Anaerobic C-ring cleavage of genistein and daidzein by Eubacterium ramulus

Abstract

Eubacterium ramulus, a flavonoid-degrading anaerobic bacterium from the human gastrointestinal tract, was tested for its ability to transform the isoflavonoids genistein-7-O-glucoside (genistin), genistein and daidzein. Genistein was completely degraded by E. ramulus via 6′-hydroxy-O-desmethylangolensin to 2-(4-hydroxyphenyl)-propionic acid. Dihydrogenistein was neither observed as an intermediate in this transformation nor converted itself by growing cells or cell-free extracts of E. ramulus. Genistein-7-O-glucoside was partially transformed by way of genistein to the product 2-(4-hydroxyphenyl)-propionic acid. Daidzein was in part degraded to O-desmethylangolensin, the corresponding metabolite to 6′-hydroxy-O-desmethylangolensin. The hydroxyl group in position 6′ of O-desmethylangolensin is crucial for further degradation.

Isoflavonoid, Genistein, Daidzein, C-ring cleavage, Eubacterium ramulus

1 Introduction

Isoflavonoids are phytoestrogens, whose molecular mass and structure resemble those of steroids. They are taken up in considerable amounts with food and predominantly occur in legumes, especially in soybeans [1], but also in beer [2]. Genistein and daidzein belong to the most common isoflavonoids found in plants. Multiple studies reported that isoflavonoids can positively influence hormone-related cancers, atherosclerosis, osteoporosis, menopausal symptoms and cholesterol levels in the blood [3]. Work on the possible mechanism of action of the isoflavonoids revealed that genistein, for example, inhibits various tyrosine-specific protein kinases [4] and topoisomerases I and II [5]. These enzymes play an important role in cell proliferation and transformation. Tyrosine kinase activity has also been implicated in oncogene expression that is associated with breast cancer [6,7]. Owing to their effects, isoflavonoids have been proposed for use in chemoprevention and therapy of hormone-dependent diseases [8].

However, the effects of the isoflavonoids are dependent on their metabolic fate in the gastrointestinal tract. Certain species of the human intestinal microbiota are able to convert isoflavonoids [9] and may thereby greatly increase or decrease their effect. For example, daidzein does not inhibit topoisomerase I or II, while equol, a transformation product of daidzein, inhibits this enzyme to a larger extent than genistein [10]. To date, only a few isoflavonoid transforming bacterial species have been identified and the reactions catalyzed by these strains are limited to deglycosylation, demethylation and an initial reduction of the aglycon [11,12]. A broader spectrum of metabolites was detected in urine samples of human volunteers challenged with soy [13] and in supernatants of isoflavonoid incubations with total fecal flora [9]. However, a cleavage of the C-ring by a single bacterial species has not been observed yet.

The anaerobic bacterium Eubacterium ramulus, which accounts on average for 0.16% of the total bacterial cell counts in the human gastrointestinal tract, is able to degrade flavonoids that belong to different subclasses such as flavonols and flavones [14,15]. We show herein that E. ramulus, moreover, is capable of cleaving the heterocyclic C-ring of the isoflavonoids genistein and daidzein.

2 Materials and methods

2.1 Medium and growth conditions

For cultivation of E. ramulus (wK1), which was previously isolated by Schneider et al. [14], the anoxic techniques applied were essentially those of Hungate [16] and Bryant [17]. Cultures were grown under strictly anoxic conditions in 16-ml tubes, fitted with butyl rubber stoppers and screw caps. The tubes contained 10 ml of ST medium, which consisted (l⁻¹) of 9 g tryptically digested peptone from meat, 1 g proteose peptone, 3 g meat extract, 4 g yeast extract, 6 g glucose, 3 g NaCl, 2 g Na₂HPO₄, 0.5 ml Tween 80, 0.25 g cystine, 0.25 g cysteine, 0.1 g MgSO₄.7 H₂O, 5 mg FeSO₄.7 H₂O, and 3.4 mg MnSO₄.2 H₂O. The pH was adjusted to 7.0 and the cultures were incubated overnight at 37°C.

2.2 Degradation experiments

The degradation experiments were carried out in 16-ml tubes with butyl rubber stoppers. The tubes contained 9.8 ml ST medium and a gas phase of N₂/CO₂ (80:20, v/v). An aliquot of 100 μl from a stock solution of genistein-7-O-glucoside (20 mM), genistein (50 mM), dihydrogenistein (50 or 5 mM) or daidzein (50 mM) in dimethylsulfoxide (DMSO) was added to the medium under anoxic conditions. The media were inoculated with 100 μl of an exponentially growing culture of E. ramulus and incubated at 37°C. Samples were taken immediately after inoculation, hourly from 2 to 12 h after and after 24 and 48 h. The incubation of E. ramulus with dihydrogenistein was continued for another 5 days. Each sample was centrifuged at 12 000×g for 5 min and the supernatant was subjected to HPLC analysis. The pellet was dissolved in methanol, centrifuged again and the supernatant analyzed by HPLC.

2.3 Preparation of cell-free extracts

E. ramulus cultures were grown overnight in ST medium and cell-free extracts were prepared under anoxic conditions at 4°C. The cells were centrifuged (10 000×g, 15 min) and washed twice with 50 mM potassium phosphate buffer (pH 7.0). Each gram of cell pellet was resuspended in approximately 1 ml of the same buffer supplemented with DNase and ruptured by passage through a French pressure cell at 130 MPa (Aminco, Silver Springs, USA). Cell-free extracts (approximately 15 mg protein ml⁻¹) were obtained by centrifugation at 18 000×g for 20 min.

2.4 Incubation of dihydrogenistein with cell-free extracts

The assay contained 500 or 50 μM dihydrogenistein (added from stock solutions in DMSO) in 50 mM potassium phosphate buffer (pH 7.0). The final DMSO concentration was 1%. The assay was supplemented with 100 μM NADH or NADPH. The reaction was started by the addition of cell-free extract (approximately 150 μg protein ml⁻¹). The assay was performed under N₂/CO₂ (80:20, v/v) at 37°C. Samples were taken immediately after starting the reaction, after 1, 2 and 3 h and mixed with one volume of methanol/H₂O/acetic acid (50:45:5, v/v) to stop the reaction. The mixture was centrifuged at 12 000×g for 5 min and the supernatant was used for HPLC analysis.

2.5 HPLC analysis

The flavonoids and their aromatic degradation products were determined by HPLC analysis according to Braune and coworkers [15]. Methanol and 2% aqueous acetic acid served as mobile phase and formed a gradient as follows: from 5 to 30% methanol in 20 min, from 30 to 50% in 5 min, from 50 to 65% in 5 min, 65% methanol maintained for 5 min, and from 65 to 100% methanol in 7 min. The flow rate was 0.8 ml min⁻¹.

2.6 Mass spectrometry (MS)

Coupled HPLC–electrospray ionization MS (ESI–MS) and collision-induced dissociation tandem mass spectrometry (MS/MS) were carried out according to Braune and coworkers on selected incubation supernatants from degradation experiments [15]. MS analysis was performed in positive ionization mode. The cone and capillary voltage used were 19 V and 3.01 kV, respectively. Product ion scans of [M+H]⁺ were carried out at low-energy collisions (15 eV).

2.7 Nuclear magnetic resonance (NMR) spectrometry

¹H NMR and ¹³C NMR spectra of the synthesized dihydrogenistein were recorded at 300 MHz on Bruker AMX 300 with Me₄Si (δ 0) as the internal standard.

2.8 Chemicals

Genistein-7-O-glucoside (genistin) was obtained from Roth (Karlsruhe, Germany). Genistein, daidzein and 2-(4-hydroxyphenyl)-propionic acid were purchased from Acros Organics (Geel, Belgium). Dihydrogenistein was obtained by hydrogenation of genistein according to Chang and Nair [10]. The identity of the product was checked by LC–MS and NMR analysis.

3 Results and discussion

3.1 Transformation of genistein

Genistein (0.5 mM) was completely transformed by growing cells of E. ramulus, yielding two compounds with retention times of 33.9 min (Gm1) and 26.0 min (Gm2) when analyzed by HPLC (Fig. 1). The time course of genistein degradation is depicted in Fig. 2a. Gm2 was identified as 2-(4-hydroxyphenyl)-propionic acid based on its retention time and UV-spectrum with maxima at 237 and 281 nm and verified by comparison with the pure compound. MS analysis gave the expected [M+H]⁺ of m/z 167. Using LC–MS/MS, the same pattern of daughter ions of m/z 167 [M+H]⁺, namely m/z 93 and m/z 121, was observed for Gm2 and the authentic 2-(4-hydroxyphenyl)-propionic acid. MS analysis of Gm1 resulted in a molecular ion peak of m/z 275 [M+H]⁺. The product ion spectrum of this molecular ion is given in Fig. 3a. It shows high intensities of m/z 121 and 149 as well as additional characteristic daughter ions of m/z 181, 229, and 257. Based on these MS data, the analyzed metabolite was identified as 6′-hydroxy-O-desmethylangolensin. The observed pattern of daughter ions corresponds well with that obtained by Coldham and coworkers, who identified this compound by evaluation of the product ion spectra derived from unlabelled and ¹⁴C-labelled ions [18].

1

HPLC elution profile of the supernatant of genistein transformation by E. ramulus for 3 h under anaerobic conditions. Gm1: 6′-hydroxy-O-desmethylangolensin; Gm2: 2-(4-hydroxyphenyl)-propionic acid.

2

Biotransformation kinetics of isoflavonoids by E. ramulus in ST medium. a: Change of concentration of substrate genistein (?) and metabolites Gm1 (□) and Gm2 (◯); b: change of concentration of substrate genistein-7-O-glucoside (▴) and metabolites Gg1 (◯) and Gg2 (?); c: change of concentration of substrate daidzein (?) and metabolite Dm1 (◯). Broken lines refer to y-axis on the right.

3

Product ion spectra generated by LC–MS/MS from the molecular ions of (a) Gm1 ([M+H]⁺m/z 275) and (b) Dm1 ([M+H]⁺m/z 259).

Based on the structural elucidation of the metabolites and the time course of genistein conversion, the transformation of genistein by E. ramulus includes fission of the C-ring to form 6′-hydroxy-O-desmethylangolensin and further cleavage resulting in the formation of 2-(4-hydroxyphenyl)-propionic acid originating from the B-ring (Fig. 4). Presumably, the A-ring is liberated as phloroglucinol. In E. ramulus, under suboptimal growth conditions, the latter is an intermediate in the breakdown of quercetin, whose structure is highly similar to that of genistein [14]. However, if used as a substrate, phloroglucinol is quickly degraded to butyrate and acetate (unpublished data) and could therefore not be detected as an intermediate in genistein transformation.

4

Conversion of genistein and daidzein by E. ramulus.

The metabolites 6′-hydroxy-O-desmethylangolensin and 2-(4-hydroxyphenyl)-propionic acid have been proposed as intermediates in genistein transformation by caecal microbiota of the rat [18]. 6′-Hydroxy-O-desmethylangolensin was also found in the urine of human volunteers challenged with soy [13,19]. Dihydrogenistein, a proposed intermediate in microbial genistein degradation in the human gastrointestinal tract [13], could not be detected in our experiments. Furthermore, 0.5 or 0.05 mM dihydrogenistein was not converted by growing cells of E. ramulus within a period of 7 days or by cell-free extracts of E. ramulus within 3 h in the presence of NADH or NADPH. These findings indicate that dihydrogenistein was not an intermediate in the conversion of genistein to 2-(4-hydroxyphenyl)-propionic acid, as previously assumed [9,13,18]. However, it cannot be excluded that we did not meet the conditions necessary for the transformation of dihydrogenistein by E. ramulus. But, since the transformation of genistein to dihydrogenistein as the final product has been shown to be catalyzed by specific bacteria in the human intestine [11], it is reasonable to assume that the dihydrogenistein observed in the reported experiments [9,13,18] represents a final product of certain bacterial species but not an intermediate undergoing further degradation.

3.2 Transformation of genistein-7-O-glucoside

The incubation of E. ramulus with 0.5 mM genistein-7-O-glucoside yielded two metabolites with retention times of 26.0 min (Gg1) and 34.7 min (Gg2; Fig. 5). The time course of genistein-7-O-glucoside transformation is depicted in Fig. 2b. Gg1 was identified by HPLC and LC–MS as 2-(4-hydroxyphenyl)-propionic acid, which had already been observed in genistein degradation. Comparison of the retention time and UV-spectrum of Gg2 with the reference substance revealed its identity with genistein. The ability of E. ramulus to hydrolyze the glycosidic bond of genistein-7-O-glucoside is in agreement with data collected by Schneider and Blaut [14]. E. ramulus was reported to cleave off the sugar moiety of flavone and flavonol glycosides in positions 3 and 7 of the ring system. The ability of an isolated bacterial species to hydrolyze the glycosidic bond of genistein-7-O-glucoside and daidzein-7-O-glucoside has been reported previously for Escherichia coli HGH21 and the newly isolated strain HGH6 [11].

5

HPLC elution profile of the supernatant of genistein-7-O-glucoside transformation by E. ramulus for 7 h under anaerobic conditions. Gg1: 2-(4-hydroxyphenyl)-propionic acid; Gg2: genistein.

3.3 Transformation of daidzein

Growing cells of E. ramulus converted 0.5 mM daidzein to a single product with a retention time of 35.4 min (Dm1; Fig. 6). The time course of daidzein degradation is depicted in Fig. 2c. Analysis of Dm1 by ESI–MS revealed a molecular ion [M+H]⁺ of m/z 259 corresponding to both O-desmethylangolensin and 4-hydroxy-equol. These compounds were already described as metabolites detected in the urine of humans challenged with soy [19] and in urine and feces of rats supplemented with daidzein [20]. Since authentic reference compounds were not available, identification was accomplished by interpretation of the MS/MS data. The structural analysis resulted in a product ion fragmentation spectrum of m/z 259 [M+H]⁺ (Fig. 3b) similar to that of 6′-hydroxy-O-desmethylangolensin (Fig. 3a), which was formed during genistein degradation. The product ions m/z 121 and 149, which were postulated to contain the B-ring of the isoflavonoid molecule [18], are found in both spectra. Since the only difference between daidzein and genistein is the additional 6′-hydroxyl group in the latter compound, the A-ring-containing daughter ions of the respective metabolite are expected to differ by a molecular mass of 16. This difference was actually observed with the postulated A-ring containing product ions (m/z 165 versus m/z 181) and the A- and B-ring containing product ions (m/z 213 and 241 versus m/z 229 and 257). In both MS/MS spectra, the fragment m/z 149 (4-HO-C₆H₄-C(CH₃)CO⁺) exhibits the highest intensity. Its formation can by explained by assuming the α-cleavage of 6′-hydroxy-O-desmethylangolensin and of the analogous ketone in the case of the product of daidzein conversion. Taken together, we conclude that the product obtained from daidzein transformation represents O-desmethylangolensin (Fig. 4). Two pathways have been suggested for the degradation of daidzein by intestinal bacteria, resulting in the formation of O-desmethylangolensin or equol [13,19]. E. ramulus transforms daidzein according to the O-desmethylangolensin pathway, and hence does not produce the highly estrogenic equol. O-Desmethylangolensin was not further degraded by E. ramulus, indicating that the 6′-hydroxyl group of 6′-hydroxy-O-desmethylangolensin is a prerequisite for cleavage by this organism.

6

HPLC elution profile of the supernatant of daidzein transformation by E. ramulus for 7 h under anaerobic conditions. Dm1: O-desmethylangolensin.

E. ramulus represents an average of 0.16% of the total fecal flora, which is comparable in number to E. coli, and may therefore be one of the predominant isoflavonoid degrading bacteria in the human gastrointestinal tract. Its characteristics in genistein and daidzein degradation are in accordance with data collected by Xu and coworkers [21], who have shown that genistein is more susceptible to breakdown by gut microorganisms than daidzein. This fact nourishes the thesis that because of the effective degradation of genistein after biliary excretion, less genistein can be reabsorbed. This could explain the lower urinary excretion of genistein (37%) after application in comparison to daidzein (52%), while plasma concentrations of genistein are in general higher than those of daidzein [22]. The overall bioavailability of genistein may therefore be reduced due to the effective degradation of this compound by intestinal bacteria such as E. ramulus.

Acknowledgements

This work was supported by the Deutsche Forschungsgemeinschaft (Grant number INK 26/B1-1). We are indebted to W. Engst (German Institute of Human Nutrition, Bergholz-Rehbrücke, Germany) for supporting us in LC–MS analysis and to A. Krtitschka for NMR analysis (University of Potsdam, Germany).