J Sci Food Agric 1991,57,417426

Distribution of Bound and Free Phenolic Acids in Oranges

(Citrus sinensis) and Grapefruits (Citrus paradisi)

Hanna Peleg," Michael Naim,"* Russell L Rouse@

and Uri Zehavi"
"Department of Biochemistry and Human Nutrition, Faculty of Agriculture, The Hebrew
University of Jerusalem, Rehovot 76100, Israel, and bCitrusResearch and Education Center,
Institute of Food and Agricultural Sciences, University of Florida, Lake Alfred,
Florida 33850, USA

(Received 29 June 1990; revised version received 20 June 1991;

accepted 12 August 1991)

A BSTR A CT

Free phenolic acids may be the precursors for vinyl phenols and off-flavours
formed in citrus products during storage. Quantitative determination of free
and bound phenolic acids in fruit parts of grapefruit (Citrus paradisi
Macfadyen) and oranges (Citrus sinensis ( L ) Osbeck) was performed by
extraction with ethyl acetate, silica gel column chromatography and HPLC
analyses of samples before and after alkaline hydrolysis. The content of
free and bound phenolic acids was further determined in juice derived from
fruit harvested early, mid and late in season. As found previously for ferulic
acid, phenolic acids occur mainly in bound forms in grapefruits and oranges.
In both fruits the peels contained the major portion of cinnamic acids
compared with the endocarp, and theJavedo was richer in hydroxycinnamic
acids than the albedo. In most cases, hydroxycinnamic acid content was in
thefollowing order:ferulic acid> sinapic acid> coumaric acid> caffeic acid.
Results showed that the content of bound cinnamic acids was unchanged or
slightly elevated from early to late season. However, the content of free
acids was reduced during that period.

Key words: Hydroxycinnamic acids, ferulic, sinapic, coumaric, caffeic,

orange, grapefruit, peel, endocarp, phenolic, off-flavour.

* To whom correspondence should be addressed.

417

J Sci Food Agric 0022-5142/91/$03.500 SCI, 1991. Printed in Great Britain

418 H Peleg, M Naim, R L Rouseff, U Zehavi

INTRODUCTION

Phenolic acids and their derivatives are widely distributed in plants. Their levels
vary dramatically, especially as influenced by factors such as germination, ripening,
storage and processing (Maga 1978). Phenolic acids are present in all plant-derived
food systems and in most diets. These compounds have been implicated as possibly
influencing the toxicological, nutritional, colouring, sensory and antioxidant
properties of foods with which they are associated (Maga 1978; Herrmann 1990a,b).
Hydroxycinnamic acids occur naturally in a wide range of combined forms.
They may be conjugated with organic acids, sugars, amino compounds, lipids,
terpenoids and other phenolics (Harborne 1980). They have been shown to be
directly involved in the biosynthesis of coumarins, flavonoids and lignin. Questions
concerning the mode of linkage between phenolic acids and carbohydrates in plant
cell walls have attracted considerable interest in recent years. The acids are
considered to be present as phenolic-carbohydrate esters because they are released
following alkaline hydrolysis. Modern chemical and physical (particularly NMR)
methods (Ishii and Hiroi 1990) have been extensively employed in structure
elucidation of phenolic-carbohydrate esters.
Stohr and Herrmann (1975a) reported the total quantity of phenolic acids in
oranges, grapefruits and lemons. However, data on the amount of free forms were
not available. We previously proposed (Naim et a1 1988; Peleg et a1 1988) that
free ferulic acid is the most likely precursor of p-vinylguaiacol (PVG), which is
believed to be a major detrimental compound contributing an ‘old fruit’ or ‘rotten’
flavour (Tatum et a1 1975). Concentration of PVG in orange juice increased with
storage time at elevated temperatures (Naim et a1 1988; Lee and Nagy 1990). The
possible mechanism for PVG production is decarboxylation of ferulic acid under
acidic conditions (Johnson and Heinz 1949; Pyysalo et a1 1977). Since ferulic acid
in citrus products occurs mainly in bound forms, it was further hypothesised (Peleg
et a1 1988) that the rate of ferulic acid release is important for PVG formation.
Indeed, linked forms of ferulic acid found in citrus fruit, eg ferulolyglucose (Reschke
and Herrmann 1981) and feruloylputrescine (Wheaton and Stewart 1965), vary
in their potential to lead to free ferulic acid (Peleg et al 1988). Other
hydroxycinnamic acids may be decarboxylated to vinyl phenols but the
contribution to citrus aroma is unknown. Limited data are available on the
distribution of bound and free phenolic acids in citrus fruit. The present
investigation was undertaken to quantify free and bound phenolic acids in
grapefruit and orange fruit parts (flavedo, albedo, endocarp and juice sacs) and
to assess the effect of time of harvesting on the level of phenolic acids in citrus juice.

MATERIALS AND METHODS

Citrus fruit
Four orange trees (Citrus sinensis (L) Osbeck var Shamuti) and four grapefruit
trees (Citrus paradisi Macfadyen var Marsh) were randomly selected from the
citrus orchard of the Faculty of Agriculture in Rehovot to serve as fruit sources.
Phenolic acids in oranges and grapefruits 419

Fruits were harvested randomly from various parts of the trees during early (20
November 1988),mid (1 5 January 1989) and late (29 March 1989) seasons. Percent
acidity was measured by titration with NaOH (0.156 M). Total soluble solids
content was measured as "Brix with an RSM-80 refractometer (Bellingham &
Stanley Ltd, Tunbridge Wells, UK).
Reagents and standards
Solvents were analytical (Frutarom, Haifa, Israel) and HPLC (Merck, Darmstadt,
FRG) grade. Laboratory deionised water was further purified using a vacuum
filter (0.45 ,UM,Schleicher & Schuell, Dassel, FRG). Ferulic, caffeic and coumaric
acids were obtained from Sigma Chemical Co (St Louis, MO, USA), and sinapic
acid was obtained from Aldrich Chemical Co (Milwaukee, WI, USA).
Preparation of samples
( a ) Distribution in fruit parts
Fruits, harvested during mid season, were separated by hand into : peel, albedo,
endocarp and juice sac. Samples of 1 kg were obtained for each fruit part. They
were extracted with water/ethanol/methanol (2:l:l v) in a blender for 5 min and
left in the same solution for 60 min. Extracts were then filtered and cold centrifuged
(8000xg, 15 min). Solvents were removed under reduced pressure. Half of the

-
residual aqueous clear solution was hydrolysed (2 M NaOH, 24 h, room
temperature 22-24°C) and then acidified. The hydrolysed and non-hydrolysed
fractions were extracted three times with 300ml ethyl acetate and the organic
phase was concentrated under reduced pressure.

(b) Juice contents

Fruits were extracted with a pneumatic extractor (Baccara, Kibbutz Geva). Juice
samples of 250ml were hydrolysed ( 2 NaOH, ~ under nitrogen, 4 h , room
temperature) and then acidified. Non-hydrolysed juice samples of 500 ml and the
hydrolysed ones were cold centrifuged (8000 x g, 15 min). The clear supernates
were extracted twice with 400 ml and 200 ml ethyl acetate for hydrolysed and
non-hydrolysed samples respectively. The organic phase was concentrated under
reduced pressure.
The ethyl acetate extracts from (a) and (b) were adsorbed on to silica gel 60
(Merck, 70-230 mesh, 1 g), chromatographed on a column of the same absorbent
(20 g) and eluted with 150 ml ethyl acetate. Fractions having typical cinnamic acid
spots on TLC when compared with markers were pooled. If necessary, the samples
were concentrated by evaporation of some of the ethyl acetate prior to analysis.
Thin layer chromatography (TLC) and high performance liquid chromatography
(HPLC) analyses
TLC analyses were carried out on silica gel 60 F254plates (Merck) developed with
ether/petroleum ether (9:l v). Spots were detected under UV (254 and 366 nm)
and by spraying first with 200 g litre-' Na,CO,, and then with a solution of
Folin-Ciocalteu reagent (diluted 1 + 3 v with water) for phenols (Keith et a1 1958;
Seikel 1964).
420 H Peleg, M Naim, R L Roustff, U Zehaoi

Quantitative HPLC determinations of cinnamic acids were conducted with a

Varian 5000 variable UV-100 detector and 4270 integrator. A LiChrospher 100
RP-18 column (5 pm; 250 mm x 4 mm ID, Merck) was used with a reversed phase
precolumn (RP-18;25 mm x 4 mm ID, Merck). The mobile phase was 15 ml litre-
acetic acid in water/methanol(77:23 v). Operating conditions were: flow rate, I ml
min- '; UV, 300 nm; chart speed, 0.5 cm min-'; injected volume, 10 pl; room
temperature. Each sample was injected twice and identified by comparison of its
retention time with that of the standard; the identification was further ascertained
by co-injection of representative samples with standards. Calibration curves were
obtained by injecting each of five standards containing different concentrations of
cinnamic acids ranging from 50 to 250 pg m1-I. Fruit maturity data were analysed
using an analysis of variance procedure and Duncan's test available on the
statistical analysis system, SAS (SAS Institute, Cary, NC, USA). Minimal
significance level was set at P < 0.05.

RESULTS AND DISCUSSION

Hydrolysis of bound phenolic acids and separation
Alkaline hydrolysis of the phenolic compounds obtained from citrus juice released
the following hydroxycinnamic acids: ferulic, coumaric, caffeic and sinapic. These
acids were relatively stable to treatment with 2 M NaOH for 4 h under nitrogen
at room temperature. Some losses of caffeic acid were to be expected due to the
reactive nature of the o-dihydroxy phenols (Cilliers and Singleton 1989). However,
most of the caffeic acid was lost during the long hydrolysis (24 h ) employed in
the experiments described in Tables 1 and 2. Following extraction with ethyl
acetate and fractionation on a silica gel column, quantitative determination of

TABLE 1
Content (mg kg-I) of phenolic acids (bound and free) in grapefruits

Fruit Sinapic Ferulic Coumaric Cafseic

part
Bound Free Bound Free Bound Free Bound Free

Peel 31.76 2.11 155.47 4.88 24.26 2.38 10.05 1.67

Albedo 1.96 1.18 24.85 1.52 6.61 1.02 5.84 1.36
Flavedo" 29.80 0.93 130.62 3.36 17.65 1.36 4.21 0.31
Juice sacs 3.51 0.66 27.76 0.48 1.68 1.48 11.11 0.72
Endocarp 5.14 0.29 24.53 0.28 2.50 0.94 9.8 1.84

Fruits harvested randomly (mid season, see under 'Materials and Methods') from various
sections offour trees to provide 1 kg material. Values derived from HPLC analyses (300 nm)
as described under 'Materials and Methods'.
Caffeic acid: 3-(3,4-dihydroxyphenyl)-2-propenoic acid; coumaric acid: 3-(4-hydroxyphenyl)-
2-propenoic acid; ferulic acid: 3-(4-hydroxy-3-methoxyphenyl)-2-propenoic acid; sinapic
acid: 3-(3,5-dimethoxyphenyl-4-hydroxyphenyl)-2-propenoic acid.
By calculation. Values obtained for caffeic acid are apparently low due to decomposition
(Cilliers and Singleton 1989).
Phenolic acids in oranges and grapefruits 42 I

TABLE 2
Content (mg kg-') of phenolic acids (bound and free) in oranges

Fruit Sinapic Ferulic Coumaric Cafeic

part
Bound Free Bound Free Bound Free Bound Free

Peel 95.10 5.37 178.40 3.15 76.65 0.48 7,29 0.19

Albedo 46.24 0.12 27.15 0.45 5.06 0.02 3.47 0-03
Flavedo" 48.86 5.25 151.76 2.15 71.59 0.46 3.82 0.16
Juice sacs 8.57 0.11 27.95 0.11 5.30 0.02 3.06 0.02
Endocarp 10.76 0.08 21.25 0.08 4.44 0.02 1.78 0.01

Fruits harvested randomly (mid season, see under 'Materials and Methods') from various
sections offour trees to provide 1 kg material. Values derived from HPLC analyses (300 nm)
as described under 'Materials and Methods'.
By calculation. Values obtained for caffeic acid are apparently low due to decomposition
(Cilliers and Singleton 1989).

phenolic acids was performed by HPLC. Separation of phenolic compounds

employing a reversed phase column is quite common (Casteelle et al 1983;
Conkerton and Chapital 1983; Seo and Morr 1984) and the procedure employed
resulted in good separation within a reasonable time. Retention times (min): caffeic
acid 1 1.3; coumaric acid 21 '1 ;ferulic acid 26-2 and sinapic acid 30.5. The coefficients
of variation (CV) of retention times for six runs of standards were less than 2.1 %.
The recoveries of the phenolic acids extracted from the juice were as follows: caffeic
acid 61%, coumaric acid 54%, ferulic acid 55% and sinapic acid 57%.
Distribution
The distribution of bound and free phenolic acids in grapefruits and oranges is
presented in Tables 1 and 2. To the best of our knowledge the present results are
the first to show the composition of free phenolic acids in grapefruits and oranges.
In all fruit parts, only minor amounts occur in the free state and most was present
in conjugated forms, which are capable of liberation by hydrolysis. In most cases
bound hydroxycinnamic acid content was in the following order: ferulic
acid > sinapic acid > coumaric acid > caffeic acid. Ferulic acid content was high
compared with the other acids. The peels contained the major quantity of cinnamic
acids compared with the endocarp, and the flavedo was richer in acids than the
albedo. These results were to be expected considering the fact that, in general,
citrus peel contains higher levels of many constituents compared with the
corresponding juice or other edible portions (Sinclair 1984). The same general
trend was found in oranges and grapefruits with both free and bound forms.
Orange peels contained more cinnamic acids than grapefruit peels but the contents
in juice sacs and endocarp of both fruits were similar. A concentration gradient
of phenolic acids appears to exist from the flavedo towards the juice.
It is known that an increase in mechanical pressure during fruit extraction
improves juice yield but may affect product taste due to penetration of peel
components, such as limonoids and flavonoids, into the juice (Maier et a1 1980;
Rouseff 1980). It seems, therefore, that the peels, and especially the flavedo, can
 P
N

TABLE 3
Total acidity and total soluble solids in grapefruit (var Marsh) and orange (var Shamuti) juice extracted from fruit harvested early,
mid and late in season

Date of Grapefruit Orange

harvest
Acidity* Brix "Brixfacid Acidity* "Brix "Brixlacid
(% w ) ratio (Yo w ) ratio

20 November 1988 1.91 f0.01" 9.47 f0.04" 4.95 f0.03" 1.28 f0.01" 10.58f0.03" 8.26f0.10"
15 January 1989 1.45f0.0Ib 9.65 0.05' 6.61 f O W b 1.02 f0.0 I b 10.89f0.09b 10.72f0.08b
29 March 1989 1.42 f0.0 1 9.25 f0.05' 6.56 fO.OSb 0.88 f0.01' 11.73 f0.0 1' 13.33f0.15'
b
m
Values are the mean and SEM of 2 4 replicates. Values not sharing the same superscript letter within each column are different 9
( P <0.05). t
* As anhydrous citric acid.

C
Phenolic acids in oranges and grapefruits 423

also be the source of phenolic acids in juice. If so, the technological processes to
which fruits are exposed during juice extraction may affect the amount of phenolic
acids transferred into the juice. These acids are the precursors for the formation
of vinyl phenols such as PVG, hence contributing to the development of
objectionable aromas during juice storage. The fact that, among phenolic acids,
ferulic acid was the most abundant also indicates the great potential for PVG
formation in orange and grapefruit juice. Our results indicate that the content of
total phenolic acids in the peels is higher than previously reported by Stohr and
Herrmann (1975a). This may be due to differences in fruit source, variety and
state of maturity as well as differences in the procedures employed for obtaining
samples and for quantitative determination of phenolic acids.
Maturity effects
The influence of harvest time on the juice total soluble solids, the total acidity
and the ratio of total soluble solids to total acidity is shown in Table 3. As expected,
during the season total acidity decreased whereas the total soluble solids and the
ratio increased. The content of bound forms of cinnamic acids in juice was not
affected by the harvest date although a trend towards an increase in the content
of bound sinapic and coumaric acids (grapefruit and orange) and caffeic (grapefruit)
was suggested (Figs 1 and 2). This may be related to biochemical changes that
occur when fruits reach their maximum size and growth is terminated (Sinclair

I
0.8

0.6
.2.
\ 0.4

1 0.2

._
0
0.0
.-
-
J

.-
c
3
SIN FER cou ClrF
L
L

SIN FER cou CAF

Phenolic acids
Fig 1. Content of free (A) and bound (B) sinapic (SIN), ferulic (FER), coumaric (COU) and caffeic
(CAF) acids in grapefruit juice (var Marsh) extracted from fruit harvested early (blank bars), mid
(hatched bars) and late (crosshatched bars) in season. Values are the mean and SEM of two replicates
(two samples, each analysed twice by HPLC). Bars not sharing the same superscript letter within each
acid are different (P ~ 0 . 0 5 ) .
424 H Peleg, M Naim, R L Rousef, U Zehavi

1.2

1 .o

0.8

.2\ 0.6

>E 0.4
v
0.2
._
V
.-3 0.0
Q
0
SIN FEA cou CAF

c
540
c
C
6 30

20

10

0
SIN FER cou w
Phenolic acids

Fig 2. Content of free (A) and bound (B) sinapic (SIN), ferulic (FER), coumaric (COU) and caffeic
(CAF) acids in orange juice (var Shamuti) extracted lrom fruit harvested early (blank bars), mid
(hatched bars) and late (crosshatched bars) in season. Values are the mean and SEM of two replicates
(two samples, each analysed twice by HPLC). Bars not sharing the same superscript letter within each
acid are different ( P < 0.05).

1984). On the other hand, the content of free acids was reduced, though statistical
significance was not evident in all cases. Most importantly, although the content
of free ferulic acid decreased during the season, that present at the end of the
season still exceeded the amount needed to form an above-taste threshold level
of PVG (0.075 ppm; Tatum et al 1975).
The changes reported by us are similar to those in strawberries (Stohr and
Herrmann 1975b), but not apples and pears (Mosel and Herrmann 1974), where
total phenolic acid contents decline during progressive growth of the fruit and
reach a steady level at maturity.
In conclusion, the maturity results indicate that, although the free acids decrease
during the season, the reservoir of bound phenolic acids remains basically
unchanged. This means that there is always a potential source of precursors for
the formation of objectionable aroma components during storage of processed juice
products.

ACKNOWLEDGEMENTS

This research was supported by the Israel Committee for Citrus Fruit Products
and by Grants No 1-1379-87 and 1-1528-88 from BARD, The United States/Israel
Binational Agricultural Research & Development Fund. We thank Mrs Martha
Phenolic acids in oranges and grapefruits 425

Levinson, Mr Avihai Greenberg and Mr Ram Eyal for excellent technical

assistance. Discussions and comments by Mr Assaf Tzur, D r Benjamin J Striem,
Mr Sam Bernhardt and M r Onn Rabinovitz were helpful.

REFERENCES

Casteelle K V, Geiger H, Van Sumere C F 1983 Separations of phenolics (benzoic acids,

cinnamic acids, phenylacetic acids, quinic acid esters, benzaldehydes and acetophenone,
miscellaneous phenolics and coumarins by reversed-phase high-performance liquid
chromatography. J Chromatogr 258 111-124.
Cilliers J J L, Singleton V I 1989 Nonenzymic autoxidative phenolic browning reactions
in a caffeic acid model system. J Agric Food Chem 37 89s896.
Conkerton E J, Chapital D C 1983 High-performance liquid chromatography separations
of the cis-trans isomers of cinnamic acid derivatives. J Chromatogr 281 326-329.
Harborne J B 1980 Plant phenolics. In: Encyclopedia of Plant Physiology, Vol8 Secondary
Plant Products, eds Bell E A & Charlwood B V. Springer-Verlag, Berlin, pp 329402.
Herrmann K 1990aSignificance of hydroxycinnamic acid compounds in food. I. Antioxidant
activity-effect on the use, digestibility, and microbial spoilage of food. Chem Mikrobiol
Technol Lebensm 12 137-144.
Herrmann K 1990b Significance of hydroxycinnamic acid compounds in food. 11. Effect
on the activity of food enzymes, alleged anti-thiamine effect, bitter taste, pharmacological
effectivenessand therapeutic uses, and significance as anti-mutagens and anti-carcinogens.
Chem Mikrobiol Technol Lebensm 12 161-167.
Ishii T, Hiroi T 1990 Isolation of feruloylated arabinoxylan oligosaccharides from bamboo
shoot cell-walls. Carbohydr Res 196 175-183.
Johnson W S, Heinz W E 1949 The acid-catalyzed decarboxylation of cinnamic acids.
J Amer Chem SOC 71 2913-2918.
Keith R W, Le Turneau D, Manlum D 1958 Quantitative paper chromatographic
determination of phenols. J Chromatogr 1 534-536.
Lee H S, Nagy S 1990 Formation of 4-vinyl guaiacol in adversely stored orange juice as
measured by an improved HPLC method. J Food Sci 55 162-163, 166.
Maga J A 1978 Simple phenol and phenolic compounds in food flavor. CRC Crit Rev Food
Sci Nutr 10 323-372.
Maier V P, Hasegawa S, Bennett R D, Echols L C 1980 Limonin and limonoids. In: Citrus
Nutrition and Quality, eds Nagy S & Attaway J A. American Chemical Society,
Washington, DC, pp 63-82.
Mosel H D, Herrmann K 1974 Changes in catechins and hydroxycinnamic acid derivatives
during development of apples and pears. J Sci Food Agric 25 251-256.
Naim M, Striem B J, Kanner J, Peleg H 1988 Potential of ferulic acid as a precursor to
off-flavors in stored orange juice. J Food Sci 53 5W503, 512.
Peleg H, Striem B J, Naim M, Zehavi U 1988 Phenolic compounds contributing to off-flavors
in citrus products. Proc 6th Int Citrus Congress, Vol 4, Tel Aviv, eds Goren R & Mendel
K. Balaban, Rehovot, and Margraf, Weikersheim, pp 1743-1748.
Pyysalo T, Torkkeli H, Honkanen E 1977 The thermal decarboxylation of some substituted
cinnamic acids. Lebensm- Wiss u Technol 10 145-147.
Reschke A, Herrmann K 1981 Vorkommen von 1-0-Hydroxycinnamyl-fi-D-glucosen im
Obst. Z Lebensm Unters-Forsch 173 458463.
Rouseff R L 1980 Flavonoids and citrus quality. In: Citrus Nutrition and Quality, eds Nagy
S & Attaway J A. American Chemical Society, Washington, DC, pp 83-108.
Seikel M K 1964 Isolation and identification of phenolic compounds in biological materials.
In: Biochemistry of Phenolic Compounds, ed Harborne J B. Academic Press, London,
pp 33-76.
426 H Peleg, M Naim, R L Rousefl, U Zehavi

Seo A, Morr C V 1984 Improved high-performance liquid chromatographic analysis of

phenolic acids and isoflavonoids from soybean protein products. J Agric Food Chem 32
53cL533.
Sinclair W B 1984 The biochemistry and physiology of the lemon and other citrus fruits.
University of California, Division of Agriculture and Natural Resources.
Stohr H, Herrmann K 1975a Uber das Vorkommen von Verbindungen der
Hydroxyzimtsauren, Hydroxybenzoesauren und Hydroxycumarine in Citrusfruchten.
Z. Lebensm Unters-Forsch 159 305-306.
Stohr H, Herrmann K 1975b The phenols of fruits. V. The phenols of strawberries and
their changes during development and ripeness of the fruits. Z Lebensm Unters-Forsch
158 341-348.
Tatum J H, Nagy S, Berry R E 1975 Degradation products formed in canned single-strength
orange juice during storage. J Food Sci 40 701-709.
Wheaton T A, Stewart I 1965 Feruloylputrescine: isolation and identification from citrus
leaves and fruits. Nature 206 62&621.