Identification of New Derivatives of Sinigrin and Glucotropaeolin Produced by the Human Digestive Microflora Using H Nmr Spectroscopy Analysis of in Vitro Incubations

Oneand two-dimensional H NMR spectroscopy were used to study the biotransformation of two dietary glucosinolates, sinigrin (SIN), and glucotropaeolin (GTL) by the human digestive microflora in vitro. The molecular structures of the new metabolites issued from the aglycone moiety of the glucosinolate were identified, and the modulation of carbon metabolism was studied by quantifying bacterial metabolites issued from the xenobiotic incubation in the presence or absence of a source of free glucose. Unambiguously and for the first time, it was shown that SIN and GTL were transformed quantitatively into allylamine and benzylamine, respectively. The comparison of the kinetics of transformation of SIN and GTL with and without glucose clearly showed that the presence of glucose did not modify either the nature of the metabolites or the rate of transformation of the glucosinolates (complete degradation within 30 h). The main end products of the glucose moiety of glucosinolates were characteristic of anaerobic carbon metabolism in the digestive tract (acetate, lactate, ethanol, propionate, formate, and butyrate) and similar to those released from free glucose. This work represents the first application of H NMR spectroscopy to the study of xenobiotic metabolism by the human digestive microflora, demonstrating allyland benzylamine production from glucosinolates. Whether these amines are produced in vivo from dietary glucosinolates remains to be established. This would reduce the availability of other glucosinolate metabolites, notably cancer-protective isothiocyanates. Glucosinolates are sulfur-containing phytochemicals present in edible cruciferous plants, such as Brussels sprouts, cabbage, radish, etc. Their common structure comprises a -thioglucose group, a sulfonated oxime moiety, and a variable side chain derived from methionine, tryptophan, or phenylalanine. Upon disruption of plant tissues during food processing or ingestion, glucosinolates are hydrolyzed by the endogenous enzyme “myrosinase” (thioglucoside glucohydrolase EC 3.2.3.1.) to yield isothiocyanates, nitriles, and other minor products (Fenwick et al., 1983). Numerous experimental studies performed in animal and cellular models implicated isothiocyanates as the main bioactive agents responsible for the anti-cancer properties of cruciferous vegetables (Musk and Johnson, 1993; Nugon-Baudon and Rabot, 1994; Verhoeven et al., 1997). In comparison, the fate of glucosinolates following ingestion, in particular their breakdown site and rate, has received little attention. When plant myrosinase is active, glucosinolates are rapidly hydrolyzed in the food or in the proximal gut (De Vos and Blijleven, 1988; Campbell et al., 1995). Should the enzyme be deactivated by cooking (Jongen, 1996), glucosinolates reach the large bowel, where they are broken down by the resident microflora (Rabot et al., 1993; Michaelsen et al., 1994). A few recent experiments performed in humans and gnotobiotic rats associated with a human fecal flora have demonstrated that microbial breakdown of glucosinolates leads to the formation of isothiocyanates (Shapiro et al., 1998; Elfoul et al., 2001). The conversion is incomplete, however, suggesting that other metabolites are produced; at present, nothing is known of the identity of these latter products. The aim of this study was to investigate the metabolism of glucosinolates by the human colonic microflora in vitro, using H NMR spectroscopy. Indeed, NMR signals are real fingerprints of molecules, and a wide range of metabolites can be measured simultaneously and without a priori hypothesis. In addition, despite a rather low sensitivity, this method is very convenient since it can be performed directly on biological samples without prior purification. Finally, quantitative data can be collected. H NMR spectroscopy has proved to be a powerful tool for analysis of the metabolite composition of biological fluids (Nicholson and Wilson, 1989; Fan, 1996; Kalic et al., 2000) and for the study of drug metabolism (Gartland et al., 1991; Holmes et al., 1995; Bollard et al., 1996; Foxall et al., 1996). More recently, this technique has been applied to the study of microbial metabolism (Matheron et al., 1998; Brecker and Ribbons, 2000; Weber and Brecker, 2000) and, in particular, microbial degradation of xenobiotics (Gaines et al., 1996; Besse et al., 1998; Combourieu et al., 1998a,b, 2000; Poupin et al., 1998; Delort and Combourieu, 2000). This research was supported by the European Community under the program FAIR CT97 3029 entitled Effects of Food-Borne Glucosinolates on Human Health and by a fellowship awarded to one of the authors (L.E.) by the French Ministry of Education, Research, and Technology. Address correspondence to: Anne-Marie Delort, Laboratoire Synthèse Electrosynthèse et Etude de Systèmes à Intérêt Biologique, UMR 6504 du CNRS, Université Blaise Pascal, 63177 Aubière Cedex, France. E-mail: [email protected] 0090-9556/01/2911-1440–1445$3.00 DRUG METABOLISM AND DISPOSITION Vol. 29, No. 11 Copyright © 2001 by The American Society for Pharmacology and Experimental Therapeutics 396/937664 DMD 29:1440–1445, 2001 Printed in U.S.A. 1440 at A PE T Jornals on A ril 8, 2017 dm d.aspurnals.org D ow nladed from In this work, we have used 1D and 2D H NMR spectroscopy to elucidate the microbial biotransformation of two structurally different glucosinolates, sinigrin and glucotropaeolin. Sinigrin (SIN; Fig. 1) is a simple aliphatic glucosinolate prevalent in a wide range of cruciferous vegetables, whereas glucotropaeolin (GTL; Fig. 1) is an aromatic glucosinolate specifically present in garden cress and papaya fruit. In comparative experiments, SIN and GTL were added to the culture media, either alone or concurrently with free glucose. In this way, we sought to determine whether the availability of a simple source of energy and carbon, as is likely to be the case in the digestive tract, would hamper glucosinolate metabolism by digestive bacteria. In addition, data were collected from the NMR spectra to identify and quantify metabolites produced by the fermentation of the glucosinolate sugar moiety. Materials and Methods Incubation Experiments. Freshly passed stools were collected from a healthy adult human subject who had not taken any antibiotics for at least 3 months preceding the study and who usually consumed a Western style diet. Feces were transferred into an anaerobic glove box where they were thoroughly mixed with the incubation buffer, using an Ultra-turrax blender (Janke and Kunkel GmbH, Staufen, Germany). Incubation buffer was 0.1 M sterile potassium phosphate, pH 7.0, with added yeast extract (Difco) 2 g l . The suspension was adjusted at 1 g 100 ml 1 by further addition of incubation buffer and divided into 20 ml fractions, transferred to amber glass vials, either alone (control without substrate) or together with SIN (Sigma-Aldrich, St. Louis, MO) and GTL (Merck, Darmstadt, Germany) (6 mM each) or SIN, GTL, and glucose (6 mM each). Sterile controls containing SIN and GTL were prepared to check the stability of the substrates in the absence of bacterial cells (controls without flora). SIN, GTL, and glucose were added as freshly prepared aqueous stock solutions sterilized by filtration (Millex-GS 0.22 m; Millipore Corporation, Bedford, MA). Vials were tightly closed with butyl-rubber stoppers and sealed with aluminum caps to maintain anoxic conditions and to avoid loss of volatile isothiocyanates. They were incubated at 37°C in a shaking bath (50 rpm). All incubations were performed in duplicate. Samples were collected by a puncture through the stoppers at 0, 3, 6, 18, and 30 h. They were centrifuged to remove bacteria (8000g, 10 min, 4°C), and supernatants were stored at 20°C until analysis. HPLC Analysis. Supernatants were analyzed for residual intact SIN and GTL using HPLC analysis of desulfoglucosinolates, as recommended by the International Standardization Organization for glucosinolate analysis in Brassica (Anonymous, 1990), except that no methanol extraction was required before the desulfation step. NMR Spectroscopy. Preparation of the samples for NMR and quantification of the metabolites were performed as previously described (Combourieu et al., 1998a). No purification was performed before analysis, and pH was adjusted to 7 to avoid changes in chemical shifts. TSPd4 (10% v/v of an 8 mM solution in D2O) constituted a reference for chemical shifts (0 ppm) and quantification. All H NMR spectra were recorded on a Bruker Avance 300 spectrometer (Bruker, Newark, DE) at 300.13 MHz at 25°C with a 5-mm H-C-N inverse probe equipped with z-gradients. 1D H NMR experiments. Water was suppressed by presaturation or by the classical double-pulsed field gradient echo sequence: WATERGATE (Price, 1999). In both cases, 64 scans were collected (relaxation delay, 5 s; acquisition time, 3.64 s; 32,000 data points). A 0.3-Hz line broadening was applied before Fourier transformation, and a baseline correction was performed on spectra before integration with Bruker software. The H NMR spectra obtained were compared with those of the controls. 2D H NMR experiments. Gradients-correlation spectroscopy was used to identify all the end products of carbon metabolism. The data were acquired as 2048 256 point files, accumulating eight transients per t1 increment. Zerofilling in t1 and unshifted sinusoidal window function in both time domains were employed before Fourier transformation. 2D phase-sensitive total correlation spectroscopy (TOCSY) experiments with water resonance suppression by a WATERGATE sequence (put at the end of the sequence) were used to assign all members of a coupled spin network. Spectral widths were adjusted in both dimensions to encompass all H signals of interest. The “mixing period” (corresponding to several cycles of MLEV-17 spin-lock sequence) was 20 to 80 ms. The responses of eight scans for each of 512 t1 increments were acquired. Zero-filling in t1 and sine window function in both dimensions were applied before 2D Fourier transformation. GC-MS Analysis. Samples were extracted three times with CDCl3, and the organic layers were dried over MgSO4. The extracts were analyzed directly by GC-MS on an HP 5890 Series II Plus GC equipped with an HP 5989 B mass spectrometer (Hewlett Packard, Palo Alto, CA). Separation was achieved on a 30-m OPTIMA-5-MS capillary column (0.25-mm i.d. and 0.25m film) using the following temperature program: 55°C, 2-min hold, increased by 5°C min 1 to 80°C, 15-min hold. Helium was used as a carrier gas at a linear velocity of 41 cm s . The temperatures of injector, interface, and source were 250, 250, and 200°C, respectively. The ionization mode was electronic impact at 70 eV, and the detection mode chosen was single ion monitoring, which increases sensitivity by a factor of 10.