N-dealkylation of an N-cyclopropylamine by horseradish peroxidase. Fate of the cyclopropyl group.

Cyclopropylamines inactivate cytochrome P450 enzymes which catalyze their oxidative N-dealkylation. A key intermediate in both processes is postulated to be a highly reactive aminium cation radical formed by single electron transfer (SET) oxidation of the nitrogen center, but direct evidence for this has remained elusive. To address this deficiency and identify the fate of the cyclopropyl group lost upon N-dealkylation, we have investigated the oxidation of N-cyclopropyl-N-methylaniline (3) by horseradish peroxidase, a well-known SET enzyme. For comparison, similar studies were carried out in parallel with N-isopropyl-N-methylaniline (9) and N,N-dimethylaniline (8). Under standard peroxidatic conditions (HRP, H(2)O(2), air), HRP oxidizes 8 completely to N-methylaniline (4) plus formaldehyde within 15-30 min, whereas 9 is oxidized more slowly (<10% in 60 min) to produce only N-isopropylaniline (10) and formaldehyde (acetone and 4 are not formed). In contrast to results with 9, oxidation of 3 is complete in <60 min and affords 4 (20% yield) plus traces of aniline. By using [1'-(14)C]-3, [1'-(13)C]-3, and [2',3'-(13)C]-3 as substrates, radiochemical and NMR analyses of incubation mixtures revealed that the complete oxidation of 3 by HRP yields 4 (0.2 mol), beta-hydroxypropionic acid (17, 0.2 mol), and N-methylquinolinium (16, 0.8 mol). In buffer purged with pure O(2), the complete oxidation of 3 yields 4 (0.7 mol), 17 (0.7 mol), and 16 (0.3 mol), while under anaerobic conditions, 16 is formed quantitatively from 3. These results indicate that the aminium ion formed by SET oxidation of 3 undergoes cyclopropyl ring fragmentation exclusively to generate a distonic cation radical (14+*) which then partitions between unimolecular cyclization (leading, after further oxidation, to 16) and bimolecular reaction with dissolved oxygen (leading to 4 and 17 in a 1:1 ratio). Neither beta-hydroxypropionaldehyde, acrolein, nor cyclopropanone hydrate are formed as SET metabolites of 3. The synthetic and analytical methods developed in the course of these studies should facilitate the application of cyclopropylamine-containing probes to reactions catalyzed by cytochrome P450 enzymes.

Identification of three protein targets for reactive metabolites of bromobenzene in rat liver cytosol.

The hepatotoxicity of bromobenzene and many other simple organic chemicals is believed to be associated with covalent binding of chemically reactive metabolites to cellular proteins. Recently, a rat liver microsomal esterase was shown to be targeted by bromobenzene metabolites formed in vitro [Rombach, E. M., and Hanzlik, R. P. (1998) Chem. Res. Toxicol. 11, 178-184]. To identify protein targets for bromobenzene metabolites in cytosol, we incubated liver microsomes and glutathione-depleted liver cytosol from phenobarbital-treated rats with [(14)C]bromobenzene in vitro. In a separate experiment, we intraperitoneally injected a hepatotoxic dose of [(14)C]bromobenzene to phenobarbital-treated rats. The cytosol fractions from both experiments were recovered and analyzed for protein-bound radioactivity. Under the conditions that were used, 2.6 and 3.9 nmolar equiv of bromobenzene/mg of cytosolic protein was bound in vitro and in vivo, respectively. Denaturing polyacrylamide gel electrophoresis of these cytosolic proteins followed by phosphor imaging analysis revealed several radiolabeled protein bands over a broad molecular mass range, the patterns observed in vitro and in vivo being generally similar to each other. Cytosolic proteins labeled in vitro were separated by ion exchange chromatography and electrophoresis, and three major radioactive bands with estimated molecular masses of ca. 14, 25, and 30 kDa were in-gel digested with trypsin, followed by on-line HPLC electrospray ionization mass spectrometry of the resulting peptide mixtures. For the three protein bands, the observed peptide masses were found to match the predicted tryptic fragments of liver fatty acid binding protein, glutathione transferase subunit A1, and carbonic anhydrase isoform III, respectively, with 83, 45, and 59% coverage of the corresponding complete sequences. The possible relationship of the adduction of these proteins to the toxicological outcome is discussed.

Roles of key active-site residues in flavocytochrome P450 BM3.

The effects of mutation of key active-site residues (Arg-47, Tyr-51, Phe-42 and Phe-87) in Bacillus megaterium flavocytochrome P450 BM3 were investigated. Kinetic studies on the oxidation of laurate and arachidonate showed that the side chain of Arg-47 contributes more significantly to stabilization of the fatty acid carboxylate than does that of Tyr-51 (kinetic parameters for oxidation of laurate: R47A mutant, Km 859 microM, kcat 3960 min-1; Y51F mutant, Km 432 microM, kcat 6140 min-1; wild-type, Km 288 microM, kcat 5140 min-1). A slightly increased kcat for the Y51F-catalysed oxidation of laurate is probably due to decreased activation energy (DeltaG) resulting from a smaller DeltaG of substrate binding. The side chain of Phe-42 acts as a phenyl 'cap' over the mouth of the substrate-binding channel. With mutant F42A, Km is massively increased and kcat is decreased for oxidation of both laurate (Km 2. 08 mM, kcat 2450 min-1) and arachidonate (Km 34.9 microM, kcat 14620 min-1; compared with values of 4.7 microM and 17100 min-1 respectively for wild-type). Amino acid Phe-87 is critical for efficient catalysis. Mutants F87G and F87Y not only exhibit increased Km and decreased kcat values for fatty acid oxidation, but also undergo an irreversible conversion process from a 'fast' to a 'slow' rate of substrate turnover [for F87G (F87Y)-catalysed laurate oxidation: kcat 'fast', 760 (1620) min-1; kcat 'slow', 48.0 (44.6) min-1; kconv (rate of conversion from fast to slow form), 4.9 (23.8) min-1]. All mutants showed less than 10% uncoupling of NADPH oxidation from fatty acid oxidation. The rate of FMN-to-haem electron transfer was shown to become rate-limiting in all mutants analysed. For wild-type P450 BM3, the rate of FMN-to-haem electron transfer (8340 min-1) is twice the steady-state rate of oxidation (4100 min-1), indicating that other steps contribute to rate limitation. Active-site structures of the mutants were probed with the inhibitors 12-(imidazolyl)dodecanoic acid and 1-phenylimidazole. Mutant F87G binds 1-phenylimidazole >10-fold more tightly than does the wild-type, whereas mutant Y51F binds the haem-co-ordinating fatty acid analogue 12-(imidazolyl)dodecanoic acid >30-fold more tightly than wild-type.

Design, synthesis, and evaluation of azapeptides as substrates and inhibitors for human rhinovirus 3C protease.

A series of azapeptides was prepared and assessed as inhibitors of the human rhinovirus 3C protease. Boc-VLFaQ-OPh was a slow-turnover substrate that gave transient (ca. 1-2 h) inhibition as it underwent hydrolysis. Boc-VLFaG-OPh gave very slow but essentially irreversible inhibition.

Detection of adducts of bromobenzene 3,4-oxide with rat liver microsomal protein sulfhydryl groups using specific antibodies.

The hepatotoxicity of bromobenzene (BB) has been attributed to covalent modification of cellular proteins by reactive metabolites generated during its oxidative biotransformation. Much of the net covalent binding which occurs originates via quinone metabolites, but bromobenzene 3,4-oxide (BBO), which is the reactive metabolite thought to be most significant toxicologically, also arylates protein side chains, although to a lesser extent. To facilitate the detection, isolation, and identification of rat liver proteins specifically adducted by BBO, we raised polyclonal antibodies capable of recognizing S-(p-bromophenyl)cysteine moieties (anti-BP) by immunizing rabbits with p-bromophenylmercapturic acid conjugated to keyhole limpet hemocyanin. The antiserum had a high titer, showed a high specificity for hapten in competitive ELISA with hapten analogues, and performed well in Western blot experiments using synthetically haptenized control proteins. When used for Western analysis of protein fractions from in vitro incubations of rat liver microsomes with [14C]BB, affinity-purified anti-BP recognized a limited number of bands, each of which also contained 14C. One of these bands corresponds to hydrolase B, a nonspecific esterase known to contain one free sulfhydryl group and previously shown to be a target protein for [14C]BB metabolites.

Imidazolyl carboxylic acids as mechanistic probes of flavocytochrome P-450 BM3.

omega-Imidazolyl carboxylic acids (C10-C12) have been used as probes of the active site and catalytic mechanism of the fatty acid hydroxylase P-450 BM3 from Bacillus megaterium. These compounds are the most potent inhibitors of P-450 BM3 yet reported. All are mixed inhibitors, increasing the Km and decreasing the kcat for laurate oxidation. All ligate the P-450 BM3 ferric heme iron, inducing a type II shift in the Soret absorbance band from 419 to 424 nm. Binding to the ferrous form is much weaker. 10-(Imidazolyl)decanoic acid was the best inhibitor (Kic = 0.9 microM, Kiu = 5.7 microM), while 12-(imidazolyl)dodecanoic acid (Kic = 1.35 microM, Kiu = 6.9 microM) was superior to 11-(imidazolyl)undecanoic acid (Kic = 7.5 microM, Kiu = 16 microM). Dissociation constants for binding to oxidized P-450 BM3 heme iron were determined spectrophotometrically as 8 microM (C12 azole) and 27 microM (C11 azole). The binding of 10-(imidazolyl)decanoic acid was too tight for an absolute Kd to be determined spectrophotometrically, but this value is <0.2 microM. The binding of different fatty acids to the enzyme was found to have distinct effects on the Kd for the azoles. Laurate induced tighter binding (Kd for the C12 azole lowered to 4.7 microM), while arachidonate weakened the affinity (Kd increased to 23 microM). Arachidonate diminished the affinity for the C10 azole sufficiently that a Kd could be determined by spectrophotometric titration (11 microM). Affinity for the C12 azole was decreased in active-site-mutants R47G (R47 tethers the fatty acid carboxylate group) and F87Y but increased in mutant F87G-indicating an important role for this residue in determining heme accessibility. The C10 azole binds much more weakly to the spin-state-insensitive F87Y (32. 2 microM), suggesting that the inhibitors may bind preferentially to different conformers of P-450 BM3. NADP+ binding in the reductase also tightened affinity of these inhibitors for P-450 BM3 (Kd for the C12 azole decreased to 2.7 microM), but this effect was not observed for FMN-deficient mutant W574D, suggesting that the interdomain effect of NADP+ on inhibitor binding was mediated via flavin mononucleotide. Resonance Raman spectroscopy indicates that the inhibitors form low-spin complexes with P-450 BM3 and that their binding induces movements of the heme vinyls relative to the ring.

Synthesis and evaluation of peptidyl Michael acceptors that inactivate human rhinovirus 3C protease and inhibit virus replication.

Human rhinovirus, the chief cause of the common cold, contains a positive-sense strand of RNA which is translated into a large polyprotein in infected cells. Cleavage of the latter to produce the mature viral proteins required for replication is catalyzed in large part by a virally encoded cysteine proteinase (3Cpro) which is highly selective for -Q approximately GP- cleavage sites. We synthesized peptidyl derivatives of vinylogous glutamine or methionine sulfone esters (e.g., Boc-Val-Leu-Phe-vGln-OR: R = Me, 1; R = Et, 2) and evaluated them as inhibitors of HRV-14 3C protease (3Cpro). Compounds 1 and 2 and several related tetra- and pentapeptide analogues rapidly inactivated 3Cpro with submicromolar IC50 values. Electrospray mass spectrometry confirmed the expected 1:1 stoichiometry of 3Cpro inactivation by 1, 2, and several other analogues. Compound 2 also proved to be useful for active site titration of 3Cpro, which has not been possible heretofore because of the lack of a suitable reagent. In contrast to 1, 2, and congeners, peptidyl Michael acceptors lacking a P4 residue have greatly reduced or negligible activity against 3Cpro, consistent with previously established structure-activity relationships for 3Cpro substrates. Hydrolysis of the P1 vinylogous glutamine ester to a carboxylic acid also decreased inhibitory activity considerably, consistent with the decreased reactivity of acrylic acids vs acrylic esters as Michael acceptors. Incorporating a vinylogous methionine sulfone ester in place of the corresponding glutamine derivative in 1 also reduced activity substantially. Compounds 1 and 2 and several of their analogues inhibited HRV replication in cell culture by 50% at low micromolar concentrations while showing little or no evidence of cytotoxicity at 10-fold higher concentrations. Peptidyl Michael acceptors and their analogues may prove useful as therapeutic agents for pathologies involving cysteine proteinase enzymes.

Azapeptides as inhibitors and active site titrants for cysteine proteinases.

Ester and amide derivatives of alpha-azaglycine (carbazic acid, H2NNHCOOH), alpha-azaalanine, and alpha-azaphenylalanine (i.e., Ac-l-Phe-NHN(R)CO-X, where X = H, CH3, or CH2Ph, respectively) were synthesized and evaluated as inhibitors of the cysteine proteinases papain and cathepsin B. The ester derivatives inactivated papain and cathepsin B at rates which increased dramatically with leaving group hydrophobicity and electronegativity. For example, with 8 (R = H, X = OPh) the apparent second-order rate constant for papain inactivation was 67 600 M-1 s-1. Amide and P1-thioamide derivatives do not inactivate papain, nor are they substrates; instead they are weak competitive inhibitors (0.2 mM < Ki < 4 mM). Inactivation of papain involves carbamoylation of the enzyme, as demonstrated by electrospray mass spectrometry. Active site titration indicated a 1:1 stoichiometry for the inactivation of papain with 8, and both inactivated papain and cathepsin B are highly resistant to reactivation by dialysis (t1/2 > 24 h at 4 degrees C). Azaalanine derivatives Ac-L-Phe-NHN(CH3)CO-X inactivate papain ca. 400- 900-fold more slowly than their azaglycine analogues, consistent with the planar configuration at Nalpha of the P1 residue and the very substantial stereoselectivity of papain for L- vs D- residues at the P1 position of its substrates. Azaglycine derivative 9 (R = H, X = OC6H4NO2-p) inactivates papain extremely rapidly (>70 000 M-1 s-1), but it also decomposes rapidly in buffer with release of nitrophenol (kobs = 0.13 min-1); under the same conditions 8 shows <7% hydrolysis over 24 h. This nitrophenol release probably involves cyclization to an oxadiazolone since 17 (R = CH3, X = OC6H4NO2-p), which cannot form an isocyanate, releases nitrophenol almost as rapidly (kobs = 0.028 min-1). Cathepsin C, another cysteine proteinase with a rather different substrate specificity (i.e., aminopeptidase), was not inactivated by 8, indicating that the inactivation of papain and cathepsin B by azapeptide esters is a specific process. Their ease of synthesis coupled with good solution stability suggests that azapeptide esters may be useful as active site titrants of cysteine proteinases and probes of their biological function in vivo.

Identification of a rat liver microsomal esterase as a target protein for bromobenzene metabolites.

The hepatotoxicity of bromobenzene and many other simple organic molecules has been associated with their biotransformation to chemically reactive metabolites and the subsequent covalent binding of those metabolites to cellular macromolecules. To identify proteins targeted by bromobenzene metabolites, we incubated [14C]bromobenzene in vitro with liver microsomes from phenobarbital-induced rats under conditions which typically led to covalent binding of 2-4 nmol equiv of bromobenzene/mg of protein. Microsomal proteins were solubilized with detergent, separated by chromatography and electrophoresis, and analyzed for 14C by phosphorimaging of stained blots. Much of the radioactivity was associated with several bands of proteins of ca. 50-60 kDa, plus another prominent band around 70 kDa, but labeling density appeared to vary considerably overall. A major radiolabeled protein was purified by preparative electrophoresis and submitted to automated Edman microsequencing. Its N-terminal sequence was found to correspond to that of a known rat liver microsomal carboxylesterase (E.C. 3.1.1.1) previously identified as a target for reactive metabolites of halothane. The extent to which covalent modification of this protein by reactive metabolites contributes to the production of hepatotoxic effects remains to be determined.

Heme-coordinating analogs of lauric acid as inhibitors of fatty acid omega-hydroxylation.

A series of omega-substituted fatty acids with potential heme-coordinating groups was synthesized as inhibitors of lauric acid omega-hydroxylation. The compounds were evaluated using liver microsomes from clofibrate (CF)-induced rats and an engineered expressed CYP4A1-derived fusion protein called f4A1. omega-Imidazolyl-decanoic acid (compound 11) and omega-aminolauric acid (compound 7) were potent Type II ligands and potent inhibitors of lauric acid omega-hydroxylation in both CF-microsomes and f4A1. Replacing their terminal amino or imidazolyl groups with other potential iron-binding groups such as omega-methylsulfinyl-, omega-cyano-, omega-azido-, or omega-formamido all greatly reduced their potency as inhibitors of omega-hydroxylation and their affinity for cytochrome P450 as measured by Ks values. In CF-microsomes, inhibition of (omega-1)-hydroxylation of lauric acid by a homologous series of omega-imidazolyl-alkanoic acids varied only 2-fold but in the same incubations inhibition of omega-hydroxylation increased 22-fold upon going from C-8 to C-12. A similar dependence of binding affinity and inhibitory potency on chain length was also seen in the f4A1 system. In contrast, chain length had little effect on activity among n-alkylamines or N-alkylimidazoles lacking a carboxyl or other polar functional group, suggesting that 7, 11, and related bifunctional compounds interact with CYP4A1 in CF-microsomes and with f4A1 in a specific bidentate fashion. Imidazoles containing phenyl, benzyl, or phenylethyl substituents at N-1 interact less strongly than related N-alkyl-imidazoles of similar carbon number and hydrophobicity, suggesting that the steric bulk and/or rigidity of the phenyl ring is not well accommodated in the active site.

Detection of benzoquinone adducts to rat liver protein sulfhydryl groups using specific antibodies.

Benzoquinone is an electrophilic metabolite of bromobenzene and other simple aromatic compounds of toxicological interest including benzene, phenol, hydroquinone, and acetaminophen. In reacting with proteins benzoquinone shows great selectivity for Michael addition to sulfhydryl groups and formation of S-(2,5-dihydroxyphenyl) protein adducts. To facilitate the specific detection and eventual isolation and identification of such adducted proteins, we prepared an antiserum capable of recognizing hydroquinone moieties by immunizing rabbits with keyhole limpet hemocyanin modified with 3-[2,5-dihydroxyphenyl)thio]propanoyl groups as haptens. The antiserum had a high titer and showed high specificity for hapten in competitive ELISA with hapten analogues. In Western blot experiments the antiserum detected not only synthetically haptenized control proteins but also several proteins from rat liver microsomes that had been incubated in vitro with [14C]bromobenzene. This binding was completely blocked by free hapten, showing that it was hapten-specific. Each of the microsomal protein bands detected in the Western blots also contained radioactivity, but not all radioactive protein bands reacted with antibody. This antiserum should prove useful in exploring the role of protein arylation by benzoquinone in cytotoxic responses to its metabolic precursors.

Effects of steric bulk and conformational rigidity on fatty acid omega hydroxylation by a cytochrome P450 4A1 fusion protein.

The action of cytochrome P450 4A1 (CYP4A1) on fatty acid substrates is characterized by a pronounced regioselectivity for omega-hydroxylation. To elucidate the chemical basis of this specificity we probed the active site of a CYP4A1 fusion protein (f4A1) with bulky and/ or rigid analogs of lauric acid, the optimum natural substrate for f4A1 and CYP4A1. f4A1 efficiently omega-hydroxylates lauric acid, epoxidizes 11-dodecenoic acid, and oxidizes 11-dodecynoic acid to 1,12-dodecanedioic acid. Medium length fatty acids having omega-terminal groups as large as t-butyl or m-tolyloxy bind tightly to f4A1 as Type I ligands and are efficiently hydroxylated on their methyl termini. omega-Phenylnonanoic acid also induces a Type I binding spectrum (Ks = 0.77 microM) but fails to undergo hydroxylation and strongly inhibits lauric acid hydroxylation by f4A1. Slightly shorter acids such as omega-phenyloctanoic acid, naproxen, and ibuprofen also strongly inhibit lauric acid hydroxylation but do not induce a Type I spectrum and do not undergo hydroxylation. Like 10-methoxydecanoic acid, the rigid and rod-like 4'-methoxy-4-biphenylcarboxylic acid is O-demethylated by f4A1, which also omega-hydroxylates m- and p-heptyloxybenzoic acids but not m- or p-amyloxyhydrocinnamic acids. The histamine antagonist cimetidine and the peroxisome proliferator perfluorodecanoic acid are both potent inhibitors of f4A1. Thus the active site of f4A1 is quite tolerant of steric bulk and rigidity around the heme region and the polar group recognition site, but perhaps less so in the midchain region. Although CYP4A enzymes are not usually regarded as "drug metabolizing P450s," the fact that commonly used therapeutic agents strongly inhibit lauric acid omega-hydroxylation by f4A1 as well as liver microsomes from clofibrate-induced rats suggests these and related agents could potentially interfere with the contribution of CYP4A enzymes to the metabolism of endogenous lipids.

Active site structure and substrate specificity of cytochrome P450 4A1: steric control of ligand approach perpendicular to heme plane.

The active site of an engineered, expressed fusion protein containing the sequences of cytochrome P450 4A1 (lauric acid omega-hydroxylase) and NADPH-cytochrome P450 reductase, has been probed with olefinic, acetylenic, aromatic, and other analogs of its normal substrate. The fusion protein omega-hydroxylates lauric acid, epoxidizes 11-dodocenoic acid, oxidizes 11-dodecynoic acid to 1,12-dodecandioic acid, but does not oxidize 9-phenylnonanoic acid. Nevertheless, the latter is a tight-binding Type I ligand (Ks = 0.77 microM) and a potent inhibitor of lauric acid hydroxylation. These and other observations are used to construct an active site model that accounts for its remarkable substrate and inhibitor specificity.

Heteroatom substitution shifts regioselectivity of lauric acid metabolism from omega-hydroxylation to (omega-1)-oxidation.

In contrast to several isoforms of cytochrome P450 (including 1A1, 1A2, 2B1, 2C2, 2C6 and 2E1), cytochrome P450 4A1 and a fusion protein derived from it, show a strong preference for hydroxylation of lauric acid (C12) at the less chemically reactive omega-CH3 group instead of the more reactive (omega-1)-CH2 group. We have explored the interplay of steric effects on substrate binding vs chemical reactivity at various substrate loci in determining the striking difference in regioselectivity of CYP2B1 and a 4A1-derived fusion protein, through studies with heteroatom substituted analogs of lauric acid, i.e., 10-methoxydecanoic acid and 10-methylthiodecanoic acid. With both enzymes the former undergoes simple omega-hydroxylation (giving 10-hydroxydecanoic acid and HCHO), but the latter undergoes facile S-oxidation at the omega-1 position instead of omega-hydroxylation. The dramatic shift in the regioselectivity of the fusion protein toward the thia analog is consistent with the greater length of C-S bonds and the greater atomic radius and polarizability of sulfur lone-pair electrons within an otherwise restrictive active site.

Bromobenzene 3,4-oxide alkylates histidine and lysine side chains of rat liver proteins in vivo.

The hepatotoxic effects of bromobenzene (BB) are correlated with and generally ascribed to the covalent modification of cellular proteins by chemically reactive metabolites, particularly BB-3,4-oxide. Previous studies revealed that quinone as well as epoxide metabolites of BB form adducts to protein sulfur nucleophiles, that the quinone-derived adducts are more abundant by a factor of ca. 7, and that collectively these sulfur adducts account for only about 10% of the total protein covalent binding [Slaughter, D. E., and Hanzlik, R. P. (1991) Chem. Res. Toxicol. 4, 349-359]. To examine the possibility that metabolically-formed BB-3,4-oxide alkylates nitrogen nucleophiles on proteins under toxicologically relevant conditions in vivo, we synthesized standards of N tau-(p-bromophenyl)histidine (7) and N epsilon-(p-bromophenyl)lysine (8) as anticipated adduct structures and used them to guide a chromatographic search for their presence in hydrolysates of liver protein from BB-treated rats. While radio-LC chromatography and GC/MS provide unequivocal evidence for their presence, the amounts of 7 and 8 observed are very low ( < 1% of total covalent binding). The apparently small net contribution of epoxide metabolites to covalent binding of BB in vivo suggests the majority of binding may arise via quinone metabolites, but this should not be construed to imply that quinone adducts are necessarily more important toxicologically than epoxide adducts; in this context the identity of the protein targets is probably at least as important as the type of electrophilic metabolite involved.

Effects of ligand homologation and ligand reactivity on the apparent kinetic specificity of papain.

Papain, a prototype cysteine proteinase, shows a pronounced kinetic preference for substrates and inhibitors based on the Ac-L-Phe-Gly-structural motif. Replacing the L-Phe at position P2 with D-Phe, or with a less hydrophobic residue such as Leu or Met, results in decreases of substrate or inhibitory activity of up to 400-fold. In this study we examined the effect of homologating the P1 glycine moiety to beta-alanine in the context of specific ester and amide substrates, peptidyl nitrile and -aldehyde transition state analog inhibitors, and peptidyl Michael acceptors as irreversible affinity labels. Papain discriminates extremely strongly (i.e., from 1000-fold to > or = 29,000-fold) against the 'homologs' based on beta-alanine at P1 compared to 'analogs' based on glycine at P1. However, with highly reactive ligands such as p-nitrophenyl esters, homolog/analog discrimination is greatly reduced (i.e., < or = 10-fold). These observations are interpreted in terms of (1) cooperativity between several non-covalent enzyme-ligand interactions and the covalent interaction of the ligand P1 moiety with Cys-25 of papain, (2) the decreased ability of homologs to utilize these cooperative interactions optimally because of their extended size, and (3) a decrease in the importance of the cooperative interactions as the intrinsic chemical reactivity of the ligand increases. Some implications of this analog vs. homolog discrimination for peptidyl disulfide and peptidyl chloromethane probes of protease specificity and mechanism are discussed.

Fatty acid discrimination and omega-hydroxylation by cytochrome P450 4A1 and a cytochrome P4504A1/NADPH-P450 reductase fusion protein.

The omega-hydroxylation of fatty acids by certain cytochrome P450 enzymes shows a degree of chain-length and regionspecificity which is remarkable in view of the conformational flexibility of these substrates, the strong similarity in properties among homologs, and the lack of polar groups (other than the carboxy terminus) with which to guide and strength enzyme-substrate interactions. To investigate the chemical basis for these features of omega-hydroxylation we designed and synthesized a series of lauric acid analogs and evaluated them as substrates and inhibitors of omega-hydroxylation catalyzed by cytochrome P4504A1 and a cytochrome P450 4A1/NADPH-P450 reductase fusion protein. Among n-alkanoic acids, lauric acid was found to have the optimum chain length for the fusion protein, as it does for native cytochrome P450 4A1. With both enzymes, chain shortening caused a precipitous drop in turnover while chain lengthening caused a gradual drop in turnover. The fusion protein omega-hydroxylated methyl laurate and lauryl alcohol about 1/10th as efficiently as lauric acid, but it did not hydroxylate lauramide. 10-Methoxydecanoic acid underwent O-demethylation (via omega-hydroxylation). The branched substrate 11-methyllauric acid was hydroxylated efficiently and selectively at the omega-position. In contrast, the cyclopropyl analog 11,12-methanolauric acid was not detectably hydroxylated, although it induced Type I binding spectrum and inhibited lauric acid omega-hydroxylation by 43% at equimolar concentrations. omega-(Imidazolyl)-decanoic acid induced a Type II heme-binding spectrum and was an especially potent inhibitor of lauric acid hydroxylation. Collectively these data suggest that the active site of cytochrome P450 4A1 has an elongated tubular shape of definite length (ca. 14 A) with a recognition site for polar groups (including but not limited to carboxyl) at its entrance and the (oxo)heme group at its terminus.