OUP user menu

Formation of allyl isothiocyanate from sinigrin in the digestive tract of rats monoassociated with a human colonic strain of Bacteroides thetaiotaomicron

Lila Elfoul, Sylvie Rabot, Nasser Khelifa, Alain Quinsac, Annabelle Duguay, Alain Rimbault
DOI: http://dx.doi.org/10.1111/j.1574-6968.2001.tb10589.x 99-103 First published online: 1 April 2001


A human digestive strain of Bacteroides thetaiotaomicron was tested for its ability to metabolise sinigrin, a glucosinolate commonly found in Brassica vegetables. Gnotobiotic rats harbouring the bacterial strain were orally dosed with 50 μmol sinigrin. HPLC analysis of the digestive contents showed that sinigrin was degraded in the large bowel, where B. thetaiotaomicron was established at a high level. Concurrently, a hydrolysis product of sinigrin, allyl isothiocyanate, was identified by GC-MS analysis, following headspace solid-phase microextraction of the digestive contents; its production peaked at ca. 200 nmol, 12 h after dosing. This is the first study to demonstrate in vivo the involvement of a human colonic predominant bacterium in the bioconversion of a dietary glucosinolate to a potentially anticarcinogenic isothiocyanate.

  • Bacteroides thetaiotaomicron
  • Gastrointestinal tract
  • Glucosinolate
  • Sinigrin
  • Allyl isothiocyanate
  • Gnotobiology

1 Introduction

Members of the genus Bacteroides are among the predominant bacteria found in the human gastrointestinal tract. In accordance with their metabolic versatility, many Bacteroides species have evolved certain enzymes to aid in the degradation of the wide range of dietary components they may encounter in the colonic environment [1]. We have isolated from human faeces a Bacteroides thetaiotaomicron strain (II8) capable of degrading in vitro glucosinolates, a group of β-thioglucosides present in Brassica vegetables [2]. Upon disruption of plant tissues occurring during food processing or ingestion, glucosinolates are hydrolysed by the endogenous enzyme myrosinase (thioglucoside glucohydrolase, EC to yield isothiocyanates, nitriles and other minor products [3]. Should the myrosinase be deactivated by cooking [4], glucosinolates reach the large bowel where they are broken down by the resident microflora [5,6]. However, it is still not known whether the degradation products formed are the same as those produced by plant myrosinase. Such information is necessary for a proper understanding of the bioavailability and, subsequently, of the physiological impact of food-borne glucosinolates. Indeed, many experiments point out isothiocyanates as the bioactive agents responsible for the anti-cancer properties of Brassica vegetables [7,8].

In this study, our aim was to investigate the ability of B. thetaiotaomicron II8 to convert glucosinolates into isothiocyanates and nitriles in vivo, using sinigrin, a simple aliphatic glucosinolate, as a model compound (Fig. 1). The digestive release of allyl isothiocyanate (AITC; Fig. 1) and allyl cyanide (ACN; Fig. 1) from artificially administered sinigrin was determined in gnotobiotic rats harbouring B. thetaiotaomicron II8 and consuming a Brassica-containing food aimed at providing a realistic dietary environment.

Figure 1

Chemical structures of sinigrin, allyl isothiocyanate and allyl cyanide.

2 Materials and methods

2.1 Experimental design

Thirty germ-free F344 adult male rats, originating from the INRA Breeding Unit and housed in Plexiglas® isolators (Ingénia, Villejuif, France), were orally inoculated with a 24-h culture of B. thetaiotaomicron II8. The strain was grown in brain-heart infusion broth 37 g l−1 (Difco), pH 7.0, supplemented with yeast extract (Difco) 5 g l−1, haemin (Sigma-Aldrich) 5 mg l−1 and sinigrin (Sigma-Aldrich) 10 mM, and incubated at 37°C in anoxic conditions. Animals were given free access to a pelleted diet (Table 1), sterilised by γ-irradiation at 45 kGy, and to tap water, sterilised by steaming (120°C, 40 min).

View this table:
Table 1

Composition of the diet (g kg−1)

Maize starch229.85
Mashed potato290.00
Isolated soy proteina80.00
Rapeseed mealb100.00
Maize oil30.00
Mineral additivec70.00
Vitamin additived10.00
Protein (N×6.25)170.6
Gross energy (MJ kg−1)16.9
  • aIsolated soy protein PP500E (Protein Technologies International).

  • bDehulled myrosinase-free and sinigrin-free rapeseed meal (cv Darmor) containing 38.5 μmol g−1 glucosinolates.

  • cThe mineral mixture provided the following (mg kg−1 diet): CaHPO4, 30 100; KCl, 7000; NaCl, 7000; MgO, 735; MgSO4, 3500; Fe2O3, 210; FeSO4.7H2O, 350; MnSO4.H2O, 170; CuSO4.5H2O, 35; ZnSO4.7H2O, 141; CoSO4.7H2O, 0.28; KI, 0.56.

  • dThe vitamin mixture provided the following (kg−1 diet): retinol, 19 800 IU; cholecalciferol, 2500 IU; thiamine, 20 mg; riboflavin, 15 mg; pantothenic acid, 70 mg; pyridoxine, 10 mg; myoinositol, 150 mg; cyanocobalamin, 0.05 mg; ascorbic acid, 800 mg; tocopherol, 170 mg; menadione, 40 mg; niacin, 100 mg; choline, 1360 mg; folic acid, 5 mg; p-aminobenzoic acid, 50 mg; biotin 0.3 mg.

After 2 weeks adaptation to the diet and the flora, two rats were killed by CO2 inhalation and the levels of bacterial population in the stomach, small intestine, caecum and colon were determined by anaerobically growing serial 10-fold dilutions of the contents [6]. The other rats were randomly allocated to seven groups of four animals each and dosed by stomach tube under light ether anaesthesia, either with 50 μmol sinigrin dissolved in 0.5 ml distilled water (six treated groups), or with 0.5 ml distilled water alone (one control group). Each dose was flushed into the stomach with 0.5 ml distilled water. Solutions used for the gavage were prepared and filter-sterilised immediately prior to administration. After dosing, rats were killed by CO2 inhalation, at 0 h for the control group and at 3, 6, 12, 18, 24 or 36 h for the treated groups. The stomach, small intestine, caecum and colon were collected separately and their contents divided into two samples. One was freeze-dried and stored at −20°C for sinigrin analysis. The other was transferred into a 12-ml amber glass vial cooled in a water–ice bath, along with an appropriate volume of chilled phosphate buffer 100 mM, pH 7.0, to allow for a 4-ml headspace volume. Vials were tightly closed with butyl-rubber stoppers, sealed with aluminium caps and stored at −20°C until AITC and ACN analysis.

All procedures were carried out in accordance with the European guidelines for the care and use of laboratory animals.

2.2 Biochemical analyses

Sinigrin was analysed following the HPLC method recommended by the International Standardisation Organisation [9].

AITC and ACN, which are volatile compounds, were analysed by gas chromatography (GC), following headspace solid-phase microextraction [10]. Briefly, a 75 μm carboxene–poly(dimethylsiloxane)-coated silica fibre (Supelco) was exposed for 10 min to the headspace of the sample maintained at 45°C in a water bath, after addition of benzyl isothiocyanate (Sigma-Aldrich) as an internal standard. Thermal desorption of the analytes extracted into the fibre coating and their subsequent GC separation were carried out using a Carlo Erba model HRGC 5300 chromatograph equipped with a flame ionisation detector. Samples were desorbed at 250°C for 20 s in the splitless injection port and separation was achieved on a CP-Sil 8 CB capillary column (25 m×0.53 mm i.d.; film thickness, 2 μm, Chrompack) with nitrogen as the carrier gas (inlet pressure 70 kPa). The oven temperature was maintained at 45°C for 3 min, ramped at 6°C min−1 to 160°C, then further ramped at 10°C min−1 to 200°C and held for 1 min. Detector temperature was 250°C. Data were collected and peaks integrated using a Shimadzu model C-R6A integrator. Identification was based on the identity of retention times with those of the authentic standards (Sigma). It was confirmed by GC-mass spectrometry (MS) analysis run on a Fisons model 800 gas chromatograph coupled to a quadrupole Fisons MD 800 mass spectrometer with an INCOS (Finnigan) data system. GC separation was achieved on a Q2 capillary column (25 m×0.25 mm i.d.; film thickness, 0.25 μm; Quadrex) and the oven was temperature-programmed from 55°C for 2 min to 80°C at a rate of 5°C min−1. Ionisation was performed by electron impact (70 eV; emission current, 0.5 mA) and masses were scanned from 50 to 300 amu.

3 Results

The population level of B. thetaiotaomicron II8 varied according to the digestive compartment. In the stomach and small intestine, average bacterial counts, expressed as log10 of the number of bacteria per g of content, were 2.58 and 7.00, respectively. In the large bowel, the strain became established at a high level, i.e. 9.60 and 9.48 in the caecum and colon, respectively.

No sinigrin was detected in the gastrointestinal tract of control rats. In treated rats, the highest recovery of intact sinigrin averaged 28 μmol, 3 h after the gavage (Fig. 2); the largest amount, accounting for 46% of the ingested dose, was present in the caecum and colon, with the stomach and small intestine containing only a low quantity (10% of the ingested dose). After 6 h, sinigrin steadily decreased in all compartments; from 18 h onwards, it was virtually absent in the stomach and small intestine and did not exceed 2 μmol in the caecum and colon.

Figure 2

Amounts of sinigrin in the gastrointestinal tract of gnotobiotic rats harbouring B. thetaiotaomicron II8, at various times after oral dosing with 50 μmol sinigrin.

Neither ACN nor AITC was detected in the gastrointestinal tract of control rats. In treated rats there was no ACN either, whereas AITC appeared as soon as 3 h after dosing (Fig. 3). Its identification was ascertained by the identity of its chromatographic retention time (Fig. 4) and of its mass spectra (Fig. 5) with those of the authentic standard. In the stomach, trace amounts (≤1 nmol) were temporarily detected; in the small intestine, a small amount, averaging 10 nmol, was present at 3 h and traces persisted until the end of the experiment. The highest quantities appeared in the large bowel, mainly from 6 to 12 h after dosing; at this time, the caecum and the colon contained 112±1 and 88±21 nmol of AITC (mean±S.E.M.), respectively. After 12 h, a steady decrease was observed but substantial amounts of the order of 30 nmol were still found at 24 h.

Figure 3

Amounts of AITC in the gastrointestinal tract of gnotobiotic rats harbouring B. thetaiotaomicron II8, at various times after oral dosing with 50 μmol sinigrin.

Figure 4

GC profile obtained for the caecal content of a gnotobiotic rat harbouring B. thetaiotaomicron II8, 24 h after oral dosing with 50 μmol sinigrin. For GC conditions, see text.

Figure 5

Mass spectra (EI, 70 eV) of AITC showing a molecular ion at m/z 99 and fragments at m/z 58 and 72. Sample was the caecal content of a gnotobiotic rat harbouring B. thetaiotaomicron II8, 24 h after oral dosing with 50 μmol sinigrin. For GC-MS conditions, see text.

4 Discussion

Glucosinolates are a major group of potentially anticarcinogenic factors naturally occurring in Brassica vegetables [11]. As animal and cell culture studies provide increasing evidence that isothiocyanates are the main glucosinolate breakdown products responsible for this beneficial effect in human health [7,8], knowledge of the bioavailability of glucosinolates and isothiocyanates following ingestion of Brassica becomes more crucial [12]. The importance of intestinal bacteria in the biotransformation and biological activities of food-borne glucosinolates is clearly established [5,6,13,14]; yet little is known of the nature of breakdown products. Palop et al. [15] report the formation of AITC upon incubation of a Lactobacillus agilis strain with sinigrin, but the relevance of this bacterial species to the human colonic microflora is unlikely. Recently, isothiocyanates have been detected upon in vitro incubation of human faeces with watercress juice of which the myrosinase activity had been deactivated by cooking [16]. In this study, we have shown that a B. thetaiotaomicron strain isolated from human faeces is capable of converting sinigrin into AITC. This biotransformation occurred in vivo, in the large intestine of gnotobiotic rats, where the B. thetaiotaomicron concentration was in accordance with the usual level of this species in the human colon [1].

Musk et al. [17] have demonstrated that AITC is selectively cytotoxic to colonic cancer cells in vitro. In rats treated with a chemical carcinogen, oral administration of sinigrin suppresses proliferation and increases apoptosis in the basal regions of colorectal crypts [18]. The delivery of AITC to the colonic mucosa following sinigrin hydrolysis by B. thetaiotaomicron supports the idea that the colonic microflora plays a key role in the protective effects of dietary sinigrin against the development of colorectal cancer.

Measurements of instantaneous amounts of sinigrin and AITC in digestive fluids do not allow us to estimate the conversion rate of sinigrin to AITC because they represent a balance between degradation, production and absorption. Nevertheless, comparison of the relative amounts of sinigrin and AITC in the large intestine of gnotobiotic rats, respectively in the order of 10 μmol and 100 nmol at their highest points, may suggest that the B. thetaiotaomicron strain has converted sinigrin into other metabolites. Absence of ACN, the other main potential product from sinigrin hydrolysis, does not support this idea and reflects that physiological conditions present in the large intestine are not favourable for the formation of nitriles; these are indeed preferentially formed in acidic conditions [3]. However, formation of ACN followed by its ready conversion into secondary metabolites, as observed in sheep rumen, cannot be totally ruled out [19]. Similarly, the B. thetaiotaomicron strain may have metabolised AITC further, although, so far, isothiocyanate transformation into secondary metabolites has been identified only in non-digestive bacteria [20]. It is possible that AITC was absorbed through the colonic mucosa, thus being able to exert systemic effects. Pharmacokinetics studies using orally administered isothiocyanates show their rapid absorption from the upper gastrointestinal tract in rat, followed by their conversion to N-acetylcysteine derivatives excreted in urine [21,22]. Urinary analysis of the N-acetylcysteine conjugate of AITC in future investigations should make it possible to estimate the extent of AITC production and absorption in the large intestine, following ingestion of sinigrin.

The quantity of intact sinigrin that was recovered throughout the gastrointestinal tract at the earliest collection point, i.e. 3 h after the gavage, accounted only for 55% of the dose ingested. This result, together with the low level of AITC at the same time, unexpectedly suggests that sinigrin is partly absorbed without prior hydrolysis in the digestive tract. Michaelsen et al. [5] have pointed out that sinigrin can be passively transported from the mucosal to the serosal side of everted sacs made from small intestine and colon of rodents. However, the absence of biological effects in germ-free rats fed on a glucosinolate-containing diet shows that this phenomenon is not accompanied by the formation of bioactive derivatives [23]. A thorough analysis of body tissues and fluids for intact glucosinolates following their ingestion would help to clarify this point.

On the whole, these findings give new insight into the fate of food-borne glucosinolates in the gastrointestinal tract, and bring out a new metabolic activity of a numerically predominant bacterial species of the human colon towards dietary compounds. Experiments with rats associated with a whole human faecal flora are in progress to extend their biological significance. Indeed, other intestinal bacteria are likely to achieve the same metabolic reaction and contribute to the overall production of bioactive isothiocyanates in the large bowel.


The authors thank Mr José Durao for taking care of the gnotobiotic rats and Mrs Solène Garrido for technical assistance in HPLC analyses. Thanks also to Dr Jacques Evrard, CETIOM, Pessac, France, for the generous gift of the rapeseed meal. Dr Jean-Luc Luisier and Mr Ramin Azodanlou, Ecole d'Ingénieurs du Valais, Sion, Switzerland, provided most helpful advice on the use of solid-phase microextraction. This research was supported by the European Community under the programme 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.


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
  22. [22].
  23. [23].
View Abstract