Postharvest Technology of Horticultural Crops, ed. by Kader AA. University of California, Oakland, CA, p. 519 (2002). del Amor FM, Yield and fruit quality response of sweet pepper to organic and mineral fertilization. Renewable Agric Food Syst 22:233238 (2006). Perez-Lopez AJ, Lopez-Nicolas JM, Nunez-Delicado E, Del Amor FM and Carbonell-Barrachina AA, Effects of agricultural practices on color, carotenoids composition, and minerals contents of sweet peppers, cv. Almuden. J Agric Food Chem 55:81588164 (2007). Meheriuk M, McKenzie DL, Neilsen GH and Hall JW, Fruit pigmentation of four green apple cultivars responds to urea sprays but not to nitrogen fertilization. HortScience 31:992993 (1996). Yildirim E, Guvenc I, Turan M and Karatas A, Effect of foliar urea application on quality, growth, mineral uptake and yield of broccoli (Brassica oleracea L., var. italica). Plant Soil Environ 53:120128 (2007). Kaya C and Higgs D, Relationship between water use and urea application in salt-stressed pepper plants. J Plant Nutr 26:1930 (2003). Harris JR and Bassuk NL, Effects of defoliation and antitranspirant treatments on transplant response of scarlet oak, green ash and Turkish hazelnut. J Arboric 21:3336 (1995). Gale J and Poljakoff-Mayber A, Effect of antitranspirant spray on the uptake and transport of rubidium by plants. Plant Cell Physiol 4:285288 (1963). del Amor FM and Rubio JS, Effects of antitranspirant spray and potassium : calcium:magnesium ratio on photosynthesis, nutrient and water uptake, growth, and yield of sweet pepper. J Plant Nutr 32:97111 (2009). Phelps E and Morelock T, Antitranspirants as a selection technique in breeding for tomato fruit cracking resistance. HortScience 22:728733 (1987). Lipe WN, Hondnet K, Gerst M and Wendt CW, Effects of antitranspirants on water use and yield of greenhouse and eld grown onion. HortScience 17:242244 (1982). Byari SH and Okeefe RB, Effect of antitranspirants on heat and drought response of potato varieties. HortScience 17:513523 (1982). del Amor FM and Gomez-Lopez MD, Agronomical response and water use efciency of sweet pepper plants grown in different greenhouse substrates. Hortscience 44:810814 (2009). K hkonen MP, Hopia AI, Vuorela HJ, Rauha JP, Pihlaja K, Kujala TS, a et al, Antioxidant activity of plant extracts containing phenolic compounds. J Agric Food Chem 47:39543962 (1999).

2721

J Sci Food Agric 2010; 90: 27162722

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org
28 del Amor FM, Cuadra-Crespo P, Varo P and Gomez MC, Inuence of foliar urea on the antioxidant response and fruit color of sweet pepper under limited N supply. J Sci Food Agric 89:504510 (2009). 29 Heath RL and Packer L, Photoperoxidation in isolated chloroplasts. I. Kinetics and stoichiometry of fatty acid peroxidation. Arch Biochem Biophys 125:189198 (1968). 30 Waters, UPLC AminoAcidAnalysis Solution. Waters Corporation, Milford, MA (2006). 31 ElsevieValero C, Crisosto CH and Slaughter D, Relationship between nondestructive rmness measurements and commercially important ripening fruit stages for peaches, nectarines and plums. Postharvest Biol Technol 44:248253 (2007). 32 King AD and Bolin HR, Physiological and microbiological storage stability of minimally processed fruits and vegetables. Food Technol 43:132135 (1989). 33 Kays SJ, Post-harvest Physiology and Handling of Perishable Plant Products. Van Nostrand-Reinhold, New York, p. 733 (1991). 34 Encyclopedia of Fruits and Nuts, ed. by Janick J and Paull RE. Purdue University, West Lafayette. IN; University of Hawaii at Manoa, HI, p. 976 (2008). 35 Von Mollendorff LJ, De Villiers OT, Jacobs G and Westraad I, Molecular characteristics of pectic constituents in relation to rmness, extractable juice, and woolliness in nectarines. J Am Soc Hortic Sci 118:7780 (1993). 36 Meir S, Rosenberger I, Aharon Z, Grinberg S and Fallik E, Improvement of the postharvest keeping quality and colour development of bell pepper (cv. Maor) by packaging with polyethylene bags at a reduced temperature. Postharvest Biol Technol 5:303309 (1995). 37 Cox KA, McGhie TK, White A and Woolf AB, Skin colour and pigment changes during ripening of Hass avocado fruit. Postharvest Biol Technol 31:287294 (2004). 38 del Amor FM, Growth, photosynthesis and chlorophyll uorescence of sweet pepper plants as affected by the cultivation method. Ann Appl Biol 148:133139 (2006). 39 Papoulias E, Siomos AS, Koukounaras A, Gerasopoulos D and Kazakis E, Effects of genetic, pre- and post-harvest factors on phenolic content and antioxidant capacity of white asparagus spears. Int J Mol Sci 10:53705380 (2009). 40 Robards K, Prenzler PD, Tucker G, Swatsitang P and Glover W, Phenolic compounds and their role in oxidative processes in fruits. FoodChem 66:401436 (1999). 41 Chimphango SBM, Musil CF and Dakora FD, Response of purely symbiotic and NO3 fed nodulated plants of Lupinus luteus and Vicia atropurpurea to ultraviolet-B radiation. J Exp Bot 54:17711784 (2003). 42 Yang SH, Tao J, Liu XF, Guo DA and Zheng JH, Effects of carbon source and nitrogen source on callus growth and avonoid content in Glycyrrhiza uralensis. J Chin Mater Med 31:18571859 (2006). 43 Shin Y, Liu RH, Nock JF, Holliday D and Watkins CB, Temperature and relative humidity effects on quality, total ascorbic acid, phenolics and avonoid concentrations, and antioxidant activity of strawberry. Postharvest Biol Technol 45:349357 (2007). 44 Padda MS and Picha DH, Effect of low temperature storage on phenolic composition and antioxidant activity of sweet potatoes. Postharvest Biol Technol 47:176180 (2008). 45 Cordenunsi BR, Genovese MI, do Nascimento JR, Hassimotto NM, dos Santos RJ and Lajolo FM, Effects of temperature on the chemical composition and antioxidant activity of three strawberry cultivars. Food Chem 91:113121 (2005). 46 Ayala-Zavala JF, Wang SY, Wang CY and Gonz lez-Aguilar GA, Effect a of storage temperatures on antioxidant capacity and aroma compounds in strawberry fruit. Lebensm Wiss Technol 37:687695 (2004).

P Cuadra-Crespo, FM del Amor

47 Wongsheree T, Ketsa S and van Doorn WG, The relationship between chilling injury and membrane damage in lemon basil (Ocimum citriodourum) leaves. Postharvest Biol Technol 51:9196 (2009). 48 Imahori Y, Kanetsune Y, Ueda Y and Chachin K, Changes in hydrogen peroxide content and antioxidative enzyme activities during the maturation of sweet pepper (Capsicum annuum L.) fruit. J Japan Soc Hortic Sci 69:690695 (2000). 49 Cadenas E and Davies JA, Mitochondrial free radical generation: oxidative stress and aging. Free Radic Biol Med 29:220230 (2000). 50 Imahori Y, Takemura M and Bai J, Chilling-induced oxidative stress and antioxidant responses in mume (Prunus mume) fruit during low temperature storage. Postharvest Biol Technol 49:5460 (2008). 51 Baker NR, Chilling stress and photosynthesis, in Causes of Photooxidative Stress and Amelioration of Defense Systems in Plants, ed. by Foyer CH and Mullineaux PM. CRC Press, Boca Raton, FL, pp. 127154 (1994). 52 Zilkah S, Wiesmann Z, Klein I and David I, Foliar applied urea improves freezing protection to avocado and peach. Sci Hortic 66:8592 (1996). 53 Lim CS, Kang SM and Cho JL, Antioxidizing enzyme activities in chilling-sensitive and chilling-tolerant pepper fruit as affected by stage of ripeness and storage temperature. J Am Soc Hortic Sci 134:156163 (2009). 54 Ju Z, Yuan Y, Liou C, Zhan S and Xiu S, Effects of low temperature on H2 O2 and heart browning of Chili and Yali pear (Pyrus bretscheideri R.). Sci Agric Sinica 27:7781 (1994). 55 Blokhina O, Virolainen E and Fagerstedt KV, Antioxidants, oxidative damage and oxygen deprivation stress: a review. Ann Bot 91:179194 (2003). 56 Wang CY, Kramer GF, Whitaher BD and Lusby WR, Temperature preconditioning increases tolerance to chilling injury and alters lipid composition in zucchini squash. J Plant Physiol 140:229235 (1992). 57 Wang CY, Effects of temperature preconditioning on catalase, peroxidase, and superoxide dismutase in chilled zucchini squash. Postharvest Biol Technol 5:6776 (1995). 58 Davey MW, Stals E, Panis B, Keulemans J and Swennen RL, Highthroughput determination of malondialdehyde in plant tissues. Anal Biochem 347:201207 (2005). 59 Posmyk MM, Bailly C, Szafranska K, Janas KM and Corbineau F, Antioxidant enzymes and isoavonoids in chilled soybean (Glycine max. (L.) Merr.) seedlings. J Plant Physiol 162:403412 (2005). 60 Wismer WV, Worthing WM, Yada RY and Marangoni AG, Membrane lipid dynamics and lipid peroxidation in the early stages of lowtemperature sweetening in tubers of Solanum tuberosum. Physiol Plant 102:396410 (1998). 61 Gualanduzzi S, Baraldi E, Braschi I, Carnevali F, Gessa CE and De Santis A, Respiration, hydrogen peroxide levels and antioxidant enzyme activities during cold storage of zucchini squash fruit. Postharvest Biol Technol 52:1623 (2009). 62 Gomes MH and Rosa E, Free amino acid composition in primary and secondary inorescences of 11 broccoli (Brassica oleracea var. italica) cultivars and its variation between seasons. J Sci Food Agric 81:295299 (2000). 63 Belitz HD and Grosch W, Amino acids, peptides, and proteins, Food Chemistry. Springer, Berlin, pp. 834 (1999). 64 Oliveira AP, Pereira DM, Andrade PB, Valenta P, Sousa C, Pereira JA, et al, Free amino acids of tronchuda cabbage (Brassica oleracea L. Var. costata DC): inuence of leaf position (internal or external) and collection time. J Agric Food Chem 56:52165221 (2008). 65 Titus JS and Kang SM, Nitrogen metabolism, translocation, and recycling in apple trees. Hortic Rev 4:204246 (1982).

2722
wileyonlinelibrary.com/jsfa
c 2010 Society of Chemical Industry J Sci Food Agric 2010; 90: 27162722

Short Communication
Received: 21 May 2010 Revised: 5 July 2010 Accepted: 9 July 2010 Published online in Wiley Online Library: 5 August 2010

(wileyonlinelibrary.com) DOI 10.1002/jsfa.4119

Factors affecting ANKOM ber analysis of forage and browse varying in condensed tannin concentration
Thomas H Terrill,a Richard M Wolfeb and James P Muirb
Abstract
BACKGROUND: Browse species containing condensed tannins (CTs) are an important source of nutrition for grazing/browsing livestock and wildlife in many parts of the world, but information on ber concentration and CTber interactions for these plants is lacking. RESULTS: Ten forage or browse species with a range of CT concentrations were oven dried and freeze dried and then analyzed for ash-corrected neutral detergent ber (NDFom) and corrected acid detergent ber (ADFom) using separate samples (ADFSEP) and sequential NDF-ADF analysis (ADFSEQ) with the ANKOM ber analysis system. The ADFSEP and ADFSEQ residues were then analyzed for nitrogen (N) concentration. Oven drying increased (P < 0.05) ber concentrations with some species, but not with others. For high-CT forage and browse species, ADFSEP concentrations were greater (P < 0.05) than NDFom values and approximately double the ADFSEQ values. Nitrogen concentration was greater (P < 0.05) in ADFSEP than ADFSEQ residues, likely due to precipitation with CTs. CONCLUSION: Sequential NDF-ADF analysis gave more realistic values and appeared to remove most of the ber residue contaminants in CT forage samples. Freeze drying samples with sequential NDF-ADF analysis is recommended in the ANKOM ber analysis system with CT-containing forage and browse species. c 2010 Society of Chemical Industry Keywords: acid detergent ber; ANKOM ; freeze drying; neutral detergent ber; condensed tannins; oven drying

INTRODUCTION
Browse species containing condensed tannins (CTs), including herbaceous legumes and woody species, are an important component of livestock and wildlife diets throughout the world1,2 but there is little information available on the nutritional adequacy of CT-containing browse species. There have been a number of reports on the potential benets of including low to medium levels of CT-containing forage and browse species in ruminant diets, including reducing bloat,3 production of greenhouse gases,4 and infection by gastrointestinal nematodes,5,6 as well as improving protein utilization efciency and increasing weight gains, milk production, and reproductive performance of livestock.7 9 Intake of high CT-containing forages has been improved by sun-drying10 or treating with polyethylene glycol to bind the tannins.11 As interest in CT forage and browse species as components of the diet for both domestic livestock and ruminant wildlife continues to grow, accurate laboratory analysis of quality indices for these forages is critical. Muir et al.12 and Wolfe et al.13 reported a wide range in CT and N concentrations of several native, herbaceous legumes from Texas, USA, but information on ber concentration and possible CTber interactions is limited for these and other browse species found throughout the world. The ANKOM detergent ber analysis system with lter bag technology14,15 has replaced the crucible method16 in many herbage nutritive value laboratories throughout the world, although both are based on the same general chemical

principles.17 Terrill and Koivisto18 reported higher ADFSEP than ADFSEQ values for oven-dried cool-season and warm-season pasture legumes high in CT using the ANKOM method, but little information is available on the application of this system to CT-containing browse species. The purpose of the current investigation was to measure the effect of herbage sample drying method and ANKOM detergent ber analysis method on resulting ber constituents from herbaceous and woody species varying in CT concentration.

MATERIALS AND METHODS

Leaf material from 10 native and introduced plant species varying in CT concentration and consumed by goats and deer in central Texas (Table 1) were collected from three replicates of 1.5 3.0 m plots at the Texas AgriLife Research Center, Stephenville, Texas.

Correspondence to: James P Muir, Texas AgriLife Research, Texas A&M System, 1229 North U.S. Highway 281, Stephenville, TX 76401, USA. E-mail: j-muir@tamu.edu

a Agricultural Research Station, Fort Valley State University, Fort Valley, GA 31030, USA

2723

b Texas AgriLife Research, Texas A&M System, 1229 North U.S. Highway 281, Stephenville, TX 76401, USA

J Sci Food Agric 2010; 90: 27232726

www.soci.org

c 2010 Society of Chemical Industry

www.soci.org

TH Terrill, RM Wolfe and JP Muir

Table 1. Names, growth habits and condensed tannin (CT) concentrations of herbage species used in the trial Species Scientic name Acacia angustissima var. hirta Desmodium paniculatum Lespedeza procumbens Lespedeza stuevei Smilax rotundifolia Quercus sinuata var.breviloba Leucaena retusa Gleditsia tricanthos Cynodon dactylon Panicum virgatum

CT concentration Common name Eastern prairie acacia Panicled tick-clover Creeping bush-clover Tall bush-clover Common smilax Shin oak Golden-ball lead-tree Common Honey-locust Common Bermudagrass Switchgrass cv. Alamo Type Herbaceous perennial legume Herbaceous perennial legume Herbaceous perennial legume Herbaceous perennial legume Perennial vine Woody perennial Woody perennial Woody perennial Perennial grass Perennial grass (g kg1 ) 5 121 66 123 34 125 13 92 Trace Trace

Total CT determined according to Terrill et al.21 using a self-standard except for species without measurable CT.

For woody perennials and Smilax, three different plants in a 10 km radius of Stephenville were used as replications, with leaf material collected below the approximate browse height for goats (1.5 m). For non-CT controls, leaf material of two species of grasses was collected (Table 1). Half of each sample was oven dried at 55 C for 48 h, and the other half was freeze dried for 72 h. The dry samples were then ground with a Wylie mill to pass a 1 mm mesh screen and stored in air-tight plastic bottles at room temperature. Samples were analyzed for dry matter using AOAC procedures,19 and neutral detergent ber (NDF), corrected acid detergent ber (ADFom) using separate samples (ADFSEP), and sequential NDFom-ADFom (ADFSEQ) concentrations using the ANKOM Model 200 and the ber bag technique developed by ANKOM (ANKOM Technology, Macedon, NY, USA). Heat-stable -amylase and sodium sulte were added to the neutral detergent solution as recommended with the ANKOM system.20 After ADFSEP and ADFSEQ extractions were completed, the ber mat was removed for each sample and analyzed for total N using an Elementar Vario Macro combustion analyzer (Elementar Americas, Inc., Mt. Laurel, NJ, USA). All NDF and ADF residues were analyzed and corrected for ash content.19 The ash-corrected values were expressed as NDFom and ADFom. The freeze-dried samples were analyzed for total condensed tannin (TCT) concentration using the method of Terrill et al.,21 with puried CTs from each species used as standards.13 The data were analyzed as a completely randomized block design with three replicates (plots or cuttings) using the GLM procedure of SAS.22 Differences were considered signicant at P < 0.05 and means were separated using the LSMeans procedure where appropriate. For NDFom data, species and drying method were included in the model, while species, drying method, and ADF analysis method were included in the model for ADFom and ADFN data.

Drying method had no effect on NDFom concentration of herbaceous legumes and grasses, but oven drying increased NDFom concentration compared with freeze drying for woody species. For herbaceous native legumes, differences in ADFom due to drying method increased as CT levels increased, with greater ADFSEP in oven-dried than freeze-dried bush-clovers, which have greater CTs than eastern prairie acacia.13 These differences were removed with sequentially analyzed ADFom for seven of the ten species tested. For woody species, ADFSEP and ADFSEQ residues were greater in oven-dried than freeze-dried material, with greater differences in the ADFSEP residues. Drying method and ADFom analysis protocol had no effect on ADFom residues of grasses. Drying method also had little effect on ADFSEP-N and ADFSEQ-N values for the species tested, but N concentrations were approximately double in ADFSEP compared with ADFSEQ residues.

DISCUSSION
The TCT levels in these plants were generally lower than TCT values reported for similar browse species analyzed using puried Quebracho CT as the standard.13 Although it is more time consuming to purify CTs from individual species to use as standards, this provides a more accurate and meaningful analysis of TCT using the butanol HCl colorimetric method.13 Forage drying method (oven dried versus freeze dried) had a greater effect on ANKOM ber analysis of woody species than herbaceous forages, but the major effect of CT in this study appears to be on ADFSEP and ADFSEQ values, with much greater ADFSEP concentrations for most of the browse species tested, regardless of drying method (Table 2). In fact, ADFSEP was greater than NDFom in six of eight herbaceous native legumes and/or woody species tested. As hemicellulose concentration of forages is estimated by subtracting ADF from NDF, these results would give negative values for this ber fraction, which is not possible. Similar results have been reported previously for herbaceous and brushy CT-containing species using the crucible method of detergent ber analysis20 but the differences were greater using the ANKOM system in the current investigation. The contaminants in ADFSEP residues were N and other materials, most likely CTs. Terrill et al.23 reported higher N and

RESULTS
Total condensed tannin concentration ranged from trace amounts in grass samples to 125 g kg1 in shin oak (Table 1). There was a wide range of TCT levels in both the herbaceous legumes (5123 g kg1 ) and woody species (13125 g kg1 ) tested. There was a cultivar drying method interaction for NDFom data, while a cultivar drying method ADF analysis method interaction was also a factor for ADFom and ADFN (Table 2).

2724

wileyonlinelibrary.com/jsfa

c 2010 Society of Chemical Industry

J Sci Food Agric 2010; 90: 27232726

Condensed tannins and ANKOM ber analysis

www.soci.org

Table 2. Ankom detergent ber analysis (g kg1 dry matter) of oven-dried (OD) and freeze-dried (FD) forage and browse species with a range of condensed tannin levels Drying Species Eastern prairie acacia Panicled tick-clover Creeping bush-clover Tall bush-clover Common smilax Shin Oak Golden-ball lead-tree Honey locust Common bermudagrass Alamo switchgrass method OD FD OD FD OD FD OD FD OD FD OD FD OD FD OD FD OD FD OD FD NDFom 250a 243a 276a 291a 322a 290a 285a 284a 391a 292b 406a 353b 336a 250b 440a 333b 611a 595a 581a 560a Constituents of the Ankom Fiber Analysis System ADFSEP 150aA 197bA 451aA 455aA 458aA 412bA 398aA 434bA 551aA 443bA 351aA 262bA 375aA 352aA 506aA 386bA 307aA 299aA 294aA 297aA ADFSEQ 93aB 96aB 181aB 194aB 206aB 196aB 201aB 207aB 304aB 227bB 240aB 216bB 186aB 159aB 285aB 231bB 280aA 274aA 262aA 264aA ADFSEP N 32.9aA 44.7bA 25.2aA 29.6bA 25.9aA 25.2aA 22.1aA 23.9aA 25.0aA 22.2aA 15.5aA 9.4bA 39.6aA 41.8aA 25.6aA 21.3bA 4.7aA 4.8aA 3.8aA 3.7aA ADFSEQ N 25.6aB 20.9bB 11.9aB 11.6aB 11.8aB 11.3aB 9.8aB 9.0aB 19.5aB 7.8bB 9.8aB 8.7aB 14.4aB 11.9bB 18.5aB 10.5bB 4.5aA 4.0aA 3.6aA 3.9aA

Standard error for cultivar drying method interaction (P < 0.001) for NDFom data = 13.2 g kg1 dry matter. Standard error for cultivar drying method ADF analysis method interactions (P < 0.05) for ADFom and ADFN data = 12.9 and 1.3 g kg1 dry matter, respectively. Means within columns for each species comparing OD and FD followed by different lower-case letters differ at P < 0.05. Means within lines under columns ADFSEP and ADFSEQ or under columns ADFSEP N and ADFSEQ N followed by different upper-case letters differ at P < 0.05. NDFom, neutral detergent ber; ADFSEP, acid detergent ber (ADFom), separate sample analysis; ADFSEQ, ADFom sequential analysis; N, nitrogen.

CT concentrations in NDF and ADFSEP residues from the high CT plant sericea lespedeza [Lespedeza cuneata (Dum-Cours.) G. Don.] after detergent ber analysis using the crucible method. In their study, the CTs were recovered in the lignin fraction. In a crucible method study with a range of CT-containing plants, PaganRiestra et al.20 reported that adding sodium sulte to NDF analysis followed by sequential ADF analysis removed most of the ber residue contaminants and gave more accurate ber estimates. The current investigation conrmed that a similar protocol should be used for the ANKOM analysis system. With this system, sodium sulte and -amylase are routinely added to the neutral detergent extraction step, and after rinsing and drying the lter bags with NDF residues, these residues are then extracted with ADF solution. This protocol greatly reduced N in ADF residues and lowered the ADF values by over 50% for both oven-dried and freeze-dried samples for several of the forage/browse species containing CT in the current investigation. As the ANKOM detergent ber analysis system continues to grow in popularity in animal nutrition laboratories around the world, as does the interest and benets of including CT-containing plants in ruminant diets, the following analytical protocols that provide realistic ber values for these forage and browse species is critical. This is particularly true for ADF, as it is often used to estimate digestibility of forages. With separate NDF-ADF analysis using the ANKOM system, inated ADF values would lead to an underestimation of TDN for CT-containing forage and browse species.

CONCLUSIONS
Separate sample ADFom analysis using the ANKOM system can increase ber estimates in CT-containing herbaceous legume and browse-type plant species. Use of freeze drying and sequential NDF-ADF analysis minimizes ADFom contaminants and gives more realistic ber estimates when using the ANKOM ber analyzer with plant material containing CTs.

REFERENCES
1 van Rooyen N, Grunow JO and Theron GK, Veld management, in Game Ranch Management, ed. by Bothma J du P. J.L. van Shaik, Pretoria, pp. 567607 (1989). 2 Berg WA, Native forb persistence under grazing of a southern Great Plains planting, in Proceedings of the Vth Int. Rangeland Congress, Society for Range Management, Denver, Colorado. Society for Range Management, Wheat Ridge CO., USA, pp. 4647 (1996). 3 Min BR, Pinchak WE, Fulford JD and Puchala R, Wheat pasture bloat dynamics, in vitro ruminal gas production, and potential bloat mitigation with condensed tannins. JAnimSci 83:13221331 (2005). 4 Puchala R, Min BR, Goetsch AL and Sahlu T, The effect of a condensed tannin-containing forage on methane emission by goats. J Anim Sci 83:182186 (2005). 5 Shaik SA, Terrill TH, Miller JE, Kouakou B, Kannan G, Kaplan RM, et al, Sericea lespedeza hay as a natural deworming agent against gastrointestinal nematode infection in goats. Vet Parasitol 139:150157 (2006). 6 Lange K, Olcott DD, Miller JE, Mosjidis JA, Terrill TH, Burke JM, et al, Effect of sericea lespedeza (Lespedeza cuneata) fed as hay, on natural and experimental Haemonchus contortus infections in lambs. Vet Parasitol 141:273278 (2006).

2725

J Sci Food Agric 2010; 90: 27232726

c 2010 Society of Chemical Industry

wileyonlinelibrary.com/jsfa

www.soci.org
7 Waghorn GC, Ulyatt MJ, John A and Fisher MT, The effect of condensed tannins on the site of digestion of amino acids and other nutrients in sheep fed on Lotus corniculatus L. Br J Nutr 57:115126 (1987). 8 Min BR, McNabb WC, Barry TN, Kemp PD, Waghorn GC and McDonald MF, The effect of condensed tannins in Lotus corniculatus upon reproductive efciency and wool production in sheep during late summer and autumn. J Agric Sci 132:323334 (1999). 9 Barry TN and McNabb WC, The implications of condensed tannins on the nutritive value of temperate forages fed to ruminants. Br J Nutr 81:263272 (1999). 10 Terrill TH, Windham WR, Evans JJ and Hoveland CS, Condensed tannins in sericea lespedeza: Effect of preservation method on tannin concentration. Crop Sci 30:219224 (1990). 11 Jones WT and Mangan JL, Complexes of the condensed tannins of sainfoin (Onobrychis viciifolia Scop.) with fraction 1 leaf protein and with submaxillary mucoprotein, and their reversal by polyethylene glycol and pH. J Sci Food Agric 28:126136 (1977). 12 Muir JP, Taylor J and Interrante SM, Herbage and seed from native perennial herbaceous legumes of Texas. Range Ecol Manage 58:643651 (2005). 13 Wolfe RM, Terrill TH and Muir JP, Season and drying method effects on condensed tannin levels in perennial herbaceous legumes. J Sci Food Agric 88:10601067 (2008). 14 Komarek AR, Robertson JB and Van Soest PJ, A comparison of methods for determining ADF using the lter bag technique versus conventional ltration. J Dairy Sci 77:2426 (1993). 15 Vogel KP, Pederson JF, Masterson SD and Toy JJ. Evaluation of a lter bag system for NDF, ADF, and IVDMD forage analysis. Crop Sci 39:276279 (1999).

TH Terrill, RM Wolfe and JP Muir

16 Van Soest PJ. Use of detergents in the analysis of brous feeds. 1. Preparation of ber residues of low nitrogen content. J Assoc Off Anal Chem 46:825829 (1963). 17 Van Soest PJ, Robertson JB and Lewis BA, Methods for dietary ber, neutral detergent ber, and nonstarch polysaccharides in relation to animal nutrition. J Dairy Sci 74:35833597 (1991). 18 Terrill TH, and Koivisto JM, Comparison of the ANKOM , brecaps, and brebags systems for detergent ber analysis of condensed tannin (CT)-containing and non-CT forages. Proc Am For Grassl Coun 12:263267 (2003). 19 Association of Ofcial Analytical Chemists, Ofcial Methods of Analysis, 14th edition. AOAC, Washington, DC, pp. 129130 (1984). 20 Pagan-Riestra S, Wolfe RM, Terrill TH and Muir JP, Effect of drying method and assay methodology on detergent ber analysis in plants containing condensed tannins. Anim Feed Sci Technol 154:119124 (2009). 21 Terrill TH, Rowan AW, Douglas GB and Barry TN, Determination of extractable and bound condensed tannin concentration in forage plants, protein concentrate meals, and cereal grains. J Sci Food Agric 58:321329 (1992). 22 SAS Institute, SAS/STAT Software: changes and enhancements. Release 6.07, SAS Technical Report, SAS Institute, Cary, NC (1992). 23 Terrill TH, Windham WR, Evans JJ and Hoveland CS, Effect of drying method and condensed tannin on detergent ber analysis of sericea lespedeza. J Sci Food Agric 66:337343 (1994).

2726
wileyonlinelibrary.com/jsfa
c 2010 Society of Chemical Industry J Sci Food Agric 2010; 90: 27232726