Correlation between activation of PPARγ and resistin downregulation in a mouse adipocyte cell line by a series of thiazolidinediones.

The present study shows significant correlations between the EC50 for PPARγ activation in a reporter gene cell line and resistin downregulation in mouse adipocytes, and between the IC50 for resistin downregulation and the already published minimum effective dose for antihyperglycemic activity in a mouse model. These correlations indicate that PPARγ mediated downregulation of resistin might promote insulin sensitivity and that downregulation of resistin in mouse adipocytes provides an adequate and possibly more direct bioassay for screening of newly developed antihyperglycemic compounds. Because of the higher throughput of the PPARγ the resistin downregulation assays seems most suitable to be used as a second tier in a tiered screening strategy.

Physiologically based biokinetic (PBBK) model for safrole bioactivation and detoxification in rats.

A physiologically based biokinetic (PBBK) model for alkenylbenzene safrole in rats was developed using in vitro metabolic parameters determined using relevant tissue fractions. The performance of the model was evaluated by comparison of the predicted levels of 1,2-dihydroxy-4-allylbenzene and 1'-hydroxysafrole glucuronide to levels of these metabolites reported in the literature to be excreted in the urine of rats exposed to safrole and by comparison of the predicted amount of total urinary safrole metabolites to the reported levels of safrole metabolites in the urine of safrole exposed rats. These comparisons revealed that the predictions adequately match observed experimental values. Next, the model was used to predict the relative extent of bioactivation and detoxification of safrole at different oral doses. At low as well as high doses, P450 mediated oxidation of safrole mainly occurs in the liver in which 1,2-dihydroxy-4-allylbenzene was predicted to be the major P450 metabolite of safrole. A dose dependent shift in P450 mediated oxidation leading to a relative increase in bioactivation at high doses was not observed. Comparison of the results obtained for safrole with the results previously obtained with PBBK models for the related alkenylbenzenes estragole and methyleugenol revealed that the overall differences in bioactivation of the three alkenylbenzenes to their ultimate carcinogenic 1'-sulfooxy metabolites are limited. This is in line with the generally less than 4-fold difference in their level of DNA binding in in vitro and in vivo studies and their almost similar BMDL(10) values (lower confidence limit of the benchmark dose that gives 10% increase in tumor incidence over background level) obtained in in vivo carcinogenicity studies. It is concluded that in spite of differences in the rates of specific metabolic conversions, overall the levels of bioactivation of the three alkenylbenzenes are comparable which is in line with their comparable carcinogenic potential.

Superinduction of estrogen receptor mediated gene expression in luciferase based reporter gene assays is mediated by a post-transcriptional mechanism.

Several estrogenic compounds including the isoflavonoid genistein have been reported to induce a higher maximal response than the natural estrogen 17ß-estradiol in in vitro luciferase based reporter gene bioassays for testing estrogenicity. The phenomenon has been referred to as superinduction. The mechanism underlying this effect and thus also its biological relevance remain to be elucidated. In the present study several hypotheses for the possible mechanisms underlying this superinduction were investigated using genistein as the model compound. These hypotheses included (i) a non-estrogen receptor (ER)-mediated mechanism, (ii) a role for an ER activating genistein metabolite with higher ER inducing activity than genistein itself, and (iii) a post-transcriptional mechanism that is not biologically relevant but specific for the luciferase based reporter gene assays. The data presented in this study indicate that induction and also superinduction of the reporter gene is ER-mediated, and that superinduction by genistein could be ascribed to stabilization of the firefly luciferase reporter enzyme increasing the bioluminescent signal during the cell-based assay. This indicates that the phenomenon of superinduction may not be biologically relevant but may rather represent a post-transcriptional effect on enzyme stability.

Identification of nevadensin as an important herb-based constituent inhibiting estragole bioactivation and physiology-based biokinetic modeling of its possible in vivo effect.

Estragole is a natural constituent of several herbs and spices including sweet basil. In rodent bioassays, estragole induces hepatomas, an effect ascribed to estragole bioactivation to 1'-sulfooxyestragole resulting in DNA adduct formation. The present paper identifies nevadensin as a basil constituent able to inhibit DNA adduct formation in rat hepatocytes exposed to the proximate carcinogen 1'-hydroxyestragole and nevadensin. This inhibition occurs at the level of sulfotransferase (SULT)-mediated bioactivation of 1'-hydroxyestragole. The Ki for SULT inhibition by nevadensin was 4 nM in male rat and human liver fractions. Furthermore, nevadensin up to 20 microM did not inhibit 1'-hydroxyestragole detoxification by glucuronidation and oxidation. The inhibition of SULT by nevadensin was incorporated into the recently developed physiologically based biokinetic (PBBK) rat and human models for estragole bioactivation and detoxification. The results predict that co-administration of estragole at a level inducing hepatic tumors in vivo (50mg/kg bw) with nevadensin at a molar ratio of 0.06, representing the ratio of their occurrence in basil, results in almost 100% inhibition of the ultimate carcinogen 1'-sulfooxyestragole when assuming 100% uptake of nevadensin. Assuming 1% uptake, inhibition would still amount to more than 83%. Altogether these data point at a nevadensin-mediated inhibition of the formation of the ultimate carcinogenic metabolite of estragole, without reducing the capacity to detoxify 1'-hydroxyestragole via glucuronidation or oxidation. These data also point at a potential reduction of the cancer risk when estragole exposure occurs within a food matrix containing SULT inhibitors compared to what is observed upon exposure to pure estragole.

Phytoestrogen-mediated inhibition of proliferation of the human T47D breast cancer cells depends on the ERalpha/ERbeta ratio.

This study investigates the importance of the intracellular ratio of the two estrogen receptors ERalpha and ERbeta for the ultimate potential of the phytoestrogens genistein and quercetin to stimulate or inhibit cancer cell proliferation. This is of importance because (i) ERbeta has been postulated to play a role in modulating ERalpha-mediated cell proliferation, (ii) genistein and quercetin may be agonists for both receptor types and (iii) the ratio of ERalpha to ERbeta is known to vary between tissues. Using human osteosarcoma (U2OS) ERalpha or ERbeta reporter cells it was shown that compared to estradiol (E2), genistein and quercetin have not only a relatively greater preference for ERbeta but also a higher maximal potential for activating ERbeta-mediated gene expression. Using the human T47D breast cancer cell line with tetracycline-dependent ERbeta expression (T47D-ERbeta), the effect of a varying intracellular ERalpha/ERbeta ratio on E2- or pythoestrogen-induced cell proliferation was characterised. E2-induced proliferation of cells in which ERbeta expression was inhibited was similar to that of the T47D wild type cells, whereas this E2-induced cell proliferation was no longer observed when ERbeta expression was increased. With increased expression of ERbeta the phytoestrogen-induced cell proliferation was also reduced. These results point at the importance of the cellular ERalpha/ERbeta ratio for the ultimate effect of (phyto)estrogens on cell proliferation.

Substrate interactions during the biodegradation of benzene, toluene, ethylbenzene, and xylene (BTEX) hydrocarbons by the fungus Cladophialophora sp. strain T1.

The soil fungus Cladophialophora sp. strain T1 (= ATCC MYA-2335) was capable of growth on a model water-soluble fraction of gasoline that contained all six BTEX components (benzene, toluene, ethylbenzene, and the xylene isomers). Benzene was not metabolized, but the alkylated benzenes (toluene, ethylbenzene, and xylenes) were degraded by a combination of assimilation and cometabolism. Toluene and ethylbenzene were used as sources of carbon and energy, whereas the xylenes were cometabolized to different extents. o-Xylene and m-xylene were converted to phthalates as end metabolites; p-xylene was not degraded in complex BTEX mixtures but, in combination with toluene, appeared to be mineralized. The metabolic profiles and the inhibitory nature of the substrate interactions indicated that toluene, ethylbenzene, and xylene were degraded at the side chain by the same monooxygenase enzyme. Our findings suggest that soil fungi could contribute significantly to bioremediation of BTEX pollution.

No evidence for binding between resistance gene product Cf-9 of tomato and avirulence gene product AVR9 of Cladosporium fulvum.

The gene-for-gene model postulates that for every gene determining resistance in the host plant, there is a corresponding gene conditioning avirulence in the pathogen. On the basis of this relationship, products of resistance (R) genes and matching avirulence (Avr) genes are predicted to interact. Here, we report on binding studies between the R gene product Cf-9 of tomato and the Avr gene product AVR9 of the pathogenic fungus Cladosporium fulvum. Because a high-affinity binding site (HABS) for AVR9 is present in tomato lines, with or without the Cf-9 resistance gene, as well as in other solanaceous plants, the Cf-9 protein was produced in COS and insect cells in order to perform binding studies in the absence of the HABS. Binding studies with radio-labeled AVR9 were performed with Cf-9-producing COS and insect cells and with membrane preparations of such cells. Furthermore, the Cf-9 gene was introduced in tobacco, which is known to be able to produce a functional Cf-9 protein. Binding of AVR9 to Cf-9 protein produced in tobacco was studied employing surface plasmon resonance and surface-enhanced laser desorption and ionization. Specific binding between Cf-9 and AVR9 was not detected with any of the procedures. The implications of this observation are discussed.

Disulfide bond structure of the AVR9 elicitor of the fungal tomato pathogen Cladosporium fulvum: evidence for a cystine knot.

Disease resistance in plants is commonly activated by the product of an avirulence (Avr) gene of a pathogen after interaction with the product of a matching resistance (R) gene in the host. In susceptible plants, Avr products might function as virulence or pathogenicity factors. The AVR9 elicitor from the fungus Cladosporium fulvum induces defense responses in tomato plants carrying the Cf-9 resistance gene. This 28-residue beta-sheet AVR9 peptide contains three disulfide bridges, which were identified in this study as Cys2-Cys16, Cys6-Cys19, and Cys12-Cys26. For this purpose, AVR9 was partially reduced, and the thiol groups of newly formed cysteines were modified to prevent reactions with disulfides. After HPLC purification, the partially reduced peptides were sequenced to determine the positions of the modified cysteines, which originated from the reduced disulfide bridge(s). All steps involving molecules with free thiol groups were performed at low pH to suppress disulfide scrambling. For that reason, cysteine modification by N-ethylmaleimide was preferred over modification by iodoacetamide. Upon (partial) reduction of native AVR9, the Cys2-Cys16 bridge opened selectively. The resulting molecule was further reduced to two one-bridge intermediates, which were subsequently completely reduced. The (partially) reduced cysteine-modified AVR9 species showed little or no necrosis-inducing activity, demonstrating the importance of the disulfide bridges for biological activity. Based on peptide length and cysteine spacing, it was previously suggested that AVR9 isa cystine-knotted peptide. Now, we have proven that the bridging pattern of AVR9 is indeed identical to that of cystine-knotted peptides. Moreover, NMR data obtained for AVR9 show that it is structurally closely related to the cystine-knotted carboxypeptidase inhibitor. However, AVR9 does not show any carboxypeptidase inhibiting activity, indicating that the cystine-knot fold is a commonly occurring motif with varying biological functions.

Structure-activity study on the quinone/quinone methide chemistry of flavonoids.

A structure-activity study on the quinone/quinone methide chemistry of a series of 3',4'-dihydroxyflavonoids was performed. Using the glutathione trapping method followed by HPLC, (1)H NMR, MALDI-TOF, and LC/MS analysis to identify the glutathionyl adducts, the chemical behavior of the quinones/quinone methides of the different flavonoids could be deduced. The nature and type of mono- and diglutathionyl adducts formed from quercetin, taxifolin, luteolin, fisetin, and 3,3',4'-trihydroxyflavone show how several structural elements influence the quinone/quinone methide chemistry of flavonoids. In line with previous findings, glutathionyl adduct formation for quercetin occurs at positions C6 and C8 of the A ring, due to the involvement of quinone methide-type intermediates. Elimination of the possibilities for efficient quinone methide formation by (i) the absence of the C3-OH group (luteolin), (ii) the absence of the C2=C3 double bond (taxifolin), or (iii) the absence of the C5-OH group (3,3',4'-trihydroxyflavone) results in glutathionyl adduct formation at the B ring due to involvement of the o-quinone isomer of the oxidized flavonoid. The extent of di- versus monoglutathionyl adduct formation was shown to depend on the ease of oxidation of the monoadduct as compared to the parent flavonoid. Finally, unexpected results obtained with fisetin provide new insight into the quinone/quinone methide chemistry of flavonoids. The regioselectivity and nature of the quinone adducts that formed appear to be dependent on pH. At pH values above the pK(a) for quinone protonation, glutathionyl adduct formation proceeds at the A or B ring following expected quinone/quinone methide isomerization patterns. However, decreasing the pH below this pK(a) results in a competing pathway in which glutathionyl adduct formation occurs in the C ring of the flavonoid, which is preceded by protonation of the quinone and accompanied by H(2)O adduct formation, also in the C ring of the flavonoid. All together, the data presented in this study confirm that quinone/quinone methide chemistry can be far from straightforward, but the study provides significant new data revealing an important pH dependence for the chemical behavior of this important class of electrophiles.

Peroxidase-catalyzed formation of quercetin quinone methide-glutathione adducts.

The oxidation of quercetin by horseradish peroxidase/H(2)O(2) was studied in the absence but especially also in the presence of glutathione (GSH). HPLC analysis of the reaction products formed in the absence of GSH revealed formation of at least 20 different products, a result in line with other studies reporting the peroxidase-mediated oxidation of flavonoids. In the presence of GSH, however, these products were no longer observed and formation of two major new products was detected. (1)H NMR identified these two products as 6-glutathionylquercetin and 8-glutathionylquercetin, representing glutathione adducts originating from glutathione conjugation at the A ring instead of at the B ring of quercetin. Glutathione addition at positions 6 and 8 of the A ring can best be explained by taking into consideration a further oxidation of the quercetin semiquinone, initially formed by the HRP-mediated one-electron oxidation, to give the o-quinone, followed by the isomerization of the o-quinone to its p-quinone methide isomer. All together, the results of the present study provide evidence for a reaction chemistry of quercetin semiquinones with horseradish peroxidase/H(2)O(2) and GSH ultimately leading to adduct formation instead of to preferential GSH-mediated chemical reduction to regenerate the parent flavonoid.

Regioselectivity and reversibility of the glutathione conjugation of quercetin quinone methide.

The chemical reactivity, isomerization, and glutathione conjugation of quercetin o-quinone were investigated. Tyrosinase was used to generate the unstable quercetin o-quinone derivative which could be observed upon its subsequent scavenging by glutathione. Identification of the products revealed formation of 6-glutathionyl-quercetin and 8-glutathionyl-quercetin adducts. Thus, in particular, glutathione adducts in the A ring of quercetin were formed, a result which was not expected a priori. Quantum mechanical calculations support the possibility that the formation of these glutathione adducts can be explained by an isomerization of quercetin o-quinone to p-quinone methides. Surprisingly, additional experiments of this study reveal the adduct formation to be reversible, leading to interconversion between the two quercetin glutathione adducts and possibilities for release and further electrophilic reactions of the quercetin quinone methide at cellular sites different from those of its generation.

Folding and conformational analysis of AVR9 peptide elicitors of the fungal tomato pathogen Cladosporium fulvum.

The race-specific elicitor AVR9, produced by the phytopathogenic fungus Cladosporium fulvum, is a 28-residue beta-sheet peptide containing three disulfide bridges. The folding of this peptide to its native conformation was examined in the presence of oxidized (GSSG) and reduced (GSH) glutathione at concentrations resembling those present in the endoplasmic reticulum. The concentrations of GSH and GSSG, and the applied temperature strongly affected the folding efficiency. The effect of temperature appeared reversible. The conditions for in vitro folding were optimized and a maximum yield of 60-70% of correctly folded peptide was obtained. In vitro folded AVR9 is equally as active as native fungal AVR9. They both display similar NMR characteristics, indicating that they have the same 3D structure and identical disulfide bridges. Thus, AVR9 can be folded correctly in vitro. This folding can be described by disulfide bridge formation leading to scrambled three-disulfide species, followed by disulfide reshuffling to acquire the native structure. The presence of urea significantly affected the folding of AVR9, indicating that noncovalent interactions play a role in directing correct folding. Protein disulfide isomerase increased the folding rate at least 15-fold, but had no effect on the yield. The folding procedure has also been applied successfully to two mutant AVR9 peptides, (K23A)AVR9 and biotinylated AVR9. We conclude that the 28-residue sequence, without the preprosequence (as present in vivo), contains sufficient information to direct correct folding and disulfide bridge formation in vitro.

Occurrence of the NIH shift upon the cytochrome P450-catalyzed in vivo and in vitro aromatic ring hydroxylation of fluorobenzenes.

The in vivo cytochrome P450-catalyzed aromatic hydroxylation of a series of fluorobenzenes was investigated with special emphasis on the importance of the fluorine NIH shift. The results obtained demonstrate a minor role for the NIH shift in the metabolism of the fluorobenzenes to phenolic metabolites in control male Wistar rats. These in vivo results could indicate that (1) the NIH shift is an inherently minor process for fluorine substituents or (2) it is a potentially significant process but the presumed epoxide that leads to formation of the NIH-shifted metabolite is lost to an alternative metabolic pathway. In contrast to the in vivo data, in vitro experiments showed a significant amount of an NIH-shifted metabolite for 1,4-difluorobenzene. This result eliminates the explanation that the NIH shift is an inherently minor process for fluorine substituents. Results of additional experiments presented in this paper show that the reduced tendency of fluorine-substituted benzenes to undergo an NIH shift in vivo can-at least in part-be ascribed to the possible existence of alternative pathways for metabolism of the epoxide, such as, for example, GSH conjugation, being more efficient for fluorinated than chlorinated arene oxides.

GSTP1-1 stereospecifically catalyzes glutathione conjugation of ethacrynic acid.

Using 1H NMR two diastereoisomers of the ethacrynic acid glutathione conjugate (EASG) as well as ethacrynic acid (EA) could be distinguished and quantified individually. Chemically prepared EASG consists of equal amounts of both diastereoisomers. GSTP1-1 stereospecifically catalyzes formation of one of the diastereoisomers (A). The GSTP1-1 mutant C47S and GSTA1-1 preferentially form the same diastereoisomer of EASG as GSTP1-1. Glutathione conjugation of EA by GSTA1-2 and GSTA2-2 is not stereoselective. When human melanoma cells, expressing GSTP1-1, were exposed to ethacrynic acid, diastereoisomer A was the principal conjugate formed, indicating that even at physiological pH the enzyme catalyzed reaction dominates over the chemical conjugation.

13C NMR analysis of the cysteine-sulfenic acid redox center of enterococcal NADH peroxidase.

In order to characterize the native Cys42-sulfenic acid redox center of the flavoprotein NADH peroxidase by NMR, an expression protocol has been developed which yields the [3-13C]Cys42-labeled protein in 100 mg quantities. Difference spectra of the labeled minus unlabeled oxidized enzyme (E) give a peak at 41.3 ppm (relative to dioxane) which represents the Cys42-sulfenic acid. Reduction of labeled E with 1 equiv of NADH gives the air-stable two-electron reduced (EH2) species, and oxidized minus reduced difference spectra give maxima and minima at 41.3 and 30.8 ppm, respectively, corresponding to the Cys42-sulfenic acid and -thiolate species. Peroxide inactivation of E, which has previously been attributed to oxidation of the Cys42-sulfenic acid to the Cys42-sulfinic and/or sulfonic acid states, gives rise to a new maximum in the difference spectrum of Einactive minus E at 57.0 ppm. A similar expression protocol was used to obtain the [ring-2-13C]His-labeled peroxidase HHAA mutant (His10His23Ala87Ala258); the spectral change over the pH range 5.8-7. 8 is attributed to deprotonation of the surface-exposed His23. Furthermore, replacement of Arg303, which is hydrogen bonded to His10, has no effect on the 13C spectrum. These results provide direct evidence in support of the peroxidase Cys42-sulfenic acid/thiol redox cycle and add significantly to our structure-based understanding of protein-sulfenic acid stabilization and function.

The race-specific elicitor AVR9 of the tomato pathogen Cladosporium fulvum: a cystine knot protein. Sequence-specific 1H NMR assignments, secondary structure and global fold of the protein.

The secondary structure and global fold of the AVR9 elicitor protein of Cladosporium fulvum has been determined by 2D NMR and distance-geometry protocols. The protein consists of three anti-parallel strands forming a rigid region of beta-sheet. On the basis of the NMR-derived parameters and distance geometry calculations, it is evident that the AVR9 protein is structurally very homologuous to carboxy peptidase inhibitor (CPI) of which the X-ray structure is known. The AVR9 protein reveals the presence of a cystine knot, which consists of a ring formed by two disulfide bridges and the interconnecting backbone through which the third disulfide bridge penetrates. This structural motif is found in several small proteins such as proteinase inhibitors, ion channel blockers and growth factors. The implications of the structural relationship between AVR9 and other biologically active proteins are discussed.

Molecular and biochemical basis of the interaction between tomato and its fungal pathogen Cladosporium fulvum.

The interaction between the biotrophic fungal pathogen Cladosporium fulvum and tomato complies with the gene-for-gene model. Resistance, expressed as a hypersensitive response (HR) followed by other defence responses, is based on recognition of products of avirulence genes from C. fulvum (race-specific elicitors) by receptors (putative products of resistance genes) in the host plant tomato. The AVR9 elicitor is a 28 amino acid (aa) peptide and the AVR4 elicitor a 106 aa peptide which both induce HR in tomato plants carrying the complementary resistance genes Cf9 and Cf4, respectively. The 3-D structure of the AVR9 peptide, as determined by 1H NMR, revealed that AVR9 belongs to a family of peptides with a cystine knot motif. This motif occurs in channel blockers, peptidase inhibitors and growth factors. The Cf9 resistance gene encodes a membrane-anchored extracellular glycoprotein which contains leucine-rich repeats (LRRs). 125I labeled AVR9 peptide shows the same affinity for plasma membranes of Cf9+ and Cf9- tomato leaves. Membranes of solanaceous plants tested so far all contain homologs of the Cf9 gene and show similar affinities for AVR9. It is assumed that for induction of HR, at least two plant proteins (presumably CF9 and one of his homologs) interact directly or indirectly with the AVR9 peptide which possibly initiates modulation and dimerisation of the receptor, and activation of various other proteins involved in downstream events eventually leading to HR. We have created several mutants of the Avr9 gene, expressed them in the potato virus X (PVX) expression system and tested their biological activity on Cf9 genotypes of tomato. A positive correlation was observed between the biological activity of the mutant AVR9 peptides and their affinity for tomato plasma membranes. Recent results on structure and biological activity of AVR4 peptides encoded by avirulent and virulent alleles of the Avr4 gene (based on expression studies in PVX) are also discussed as well as early defence responses induced by elicitors in tomato leaves and tomato cell suspensions.

5-Fluorouracil metabolite patterns in viable and necrotic tumor areas of murine colon carcinoma determined by 19F NMR spectroscopy.

High-resolution 19F NMR spectroscopy at 9.4 T was used to study the difference in the metabolite pattern of 5-fluorouracil (5-FU) between viable and necrotic tissues of C38 murine colon tumors grown in C57BI/6 mice. Studies were performed on perchloric acid extracts of these tumor fractions after 5-FU treatment. The 19F nuclear magnetic resonance spectra exhibited resonances representing 5-FU, the catabolites alpha-fluoro-beta-ureidopropionic acid and alpha-fluoro-beta-alanine, as well as several fluoronucleotide anabolites. The absolute concentrations of anabolites and catabolites and the anabolite-to-catabolite ratio were significantly lower in the necrotic fraction than in the viable tumor fraction 50 min after administration of 5-FU, whereas the absolute concentration of 5-FU was the same. Therefore, in 5-FU metabolism studies with NMR spectroscopy, it is important to consider the necrotic contribution to the tumor volume.

Influence of the halogen-substituent pattern of fluoronitrobenzenes on their biotransformation and capacity to induce methemoglobinemia.

In the present study both the biotransformation patterns and the capacity to induce methemoglobinemia of a series of fluoronitrobenzenes were investigated. This was done to investigate to what extent variation in the number and position of the halogen substituents influence the metabolic fate of the fluoronitrobenzenes, thereby influencing their capacity to induce methemoglobinemia. The results obtained were compared to the effect of the fluorine substituent patterns on the calculated electronic characteristics and, thus, on the chemical reactivity of the fluoronitrobenzenes. Analysis of the in vivo metabolic profiles demonstrates a dependence of the extent of nitroreduction, of glutathione conjugation, and of aromatic hydroxylation with the pattern of halogen substitution. With an increasing number of fluorine substituents at electrophilic carbon centers, 24-hr urine recovery values decreased and fluoride anion elimination increased, due to increased reactivity of the fluoronitrobenzenes with cellular nucleophiles. In vitro studies even demonstrated a clear correlation between calculated parameters for the electrophilicity of the fluoronitrobenzenes and the natural logarithm of their rate of reaction with glutathione or with bovine serum albumin, taken as a model for cellular nucleophiles (r = 0.97 and r = 0.98, respectively). Increased possibilities for the conjugation of the fluoronitrobenzenes to cellular nucleophiles were accompanied by decreased contributions of nitroreduction and aromatic hydroxylation to the overall in vivo metabolite patterns, as well as by a decreased capacity of the fluoronitrobenzenes to induce methemoglobinemia. In vitro studies on the rates of nitroreduction of the various fluoronitrobenzenes by cecal microflora and rat liver microsomes revealed that the changes in the capacity of the fluoronitrobenzenes to induce methemoglobinemia were not due to differences in their intrinsic reactivity in the pathway of nitroreduction, leading to methemoglobinemia-inducing metabolites. Thus, the results of the present study clearly demonstrate that the number and position of fluorine substituents in the fluoronitrobenzenes influence the capacity of the fluoronitrobenzenes to induce methemoglobinemia, not because their intrinsic chemical reactivity for entering the nitroreduction pathway is influenced. The different methemoglobinemic capacity must rather result from differences in the inherent direct methemoglobinemic capacity and/or reactivity of the various toxic metabolites and/or from the fact that the halogen substituent pattern influences the electrophilic reactivity, thereby changing the possibilities for reactions of the nitrobenzenes with glutathione and, especially, other cellular nucleophiles. When the number of fluorine substituents increases, the electrophilicity of the fluoronitrobenzenes can become so high that glutathione conjugation is no longer able to compete efficiently with covalent binding of the fluoronitrobenzenes to cellular macromolecules. As a consequence, it can be suggested that with an increasing number of fluorine substituents at electrophilic carbon centers in a nitrobenzene derivative, a toxic end point of the nitrobenzene other than formation of methemoglobinemia can be foreseen.