Food

Chemistry
Food Chemistry 105 (2007) 325–333
www.elsevier.com/locate/foodchem

Analytical, Nutritional and Clinical Methods

Comparison of extraction solvents and techniques used for the assay

of isoflavones from soybean
Devanand L. Luthria a,*, Ronita Biswas a, Savithiry Natarajan b
a
Food Composition Laboratory, Beltsville Human Nutrition Research Center, ARS, USDA, Room 202 B, Building 161, BARC-E,
Beltsville, MD 20705-3000, United States
b
Soybean Genomics and Improvement Laboratory, Plant Sciences Institute, ARS, USDA, Beltsville, MD 20705-3000, United States

Received 1 November 2006; received in revised form 13 November 2006; accepted 16 November 2006

Abstract

The impact of extraction solvents and techniques on the assay of isoflavones from soybean was investigated. This systematic study
was undertaken to address substantial variations in the solvents and procedures used for the extraction of isoflavones from soybeans
by different research groups as described in the recent peer-reviewed published literature. Comparison of four previously optimized
and commonly deployed solvent mixtures (acetonitrile:water 58:42 (v/v); ethanol:water 70:30 (v/v); methanol:water 90:10 (v/v); super-
heated pressurized water) was carried out for the extraction of isoflavones. In addition, we also examined the extraction efficiencies
of three additional new solvent mixtures (dimethyl sulphoxide:acetonitrile:water, 5:58:37 (v/v/v); dimethyl sulphoxide:ethanol:water,
5:70:25 (v/v/v); Genapol:water 5:95 (v/v)) for the extraction of isoflavones from soybeans. Assessments of six commonly used extraction
techniques (shaking, vortexing, sonication, stirring, Soxhlet, and pressurized liquid extraction (PLE)) with an optimized extraction sol-
vent mixture was also performed. Both, the total isoflavones content and the isoflavones HPLC profile varied significantly with different
extraction solvents and techniques. Optimum total isoflavones recoveries from soybean samples were obtained with dimethyl sulphox-
ide:ethanol:water (5:75:25, v/v/v) solvent mixture using a PLE. Intermediate extraction recoveries of total isoflavones from soybean sam-
ples were obtained with the other extraction solvent mixtures and techniques tested. The extraction efficiencies of isoflavones with shaker,
vortex, stirring, and Soxhlet were between 65% and 70% as compared to PLE. Total isoflavones extracted by the sonication procedure
was 93.3% as compared to PLE.
Published by Elsevier Ltd.

Keywords: Soybeans; Isoflavones; Extraction solvent optimization; Stirring; Shaking; Vortexing; Pressurized liquid extraction; Sonication; HPLC

1. Introduction Rao, 1997; Lee, Wang, Murphy, & Hendrich, 1995; Slavin,
Jacobs, & Marquart, 1997; Vesper et al., 1999). Isoflavones
Interest in soybeans and soy based products has grown enriched extracts have been evaluated in the prevention of
significantly in the past few decades due to the association a wide range of health problems associated with meno-
of soybean consumption with variety of health protective pause, cardiovascular disease, osteoporosis, and in breast,
effects (Hasler, 1998). Soybeans are known to contain a prostate, and colon cancers (Anderson & Garner, 1997;
large number of bioactive phytochemicals such as isoflav- Caragay, 1992; Messina, 1999). Isoflavones are widely dis-
ones, saponins, phytosterols, protease inhibitors, inositol tributed in the plant kingdom, but accumulate predomi-
hexaphosphates, sphingolipids, phenolic acids, and Bow- nantly in plants of the Leguminosae family. The best
man-Birk trypsin inhibitors (Kennedy, 1998; Koratkar & natural source of isoflavones is soybeans, which have been
a major part of the traditional diet for eastern Asian pop-
ulations for centuries. The global annual consumption of
*
Corresponding author. Tel.: +1 301 504 7247; fax: +1 301 504 8314. soybeans has increased from 114 to 170 million tons during
E-mail address: luthriad@ba.ars.usda.gov (D.L. Luthria). the past decade (Klejdus et al., 2004). Soybeans contain

0308-8146/$ - see front matter Published by Elsevier Ltd.

doi:10.1016/j.foodchem.2006.11.047
326 D.L. Luthria et al. / Food Chemistry 105 (2007) 325–333

1.2–2.4 mg of total isoflavones per gram of sample (Rost- tion solvent composition (methanol, ethanol, and
agno, Palma, & Barroso, 2004). This hundred percent var- acetonitrile with different proportions of acidified and
iation in isoflavones content in soybeans is due to variation non- acidified water) has been used. Rostagno et al.
in genotypes, environment, location, post harvest storage, (2004) compared the extraction efficiency of isoflavones
and assay procedures. with different ethanol and methanol water mixtures (30–
Isoflavones are oxygen heterocycles containing a 80%) using a pressurized liquid extractor (PLE). They
3-phenylchroman skeleton that is hydroxylated at 40 and reported that optimum yields of isoflavone content were
7 positions (Penalvo, Nurmi, & Adlercreutz, 2004) obtained with 70% ethanol in water. In another study,
(Fig. 1). Based on the substitution pattern on carbons 5 Klejdus et al. (2004) reported that methanol:water
and 6, three aglycon forms of isoflavones commonly found (90:10%, v/v) was the best solvent mixture for extracting
in soybeans are daidzein, genistein, and glycitein. These isoflavones from soybean samples. At about the time,
three isoflavones can also exist in conjugated forms with Lin and Giusti (2005) compared the extraction efficiencies
glucose (daidzin, geinstin, glycitin), malonylglucose (malo- of different aqueous (acidified and non-acidified) mixtures
nyldaidzin, malonylgeinstin, malonylglycitin), and acetyl- of methanol and acetonitrile solvent mixtures and
glucose (acetyldaidzin, acetylgeinstin, acetylglycitin) units. reported that 58% non-acidified acetonitrile provided
Thus 12 free and conjugated forms of isoflavones have optimum extraction of isoflavones. Thus wide variations
been isolated from different soybean samples (Fig. 1). in the usage of different solvent mixtures and techniques
As 12 different soy isoflavones have a wide variation in have been applied for the extraction of isoflavones from
polarities, development of an optimized extraction proce- soybeans and other plant matrices.
dure for all isoflavones has been a challenging task. Dif- The amounts of isoflavones in plants, fruits, vegetables
ferent researchers have deployed various solvents and and herbs are influenced by genotypes, agronomic prac-
techniques for extraction of isoflavones from soybeans. tices, maturity at harvest, and post-harvest, storage, cli-
The data presented in Table 1 summarizes some of the matic, regional, and processing conditions (Achouri,
most popular methods for extraction of isoflavones used Boye, & Belanger, 2005; Griffith & Collison, 2001; Kao,
by various research groups found in peer-reviewed litera- Su, & Lee, 2004; Klejdus et al., 2005; Lee et al., 2004; Mur-
ture during the past five years. Researchers have utilized phy, Barua, & Hauck, 2002; Wu, Wang, Sciarappa, &
different extraction equipment or techniques such as Simon, 2004). To minimize these variations and to evaluate
multi-wrist shaker, rotary shaker, stirring, sonicator, the influence of sample preparation (extraction solvents
Soxhlet, supercritical fluid extractor, and pressurized and techniques), all analyses were performed on one
liquid extractor for extraction of isoflavones from soy- homogenous sample obtained by grinding soybeans pro-
beans (Table 1). In addition, a wide variation in extrac- cured from a single source.

R3O 8
O
2

R2 5
4’
R1 O OH

Name R1 R2 R3

Daidzein H H H
Glycitein H OCH3 H
Genistein OH H H
Daidzin H H Glu
Glycitin H OCH3 Glu
Genistin OH H Glu
Acetyldaidzin H H Glu-COCH3
Acetylglycitin H OCH3 Glu-COCH3
Acetylgenistin OH H Glu-COCH3
Malonyldaidzin H H Glu-COCH2COOH
Malonylglycitin H OCH3 Glu-COCH2COOH
Malonylgenistin OH H Glu-COCH2COOH

Fig. 1. Chemical structures of 12 isoflavones isolated from soybeans.

 D.L. Luthria et al. / Food Chemistry 105 (2007) 325–333 327

Table 1
Summary of commonly deployed procedures used by various authors for extraction of isoflavones soybean samples
Serial # Extraction procedure Best solvent Publication Reference
year
1 Ground manokin soybeans (2 g) were extracted with 12 mL of six different 58% ACN 2005 Lin & Giusti
extraction solvents (83% ACN, 83% acidified ACN, 80% MeOH, 80%
acidified MeOH, 58% ACN, and 58% acidified ACN). Extractions of
isoflavones were carried out at RT for 2 h
2 Soy meal was extracted with acidified ACN (58%) by shaking sample for One solvent 2004 Lee et al.
2 h at RT
3 Comparison of four different sample preparation (extraction and 80% EtOH containing 2004 Penalvo et al.,
hydrolysis) procedures with three different solvents (80% EtOH, 80% 1 M HCl
acidified EtOH, and 80% MeOH) was carried out at various temperature
(60–100 °C) for different time intervals (30–240 min)
4 Two grams of different food samples were extracted with four different 53% ACN 2002 Murphy
acidified solvents 53% ACN, 53% MeOH, 53% EtOH and 53% acetone. et al.,
The mixture was stirred at room temperature for 2 h
5 Defatted soy samples were extracted with 80% ACN, 80% MeOH and 80% Similar yields with all 2005 Achouei
EtOH by stirring and sonicating samples. The authors suggest replicate three solvents et al.,
extraction are essential for optimum extraction. Significant decrease in
isoflavone yields were obtained with one extraction cycle
6 Freeze dried soybeans were extracted with pressurized liquid extractor 70% EtOH 2004 Rostango
(PLE) with different proportions of (30–80%) of aqueous EtOH and MeOH et al.,
solvent mixtures at different temperature, pressures, solid-to-solvent ratios.
Optimization of PLE conditions are described
7 The authors used a high pressure hot water process to extract five One solvent 2004 Li-Hsun
isoflavones from detffated soy flakes An experimental design was adopted et al.,
to optimize extraction conditions, and five isoflavones were purified. The
optimum conditions were determined as 110 °C and 641 psig over 2.3 h of
extraction
8 Comparison of different extraction conditions (extraction time, 90% MeOH 2004 and 2005 Klejdus et al.,
temperature, pressure, cycles, solvents) and techniues (PLE,
PLE + sonication, sonication, and Soxhlet)

This paper provides a direct comparison of four com- aliquots in amber bottles. Each bottle was flushed with
monly used extraction solvents or solvent mixtures (aceto- nitrogen, and stored in a freezer (<60 °C) until analyzed.
nitrile:water (58:42, v/v); ethanol:water (70:30, v/v);
methanol:water (90:10, v/v); superheated pressurized 2.2. Chemicals
water) used for the extraction of isoflavones from soybeans
as documented in the peer reviewed literature (Table 1). We HPLC grade acetonitrile and methanol were purchased
have also examined efficiencies of three additional solvent from Fisher Chemicals (Fair Lawn, NJ, USA), and dena-
mixtures (dimethyl sulphoxide:ethanol:water (5:70:25, v/ tured anhydrous ethanol was obtained from Mallinckrodt
v/v), dimethyl sulphoxide:acetonitrile:water, (5:58:37, v/v/ (Paris, KY, USA). HPLC-grade formic acid and dimethyl
v) and 5% Genapol, a nonionic surfactant oligo (ethylene- sulphoxide was procured from Aldrich Chemical Com-
glycol) monoalkyl ether in water) for extraction of isoflav- pany (Milwaukee, WI, USA). Diatomaceous earth (Celite
ones from soybeans. In addition, we have compared 545) and Ottawa sand were purchased from Fisher Chem-
extraction efficiencies of six commonly used extraction icals (Fair Lawn, NJ, USA). Deionized water (18 X) was
techniques (stirring, Soxhlet, sonication, PLE, vortexing prepared using a Millipore Milli-Q purification system
and mechanical wrist shaker) for the extraction of isoflav- (Millipore Corp., New Bedford, MA, USA). Genapol
ones from soybeans. X-080 was purchased from Sigma–Aldrich Chemie Gmbh
(Munich, Germany). Polyvinylidene difluoride (PVDF)
2. Materials and methods syringe filters with pore size 0.45 lm were purchased from
National Scientific Company (Duluth, GA, USA). Isoflav-
2.1. Plant material one standards (Genistein, Daidzin, Genistin, Daidzein,
and Glycitin) were purchased from LC Laboratories
Soybean seeds were purchased from a local grocery (Woburn, MA, USA).
store (Mom in Beltsville, Maryland). Immediately, after
receipt, the material was ground in a home style coffee grin- 2.3. Comparison of extraction solvents
der. The ground material was then passed through a stan-
dard 20-mesh sieve (particle size <0.825 mm). The ground All extractions for the comparison of extraction efficien-
sample was mixed thoroughly and subdivided into multiple cies of different solvent mixtures were carried out with an
328 D.L. Luthria et al. / Food Chemistry 105 (2007) 325–333

Accelerated Solvent Extractor (ASE) from Dionex Corpo- in a sonicator bath (Branson 2510, Branson Ultrasonic
ration (Model ASE 200, Dionex Corporation, Sunnyvale, Corporation, Danbury, CT, USA) at ambient temperature
CA, USA). Aliquots of 500 ± 1 mg of ground soybean for 15 min. Extraction with vortexing was performed by
seeds were placed in an 11 ml stainless steel extraction cell. vortexing the vials for 1 min (three times with each cycle)
Two circular cellulose filters (size 1.983 mm, Dionex Cor- on a Daigger Vortex Genie 2 (Scientific Industries, Inc.,
poration) were placed at the bottom of the extraction cell Bohemia, NY, USA). After extraction with the different
in order to prevent suspended particles from entering the procedures, the mixture was centrifuged at a low speed
collection vials. The remaining void volume in the cell (838g) for approximately 10 min. The supernatant was
was filled with Ottawa Sand. Both extraction cells and transferred into a 25 ml volumetric flask. The residue
extract collection amber vials were arranged appropriately was re-suspended in the fresh extraction solvent mixture
in the two designated carousels. Extractions were carried (8 ml dimethyl sulphoxide:acetonitrile:water (5:58:37, v/v/
out using seven different solvent mixtures 1. acetoni- v)) each time and the extraction process was repeated
trile:water (58:42, v/v); 2. dimethyl sulphoxide:acetoni- twice. A total of three extraction cycles were performed
trile:water, (5:58:37, v/v/v); 3. ethanol:water (70:30, v/v); with each procedure. The supernatant from the three
4. methanol:water (90:10, v/v); 5. dimethyl sulphoxide:eth- extraction cycles were pooled and the volume of the com-
anol:water, (5:70:25, v/v/v); 6. water; and 7. genapol:water bined extract was adjusted to 25 ml with extraction solvent
(5:95, v/v). Extractions were performed under previously and appropriate aliquots of extracts were filtered through
optimized conditions as reported by Rostagno et al. a 0.45 lm PVDF syringe filter for isoflavone analysis by
(2004). In short, all extractions were carried out at HPLC analysis. Four replicate extractions and duplicate
1000 psi at 100 °C, with a 5 min equilibration time, a HPLC analyses of each extract were carried out for each
7 min static time, and a 90 s purge time for each extraction sample.
cycle. A total of three extraction cycles were performed for
each sample. The extraction temperature was set to 115 °C 2.4.2. Soxhlet extraction of isoflavones from soybean
only in the case when water was the extraction solvent. The samples
extracts were collected in 60 ml amber sample vials fitted Soxhlet extractions were carried out by placing ground
with Teflon coated rubber caps (I-CHEM, New Castle, soybean sample (3 g ± 1 mg) in a Whatman filter paper #
DE, USA). Each extract was transferred to a 25 ml volu- 4. The filter paper containing sample was wrapped and
metric flask and the total volume was adjusted to 25 ml placed in a thimble. The thimble was than placed in a Soxh-
with the appropriate extraction solvent mixture. These ali- let extraction apparatus. Approximately 125 ml of extrac-
quots of soybean extracts were filtered through a 0.45 lm tion solvent (8 ml dimethyl sulphoxide:acetonitrile:water
PVDF syringe filter prior to analysis of isoflavones by (5:58:37, v/v/v)) was added into the Soxhlet apparatus.
HPLC assay. Four replicate extractions and duplicate All samples were extracted with hot boiling solvent mixture
HPLC analyses of each extract were carried out for each for 3 h. After 3 h, extraction apparatus was rinsed twice
sample. with additional 12.5 ml extraction solvent mixture. The
total volume of the soybean extract was adjusted to
2.4. Comparison of extractions techniques 150 ml. Appropriate aliquots of extracts were filtered
through a 0.45 lm PVDF syringe filter for isoflavone anal-
Comparison with five other extraction techniques (vor- ysis by HPLC analysis. Four replicate extractions and
texing, shaking, stirring, sonication and Soxhlet) com- duplicate HPLC analyses of each extract were carried out
monly used for extraction of isoflavones from soybean for each sample.
samples was carried out using the single optimized extrac-
tion solvent mixture (dimethyl sulphoxide:acetonitrile: 2.5. Determination of phenolic compounds by HPLC
water, (5:58:37, v/v/v)) with the same solid-to-solvent ratio.
Analysis of isoflavones from all extraction procedures
2.4.1. Extraction of isoflavones from soybean samples by and seven different extraction solvents was carried out
stirring, shaker, sonication, and vortexing procedures using an Agilent 1100 HPLC system consisting of a binary
Approximately 500 ± 1 mg of ground soybean sample pump with a vacuum degasser, a thermostatted column
were placed in 16  125 mm screw cap vial and 8 ml of compartment, an auto-sampler, and a diode array detector
extraction solvent mixture was added to each tube. Stirring (DAD, Agilent Technologies, Palo Alto, CA). Separation
was carried out by adding one 8  1.5 mm PVDF coated of isoflavones was achieved using a reversed phase
magnetic bar to each vial and placing vials in a 100 ml bea- C18Luna column (Phenomenex, Lorance, CA, USA,
ker on a Corning hot plate/stirrer (Model PC 351). For 150  4.6 mm; particle size 5 lm), preceded by a guard col-
shaker, extractions were carried out by placing extraction umn (Phenomenex, 4  3.0 mm) of the same stationary
vials containing ground soybean sample with 8 ml extrac- phase. The HPLC gradient and conditions were similar
tion solvent on a Burrell Wrist Shaker (Model 75, Burrell to those described by Lee et al. (2004) and Murphy et al.
Corporation, Pittsburg, PA, USA) at a high speed (setting (1997) with minor modifications. Solvent A consisted of
# 10) for 60 min. For sonication, sample vials were placed 1% (v/v) acetic acid in water and solvent B was 100% ace-
 D.L. Luthria et al. / Food Chemistry 105 (2007) 325–333 329

tonitrile. Flow rate was set to 0.6 ml/min and the column spectral analysis as reported in literature (Griffith & Colli-
and the guard column were thermostatically controlled at son, 2001; Klejdus et al., 2004; Wu et al., 2004). The twelve
25 °C. For the first 5 min gradient was maintained at isoflavones were identified as 1. daidzin, 2. glycitin, 3. gen-
15% B. The gradient elution was changed from 15% to istin, 4. malonyldaidzin, 5. malonylglycitin, 6. acetyldaid-
45% B in a linear fashion for a duration of 44 min. Column zin, 7. acetylglycitin, 8. malonylgenistin, 9. daidzein, 10.
was washed at 100% B for 5 min and equilibrated for addi- glycitein, 11. acetylgenistin, and 12. genistein. Table 2 sum-
tional 15 min at 15% B. Total analysis time per sample was marizes the HPLC peak areas of each of the twelve identi-
70 min. HPLC chromatograms were detected using a photo fied isoflavones with four different solvent mixtures
diode array UV detector. The structures for all isoflavones commonly deployed for extraction of isoflavones as
were confirmed by comparison of retention time, UV spec- reported in the peer-reviewed literature (Table 1). We have
tra and mass spectral analysis. Peak areas were integrated evaluated extraction efficiencies of isoflavones from soy-
for quantitation. As authentic purified standard of for all beans with three additional solvent mixtures (dimethyl sul-
twelve isoflavones were not easily available. Comparison phoxide:ethanol:water (5:70:25, v/v/v); dimethyl
of extraction efficiencies was achieved by comparing HPLC sulphoxide:acetonitrile:water (5:58:37, v/v/v); and Gen-
peak areas. apol:water (5:95, v/v)). Addition of dimethyl sulphoxide
to the two previously optimized extraction solvent mixtures
2.6. Statistical analysis (ethanol:water (70:30, v/v); acetonitrile:water (58:42, v/v))
was carried out based on the solubility of isoflavones
All statistical analyses were conducted using ANOVA reported at the Sigma-Aldrich website (http://www.sigma-
and t-test using a statistical analysis software (SASÒ v aldrich.com). In addition, Griffith and Collison (2001)
9.1.2 from SAS Institute, Inc., Cary, NC, USA). Individual reported improvement (0.7–10.6%) in the extraction effi-
peak areas and the total areas were compared with each ciency of different isoflavones from soy samples extracted
other using least significant difference (LSD) method to with internal standard apigenin dissolved in dimethyl sul-
determine any significant difference (p < 0.001) among phoxide. The authors suggested that marginal increase in
them. isoflavones content might be attributed to the lack of acid
or to the presence of small quantity of dimethyl sulphoxide.
3. Results and discussion The authors suggested that additional research was needed
to evaluate the influence of dimethyl sulphoxide on extrac-
3.1. Effect of solvent composition on extraction of isoflavones tion efficiency of isoflavones.
In the current study, as purified standards for all 12
A typical HPLC chromatogram of the isoflavones isoflavones were either not commercially available or very
extracted from soybean sample is shown in Fig. 2. The expensive, evaluation of extraction efficiency was carried
structures of the 12 isoflavones were confirmed by compar- out by comparison of HPLC areas. Individual peak areas
ison of the UV spectra, retention times as well as mass and the HPLC profiles for different isoflavones were

DAD1 A, Sig=260,4 Ref=700,100 (R060720A\030-4101.D)

mAU
3

50 8

40
1 9
30 4

20 11
6
5
10 2
7 10 12

0 1
5 10 15 20 25 30 35 40 mi

Fig. 2. A typical HPLC profile of isoflavones extracted from soybean. The peaks are identified as 1. Daidzin, 2. Glycitin, 3. Genistin, 4. Malonyldaidzin,
5. Malonylglycitin, 6. Acetyldaidzin, 7. Acetylglycitin, 8. Malonylgenistin, 9. Daidzein, 10. Glycitein, 11. Acetylgenistin, and 12. Genistein.
 330
Table 2
Average HPLC peak areas from four replicate soybean extracts analyzed in duplicate for 12 isoflavones extracted with eight different solvents mixtures
Solvents Areas
Daidzin Glycitin Genistin Malonyl Malonyl Acetyl Acetyl Malonyl Daidzein Glycitein Acetyl Genistein SUM
Daidzin Glycitin Daidzin Glycitin Genistin Genistin
58% ACN 265.8d 62.0d 363.8e 174.6c 0.0c 46.8d 295.4d 32.5b,c 0.0c 51.7c 65.6d 46.5c 1401.8e
58% ACN, 5% DMSO 175.3f 53.0d 662.6d 351.2a 0.0c 157.2a 606.1b 30.5c 0.0c 73.2a,b 213.2a 84.0b 2406.4d
70% ETOH 571.7c 88.8c 784.4c 386.3a 226.6b 78.6c 58.1e 38.1b 752.8b 61.0b,c 67.6d 68.3b 3182.7c
70% ETOH, 5% DMSO 662.2b 111.2b 920.4b 382.6a 235.5a 117.2b 707.5a 69.2a 1054.4a 76.9a 172.8b 83.9b 4594.3a
90% MEOH 1065.1a 179.9a 1476.2a 243.8b 0.0c 109.5b 430.4c 0.00d 0.0c 83.1a 152.0c 106.0a 3846.2b
Water 243.9d,e 37.1e 309.2e,f 65.2d 0.0c 25.5e 95.4e 0.00d 0.0c 22.4d 25.8e 12.4d 837.1f

D.L. Luthria et al. / Food Chemistry 105 (2007) 325–333

5% genipol 191.0e,f 31.5e 230.4f 42.3d 0.0c 21.4e 61.4e 0.00d 0.0c 15.2d 19.4e 15.5d 628.4g
LSD 68.2 11.2 86.3 58.4 7.6 15.8 63.3 7.0 6.5 12.2 13.2 17.9 128.4
The average areas of the peaks with any identical letter are not significantly different with each other at the a = 0.05 level by least significant difference test; n = 3. ACN, acetonitrile; DMSO, dimethyl
sulphoxide; ETOH, ethanol; MEOH, methanol; LSD, least significant difference.

Table 3
Comparison of the recovery of individual isoflavones extracted from soybeans using different solvent mixtures
Solvents Areas (%)
Daidzin Glycitin Genistin Malonyl Malonyl Acetyl Acetyl Malonyl Daidzein Glycitein Acetyl Genistein SUM
Daidzin Glycitin Daidzin Glycitin Genistin Genistin
58% ACN 25.0 34.5 24.6 45.2 0 29.77 41.8 47.0 0.0 67.2 30.8 43.9 30.5
58% ACN, 5% 16.5 29.5 44.9 100.0 0 100.0 85.7 44.1 0.0 95.2 100.0 79.2 52.4
DMSO
70% ETOH 53.7 49.4 53.1 100.0 96.2 50.0 8.2 55.1 71.4 79.3 31.7 64.4 69.3
70% ETOH, 5% 62.2 61.8 62.3 100.0 100.0 74.6 100.0 100.0 100.0 100.0 81.1 79.2 100.0
DMSO
90% MEOH 100.0 100.0 100.0 63.1 0 69.7 60.8 0.0 0.0 100.0 71.3 100.0 83.7
Water 22.9 20.6 20.9 16.9 0 16.2 13.5 0.0 0.0 29.1 12.1 11.7 18.2
5% genipol 17.9 17.5 15.6 11.0 0 13.6 8.7 0.0 0.0 19.8 9.1 14.6 13.7
ACN, acetonitrile; DMSO, dimethyl sulphoxide; ETOH, ethanol; MEOH, methanol.
 D.L. Luthria et al. / Food Chemistry 105 (2007) 325–333 331

significantly dissimilar with different extraction solvent flavones (malonyldaidzin, malonylglycitin, acetyldaidzin,
mixtures. Optimum extraction of the total isoflavones acetylglycitin, malonylgenistin, daidzein, glycitein, acetyl-
content as measured by the HPLC peak areas was genistin, and genistein) were obtained with (dimethyl sulph-
obtained when extractions were performed with dimethyl oxide:ethanol:water (5:70:25, v/v/v)) (Fig. 3). All twelve
sulphoxide:ethanol:water (5:70:25, v/v/v) solvent mixture. isoflavones were detected in measurable quantities only with
The extraction efficiency of all other solvent mixtures were two different extraction solvent mixtures (dimethyl sulphox-
compared with optimized extraction solvent mixture ide:ethanol:water (5:70:25, v/v/v) and ethanol:water (70:30,
(dimethyl sulphoxide:ethanol:water (5:70:25, v/v/v)). Low- v/v)). The two isoflavones, daidzein and malonylglycetin
est yields (13.7 and 18.2%) of total isoflavones were were not extracted in measurable amounts with five other
obtained with water and 5% Genepol in water. Evaluation extraction solvent mixtures acetonitrile:water (58:42, v/v);
with nonionic surfactant oligo (ethyleneglycol) monoalkyl dimethyl sulphoxide:acetonitrile:water (5:58:37, v/v/v);
ether (Genapol X-080) was carried out based on the methanol:water (90:10, v/v); water; and Genapol:water
recent study on extraction of isoflavones diadzein from (5:95, v/v) evaluated in the current study. Similarly, Klejdus
Puerariae radix (He et al., 2005). The authors obtained et al. (2005) extracted trace quantity of daidzein (1.2% of
optimum extraction of isoflavones with 5% Genapol total isoflavones extracted) and did not detect presence of
X-080 (w/v). However, in the present study significant any malonyl glycitin from soybeans with methanol:water
decrease in isoflavones yields were obtained with 5% Gen- (90:10,% v/v) as extraction solvent. In another study, Lin
apol in water. This variation in extraction efficiencies may and Giusti (2005) reported trace quantities (1.2% and
be attributed to the differences in the sample matrix, iso- 1.3%) of the two isoflavones daidzein and malonylglycetin
flavones content and composition. Only 30.5% of the iso- with acetonitrile:water (58:42, v/v) solvent mixture. Insig-
flavones were extracted with acetonitrile:water (58:42, v/v) nificant amounts of the above two isoflavones (daidzein
solvent mixture. The yield of isoflavones content increased and malonyl–glycetin) were extracted from soybeans with
to 52.3%, when extractions were performed with dimethyl superheated water at elevated pressures (Li-Hsun, Ya-
sulphoxide:acetonitrile:water (58:37:5, v/v/v) solvent mix- Chuan, & Chieh-Ming, 2004).
ture. Similar increase in extraction yields of isoflavones
were obtained when 5% DMSO was added to etha- 3.2. Effect of extraction techniques on assay of isoflavones
nol:water (70:30, v/v) solvent mixture. Intermediate yields from soybeans
(83.7%) of total isoflavones were extracted with (metha-
nol:water, 90:10, v/v). Various extraction techniques have been used by differ-
There were significant differences (p < 0.001) in individ- ent research groups for extraction of isoflavones from soy-
ual isoflavones content extracted from a soybean sample beans (Table 1). Comparison of extraction of isoflavones
with different solvents mixtures (Table 3). Maximum yields with single optimized extraction solvent mixture (dimethyl
of the three glycosylated isoflavones (daidzin, glycitin and sulphoxide:acetonitrile:water, 5:58:37, v/v/v) with same
genistin) were obtained with methanol:water (90:10, v/v). solid-to-solvent ratio was carried out using six commonly
However, the optimum yields for all the other nine iso- used extraction techniques (PLE, sonication, Soxhlet,

5000
58 % ACN
58 % ACN, 5 % DMSO
70 % ETOH
4000
70 % ETOH, 5 % DMSO
90 % MEOH
Water
5 % genipol
3000

2000

1000

0
n

tin
n

n
in

in
n
in

tin

es
in

ti

ei
ti

zi
ti

ei
st

ite
z

dz

ci

is
ci

n
is
ci

st
id
id

z
i

vo
ly

en
ly
en

c
ly

id
en
ai

i
Da
Da

en
ly
G
G

lG

fla
Da
lD

G
G

G
yl

yl

so
yl
ny

ny

yl
et

et

-1000
t
on

I
ce
o

Ac

Ac

al
al

al

al

t
M

To
M

Fig. 3. Influence of extraction solvent composition of assay of 12 isoflavones from soybeans.

332 D.L. Luthria et al. / Food Chemistry 105 (2007) 325–333

Table 4
Isoflavone peak areas obtained by extraction with different methods such as PLE, sonication, Soxhlet, shaker, vortex, and stirring using 70% ETOH, 5%
DMSO from Organic soybeans
Extraction Areas
method
Daidzin Glycitin Genistin Malonyl Malonyl Acetyl Acetyl Malonyl Daidzein Glycitein Acetyl Genistein SUM
Daidzin Glycitin Daidzin Glycitin Genistin Genistin
PLE 662.3b 111.2b 920.5b 382.7c 235.5a 117.2a 69.2a 707.5c 1054.5a 76.9b,c 172.8a 83.9b 4594.3a
Sonication 426.7d 87.6c 648.8d 676.8b 213.5a 14.7c 6.96c 969.1b 1030.8b 95.5a 30.0c 83.7b 4284.1b
Soxhlet 804.1a 124.5a 1133.2a 302.1d 0.0c 61.1b 0.0d 460.9d 0.0c 73.1c 94.2b 85.5b 3138.7c
Shaker 478.4c 82.7c 688.6c 715.5a 0.0c 0.0d 35.9b 1076.5a 0.0c 76.1b,c 0.0d 85.1b 3238.7c
Vortex 451.2c,d 76.9d 655.1d 679.3b 0.0c 0.0d 32.0b 958.7b 0.0c 74.6b,c 0.0d 83.6b 3011.5d
Stirring 470.5c 74.1d 690.7c 728.6a 0.0c 0.0d 31.5b 1048.0a 0.0c 79.2b 0.0d 90.6a 3213.2c
LSD 34.8 5.3 33.1 21.8 5.9 5.0 6.9 64.3 3.0 5.6 6.0 4.2 121.8
The average areas of the peaks with any identical letter are not significantly different with each other at the a = 0.05 level by least significant difference test;
n = 3. PLE, accelerated solvent extraction LSD, least significant difference.

shaker, vortex, and stirring). Table 4 shows the result of & Gonzalez, 2001; Luthria, 2006; Naczk & Shahidi,
individual and total isoflavones content isolated by six dif- 2004) clearly illustrates that it is essential to systematically
ferent extraction procedures. The results indicate that opti- evaluate and optimize the extraction solvent composition,
mum yields of total isoflavones were extracted by PLE extraction procedure and conditions for accurate and
procedure. Only marginal decrease (6.7%) in yields of iso- reproducible estimation of structurally diverse phenolic
flavones content was obtained when extractions were per- compounds from different food matrices. In the present
formed with sonication. The extraction efficiency of the study, optimum isoflavones recoveries from soybean sam-
other four procedures (stirring, Soxhlet, shaker and vortex) ples were obtained with dimethyl sulphoxide:ethanol:water
was between 65.6% and 70.4% as compared to PLE. Rost- (5:70:25, v/v/v) solvent mixture using a PLE. Several thou-
agno, Araujo, and Sandi (2002) had compared the extrac- sands different phenolics with wide range of polarities,
tion efficiencies of soybean isoflavones by three extraction forms (aglycon, glycosylated, acetylated/malonylated,
techniques (sonication, Soxhlet, and supercritical fluid esterified to acids, etc.) have been reported in literature.
extraction). The authors reported that optimum extrac- In addition, phenolic compounds may be associated with
tions were obtained with sonication procedure. The extrac- different cellular components such as cell walls, proteins,
tion efficiency for Soxhlet and supercritical fluid extraction lipids etc. It is critical to optimize sample preparation pro-
were 68.3% and 27.7% respectively as compared to the son- cedures for accurate analysis of different classes of phenolic
ication procedure. Similar extraction recoveries for total compounds isolated from various food matrices.
isoflavones were obtained with Soxhlet procedure (73.2%)
as compared to sonication in the present study. However, Acknowledgements
in another study, Delmonte, Perry, and Rader (2006)
reported similar recoveries for total isoflavones by sonica- The authors would like to thank Mr. Andrew Anderson
tion and shaker techniques. These results are different form for his help in sample preparation and extraction. We
our present study as we observed that extraction efficiencies would like to thank Dr. Mustafa Oczan and Dr. Pei Chen
with total isoflavones by shaker procedure were only 75.6% of our laboratory for their help with LC–MS analysis. We
efficient as compared to sonication procedure. This varia- also like to thank Mr. Bruce Richter from Dionex Corpo-
tion in isoflavones extraction yields from soybeans by Del- ration for providing supplies for ASE extraction.
monte et al. (2006) may be attributed to differences in
extraction solvents, solid-to-solvent ratios, and tempera- References
ture used during extraction by shaker and sonication pro-
cedures. However in the present study, same extraction Achouri, A., Boye, J. I., & Belanger, D. (2005). Soybean isoflavones:
solvent mixture, same solid-to-solvent ratio and same Efficacy of extraction conditions and effect of food type on extract-
extraction conditions were used for sonication and shaker ability. Food Research International, 38, 1199–1204.
Anderson, J. J. B., & Garner, S. C. (1997). Phytoestrogens and human
procedures.
function. Nutrition Today, 32, 232–239.
Antolovich, M., Prenzler, P., Robards, K., & Ryan, D. (2000). Sample
4. Conclusion preparation in the determination of phenolic compounds in fruits.
Analyst, 125, 989–1009.
This study as well as other previously reported publica- Caragay, A. B. (1992). Cancer preventive foods and ingredients. Food
Technology, 46(4), 65–68.
tions (Luthria & Mukhopadhyay, 2006; Luthria, Mukho-
Delmonte, P., Perry, J., & Rader, J. I. (2006). Determination of
padhyay, & Kwansa, 2006; Mitra & Brukh, 2003; isoflavones in dietary supplements containing soy, Red Clover and
Mukhopadhyay, Luthria, & Robbins, 2006) and reviews kudzu: Extraction followed by basic or acid hydrolysis. Journal of
(Antolovich, Prenzler, Robards, & Ryan, 2000; Escarpa Chromatography A, 1107, 59–69.
 D.L. Luthria et al. / Food Chemistry 105 (2007) 325–333 333

Escarpa, A., & Gonzalez, M. C. (2001). An overview of analytical Luthria, D. L. (2006). Significance of sample preparation in developing
chemistry of phenolic compounds in foods. Critical Review in analytical methodologies for accurate estimation of bioactive com-
Analytical Chemistry, 31, 57–139. pounds in functional foods. Journal of Science of Food and Agriculture,
Griffith, A. P., & Collison, M. W. (2001). Improved methods for the 86(14), 2266–2272.
extraction and analysis of isoflavones from soy-containing foods and Luthria, D. L., & Mukhopadhyay, S. (2006). Influence of sample
nutritional supplements by reversed-phase high-performance liquid preparation on assay of phenolic acids from eggplant. Journal of
chromatography and liquid chromatography–mass spectrometry. Agricultural and Food Chemistry, 54, 41–47.
Journal of Chromatography A, 913, 397–413. Luthria, D. L., Mukhopadhyay, S., & Kwansa, A. L. (2006). A systematic
Hasler, C. M. (1998). Functional Foods: Their role in disease prevention approach for extraction of phenolic compounds using parsley (Petro-
and health promotion. Food Technology, 52(1111), 63–70. selinum crispum) flakes as a model substrate. Journal of Science of Food
He, J., Zhao, Z., Shi, Z., Zhao, M., Li, Y., & Chang, W. (2005). Analysis and Agriculture, 86(9), 1350–1358.
of isoflavone diadzein in Puerariae radix with micelle-mediated Messina, M. (1999). Soy, soy phytoestrogens (isoflavones), and breast
extraction and preconcentration. Journal of Agricultural and Food cancer. American Journal of Clinical Nutrition, 70, 574–575.
Chemistry, 53, 518–523. Mitra, S., & Brukh, R. (2003). Sample preparation: an analytical
Kao, F.-J., Su, N.-W., & Lee, M.-H. (2004). Effect of water-to-bean ratio perspective. In S. Mitra (Ed.), Sample preparation techniques in
on the contents and compositions of isoflavones in Tofu. Journal of analytical chemistry (pp. 1–36). New York, NY: John Wiley and
Agricultural and Food Chemistry, 52, 2277–2281. Sons, Inc.
Kennedy, A. R. (1998). The Bowman–Birk inhibitor from soybeans as an Mukhopadhyay, S., Luthria, D. L., & Robbins, R. J. (2006). Optimization
anticarcinogenic agent. American Journal of Clinical Nutrition, 68, of extraction process for phenolic acids from Black cohosh (Cimicifuga
1406S–1412S. racemosa) by pressurized liquid extraction. Journal of Science of Food
Klejdus, B., Mikelova, R., Adam, V., Zehnalek, J., Vacek, J., Kizek, R., and Agriculture, 86, 156–162.
et al. (2004). Liquid chromatographic–mass spectrophotometric deter- Murphy, P. A., Barua, K., & Hauck, C. C. (2002). Solvent extraction
mination of genistin and daidzin in soybean food samples after selection in the determination of isoflavones in soy foods. Journal of
accelerated solvent extraction with modified content of extraction cell. Chromatography B, 777, 129–138.
Analytica Chimica Acta, 517, 1–11. Murphy, P. A., Song, T., Buseman, G., & Barua, K. (1997). Isoflavones in
Klejdus, B., Mikelova, R., Petrolova, J., Potesil, D., Adam, V., Stiborova, soy-based infant formulas. Journal of Agricultural and Food Chemistry,
M., et al. (2005). Evaluation of isoflavone aglycon and glycoside 45, 4635–4638.
distribution in soy plants and soybeans by fast column high-perfor- Naczk, M., & Shahidi, F. (2004). Extraction and analysis of phenolics in
mance liquid chromatography coupled with a diode-array detector. food. Journal of Chromatography A, 1054, 95–111.
Journal of Agricultural and Food Chemistry, 53, 5848–5852. Penalvo, J. L., Nurmi, T., & Adlercreutz, H. (2004). A simplified HPLC
Koratkar, R., & Rao, A. V. (1997). Effect of soya bean saponins on method for total isoflavones in soy products. Food Chemistry, 87,
azoxymethane-induced preneoplastic lesions in the colon of mice. 297–305.
Nutrition and Cancer, 27, 206–209. Rostagno, M. A., Araujo, J. M. A., & Sandi, D. (2002). Supercritical fluid
Lee, J. H., Renita, M., Fioritto, R. J., Martin, S. K., Schwartz, S. J., & extraction of isoflavones from soybean flour. Food Chemistry, 78,
Vodovotz, Y. (2004). Isoflavone characterization and antioxidant 111–117.
activity of Ohio soybeans. Journal of Agricultural and Food Chemistry, Rostagno, M. A., Palma, M., & Barroso, C. G. (2004). Pressurized liquid
52, 2647–2651. extraction of isoflavones from soybeans. Analytica Chimica Acta, 522,
Lee, K. W., Wang, H. J., Murphy, P. A., & Hendrich, S. (1995). Soybean 169–177.
isoflavone extract suppresses early but not later promotion of Slavin, J., Jacobs, D., & Marquart, L. (1997). Whole grain consumption
hepatocarcinogenesis by Phenobarbital in female rat liver. Nutrition and chronic disease: Protective mechanisms. Nutrition and Cancer, 27,
and Cancer, 24, 267–278. 14–21.
Li-Hsun, C., Ya-Chuan, C., & Chieh-Ming, C. (2004). Extracting Vesper, H., Schmelz, E. M., Nikolova-Karakashian, M. N., Dillehay, D.
and purifying isoflavones from defatted soybean flakes using L., Lynch, D. V., & Merrill, A. H. Jr., (1999). Sphingolipids in food
superheated water at elevated pressures. Food Chemistry, 84, 279– and the emerging importance of spingholipidsto nutrition. Journal of
285. Nutrition, 129, 1239–1250.
Lin, F., & Giusti, M. M. (2005). Effects of solvent polarity and acidity on Wu, Q., Wang, M., Sciarappa, W. J., & Simon, J. E. (2004). LC/UV/ESI–
the extraction efficiency of isoflavones from soybeans (Glycine max). MS Analysis of isoflavones in Edamame and Tofu Soybeans. Journal
Journal of Agricultural and Food Chemistry, 53, 3795–3800. of Agricultural and Food Chemistry, 52, 2763–2769.