....... A MODEL SYSTEM ANALYSIS "OF THE! MECHANISM (IF ' 2*KETDS3r-DE0XY-6 PHOSPIIOGLUCONIC I _ 7 ACID ALDOLASE: THE CATALYTIC ROLE 0F ' THE SCHIFF BASE INTERMEDIATE ' Thesis for the Degree of. Phfi ll ’ MICHIGAN STATE UNIVERSITY MARK A; ROSEMAN 1970 This is to certify that the thesis entitled A MODEL SYSTEM ANALYSIS OF THE MECHANISM OF 2-KETO-3-DEOXY-6-PHOSPHOGLUCONIC ACID ALDOLASE: THE CATALYTIC ROLE OF THE SCHIFF BASE INTERMEDIATE presented by Mark A . Ros eman has been accepted towards fulfillment of the requirements for Ph. D. degree in Biochemistry /, . / ' ,. 7 ./ 1 7 1/ I/ I fl/[y f 73”? Major professor Date December 19 70 0-7639 LIBRARY ' Michigan State University ABSTRACT A MODEL SYSTEM ANALYSIS OF THE MECHANISM OF 2-KETO-3-DEOXY-6-PHOSPHOGLUCONIC ACID ALDOLASE: THE CATALYTIC ROLE OF THE SCHIFF BASE INTERMEDIATE BY Mark AIIRoseman The rate constants of enamine formation from ket- imines of pyruvate, and the cyclic ketimine Al-piperidine- 2-carboxy1ic acid, were determined in aqueous solution using rapid iodination of the enamine to measure the reactions. The constant for ketimines of pyruvate was determined indirectly by comparing the kinetics of iodi- nation of pyruvate in primary amine buffers to the kinetics in imidazole and phOSphate buffers. In all buffers, the following rate law was obtained: k = k + k B + k A. o a b The constant ka was greater for primary amine catalysis obs than for imidazole or phosphate catalysis. This result suggested that the kaA term results from water-catalyzed tautomerization of the protonated ketimine to the enamine. From this, the constant for enamine formation was calcu- lated; ke = 0.304 sec-1. Mark A. Roseman The tautomerization of Al—piperidine-2-carboxylic acid could be measured directly. The kinetics indicated that tautomerization of the protonated ketimine predomi- nates, and is strongly catalyzed by general bases. The rate constants obtained with this compound differ markedly with those of the ketimines of pyruvate. The discrepancies are discussed as well as the application of these results in determining the catalytic role of the ketimine intermediates in enzyme-catalyzed enolizations. A MODEL SYSTEM ANALYSIS OF THE MECHANISM OF 2-KETO“3-DEOXY-6-PHOSPHOGLUCONIC ACID ALDOLASE: THE CATALYTIC ROLE OF THE SCHIFF BASE INTERMEDIATE BY (\ \0/ Mark A. Roseman A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Biochemistry 1970 C(5- 22‘4/ 7 T0 T0 To To To To To T0 T0 DEDICATION my junior high school teachers who predicted I would be a high school drop-out; my high school counselor who predicted I would never go to college; Dean Anderson who predicted I would never graduate from the University of Michigan; my father who thought they may be right; my mother who knew better; my mother-in-law who wanted me to become a doctor; my father-in-law who wonders when I'll get a job; my wife, the overseer, who kept me working; and my son, whose financial support encouraged hasty completion of this work. ii ACKNOWLEDGMENTS I would like to extend my sincere gratitude to Dr. W. A. Wood for the patience he showed me and the courage to support a research project in an area far removed from the general program of his laboratory. I wish to acknowledge financial support as a trainee under the Training Grant (GM 1096) awarded to the Department of Biochemistry. II also extend my lasting appreciation to my wife Jo Ellen, for her understanding and encouragement through- out the course of this work, and assistance in the prepar- ation of this manuscript. iii VITA Mr. Roseman was born October 30, 1944, in Brooklyn, New York. Shortly thereafter, he was moved to the Interior, where he has lived ever since. His formative years were spent in Ann Arbor, Michigan, where he attended Ann Arbor High School and the University of Michigan. He majored in Cellular Biology at the University of Michigan, and graduated with a Bachelor of Science degree in May, 1965. Since that time he has attended Michigan State University, where he has received his graduate education in biochemistry. He is.married and has a son. iv LIST OF LIST OF Chapter I. II. III. IV. TABLE OF CONTENTS TAB LE S O O O O C C C O O O O O FIGURES. . . . . . . . . . . . INTRODUCTION . . . . . . . . . . LITERATURE REVIEW. . . . . . . . . Schiff Base Catalysis: A Perspective . The Mechanism of Schiff Base Formation . Aldol Reactions: Nonenzymatic. . . . Aldol Reactions: Enzymatic. . . . . B-Decarboxylation: Nonenzymatic . . . B-Decarboxylation: Enzymatic . . . . Enolization: Nonenzymatic . . . . . Enolization: Enzymatic . . . . . . METHODS . . . . . . . . . . . . Synthesis of Al-Piperidine—Z-Carboxylic Acid . . . . . . . . . . . . Buffers . . . . . . . . . . . Iodination Procedures. . . . . . . pH-Stat Iodination Procedures . . . . Determination of pKa of n-Butylamine. . Determination of the EquIlibrium Con- stant of Schiff Base Formation. . . . Stopped Flow Measurements . . . . . Enzymatic. . . . . . . . . . . RESULTS C O O O I O O O O O O 0 Introduction. . . . . . . . . . Demonstration of Schiff Base Formation . Kinetics of Schiff Base Formation. . . Page vii viii 10 10 12 15 19 21 23 24 35 39 39 48 51 54 55 56 64 66 66 67 70 Chapter Enolization of P ruvate Enolization of A -Piperidine-2- Carboxylic Acid. V. DISCUSSION REFERENCES. vi Page 75 94 115 133 LIST OF TABLES Table Page 1. Catalytic Constants in Rate Equation Describing Enolization of Isobutyral- dehyde in Methylamine Buffers . . . . 32 2. Equilibrium Constant of Schiff Base Formation. . . . . . . . . . . 62 3. Rate Constants for Schiff Base Formation from Pyruvate and Butylamine . . . . 74 4. Iodination of Pyruvate in Imidazole, Phosphate and Amine Buffer: Summary of Catalytic Constants ka and kb . . . 83 5. pH—Rate Profile of the Iodination of Al-Piperidine-2-Carboxylic Acid by the pH-Stat Method . . . . . . . . . 103 6. Complete Rate Constant Table . . . . . 113 7. Comparison of the Enzymatic Rate Constant of Enolization with the Corresponding Pseudo First Order Rate Constants for Pyruvate and Al-PCA at pH 8.0 . . . . 123 8. Comparison of the Enzymatic Enolization with Imidazole-Catalyzed Enolization of P ruvate and Protonated Ketimine of A -PCA. . . . . . . . . . . . 124 vii Figure l. 2. 12. 13-l3a. 14. 15. LIST OF FIGURES Synthesis of Al-Piperidine-2-Carboxylic ACid O O O O O O O O O O O O 0 Absorption Spectra of Al-Piperidine-Z— Carboxylic Acid in Water and 0.1 M NaOH . Difference Spectra of Pyruvate in nfButylamine Buffers. . . . . . . . Difference Spectra of Pyruvate-KDPG Aldolase Mixtures. . . . . . . . . . . . A Typical Reaction of Pyruvate with Butylamine to Form Schiff Base as Measured by Stopped Flow Techniques. . . . . . Determination of the Constants k0, ka' and kb from the Iodination of Pyruvate Accord- ing to the Experimental Rate Equation Rate = kobsP = (k0 + ka + kaA)P. . . . Typical Iodination Reaction of Pyruvate by the pH-Stat methOd O C O O C O O 0 Determination of the Rate Equation for the Iodination of Pyruvate in the Absence of Buffer . . . . . . . . . . . . Typical Iodination Reaction of Al— Piperidine—Z-Carboxylic Acid Measured with the Stopped Flow Apparatus . . . . Iodination of Al-Piperidine-2-Carboxylic Acid in Imidazole Buffer by Stopped Flow Techniques: Plot of Rate Versus Buffer Concentration . . . . . . . . . . viii Page 41 50 58 69 73 77-82 89 92-93 97 99 Figure 16. 17. 18. 19. Typical Iodination Reaction of Al-Piperidine- 2-Carboxylic Acid by the pH-Stat Method. Determination of the Rate Law for the Iodi- nation of A -Piperidine-2—Carboxylic Acid in the Absence of Buffer. . pH-Stat Iodination of Al-Piperidine-Z- Carboxylic Acid. . . . . Iodination of Al-Piperidine-2-Carboxylic Acid by the pH-Stat Method; Determination of the Constants kOH- and k0 ix Page 102 107 110 112 CHAPTER I INTRODUCTION Enzyme catalysis is unquestionably the most funda- mental and most important biological process. Still there is no adequate explanation for it; it is generally acknowl- edged that the catalytic process has not been satisfacto- rily described for even a single enzyme. However, as Koshland points out (1), one important conclusion has been reached: enzyme catalysis can probably be explained in terms of processes familiar to the organic chemist. This conclusion has been the stimulus for the "model system" approach to enzyme catalysis. A model system is essentially a nonenzymatic ana- log of an enzyme-catalyzed reaction. Usually, the model reaction is studied under a number of conditions in order to determine the conditions and catalysts which efficiently accelerate the rate of reaction. In this way, one gathers a list of potential catalytic forces which might be uti- lized by an enzyme. Indeed, a number of effective cata- lytic forces have been discovered and characterized this way. These include proximity effects, electrostatic interactions, hydrophobic interactions, general acid- general base effects, charge-transfer-complex formation, hydrogen-bond formation, bond strain, facilitated dif- fussion, and orbital steering. It should be quite evident that if enzyme cata- lysis operates by processes familiar to organic chemistry, we shall never understand enzyme catalysis any better than we do simple nonenzymatic catalysis. For this reason, model system studies are essential. Unfortu- nately, they have one serious limitation: it is very difficult to show that a catalytic process which works for a model system actually exists for its enzymatic counterpart. For example, it is easy to show that bond strain accelerates a nonenzymatic reaction; but it is not so easy to show that a particular bond in a substrate is strained in the active site complex. Obviously, mean- ingful comparisons between model and enzymatic reactions can only be made if a fairly accurate description of enzyme-substrate complexes is at hand. In the absence of such information, model systems often seem to be an endless variety of hypothetical situations. Fortunately, one class of enzymes provides some exception to this 1imitation--those which form covalent complexes with substrate. In such cases we have a fairly accurate picture of at least one of the inter- actions between enzyme and substrate. Undoubtedly, the covalent bond is not the only interaction, but it does provide a reasonably firm starting point for a model system analysis. The research presented here seeks to determine the catalytic effect of Schiff base formation in the bacterial enzyme 2-keto-3-deoxy-6-phosphogluconic acid aldolase (KDPG aldolase). This enzyme catalyzes the reversible aldol cleavage of KDPG to pyruvate and glyceraldehyde-3- phosphate (G3P): $00— $00" C=O C=O Pyruvate I I fiHz CH3 —-—+ HfiOH +—-—— + HCOH H ,o \ / = c H2C0P03 HCOH G3P l _ KDPG Hzcopo3 Mechanistically, the reaction has been shown to proceed through an obligatory Schiff base intermediate (2, 3) as illustrated on the next page in the direction of conden- sation. If pyruvate is incubated with enzyme in the absence of G3P, a rapid exchange of the methyl hydrogens with water is observed. That is, the aldolase catalyzes the enolization of pyruvate. In every sense, this simple "partial" reaction is an enzyme-catalyzed process. Since it is simple, and proceeds through a co- valent intermediate, this reaction is ideal for model coo' coo’ + C00 | -H20 E -H | C=O + NHZE :9 C=H-E <———_F- C-H-E I +H20 | +H H CH3 CH3 CH2 + COO- coo— HC¢O | I + I c=o C=§-E HCOH I I = CH +H 0 CH :2; H copo 2 2 2 2 3 | + NHZE +——— | HCOH -H20 HTOH I HCOH HCOH l __ l _ HZCOPO3 H2COPO3 system studies. For these reasons, the Schiff base- catalyzed enolization of pyruvate is the subject of this research. The fundamental problem then, and the purpose of this research is determining the contribution of the Schiff base to the overall catalysis. This would be done by comparing the enzymatic rate of enolization to that of pyruvate and a Schiff base of pyruvate. If Schiff base formation is not sufficient to account for the catalytic rate, it is important to con- sider the popular vieWpoint that enzyme catalysis results from several catalytic forces working together. If so, how readily can the tautomerization of the Schiff base itself be catalyzed? All too often the argument is made that if a single catalytic force accelerates a reaction by a factor, x, and another force independently accelerates the reaction by a factor, y, the combined effect of both forces working together will be an acceleration (x)(y). Is this necessarily correct? All these questions shall be considered. To determine the contribution of the Schiff base to the overall catalyses, the rate constant for the tautomerization of a ketimine of pyruvate1 to the enamine in aqueous solution had to be determined: 1At this point, the standard nomenclature for Schiff base compounds should be presented. The generic term for a Schiff base is imine for which the general structure is R R" \C=N / RI/ \ Rm If R or R' is hydrogen, the imine has been formed from an aldehyde and is termed an aldimine; likewise, if R and R' are alkyl groups, the imine has been formed from a ketone and is termed a ketimine. Both aldimines and ketimines can tautomerize to an enol-like structure called an enamine: Lmines can also form addition compounds similar to those formed by aldehydes and ketones: R R" R'——C——N——Rm X H where X is some nucleophile (such as OH, CN, S, etc.). If X is OH, the addition compound is a carbinolamine; otherwise it is usually a substituted imine. C00 C00 I Is, I C=N-R C-N-R H -H" II H CH3 CH2 ketimine enamine Assuming that the ketamine can be produced, this seems to be a relatively straightforward affair. However, a consideration of the mechanism of Schiff base formation reveals the complications: IOO— coo- I00- -H O H H H+ Go A :NH R ———>( HOC-E-R <_——2—>. C=N-R 2 + | + +H20 | CH3 CH3 CH3 pyruvate amine carbinolamine ketimine It is seen that Schiff base formation involves a dehy- dration; consequently, the Schiff base is unstable in aqueous solution. In fact, Schiff base hydrolysis is usually much faster than enamine formation. The equi- librium constant for Schiff base formation is large enough, however, that a significant amount of Schiff base can be formed in rapid equilibrium with pyruvate if fairly high levels of amine are used. Unfortunately, such an equilibrium mixture causes another serious problem: the amine can function effectively as a general acid-base catalyst for the enolization of free pyruvate. There- fore, the enolization of pyruvate occurs by two routes functioning simultaneously: l. ketimine catalysis coo" $00' $00" C=O + HZN-R =3 C=N-R —> fi-M-R (via neutral imine) I I CH3 CH3 CH2 coo’ coo“ C00’ | + HZN-R I H . . =0 + :23 C=g-R -—9 C-N-R (via protonated imine) . + I In - CH3 CH3 2 2. general acid-base catalysis I00- $007 I’a‘NH .—+ fi-O- (general base) H ~CiHe—- NR CH 2 H coo’ foo‘ H Céa‘x‘ H<§+-R -—+ COH (general acid) H H CH2 CH2 Any method used to measure enolization will not distin- guish between the enamine and enol of pyruvate. Kinetically, the two mechanisms are equivalent; that is they both obey the same experimental rate equation. Consequently, a certain amount of kinetic gymnastics is required to distinguish one from the other. A complete description of such analyses will be given in the Literature Review. It is only necessary at this point to say that any such analysis has weaknesses. For this reason, an independent method for measuring ketimine-enamine con- version was sought to supplement the method described above. It is clear that all the complications of this system arise from the inescapable requirement to work with equilibrium mixtures. What is needed, then, is a ketimine which is similar to the pyruvyl ketimine but stable to hydrolysis in aqueous solution. Fortunately, such a compound exists, namely Al-piperidine-Z-carboxylic acid: 2 H H / 2 \H H2 C-COOH N H+ Conceivably, this compound could exist in equilibrium with the open chain form: However, Macholan and Svatek (4) have shown that the com- pound exists almost exclusively in the ring form over the entire pH scale. Furthermore, they have demonstrated the ketimine-enamine conversion: /H /H ‘\H ;;:i _ \ _ - COO N1/C COO N +H +H While the use of this compound has the advantage of pro- viding a direct measure of ketimine-enamine conversion, there is now the disadvantage of not knowing the effect the ring might have on enamine formation. Therefore, the effect of the ring on enamine formation will be seriously considered when comparing the reactivity of this compound to the acyclic ketimines of pyruvate. CHAPTER II LITERATURE REVIEW Schiff Base Catalysis: A Perspective The reactions of carbonyl-containing compounds are numerous and apparently diverse. There is, however, a satisfying mechanistic consistency to them all: the carbonyl group is an electrophilic center for nucleophilic attack or intramolecular electron rearrangement. Two examples of intramolecular rearrangement shown below are B-decarboxylation and dealdolization: l. B-decarboxylation CH -C-C-C'-e/ --> CH -C=CH + C=O 2. dealdolization CH CH 3 l 3 CH3-CIERZLC-CH —+’ CH3-C + CH =C-CH3 | u 3 \\ 2 l- 35 q) 0 O \ H 10 11 As early as 1932 Pedersen recognized that the protonated form of the carbonyl group (:C=6H) should serve this electrophilic function far more effectively than the unprotonated form. While this is quite true, protonation is significant only at low pH, since the pKa of the protonated carbonyl is quite low. There are, how— ever, two means for increasing the electrophilic character of the carbonyl under neutral or alkaline conditions-- metal ion activation and Schiff base formation. Only Schiff base formation will be discussed here. The Schiff base can exist in either neutral or protonated form with a pKa of 7.00-8.00: The similarity to the neutral and protonated forms of the carbonyl group is clear; however, unlike the carbonyl, a significant fraction of Schiff base is protonated under neutral conditions. Schiff base activation of the carbonyl is therefore a particularly attractive mechanism for enzymes which normally operate near neutrality. There is one other important property of the Schiff base which distinguishes its chemistry from that of the carbonyl. Under appropriate conditions, the .position of the carbon-nitrogen double bond can shift: 12 I H I I +HKEF-N-(f-R —-> H-C-N=(|I-R + H H I This double bond rearrangement is an essential feature of many reactions involving pyridoxal phosphate. A large number of enzymatic and nonenzymatic Schiff base-catalyzed reactions have now been studied in detail. These include: aldol reactions, a-decarboxy- lations, B—decarboxylations, and enolizations. Perhaps the most important class of Schiff base reactions are those involving pyridoxal phosphate. This review, however, is intended to cover only those areas of Schiff base catalysis which are applicable to the present research. Two excellent reviews of pyridoxal catalysis have been published by Bruice (S) and Snell (6). The Mechanism of Schiff Base Formation Although the mechanism of Schiff base formation has been exhaustively studied, only those aspects which are relevant to this research will be discussed. Compre- hensive reviews on the subject have been published by Bruice (7) and Jencks (8). Early studies of the kinetics of oxime formation showed the reaction to be catalyzed by acid and alkali q .-'I .. A, .Qw in iv. . .g- .- . ovodov . a in u‘n. .sgq... A viao‘ .' "-va a .- 'V.¢~v . ..'_. . .- 0...“_ “O U“ Iv. u. .“ "oou . Q» n" on... . -- V i _v«.. “- . n. ‘ d l3 (9, 10). However, the rate showed a peculiar dependence on acid concentration: there appeared to be an optimum acidity between the one extreme where hydroxylamine is totally unionized and the other where it is totally pro- tonated. Although the explanations put forth are not correct in detail, they did suggest the intermediate formation of carbinolamine species H I R—N-C-OH which could have different ionic forms at different levels of acid. The first definitive evidence for the carbinolamine intermediate came from Bodforss's work on phenylhydrazone formation (11). He found that the disappearance of phenylhydrazine occurs faster than formation of phenyl— hydrazone; this indicates rapid accumulation of an inter— mediate. Later, Olander (12) also concluded from the kinetics of acetoxime formation that a carbinolamine must be an obligatory intermediate. Bartlett and Conant (13) also described a maximum in the pH—rate profile of semicarbazone formation. :Furthermore, they demonstrated that the reaction is catalyzed by general acids. Unfortunately, they failed ix: Carry out the experiments necessary to distinguish the effect of pH from the effect of buffer catalysis. For 14 this reason their interpretations of the pH~rate profile (which will not be discussed here) are somewhat incon- clusive. The overall features of the mechanism were sub- stantially clarified by Jencks (14). Using spectral tech- niques he found that at the alkaline side of the pH-rate profile a rapid decrease of carbonyl absorbance is first observed, followed by the slow appearance of imine ab- sorbance. This means that carbinolamine is formed in a rapid equilibrium followed by the rate determining de- hydration to imine. Under these conditions, he also showed the dehydration of carbinolamine to be catalyzed by general acids (which contrasts with the previous inter- pretations that carbinolamine formation is general acid- catalyzed). On the acid side of the pH-rate profile, the situation is reversed: dehydration is fast and carbinol- amine formation is rate determining. The reason for the change in rate determining step is then quite clear. Under acid conditions the acid-catalyzed dehydration is, as one would expect, fast; whereas the rate of carbinol- amine formation, which depends on the concentration of the conjugate base of the amine, is slow. Conversely, under alkaline conditions the concentration of conjugate base is high so that carbinolamine formation is fast; Vfluareas the rate of acid-catalyzed dehydration is low. 15 The maximum rate occurs between these extremes where a suitable compromise is achieved. Aldol Reactions: Nonenzymatic The first and perhaps most definitive demonstration of Schiff base catalysis was accomplished by Westheimer and Cohen in 1938 (15). These workers sought to resolve an apparent discrepancy in the literature concerning the -COH-CH dealdolization of diacetone alcohol [(CH COCH3]. 3)2 2 Previously, French (16) had studied the effect of phenol- phenolate buffer on the dealdolization reaction. He found the reaction to be dependent only on the hydroxide ion concentration of the buffer and not the concentration of buffer components. In other words, the reaction appeared to be insensitive to general acid—base catalysis. On the other hand, Miller and Kilpatrick (17) found the reaction to be catalyzed not only by hydroxide ion, but also primary and secondary amines according to the rate law rate kobs(diacetone alcohol) [k0 + kOH(OH) + kb(B)] diacetone alcohol where (B) is the concentration of the basic form of the amine, k k ko are the constants for catalysis by b' on" amine, hydroxide ion, and water, respectively, and kobs is the pseudo first order rate constant at constant pH l6 and buffer concentration. This linear dependence of the rate on [B] is characteristic of general-base catalysis. Westheimer and Cohen confirmed this effect of pri- mary and secondary amines but then showed the reaction to be insensitive to tertiary amines. Since tertiary amines should be as effective as general bases, the only reason- able explanation for these results is that primary and secondary amines catalyze the reaction through an inter-- mediate Schiff base mechanism rather than a general-base mechanism; of course, tertiary amines are incapable of forming Schiff bases. The mechanism for hydroxide-catalyzed dealdoli- zation is shown below. This is typical specific-base catalysis: OH 0 q? o H H _ fast I " slow CH3-C-C-C-CH + OH :22? CH -c:cic-CH 4—__ IH 3 3|H2 3 CH3 CH3 0' CH3 | ;c=o + c=c -> 2CH3-C=O CH3 CH3 | CH3 The Schiff base-catalyzed mechanism is probably the following: l7 OH 0 1 CH3\| ll C-CHZ-C-CH3 _ CH3 0.! DOH + :2, CH3 I )+ RNH2 + H ‘— \C-CHZ-C-CH3 K111 CH3/ V' - HNR ‘u’ if + CH 3\\ CH /C-CH2-C CH3 3 .— DO J CH H CH fast CH ___*r.dk.s. 3\C=o + C=C/ 3 >——> 2 O=C/ 3 + RNH2 + H+ CH / H \N-R \CH 3 H 3 where DOH is diacetone alcohol; DO-, the ionized diacetone alcohol; I, the Schiff base intermediate; K1' the ioni- zation constant of diacetone alcohol; K the equilibrium 2: constant for Schiff base formation; and k, the rate con- stant for carbon-carbon bond cleavage. Notice that I is the protonated Schiff base of the ionized form of di- acetone alcohol. Since the problem of kinetically equivalent mechanisms occurs throughout this thesis, it is important to show mathematically why this mechanism would obey a rate equation identical to that for general-base catalysis. For simplicity, assume DOH, DO-, and I are in rapid equilibrium and DO- and I are in much smaller concen- trations than DOH. 18 Then K = (Do‘)(H+) 1 (DOH) K = I 2 (RNH2)(DO‘) + HCOH l HC=O HCOH | | = HCOH H2COPO3 H cope: 2 3 18 When KDPG labeled with 0 in the carbonyl oxygen was irreversibly converted to products, no 018 appeared in pyruvate. This result strongly supports the Schiff base mechanism. Finally, cyanide inhibits muscle aldolase in the presence of DHAP (23), presumably by forming an addition compound with the Schiff base (sbustituted ketimine) intermediate: CH20® CH o® | 2 Nsc" C=N-E :2: NEC-f-N-E | H HC-OH HCOH H H B-Decarboxylation: Nonenzymatic Pedersen (24) found the decarboxylation of a, a, dimethylacetoacetic acid in aniline or gfchloroaniline buffer followed a pH-rate profile corresponding to the rate equation 22 0 v = k[RNH2][-C-COO-][H+]. From previous discussions, it is clear that this term is consistent with Schiff base-catalyzed decarboxylation. In a later study of the amine-catalyzed decarboxy- lation of oxalacetic acid Pedersen (25) found the reaction to be catalyzed by ammonium, ethylammonium, and anilinium ions. Of these, the anilinium ions were most effective in acid solution. Again, he attributed the catalysis to an intermediate Schiff base. However, this interpretation was convincingly challenged by Hay (26) who showed with model compounds that Schiff base formation between oxal— acetate and aniline does not occur at all in aqueous solution. Surprisingly, a stable carbinolamine is formed rapidly and in high concentration. Catalysis is then seen as a concerted elimination of the anilinium ion: - -—> PhNH2 + OZCCOCHZCOZH +__ OH 0 “o C C CH fl‘?’ '0 cc CH co PhNH - - - - -—+ = + + 2 2 2 2 2 2 |\\-// l (EHZPh OH + Hay went on to show however that Schiff base formation is favorable in ethanol and probably accounts for the cataly- sis under these conditions. 23 On the other hand, Westheimer (27) showed that in aqueous solution cyanomethylamine catalyzes the decarboxy- lation of acetoacetate at the same rate as it forms Schiff base with ethylacetoacetate. Here the ester was used to measure Schiff base formation independently of decarboxy- lation. Assuming that the acid forms a Schiff base at the same rate as the ester, these results support the Schiff base mechanism. This reviewer agrees with Jencks's view (28) that there is no real contradiction in all these findings, but merely that different mechanisms operate with different substrates and conditions. Indeed, the carbinolamine- catalyzed pathway may be peculiar to aniline; otherwise it is difficult to see why tertiary alkyl amines are in- effective catalysts. B-Decarboxylation: Enzymatic In a study of the acetoacetic acid decarboxylase 18 from reaction, Westheimer (29) followed the loss of 0 the C-3 carbonyl oxygen of acetoacetic acid upon con- version to acetone. He found none of the label retained in acetone, which suggests a Schiff base mechanism for decarboxylation. (This complete exchange also argues 18 against a carbinolamine mechanism by which some 0 should be retained.) 24 These interpretations were supported by subsequent borohydride inactivation of the enzyme in the presence of substrate followed by isolation of the expected adduct e-N6-i50propyllysine (30, 31). Furthermore, cyanide was shown to inhibit the enzyme in the presence of substrate, presumably by form- ing a substituted ketimine. It has also been shown that certain Schiff base- forming aldolases are capable of catalyzing decarboxylation reactions (32, 33). Enolization: Nonenzymatic The only significant investigations in this area have been Bender's studies on the amine-catalyzed enoli- zation of acetone (34) and Hine's studies of the amine- catalyzed enolization of isobutyraldehyde (35). Since these investigations bear most directly on the research presented here, they will both be discussed in some detail. Bender measured the enolization of acetone in the presence of a large number of amines. The kinetics followed a simple rate law: rate Of = k (acetone) enolization obs [(k0 + ka(A) + kb(B) + kab(A)(B)](acetone) 25 where k0 is the pseudo first order rate constant under bs conditions of constant buffer concentration and pH; k0 is the constant for catalysis by water, hydroxide ion, and hydronium ion at fixed pH; ka is the constant for catalysis by the acid species of the buffer; kb is the constant for catalysis by the basic species of the buffer; kab is the constant for "concerted" catalysis by both the acid and base species; (A) and (B) are the concentrations of acid and base species, respectively. Only some of the amines showed a kab term, but all showed ka and kb terms. Since the important interpretations were done with methylamine buffer which does not show such a kab term, only the simplified rate law kobs = k0 + kaA + ka shall be considered for the moment. All the constants in this equation can be deter- mined as follows. kobs is determined from a plot of acetone concentration versus initial rate at fixed pH and buffer concentration; k0 is determined from the intercept of a plot of buffer concentration versus kObs at fixed pH; ka and kb are determined from the following algebraic manipulations: rak = k S(acetone) ob [k0 + kaA + ka](acetone) kobs = R0 + kaA + ka 26 Dividing by A, defining the variable r = A/B, followed by inconsequential rearrangements gives: k' = kobs - ko r + l A + B r _ 1 -'ka + kb(r) Thus, a plot of k' versus l/r gives kb as the lepe and ka as the intercept. The overall form of the rate law is apparently uninformative since it is identical to that for simple general acid-base catalysis; the effect of Schiff base catalysis is not immediately evident. (As a matter of review, the mechanisms for general acid and base catalysis by amine is shown below: general base CH CH l 3 k l 3_ 1|: C=O b C-O |\"fi H -—+ H + H-N-R H-C-H b :NR H-C +| | H | H H H general acid CH H CH C=/o\ H-N;R —-> C-OH + RNH \I II 2 l H—C'S-H H H-C H H 27 However, Bender utilized the magnitude of the constants kb and ka rather than the form of the equation to determine the existence of Schiff base catalysis. This was done by use of the Bransted relation, which relates the strength of an acid or base (by pKa or pr) to its catalytic capacity, ka or kb‘ When the kb values for the various amines were compared to those for other general bases, the catalytic ability of the basic form of the amine corresponds very well to that predicted for general base catalysis; therefore, the term kb (B)(acetone) in the rate law was interpreted as simple base catalysis. However, when the ka values for the amines were compared in the same way it was found that ka was as much as one million times larger than the value predicted by the acid strength of the protonated amine. Bender inter- preted these results to mean that the ka (acetone)(A) term in the rate equation actually represents water-catalyzed enolization of the protonated ketimine. This mechanism is, of course, kinetically equivalent to general acid catalysis insofar as the overall form of the rate law is concerned. In support of this interpretation, Bender found that the rate law with trimethylamine, which cannot form a Schiff base, contains a kb term but no ka term. Further- more, with methylamine buffer the plot of k' versus l/r deviates from linearity at lower pH. Further experiments 28 showed that this deviation could be accounted for by a change in the rate determining step from enolization to Schiff base formation itself. The rate constant for the enolization of the pro- tonated ketimine can then be calculated if the following constants are known: ka; the equilibrium constant for Schiff base formation; and the pKa for the protonated ketimine. Bender found the enolization constant to be 2.6 x 10'.2 M-lsec-l. This value is approximately four hundred times less than the aldolase-catalyzed enolization of dihydroxyacetone phosphate, but 109 times greater than the water-catalyzed enolization of acetone. While the basic interpretations are probably cor- rect, there are a few criticisms that could be made about this work. First, the pKa for the protonated ketimine could not be experimentally determined and was therefore approximated as 7.6. Second, it would have been most desirable to measure the enolization on both sides of the pKa, as is usually done, to show that one ionized form of a substrate is active. Bender apparently tried this, but the system was intractable. Third, the method for deter- mining the important constant ka requires that ka be measured relative to k (i.e., as the intercept of the k' b versus l/r plot). It turns out that the value of ka is only 1.5% the value of kb; that is, a small change in the slope of the k' versus 1/r plot results in a large change in the intercept. Conceivably, an unknown systematic 29 error in the determination of kb could result in a posi- tive intercept in such a plot where none should really exist. Fourth, it is a bit strange that in methylamine buffer no term in the rate equation was seen for base- catalyzed enolization of the protonated ketimine: ‘EHB T“3 H H C=N-CH3 kab C-N-CH3 + H | 3‘+ ___, a + CH3NH3 CH3N:‘——>H-C-H H-C H I | H H _ '1' rate — kab(acetone)(RNH3)(RNH2) The absence of such a term implies that the protonated ketimine is not susceptible to further catalysis. This creates the puzzling question as to how an aldolase makes up the four hundred-fold difference in rate. Bender touches on this question but not satisfactorily. Further- more, this result contrasts with Hine's studies on the enolization of isobutyraldehyde wherein the predominant route was found to be base-catalyzed enolization of the protonated ketimine. However, some of Hine's interpre- tations are questionable. These studies will now be discussed. The major difference in this system, compared to Bender's, is that the equilibrium constant for imine formation with isobutyraldehyde is so large that most of 30 the aldehyde exists as the imine in the presence of moderate concentrations of methylamine buffer. For this reason, Hine considers this system superior to any pre- viously studied for imine catalysis. However, the high concentration of imine made it necessary to consider many more factors which might influence the overall rate than was necessary for the acetone system. This thoroughness causes the kinetic treatment to be extremely complicated but serves to illustrate the difficulties in determining the desired constants. Hine measured the enolization by following the loss of deuterium from isobutyraldehyde-Z-d. When this was done under conditions where the concentration of methylamine and imine were held constant, the rate of exchange increased with increasing concentrations of methylammonium ions. As with the acetone system, the dependency on methylammonium ions is best attributed to enolization via the protonated imine rather than via general acid catalysis. In contrast to the acetone system, where only a fraction of the total enolization could be accounted for this way, most of the enolization occurs through the imine under Hine's conditions. How- ever, it still remains to be determined precisely how much exchange proceeds through the protonated imine and how much through other routes. For this purpose, Hine considers every possible mechanism for exchange. 31 There are three species which are capable of exchange: the aldehyde, the neutral imine, and the pro- tonated imine. For each of these species, enolization could be catalyzed by any of the bases in solution-- methylamine, the neutral imine, hydroxide ion, and water. The resulting rate equation contains ten terms. After suitable algebraic substitution and rearrangement we have the following equation: . + . + k = EE.+ killm] + kaMH[MeNH3] + kiKMHIIMJIMeNH3] cor I< KIMeNsz KIH KIH[MeNH2] + I I th kWKMHIHZO] [MGNH3] K + KIHIMGNHZJ + kaMeNHZ] + ki[Im] + thWIMeNHz] 4. KMH [MeNH3] where each small letter k on the right hand side of the equation is a rate constant for one of the base-catalyzed routes as listed in Table 1. Each capital letter, K, is an equilibrium constant as defined below: [H+][MeNH2] K = MH [MeNHg] Kw = ion product of water K _ (imine) (MeNH2)(aldehyde) K (protonated imine) IH = (H¥)(neutral imine) 32 TABLE l.--Cata1ytic Constants in Rate Equation Describing Enolization of Isobutyraldehyde in Methylamine Buffers. Rate Constant Species Base km aldehyde MeNH2 ki " neutral imine kh " OH- ké neutral imine MeNH2 k; " neutral imine kg " OH- kfi protonated imine MeNH2 ki " neutral imine kfi " OH- ké " H20 33 k kh Kw The constant k = _E._ + cor fI K KMHIMeNH3] pseudo first order rate constant from where kp is the -d(imine - d) dt = k (imine - d) P and fI is a factor which corrects for the loss of deuterium from the imine as a result of deuterium ex- change from the aldehyde in rapid equilibrium. As it stands, the equation is too complex to solve for the desired constants. The only recourse is to dis- card certain terms as "negligible“ based on carefully chosen arguments. First, there are two terms that never entered the equation at all: the water-catalyzed enolization of the aldehyde and the water-catalyzed enolization of the neutral imine. Previous experiments showed the water- catalyzed enolization of the aldehyde to be very small; from this Hine reasoned that the neutral imine would be even less susceptible to water catalysis and could also be ignored. The next term to be discarded was the last in the above equation by realizing that if the equation is put in the form kg Kw[MeNH2] + KMH[MeNH3] +- _. I kcor — c + c [MeNH3] + 34 a plot of kc versus MeNH; at constant MeNH2 and imine or will be linear only if the last term is negligible. This was found to be the case. Since this discarded term represents enolization of the neutral imine by hydroxide ion, Hine reasoned that the other weaker bases should be even less capable of catalyzing the enolization of neutral imine. There- fore, the terms containing k“, kg, k; can also be dis- carded. Finally, if the neutral imine functions as a general base catalyst, kcor should increase with increas- ing imine. Since this does not occur to a significant extent the terms containing ki, ki, and k; can be neglected. With this weeding process complete, the simpli- fied equation is: + I I kh Kw + km KMHIMeNH3] K K IH IH cor Thus, the slope of kcor versus (MeNH3) gives kfi KMH/KIH from which kfi, the constant for the methylamine-catalyzed enolization of the protonated inine, can be calculated. For this, KIH must be known--but it is not. Just as with the acetone system, this constant had to be estimated. The intercept gives the sum of the two kinetically equivalent rOutes--the amine-catalyzed enolization'of the 35 aldehyde and the hydroxide ion-catalyzed enolization of the protonated imine. Hine concluded then, that enolization occurs pre- dominantly via amine-catalyzed enolization of the pro- tonated imine with a rate constant being three hundred times greater than the amine-catalyzed enolization of the aldehyde. Hence: as expected, the protonated imine is more reactive than the aldehyde. In evaluating this work, the arguments used for discarding the terms in the rate law must be questioned. But aside from this, this reviewer is most puzzled by the lack of any arguments for discarding the term for the water-catalyzed enolization of the protonated imine (kw). Somehow, it just vanished. Recall that with acetone in methylamine buffer Bender found only water-catalyzed enolization of the Schiff base. Without some eXplanation for this omission, it is difficult to see how the linearity of kcor with (MeNHB) can be interpreted as it has. If the kw term is not omitted, the slope becomes the sum of the water-catalyzed and amine-catalyzed routes with no means for distinguishing the individual contribution by each. Enolization: Enzymatic Rose and Rieder (36) first discovered that in tritiated water, muscle aldolase catalyzed the exchange 36 of tritium with the pggfgfhydrogenz from the hydroxymethyl of dihydroxyacetone phosphate (DHAP). Since DHAP has two equivalent hydrogens at the hydroxymethyl position, the aldolase can apparently distinguish one from the other. This conclusion is supported by the observation that hexose diphOSphate (HDP) which results from the conden- sation of DHAP with G3P, incorporated no tritium; since the hydrogen at the C-3 position of HDP derives from DHAP, random exchange of DHAP hydrogens would necessarily have produced labeled HDP. In a later detailed study (37) these workers con- sidered the possibility that the exchange reaction might actually depend on contaminating levels of aldehyde. By this view, the aldehyde forms the condensation product which, upon subsequent reversion to trioses, produces labeled DHAP. They dispelled this possibility by showing that HDP actually inhibits the exchange reaction of DHAP. They went on to show that the kinetics of exchange and condensation were consistent with the following scheme: 2This stereochemical designation is discussed by K. R. Hanson, J. Am. Chem. Soc. 88, 2731 (1966). 37 DHAP + E -—l* E-DHAP (exchange) k k ‘ G3P + E-DHAP -2* E—HDP -—3~ E+HDP (condensation) That is, exchange and condensation occur at the same site on the enzyme, which implies that they are mechanistically linked. According to this scheme, the exchange from tritiated DHAP (in the absence of GBP) should obey the rate law ”Pficom where vx is the rate of exchange, KX the dissociation con- stant for E-S complex, and D the concentration of DHAP. As predicted from this rate equation, a plot of l/vx versus l/D was linear. The scheme also predicts that as G3P is increased, the rate of exchange from tritiated DHAP will become equal to the rate of condensation. This should occur because at higher levels of G3P, condensation with E-DHAP competes ever more favorably with dissociation of E-DHAP. The re- sults of the experiment were in accord with these pre- dictions. 38 Since that time, other enzymes have been shown to catalyze similar exchange reactions (38, 39): KDPG aldolase, u-keto glutaric acid aldolase, and acetoacetic acid decarboxylase. CHAPTER I I I METHODS Synthesis of Al-Piperidine-2-Carboxylic Acid The synthesis of Al-piperidine-Z—carboxylic acid (Al-PCA) has been accomplished by Meister according to the following procedure (40): First, N-e-amino-carbobenzo- xylysine is converted to a-keto carbobenzoxylysine (a-keto-CBZlysine) by Beamino acid oxidase. Since hydro- gen peroxide is a product of the reaction, catalase is included to prevent the decarboxylation of the a-keto acid. The a-keto CBZlysine is isolated from the mixture by acid extraction into ethylacetate followed by evaporation of the solvent and crystallization with petroleum ether. The carbobenzoxy group is removed with acetic acid--HBr reagent from which the transient product e-amino-a-keto caproic acid is produced. This compound spontaneously cyclyzes to the hydrobromide salt of Al-PCA. The entire sequence is illustrated in Figure 1. Unfortunately, this author could not realize this synthesis when Meister's procedures were rigorously followed. It has come to his attention that other 39 40 .Uflom oflamxonumolmlmcwpflummflmla< mo mammnucmmll.a musmflm ~om+owx $226; ~oo+emzz ~o~z 28 £389.60 + + -~-oc§§t._< 05230-2»..-8 23233123..— .mu $1sz ., 39.2%: 30.2%: :ooo.o\ «mm «:w ~zw + NH .5 4.. .50 NI_U .540 «1.0 NImv on. cum «zzw: . :08 . :08 :80 42 laboratories have also met with failures. Macholan and Svatek (4) did achieve the synthesis by what they called a "modified" Meister procedure, upon which they did not elaborate. Because of these difficulties, this author feels it necessary to describe in detail the problems he en- countered and the modifications required to successfully synthesize the compound. At first, the Meister procedure was attempted as directed. 5.0 g a-keto-CBZ lysine were suspended in 200 ml triply distilled water and adjusted to pH 7.0 with 2 E NaOH. After equilibration to 37°C in a water bath, 40 units of Efamino acid oxidase were added. The solution was carefully adjusted to pH 7.0 again, and replaced in the water bath. Oxygen was continuously bubbled through the solution during the course of the reaction. Stirring was effected with a submersible magnetic stirring motor. Foaming was controlled by occasional addition of Corning Antifoam. The reaction was followed by acid-extraction of the product from 2 ml aliquots with ethylacetate and measuring the characteristic ultraviolet absorption spectrum of the carbobenzoxy (CBZ) group. The reaction did not proceed as fast as Meister indicated; after 48 hours, the spectral determination indicated that only 2 g a-keto CBZ lysine had been converted. 43 The reaction was terminated at this point by acidification of the mixture to Congo Red followed by extraction with several portions ethylacetate. The combined extracts were dried overnight (NaZSO4). The solvent was then removed in 33222, which produced a large amount of white crystal- line material. This material was redissolved in a small volume of ethylacetate and recrystallized by addition of petroleum ether (bp 40-55°C). After standing at —10°C overnight, the product was collected by suction filtration and washed with petroleum ether. The melting point was 105-106.5°C (reported mp a—keto-CBZlys; 109°C); yield = 1.3 g (26%). The CBZ group was removed as directed by addition of 1.0 ml HOAc--HBr reagent to 250 mg a-keto-CBZlysine. After an hour when the vigorous reaction had subsided, the white crystalline product was precipitated and washed with ethylether; yield, 80 mg (41%). However, the product melted at 105°C, far from the reported melting point of 190°C for Al-PCA. In addition, the melting point was quite sharp, indicating that a pure compound had been synthesized. Comparison of the uv- visible absorption spectrum of this material to those published by Macholan showed definitely that the compound was not Al-PCA. Although the shape of the spectrum could-- with some imagination--possibly resemble that of the authentic compound, the extinction coefficient at 256 mu 44 was only about 3.0 instead of the reported value 725 mu. The spectrum did show, however, that the CBZ group had been removed. The HOAc--HBr treatment was first examined as a source of problems. It has been reported that bromination sometimes occurs as a side reaction (41). However, bromi- nation can successfully be suppressed by addition of phenol to the reaction mixture. When this modification was tried, however, the product had the same character- istics as before. It was therefore concluded that the source of trouble was elsewhere, perhaps in the oxidation of Ere-amino-CB21ysine itself. Since an extractable product was formed, oxi- dation had certainly occurred. However, if for some reason peroxide had not been sufficiently removed by catalase the a-keto derivative could have been quanti- tatively decarboxylated. In that case, the material isolated would actually be e—amino-CBZ valeric acid: COOH T=O COOH CH H O CH I2 22 '2 +c02 CH CH I2 l2 THZ CH2 HCNCBZ H CNCBZ 2 45 To test this possibility, the extractable material was tested for a carbonyl function with 2,4-dinitrophenyl- hydrazine and semicarbazide. Meister had reported a 2,4-dinitrophenylhydrazone derivative of the a-keto-CBZ- lysine. However, all attempts to form this derivative failed. Likewise, the compound was unreactive to semi- carbazide. As a further test, the titration curves for this compound and the one recovered from HOAc—-HBr treatment were determined. If the extractable material is Nee- aminoCBZvaleric acid, the acid group should have a pKa between 4 and 5, whereas that for the a-keto product should have a pKa between 2 and 3. The curve clearly showed only one titratable group with a pKa between 4 and 5. The HOAc-—HBr treated material showed two titrat- able groups, one with a pKa = 4-5, and the other with a pKa greater than 10. This is consistent with the expected product N-e-aminovaleric acid. The above results provide good evidence that N-e-amino CBZlysine had first been converted to a-keto- CBZlysine then decarboxylated to N-e-amino-CBZvaleric acid. The decarboxylation problem was eventually solved by using very large amounts of catalase. But in addition, other objectionable features of the synthesis had to be dealt with. First, the CBZlysine is not only insoluble 46 in water, but floats as a large foam over the solution. This problem is exacerbated by the magnetic stirrer. The most diligent attempts to wet this material down during the course of the reaction were not successful. Second, the pH is not well controlled under these conditions; CBZlysine is not a good buffer near neutrality. Third, when the HOAc-—HBr treatment is carried out as directed the solution turns very dark yellow or orange. When ether is added, a small yellow layer settles to the bottom and is very difficult to remove from the product. Fourth, as mentioned earlier, the reaction rate is not as fast as it is reported to be. With all these objections in mind, the successful synthesis will not be presented. The modifications will be discussed as they arise. First, the magnetic stirrer was replaced by a shaking water bath. Since shaking is not as effective in distributing the bubbled oxygen through the solution the reaction volume was scaled down ten-fold. Shaking was much more effective in keeping the CBZlysine in suspension. The pH was maintained with 0.1 g HEPES buffer, pH 7.5. 500 mg CBZlysine in 20 ml HEPES buffer was placed in a 50 m1 Erlynmeyer flask which was then stoppered and vigorously shaken by hand to suspend the CBZlysine as well as possible. After this mixture had reached 37°C on the water bath, the reaction was initiated by the addition of 20 units of Lfamino acid oxidase (Worthington) and 10 pl 47 catalase suspension containing 206,500 U/ml. The level of amino acid oxidase is much higher than the Meister procedure calls for, in order to increase the rate of reaction. More importantly, it was found after repeated attempts, that extraordinary amounts of catalase are needed to prevent decarboxylation. For this reason, 10 ul catalase was added every 15 min throughout the course of the reaction. Water-saturated oxygen was continuously bubbled through the solution. Foaming was controlled by an occasional drop of l-octanol. The reaction was terminated after approximately fourteen hours even though it had not proceeded to com- pletion. After acidification to pH 3.0, the extraction and crystallization was carried out according to Meister's procedure with little complication. Approximately 100 mg product were recovered. Because of the small amount of material, the recrystallized product was collected and washed by centrifugation instead of suction filtration. The product melted at 100-103 (g,£. 106°). This time the product gave a positive semicarbazide test and readily formed a phenylhydrazone derivative, mp 145-155 (g,£. mp 160°C). The HOAc—-HBr treatment was performed as follows: 2 m1 HOAC--HBr was pretreated with 50-100 mg phenol, which later prevented formation of the yellow contaminant. 0.5 ml of this reagent was added to 25 mg a-keto-CBZlysine. 48 After one hour, when CO evolution had ceased, the product 2 was precipitated and washed several times with ether by centrifugation. After perhaps the second ether-wash, the product sticks tenaciously to a glass stirring rod. This makes washing impossible, but can be used to advantage by transferring the precipitate on the glass rod to a fresh tube of anhydrous ether. After 15 min of agitiation, the product comes loose from the rod and settles to the bottom of the tube. From this point on, washing can be continued without problems. The product was dried in 32229 overnight (mp 190°C, c.f. 192°C). Absorption spectra of the compound in water and 0.1 E_NaOH (Figure 2) were identical to those published by Macholan. In 0.1 §_NaOH, the extinction coefficient at 256 mu (3 = 725) usually agreed with the reported value within 5%. Buffers All catalysts except potassium phosphate were recrystallized before use as follows: nfButylamine hydrochloride (Eastman) was dis- solved in hot alcohol then allowed to cool to room temper- ature. When the initial precipitation was nearly complete an excess ethyl ether was added which precipitated the remaining material. The mixture was then heated on a steam bath until just enough ether had evaporated to allow dissolution of all the material. At this point, the flask 49 Figure 2.--Absorption spectra of Al-piperidine-Z- carboxylic acid in water and 0.1 E NaOH. The spectra were determined with a Cary Model 15 recording spectro- photometer. Each solution contained 0.1433 mg per ml of the hydrobromide salt of Al-piperidine-2-carboxylic acid. 50 0.8- 0.6- 3 0.IN NaOH 1: /'H20‘ 0 .g 0.4 / m .D 4 0.2 , o 250 360 Wavelength , m p 51 was quickly stoppered and allowed to cool. Imidazole (Sigma) was recrystallized several times from benzene. n-e-Aminocaproic acid was recrystallized from EtOH--HZO-- ether. The buffers were adjusted to the desired pH with NaOH or HCl. For most determinations a Beckman pH meter with expanded scale was used. All buffers contained 0.2 g KI and NaCl to maintain the ionic strength at 0.4. Iodination Procedures The enolization of ketones (or ketimines) can be followed by rapid halogenation of the enol (or enamine). Under proper conditions, halogenation is zero order with respect to halogen concentration so that the rate of halogenation equals the rate of enolization. In the present case, iodination was used. In a concentrated solution of KI, I is mostly 2 O —_A- 3.12+1 ‘_I3. Since the I3 species has an intense absorbance maximum converted by a rapid equilibrium to I at 351 mu (6 = 26,400), the iodination can be measured spectrophotometrically by following the loss of absorbance at this wavelength. The iodinations were performed as follows: 1.0 ml buffer was placed in a cuvette and equilibrated in the cuvette compartment of a Gilford Model 2000 spectrOphoto- meter with automatic cuvette changer. A Haake circu- lating water bath was used to maintain the temperature 52 at 25°C. After equilibration, iodine in 0.2 §_KI was added with a Hamilton microsyringe to a final concen- tration of approximately 3 x 10.5 g. A blank rate was determined prior to addition of pyruvate. The reaction was initiated by the addition of pyruvate in large excess over iodine, with a Hamilton microsyringe. The rates were linear except near the very end where most of the iodine had been used. This linearity indicated that the re- action was indeed zero order in iodine. As a further check, the reactions were occasionally repeated at differ- ent iodine concentrations; in all cases the results were identical. The rate of enolization was calculated from the rate of iodination on the basis of an extinction coef- ficient 2.6 x 104 for the triiodide ion (42). Although this method is well established by now, Coward and Bruice (43) have vigorously assaulted the use of halogenation to measure enolization. In particular, some of Bender's work with acetone was questioned. Bruice and Coward were study- ing the enolization of B-amino ketones of the form 53 Specifically, they were looking for intramolecular catalysis by the tertiary amino group. They found the reaction of these compounds with iodine to be so fast that the above initial-rate method (in which substrate is in excess over iodine) could not be used. Instead, they followed the first order reaction of substrate with iodine in excess. Since the high concentration of iodine pro- hibited spectrophotometric methods under these conditions, a pH-stat assay was used to measure the hydrogen ion pro- duced upon iodination. These workers raised an additional objection to the initial-rate method because such a small fraction of substrate is reacted during the determination. Conse- quently, they reasoned, a small amount of highly reactive impurity could seriously interfere with the results. While this is true, this objection could be made for many kinetic analyses which utilize initial-rate measurements. They also found the first order method objection- able because the iodine in such high concentration reacts with the buffer components at a rate comparable to the reaction with substrate. According to them, the reaction of iodine with phosphate buffer is so fast that they questioned the validity of Bender's results with this buffer. However, this author, who has also used phosphate buffer for measuring the enolization of acetone and pyruvate, found the blank rate to be virtually 54 immeasurable with this buffer in the pH range 5.75-6.50. It is true that reaction with buffer and hydroxide ion does become increasingly bothersome at higher pH, and especially with amine buffers. However, the enolization of acetone and pyruvate is sufficiently fast so that the reaction of iodine with buffer can be subtracted without a severe loss of accuracy. What is most puzzling to this author is that Coward and Bruice could obtain any meaningful data at all with tertiary amines by iodination procedures. Bender reported that the enolization of acetone in trimethylamine could not be measured by iodination since trimethylamine reacts rapidly and irreversibly with iodine to form an N-iodo complex. (Primary and secondary amines also form this complex but to a lesser degree which does not inter- fere with the reaction.) Instead, Bender had to use deuterium exchange to measure the enolization with this buffer. pH—stat Iodination Procedures Iodination of pyruvate and Al-piperidine-Z- carboxylic acid in the absence of buffer was performed on a Sargent Recording pH-stat in order to maintain con- stant pH. Normally, 0.001 or 0.005 N NaOH was used for this purpose. 55 Prior to addition of substrate, a solution con- taining 0.2 M KI, 0.2 M_NaCl and 3 x 10-5 E'IZ in 50 ml was equilibrated to 25°C on the pH-stat and adjusted to the desired pH. Although the pH-stat was used to maintain pH, all pH readings were made with a Beckman pH meter with expanded scale and combined glass electrode. The reaction was initiated by the addition of substrate in 0.5 ml and the iodination followed by with- drawing aliquots at timed intervals and measuring the absorbance at 351 mu with a Beckman DU spectrophotometer. Determination of pKa of Q-butylamine — Since the dissociation constant of any acid is sensitive to the ionic strength and, sometimes, the par- ticular ions in solution, the pKa of nebutylamine had to be determined under iodination conditions (0.2 M.KI + NaCl; I = 0.4). This is done in principle by half neutralizing a solution of n-butylaminehydrochloride with NaOH so that the concentration of conjugate acid equals the concen- tration of conjugate base. Then, according to the n'- BuNH2 Henderson-Hasselbach equation (pH = pKa + log ——————7¥§) n - BuN the pKa equals the pH of the solution. Following this procedure, the pH of a solution containing 0.100 M butylaminehydrochloride, 0.2 M KI, 0.1 M NaCl, and 0.05 N NaOH was found to be 10.85. For this determination, a Beckman Model G pH meter was used. 56 The validity of the technique was assured by com- paring the pKa of nfbutylamine in water alone to the literature value (44). The uncorrected value was 10.71 at 25°C, which compared very well to the reported value of 10.66. However, the pKa determined by this method has to be corrected for the amount of NaOH needed to achieve pH 10.85 in the absence of amine. This was done with a blank titration of a solution containing 0.2 M NaCl and 0.2 M_KI. After appropriate corrections, the pKa of nfbutylamine at 25°C was found to be 10.896. Determination of the Equilibrium Constant of Schiff Base Formation Perhaps the best method for determining the equilibrium constant for Schiff base formation (KS) is the polarigraphic method used by Zuman (45). Zuman deter- mined KS for a number of aldehydes and ketones (including pyruvate) with glycine, alanine, ammonia, and histamine. Unfortunately, simple amines like nebutylamine were not included. Since polarigraphic apparatus was not available to this author, a spectrophotometric method was devised. When pyruvate is placed in amine buffer, the pyruvyl ab- sorption at 320 mu is decreased and a new absorption appears at about 250 mp. Notice that the spectra in Figure 3 are difference spectra. If the absorption of either the Schiff base or pyruvate can be measured 57 .mm may CHmyCHmE oy ymmmsa mCHEm may CH pmpCHUCH mmz mymammoam 2 mo. 0 .o. m 0y m mmCmy mm may CH .Hmmmaa mCHEm may no mm may 0y pmymsnpm ymmmsa mymammoam 2 mo. 0 CH mym>summ z mIoH x mm. m UmCHmyCoo ymayo may .ymmmsa Nmzsmla cmCHmyCoo mCO .Emma moCmymmmy may CH UmomHm mums EopCmy CH mmyym>so mo yHmm m .mmHSmxHH .mym>5y>m z mIOH x mm. m cam ummmsa mmzsm z o. H .HE mCo CH meHmyCoo ymayo may Cam ymmmsa mymammoam vaHmyCoo yHmm may mo myym>50 mCO .HE m. H mo myHommmo mECHo> m UCm aymm yamHH ymymEHyCmo mCo m mCH>ma aomm .Emma mHmEMm may CH UmOMHm mym3 wmyym>so Eocamy 039 .ymymEoyoamouyommm mH Hmpoz ammo m ayH3 UmCHEymymv mymz myyommm maB .mymwmsa mCHEmHmysaIC CH mym>5ymm Ho muyommm moamymmmHQII.m mysmHm 58 as .59.... 263 awn on» o I‘IL‘E ‘ g \\\ no \ totem , \ to 656253-... c. 232?. .6 3.2.3 35333 6.0 aauoqaosqv v 59 independently of the other, the equilibrium constant can be readily calculated. Assuming the absorption spectrum of the Schiff base is roughly equivalent to that of Al-piperidine—Z-carboxylic acid, it can be seen that the absorption spectra of pyruvate and Schiff base overlap significantly at about 250 mu, whereas only the pyruvate spectrum shows measurable absorbance at 320 mu. However, this could be an illusion. The extinction coefficient of pyruvate at 320 mu is so low (a = 20) that the “tail" of the more highly absorbant Schiff base peak might overlap significantly with the pyruvate maximum at 320 my. For this reason the spectral method, which will now be de- scribed, was tested with glycine for which Ks could be compared with the polarigraphically determined value. The equilibrium expression is K: [S] s IPiIAm:i where K8 is the equilibrium constant; P is the concen- tration of pyruvate at equilibrium; Am: is the concen- tration of the conjugate base of the amine. The following terms are now introduced: P0 = initial concentration of pyruvate A0 = absorbance of PC at 320 mu. 6 = extinction coefficient of pyruvate at P 320 mu. A = absorbance of P at 320 mu. 60 It can be seen that which is substituted into the equilibrium expression to give PO-P Ks = Am: P Also A i=_€_=é_ P A A o _g o 6 So that ._ A P (A ) Po 0 Substituting into the previous equilibrium expression: A ”’0". (7;) (Po) 0 o —A-— (Am.> (A0) (PO) 159%) . . 0 (Am) (it) 0 from which KS can be determined. Am: is calculated from the known pKa of the amine. 61 KS for the glycine-pyruvate system was determined by measuring the loss of 320 mu absorbance of a solution containing in 1.0 ml, 1.0 M glycine, and 0.01 M pyruvate adjusted to pH 9.85. The readings were made with a Gilford Recording Spectrophotometer. Equilibrium was essentially attained in less than five minutes. AC was determined from the absorbance of pyruvate at 320 mu in 0.05 bi- carbonate buffer pH 9.85. Using this method, KS was found to be 2.54, in excellent agreement with the polarigraphic value of 2.47. This result supports the validity of the spectral method. In a similar fashion, KS was determined for the nfBuNHZ-pyruvate system at five pH values between 9.55 and 10.65 from equilibrium mixtures containing 0.01 M pyruvate and 0.4 M BuNH The results are shown in 2. Table 2. For some reason, KS decreases with increasing pH; this trend was found to be reproducible. Apparently, another pH—dependent ionization is taking place. It should be mentioned that Zuman did not study the pH- dependence of the equilibrium, but simply measured Ks near the pKa of the amine. 62 TABLE 2.--Equilibrium constant of Schiff base formation. -1 pH KS(M ) 9.55 14.28 9.75 13.375 10.04 12.182 10.30 10.65 10.65 7.96 63 Stopped Flow Measurements All stopped flow determinations were performed with a Durrum Gibson stopped flow apparatus fitted with a Beckman DU Optical system. The change in transmittance was recorded with a Tektronix 564 Storage oscilloscope from which photographs of the traces could be made. The temperature was maintained at 25°C with a circulating water bath. The rate constants for Schiff base formation and hydrolysis were determined from the increase in absorbance at 256 mu upon mixing equal volumes of 0.01 M pyruvate and 0.4 M nfBuNH buffer. Since the amine is in great 2 excess, the forward reaction is pseudo first order in pyruvate; the reverse reaction, of course, is truly first order in Schiff base. Therefore, the rate data can be expressed in terms of a reversible first order process in both directions according to the integrated equation: A - A L911 = _ ln( A _ A ) (kf + kr)t — kO t eq bs where A0 is the absorbance at time zero, Ae is the absorbance at equilibrium, A is the absorbance at time t, kf is the pseudo first-order rate constant for the forward reaction, and kr is the first-order rate constant for the reverse reaction. 64 A - A o e A - A versus t. kf and kr can then be calculated from the two equations kobs = kf + kr and Ke k is then determined from a plot of 1n obs q = kf/kr' where Ke is the equilibrium constant for Schiff base formation. q The stopped flow iodinations of Al-piperidine-Z- carboxylic acid were performed by mixing in equal volumes a solution containing Al-PCA in imidazole buffer with a solution containing buffer and iodine. As usual, the buffers contained 0.2 M_KI and NaCl to maintain the ionic strength at 0.4. Enzymatic Crystalline KDPG-aldolase was prepared according to the method of Meloche and Wood (46). Briefly, the purification involves the following steps. First, 200 g Pseudomonas putida cells are disrupted by sonic oscillation and the cell debris removed by centrifugation. Since the aldolase is uncommonly acid stable, the bulk of protein is then removed by acidification to pH 2.0 followed by centrifugation. The supernatant is further fractionated with ammonium sulfate. The final ammonium sulfate step precipitates the aldolase, which is redissolved in a small volume of water. The last step prior to crystallization is fractionation with calcium phosphate gel. Crystalli- zation is effected with ammonium sulfate. 65 The enzyme is assayed by the method of Meloche and Wood (46) which employs a lactic dehydrogenase coupling system: KDPG EAQEAEES- pyruvate LDH lactate + /\ G3P DPNH DPN The disappearance of DPNH is followed at 340 my in a Gilford Recording Spectrophotometer. CHAPTER IV RESULTS Introduction As discussed in the Introduction, two systems were used to study ketimine-enamine tautomerization: pyruvate in amine buffer and the cyclic ketimine analogue Al- piperidine-Z-carboxylic acid. The tautomerization of Al-piperidine-Z—carboxylic acid is simply measured directly by iodination as a function of pH. From these measurements, the rate law is elucidated and the desired constants calculated. The pyruvate-amine mixtures, however, can only be used for this purpose after certain fundamentals have been established: (1) Schiff base formation must indeed occur, and (2) ketimine-enamine tautomerization must be rate limiting. Once these have been established, the rate law can be elucidated and the constants determined. These constants are then compared to the corresponding constants obtained in buffers which do not form a Schiff base, in order to determine the contribution of the Schiff base towards the overall rate of enolization. This system will be discussed first. 66 67 Demonstration of Schiff Base Formation First, it is necessary to demonstrate that a Schiff base is formed from pyruvate and_simple alkyl amines in aqueous solution. This was done by comparing the absorption spectra of mixtures containing pyruvate and amine buffer (Figure 3) to those of the Schiff base Al-piperidine-2-carboxylic acid (Figure 2). As expected, in amine buffer the carbonyl absorbance of pyruvate at 320 mu decreases while a new absorbance appears below 300 mu. Furthermore, the pH dependence of the spectra is very similar to that observed for Al-piperidine-Z- carboxylic acid. This conclusion that the spectral changes are characteristic of a Schiff base linkage is also supported by the fact that the spectrally determined equilibrium constant for Schiff base formation from pyruvate and glycine agrees well with the polarigraphically determined value (see Methods, p. 59). It has been hoped, that the Schiff base formed from pyruvate and amine resembles that formed from pyru- vate and the lysine e-amino group in KDPG aldolase. For this reason, a spectrum of the pyruvate-enzyme complex was determined, as shown in Figure 4. It can be seen (Figures 3 and 4) that the spectrum of the pyruvate- enzyme complex does resemble that of the model-system Schiff base. This does not mean that the enzymatic and 68 I .mym>summ cam .mymammoam z H.o .m.m mm ummmsa mymyoa m m.o .Hm m.o CH meHmyCoo ymayo maBl .mmmHova Omox mo myflcs comm can .Cmuusn mymnmmonm z H.o .m.m mm umuusn mymuog z m.o .He m.o cy UmCHmyCoo mCo .Emma moamymmmy may CH UmOMHQ mumB EOCCmy CH mmyym>so mo yHmm m .mmHB ImaHH mm «IOH x NHIv CoHymyyCmoCoo mo mym>su>m pCm .mfiKD ooo.OH wa>Hyom UHHHommm I mo mmMHopHm Odom mo myHCC comm UCm Hmmwda mymammoam z H.o .m.m Mm ymmmaa mymyoa z m.o .HE m.o CH UmCHmyCoo umayo may mmmyma3 .m.m mm .ymmmsa mymyoa z m.o pCm ymmmsa mymammoam m H.o .HE m.o CH UmCHMyCoo HHmm may mo myym>so mCO .HE m.H mo myHommmo mEdHo> m can aymm yamHH HmymEHyCmu mCo m mCH>ma aomm .Emma mHmEMm may CH CmUMHm mumz mmyym>so SOCCmy 039 "monHom mm ymymEoyoamoyyommm mH Hmpoz mymo m CH meHEymymp mymz myyommm maa .mmydnyE mmmHomHm omoxlmym>5y>m mo myyommm moCmymmmHoll.v mquHm 69 as £35.26; 8» H.» 8.” RN TIE 2253a 2 v0. x c \ 222$ : ¢o_xo.\ 225.5. 2 on. .3. 00.0 . 5.0 .86 $0.0 aouoqmsqv v 70 nonenzymatic Schiff bases are identical in ionic form and environment, but the spectra are at least consistent with the notion that both Schiff bases are similar in many respects. Kinetics of Schiff Base Formation Before enolization experiments with pyruvate- amine mixtures can be realized, it is necessary to assure that in amine buffers Schiff base formation is much faster than enolization; if it is not, Schiff base for- mation could be rate limiting, and the iodination assay for enolization would obviously be a measure of Schiff base formation rather than the ketimine-enamine con— version. As can be seen from Figures 2 and 3, the rate of Schiff base formation can be measured by the increase in absorbance at 256 mu. By this method, it was readily observed that Schiff base formation was indeed much faster than the rate of enolization (as measured by iodination) over the pH range 9.2 to 10.6. In fact, the reaction is too fast to permit determination of the rate constant by standard spectrophotometric means. These simple--and somewhat qualitative--observations provide sufficient evidence that enamine formation is rate limiting. For the experiments planned, it is not neces- sary to determine the rate constant for Schiff base formation. Nonetheless, it may be of some value to 71 determine this constant. The rate of Schiff base for- mation is apparently very fast; how then might it compare to the enzymatic rate of Schiff base formation? To answer this question, the rate constants for Schiff base formation and hydrolysis of the model reaction in the pH range 9.2 to 10.6 were determined by stopped flow techniques as discussed in the Methods section. Plots of ln(f%%;;;;§1) versus t are shown in Figure 5. The rate constaigs so obtained are summarized in Table 3. Although a full discussion of these results will be deferred to the Discussion section, it may be stated here that the rate constant for Schiff base for- mation between pyruvate and butylamine is at least two or three orders of magnitude less than the enzymatic rate constant for the enolization of pyruvate. Since the enzymatic rate of Schiff base formation must be as fast as enolization, the enolization constant of the enzyme is a lower limit of the constant for Schiff base for- mation. Therefore, the enzymatic rate of Schiff base formation is at least two or three orders of magnitude greater than the nonenzymatic rate. 72 .Oomm ym UmCHmyCHmE .mysymymmEmy may Cam «om.OH mH CoHyommH CMHsoHyymm mHay mo mm mas .18 mmm ym moamaHOmam CH mmmmyoCH may ma Um30HH0m mmz CoHyommy maB .mym>CH>m m moo.o Cam ymmmsa mmzsm m N.o .mCHxHE Hmymm .UmCHmyCoo mysnyE COHyommu maB .mvaHCaUmy onm Cmmmoym ma Umysmmmfi mm mmma HHHaom Enom oy mCHEmHmyCa ayH3 mym>sywm mo CoHyommH HMOHQMy all.m myCmHm 73 .mucooomv we: m H m. N. . _ o o . o w: \I/ o -QOWAWW W V O b a ..D ( o .. 30.. Co 5:65.523 .3 "8:088“. $5 :28. 74 TABLE 3.--Rate Constants for Schiff Base Formation from Pyruvate and Butylamine. [The reaction mixture contained, after mixing in the stopped flow apparatus, 0.2 M BuNH buffer and 0.005 M_pyruvate. The reaction was followe by the increase in absorbance at 256 mu.) pH kobs kf Mmlsec-l kr sec-1 (at 0.2 M BuNHz) 10.56 0.1729 0.695 0.034 10.23 0.1264 0.472 0.0315 9.87 0.1039 0.349 0.0342 9.51 0.0939 0.229 0.0458 9.20 0.0940 0.171 0.0597 75 Enolization of Pyruvate Since the concentration of Schiff base inter- mediate is small compared to the concentration of pyruvate, a kinetic analysis similar to that used by Bender is more appropriate than an analysis used by Hine (see Literature Review). Therefore, the enolization of pyruvate in buffers which cannot form a Schiff base was compared to the enoli- zation in primary amine buffers. For this comparison, phosphate and imidazole buffers were used. The results of these experiments are shown in Figures 6-11 and the constants so determined are summarized in Table 4. Most of the studies with amine were done with n-e-aminocaproic acid since this amine is not volatile. To assure that the carboxyl group does not contribute to the reaction, the kinetics in this buffer were later corroborated with the kinetics in n-butylamine buffer. Generally speaking, the kinetics of enolization in phosphate, imidazole, and amine are very similar to those observed for acetone; in all cases, the kinetics could be expressed in terms of the simple rate equation rate = k SP = (k0 + k B + kaA)P. Most importantly, the rate ob b law for amine-catalyzed enolization shows a term corres- ponding to general-acid catalysis, whose catalytic con- stant ka is very much larger than the corresponding constants observed for imidazole or phosphate. As discussed earlier (Literature Review, p. 25) this term is best attributed to water-catalyzed enolization of the 76 .Oomm ym meHmy ICHmE mm3 mmmmo HHm CH mysymymmfimy mae .HE mCo mo mECHo> m CH mym>CH>m pCm v.0 ym aymCmuym UHCoH may CHmyCHmE oy Homz .NH .HM m «.0 UmCHmyCoo mmysnyE CoHyommy HHC .CoHymyyCmOCoo Hmmmsa pCm mm WCmymCou ym CoHymyyCmoCoo mym>symm msmum> mymy no myon Sony UmCHEymymp mHmB ma a you mmsHm> maB .yxmy may CH pmaHyommp CoHymCUm mymu HmmCHH may mo COHymEy0mmCmyy may oy mCHUHooom ax UCm ma mm>Hm amHa3 mmsmuglhl m+< u; H o m o H H + x n n x mo yOHm m mH UCoomm may UCM .CoHymyyCmOCoo ymmmsa msmym> maoa mo yOHm m mH ymyHm maB .C3oam mum myon 03y .Hmmmsa aomm mom a mac 6 0 CAC x + m x + xv u C x u mymu COHyvam mymy HmyCmEHymmxm may 0y mCHCHooom mym>sy>m mo CoHymC IHUoH may Eoym ax CCm .mx .ox myCmymCoo may no CoHyMCHEHmmeII.HHIm mmHCmHm 77 83 0on as: 22.82880 . totam fied ‘ 8.0.0 5:228:60 Exam ,myofimocm 330.. o i 3 so I,wQOqun ¢\'! ¢ 093 39 un c3. [s m mhsmHm 78 . .5 08.0 8.8 8.3 83 . . .o . H . \ , \\\ A . \ hm Ewomnb.xmmn.o «6......8225 . ..o.n _...2 .ma» ..o_x¢.w~ ..mxuoqgm . -08 O 5. use .3. . 6.2 .8 8.8583me 5:5 26:32... by. h myCmHm . QOI X RI 79 00.0 .2. 8.3.2880 Elam fied 80.0 mNQo . _ 5:228:00 totem 330.. Exam «.0325. 0 N <5 . v- .-w .9es ~90: qu°>l m muumHm 80 m.» 5:56 no. x m .o u a. 1822:. ....2 as .0. x N .9 “a. u 8% 3. .23 3. .0 5.8555me :55 23026.. m mHCmHm 81 .5: 00.80.5080 .825 00.0 00.0 000.0 0000 O O oo.m IQ O O o o 00.0 :0 o o o o 00.0 :0 o . O .o 00.0 ..a o _ . . 022.5008 .035 «>300. 0 0:5 0.04 0.0.08.0... .. Y: 00.0. :0 o z 0H 00:0H0_ 00‘ 82 .x. p 0.0 0.0 0.06 0.00 0.0.0 «0.0 0 r52 rs. m. ..0 ..u.. n 3022:. 70...... .s. no.0 a... 000.0 0x 000 00. 00 00:05:...38 .mtam 0.00.0000...2..Y.c HH muzmHm 83 TABLE 4.--Iodination of Pyruvate in Imidazole, Phosphate, and Amine Buffer: Summary of Catalytic Constants ka and kb. Buffer k k a b M.1 sec—l MDl sec-l Phosphate < 5.00 x 10’6 2.84 x 10"4 Imidazole < 1.00 x 10'5 2.53 x 10'4 3 N—e-aminocaproate 1.88 x 10- 0.0971 84 protonated ketimine rather than the general catalysis, for which the following mechanism can be written: coo" C=O + RNH CH3 pyruvate COO C=N-R + H CH neutral Schiff base + k1 COO k2 k2 HOC-NR k2 | H CH3 carbinolamine K coo” coo" —¥§ + ke I C=NR 5 C-N-R I H H H CH3 CH2 protonated Schiff base enamlne where enamine formation is rate limiting. Solving the rate equation for this mechanism in terms of pyruvate and protonated amine gives rate of iodination by ketimine SO and (53>(-5153— K2 k-lk-Z + )kGIRNH3JIPI Ka + (§;)(Kl)ke[RNH3IIPI K — .2 ka (K )(Kl)ke 2 k = kaKZ e K K 85 From this, the rate constant for the conversion of ketimine to enamine, ke' can be calculated if Ka' K1' and K2 are known. Ka’ the ionization constant of the amine was determined as described in Methods, p. 55. K1' the equilibrium constant for Schiff base formation was determined as described in Methods, p. 56. Since, as noted, the equilibrium constant varied somewhat over the l was chosen. pH range studied; an average value of 10 M7 As mentioned in the Literature Review, the ioni- zation constant for the Schiff base, K2, could not be experimentally determined for the acetone or isobutyralde- hyde systems. Neither could it be determined for the pyruvate system in this study. However, the stability of the cyclic Schiff base Al-piperidine-2-carboxy1ic acid to hydrolysis has permitted Macholan to determine the ionization constant of this compound with good accuracy. The value he obtained was pKa = 7.6. Since this compound is a cyclic analog of the pyruvyl ketimine, this value can properly be applied to those ketimines of pyruvate formed in amine buffers. With the above constants deter- mined, the value of the model system enolization constant, ke, is calculated to be 0.304 sec.1 as compared to the enzymatic rate constant lZOsec-l.3 3The rate constant for the enzyme-catalyzed enolization was calculated from tritium exchange data previously obtained in this laboratory (58). The value 0.25 uatom/min/International Unit of aldolase at 28°C (measured by the cleavage reaction described in Methods, 86 Assuming that all of the kinetic interpretations are correct, this result shows quite simply that Schiff base formation between pyruvate and KDPG aldolase is not sufficient to account for the enzymatic rate of enoli- zation. This will be discussed more fully later. As stated in the Introduction, the purpose of this research is to determine the contribution of the Schiff base to the enzymatic rate. For this purpose, the rate constants for the following three processes must be known: (1) the enzymatic-catalyzed enolization of pyru- vate; (2) the nonenzymatic Schiff base-catalyzed enoli- zation of pyruvate; and (3) the "uncatalyzed" enolization of pyruvate, which may be defined as the enolization of pyruvate in water at a pH near the pH-optimum of the enzyme. The rate constants for the first two processes have already been presented, leaving only that of the uncatalyzed reaction to be determined. Normally, the p. 65) agreed very well with the value reported by Rose, 0.41 natom/min/International Unit of enzyme (22). After correcting for the isotOpe effect, Rose estimated the rate of hydrogen exchange to be 1.92 uatoms/min/Inter- national Unit of enzyme (unfortunately, Rose did not re- port the temperature of the reaction). Presumably, Rose used this value to calculate the enolization constant (k6 = 400 sec'l), which appeared in a publication by Westheimer and Tagaki (38) as a personal communication to these authors. Our calculations indicate that this value is correct (we obtain ke = 423 sec-1) if it is assumed that there is one catalytic site per molecule of enzyme. However, studies in this laboratory have shown that there are three catalytic sites per molecule (59). Therefore, the enolization constant is 423 + 3 = 141 sec-1. At 25°C, this constant is estimated to be 120 sec‘l. 87 uncatalyzed reaction is simply determined from the inter- cept of a plot of buffer concentration versus rate of constant pH. However, the uncatalyzed enolization was found to be so small compared to the buffer-catalyzed enolization that this method proved useless. The only recourse was to find a method for measuring the enoli- zation in the absence of buffer. The obvious complication in such an experiment is maintaining the pH. It can be seen that iodination of pyruvate liberates a proton, which causes the pH to drop during the course of the reaction. This problem was readily solved, however, by controlling the pH with a pH-stat, as described in the Methods section. The reaction was measured over the pH range 7.2 to 8.2, which includes part of the pH-optimum range of the aldolase. A typical reaction is shown in Figure 12. The rate constants were determined as follows. Conceivably, the enolization can be catalyzed significantly by every acid and base species in solution. In that case the rate law would be _ - + rate - [kH O(H20) + kOH(OH ) + kH+(H ) + kP(P) 2 + . . . etc.]P 88 .oomm um Umcamuswme mm3 musumummEmu on» can m.h um omcflmucfime was cofluommu “Manoeuumm was» mo no one .HE om mo mEdHo> Hmuou m ca mum>su>m m T3 as... Homz m m6 .NH .51 m «5 35380 33st coflommu was 6939: umumumm may wn mum>su>m mo cofluomwu coflumcfipofl HMUHQ>B|I.NH musmflm 89 32:55 2:: 8 n 8 b 225:8 ca 5:058. 5. o T 9.- rfm (gt; ' aouoqmsqv 90 The last term in the equation represents the following: coo' coo’ | CH 3 l __ c=’0‘\ l kP c-o l3. C=O ——) H H-C H | CH l C 2 H / \\ e0 0 Although all the constants can be determined, it is much simpler to first assume that some of the terms in the above equation are negligible and plot the data accord- ingly. If the data fit the simplified equation, there is (no need to proceed any further. It appears by inspection 'that the reaction is predominantly catalyzed by hydroxide .ion for which the following rate law can be written: rate = kobSP = kOH(OH )(P) .EPigure 13 shows that a plot of hydroxide ion concentration (kactually hydroxide ion activity) versus rate is linear ‘aJld.passes through the origin. This result substantiates tllea simplified rate law shown above. The rate constant fc>1r the hydroxide ion catalyzed enolization as determined frtaan the slope of this plot is kOH = 0.360M-lsec_l. This ratzea constant enables the calculation of uncatalyzed rate at Etny pH value within this pH range. At pH 8.0, where the: fluom msmum> mnox mo uoam m m30£m ma musmfim .NH musmflm cw czocm cam uxmu mnu CH omnauommp mm ponumfi umumlmm map >3 panama Ibo was mcofiumsflEkump mmmnu How mumo one .ummmsn mo mosmmnm may GH mum>5n>m mo coflumcflpofi map How coflumsqm mumu map mo coHuMCHEHmumoll.ma musmflm 92 .2 ~.9. x Lac 9 o. ‘ m on. x. no; 1823:. ..2 .55 m._~m:o._u8o_m 0 ES ..xoo 333. 225.5 do 5553. .._ sqo o. 93 .lrgonv "wav— 3. 8 mm. no.0 ... 82m 2253a co cozosuo. 10¢ ..od mma muzuflm 94 The rate constants for all three processes de- scribed earlier can then be summarized below: 1. Enzyme-catalyzed enolization = 120 sec-1 2. Schiff base-catalyzed enolization = 0.304 sec-1 3. "Uncatalyzed" enolization = 0.36 x 10.6 sec-1 From this it can be seen that enolization of the protonated 6 Schiff base is 10 fold greater than the uncatalyzed enolization but 3 x 102 fold less than the enzyme cata- lyzed enolization. Schiff base catalysis, therefore, is highly effective, but only accounts for0.25% of the enzymatic rate. Enolization of A'-piperidine- 2-carboxylic Acid The shortcomings of the foregoing analyses have already been discussed in the Introduction and Literature Review. To obviate these shortcomings, a direct measure of enamine formation was sought with the stable Schiff base, Al-piperidine-2-carboxylic acid. It was hoped that the rate constant for enamine formation of this compound would agree well with that determined from the enolization of pyruvate in amine buffers. Recall that in amine buffer, an acid-catalyzed term was observed which was interpreted as water-catalyzed enolization of the protonated ketimine: coo' coo- | + H20 | H C=N-R ——> C-N-R l H II CH CH 3 2 95 For comparison, the same constant is desired for Al- piperidine-Z-carboxylic acid: “COO- C-COO 2122+ 522 However, preliminary experiments showed immedi- ately that there is a striking difference between the two systems: in imidazole and phosphate buffers, the iodi- nation of Al-PCA is strongly buffer-catalyzed, whereas no buffer-catalyzed enolization of the protonated ketimine was observed with pyruvate in amine buffer. Furthermore, the reaction is too fast in the presence of buffer to allow accurate determinations of the rate constants by ordinary means. Therefore, the iodination in imidazole buffer was studied by stopped flow techniques as described in Methods. The results at pH 7.25 are shown in Figures 14 and 15. It can be seen that the rate increases linearly with buffer concentration, but the rate of reaction at zero buffer concentration is negligible compared to the buffer-catalyzed rate. Therefore, this method is useless for determining the water-catalyzed enolization. 96 .momuu mmoomoHHHomo may ma cm>Hm mocmuuflfimcmuu uswouwm on» on mcflpcommmuuoo :E Hmm um mononuomnm omumasoamo map ma m>uso mau co ucwom comm .Uwom ceameQHMflHNImcapfinmawmla< m vloa x n.m mam .v.o um AWmcwHum cacOH ms» cflmncflms on Homz .NH 2 muoa .Hx z ~.o .m~.~ mm “manna maoumcflsa z mo.o nmcflm» Icon musuxHE :ofluommu map .mcwxflfi Hound .wsumummmm 30Hm commoum map nufl3 cmHSmmmE Uflom oHamxonumolmImcfipflummflmla< mo cowuommu coaumcflpofi HMUHQ>BII.eH musmflm 97 o. , “358$ 05: Boo 33.8980 852.35.? ,3 c3853. ‘3 <5. -0.— rfu: [92 ‘ aouoqxosqv 98 Figure 15.--Iodination of Al—piperidine-Z- carboxylic acid in imidazole buffer by stopped flow techniques: plot of rate versus buffer concentration. AA--...____. —- 99 / sec g rote . A Absorbance 3 9 ,9 9 Iodination 0f AI-Piperidine Z-Carboxylic acid 0 00'25 x 0050 003/5 Buffer Concentration M 0300 100 Fortunately, the reaction is slow enough in the absence of buffer to allow measurements to be made by the pH-stat procedure used for pyruvate. The iodination was studied over the pH range 7.2 to 8.2, as shown in Figure 16. Table 5 shows the rate of iodination as a function of pH (pH-rate profile) over this range. The rate data is analyzed in much the same way that it was for pyruvate, only in this case the situation is somewhat more complicated, for the following reason. Since the pKa for the protonated Schiff base is 7.6, the ratio of protonated to unprotonated form changes dramati- cally over the pH range studied. And since the enolization constant for the two forms is undoubtedly very different, the complete rate equation must take into account the concentration of each form as a function of pH. The analysis can be simplified considerably by making the reasonable assumption that enolization of the neutral form is negligible compared to enolization of the protonated form. (If this assumption is not correct, the simplified rate equation will not satisfactorily describe the experimental data.) Then, if enolization is attributed entirely to the protonated Schiff base, the resulting rate equation is: rate = [kHZO + kOH-(OH') + kS(S)](SH+) 101 .Uomm um .mnsumummfimu map can “v.5 um omcwmucfime mmz cofluommu umHsoHuumm mficulmo mm 029 .HE om mo mEdHo> Hmuou m an chow unasxonumoumumcnofinmmnmnaa z vuoa x a can .manwon .Homz m m.o .H& z ~.o nwcfimucoo musuxwe cofluommu 039 .oonumfi umumlmm on» an vaom owmeonumoumlmcapfluwmfimua< mo cofluommu coaumcfioofl H00flmmall.ma musmflm 102 32:55 as: _ t _ . .200 238960 :Nu0c_2..ma_dn.< .6 22853. 10.0 . rim 19c; ‘GOUDQJOSQV. 103 TABLE 5.--pH-rate Profile of the Iodination of Al-piperi- dine-Z-carboxylic acid by the pH-stat Method. ' 3 pH rate in A351 mu/sec x 10 7.15 1.76 7.4 2.13 7.625 2 6 7.90 3.19 8.25 3.76 104 where the last term represents catalysis by the neutral form of the Schiff base: \9 + Nice-c009 ‘cooe écoo- N N (‘9 H G) H By convention, the rate equation is usually expressed in terms of the total concentration of Schiff base, SO: + ' kH20 Ks(H ’ kOHKé Kw rate = S + S , + O , + 0 [1 + KS(H )1 [1 + KS(H )1 , + ksKs(H ) 2 + + 2 So [1 + K;(H >1 where K; is the dissociation constant of the protonated Schiff base and Kw the ion product of water. As discussed earlier for the enolization of pyru- vate under these conditions, it is possible that only one or two of the terms in the rate equation are significant. Assuming first that hydroxide-ion catalysis predominates, the data were plotted according to the rate law 105 l kOH Ks Kw kobs = + 1 + K;(H ) as shown in Figure 17. One of two results is expected. If hydroxide-ion K'K s w catalysis is predominant, the plot of + versus 1 + K;(H ) kobs should be linear and pass through the origin. If, on the other hand, other terms in the rate law are signifi- cant, the plot should not be linear. However, as seen in Figure 17, this plot appears linear and shows a definite positive intercept. At first this is somewhat puzzling since there is no readily appar- ent way that the complete rate equation could obey such a relationship. Normally, the intercept in such a plot corresponds to a pH-independent reaction; but all three terms in the rate law are pH-dependent. The answer to this dilemma lies in a rather subtle feature of the last term in the rate expression. A simple calculation shows that while this term is indeed pH—dependent, its value does not change greatly in a pH region about the pKa of the protonated Schiff base. K'K s w Therefore, its effect on a plot of k versus obs , + l + KS(H ) is not great enough to cause a measurable deviation from linearity; instead, this term appears to be pH—independent when measured against the hydroxide ion catalysis. 106 Figure l7.--Determination of the rate law for the iodination of A1-piperidine-2—carboxylic acid in the absence of buffer. All data were obtained by the pH-stat method described in the text, as shown in Figure 16 and Table 5. The data so obtained are plotted here according to the simplified rate law + — rate — kOH-(SH )(OH ) k», x 1033“" M" 107 108 The catalytic constant for this term, ks, can then be determined from a plot of rate/SO versus So at constant pH according to the expression: + k K'(H ) . . + £29: .... H20 S + kOHKssKw + (So)ksKs(H ’ ' + | + I + 2 o [1+KS(H )] [1+KS(H)] [1+KS(H)] This plot is shown in Figure 18. Once kS is determined, the other two constants can be determined from a plot of (rate)[l + Ké(H+)] k K'S 1 " versus ‘— ;L s s 0 so [n+1 [1 + Ké(H+)] H+ according to the following rearrangement of the rate expression as shown in Figure 19: I + I | _£-rate[1 + KS(H )] _ kSKSSo = kOHKSKw S + o [H J [1 + K;(H+)] H+ H20 5 where kOH can be determined from the slope and kH 0 from 3 i -1 the intercept. kOH was found to be 1.14 x 10 M. sec 5 -l and k less than 6 x 10- sec . H20 These results are quite different than those obtained from the pyruvate amine mixtures. This will be discussed in detail in the Discussion. For convenience, all the catalytic constants determined in this research are tabulated in Table 6. In a number of cases, constants determined from inter- cepts could not be distinguished from zero with 109 .Uomm um .mysumummfimu may cam no.5 um omcHMMCHmE mMB mm one .HE om mo mEdHo> Hmuou m as Humz z ~.o cam .NH .Hx z m.o .wumuumndm ou cofipflopm CH omcfimucoo musuxfle cofluommu was .coflmmmumxm mums on» as mfiumu oBu umuflm 030 MO 53m map ma umwoumusfl 039 m HA+EV.M + Ha m m 1+5 .x x >uflucmsv may ma mmoam one m m m mfix+mv.x + AH Hx+mv.s + H1 Hx+mv.x + H1 0m o m m + 3 m + m u mums mo mA+mv.x x M.M x A+mvmxo mx coflumsvm m mamum> om\0umu no uon m an mx unnumcoo 030 m0 coaumc H< mo coflumcwoow umumlmmln.ma musmflm mama mzu on mcfionooom .o IHEkumo .Uwom oaaaxonumolmlmcflcflummflml 110 to md 2 no. x cm N6 ..0 °s/a:oa 111 Figure 19.--Iodination of Al-piperidine-Z- carboxylic acid by the pH-stat method; determination of the constants kOH- and k0 from a plot of 0 + U _1 rate [1 + KS (H )J - kSKSSO versus —]—-' So (H+) [1 + KS(H+)] H+ according to the rate equation shown in the figure. The slope is given by kOH-KsKw and the intercept by kH OKé. The data for this plot have been presented in Figure 18 and Table 5. 112 K's 4 (I+K§[H+ )] " '0 f 0.‘ "3 [H rate (I+K '[H‘]) ’11—- q: Determination of "an and kH20 1'2 :14 I6 113 TABLE 6.--Complete Rate Constant Table. Substrate Catalyst Constant _ -6 . -l Pyruvate H O k - cyclopentanone > cyclobutanone. However, 119 the order of reactivity was reversed in the base-catalyzed route. Furthermore, the base-catalyzed enolization of the cyclic ketones was much faster than the base-catalyzed enolization of corresponding acyclic ketones. Schriesheim EE.§1° (48) observed a similar order of reactivity for the isomerization of cyclic olefins. Two explanations were offered by the two groups of investigators. Schechter 23 21. (49) suggested that two principles governed the reactivity of these compounds: (1) the transition state for an acid-catalyzed enolization most resembles the enol, whereas the transition state for a base-catalyzed reaction most resembles the ketone; (2) the acidity of a proton alpha to a carbonyl increases as the size of the ring de- creases. Both of these principles originated in the work of other investigators (50-52). The order of reactivity can then be explained as follows: In the acid-catalyzed route, the ring is forced to accommodate a substantial amount of double bond charac- ter; this becomes ever more difficult as the size of the ring decreases. Therefore, larger cyclic ketones should undergo enolization more readily than do smaller cyclic ketones. On the other hand, since the transition state for base-catalyzed enolization most resembles the ketone, the ring does not suffer a great increase in strain in the transition state. Therefore, the most important factor governing base-catalyzed enolization is the acidity of the alpha proton, which increases as the ring size decreases. 120 As a result, the base-catalyzed enolization of smaller rings should be more favorable than that of larger rings. An alternative explanation for the order of reactivity in base-catalyzed enolization was offered by Schriesheim gt 21. (48). These workers showed with molecular models that as the ring size decreases, the orientation of the carbon-hydrogen bond in relation to the plane of the carbonyl approaches the perpendicular. Since this perpendicular relation is known to be most favorable for enolization, the rate of enolization should increase with decreasing ring size. Whatever the explanation, the important experi- mental findings of these studies which pertain to the present research is that base-catalyzed enolization of a cyclic six-member ketone is significantly larger than the corresponding base-catalyzed enolization of an acyclic ketone by a factor of perhaps ten. This is in sharp contrast to results obtained in this study which indicate that the water-catalyzed enoli- zation of the protonated ketimine of pyruvate is 2 x 104 times greater than the water-catalyzed enolization of the protonated ketimine of Al-piperidine-2-carboxylic acid. Now that all these systems have been discussed in detail, it is appropriate to attempt to reach some con- clusions about the results of the present research. In 121 order to do this, the following dilemma must be considered: the system with pyruvate in amine buffer provides pre- cisely the ketimine which is to be studied, but the kinetics may be ambiguous. On the other hand, Al-piperidine- 2-carboxy1ic acid system provides a ketimine that is somewhat different than the one intended for study, but the kinetics are unambiguous. Which system, then, is a more reliable measure of the ketimine-enamine tautomerization of an acyclic ketimine of pyruvate? In the opinion of this author, the piperidine system is the more reliable. As discussed earlier, the only objection to this system is the possibility that the enolization of a cyclic ketimine might be greatly different than the enolization of an acyclic ketimine. If this explanation is used to account for the differences of the two systems, it would be necessary to conclude that the tautomerization of a cyclic six-membered ketimine is 2 x 104 less than that of an acyclic ketimine. This con- clusion could not differ more, in direction and magni- tude, from the results obtained with cyclic ketones, where the base-catalyzed enolization of cyclohexanone was found to be greater than that of the acyclic ketone by a factor of approximately ten. To make the comparison complete, we must consider differences in the rate at which methyl and methylene hydrogens (alpha to a carbonyl) undergo exchange. Rappe 122 and Sachs (53) showed that the exchange rate of methylene hydrogens is 0.6 to 0.23 that of methyl hydrogens. This effect should nearly compensate for the stimulation effect of the ring. Therefore, it seems most reasonable to conclude, tentatively at least, that the results obtained with the pyruvate and acetone systems are questionable, while those obtained with the Al-piperidine system are probably a good measure of the tautomerization of acyclic ketimines. Using Al-piperidine-2-carboxylic acid as a model for the aldolase reaction, the role of the Schiff base in the enzymatic process can now be discussed. These studies have established the contribution of Schiff base catalysis alone to the overall enzymatic rate; clearly, Schiff base formation alone can only account for a fraction of the catalysis, as shown in Table 7 which compares the enzymatic rate constant with the pseudo first order rate constants for pyruvate and Al-piperidine-Z- carboxylic acid at pH 8.0. The enolization of Al-piperidine- 2-carboxylic acid is approximately 103 times greater than pyruvate but 2.5 x 105 times less than the enzymatic enolization. However, the enolization of pyruvate and Al-piperidine-2-carboxylic acid is strongly catalyzed by general bases. Table 8 compares the enzymatic enolization with the imidazole-catalyzed enolization of pyruvate and protonated ketimine of Al-PCA. Once again, the enolization 123 TABLE 7.--Comparison of the Enzymatic Rate Constant of Enolization with the Corresponding Pseudo First Order Rate Constants for Pyruvate and A -PCA at pH 8.0. Rate Constant of compound Enolization* sec-1 Enzyme 120 Al-piperidine-Z-carboxylic acid 5 x 10‘.4 Pyruvate 3.6 x 10-7 *There is some difficulty comparing the rate con- stant of the enzymatic enolization to the rate constants for Al-PCA and pyruvate. For both of these nonenzymatic reactions, the rate constants have been expressed in terms of 1.0 M pyruvate or Al-PCA and 10"6 M hydroxide (pH 8.0). Therefore, the nonenzymatic rate constants are clearly pseudo first order constants. The order of the enzymatic process is not so clear, since it is not known whether water, hydroxide ion, or an amino acid residue catalyzes the tautomerization of the pyruvate-enzyme Schiff base. Therefore, the enzymatic rate constant could only be calculated from the rate of tautomerization of 1.0 M pyruvate-enzyme complex. Consequently, the enzymatic constant can only be expressed as a true first order con- stant, although the reaction may be of higher order. But as long as the physical significance of these constants has been stated, the comparisons can be made, and evalu- ated accordingly. 124 TABLE 8.--Comparison of the Enzymatic Enolization with Imidazole-Catalyzed Enolization of Pyruvate and Protonated Ketimine of Al—PCA. Compound Rate Constant -1 Enzyme 120 sec 1 . . . . . -1 -l A -piperid1ne-2-carboxylic aCld 0.41 sec M 3 -l -1 Pyruvate 0.253 x 10. sec M 125 of Al-piperidine-Z-carboxylic acid is approximately 2 x 103 times greater than pyruvate, but under these conditions it is only 3 x 102 times less than the enzymatic enolization. From this information, it is possible to con- struct a hypothetical model for the enzyme-catalyzed reaction by proposing that catalysis is effected by an amino acid side chain functioning as a general base to catalyze the enolization of the protonated ketimine of pyruvate. If the reasonable assumption is made that the base is held in close proximity to the methyl hydro- gens, the effective base concentration relative to the methyl protons could reasonably be assigned a value of ten (if not higher). So, if the imidazole group of histidine were the base, the rate constant for this enolization would be ten times the value of the rate constant determined with imidazole buffer, £33., 4.1 sec-l. This value is now one hundred times less than the enzyme- catalyzed rate. But if the base is stronger than imid— azole, such as the g-amino group of lysine, thiolate anion of cysteine, or phenolate anion of tyrosine, the rate constant would be as large as that observed for the enzyme-catalyzed reaction. This conclusion derives from a rough interpolation of the rate constants for the imidazole-catalyzed enolization and the hydroxide ion- catalyzed enolization. Of course, this model must remain Ill-1'11! Ii lliiolllilll‘ll" I. [if 126 speculative until it can be shown that an amino acid residue does function as a general base in the catalytic reaction. However, there is some circumstantial evidence for the involvement of a general base in aldolase reactions (aside from the usual plethora of "active site" studies involving the systematic destruction of every possible amino acid side chain). Muscle aldolase catalyzes the enolization of dihydroxyacetonephosphate by a Schiff base mechanism analogous to that of the KDPG aldolase. However, only one of the hydroxymethyl hydrogens is exchanged. If Schiff base formation alone provided sufficient activation of the enolization reaction, it would be difficult to explain the stereospecificity of the exchange reaction, since both hydroxymethyl hydrogens should undergo spon- taneous exchange. The simplest explanation for the stereoselectivity is that binding of the substrate places only one of these hydrogens in proximity with a general- base catalyst in the site. Moreover, it should not be overlooked that dihydroxyacetonephosphate has Eggr potentially exchangeable hydrogens when those at the carbon 1 position are included. Since these hydrogens do not exchange at all, it seems that Schiff base catalysis alone is not sufficient. Although this research has been concerned with Schiff base-catalyzed enolizations, it is worth discussing the problem of Schiff base formation itself. Both the I Jr. | .I-l Jill]. I 4' ‘1]!!! II II. '1! I! [If ['1' {.l ' 127 kinetic and equilibrium constants obtained for Schiff base formation show clearly that this part of the enzyme- catalyzed reaction must itself be enzyme-catalyzed. For example, the equilibrium constant of Schiff base for- mation from pyruvate and butylamine was found to be 10 M—l; by comparison, the binding constant of pyruvate to KDPG aldolase is 2 x 104 M-1 (determined from the Km). But this is not at all surprising. Most likely, pyruvate first forms an energetically favorable noncovalent complex with the enzyme prior to Schiff base formation, so that the overall equilibrium constant of Schiff base formation would be the product of the equilibrium constant for non- covalent binding and the intrinsic equilibrium constant of Schiff base formation. Other factors may play a role in enhancing enzymatic Schiff base formation, such as maintenance of an uncharged lysine at neutral pH. From the extensive investigations of Schiff base formation, even more factors might be imagined such as acid-catalyzed dehydration of the carbinolamine at the active site. From this very general description, it is obvious that the enzyme-catalyzed formation of Schiff base is every bit as mysterious as the enolization, if not more so. And, apparently, it would be even more difficult to construct a nonenzymatic model of this reaction than for the enolization reaction. However, the properties of Al-piperidine—Z— carboxylic acid may well have some bearing on Schiff base 128 formation as well as enolization. Consider the remarkable stability of this compound to hydrolysis, as illustrated in the following example. The equilibrium constant for Schiff base for- mation from pyruvate and butylamine is 10 M-l. Then at pH 7.6 where the concentration of protonated imine equals that of neutral imine we can write K = (ST) s mum; where ST is the total concentration of imine. In a solution containing 1.0 M_pyruvate and 1.0 M amine, the concentration of neutral amine will be approximately 10.3 M. From this, the ratio ST/P is readily calculated to be 2 x 10-2. In other words, the equilibrium ratio of Schiff base to pyruvate is only one to fifty. However, the equilibrium ratio of Al-piperidine—Z-carboxylic acid to the open chain form at this pH lies almost exclusively (within experimental measurement) in the direction of the Schiff base. Another simple calculation reflects the possible implications of this phenomenon. Let us assume, just to make the point, that the equilibrium constant for the ring closure can be calcu- lated by the same equation used for the pyruvate amine mixture. The results of this calculation are startling: 4 the value obtained for the equilibrium constant is l x 10 the value of the dissociation constant is the reciprocal, o I 129 l x 10-4, which is comparable to the Km of pyruvate for KDPG aldolase! The question to be answered is how this intra- molecular formation of Schiff base can be so much more favorable than the intermolecular reaction between pyru- vate and amine. The first explanation that comes to mind is that the "effective concentration" of reacting groups is much higher in the intramolecular reaction. However, a simple calculation shows that if the intrinsic equi- librium constant for Schiff base formation in the intra- molecular reaction is the same as that for the inter- 1), the “effective concentration" molecular reaction (10 MT of amine would have to be an impossible 1000 M’to account for the stability of the cyclic Schiff base at pH 7.6. This phenomenon is actually fairly common for intramolecular reactions. Since this subject has been discussed in detail by Jencks (54) and others referenced within, a detailed discussion of possible explanations for the behavior of intramolecular reactions will not be presented here. Suffice it to say that factors other than local concentration effects are involved. Three such factors which have been suggested are rotamer orien- tation, changes in salvation, and orbital overlap. Two of the more impressive studies of intramolecular reactions are worth mentioning. Bruice and Pandit (55, 56) found that the rate of nucleophilic attack of carboxylate on an ester increases dramatically as the groups are forced 130 together by the geometrical constraints of the molecule. Similarly, Storm and Koshland (57) found enormous acceler- ations in the rate of lactone formation when the reacting groups were forced into close proximity with the prOper orientation. Whatever the explanation for the rate accelerations, both studies show clearly that maximal rate accelerations only occur when severe limitations are imposed on the number of possible orientations the reacting groups can assume in relation to one another. Presumably, then, this precise orientation is effected in an enzyme catalyzed reaction. But the stability of Al-piperidine-Z-carboxylic acid indicates that this may not always be necessary. The only apparent restriction of the relative positions of the reacting groups in the Open chain form is that they cannot be separated by more than about ten angstroms. Therefore, the highly favorable ring closure reaction must be ex- plained in other ways. This may have important consequences regarding not only enzymatic Schiff base formation, but also substrate binding and enzyme catalysis in general. Specifically, substrate binding and catalysis may be effected by a more subtle mechanism than directly jamming the reacting or binding groups together in a unique orientation. Perhaps the initial binding of substrate is weak, but the presence of substrate causes the entire active site to change to an 131 energetically more favorable configuration that includes the substrate in much the same way that the rather distant amino and carbonyl groups of the open chain form of Al-piperidine-2-carboxylic acid are combined to form the more stable Schiff base. This view, of course, is essentially the Koshland induced—fit model. It seems, then, that Al-piperidine-2-carboxylic acid may well be a good model for many enzymatic processes. Finally, the possible role of the Schiff base in substrate binding should be mentioned. Throughout this work the catalytic role of the Schiff base has been emphasized, but it is difficult to overlook the possibility that the enzyme might use this covalent bond to assist the binding of pyruvate. Moreover, the covalent bond is an ideal way to bind intermediates. Consider the following scheme for the enzymatic aldolization of pyruvate and GBP: --H+ _ +G3P E + P :— E-PiEP ‘1‘ EP —> product +H+ noncovalent Schiff enol(ate) interaction base Since this is an ordered mechanism, the enzyme must efficiently bind the enol (enolate) form of pyruvate as well as pyruvate itself; otherwise, the enol would readily dissociate from the enzyme. In order to 132 effectively bind pyruvate in both keto and enol forms, the enzyme might have to undergo a conformational change upon enolization. However, this is not necessary with a Schiff base mechanism since the enol form of pyruvate is a covalently bound enamine which cannot dissociate directly from the enzyme. This argument, of course, may be teleology at its worst, but still it is an attractive possibility. 10. 11. Koshland, Ingram, J. REFERENCES D. E., Jr., and Neet, Biochem., 31, 359 (1968). M., and Wood, W. A. 240, 4146 (1965). Rose, I. A., and Biophys., 118, 758 (1965). and O'Connell, E. L. Macholan, L., and Svatek, E. Comm., 3;, 2564 (1960). Bruice, T. C., Mechanisms. 1966, p. Snell, E. E., Fasella, P. M., Fanelli, A. of Pyridoxal Catalysis. 181. Rress,1963. Bruice, T. C., Mechanisms. 1966. Jencks, W. P. and R. W. Taft (eds.). R. and Benkovic, S. K. E. Ann. Rev. J. Biol. Chem., Arch. Biochem. Coll. Czech. Chem. In Bioorganic New York: W. A. Benjamin, Inc., and Benkovic, S. Organic Chemistry, Vol. 2. science Publishers, Inc., 1964, p. Acree, S. F., and Johnson, J. M. 221 308 (1907). Barret, E. I and Lapworth, A. 85 (1908). Bodforss, S. Z. Physik. Chem., 133 Braunstein, A., and Chemical and Biologigal Aspects New York: Pergamon In Bioorganic New York: W. A. Benjamin, Inc., In S. G. Cohen, A. Streitweish, Jr., Progress in Physical New York: Inter- 1 9, Am. 63. Chem. J., J. Chem. Soc., 22, 223 (1924). 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 134 Olander, A. Z. Physik. Chem., 129, l (1927). Bartlett, P. D., and Conant, J. B. J. Am. Chem. Soc., 2881 (1932). Jencks, W. P. J. Am. Chem. Soc., 8;, 475 (1959). Westheimer, F. H., and Cohen, H. J. Am. Chem. Soc., ‘69, 90 (1938). French, C. C. J. Am. Chem. Soc., 51, 3215 (1929). Miller, J. G., and Kilpatrick, M. J. Am. Chem. Soc., 53, 3217 (1951). Speck, J. C., Jr., and Forist, A. A. J. Am. Chem. Soc., 12, 4659 (1957). Pontremoli, S. Proc. Natl. Acad. Sci., 41, 1942 (1961). Grazi, E., Cheng, T., and Horecker, B. L. Bioch. and Biophys. Res. Comm., 1, 250 (1962). Speck, J. C., Jr. J. Am. Chem. Soc., 85, 1012 (1963). Rose, I. A., and O'Connell, E. L. Arch. Bioch. Biophys., 118, 758 (1967). Cash, D. J., and Wilson, I. B. J. Biol. Chem., 241, 2490 (1966). Pedersen, K. J. J. Am. Chem. Soc., 60, 595 (1938). Pedersen, K. J. Acta. Chem. Scand., 8, 710 (1954). Hay, R. W. Aust. J. Chem., lg, 337-51 (1966). Guthrie, J. P., and Westheimer, F. H. Fed. Proc., 36, 562 (1967). Jencks, W. P. In Catalysis in Chemistry and Enzymology. New York: McGraw Hill, 1969, p. l . Hamilton, G. H., and Westheimer, F. H. J. Am. Chem. Soc., 21, 6332 (1959). Fridovich, I., and Westheimer, F. H. J. Am. Chem. Soc., 84, 3209 (1962). Warren, S., Zerner, B., and Westheimer, F. H. Biochemistry, 5, 817 (1966). 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 135 Kobes, R. D., and Dekker, E. E. Bioch. and Biophys. Res. Comm., 51, 607 (1967). Ingram, J. M., and Wood, W. A. J. Biol. Chem., 241, 3256 (1966). Bender, M. L., and Williams, A. J. Am. Chem. Soc., 55, 2502 (1966). Hine, J., Menon, B. C., Jenson, J. H., and Mulders, J. J. Am. Chem. Soc., 55, 3367 (1966). Rose, I. A., and Rieder, S. V. J. Am. Chem. Soc., '11, 5764 (1955). Rose, I. A., and Rieder, S. V. J. Biol. Chem., 231, 315 (1958). Tagaki, W., and Westheimer, F. H. Biochemistry, 1, 901 (1968). Kobes, R. D. Thesis. University of Michigan, 1967. Meister, A. J. Biol. Chem., 206, 577 (1954). Weisblat, D. I., Magerlein, B. J., and Meyers, D. R. J. Am. Chem. Soc., 15, 3630 (1953). Harper, E. T., and Bender, M. L. J. Am. Chem. Soc., 51, 5625 (1965). Coward, J. K., and Bruice, T. C. J. Am. Chem. Soc., 55, 5339 (1969). Hall, N. F., and Sprinkle, M. R. J. Am. Chem. Soc., 55, 3469 (1932). Zuman, P. Coll. Czech, Chem. Comm., 15, 839 (1950). Meloche, H. P., and Wood, W. A. J. Biol. Chem., 339, 3515 (1964). Hine, J., Menon, B. C., Jenson, J. H., and Mulders, J. J. Am. Chem. Soc., 55, 3367 (1966). Schriesheim, A., Muller, R. J., and Rowe, C. A., Jr. J. Am. Chem. Soc., 55, 3164 (1962). Shechter, H., Collis, M. J., Dessy, R., Okuzumi, Y., and Chen, A. J. Am. Chem. Soc., 55, 2905 (1967). 50. 51. 52. 53. 54. 55. 56. 57. 58. 59. 136 Swain, C. G., Stivers, E. C., Reuwer, J. F., Jr., and Schaad, L. J. J. Am. Chem. Soc., 55, 5855 (1958). Tamres, M., and Searles, S. J. Am. Chem. Soc., 55, 2100 (1959). Campbell, H. J., and Edwards, J. T. Can. J. Chem., 55, 2109 (1960). Rappe, C., and Sachs, W. H. J. Org. Chem., 55, 4127 (1967). Jencks, W. P. In Catalysis in Chemistry and Enzymology. New York: McGraw Hill, 1969. Bruice, T. C., and Pandit, U. K. J. Am. Chem. Soc., £3: 5328 (1960). Bruice, T. C., and Pandit, U. K. Proc. Natl. Acad. Sci., 55, 402 (1960). Storm, D. R., and Koshland, D. E., Jr. Proc. Natl. Acad. Sci., 55, 445 (1970). Barran, L. R. Thesis. Michigan State University, 1969. Hammerstedt, R. H., M6h1er, H., Decker, K. A., and Wood, W. A. Manuscript received for publication. T ”ilifiliiiiflifliifiiiiijiflifiiujiifliilifi