THE CHEMISTRY AND PHYSICS OF ENZYME CATALYSIS S. Doonan, B.Sc., Ph.D. . . . . . . .. . . . . .. 117 Deportment of Chemistry, University College, Gower Street, London W C I Enzyme structure . . Mechanism and catalysis . . . . . . .. . . .. .. 123 Possible factors involved in enzyme catalysis . . . . . . . . 127 Studies of individual enzymes . . . . . . . . .. . . 131 Lysozyme, 13 1 Chymotrypsin, 136 Carboxypeptidase A, 138 Conclusion . . . . . . . . .. .. . . . . .. 140 Acknowledgments . . .. . . .. .. . . .. . . 141 References . . .. . . . . . . .. . . .. .. 141 Ten years ago, in conclusion to a review of current knowledge concerning the mechanism of enzymic catalysis, Lumryl stated that ‘there is much ill-defined swamp and little firm ground’. Progress in our understanding of the mode of action of enzymes in this decade has been spectacular and it is now possible, for a small number of enzymes, to describe both the mechanisms of the reac- tions which they catalyse and the factors involved in the catalysis of these reactions : the necessary distinction between mechanism and catalysis will be discussed below.In the present review, the nature of the problem of enzyme catalysis will be defined, some of the more important factors which have been proposed to account for the activity of enzymes will be discussed and, finally, the results of recent chemical and physical investigations of selected enzymes will be described briefly to illustrate these general considerations. More extensive coverage of these topics may be found in recent reviews by Doonan, Vernon and Banks2 and Koshland and Neet.3 ENZYME STRUCTURE Enzymes are globular proteins and hence are polymers constructed from the twenty L-a-amino acids whose characteristic side chains are listed in Table 1, plus the imino acid proline. The amino acids are joined together by their amino and carboxyl functional groups to produce a polypeptide chain of the type shown, where R1, R2 etc. represent the distinguishing side chains: I I I H2NCHCONHCHCO... . . NHCHCONHCHCOOH I Rl R n R2 R”- I A free amino group occurs at one end of the chain (the N-terminus) and a free carboxyl group at the other (the C-terminus). In addition to amino-acid residues, some enzymes contain either a metal ion or a small organic molecule which is essential for enzymic activity. Examples are the zinc ion in carboxypeptidase and the pyridoxal-5’-phosphate cofactor of aspartate aminotransferase ; these enzymes are discussed on Doonan 117 9 pages 138-140 and 124-126 respectively.However, even in these cases, it is the protein part of the enzyme which is primarily responsible for its catalytic capabilities and it is with the relationship between protein structure and enzymic catalysis that this review is concerned. ~~ ~~ ~ R Side chain H - C H3- cH3>CHCH2- CH3 CH3CH2 >CH - CH3 HOCH, - CH3CH - I OH Table 1. Structures of the side chains of the amino acids occurring in enzymes NH2CHCOOH R I Amino acid Glycine Alanine Valine Leucine lsoleucine Serine Th reon i ne Cystine Cysteine (Half cystine) Methionine NHzCHCOOH R I Amino acid Phenylalanine Tyrosi n e Tryptophan Aspartic acid Asparagine Glutamic acid Glutamine Lysine Arginine Histidine SCH, - I SCH2 - HSCH2 - CH3SCH2CH2 - The imino acid proline has the structure 'i H Protein structure is conveniently considered at two levels.Since the struc- tures of the constituent amino acids are known, and it can be assumed that the amino-acid residues are joined by peptide bonds, then once the numbers of residues of each type and the sequence in which they occur in the poly- peptide chain are known, the covalent structure of the protein molecule is completely specified. This is usually referred to as the primary structure.More important for the correlation of structure and activity, however, is the HO 0 CH2 - a ! - c H 2 - - CH2COOH - CHzCONHz - CHlCH2COOH - CH2CH2CONHl - CH2CHzCH2 N HC N H2 II +NH CH; /- 0 COOH 118 R Side chain R. I.C. Reviews conformation of the polypeptide chain and the arrangements of the individual amino-acid side-chains in space. These two aspects of protein structure will be considered separately. The chemical methods for 'the determination of the primary structures of proteins have been exhaustively reviewed.4 In outline, the basis of currently used methods is the Edman degradation process5 in which the N-terminal amino acid is removed from the polypeptide chain by the sequence of reac- tions shown below.The terminal amino acid is isolated and identified as the I I NHPh PI1 - N = C = 5 + NH2CHCONHCHCO.. . .- P" 9 Ph - NHCNHCHCONHCHCO . . . . I I R2 Rl HN, I R2 II s R, R,-CH -CO I \ /NPh C I I 3 phenylthiohydantoin derivative, whilst the remainder of the peptide chain is left intact. Repeated application of the degradation process yields the primary sequence of the polypeptide. In practice, technical difficulties limit the application of this technique to polypeptides containing about 20-30 amino-acid residues. For example, the structures of the 18-residue peptide apamin6 and the 26-residue peptide melitin,7 both isolated from bee venom, have been determined by repeated application of the Edman degradation technique.Since the amino-acid chains of enzymes are very much longer than those of the peptides mentioned above, and may contain from 100 to 1000 amino- acid residues, their primary structures are not immediately determinable by application of the Edman technique. In general, the polypeptide chain is first degraded into smaller peptides, these smaller fragments are separated from one another, and their primary structures are determined. The problem of the order in which these smaller fragments occurred in the native protein must then be solved; this is done by the method of overlaps, illustrated in Fig. 1. The chain is represented by a series of amino-acid residues (R11, R12 etc.) and the dotted lines represent points of hydrolysis.The first hydrolysis yields a series of peptides PI, PZ etc. whose structures may be determined using the Edman method. It is then necessary to decide in which order these 119 Doonan I Fig. I . The method of overlaps applied t o the determination of the primary structures of proteins. peptides occurred in the original structure; that is, whether the original sequence was PlP2P3P4.. . or P I P ~ P ~ P ~ . . . or P3P4PlP2.. . and so on. A second hydrolysis is carried out, under different conditions from the first, to give a new set of peptides Pi, Pi, Pi etc. These new peptides are examined for portions occurring in two or more of the first set. For example in Fig. 1, peptide Pi will contain amino acids both from PI and P2.Similarly Pi con- tains parts of P2 and P3 and Pi contains parts of P3 and P4. Hence the original sequence must have been P1P2P3P4 . . . . In practice, three or four different methods of hydrolysis are usually required to provide all the overlaps. The methods of hydrolysis used depend on the selectivity of enzymes which catalyse the hydrolysis of peptide bonds. For example, the enzyme trypsin catalyses the hydrolysis of peptide bonds involving the amino acids lysine and arginine whilst the similar enzyme chymotrypsin is most active in the hydro- lysis of peptide bonds involving the aromatic amino acids phenylalanine, tyrosine and tryptophan; an account of methods used in the hydrolysis of proteins is given in section V of reference 4.Primary sequence determinations are technically difficult and require a large amount of starting material. For these reasons, complete structures are known for only a few rather small enzymes which are available in large quantities. An example is the enzyme chymotrypsin which is disussed in detail later (pp. 136-138). The primary sequence has been determined by Hartley8 and by Keil and S6rm.g The primary structure, as presented by Matthews et aZ.10 is given in Fig. 2; the amino acids are represented by the first three letters of their names except in the cases of isoleucine, asparagine and gluta- mine which are abbreviated to ile, asn and gln respectively. The enzyme con- tains 241 amino-acid residues (the numbering system in Fig.2 is that for a 120 R. I.C. Reviews NI Fig. 2. The primary structure of a-chymotrypsin (from reference 10). c precursor, chymotrypsinogen, which has four more amino-acid residues than chymotrypsin) in three polypeptide chains. The chains are joined together by disulphide bridges from the amino acid cystine (Table 1); similarly there are three intra-chain disulphide bridges. Primary sequences such as that in Fig. 2 tell us very little about the activity of enzymes. For example, chemical evidence (p. 136) has shown that two important amino-acid side-chains for the activity of chymotrypsin are those of histidine-57 and serine-195. These residues are widely separated in the primary structure and it is not possible to see from the sequence given how they contribute to the activity of the enzyme. In order to solve this problem it is necessary to determine the conformation of the peptide chain and the orientations of the amino-acid side-chains in space.Information of this sort can only be obtained by x-ray diffraction analysis. Doonan 121 Fig. 3. Conformation of the peptide chain of a-chymotrypsin. The determination of the structures of small molecules by x-ray crystallo- graphy is now largely a routine matter, but this is not the case in protein structure analysis due to the size and complexity of protein molecules. Indeed, the classical methods of x-ray diffraction analysis are not applicable in these cases. The method of isomorphous replacement was developed by Kendrew and his associates to overcome these problems, and was used to determine the complete three dimensional structure of the oxygen-carrying protein myoglobin.16 Lack of space precludes even an outline description of x-ray diffraction studies of proteins, but the subject has been extensively reviewed.l2> l 3 An example of the type of results obtained is shown in Fig.3. This figure R.I.C. Reviews 122 gives a simplified diagram of the structure of chymotrypsin determined by Matthews, Sigler, Henderson and Blow.lo Only the conformation of the peptide chain is shown together with the side chains of residues histidine-57 and serine-195 (which in this structure carries a tosyl group). It can be seen that the conformation of the peptide chain brings these two important amino- acid side-chains into close proximity.A detailed discussion of the relationship of this structure to the activity of chymotrypsin is given later (pp. 136-138). At the present time complete three-dimensional structures of only a very small number of proteins have been determined and very few general observa- tions about the relationships between primary and higher order structures can be made. In general, the non-polar amino-acid side-chains are found in the interior of the molecule where they are well shielded from the solvation shell of the protein; the polar side chains are on the surface of the molecule and interact directly with the solvation shell. The main forces stabilizing the struc- ture appear to be van der Waals’ interactions between non-polar side chains, and hydrogen bonds which are formed between the carbonyl oxygen and amide hydrogen atoms of peptide bonds which are brought into the correct juxtaposition by the folding of the polypeptide chain.Determination of the three-dimensional structure of a protein requires prior knowledge of the primary structure and since the latter is more easily determinable, it would obviously be of great advantage if the three-dimensional structure could be calculated from the primary structure. It is widely held that a connection between these two levels of structure must exist. The genetic material of any cell, that is, the nucleic acids which it contains, is thought to specify only the primary sequences of the proteins which the cell produces.Once the proteins are synthesized, they then fold spontaneously into the correct three-dimensional configuration. If it is assumed that the configura- tion achieved is that of maximum thermodynamic stability then, in theory at least, it is possible to calculate this configuration. However, recent evidence2 tends to suggest that the formation of a particular three-dimensional structure is determined by kinetic rather than thermodynamic factors. If this is so, then there is no possibility, in the foreseeable future, of predicting the complete structure of a protein from its known primary sequence. This being the case, the structural information which is required for the interpretation of enzyme activity must be obtained by the difficult and time- consuming techniques outlined above, and the rate at which new structures are forthcoming is likely to be small.However, the few structures which have been reported provide a wealth of information for the correlation of enzyme structure with activity, the problem to which the remainder of this review is devoted. MECHANISM AND CATALYSIS In outline, the events occurring during an enzyme catalysed reaction may be described in terms of three distinct stages, namely: (a) combination of the molecule or molecules (usually termed substrates) which are to undergo reaction with a particular region on the enzyme surface (the active site) to form the so-called Michaelis complex ; (b) covalency changes in the substrate ; (c) diffusion from the protein of the products of reaction.Given such a Doonan 123 sequence of events, then the term mechanism should properly be used to describe the interactions which the substrate makes with amino-acid side- chains of the protein to form the Michaelis complex, the number and struc- tures of the intermediates lying on the reaction pathway between the Michaelis complex and the formation of products, and the order in which the products diffuse from the enzyme surface. A simple, generalized scheme for an enzyme catalysed reaction is given below, where E, S, and A and B represent the enzyme, substrate and products respectively. Mechanism then describes the structures of the species ES, EAB and EA, the interactions which the substrate and products make with the protein in these species, and rates of the unit steps in the reaction sequence.In these terms, the mechanism of action of a variety of enzymes is, at least in part, understood; the qualification is necessary since in no case are the values of all the rate constants associated with unit steps in the process known. It must be emphasized that such a description of an enzyme catalysed reaction tells us what happens but not why these events occur at rates which are so much greater than those of similar processes in non-catalysed reactions ; this is the problem of catalysis. Catalysis is not simply derivable from mechanism and the information required to explain the catalytic effect of an enzyme is not included in a description of its mechanism of action.This distinction between mechanism and catalysis may be illustrated by a specific example, namely the enzyme aspartate aminotransferase which catalyses reaction 2. The enzyme from pig heart muscle has been purified and COOH I COOH I COOH I CHNH, co COOH I CHNH, I 1 CH2 + I CH2 CH2 I I I COOH aspartic acid co e ‘ + ( 3 4 2 I CH2 I COOH glutamic acid CH2 I COOH oxaloacetic acid CHO COOH a - ketoglutaric acid characterized,14J5 and a complete analysis of the steady-state kinetics of the enzyme catalysed reaction has been carried 0ut.l6J7 This enzyme depends for its catalytic activity on the presence of one or other of the vitamin B g deriva- tives pyridoxal-5’-phosphate, I I 0 or pyridoxamine-5’-phosphate.124 R. I.C. Reviews 0 II H o Z c H 2 ~ O H I These species are tightly bound to the enzyme surface and are referred to as cofactors for the enzyme. There is abundant evidence (for a summary see reference 18) that four intermediates, excluding Michaelis complexes, lie on the reaction pathway and that the mechanism of the reaction may be repre- sented by equations 3 and 4, where E-CHO and E-CHzNHz represent the k2 + H 2 0 - H,O - L kl COOH I E-CHO + NH2CH I CH2 E - CHl- N =C I COOH + H2O COOH I d k5 Y k6 - H,O - H,O - k 7 k8 +H,O L +H20 - k l l A I CH2 I CH2 CH2 I COOH I COOH E - CYNH, + CO I I COOH COOH I E - CH=N -CH I CH2 I k I2 -H,O CH2 I COOH Doonan CH2NH2 CH3 COOH k3 I k4 CH2 E-CH=N-CH I I COOH COOH I E-CH2NH2 + CO I CH2 I COOH k9 L 1 kI0 COOH E - CHO + NH2CH I COOH I CH2 E - CH2- N=C I CH2 I I COOH I CH2 I CH, I COOH 125 Factor ~~ Table2.Comparison of rate constants in the enzyme catalysed and non-enzymic transamina- tion reactions Enzymic reaction Rate constant Value (s-1) Non-enzymic reaction 2.0 x 10-6 250 I100 I00 1.25 x los - 1 . 1 x 107 1 . 1 x 109 - 9.1 x 10-6 0.8 x 10-6 k3 k4 k9 k o 900 forms of the enzyme containing pyridoxal- and pyridoxamine-5’-phosphates respectively. The rate determining steps in reaction sequences 3 and 4 are the prototropic shifts characterized by rate constants k3, k4, kg and klo; values of these four rate constants have been determined2J6J7 and are given in Table 2.After the initial demonstration by Snell and his co-workerslg of a trans- amination reaction when pyridoxal was incubated with amino acids in the absence of the enzyme, an extended investigation of the non-enzymic reaction between the natural substrates and cofactors of aspartate aminotransferase was carried out.16 The reaction mechanism in this model system parallels precisely that of the enzyme catalysed reaction; that is, the sequence of events is the same as that given in equations 3 and 4 but with E-CHO and E-CH2NH2 replaced by the free cofactors.Moreover, the values of nine of the 12 rate constants in the reaction sequence were evaluated; the values of k3, kg and klo are given in Table 2. Comparison of the figures in Table 2 yields information of considerable interest since they provide a direct measure of the catalytic effect of the protein part of aspartate aminotransferase on individual steps in the reaction seq- uence; these catalytic factors can be seen to range from 107 to 109. Hence the situation arises where the presence of the catalytic protein enhances the rates of processes which are mechanistically similar by very large factors. There is nothing in the formulation of the mechanism of the enzyme-catalysed reaction given in equations 3 and 4 which accounts for such a rate enhancement; thus the problem of catalysis is not solved by a statement of mechanism.The description of the mechanism of enzyme catalysed transamination given above is incomplete insofar as the groups in the protein which interact with the bound substrate have not been specified. It might, then, be assumed that once these groups have been identified, the origin of the catalytic effect of the protein will become obvious from the complete mechanistic description. This, however, is not the case. If attention is focused on the prototropic shift in reaction sequence 3, it may reasonably be supposed that the process involves suitably positioned acidic (-AH) and basic (-B) amino-acid side- chains in the protein which function as proton donors and acceptors.Accord- ing to the currently accepted formalism of enzymology the process would then be written as in Fig. 4, where the continuous line represents the enzyme surface, @ denotes the phosphate group and R is the substrate side chain. This more complete formulation of the reaction mechanism does contain 126 R. I . C. Reviews L 1 Fig. 4. The rate-determining step in enzymic transamination. some information concerning catalysis but only in the relatively trivial sens.: that the reaction is acid-base catalysed; it does not give any indication as to why the rate enhancement factors are so large. General acid and general base catalysis are well known in physical organic chemistry but the rate enhance- ments observed rarely exceed a factor of 10.For example, in the imidazole catalysed hydrolysis of acetyl tyrosine ethyl ester, the rate of the general base catalysed hydrolysis was found to be only tenfold greater than that of the non-catalysed hydrolysis ;20 this rate enhancement is several orders of magni- tude less than that observed for the catalysis of hydrolysis of this substrate by the enzyme chymotrypsin (see later). In general, the problem remains as to why acid-base catalysis by the side-chain groups in the active site of enzymes is so remarkably effective. One explanation of this effectiveness, which has frequently been assumed, arises from a misconception of the significance to be attached to formal mechanistic diagrams such as that in Fig. 4.Such diagrams are frequently interpreted in terms of a synchronous process (the push-pull hypothesis21), that is, in terms of a concerted proton donation by -AH and abstraction by -B. Large catalytic factors are then assumed to arise from the acid and base groups acting in concert rather than in a sequential fashion. In fact the arrows in Fig. 4 simply indicate which bonds are made and broken in the passage from one enzyme-substrate intermediate to the other and contain no informa- tion about the timing of these events. Concerted acid-base catalysis has not been conclusively demonstrated for any enzyme-catalysed reaction, and even if such processes were concerted there is no reason to suppose that acid and base groups acting together would be more effective in catalysis than the same groups acting sequentially.POSSIBLE FACTORS INVOLVED IN ENZYME CATALYSIS In the previous section a distinction was drawn between mechanism and catalysis in enzymic reactions and in particular it was argued that acid-base catalysis cannot account for the large rate enhancement effects of enzymes. A similar point must be made in connection with the possible catalytic signifi- cance of formation of covalent enzyme-substrate intermediates. This is a feature of the mechanism of action of a variety of enzymes,3 but again one which cannot contribute significantly to their catalytic effects. An example should help to clarify this point. Doonan 127 Several enzymes have been studied (chymotrypsin, trypsin, elastase, throm- bin, subtilisin) which catalyse the hydrolysis of ester and amide linkages (reaction 5, where X is -OR’ or -NHR’).There is good evidence,2J2 particu- larly in the case of chymotrypsin, that the reaction pathway includes an inter- RCOOH + HX RCOX + H,O- i BH R R ( 5 ) mediate formed between the acyl group of the substrate and the hydroxyl group of a serine residue in the protein. The process may be written as in is B: - X - B H + +H20 k 2 histidine residue R (see later). has equation a basic 6 group where which -OH represents been o - c = o the identified R I serine side as the chain imidazole r hydroxyl 0 - H + side + group R C = O OH 1 I chain and of -B (6) a A hypothetical alternative scheme, still involving group B but not covalent intermediate formation is shown in equation 7.Here, the reaction sequence involves direct attack of a water molecule on the substrate. It is evident that, for the formation of a covalent intermediate to contribute to catalysis, both kl and k2 in equation 6 must be much greater than k3 in equation 7; this could only be the case if the serine hydroxyl group was both a better attacking group than water and a better leaving group than -X (-OR’ or -NHR’). In general, serine hydroxyl groups in proteins are not good nucleophiles or leaving groups, and, given the reaction mechanism in equation 6, the problem of catalysis by chymotrypsin and similar enzymes remains.Another possible catalytic factor which may be proposed on the basis of reaction schemes such as those in Fig. 4 and equations 3,4, 6 and 7 is the so- called proximity effect.3 These reaction sequences involve formation of complexes between either two catalytic groups and one substrate molecule or one catalytic group and two substrate molecules. It might be supposed that the formation of such ternary complexes would lead to an improved collision probability (i.e. an increased relative concentration of reactants) with con- comitant rate enhancement. An analogous effect is well known in physical organic chemistry in the phenomenon of ‘intramolecular catalysis’.2392* To give a single example, hydrolysis of the propanol thiolester of 4-(4’-imidazoly1)- butyric acid proceeds with intramolecular catalysis by the imidazole group (equation 8) and is about 3 x lo6 times as fast as similar reactions in which 128 R.I.C. Reviews + PrSH imidazole groups do not p a r t i ~ i p a t e . ~ ~ The analogy between such reactions and those occurring on the surface of an enzyme is obvious, and it might be thought that the proximity effect would lead to similar large rate enhance- ment factors in enzyme-catalysed reactions. Koshland26~~7 has calculated, however, that the proximity effect will lead to a decrease in rate for many enzyme-catalysed reactions and for those cases in which rate enhancement is to be expected the calculated factor is far less than that which is observed in practice. Koshland and Neet3 explain this effect in terms of the increased relative concentrations of substrates and reactive groups being partially or completely offset by the very low absolute concentration of enzyme active sites. Thus, in general, although proximity effects may contribute significantly to catalysis for some enzymes the effect is not of dominant importance.It is clear from the examples given that a complete mechanistic description of an enzyme-catalysed reaction provides information about the types of catalytic processes involved but does not answer the question as to why the catalysis is so great compared with that observed in non-enzymic model systems. The problem is to explain why the mutual reactivity of a particular set of substrates and catalytic groups is so much greater on the enzyme surface than in free solution.For example, in enzymic transamination (Fig. 4) why are the groups -AH and -B so effective in donating and removing protons from the bound substrate in the environment provided by the active site of the enzyme. Similarly, in the case of catalysis by chymotrypsin (equation 6) an explanation must be provided for the enormous increase in nucleo- philicity of the serine hydroxyl group in the enzyme-substrate complex compared with the nucleophilicity of hydroxyl groups in free solution; in addition the possibility that the carbonyl group of the substrate is activated in the enzyme-substrate complex must be considered. Recent evidence2 suggests that the fundamental factor in enzyme catalysis is the physical environment of the active site and the effects which a particular environment produces on the properties of catalytic groups and substrates in the active site.Two main ways in which environmental effects could modify the behaviour of substrates and catalytic groups may be proposed, both of which have analogies in physical organic chemistry; these are strain and solvent effects. The strain effect1928 arises from attractions (ionic, non-polar and hydrogen bonding) between the substrate and protein side chains which physically distort the substrate molecule, thus enhancing its reactivity. That large rate enhancements can arise from such physical distortions is well known. For example, Haake and Westheimerzg have shown that the base catalysed hydro- lysis of the monoanion of ethylene phosphate (equation 9) proceeds 107 times Doonan 129 C H I - 0 CH20PO:- 0 (9) +OH--- 0- CH2- 0 CH,OH (10) , P , f OH-- 1 CH3-0 ' 4 0 C H 3 - 0 I CH30POi-+CH30H CH ,Cl + I- (1 1) (CH3)zNCHO ' + ,P, 0- faster than the hydrolysis of the corresponding anion of dimethylphosphate (equation 10); this rate difference was attributed solely to bond strain in the 5-membered ring system of ethylene phosphate.Allowing for differences in position of bond cleavage, a rate difference of at least 108 was calculated for hydrolysis of the P-0 bond in the two cases. It is easy to see how such an effect could contribute to enzyme catalysis and its possible contribution to catalysis by the enzyme lysozyme will be discussed later.The second environmental effect, and the one which may prove to be by far the most important factor in enzyme catalysis, depends on the ability of enzymes to exclude water from their active sites on formation of enzyme- substrate complexes. Thus the catalytic groups and bound substrates are removed from the bulk solvent (water) and transferred to an environment which is a function of the particular amino-acid side-chains forming the active site region; this environment can then be 'tailor-made' to suit the requirements of the particular reaction. Again, recourse to physical organic chemistry provides examples of the profound rate changes which frequently accompany changes in solvent systems.Water, and other hydrogen-bonding solvents, are poor media for many organic reactions due to their ability to solvate ionic species. For example, an anion in aqueous solution is strongly solvated so that the negative charge is effectively distributed over the ion and solvation shell ; thus its nucleophilicity is drastically reduced. In dipolar aprotic solvents, however, where solvation is much lower, an anion is a much stronger nucleophile.30 The S N ~ reaction shown in equation 11 has been studied in a variety of solvents31 and some relative rate data are given in Table 3. CH, I +CI-- Table 3. Relative rates for reaction I I at 25"31 CH3NHCHO NHzCHO CH30H Solvent Relative rate 12.5 I 54.3 1.2 x 106 A change from methanol to the dipolar aprotic solvent dimethyl formamide results in a rate increase of lo6 fold.Formamide and methyl formamide, which are relatively good hydrogen bonding solvents, do not cause a com- parable increase in rate. Similarly, anions in dipolar aprotic solvents act as much stronger bases than the same anions in polar hydrogen bonding solvents. For example, the R.I.C. Reviews 130 methoxide ion catalysed racemization of (+)-2-methyl-3-phenylpropionitrile is 109 times faster in dimethyl sulphoxide as solvent than the same reaction in methanol;32 since the rate limiting step in this reaction is the removal of a proton from the asymmetric carbon atom, this rate difference reflects the increased basicity of the methoxide ion in dimethyl sulphoxide as solvent.An even larger effect has been observed in the racemization of 1-phenylmethoxy- ethanol by t-butoxide ions in dimethyl sulphoxide which proceeds 1012 times faster than the reaction in methanol catalysed by methoxide ions.33 From this very small selection of examples of the effect of solvent changes on reaction rates (further examples are discussed in, for example, reference 30) it can be seen that the rate enhancements observed are of the same order of magnitude as the catalytic effect of enzymes. Hence, given the ability of an enzyme to exclude water from the active site and thus effectively to change the solvent in which the reaction is taking place, it is reasonable to propose that environmental effects are largely, if not entirely, responsible for enzyme catalysis. Lysozyrne Lysozyme catalyses the hydrolysis of /3(1 + 4) glycosidic linkages in poly- saccharides of the type shown in Fig.5. The natural substrate is a poly- saccharide, present in the cell walls of certain bacteria, which consists of alternating units of N-acetylmuramic acid (NAM) and N-acetylglucosamine (NAG) ;42 hydrolysis occurs exclusively between NAM and NAG residues and not between NAG and NAM. The enzyme also catalyses the hydrolysis CH2OH CH2 OH CHIOH NHCOCH, NHCOCH, NHCOCH, STUDIES OF INDIVIDUAL ENZYMES Many studies of the mechanism of action of enzymes have been reported,34 but in only a small number of cases is information pertinent to a description of catalysis available.The reason for this is that the information required can only be derived from a knowledge of the complete three-dimensional struc- tures of enzymes. Such structure determinations, which, as has already been pointed out, present very formidable problems in the application of x-ray crystallography, have so far been reported for a rather small number of enzymes (lysozyme35J6 ribonuclease,37$3* chymotrypsin,loJg carboxypepti- dase40 and papain41). Of these, lysozyme, chymotrypsin and carboxy- peptidase are perhaps best understood and these three examples will be used to illustrate the general considerations of the previous section. CH2OH CH20H NHCOCH, NHCOCH, R = - H (N- acetylglucosamine) ,CH3 or R=-CH, (N - acetylrnuramic acid) COOH Fig.5. Substrates for lysozyme. 131 Doonan x1-x2 dimer 0.003 X1-Xz-X3 trimer I 4 4 X1-Xz-X3-X4 X1-Xz-X3-X4-X5 Table 4. Hydrolysis of oligosaccharides of NAG by lysozyme44 Main cleavage points J. I Relative rare 8 4000 30 000 Saccharide tetramer pentamer hexamer C 9 ? $ 4 xl-x2-x3-x4-X5-x6 - C of chitin,43 the polysaccharide in which only NAG occurs. Oligosaccharides containing two to four residues of NAG are hydrolysed very slowly, the pentamer much more rapidly and the enzyme is maximally active with the hexamer ; with the pentamer and hexamer, cleavage occurs principally between residues four and five (Table 4).44 Phillips and his coworkers35J6 have determined the complete three-dimen- sional structure of lysozyme and also that of a complex between lysozyme and the inhibitor (or rather, very poor substrate), tri-NAG.From these studies, the active site of the enzyme has been located and a plausible hypothesis for the mechanism of action and catalytic effect of the enzyme has been advanced. Lysozyme is a compact molecule with a pronounced cleft running along one side; in the complex between the enzyme and tri-NAG, the trimer is bound into this cleft (shown diagramatically in Fig. 6a) with the reducing end near the centre. Since the trimer inhibits the binding of hexa-NAG to the enzyme, Phillips and his coworkers assumed that it occupied part of the binding site of the hexamer and attempted to fit further NAG residues into their model of the enzyme-tri-NAG complex.It was found that three more residues could be fitted (Fig. 6b) provided that residue D was distorted some- what from the normal chair conformation. Assuming that the model repre- sented the Michaelis complex between lysozyme and hexa-NAG, it was possible to locate the catalytic groups in the active site. Rupley's work on the hydrolysis of oligomers of NAG suggested that cleavage occurred between residues D and E. This postulate was strongly supported by attempts to build asp-52 / 0- \O .1 ? C ? 6 F f I t b) Binding of tri - NAG Binding of hexa - NAG Arrangement of catalytic groups Fig. 6. Diagrammatic representation of the substrate binding site of lysozyme. 132 R. I . C. Reviews a NAM-NAG polymer into the active site.It was found that NAM residues could not occupy positions C and E because of interactions between the lactyl side chain on C3 of the hexose ring and amino-acid side-chains; the hexamer could only be fitted if it had the sequence NAG. NAM .NAG. NAM . NAG. NAM. From the known specificity of the enzyme, possible sites of hydrolysis were therefore limited to the links between residues B and C or D and E. Since in the unreactive complex with tri-NAG sites A, B and C are occupied, it seems that hydrolysis must occur between residues D and E. Moreover, two likely catalytic groups, namely the carboxyl side chains of residues glutamic-35 and aspartic-52, are located on either side of the linkage between residues D and E (Fig.6c). Close examination of the lysozyme-hexa-NAG model has yielded informa- tion about the mechanism and catalysis of the hydrolytic reaction. Of greatest significance are the micro-environments of the catalytic side chains of aspartic- 52 and glutamic-35. The former is located in a polar region of the enzyme whereas the environment of glutamic-35 is essentially non-polar ; under the conditions of pH at which lysozyme is active, aspartic-52 will exist as the carboxylate ion whilst glutamic-35 will be in its uncharged protonated form (Fig. 6c). The carboxyl group of glutamic-35 is situated about 0.3 nm from CH3CONH I CH3CONH C 0'" \ asp Fig. 7. The mechanism of action of lysozyme (continued ovedeof). 133 Doonan 10 CHlOH I 0- CH20H A - B - C CH3CONH I CH,CONH C 04 \ - a* P 0-H A - B-C Go :- I CH,CONH I 0GC\ asp CH,CONH C i 04 \ the glycosidic oxygen atom between residues D and E and is, therefore, ideally positioned to participate in acid catalysis of the hydrolytic reaction.On the basis of the information given above, and the established mechanisms of non-enzymic hydrolysis of glycosides, Vernon45 has proposed the scheme shown in Fig. 7 for the enzyme-catalysed hydrolysis of hexa-NAG. Glutamic-35 protonates the glycosidic oxygen atom, after which hetero- lysis of the bond between the protonated oxygen and C1 of residue D occurs with the generation of a carbonium ion. The dimer EF leaves the active site and is replaced by a water molecule. The water molecule reacts with the carbonium ion to generate the product tetra-NAG which then diffuses away from the enzyme.Arguments in support of this mechanism rather than other 134 R. I. C. Reviews possibilities involving bimolecular cleavage of the glycosidic bond or intra- molecular participation by the acetamido group on C2 of residue D have been discussed .2 Given the mechanism in Fig. 7, it remains to explain the rate enhancement produced by the enzyme, the estimated magnitude of which is 1010 fold or greater.46 Of the factors which may contribute to catalysis, that arising from protonation of the glycosidic oxygen atom by glutamic-35 is the most difficult to assess. The catalytic factor from this source will depend on the proton donating and accepting properties of the acidic group and glycosidic oxygen atom respectively in the particular environment provided by the enzyme; there is, of course, no way of measuring this.However, the acid strength of the glutamic residue in its essentially non-polar environment would be expected to be low, and consequently acid catalysis is not likely to contribute a large rate enhancement factor to the overall reaction. Similarly, the distortion of residue D which occurs on binding of the substrate to the enzyme surface is a contributing, but not a dominant, factor in catalysis. Model building studies suggest that substrate binding is accompanied by distortion of the ring of residue D towards the so-called half-chair conformation in which atoms C1, C2, C4 and the ring oxygen are coplanar; this is the preferred conforma- tion of carbonium ions derived from pyranosyl ring systems since it enables the positive charge to be distributed over both C1 and the ring oxygen atom.47 Distortion of the substrate on binding lowers the energy barrier between the enzyme-substrate complex and the transition state of the reaction, thus increasing the reaction rate.It has been argued,2 however, that the rate enhancement due to ring distortion is unlikely to be more than 104 fold, which still leaves the major part of the catalysis to be explained. The feature of the active site of lysozyme which is responsible for the major part of the catalytic activity of the enzyme is the negative charge on aspartic-52, the side chain oxygen atoms of which approach to within about 0.3 nm of C1 and the ring oxygen atom of residue D.It seems that the energy required for formation of the carbonium ion intermediate, the rate limiting step in the reaction sequence, is provided by the electrostatic interaction between the negative charge on aspartic-52 and the developing positive charge on the carbonium ion. It can be shown that the energy of interaction between two point charges separated by 0.3 nm in vacuo is about 5 eV*,45 whereas the difference in free energy of activation required to produce the estimated rate increase in the enzyme catalysed reaction is only about 0.5 eV; obviously on this simple model the electrostatic energy is more than enough to account for catalysis by the enzyme.The point-charge model overestimates the energy of interaction since both of the charges will be dispersed to some extent-that on aspartic-52 over both carboxyl oxygen atoms and that of the carbonium ion over C1 and the ring oxygen-and in addition the dielectric constant of the environment of the charges will be greater than unity. Even allowing for these effects, the energy involved is certainly large enough to account for the catalytic effect of the enzyme. A requirement for this catalytic effect to operate is the exclusion of bulk solvent (water) from the active site in the enzyme-substrate complex: * 1 eV is approximately equal to 1.602 x 10-19J. 135 Doonan water, by virtue of its high dielectric constant, would reduce the electrostatic interaction to a value less than 0.1 eV thus providing negligible rate enhance- ment.In summary, the major part of the catalytic activity of lysozyme derives from the energy of interaction of a negative charge in the enzyme and a developing positive charge on the substrate, the function of the enzyme being to provide a physical environment in which these charges will form readily and interact strongly. Ch ymo t rypsiiz Chymotrypsin is an enzyme which catalyses the hydrolysis of a variety of esters and amides (equation 5). The mechanism of action of the enzyme has been the subject of extensive investigation~~~9~~ and was in large part under- stood before the three-dimensional structure was determined from x-ray diffraction studieslOJ9 (see p.122); these studies have, however, led to an understanding of the catalytic properties of the enzyme. Recognition of the importance of a serine residue for the activity of the enzyme arose from the work of Balls and J a n ~ e n ~ ~ who showed that the enzyme is completely inhibited by treatment with diisopropylphosphofluoridate (DFP) in stoicheiometric amounts. Subsequent work50 showed that DFP reacts with a unique serine residue (serine-195 in the sequence published by Hartleys) and with none of the other 28 serine residues in the enzyme. These observations were taken to show that serine-195 is involved in the activity of the enzyme. It has now been established with some certainty that for sub- strates other than esters or amides of aromatic amino acids, hydrolysis occurs by the so-called double displacement reaction sequence (equation 12) ;22948 that is, the hydroxyl group of serine-195 (shown as E.OH in equation 12) H O E.OH -k RCOX (12) + HX E.OCOR 2 E.OH + RCOOH displaces group -X from the substrate and is itself acylated, after which hydrolysis of the acyl enzyme intermediate liberates the free enzyme and carboxylic acid.Although the evidence that reaction sequence 12 is applicable to the hydrolysis of esters and amides of aromatic amino acids is less com- plete,2 it will be assumed for the purposes of the present discussion that equation 12 is a general representation of chymotrypsin catalysed reactions. The second reactive amino-acid side-chain to be implicated in chymo- trypsin activity was the imidazole ring of histidine-57.Studies of the depen- dence of reaction rate on pH showed that activity was controlled by a group with pKa of about 7 and it was assumed that this group was the side chain of a histidine residue.51 Further, Schoellmann and Shaw52 showed that inhibi- tion of the enzyme by the substrate analogue tosyl-L-phenylalanylchloro- methane resulted from reaction of the reagent with an imidazole side chain, and Smillie and H a r t l e ~ ~ ~ identified this as the side chain of histidine-57. On the basis of these observations, a possible reaction mechanism may be written as in equations 13 and 14, where -OH and >N: represent the hydroxyl group of serine-195 and one of the ring nitrogen atoms of histidine- 57 respectively.R.I.C. Reviews 136 The sequence of reactions in equations 13 and 14 does not, of course, account for catalysis by the enzyme. The rate limiting steps in the sequence are presumably the nucleophilic attacks by serine on the substrate in 13 and by water on the acyl enzyme intermediate in 14; hence catalysis must arise from some factor which increases the nucleophilicity of these species. It was pointed out earlier (p. 130) that the most effective nucleophiles are unsolvated anions; hence it is reasonable to assume that the enzyme functions by pro- ducing the alkoxide ion of serine-195 (in reaction 13) and a hydroxide ion (in reaction 14) under conditions where they are essentially unsolvated, and that the role of histidine-57 is to assist in the deprotonation reaction producing these ions.The central problem is to explain how the base strength of histi- dine-57, in the particular environment of the enzyme active site, is increased to the extent that it can effectively deprotonate the serine residue or a water molecule. The solution to this problem has come from the x-ray diffraction studies carried out by Blow and his colleagues.l0J9 They found that residues histi- dine-57 and serine-195 are properly positioned to interact in the active site, but, in addition, histidine-57 is in close contact with the carboxyl side chain of aspartic-102,5* a residue which had not previously been implicated in the activity of the enzyme.These three residues are arranged in the enzyme in such a way that a pair of protons can interact with four binding sites (namely the carboxylic oxygen, two nitrogen atoms of the imidazole ring and the serine oxygen) to produce a system with the extreme forms shown in Fig. 8; his his v Fig. 8. Charge relay system at the active site of a-chymotrypsin. Doonan 137 these species may be considered either as tautomers in equilibrium or as canonical forms of a resonance structure depending on whether proton movements occur. If the form on the right of Fig. 8 predominates, then the serine oxygen atom will carry a large fraction of a negative charge; that is, the coupling of aspartic-102 and histidine-57 will increase the base strength of the histidine residue sufficiently to deprotonate serine-195.In fact, aspartic- 102 is situated in a non-polar environment in which its ionization will be suppressed and hence the form of the aspartic-histidine-serine system in which aspartic-102 is uncharged and serine-102 exists as the alkoxide ion will predominate. Thus the particular arrangement of aspartic-102 and histidine-57 constitutes a device for activating serine-195 in reaction 13 and, by analogy, the water molecule involved in the hydrolysis of the acyl enzyme inter- mediate (reaction 14). Generation of a negative charge on swine-195 is a necessary but not a sufficient condition for catalysis of reaction 13. The second requirement is that binding of the substrate to the active site displaces water from the region of serine-195, thereby desolvating the partial negative charge formed by inter- action with histidine-57.The fact that serine-195 exhibits high nucleophilicity only towards substrates and substrate analogues which bind at the active site supports the idea that binding promotes desolvation of the active serine residue. The reactivity of the desolvated alkoxide ion should then be great enough to account for the catalytic effect. I (15) I Rn- I Carboxypeptidase A The mechanism of action and catalytic features of carboxypeptidase A are not so well understood as are those of lysozyme and chymotrypsin, but recent x-ray diffraction studies40 have revealed interesting features of the interaction between the enzyme and its substrate.The active site of an enzyme is fre- quently pictured as a rigid region of the enzyme surface in which the catalytic groups are correctly orientated to react with the substrate once the latter has become attached to the appropriate binding sites. It was suggested some time ago, however, that in some cases the catalytic groups may be forced into the correct orientation as a result of substrate binding; that is, the active site should be considered as a flexible structure which can be moulded into the correct 'shape' by the act of binding the substrate.55~56 Evidence that this effect plays an important part in the mechanism of action of carboxypeptidase A has been 0btained.~0 Carboxypeptidase-A catalyses the hydrolysis of the C-terminal peptide bond of peptides and proteins (reaction 15).The enzyme contains one gram I . . . CHCONHCHCOOH + H,O - . . .CHCOOH + NHzCHCOOH I Rn-1 R" R n atom of zinc per mole of protein, and it has been shown that the zinc ion is essential for catalytic a~tivity.5~ Similarly, a tyrosine residue has been impli- cated in the activity of the en~yrne.5~959 Lipscomb and his coworkers40 have described the events which occur on binding of a poor substrate, glycyl-L-tyrosine, to the active site of the enzyme R.I.C. Reviews 138 \ \ I I I I * I I I OH / I ____/ - / NH,CH,C Go NHCHCOO- I I1 /I\ OH - _ I , active enzyme - substrate complex Fig. 9. Substrate induced conformational changes in carboxypeptidase-A. inactive enzyme- substrate complex (Fig.9). The tyrosine residue of the substrate fits into a 'pocket' in the enzyme surface and is held in position by interaction between the carbonyl oxygen atom and the protein-bound zinc ion. An interaction occurs between the negatively charged carboxyl group on the substrate and a positively charged arginine residue in the protein (arginine- 145) which causes the latter residue to move by about 0.2 nm towards the substrate. This small movement has a profound effect on other regions of the enzyme chain, and in particular it causes the hydroxyl group of residue tyrosine-248 to move by about 1.20 nm from a position pointing into the solvent shell of the enzyme to a new position directly above the peptide bond of the substrate (Fig.9). This appears to be a direct demonstration of Koshland's induced fit the0ry.~~?56 These induced conformational changes also explain the observation that carboxypeptidase-A does not catalyse the hydrolysis of peptides in which the terminal carboxyl group is amidated even though such amidated peptides bind readily to the enzyme. Amidated peptides lack a negative charge on the terminal carboxyl group and hence cannot initiate the conformational changes which lead to the catalytically active conformation of the active site. The events which occur after formation of the active enzyme-substrate complex are a matter of speculation. It seems likely that the function of tyrosine-248 is to protonate the peptide nitrogen atom thus facilitating either unimolecular or bimolecular cleavage of the peptide bond ; polarization of the peptide carbonyl group by binding to the zinc ion will also assist in either mode of bond cleavage.Crystallographic studies4() have shown that the carboxyl group of residue glutamic-270 is orientated directly towards the peptide bond of the substrate. This group could participate in the reaction by nucleophilic attack on the carbonyl group to produce a mixed anhydride, subsequent hydrolysis of which would yield the desired product (Fig. 10). Alternatively, glutamic-270 may participate by stabilization of an inter- mediate carbonium ion formed by unimolecular cleavage of the peptide bond ; the situation would then be very similar to that found with lysozyme.Doonan 139 I I / I :H-COO- - 1 Rn I Rn-1 O Rn-1 0 I I --L - CH - c -OH NH~CHCOO- I1 Rrl I I Zn2+ Zn2+ A Fig. 10. A possible reaction mechanism for carboxypeptidase-A. Whichever of these two mechanisms is operative, the general requirements for catalysis seem clear. Firstly, the environments of the tyrosine hydroxyl group and the peptide nitrogen atom must be such that proton transfer from one to the other is essentially complete. Secondly, the carboxyl group of glutamic-270 must be shielded from solvent water so that the effectiveness of the group either as a nucleophile or in stabilizing a carbonium ion by electro- static interactions is maximized. A precise description of these catalytic factors must await more detailed information about the active site of the enzyme.CONCLUSIONS The central theme of this article has been that a description of catalytic factors requires more information than is included in the most detailed statement of the mechanism of an enzyme-catalysed reaction. In particular, information is required about the physical environments of catalytic groups in the active site and the way in which environmental factors change the chemical behaviour of these groups. This sort of information is only available at the present time for a very small number of enzymes, but it may reasonably be hoped that the principles outlined above to explain catalysis by lysozyme, chymotrypsin and carboxypeptidase will be found to be of much more general application.R. I . C. Reviews 140 5 P. Edman, Proc. R . Anst. chem. Inst., 1957, 434; Ann. N. Y. Acad. Sci., 1960, 88, 602. ACKNOWLEDGMENTS Figure 2 is reproduced by courtesy of Dr D. M. Blow and the editor of Nature; Figure 3 by courtesy of Dr Blow and the Medical Research Council. REFERENCES 1 R. Lumry, in The enzymes (ed. P. D. Boyer, H. Lardy and K. Myrback), vol. 1, chapter 4. New York: Academic Press, 1959. 2 S. Doonan, C. A. Vernon and B. E. C. Banks, Prog. Biophys. molec. Biol., 1969, in press. 3 D. E. Koshland and IS. E. Neet, A. Rev. Biochem., 1968,37, 359. 4 C. H. W. Hirs (ed.), Methods in enzymology. New York: Academic Press, 1967. 6 R. Shipolini, A. F. Bradbury, G. L. Callewaert and C. A. Vernon, Chem. Commun., 7 G.L. Callewaert and R. Shipolini, unpublished work. 8 B. S. Hartley, in Structure and activity of enzymes, (ed. T. W. Goodwin, J. I. Harris and B. S. Hartley), p. 47. London: Academic Press, 1964. 9 B. Keil and F. Sbrm, in Structure and activity of enzymes (ed. T. W. Goodwin, 652. 1967, 679. J. I. Harris and B. S. Hartley), p. 37. London: Academic Press, 1964. 10 B. W. Matthews, P. B. Sigler, R. Henderson and D. M. Blow, Nature, Lond., 1967, 214, 11 J. C. Kendrew, H. C. Watson, B. E. Strandberg, R. E. Dickerson, D. C. Phillips and 5, 528. V. C. Shore, Nature, Lond., 1961, 190, 666. 12 L. Stryer, A . Rev. Biochem., 1968, 37, 25. 13 D. C. Phillips, Prog. Biophys. molec. Biol., 1969, in press. 14 B. E. C. Banks, S. Doonan, A. J. Lawrence and C.A. Vernon, Eur. J . Biochem., 1968, 15 B. E. C. Banks, S. Doonan, J. Gauldie, A. J. Lawrence and C. A. Vernon, Eur. J. 16 B. E. C. Banks, M. P. Bell, A. J. Lawrence and C. A. Vernon, in Pyridoxal catalysis: enzymes and model systems (ed. E. E. Snell, A. E. Braunstein, E. S. Severin and Yu. M. Torchinsky), p. 191. New York: Interscience, 1968. Biochem., 1968, 6, 507. 17 B. E. C. Banks, A. J. Lawrence and C. A. Vernon, Eur. J. Biochem., 1969, in press. 18 S . F. Velick and J. Vavra, in The enzymes (ed. P. D. Boyer, H. Lardy and K. Myrback), vol. 6, chapter 15. New York: Academic Press, 1962. 19 D. E. Metzler, M. Ikawa and E. E. Snell, J. Am. chem. Soc., 1954, 76, 648. 20 J. Kirsch and W. P. Jencks, J. Am. chem. SOC., 1964, 86, 837.21 C. G. Swain and J. F. Brown, J. Am. chem. SOC., 1952, 74, 2538. 22 M. L. Bender and F. J. Kkzdy, A. Rev. Biochem., 1965, 34,49. 23 M. L. Bender, J. Am. chem. Soc., 1957, 79, 1258. 24 B. Capon, Q. Rev. chem. SOC., 1964, 18,45. 25 T. C. Bruice, J. Am. chem. SOC., 1959, 81, 5444. 26 D. E. Koshland, J. theor. Biol., 1962, 2, 75. 27 D. E. Koshland, J. cell. comp. Physioi., 1956, 47 suppl. 1, 217. 28 W. P. Jencks, in Current aspects of biochemical energetics (ed. N. 0. Kaplan and E. P. Kennedy), p. 273. New York: Academic Press, 1966. 29 P. C. Haake and F. H. Westheimer, J. Am. chern. SOC., 1961, 83, 1102. 83, 3678. 30 A. J. Parker, Q. Rev. chem. SOC., 1962, 16, 163; Chem. Rev., 1969, 69, 1. 31 A. J. Parker, J. chem. Soc., 1961, 1328. 32 D. J. Cram, B. Rickborn, C. A. Kingsbury and P. Haberfield, J. Am. chem. Soc., 1961, 33 D. J. Cram, C. A. Kingsbury and B. Rickburn, J. Am. chem. SOC., 1961, 83, 3688. 34 P. D. Boyer, H. Lardy and K. Myrback (eds), The enzymes, vols 1-8. New York: Academic Press, 1959-63. 35 C. C. F. Blake, G. A. Mair, A. C. T. North, D. C. Phillips and V. R. Sarma, Proc. R. 36 C. C. F. Blake, L. N. Johnson, G. A. Mair, A. C. T. North, D. C. Philips and SOC., 1967, B167, 365. 37 G. Kartha, J. Bello and D. Harker, Nature, Lond., 1967, 213, 862. V. R. Sarma, Proc. R . SOC., 1967, B167, 378. 38 H. W. Wyckoff, K. D. Hardman, N. M. Allewell, J. Inagami, D. Tsernoglou, L. N. Johnson and F. M. Richards, J. b i d . Chem., 1967, 242, 3749. 39 P. B. Sigler, D. M. Blow, B. W. Matthews and R. Henderson, J. mol. Biol., 1968, 35, 143. Doonan 141 40 G. N. Reeke, J. A. Hartsuck, M. L. Ludwig, F. A. Quiocho, T. A. Steitz and W. N. Lipscomb, Proc. natn. Acad. Sci. U.S.A., 1967, 58, 2220. 41 J. Drenth, J. N. Jansonius, R. Koekoek, H. M. Swen and B. G. Wolthers, Nature, Lond., 1968, 218, 929. 42 M. R. J. Salton and J. M. Ghuysen, Biochim. biophys. Acta, 1959,36, 552; 1960,45, 355. 43 L. R. Berger and R. S . Weiser, Biochim. biophys. Acta, 1957, 26, 517. 44 J. A. Rupley, Proc. R. SOC., 1967, B167, 416. 45 C. A. Vernon, Proc. R. Soc., 1967, B167, 389. 46 L. N. Johnson, D. C. Phillips and J. A. Rupley, Brookhaven Symp. Biol., 1968, in press. 47 R. V. Lemieux and G . Haber, Can. J. Res., 1955, 33, 128. 48 T. C. Bruice and S. J. Benkovic, Bioorganic mechanisms, vol. 1, chapter 2. New York : Benjamin, 1966. 49 A. K. Balls and E. F. Jansen, Adv. Enzymol., 1952, 13, 321. 50 R. A. Oosterbaan, M. van Adrichem and J. A. Cohen, Biochim. biophys. Acta, 1962, 63, 204. 51 D. E. Koshland, D. H. Strumeyer and W. J. Ray, Brookhaven Symp. Biol., 1962,15, 101. 52 G. Schoellmann and E. Shaw, Biochemistry, 1963, 2, 252. 53 L. B. Smillie and B. S. Hartley, Fed. Eur. Biochem. Socs, 1964, abstract A 30. 54 D. M. Blow, J. J. Birktoft and B. S . Hartley, Nature, Lond., 1969, 221, 337. 55 D. E. Koshland, Proc. natn. Acad. Sci. U.S.A., 1958, 44, 98. 56 D. E. Koshland, J. cell. comp. Physiol., 1959, 54, 235. 57 B. L. Vallee, J. A. Rupley, T. L. Coombs and H. Neurath, J. biol. Chem., 1960,64,235. 58 0. A. Roholt and D. Pressman, Proc. natn. Acad. Sci. U.S.A., 1967, 58, 280. 59 R. T. Simpson, J. F. Riordan and B. L. Vallee, Biochemistry, 1963, 2, 616. 142 R. I.C. Reviews