酶的反应机制

MECHANISM OF ENZYME ACTION1,~ BY WILLIAM P. JENCKS Brandeis University, Waltham, Massad~usetts Clearly, a review of this subject must be highly selective. The reviewer has chosen only a few topics for discussion and has attempted to follow the editors’ request for an interpretive, rather than a comprehensive, review. He has not hesitated to climb out on a limb in a number of cases, in the hope that attempts to saw off the limb by interested or irritated investigators will advance knowledge in this area. A fact that is often glossed over should be emphasized at the outset: In spite of the large number of enzyme "mechanisms" which have been proposed, there is no instance in which the mechanism of action of an enzyme is understood. If the emphasis in this review seems to favor mechanisms of nonenzymic catalysis, this is because nonezymic systems are so much simpler to study and, in some cases, to understand. Hopefully, an improved under- standing of the mechanisms of catalysis which are available to enzymes may aid in elucidating the mechanism of enzymic catalysis. Some small but promising steps in this direction have been taken. In the past year four more or less comprehensive monographs concerned with this subject have appeared (1 to 4). In addition, Bender’s recent review (5) is an excellent source of information on acyl group transfer reactions and detailed discussions of individual enzymes may be found in the multi- volume second edition of The Enzymes. Of the many influences an enzyme might exert to increase the rate of a reaction, sufficient information is available for discussion of examples of (a) catalysis by approximation, (b) covalent catalysis, (c) general acid-base catalysis and (d) catalysis by distortion. Clearly, in many instances there overlap between these mechanisms. 1. Catalysis by approximation.--Koshland (6) has extended his earlier calculations (7) on the amount of rate acceleration which might be expected by simple approximation of reacting molecules to include a requirement for proper mutual orientation of the reacting molecules and a consideration of the increase in rate which might be expected if a number of catalysts were to act in a concerted manner. The introduction of reasonable orientation requirements does not alter the previous conclusion that little rate accelera- tion can be obtained by simple approximation of the molecules, except in very dilute solutions. This is principally because of the very low concentra- tion of reactive sites which is present in a dilute enzyme solution. Since a 1 The survey of the literature pertaining to this review was completed in October 1962. ~ The following abbreviations will be used: DFP (diisopropyl fluorophosphate); NADH (diphosphophyridine nucleotide, reduced form). 639 Annual Reviews www.annualreviews.org/aronline A nn u. R ev . B io ch em . 1 96 3. 32 :6 39 -6 76 . D ow nl oa de d fro m ar jou rna ls. an nu alr ev iew s.o rg by F lo rid a St at e U ni ve rs ity o n 10 /2 7/ 08 . F or p er so na l u se o nl y. 640 JENCKS number of intramolecular reactions proceed at rates which are very much faster than their intermolecular counterparts (at any attainable concentra- tion of reactants) and since it is likely that enzymes do orient molecules in position for reaction with each other, this requires that the specific rate constant for the intramolecular or enzymic reaction must be increased. This could result from steric compression or distortion, or from the action of catalysts. Koshland shows that for a concerted reaction of two molecules with three catalysts the rate constant will be increased by a factor of 10~3 to 10lr over that in free solution if the catalytic groups are properly oriented about an active site to which the substrate molecules are bound. While this is a large factor, it should be remembered that the rate of a five-molecule reaction in solution is so slow as to have never been observed, so that the rate constant of the catalyzed reaction may still not be large. The number of published examples of intramolecular reactions which are proposed as models for enzymic reactions has continued to increase rapidly; many examples may be found in recent reviews (1, 5, 8) and only a few will be noted here. The rapid solvolysis of half-esters of succinate, glutarate and phthalate derivatives illustrates many of the conclusions to be drawn from this class of reaction (9 to 17). The reactions proceed at a rate prop- ortional to the concentration of the monoanlon species and involve attack of carboxylate ion at the adjacent carboxyl group to form the cyclic I II III anhydride. In one instance the cyclic anhydride has been identified as an intermediate (16). The introduction of methyl or isopropyl groups on the intermediate carbon atoms forces the carboxyl groups to approach each other more closely (I) and results in rate accelerations of the order of 100-fold (14, 15, 16). A still larger rate-acceleration is observed if the reacting groups are fixed in apposition by attachment to a rigid bicyclic ring system (15, 16). This result of gem substitution and approximation has been described in terms of a decrease in unprofitable rotamer distribution (16), but may also involve a rate increase due to a forced approximation of the reacting groups to a position close enough to overcome some of the energy barrier to reac- tion (I). Analogous rate accelerations are found in polymers containing Annual Reviews www.annualreviews.org/aronline A nn u. R ev . B io ch em . 1 96 3. 32 :6 39 -6 76 . D ow nl oa de d fro m ar jou rna ls. an nu alr ev iew s.o rg by F lo rid a St at e U ni ve rs ity o n 10 /2 7/ 08 . F or p er so na l u se o nl y. MECHANISM OF ENZYME ACTION 641 carboxylate, carboxylic acid and ester groups, and there is evidence that both the acid and the anion species may accelerate the hydrolysis (14, 18, 19, 20). Several examples are known of rate accelerations which apparently are the result of internal general acid catalysis. The alkaline hydrolysis of esters (and thiol esters) of the protonated species of dimethylaminoethanol (and dirnethylaminoethanethiol) is some 100 to 200-fold faster than the corre- sponding choline (or thiocholine) derivatives, which are structurally similar to the dimethylamino compounds except for the substitution of a methyl group for hydrogen (21, 22, 23). It has been suggested that these reactions involve general acid catalysis by the proton of the alkylammonium group (II), but an alternative explanation is that they involve intramolecular nu- cleophilic attack of the free amino group (III). A rapid hydrolysis and reac- tion with acetate is found in the analogous case of methyl pyrrolidylacetyl- salicylate hydrochloride (24). More definitive evidence for internal general acid-base catalysis is found in the solvolysis of o-carboxyphthalimide, studied by Zerner & Bender (25), which proceeds at a decreased rate in D20. The mechanism of the rate acceleration of ester solvolysis produced by a properly oriented hydroxyl group is still not clear, in spite of concentrated study (26, 27, 28). These reactions generally proceed only a few times faster than the hydrolysis of corresponding compounds in which a methoxyl group is sub- stituted for hydroxyl, and the inductive effect of oxygen substitution in allcyclic compounds may be considerably larger than the hydrogen bonding effect introduced by the hydroxyl group, at least in aqueous solution. The increased rate of semicarbazone formation from o-OH, compared to p-OH benzaldehyde, which might have been ascribed to intramolecular hydrogen bonding, is found to an even greater extent in the o-OCH3 compound and is apparently due to a difference in resonance stabilization of the aldehyde by ortho and para substituents (29). Sheehan & McGregor (30) have reported that the cyclic peptide, cyclo- gly-hist-ser-gly-hist-ser-, which contains several of the groups which have been implicated in the action of certain proteolytic enzymes, undergoes a self-induced hydrolysis in water to give serine and a diketoplperazine. The mechanism of this interesting reaction is unexplained. The intramolecular reactions so far considered suffer the disadvantage, as enzyme models, that the chemist has built in the approximation of the reacting groups when the molecule was synthesized; an enzyme does not have all the tools and energy of the chemist available to it and must utilize its available binding forces to approximate the reactants. While this is presumably what does happen on the enzyme surface, it requires energy to occur, and a more satisfactory model would include binding of the substrate by weak forces to a catalyst or another reactant to facilitate the reaction and, perhaps, to confer specificity on the reaction. Clear-cut examples of this kind of catalysis for reactions in solution are conspicuous by their rarity, although a number are known in heterogeneous catalysis and for non- Annual Reviews www.annualreviews.org/aronline A nn u. R ev . B io ch em . 1 96 3. 32 :6 39 -6 76 . D ow nl oa de d fro m ar jou rna ls. an nu alr ev iew s.o rg by F lo rid a St at e U ni ve rs ity o n 10 /2 7/ 08 . F or p er so na l u se o nl y. 642 JENCKS catalytic reactions. The best example, unfortunately, involves negative catalysis, but serves to illustrate the principle: Higuchi & Lachman (31) found that caffeine markedly inhibits the hydrolysis of benzocaine in water by forming an unreactive complex bet~veen the two molecules. The forces which are responsible for the equilibrium constant of 60 M-1 for the associa- tion of these molecules presumably fall into the category loosely designated "charge transfer forces." Attempts by Marburg, in the author’s laboratory, to utilize such forces to demonstrate positive catalysis in reactions of com- pounds of this kind have been unsuccessful. Ross & Kuntz (32) described somewhat similar situation in the reaction of aniline with 1-chloro-2,4- dinitrobenzene. The rate constant for this reaction decreases with increasing concentration of the reactants because of the formation of an unreactive, yellow molecular complex between the reactants; no such rate decrease or complex formation was found for the reaction with the aliphatic compound, dibutylamine. Rate accelerations of up to 3-fold have been noted by Bell et al. (33) in the carboxylate ion catalyzed enolization of ketones when the catalyst and the ketone both contain large nonpolar groups which could interact by "hydrophobic" (34) forces. Swain & Taylor (35) found a similar small rate acceleration by comparison of the rates of reaction of phenoxide and hydroxide ions with benzyldimethyl- and trimethylsulfonium salts. Al- though hydroxide ion reacts twice as rapidly as phenoxide with the tri- methyl compound, phenoxide reacts three times more rapidly than hydroxide with the benzyl compound. The inhibition by aliphatic amines of the rate of alkaline decomposition of imido esters in aqueous solution appears to be a further example of this type of interaction (36). A clear-cut example of the operation of such hydrophobic forces to cause ion-pair formation under condi- tions in which smaller ions do not associate has been reported by Packter & Donbrow (37), who have measured the equilibrium constants for ion pair formation between ions such as decyltrimethylammonium and azo- benzene-4-sulfonate. There is a regular increase in the equilibrium constant for ion-pair formation as the size for the interacting nonpolar groups is increased. Bunnett and co-workers (38, 39, 40) have attributed the espe- cially rapid reaction of highly polarizable nucleophilic reagents with aro- matic substrates and benzyl chlorides containing polarizable substituents in the ortlw position to a London force interaction between the nucelophilic reagent and the substrate; however, it is difficult to explain the high reactiv- ity of p-substituted compounds by this mechanism. An example of catalysis in which the catalyst apparently serves only to approximate the reactants has been reported by Peer (41). The o-hydroxyla- tion of phenol by formaldehyde in benzene is specifically promoted by borate, which presumably acts by virtue of its ability to rapidly and reversibly form esters with hydroxyl groups, to give a mixed diester of borate with the hydroxyl groups of phenol and of (hydrated) formaldehye and thus promote hydroxymethylation in the ortho position. Borate is ineffective in aqueous solution, presumably because of a decreased tendency to form the phenyl Annual Reviews www.annualreviews.org/aronline A nn u. R ev . B io ch em . 1 96 3. 32 :6 39 -6 76 . D ow nl oa de d fro m ar jou rna ls. an nu alr ev iew s.o rg by F lo rid a St at e U ni ve rs ity o n 10 /2 7/ 08 . F or p er so na l u se o nl y. MECHANISM OF ENZYME ACTION 643 ester in water, but a number of metal hydroxides are effective in both solvents and presumably act in a similar manner, by binding to both the phenol and formaldehyde. Grant et al. (43) have made the remarkable observation that imidazole catalyzes the decomposition of penicillin if the solution is frozen, but not if it is in the liquid state. The catalysis is less rapid in frozen D~O. This un- usual catalysis might be caused by an approximation of the reactants in the crystal lattice, but it was also suggested that the rapid proton transfer which can occur in the crystal structure of ice may be invoved. The incorporation of a reactant into a charged micelle may either increase or decrease the rate of its reaction compared to the rate in aqueous solution. Duynstee & Grunwald (44) found that the reaction of cationic dyes, such as crystal violet, with hydroxide ion is accelerated 4 to 50 times if the dye is incorporated into a cationic micelle of cetyltrimethylammonium bro- mide, but is decreased in a mieelle of sodium lauryl sulfate, as might be expected from electrostatic considerations. The hydrolysis of the hydrophobic Schiff base, benzylideneanillne, is decreased some 25-fold when it is incor- porated into a micelle of cetyltrimethylammonium bromide (45) and the rate of alkaline decomposition of imine dyes is decreased by several orders of magnitude in micelles (46). Bender & Chow (47) have reported that the reaction of o-nitrophenyl oxalate anion with 2-aminopyridinlum cation (which might, alternatively, be the kinetically indistinguishable reaction of o-nitrophenyl hydrogen oxa- late with 2-aminopyridine) occurs at a considerably faster rate than might be expected; no reaction is found, for example, of 2-aminopyridinium ion with the uncharged ester, o-nitrophenyl acetate. The interpretation of this as an example of facilitation by interaction of opposite charges is clouded by the above-mentioned ambiguity as to the nature of the reacting species and by the fact that the 4-aminopyridinium ion also reacts readily. Similar types of interaction have been searched for in the reactions of polymers with small molecules. Ladenheim & Morawetz (48) found that bromoacetamide reacts 4- to 10-fold more rapidly with a partially ionized methacrylic acid polymer than with the corresponding monomer; this is attributed to an interaction vdth the carboxyllc acid groups. Both bromoacetamide and bromoacetate react with poly(vinylpyridine betaine) under conditions which there is no reaction with the monomer; this must be due to a concen- tration of both alkylating agents in the vicinity of the polymer, rather than an electrostatic effect, since both the anion and the amlde react. Letslnger & Savereide (49, 50) have shown that poly(4-vinylpyridine) reacts specifi- cally with the anion, 3-nitro-4-acetoxybenzene sulfonate, compared to the uncharged substrate, dinitrophenyl acetate. The reaction shows a pH-rate maximum and apparently involves protonated nitrogen atoms, which act to promote electrostatic binding of substrate, and free nitrogen atoms, for nucleophille or general-base-catalyzed hydrolysis. Per~nyi (51) has shown that cysteamlne is less effective than the corresponding polymer in a reac- Annual Reviews www.annualreviews.org/aronline A nn u. R ev . B io ch em . 1 96 3. 32 :6 39 -6 76 . D ow nl oa de d fro m ar jou rna ls. an nu alr ev iew s.o rg by F lo rid a St at e U ni ve rs ity o n 10 /2 7/ 08 . F or p er so na l u se o nl y. 644 JENCKS tion with p-nitrophenyl phosphate; the explanation of this observation is not clear. The many recent advances in the synthesis of polymers with specifically oriented groups have indicated that a highly specific interaction must occur between the monomer undergoing addition to the growing poly- mer and the catalyst and polymer itself to account for the high degree of optical and geometric specificity of the products. Specificity of addition is found with both heterogeneous and homogeneous catalysis. The theory that the configuration of the terminal polymer unit to which addition takes place determines the specificity of the next addition finds no support in the observation of Fray & Robinson (52) that optically active trialkylaluminum does not serve as a primer for the synthesis of optically active polymers. This subject is considered in detail in recent reviews (53, 54). Ion exchange resins are, in principle, similar to soluble polymers and show a similar, but moderate, degree of selectivity. For example, sulfonic acid resins preferentially catalyze the hydrolysis of positively charged peptides and polysaccharides (55, 56 and references therein). Finally, the reactions which occur in inclusion compounds should be mentioned. A good example is the decomposition of substituted sym- diphenylpyrophosphates in the presence of cyclodextrins, studied by Cramer and co-workers (57, 58). These reactions, which result in phosphorylatlon of the cyclodextrins, rather than hydrolysis, are of interest more because of their specificity than because of their rather small rate accelerations. With a-, ~-, and ~’-cyclodextrins, which have different size hole.s in the center, there is a considerable degree of specificity with respect to the particular diphenyl pyrophosphate which will undergo reaction, presumably reflecting the fit of the substrate into the cyclodextrin. Although inclusion compounds are generally studied as solids, they can exist in aqueous solution. Schlenk & Sand (59), for example, have shown that the solubility of C~ to C12 straight- chain carboxylic acids is increased in the presence of a- and fl-cyclodextrins, with which they can form soluble inclusion compounds. It is of interest, in respect to the hypothesis that combination with a substrate may change the conformation of an enzyme, that cyclodextrins show a change in optical rotation on combination with fatty adds. Relatively little experimental information is available in respect to the means by which specificity is introduced in the interaction of small mole- cules with well-defined large molecules. The synthesis of cyanohydrins from optically inactive ketones gives an optically active product when it is catalyzed by optically active quinines; this result requires a stereospecific interaction between the substrate and the optically active catalyst at the moment of cyanide attack (60). A few similar reactions are known (61). studies of Molyneux & Frank (52) on the binding of small molecules polyvinylpyrrolldone in aqueous solution are of interest in this connection. This binding takes place through a "hydrophobic" interaction, and is accompanied by an increase in entropy, because of the decreased number of Annual Reviews www.annualreviews.org/aronline A nn u. R ev . B io ch em . 1 96 3. 32 :6 39 -6 76 . D ow nl oa de d fro m ar jou rna ls. an nu alr ev iew s.o r