11 Enzyme Chemistry By D. GANl Department of Chemistry The University Southampton SO9 5NH 1 Introduction As in previous years,**’ the objective of this review is to highlight some of the more important findings of the recent past with special emphasis on published work which increases our understanding of enzyme mechanism and on recent progress in the design and production of catalytic antibodies. These new synthetic entities are enzymes of a sort and offer enormous potential for the future design of sequence- specific endo-proteases the protein-cutting equivalents of the DNA restriction enzymes or functional group specific hydro-lyases e.g. esterases and phosphatases. Before embarking on this review it might be interesting and informative to read Kornberg’s article relevant to the foundations of Enzyme Chemistry ‘The Two Cultures Chemistry and Bi~logy’.~ The article is adapted from a lecture presented at a meeting of the American Association for the Advancement of Science.2 DNA Polymerase I Research on the mechanism of E. coli DNA polymerase I Klenow fragment was reviewed last year.2 Benkovic and co-workers have now determined the minimal kinetic scheme of the DNA polymerization reaction using short DNA oligomers of defined sequence The key feature of the scheme is a minimal two-step sequence that interconverts the ternary E.DNA;dNTP and E.DNA,+I.PP complexes. The rate is not limited by the actual polymerization but by a separate step which is probably important in ensuring fidelity. Wallace and co-workers have shown that the Klenow fragment accepts both (5R)-and (5s)-5,6-dihydrothymidine triphosphates as substrates for DNA polymeriz- ation.Using exonuclease 111-activated DNA (salmon testes) as a template the two diastereomers ofdihydrothymidine phosphate were incorporated into the synthe- sized DNA in a manner which reflected the ratio of the isomers present in the nucleotide ~001.~ Modak has shown that pyridoxal5’-phosphate (PLP) is a substrate binding site directed reagent for E. coli pol I.6 The covalent attachment of PLP ‘ D. Gani Ann. Rep. Prog. Chem. Sect. B 1985 82 287. D. Gani Ann. Rep. Prog. Chem. Sect. B 1986 83 303. A. Kornberg Biochemistry 1987 26 6888. R. D. Kuchta V. Mizrahi P. A. Benkovic K. A. Johnson and S. J. Benkovic Biochemistry 1987,26,8410.H. Ide R. J. Melamede and S. S. Wallace Biochemistry 1987 26 964. A. Basu and M.J. Modak Biochemistry 1987,26 1704. 279 280 D. Gani causes inactivation of the enzyme and loss of substrate binding ability. The inactiva- tion is dependent upon the presence of a divalent metal ion. The enzyme reacts with 4 mols of PLP in the absence of substrate and with 3 mols of PLP in the presence of substrate. In order to identify the substrate binding region of the protein the PLP-lysine imines were reduced with borotritide and the resulting labelled protein was subjected to tryptic digestion. The peptides were analysed and Lys-758 was identified as the site for PLP binding within the substrate binding region. It was concluded that Lys-758 is the site of binding for the metal chelate form of nucleotide substrates in E.coli DNA pol I. 3 Amino Acyl tRNA Synthetases The mechanism of the reaction catalysed by tyrosyl-tRNA synthetase has been reviewed recently.'.2 Fersht and co-workers have continued to investigate the reaction in detail.'-'' The enzyme catalyses the aminoacylation of tRNA in a two-step reaction (Scheme 1). E + Tyr + ATP + E.Tyr-AMP + PP E.Tyr-AMP + tRNA + Tyr-tRNA + AMP Scheme 1 Fersht has used the gradient (p values) of linear free energy plots obtained by studying the kinetics of mutant enzymes to determine the fraction of the overall binding energy used in stabilizing particular complexes during catalysis. The forma- tion of E-Tyr-AMP from E-TyrSATP results in an increase in binding energy between the enzyme and the side-chain of tyrosine and the ribose ring of ATP.Linear free energy plots of enzymes modified at these positions give the fraction of the binding energy change that occurs on formation of the transition state for the chemical reaction and various complexes. Fersht showed that groups that specifically stabilized the transition state are characterized by p values >> 1. These elegant studies which are derived from a vast amount of kinetic and thermodynamic data for several series of mutant enzymes are accompanied by strong warning^.^ These mainly concern the nature of specific mutations which Fersht has categorized into six groups according to the 'disruptive' effect of the mutation.Following on from these studies Fersht showed that Thr-40and His-45 provide a binding site for the pyrophosphoryl portion of the transition state for the formation of tyrosyl adenylate from tyrosine and ATP and for pyrophosphate in the reverse direction. Deletion of the side chains in mutant enzymes destabilized the transition state by 4.9 kcal mol-' (His-Ma-40) or by 4.1 kcal mol-' (His-Ma-45) confirming the loss of charged hydrogen-bonding interactions? In order to examine the role of His-45 further potentially conservative Gln-45 and Asn-45 mutants were constructed. Both mutant enzymes were debilitated compared to the wild-type enzyme but Gln-45 was more ' A. R. Fersht R. J. Leatherbarrow and T. N. C. Wells Biochemistry 1987 26 6030. * D.M.Lowe G.Winter and A. R. Fersht Biochemistry 1987 26,6038. R.J. Leatherbarrow and A. R. Fersht Biochemistry 1987 26,8524. Enzyme Chemistry 28 1 active and Asn-45 less active than Ala-45 indicating that asparagine causes des- tabilization of the transition state compared to alanine. Fersht considered that the location of the amide NH2 group of glutamine was similar to that of the imidazole E-NH of histidine whereas the amide NH2 of asparagine is comparable to the imidazole 8-NH2 in the wild-type enzyme. The results are used to suggest that the E-NH rather than the 8-NH group of His-45 is involved in transition-state stabiliz- ation in the reaction catalysed by the wild-type enzyme (Figure 1). The unexpected low activity of the Asn-45 mutant is used to illustrate the dangers of introducing groups into positions which cause alternative interactions.Figure 1 (Reproduced by permission from Biochemistry 1987 26 6030) Fersht has also examined heterodimeric tyrosyl-tRNA synthetase by altering the two Phe-164 residues of the wild-type hornodimeric enzyme which are on the axis of symmetry and interact in a hydrophobic region at the subunit interface to aspartate (or glutamate) and lysine." The salt bridges engineered into the hydrophobic subunit interface are weak but sufficient to direct specificity in dimerization when the separately produced inactive monomers Asp-164 (or Glu-164) and Lys-164 are mixed in equimolar amounts. A method for the kinetic analysis of dimeric enzymes which reversibly dissociate into inactive subunits is also presented." Fersht has recently compared the deduced amino acid sequence of Bacillus stearothermophilus valyl-tRNA synthetase with isoleucyl-tRNA synthetase from E.lo W. H.J. Ward D. H.Jones and A. R.Fersht Biochemistry 1987 26 4131. 282 D. Gani coli and reports that there is 25% homology. There are several regions which are highly conserved." One region the HIGH region His-Ile-Gly-His near the N- terminus is conserved in many aminoacyl-tRNA synthetases although Ile is some- times replaced by other hydrophobic residues Leu or Met. In tyrosyl-tRNA syn- thetase the first histidine residue is His-45 which has been shown to form part of a binding site for the y-phosphate of ATP in the transition state for the reaction vide supra.Using site-specific mutagenesis Fersht has shown that for B. stearother-mophilus valyl-tRNA synthetase Thr-52 and His-56 serve similar functions to residue Thr-40 and His-45 in tyrosyl-tRNA synthetase and are involved in binding to the transition state for the reaction but not to either of the substrates valine and ATP or the product.12 Leon and S~hulman'~ have covalently coupled the minor base 3-(3-amino-3- carboxypropyl) uridine of the variable loop of the elongator methionine tRNA to lysine-596 of E. coli methionyl-tRNA synthetase. To date five peptides in the primary sequence of native methionyl tRNA synthetase that are covalently coupled to methionine tRNA have been identified. The 'molecular recognition' of isoleucine and valine by energy dissipation for isoleucyl tRNA synthetase has been reviewed by Cramer and Frei~t.'~ 4 Dihydrofolate Reductase Dihydrofolate reductase catalyses the NADPH-dependent reduction of dihydrofolic (H2F) acid to give tetrahydrofolic acid (H4F).The kinetics of E. coli reductase have been investigated by Benkovic and co-workers and a scheme that predicts the steady-state kinetic parameters and full time course kinetics under a variety of conditions has been pre~ented'~ (Scheme 2). The binding kinetics suggest that during steady-state turnover product dissociation follows a specific preferred pathway in which H4F dissociaton occurs after NADPH replaces NADP+ in the ternary complex. The dissociation is thought to be the rate-limiting step at low pH because k, = V,, .The transfer of hydride from NADPH to H2F shows a deuterium isotope effect of 3 and is rapid essentially irreversible and pH-dependent; the pK = 6.5 reflects the ionization of a single group in the active site. The scheme accounts for the apparent pK = 8.4 observed in the steady state which may be attributed to a change in the rate-determining step from H4F release at low pH to hydride transfer at high pH. The role of Phe-3 1 was assessed by preparing Tyr-3 1 and Val-3 1mutant enzymes.16 Phe-31 is a strictly conserved residue which interacts with the pteroyl portion of H2F in a hydrophobic pocket. The kinetic properties of the mutant enzyme were similar to those of the wild-type enzyme. The rate of hydride transfer was slightly decreased and the rate of H4F dissociation increased in the mutant proteins so that V,, was increased twofold over the wild-type enzyme.T. J. Borgford N. J. Brand T. E. Gray and A. R. Fersht Biochemistry 1987 26 2480. 12 T. J. Borgford T. E. Gray N. J. Brand and A. R. Fersht Biochemistry 1987 26 7246. l3 0. Leon and L. H. Schulman Biochemistry 1987 26 1933. 14 F. Cramer and W. Freist Acc. Chem. Res. 1987 20 79. l5 C. A. Fierke K. A. Johnson and S. J. Benkovic Biochemistry 1987 26 4085. 16 J. Chen K. Taira C. D. Tu and S. J. Benkovic Biochemistty 1987 26 4093. Enzyme Chemistry 5 pM-' 11 50 I1 3 ENADPH -EH~F-950 H~F 0.6 1.7 20 g M-' L ENADPH ENADPH -85 . EH4F 1.4 E-H4F 3.5 12.5 8 MM-' 25 gM-' pH-Independent kinetic scheme for E.coli dihydrofolate reductase at 25 "C rate constants s-l Scheme 2 In order to explore the substrate protonation mechanism of the reductase Howell and co-workers constructed a double mutant in which the usual proton donor Asp-27 was replaced by serine and in which a nearby threonine residue 113 was replaced by the alternative proton donor glutamic acid.I7 The double mutant was threefold more active than the single mutant Ser-27,I8 but k,, was 25-fold lower than for the wild-type enzyme. It was concluded that the double mutant does not stabilize the transition state for the hydride reduction through initial protonation at 0-4 of the substrate" and that the most likely site for protonation is N-1. wild-type D27S + T113E mutant Scheme 3 Mutant enzymes with specific amino acid replacements in two helices and in two strands of the central @-sheet of dihydrofolate reductase have also been used to probe the folding and stability of the E.coli 5 PLP-Dependent Enzymes Site-specific mutagenesis has now been extensively used to study the mechanism of aspartate aminotransferase. Substitution of Arg-292 which normally binds the p-and y-carboxyl groups of aspartic and glutamic acid respectively for aspartic acid gave an a-transaminase capable of acting upon ornithine and arginine albeit slowly.22 Replacement of the active site Lys-258 by alanine gave an inactive enzyme '' E. E. Howell. M. S. Warren C. L. J. Booth. J. E. Villefranca and J. Kraut Biochemisrry 1987 26 8591. E. E. Howell J.E. Villefranca M. S. Warren S. J. Oatley and J. Kraut Science 1986 231 1123. 19 J. E. Gneady Biochemistry 1985 24 4761. 20 N. A. Touchette K. M. Perry and C. R. Matthews Biochemistry 1986 25 5445. 'I K. M. Perry J. J. Onufper N. A. Touchette C. S. Herndon M. S. Gittelman C. R.Matthews J. Chen R. J. Mayer K. Taka S. J. Benkovic E. E. Howell and J. Kraut Biochemistry 1987 26 2674. 22 C. N. Cronin B. A. Malcolm and J. R. Kirsch J. Am. Chem. Soc. 1987 109 2222. 284 D. Gani Tyr-70 Lys-258 I / + H-0 HZN Arg-292 ...02c~C02-+Arg-386 Figure 2 which acted as an oxaloacetate decarboxylase when pyridoxamine S'-phosphate (PMP) was used as the coenzyme.23 Unlike the mitochondria1 and E. coli wild-type enzyme the Ala-258 E. coli mutant is unable to labilize the C-4' pro-S hydrogen of PMP.24Kirsch has shown that the replacement of Tyr-70 (the other possible candidate of the active-site base) by alanine gives an enzyme which retains 17% of the activity of the wild-type enzyme.Collectively these results strongly support the notion that Lys-258 is the proton abstracting-donating group in physiological transamination reactions. Kuramitsu et al. have shown that the replacement of Lys-258 by arginine gives an enzyme which is 3% active,25 and Metzler and co-workers have shown that 3'-0-methyl- pyridoxal 5'-phosphate functions as a poor coenzyme for aspartate aminotransferase- catalysed transamination half-reactions.26 Walsh has studied the time-dependent inactivation of PLP-dependent 1 -aminocyc- lopropanecarboxylate deaminase and alanine racemase by 1 -aminocyc- lopropanepho~phonate~'( 1).The physiological role of the deaminase is to catalyse H$.@ ACPP -0.,P=0 -0 (1) the formation of a-ketobutyrate and ammonia from 1-aminocyclopropanecarboxylic acid' and the role of alanine racemase is well established.',2 The enzymes were inactivated with KM/Ki ratios of 500 and 2000 respectively and in each case inhibition was characterized by a slow-binding slow-dissociating behaviour. Analysis 23 B. A. Malcolm and J. F. Kirsch Biochem. Biophys. Res. Cornmun. 1985 132 915. 24 S. Kochhar N. L. Finlayson J. F. Kirsch and P. Christen J. Biol. Chern. 1987 262 11446. 25 S. Kurarnitsu Y. Inoue S. Tanase Y. Morino and H. Kagarniyama Biochem.Biophys. Res. Commun. 1987 146 416. 26 V. J. Chen D. E. Metzler and T. W. Jenkins J. Biol. Chem. 1987 262 14422. 27 M. D. Erion and C. T. Walsh Biochemistry 1987 26 3417. Enzyme Chemistry k, E + ACPP F= E.ACPP E = ACPC deaminase k2 ki = 5pM Scheme 4 k, E + ACPP & E.ACPP eE.ACPP* kz k4 E = alanine racemase k = 8mM KF = 2pM Scheme 5 of the pre-steady-state kinetics revealed a kinetically detectable intermediate E.1 complex in the inhibition mechanism for the racemase but not for the deaminase. 6 Other Enzymes Blundell and Szelke and co-workers have recently reported on the modes of binding of transition-state inhibitors at the active-site of the aspartic protease endothiapep- sin.28 The most potent inhibitor H261 contains the grouping -CHOH-CH2-in place of the scissile carboxamide functionality of the substrate.The compound binds so that the hydroxyl group occupies a spatial position similar to that occupied by the small molecule (probably water) which is hydrogen bonded to the two aspartate residues in the native protein. As yet it is not known which hydroxyl group Figure 3 (Reproduced by permission from Biochemistry 1987 26 5585) T. L. Blundell J. Cooper S. I. Foundling D. M. Jones B. Atrash and M. Szelke Biochemistry,1987 26 5585. 286 D. Gani carbonyl-derived or water-derived in the transient geminal aminodiol is mimicked by the inhibitor. and also Polgar3’ have recently discussed the mechanism of aspartic proteases while COI-VO~~~ and co-workers and Pals32 and co-workers have investigated the kinetics of the inhibition of human renin and other aspartic proteases using statine and hydroxyethylene containing peptide analogues.Knowles has studied enzyme relaxation in the reaction catalysed by triosephos- phate is~merase.~~ Using the tracer perturbation method of Britton which involves measuring the time-dependent distribution of radio-labelled substrate and product after the system is perturbed from an initial equilibrium by the addition of a relatively large amount of unlabelled material it was shown that the enzyme exists in two forms. One form binds and isomerizes (R)-glyceraldehyde 3-phosphate and the other form binds and isomerizes dihydroxyacetone 3-phosphate. The rate of the interconversion of free (unbound) forms of the enzyme is lo6s-’.Knowles points out that it is improper and possibly erroneous to presume that the rates of the interconversion of the unliganded substrate-binding and product-binding forms of an enzyme ( k4 and kP4) are not kinetically significant. ‘SOH Y OP Scheme 6 Gani and co-~orkers~~-~~ have shown that in the presence of ammonia 3-methylas- partase converts a range of halogeno- and alkyl-fumaric acids into the corresponding 3-substituted L-aspartic acids via anti-addition. Further mechanistic studies of the enzyme revealed that the deamination of the physiological substrate (2S,3S) -3-methylaspartic acid showed a primary isotope effect upon both V,, and V of 1.7 while the substrates (2S,3S)-3-ethylaspartic acid and aspartic acid showed ”( V/K) and ”(V) effects of 1.15 and 1.0 respectively.The results were rationalized in terms of transition-state binding where the transition states for non-physiological substrates would interact with the active-site weakly. It was argued that this would cause C-N bond cleavage to become kinetically more important than for the physiological substrate and thus obscure the deuterium isotope effects. 29 L. H. Pearl FEBS Lett. 1987 214 9. 30 I-. Polgar FEBS Left. 1987 219 1. 3’ F. Cumin D. Nisato J. Gagnol and P. Corvol Biochemistry 1987 26 7615. 32 W. M. Kati D. T. Pals and S. Thaisrivongs Biochemistry 1987 26 7621. 33 R. T. Raines and J. R. Knowles Biochemistry 1987 26 7014. 34 M. Akhtar M.A. Cohen and D. Gani Tetrahedron Lett. 1987 28 2413. 3s M. Akhtar M. A. Cohen N. P. Botting and D. Gani Tetrahedron 1987 43 5899. 36 N. P. Botting M. A. Cohen M. Akhtar and D. Gani J. Chem. SOC.Chem. Commun. 1987 1371. Enzyme Chemistry Enz A H H u R =H,Et Scheme 7 7 Catalytic Antibodies Over the past two years the ability of antibodies to catalyse chemical reactions in an almost analogous fashion to enzymes has been realised. Antibodies bind biological macromodels as well as small synthetic molecules with enzyme-like affinities and specificities. Also antibodies can be generated selectively towards any molecule of theoretical interest. The large rate enhancement observed for enzyme-catalysed reactions over and above the rate for the corresponding chemical reactions is ascribed to the ability of enzymes to bind the transition state for the reaction more tightly than the reactants or products.Hence the activation energies for chemical steps are lowered compared to reactions where the medium interacts essentially equally with each of the species. This analysis of enzymic catalysis led to the advent of transition-state inhibitors which are now well established as potent tight-binding competitive inhibitors. With this background and the availability of transition-state inhibitors Lerner and co- workers and others investigated the possibility of raising antibodies to transition-state analogue haptens. Lerner and ~o-workers,~~~~~ chose to create and investigate esterase activity and therefore prepared phosphonate monoesters [e.g.(2)] for incorporation into antigens. It was reasoned that the transition state for potential ester hydrolysis with its developing negative charge would be accurately mimicked by the tetrahedral phosphate monoester anion. The phosphonate was conjugated with keyhole limpet hemocyanin to produce the actual antigen which was used to immunize mice. The monoclonal antibodies were prepared and were assayed for esterase activity using compound (2). Some of the antibodies reacted stoicheiometrically to release the fluorescent phenol (7-hydroxycoumarin) and under basic condition acted catalytically (Scheme 8). The observed reaction rates were very small but the haptenic phosponate monester (2) was found to be a potent competitive inhibitor.It is proposed that the antibody catalyses transacylation of itself but that deacylation is not catalysed. 37 A. Tramontano K. D. Janda and R. A. Lerner Roc. Nafl.Acad. Sci. USA 1986,83 6736. 38 A. Tramontano K. D. Janda and R. A. Lerner Science 1986 234 1566. 288 D. Gani 0 H R=COMe or N-Keyhole limpet hemocyanin 0 Schultz and co-w~rkers~~~~~ have also used tetrahedral phosphonate transition- state analogues to raise antibodies. The antibodies catalyse carbonate hydrolysis at appreciable rates and show the expected kinetic properties in the presence of transition-state analogue inhibitors. Finally Benkovic4' has used a cyclic phosphon- ate ester (4) to elicit the production of a S-lactone synthase antibody (Scheme 9).Higher pH CFSCONH HOa. I 3 COzH + IgG Scheme 8 39 S. J. Pollack J. N. Jacobs and P. G. Schultz Science 1986 234 1570. 40 J. Jacobs and P. G. Schultz J. Am. Chem. Soc. 1987 109 2174. 41 A. D. Napper S. J. Benkovic A. Tramontano and R. A. Lerner Science 1987 237 1041. Enzyme Chemistry The antibody catalysed the cyclization of the racemic hydroxyester to give the lactone in 94% enantiomeric excess. The rate acceleration factor was 167. S-o OPh ___ 24811 H NHAC NHAc OH (*I) 94% e.e. kc,,lk,,c,t = 167 Scheme 9 Clearly the full potential of catalytic bodies is enormous and given that only half a dozen or so papers have been published so far the immediate future promises to be extremely exciting.