Variations in metabolism of the soy isoflavonoid daidzein by human intestinal microfloras from different individuals
Isoflavonoids found in legumes, such as soybeans, are converted by intestinal bacteria to metabolites that might have increased or decreased estrogenic activity. Variation in the effects of dietary isoflavonoids among individuals has been attributed to differences in their metabolism by intestinal bacteria. To investigate this variation, the metabolism of the isoflavonoid daidzein by bacteria from ten fecal samples, provided at different times by six individuals on soy-containing diets, was compared. After anaerobic incubation of bacteria with daidzein for 2 weeks, four samples had metabolized daidzein and six samples had not. Three of the positive samples were from individuals whose microflora had not metabolized daidzein in previous samples. Dihydrodaidzein was observed in one sample, dihydrodaidzein and equol in another sample, and equol and O-desmethylangolensin in two other samples. These results corroborate the hypothesis that the microflora of the gastrointestinal tract of an individual influences the particular isoflavone metabolites produced following consumption.
KeywordsIsoflavonoids Daidzein Equol Intestinal bacteria Soybean
Isoflavonoids are diphenols found in leguminous plants and have natural roles in plant defense and root nodulation (Peters et al. 1986; Phillips 1992). There is growing interest in the consumption of soy products because of the beneficial effects of isoflavonoids (Adlercreutz et al. 1993; Coward et al. 1993; Messina et al. 1994; Adlercreutz 1995; Anderson et al. 1998; Messina 2000) in chemoprevention and therapy of hormone-dependent diseases.
After ingestion, isoflavonoids may be converted to metabolites by the human intestinal microflora. This conversion is essential for the absorption, bioavailability, and estrogenic activities of these compounds (Xu et al. 1995; Lampe et al. 2001; Setchell et al. 2002). In vivo studies have shown variations in health benefits of phytoestrogens among individuals, which has been attributed to dissimilarities in the populations of colonic bacteria responsible for isoflavonoid conversion (Setchell et al. 2002). However, the variation of isoflavone metabolism in vitro by the colonic bacteria of different individuals has not been shown.
Daidzein (4′,7-dihydroxyisoflavone) is one of the principal isoflavonoids and is found in soybeans as the glucoside daidzin (Coward et al. 1993; Adlercreutz 1995). It binds to estrogen receptors, induces cell transcription and elicits biological activities similar to those of other natural estrogens (Setchell et al. 2002). In vivo studies have shown that, after consumption of soy products rich in daidzein, the intestinal bacteria of some individuals produce either the highly estrogenic compound equol, a reduction product of daidzein, or a nonestrogenic ring-cleavage product, O-desmethylangolensin (O-DMA) (Joannou et al. 1995; Kelly et al. 1995). In addition, dihydrodaidzein, tetrahydrodaidzein, 2-dehydro-O-desmethylangolensin, O-DMA and equol have been found in urine after consumption of soy products and are considered to be metabolites of daidzein (Joannou et al. 1995; Kelly et al. 1995).
Dihydrodaidzein, 3-(4-hydroxyphenyl)-benzopyran-4,7-diol, O-DMA, and equol have been detected after incubation of daidzein with fecal bacteria under anoxic conditions (Chang and Nair 1995; Hur et al. 2000, 2002; Schoefer et al. 2002).
Although the effectiveness or failure of dietary phytoestrogens for prevention of hormone-dependent diseases has been attributed to variations in the metabolism of these compounds by colonic bacteria of different individuals, there is no study demonstrating this variation in vitro. In this study, we compared the metabolism of daidzein by intestinal microflora of several individuals consuming soy diets and identified the metabolites by HPLC and mass spectrometry.
Materials and methods
Metabolism of daidzein
Metabolism of daidzein by bacteria from fecal samples received either from different individuals or from the same individual at different times. ND Not done
Age and gender
1B/equol and dihydrodaidzein
2C/equol and O-DMA
5A/equol and O-DMA
After the samples were received (0–2 h after collection), they were transferred to an anaerobic glove box. Each sample was used to inoculate two bottles containing 100 ml of brain heart infusion (BHI) medium under a CO2/ H2/ N2 atmosphere (5/10/85, v/v/v ). To one bottle of medium from each set 5 μg daidzein/ml was added; the cultures were incubated under CO2/ H2/ N2 at 37 °C. As a control, sterile BHI medium was incubated with 5 μg daidzein/ml. Every day for 14 days, 2.5 ml from each of the cultures and controls was removed aseptically and extracted with ethyl acetate, which was dried over anhydrous Na2SO4 and evaporated. The residues were dissolved in 90% acetonitrile:10% water for HPLC analysis.
Synthesis of daidzein metabolites for standards
The Star HPLC system from Varian (Palo Alto, Calif., USA) consisted of a model 230 pump, a model 430 autosampler with a 100-μl loop and a model 330 photo diode array spectrophotometer. A Spherisorb C18 reversed-phase column (4.6×250 mm, S5, ODS 2 Phase Sep, Clwyd, Wales, UK) was used. The mobile phase components were 10% acetonitrile, 0.1% acetic acid, and 90% water (A), and 90% acetonitrile, 0.1% acetic acid and 10% water (B). After sample injection, the column was washed with 100% A for 10 min followed by elution with a linear gradient of 10% B to 90% B for 50 min. The detector was monitored at 280 nm and UV spectra of the peaks were scanned from 220 to 450 nm. The flow rate was 1 ml/min.
Liquid chromatography mass spectrometry
Analyses were carried out on a Hewlett-Packard 5989B (Palo Alto, Calif., USA) mass spectrometer with a Hewlett-Packard 1090L/M HPLC and Prodigy ODS (3) 2.0×250 mm 5 μm 100A HPLC column (Phenomenex, Torrance, Calif., USA). The mass spectrometer was operated in the negative electrospray ionization (ESI) mode. With the capillary exit voltage variable, full scans were acquired from m/z 200 to 300. With the capillary exit voltage at −200 V for in-source collision-induced dissociation (CID), full scans were acquired from m/z 50 to 300. The mobile phase, delivered at 0.2 ml/min, was a linear gradient from 20% aqueous acetonitrile to 80% aqueous acetonitrile in 40 min with constant 3 mM ammonium formate.
Nuclear magnetic resonance spectroscopy
1H nuclear magnetic resonance (NMR) spectra were determined on a Bruker 500 MHz NMR spectrometer operated at 301 K. The peaks were collected from HPLC injections, dried in vacuo and dissolved in deuterated methanol. The chemical shifts were defined by assigning the deuterated methanol resonance peak to 3.31 ppm. The spectral width was 7,500 Hz with a 1.0-s delay time. For every proton resonance, nuclear Overhauser effect and homonuclear decoupling experiments were carried out.
Daidzein and equol were from Sigma (St Louis, Mo., USA) and Indofine (Somerville, N.J., USA), respectively. HPLC-grade acetonitrile and methanol were from J.T. Baker (Phillipsburg, N.J., USA).
All of the cultures grew equally well with or without 5 μg daidzein/ml under anoxic conditions. The elution profiles of each of the cultures incubated with and without daidzein were compared and the spectra of the individual peaks were determined.
Chromatograms of sample 1B incubated with daidzein showed two peaks that were different from the same sample without daidzein (Fig.1B). One peak had a retention time and a UV/visible spectrum similar to those of dihydrodaidzein (Hur et al. 2000), and the other peak, which eluted later from the column, had a retention time and a UV spectrum similar to those of equol (Fig. 1B). No daidzein metabolites were detected from sample 1A, which had been obtained from this individual 9 months earlier, either at the time of sample collection or after storage at −70 °C.
In sample 3B, one peak was observed after incubation with daidzein that was absent from the chromatograms of sample 3B incubated without daidzein and from sample 3A, which had been obtained from the same individual 9 months earlier and incubated with daidzein at the time of sample collection. This peak had a UV spectrum similar to that of dihydrodaidzein (Hur et al. 2000).
Two other cultures, 2C and 5A (Fig. 1C), had similar chromatograms. Each had one peak with a retention time and a UV/visible spectrum similar to those of equol and another peak, which eluted later, with a retention time and a UV/visible spectrum similar to those of O-DMA (Hur et al. 2002). Twice as much O-DMA was produced in sample 2C as in 5A. Sample 2C was from the same individual, who had provided one sample (2A) 9 months earlier and another (2B) 1 month earlier. Those previous samples did not metabolize daidzein.
1H-NMR analysis of dihydrodaidzein. All experiments were carried out at 500 MHz using a Bruker NMR spectrometer. The proton numbering scheme refers to Fig. 3. The solvent was deuterated methanol and the methanol peak was set to δ=3.31 ppm. Integration is the normalized volume for chemical shift, multiplicity is the peak splitting pattern, and J(Hz) is the first-order coupling constant
C2′ -H, C6′ -H
C3′ -H, C5′ -H
1H-NMR analysis of equol. All experiments were carried out at 500 MHz using a Bruker NMR spectrometer. The conditions are the same as reported in Table 2
The rates of metabolism of daidzein and the production of metabolites by these fecal cultures differed. Sample 5A had completely metabolized 5 μg daidzein/ml after 6 days and sample 2C had metabolized it after 10 days. However, some daidzein was left even after 14 days of incubation with sample 1B.
Daidzein is a weakly estrogenic compound with potential health benefits for several conditions, including hypercholesterolemia and osteoporosis (Adlercreutz 1995; Adlercreutz et al.1993; Wilcox 1990; Lampe and Messina 1998; Lichtenstein 1998; Messina 2000; Messina and Messina 2000). However, wide variation in individual capacity to metabolize these compounds may influence the benefit obtained by the consumption of isoflavonoids (Xu et al. 1995; Joannou et al. 1995; Setchell et al. 2002). In vivo, this difference has been attributed to possible differences in metabolism by the intestinal microflora (Kelly et al. 1993, 1995, Joannou et al. 1995; Setchell et al. 2002). However, there have been no in vitro studies showing differences in daidzein metabolism by the microfloras of different individuals. In this study, we found variation in phytoestrogen metabolism not only in the fecal bacteria of different individuals but also in samples from the same individual obtained at different times. Since all of the fecal samples were transferred to 10% milk, which supports the bacteria during transport and storage, and were used to inoculate the BHI medium within 2 h after collection, this difference could not be attributed to a loss of daidzein-metabolizing bacteria in some samples and not in others.
Dihydrodaidzein was produced by the intestinal microflora of a child whose microflora had not metabolized daidzein when sampled previously. Dihydrodaidzein and equol were found in sample 1B, from another individual whose microflora had not metabolized daidzein when sampled 9 months earlier.
Dihydrodaidzein is an intermediate metabolite, produced by reduction of daidzein, and eventually may be converted to equol (Kelly et al. 1993; Joannou et al. 1995). We previously isolated a bacterium from the intestinal microflora that produces dihydrodaidzein but does not convert it further to equol (Hur et al. 2000). It appears that different bacteria are involved in the different steps of the postulated conversion of daidzein to equol.
Three samples from another individual also showed differences in the metabolism of daidzein. The earlier samples (2A and 2B) did not metabolize daidzein but the later sample (2C) produced O-DMA and equol. Both of the adults providing samples 1B and 2C reported that their diets had changed to contain more fruit and vegetables. Subject no. 1 consumed one to two servings of fruit (one apple, one orange or both) per day. Subject no. 2 consumed at least four servings of fruits (one apple, one orange, one pear and one or two plums) everyday. The child's diet was changed to include pasta and tofu.
The subjects were healthy individuals who had not used antimicrobial agents that could affect the microbial population prior to or during the study. Only one sample from the other subjects (5A) produced both equol and O-DMA. The samples from two other individuals consuming soy diets did not metabolize daidzein, indicating that there is variation in the metabolism of these compounds even with continuous consumption of isoflavonoids.
Equol, which is found in the urine of 30–40% of individuals who consume soy isoflavones, is more estrogenic than either daidzein or O-DMA (Joannou et al. 1995; Sathyamoorthy and Wang 1997; Setchell et al. 2002). The amount of equol that circulates in unbound form in the blood is higher than that of daidzein (49.7% equol and 18.7% daidzein), and equol is excreted in the urine for a longer time than daidzein (Kelley et al. 1995; Nagel et al. 1998; Watanabe et al. 1998; Duncan et al. 2000; Setchell et al. 2000; Lampe et al. 2001). Equol also has superior antioxidant activity when compared with other isoflavones (Setchell et al. 2002). Therefore, it has been assumed that among individuals who regularly consume soy products, those who produce equol have longer exposure to strongly estrogenic compounds. Premenopausal women who excrete equol have plasma hormone profiles associated with a lower risk of breast cancer (Duncan et al. 2000).
Different bacteria may convert dihydrodaidzein to either O-DMA or equol. Two samples in our study had both types of bacteria, which shows that equol producers (Joannou et al. 1995; Sathyamoorthy and Wang 1997; Setchell et al. 2002) may also convert daidzein to metabolites with no estrogenic activity, such as O-DMA, and that individuals whose microflora sometimes produces equol may not produce it at other times.
Although the number of samples obtained was limited by the degree of cooperation of individuals providing them, samples obtained at different times from three participants out of six varied in the metabolism of daidzein. The other three individuals provided only one sample each, and one of these had daidzein-metabolizing bacteria.
Rowland et al. (2000) found that good equol excretors consumed less fat and more carbohydrate as a percentage of energy than poor excretors. In this study, we found that after two of the subjects began consuming more fruits and vegetables, their microfloras produced equol in vitro. We suggest that the prevalence of the bacteria producing equol in fecal samples changed so that equol was produced in some cultures and not in others. The fecal flora in the swab samples used in this study was diluted in comparison to that of actual fecal samples. However, considering that up to 1011 bacterial cells are found per gram of colonic content, a large amount of bacteria of various species should be present in a fecal swab taken from any individual at any time. Since the conditions of transport and growth for all of the samples were the same, the differences observed in the metabolism of daidzein by different samples may be attributed to variation in the prevalence of bacteria metabolizing this compound at different times. Thus equol-producing bacteria were abundant only in some of the samples provided.
Special thanks to K. Barry Delclos for reviewing this manuscript, Dr. John B. Sutherland for discussion and suggestions during manuscript preparation, and Dr. Carl E. Cerniglia for his research support. This work was supported in part by an appointment of (MP and CD) at the National Center for Toxicological Research administered by the Oak Ridge Institute for Science and Education through an interagency agreement between the U.S. Department of Energy and the U.S. Food and Drug Administration and University of Arkansas at Little Rock, Arkansas (CD).
- Adlercreutz H (1995) Phytoestrogens: epidemiology and possible role in cancer protection. Environ Health Perspect 103:103–112Google Scholar
- Coward L, Barnes N, Setchell KDR, Barnes S (1993) Genistein and daidzein, and their β-glycoside conjugates: anti-tumor isoflavones in soybean foods from American and Asian diets. J Agric Food Chem 41:1961–1967Google Scholar
- Duncan AM, Merz-Demlow BE, Xu X, Phipps WR, Kurzer MS (2000) Premenopausal equol excretors show plasma hormone profiles associated with lowered risk of breast cancer. Cancer Epidemiol Biomarkers Prevent 9:581–586Google Scholar
- Lampe J, Messina M (1998) Are phytoestrogens nature's cure for what ails us? A look at the research. J Amer Diet Assoc 98:974–976Google Scholar
- Lampe JW, Skor HE, Li S, Wähälä K, Howald WN, Chen C (2001) Wheat bran and soy protein feeding do not alter urinary excretion of the isoflavan equol in premenopausal women. J Nutr. 131:740–744Google Scholar
- Messina M, Messina V (2000) Soyfoods, soybean isoflavones, and bone health: a brief overview. J Renal Nutr 10:63–68Google Scholar
- Phillips, DA (1992) Flavonoids: plant signals to soil microbes. Annu Rev Phytochem 26:201–231Google Scholar
- Wilcox G, Wahlqvist ML, Burger HG, Medley G (1990) Oestrogenic effects of plant foods in postmenopausal women. Brit Med J 301:905–906Google Scholar