Industrial Crops and Products 69 (2015) 385394

Contents lists available at ScienceDirect

Industrial Crops and Products

journal homepage: www.elsevier.com/locate/indcrop

Characterization of phenolic compounds, anthocyanidin, antioxidant

and antimicrobial activity of 25 varieties of Mexican Roselle
(Hibiscus sabdariffa)
I. Borrs-Linares a,b , S. Fernndez-Arroyo a,b , D. Arrez-Roman a,b, , P.A. Palmeros-Surez c ,
R. Del Val-Daz c , I. Andrade-Gonzles c , A. Fernndez-Gutirrez a,b , J.F. Gmez-Leyva c ,
A. Segura-Carretero a,b
a
Department of Analytical Chemistry, Faculty of Sciences, University of Granada, Avda. Fuentenueva s/n, C.P. 18071 Granada, Spain
b
Research and Development Functional Food Centre, Health Science Technological Park, Avda Conocimiento s/n, C.P. 18016 Granada, Spain
c
Laboratory of Molecular Biology, Technological Institute of Tlajomulco, Km 10 Carretera a San Miguel Cuyutln, C.P. 45640 Tlajomulco de Zniga,
Jalisco, Mexico

a r t i c l e i n f o a b s t r a c t

Article history: This study reports the phytochemical prole (phenolics, avonoids and anthocyanins), the antioxidant
Received 9 October 2014 capacity and the antibacterial activity of the ethanolic extracts of a collection of 25 Mexican Hibiscus
Received in revised form 5 February 2015 sabdariffa (Hs) varieties with different calyx color intensities, from greenyellow to deep red, culti-
Accepted 22 February 2015
vated in the same condition. A great variety of phenolic compounds were identied in the different
extracts using HPLC-DAD-ESI-TOF-MS, mainly phenolic acids, avonoids, and anthocyanidin, some of
Keywords:
which have been described for the rst time. The total phenolic, avonoid and anthocyanidin contents
Hibiscus sabdariffa
showed great variation in the different Hs extracts, ranging from 2400 300 to 10,000 400 mg gallic
Anthocyanins
Phenolic compounds
acid equivalents (GAE)/100 g dry calyx (dc), from 419 2 to 2260 70 mg quercetin/100 g dc, and from 0
Antimicrobial activity to 4408 mg/100 g dc with different ratios of delphinidin:cyanidin, respectively. Furthermore, the antiox-
HPLC-AD-ESI-TOF-MS idant capacity determined by 2,2-diphenyl-1-picrylhydrazyl (DPPH) assay presented the same behavior,
with values that varied from 27.4 0.3 to 112 8 mol equivalent trolox/g dc. In addition, the antibacte-
rial activity of the HS extracts was assayed against Gram-negative (Escherichia coli, Salmonella enteritidis)
and Gram-positive (Staphylococcus aureus, Micrococcus luteus) microorganisms, demonstrating that the
ethanol extracts were effective against all the bacterial strains tested, showing a greater effect against
Gram-positive bacteria. Finally, a multivariate analysis for the classication of the Hs varieties has been
carried out based on the anthocyanidin content. The results reported in this paper shows for the rst
time antibacterial and phytochemistry diversity of the different varieties of Hs, which highlights the
importance of a correct description of Hs materials used in research papers.
2015 Elsevier B.V. All rights reserved.

1. Introduction
Hibiscus sabdariffa L. (Hs; Malvaceae), commonly known as
roselle, red sorrel or karkade, is an annual herbaceous subshrub
belonging to the family Malvaceae. This plant is native to Africa and
grows in tropical and sub-tropical regions such as Sudan, southern
Asia, and America. Fleshy calyces (sepals) are commercially impor-
tant for the production of beverages, juices, jams, and syrup in the
Corresponding author at: Department of Analytical Chemistry, Faculty of Sci-
ences, University of Granada, Avda. Fuentenueva s/n, Granada C.P. 18071, Spain.
Tel.: +34 958248409; fax: +34 958243328.
E-mail address: darraez@ugr.es (D. Arrez-Roman).
food industry. Furthermore, these calyces are a good source of nat-
ural food colorants because of their high pigment content (Bridle
and Timberlake, 1997). Moreover, the dried calyces are consumed
worldwide in hot infusions and in cold drinks (Hervert-Hernndez
and Goni, 2012). Besides its extended consumption as a beverage
and its uses in food industry, roselle is also used in animal feed,
nutraceuticals, cosmetics and pharmaceuticals (Wang et al., 2012).
Hs plant contains proteins, fats, carbohydrates, acids, min-
erals, vitamins, as well as many types of phenolic compounds.
Mucilaginous polysaccharides and pectins also form part of its
composition. Furthermore, Hs seeds contain a wide range of less
polar compounds, such as sterols (sitosterol, ergosterol, campes-
terol). Although the composition of this plant has been thoroughly

http://dx.doi.org/10.1016/j.indcrop.2015.02.053
0926-6690/ 2015 Elsevier B.V. All rights reserved.
386 I. Borrs-Linares et al. / Industrial Crops and Products 69 (2015) 385394

studied, one notable (and underreported) group of compounds
present in this plant is phenolic compounds, which recently have
attracted a great deal of attention due to their benecial effects in
the promotion of human health and well-being (Ali et al., 2005;
Crozier et al., 2009; Rodrguez-Medina et al., 2009).
The phenolic compounds found in this plant include organic
and phenolic acids, such as citric acid, hydroxycitric acid, hibiscus
acid, and protocatechuic acid. Flavonoids such as quercetin, lute-
olin or gossipetin, and their respective glycosides are also present.
Anthocyanins, detected in high amounts in the calyces, are respon-
sible for the bright red color. The most frequent anthocyanins
of Hs owers are cyanidin-3-glucoside, delphinidin-3-glucoside,
cyanidin-3-sambubioside, and delphinidin-3-sambubioside (Ali
et al., 2005; Rodrguez-Medina et al., 2009).
Numerous pharmacological properties have been attributed to
Hs calyx (Ali et al., 2005). The most signicant ones are its car-
dioprotective action, antihypertensive action, effectiveness against
low-density lipoprotein oxidation, and hyperlipidemia (Alarcn-
Aguilar et al., 2007; Beltrn-Debn et al., 2010; Gurrola-Daz et al.,
2010; Fernndez-Arroyo et al., 2012; Herranz-Lpez et al., 2012;
Kuo et al., 2012; Laikangbam and Damayanti Devi, 2012; Hopkins
et al., 2013). Also potent effects in the reduction of urinary concen-
trations of creatinine, uric acid, citrate, tartrate, calcium, sodium,
potassium, and phosphate have been demonstrated (Kuo et al.,
2012; Laikangbam and Damayanti Devi, 2012). Moreover, its anti-
inammatory, antioxidative, hepatoprotective, and antitumoral
effects have been highlighted (Fernndez-Arroyo et al., 2011; Lin
et al., 2012; Wang et al., 2000).
Most of the research on Hs does not specify the origin of the
variety and the crop site, making it difcult to make comparisons
between the phytochemical prole and biological properties of
extracts obtained in different studies due to the inherent vari-
ability of the variety and harvesting location of a crop. The aim
of the present work has been to identify the phenolic com-
position by HPLC-DAD-ESI-TOF-MS of the ethanolic extract in
a collection of 25 different varieties of Hs from Mexico and
determining the content of phenolic, avonoid, antocyanidin and
antioxidant activity as well as their antibacterial activity against
Escherichia coli, Salmonella enteritidis, Staphylococcus aureus and
Micrococcus luteus.
2. Material and methods
2.1. Chemicals
All chemicals used were of analytical HPLC reagent grade.
Formic acid, triuoroacetic acid, methanol and acetonitrile,
used for preparing mobile phases, were purchased from Fluka,
SigmaAldrich (Steinheim, Germany) and Lab-Scan (Gliwice,
Sowinskiego, Poland), respectively. Water was puried by a Milli-
Q system from Millipore (Bedford, MA, USA). The standards, for
the comparison with the compounds detected, chlorogenic acid,
cyanidin chloride, delphinidin chloride, quercetin, and quercetin-
3-glucoside were purchased from Fluka, SigmaAldrich (Steinheim,
Germany) and Extrasynthese (Genay Cedex, France). The stock
solutions containing these analytes were prepared in methanol
(Lab-Scan), and the hydrochloric acid (HCl) used for anthocyanin
determination was purchased from Lab-Scan (Gliwice, Sowin-
skiego, Poland).
2.2. Sample collection and processing
Seeds of 25 different varieties of Hs were collected from the main
producing areas in Mexico during the years 20072010 (Table 1).
The different varieties were established in the eld under the
same conditions until physiological maturity and then the calyces
were collected and dried in a forced convection hot-air oven. A
total of 100 g of dried calyxes of each Hs variety were grounded
and extracted using acidied ethanol (70% ethanol plus 0.1% HCl)
with a solvent to sample extract ratio of 1:10 (w/v). After 72 h of
extraction, solvent was evaporated at 40 C using a rotary evapora-
tor (Bchi R-215, Bchi Labortechnik AG, Switzerland) under high

Table 1
Content of total phenolic compound, avonoids, anthocyanin, delphinidin, cyanidin and antioxidant capacity in H. sabdariffa extracts.

HS extract Variety Origin Total phenolic Flavonoid Anthocyanidin Delphinidin Cyanidin content Antioxidant
content content content** content capacity
(mg GAE/g dc) (mg QE/ 100 g dc) (mg/100 g dc) (mg/100 g dc) (mg/100 g dc) (mol ET/ g dc)

HS 1 Americana Jalisco 30 6hi 813 5i 969m 444k 525 h 104.0 9a

HS 2 Tepalcatepec Michoacn 90 20b 692 8jk 1673i 1083g 589f 83.0 0.6dc
HS 3 Diamante Jalisco 98 1a 1695 8b 2967d 2232c 734c 112.0 8a
HS 4 Colima Colima 94 9b 1310 40 d 2306f 1658e 648d 60.1 1fe
HS 5 Tempranilla Jalisco 41 3g 429 5n 4408a 3535a 873a 40.0 2ji
HS 6 Talpa Jalisco 38 5g 419 2n 4085c 3300b 784b 45.0 8i
HS 7 Violenta Jalisco 60 10e 720 30j 4246b 3495a 750c 48.6 0.4i
HS 8 Quesera Colima 57 9ef 530 20l 1781h 1223f 558g 105.8 0.3a
HS 9 CriollaTala Jalisco 64 2e 890 10hi 1369j 825h 543g 80.0 2dc
HS 10 Tecoman Colima 42 4g 470 10n 1013m 511k 501i 37.0 1k
HS 11 El Bordo Colima 34 3h 1030 70gh 1308k 783h 525h 76.6 0.3ed
HS 12 Tempranilla Colima 45 9fg 1140 50fe 2613e 1956d 657d 84.0 2bcd
HS 13 Sudan Colima 43 4g 488.0 0.6m 929m 436k 492i 36.3 0.3k
HS 14 M. Luna Colima 70 10d 681 7k 1500j 940h 560g 38.5 0.1j
HS 15 Pisila Colima 68 3d 740 20ji 1361k 791h 569g 56.0 2fgh
HS 16 JB 01 SM Guerrero 83 1c 1050 30g 0n 0l 0j 84.0 8bcd
HS 17 JCP 01T Guerrero 88.8 0.6cb 950 30h 1058L 552j 505i 56.0 3fg
HS 18 Blanca Guerrero 81 1c 1272 10 0n 0l 0j 49.2 0i
HS 19 JJ 01 SM Guerrero 83 1c 810 50i 1216L 663i 553g 48.6 0.5i
HS 20 JR 01 C Guerrero 37 5gh 1070 10fg 1147l 633i 514h 85.0 0.9bcd
HS 21 Americana Colima 51 7f 504 4m 1815h 1205f 609e 35.1 0.1kl
HS 22 P. Anzar Colima 28 4i 632 5kl 1178l 678i 500i 97.3 8bc
HS 23 Blanca Jalisco 24 3j 855 8hi 0n 0l 0j 27.4 0.3l
HS 24 Real Veracruz 100 4a 1500 70c 1280k 707i 572f 101.0 6b
HS 25 Reyna Nayarit 70 1d 2260 70a ND ND ND 97.0 4bc
*
Mean values in a column superscripted by the same letter are not signicantly different at p < 0.05.
ND: Not determined.
**
Anthocyanidin content = Delphinidin + Cyanidin.
 I. Borrs-Linares et al. / Industrial Crops and Products 69 (2015) 385394
vacuum. The resulting extracts were used for all the determinations
during the study.
2.3. Total phenolic content determination
Total phenolic content of ethanol extracts of 25 calyces was
determined by the FolinCiocalteu method (Singleton and Rossi,
1956). An aliquot of 450 L fresh Folin reagent 1N in water (1:1 v/v)
was mixed with 50 L of different concentrations of each extract
or different concentrations (20180 g/mL) of gallic acid used as
standard. Further, after incubation for 5 min at room temperature,
500 L of (15% Na2 CO3 ) sodium carbonate solution was added and
the solution was mixed thoroughly and incubated for 60 min at
room temperature. The absorbance was measured at 765 nm. All
results were expressed as milligram of Gallic Acid Equivalent per
100 g of dried calyx (mgGAE/100 g dc).
2.4. Total avonoid content determination
A colorimetric method was used to determine the avonoid
content (Chang et al., 2002). An aliquot of 0.5 mL of each extract
(50 mg/mL) was separately mixed with 1.5 mL of methanol, 0.1 mL
of 10% aluminum chloride, 0.1 mL of 1M potassium acetate and
2.8 mL of distilled water. It was kept at room temperature for
30 min. The absorbance of the reaction mixture was measured at
415 nm with a Microplate spectrophotometer UVvis (MultiskanTM
GO, Thermo Scientic, Waltham, MA, USA). The calibration curve
was recorded using quercetin solutions in methanol with concen-
trations ranging from 10 to 100 g/mL. The total avonoid content
was expressed in terms of milligram of quercetin equivalent per
100 grams of dried calyx (mg QE/100 g dc).
2.5. Anthocyanin, delphinidin and cyanidin content
determination
The determination of these analytes was performed by HPLC
as reported by (Gurrola-Daz et al., 2010). Before the determi-
nation Hs extracts were acid hydrolyzed as described by (Zhang
et al., 2004). Briey, 100 mg of Hs extract was dispersed in a 1:1
(v/v) water/methanol solution containing 2N HCl, syringe-ltered
through 0.22 mm nylon membrane (Millipore) and incubated
at 100 2 C for 60 min, immediately cooled to room temper-
ature for analysis. A volume of 25 L was injected onto a
Zorbax C18 column (250 4.6 mm, 5 m particle size). Antho-
cyanins were eluted by an isocratic gradient of 50:49.99:0.01 (v/v)
methanol:water:triuoroacetic acid for 20 min at a ow rate of
1 mL/min using a Thermo Finnigan Surveyor HPLC system (Thermo
Finnigan, CA, USA). Detection was performed at 522 nm, and spec-
tral data from 490590 nm were also collected, using a photo-diode
array detector (UV6000, Thermo Finnigan, CA, USA). Anthocyani-
dins were identied by comparing the sample retention times and
peaks with those of known standards for cyanidin and delphini-
din chlorides. Reference standards ranging between 0.2 and 100
mg/L were prepared by dissolving either cyanidin or delphinidin
chlorides in 1% HCl solution. Linear regression analysis was car-
ried out on the data for peak area versus concentration, obtaining a
linear calibration with 98% accuracy. The cyanidin and delphinidin
content (mg/100 g dried calyx) was calculated using its correspond-
ing standard curve. Both delphinidin and cyanidin sambubiosides
constituted more than 99% of the total anthocyanins.
2.6. Determination of antioxidant activity by the DPPH
radical-scavenging method
The scavenging capacity of the DPPH radical was determined
using the stable 2,2-diphenyl-1-picrylhydrazyl radical (DPPH ) as
387
previously described (Re et al., 1999). Briey, an aliquot (50 L) of
extract was placed in 96-well microplates, and 200 L of 0.1 mM
methanolic solution of DPPH was added and allowed to react in
the dark at room temperature. The decrease in absorbance of DPPH
at 515 nm was measured until the absorbance stabilized (30 min).
Methanol at 80% was used as a blank solution, and DPPH solution
without test samples served as the control. The antioxidant activity
of the samples at various concentrations was determined using a
trolox standard curve at concentrations of 0, 25, 50, 75, 100, 125,
and 500 g/mL, and expressed in terms of mol equivalent of trolox
per gram of dried calyx (mol ET/g dc).
2.7. Instrumentation
2.7.1. HPLC-DAD-ESI-TOF analysis
In order to study the qualitative composition of the extracts,
dried extracts of Hs calyces obtained as described below were
diluted with ultrapure water up to a concentration of 25 mg/mL
while stirring in a vortex. Prior to analysis, the solutions were l-
tered with single-use lters (0.45 m) and directly injected into the
HPLC system.
The analyses were carried out using an Agilent 1200 RRLC sys-
tem (Agilent Technologies, Palo Alto, CA, USA) of the Series Rapid
Resolution equipped with a vacuum degasser, an autosampler, a
binary pump, and a UVvis detector. The chromatographic separa-
tion was performed on a Zorbax Eclipse Plus C18 analytical column
(4.6 150 mm, 1.8 m particle size) at a ow rate of 0.5 mL/min.
The temperature of the column was maintained at 25 C and the
injection volume was 10 L. Prelters were used as precolumn,
RRLC in-line lters, 4.6 mm, 0.2 m (Agilent Technologies, Palo
Alto, CA, USA). The mobile phases used were water:acetonitrile
(90:10, v/v) plus 1% of formic acid as mobile phase A and acetoni-
trile as mobile phase B. The separation was conducted using the
following multi-step gradient: 0 min, 5% B; 20 min, 20% B; 25 min
40% B; 30 min 5% B; and nally a conditioning cycle of 5 min with
the initial conditions for the next analysis. The UVvis detection
was performed in a wavelength range 190950 nm.
For MS detection a microTOFTM mass analyzer (Bruker Daltonik,
Bremen, Germany) was used coupled to the HPLC system via an ESI
interface (model G1607A, Agilent Technologies, Palo Alto, CA, USA)
operating in negative ionization mode showing the molecular ions
[MH] . Due to the high ow rate of the separation, to ensure sta-
ble ionization conditions and reproducible results, the efuent from
the HPLC system was split before it was introduced into the mass
spectrometer (split ratio 1:3). External mass spectrometer calibra-
tion was performed with a sodium formate cluster solution (5 mM
sodium hydroxide and 0.2% formic acid in water/2-propanol (1:1,
v/v)) in quadratic plus high-precision calibration (HPC) regression
mode. The calibration solution was injected at the beginning of the
run and all the spectra were calibrated prior to identication. This
external calibration eliminates the necessity of a dual sprayer setup
for internal mass calibration due to the compensation of tempera-
ture drift in the micro-TOF analyzer, providing accurate mass values
(better than 5 ppm).
Source and transfer parameters were optimized as previously
described (Fernndez-Arroyo et al., 2011) for good sensitivity and a
reasonable resolution of the mass range for the compounds of inter-
est (501000 m/z) in order to improve ionization performance. The
optimum values of source parameters were: capillary voltage of
4500 V; drying gas temperature, 200 C; drying gas ow, 7 L/min;
nebulizing gas pressure, 1.5 bar, and end plate offset 500 V. The
values of transfer parameters were: capillary exit, 120 V; skim-
mer 1, 40 V; hexapole 1, 23 V, RF hexapole, 50 Vpp; skimmer 2,
22.5 V; transfer time, 50 s, and 3 s pre-pulse storage time.
The accurate mass data of the molecular ions were pro-
cessed through the software Data Analysis 4.0 (Bruker Daltonik),
388 I. Borrs-Linares et al. / Industrial Crops and Products 69 (2015) 385394
which provided a list of possible elemental formulas by using
the Generate Molecular Formula EditorTM . The editor uses a
CHNO algorithm, which provides standard functionalities such
as minimum/maximum electron conguration, elemental range,
and ring-plus double-bond equivalent, as well as a sophisticated
comparison of the theoretical with the measured isotopic pattern
(SigmaValue) for increased condence in the suggested molecular
formula. The widely accepted accuracy threshold for conrmation
of elemental compositions was established at 5 ppm. Many chemi-
cally possible formulae are obtained even with high mass accuracy;
although the use of isotopic abundance patterns as a single fur-
ther constraint removes more than 95% of the false candidates. This
orthogonal lter can condense several thousand candidates down
to only a small number of molecular formulas.
2.8. Biological material
The potential antibacterial activities of all extracts were ana-
lyzed against the following bacteria: Escherichia coli (ATCC 25,922)
Salmonella enteritidis (ATCC 14,028), Staphylococcus aureus (ATCC
25,923), and Micrococcus luteus (ATCC 9341). All bacterial isolates
were maintained in nutrient agar media and incubated at 37 C
for 24 h. For antimicrobial activity determination, a single colony
was selected from fresh culture from all strains and grown in 5 mL
nutrient broth for 24 h at 37 C.
2.8.1. Antimicrobial activity assay
The antimicrobial activity of Hs ethanol extracts were assayed
by the agar well diffusion method according to (CLSI, 2005).
Mueller-Hinton agar medium was poured into aseptically Petri
plates and allowed to solidify. The tests of the bacterial strains
were prepared by the inoculation of a 24-h culture at a density
of 108 UFC/mL of bacterial strains by the spread-plate method. For
this, 20 L of ethanol extract equivalent to 7.6 mg of total solids
were placed in sterile paper lter disks 6 mm wide. Five antibiotics
were used as a positive control: gentamicin (10 g), erythromycin
(15 g), vancomycin (30 g), tetracycline (30 g), and clindamycin
(2 g) (Sensi-Disc BBL, USA). Ethanol discs were used as a negative
control. Agar plates were incubated at 37 C for 24 h. Diameter of
clear zone of inhibition in millimeters around the disks containing
the Hs extracts was taken as an index of the degree of antibacterial
sensitivity considering the diameter of inhibition as: Susceptible
21 mm, Intermediate 1720 mm and Resistant 16 mm according
to (CLSI, 2005).
2.9. Statistical analysis
Data was analyzed using Origin (version Origin Pro 8 SR0,
Northampton, MA, USA) to perform one-way-analysis of variance
(ANOVA) at a 95% condence level (p 0.05) to identify signicant
differences among phenolic, avonoid, anthocyanin, delphinidin,
and cyanidin content, as well as for antioxidant and antimicro-
bial activities of Hs extracts. Pearson correlation coefcients were
also calculated with this software in order to measure the linear
correlation between variables (Table 5).
2.10. Multivariate data analysis
The similarities between the 25 varieties of Hs were ana-
lyzed using two different techniques of multivariate analysis. The
principal component analysis (PCA) was performed in order to
determine the data variance of the six analyzed variables (Contents
of total phenolic compounds, avonoids, antocyanins, delphinidin,
cyanidin and antioxidant capacity). PCA was performed using the
software Statgraphics Plus 5.0 (Statistical Graphics Corp., Rockville,
MD). Clustering analysis was conducted using InfoStat Version
2013, which generates a dendrogram of the Hs varieties based on
the euclidean distance.
3. Results and discussion
3.1. Content of total phenolic compound, avonoids,
anthocyanidin, and antioxidant activity in Hs extracts
The amounts of total phenolic compounds of Hs extracts, deter-
mined by FolinCiocalteau method, are listed in Table 1. The total
phenolic content showed great variation in different Hs extracts,
ranging from 2400 300 to 10,000 400 mg GAE/100 g dc. The
highest level of phenolics was found in Hs 24 from Veracruz
(Real variety), followed by Hs 3 from Jalisco (Diamante variety)
(10,000 400 and 9800 100 mg GAE/100 g dc, respectively), and
no signicant differences were found among them. On the other
hand, the lowest phenolic content corresponded to Hs 23 from
Jalisco (Blanca variety).
With respect to the avonoid content (Table 1), a similar behav-
ior was observed; the values ranged from 4192 2 to 2260 70 mg
QE/100 g dc. The highest content was found in Hs 25 (Reyna vari-
ety) from Nayarit, followed by Hs 3 (Diamante variety) from Jalisco,
which is signicantly lower (2260 70 and 1695 8 mg QE/100 g
dc, respectively). The minimum content was found in Hs 6 (Talpa
variety) from Jalisco, followed by Hs 5 (Tempranilla variety) and
Hs 10 (Tecoman variety), although no signicant differences were
present between them.
In the determination of the total anthocyanidin content
(Table 1), the highest levels of these compounds were found in
the extracts Hs 5 (Tempranilla variety) and Hs 7 (Violenta vari-
ety), both harvested in Jalisco. The same behavior was found for
the delphinidin content, where these extracts presented the high-
est concentration. Nevertheless, the determination of the cyanidin
content revealed that along with Hs 5, which possessed the high-
est concentration in cyanidin, Hs 6 from the variety Talpa collected
in Jalisco was one of the richest extracts in cyanidin instead of Hs
7. On the other hand, in extracts Hs 16, 18, 23 and 25 no antho-
cyanin, delphinidin or cyanidin were detected. These extracts were
obtained from white varieties of Hs and, as mentioned before, since
anthocyanins are responsible for the red color of this plant, it was
expected that these compounds were not present in their calyces.
The contents of these phytochemicals (phenolic compounds,
avonoids, anthocyanidin, cyanidin and delphinidin) found in our
study have found to be higher than previous results reported by
other authors. In a study conducted by Christian and Jackson (2009)
the analysis of dried calyx from two Hs varieties, red and colorless,
revealed a total phenolic content which varied from 5002400 mg
GAE/100 g dc, while the monomeric anthocyanin content uctu-
ated from 2 to 345 mg/100 g dc. Furthermore, Mohd-Esa et al.
(2010) reported a total phenolic content of 185291 mg GAE/100 g
dc, which were too much lower than our results. Moreover, Sindi
et al. (2014) also reported lower values for total phenolic con-
centration and total monomeric anthocyanins, ranging from 0 to
2200 mg GAE/100 g dc and from 0 to 600 mg/100 g dc, respectively.
In the same study, the individual concentrations of delphinidin
(0426 mg/100 g dc) and cyanidin (0427 mg/100 g dc) had a very
similar behavior than total contents, being lower than the reported
in the present research.
Nevertheless, these differences could be explained because of
different factors. One aspect to be considered respect to the con-
tent of these phytochemicals is the extraction process of the calyx.
Recently, Sindi et al. (2014) evaluated the inuence of the extrac-
tion solvent in the contents of total phenolic and anthocyanins,
as well as antioxidant capacity, using a variety of Hs. The results
showed that there were a signicant difference in the content of
 I. Borrs-Linares et al. / Industrial Crops and Products 69 (2015) 385394
Fig. 1. Typical HPLC-DAD-ESI-TOF-MS prole of different H. sabdariffa extracts. The numbering of the peaks correspond to those listed in Table 1.
these compounds and the antioxidant capacity depending on the
extraction procedure. These ndings highlight the difculty for the
comparison of results obtained in different studies due to inconsis-
tencies in the used methodology.
Additionally, as our results suggested, other factor that inuence
the content of these phytochemicals was the Hs variety (Wang et al.,
2014). These results highlights the need to specify the Hs origin
genotype, due to most of publications only report that Hs were
purchased in supermarkets or merely referred as Hs, which can lead
to a signicant variation in results and make comparisons difcult.
The antioxidant capacity of Hs extracts (Table 1) was deter-
mined by the DPPH method. The values ranged from 27.4 0.3 to
112 8 mol ET/g dc. The highest antioxidant capacity was found
in Hs 3 (Diamante variety), followed by Hs 8 (Quesera variety) and
Hs 1 (Americana variety), although no signicant differences were
detected among them. Meanwhile, the lowest antioxidant capacity
corresponded to extract Hs 23, a white variety from Jalisco, fol-
lowed by Hs 21 from Colima (Americana variety). The antioxidant
capacity of some Hs extracts was in the range of the activity exerted
by grape wine, one of the most common antioxidant natural source
(Xia et al., 2010). Other authors have reported the antioxidant activ-
ity of various extracts of Hs, in fact, recently this plant has gained
attention due to the growing interest in natural antioxidants (Tsai
et al., 2002; Wang et al., 2014).
Pearson correlation coefcients showed no direct relationship
between the antioxidant capacity and antocyanidin or pheno-
lic content, on the contrary to avonoids (r = 0.516, signicance
p < 0.05). These results are contradictory with several reports in
which a direct correlation between phenolic or anthocyanin con-
tent and antioxidant capacity was observed (Tsai et al., 2002;
Djeridane et al., 2006; Wojdylo et al., 2007). However, these differ-
ences might be explained by the fact that the antioxidant activity
of the Hs extracts could be attributed to the presence of individual
or specic phenolic compounds, and not to the total amount of this
group of compounds. Moreover, as commented before, the vari-
ety of Hs is crucial respect to the total contents and distribution of
these phytochemicals (phenolic, avonoids and anthocyanins), so
different varieties probably possessed different chemical structures
and concentrations of individual phenolic acids, avonoids and
anthocyanins, which are directly correlated with the antioxidant
activity. Accordingly, the different results in these studies could be
389
explained because it is difcult to make a direct comparison due to
diversity of Hs varieties, as mention above.
The study revealed that the content of the compounds analyzed
(phenolic, avonoid and anthocyanins) and the antioxidant activ-
ity found for the 25 different Hs extracts have a great variation
between varieties, since the seeds collected in different parts of
Mexico were established under the same geographic condition and
fertility, which allows a precise comparison between these vari-
eties. The results obtained highlight the potential uses of several
Hs varieties studied. In particular, the extracts Hs 3, Hs 24, Hs 25,
the richest extracts in total phenolic compounds and avonoids,
could be used as potential sources of these compounds for the
development of functional food and nutraceuticals, although stud-
ies regarding bioavailability need to be performed (Patel, 2014).
The same application could be concluded for the extracts with the
highest antioxidant capacity, in particular for Hs 1, Hs 3 and Hs 8
(Wang et al., 2014). On the other hand, the extracts with highest
contents of total anthocyanins, cyanidin and delphinidin (Hs 5, Hs
6 and Hs 7) could be used in the food or pharmaceutical industry
as source of natural pigments, a highly valued group of substances
(Frimpong et al., 2014).
3.2. Characterization of phenolic compounds in Hs varieties by
HPLC-DAD-ESI-TOF-MS
The analysis of the Hs extracts by HPLC-DAD-ESI-TOF-MS
revealed the presence of a wide variety of polyphenols and other
polar compounds (Fig. 1). Ions detected were tentatively identi-
ed by their generated molecular formula, the use of standards
when available and after a thorough survey of the literature on
Hs. Thus, Table 2 lists the peak number, retention time, observed
m/z, the generated molecular formula and the proposed compound
detected in the different Hs varieties. The number of compounds
identied was in the range of 2442 depending on the variety.
The presence of hibiscus acid derivatives was notable, as in the
case of hibiscus acid hydroxyethylesther, hibiscus acid dimethyles-
ther, and hibiscus acid hydroxyethyldimethylesther, besides a
nitrogenated compound called N-feruloyltyramine in all the Hs
varieties.
On the other hand, hydroxycitric acid (peak 1) was detected in
all extracts except in Hs 2, 11, 16, 20, 23, and 25, while hibiscus acid
390 I. Borrs-Linares et al. / Industrial Crops and Products 69 (2015) 385394

Table 2
Compounds identied in H. sabdariffa extracts.

Peak RT (min) Theoretical m/z Mol. formula Proposed compound HS extracts

1 4.10 207.0146 C6 H8 O8 Hydroxycitric acid 1, 310, 12-15, 1719, 21, 22, 24, 25
2 4.87 189.0041 C6 H6 O7 Hibiscus acid 124
3 5.28 221.0303 C7 H10 O8 3-Deoxy-d-lyxo-heptulosaric acid 1, 3, 6, 813, 1625
4 5.50 235.0459 C8 H12 O8 Hibiscus acid hydroxyethylesther 125
5 6.13 353.0878 C16 H18 O9 Chlorogenic acid (isomer 1) 3, 5, 12, 13, 19, 23, 24
6 6.36 249.0616 C9 H14 O8 Trimethylhydroxycitric acid (isomer 1) 1-2, 611, 1425
7 6.69 249.0616 C9 H14 O8 Trimethylhydroxycitric acid (isomer 2) 1-2, 611, 1419, 21, 22, 24, 25
8 7.35 217.0354 C8 H10 O7 Hibiscus acid dimethylesther 125
9 7.75 249.0616 C9 H14 O8 Trimethylhydroxycitric acid (isomer 3) 6, 9, 10, 19, 2123
10 7.85 353.0878 C16 H18 O9 Chlorogenic acid 35, 13, 19, 20, 23, 24
11 8.31 353.0878 C16 H18 O9 Chlorogenic acid (isomer 3) 25, 8, 12, 13, 18, 19, 20, 23, 24
12 9.13 263.0772 C10 H16 O8 Hibiscus acid hydroxyethyldimethylesther 125
13 9.50 369.0463 C15 H14 O11 2-O-trans-caffeoyl-hydroxycitric acid 13, 8, 13, 19, 2225
14 9.70 367.1034 C17 H20 O9 Methylchlorogenate (isomer 1) 2, 612, 14, 1625
15 11.07 337.0929 C16 H18 O8 Coumaroylquinic acid 12
16 11.40 507.1144 C23 H24 O13 Syringetin-3-O-glucoside 17, 21, 22
17 11.80 263.0772 C10 H16 O8 Kinsenoside 611, 1418, 2125
18 12.30 367.1034 C17 H20 O9 Methylchlorogenate (isomer 2) 1, 2, 6, 8, 1113, 1625
19 13.50 367.1034 C17 H20 O9 Methylchlorogenate (isomer 3) 11, 1620, 2225
20 13.77 381.1191 C18 H22 O9 Ethylchlorogenate (isomer 1) 125
21 14.00 335.0772 C16 H16 O8 5-O-caffeoylshikimic acid 3, 5, 12, 19, 2025
22 14.44 475.0882 C22 H20 O12 Kaempferol-3-O-glucuronic acid methyl esther (isomer 1) 21
23 15.11 475.0882 C22 H20 O12 Kaempferol-3-O-glucuronic acid methyl esther (isomer 2) 21
24 15.46 461.0725 C21 H18 O12 Kaempferol 3-O-glucuronide 1, 412, 1424
25 15.50 463.0882 C21 H20 O12 Quercetin-3-glucoside 2, 3, 13, 19
26 16.22 381.1191 C18 H22 O9 Ethylchlorogenate (isomer 2) 125

27 17.10 381.1191 C18 H22 O9 Ethylchlorogenate (isomer 3) 125

28 17.85 475.0882 C22 H20 O12 Luteolin-7-O-D-glucuronidemethylester 10, 11, 15, 17, 21
29 18.00 319.0823 C16 H16 O7 Methyl-epigallocatechin 3, 12
30 18.60 489.1038 C23 H22 O12 Kaempferol 3-O-(6-O-acetyl) glucoside 13, 12, 11, 13, 19, 24
31 21.75 317.0303 C15 H10 O8 Myricetin 2, 3, 12, 13
32 21.90 489.1038 C23 H22 O12 4-O-Acetylquercitrin 1, 4, 5, 15, 17
33 22.95 609.125 C30 H26 O14 Prodelphinidin B3 3, 23-25
34 24.70 385.0919 C20 H18 O8 Cleomiscosin (isomer 1) 2, 8, 18, 2225
35 24.92 312.1241 C18 H19 NO4 N-feruloyltyramine 125
36 26.40 385.0929 C20 H18 O8 Cleomiscosin (isomer 2) 1, 510, 1225
37 27.70 301.0354 C15 H10 O7 Quercetin 15, 8, 1113, 15, 1820, 2325

(peak 2) was missing only in the last extract. Also, three isomers of
trimethylhydroxycitric acid, another hibiscus acid derivative, were
detected in most of the extracts. The rst and the second isomers
appeared at 6.36 and 6.69 min, respectively, but neither compound
was detected in the extracts 35, 12, 13, or in the case of the second
isomer in the extracts 20 and 23. The third isomer, which eluted at
7.75 min, was present in the extracts 6, 9, 10, 19, 2123.
Furthermore, chlorogenic acid (5-caffeoylquinic acid), a com-
pound with proven antimicrobial activity, was characterized
in Hs together with two isomers, called neochlorogenic acid
(3-caffeoylchlorogenic acid) and cryptochlorogenic acid (4-
caffeoylquinic acid). Chlorogenic acid was conrmed by compar-
ison of its retention time, 7.85 min, and its UVvis and MS spectra
with the standard. Its isomers, at retention times 6.13 and 8.31 min
for neochlorogenic and cryptochlorogenic acid respectively, were
differentiated based on the relative intensities in MS spectra and the
literature (Vallverd-Queralt et al., 2010). Cryptochlorogenic acid
showed a m/z 173 which corresponds to [quinic-H-H2 O] whereas
neochlorogenic acid does not present this loss. The rst isomer
(neochlorogenic acid) was found in the extracts 3, 5, 12, 13, 19,
23, and 24; while chlorogenic acid appeared in 3, 4, 5, 13, 19, 20,
23, and 24. Cryptochlorogenic acid was the most frequent, being
present in all the extracts in which the other two isomers appeared,
plus extracts 2, 8, and 18. Related to this antimicrobial agent, the
extracts contained two compounds, i.e., methyl esther chlorogenic
acid and ethyl esther chlorogenic acid, commonly called methyl
chlorogenate and ethyl chlorogenate. It is remarkable that these
compounds are the methyl and the ethyl esther forms of chloro-
genic acid, so in the extracts, there are three isomers of chlorogenic
acid and three isomers of these esther forms. Regarding methyl
chlorogenate, the isomers have been found at retention times 9.70,
12.30, and 13.50 min. The rst isomer was not found in extracts
15, 13, or 15 whereas the second was not present in extracts 35,
7 9, 10, 14, or 15. Lastly, the third isomer was not detected in
the extracts 110, 1215, or 21. Otherwise, the three isomers cor-
responding to ethyl chlorogenate were detected in all extracts at
retention times 13.77, 16.22, and 17.10 min.
Moreover, other phenolic compounds present in this
plant matrix have also been detected, such as 2-O-trans-
caffeoyl-hydroxycitric acid at 9.50 min (extracts 13, 8, 13,
19, 2225), coumaroylquinic acid at 11.07 min (extract 12), 5-O-
caffeoylshikimic acid at 14.00 min (extracts 3, 5, 12, 1921, 2325),
quercetin-3-glucoside at 15.5 min (extracts 2, 3, 13, and 19),
methylepigallocatechin at 18.00 min (extracts 3 and 12), myricetin
at 21.75 min (extracts 2, 3, 12, and 13), prodelphinidin B3 at 22.95
min (extracts 3, 23 25) and nally quercetin with a retention time
of 27.7 min, which was present in almost all extracts except in 6,
7, 9, 10, 14, 16, 17, 21, and 22.
Besides these well-known Hs compounds, other new
compounds were detected in the extracts: 3-deoxy-D-lyxo-
heptulosaric acid (in all extracts at 5.28 min apart from 2, 4, 5, 7,
14, and 15), syringetin-3-O-glucoside (in extracts 17, 21, and 22 at
11.40 min), kinsenoside (at 11.80 min in all Hs extracts except 15,
12, 13, 19, and 20), luteolin-7-O--D-glucuronide methyl ester (in
extracts 10, 11, 15, 17, and 21 at 17.85 min), 4 -O-acetylquercitrin
(in extracts 1, 4, 5, 15, and 17 at 21.90 min) and nally 2 isomers of
cleomiscosin, commonly named cleomiscosin A and B (at 24.70 min
in extracts 2, 8, 18, 2225, and at 26.40 min in all extracts except in
extracts 24, and 11, respectively). Moreover, different kaempferol
derivatives were found in different extracts, specically two
 I. Borrs-Linares et al. / Industrial Crops and Products 69 (2015) 385394 391

Table 3
Unknown compounds from Hibiscus sabdariffa extracts.

Peak RT (min) Measured m/z Theoretical m/z Mol. Formula Error (ppm) mSigma HS extracts

a 10.76 217.0344 217.0354 C8 H10 O7 4.4 5.3 1, 2, 6, 7, 9, 10, 11, 1417, 21, 22, 25
b 18.14 397.0787 397.0776 C17 H18 O11 2.7 1.5 2, 8, 24, 25
c 19.40 365.1248 365.1242 C18 H22 O8 1.8 19.9 2, 4, 5, 8, 9, 11, 12, 1417, 2022, 24, 25
d 20.29 357.1186 357.1191 C16 H22 O9 1.5 7.3 2
e 21.96 429.0478 429.0463 C20 H14 O11 3.3 25.3 1, 4, 5, 6, 819, 21, 22, 24, 25
f 22.55 303.1074 303.1085 C13 H20 O8 3.6 12.4 1, 8, 11, 1418, 21, 22, 25
g 23.15 411.0923 411.0933 C18 H20 O11 2.3 15.2 2
h 23.65 399.0925 399.0933 C17 H20 O11 2.1 14 2, 6, 812, 14-25
i 26.21 443.0673 443.0679 C14 H20 O16 1.2 13.5 1, 525
j 27.16 206.0213 206.0207 C8 H5 N3 O4 2.9 16.4 2

isomers of kaempferol-3-O-glucuronic acid methyl esther which
appeared at 14.44 and 15.11 min (only in extract 21), kaempferol
3-O-glucuronide at 15.46 min in all extracts apart from 2, 3, 13, 25,
as well as kaempferol 3-O-(6 -O-acetyl) glucoside (in Hs extracts
13, 1113, 19, and 24 with a retention time of 18.60 min).
Besides the effort to characterize the Hs extracts, some com-
pounds remain unknown due to lack of evidence to determine
the structure. Table 3 provides information gained by the TOF-MS
related to these ions, including their retention time, measured and
theoretical m/z, generated molecular formula, mass error (ppm),
mSigma values, and Hs extract where they were detected. In this
sense, further studies need to be performed to identify these
unknown compounds, such as MS/MS experiments.
In most cases there is a very similar composition in extracts
from different origins and varieties, such as Hs 20 (variety JR 01
SM) and 23 (Blanca variety) from Guerrero and Jalisco respectively,
as reected in Fig. 1. Nevertheless, some extracts have a more
complex phenolic composition, such as Hs 2 (variety Tepalcate-
pec, from Michoacn), compared to other extracts such as Hs 20 or
23.
3.3. Antimicrobial activity of Hs varieties
The results of the antimicrobial assays showed that all ethanol
Hs extracts have great potential as antibacterial agents against
Gram-positive and Gram-negative bacteria. Some had similar
antimicrobial capacity compared to the antibiotics used as positive
control (Table 4). In addition, it was observed that the response for
each microorganism tested was different. The Hs extracts revealed
greater antimicrobial activity against S. aureus and M. luteus than
E. coli and S. enteritidis. We found that the most sensitive organ-
isms were Gram positive S. aureus and M. luteus with an inhibition
range from 16 to 22 mm and 10 to 18 mm, respectively, while the
least sensitive were E. coli and S. enteritidis with a clear zone of
1016 mm.
The most effective Hs extracts in the inhibition of bacterial
growth were Hs 2 (Tepalcatepec variety), Hs 3 (Diamante variety)
and Hs 20 (JR 0001C variety) for E. coli, Hs 11 (El Bordo variety) for
S. enteritidis, Hs 23 (Blanca variety) for S. aureus, and Hs 2 (Tepal-
catepec variety) for M. luteus. It is important to note that three of
these extracts belong to red varieties of this plant, while only Hs

Table 4
Antimicrobial activity of Hibiscus sabdariffa extracts.

HS extract Variety Origin Inhibitory zone (mm)

E. coli S. enteritidis S. aureus M. luteus

HS 1 Americana Jalisco 14 0b 13.7 0.6def 13.7 0.6cdef 12.7 0.6fghi

HS 2 Tepalcatepec Michoacn 16 0a 15.3 0.6bc 16.0 0bc 21 1a
HS 3 Diamante Jalisco 15.7 0.6a 14.7 0.6cd 15 1bcd 17 1bc
HS 4 Colima Colima 10.3 0.6e 12 0ghi 10 0gh 10.3 0.6i
HS 5 Tempranilla Jalisco 10 0e 10.7 0.6ij 12 2e 11.3 0.6ghi
HS 6 Talpa Jalisco 11.7 0.6d 14.3 0.6cde 16 0bc 12.3 0.6fgh
HS 7 Violenta Jalisco 12 0cd 14 0cdef 10.3 0.6gh 14.3 0.6de
HS 8 Quesera Colima 10.3 0.6e 14.7 0.6cd 8.7 0.6h 12 0fghi
HS 9 CriollaTala Jalisco 14 0b 14 0cdef 13 1def 18 0b
HS 10 Tecoman Colima 10.3 0.6e 11 0ij 14 2cdef 13.3 0.6ef
HS 11 El Bordo Colima 14 0b 18.0 1a 14.3 0.6bcde 16 0cd
HS 12 Tempranilla Colima 14 0b 13.7 0.6def 15.7 0.6bc 12 0fghi
HS 13 Sudan Colima 10 0e 10 0j 10.3 0.6gh 10 0i
HS 14 Media Luna Colima 10 0e 13 0efg 10 0gh 10.7 0.6hi
HS 15 Pisila/Colima Colima 12 0cd 16.3 0.6b 15 1bcd 17.7 0.6bc
HS 16 JB 00001 SM Guerrero 12.7 0.6c 15 0 bc 14 0 cdef 16 0 cd
HS 17 JCP 0001 T Guerrero 10 0 e 13 0efg 10 0gh 13 0efg
HS 18 Mutante Blanca Guerrero 14 0b 14.7 0.6c 15 1bcd 18.7 0.6b
HS 19 JJ 00001 SM Guerrero 10 0e 13 0efg 11.7 0.6fg 12.7 0.6fg

HS 20 JR 00001 C Guerrero 16 0a 12.7 0.6fgh 16.7 0.6b 10.3 0.6i

HS 21 Americana Colima 12 0cd 11.3 0.6hij 15.7 0.6bc 13.7 0.6e
HS 22 Puerta de Anzar Colima 12 0cd 14.7 0.6c 15 1bcd 12 0fghi
HS 23 Variedad Blanca Jalisco 12 0cd 11.3 0.6hij 22 0a 13 0efg
HS 24 Real Veracruz 10 0e 10.3 0.6j 10 0gh 10.3 0.6i
HS 25 Reyna Nayarit 12 0cd 10.7 0.6ij 13 0def 13.3 0.6ef

Antibiotic Clindamycin 0 22 0 33
Erythromycin 0 24 0 35
Gentamicin 20 16 14 23
Tetracycline 24 28 13 28
Vancomycin 2 18 0 21
392 I. Borrs-Linares et al. / Industrial Crops and Products 69 (2015) 385394

Table 5
Pearson correlation coefcients of all studied variables.

Phenolics Flavonoids Anthocyanidins Delnidin Cyanidin DPPH Ant. Act. Ant. Act. Ant. Act. Ant. Act.
(E. coli) (S. enteritidis) (S. aureus) (M. luteus)

Phenolics 1
Flavonoids 0.477* 1
Anthocyanidins 0.043 0.307 1
Delnidin 0.045 0.269 0.992** 1
Cyanidin 0.029 0.411 0.843** 0.769** 1
DPPH 0.205 0.516* 0.115 0.133 0.014 1
Ant. Act. (E. coli) 0.039 0.281 0.067 0.057 0.096 0.462* 1
Ant. Act. (S. enteritidis) 0.038 -0.061 0.045 0.041 0.052 0.320 0.510* 1
Ant. Act. (S. aureus) 0.421 0.001 0.205 0.166 0.332 0.096 0.5764* 0.178 1
Ant. Act. (M. luteus) 0.3453 0.106 0.132 0.108 0.207 0.114 0.580* 0.608* 0.341 1
*
Signifance p < 0.05.
**
Signicance p < 0.001.

23 is from a color less variety. Pearson correlation coefcients
did not demonstrate a signicant correlation between antimicro-
bial activity of these extracts and the total content of anthocyanins
or cyanidin and delphinidin contents.
Plant polyphenols are considered to have antimicrobial activ-
ity, generally by the disturbance of the function of bacterial cell
membranes which retards bacterial growth or multiplication. Nev-
ertheless, other compounds such as quercetin, could act essentially
by enzyme inhibition of DNA gyrase (Cushnie and Lamb, 2005).
However, according to the antimicrobial activity of Hs extracts,
a high total content in phenolic compounds, including avonoid,
anthocyanin, cyanidin, and delphinidin, are not always correlated
to high antibacterial activity. In fact, the most potent Hs extracts
against the studied microorganisms did not exhibit the highest con-
tent in these compounds, such as Hs 2, 11, and 23. Therefore, the
antibacterial activity exhibited by these extracts could be attributed
to the presence of specic phenolic compounds in their composi-
tion and to the possible existence of synergistic effects with others
non phenolic compounds present in the Hs extracts.
The inhibition zones founded in our research for Hs extracts
agreed with those reported by Jung et al. (2013), who used
the disk diffusion method for evaluating the antimicrobial activ-
ity of an unidentied variety of Hs at different concentrations
(25200 mg/L). The study reported the inhibition zones for Bacillus
subtilis (1018 mm), E. coli (1323 mm) and S. aureus (1634 mm),
results which were in concordance with the higher sensibility
found for S. aureus in our research.
Antimicrobial activity of these Hs extracts could be used to
characterize and develop new healthy food ingredients, medical
compounds or pharmaceuticals used to control the growth of food
and human pathogens (Higginbotham et al., 2014; Higginbotham
et al., 2014).
Due to the complexity of the composition found for the Hs
extracts, it is difcult to attribute the potent antimicrobial activ-
ity of some of the extracts to a specic component or group of
compounds. Antimicrobial activity of these extracts is likely to be
caused by multiple mechanisms and synergies of various com-
pounds with different chemical forms. On this sense, future studies
need to be performed for the determination of the antimicrobial
activity of individual Hs compounds and different combination of
them, in order to stablish the compound or groups of compounds
responsible for the exerted bioactivity against the studied microor-
ganisms and the possible existence of synergic or antagonist effects.
3.4. Multivariate data analysis
The main concept of PCA is to project experimental data from
high dimensional space onto a lower dimensional one, replacing
the large number of variables by a small number of uncorrelated
principal components that can sufciently explain data structure.

Fig. 2. Biplot graph showing the grouping of 25 H. sabdariffa extracts.

 I. Borrs-Linares et al. / Industrial Crops and Products 69 (2015) 385394
Fig. 3. Clustering analysis for 25 H. sabdariffa extracts based on data derived of anthocyanidin, cyanidin and delphinidin contents.
The experimental data are ordered in such a way that the vari-
ance explained by the rst principal component is the greatest; the
variance explained by the second principal component is smaller,
and so on. In this study, after using PCA, the two principal com-
ponents (PC1 and PC2) explained 98.7% of the total variance of
the samples. The principal component analysis revealed that the
differences observed in the Hs varieties were derived from the
composition of anthocyanins. For PC1 signicant variables were
the contents in delphinidin and anthocyanin, while for PC2 signif-
icant variables were avonoid content and total phenolic content.
Examining the two dimensional PCA plot (Fig. 2), only Hs 57 were
well separated from the other extracts and composed a recogniz-
able cluster. Apparently, Hs 16, 18, 23 and 25 were also separated,
but did not congregate as closely as the other group. No separation
or clusters were observed on the PCA plot for the rest of Hs extracts.
Furthermore, clustering analysis of 25 Hs varieties (Fig. 3)
showed that the samples are divided into four well-dened groups
(AD) depending on the anthocyanin content. The different groups
(AD) present a proportion of delphinidin:cyanidin of 4:1, 3:1,
2:1 and 0:0, respectively. This relation agreed with the calyx col-
oration of the different varieties of Hs analyzed, suggesting that
delphinidin-3-sambubioside is the pigment that contributes the
most to the color of the calyx. This result also agreed with the corre-
lation of the total anthocyanin content with the individual contents
of delphinidin and cyanidin (r = 0.992 and r = 0.843, respectively).
4. Conclusions
The analytical platform used permitted the tentatively char-
acterization of seven new compounds in Hs extracts: 3-deoxy-
D-lyxo-heptulosaric acid, syringetin-3-O-glucoside, kinsenoside,
luteolin-7-O--D-glucuronide methyl ester, 4 -O-acetylquercitrin
and nally 2 isomers of cleomiscosin, commonly named cleomis-
cosin A and B. The present study reported for the rst time the
chemical composition, antioxidant activity, and antibacterial activ-
ity of 25 major varieties of roselle harvested in Mexico. The results
found highlighted the potential uses of some of the extracts as
natural sources of pigments, antioxidants and antimicrobial com-
393
pounds. Furthermore, this work is the rst report available of the
classication of 25 different Hs varieties based on a multivariate
analysis, which could be useful for varieties classication according
to their anthocyanin content in relation to content of delphinidin-
3-sambubioside and cyanidin-3-ambubioside.
Acknowledgements
The author IBL acknowledges nancial support from the
Spanish Ministry of Education and Science (FPI grant, BES-2009-
028128) and the Spanish Ministry of Economy and Competitiveness
(MINECO) and the European Social Fund (FSE) for the contract PTQ-
13-06429. The authors are also grateful to the Spanish Ministry of
Education and Science for the project AGL2011-29857-C03-02, to
Andalusian Regional Government Council of Innovation and Sci-
ence for the excellence projects P09-CTS-4564, P10-FQM-6563 and
P11-CTS-7625, as well as GREIB.PT.2011.18. The authors wanted to
thank to Aarn Morn from INIFAP Campus Tecomn for providing
some varieties of Hibiscus sabdariffa.
References
Alarcn-Aguilar, F.J., Zamilpa, A., Prez-Garca, M.D., Almanza-Prez, J.C.,
Romero-Nnez, E., Campos-Seplveda, E.A., Vzquez-Carrillo, L.I.,
Romn-Ramos, R., 2007. Effect of Hibiscus sabdariffa on obesity in MSG mice. J.
Ethnopharmacol. 114, 6671.
Ali, B.H., Al Wabel, N., Blunden, G., 2005. Phytochemical, pharmacological and
toxicological aspects of Hibiscus sabdariffa L.: a review. Phytother. Res. 19,
369375.
Beltrn-Debn, R., Alonso-Villaverde, C., Aragons, G., Rodrguez-Medina, I., Rull,
A., Micol, V., Segura-Carretero, A., Fernndez-Gutirrez, A., Camps, J., Joven, J.,
2010. The aqueous extract of Hibiscus sabdariffa calices modulates the
production of monocyte chemoattractant protein-1 in humans. Phytomedicine
17, 186191.
Bridle, P., Timberlake, C.F., 1997. Anthocyanins as natural food colours selected
aspects. Food Chem. 58, 103109.
CLSI, 2005. Performance standards for antimicrobial susceptibility testing. CLSI
approved standard M100-S15.
Chang, C., Yang, M., Wen, H., Chern, J., 2002. Estimation of total avonoid content
in propolis by two complementary colometric methods. J. Food Drug Anal. 10,
178182.
394 I. Borrs-Linares et al. / Industrial Crops and Products 69 (2015) 385394
Christian, K.R., Jackson, J.C., 2009. Changes in total phenolic and monomeric
anthocyanin composition and antioxidant activity of three varieties of sorrel
(Hibiscus sabdariffa) during maturity. J. Food Compos. Anal. 22, 663667.
Crozier, A., Jaganath, I.B., Clifford, M.N., 2009. Dietary phenolics chemistry
bioavailability and effects on health. Nat. Prod. Rep. 26, 10011043.
Cushnie, T.P.T., Lamb, A.J., 2005. Antimicrobial activity of avonoids. Int. J.
Antimicrob. Agents 26, 343356.
Djeridane, A., Yous, M., Nadjemi, B., Boutassouna, D., Stocker, P., Vidal, N., 2006.
Antioxidant activity of some Algerian medicinal plants extracts containing
phenolic compounds. Food Chem. 97, 654660.
Fernndez-Arroyo, S., Rodrguez-Medina, I.C., Beltrn-Debn, R., Pasini, F., Joven, J.,
Micol, V., Segura-Carretero, A., Fernndez-Gutirrez, A., 2011. Quantication of
the polyphenolic fraction and in vitro antioxidant and in vivo anti-hyperlipemic
activities of Hibiscus sabdariffa aqueous extract. Food Res. Int. 44, 14901495.
Fernndez-Arroyo, S., Herranz-Lpez, M., Beltrn-Debn, R., Borrs-Linares, I.,
Barrajn-Cataln, E., Joven, J., Fernndez-Gutirrez, A., Segura-Carretero, A.,
Micol, V., 2012. Bioavailability study of a polyphenol-enriched extract from
Hibiscus sabdariffa in rats and associated antioxidant status. Mol. Nutr. Food
Res. 56, 15901595.
Frimpong, G., Adotey, J., Ofori-Kwakye, K., Lugrie Kipo, S., Dwomo-Fokuo, Y., 2014.
Potential of aqueous extracts of Hibiscus sabdariffa calyces as coloring agent in
three pediatric oral pharmaceutical formulations. J. Appl. Pharm. Sci. 4,
001007.
Gurrola-Daz, C.M., Garca-Lpez, P.M., Snchez-Enrquez, S., Troyo-Sanromn, R.,
Andrade-Gonzlez, I., Gmez-Leyva, J.F., 2010. Effects of Hibiscus sabdariffa
extract powder and preventive treatment (diet) on the lipid proles of patients
with metabolic syndrome (MeSy). Phytomedicine 17, 500505.
Herranz-Lpez, M., Fernndez-Arroyo, S., Prez-Sanchez, A., Barrajn-Cataln, E.,
Beltrn-Debn, R., Menndez, J.A., Alonso-Villaverde, C., Segura-Carretero, A.,
Joven, J., Micol, V., 2012. Synergism of plant-derived polyphenols in
adipogenesis: perspectives and implications. Phytomedicine 19, 253261.
Hervert-Hernndez, D., Goni, I., 2012. Contribution of beverages to the intake of
polyphenols and antioxidant capacity in obese women from rural Mexico.
Public Health Nutr. 15, 612.
Higginbotham, K.L., Burris, K.P., Zivanovic, S., Davidson, P.M., 2014. Aqueous
extracts of Hibiscus sabdariffa calyces as an antimicrobial rinse on hot dogs
against Listeria monocytogenes and methicillin-resistant Staphylococcus aureus.
Food Control 40, 274277.
Hopkins, A.L., Lamm, M.G., Funk, J.L., Ritenbaugh, C., 2013. Hibiscus sabdariffa L. in
the treatment of hypertension and hyperlipidemia: a comprehensive review of
animal and human studies Hibiscus sabdariffa L. in the treatment of
hypertension and hyperlipidemia: a comprehensive review of animal and
human studies. Fitoterapia 85, 8494.
Jung, E., Kim, Y., Joo, N., 2013. Physicochemical properties and antimicrobial
activity of roselle (Hibiscus sabdariffa L.). J. Sci. Food Agric. 93, 37693776.
Kuo, C., Kao, E., Chan, K., Lee, H., Huang, T., Wang, C., 2012. Hibiscus sabdariffa L.
extracts reduce serum uric acid levels in oxonate-induced rats Hibiscus
sabdariffa L. extracts reduce serum uric acid levels in oxonate-induced rats. J.
Funct. Foods 4, 375381.
Laikangbam, R., Damayanti Devi, M., 2012. Inhibition of calcium oxalate crystal
deposition on kidneys of urolithiatic rats by Hibiscus sabdariffa L. extract. Urol.
Res. 40, 211218.
Lin, H., Chan, K., Sheu, J., Hsuan, S., Wang, C., Chen, J., 2012. Hibiscus sabdariffa leaf
induces apoptosis of human prostate cancer cells in vitro and in vivo Hibiscus
sabdariffa leaf induces apoptosis of human prostate cancer cells in vitro and in
vivo. Food Chem. 132, 880891.
Mohd-Esa, N., Hern, F.S., Ismail, A., Yee, C.L., 2010. Antioxidant activity in different
parts of roselle (Hibiscus sabdariffa L.) extracts and potential exploitation of the
seeds. Food Chem. 122, 10551060.
Patel, S., 2014. Hibiscus sabdariffa: an ideal yet under-exploited candidate for
nutraceutical applications Hibiscus sabdariffa: an ideal yet under-exploited
candidate for nutraceutical applications. Biomed. Prevent Nutr. 4,
2327.
Re, R., Pellegrini, N., Proteggente, A., Pannala, A., Yang, M., Rice-Evans, C., 1999.
Antioxidant activity applying an improved ABTS radical cation decolorization
assay. Free Radical Bio. Med. 26, 12311237.
Rodrguez-Medina, I.C., Beltrn-Debn, R., Molina, V.M., Alonso-Villaverde, C.,
Joven, J., Menndez, J.A., Segura-Carretero, A., Fernndez-Gutirrez, A., 2009.
Direct characterization of aqueous extract of Hibiscus sabdariffa using HPLC
with diode array detection coupled to ESI and ion trap MS. J. Sep. Sci. 32,
34413448.
Sindi, H.A., Marshall, L.J., Morgan, M.R.A., 2014. Comparative chemical and
biochemical analysis of extracts of Hibiscus sabdariffa. Food Chem. 164,
2329.
Singleton, V.L., Rossi, J.A., 1956. Colorimetry of total phenolics with
phosphomolybdic-phosphotungstic acid reagent. Am. J. Enol. Viticult. 16,
144158.
Tsai, P., McIntosh, J., Pearce, P., Camden, B., Jordan, B.R., 2002. Anthocyanin and
antioxidant capacity in Roselle (Hibiscus sabdariffa L.) extract. Food Res. Int. 35,
351356.
Vallverd-Queralt, A., Juregui, O., Medina-Remn, A., Andrs-Lacueva, C.,
Lamuela-Ravents, R.M., 2010. Improved characterization of tomato
polyphenols using liquid chromatography/electrospray ionization linear ion
trap quadrupole orbitrap mass spectrometry and liquid
chromatography/electrospray ionization tandem mass spectrometry. Rapid
Commun. Mass Spectrom. 24, 29862992.
Wang, C., Wang, J., Lin, W., Chu, C., Chou, F., Tseng, T., 2000. Protective effect of
Hibiscus anthocyanins against tert-butyl hydroperoxide-induced hepatic
toxicity in rats. Food Chem. Toxicol. 38, 411416.
Wang, M.L., Morris, B., Tonnis, B., Davis, J., Pederson, G.A., 2012. Assessment of oil
content and fatty acid composition variability in two economically important
Hibiscus species. J. Agric. Food Chem. 60, 66206626.
Wang, J., Cao, X., Jiang, H., Qi, Y., Chin, K.L., Yue, Y., 2014. Antioxidant activity of leaf
extracts from different Hibiscus sabdariffa accessions and simultaneous
determination of ve major antioxidant compounds by LC-QTOF-MS.
Molecules 19, 2122621238.
Wojdylo, A., Oszmianski, J., Czemerys, R., 2007. Antioxidant activity and phenolic
compounds in 32 selected herbs. Food Chem. 105, 940949.
Xia, E., Deng, G., Guo, Y., Li, H., 2010. Biological activities of polyphenols from
grapes. Int. J. Mol. Sci. 11, 622646.
Zhang, Z., Kou, X., Fugal, K., McLaughlin, J., 2004. Comparison of HPLC methods for
determination of anthocyanins and anthocyanidins in bilberry extracts. J.
Agric. Food Chem. 52, 688691.