Reversible and time-dependent inhibition of the hepatic cytochrome P450 steroidal hydroxylases by the proestrogenic pesticide methoxychlor in rat and human

Methoxychlor, a currently used pesticide, is demethylated and hydroxylated by several hepatic microsomal cytochrome P450 enzymes. Also, methoxychlor undergoes metabolic activation, yielding a reactive intermediate (M*) that binds irreversibly and apparently covalently to microsomal proteins. The study investigated whether methoxychlor could inhibit or inactivate certain liver microsomal P450 enzymes. The regioselective and stereoselective hydrox-ylation of testosterone and the 2-hydroxylation of estradiol (E₂) were utilized as markers of the P450 enzymes inhibited by methoxychlor. Both reversible and time-dependent inhibition were examined. Coincubation of methoxychlor and testosterone with liver microsomes from phenobarbital treated (PB-microsomes) male rats, yielded marked diminution of 2α- and 16α-testosterone hydroxylation, indicating strong inhibition of P4502C11 (P450h). Methoxychlor moderately inhibited 2β-, 7α-, 15α-, 15β-, and 16β-hydroxylation and androstenedi-one formation. There was only a weak inhibition of 6β-ydroxylation of testosterone. The methox-ychlor-mediated inhibition of 6β-hydroxylation was competitive. By contrast, when methoxychlor was permitted to be metabolized by PB-microsomes or by liver microsomes from pregnenolone-16α-car-bonitrile treated rats (PCN-microsomes) prior to addition of testosterone, a pronounced time-dependent inhibition of 6β-hydroxylation was observed, suggesting that methoxychlor inactivates the P450 3A isozyme(s). The di-demethylated methoxychlor (bis-OH-M) and the tris-hydroxy (ca-techol) methoxychlor metabolite (tris-OH-M) inhibited 6β-hydroxylation in PB-microsomes competitively and noncompetitively, respectively; however, these methoxychlor metabolites did not exhibit a time-dependent inhibition. Methoxychlor inhibited competitively the formation of 7α-hydroxytestosterone (7α-OH-T) and 16α-hydroxy-testosterone (16α-OH-T) but exhibited little or no time-dependent inhibition of generation of these metabolites, indicating that P450s 2A1, 2B1/B2, and 2C11 were inhibited but not inactivated. Methoxychlor inhibited in a time-dependent fashion the 2-hydroxylation of E2 in PB-microsomes. However, bis-OH-M exhibited solely reversible inhibition of the 2-hydroxylation, supporting our conclusion that the inactivation of P450s does not involve participation of the demethylated metabolites. Both competitive inhibition and time-dependent inactivation of human liver P450 3A (6β-hydroxylase) by methoxychlor, was observed. As with rat liver microsomes, the human 6β-hydroxylase was inhibited by bis-OH-M and tris-OH-M competitively and noncompetitively, respectively. Testosterone and estradiol strongly inhibited the irreversible binding of methoxychlor to microsomal proteins. This might explain the “clean” competitive inhibition by methoxychlor of the 6β-OH-T formation when the compounds were coin-cubated. Glutathione (GSH) has been shown to interfere with the irreversible binding of methoxychlor to PB-microsomal proteins. The finding that the coincubation of GSH with methoxychlor partially diminishes the time-dependent inhibition of 6β-hydroxylation provides supportive evidence that the inactivation of P450 3A isozymes by methoxychlor is related to the formation of M*.     

Purification and characterization of the hepatic CYP2C and 3A isozymes from phenobarbitone pretreated rhesus monkey

Hepatic P450s, named M-3 and M-4 were purified from phenobarbitone pretreated rhesus monkey. These demonstrated polypeptide molecular mass of 50 and 52.5 kDa and specific content of 12 and 20 nmol P450/mg protein, respectively. Both the isozymes demonstrated low spin state of heme. Antibodies raised against M-3 inhibited the activity of aminopyrine, erythromycin and ethylmorphine N-demethylase in the microsomes obtained from PB pretreated rhesus monkey by 76, 40 and 35%, respectively. M-4 did the same by 69, 85 and 79%, respectively. These observations indicated M-3 and M-4 to be the members of CYP2C and 3A subfamilies, respectively. These results were substantiated by the observations that M-3 metabolized aminopyrine whereas M-4 metabolized aminopyrine, erythromycin and ethylmorphine in the reconstituted system. Microsomal lipids and cytochrome b₅ enhanced the rate of these reactions. Further confirmation to the identity of these isozymes was provided by N-terminal amino acid sequences. The first 10 N-terminal amino acid residues of M-3 were 90% similar to CYP2C20 and 2C9 and that of M-4 were 100 and 90% similar to CYP3A8 and 3A5, respectively. In conclusion, two isozymes of hepatic P450 purified from PB pretreated rhesus monkey belong to CYP2C and 3A subfamilies.     

cDNA cloning and characterization of feline CYP1A1 and CYP1A2

Deficiency of drug glucuronidation in the cat is one of the major reasons why this animal is highly sensitive to the side effects of drugs. The characterization of cytochrome P450 isoforms belonging to the CYP1A subfamily, which exhibit important drug oxidation activities such as activation of pro-carcinogens, was investigated. Two cDNAs, designated CYP1A-a and CYP1A-b, corresponding to the CYP1A subfamily were obtained from feline liver. CYP1A-a and CYP1A-b cDNAs comprise coding regions of 1554 bp and 1539 bp, and encode predicted amino acid sequences of 517 and 512 residues, respectively. These amino acid sequences contain a heme-binding cysteine and a conserved threonine. The cDNA identities, as well as the predicted amino acid sequences containing six substrate recognition sites, suggest that CYP1A-a and CYP1A-b correspond to CYP1A1 and CYP1A2, respectively. This was confirmed by the kinetic parameters of the arylhydrocarbon hydroxylase and 7-ethoxyresorufin O-deethylase activities of expressed CYPs in yeast AH22 cells and by the tissue distribution of each mRNA. However, theophylline 3-demethylation is believed to be catalyzed by CYP1A1 in cats, based on the high V(max) and low K(m) seen, in contrast to other animals. Because feline CYP1A2 had a higher K(m) for phenacetin O-deethylase activity with acetaminophen, which cannot be conjugated with glucuronic acid due to UDP-glucuronosyltransferase deficiency, it is supposed that the side effects of phenacetin as a result of toxic intermediates are severe and prolonged in cats.     

Inhibition of human hepatic CYP isoforms by mGluR5 antagonists

Characterization of new chemical entities for their potential to produce drug-drug interactions is an important aspect of early drug discovery screening. In the present study, the potential for three metabotropic glutamate receptor antagonists to interact with recombinant human CYPs was investigated. 2-Methyl-6-(phenylethenyl) pyridine (SIB-1893), 2-methyl-6-(phenylethynyl) pyridine (MPEP) and 3-[2-methyl-1,3-thiazol-4-yl) ethynyl]-pyridine (MTEP) were moderate competitive inhibitors of recombinant human CYP1A2 (Ki, 0.5-1 microM). SIB-1893, but not MPEP or MTEP, was also a moderate competitive inhibitor of CYP1B1. MPEP and MTEP were weak inhibitors of CYP2C19. None of the three compounds tested were significant inhibitors (IC(50) values >50 microM) of CYP3A4, 2C9, 2D6, 2A6, 2B6 or 2E1. The results suggest that MTEP is a selective inhibitor of CYP1A2 and may prove to be a useful tool in studying drug-drug interactions involving this enzyme.     

CYP3A4 and CYP2D6 inhibitory activities of Indonesian medicinal plants

Thirty samples of Indonesian medicinal plants were analyzed for their capacity to inhibit in vitro metabolism by human cytochrome P450 3A4 (CYP3A4) and CYP2D6 with a radiometric assay. The MeOH-soluble fractions of 25 samples, prepared from water extracts, demonstrated inhibitory activity more than 50% on the metabolism mediated by CYP3A4, and 21 samples on the metabolism mediated by CYP2D6. Among the MeOH-soluble fractions, Piper nigrum leaf showed the highest inhibitory activity against CYP3A4 (91.7%), and Punica granatum against CYP2D6 (98.1%). The water extracts of which MeOH-soluble fraction showed inhibitory activity more than 70% were fractionated with EtOAc. From the EtOAc-soluble fractions, Curcuma heyneana (67.0%), Pi. cubeba (75.0%), Pi. nigrum fruit (84.0%), Pi. nigrum leaf (85.8%), and Zingiber aromaticum (75.3%) demonstrated inhibitory activity more than 50% on the metabolism mediated by CYP3A4, but only Pi. nigrum fruit (72.8%) and Pi. nigrum leaf (69.1%) showed strong inhibitory activity against CYP2D6. For samples that showed more than 70% inhibition, their IC(50) values were determined. The most potent inhibitory activity against CYP3A4 (IC(50) value of 25 microg/ml) was found for the extract of Pi. nigrum leaf, while that of Catharanthus roseus showed the most potent inhibitory effect against CYP2D6 (IC(50) value of 11 microg/ml). These results should indicate once more the possibility of potential medicinal plant-drug interactions.     

Expression and induction of CYP3As in human fetal hepatocytes

CYP3A4 and CYP3A7 mRNA expression levels were markedly up-regulated by dexamethasone (DEX), but not by rifampicin (RIF). CYP3A5 mRNA level was not increased significantly by DEX, RIF, or phenobarbital. Testosterone 6beta-hydroxylase activity was induced to about 2-fold of control by DEX. However, concomitant treatment with RIF did not alter DEX-mediated induction of CYP3A mRNA expression and testosterone 6beta-hydroxylase activity. DEX-mediated induction of CYP3A mRNA was suppressed in a dose-dependent manner by RU486, a glucocorticoid receptor (GR) antagonist. At 5microM RU486, DEX-mediated induction of CYP3A4, CYP3A5, and CYP3A7 mRNA expression was inhibited almost completely. These results suggest that, in human fetal hepatocytes, PXR is not involved in DEX-mediated induction of CYP3A4 and CYP3A7, and that the induction is mediated directly by GR.     

Metabolic pathways of dioxin by CYP1A1: species difference between rat and human CYP1A subfamily in the metabolism of dioxins

Metabolism of polychlorinated dibenzo-p-dioxins by CYP1A subfamily was examined by using the recombinant yeast microsomes. In substrate specificity and reaction specificity, considerable species differences between rats and humans were observed in both CYP1A1- and CYP1A2-dependent metabolism of dioxins. Among four CYPs, rat CYP1A1 showed the highest activity toward dibenzo-p-dioxin (DD) and mono-, di-, and trichloroDDs. To reveal the mechanism of dioxin metabolism, we examined rat CYP1A1-dependent metabolism of 2-chloro-dibenzo-p-dioxin. In addition to hydroxylation at an unsubstituted position, hydroxylation with migration of a chloride substituent, hydroxylation with elimination of a chloride substituent, and cleavage of an ether linkage of the dioxin ring were observed. In particular, the cleavage of an ether linkage of the dioxin ring appeared most important for the detoxication of dioxins. Based on these results, the metabolic pathways of 2-chloro-dibenzo-p-dioxin by rat CYP1A1 were proposed. The metabolic pathways contain most of the metabolites observed in vivo using experimental animals, suggesting that P450 monooxygenase systems including CYP1A1 are greatly responsible for dioxin metabolism in vivo.     

Influence of single and concurrent clofibrate and phenobarbital administration on cytochrome P450-dependent mixed function oxidase activities and peroxisome proliferation in male rat liver

The influence of both single and concurrent administration of phenobarbital and clofibrate on hepatomegaly, cytochrome P450-depen-dent mixed function oxidase activities, and peroxisome proliferation in male rat liver have been studied. Both xenobiotics separately increase the liver :body weight ratio and their combined administration results in greater hepatomegaly than either compound alone. Both compounds induce NADPH-cytochrome c(P450) reductase activity and laurate ω- and ω-1-hydroxylase activities, but only phenobarbital induces pentoxyresorufin-O-de-alkylase. None of the drug treatments induced microsomal cytochrome b5. Phenobarbital did not cause peroxisome proliferation and inhibited the corresponding clofibrate-dependent proliferation. Taken collectively, our studies have demonstrated that concomitant treatment with phenobarbital and clofibrate are largely permissive with respect to the hepatic mixed function oxidase system but have opposing effects on the phenomenon of peroxisome proliferation in the same tissue.     

Hepatic cytochrome P450 2B induction by ethyl/ phenyl-substituted congeners of phenobarbital in the B6C3F1 mouse

The abilities of structural congeners of phenobarbital to induce immunoreactive hepatic cytochrome P450 2B (CYP2B) protein and associated catalytic activity (benzyloxyresorufin O-dealkylation) in the male B6C3F1 mouse were examined. Interspecies differences in inducing ability were examined through comparison of the results with induction data obtained previously with the male F344/NCr rat. The congeners were administered in the diet for 2 weeks at concentrations equimolar to 500 ppm of the prototype CYP2B inducer, phenobarbital. Of the series of compounds tested, phenobarbital was the most effective inducer of benzyloxyresorufin O,-dealkylation and immunoreactive CYP2B protein, with 2-ethyl-2-phenylsuccinimide, 5-ethyl-5-phenylhydantoin, primidone, and glutethimide being only 19–42% as effective. 5-Ethyl-5-phenyloxazolidinedione and the ring-opened and decarboxylated congeners, N-(2-ethyl-2-phenylacetyl)urea and 2-ethyl-2-phenyl-malonamide, displayed minimal induction of these catalytic activities. Dose-response experiments performed with 5-ethyl-5-phenylhydantoin indicated that the intrinsic CYP2B-inducing activity of this congener was as great as that of phenobarbital in the mouse, although a fourfold greater dietary concentration of this hydantoin (2000 ppm) was required to elicit a response equivalent to that caused by 500 ppm phenobarbital. When extent of induction was related to serum total xenobiotic concentration rather than to administered dietary concentration, the potencies of the two congeners were determined to be more similar (58 vs. ≥78 μ M for phenobarbital and 5-ethyl-5-phenylhydantoin, respectively).     

Localization of CYP4B1 in the rat nasal cavity and analysis of CYPs as secreted proteins

CYP4B1 is highly expressed in rat nasal respiratory mucosa, and to a lesser extent in olfactory mucosa. Examination of high-power photomicrographs suggests that CYP4B1 may be a secreted protein, based on the fact that immunoreactivity appears to be present in the lumens of ducts of Bowman's glands (rather than intracellular localization, as we observed with an antibody recognizing CYP2F4) and in secretory granules in respiratory mucosa. Furthermore, anti-CYP4B1 immunoreactivity is present on the surface of both respiratory and olfactory mucosa. We used SignalP 3.0 analysis to ascertain the likelihood that rat CYP4B1 is a secreted protein. While this analysis does not suggest that rat CYP4B1 is a secreted protein, several other cytochrome P450 enzymes were predicted to be secreted proteins. The observation that multiple human cytochrome P450s appear to be secreted proteins helps to explain the appearance of anti-cytochrome P450 antigens in cases of human autoimmune liver diseases.     

19F-NMR reveals substrate specificity of CYP121A1 in Mycobacterium tuberculosis

Cytochromes P450 are versatile enzymes that function in endobiotic and xenobiotic metabolism and undergo meaningful structural changes that relate to their function. However, the way in which conformational changes inform the specific recognition of the substrate is often unknown. Here, we demonstrate the utility of fluorine (¹⁹F)-NMR spectroscopy to monitor structural changes in CYP121A1, an essential enzyme from Mycobacterium tuberculosis. CYP121A1 forms functional dimers that catalyze the phenol-coupling reaction of the dipeptide dicyclotyrosine. The thiol-reactive compound 3-bromo-1,1,1-trifluoroacetone was used to label an S171C mutation of the enzyme FG loop, which is located adjacent to the homodimer interface. Substrate titrations and inhibitor-bound ¹⁹F-NMR spectra indicate that ligand binding reduces conformational heterogeneity at the FG loop in both the dimer and in an engineered monomer of CYP121A1. However, only the dimer was found to promote a substrate-bound conformation that was preexisting in the substrate-free spectra, thus confirming a role for the dimer interface in dicyclotyrosine recognition. Moreover, ¹⁹F-NMR spectra in the presence of substrate analogs indicate the hydrogen-bonding feature of the dipeptide aromatic side chain as a dicyclotyrosine specificity criterion. This study demonstrates the utility of ¹⁹F-NMR as applied to a multimeric cytochrome P450, while also revealing mechanistic insights for an essential M. tuberculosis enzyme.     

Inhibition of CYP2B4 by the mechanism-based inhibitor 2-ethynylnaphthalene: inhibitory potential of 2EN is dependent on the size of the substrate

2-Ethynylnaphthalene (2EN) is a mechanism-based inhibitor of CYP2B4 with two components to the inhibition, (1) enzyme inactivation, which requires covalent binding of the 2EN metabolite, and (2) reversible inhibition by 2EN itself. Both inhibitory components were examined using several different CYP2B4 substrates. Preincubation of CYP2B4 with 2EN led to a time-dependent inactivation of each of the CYP2B4-dependent activities examined; however, the ability of 2EN to reversibly inhibit CYP2B4 depended on the substrate employed, which is inconsistent with classical inhibition patterns. The degree 2EN's reversible inhibition was shown not to correlate with the substrate affinity for the active site, but with parameters related to the molecular size of the substrate. The results are consistent with 2EN and the smaller substrates simultaneously fitting in the CYP2B4 active site, leading to very little inhibition. Larger substrates exhibited greater degrees of inhibition because of their inability to co-bind with inhibitor in the active site.     

Generation of 2,3,7,8-TCDD-metabolizing enzyme by modifying rat CYP1A1 through site-directed mutagenesis

Polychlorinated dibenzo-p-dioxins (PCDDs) are known as g environmental contaminants on account of the extreme toxicity. Among these compounds, 2,3,7,8-tetrachlorodibenzo-p-dioxin (2,3,7,8-TetraCDD) is regarded as the most toxic one. The extremely high toxicity of 2,3,7,8-TetraCDD is based on its high affinity for Ah receptor and nearly undetectable metabolism in mammalian body. Based on our previous studies, we assumed that enlarging the space of substrate-binding pocket of rat CYP1A1 might generate the catalytic activity toward 2,3,7,8-TetraCDD. Large-sized amino acid residues located at putative substrate-binding sites of rat CYP1A1 were substituted for alanine by site-directed mutagenesis. Among eight mutants examined, the mutant in the putative F-G loop, F240A, showed metabolic activity toward 2,3,7,8-TetraCDD. HPLC and GC-MS analyses strongly suggested that the metabolite was 8-hydroxy-2,3,7-TriCDD. Ah receptor assay revealed that the affinity of 8-hydroxy-2,3,7-TriCDD for Ah receptor was less than 0.01% of 2,3,7,8-TetraCDD, indicating that the F240A-dependent metabolism resulted in remarkable detoxification of 2,3,7,8-TetraCDD. The novel 2,3,7,8-TetraCDD-metabolizing enzyme could be applicable to bioremediation of contaminated soils with dioxin, elimination of dioxin from foods, and clinical treatment for people who accidentally take dioxin into their systems.     

Rational design of novel CYP2A6 inhibitors

Inhibition of CYP2A6-mediated nicotine metabolism can reduce cigarette smoking. We sought potent and selective CYP2A6 inhibitors to be used as leads for drugs useful in smoking reduction therapy, by evaluating CYP2A6 inhibitory effect of novel formyl, alkyl amine or carbonitrile substituted aromatic core structures. The most potent CYP2A6 inhibitors were thienopyridine-2-carbaldehyde, benzothienophene-3-ylmethanamine, benzofuran-5-carbaldehyde and indole-5-carbaldehyde, with IC₅₀ values below 0.5 μM for coumarin 7-hydroxylation. Nicotine oxidation was effectively inhibited in vitro by two alkyl amine compounds and benzofuran-5-carbonitrile. Some of these molecules could serve as potential lead molecules when designing CYP2A6 inhibitory drugs for smoking reduction therapy.     

Regulation of the expression of two female-predominant CYP3A mRNAs (CYP3A41 and CYP3A44) in mouse liver by sex and growth hormones

A second female-predominant murine CYP3A, CYP3A44, was isolated from liver and its mRNA expression was compared with that of the previously described CYP3A41. The expression of CYP3A44 was relatively constant after birth in females, whereas it gradually declined in males after 5 weeks of age. The expression of CYP3A41 increased with age in females after 3 weeks of age, whereas it gradually declined in males after 5 weeks of age. Hypophysectomy and growth hormone replacement indicated that expression of both CYP3A mRNAs in females was dependent on the feminine plasma growth hormone profile. Estradiol induced the expression of both mRNAs and the effect was dependent on the presence of the pituitary gland. These observations suggest that endocrine control of expression might be similar, but not identical, for two female-predominant CYP3A mRNAs.     

Loss of CYP3A7 gene induction by 1,25-dihydroxyvitamin D3 is caused by less binding of VDR to the proximal ER6 in CYP3A7 gene

Cytochrome P450 3A4 and 3A7 (CYP3A4 and CYP3A7, respectively) are predominant forms in the human adult and fetal liver, respectively. 1,25-Dihydroxyvitamin D(3) (1,25(OH)(2)D(3)) is known to be a potent inducer of CYP3A4 in human colon carcinoma Caco-2 via vitamin D receptor (VDR). However, whether CYP3A7 is inducible by 1,25(OH)(2)D(3) has not yet been elucidated. In the present study, we examined the effect of 1,25(OH)(2)D(3) on CYP3A7 gene expression in Caco-2 cells, which express CYP3A4 and CYP3A7 mRNAs. 1,25(OH)(2)D(3) hardly induced the expression of CYP3A7 mRNA in contrast to the marked induction of CYP3A4 mRNA. Reporter assay using 5'-franking region CYP3A4 and CYP3A7 genes also revealed that 1,25(OH)(2)D(3) activates CYP3A4 promoter, but not CYP3A7 promoter, which has two mutations in the proximal ER6 site compared with CYP3A4 promoter. In addition, we found that the binding of VDR to the proximal ER6 in CYP3A7 gene was markedly less than that to the proximal ER6 in CYP3A4 gene using gel shift assay. Taken together, the decrease of VDR binding to the proximal ER6 caused by the mutation results in the loss of CYP3A7 gene activation by 1,25(OH)(2)D(3).     

Molecular modelling of human CYP1B1 substrate interactions and investigation of allelic variant effects on metabolism

Molecular modelling of human CYP1B1 based on homology with the mammalian P450, CYP2C5, of known three-dimensional structure is reported. The enzyme model has been used to investigate the likely mode of binding for selected CYP1B1 substrates, particularly with regard to the possible effects of allelic variants of CYP1B1 on metabolism. In general, it appears that the CYP1B1 model is consistent with known substrate selectivity for the enzyme, and the sites of metabolism can be rationalized in terms of specific contacts with key amino acid residues within the CYP1B1 heme locus. Furthermore, a mode of binding interaction for the inhibitor, alpha-naphthoflavone, is presented which accords with currently available information. The current paper shows that a combination of molecular modelling and experimental determinations on the substrate metabolism for CYP1B1 allelic variants can aid in the understanding of structure-function relationships within P450 enzymes.     

Identification of Val117 and Arg372 as critical amino acid residues for the activity difference between human CYP2A6 and CYP2A13 in coumarin 7-hydroxylation

Human cytochrome P450 (CYP) 2A6 and 2A13 play an important role in catalyzing the metabolism of many environmental chemicals including coumarin, nicotine, and several tobacco-specific carcinogens. Both CYP2A6 and CYP2A13 proteins are composed of 494 amino acid residues. Although CYP2A13 shares a 93.5% identity with CYP2A6 in the amino acid sequence, it is only about one-tenth as active as CYP2A6 in catalyzing coumarin 7-hydroxylation. To identify the key amino acid residues that account for such a remarkable difference, we generated a series of CYP2A6 and CYP2A13 mutants by site-directed mutagenesis/heterologous expression and compared their coumarin 7-hydroxylation activities. In CYP2A6, the amino acid residues at position 117 and 372 are valine (Val) and arginine (Arg), respectively; whereas in CYP2A13, they are alanine (Ala) and histidine (His). Kinetic analysis revealed that the catalytic efficiency (Vmax/Km) of the CYP2A6 Val(117)--> Ala and Arg(372)--> His mutants was drastically reduced (0.41 and 0.64 versus 3.23 for the wild-type CYP2A6 protein). In contrast, the catalytic efficiency of the CYP2A13 Ala(117) --> Val and His(372) --> Arg mutants was greatly increased (2.65 and 2.60 versus 0.31 for wild-type CYP2A13 protein). These results clearly demonstrate that the Val at position 117 and Arg at position 372 are critical amino acid residues for coumarin 7-hydroxylation. Based on the crystal structure of CYP2C5, we have generated the homology models of CYP2A6 and CYP2A13 and docked the substrate coumarin to the active site. Together with the kinetic characterization, our structural modeling provides explanations for the amino acid substitution results and the insights of detailed enzyme-substrate interactions.