Quantification of the group B soyasaponins by high-performance liquid chromatography

High-performance liquid chromatographic methods were developed for the isolation and quantitative determination of the group B soyasaponins, including 2,3-dihydro-2,5-dihydroxy-6-methyl-4H-pyran-4-one (DDMP)-conjugated soyasaponins alphag, betag, and betaa, and their non-DDMP counterparts, soyasaponins V, I, and II, respectively, with formononetin used as the internal standard. The limits of quantification for soy products were 0.11-4.86 micromol/g. The within-day and between-days assay coefficients of variation were <9.8 and < 14.3%, respectively. The group B soyasaponin concentrations in 46 soybean varieties ranged from 2.50 to 5.85 micromol/g. Soy ingredients (soybean flour, toasted soy hypocotyls, soy protein isolates, textured vegetable protein, soy protein concentrates, and Novasoy) and soy foods (commercial soy milk, tofu, and tempeh) contained the group B soyasaponins from 0.20 to 114.02 micromol/g. There was no apparent correlation between isoflavone and soyasaponin concentrations in the soy products examined.     

Evaluation of soyasaponin, isoflavone, protein, lipid, and free sugar accumulation in developing soybean seeds

A combination of analytical techniques was used to examine and quantify seed compositional components such as protein, lipid, free sugars, isoflavones, and soyasaponins during soybean development and maturation in two Korean soybean cultivars. Protein accumulation was rapid during reproductive stages, while lipid content was only relatively moderately increased. The major carbohydrate saccarides sucrose and stachyose constantly increased during the reproductive stage. Previously published results suggest that the free sugar and lipid content reached their maximal concentrations at a relatively early stage of seed development and remain constant in comparison to other chemical components. The malonylglucosides were the predominant isoflavone form followed by the glucosides, acetyl glucosides, and aglycone forms. As soybean seed matures, total soyasaponin concentration was constantly decreased until the R8 stage. Soyasaponin beta(g) was the major soyasaponin in DDMP-conjugated group B soyasaponins, followed by the non-DDMP counterpart soyasaponin I and soyasaponin A1. The ratio of total isoflavone to total soyasaponin in the developing soybean increased from 0.06 to 1.31. Protein, lipid, and free sugar contents in the developing soybean seeds showed significant positive correlations with conjugated isoflavones and total isoflavone concentration, while the lipid contents showed a negative correlation with the isoflavone aglycone. Protein, lipid, and free sugar contents showed a negative correlation with total group A and B soyasaponins and total soyasaponins; however, only the soyasaponin A content was significantly negatively correlated with free sugar content. Total soyasaponin content was negatively correlated with isoflavone content (r = -0.828 at p < 0.01).     

Complete quantification of group A and group B soyasaponins in soybeans

A combination of high-pressure extraction and preparative chromatography was used to purify the group A and group B soyasaponins from soy germ for use as analytical standards and for use in biological assays. A standardized sample preparation and extraction method was developed for the analysis of phytochemicals found in soy and processed soy products, which is reproducible in other laboratories. The extracts can be analyzed with standard liquid chromatography-mass spectrometry and high-performance liquid chromatography methods to identify and quantitate the group A and group B forms of the soy saponins, as well as the soy isoflavones. Complete saponin analysis of the extracts prepared from soy germ (hypocots), hulls, and cotyledons shows that a significant portion of the saponins is concentrated in the germ. The germ contains nearly all of the group A soyasaponins, while the group B soyasaponins are nearly equally distributed between the germ and the cotyledons. The hulls contain little of either isoflavones or saponins. Whole (full fat) soybeans grown on a tract in central Illinois in 2003 contain approximately 4-6% saponins on a weight basis, of which about one-fifth or less of the total saponin content are group A soyasaponins; the balance is group B soyasaponins.     

Use of a simplified HPLC-UV analysis for soyasaponin B determination: study of saponin and isoflavone variability in soybean cultivars and soy-based health food products

Soyasaponins are phytochemicals of major interest for health. Their identification and quantification remain difficult owing to the large number of structural isomers in soybeans and the lack of stable standards. In this study, a rapid method using high performance liquid chromatography (HPLC) using a UV detector (205 nm) was developed to identify and quantify soyasaponins belonging to group B and compare them with isoflavones in different soy materials. 2,3-Dihydro-2,5-dihydroxy-6-methyl-4H-pyran-4-one (DDMP)-conjugated soyasaponins were determined using external calibration or a molecular mass ratio after alkaline hydrolysis to cleave their DDMP moieties. The detection limit of soyasaponin I, used as a reference molecule to simplify the analysis, was 0.065 micromol/g. Soyasaponin contents in seven soybean varieties ranged from 13.20 to 42.40 micromol/g in the germ and from 2.76 to 6.43 micromol/g in the cotyledons. The within-day and between-days variation coefficients did not exceed 7.9 and 9.0%, respectively, for the major soyasaponins. Soyasaponin B quantification in different soy-based health supplements was reported along with measurements of their isoflavone content to provide information on the variability of these bioactive compounds among different types of soy food materials.     

Determination of soyasaponins in soy with LC-MS following structural unification by partial alkaline degradation

High-performance liquid chromatography coupled with electrospray ionization mass spectrometer was used to study the soyasaponins in soy. It was found that each soyasaponin belonging to group A existed mainly in their genuine acetylated forms. The partially to fully deacetylated structures coexisted in various proportions. Likewise, the soyasaponins belonging to group B in soy were detected as both 2,3-dihydro-2,5-dihydroxy-6-methyl-4-pyrone (DDMP) conjugated forms and non-DDMP forms. The structural diversity of soyasaponins hinders the separation and determination of the individual compounds in soy. In the present studies, the soyasaponins extracted from soy were treated with sodium hydroxide under mild conditions to cleave the acetyl groups from soyasaponins in group A as well as the DDMP from soyasaponins in group B, while the glycoside structures remained unaffected. By doing so, all soyasaponins originating from the same initial structures were unified into well-defined structures and then quantified individually using the selective ion recording of their [M-H](-) ions. The pure deacetyl and non-DDMP soyasaponins were used as the external standards. The quantification limits of soyasaponins in group A and group B were 1.74 and 1.89 ng injected on column with recovery rates of 94.1% +/- 4.2% and 96.9% +/- 2.9%, respectively.     

Effect of soyasapogenol A and soyasapogenol B concentrated extracts on HEP-G2 cell proliferation and apoptosis

The growth inhibition and the induction of apoptosis brought about by soyasaponins extracted from soy flour ( Glycine max (L.)) and concentrated for soyasapogenols A and B formed by hydrolysis were tested for cytoactivity in the human hepatocellular carcinoma cell line Hep-G2. Concentrated soyasapogenol A (SG-A) and soyasapogenol B (SG-B) extracts contained approximately 69.3% and 46.2% of their respective aglycones (soyasapogenols) assessed by HPLC and ESI-MS, while the soyasaponin extract (TS), derived from crude methanol extraction, did not contain any detectable amounts of SG-A or SG-B. An MTT viability assay showed that all three extracts had an effect on Hep-G2 proliferation in a dose-response manner with 72 h LC50 values of 0.594+/-0.021 mg/mL for TS, 0.052+/-0.011 mg/mL for SG-A, and 0.128+/-0.005 mg/mL for SG-B. Apoptotic cells were determined by flow cytometry cell cycle analysis and confocal laser scanning microscopy (CLSM). Cell cycle analysis indicated a significant ( P< 0.05) greater sub-G1 buildup of apoptotic cells at 24 h (25.63+/-2.1%) and 72 h (47.1+/-3.5%) for the SG-A extract compared to SG-B, whereas the TS extract produced only a minor buildup of sub-G1 cells. CLSM confirmed a morphological change of all treatments after 24 h, at the respective LC50 concentrations. These results show that the samples that contained mainly soyasapogenols A and B showed a greater ability to inhibit proliferation of cultured Hep-G2 when compared to a total soyasaponin extract that did not contain any soyasapogenols.     

Group B saponins in soy products in the U.S. Department of Agriculture--Iowa State University isoflavone database and their comparison with isoflavone contents

Isoflavones in soy protein foods are thought to contribute to the cholesterol-lowering effect observed when these products are fed to humans. The group B saponins are another ethanol-soluble phytochemical fraction associated with soy proteins and isoflavones and have also been associated with cholesterol-lowering abilities. We measured the group B soyasaponin concentrations in a variety of soy foods and ingredients in the U.S. Department of AgricultureIowa State University Isoflavone Database. We compared the isoflavone and soy saponin concentrations and distributions in intact soybeans, soy ingredients, and retail soy foods. Group B saponins occur in six predominant forms. There appears to be no correlation between saponin and isoflavone concentrations in intact soybeans ranging from 5 to 11 mumol isoflavones/g soybean and from 2 to 6 mumol saponin/g soybean. Depending upon the type of processing, soy ingredients have quite different saponins/isoflavones as compared to mature soybeans. In soy foods, the saponin:isoflavone ration ranges from 1:1 to 2:5, whereas in soy protein isolates, the ratio is approximately 5:3. Ethanol-washed ingredients have very low saponins and isoflavones. These very different distributions of saponins and isoflavones in soy products may affect how we view the outcome of feeding trials examining a variety of protective effects associated with soy consumption.     

Human fecal metabolism of soyasaponin I

The metabolism of soyasaponin I (3-O-[alpha-L-rhamnopyranosyl-beta-D-galactopyranosyl-beta-D-glucuronopyranosyl]olean-12-ene-3beta,22beta,24-triol) by human fecal microorganisms was investigated. Fresh feces were collected from 15 healthy women and incubated anaerobically with 10 mmol soyasaponin I/g feces at 37 degrees C for 48 h. The disappearance of soyasaponin I in this in vitro fermentation system displayed apparent first-order rate loss kinetics. Two distinct soyasaponin I degradation phenotypes were observed among the subjects: rapid soyasaponin degraders with a rate constant k = 0.24 +/- 0.04 h(-)(1) and slow degraders with a k = 0.07 +/- 0.02 h(-)(1). There were no significant differences in the body mass index, fecal moisture, gut transit time, and soy consumption frequency between the two soyasaponin degradation phenotypes. Two primary gut microbial metabolites of soyasaponin I were identified as soyasaponin III (3-O-[beta-D-galactopyranosyl-beta-D-glucuronopyranosyl]olean-12-ene-3beta,22beta,24-triol) and soyasapogenol B (olean-12-ene-3beta,22beta,24-triol) by NMR and electrospray ionized mass spectroscopy. Soyasaponin III appeared within the first 24 h and disappeared by 48 h. Soyasapogenol B seemed to be the final metabolic product during the 48 h anaerobic incubation. These results indicate that dietary soyasaponins can be metabolized by human gut microorganisms. The sugar moieties of soyasaponins seem to be hydrolyzed sequentially to yield smaller and more hydrophobic metabolites.     

MALDI-TOF MS analysis of isoflavones in soy products

Matrix-assisted laser desorption/ionization time-of-flight mass spectrometry (MALDI-TOF MS) is a new technique having a number of advantages for food analysis. This study is the first to demonstrate the use of MALDI-TOF MS to identify isoflavones in soy samples. 2',4',6'-Trihydroxyacetophenone (THAP) and 2,5-dihydroxybenzoic acid (DHB) were both good matrices for isoflavones, but DHB was chosen as the best because it worked well for sample extracts, with good spot-to-spot repeatability. Isoflavones were predominantly ionized in a protonated form with a very small amount of sodium or potassium adduct ions. Fragmentation occurred only through loss of glycosidic residues. Daidzin showed more than twice the response of genistin using MALDI-TOF MS. A simple solid phase extraction of isoflavones from soy samples was developed for MALDI-TOF MS analysis. MALDI-TOF MS can provide an isoflavone profile in 2 min and serves as a powerful tool to identify and study processing changes of isoflavones in soy products.     

Insecticidal components from field pea extracts: soyasaponins and lysolecithins

Extracts from field peas (Pisum sativum L.) have previously been shown to have a utility to control insect pests. To identify potentially new bioinsecticides in field crops, we describe the fractionation of impure extracts (C8 extracts) derived from protein-rich fractions of commercial pea flour. The activity of separated fractions was determined by a flour disk antifeedant bioassay with the rice weevil [Sitophilus oryzae (L.)], an insect pest of stored products. Bioassay-guided fractionation showed that the triterpenoid saponin fractions were partly responsible for the antifeedant activity of C8 extracts. Soyasaponin I (soyasaponin Bb), isolated from peas and soybeans, and mixtures of soyasaponins, comprised of soyasaponins I-III and isolated from soybeans, were inactive antifeedants, but dehydrosoyasaponin I (the C-22 ketone derivative of soyasaponin I), a minor component found in C8 extracts, was shown to be an active component. Dehydrosoyasaponin I (soyasaponin Be) and soyasaponin VI (soyasaponin betag) coeluted under conditions of silica gel thin-layer chromatography and C18 high-performance liquid chromatography. However, dehydrosoyasaponin I could be isolated from saponin-enriched fractions with a reversed phase column of styrene/divinylbenzene operated at alkaline pH. Phospholipids of the lysolecithin type were also identified in saponin fractions of C8 extracts from peas. Three of the lysolecithins were inactive alone against rice weevils, but mixtures of these phospholipids enhanced the insecticidal activity of dehydrosoyasaponin I.     

Characterization of isoflavones and their conjugates in female rat urine using LC/MS/MS

Isoflavone phytoestrogens found in soybeans are the most widely studied phytochemicals in human diets and soy infant formulas. The health benefits of the isoflavones daidzein and genistein have been reported, and concerns about potential adverse effects have also been raised. However, the results of direct analysis of isoflavones and their metabolites in biological fluids after consumption of soy-containing diets are scarce. This study describes an LC/MS/MS method for the analysis of isoflavones and their metabolites in the urine of female rats fed diets made with soy protein isolate. Five isoflavones (daidzein, genistein, glycitein, dihydrodaidzein, and O-desmethylangolensin) were identified by comparison with authentic standards. Seventeen conjugates of isoflavones were characterized in the urine, the most unusual being genistein 5-glucuronide and four glucuronide conjugates of reductive metabolites of daidzein. The application of LC/MS/MS to analyze isoflavone metabolites is simple and sensitive, and appears to be an excellent method for determining the bioavailability and metabolism of food phytochemistry.     

Determination of saponins in legumes by direct densitometry.

Research has shown that dietary saponins may have health benefits. A simple, rapid method for the determination of saponins in legumes, using densitometry, is described. Saponin preparations, after pretreatment to remove nonsaponin components, are spotted in rows on a thin-layer chromatography plate, along with soyasaponin standards. The plate, without solvent development, is directly treated with sulfuric acid and heated. Violet spots develop which have a density proportional to the amount of saponin present. The standard curve has a correlation coefficient of 0.99 and is linear over the range of 1.25 to 10 microg of soyasaponins applied. The method has a coefficient of variation of less than 3% and compares favorably with quantitative thin-layer chromatography. Using this method the saponin contents of defatted soy flour (0.58%), dried navy beans (0.32%), and dried kidney beans (0.29%) were determined, and these results were found to be consistent with previous reports in the literature.     

LC/UV/ESI-MS analysis of isoflavones in Edamame and Tofu soybeans

High-performance liquid chromatography coupled with ultraviolet and electrospray ionization mass spectrometry (HPLC/UV/ESI-MSD) was applied to the study of isoflavones in both Edamame and Tofu soy varieties, from which the immature fresh soybeans or the mature soybean seeds are consumed, respectively. Positive atmospheric pressure interface (API) MS and MS/MS were used to provide molecular mass information and led to the identification of a total 16 isoflavones, including three aglycones, three glycosides, two glycoside acetates, and eight glycoside malonates. The major isoflavones in soybean seeds were daidzein and genistein glycoside and their malonate conjugates. Trace levels of daidzein and genistein acetyl glycosides were found only in the mature dry soybean seeds. To facilitate quantitative analysis, acid hydrolysis during extraction of soy samples was selected to convert the various phytoestrogen conjugates into their respective isoflavone aglycones, allowing accurate quantitation of total phytoestrogens as aglycones. On the basis of HPLC combined with UV and MS detection, all three targeted soy isoflavone aglycones, daidzein, genistein and glycitein in hydrolyzed extracts were successfully quantified within 25 min with formononetin used as the internal standard. The standard curves of UV detection were fitted in the range of 14.16-29000 ng/mL for daidzein, 15.38-31500 ng/mL for genistein, and 11.72-24000 ng/mL for glycitein. For MS detection, the standard curves were established in the range of 3.54-1812.5 ng/mL for daidzein, 3.85-1968.75 ng/mL for genistein, and 2.93-1500 ng/mL for glycitein. Good linearities (r(2) > 0.999 for UV and r(2) > 0.99 for MS) for standard curves were achieved for each isoflavone. The accuracy and precision (RSD) were within 10% for UV detection and 15% for MS detection (n = 10). Using this method, the phytoestrogen levels of total isoflavone aglycones from 30 soybean seed varieties were then evaluated for confirmation of the technique. Total isoflavones ranged across the varieties from 0.02 to 0.12% in the Edamame varieties, which are harvested while the seeds are still immature, and from 0.16 to 0.25% in Tofu varieties, harvested when the seeds are physiologically mature. While the literature has focused on the isoflavone content of soy products and processing soy, this report provides a reliable analytical technique for screening of authenticated fresh immature Edamame soybeans and Tofu soybeans.     

Soyasaponins resist extrusion cooking and are not degraded during gut passage in Atlantic salmon (Salmo salar L.)

The stability of soyasaponins in fish feed formulations was investigated. The level of soyasaponin Ab, Bb, Bc, Ba-2,3-dihydro-2,5-dihydroxy-6-methyl-4H-pyran-4-one (Ba-DDMP), Bb-DDMP, and Bc-DDMP was quantified in 15 samples of defatted soybean meal, two full fat soybean meals, and two soybean protein concentrates by reverse phase high-performance liquid chromatography. The total level of saponins in the 15 samples of commercial defatted soybean meal ranged from 4.8-6.8 micromol/g (5.1-7.0 g/kg). The two full fat meals contained 4.4 and 4.7 micromol/g whereas no saponins could be detected in the alcohol-extracted soybean protein concentrates. Fifteen batches of fish feed containing 20% defatted soybean meal were produced by twin-screw extrusion from the 15 different samples of defatted soybean meal. Extrusion did not reduce the total level of group B saponins, but the ratio between DDMP-conjugated group B saponins and non-DDMP-conjugated group B saponins was slightly reduced. A soybean-containing diet was fed to seawater adapted Atlantic salmon for 9 weeks. Yttrium oxide was included in the feed as an inert marker in order to estimate the disappearance of saponins during gut passage. High levels of intact non-DDMP-conjugated group B soyasaponins were found in feces whereas only low levels of DDMP-conjugated saponins could be detected. The overall disappearance of saponins was close to zero, and the concentration of intact saponins in dry feces reached levels several fold higher than dietary levels. The present work demonstrates that non-DDMP-conjugated group B soyasaponins resist extrusion cooking and remain intact during gut passage in Atlantic salmon. The latter is contrary to earlier findings in endothermic animals.     

LC-MS/MS analysis of lysophospholipids associated with soy protein isolate

Approximately 20-25% of babies consuming formula are fed formulas containing soy protein as the sole protein source. Isoflavone phytoestrogens are considered as major bioactive components present in soy protein isolate (SPI) and the most widely studied phytochemicals in human diets and soy infant formulas. In the present study, LC-MS/MS analysis of SPI phytochemical extract resulted in identifications of fifty-six lysophospholipids (lyso-PL), including eighteen lysophosphatidylcholines (lyso-PC), twelve lysophosphatidylethanolamines (lyso-PE), eleven lysophosphatidylinositols (lyso-PI), eleven lysophosphatidic acids (lyso-PA) and four lysophosphatidylglycerols (lyso-PG). The LC-MS/MS conditions were first developed using commercially available standard mixtures. Under these conditions, lyso-PL compounds could be separated by LC and yielded unique fragments and neutral losses in the electrospray ionization mass spectrometry (ESI-MS) with both positive- and negative-ion modes, quite indicative of the polar headgroups for different subclasses. This is the first characterization of lyso-PL in SPI. Lyso-PLs are important cell signaling and growth factor molecules experimentally linked to cardiovascular function and disease, cancer, and neurological disorders. These findings suggest that biological effects of PLs associated with SPI in infants clearly needs to be considered as part of an overall evaluation of potential health benefits.     

Fractionation of lentil seeds (Lens culinaris Medik.) for insecticidal and flavonol tetraglycoside components

Crude methanol extracts from four cultivated varieties of mature lentil seeds (Lens culinaris Medik.) were found to possess antifeedant and insecticidal properties in laboratory tests with the rice weevil (Sitophilus oryzae L.), an insect pest of stored products. Flash chromatography with silica gel on active Diaion HP-20 methanol extracts gave flavonol, lysolecithin, soyasaponin, and peptide fractions, as determined by HPLC and electrospray ionization LC/MS. The flavonol fraction was shown by high-resolution NMR experiments to contain a mixture of kaempferol 3-O-beta-glucopyranosyl(1-->2)-O-[alpha-rhamnopyranosyl(1-->6)]-beta-galactopyranoside-7-O-alpha-rhamnopyranoside and, tentatively, kaempferol 3-O-beta-glucopyranosyl(1-->2)-O-[alpha-rhamnopyranosyl(1-->6)]-beta-glucopyranoside-7-O-alpha-rhamnopyranoside. These inactive tetraglycosides, although inseparable under the reported HPLC conditions, were detected by NMR spectroscopy in nearly equal proportions. Three lysolecithins were identical to those previously identified in pea extracts. Soyasaponin I (soyasaponin Bb) and soyasaponin VI (soyasaponin betag) were found in Diaion HP-20 methanol extracts. An insecticidal lentil peptide with a mass of 3881 Da, isolated from an Eston variety in small quantities by anion exchange chromatography, was related to the cysteine-rich pea albumin 1b class of botanical insecticides. Binary mixtures of the insecticidal lentil peptide and soybean soyasaponin I were synergistic in tests with S. oryzae.     

Changes of astringent sensation of soy milk during tofu curd formation

The effect of isoflavone on soy milk and tofu astringency was investigated, and no consistency was found between an undesirable astringent taste and isoflavone contents. Isoflavone-enriched extract (approximately 39% isoflavones) showed no astringency. Soybean foods having high amounts of isoflavones showed less astringency. About 80% of isoflavones exist freely in both soy milk and tofu, but 55% of phytates (which play an important role in the formation of the tofu curd network) exist freely in the soy milk, and 6-13%, on the basis of coagulation, existed freely in the tofu curds. A 1% potassium phytate solution at pH 7 showed the very same astringency as soy milk; however, calcium phytate at the same concentration and pH showed no undesirable sensation. Thus, it is assumed that the astringent characteristics caused by phytic ions in soy milk are lost upon conversion of phytic ions to their insoluble salt forms during soy milk coagulation.     

Estrogenic activity of glycitein, a soy isoflavone

Glycitein (4',7-dihydroxy-6-methoxyisoflavone) accounts for 5-10% of the total isoflavones in soy food products. The biological activity of this compound has not been reported to date, although numerous studies have been performed with the other soy isoflavones, daidzein and genistein. Glycitein was isolated from soy germ to 99% purity. Weaning female B6D2F1 mice were dosed with glycitein (3 mg/day), genistein (3 mg/day), and diethylstilbestrol (DES) (0.03 microg/day) in 5% Tween 80 by gavage for 4 days. A control group received an equal volume of 5% Tween 80 solution daily. The uterine weight increased 150% with glycitein (p < 0.001), 50% with genistein (p < 0. 001), and 60% with DES (p < 0.001) compared with the control group. DES, 17beta-estradiol, and three isoflavones (daidzein, genistein, and glycitein) were examined for their competitive binding abilities with 17beta-((3)H)estradiol to the estrogen receptor proteins of the B6D2F1 mouse uterine cytosol. The concentrations of each compound required to displace 50% of the ((3)H)estradiol at 5 nM in the competitive binding assay were 1.15 nM DES, 1.09 nM 17beta-estradiol, 0.22 microM genistein, 4.00 microM daidzein, and 3.94 microM glycitein. These data indicated that glycitein has weak estrogenic activity, comparable to that of the other soy isoflavones but much lower than that of DES and 17beta-estradiol.     

Isoflavones in soy-based foods consumed in Brazil: levels,distribution, and estimated intake

The isoflavone content and profile in processed soy-based products consumed in Brazil were determined by high-performance liquid chromatography and photodiode array detection of the intact isoflavones (naturally occurring aglycons, malonyl, acetyl, and beta-glycosides derivatives). Total isoflavone content varied significantly among products, from 2 to 100 mg/100 g (wet basis, expressed as aglycons), with the lowest content being found for soy-based enteral/oral diets and the highest found for textured soy proteins. For soy beverages isoflavone content varied from 12 to 83 mg/L. Soy sauce, miso, and tofu had isoflavone contents of 5.7 mg/L, 20 mg/100 g, and 7 mg/100 g, respectively. The beta-glycosides were the predominant form of the isoflavones in the products analyzed, except for miso, shoyu, and "Diet Shake" in which the aglycons were present in the highest proportions. On the basis of these data, the daily intake of isoflavone from soy products was estimated: the highest values were found for infants fed soy-based formulas, from 1.6 to 6.6 mg/kg of body weight.