Formation of DNA Adducts by 1-Methoxy-3-indolylmethylalcohol, a Breakdown Product of a Glucosinolate, in the Mouse: Impact of the SULT1A1 Status-Wild-Type, Knockout or Humanised.

We previously found that feeding rats with broccoli or cauliflower leads to the formation of characteristic DNA adducts in the liver, intestine and various other tissues. We identified the critical substances in the plants as 1-methoxy-3-indolylmethyl (1-MIM) glucosinolate and its degradation product 1-MIM-OH. DNA adduct formation and the mutagenicity of 1-MIM-OH in cell models were drastically enhanced when human sulfotransferase (SULT) 1A1 was expressed. The aim of this study was to clarify the role of SULT1A1 in DNA adduct formation by 1-MIM-OH in mouse tissues in vivo. Furthermore, we compared the endogenous mouse Sult1a1 and transgenic human SULT1A1 in the activation of 1-MIM-OH using genetically modified mouse strains. We orally treated male wild-type (wt) and Sult1a1-knockout (ko) mice, as well as corresponding lines carrying the human SULT1A1-SULT1A2 gene cluster (tg and ko-tg), with 1-MIM-OH. N2-(1-MIM)-dG and N6-(1-MIM)-dA adducts in DNA were analysed using isotope-dilution UPLC-MS/MS. In the liver, caecum and colon adducts were abundant in mice expressing mouse and/or human SULT1A1, but were drastically reduced in ko mice (1.2-10.6% of wt). In the kidney and small intestine, adduct levels were high in mice carrying human SULT1A1-SULT1A2 genes, but low in wt and ko mice (1.8-6.3% of tg-ko). In bone marrow, adduct levels were very low, independently of the SULT1A1 status. In the stomach, they were high in all four lines. Thus, adduct formation was primarily controlled by SULT1A1 in five out of seven tissues studied, with a strong impact of differences in the tissue distribution of mouse and human SULT1A1. The behaviour of 1-MIM-OH in these models (levels and tissue distribution of DNA adducts; impact of SULTs) was similar to that of methyleugenol, classified as "probably carcinogenic to humans". Thus, there is a need to test 1-MIM-OH for carcinogenicity in animal models and to study its adduct formation in humans consuming brassicaceous foodstuff.

Feeding Brassica vegetables to rats leads to the formation of characteristic DNA adducts (from 1-methoxy-3-indolylmethyl glucosinolate) in many tissues.

Juices of Brassica vegetables are mutagenic and form characteristic DNA adducts in bacteria and mammalian cells. In this study, we examined whether such adducts are also formed in vivo in animal models. Rats fed raw broccoli ad libitum in addition to normal laboratory chow for 5 weeks showed one major adduct spot and sometimes an additional minor adduct spot in liver, kidney, lung, blood and the gastrointestinal tract, as determined by 32P-postlabelling/thin-layer chromatography. Adducts with the same chromatographic properties were formed when herring sperm DNA (or dG-3'-phosphate) was incubated with 1-methoxy-3-indolylmethyl glucosinolate (phytochemical present in Brassica plants), in the presence of myrosinase (plant enzyme that hydrolyses glucosinolates to bioactive breakdown products). UPLC-MS/MS analysis corroborated this finding: 1-Methoxy-3-indolylmethyl-substituted purine nucleosides were detected in the hepatic DNA of broccoli-fed animals, but not in control animals. Feeding raw cauliflower led to the formation of the same adducts. When steamed rather than raw broccoli was used, the adduct levels were essentially unchanged in liver and jejunum, but elevated in large intestine. Due to inactivation of myrosinase by the steaming, higher levels of the glucosinolates may have reached the large bowl to be activated by glucosidases from intestinal bacteria. In conclusion, the consumption of common Brassica vegetables can lead to the formation of substantial levels of DNA adducts in animal models. The adducts can be attributed to a specific phytochemical, neoglucobrassicin (1-methoxy-3-indolylmethyl glucosinolate).

Methyleugenol DNA adducts in human liver are associated with SULT1A1 copy number variations and expression levels.

Methyleugenol is a rodent hepatocarcinogen occurring in many herbs and spices as well as essential oils used for flavoring. Following metabolic activation by cytochromes P450 (CYPs) and sulfotransferases (SULTs), methyleugenol can form DNA adducts. Previously, we showed that DNA adduct formation by methyleugenol in mouse liver is dependent on SULT1A1 expression and that methyleugenol DNA adducts are abundant in human liver specimens. In humans, SULT1A1 activity is affected by genetic polymorphisms, including single-nucleotide polymorphisms (SNPs) and copy number variations (CNVs). Here we investigated the relationship between individual methyleugenol DNA adduct levels and SULT1A1 in human liver samples. Using isotope-dilution ultraperformance liquid chromatography coupled with tandem mass spectrometry, we quantified methyleugenol DNA adducts in 121 human surgical liver samples. Frequent CNVs, including deletions (f = 3.3%) and duplications (f = 36.4%) of SULT1A1, were identified using qPCR and TaqMan assays in the donors' genomic DNA. SULT1A1 mRNA and protein levels were quantified using microarray data and Western blot analysis, respectively. Methyleugenol DNA adducts were detected in all 121 liver samples studied. Their levels varied 122-fold between individuals and were significantly correlated to both mRNA and protein levels of SULT1A1 (r s = 0.43, and r s = 0.44, respectively). Univariate and multivariate statistical analysis identified significant associations of SULT1A1 CNVs with mRNA (p = 1.7 × 10-06) and protein (p = 4.4 × 10- 10) levels as well as methyleugenol DNA adduct levels (p = 0.003). These data establish the importance of SULT1A1 genotype for hepatic methyleugenol DNA adducts in humans, and they confirm a strong impact of SULT1A1 CNVs on SULT1A1 hepatic phenotype.

Chalcone Isomerase from Eubacterium ramulus Catalyzes the Ring Contraction of Flavanonols.

The enzyme catalyzing the ring-contracting conversion of the flavanonol taxifolin to the auronol alphitonin in the course of flavonoid degradation by the human intestinal anaerobe Eubacterium ramulus was purified and characterized. It stereospecifically catalyzed the isomerization of (+)-taxifolin but not that of (-)-taxifolin. The Km for (+)-taxifolin was 6.4 ± 0.8 µM, and the Vmax was 108 ± 4 µmol min-1 (mg protein)-1 The enzyme also isomerized (+)-dihydrokaempferol, another flavanonol, to maesopsin. Inspection of the encoding gene revealed its complete identity to that of the gene encoding chalcone isomerase (CHI) from E. ramulus Based on the reported X-ray crystal structure of CHI (M. Gall et al., Angew Chem Int Ed 53:1439-1442, 2014, http://dx.doi.org/10.1002/anie.201306952), docking experiments suggest the substrate binding mode of flavanonols and their stereospecific conversion. Mutation of the active-site histidine (His33) to alanine led to a complete loss of flavanonol isomerization by CHI, which indicates that His33 is also essential for this activity. His33 is proposed to mediate the stereospecific abstraction of a proton from the hydroxymethylene carbon of the flavanonol C-ring followed by ring opening and recyclization. A flavanonol-isomerizing enzyme was also identified in the flavonoid-converting bacterium Flavonifractor plautii based on its 50% sequence identity to the CHI from E. ramulus IMPORTANCE: Chalcone isomerase was known to be involved in flavone/flavanone conversion by the human intestinal bacterium E. ramulus Here we demonstrate that this enzyme moreover catalyzes a key step in the breakdown of flavonols/flavanonols. Thus, a single isomerase plays a dual role in the bacterial conversion of dietary bioactive flavonoids. The identification of a corresponding enzyme in the human intestinal bacterium F. plautii suggests a more widespread occurrence of this isomerase in flavonoid-degrading bacteria.

Identification and functional expression of genes encoding flavonoid O- and C-glycosidases in intestinal bacteria.

Gut bacteria play a crucial role in the metabolism of dietary flavonoids and thereby influence the bioactivity of these compounds in the host. The intestinal Lachnospiraceae strain CG19-1 and Eubacterium cellulosolvens are able to deglycosylate C- and O-coupled flavonoid glucosides. Growth of strain CG19-1 in the presence of the isoflavone C-glucoside puerarin (daidzein 8-C-glucoside) led to the induction of two proteins (DfgC, DfgD). Heterologous expression of the encoding genes (dfgC, dfgD) in Escherichia coli revealed no C-deglycosylating activity in the resulting cell extracts but cleavage of flavonoid O-glucosides such as daidzin (daidzein 7-O-glucoside). The recombinant DfgC and DfgD proteins were purified and characterized with respect to their quaternary structure, substrate and cofactor specificity. The products of the corresponding genes (dfgC, dfgD) from E. cellulosolvens also catalysed the O-deglycosylation of daidzin following their expression in E. coli. In combination with three recombinant proteins encoded by adjacent genes in E. cellulosolvens (dfgA, dfgB, dfgE), DfgC and DfgD from E. cellulosolvens catalysed the deglycosylation of the flavone C-glucosides homoorientin (luteolin 6-C-glucoside) and isovitexin (apigenin 6-C-glucoside). Even intact cells of E. coli expressing the five E. cellulosolvens genes cleaved these flavone C-glucosides and, also, flavonoid O-glucosides to the corresponding aglycones.

Determination of sulfotransferase forms involved in the metabolic activation of the genotoxicant 1-hydroxymethylpyrene using bacterially expressed enzymes and genetically modified mouse models.

1-Methylpyrene, a carcinogenic polycyclic aromatic hydrocarbon, forms benzylic DNA adducts, in particular N2-(1-methylpyrenyl)-2'-deoxyguanosine, in mice and rats. It is bioactivated via 1-hydroxymethylpyrene (1-HMP) to electrophilic 1-sulfooxymethylpyrene (1-SMP). In this study, we explored the role of individual mouse sulfotransferase (SULT) forms in this activation. First, we showed that all nine mouse SULTs tested were able to activate 1-HMP to a mutagen in the his- Salmonella typhimurium reversion test. Some activation was even observed with Sult2a3 and Sult5a1, orphan forms for which no substrates were identified hitherto. Subsequently, we used cytosolic preparations from tissues of four mouse lines (wild-type, Sult1a1-, Sult1d1-, and transgenic for human SULT1A1/2) for the activation of 1-HMP in the mutagenicity assay. The most prominent impacts of the genetic SULT status were 96% decrease in hepatic activation by Sult1a1 knockout, 99% decrease in renal activation by Sult1d1 knockout, and 100-fold increase in pulmonary activation by transgenic human SULT1A1/2. Finally, we treated the various mouse lines with 1-HMP (19.3 mg/kg, intraperitoneally), and then determined 1-SMP levels in plasma and DNA adducts in tissues. Transgenic human SULT1A1/2 strongly enhanced 1-SMP plasma levels and DNA adduct formation in the liver, lung, heart, and kidney but not in the colon. Sult1a1 and Sult1d1 knockout reduced plasma 1-SMP levels as well as DNA adduct formation in some tissues (strongest effects: 97% decrease in 1-SMP and 89% decrease in hepatic adducts in Sult1a1- mice). The adduct levels detected in various tissues did not accurately reflect the activation capacity of these tissues determined in vitro, probably due to the distribution of the reactive metabolite 1-SMP via the circulation. In conclusion, we demonstrated that many mouse SULT forms are able to activate 1-HMP. In vivo, we verified a prominent role of Sult1a1 in hepatic and renal adduct formation and a smaller but unambiguous role of Sult1d1, and demonstrated the strong impact of transgenic human SULT1A1/2.

Identification and quantification of protein adducts formed by metabolites of 1-methoxy-3-indolylmethyl glucosinolate in vitro and in mouse models.

1-Methoxy-3-indolylmethyl (1-MIM) glucosinolate (GLS) occurring in cabbage, broccoli, and other cruciferous plants is a potent mutagen requiring metabolic activation by myrosinase present in plant cells and intestinal bacteria. We previously reported that 1-MIM-GLS and its alcoholic breakdown product 1-MIM-OH, which requires additional activation by sulfotransferases, form DNA adducts in mice. In the present study, the formation of protein adducts was investigated. First, two major adducts obtained after incubation of individual amino acids, serum albumin, or hemoglobin with 1-MIM-GLS in the presence of myrosinase were identified as τN-(1-MIM)-His and πN-(1-MIM)-His using MS and NMR spectroscopy. After the development of a specific detection method using isotope-dilution UPLC-ESI-MS/MS, adduct formation was confirmed in mice after oral treatment with 1-MIM-GLS. Adduct levels were highest in the cecum and colon, somewhat lower in serum albumin and the liver, and also readily detectable in the lung and hemoglobin. On the contrary, oral treatment with 1-MIM-OH produced the highest adduct levels in the liver. The higher ratio of albumin to hemoglobin adducts in 1-MIM-OH- compared to 1-MIM-GLS-treated animals (8.1 versus 3.5) suggests that in 1-MIM-OH-treated animals albumin adducts were produced mostly in the liver, the site of albumin synthesis. The formation of adducts was approximately linear over a range of single oral doses from 20 to 600 µmol/kg body mass. Repeated oral administration of 1-MIM-OH (up to 40 treatments, thrice per week) led to continuous accumulation of hemoglobin adducts, whereas the level of serum albumin adducts remained rather constant, which reflects the different turnover rates of these proteins (t1/2 nearly 1.9 d for serum albumin and 25 d for hemoglobin in the mouse). Accumulation of adducts was also noticed in the lung. Adduct levels were higher, but their accumulation was weaker in the liver and kidney. The method developed will be useful to assess the exposure of humans to reactive metabolites formed from 1-MIM-GLS present in many foods.

Formation of hepatic DNA adducts by methyleugenol in mouse models: drastic decrease by Sult1a1 knockout and strong increase by transgenic human SULT1A1/2.

Methyleugenol--a natural constituent of herbs and spices--is hepatocarcinogenic in rodent models. It can form DNA adducts after side-chain hydroxylation and sulfation. We previously demonstrated that human sulfotransferases (SULTs) 1A1 and 1A2 as well as mouse Sult1a1, expressed in Salmonella target strains, are able to activate 1'-hydroxymethyleugenol (1'-OH-ME) and 3'-hydroxymethylisoeugenol (3'-OH-MIE) to mutagens. Now we investigated the role of these enzymes in the formation of hepatic DNA adducts by methyleugenol in the mouse in vivo. We used FVB/N mice [wild-type (wt)] and genetically modified strains in this background: Sult1a1 knockout (ko), transgenic for human SULT1A1/2 (tg) and the combination of both modifications (ko-tg). Methyleugenol (50mg/kg body mass) formed 23, 735, 3770 and 4500 N (2)-(trans-methylisoeugenol-3'-yl)-2'-deoxyguanosine adducts per 10(8) 2'-deoxyribonucleosides (dN) in ko, wt, ko-tg and tg mice, respectively. The corresponding values for an equimolar dose of 1'-OH-ME were 12, 1490, 12 400 and 13 300 per 10(8) dN. Similar relative levels were observed for the minor adduct, N (6)-(trans-methylisoeugenol-3'-yl)-2'-deoxyadenosine. Thus, the adduct formation by both compounds was nearly completely dependent on the presence of SULT1A enzymes, with human SULT1A1/2 producing stronger effects than mouse Sult1a1. Moreover, a dose of 0.05 mg/kg methyleugenol (one-fourth of the estimated average daily exposure of humans) was sufficient to form detectable adducts in humanized (ko-tg) mice. Although 3'-OH-MIE was equally mutagenic to 1'-OH-ME in Salmonella strains expressing human SULT1A1 or 1A2, it only formed 0.14% of hepatic adducts in ko-tg mice compared with an equimolar dose of 1'-OH-ME, suggesting an important role of detoxifying pathways for this isomer in vivo.

Mild gut inflammation modulates the proteome of intestinal Escherichia coli.

Using interleukin 10-deficient (IL-10(-/-) ) and wild-type mice monoassociated with either the adherent-invasive Escherichia coli UNC or the probiotic E. coli Nissle, the effect of a mild intestinal inflammation on the bacterial proteome was studied. Within 8 weeks, IL-10(-/-) mice monoassociated with E. coli UNC exhibited an increased expression of several proinflammatory markers in caecal mucosa. Escherichia coli Nissle-associated IL-10(-/-) mice did not do so. As observed previously for E. coli from mice with acute colitis, glycolytic enzymes were downregulated in intestinal E. coli UNC from IL-10(-/-) mice. In addition, the inhibitor of vertebrate C-type lysozyme, Ivy, was upregulated on messenger RNA (mRNA) and protein level in E. coli Nissle from IL-10(-/-) mice compared with E. coli UNC from these mice. Higher expression of Ivy in E. coli Nissle correlated with an improved growth of this probiotic strain in the presence of lysozyme-ethylenediaminetetraacetic acid (EDTA). By overexpressing Ivy, we demonstrated that Ivy contributes to a higher lysozyme resistance of E. coli, supporting the role of Ivy as a potential fitness factor. However, deletion of Ivy did not alter the growth phenotype of E. coli Nissle in the presence of lysozyme-EDTA, suggesting the existence of additional lysozyme inhibitors that can take over the function of Ivy.

A secondary metabolite of Brassicales, 1-methoxy-3-indolylmethyl glucosinolate, as well as its degradation product, 1-methoxy-3-indolylmethyl alcohol, forms DNA adducts in the mouse, but in varying tissues and cells.

1-Methoxy-3-indolylmethyl (1-MIM) glucosinolate, a secondary metabolite of Brassicales species, and its breakdown product 1-MIM alcohol are mutagenic in cells in culture after activation by plant ß-thioglucosidase and human sulphotransferase, respectively. In the present study, we administered these compounds orally to mice to study time course, dose dependence, tissue distribution and cellular localization of the 1-MIM DNA adducts formed. We used isotope-dilution ultra-performance liquid chromatography-tandem mass spectrometry to quantify the adducts and raised an antiserum for their immunohistochemical localization. Both compounds formed the same adducts, N(2)-(1-MIM)-2'-deoxyguanosine and N(6)-(1-MIM)-2'-deoxyadenosine, approximately in a 3.3:1 ratio. 1-MIM glucosinolate primarily formed these adducts in the large intestine, with a luminal-basal gradient, probably due to activation by thioglucosidase from intestinal bacteria. 1-MIM alcohol formed higher levels of adduct than the glucosinolate. Unlike after treatment with the glucosinolate, luminal and basal enterocytes were similarly affected in caecum, and liver and stomach were additional important target tissues. Maximal adduct levels were reached 8 h after the administration of both compounds. The hepatic DNA adducts persisted for the entire observation period (48 h), whereas those in large intestine rapidly declined due to cell turnover, as verified by immunohistochemistry. Hepatic adduct formation was focused on the periportal hepatocytes with concomitant depletion of glycogen, p53 activation and p21 induction. Adduct formation in caecum was associated with massive apoptosis, p53 activation and p21 induction, in particular after treatment with 1-MIM alcohol. It remains to be studied whether similar effects occur in humans after the consumption of Brassicales species.

Identification and expression of genes involved in the conversion of daidzein and genistein by the equol-forming bacterium Slackia isoflavoniconvertens.

Gut bacteria play a key role in the metabolism of dietary isoflavones, thereby influencing the availability and bioactivation of these polyphenols in the intestine. The human intestinal bacterium Slackia isoflavoniconvertens converts the main soybean isoflavones daidzein and genistein to equol and 5-hydroxy-equol, respectively. Cell extracts of S. isoflavoniconvertens catalyzed the conversion of daidzein via dihydrodaidzein to equol and that of genistein to dihydrogenistein. Growth of S. isoflavoniconvertens in the presence of daidzein led to the induction of several proteins as observed by two-dimensional difference gel electrophoresis. Based on determined peptide sequences, we identified a cluster of eight genes encoding the daidzein-induced proteins. Heterologous expression of three of these genes in Escherichia coli and enzyme activity tests with the resulting cell extracts identified the corresponding gene products as a daidzein reductase (DZNR), a dihydrodaidzein reductase (DHDR), and a tetrahydrodaidzein reductase (THDR). The recombinant DZNR also converted genistein to dihydrogenistein at higher rates than were observed for the conversion of daidzein to dihydrodaidzein. Higher rates were also observed with cell extracts of S. isoflavoniconvertens. The recombinant DHDR and THDR catalyzed the reduction of dihydrodaidzein to equol, while the corresponding conversion of dihydrogenistein to 5-hydroxy-equol was not observed. The DZNR, DHDR, and THDR were expressed as Strep-tag fusion proteins and subsequently purified by affinity chromatography. The purified enzymes were further characterized with regard to their activity, stereochemistry, quaternary structure, and content of flavin cofactors.

Abundance of DNA adducts of methyleugenol, a rodent hepatocarcinogen, in human liver samples.

Methyleugenol is a genotoxic carcinogen in mice and rats, the liver being the primary target tissue. Methyleugenol occurs in fennel and many herbs and spices. Furthermore, methyleugenol-containing plant extracts and chemically prepared methyleugenol are used as flavoring agents. We analyzed surgical human liver samples from 30 subjects for the presence of DNA adducts originating from methyleugenol using isotope-dilution ultra-performance liquid chromatography-tandem mass spectrometry (UPLC-MS/MS). Twenty-nine samples unambiguously contained the N (2)-(trans-methylisoeugenol-3'-yl)-2'-deoxyguanosine adduct. A second adduct, N (6)-(trans-methylisoeugenol-3'-yl)-2'-deoxyadenosine, was also found in most samples, but at much lower levels, in agreement with the results from experimental models. The maximal and median levels of both adducts combined were 37 and 13 per 10(8) nucleosides (corresponding to 4700 and 1700, respectively, adducts per diploid genome). This is the first demonstration of DNA adducts formed by a xenobiotic in human liver using UPLC-MS/MS, the most reliable method available. It has been estimated for diverse rat and mouse hepatocarcinogens that 50-5500 adducts per 10(8) nucleosides are present after repeated treatment at the TD50 (daily dose that halves the probability to stay tumor-free in long-term studies). We conclude that the exposure to methyleugenol leads to substantial levels of hepatic DNA adducts and, therefore, may pose a significant carcinogenic risk.

Contribution of gut bacteria to the metabolism of cyanidin 3-glucoside in human microbiota-associated rats.

Cyanidin 3-glucoside (C3G) is one of the major dietary anthocyanins implicated in the prevention of chronic diseases. To evaluate the impact of human intestinal bacteria on the fate of C3G in the host, we studied the metabolism of C3G in human microbiota-associated (HMA) rats in comparison with germ-free (GF) rats. Urine and faeces of the rats were analysed for C3G and its metabolites within 48 h after the application of 92 µmol C3G/kg body weight. In addition, we tested the microbial C3G conversion in vitro by incubating C3G with human faecal slurries and selected human gut bacteria. The HMA rats excreted with faeces a three times higher percentage of unconjugated C3G products and a two times higher percentage of conjugated C3G products than the GF rats. These differences were mainly due to the increased excretion of 3,4-dihydroxybenzoic acid, 2,4,6-trihydroxybenzaldehyde and 2,4,6-trihydroxybenzoic acid. Only the urine of HMA rats contained peonidin and 3-hydroxycinnamic acid and the percentage of conjugated C3G products in the urine was decreased compared with the GF rats. Overall, the presence of intestinal microbiota resulted in a 3·7% recovery of the C3G dose in HMA rats compared with 1·7% in GF rats. Human intestinal bacteria rapidly degraded C3G in vitro. Most of the C3G products were also found in the absence of bacteria, but at considerably lower levels. The higher concentrations of phenolic acids observed in the presence of intestinal bacteria may contribute to the proposed beneficial health effects of C3G.

Optimized enzymatic hydrolysis of DNA for LC-MS/MS analyses of adducts of 1-methoxy-3-indolylmethyl glucosinolate and methyleugenol.

Mass spectrometric analyses of DNA adducts usually require enzymatic digestion of the DNA to nucleosides. The digestive enzymes used in our laboratory included a calf spleen phosphodiesterase, whose marketing was stopped recently. Using DNA adducted with bioactivated methyleugenol and 1-methoxy-3-indolylmethyl glucosinolate-each forming dA and dG adducts-we demonstrate that replacement of calf spleen phosphodiesterase (Merck) with bovine spleen phosphodiesterase (Sigma-Aldrich) leads to unchanged results. Enzyme levels used for DNA digestion are extremely variable in different studies. Therefore, we sequentially varied the level of each of the three enzymes used. All dose (enzyme)-response (adduct level) curves involved a long plateau starting below the enzyme levels employed previously. Thus, we could reduce the amounts of micrococcal nuclease, phosphodiesterase, and alkaline phosphatase for quantitative DNA digestion by factors of 4, 2, and 333, respectively, compared to our previous protocols. Moreover, we observed significant phosphatase activity of both phosphodiesterase preparations used, which may affect the recovery of adducts with methods requiring digestion to 2'-deoxynucleoside-3'-monophosphates (e.g., (32)P-postlabeling). In addition, the phosphodiesterase from Sigma-Aldrich, but not that from Merck, deaminated dA. This was irrelevant for the dA adducts studied, involving bonding at N(6), but might complicate the analysis of other dA adducts.

Detection of DNA adducts originating from 1-methoxy-3-indolylmethyl glucosinolate using isotope-dilution UPLC-ESI-MS/MS.

1-Methoxy-3-indolylmethyl (1-MIM) glucosinolate, present at substantial levels in several food crops (e.g., broccoli and cabbage), forms DNA adducts in vitro and is mutagenic to bacterial and mammalian cells after activation by the plant enzyme myrosinase. Moreover, a breakdown product, 1-MIM alcohol, is metabolized to a secondary reactive intermediate by some mammalian sulfotransferases (SULTs). First, we incubated herring-sperm DNA with 1-MIM glucosinolate in the presence of myrosinase. We identified and synthesized the predominant adducts, N(2)-(1-MIM)-dG and N(6)-(1-MIM)-dA, and developed an UPLC-ESI-MS/MS method for their specific detection using isotopic dilution. Second, we demonstrated both DNA adducts in target cells (Salmonella typhimurium TA100 and Chinese hamster V79) of standard mutagenicity tests treated with 1-MIM glucosinolate/myrosinase as well as in 1-MIM alcohol-treated Salmonella and V79 cells engineered for expression of human SULT1A1. Similar excesses of N(2)-(1-MIM)-dG over N(6)-(1-MIM)-dA adducts were found in all cellular models independent of the test compound (1-MIM glucosinolate or alcohol), whereas dA adducts predominated in the cell-free system. Finally, we detected both DNA adducts in colon tissue of a mouse orally treated with 1-MIM glucosinolate. We are going to use this specific and sensitive method for investigating genotoxic risks of food-borne exposure to 1-MIM glucosinolate in animal and human studies.

Metabolism of methyleugenol in liver microsomes and primary hepatocytes: pattern of metabolites, cytotoxicity, and DNA-adduct formation.

Methyleugenol (1) is a constituent of many foods, in particular of herbal spices, and is used as flavoring agent in foodstuffs and as fragrance in cosmetics. 1 has been found to be carcinogenic in rodents, its metabolite, 1-hydroxymethyleugenol (2) acting as proximate DNA-binding carcinogen. We incubated 1 with liver microsomes of rat, bovine, and human origin. We found 2, 3-hydroxymethylisoeugenol (3), and 6-hydroxymethyleugenol (4) as major metabolites, and 1-oxomethyleugenol (5), 3-oxomethylisoeugenol (6), eugenol (9), chavibetol (11), and (RS)-2,3-dihydroxy-2,3-dihydromethyleugenol (7) as minor metabolites. Methyleugenol-2,3-epoxide (8), probably the precursor of 7, could not be detected. Incubations with synthetic metabolites were applied in order to uncover metabolic pathways. Incubations with primary rat hepatocytes revealed mainly nonconjugated 2 and conjugated 4, and minor amounts of partly conjugated 7 and conjugated 9 + 11. The "reactive metabolites" 3, 5, 6, and 8 were not detectable, possibly due to rapid reaction with cellular macromolecules. The highest cytotoxicity (resazurin reduction assay and lactate dehydrogenase leakage assay) was observed for the main metabolite 2 and its secondary metabolite 5 with EC(50) values of 50 and 10 µM, respectively. Deoxyadenosine or deoxyguanosine adducts were formed by incubating 1 or metabolites with rat hepatocytes. The rank order of adduct formation was 2 > 1 > 3 > 6, whereas 4, 5, and 8 were inactive. In conclusion, we present a virtually complete pattern of microsomal (rat, bovine, and human) and hepatocellular (rat) metabolites of 1 suggesting the formation of several reactive metabolites possibly involved in carcinogenicity, organ toxicity, and immune reactions.

Mutagenicity of 5-hydroxymethylfurfural in V79 cells expressing human SULT1A1: identification and mass spectrometric quantification of DNA adducts formed.

5-Hydroxymethylfurfural (HMF), a heterocyclic product of the Maillard reaction, is a ubiquitous food contaminant. It has demonstrated hepatocarcinogenic activity in female mice. This effect may originate from sulfo conjugation of the benzylic alcohol yielding 5-sulfooxymethylfurfural (SMF), which is prone to react with DNA via nucleophilic substitution. Indeed, we showed that HMF induces gene mutations in Chinese hamster V79 cells engineered for the expression of human (h) sulfotransferase (SULT)1A1 but not in parental V79 cells. In order to identify potential DNA adducts, we incubated DNA samples with SMF or HMF in aqueous solution. Modified DNA was digested and surveyed by liquid chromatography coupled with tandem mass spectrometry (LC-MS/MS) for adducts that may be formed by nucleosides either via nucleophilic substitution at the electrophilic carbon atom of SMF or via imine formation with the aldehyde group present in HMF and SMF. The most abundant adducts formed from SMF, N(6)-((2-formylfuran-5-yl)methyl)-2'-deoxyadenosine (N(6)-FFM-dAdo) and N(2)-((2-formylfuran-5-yl)methyl)-2'-deoxyguanosine (N(2)-FFM-dGuo), were synthesized, purified, and characterized by (1)H NMR. Imine adducts were only detected when DNA was incubated with very high levels of HMF following reduction of the imines to corresponding secondary amines by NaBH(3)CN. Sensitive techniques based on LC-MS/MS multiple reaction monitoring for the quantification of the adducts in DNA samples were devised using isotope-labeled [(15)N(5)]N(6)-FFM-dAdo and [(13)C(10),(15)N(5)]N(2)-FFM-dGuo as internal standards. Both 5-methylfurfuryl adducts were detected in DNA from V79-hSULT1A1 treated with HMF but not in DNA from V79 control cells. Considering the lack of other known mutagenic metabolites, we hypothesize that the hepatocarcinogenic potential of HMF originates from the formation of mutagenic SMF.

Impact of nutritional factors on the proteome of intestinal Escherichia coli: induction of OxyR-dependent proteins AhpF and Dps by a lactose-rich diet.

To study the impact of nutritional factors on protein expression of intestinal bacteria, gnotobiotic mice monoassociated with Escherichia coli K-12 were fed three different diets: a diet rich in starch, a diet rich in nondigestible lactose, and a diet rich in casein. Two-dimensional gel electrophoresis and electrospray-tandem mass spectrometry were used to identify differentially expressed proteins of bacteria recovered from small intestine and cecum. Oxidative stress response proteins such as AhpF, Dps, and Fur, all of which belong to the oxyR regulon, were upregulated in E. coli isolates from mice fed the lactose-rich diet. Luciferase reporter gene assays demonstrated that osmotic stress caused by carbohydrates led to the expression of ahpCF and dps, which was not observed in an E. coli ΔoxyR mutant. Growth of ahpCF and oxyR deletion mutants was strongly impaired when nondigestible sucrose was present in the medium. The wild-type phenotype could be restored by complementation of the deletions with plasmids containing the corresponding genes and promoters. The results indicate that some OxyR-dependent proteins play a major role in the adaptation of E. coli to osmotic stress. We conclude that there is an overlap of osmotic and oxidative stress responses. Mice fed the lactose-rich diet possibly had a higher intestinal osmolality, leading to the upregulation of OxyR-dependent proteins, which enable intestinal E. coli to better cope with diet-induced osmotic stress.

Identification of human and murine sulfotransferases able to activate hydroxylated metabolites of methyleugenol to mutagens in Salmonella typhimurium and detection of associated DNA adducts using UPLC-MS/MS methods.

Methyleugenol, a secondary metabolite present in many herbal spices, is carcinogenic in various tissues of mice and rats but negative in standard in vitro mutagenicity tests. Several observations indicate that hydroxylation followed by sulfation is an important activation pathway in the carcinogenicity and DNA adduct formation by methyleugenol and other alkenylbenzenes in animal models. However, sulfation is not taken into account in standard in vitro tests. Therefore, we have studied whether expression of murine or human sulfotransferases (SULTs) in the target strain, Salmonella typhimurium TA100, leads to the activation of hydroxylated metabolites of methyleugenol [(+)-1'-hydroxymethyleugenol, (-)-1'-hydroxymethyleugenol and (E)-3'-hydroxymethylisoeugenol]. Human SULT1A1 (a form expressed at high levels in many tissues) and SULT1C2 (expressed primarily in foetal tissues) activated all three compounds even at very low substrate concentrations. At higher concentrations, activation was also observed with human SULT1A2 and SULT1E1. Murine Sult1a1 required higher substrate concentrations than its human orthologue. Other SULT forms (human 1A3, 1C1, 1C3, 2A1 and 2B1b as well as murine 1d1) did not activate any methyleugenol metabolites studied. Furthermore, we developed isotope-dilution mass-spectrometric methods for the sensitive and specific detection of DNA adducts formed by methyleugenol metabolites. All three hydroxylated metabolites formed the same DNA adducts in S. typhimurium TA100-hSULT1A1: high levels of N (2)-(trans-methylisoeugenol-3'-yl)-2'-deoxyguanosine and modest levels of N (6)-(trans-methylisoeugenol-3'-yl)-2'-deoxyadenosine. Adduct levels correlated with the mutagenic effects induced. No adducts were formed by the test compounds in the SULT-deficient standard strain TA100. In conclusion, several methyleugenol metabolites are activated to DNA-reactive mutagens in S. typhimurium upon incorporation of appropriate sulfation capacity. We have identified human and murine SULT forms able to catalyse this activation. Methods were developed that may be utilised to analyse DNA samples from human tissues specifically for the possible presence of methyleugenol adducts.

Dextran sodium sulfate-induced inflammation alters the expression of proteins by intestinal Escherichia coli strains in a gnotobiotic mouse model.

To identify Escherichia coli proteins involved in adaptation to intestinal inflammation, mice were monoassociated with the colitogenic E. coli strain UNC or with the probiotic E. coli strain Nissle. Intestinal inflammation was induced by treating the mice with 3.5% dextran sodium sulfate (DSS). Differentially expressed proteins in E. coli strains collected from cecal contents were identified by 2-dimensional difference gel electrophoresis. In both strains, acute inflammation led to the downregulation of pathways involved in carbohydrate breakdown and energy generation. Accordingly, DSS-treated mice had lower concentrations of bacterial fermentation products in their cecal contents than control mice. Differentially expressed proteins also included the Fe-S cluster repair protein NfuA, the tryptophanase TnaA, and the uncharacterized protein YggE. NfuA expression was 3-fold higher in E. coli strains from DSS-treated than from control mice. Reporter experiments confirmed the induction of nfuA in response to iron deprivation, mimicking Fe-S cluster destruction by inflammation. YggE expression, which has been reported to reduce the intracellular level of reactive oxygen species, was 4- to 8-fold higher in E. coli Nissle than in E. coli UNC. This was confirmed by in vitro reporter gene assays indicating that Nissle is better equipped to cope with oxidative stress than UNC. Nissle isolated from DSS-treated and control mice had TnaA levels 4- to 7-fold-higher than those of UNC. Levels of indole resulting from the TnaA reaction were higher in control animals associated with E. coli Nissle. Because of its anti-inflammatory effect, indole is hypothesized to be involved in the extension of the remission phase in ulcerative colitis described for E. coli Nissle.