The Mechanism of Action of Adenosylcobalamin-Dependent Diol Dehydratase. I. Glycerol and other Substrate Analogs as Substrates and Inactivators of Diol Dehydratase: Kinetics, Stereospecificity, and Mechanism. II. Hydrogen Transfer in the Inactivation of Diol Dehydratase by Glycerol. III. Hydrogen Transfer in Catalysis by Diol Dehydratase Citation Moore, Kevin William (1979) The Mechanism of Action of Adenosylcobalamin-Dependent Diol Dehydratase. I. Glycerol and other Substrate Analogs as Substrates and Inactivators of Diol Dehydratase: Kinetics, Stereospecificity, and Mechanism. II. Hydrogen Transfer in the Inactivation of Diol Dehydratase by Glycerol. III. Hydrogen Transfer in Catalysis by Diol Dehydratase. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/efx4-ad25. https://resolver.caltech.edu/CaltechTHESIS:03162026-223828895 Abstract Chapter I A number of vicinal diols were found to react with propanediol dehydratase, typically resulting in the conversion of enzyme-bound adenosylcobalamin to cob(II)alamin and formation of aldehyde or ketone derived from substrate. Moreover, most are capable of effecting the irreversible inactivation of the enzyme. The kinetics and mechanism of product formation and inactivation were investigated. Glycerol, found to be a very good substrate for diol dehydratase as well as a potent inactivator, atypically, did not induce cob(II)alamin formation to any detectable extent. With glycerol, the inactivation process was accompanied by conversion of enzyme-bound adenosylcobalamin to an alkyl or thiol cobalamin, probably by substitution of an amino acid side chain near the active site for the 5'-deoxy-5'-adenosyl ligand on the cobalamin. The inactivation reaction with glycerol as the inactivator exhibits a deuterium isotope effect of 14, strongly implicating hydrogen transfer as an important step in the mechanism of inactivation. The isotope effect on the rate of product formation was found to be 8. Experiments with isotopically substituted glycerols indicate that diol dehydratase distinguishes between "R" and "S" binding confonnations, the enzyme-(R)-glycerol complex being predominantly responsible for the product-forming reaction, while the enzyme-(S)- glycerol complex results primarily in the inactivation reaction. A comparative study of the reaction of IIieso- and dl-2,3-butanediols with diol dehydratase shows that the stereospecificity of hydrogen abstraction during inactivation is the same as that of catalysis, suggests that hydrogen abstraction from C-1 of substrate may be concerted with cleavage of the carbon-cobalt bond, and further suggests that formation of a carbon-cobalt bond between coenzyme and substrate is not obligatory for catalysis. Mechanistic implications are discussed. A method for removing enzyme-bound hydroxycobalamin that is nondestructive to the enzyme and a technique for measuring the binding constants of (R)- and (S)-1,2-propanediols are presented. Chapter II We have investigated the kinetic characteristics of the inactivation of the adenosylcobalamin-dependent enzyme propanediol dehydratase by glycerol, (RS)-1,1-dideuterioglycerol, (R)-1,1-dideuterioglycerol, (S)-1,1-dideuterioglycerol, and perdeuterioglycerol in the absence of substrate and in the presence of 1,2-propanediol and 1,1-dideuteriopropanediol. Substitution of deuterium for hydrogen at C-1 of 1,2-propanediol or at the pro-(R) hydroxymethyl group of glycerol has no effect on the apparent dissociation constants K I and K G , respectively, of the holo-enzyme- substrate complexes. In contrast, the same substitution at the pro-(S) hydroxymethyl group of glycerol increases the value of the apparent K G . The results suggest that the binding of only the pro-(S) conformation is perturbed by this latter substitution. In the presence of a competL1g substrate, this K G is observed to vary as a function of glycerol concentration in a manner indicating that kinetic terms involving steps of the reaction subsequent to binding of substrate contribute to determining its value. The inactivation kinetics obtained with these glycerols in the presence of 1,2-propanediol and 1,1-dideuterio-1,2-propanediol imply that hydrogen (or deuterium) attached to C-1 of 1,2-propanediol participates in the inactivation process and contributes to the expression of a kinetic isotope effect on the rate of inactivation. The mechanism for this inactivation must involve the cofactor as an intermediate hydrogen carrier, presumably in the form of 5'-deoxyadenosine. Moreover, a mechanism involving a rate determining transfer of hydrogen from an intermediate containing three equivalent hydrogens accounts for the results. When diol dehydratase holoenzyme is inactivated by [1- 3 H]-glycerol, 5'-deoxyadenosine which is enriched in tritium by a factor of 2.1 over that in glycerol can be isolated from the reaction mixture. Chapter III Studies (Chapter II; Bachovchin et al. (1978) Biochemistry 17, 2218) of the mechanism of inactivation of adenosylcobalamin-dependent dial dehydratase have led to the development of a general method to describe the kinetics of a reaction pathway containing a reservoir of mobile hydrogen. Analysis by this method of catalytic rate measurements for mixtures of 1,2-propanediol and 1,1-dideuteriol, 2-propanediol supports a mechanism involving an intermediate with three equivalent hydrogens, in which hydrogen transfer from this intermediate to product is the major rate-contributing step. The results of this analysis predict that diol dehydratase should exhibit a tritium isotope effect of 4.5, a prediction which was verified experimentally. Other results using tritium as a trace label (Essenberg et al. (1971) J. Am. Chem. Soc. 93, 1242) are discussed in light of these deuterium isotope studies. Results of analogous experiments employing mixtures of glycerol and perdeuterioglycerol as substrates suggest that during catalysis with this substrate, neither hydrogen abstraction from C-1 of glycerol nor from the intermediate is strictly rate determining. Item Type: Thesis (Dissertation (Ph.D.)) Subject Keywords: (Chemistry) Degree Grantor: California Institute of Technology Division: Chemistry and Chemical Engineering Major Option: Chemistry Thesis Availability: Public (worldwide access) Research Advisor(s): Richards, John H. Thesis Committee: Unknown, Unknown Defense Date: 30 April 1979 Funders: Funding Agency Grant Number NIH UNSPECIFIED California Institute of Technology UNSPECIFIED Record Number: CaltechTHESIS:03162026-223828895 Persistent URL: https://resolver.caltech.edu/CaltechTHESIS:03162026-223828895 DOI: 10.7907/efx4-ad25 Default Usage Policy: No commercial reproduction, distribution, display or performance rights in this work are provided. ID Code: 18428 Collection: CaltechTHESIS Deposited By: Benjamin Perez Deposited On: 18 Mar 2026 17:36 Last Modified: 18 Mar 2026 17:52 Thesis Files PDF - Final Version See Usage Policy. 75MB Repository Staff Only: item control page

THE MECHANISM OF ACTION OF ADENOSYLCOBALAMINDEPENDENT DIOL DEHYDRATASE I. GLYCEROL AND OTHER SUBSTRATE ANALOGS AS SUBSTRATES AND INACTIVATORS OF DIOL DEHYDRATASE: KINETICS, STEREOSPECIFICITY, AND MECHANISM. II. HYDROGEN TRANSFER IN THE INACTIVATION OF DIOL DEHYDRATASE BY GLYCEROL. III. HYDROGEN TRANSFER IN CATALYSIS BY DIOL DEHYDRATASE. Thesis by Kevin William Moore In Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy California Institute of Technology Pasadena, California 1979 (Submitted April 30, 1979) ii To 'Toots' iii ACKNOWLEDGMENTS I am grateful to Professor John H. Richards for his advice, a ,1d financial and moral support. two rare but highly appreciated qualities: He possesses he doesn't mind being pestered by students , yet rarely pesters them in return. Pr •.:>fessor David A. Evans went out of his way to be helpfu :_ to me in the synthesis of (S) - 1, l - dideuterioglyce1.ol (Chapter I), for which I am most appreciative. -~ am also grateful to Drs. R. G. Eagar, Jr. and W. W. Bachovchin for their congenial collaboration on the work described in this thesis. The members of the Richards' gro1p have made Caltech an exciting, very pleasant, rewa: ding and amusing place to work. I also thank Mrs. DeJorah Chester for typing large parts of this thesis. Th ose are the parts with no typographical errors or omitted phrases. I acknowledge the National Institutes of Health (Predoctoral Traineeship, 1974-1978) and the Calif0rnia Institute of Technology (supplemental Graduate Teaching and Research Assistantships, 1974-1978; Graduate Research Assistantship, 1978-1979) for generous support. Most important, E. Moore, family, my par en ts, Harry L. and Aileen and my fantastic wife, Hsiao-ping Hsu have been a constant source of love, attention, and good iv humor. This must also be said of my best friends and in- laws, Shi-ping and Jin-Chen Hsu . speakably grateful~ To them all I am un- V TABLE OF CONTENTS Page Introduction 1 Background Adenosylcobalamin-dependent enzymes Hydrogen transfer Cleavage of the carbon-cobalt bond Migration of group X References 1 5 12 17 19 23 Chapter I Abstract xiii Introduction 29 Experimental 30 Enzyme preparation Assays Determination of Inactivation Rates Determination of 8-Hydroxypropionaldehyde [ 1- 1 '+c}-Glycerol Labeling of Diol Dehydratase Subs trttes [1- 4 C]-Glycerol (R)-1,1-Dideuterioglycerol (RS)-1,1-Dideuterioglycerol (S)-1,1-Dideuterioglycerol (RS)-1,1-Dideuterio-l,2-propanediol 3,3,3-Trifluoro-l,2-propanediol Isobutylene glycol 1,2-Butanediol 3-Chloro-1,2-propanediol Thioglycerol • meso-2, 3-Butanediol 'a-'2,3-Butanediol al-2,3-Butanediol Results 30 31 31 33 34 35 35 35 36 38 38 38 38 39 39 39 39 39 41 Kinetics of Glycerol Inactivation in the Absence 41 of 1,2-Propanediol Kinetics of Glycerol Inactivation in the Presence 45 of 1,2-Propanediol 51 Kinetics of Inactivation - Other Substrates Deuterium Isotope Effects on the Rate of 52 Inactivation by Glycerol 56 Product Formation from Glycerol vi Turnover Number Rate of Formation of s-Hydroxypropionaldehyde Deuterium Distribution in Products Product Formation - Other Substrates Spectral Observation of the Holoenzyme Glycerol Complex Spectral Observations of Enzyme-Bound OH-Cb1 Spectral Observation - Other Substrates Recovery of Active Apoenzyme from Inactive Apoenzyme - OH-Cbl Complexes Reactivation of 02 - Inactivated Holoenzyme Reactivation of Glycerol - Inactivated Holoenzyme Inactivation by [1- 14 C]-Glycerol • SDS-Gel Electrophoresis Discussion (i) 61 65 68 72 72 75 75 77 77 Mechanism of Inactivation and Catalysis Kinetics of Catalysis and Inactivation Kinetic Isotope Effect Spectral Observation of Enzyme-Bound Cofactor Stereospecificity of Catalysis and Inactivation (ii) Page 56 57 61 61 Glycerol 2,3-Butanediols Substrate Specificity Relation of Diol Dehydratase to Glycerol Dehydratase 77 78 79 79 84 84 90 93 95 97 References Chapter II Abstract xv Introduction 102 Experimental 103 Enzyme Preparations S'-Deoxyadenosine Substrates Determination of Inactivation Rates Determination of Michaelis Constants KG Inactivation by [l- 3 H]-Glycerol 103 104 104 104 105 105 vii Results 108 Kinetics of Inactivation by Perdeuterioglycerol 108 Kinetics of Inactivation by (S)-, (RS)-, and (R)-1,1-Dideuterioglycerols 109 Kinetics of Glycerol Inactivation. Competition with (RS)-1,1-Dideuterio-l,2-Propanediol 113 Kinetics of Glycerol Inactivation. Competition with (RS)-1,2-Propanediol 124 Isolation and Identification of 5'-Deoxyadenosine 128 Discussion 131 Variation of KG Variation of ki Summary and Conclusions References 131 135 146 149 • Chapter III Abstract xvii Introduction 152 Experimental 153 Propanediol Dehydratase Assays Tritium Isotope Effect Substrate (RS)-1,1,2-Trideuterio-l,2-propanediol (RS)-Perdeuterio-1,2-propanediol (RS)-[1- 14 C]-l,2-Propanediol (RS)-[1- 3H]-l,2-Propanediol Perdeuterioglycerol Computing Results Derivation of Rate Expression Glycerol as Substrate Tritium Isotope Effect 153 153 154 155 155 155 155 155 155 156 156 161 169 174 viii 177 Discussion Models for Hydrogen Transfer Rate Determining Step Glycerol as Substrate Relation to Microscopic Rate Constants Relation to Previous Studies Conclusions 178 180 182 184 185 195 197 References Appendix Growth of Klebsiella pneumon:iae and Isolation of Propanediol Dehydratase 201 The Culture The Harvest The Cells The Sonicate The Assay Enzyme Purification 202 205 206 206 207 208 Suggested Further Studies 216 Propositions Abstracts xviii Proposition I 219 Proposition II 236 Proposition III 243 Proposition IV 249 Proposition V 255 ix LIST OF TABLES Chapter Table Title Pa ge I I Kinetic Parameters of Substrates , Inhibitors , and Inactivators of Diol Dehydratase. 53 kp and ki fo r Glycerol and its D~uterated Analogs. 55 Reactivation Data . 73 I Chromatography of 5'-Deoxy adenosine. 107 II Limiting Values of ki and Isotope Effects. 1 39 I Values of Rate Constants Obtained with f Constant. 168 II Values of Rate Constants Obtained with f Allowed to Vary. 170 II III II III III Variation ,of Least Squares Criterion as a Function of n and f with Unvarying Rate Constants . 171 IV Limiting Values of Overall Isotope Effects for a Two-Step Reaction. 18 6 X LIST OF FIGURES Chapter Figure Introduction 1 The Structure of Cobalamin 2 2 Adenosylcobalamin-Dependent Reactions 7 Mechanism of AdenosylcobalaminDependent Rearrangements. 10 Possible Mechanisms for Migration of Group X. 13 Inactivation Rate Constant k. b as a Function of Glycerol Cofi~o s centration. 43 Double-Reciprocal Plot of ki ,o b s . as a Function o f Glycerol Concentration in the Presence of (RS)1,2-Propanediol. 46 3 Temperature Dependence of k.1,0b S 48 4 Production of e-Hydroxypropionaldehyde as a Function of Time. 58 The Effects of Glycerol and 1,2Propanediol on the UV-Visible Spectrum of Enzyme-Bound Adenosylcobalamin. 63 Absorption Spectra of OH-Cbl in Solution and Bound to Diol Dehydratase. 66 The Effect of Several Substrate Analogs on the Absorption Spectrum of Enzyme-Bound Adenosylcobalamin. 69 Absorption Spectra of Holoenzyme~ 2,3-Butanediol Complexes. 69 Double-Reciprocal Plot of ki obs as a Function of (S)-1,1' Dideuterioglycerol Concentration. 111 3 4. I 1 2 5 6 7a 7b II 1 Title Page xi 2 Double-Reciprocal Plot of ki b as a Function of (RS)-1,1- ,o s Dideuterioglycerol Concentration in the Presence of dl-2,3-Butanediol 114 3. Double-Reciprocal Plot of ~i obs as a Function of Perdeuterioglycerol Concentration in the Presence of (RS)-l,l-Dideuterio-1,2-Propanediol. 4 Double-Reciprocal Plot of ki,obs as a Function of Inactivator Concentration in the Presence of (RS)-1,1Dideuterio-1,2-Propanediol. 119 5 Observed Isotope Effect as a Function of the Reciprocal of Inactivator Concentration in the Presence of (RS)l,l-Dideuterio-1,2-Propanediol. 122 6 Observed Isotope Effect as a Funetion of the Reciprocal of Inactivator Concentration in the Presence of (RS)-1,2-Propanediol. 126 Radioactivity Associated with 5'Deoxyadenosine as a Function of the Amount of Apoenzyme Inactivated by Radioactive Glycerols . 129 Variation of k t 0 bs with Mole Fraction of ca • Deuterium in 1,2-Propanediol. 157 Temperature Dependence of k t b for 1,2-Propanediol and l,l~a ,o s Dideuterio-1,2-propanediol. 159 Reaction Scheme for a Hypothetical Mechanism in which the Intermediate Carries Two Equivalent Hydro gens. 162 Variation of k t 0 b of Glycerol as a Function ca • s of the Mole Fraction of Perdeuterioglycerol in the Substrate . 172 7 III 117 1 2 3 4 xii 5 Appendix 1 Relative 3tt/14c dpmRatio of Residual 1,2-Propanediol as a Function of the Amount of Substrate Consumed. 175 Elution Profile of Diol Dehydratase During Cellex D Chromatog raphy in Buffer G. 214 xiii Chapter I ABSTRACT A number of vicinal diols were found to react with propanediol dehydratase, typically resulting in the conversion of enzyme-bound adenosylcobalamin to cob(II)alamin and formation of aldehyde or ketone derived from substrate. Moreover, most are capable of effecting the irreversible inactivation of the enzyme. The kinetics and mechanism of product formation and inactivation we re investigated. Glycerol, found to be a very good substrate for diol dehydratase as well as a potent inactivator, atypically , did not induce cob(II)alamin formation to any detectable extent. With glycerol, the inactivation process was ac - companied by conversion of enzyme-bound adenosylcobal amin to an alkyl or thiol cobalamin, probably by substitution of an amino acid side chain near the a ctive site for the 5' -deo·.x y-5 '-adenosyl ligand on the cobalamin. The i nacti - vation reaction with glycerol as the inactivator exhib its a deuterium isotope effect of 14, strongly implicating hydrogen transfer as an important step in the mechani sm of inactivation. The isotope effect on the rate of product formation was found to be 8. Experiments with isotopically substituted glycerols indicate that diol dehydratase dis tinguishes between "R" and "S" binding confonnations, the enzyme-(R)-glycerol complex being predominantly res pons ibl for the product-forming reaction, while the enzyme-(S)- 0 ~ xiv glycerol complex results primarily in the inactivation reaction. A comparative study of the reaction of IIieso- and dl-2,3-butanediols with diol dehydratase shows that the stereospecificity of hydrogen abstraction during inactiva·· tion is the same as that of catalysis, suggests that hydro. gen abstraction from C-1 of substrate may be concerted with cleavage of the carbon-cobalt bond, and further suggests that formation of a carbon-cobalt bond between coenzyme and substrate is not obligatory for catalysis. implications are discussed. Mechanistic A method for removing enzyme-bound hydroxycobalamin that is nondestructive to the enzyme and a technique for measuring the binding constants of (R)- and (S)-1,2-propanediols are presented. xv Chapter II ABSTRACT We have investigated the kinetic characteristics of the inactivation of the adenosylcobalamin-dependent enzym1:~ propanediol dehydratase by glycerol, (RS)-1,1-dideuterioglycerol, (R)-1,1-dideuterioglycerol, (S)-1,1-dideuterioglycerol, and perdeuterioglycerol in the absence of substrat.e and in the presence of 1, 2-propanediol and 1, 1-dideuteriopropanediol. Subs ti tut ion of deuterium for hydro-· gen at C-1 of 1,2-propanediol or at the pro-(R) hydroxymethyl group of glycerol has no effect on the apparent dissociation constants KI and KG, respectively, of the holoenzyme-substrate complexes. In contrast, the same substi- tution at the pro-(S) hydroxymethyl group of glycerol increases the value of the apparent KG. The results suggest that the binding of only the pro-(S) conformation is pertur bed by this latter substitution. In the presence of a competL1g substrate, this KG is observed to vary as a function of glycerol concentration in a manner indicating that kinetic terms involving steps of the reaction subsequent to binding of substrate contribute to determining its value. The in- activation kinetics obtained with these glycerols in the presen::!e of 1,2-propanediol and l,l - dideuterio-1,2-pro panediol imply that hydrogen (or deuterium) attached to C-1 of 1,2-propanediol participates in the inactivation xvi process and contributes to the expression of a kinetic isotope effect on the rate of inactivation. The mechanism for this inactivation must involve the cofactor as an intermediate hydrogen carrier, presumably in the form of 5' deoxyadenosine. Moreover, a mechanism involving a rate- determining transfer of hydrogen from an intermediate containing three equivalent hydrogens accounts for the re sults. When diol dehydratase holoenzyme is inactivated by [1-3H]-glycerol, 5'-deoxyadenosine which is enriched in tritium by a factor of 2.1 over that i n glycerol can be isolated from the reaction mixture. xvii Chapter III ABSTRACT Studies (Chapter II; Bachovchin et al . (1978) Biochemistry 17, 2218) of the mechanism of inactivation of adenosylcobalamin-dependent dial dehydratase have led to the development of a general method to describe the kinetics of a reaction pathway containing a reservoir of mobile hydrogen. Analysis by this method of catalytic rate mea- surements for mixtures of 1,2-propanediol and 1,1-dideuteriol,2-propanediol supports a mechanism involving an intermediate with three equivalent hydrogens, in which hydrogen transfer from this intermediate to product is the major rate-contributing step. The results of this analysis pre- dict that diol dehydratase should exhibit a tritium isotop e effect of 4.5, a prediction which was verified experimental ly. Other results using tritium as a trace label (Essenberg et al. (1971) J. Am. Chem. Soc. 93, 1242) are discussed in light of these deuterium isotope studies. Results of analogous experiments employing mixtures of glycerol and perdeuterioglycerol as substrates suggest that during catalysis with this substrate, neither hydrogen abstractioLt from C-1 of glycerol nor from the intermediate is strictly rate determining. xviii ABSTRACTS OF PROPOSITIONS Proposition I In light of the studies reported in the body of this thesis, further investigations on the mechanism of adenosyl cobalamin-dependent diol dehydratase are proposed. Proposition II Metastasis of cancer cells can be inhibited by intro duction into the cancer host of microorganisms such as Bacillus Calmette-Guerin and Corynebacterium parvum. Experiments are proposed to study the possible interactions of these agents with the irmnune system in inhibiting metastasis. Proposition III .Complement fixation via the classical pathway is initiated by interaction of the complement protein Clq with the heavy chain constant (Si) region of immunoglobulin M and certain classes of immunoglobulin G. Certain CH amino acids, in particular tryptophan and tyrosine, have been implicated in this interaction. A study of the pH-depen- dence of complement fixation by irmnunoglobmlins , and of inhib.ition of complement fixation by peptides, containing 3-nitrotyrosine is proposed. xix Proposition IV A recombinant DNA study of the structure of genes coding for irrrrnunoglobulin heavy chains is proposed. In particular, sequence analysis of the "introns", untranslated segments of DNA which separate coding regions, may prove enlightening in relation to antibody gene evolution. ~roposition V Release into the environment of large amounts of synthetic, nonbiological organic compounds is a major concern. The use of bacterial mutagenesis techniques for creation of mutant microorganisms with enzymes capable of converting these compounds into biodegradable products is proposed. 1 INTRODUCTION Background The red, crystalline substance known as vitamin B12 was purified in 1948 at the Merck (Rickes et al., 1948) and Glaxo Laboratories (Smith & Parker, 1948) after a twentyyear search for the anti-pernicious anemia factor present in liver. While this success ended a protracted period of intensive effort, it also marked the birth of a singularly fascinating area of research, which has made important con -tributions to both the biological and chemical sciences. The structure of vitamin B12 was announced by Hodgkin et al. (1955) (Fig. 1). The molecule contains a cobalt atom (co3+, low spin d 6 ) coordinated to the four nitrogen atoms of corrin, a tetrapyrrole macrocycle. Corrin is superficially similar to porphyrin, but lacks the latter's completely conjugated TI -electron system and strictly plana x.· configuration, and has a methine instead of the usual methylene bridge between two of the rings ("A" and "D"). The lower (a) ligand site is occupied by an unusual nucleo ,. tide, 5,6-dimethylbenzimidazole, while the 8-ligand site may be occupied by a ligand such as those shown. Of the a-ligands indicated, all but cyanide occur naturally ; the latter is introduced as an artifact of the purification procedure . The species with the nucleoside (adenosyl- cobalamin, AdoCbl, coenzyme B12 ) and the methyl group (methylcobalamin, MeCbl) as a-ligands are unique in their 2 Figure 1 (a) Derivatives of cobalamin. R = CH 2 CONH 2 , R' = CH 2CH2CONH2 X = OH CN (hydroxycobalamin , OH - Cbl) (cyanocobalamin, CN-Cbl) CH 3 (methylcobalamin, Me-Cbl) S'-deoxy-S'adenosyl (adenosylcobalamin, AdoCbl) (b) Corrin. roline rings. Letters A-D identify the individual pyr- · The numbering system corresponds to that of the porphyrin ring; hence the number 20 has been omitted (Hogenkamp, 1975). 3 (a) (b) Figure 1 4 being the only known biologically occurring compounds (besides the alkyl mercurys) with a carbon-metal bond. The structure of adenosylcobalamin was announced by Lenhert & Hodgkin (1961). The structural complexity of the cobalamins is mirrored in the roles which they play in biology. In man and other mammals, the importance of B12 is evinced by the symptoms which occur in its absence. The most corrnnon of these are pernicious (megaloblastic) anemia, and neurological abnormalities associated with the appearance of lesions on and the demyelination of the spinal cord. As little as 1 µg/day of cobalamin is required by man to avert these symptoms. While the precise metabolic role played by co- balamins in ameliorating these conditions remains unclear, evidence has been obtained implicating their direct or indirect involvement in the biosynthesis of deoxyribonucleotides and metabolism of fatty acids (Beck, 1975; Kishimoto , et al., 1973; Cardinale et al., 1970; Barley et al., 1972; Green et al., 1975). In addition, the methyl derivative acts as a methyl transfer agent in the conversion of homocysteine to methionine, and may also participate in nucleotide biosynthesis. Mammals do not synthesize cobalamin, and have developed an elaborate absorption and transport system to extract it from their diet (Ellenbogen , 1975) . Cobalamin which has been freed by the digestive process is bound to a protein 5 termed intrinsic factor. The intrinsic factor-cobalamin complex then binds to the microvilU. of the distal ileum in a process which requires ca2+ and pH above 6 . 4 . Once the complex is bound, cobalamin is released into the cell in an energy-requiring step, from which it enters the bloGd and binds to globulins called transcobalamins , for trans port and storage. All cobalamin is synthesized exclusively by micro organisms, which also use this molecule to pe r form a variety of chemical reactions. These reactions encompass such diverse processes as isomerizations and synthesis of deoxy ribonucleotides. The involvement of cobalamin in the fun cia - mental process of nucleic acid biosynthesis bespeaks an evolutionary antiquity which constitutes an apparently marked contrast to its structural complexity . Clearly, the evolutionary history of cobalamin, if and when elucidated, will present a fascinating picture. Adenosylcobalamin- Dependent Enzymes Study of the cobalamin-dependent reactions which occur in microorganisms, and include analogs of those known or suspected to occur in mannnalian metabolism, has provided the bulk of current knowledge about the mechanism of action of cobalamin as a prosthetic group . Perhaps the most intriguing of these reactions are those illustrated in 6 Fig. 2 : a family of eleven AdoCbl-dependent reactions , ten rearrangments and a reduction. The apparent variety of the rearrangements is somewhat illusory, and all may be generalized by the following scheme: a vicinal interchange of hydrogen and a group "X". Varia- tion of the latter, which may be an --OH, -NH 2 , alkyl, or acyl group, is responsible for the observed diversity of the reactions. Mechanistic studies have shown that, unlike most enzyme-catalyzed reactions, these rearrangements have essentially no counterpart in laboratory chemistry. Figure 3 summarizes the current, generally accepted mechanism of adenosylcobalamin-dependent rearrangements. The first step is homolytic cleavage of the carbon-cobalt bond, a reaction which appears to be facilitated by binding of cofactor to enzyme, even in the absence of substrate. Subsequently, a hydrogen atom is transferred from C-1 of substrate to C-5' of the nucleoside moiety. This hydrogen atom becomes stereochemically equivalent to those initially present at C-5' by rotation about the C-4'-C-5' bond. The migration of group X from C1 to C2 is accomplished by a process which is not understood. (Some mechanisms suggested for this rearrangement are listed in Fig. 4. As discussed in Chapter III we currently favor the third one listed.) 7 Figure 2 Adenosylcobalamin-Dependent Reactions. have been found in mammalian tissues. tThese enzymes 8 Enzyme 1. Reaction Glutamate mutase COSCoA 2. Methylmalonyl CoA mutaset 3. a~Methyleneglutarate mutase 4. Ethanolamine deaminase 5. Biol dehydratase 6. Glycerol dehydratase I CH3-CH I I I CH 2 -CH 2 I COOR OH COSCoA COOR OH I CH2-CH-CH2 I OH 9 Reaction Enzyme NH2 I HOOC-CH 2 -CH-CH 2-CH2-CH 2NH2 t 7. NH 2 ~-a-Lysine mutase I HOOC-CH 2 -CH-CH 2-CH-CH 3 I NH 2 NH 2 ➔ I HOOC-CH-CH 2-CH 2-CH2-CH 2NH2 + 8. £-a-Lysine mutase NH2 I HOOC-CH-CH2-CH2-CH-CH3 I NHz NH 2 I HOOC-CH-CH2-CH2-CH2NH2 t 9. d-Ornithine mutase NH2 I HOOC-CH-CH3-CH-CH3 I NH2 10. Leucine 2,3-Aminomutase (Poston, 1976; 1977).t 11. Ribonucleotide reductase 10 Figure 3 Mechanism of Adenosylcobalamin-Dependent Rearrangements. CH2Ad I Adenosylcobalamin is represented schematically as >Co<, substrate as S, and product as P. 11 CH 2 Ad I SH >Co< J1' ·CH 2 Ad SH >Co< /-" ,,, ~ CH3Ad S· ·CH 2 Ad PH . >Co< ! \. ? . ,' >CO< I V, ~ P· - - - --- Figure CH 3Ad >Co< 3 12 To complete the rearrangement, one of the three hydrogens at c-5' of the cofactor is transferred to C-2 of substrate. With the enzymes diol dehydratase , glycerol dehydratase , and ethanolamine deaminase, an HX group is then eliminate d stereospecifically from C-1 prior to r elease of product . Hydrogen transfer. Because hydrogen transfer reactions lend themse1ves to studies involving i sotopic substitution, this aspect of the catalytic mechanism has been examined thoroughly. The migration of hydrogen occurs without ex - change with solvent in these rearrangements (Brownstein & Abeles, 1961; Babior, 1969; Somack & Costilow, 1973; Erfle et al., 1964; Iodice & Barker, 1963; Kung & Tsai, 1971; Wood et al., 1964). All of these reactions exhibit a primary isotope effect when hydrogen-isotope labeled substrates are used, indicating that cleavage of a C-H bond is an important rate-contributing step. The fate of the hydrogen derived from C-1 of substrate has been studied in detail. Early observations with diol dehydratase (Brownstein & Abeles, 1961) showed that hydrogen removed from C-1 appeared at C-2 of the product, thus seeming to indicate that the rearrangement is intramolecular However, the key experiment which showed otherwise was carried out by Abeles & Zagalak (1966) . Diol dehydratase was allowed to react with mixtures of [1- 3 H]-l,2-propanediol and unlabeled ethylene glycol, and tritium was found 13 Figure 4 possible Mechanisms for Migration of Group X. 14 X X 1. a -x 0 I I CH2-CH2 ~ CH2-CH2 • I >CO< III >CO< . II +Xe CH 2-CH 2 ➔ "'--, ,+i 'J II >CO< x > CH -CH I I 2 X I CH 2-CH 2 ➔ 2 III >CO< X 2.b I CH2-CH2 H(!) .II >Co< ~ I I ' I > CH 2 £!. CH 2 (£)XH I CH 2-CH2 -H(i) X . I > CH 2-CH2 X X CH2CH2 XH \ > CH 2-CH2 . 3.c EB HX@ X I > CH2-CH2 (£) .II >CO< ➔ 0 I >CO< X > I CH 2-CH 2 . II >CO< a) Silverman & Dolphin (1976). b) Golding & Radom (1976). c) Eagar et al. (1972); Halpern (1974). I CH 2 CH 2 @ 0 I >Co< 15 at c-2 of acetaldehyde, the product derived from ethylene glycol. This result implicated the holoenzyme as an inter- mediate hydrogen carrier. In later work, the 5'-carbon of AdoCbl was found to play this role in the diol dehydratase (Frey & Abeles, 1966), ethanolamine deaminase (Babior, 1968), and several other rearrangements . During catalysis with tritiated substrate, AdoCbl containing tritium at C~S' can be isolated by denaturing the holoenzyme complex with trichloroacetic acid. thetically Furthermore, both this biosyn- labeled coenzyme and the synthetically labeled species,when added to apoenzyme transfer all of their tritium to product during catalysis (Frey et al., 1967a). Particularly interesting is the observation that both C-5' hydrogens participate in this reaction (Frey et al., 1967b), an unexpected result in view of the observed stereospecificity of enzymatic reactions. Presumably, this would only occur if the two C-5' hydrogen .a toms become stereochemically equivalent during the course of catalysis. Futhermore, experiments carried out with mix- tures of [1-~HJ,-1,2-propanediol and unlabeled ethylene glycol as substrate indicated (Frey et al., 1967a) that even with infinite amounts of the latter, tritium would still be found at C-2 of propionaldehyde product. Thus, there is a finite probability for hydrogen to be returned to C-2 of the same molecule from which it was abstracted. These 16 observations suggest the intervention of an intermediate containing at least three stereochemically equivalent hydrogens, one of which is transferred to C-2 of the product molecule. The presence of this three-hydrogen intermediate along the catalytic pathway has been inferred in a number of studies. Use of 3 H as a trace label in the dioldehydratase (Frey et al., 1967a; Essenberg et al., 1971) and ethanolarnine deaminase (Babior, 1969; Weisblat & Babior, 1971) reac- tions has supplied such evidence. The methylmalonyl CoA mutase (Miller & Richards, 1969) and glutamate mutase (Eagar et al., 1972) reactions were studied by allowing mixtures of unlabeled and deuterated substrates to proceed partially toward equilibrium, after which the distribution of deuterated substrate and product molecules was determined by mass spectroscopy for comparison to distributions calculated for various models of the reaction. The results of these and other studies point to the existence of an intermediate in which the hydrogen derived from substrate becomes equivalent to the two hydrogens present at C-5' of AdoCbl. A likely candidate for this intermediate is S'-deoxyadenosine, as first suggested by Ingraham (1964). A plausible mode of its formation would be rupture of the carbon-cobalt bond of AdoCbl, followed by hydrogen transfer 17 to the 5'-carbon atom of the resulting nucleoside. Stereo- chemical equivalence of the three hydrogens could beeffected by rapid rotation of the C-5' methyl group, as has been observed by 1H NMR for nucleosides bound to ethanolamine deaminase (Hull et al., 1975). In fact, formation of 5'-deoxyadenosine has been demonstrated in the diol dehydratase (Wagner et al., 1966; Finlay et al., 1972), methylmalonyl CoA mutase (Babior et al., 1973), !-e-lysine mutase (Baker et al., 1973), and ethanolamine deaminase (Babior, 1970a).reactions. However, in these cases 5'- deoxyadenosine was formed during reaction with substrate analogs which irreversibly inactivate holoenzyme; hence, 5'-deoxyadenosine could be construed as the product of a destructive side reaction, rather than of catalysis. In contrast, results obtained from further studies of ethanolamine deaminase (Babior, 1970b; Babior et al., 1974a), in which 5'-deoxyadenosine was detected upon interruption of catalysis, showed that the amount of 5'-deoxyadenosine recovered varied with the denaturing agent used. This result was interpreted to indicate reversible formation of this moiety during catalysis. Thus, the role of 5'- deoxyadenosine as an intermediate hydrogen carrier in AdoCbl-dependent rearrangements is fairly well established. Cleavage of the Carbon-Cobalt Bond. In principle, the carbon~cobalt bond of AdoCbl may be cleaved in three ways to produce a 5'-deoxyadenosyl fragment: 18 © CHzAd ·CHzAd 0 CH 2 Ad >co< (CblIII) Each type of cleavage would produce a cobalt atom of a different, stable electronic structure (Co I , d 8 ; Go II , d 7 ; or co 111 , d 6). In addition, this reaction presumably determines the manner in which hydrogen is transferred, i.e., as hydride, atom, or proton. Thus, the nature of this reaction has been the subject of considerable effort. Current evi- dence, mostly spectroscopic, favors the second of these, hemolytic cleavage to yield a Cbl 11 species and S'-deoxyadenosyl radical. When diol dehydratase (Wagner et al., 1966; Eagar et al., 1975) and ethanolamine deaminase (Joblin et al., 1975) holoenzymes are allowed to react with their respective substrates, the uv-visible absorption spectrum of the holoenzyme complex changes from an AdoCbl spectrum to one very similar to that of Cbl 11 (low spin d 7), at a rate kinetically competent with catalysis. Upon deple- tion of added substrate, the spectrum changes to one characteristic of OH-Cbl (Fig . 1, CblIII) in an oxygen-dependent process which results in destruction of enzyme-bound cofactor (Wagner et al., 1966) and inactivation of holoenzyme. A similar reaction occurs in the presence of oxygen with CblII free in solution. In efforts to detect and characterize unpaired electrons 19 in these systems, electron spin resonance experiments have been carried out with ethanolamine deaminase (Babier et al., 1972; 1974b), diol dehydratase (Finlay et al., 1973), and glycerol dehydratase (Cockle et al., 1972). When solutions of these holoenzymes are frozen during reaction with substrates or substrate analogs, spectra are observed with a signal at g = 2.2 (assigned to Cbl 11 ), and a narrower doublet signal at g"" 2, attributed to a free radical. The appearance of these signals requires the presence of both holoenzyme and substrate or analog , and occurs on a time scale sufficiently rapid to represent actual catalytic intermediates (Valinsky et al., 1974). These signals have been simulated in terms of a weak electrostatic exchange interaction between two unpaired electrons (Schepler et al., 1975). Thus, a variety of evidence for homolysis of the cobaltcarbon bond of AdoCbl has been accumulated; the participa- tion of this species as a catalytic intermediate seems to be on solid ground. Recently, a paper has appeared (Toraya et al., 1979) which indicates that cleavage of this bond is effected by interaction of the peripheral amide side chains of the corrin ring with the protein in the diol dehydratase reaction. • Migration of Group X. While the above aspects of the Iilechanism of AdoCbl-dependent rearrangements (Fig. 3) are fairly well established, very little is known about the 20 nature of group X migration. Elegant stereochemical studies employing 18 0-labelled substrates enabled Retey et al. (1966b) to infer the existence in the diol dehydratase reaction of a 1,1-propanediol intermediate which is dehydrated stereospecifically to aldehyde by the enzyme. The car- bon from which group X migrates undergoes a net inversion of configuration in the diol dehydratase (Retey et al., 1966a; Zagalak et al., 1966) and glutamate mutase (Sprecher & Sprinson, 1964) reactions, while retention was observed with methylmalonyl CoA mutase (Sprecher et al., 1964; Retey & Zagalak, 1973) and racemization with ethanolamine deamin- ase (Retey et al., 1974). A standard enzymological technique used in approaching such problems is to demonstrate an enzyme-catalyzed exchange between the reaction products and compounds thought to represent the form of the intermediate in question. These experiments have enjoyed no success in studies of AdoCbl-dependent enzymes . Moreover, attempts to generate active holoenzyme by allowing apoenzyme to interact with 5'-deoxyadenosine and several cobalamin derivatives postulated to represent catalytic intermediates (Carty, 1973; Eagar, 1974; Krouwer & Babior, 1977; Krouwer et al., 1978) have also failed. In view of the lack of information concerning group X migration which is available from study of the enzyme mechanisms, chemists have examined model reactions in 21 efforts to explain this aspect of the reaction. Some of these studies and the proposed mechanisms arising from them (Fig. 4) have been reviewed (Schrauzer, 1971, 1976; Brown, 1973; Abeles & Dolphin, 1976). More recently, Gold- ing & Radom (1976) have suggested that 1,2-rearrangements of radicals can be facilitated by protonation of the migrating group. Corey et al. (1977) have suggested a novel mechanism involving oxidative addition and reductive elimination with substrate and Cbl 1 in which the corrin ring undergoes electrocyclic cleavage; however, several features of the enzymatic reactions are inconsistent with this mechanism. For example, no evidence has been found for the existence of an intermediate in which substrate becomes covalently attached to cobalt; the Corey mechanism does not account for the observation of ESR spectra exhibiting Cbl 11 and organic radical resonances; the acetamide group at C-2 of corrin, postulated to initiate electrocyclic cleavage, is at best a questionable nucleophile. At present, no genuinely satisfactory picture of the mechanism of these rearrangements has arisen from model reactions studied thus far. The work presented in this thesi.s describes studies on a new class of reactions of diol dehydratase. Diol dehy- dratase catalyzes the conversion of a number of vicinal diols other than 1 , 2- propanediol to the corresponding carbonyl compounds. In addition, in a concomitant and 22 competing reaction, these substrates irreversibly inactivate holoenzyme. We have carried out a detailed mechanistic study of the inactivation reaction (Chapters I and II). The results of these experiments suggested further experi-ments on the mechanism of catalysis, which are reported in Chapter III. The results, which indicate that the mechanism of inactivation is strikingly similar to that of catalysis, and suggest an important role for the protein in this rearrangement, are of value in enhancing our understanding of the mechanism of action of adenosylcobalamin as a cofactor in enzymatic processes. Some of the work described has already appeared (Bachovchin et al., 1977, 1978; Moore et al., 1979; Moore & Richards, 1979). 23 REFERENCES Abeles, R. H. ' & Dolphin, D. (1976) Acc. Chem. Res . ~. llL~ . Abeles, R. H.' & Zagalak, B. (1966) J. Biol. Chem. 241, L~45 . Babior, B. (1968) Biochim. Biophys. Acta 167, 456. Babior, B. M. (1969) J. Biol. Chem. ?4~, 449. Babior, B. M. (1970a) J. Biol. Chem. 245_, 17.55; (1970b ) J . Biol. Chem. 245, 6125. Babior, B. M., Moss, T. H., & Gould, D. C. (19 72) J . Biol . Chern . 247, 4389. Babior, B. M. , Woodams, A. D., & Brodie , J . D . (1973) J. Biol. Chem. 248, 1445 . Babior, B. M., Carty, T . J., & Abeles, R. H. (1974a) J. Biol . Chem. 249, 1689. Babior , B. M., Orme -Johnson, W. H., Beinert, H., & Moss, T. H. (197L}b) J. Biol. Chem. J_49, 4.537. Bachovchin, W.W., Eagar, R. G., Jr., Moore, K. W., & Richards, J. IL (1977) Biochemistry 16_, 1082 . Bachovchin , W: w:, Moore, K. W., & Richards, J. H. (1978) Biochemistry 1~, 2218. Baker, J. J., van der Drift, C., & Stadtman , T. C. (1973 ) Biochemistry 12, 1054.. Barley, F. W., Sato, G. H., & Abeles, R.H. (1972) J. Biol. Chem. ~47, 42.70. Beck , W. S. (1975) in Cobalamin, B. Babier, Ed., Wiley and Sons, New York, Ch. 9. Brown, D. G. (1973) Prog. Inor g. Chem. 18, 177 . Brownstein, A. M., & Abeles, R. H. (1961) J. Biol. Chem. 236_, 1199. 24 Cardinale, G. J., Carty, T. J., & Abeles, R.H. (1970) J. Biol. Chem. ~45 , 3771 . Carty, T. J. (19 73) Ph . D. Dissertation, Brandeis University. Cockle, s. A. , Hill, H. A. 0. ' Williams, R. J. P., Davies , s. P .' & Fos t er, M. A. (1972) J. Am . Chem. Soc. 94, r ·5_ Corey, E. J.' Cooper, N. J., & Green, M. L. H. (1977) Proc. Natl. Acad. Sci. USA 74, 811. Eagar, R. G., Jr. (1974 ) Ph. D. Dissertation, California Institute of Technology. Eagar, R. G., J r., Baltimore, B . G ., Herbst , M. M., Barker.·, H. A., & Richards , J. H. (1972) Biochemistry g, 253. Eagar, R. G., Jr., Bachovchin, W.W. , & Richards, J. H. (1975) Biochemistry 14, 5523 . Ellenbogen, L. (1975) in Cobalamin , B. Babior, Ed., Wiley and Sons, New York, Ch. 5. Erfle , J. D., Clark, J . M., Nystrom, R. F., & Johnson, B. C. (1964) J . Biol. Chem. 23~, 1920. Essenberg, M. K., Frey, P . A., & Abeles, R.H . (1971) J. Am.Chem. Soc . 93, 1242. Finlay, T. H., Valinsky, J., Sato, K., & Abeles, R.H. (1972) J. Biol. Chem. 247, 4197. Finlay, T. H., Valinsky, J., Mildvan, A. S., & Abeles, R. H. (1973) J. Biol. Chem. 248, 128.5. Frey, P. A., & Abeles, R. H . (1966 ) J. Biol. Chem. 241:_, 2732. Frey, P . A., Essenberg, M. K. , & Abeles , R. H. (1967a ) J . Biol. Chem . 242 , 5369. Frey, P. A., Kerwar, S . S . , & Abele.s, R. H. (1967b) Biochem. Biophys. Res. Commun . ~9, 873. 25 Golding, B. T., & Radom, L. (1976) J. Am. Chem. Soc. 98, 6331. Green, R., van Tonder, S. V. , Oettle , G. J ., Cole, G., & Metz, J. (1975) Nature 254, 148 . Halpern, J. (1974) Ann. N. Y. Acad. Sc i. 239, 2. Hodgkin, D. C., Pickworth, J., Robertson, J. H., Truebloo d , K. N., Prosen, R. J., White, J. G., Bonnett, R., Cannon, J . R., Johnson , A. W., Sutherland , I., Todd, Sir A. , 6, Smith, E. L. (1955) Natur e 176, 325 . Hogenkamp, H.P. C. (1975) in Cobalamin, B. Babier, Ed., Wiley and Sons, New York , Ch. 1 . Hull, W. E., Mauck, L., & Babier, B. M. (1975) J. Biol. Chem. 250, 8023. Ingraham, L . L. (1964) Ann . N. Y. Acad. Sci. 112, 713. Iodice, A. A., & Barker, H. A. (1963) J. Biol. Chem. 238, 2094. Joblin, K. N., Johnson, A. W., Lappert, M. F., Hollaway, M. R., & White, H. A. (1975) FEBS Lett. 53, 193. Kishimoto, Y., Williams, M., Moser, H. W. , Hignite, C., & Biemann, K. (1973) J. Lipid Res. 14, 69. Krouwer, J. S., & Babier, B. M. (1977) J . Biol. Chem. 252, 5004. Krouwer, J . S., Schultz, R. M., & Babior, B. M. (1978) J. Biol. Chem. 253, 1041. Kung, H. F., & Tsai, L. (1971) J . Bio l. Chem. 246, 6436. Lenhert, P. G., & Hodgkin, D. C. (19 61) Nature 192, 937. Miller, W.W . , & Richards, J . H. (1969) J. Arn. Chem. Soc. 91 , 1498 . Moore, K. W. , Bachovchin, W. W. , Gunter, J. B. , & Richards . J. H. (1979) Biochemistry , in press . 26 Moore, K. W., & Richards, J. H. (1979) Biochem. Biophys. Res. Connnun., in press. Poston, J.M. (1976) J. Biol. Chem. 251, 1859; (1977) Science 195, 301. Retey, J., & Zagalak, B. (1973) Angew. Chemi Retey, J . , Umani-Ronchi, A., ientia 22, 72. & Arigoni, Retey, J., Umani-Ronchi, A., Seible, J., (1966b) Experientia 22, 502. 85, 721. D. (1966a) Exper& Arigoni, D. Retey, J., Suckling, C. J . , Arigoni , D., & Babior, B. M. (1974) J. Biol. Chem. 249, 6359. Rickes, E. L., Brink, N. G., Koniuszy, F. R., Wood , T . R., & Folkers, K. (1948) Science 107, 396. Schepler, K. L., Dunham, W.R., Sands, R.H., Fee, J. A., & Abeles, R. H. (197 5) Biochim. Biophys. Acta 397, 510 . Schrauzer, G. N. (1971) Adv. Chem. Ser. 100, l; (1976) Angew. Chemie Int. Ed. Engl. 15, 417. Silverman, R. B., 98, 4626. Smith, E. L., & & Dolphin, D. (1976) J. Am. Chem. Soc. Parker, L. F. J. (1948) Biochem. J. 43, viii . Somack, R., & Costilow, R. N. (1973) Biochemistry 12, 2597 " Sprecher, M., & Sprinson, D. B. (1964) Ann . N. Y. Acad. Sci. 112, 655. Sprecher, M., Clark, M. J., & Sprinson, D. B. (1964) Biochem. Biophys. Res. Commun. 15, 581. Toraya , T. , Krodel, E., Mildvan, A. S., & Abeles, R. H. (1979) Biochemistry 18, 417. Valinsky, J., Abeles, R.H., & Fee, J. A. (1974) J. Am. Chem . Soc. 96, 4709. 27 Wagner, 0. W., Lee, H. A., Jr., Frey , P. A., & Abeles, R. H. (1966) J. Biol. Chem. 241, 1751. Weisblat, D. A., & Babior, B. M. (1 971 ) J . Biol. Chem . .• 246, 6064. Wood, H. G., Kellermeyer, R. W. , Stjernholm, R., & Allen , S. H. G. (1964) Ann. N. Y. Acad. Sci. 112, 661. Zagalak, B., Frey, P.A . , Karabat sos, G. L., & Abeles, R.H. (1966) J. Biol. Chem. 241, 3028 . 28 CHAPTER I GLYCEROL AND OTHER SUBSTRATE ANALOGS AS SUBSTRATES AND INACTIVATORS OF DIOL DEHYDRATASE: KINETICS, STEREOSPECIFICITY, AND MECHANISM. 29 INTRODUCTION Propanediol dehydratase ((RS)-l,2 - propanediol hydro lyase; EC 4 .2. 1.28) from Klebsiella EEeumonia~. (ATCC 8724) is one of ten known enzymes which utilize adenosylcobal.amia as a cofactor to catalyze 1, 2 ,~ rearrangements of hydrogen and some group X, where X can be an alkyl, acyl, amino, or hydroxyl substituent (Hogenkamp, 1968; Bab:i.or, 1975a,b; Abeles & Dolphin , 1976) . These enzymes exhibit a r emark- able specificity for their respective substrates. For example, although diol dehydratase will catalyze the conversion of both (R)- and (S)-1,2 -propanediols (albeit with different rates and different eventua l fates for the C-2 hydroxyl oxygen of the two isomeric substrates (Retey et al . , 1966)), relatively minor structural modifications car. abolish catalytic or inhibitory activity (Lee & Abeles , 1963). Toraya & Fukui (1972) found that 1,2-butanediol ar. i styrene glycol, although catalytically inert, are weak com petitive inhibitors. Eagar et al. (1975) showed that 3-fluoro-1,2 -p ropanediol is a substrate for diol dehydratase with catalytic constants comparable to those of 1,2propanediol. Toraya et al . (1976) also reported that glyce rol functions both as a substrate and an inactivator of diol dehydratase . This chapter reports results with a number of modified vicinal dials including glycerol, thioglycerol, 3-chloro1,2-propanediol, 1,2-butanediol, isobutylene glycol, 30 3 ,3, 3-trifluoro-1,2-propanediol, and meso-, d-, and -dl--2,3-butanediols. In general, these substances act as substrates (leading to product formation by the usual rearrangement), and also bring about i rreversible inactivatior, of holoenzyme. The inactivation reactions are all depen- dent on both apoenzyme and cofactor, and independent of the presence of oxygen, which in the absence of substrate is a potent inactivator of the holoenzyme. We have studied the mechanism of catalysis and inactivation for glycerol in some detail because of its close structural resemblance to the normal substrate, 1,2-propanediol, its presence in many biological environments (glycerol is in fact a component of the growth medium use d to induce propanediol dehydratase), and because of the existence of an adenosylcobalamin-dependent glycerol dehy dratase that has been reported to be inactivated by its substrate, glycerol (Schneider & Pawelkiewicz, 1966). Moreover, even though glycerol is such a potent inactivator that significant product can be observed only when large amounts of enzyme are used, the rate constant for product formation exceeds that observed for the noninactiva- ting substrate 1,2-propanediol . EXPERIMENTAL Enzyme Preparations . Diol dehydratase was obtained from Klebsiella pneumoniae by a procedure similar to that 31 of Lee & Abeles (1963). Fraction E-8 with a specific acti- vity of between 25 and 60 was used for all determinations. 1,2-Propanediol-free enzyme was prepared by dialysis at 4° c against several changes of 10 mM K2HP0 4 buffer (Frey et al., 1967). Adenosylcobalamin. Adenosylcobalamin (AdoCbl) was pur- chased from Sigma Chemical Co. Hydroxycobalamin. Hydroxycobalamin (0H-Cbl) was pur- chased from Sigma Chemical Co. Assays. All assays for product aldehyde were carried out in the dark or in dim red light at 37° C. Two dif- ferent methods were used and are described below. In sev- eral cases, results were duplicated by both methods. DNP Assay. Aldehyde was assayed colorimetrically as the 2,4-dinitrophenylhydrazone using the method described by Eagar et al. (1975). Alcohol Dehydrogenase-~-Nicotinamide Adenine Dinucleotide Assay. Yeast alcohol dehydrogenase (ADH, Sigma) and a-nicotinamide adenine dinucleotide (reduced form, NADH, Sigma) reduce propionaldehyde to 1-propanol. The maximum velocity of this reaction is 30 % of that for acetaldehyde reduction (Bruennner & Roe, 1971). The production of pro- pionaldehyde by diol dehydratase can be measured by monitoring the decrease of absorbance at 340 nm, due to oxidation of NADH to NAD in the presence of excess ADH. Determination of Inactivation Rates. Coupled Enzyme 32 Assay. Reaction mixtures consisted of: ADH, 15 units; NADH, 0.2 mM; diol dehydratase, 0.1 unit; bovine serum albumin, 0.06 mg; potassium phosphate buffer, pH 8.0, 30 ml:1; AdoCbl, 0.02 rnM; and the desired amount of 1,2 =propanediol or in.activator. Total volume 2.3 ml. The reaction mixture was incubated at 37° C with stirring. The inactivation reac- tion was started by addition of AdoCbl and stopped by addition of 0.05 ml 6 M 1,2 .. propanediol solution. Enzyme acti- vity (measured by the rate of product ion of propi.onaldehyct 2) was determined at a minimum of four different inactiva.tioL times; the slope of the corresponding semilog plot represents the observed inactivation rate constant (k i,obs ) for the given concentration of inactivator and 1,2-propanediol. The reproducibility of k.J., 0 b S as measured by this method is quite reliable, usually within three to five percent. However, when k.1., 0 b S is quite small ( less than about 0.06 min - 1 ) or large (greater than about 1.2 min -1 ) the uncertainty can in some cases approach eight or nine percent. For rate determinations under anaerobic conditions, both the reaction mixture and AdoCbl solution were thoroughly deoxygenated under argon. AdoCbl was transferred to the reaction mixtures using a gas-tight syringe. DNP Method. Reaction mixtures consisted of the follow- ing: diol dehydratase, about 0.06 unit; potassium phosphate buffer, pH 8.0, 30 mM; AdoCbl, 0.013 mM; bovine serum 33 albumin, 0.4 mg; and the desired amount of substrate and/or inactivator. Total volume 2.0 ml, 37° C. The reaction was started by addition of AdoCbl and the inactivation quenched with 0.05 ml 6 M 1,2-propanediol. The amount of enzyme remaining active as a function of time after the addition of AdoCbl was determined by measuring the production of propionaldehyde at 1 min intervals for 8 min after the addition of 1, 2-propanediol. The rate of propionaldehyde pro •- duction was linear for at least 8 min after the addition of 1,2-propanediol. Determination of !-Hydroxypropionaldehyde. The produc- tion of s-hydroxypropionaldehyde was determined by the acrolein-specific assay of Circle et al. (1945). Reaction mixtures consisted of diol dehydratase,"'3 units, potassium phosphate buffer, pH 8.0, 30 mM; bovine serum albumin, 0.4 mg; AdoCbl, 0.013 mM; and glycerol, 70 mM. ml, 37° C. Volume 2 .0 The reaction was started by the addition of AdoCbl and quenched by the addition of 0.1 ml 2 N HCl. S-Hydroxypropionaldehyde can also be determined satisfactorily using the DNP assay as described for propionaldehyde. The reaction mixtures used were the same as above with 0 . 6 unit of diol dehydratase giving the best results. Absorbance was measured at 475 nm. The extinction coeffi - cient at 475 nm from the dinitrophenylhydrazone of s-hydroxypropionaldehyde is somewhat smaller than that from the derivative of propionaldehyde. 34 [l- 14 C]-Glycerol Labeling of Diol Dehx:dratase. Reac - tion mixtures consisted of the fol l owing: apoenzyme, 40 units; [l- 14 C]-glycerol (0 . 47 mGi/mmole), 0.14 M; potassiu;\L phosphate buffer, pH 8.0, 30 mM; AdoCbl , 0.03 mM; volume 0.5 ml. Two controls were performed, one in which AdoCbl was omitted from the reaction and one in which OH- Cbl was substituted for AdoCbl. The reaction mixtures were incu- bated at 37° C for 30 min. Two basic p r ocedures were use L: for removing glycerol, trichloroacetic acid (TCA) precipi -· tation of the protein, or gel chromatography, in each cas P followed by dialysis. Following the 30 min incubation pe r - iod, 0.04 ml 20 % TCA was added. After centrifugation , the supernatant was removed and the protein pellet was washed in 0.5 ml H2 o and centrifuged. The resulting protein pel• let was suspended in 1.0 ml H2 0 and dialyzed successively against O.13 1 10 mM K2 HP0 4 , 3 x O.13 1 of a solution con •taining 8 M urea; 50 mM Tris citrate, pH 4; 1 mM disodium EDTA; and 0.1 % sodium dodecyl sulfate, and finally against 3 x O. 13 1 10 mM K 2 HPOL~. When the radioactivity of the dialysis solution had diminished to background, 0.45 ml of the sample and 0.45 ml H2 o were added to 10 ml PCS:xylene (2:1) and assayed for 14 C. A 0.1 ml aliquot was assayed for protein content by the method of Lowry et al. (1951). Alternatively, the 0.5 ml sample was applied to a 2 . 6 x 15 cm BioGel P-150 column and eluted with a buffer 35 containing 10 mM K2 HP0 4 , 0.1 M g lycerol, and 0.01 % NaN . 3 The protein-containing fractions were pooled and dialyzed as described above. A 0.9 ml aliquot of the dialysate was assayed for radioactivity and a 0.4 ml aliquot for protein. Substrates. Perdeuterioglycer ol (gly-d 5 ) was purchased from Merck, Sharp, and Dohme, Ltd ., and used without further purification. [l - 14 CJ-Glycerol was purchased from Amersha.m. (R) -.!., 1:_-Dideuteriog_!ycerol ( (R) =gly-d 2 ) was synthes iz-:~d from ~-mannitol via the intermediate pre paration of the 1,2,5,6-diisopropylidene derivative (Baer, 1952), and of d-2,3-isopropylidene glycerol (Ghangas & Fondy, 1971). The experimental details of the following steps are given by Eagar et al. (1975). Potassium s!_~2 , 3-isopropylidene glycerate was prepared by alkaline potassium permanganate oxidation of d-2,3-isopropylidene glycerol and was esteri ·· fied with methyl iodide in hexamethylphosphoramide . The methyl ester was then reduced with lithium aluminum deute:i.:ide (99 % D, Stohler Isotope Chemicals). The ketal was hydrolyzed in ethanol-sulfu.ric acid and the product distilled to give (R)-1,1-dideuterioglycerol with [a]l 5 -0 .14° 1 . Deuterium content was at least 98 % as determined by protcn NMR. (RS) -!_ 1 !_-Dideuterioglycerol ( (RS) •" g ly··d.z) was preparea 1. Optical rotations are uncorrected for water content (Huff, 1961) . 36 from (RS)-1,2-isopropylidene glycerol (Ghangas & Fondy, 1971) by a method analogous to that described above. (S)-1,1-Dideuterioglycerol ((S)-gly-d 2 ) was prepared from l,l-dideuterio-(d)-2,3-isopropylidene glycerol (Eagar et al., 1975). The alcohol (47g, 0.351 mole) was dissolved in 250 ml H2 0 and a few crystals of p - toluene sulfonic acid added. Water and acetone were removed by evaporation at"' 50°c under reduced pressure. To the resulting syrup was added a 10% molar excess of distilled benzaldehyde anc· 250 ml benzene. Water was removed by repeated azeotropic distillation with benzene under reduced pressure. The syrup was dissolved in 150 ml benzene and washed twice with 2% K2 C0 3 . The combined aqueous washings were extracted with benzene, which was added to the organic fraction. The benzene solution was dried and evaporated to a syrup, which contained a mixture of 1,2- and 1,3-benzylidene ketals of (R)-gly-d 2 (Sedarevich, 1966; Cambie et al., 1972). Several bubbles of dry HCl were passed through tht- 0 syrup, which was then dissolved in"' 500 ml of 4:3 ligroin: benzene. The solution was allowed to cool to 4°c. Addi- tion of a tiny seed crystal initiated crystallization of cis-l,3-benzylidene-(R)-1,1- dideuterioglycerol, which was collected by filtration and washed with cold solvent. Several repetitions of this procedure, followed by recrystallization of the collected material, yielded 26.7g, 37 (0.147 mole) of the cis ketal, mp 83° (Cambie et al., 1972). The ketal was inverted to trans-l,3-benzylidene-(S)1,1-dideuterioglycerol-2-benzoate by the following procedure, adapted from Mitsunobu & Eguchi (1971) and Alfredsson & Garegg (1973). The ketal (26g) and triphenylphosphine (75g) were dissolved in 250 ml dry tetrahydrofuran (THF) in a flame-dried flask under argon while cooling in an ice batL. A solution of 35g benzoic acid and 50g diethylazodicarboxylate, in dry THF, was added dropwise with stirring. After stirring overnight, the solution was washed with 5% K2 C0 3 , dried over Na 2 S0 4 , and evaporated to dryness under reducec: pressure. The solid was dissolved in a minimum amount of 4:1 toluene:ethyl acetate and chromatographed on silica gel in this solvent. The desired product eluted first, and was evaporated to dryness. The solid was dissolved in a minimum amount of acetone and allowed to dry slowly. The resulting crystalline mass was washed with methanol and dried in vacuum over KOH. Yield 22.8g (80 nnnole), mp 103-1040. The benzoate ketal was refluxed for 3 hr in 1 N HCl. The acid solution was cooled and extracted with chloroform and ethyl acetate . The aqueous layer was concentrated under reduced pressure. Distillation under vacuum at a bath temperature of 155-170° gave 4.6g (49 mmol) (S)1,l-dideuterioglycerol, [a]~ 5 +0.14. 0 38 (RS)-l,l-Dideuterio-1,2-propa~ediol was prepared by lithium aluminum deuteride reduction of dl-ethyl lactate (Fieser & Fieser, 1967). 98%, as determined by Deuterium content was at least 1 H NMR. The isotope effect on cata y - sis was 13, somewhat greater than the value of 10-12 reported by Abeles (1972). The UV- visible spectrum of the holoenzyme-dideuteriopropanediol complex was identical to those rep·orted for 1, 2-propanediol (Wagner et al., 1966) and 3-fluoro-1,2-propanediol (Eagar et al., 1975). 3,3,3-Trifluoro-l,2-propanediol was prepared from 3,3,3-trifluoro-l-bromacetone by lithium aluminum hydride reduction to the corresponding alcohol (McBee & Burton, 1952), which was collected by fractional distillation at 124-124.SOC. The alcohol (12.8g, 0.0674 mole) was then added to a solution of 4g of NaOH in 10 ml of water, and 3,3,3-trifluoro-l,2-epoxypropane distilled out of the reaction mixture in 95% yield. The epoxide was refluxed overnight in 1% H2so 4 , the solution was extracted five times with ethyl ether, and the ether was evaporated to give the diol. Isobutylene glycol was prepared by acid hydrolysis of isobutylene oxide (Columbia Organic Chemical Co.) (Pattison & Norman, 1957). 1,2.-Butanediol was prepared by acid hydrolysis of 1butene oxide. 39 3-Chloro-1,2-propanediol was purchased from Calbiochem and distilled before use. Thioglycerol (3-mercapto-l,2-pro2!?,nediol) was purchased from Aldrich Chemical Co. and distilled before use. meso-2,3-Butanediol was purchased as a mixture of the - meso and d isomers from Aldrich and distilled before use. - A 1 M solution showed virtually no optical rotation, so little or no~ isomer is present. The 1 isomer is not a natural product; as the Aldrich material is prepared biologically, it was assumed to contain no 1 isomer. d-2,3-Butanediol was purchased from Burdick & Jackson and had [a]5 5 - 11. 25° (c = 1 M). This compound contained approximately 1.6 mole% 1,2-propanediol as an impurity, as determined by kinetic and chemical analysis with diol dehydratase holoenzyme. This impurity was removed en- zymatically prior to use in kinetic experiments. dl-2,3-Butanediol was prepared by oxidation of cis2-butene with performic acid. 22g cis-2-Butene was condensed in a 1-1 3-neck flask at -10°c. (125 ml) was added with stirring. Chloroform A solution was prepared by mixing 41g 90% formic acid and 44.5g 30% H2o2 and was added dropwise with stirring, with the temperature kept between -5° and -10° C. After 15 hr, the solution was concentrated to~ 70 ml under reduced pressure, and neutralized carefully with 6 N NaOH. The pH was adjusted to 7-7.2, and the solution was evaporated to near dryness. The 40 sticky solid was extracted with chloroform. The chloroform was dried over Na 2so 4 , decolorized with charcoal, filter ed, and evaporated under reduced pressure. From the resul ting brown syrup was distilled 12g dl-2,3-butanediol under vacuum at 51° C. 1,2-Propanediol, glycerol, and ethylene glycol were purchased as reagent grade chemicals and distilled before use. (R) -· and (S)-1, 2-Propanediols were synthesized by a method previously described (Eagar et al., 1975). Con- centrations of aqueous solutions of these glycols were determined by bichromate oxidation using a method adapted from Englis & Wollerman (1952). NMR spectra were obtained at 60 MHz in the Fourier transform mode on a Varian T-60 spectrometer in deuteriochloroform or 2 H 0. 2 UV-Visible spectra of holoenzyme-substrate complexes were obtained between 300 and 600 nm at 27°and 37°C using a Beckman Acta III spectrophotometer. The sample solution was assayed for residual enzyme activity to confirm that apoenzyme was present in excess of cofactor. Sodium dodecyl sulfate (SDS) gel electrophoresis was carried out with materials purchased from Bio-Rad. 12% and 7.5% crosslinked gels were employed. Isoelectric focussing was carried out on pH 3.5-9 PAG plates from LKB. Diol dehydratase apoenzyme and o2- 41 and glycerol-inactivated holoenzyme were not denatured before application to the focussing gel. RESULTS Kinetics of Glycerol Inactivation in_ the Absence of ~,~~?ropanediol. In the presence of glycerol and in the absence of 1,2-propanediol and oxygen, diol dehydratase holoenzyme undergoes rapid, irreversible inactivation such that catalytic activity cannot be regenerated by addition of either 1, 2-propanediol or AdoCbl. or by extensive dialy sis. Moreover, addition of saturating amounts of 1 , 2- pro · panediol will instantly stop the inactivation react i on . This provides a method for quantitatively measuring the kinetics of the inactivation reaction. Saturating amountE of 1, 2-propanediol are added after variable periods of tin,e to a holoenzyme-glycerol mixture, thus quenching the inac--· tivation; the amount of enzyme remaining catalytically active can then be determined by measuring the sub sequent rate of propionaldehyde production. The rate of inactivation follows a first-..Jrder rate law through at least three half-lives [E where a J = [E o Je -k. i,obs (t) (1) is the amount of active enzyme remaining at time t, E0 is the initial amount of enzyme present, and k l,O . bS is the observed first-order rate constant for inactivation . 42 Thus, a plot of ln[Ea] as a function of time is linear and the slope gives a value of k.l., 0 b S for a given glycerol con= centration . A plot of k.1., ob s as a function of glycerol con- centration (Fig. 1) shows Michaelis-Menten type saturation behavior. The inactivation process can, accordingly, be described in terms of reversible association between enzyme and glycerol, followed by irreversible inactivation: k.l. E I + (2) where I represents irreversibly inactivated holoenzyme. The rate of inactivation is = (3) This differs from the classical situation because [EG] itself does not remain constant. However, if one considers only that enzyme which is still active, [E ] a = [E J + [ EG J , where Eis active enzyme free in solution, with [EG] = the race of inactivation (eq 3) becomes 43 Figure 1 Dependence of the observed rate of inactivation , k 1. , 0 b S , on [G] under anaerobic conditions in the absence of 1,2-propanediol . Values of ki ,o b s were obtained from the linear least squares plot of ln% residual enzyme activity vs time for each glycerol concentration. The correspond ing . -1 double reciprocal plot is linear and gives ki == 1 . 3 min and KG = 1. 6 mM. .0 0 r/l ....sC: ~ .... .:,,: I ~ 0.2 0.4 0.6 I 0.8 1.0 1. 2 I 2.0 [ G ) X 6.0 Figure 4.0 1 10+ 3 8.0 10.0 12.0 ~ ~ 45 -d[Ea] = dt ki[Ea] (4) ~[ Gl Integration gives [Ea] = -k [Ea]oe i,ob s (t) (5) where k. k.l., 0 b S = Thus, a plot of 1/ki ,o b s vs . l. (6 ) 1 + KG [ G] 1/[G) should , and does give a straight line (Fig . 2); they-intercept gives 1/ki; the slope or the x-intercept determines KG. The values thus determined appear in Table I. Thermodynamic Parameters. The rate constants ki fo r glycerol inactivation and k p for catalysis with 1 ,2- pr opanediol were measured as a function of temperature over the range 12° - 38° C at saturating concentrations of thes 0· substrates. The results for glycerol ar e shown in Fig . 3 as a plot of ln(ki) vs 1/T. The resulting activation energies and entropies (Moore, 1962) are Ea= 9 .2 + 1 and 16 + 1 kcal/mole and (at 37° C) ~S t= - 17.8 and - 16 . 4 e.u. for propanediol and glycerol , respec tive l y . The value of ki for glycerol was independent of pH in the range 6 . 0 - 8.6. Kinetics of Glycerol Inactivation in the Pr esence of 46 Figure 2 The reciprocal of the observed inactivation rate constant vs the reciprocal of glycerol concentration in the presence of the following concentrations of (RS ) -1,2-pro panediol: ■, 1.9 mM; • , 1.34 mM; A, 0.86 mM; ♦, 0. 4-8 mM; 0, no (RS)-1,2-propanediol. Although only one point can be depicted for the substrate - free case, a straight line is obtained with the illustrated slope and intercept " N (l) Ii ~ (JQ t-'• 1-rj c:: .... •r-l .:.:: ~ - .0 0 If) ~ E: •r-l I .... I 2 3 4 5 6 I 20 I / 40 80 100 ~ 1/(Glycerol] , M-l 60 / 120 140 160 ~ -..J 48 Figure 3 A plot of ln(k.i,o b) s as a function of 1/T f or i nactiva tion of diol dehydratase by gl ycerol . Un l e s s otherwis e indicated, the error bars are ap proximately the same s ize as the illustrated data points . 49 1.0 ~ 0 I'll .g ~o ~ - ~ 'M ~ s::: -1.0 ..-4 o""' ¢~ § -2.0 3.2 3.3 3.4 1/T, oK-1 X 10 3 Figure 3 3.5 50 1_,_ 2-Propartediol. 1, 2-Propanediol protects against glycerol inactivation in a manner indicative of simple competition between glycerol and 1,2-propanediol for the active site. The kinetics which obtain are analogous to those of simple competitive inhibition, with 1,2-propanediol as the "inhibitor" of inactivation. Consider: G! E + E where G = glycerol, + k. EG ➔ 1 S t. ES ➔ E I + P s = 1,2-propanediol, I = inactive en- zyme, and ki is the first-order inactivation rate constant. Using the same treatment just described for the substratefree case to allow for change of [EG] with time, one can show that the inactivation kinetics remain first-order in the presence of 1,2-propanediol, and the amount of active enzyme at any time is given by [E ] = a [E] e a o -k. i,obs (t) where k.l.,O b S = (7) 1 + Values of k.l.,O b S in the presence of 1,2-propanediol were determined experimentally at various concentrations of glycerol and 1,2-propanediol as previously described. The 51 resulting double-reciprocal plo t s (1/k 1.. , 0 b S· vs 1/[G]) at each fixed 1,2-propanediol concentration are linear wi th a common y-intercept, illustrating the pur ely competitive nature of the "inhibition" of glycerol inactiva t ion by 1 , 2propanediol (Fig. 2). Furthermore, Fig. 2 also demonstra tes that eq 7 is a valid description of the r e sults obtained over a wide range of substrate and glycerol concentration s. The value of ki obtained from they - intercept of Fig . 2 agrees well with the experimentally determined value (Table I) and lends further validity to this treatment. Thus, the above equation allows what we believe to be the first reliable determination of the Michaelis constan t for 1,2-propanediol and diol dehydratase (Table I) . The S isomer of 1, 2-propanediol is more effective at inhibit L 1g inactivation than is the RS mixture, while (R) -1, 2-proparn> diol is a much less effective inhibitor than the RS mixtu.,·e. That the Rand S isomers differ in their ability to prote c t diol dehydratase from glycerol inactivation reflects a difference in their respective values of~ (Table I) . The values of Km obtained in this way are really K1 . How- ever, the ratio of K1 values for (R)- and (S)-1,2 - propane·- diol determined by this method agrees well with the value of 3.2 reported by Jensen & Neese (1975 ) and suggests that in this case K1 = ~Kinetics of Inactivation - Other Substrates. strate analogs listed in Table I, except~- and All s ut - 52 - dl-2,3-butanediols, exhibit inactivation behavior wh i ch can b e described by the abov~ equations deve l oped for i nactivation by glycerol in both the presence and absence of 1 , 2Values of ki and Kr or~ for each of thes e propanediol. inactivators were determined for r e ason s of c onvenien ce by measurement of the rate of inactiva tion i n the pre sence of 1, 2-propanediol and are listed in Table I. d -· and dl - 2, 3- butanediols act as purely competit i ve inhibi to r s of ina c t ivation and catalysis by other substrates. Deuterium Isotope Effects on the Rate of Inact i vation ~ Glycerol. The rates of enzyme inactivation at satura- ting concentrations of (R)-gly-d 2 , (RS)-gly - d 2 , (S) - gly -· d;;; , and gly-d 5 were measured by the method just described, and were considerably slower than that of the undeuterated substrate. The ratio of the rate of inactivation observe o for glycerol to that observed for a saturating concentr a tion of one of the deuterated glycerols reflects an isotor, e effect on ki. Thus, for rates measured at large concen- trations of glycerol and its deuterate d derivatives, Table II lists the various kinetic p ar ameters obta ined in this way. The k. b 1.,0 S values determined in these experiments should be within 10 % of the true ki values (becau s e a t 53 Table I Kinetic parameters of substrates and inactivators of diol dehydratase. 8values of K1 for the natural substrates were determined in competition with glycerol with the natural substrates serving as "inhibitors" of inactivation as described in Results. bToraya et al. (1976) have also re- ported glycerol to be a substrate and inactivator of diol dehydratase. A value of KG for glycerol was given that agrees with this value. cProduct formation does occur, but the rate was not quantitated. dDetermined from the amount of product formed before complete inactivation . eDirect comparison of the initial rate to that of 1,2-pro panediol. fBased on a molecular weight of 250,000 and a specific activity of 60 unit/mg with 1,2-propanediol as substrate. dl isomers. gCalculated from the K1 values of the~- and (S)-1,2-Propanediol (RS)-1,2-Propanediol (R)-1,2-Propanediol Glycerol 3-Chloro-1,2-propanediol Trifluoropropanediol Thioglycerol Ethylene glycol 1 , 2-Butanediol Isobutylene glycol meso-2,3-Butanediol ~-2,3-Butanediol dl-2,3-Butanediol ~-2,3-Butanediol Substrate Table 4.1 1. 24 (KI) 0.1 6.6 5.9 (KI) 8.8 2.0 (KI) 5.5 (KI) 1.07 (K 1 ) 0.59 (K )g 1 0.19 (K1 )a 0.30 (K )a 1 0.60 (KI) a 16b Km x 10 4 M I 0 0 0 150d,e >Oc 18d 4d 1\,0 191 250 340 412d >Oc >Oc k£(sec-l)f 0 0 0 1.30 0.36 0.016 1. 74 0.098 0.054 0.47 0 . 068 0 0 0 k.1. (min- 1 ) V, ~ 412 52 190 61 720 __L --- - - k (sec-l)b 1 8 2.2 6.8 0.6 (Kn)P ~ 1.25 0.089 0.66 0.69 0.77 (min-l)c k.1., 0 b S crGl = 0.04 M. 0:, bCalculated from eq 8 using k.1. and P /E 0 . aCompetitive turnover numbers based on a specific activity of 60 unit/mg and a molecular weight of 250,000 (Essenberg et al., 1971). 19,000 33,200 16,300 5,300 51,000 Competitive Turnover No. (I_=_ _§_O min) a Catalytic and Inactivation Rate Constants for Glycerol and its Deuterated Analogs. Glycerol Glycerol-d 5 (RS)-Glycerol-d 2 (R)-Glycerol-d 2 (S)-Glycerol-d 2 Substrate. Table II. 1 14 1. 9 1.8 1. 6 D 1. kH (E). V1 V1 56 o.04 M inactivator with KG= 1 . 6 mM , KG/[G] ~0 . 04 << 1, see eq 6) and can therefore represent experimentally determined values of ki for each deuterated glycer ol. With gly-d 5 , we observe a large isotope effect (14) on ki, and smaller, nearly equal values for (R)-gly - d 2 (1 . 8) , (RS)-gly-d 2 (1.9), and (S)-gly-d 2 (1 . 6) . Product Formation from Glycerol . Diol dehydratase holoenzyrne also catalyzes the dehydr at i on of glycerol to 13-hydroxypropionaldehyde in addition to e ffecting irrever -· sible inactivation. 13-Hydroxypropionaldehyde was identi - fied (1) by the 1H NMR of its 2,4-dinitrophenylhydrazone derivative, (2) by comparative silica gel thin-layer chromatography of the 2,4-dinitrophenylhydrazone with an authentic sample prepared by the method of Hall & Stern (1950), and (3) by conversion to acrolein, which was then determined by the acrolein-specific assay of Circ l e et al . (1945). Turnover Number. The total amount of S-hydroxypropion- aldehyde formed relative to the amount of enzyme initially present was determined for glycerol, (R)-gly-d 2 , (RS)gly-d 2 , (S)-gly-d 2 , and gly-d 5 as substrates . A known amount of diol dehydratase was allowed to react for 60 mir with saturating amounts of each of the above glycerols at 37° C. No further aldehyde was produced during an addi - tional 60 min period, indicating that complete inactivatic n of the enzyme had occurred during the initial 60 min 57 period. The total amount of s ~hydroxypropionaldehyde formed was determined by the acrolein specific assay and the re sults are presented in Table II as "competitive turnover numbers," the average number of molecules of product formed by each molecule of holoenzyme before its inactivation . Although both (R)-gly-d 2 and (RS)-gly - d 2 yield less pro duct than glycerol itself before inactivating holoenzyme, gly-d 5 and (S)-gly-d 2 form considerably more product than glycerol. Rate of Formation of ~- Hydroxypropionaldehyde. The production of s-hydroxypropionaldehyde was followed for several minutes with glycerol, gly-d 5 , and (R)-gly-dz each serving as substrate. The results in Fig. 4 show that, although the initial rate of product formation is much greater for glycerol than for gly-d 5 , the rate at which glycerol inactivates the enzyme relative to inactivation by gly-d 5 is greater by an even larger factor , thus resulting in a net enhancement of product formation from gly-d 5 . Because of competition from irreversible inactivation, the rate constant for product formation (kp) cannot be obtained with any degree of confidence by extrapolation of the results in Fig. 4; the rate changes too drastically near t = 0. However, these values can be calculated if one recognizes that inactivation is a first-orde r p r ocess in [EG], as previously shown. The rate constant for product formation (kp) is the 58 Figure 4 Production of $-hydroxypropionaldehyde as a function of time for: A., glycero 1; Ill , gly-d 5 ; ~, (R) -gly-d 2 . ~-Hy-droxypropionaldehyde was determined by the DNP assay as described and is depicted here as the competitive turnover number, P.T/E 0 , for uniformity . E0 based on 60 unit/mg and a molecular weight of 250,000 g/mole (Essenberg et al., 1971). The uncertainty in the values of P1/E x 10- 3 is 0 approximately± 1.5. w >< ~ 0 0. I-' ....... I t"'l 0 5 I 10 15 20 25 I 1.0 2.0 3.0 Figure 4 5.0 minutes 4.0 ✓ 6.0 7.0 8.0 V'I \0 60 analog of the inactivation rate constant ki and d[P]/dt = k [EG]. p Under conditions whe r e the enzyme is saturated with substrate, As before, [Ea] is the concentra t ion of tota l active enzyme at any time, a quantity which continuous l y decreases becaus F of the inactivation process. [E ] a From eq 5 , = whence d[P]/dt = k p [E O ]exp(-k.t). 1. The total amount of pro - duct formed at any time is given by the integrated form of this equation: k -k.t e 1. ) if(l 1. where [PT]/[E 0 ] is the competitive turnover number at any time. At large t, the expression reduces to k = t!- (8) l. where [P ]/[E0 ] is the total amount of product formed rel00 ative to the initial amount of enzyme after complete inactivation of the enzyme. The values of [P ]/[E 0 ] have been de t e r mined as have 00 those of k. for glycerol and each of its deuterated der i 1. vatives (Table II) so that values of kp can be calculated ; they are also listed in Table II. Indeed, one can plot 61 [PT]/[E 0 ] as a function of time using the values of ki and k p in Table II and accurately reproduce the experimental graphs of Fig. 4. Deuterium Distribution in Products. The 2,4 - dinitrophe - nylhydrazones of s-hydroxypropionaldehydes formed from the variously deuterated glycerols were pur i fied by thin-layer chromatography on silica gel with chloroform as the eluting 1 solvent. The H NMR spectra of these derivatives showed that (R)-gly-d 2 gives product with essentially no deuterium at C-3, while (RS)-gly-d 2 and (S) - gly-d 2 give product with approximately 60% and at least 90% deuterium at C-3, resp ;;_,c tively. Product Formation - Other Substrates. With the exception of 1,2-butanediol, thioglycerol, and~- a.nd dl-2,3 - butane .. diols, the substrate analogs listed in Table I gave easily detectable amounts of product. Products from isobutylene glycol and meso-2,3-butanediol were identified as isobuty1aldehyde and 2-butanone, respectively , by thin-layer chro - matography of the 2,4-dinitrophen.ylhydrazones on silica gel with chloroform as the eluting solvent. Aldehydes from 3,3,3-trifluoro-l,2-propanediol and 3-chloro-1,2-propanediol were detected by the DNP assay but were not rigorously characterized. Spectral Observation of the Holoenzyme-Glycerol Complex. Figure 5 compares the visible spectrum of AdoCbl bound to diol dehydratase in the presence of glycerol to that 62 obtained in the presence of 1,2 - propanediol. Incubation 11 of 1,2-propanediol and holoenzyme produces a Cbl - like spectrum. In contrast, incubation of holoenzyme and gly - cerol produces an alkylcobalamin-like spectrum very simila r to that of enzyme-bound AdoCbl in the absence of substrat,i. This spectrum, which is obtained after complete inactivation by glycerol, remains unchanged for up to 24 hr in the dark at room temperature; after this time, slow formation of OH-Cbl can be detected. As the half-life for inactiva- tion of the enzyme by glycerol under these conditions is about 30 sec, single wavelengths at 485 nm (Amax for Cbl 11 ) or 525 nm (Amax for AdoCbl and other alkylcobalamins) were monitored immediately following addition of glycerol; no changes were detected. Even though l,l-dideuterio-1,2-propanediol reacts more slowly by a factor of 13 than 1,2-propanediol, use of the deuterated substrate does not measurably affect the amount of enzyme-bound AdoCbl observable as Cbl II . Because gly-d 5 is a reasonably good "substrate" ' but inactivates much more slowly than glycerol, we looked for spectral changes when gly-d 5 instead of glycerol was added to ho l oenzyme; none were observed. Thus, with either form of glycerol , we could see no evidence for Cbl 11 . However, Cbl 11 can be generated by addition of 1 , 2-p r o - panediol after addition of glycerol. If the addition of 1,2-propanediol almost immediately follows that of 63 Figure 5 The effects of glycerol and 1,2- propanediol on the uv-visible spectrum of enzyme-bound AdoCbl . (1) Diol dehydratase, 37 units; AdoCbl, 2.3 nmole; K2HPo 4 , 10 mM ; volume 0.3 ml. (2) 60 µmole glycerol added in a 0.03 ml aliquot to sample and reference of (1); volume 0.33 ml. (3) 18 µmole 1,2-propanediol added to sample and reference of (1); volume 0.33 ml. T = 27 0 C. Spectrum (1) has been displaced vertically somewhat for the sake of clarity. 64 0.10 0.08 0.06 0.04 0.02 400 500 wavelength, nm Figure 5 600 65 glycerol, a large fraction of enzyme-bound AdoCbl is ob11 served as Cbl . If, however, the enzyme-glycerol mixture is allowed to stand for 4 min, the addition of 1,2-propanediol does not generate any Cbl 11 in the spectrum . TCA precipitation of the protein from a solution containing holoenzyme and glycerol leads to recovery of only OH-Cbl. Addition of TCA to an identical reaction mixture containing 1,2-propanediol instead of glycerol results in nearly complete recovery of AdoCbl. In further attempts to isolate and characterize the cobalamin species formed during inactivation by glycerol, inactivated holoenzyme was denatured by heat; 5 M guanidinium chloride, pH 8; 8 M urea solution, pH 4; or digested with a-lytic protease. In each case, only OH-Cbl was recovered. Exposure of a holoenzyme-glycerol solution to a 100 watt bulb at 18 in. for 30 min resulted in nearly complete formation of enzyme-bound OH-Cbl. Spectral Observations of Enzyme-Bound OH-Cbl. Al- though the general character of the OH-Cbl spectrum remains unchanged on binding to diol dehydratase, small bathochromic shifts of the bands above 300 nm do occur and serve to distinguish free OH-Cbl from the enzyme-bound species (Fig. 6). OH-Cbl also binds to bovine serum albumin, for example, but the resulting electronic spectrum is identical to that of free OH-Cbl. For OH-Cb~ in 10 mM K2HP0 4 buffer at pH 8, we observe ·:trnax A. at 525, 510, and 352 nm, while 66 Figure 6 The absorption spectra of 0H-Cbl in solution and bound to diol dehydratase. 10 mM; volume 0.32 ml. (--): 0H-Cbl, 1.4- nmole; K HP0 , 2 4 (----): apoenzyme, 30 units; 0H-Cbl, 1.4 nmole; K2HP0 4 , 10 rnM; volume 0.32 ml. The reference cell was identical to the sample cell but lacked 0H-Cbl. T = 27° C. 67 0.10 I I I I I I I \ I I I i I 0.05 I I I I I f I I \ \ I I I I I \ I \ / I -... -, \ \ \ ' \ I \ \ '' 300 400 500 wavelength, nm Figure 6 600 68 for OH-Cbl bound to diol dehydratase under identical conditions, the corresponding values are 540, 515, and 362 nm, respectively. Spectral Observations - Other Substrates. UV-Visible absorption spectra of holoenzyme-substrate complexes with ethylene glycol, trifluoropropanediol, isobutylene glycol, and thioglycerol are nearly indistinguishable from that observed in the presence of 1 , 2- propanediol (Fig. 7a). In the case of ethylene glycolp after about seven half-lives the spectrum undergoes a rapid change indicating formation of OH-Cbl; at this time, only about 0.8 % of the original activity is present. With trifluoropropanediol, isobutyl-ene glycol, and thioglycerol, the Cbl 11 spectrum persists for hours after complete inactivation of holoenzyme . The uv-visible spectra of holoenzyme complexes with meso- and dl-2,3-butanediols are shown in Fig . 7b. With meso-2,3-butanediol no Cbl 11 spectrum is observed; this substrate induces conversion of an AdoCbl spectrum to one resembling that of OH-Cbl at a rate comparable to that of inactivation. dl- and ~-2,3-butanediols, in contrast, elicit no change in the spectrum. Furthermore, exposure of a solution containing the holoenzyme-~1-2,3-butanediol complex to a 60 watt bulb at 10 cm for 15 min causes less than a 15 % increase in A360 due to OH-Cbl formation, compared to a ~40 % increase observed in the absence of substrate. Addition of 1,2-propanediol to a solution 69 Figures 7a and 7b 7a: The effect of several substrate analogs on the ab ·sorption spectrum of enzyme-bound AdoCbl. ( - ): apoen- zyme, 30 units; K 2HPo 4 , 10 mM; AdoCbl, 1.8 nmole; volume 0.35 ml. Reference cell was identical but lacked AdoCbl. To the sample and reference cuvets was added 0 . 05 ml of 0.5 M solutions of ethylene glycol (- - ); thioglycerol (· ·); trifluoropropanediol (-· · ); or isobutylene glycol (- · -) _ Volume 0.4 ml, T = 27° C. 7b: Absorption spectra of holoenzyme-2,3-butanediol complexes. (a) dl-2,3-butanediol; apoenzyme, 100 units. (b) meso-2,3-butanediol; apoenzyme, 80 units. The sample and reference solutions were identical except for AdoCbl, and contained 20 mM K 2HP0 4 , pH 8. Spectrum (b) has been displaced vertically for the sake of clarity. mer gave the same results as the dl. 37° C. The d iso - Vblume 1 . 1 ml, 70 QJ () g -e 0 (/) ~ ~ •r-1 - -- ,I.) Cd ,-I QJ ~ ' L 350 550 450 wavelength, nm Figure 7a 71 .08 (I) 0 ~ ,e .06 0 Cl) ~ (I) > •.-l .µ <II .04 b ~ (I) ~ .02 a 350 400 500 450 wavelength, nm Figure 7b 550 600 72 containing holoenzyme and~- or dl - 2,3-butanediol produces a peak at 480 nm characteristic of Cb l 11 , while identical treatment of a solution containing holoenzyme and the meso substrate does not detectably alter the observed OH-Cbl spectrum. Recovery of Active Apoenzym<:, from Inactive Apoenzymeoil-Cbl Complexes. Dialysis of an enzyme-OH - Cbl complex against a so 1ution containing Mg 2+ and so 2- leads to re ·~ -- 3 covery of OH-Cbl in the dialysate. Subsequent dialysis to remove SO~- (which is apparently a potent inhibitor of diol dehydratase) leaves apoenzyme, which upon addition of AdoCbl is catalytically active (up to 90 % of the original activity can be recovered). Treatment with either reagent alone or both sequentially does not remove enzyme-bound OH-Cbl. Table III gives details of this procedure, which is a modification of a procedure for exchanging OH - Cbl and AdoCbl with glycerol dehydratase (Schneider et al., 1970) , and some representative results. Reactivation of ~ 2 -Inactivated Holoenzyme. Exposure of diol dehydratase holoenzyme to oxygen in the absence of 1,2-propanediol leads to inactivation and formation of enzyme-bound OH-Cbl; this reaction can be followed spectrophotometrically. OH-Cbl can be removed from oxygen-inactivated holoenzyrne by the above method; however, only very little activity is regenerated. Continued dialysis of the resulting apoenzyme 73 Table III Reactivation of several forms of inactivated diol dehydratase. aReaction solution 1fa3 contained O. QL~ M glycerol , otherwise the four reaction solutions were identical and con- tained: K2HP0 4 buffer, pH 8, 40 mM ; bovine serum albumin, 0.02 mg; diol dehydratase, 17 units. Volume 2.1 ml . Enzyme activity was determined by assaying a 0.01 ml aliquot for activity as described in Experimental . binactivation was carried out by addition of: 40 µg OH-Cbl to solution ://1; 40 µg AdoCbl to solutions #2 and #3, and 0.04 ml H20 to solution #4, followed by 15 min incubation in the dark at 37° C. An aliquot was removed for determination of activity. cReactivation: Treatment A. After inactivation, each of the four solutions was subjected to dialysis for 12 hr against each of two changes of 2 1 of solution to which had been added: K2so 3 , 100 mmole; Mg(CH 3co 2 ) 2 , 60 mmole; 1,2-propanediol, 200 mmole; and K2HPo 4 , pH 8.0, 20 mm.ale. rhis was followed by dialysis for 3 hr against 2 x 2 1 of a buffer containing 0.1 M 1,2-propanedi ol and 10 mM K2HP0 4 buffer , pH 8.0. dTreatment B refers to repetition of treatment A. OH-Cbl 02 Glycerol Control 1 2 3 4 Reactiona Method ofb Solution Inactivation III 15 17 17 Table 2.0 0 17 2.8 0 17 13 11 4.0 4.6 9 Afterc Afterd Treatment A Treatment B 0 After Inactivation 17 Before Inactivation Enzyme Activity, Units --.J +:-- 75 does lead to recovery of up to 30 % activity (Table III). The recovery of active apoenzyme was not enhanced by allowing o2 -inactivation to occur in the presence of sodium di ·thionite, dithiothreitol, or potassium borohydride, nor by the presence of these agents in the dialysis solution at concentrations of 2 mM. Reactivation of_ Glycerol-Inactivated Holoenzyme. In the presence of light, the cobalamin of the inactivated holoenzyme is converted to OH-Cbl, which, as in the case of oxygen inactivation, can be removed without simultaneous recovery of enzymatic activity. In the absence of light, nruch longer periods of dialysis are required even to remm:,~ the enzyme-bound cobalamin; again, significant enzymatic activity is not recovered. The presence of the three re- ducing agents listed above had no effect in this case, either. Inactivation £Y_ [l- 14 C]-Glycerol. When diol dehydrat~se holoenzyme is inactivated by [l- 14 C]-glycerol, up to 40 or 50 radioactive molecules become str ongly associated with each protein molecule in an AdoCbl-dependent process. Con- trol experiments using apoenzyme alone or OH-Cbl-inactiva-· ted enzyme also give evidence of labeling, although only to the extent of about one or two l abels per enzyme molecule. Experiments in which the number of enzyme active sites exceeded the amount of AdoCbl by factors of two to seven also yielded approximately 40 - 50 radioactive 76 labels per protein molecule. Accordingly, we conclude that virtually all of these labels are due to s-hydrox.ypropionaldehyde, the product of catalysis, and may arise via Schiff's base formation involving nucleophilic groups on the protein. The [ 14 CJ labels could not be removed from the protein by gel chromatography; dialysis against 5 M guanidinium chloride, pH 8; 8 M urea solution, pH 4; or 0 . 5 M hydroxylamine, or by TCA precipitation of the protein. In attempts to reduce the number of labels, several strategies were tried. The inactivation reaction was carried out in the presence of nucleophiles such as imidazole or dithiothreitol (these do not significantly affect the catalytic properties of diol dehydratase). The reaction was carried out in the presence of KBH 4 in an attempt to reduce S-hydroxypropionaldehyde irrrrnediately upon its release from the active site. Alternatively, apoenzyme was preincubated with propionaldehyde or acetaldehyde prior to addition of AdoCbl and radioactive glycerol. Unlabeled s-hydroxypropionaldehyde was also generated in situ by adding a substoichiornetric amount of AcloCbl to an apoenzyme-glycerol solution; following completion of th:ls reaction, [l- 14 CJ - glycerol and excess AdoCbl were added. We have observed that the presence of these labels does no t detectably alter the catalytic properties of residual, active diol dehydratase. This approach was most successful, 77 resulting in up to a ten-fold reduction in the amount of label associated with protein , ~-Gel Electro;ehoresis of diol dehydratase which has been inactivated by oxygen and glycerol does not reveal changes in the four-band pattern which has been described by Poznanskaya et al. (1977). Preliminary isoelectric focussing data suggest that the pI of diol dehydratase, in the neighborhood of 4 - 4.5, is not measurably changed by inactivation by oxygen or glycerol. DISCUSSION Mechanism of Inactivation and Catal~sis. The following mechanistic scheme has evolved for the conversion of 1,2propanediol to propionaldehyde by diol dehydratase (Abeles & Dolphin, 1976): SH + AdoCbl-E ~ SH-AdoCbl-E ~ S · HA<l -Cbl (I) ·Ad -Cbl (III) (II) PH + AdoCbl-E (IV) Intermediate I represents the reversibly associated enzymesubstrate c omplex; II , an intermediat e in which hydrogen ha.1 been removed from substrate to yield 5'-deoxyadenosine (HAd); III represents one or more intermediates involving 78 rearrangement; IV is a species in which hydrogen has been returned to substrate from the 5' carbon of 5'-deoxy- adenosine, presumably completing formation of product. The conversion of I to II is thought to be irreversible; attempts to demonstrate hydrogen exchange between cofactor and unreacted substrate have been unsuccessful (Carty et al., 1971; Frey & Abeles, 1966) . Four areas of experimental results of this work can be used to assess possible molecular mechanisms of inactiva tion and catalysis: (1) kinetics of catalysis and inactiva- tion, (2) kinetic isotope effects, (3) spectral observatio1s of enzyme-bound cofactor before and after inactivation, (4) stereospecificity of catalysis and inactivation. These four aspects were studied in considerably greater detail with glycerol as substrate inactivator. However , results obtained with this substrate may also apply to the others, in view of their similar kinetic behavior. • Kinetics of Catalysis and Inactivation. Equations of the Michaelis-Menten type, modified to account for decreasing enzyme activity with time due to irreversible inactivation, accurately describe all catalysis and inactiva tion kinetics presented here. The kinetics of inactivation for all substrates studied are similar in several respects : (1) the rate of inactivation is first ~order with respect to enzyme, (2) saturation behavior with respect to substrate is observed , and (3) 1,2-propanediol acts as a pure ly 79 competitive inhibitor of inac tivation . These similarities suggest that the mechanism of inactivation may be the same or very similar for each inactivator . Furthermore, they indicate that inactivation is a phenomenon associated with the active site . Kinetic Isotope Effect. Results obtained with gly-d 5 • H/V . (Vmax maxD ~ 8) demonstrate that cleavage of a carbon- hydrogen bond is an important rate - contributing step in the enzyme-catalyzed formation of product from glycerol. This suggests that the mechanism of catalysis is probably very similar to that of the normal substrate, 1, 2-propanediol. The isotope effect of 14 on inactivation suggests that the mechanism of inactivation may in part be very similar to that of catalysis. However, we have verified that s-hy- droxypropionaldehyde neither inhibits nor inactivates diol dehydratase holoenzyme. Inactivation, then, must occur at some point between initial binding of glycerol and the release of product. Spectral Observations of Enzyme-Bound Cofactor . Sub- strate-induced transformation of enzyme-bound AdoCbl to a species spectrophotometrically indistinguishable from CblII has been observed to occur with both diol dehydratase and ethanolamine deaminase (Wagner et al., 1966; Abeles & Lee , 1964; Babier, 1969). This transformation is asso- ciated with catalysis and presumably involves hemolytic cleavage of the carbon-cobalt bond. 80 A number of the modified substra te- inactivators we have examined can induce the transformation described above in the uv- visible spectrum of enzyme -bound AdoCbl, suggesting the presence in these cases of intermediate II (Cbl 11 ). The absence of a detectable Cbl 11 species with glycerol and - meso-2,3-butanediol as substrates therefore more likely represents a change of the rate-limiting step, resulting 11 in the predominance of an intermediate which is not Cbl , rather than a change in the catalytic mechanism (i.e., a CblII species still intervenes, but in a concentration too low to be detected by the techniques used). The inactivation reaction may be the result of the formation of a highly reactive intermediate such as II or III which, due to modification of the substrate, lacks the proper electronic or steric configuration necessary for normal catalysis. Destruction of cofactor and/or enzyme by some noncatalytic reaction of the above intermediate may be responsible for inactivation. Destruction of enzyme-bound cofactor is commonly associated with inactivation of AdoCbl-dependent enzymes and is usually characterized by formation of OH-Cbl. For example, when enzyme and AdoCbl are incubated together in the absence of substrate, formation of enzyme-bound OH-Cbl from enzyme-bound AdoCbl occurs at a rate apparently equal to that of inactivation (Wagner et al., 1966), probably by 11 reaction of o with Cbl 11 . 1be formation of Cbl in this 2 81 case cannot be observed spectrophotometrically, but the presence of a small amount of some paramagnetic species has been demonstrated even in the absence of substrate (Finlay et al., 1973). Of the inactivators listed in Table I , only with meso2,3-butanediol is inactivation accompanied by transformation of enzyme-bound cofactor to enzyme-bound OH-Cbl at a rate similar to that of inactivation. With this inactiva ... tor, as in the case of oxygen inactivation, formation of CblII cannot be detected spectrophotometrically. (In con- trast, the~ and dl isomers (Fig. 7b) do not alter the spectrum of enzyme-bound AdoCbl. The mechanistic signifi- cance of this result will be discussed in the next section-) However, of the inactivators that do induce formation 11 of Cbl 11 , none exhibit a transformation of the Cbl spec · trum to that of OH-Cbl concomitantly with inactivation. The formation of enzyme-bound OH-Cbl follows more rapidly after inactivation with ethylene glycol as substrate than with any of the other inactivators that induce CThl 11 form&tion. This may be the result of a diminished ability of ethylene glycol, the smallest of these substrates, to pro• • t ect t h e inactive enzyme- CblII comp 1 ex f ram oxygen. Glycerol is unique in that little change in the visible spectrum of enzyme-bound AdoCbl can be detected, yet glycerol is both a very good substrate and powerful 82 inactivator. (Slight changes in the spectrum from that ob- served for AdoCbl bound to enzyme in the absence of substrate or inactivator appear in Fig. 5. The shoulder at 370 nm characteristic of an alkylcobalamin is diminished slightly, while a new peak at 363 nm appears.) Initially , the spectrum probably represents enzyme-bound AdoCbl. Fol - lowing inactivation, however, the uv-visible spectrum must represent either a "trapped" intermediate or a new alkyl or thiocobalamin not normally observed during catalysis. AdoCbl-dependent rearrangements have been postulated to take place via a transalkylation reaction (intermediate IIIa is presumably formed from intermediate II), although no concrete evidence has yet been obtained for such a reaction occurring with any AdoCbl-dependent enzyme. The uv-visible spectrum of glycerol-inactivated holoenzyme may be due to a trapped intermediate such as IIIa or IIIb. OH OH I I R - CH - CH I >CO< Illa -+ 9H R - CH - CHOH I >CO< IIIb IIIb is more likely in view of the observed lability of the intermediate, secondary alkylcobalamins being generally less stable than primary alkylcob al amins (Hogenkamp, 1975). However , the above hypothes i s does not account for the observed inactivity of apoenzyrne upon removal of cobalamin and the apparent covalent labeling of the enzyme by 83 glycerol. In view of these results, a more likely scenario is attack of a reactive intermediate derived from substrate ,:such as II) on the enzyme, followed by or concomitant with 1 reaction between Cbl II and an amino ac i d residue at the active site. This would account for the failure to re- move the light-sensitive cobalamin intact from the active site, the inactivity of apoenzyme upon removal of cobalamin, and a small amount of the observed covalent labeling of the enzyme by glycerol. The presence of a reactive apoenzyme moiety at the active site is also indicated by the observation that removal of OH-Cbl from oxygen-inactivated holoenzyrne does not regenerate active apoenzyme . This substrate-free inac- tivation has previously been attributed only to destruction of enzyme-bound cofactor which then remains irreversibly bound to enzyme (Wagner et al . , 1966). In this respect, the mechanisms of inactivation by oxygen and glycerol may be somewhat similar . A further similarity between glycerol and oxygen-inactivated holoenzyme is seen in Table III. Continued dialysis against the Mg 2 +/SO~- solution beyond that required to remove enzyme-bound cobalamin restores some activity. Moreover, the rate of restoration appears to be about the same for both the glycer ol and oxygen-inactivated apoenzymes. This reactivation after prolonged dialysis is perhaps due to slow reduction of some oxidized species on the enzyme by so 23 However, the presence of 84 dithiothreitol, sodium dithionite, or potassium borohydride did not enhance this reactivation. The above results indicate that ina ctivation almost certainly involves the breakdown of catalysis at some intermediate step. Destruction of enzyme-bound cofactor would first be suspected as a likely cause of inactivation ; however, in most cases the enzyme-bound cobalamin species we observe (alkylcobalamin or CblII) is on e either observed during normal catalysis or associated with active holoenzyme. Furthermore, removal of the cobalamin species does not regenerate active apoenzyme. Although enzyme-bound OH-Cbl does eventually result with all inactivated holoenzyme complexes, such formation of OH-Cbl is probably the result, rather than the cause, of inactivation. This indicates that the protein must play an active role in inactivation, and , by implication , an active role in catalysis. Stereospecificity of Catalysis and Inactivation. (i) Glycerol . The two hydroxymethyl groups attaced to the prochiral (C - 2) carbon of glycerol are enantiotopic-paired substituents. This makes possible two chemically distinc t diastereomeric combinations of enzyme with glycerol , designated for purpose of discussion as EG 8 and EGR. 85 EGs represents glycerol bound to enzyme with the (pro~S)hydroxymethyl group at the site invol?ed in hydrogen abstration ("A") , while EGR represents the corresponding enzyme-glycerol complex with the (pro-R)-hydroxymethyl group so oriented . In the following discussion, we shall use results obtained with the deuterated glycerols to argue that product formation occurs mainly via a pathway through EGR, while inactivation occurs via a pathway through EG 8 . If one assumes the mechanism of the enzyme-catalyzed reaction with glycerol to be the same as that with 1,2-propanediol, the minimum kinetic scheme that must be constructed is ➔ /;, EGR + G ,{/ E + EGR' ~EG EG , 8 s + +- EGR" + +- EPR + E + PR (10) ➔ ➔ + EGS" + EPS + E + PS EGX, EGX I EGX I I ' and EPx are analogous to intermediates ' I-IV in eq 9. To present the following largely qualitative argumei:1t, we shall further abbreviate the kinetic scheme above to 86 EG E + k (R) R p E + PR // k;(Ri'- I (11) ~k:7 EG S kp(S) E + PS recognizing that kp(R) and kp(S) are composite constants representing many steps, including at least two steps involving hydrogen transfer (I-II and III-IV, eq 9). Each of these two steps can contribute to the expression of k.l. (S) and k.1. (R) an isotope effect on kp (R) or kp (S ). are the rate constants for inactivation of the respective enzyme-glycerol complexes. Consider first the products formed from (R)-gly-d 2 ; four are possible: PR PS 0 0 ball C CDzCHzCH a b cl) CH 2 CHDCD OH OH I I A C 0 0 b cH CH 2 CH 2 CD ball CD 2CHDCH a OH OH C I B I D Products A and Bare derived from EG 8 , while C and Dare 87 derived from EGR. (Products Band D result from inter - molecular hydrogen transfer . That this occurs has been demonstrated by Essenberg et al. (1971) and involves C-5' of cofactor as the intermediate hydrogen carrier.) 1 H NMR of the 2,4-dinitrophenylhydrazones of the above products indicates that better than 90% of products C and Dare formed. This result is particularly striking, since products C and D should be disfavored relative to A and B due to kinetic isotope effects on product formation. Moreover, lH NMR of the 2,4-dinitrophenylhydrazones of product derived from (S)-gly-d 2 (deuterium at carbon "a") shows virtually no hydrogen attached to the hydroxyl-bearing carbon. These results constitute strong evidence that product aldehyde originates almost entirely from EGR. The observed kinetic isotope effects further support this conclusion. For example (R)-gly-d 2 and (RS)-gly-d 2 contain the same amount of deuterium. (R)-gly-d 2 , however, exhibits a large isotope effect on product formation (6.8), while a much smaller effect (2.2) is observed with (RS)gly-d 2 . This result indicates that deuterium on carbon "c" has a greater effect on the rate of product formation than deuterium on carbon "a". The inverse isotope effect on k p obtained with (S)- gly-d 2 (kH/kn = 0.6) is actually due to a perturbation of KG caused by substitution of deuterium for hydrogen at the 88 pro-(S) hydroxymethyl group (Chapter II). result is a weakening of the binding of the The apparent EG8 con- formation of (S)-gly-d 2 , with little concomitant effect on that of the EGR conformation. Thus, the catalytic EGR conformation competes more favorably for the active site of diol dehydratase. The same effect occurs with gly-d 5 , and the observed (kH/kD)p = 8 is in reality a combination of two effects: reduction of kp due to partici- pation of heavy isotope, and enhancement of the rate due to more favorable comp~tition of the for the catalytic site. EGR conformation Hence, the actual isotope effect on catalysis is probably about 12. Accordingly, the re- duction of this isotope effect to 6.8 for (R)-gly-d 2 must be due to some contribution from hydrogen located on carbon "a". The contribution may be in the form of a small amount of product formation from the EG 8 form or pos -sibly some very slow exchange between the deuterium of enzyme-bound coenzyme and hydrogen on the "a" carbon. Onl.7 a very small contribution (2--8%) from hydrogen on the "a" carbon is needed to account for the decrease of the isotope effect from 8-12 to 6.8. Analogous arguments support the conclusion that inactivation occurs mainly via EG 8 . For inactivation, glycerol- d5 shows an isotope effect of 14; this must also represent the full extent to which an isotope effect can be expressed. 89 were hydrogen at carbon "c" as important to inactivation as to catalysis, an isotope effect on inactivation much larger than the observed 1 . 8 would be exhibited by (R) gly-d 2 . Accordingly, hydrogen at carbon "a" must make a significant contribution to inactivation. Does the isotope effect of 1.8 for (R)-gly-d 2 reflect some inactivation via EGR? There is an alternative explanation. When holoenzyrne is inactivated by glycerol (no deuterium) in the presence of l,l-dideuterio-1,2-propanediol, an isotope effect is expressed on the inactivation rate (Bachovchin et al., 1978; Chapter II). Thus, deuterium located on 1,2-propanediol can participate in the inactivation reaction and contribute to the expression of an isotope effect on ki. The mechanism of this effect most likely involves the S' carbon of cofactor as the intermediate hydrogen carrier. The importance of this result to the present discussion is that a similar effect must occur with (R)-gly-d 2 . The EGR binding conformation can behave as a deuterated substrate contributing deuterium to the coenzyrne during catalysis. Moreover, that the coenzyrne should be rich in deuterium follows from the much greater reactivity of the "c" carbon deuteriums in product formation and the much greater rate of catalysis compared to that of inactivation. Thus, the small isotope effect of 1.8 with (R)-gly-d 2 as inactivator indicates inactivation predominantly via EG 8 ; little, if any, inactivation occurs 90 via EGR· These conclusions are in stereochemical agreement with results obtained with 1,2-propanediol and 3-fluoro-1,2propanediol (Eagar et al., 1975). strates, it was With both of these sub - found that the 11 R" binding conformation is more favorable for catalysis than is the "S" binding conformation. Furthermore, substitution of fluorine at c-3 of (S)-1,2-propanediol reduces catalysis by a factor of two, while the same modification of the R isomer results in almost no changes in the catalytic rate. The present observation is that a hydroxyl group substituted at C-3 of (R)-1,2-propanediol has no detrimental effect on cataly sis, while the same modification of the deleterious to catalysis. 11 S" ·. isomer is very Thus, a pattern is revealed indicating that the properties of the holoenzyme-substrate complex are such that a substituent on the methyl group directed away from the site of hydrogen abstraction is more readily tolerated for EGR than for EG 8 . (ii) 2,3-Butanediols. The hydrogen abstracted from C-1 of 1,2-propanediol during catalysis is that located "opposite" the C-3 methyl groups in the following diagram (Zagalak et al., 1966) (arrows): 91 (R)-1,2-propanediol (S) - 1,2-propanediol - Meso, ~. and ~-2,3-butanediols have the following configurations (Rubin et al., 1952). OH OH OH OH ,C--C, ~--c, OH OH I I ,C---C, H.' j j '-H H C 3 CH 3 meso I I H,. l ! '•cH 3 H C H 3 d I I H3C,' l ' ' ,. H CH 3 H l The configurations of the d and~ isomers correspond to replacement of the reactive hydrogen at C-1 of (R)and (S)-1~2-propanediol, respectively, by a methyl group. The kinetic data reported here indicate that this modifica •• tion precludes both catalysis and inactivation, which are known to require abstrac t ion of hydrogen from this carbon atom. Furthermore , the reactivity of t h e ~ substrate indicates that the mere steric bulk of a methyl group does not prevent catalysis . Accordingly, we conclude that the catalysis and inactivation r eactions with the 2,3-butanediols, and presumably other substrate analogs, both exhibit a high degree of stereoch1:mical discr i.mina tion with respect to hydrogen attached to C-1 of substr ate . 92 These results illustrate the remarkable stereospecificity of the dial dehydratase reaction., and serve as additional evidence for the assertion that the catalysis and inactivation mechanisms of these and other substrates are similar . £ and dl-2,3-butanediols form Michaelis-Menten complexes with holoenzyme, but do not alter the uv-visible spectrum of enzyme-bound AdoCbl. In contrast, binding of 1,2- propanediol and other substrate analogs to diol dehydratase is cormnonly associated with cleavage of the carbon-cobalt bond of AdoCbl to generate a species with an absorption spectrum characteristic of CblII (in a few cases, this spectrum rapidly converts to one characteristic of OH-Cbl) (Wagner et al., 1966; Eagar et al., 1975). Other substances less structurally similar to 1,2-propanediol, such as glycolaldehyde (Wagner et al., 1966) and chloroacetaldehyde (Finlay et al., 1972) behave similarly. In this respect, the behavior of dl-2,3-butanediol, with no hydrogen stereochemically available for abstraction from C-1, constitutes a notable departure from that of other diol dehydratase substrates, all of which present hydrogen for transfer to C-5' of AdoCbl. These results imply that simple binding of substrate to holoenzyme is not sufficent to effect cleavage of the carbon-cobalt bond. On this basis, we suggest that hydrogen abstraction from C-1 of substrate may be concerted with homolysis of the carbon-cobalt bond of AdoCbl. 93 §_ubstrate Specificity. Although the results presented here demonstrate that diol dehydratase holoenzyme binds and reacts with a wider array of substrates than previously thought, the specificity for 1,2-propanediol is still rather high. Inspection of the data in Table I reveals that any modification of the normal substrate results, generally, in a decreased kp, increased~• and the appearance of ki. Glycerol is somewhat unusual in having an in - creased value of~ relative to that of 1,2-propanediol, though this is offset by its large values of~ and ki. Replacement of the C-3 methyl group by hydrogen (ethylene glycol) results in a 20-fold increase in~• as well as in irreversible inactivation. Thus, the methyl group plays an important role in both binding and in preventing inactivation. Since the rate of inactivation ob- tained with ethylene glycol is independent of the presence of oxygen (as it is with all inactivators listed in Table I), the methyl group does not prevent inactivation by pro~ tecting the active site from exposure to oxygen. The methyl group must somehow play a role in the proper positioning or stabilization of the substrate for catalysis. Comparison of the effects on~• kp, and ki observed with ethylene glycol, isobutylene glycol, 2j3-butanediols, and 1,2-butanediol indicates that these parameters are independently affected by the addition or deletion of a methyl group to the normal substrate skeleton, suggesting that 94 steric factors important for binding are different from those that influence catalysis and inactivation. Substitution for hydrogen at C-3 of 1,2-propanediol by hydroxyl, thiol, fluorine, chlorine, methyl, or three fluorines substantially affects binding, catalysis, and inactivation. A quantitatively consistent picture relating these effects to the size or chemical nature of the substituent, however, does not emerge. A possible reason for this may be that each of the above ans.logs is, in reality, a mixture of two substrates, the Rand S isomers. A·;;gi'Wtl. substituent may have an entirely different effect on the binding, catalysis, and inactivation behavior of each isomer, as was shown for 3-fluoro-1,2-propanediol (Eagar et al., 197 5) and glycerol. Surprisingly, substitution of a methyl group at C-1 (meso-2,3-butanediol) does not entirely preclude catalysis, which must in this case involve abstraction of hydrogen from a secondary carbon atom, followed by rearrangement to give the corresponding ketone. With this substrate, initial transfer of hydrogen to C-5' of AdoCbl would generate a tertiary radical: OH OH I I CH 3 -CH-~-CH3 Subsequent formation of a covalent bond by this species and Cblrr. as described by Babior (1970) and Essenberg et al. (1971) and supported by model studies (Silverman & Dolphin, 95 1976), would result in a tertiary alkylcobalamin, an exceedingly unstable molecule whose formation has not been observed (Hogenkamp, 1975). Thus, the ability of meso- 2,3-butanediol to function as a substrate for diol dehydratase suggests that an intermediate with a covalent carboncobalt bond between Cbl and substrate is not obligatory for catalysis. Some results for the ethanolamine deaminase reaction have been similarly interpreted (Krouwer & Babier, 1977: Krouwer et al., 1978). Relation of Dial Dehydratase to Glycerol Dehydratase. Glycerol dehydratase is an AdoCbl-dependent enzyme isolated from a different strain of the same bacterium that produces dial dehydratase (Zagalak & Pawelkiewicz, 1962); there are many similarities between these two enzymes. These results reveal additional similarities. Like glycerol dehydratase, dial dehydratase also converts gly cerol to a-hydroxypropionaldehyde and undergoes simultaneous inactivation. Moreover, values of~ for glycerol and ethylene glycol reported with glycerol dehydratase (1.5 x 10- 3 Mand 0.66 x 10- 3 M, ; respectively (Yakusheva et al., 1974)) are similar to those obtained with diol dehydratase (Table I). However, the rate constant for inactivation of diol dehydratase by glycerol (1.3 min- 1) differs significantly from that for glycerol dehydratase (0.35 min- 1 )(Poznanskaya et al., 1972). Thus, the main kinetic difference between the two enzymes may be a ·greater rate 96 for the inactivation of diol dehydratase by glycerol t han for the inactivation of glycerol dehydratase; in other respects, these two enzymes are even more similar than previously thought. 97 REFERENCES Abeles, R.H. (1972) Proc. Robert A. Welch Found. Conf. Chem. Res. 15. Abeles, R.H., & Dolphin, D. (1976) Acc . Chem . Res.~ , 114. Abeles, R.H., & Lee, H. A. , Jr . (1964) Ann. N.Y. Acad. Sc i . 112, 695. Alfredsson, D., & Garegg, P. J. (1973) Acta Chem. Scand. 27, 724. Babier, B. M. (1969) Biochim. Biophys. Acta 178, 406 . Babior, B. M. (1970) J. Biol. Chem. 245, 6125 . . Babier, B. M. (1975a) Acc . Chem. Res. §_, 376. Babier, B. M. (1975b) in Cobalamin, B. Babior, Ed., New York, N. Y., Wiley, Chapter 4. Bachevchin, W. W., Moore, K. W.~ & Richards, J. H. (1978) Biochemistry 17, 2218. Baer, E. G. (1952) Biochem. Prep. ~. 31 . Bruemmer, J . H., & Roe, B. (1971) J. Agric. Food Chem. !2_, 266. Cambie, R. C. , Denney, W. A., & Mathai, K. P . (1972) Aust. J. Chem. 25, 1363. Carty, T. J., Babior, B. M., & Abeles, R.H. (1971) J . Biol. Chem. 246, 6313. Circle , S. J . , Stone, L., & Boruff , C. S. (1945), Ind. Eng. Chem . , Anal. Ed., 17, 259. Eagar, R. G. , Jr., Bachovchin, W.W., & Richards, J. H. (1975) Biochemistry 14, 5523. 98 Englis, D. T., & Wollerman, L. A. (1952) Anal. Chem. 24 , 1983. Essenberg, M. K., Frey, P.A . , & Abeles , R. H. (1971) J. Am. Chem. Soc. 93, 1242. Fieser, L. E., & Fieser, M. (1967) Re~ents fo:r: ~ i c Synthesis, New York, N.Y., Wiley, p. 38L~ . Finlay, T. H., Valinsky, J., Sato, K., & Abeles, R.H. (1972) J. Biol. Chem. 247, 4197. Finlay, T. H., Valinsky, J., Mildvan, A. S ., & Abeles, R. H. (1973) J. Biol. Chem. ~ . 1285 . Frey, P.A., & Abeles, R.H. (1966) J. Biol. Chem.~. 2732. Frey, P.A., Essenberg, M. K., & Abeles, R.H. (1967) J. Biol. Chem. 242, 5369. Ghangas, G. S., & Fondy, T. P. (1971) Biochemistry 10, 3204. Hall, R.H., & Stern, E. S. (1950) J. Chem. Soc., 490. Hogenkamp, H. P. C. (1968) Ann. Rev. Biochem. 37, 225. Hogenkamp, H. P. C. (1975) Cobalamin, Babior., B. M., Ed., New York, N.Y., Wiley, Chapter 1. Huff, E. (1961) Biochim. Biophys. Acta 48, 506. Jensen, R. F., & Neese, R. A. (1975) Biochem. Biophys. Res. Connnun . 62, 816. Krouwer, J. S.,& Babier, B. M. (1977) J. Biol. Chem. 252, 5004. 99 I{rouwer, J . S., Schultz, R. M. 1 & Babier, B. M. (1978) J. Biol. Chern. 253, 1041. tee, H. A., Jr., &Abeles, R. H. (1 96 3) J. B:i.ol . Chem. ;238 , 2367 . Lowry, o. H.' Rosebrough, N. J . ' Farr , A. L . ' & Randall , R. J. (1951) J. Bio l. Chern . 193 , 26.5 . McBee, F. T. & Burton, T. M. (195 2) J. Arn . Chem. Soc. 74 3022 . Mitsunobu, O., & Eguchi, M. (1971) Bull . Chem. Soc. Japan 44, 3427. Moore, W. J. (1962) Physical Chemistry, 3r d Ed., London, Prentice Hall, pp. 297-298. Pattison, F. L. M., & Norman, J . J. (1957) J. Am . Chern. Soc. 79 _, 2515 . Poznanskaya, A. A.' Korsova, T. L . , Malakhov, A. A. , Yakovlev (1972) Vopr. Med. Khim . & ~. 145. Poznanskaya, A . A . , Tanizawa, K., Tor aya , T . ' Soda , K . , & Fukui, S. (1977) Bioorg. Khim . 1, 111. Retey, J., Urnani-Rochi , A. , Seihle, J . , & Ar i goni , D. ( 1966 _1 Experientia 22, 502. Rubin, L. J., Lardy, H. A., & Fischer, H. 0 . L. (1952) J. Am. Chern. Soc. 74, 425. Schneider, Z. , Larsen, E. G., Jacobs en, G. , Johnson , B. C. , & Pawelkiewicz, J. (1970) J. Bi ol . Chem. ~~~. 3388. Schneider, Z. , & Pawelkiewicz, J. (1966 ) Ac ta Biochirn . Polen. 13, 1. 100 sedarevich, B. (1966) J. Am. Oil Chem. Soc. 44 , 381. Silverman, R. B., & Dolphin, D. (1976) J . Am. Chem. Soc. ~. 4626 . Toraya, T., & Fukui, S. (1972) Biochim. Biophys. Acta.284, 536. Toraya, T., •Shirakashi, T. , Kosoga, T. , & Fukui, S . (1976) Biochem. Biophys. Res. Commun. 69, 475. Wagner, 0. W., Lee , H. A. , Jr., Frey, P. A., & Abeles, R. H. (1966) J. Biol. Chem. 241, 1751. Yakusheva, M. I., Malakhov, A. A., Poznanskaya, A. A., & Yakovlev, V. A. (1974) Anal. Biochem. 6~, 293. Zagalak, B., & Pawelkiewicz, J. (1962) Life Sci. ~, 395. Zagalak, B., Frey, P.A., Karabatso s, G. I., & Abeles, R. H. (1966) J. Biol. Chem . 24_!_, 3028. 101 CHAPTER II HYDROGEN TRANSFER IN THE INACTIVATION OF DIOL DEHYDRATASE BY GLYCEROL 102 INTRODUCTION Until recently, propanediol dehydra t a.se, which cataly zes the rearrangement of both (R) - and (S)~l,2-pro panedio l to propionaldehyde, was thought to be i nactive toward glycerol (Toraya et al., 197 6; Bachovchin et al., 1977; Chap ter I). This behavior distinguished diol dehydrat ase from another very similar AdoCbl --• dependent enzyme , glycero A. dehydratase, which converts glycero l to s-hydroxypropionaldehyde as well as acting upon propanediol and ethylene glycol . In fact, diol dehydratase holoenzyrne does cata-· lyze the conversion of glycerol to a-hydroxypropionaldehyde. However, glycerol also irreversibly inactivates holoenzyme, thus explaining why catalytic activity toward glycerol was not detected previously (Lee & Abeles, 1963) . Both the inactivation and catalys is reactions exhibit large primary deuterium isotope eff ects (kH/kD = 14 and 8, respectively); thus hydrogen transfer is an important rate-contributing step in each case. Experiments with var • ious isotopically substituted glycerols demonst rate d tha t two diastereorneric combinations are possible between the enzyme and glycerol; the enzyme - glycerol complexe s have been designated 11 EGR" and 11 EG 8 . 11 When glycer ol is bound in the EGR complex , a hydrogen is abstract ed f r om the pro -(R) carbon and the catalyt i c rate actual ly exceeds that observed with 1,2-propanediol as substrate. When glycero l is bound in the EGs complex , hydrogen is abstracted from 103 the pro-(S) carbon and inactivation ~ rather than catalysis, follows as the dominant event (Chapter I). 1 , 2-Propanediol act s as a purely competitive inhibitor of inactivation .over a wide range of both glycer ol and propanediol concentrations. Irreversible inactivation has been observed with ethylene glycol and diol dehydratase and with glycerol and glycerol dehydratase, although the inactivat ion rates are con• siderably slower than that of glycerol i nac t ~va tion of diol dehydratase (Lee & Abeles, 1963; Poznanskaya et al., 1972). Thus, irreversible inactivation by substrate ap- pears to be a common phenomenon associated with these enzymes and a study of its mechanism may advance our understanding of the normal catalytic mechanism. Accordingly, a detailed study of the mechanism of inac tivation was undertaken. The kinetics of glycerol inac- tivation of diol dehydratase holoenzyrne in the presence of 1,2-propanediol and l,l-dideuterio - 1,2-propanediol were invest igated. The results and their bearing on the inacti ·-· vation and catalysis mechanisms are discussed. EXPERIMENTAL Enzyme Preparations. Propanediol dehydratase was ob - tained from Klebsiella pneumoni~ (ATCC 8724) as described in Chapter I and by Lee & Abeles (1963) . 1,2-Propanediol -free enzyme was prepared as previously reported in lOL~ Chapter I. ~denosilcobalamin (AdoCbl) was purchased from Sigma Chemical Co. ~ 1 -Deoxyadenosine was prepared by the intermediate syn- thesis of N-acetyl-5'-iodo-2',3 1 - isopropylidene adenosine by a previously reported method (Jahn, 1965; McCarthy et al., 1968). 5'-Deoxyadenosine was also prepared directly from 5'-deoxy-5'-iodoadenosine (Aldrich) by catalytic hydrogenat ion as described by McCarthy et al. (1968) . The materials obtained by both methods were identical. Substrates. 1, 2-Propanedi.ol, 1, 1~-dideuterio-1, 2-pro- panediol, glycerol, perdeuterioglycerol , (RS) -1,l- dideuterioglycerol, (R)-1,1-dideuterioglycerol, (S)-1,l-dideuterioglycerol, and [l- 14 C]-glycerol were obtained as described (Bachovchin et al., 1977; Chapter I). [ 1- 3 H]-gly- cerol was purchased from Amersha.m. Determination of Inactivation Rates. Inactivation rates were determined using the coupled enzyme assay (Chapter I) . Reaction mixtures in general cons is ted of: yeast alcohol dehydrogenase (ADH, Sigma), 15 units; NADH, 0.2 mM; apodiol dehydratase, 0.05 - 0 .25 unit; bovine serum albumin, 0 .12 mg; potassium phosphate buffer, pH 8, 30 mM; AdoCbl, 0.02 nu~; and the desired amount of 1,2-propanediol and/or glycerol. The total volume was 2.3 ml. The reaction mixture was incubated at 37° C with stirring. The inactivati.on reaction was started by the addition of 105 AdoCbl and stopped by addition of 0.05 ml of a 6 M 1,2propanediol solution. Enzyme activity (measured by the rate of production of propionaldehyde) was determined at e, minimum of four different inactiva tion times; the slope of the corresponding semilog plot wa s taken a.s the observe d inactivation rate constant (k 1 ,o b) for the given concenS tration of glycerol and lr2-propanediol . As in Chapter I, uncertainties in the measured :i.na.c tivation rates were in the range 3 - 8 %. Determination of Michaelis Constar1ts ~- The values of the Michaelis constants KG for the various glycerols were determined from the variation of k. J.,O b, the observed S inactivation rate, with glycerol concentration in the absence of competing inhibitor or substrate and in the presence of the competitive inhibitors d- or ~1.-2,3-butanediol. KG For data obtained in the presence of the inhibitors .· was obtained from the slope of the double reciprocal plot. KG was calculated from rate data obtained in the absence of competing sub strate or inhibitor as described in Results. Inactivation !?.l_ [l - 3 H]-~ro1_. A typical solution was composed of the following: apoenzyme , 40 units; [l _3H]-glycerol. 0 . 49 mCi/rnmole , 0 .14 2 M; K2 HPOLi. buffer, pH 8, 10 mM; AdoCbl , 0.025 mM; total volume 0.5 ml, 37° C. In control experiments the apoenzyme solution was either boiled for 2 min or inactivated by the addition of 0.01 ml 106 of 6.4 mM CN-Cbl prior to add i tion of AdoCbl . The holoen - zyme-[l - 3 H]-glycerol solution was allowed to react for 1 hr in the dark . 5' - Deoxyadenosine (0 .15 ml of a 2 mg/ml solution) was added as carrier, and 0.0 5 ml 6.4 rnM CN-Cbl was added to visualize subsequent chromatography . The sol ,J- tion was irrnnersed in boiling water fo r 2 m:J.n and then ap -· plied to a 2.6 x 16 cm column of BioGel P-2 (200 - 400 mesh) and eluted in 4 ml fractions w:Lth l: 5 mM NH /0.02 % 3 NaN buffer in the dark . The nucleoside~conta:1..ning frac 3 tions were pooled, acidified to pH 3.0 with 2 N HCl, and applied to a 1 x 9 cm column of AG50W- X2 (200-400 mesh) previously equilibrated with 1 mM HCl . with 100 The column waswash~d ml water to remove radioactive glycerol. The nucleoside was eluted with 0.1 N NH 3 , lyophilized, dissolved in a small amount of methanol, and applied to strips of Whatman 3MM paper or to silica gel plates. Pe.per chro- matograms were developed by the desc·eriding method in the solvent systems given in Table I; silica gel plates were developed by the ascending method. The nucleoside was visualized by its ultraviolet extinction and eluted from the chromatogram with methanol. The me t hanol was removed under vacuum an d the residue was taken up in l ml water and added to 10 ml scintisol complete or PCS:xylene (2:1) for radioactivity as s ay. ardization Stand- was accomplished by using [a- 3 H]- and [a- 14 CJ -toluene of known specific a c tivities. 107 Table I: Chromatography of 5 1 -Deoxyadenosine. Method/solvent Paper n-Bu0H:H 20 (43:7) n-BuOH:HOAc:H 2 0 (12:3:5) n-BuOH:HOAc:H 2 0 (4:1:5) 1 M NH40Ac:95 % EtOH:HzO (9:20:1) 0.34 0.60 0 . 77 0.75 Silica gel H20-saturated n-BuOH:l % NH4 0H sec-Bu0H:H 20:NH4 0H (50:18:7) 0.35 0.27 108 RESULTS Kinetics of Inactivation !?z Perdeuterioglycerol. The rate of inactivat~on of diol dehydratase by perdeuterio glycerol (gly-d 5 ) was determined as a function of the con centration of gly-d 5 in the absence of 1,2-propanediol under anaerobic conditions at 37° C. As wi th glycerol, the inactivation r~tes are always first-order in active holoenzyme. The concentration of gly-d 5 was assumed to be es - sentially constant throughout the inactivation rea ction, as only a negligible fraction of the ini tial gly - d 5 was converted to 8-hydroxypropionaldehyde. Accordingly, we can apply simple Michaelis-Menten equations which have been suitably modified to account for a decreasing amount of catalytically active holoenzyme (Chapter I). = [E] e a o Thus: -k. (t) 1.,obs (1) where [E] is the concentration of active enzyme present a o initially, [Ea] is the concentration of enzyme, both free and with inactivator bound in a Michaelis-Menten complex, which remains active at any time t, k.1., 0 b S is the observed first-order inactivation rate constant, given by k.l., 0 b S, k 1.. where k.:l is the true first-o r der rate constant for (2) 109 inactivation and KG is the Michaelis constant for the enz,rme-glycerol complex . J- A plot of 1/k.1.,0 b S vs l/[gly-d5 ] is linear; from the slope and intercept values, KG and ki were obtained. The KG value determined for gly - d 5 (3.9 rnH) differs from that of glycerol (1.6 mM, Chapter I) obtaine d by the same method. The value of k.J_ obtained from theyintercept is 0.085 min- 1 , indicating an isotope effect of 15 on the inactivation rate constant. This represents rea- sonable agreement with the value of 14 obtained in Chapter I by comparison of the observed rate constants for gly - d 5 and glycerol at 0.04 Min the absence of 1,2-propanediol; the discrepancy can nevertheless be quantitatively explained by the different contributions made by oxygen inactivation to ki,obs due to the different KG values. Kinetics of Inactivation !?z ill_-, ~RS)-p and (R)-1,lDideuterioglycerols. In the absence of substrate and in the presence of oxygen, diol dehydratase holoenzyme undergoes irreversible inactivation. The rate of this inacti- vation by oxygen is approximately 0 . 12 min-l (W. W. Bachovchin, unpublished results). In the presence of glycerol under aerobic conditions, the inactivation rate is given by the following: (3) where k 0 and k.1. are the rate constants for inactivation by oxygen and glycerol, respectively, and 110 = [E] + [EG] , the total amount of active enzyme pres ent a t a given time. With KG= [E] [G]/[EG]~ we have by subst itution and rear rangement [E] = [EG] [E ] a = Substitution of these expressions into eq 3 and integratio11 give e = [E] a o -k. 1.,obs (t) (4) where k.1.,0 b S = The value k 0 + k l.. of k 1,0 . b S was measured for (S)-l,l-dideuter - ioglycerol ((S)-gly-d 2 ) as a function of concentration of this inactivator (Fig. 1). The values of k 0 , ki, and KG which provide the best fit of eq 4 to these data were determined by the method of Marquardt (1963 ) to be k 0 = 0.14 min- 1 , kl..= 0 . 82 min- 1 , and KG= 3.6 mM. k . b for (S)1. ,0 S gly-d 2 was also measured in the presence of 1 mM £-2,3butanediol (KI= 0.55 mM , Chapter I); in th i s case, the inactivation kinetics are described by (Chapter I) 111 Figure 1 Double reciprocal plot for inactivation of diol dehy~ dratase by (S)-gly-d 2 in the presence of oxygen. The dotted line represents the best fit of eq 4 to these data as . -1 described in the text with k 0 = 0.14 min- 1 , k 1 = 0.82 min, and KG= 3.6 mM. The uncertainties ink.i , 0 b S are approxi- mately indicated by the size of the data points. 112 \ \ \ r--1 b 0 0 0 \ r--1 \ '' I ~ ...... N "'O I >-. (1) bO I ,,,-... •r-1 r--1 \ ' Cl) \ ~\ ....... r--1 0 0 \ \ 1./"'1 \ b \ q ' \ \ \ N sqo',:• }l / 1 r--1 ~ bO ~ 113 k.1,0 b S = 1 + From the slope and intercept of the double reciprocal plot of the resulting data we obtain ki = 0.8 min -1 and KG= 3.8 mM, in good agreement with the results described above. The value of KG for (RS)-1,1-dideuterioglycerol ((RS)gly-d 2 ) was similarly determined in the presence of 1.74 mM ~-2,3-butanediol (KI= 0.107 mM, Chapter I), and the results are shown in Fig. 2. The values of the slope and intercept gave ki = 0.69 min -1 and KG= 2.1 mM. With (R)-1,1-dideuterioglycerol ((R)-gly-d 2 ), we find -1 ki = 0.72 min and KG= 1.6 mM, the same as the KG of glycerol itself (Chapter I). Kinetics of Glycerol Inactivation. Competition with (RS)-l,l-Dideuterio-l,2-propanediol. The rate of inacti- vation of holoenzyme in the presence of a fixed concentration of (RS)-1,1-dideuterio-1,2-propanediol (diol-d 2 ) was determined as a function of inactivator concentration for glycerol (gly), (RS) -gly-d 2 , (R)-gly-d 2 , (S)-gly-d~ and gly-d 5 . A fixed concentration of diol-d 2 of 1.08 x 10 -4 M was chosen for two reasons: (1) This concentration is sufficiently large to validate the assumption that t,(S] 'v0 over the time interval during which the inactivation rates were determined. (2) This concentration is sufficiently 114 Figure 2 Double reciprocal plot of ki ,o b s vs inactivator conc en- tration for (RS)-gly-d 2 in the presence of 1. 74 mM dl-2, 3 -, butanediol. 115 6 5 4 Cl) ,0 0~ 3 '" .!ii: ...... r-4 2 1 25 75 50 l/[(RS)-gly-d 2 J , M-l Figure 2 100 116 small to allow large inactivation rates to occur at reason able concentrations of inactivators. Figures 3 and 4 illustrate the results . With gly-d 5 as inactivator (Fig. 3), the double reciprocal plot is linear over the entire range of inactivator concentration . Moreover, in this case they-intercept is nearly identica l to that obtained for the substrate-free case. Thus, eq 5, derived in Chapter I to describe the inactivation kinetics of gly in the presence of undeuterated 1 , 2-propanediol (diol) should again apply. = KG(gly) [d' lf 1 + --=-cg._l-y-:-]- (1 + K (dio) io ) (5) 1 The slope of the double reciprocal plot with gly-d 5 as in activator should be given by slope = KG(gly-d 5 ) [diol-dz] ki(gly-d 5 ) (l + K1 (dio1-d 2 )) . (Sa) Since values of KG(gly-d 5 ) and ki(gly-d 5 ) are available, a value of K (diol-d 2 ) of 3.0 x 10 -5 M can be calculated. 1 This is the same as the value of K1 for diol previously obtained in an analogous manner with glycerol as inactivator (Chapter I) and demonstrates that there is no deuterium isotope effect on K1 for the competitive inhibition by 1,2 propanediol of inactivation of holoenzyme by glycerol. In contrast to data obtained for gly-d 5 , double 117 Figure 3 Double reciprocal plot for inactivation by gly-d 5 in competition with 1.16 x 10- 4 M diol-d 2 . Slope= 0 . 22; inteicept = 11.8. 118 ,-4 I ·S:: -g N 20 9t, I ,-4 0 •.-4 9t, "9t,"' I >-. ~ bO ....., {/) .c 15 t . 0 o,-4 ~ ...... ~ 10 20 40 30 1/[gly-d 5 ] Figure 3 50 119 Figure 4 Double reciprocal plot of k.1.,0 b S in the presence of ' -4 1.08 x 10 M diol-d 2 as a function of inactivator concentration: ( 0), glycerol; ( D) (R)-gly - d 2 ; ( l:).) , (RS) -glyd ; ( •) (S)-gly-d 2 . The values depicted at infinite in 2 activator concentration represent ki ,o b s measured exper imentally at 40 mM inactivator concentration in the absence of 1,2-propanediol. are marked. Representative "error bars" 120 20 18 16 14 12 10 rt, .,0 "" I I 6 I .-1 I 2 / / / 0 /ti. I 4 . I 0 ~ ...... ,. , , I 8 /A / -~ ,,, / • ,, .,6 ,, 0 100 200 300 500 400 1/[Inactivator] Figure 4 600 121 reciprocal plots for (R)-. (S)-, and (RS)-gly-d 2 appear to become linear only at low inactivator concentrations (Fig. 4). Data for gly appear linear as plotted in Fig. 4 , however, they actually exhibit behavior similar to that of (R)-, (S)-, and (RS)-gly-d 2 : downward curvature at high in activator concentration and linearity at low inactivator concentration (Fig. 5 better illustrates this behavior, which is discussed later in detail). In the linear regions , data for gly and (R)-gly-d 2 define the same line, whereas data for (S)- and (RS)-gly-d 2 give lines with different slopes and y-intercepts. The slopes of the linear regions should be given by the appropriate form of eq Sa. they-intercept Assuming obtained by extrapolation of the linear region to infinite inactivator concentration to be the reciprocal of the respective ki, one may use the value of K1 for diol-d 2 to calculate a value of KG for each inacti vator. In each case, this value is 1.6 mM, the same as that previously determined for gly independent of the presence of 1,2-propanediol (Chapter I). Thus, while values of KG determined for (R)-, (RS)-, and (S)-gly-d 2 in the absence of 1,2-propanediol differ, these inactivators all exhibit the same apparent KG in the linear regions of Fig. 4, where concentrations of each glycerol and of diol-d 2 are such that [EG]/[ES] < 1. A plot of the observed isotope effect vs 1/[inactivator ] (Fig. 5) more informatively illustrates the variation of 122 Figure 5 Plot of isotope effect on k.1., 0 b S as a function of the re~iprocal of inactivator concentration in the presence of 1.08 x 10 -4 M diol-d 2 : (0) glycerol; ( □) (R)-gly-d 2 ; (RS)-gly-d 2 ; ( •) (S)-gly-d 2 . The isotope effects were calculated from the experimentally determined ki obs ' as described in Results. (A) -123 100 200 300 400 1/[Inactivator] Figure 5 500 600 124 The isotope effects were calculated by comparing l<i,obs · ,, measured at each concentration of a given inactiva ""i, obs tor in the presence of diol-d 2 to the k 1.,0 . b S obtained fo r the corresponding concentrations of undeuterated glycerol and propanediol. For example, the isotope effect on inac - tivation at a given concentration of (R)-gly-d 2 would be defined as: ki obs(gly)diol = k. 1.,0 (6) b ((R)-gly-d 2 )d. 1 d is the experimentally determined s 1.0 - 2 "observed" inactivation rate constant at the particular concentration of (R)-gly-d 2 in the presence of the fixed concentration of diol-d 2 and ki,obs(gly)diol is the corresponding observed rate constant for gly in the presence of undeuterated diol. Use of eq 5 and the fact that [gly) = [(R)-gly-d 2 ] and [diol] = [diol-d 2 J reduce eq 6 to = The data in Fig. 5 show that for all four inactivator .s a limiting maximum isotope effect is approached at low in activator concentration. For (S)-gly-d 2 , this value is 6.0; for (RS)-gly-d 2 , 4.3; for gly and (R) - gly-d 2 , 3 . 0. Kinetics of Glycerol Inactivation . Competition with (RS)-l,~-Propanediol. The rate of inactivation in the 125 presence of a fixed concentration of undeuterated propane diol was determined as a function of inactivator concentration for the following inactivators: (R)-, (RS)-, and (S) gly-d 2 , and gly-d 5 . The analogous experiment with undeut r:r - ated glycerol has been described in Chapter I; the kinetics are described by eq 5 with ki=l.3 min-land KG= 1.6 rnM. Figure 6 illustrates the results as a plot of the expressed isotope effect on inactivation vs inactivator concentration. the reciprocal of Isotope effects were calculat ed as described in the previous section. At low inactivator concentrations, the data of Fig. 6 indicate that a limiting minimum isotope effect is approached for (R)-, (RS)-, and (S)-gly-d 2 ; these values are 1.0 - 1.1 for (R)- and (RS)gly-d 2 ~ and 1.4 - 1.5 for (S)-gly-d 2 . At high inactivator concentrations (20 - 40 rnM) , the isotope from eq iffects for (RS)- and (S)-gly-d 2 (as calculated 5 and 6) actually exceed those measured in the absence of 1, 2-propanediol. However, if in calculating t h ,: isotope effect on k.1., 0 b S in this region the KG values obtained in the absence of 1 , 2-propanediol are used, the re sulting "isotope effects" are equal to those observed in the absence of substrate. Thus, with (S)- and (RS)-gly-d 2 (and possibly gly-d 5 ) as inactivators in the presence of diol, the apparent KG value observed at large inactivator concentrations is similar to that measured in the absence of substrate, and 126 Figure 6 Plot of the isotope effect on k.l., 0 b S as a function of the reciprocal of inactivator concentration in the presence of 1. 4 mM diol: ( 0) gly-d 5 ; ( e ) (S)-gly-d 2 ; (~) (RS)- gly-d ; ( □) (R)-gly-d 2 . Isotope effects were calculated 2 as described under Results. 127 8 7 6 .., t) Q) '+-I '+-I 5 ~ Q) p. ..,0 4 0 Cl) H 3 •+ * • 50 • 100 150 200 1/[Inactivator] Figure 6 128 declines to a value identical to that of undeuterated glycerol at low inactivator concentration. Isolation and Identification of ~'-Deoxyadenosine. When [l - 3 H]-glycerol inactivates diol dehydratase holoen -· zyme, tritium is transferred to the nucleoside moiety of AdoCbl. This nucleoside cochromatographs with authentic 51-deoxyadenosine on silica gel and Whatman 3MM paper in a number of solvent systems. The amount of tritium in the 5'-deoxyadenosine isolated after each reaction is directly proportional to the amount of holoenzyme inactivated (Fig. 7); control experiments using [l - 14 C]-glycerol or heatinactivated apoenzyme yielded no radioactivity associated with 5'-deoxyadenosine. A least squares fit of 3 H dpm in 5'-deoxyadenosine as a function of the amount of enzyme indicates that 5'-deoxyadenosine is enriched in tritium by a factor of approximately 2.1 over substrate, on a per mole basis. This calculation assumes an essentially quan - titative liberation of 5'-deoxyadenosine from holoenzyme . ·A· control experiment in which the inactivated holoenzyme was not denatured prior to chromatography on BioGel P-2 showed that the nucleoside remains bound to the protein and does not exchange with unlabeled 5'-deoxyadenosine in solution. 129 Figure 7 Radioactivity associated with 5'-deoxyadenosine from diol dehydratase holoenzyme which has been inactivated by radio labeled glycerol. (6) [ 1 - 3H]-glycerol, 0. 49 mCi/ mmole; ( O) [ 1 - 14 C] -glycerol, 3. 7 mCi/mrnole; ( D) [l - 3H]-glycerol, 0.49 mCi/mrnole and boiled apoenzyme. 130 8 7 6 ('t') I 0 .-I ~ 5 Cl) p •r-1 Cl) 0 p Cl) re, 4 <ti :>-. >< 0 Cl) re, I 3 10 p •r-1 s .g. 2 1 10 30 20 holoenzyme, units Figure 7 40 50 131 DISCUSSION The kinetics of inactivation when both inactivator and inhibitor are "homogeneous" with respect to transferrable hydrogens (i.e., undeuterated glycerol with undeuterated propanediol, or gly-d 5 with diol-d 2 ) obey eq 5, with appropriate constant values for KG , KI , and ki, for all concen .. trations of inactivator at the fixed "inhibitor" concentration. This is not true, however , for the isotopically mixed cases (i.e ., undeuterated glycerol, (RS) -, (R)-, anc (S)-gly-d 2 with diol-d 2 ; or gly-d 5 and these dideuterated glycerols with diol) as illustrated by the double recipro-cal plots (for example, Fig. 4). Possible sources for this divergent behavior may be variation of KG, KI, and /or ki with inactivator concentration. A change of Kr for diol and diol - d 2 may be considered as an explanation for the shape of the double reciprocal plots in the isotopically mixed cases. however, appears unlikely. This explanation, The inhibitor constant KI for 1, 2-propanediol is unaffected by deuterium substitution , a c, . is that of the substrate 3-fluoro-1 , 2-propanediol (Eagar et al., 1975). Moreover, the effects observed when undeu- terated diol protects against inact i vation by gly-d 5 are too large to be explained in terms of effects on K1 . Ac- cordingly, an explanation for the observed behavior must b,~ sought in effects on KG and ki . Variation of~- In the absence of 1,2 -prop anediol, 132 the values of KG, the Michaelis constant of the holoenzyrne glycerol complex, are 1.6 rnM for gly and (R)-gly-d 2 , 2.1 tnM for (RS)-gly-d 2 , and ~3.7 rnM for both (S)-gly-d 2 and gly-d 5 . These data suggest that substitution of deuterium for hydrogen at the pro-(S) hydroxymethyl group of glycero l (gly-d 5 , (S)-gly-d 2 , and (RS)-gly-d 2 ) adversely affects binding affinity, as measured in the manner described in Results, while the corresponding substitution at the pro(R) hydroxymethyl group has no measurable effect. The KG value exhibited by (RS)-gly-d 2 is the same as that which would be predicted for a 1: 1 mixture of (R)- and (S)-gly- ' d2 . Eagar et al. (]975) have analyzed the kinetics of such a two-substrate mixture; the apparent Km of the racemic substrate in terms of the respective constants for the (R)and (S)-stereoisomers is In the present case, the predicted KG for (RS)-gly-d 2 is 2.2 mM, very close to the observed value of 2.1 mM. Thus, (R)- and (S)-gly-d 2 appear to compete for the same active site of diol dehydratase. In this light, other results (Chapter I) suggest that substitution of deuterium at the pro-(S) hydroxymethyl group of glycerol adversely affects the binding of only th,i EG conformation . The isotope effect on catalysis (kH/kD) > 8 for (S)-gly-d 2 is 0.6; this substrate is converted to 133 product some 65 % faster than glycer ol itself. As we know of no precedent for such a large inver se primary deuterium isotope effect , we attribute this result to weakene d bind ing of the EG 8 binding conformat i on of (S)-gly-d 2 , with little or no concomitant effec t on that of EGR . Thus, the catalytic EGR conformation competes more effectively for the active site with (S)-gly-d 2 as s ubst rat e; the actual microscopic turnover rate probably changes very little . This effect should also occur with gly -d 5 ; the observed (kH/kD)p = 8 (Chapter I) is in reali ty a combination of two processes: reduction of k p due to participation of heavy isotope, and enhancement of the catalytic rate due to more favorable competition of the EGR conformation for the active site. Thus, the true deuter ium isotope effect on catalysis is probably about 12. Similar reasoning suggest , that the true deuterium isotope effect on inactivation (kH/kD)i may be somewhat smaller than the value of 15 observed with gly-d 5 as inactivator. In the presence of diol or diol-d 2 as the inhibitor of inactivation , the apparent KG values of (RS) - and (S) gly-d 2 change, from the values measured above when [inactivator]/[diol] is relatively large, to the same value as gly and (R)-gly-d 2 (1.6 mM) when this ratio is relativeHowever , this change is not obly small (Fig. 4, 5, 6). served when the inhibitor of inactivation is d- or dl - 2,3butanediol, purely compe ti t ive inhibitors which do not 134 undergo catalysis or inactivation. Thus, this effect is associated in some fashion with the ability of diol and diol-dz to undergo catalysis. The nature of this phenom- was further probed by measuring k.1., 0 b S of (S)-gly-d 2 as a function of inactivator concentration in the presence enon of 1,2-propanediol of an isotopic composition (2.25 % D at c-1) chosen to result in a hydrogen isotopic composition at C-5' of AdoCbl (Moore et al., 1979; Chapter III ) which is approximately the same as that generated by (S)-gly - d 2 itself. The results indicate that in this case, the ap- parent KG of (S)-gly-d 2 is 3.7 rnM, whereas in the presence of diol or diol-d 2 , it is observed to be 1.6 rnM. These results suggest that KG is probably not a simple dissociation constant; rather, it is more analogous to a Michaelis constant ("~"). E + s Consider the simple mechanism k3 ES-+ E + P . For this reaction, Km~ (k 2 + k 3 )/k 1 , and we see that the catalytic rate constant contributes to determining the value of Km. In the present case, KG is clearly influenceJ by reaction steps subsequent to bind ing in which hydrogen attached to C-5' participates. However, an explicit de- scription of this phenomenon in kinetic terms has not yet emerged. 135 Variation of ki. One can construct a mechanism whereby ki may vary by drawing upon the currently accepted catalytic mechanism and the supporting experimental data. Catalysis by diol dehydratase holoenzyme and other AdoCbldependent enzymes involves two hydrogen-transfer steps: (1) from C- 1 of substrate to C-5' of AdoCbl and (2) from C-5' of AdoCbl to C-2 of product (Abeles, 1971). Evidence for the formation of 5'-deoxyadenosine as an intermediate and evidence for the equivalence of the C-5' hydrogens, although not conclusive, are quite extensive (Babior, 1975b; Frey et al., 1967; Miller & Richards, 1969; Eagar et al., 1972). Thus, both intramolecular and intermolecular transfers of hydrogen are possible (Abeles & Zagalak, 196~). Moreover, the isotope effect on the second hydrogen-transfer step (cofactor+ product) considerably exceeds that for the first hydrogen-transfer step; thus, deuterium (or tritium) accumulates at the 5'-carbon of cofactor during catalysis. Indeed, enrichment in tritium of enzyme.- bound AdoCbl by as much as 20-fold over that in substrate has been observed during catalysis (Essenberg et al., 1971). Thus, a mechanism exists for the "i.nhibitori: of inactivatj ·:m to contribute a hydrogen (or deuterium) to the inactivatio::i pathway. Furthermore, the isotopic composition of the 5' carbon of cofactor, and thus the isotope effect expressed on hydrogen transfer from cofactor, should vary with inactivator concentration. For the "homogeneous" cases, no 136 such variation should occur, since only hydrogen or only deuterium is involved in the reaction. Thus, t he diver - gent behavior observed between the homogeneous and mixed cases can be explained in terms of the variation of k.i and requires only that a rate -contributing or ra telimiting step for inactivation occur at some point along the pathway after the initial abstraction of hydrogen from C-1 of glycerol. This step could be hydrogen abstraction either from C-5' of the resulting 5'-deoxyadenosine intermediate or from C-1 of glycerol after equilibration between this hydrogen and those attached to C-5' of AdoCbl. However, hydrogen exchange between cofactor and unreacted substrate could not be demonstrated in the dial dehydratase (Frey et al., 1967) and ethanolamine deaminase (Carty et al., 1971) reactions. In the case of diol dehydratase, ethylene glycol, a substrate which also inactivates holoenzyme was used. Thus, equilibration of hydrogen between AdoCbl and reversibly bound substr ate is unlikely. If such exchange occurs after an irreversible step such as reaction 1 above, no distinction can be drawn b etween the~~ two suggested mechanisms; both require formation of an intermediate containing three equival ent hydrogens. We shall therefore argue that not only does transfer of hydrogen from cofactor contribute to the observed inactivation rate constant, thereby implicating the presence of an intermediate such as S'-deoxyadenosine in the inact i- 137 vation pathway, but that transfer of hydrogen from this intermediate is the rate-determining step in the inactiva-tion process. Virtually all our kinetic data accor d with this hypothesis; in particular , the limit ing value of th e isotope effects observed at low inact~vator concentrations (Fig. 5 and 6) can with perhaps one excep t ion be quantitatively explained by this hypothesis. We shall consider first the "mixed 11 c ases in which diol-d 2 functions as the inhibitor of inactivation . At very high concentrations of inactivator, diol-d 2 should be effectively excluded from the holoenzyme, and the rate of inactivation should, and does, extrapolate to that observed for inactivation in the absence of dial inhibi t or (Fig . 4 and 5; Table II lists the rates of inactivation and the corresponding isotope effects ob served for each i nactivato ~ in the absence of propanediol) . As the concentration of in · activator decreases, an increasing proportion of h.oloenzyrr, ~ interacts with diol-d 2 , and enzyme - bound AdoCbl becomes increasingly rich in deuterium. Thus, the inactivation rate constant reflects an increas ing isotope effec t (Fig. 5) . Figure 5 also illustrates that, as i nactivator concentration decreases, the observed i s o tope effe c t reaches a limiting maximum value fo r each inactivator, a beh avior consi s t ent with our hypothe sis. Considering the much larger isotope effect on the second hydrogen-transfer steJ 138 (cofactor+ product) one might expect that, as the concentration of inactivator relative to inhibitor is decreased, a point should be reached where essentially all transferrable hydrogens at C-5' of enzyme-bound AdoCbl have been replaced with deuterium . diol-d 2 Tinactivator] ~ + 0 (x and y represent the isotopic composition that would characterize the cofactor.) Thus, when the cofactor is saturated with deuterium, no further increase in the isotope effect with decreasing inactivator concentration would be possible, and a maximum value for the i sotope effect for each particular inactivator would be reached as demonstrated experimentally in Fig. 5. The inactivator concentration at which this maximum , limiting value is expressed differs f or each inactivator. With (R)-gly-d 2 , this " saturation 11 effect occurs at a relatively high concentration of (R)-gly-d 2 ; with (RS)gly-d 2 and (S)-gly - d 2 somewhat lower concentrations are required; with undeuterated glycer ol, this saturation effect is observed only at very low concentrations of inactivator . This order likewise ac cords with the prop os E.i hypothesis. We have previously shown that glycerol not only inactivates holoenzyme , it can also serve as a 1. 8 1.9 1. 6 0.72 0.68 0.82 0.085 (R)-Glycerol-d 2 (RS)-Glycerol-d 2 (S)-Glycerol-d 2 Glycerol-d 5 6.0 0.22 15 4.3 0.30 0 . 087 3.0 3 .0 0.43 0.43 ~/kn f\.,1 .4 1.4 0.9 9 >l.O <l.30 f\.,o. 1.0 1.0 (~/kD)i 1. 30 1. 30 ki, l.4Xl0- 3M diol a 1,2-Propanediol aThe k. values listed in the competition with substrate experiments are the constant ki 1 s i fitting eq 5 at low inactivator concentrations. 15 1.0 1. 3 Glycerol ki(min- 1 ),a 1.osx10- 4 M diol-d2 l,2-Propanediol-d 2 Constants and Resulting Isotope Effects Comparison of the Limiting Values of the Inactivation Rate Substrate free ki(min- 1 ) <I<ii7k0 )i Table II: \0 .... w 140 substrate (involving predominantly the pro-(R) end of the glycerol molecule) (Bachovchin et al . , 1977). (R)-gly-d 2 contains only deuterium on the pro-(R) carbon ; thus, catalysis involving this glycerol as substrate shoul d more readily enrich enzyme-bound AdoCbl with deuterium than would catalysis involving (RS)-gly - d 2 , which contains both deuterium and hydrogen in equal amounts on the pro (R) carbon. Undeuterated glycerol and (S) - gly - d 2 acting as substrates should actually retard the satura tion of enzyme-bound AdoCbl with deuterium . Thus, the observed order follows from the hypothesis and the ability of glycerol to act both as substrate and inactivator. The maximum limiting values of the isotope effects are 3.0 both for undeuterated glycerol and for (R)-gly-d 2 , while for (RS)-gly-d 2 the experimental value is 4.3 (Fig. 5). These values can be quantitatively explained by the hyptohesis. At very low concentrations, glycerol in the presence of diol-d 2 finds enzyme- bound AdoCbl saturated with deuterium. /Ad !Dz gly-H + Co -+ where x =Dor H / cx 2 Ad 141 Since glycerol contains only hydrogen, the resulting 5 'deoxyadenosine must be -CD 2H. Using 15 as the value for the primary isotope effect in transfer away from 5'-deoxy -adenosine and assuming that the three hydrogens have a statistically equal chance to be removed in this rate-detE.r-· mining step, one calculates that the isotope effect should be: This agrees reasonably well with the experimental value of 3.0. With (R)-gly-d 2 , again only hydrogen can be transferred to AdoCbl during inactivation, since inactivation occurs predominantly from the pro-(S) end of glycerol (Bachovchin et al., 1977); thus, the calculated value is the same as that for undeuterated glycerol which agrees with the experimental value. However, with (RS)-gly-d 2 either deuterium or hydrogen can be transferred. so that two isotopically substituted types of 5'-deoxyadenosine are possible: OH I RCHz or RCDz I OH + /Ad CDz I Co /Ad CD 2H or CD 3 ~Ad k.]. ---. 1 Assuming an equal contribution of both types of 5 -deoxy adenosine to the rate and the statistical equivalence of 142 the three hydrogens at C-5', one calculates that the isotope effect in this case should be ) = [ 0. 5 + 0 • 5]-1 4 5 (lc_/k ·rt Di [2/3)(1/15) + (l/3)(l)J-1 IT = • which agrees well with the experimental value. In contrast, the isotope effects calculated for a mechanism involving an intermediate with two equivalent hydrogens would be 1.9 for gly and (R)-gly-d 2 and 3.3 for (RS)-gly-d 2 . Thus, the experimental data favor a mechanism involving the intermediacy of a species with three equivalent hydrogens. With (S)-gly-d 2 as inactivator, the above treatment predicts a limiting maximum isotope effect of~ 14. However, a value of 6 is observed (Fig. 5). A possible ex- planation for this result is that the (S)-gly-d 2 as synthe ·sized (Chapter I) is partly(~ 20%) racemic, and contains some of the R isomer. As deuterium is the only source of stereoisomerism after removal of the isopropylidene protecting group in step six of the eight-step synthesis, there is no sufficiently sensitive method for detecting such a small degree of racemization during either of the last two steps. However, in control experiments, all kinetic. data for inactivation and product formation obtained with (RS)-gly-d 2 were reproduced by an equimolar mixture of (R)- and (S)-gly-d 2 ; accordingly we believe this explanation to be unlikely. Alternatively, as 143 discussed in the previous section, t he true isotope effect on inactivation may be less than the value of 15 observed with gly-d 5 as inactivator . If fo r example the true iso - tci>.pe effect is 11-12, the observed l i miting maxi mum isotop,,,: effect of 6 observed for (S)-gly - d 2 in the pr es ence of diol-dz could be explained by a small fraction( ~ 15 %) o f inactivation occurring via the pro-(R) conformation. fact, this In possibility would also account for the limit - ing maximum isotope effect observed with (R)-gly - d 2 , which as noted is somewhat higher than predi cted on the basis of the proposed mechanism (no change in the value observe d for (RS)-gly-d 2 would be expected). We recognize that thi s explanation is at root conjectural, and cannot really be verified. However, of these limiting maximum isotope effects, that observed with (S)-gly-d 2 is by far the most sensitive to small deviations from the behavior pr edict e d by the (admittedly simplified) hypothesized inactivation mechanism. Further support for the proposed mechanism comes from evaluation of the Michaelis constants for inactivator . KG, and inhibitor, K , as measured at low inactivator con1 centrations in the linear regions of Fig. 4. With k l. constant, eq 5 should apply, and in fact double -recipro cal plots of these data points are linear with slopes and y-intercepts indica~ing the a ppr opriate new values of ki 144 while indicating no effect on either KG or K for each 1 inactivator (Fig. 4) . Thus the proposed hypothesis account s for the features of the experimental results presented in Fig. 4 and 5. We now consider the corresponding "mixed" cases in which undeuterated propanediol is the inhibitor of inacti\;- ation. In these cases, the deuterium content of enzyme- bound AdoCbl should decrease with decreasing concentration of inactivator and the observed isotope effect should also decrease. ~Ad CH D IX y Co diol [inactivator] ➔ 0 I /Ad CHz I I Co Figure 6 illustrates the agreement between this expectation and the experimental results . This effect is more clearly illustrated with gly-d 5 as inactivator; at high concentra tions of gly-d 5 , the isotope effect is quite large, but with decreasing concentrations of gly-d 5 the observed isotope effect decreases dramatically. Again, as one lowers the concentration of inactivator relative to diol inhibitor, enzyme-bound AdoCbl should become saturated with the isotope present in the diol, which in this case is hydrogen rather than deuterium. Thus, one anticipates a limiting minimum value for the isotope effect . The results in Fig. 6 suggest that such a limit is in fact 145 approached for each inactivator . However, in practice, re- liable experimental data in this region of "constant k 1.. " are difficult to obtain; the value of {S] (1.4 mM) chosen to ensure b. [S] "' 0 was so large that the observed inactiva. - tion rates are quite small. Nevertheless, limiting minimum values for the isotope effects expected for each inactivator can be calculated on the basis of the above hypothesis; these values agree rather well with those approached experimentally (Fig. 6). For example, (R)-gly-d 2 with only hydrogen at the pro-(S) carbon can generate only an Ad-CH 3 intermediate; the calculaced isotope effect for inactivation is thus 1.0. The calculated value for (RS)-gly-d 2 is only slightly larger (1.2). With both these inactivators, the experimental value appears to approach 1-1.1 as the limiting minimum value. (S)-Gly-d 2 and gly-d 5 , however, should only contribute deuterium; the intermediate in the limiting case is Ad-CH 2D, and the iso tope effect calculated for this intermediate is 1.4. The data for (S)-gly-d 2 do indeed seem to attain approximately this value at low inactivator concentrations. As this iso·· tope effect is considerably smaller than the value of 15 observed at large gly-d 5 concentrations, the data obtained for gly-d present the theory with a challenging test . 5 In fact, these experimental data do indicate such a large change, and seem to approach a greater than 1 (Fig. 6). minimum value somewhat 146 5'(JMMARY AND CONCLUSIONS The results of these kinetic investigations may be summa:i ized as follows. First, while substitution of deute~itllll for hydrogen at C-1 of 1,2-propanediol has no effect on the K1 of this substrate, the same substitution at the pro-(S) (but not the pro-(R)) hydroxymethyl group of ~lycerol causes an increase in the value of KG. In the predence of another substrate which undergoes catalysis, this value of KG is observed to vary as a function of inactivator concentration. The characteristics of this variation suggest that KG is not a simple dissociation cc;nstant, and that reaction steps subsequent to formation cf the holoenzyme-glycerol complex, involving the inter1.~diate hydrogen reserovir at C-5' of AdoCbl, contribute i:o determining its value. Second, the results of this work indicate the occur rence of an intermediate such as 5'-deoxyadenosine along the inactivation pathway and, furthermore, that transfer of hydrogen from such an intermediate containing three equivalent hydrogens is the rate-determining step in the inactivation process. This hypothesis can account quantitatively for the limiting values of the isotope effects on k.]_ observed at low inactivator concentrations in the presence of fixed concentrations of diol and diol-d 2 , which function as competitive inhibitors of inactivation. 147 The hypothesis also accounts for the varia t ion of ki with the concentration of inactivator in t he cases which are isotopically "mixed" and for the invariance of k.]. in the isotopically "homogeneous" cases. Accordingly, we condlud~ that the mechanism of inactivation, like that of catalysis, involves as the first step abstraction of hydrogen from C- 1 of glycerol and transfer of this hydrogen to the 5 '- carbon of cofactor. The three hydrogens ther eby attached to C- 5' of deoxyadenosine become equivalent and r emoval of one of them is involved in the rate-determi ning step for i n activa tion. Whether this transfer from 5'-deoxyadenosine invo l ve s readdition of hydrogen to the glycerol skeleton or transfe r to some other group or species associated with the enzyme 0r coenzyme remains unclear, as does the nature of the inacti vated holoenzyme complex. The identification of the nucle> - side in the inactivated complex as 5'-deoxyadenosine suggests that the 5'-deoxyadenosyl moiety which obtains after the second hydrogen-transfer step i s capable of acquiring a third hydrogen, possibly either from substrate or a near by amino acid residue, since intact AdoCbl cannot be recovered following glycerol inactivation (Bachovchin et al., 1977) . A 5'-deoxyadenosine moiety con tai n ing three equivalen t hydrogens has long been considered as an intermediate in catalysis. The significance of these results for the mechanism of catalysis is that they establish kinetically 148 the feasibility of such an intermediate in the mechanism of glycerol inactivation, a mechanism which is very similar to and may in fact be identical catalysis. to the mechanism of The major unique feature of inactivation coulc; be that enzyme-bourid "product" formed in the normal way from glycerol bound in an "EG8 11 fashion effects irreversible inactivation of the holoenzyme complex. 149 REFERENCES Abeles, R.H. (1971) Adv. Chem . Ser. 100, 346. Abeles, R.H., & Zagalak, B. (1966) J. Biol. Chem. 2L~l, 1245 . Babior, B. M. (1975a) Acc. Chem. Res.~. 376. Babior, B. M. (1975b) in Cobalamin, 3 . Babior, Ed., New York, N. Y., Wiley, Chapter 4. Bachovchin, W. W. , Eagar, R. G., Jr . , Moore, K. W., & Richards, J. H. (1977) Biochemistry 16, 1082 . Carty, T. J., Babior, B. M., & Abeles , R . H. (1971) J. Biol. Chem. 246, 6313. Eagar, R. G., Jr., Baltimore, B. G., Herbst, M., Barker, H. A., & Richards, J. H. (1972) Biochemistry g, 253. Eagar, R. G., Jr., Bachovchin, W.W . , & Richards, J . H. (1975) Biochemistry 14, 5523. Essenberg, M. K., Frey, P.A., & Abeles, R.H. (1971) J. Am. Chem. Soc. 93, 1242. Frey, P.A., Essenberg, M. K., & Abeles, R. H. (1967), J . Biol. Chem. 242, 5369. Hogenkamp, H. P. C. (1968) Ann. Rev. Biochem. 11., 225 . Jahn, W. (1965) Chem. Ber. 98, 1705. Lee, H. A., Jr., & Abeles, R.H. (1963) J. Biol. Chem. 238 2367. Marquardt, D. W. (1963) J. Soc. Indust. Appl. Math . .th, 43 .. McCarthy, J. R. , Jr. , Robins, R. K. , & Robins, N. J . (1968 ) J. Am. Chem. Soc. 90, 4993. .150 Miller, W.W., & Richards, J. H. (1969) J. Am. Chem. Soc. 91, 1498. Moore, K. W., Bachovchin, W.W ., Gunter, J. B., & Richards, J . H. (1979) Biochemistry 18 , in pr ess . Poznanskaya, A. A., Korsova, T. L . , Malakhov, A. A., & Yakovlev, V. A. (1972) Vopr . Med. Khim . ~' 145. Toraya, T., Shirakashi, T., Kosaga, T., & Fukui, S. (1976) Biochem . Biophys. Res. Corrnnun. 69, /+75. 151 CHAPTER III HYDROGEN TRANSFER IN CATALYSIS BY DIOL DEHYDRATASE 152 INTRODUCTION Diol dehydratase is an adenosylcobalamin-dependent enzyme which catalyzes the conversion of 1,2-propaned:i.ol to propionaldehyde, of ethylene gly col to acetaldehyde (Lee & Abeles, 1963), and of glycerol t o s-hydroxypropionaldehyde (Toraya et al., 1976 ; Bachovchin et al. , 19 77 ; Chapters I, II). The purpose of the work des c ribed in this chapter is to investigate the nature of the intermediate(s) on t he catalytic pathway between the 1,2-diol s ubstrate and the 1,1-diol intermediate which is subsequent ly dehydrated to form the ald e hyde. Experimenta l values of k ca t were deter- mined for several mixtures of l ~ 2-propanediol a.nd 1 , l-di-deuterio-1, 2-propanediol. These data wer e then analyzed in terms of an equation derived to describe the kinetics which would result for several different mechanisms. This method facili t ates discrimina tion among the suggested mech .. anisms, and als o allows calculation. of isotope effects on k ca t caused by substitution of deute rium for hydrogen in a single step of the reaction pathway. Furthermore, the re- sults of this analysis were used to predict a value for th2 tritium isotope effect in the diol dehydratase reaction, a :1d this pr edicted isotope effect was measured experimentally. Analog ous experiments wer e also carried out with mixtures of glyce rol and perdeuterioglycerol as substrate. The experiments and analysis of this work a re somewha ': 153 similar to those previously applied to methylmalonyl CoA mutase (Miller & Richards, 1969) and to glutamate mutase (Eagar et al., 1972). Moreover, these methods should pro,·e applicable to the study of intermedia tes in other enzymatLc processes. EXPERIMENTAL yropanediol dehldratase was obtained from Klebsiella l~neumoniae by a procedure adapted from Lee oc Abeles ( 1963) ~s described in Chapter I. Fraction E-8 with a specific activity of 35 - 58 was used for all determinations. Adenosylcobalamin (AdoCbl.) was purchased from Sigma Chemical Co. Assays were performed as previously des cri.hed (Bachov -~ chin et al., 1977; 1978; Chapter I) . Solutions generally contained: yeast alcohol dehydrogenase (.A.DH, Sigma), 10 units; NADH, 0.2 mM; apodiol dehydratase, 0.15 unit; bovine serum albu111in , 0.1 mg; potass ium phosphate buffer, pH 8, 30 mM; AdoCbl, 0.02 m.M; and the desired mixture of deuterated and undeuterated substrate, 20 m.M . Total volume was 2.3 ml and all reactions were carried out at 37° C. The reaction was initiated by addition of AdoCbl in a 0.03 ml ali quot . The rate for each substra te rn.ixture was det er- mined from at least three ki.netic experiments. Values of kcat for glycerol -perdeuterioglycerol mixtures were obtained as described in Chapter I from 154 experimentally determined values of k. 1., 0 tion rate constant, and P 00 , b , the inactiva -· S the amount of product s - hydrox- ypropionaldehyde formed before comple te i.nact:Lvation . TLe glycerol concentration employed was at: le ast 50 mM, approximately 14 - fold greater than the KG of pe rdeuterioglycerol. At most 7 % of the substrat e was converted to pro- duct during any determination. Tritium Isotope Effect . The tritium isotope effect was determined by measuring the increas e of the 3 B/ 11+c ra . tio of residual 1 , 2-propanediol during catalysis by diol dehydratase. Reaction mixtures contained : apodiol dehy- dratase, 2.3 units; bovine serum albumin, 0.09 mg; potassium phosphate buffer, pH 8, 50 mM; AdoCbl, 0 .03 mM; and 1,2-propanediol, 20 mM , containing 0.13 µCi 3H and 0.04 µCi 14 C. Volume 0.45 ml, 37° C. The reaction was initiated by addition of AdoCbl in a 0.01 ml aliquot and quenched by addition of 0.05 ml 2 N HCl. Three 0 .02 ml aliquots were removed for aldehyde assay by colorimetric determination of propionaldehyde 2,4 - dinitrophenylhydrazone (Eagar et al., 1975) to determine the extent of re act ion. A 0.15 aliquot of 7 M 1 , 2-propanediol was added to the sample as carrier, and 0.5 ml (10 µmole) of methanolic 2 , 4-dinitro phenylhydra.zine (DNP) added. The solution was allowed to stand for 2 - 3 hr and then centrifug ed. The supernatant was extracted with three 0.36 ml portions of chloro form. The aqueous layer was added to 20 ml chloroform and dried 155 with stirring over excess so dium sulfate. filtered and evaporated under vacuum. This solution was The yellow, oily res- idue was t aken up in 1 ml water , decolorized with charcoal , and centri f uged . An 0 .9 ml al i quot was added to PCS:xylene (2: 1) and assayed f or radi oactivity by double iso tope coLn.t- ing methods . Acid was ad ded to one sample p·r.:ior to AdoCb l as a control . Subs tr ate. (RS ) -1 , 2·-'P:r.opane.diol. and (RS) - 1, l-dideut ,~r- io - 1, 2- prop anediol (diol - d 2) were obtained as describ ed in Chapter I. (RS)-.!,.!_, ~- Trideuterio - _!, ~-12..ro:e anedi ol ( di ol-d 3 ) was prepared by reduction of ethyl pyr uv at e with lithium a l um1.num deuteride (Stohler, 99 % D). (RS)-Perdeuterio-1,~-propanedio l (di ol~ d 6 ) was prepa1ed by lithium aluminum deuteride reduction of ethyl 3 ,3 , 3- tri~ deuteriopyruvate which had been s ynthes i zed from Cd (CD 3) 2 and ethyl oxalyl chloride (Aldr ich) by the general procedur ,: outlined by Ko llonitsch (1966a ,b). CD 3 I (St ohl er) was used in the synthesis of the or ganoca dmium reagent . (RS)-[l - 14 CJ - .!,~- Pr opane dio l was pur chased from Dhom. (RS )- (1 - 3H] - l,~-Pr opane dio l was synthesized by reduc · tion of dl - ethyl lactate with LiAl 3HL. (New England Nucl ear: - . Radioch emi c a l pur i ty was at least 98 % as determine.d by t hin-l ayer chromatography on silica gel with 5 : 1 chlorofont1 :methanol as eluting so l ven t . Perdeut erioglycerol was purchased from Merck, Sharp , 156 and Dohme, Ltd . GompU:ting . Computer programs based on the al gorithms described by Marquardt (1963) were used to fit the experi mental rate data to an equation der ived to describe the kinetics of AdoCbl-dependent reactions. RESULTS Figure 1 illustrates the varia tion :Ln the observed catalytic rate constant as a functi on of mole fracti on of deuterated substrate in mixtures of l, 2.,·prop anediol with Rate data. obtained for thE,e diol-dz, diol-d 3 , or diol-d . 6 three substrates are virtually identical which indicates that substitution of deuterium for hydro gen at C-2 or C-3 of 1,2-propanediol does not significantly a ff ect the reaction. Several substrate concentrations were used; since all were at least 250 - fold greater than Km (Chapters I and II), saturation kinetics were observed in each case. Moreover, as a maximum of 1.7 % of the o ri ginal substrate was converted to product during any particul ar experiment , we eliminated complications resul ting from preferential depletion of undeuterated substr ate, with a concomi tant change in mole-fraction of deuterated substrate. Figure 2 shows the temperature dependence of k cat fo:r 1,2- propanediol and diol - d 2 . Values of the activa tion energy Ea and entropy 6St ( 37° C) obtained from these data are 9 . 2 + l kcal/mole and -17 .8 + 2 e.u. for 1,2- propanedi ol 157 Figur e l k ,~ t b as a f u n c t i o n o f )C::!n , the mole; frs.cti.on of ~a ,o s ~deuterated subs tr a te, for nd.xt nres o:E l, 2."·J}"ro p ane d.i.ol ·wi th dial - dz, diol - d , and dio1 -.. d at d i:[~ f 2.:i:·2nt: s ·l..ibs tra. te c on 3 6 0 • d..1.0 l - o. , 20 mM , 0,:. 'l • .,.. f"J' d J.o ' 1.. --o.J , c e n t r a t .1.ons : ._)· ,· d' .:\.o.,.,,cJ. 2 • 1., mu , ~ ; 2 3 23.4 rnM, 6 ; d:i.o l-~d , 10 . 3 rnM, C1 6 0 < 1 158 250 200 '' '' 150 ,-I '' I C) Q) '' fJ. (/) (/) ..0 0 ~ 100 ' '' .µ cu C) ~ "' ~ 50 ' ·, '' ' -~' ~ ~ ~ -, i--· .1 .2 .3 .4 .5 .6 .7 .8 .9 mole fra c tion deuterated subst r ate (XSD) Figure 1 J. . 0 159 Figure 2 Plot of ln(k ca t ,o b) s as a function of 1/T for 1,2-propanediol ( O , right) , and diol-d 2 ( ® , left) . 160 -- 5 . 6 5. 2 3.0 4.8 2.6 ,,...., {/) ,.0 0 ~ 2.2 .µ m ~ CJ '-' 1. 8 i::: r--l - - - - - - - - - - --.~~- -- - - -13.2 3.3 1/T, oK-l x 1000 Figure 2 3 .5 161 a.nd 9.8 ± 1 kcal/mole and -20 . 8 + 2 e . u . for diol - d 2 (Moore, 1962). Derivation of Rate Expres sion: A quantitative inter• pretation of the data of Fig. 1 requires an equation which relates the observed catalytic rate to the iso topic composition of substrate. Such an equation may be derived by consideration of the generally accepted machanism of cE talysis, which involves a minimum of two steps: ( i ) t).:-ans · fer of hydrogen (or deuterium) from C~· 1 of subs tr ate to an intermediate reservoir of hydrogen, and (ii) transfer of hydrogen (or deuterium) from this reservoir t o C-· 2 of product. Figure 3 outlines a simple hypothetical mechani ~m in which hydrogen abstracted from substrate becomes one oi. two equivalent hydrogens in an intermediate reservoir. Substrate may contain either deuterium (mo le fraction x8 D) or hydrogen (mole fraction XSH = 1 -~ x8 D) , and the reservcir may likewise have either deuterium (mol e fraction XD) or hydrogen (mole fraction XiI)· As illustra ted in Fig. 3, conversion of one molecule of substrate to product may occur by any one of six milcroscopic pathways (a-f), each of which is characterized by a part icul ar overall rate constant (kHH' kHD' kDH' or kDD) . The contribution of each pathway to the overall observed ra,t e constant (kcat, obs) j s the rate constant appropria t e for that particular path multiplied by the probability that such a path will b e fol lowed. Accordingly, the catalytic rate ob served fo r a 162. Figure 3 Reaction scheme for a hypothe t:ica.1. m.e chan:Lsm in which the intermediate carries two equ:!.valent hydro gens. Contr ibu tion to kcat Rate Rlthway (ii) Intermediate (i) XHXSHkHH kHH a Blz-H -H B12-Hz +H Blz-H 2 3 2 Figure kHH e B12 -D +H XDXSHkHH --z Bu -HD XDXSHkHD kHD d Blz-H XHXSDkDH kDH C Blz-D 2 B12 -HD XHXSDkDD kDD b B12 -H +D -D D 12 XDXSDkDD kDD f B1z-D -D B12-Dz B t--' °' w 164 given isotopic composition of substrate will be the sum of terms at the bottom of Fig. 3. kcat,obs = Xii [xsHkHH + XsD (kDH ; .~DD)] + Xn [ XSH ( 7 kHH + kHD '\ t. • -2 ) -, XsDkDDJ • (l) This expression can be simplifled by der.i.ving XD (and XH) in terms of Xsn• assuming that t h e i sotopic composition of the coenzyme has reached a steady state. (This is ar ea- sonable expectation, as each molecule of holoenzyme turns over thousands of molecules of subs trate even during the very early stages of the reaction. ) In the mechanism of Fig. 3, only pathways c and d affect the isotopic composition of hydrogen attached to the coenzyme. At steady state, their contributions to kcat obs become equal, so that: ' This leads, upon s ubstitution o f }~ = 1 - XD a nd r earr ange ment, to: (2) Equation 2 i ndica tes that coenzyme -will become enriched i n deuter~um i f kHD < kDH ; f or 11ID > kDH, the intermediate will be depleted in heavy isotope relat i v e to substrate . Equation 1 may general iz ed to describe a me chanism in which the intermediate hydrogen r eservoir contains n 165 hydrogens and in which that hydrogen transferred to the reservoir from substrate has a statiscal probability f/n. of being returned to product in step (ii) (f = 1 for a mechanism with n equivalent hydrogens and f = 0 for a "meJry mechanism in which hydrogen can never be re- -go-roun d " turned to the same substrate molecule from which it was abstracted (Miller & Richards, 1969; Eagar et al., 1972)) . _ m (n-1°' kcat,obs - i;,l i-1) ~ + [n-i] XD [i -- 1) { ["' f .. XSH ,_ (n + B)kHH + Ck!·m] XSD [skDH + (~ + C)knnJ ) (3) where B = (n-i)(n-f) ri"{n-1) - C = (i-1) (n-f) n (n-1) and (~=~ is the usual combinatorial no ta tion. In a manner analogous to that used earlier in the case for f = 1 n = 2' ' one can show that eq 2 also applies to the general case. Equation 3 includes several assumptions. Hydrogen in the intermediate reservoir may not exchange with product. In our experiments both the high substrate concentrations employed (which should saturateholoe:nzyme with propanediol) and the immediate reduction by ADH/NADH of propionaldehyde released from the enzyme preclude significant recombination between enzyme and product . In addition , hydrogen in the 166 reservoir may not exchange with unreacted substrate to such an extent that the isotopic compositions of either are appreciably affected. The absence of appreciable exchange of hydrogen between cofactor and unreacted substrate has previously been demonstrated (Frey et al., 1967a. ; Carty et al., 1971). Furthermore , the finding that AdoCbl becomes considerably enriched in tritium relat:tve to substrate during catalysis (Essenberg et al., 1971) indicates that the hydrogen of substrate does not equilibrate with that of AdoCbl. Equation 3 further requires that all (n-1) hydro- gens initially in the reservoir become chemically equivalent in the intermediate. this requirement. At least two results support Both hydrogens attached to C-5i of AdoCbl can participate in the reac t ion (Frey et al., 1967b ) and 1n NMR studies of ethanolamine deaminase (Hull et al., 1975) indicate that the methyl group of enzyme-bound 5'deoxyadenosine. the putative form of the coenzyme during the intermediate stages of t he enzymatic reaction, rotates at a rate exceeding 10 7 sec ·- 1 , which :Ls many orders of ma.g• nitude larger than the rate of hydrogen transfer from the intermediate to product. Lastly, eq 3 neglects any secon- dary isotope effect; for example, abstraction of hydrogen should occur with equal facility from intermediates CH~Ad, ..., CH 2 DAd, and CHD 2Ad. Although no definitive support exists for this assumption , such secondary isotope effects should be small relative to the primary isotope effects included 167 in eq 3 (Richards, 1970). Moreover , the kinetics of inac- tivation of diol dehydratase by glycerol of varying isoto pic composition can be explained quantitatively without postulating such isotope effects (Chapter II). The parameters kHH' kHD' kDH' and kDD in eq 3 we re varied by the method of least squares estimation of nonlinear parameters to obtain t he best fit to this equation of the experimental ra te data obtained with various mixtures of 1,2-propanediol and diol-d 2 . A number of mechanistic schemes were considered, including those with 2, 3, 4, or 5 "equivalent" hydrogens (f = 1, and n::;:: 2 , 3 , 4, 5) and "merry-go-round" schemes ( f = 0 and n = 2, 3 , t~, 5) . The results of these comparisons are surmnarized in Table I. The uncertainties listed for each rate constant are stan-· <lard deviations obtained from the least squares fitting program. CJ = The value of a (column 7) was calculated from d 1/2 k . )2 E (k cac,obs i=l cat,calc (4) d p where k ca t ,ca 1c is the value of k ca t for a given value of x80 calculated by optimization of eq 3; kcat obs i s the ' experimentally observed rate fo r substrate with this par- ticular isotopic composition; dis the number of data points used in the analysis ( d = 12 for l ,2 ~pr.opane diol / diol - d 2 mixtures); and pis the number of paramet er s which were varied during the optimization (p = 4 for kHH' kHD' 168 Table I n f kHH kHD kDH kDD (J Equivalent Hyd1:£gen Models 2 1 250±1 -11.6±2.8 -56±11 20.7±1 1.22 3 l 250±1 23.1±2.1 112.±16 20.7±1 1. 22 4 1 250±1 34.7±1.8 168±18 20,7±1 1. 22 5 1 250±1 40 .4±1. 7 196±19 20.7±1 1.22 20.7±1 1.22 Merry-Go-Round Models 2-5 0 250±1 57 .8±1.4 280±23 Optimized rate constants (sec- 1 , based on a molecula-· weight of 250,000 and specific activity 60 units/mg (Essenberg et al., 1971)) for a least squares fit to eq 3 of rate data obtained with l,2-pr:opanediol/diol-d 2 mixtures. Identical results were obtained for 2,3,4, and 5 hydrogen merry-go-round schemes. 169 l<pH' and kDD) . In a second analysis the parameter f, wh i ch is pr oportional to the likelihood of t he hydrogen derived fro m sub -· strate being returned to product, was also allowed t o var y systematically . Table II s umma rizes these calculations . Although the value o f a is lar ger (a= 1 . 30) than f or th e data in Table I (a= 1.22), th e a g reement betwe e n experimental and calculated rates i s actually slightly improved, Moreover , in the compar i sons of Table II, p = 5 whereas in Table I, p = 4, and this decrea s es the denominator of eq 4 which will, accordingly , inc reas e th e ca lcul at e d values of a in Table II. Table III surmnarizes the varia t i on of E (k c a t,c a l c 2 kcat,obs) with changing values of n and f fo r a given set of rate constants (data from Tabl e I , r ow 2). These c om- parisons emphasize that, fo r a g iven s e t o f r a te const ants, the accuracy wi t h which e q 3 de scribes the experimental da ta depends sensitive l y on the ass ume d val u es o f n and f , and thereby on the ass umed mechanistic scheme. Glycerol as. Substrate . of k ca Fi gure l+ shows the varia t i on t with sub s t r ate iso t opic composition f or glyce rol, subs t r ate - i nactiv ator of d :i.ol dehydratase . ·1 Wh ile th e s e d ata ar e qua litat ive ly s i mi l a r to those of 1, 2-propanediol, the upwar d c ur v ature o f the data is much less marked than t hat of Fig . 1. The c orr esponding k cat value for (RS)- 1,1 - dideut eri oglycero l i s also shown i.n Fig . 4; i t i s t he 22.9±3.4 24.9±3.1 24.1±3.1 250±1 250±1 250±1 3 4 5 ±10 ±15 121 116 20.7±2 ±18 110 20.7±1 20.7±1 20.7±1 - kDD 99.5±36 kDH 1.63±0.14 1.29±0.11 1.0 ±0.075 0.70±0.11 -f 3.1 3 .1 3.0 2.9 n/f intermediate containing n hydrogens. Optimized rate constants (sec- 1 ) and f for mechanisms with an 20.6±8 250±2 2 - - ~D ~H n- Table II 1.30 1.30 1. 30 1.30 -0 I-' -....J 0 171 Table III E(k -k n f 2 0 5814 2 1 1462 3 l 11.8 4 1 375 5 1 941 cat,obsd cat 1 calc )2 Least squares criterion as a function of n and f for fixed values of the rate constants kHH = 250 sec - 1 , kHD = 23.1 sec - 1 , kDH = 112 sec- 1 , and kDD = 20.7 sec- 1 . 172 Figure 4 kcat,obs as a function of XSD ' the mole fraction of perdeuterioglycero l, for glycerol/perdeuterioglycerol mixtures as substrate ( [G] = 50 mM). ( A ) glycerol + perdeu- terioglycerol; ( O ) (RS )- 1, 1-dideut er io glycerol. 173 400 r-1 I 30 0 C) <I) Cl) Cl) .g ~ .µ rd 200 C) ~ 100 0.5 mole fraction perdeuterioglycerol in sub strate (Xsn> Figure 4 1.0 174 same as that expected for an equimolar mixture of glycerol and perdeuterioglycerol. As the values of KG, the enzyme - glycerol Michaelis constant, are different for glycerol and perdeuterioglycerol , these dat a could not be analyzed meaningfully in the manner described above for the data of Fig. 1. However, a quali tative analysis was carried out as described in the next section. Tritium Isotope Effect. The results of the tritium isotope effect determination are shown in Fig. 5, a plo t of the relative 3 H/ 14 C (dpm) ratio of res i dual 1,2-propane- diol as a function of the fract ion of substrate converted to product, x. As only one of the t-wo C-1 hydrogens of 1,2-propanediol can be abstracted duri.ng catalysis (Zagalak et al., 1966), only 50 % of the tritium in [l - 3H]-l,2propanediol participates in the reaction. The specific activity of a tritia.ted substrat e as a function of xis (Melander, 1960) where ST and SH are the amounts of reactive tritium and hy ·· drogen. However, when a fractio n m of the total tritium T in substrate do es no t participate in t he reaction, To (1 - m) T - m(l - x)To, 175 Figure 5 Relative 3 H/ 14 C ratio (dpm) of residual 1,2-propane- diol as a function of the amoun t of substrate consumed, x. The dashed and dotted lines are the calculated curves for kif/kT = 3 . 5 and 5.5, respe ctively. 176 12 • I ....-I 0 °' •I 11 :I •r-l .. ,I 'O (1) s:: m p.. 0 1-l p.. I 10 •I •I •I I 9 .I N .I , ....-I s:: •I 8 •r-l ~I I t.) -.:t" ....-I ........ ' •' •I • 7 ::z:: • Cl") (1) > •r-l 6 .I ,, . I .j.J m ....-I (1) 5 1-l ' I 4 ... ,I I I 3 ·i.1 I . (j.,, ; .I'!, .,, 2 . ,·,...-., ., ,; ., -·-1 I 1 .1 .5 .6 .7 .8 extent of reaction, x Figure 5 .9 1.0 177 and the above relation becomes = kT k -- - lJ (1 - m)(l - x) ~H + m. In the present case , m = 0.5 for [l - 3 H]-l , 2- propane diol. The dashed and dotted curves for Fi g; 5 are theoretical curves for (kH/kT) = 3.5 and 5.5, respectively . As the ex- perimental data lie between thes e curves, we conclude that for diol dehydratase, (kH/kT) = 4.5 ± 1. DISCUSSION A number of explanations can be considered to account for the nonlinear variation of kcat,obs (Fig. 1) with the isotopic composition of 1,2-propanediol. ations can be eliminated. Two such explan- A difference in enzyme-1,2-pro- panediol association constants as a f unction of the isotopic composition of substrate could account for our observations but seems unlikely in view of previous resul t s showing that isotopic composition does not affe c t binding of 1,2-propanediol (Chapter II) or of 3- fluoro - 1,2- propanediol (Eagar et al., 1975). A second possibility involves a reaction pathway with concerted transfer of two or more h y drogens . A relat i on between the isotopic composition of substra t e and reactiou velocity has been derived (Kresge, 1964; Gold, 1969) for such reactions. In the present case , in which no exchange 178 of substrate hydrogens with water occurs (Abeles & Lee, 1962), such an explanation would require that (kcat)l/n, where n is the number of hydrogens simultaneously in flight, should be a linear function of Xsn· Such behavior has in fact been reported in studies of the mechanism of action of serine proteases (Hunkapiller et al., 1976) and amidohydrolases (Schowen, 1977). dependence. Our results do not show such a Moreover, concerted transfer of two or more hydrogens, from C-1 of substrate to C- 5' and from C- 5' to C-2 of product, appears unlikely in view of the finding that hydrogen abstracted from C-1 of substrate has a significant probability of being returned to its parent sub~ strate molecule (Frey et al., 1967a). The third explanation, which we favor, attributes the shape of the curve in Fig. 1 to the existence of an inter mediate which is enriched in deuterium relative to substrate. Hydrogen transfer from this intermediate is a major ratedetermining step of the reaction. Substantial evidence for the existence of such an intermediate, in which the cofactor is present as 5'-deoxyadenosine, has been reported (Babior, 1975; Abeles & Dolphin, 1976). Furthermore, that this intermediate should be enriched in heavy isotope relative to substrate is supported by studies with tritiated substrates , and AdoCbl (Essenberg et al., 1971; Weisblat & Babior, 1971). • Models for Hydrogen Transfer. Equation 3 allows 179 evaluation of pathways with intermediates containing n hydrogens in which hydrogen derived from substrate has a probability f/n of participating in the second transfer step (Fig. 3, reaction (ii)). The best fit of the experimenta l rate data in Fig. 1 to this equation was determined by s ystematic variation of the rate constants kHH, kDH' kHD' and kDD by the method of least squares estimation of nonlinear parameters for 2 to 5 equivalent and 2 to 5 hydrogen merrygo-round mechanisms (Table I). In Table II, these rate con- stants, as well as the parameter f were allowed to vary. The rate constants ~Hand kDD are turnover rates when boi:h transfer steps involve only hydrogen or only deuterium, respectively; kHD and kDH are turnover rates in which deuterium is transferred only in the second, or only in the first step, respectively. As shown in Table I, row 1, a close fit for a mechani.sm involving two equivalent hydrogens requires the physicall·;meaningless condition that kHD and kDH assume negative values. On this basis we may eliminate this proposed mechan~ ism. Similar evidence also argues against the merry-go1 round mechanisms in which kDH"' 280 sec- , compared to kHH "'250 sec-l (kHH/kDH"' 0.89). Such an inverse primary is o- t ope effect has no precedent in studies of AdoCbl-dependeLt enzymes and little precedence in enzymology in general (Jencks, 1969; Richards, 1970). adequately describing Thus , the mechanisms most experimental data obtained with 180 1,2-propanediol are those involving an intermediate with three or more equivalent hydrogens (Table I, rows 2 - 4). The analysis in Table II, in which f is allowed to vary as well as the rate constants, indicates that, even when n differs from 3, hypothetical mechanisms in which the ratio n/f 1\,3 still provide the best fit to the experimental data. (In these cases, however, hydrogen derived from sub- strate is not considered statistically equivalent to those initially present at C-5' .) The improvement in this second analysis of the agreement between calculated and experimen~ tal data is, however, slight, in view of uncertainties inherent in the rate measurements. yses Thus, these data and anal- provide evidence for a mechanism involving an inter- mediate containing three equivalent hydrogens, one of which was abstracted from substrate. The possibility of more than three equivalent hydrogens cannot however be rigorously eliminated on the basis of our data alone. Rate-Determining Step. The value of kDD (20.7 sec -1 ~H/kDD = 12) indicates that the transfer of hydrogen is an important rate-contributing step of the reaction. Which step, (i) or (ii) is slower can be ascertained from the values of kDH and kHD in Table I. In the three equivalent hydrogen mechanism, substitution of deuterium for hydrogen in step (i) causes the turnover rate to decrease from 250 sec -1 (kHH) to 112 sec -1 (kDH ) for an isotope ef f ect o f kHH/kDH = 2.2. The same substitution in reaction (ii) 181 causes a much larger decrease, to kHD = 23 sec- 1 for ~H/kHD = 10.8, an isotope effect similar to that observed when deuterium participates in both reactions (kHH/kDD = (We emphasize that the rate constants l<-iJn and kDH 12). represent turnover rates when deuterium replaces hydrogen in only one of steps (i) or (ii); they are not true "microscopic" rate constants for a particular step of the reaction.) These data do nevertheless clearly identify reac- tion (ii), the transfer of hydrogen from cofactor intermediate to product,as the slowest step in the conversion of 1,2-propanediol to propionaldehyde catalyzed by diol dehydratase. This accords with conclusions drawn for ethanol- amine deaminase (Weisblat & Babior, 1971) and for inactivation of diol dehydratase by glycerol (Chapter II), but differs from a previous conclusion (Essenberg et al., 1971). The possibility that the rate-limiting step is =changed by participation of heavy isotope was tested by determining the energies of activation Ea for catalysis with diol and diol-d 2 as substrates. Presumably, a change in the rate-determining step effected by the presence of deuterium would involve a concomitant change in Ea (Weisblat & Babior, 1971). Of course, this assumption would not be cor- rect if the two reaction steps had similar activation energies, or if substitution of deuterium for hydrogen caused an activation energy change much larger than expected from a change in the zero-point energy of a carbon-hydrogen 182 bond. However, in the former case, neither step could be considered "rate-determining." In the latter case, quan- tum mechanical tunneling (Lewis & Robinson, 1968) could conce i vably cause such a large energy change, but even in this event , failure to observe a change in Ea would be quite fortuitous. The observed Ea values (Results) were virtually iden-; tical for the two substrates, and we conclude that hydrogen transfer from AdoCbl to C-2. of product i s the pr edom·· inantly rate-determining step of the reaction. Glycerol as Substrate. The data of Fig. 4 show much less marked nonlinear character than those of 1,2-propa.nediol (Fig. 1). However, the variation of KG of glycerol with isotopic composition (Chapter II) precludes an analysis of these data such as that carried out for 1,2-propanediol (Table I) . Values for KG are 1.6 mM for glycerol (and (R)-1,1-dideuterioglycerol) and about 3.8 mM for perdeuterioglycerol (and (S)-1,1 -dideuterioglycerol). The "apparent 11 KG for a mixture of two stereoisomeric s ubstrates Rand Sin relative amounts xR and xS is given by = with XR + XS = 1. Thus, in the neighborhood of XS = 1 (XSD = 1, Fig . 4) and xR = 1 (XSD = 0' F'ig. 4)' KG changes relatively slowly, and the variation of k ca t ,o.b s wi t h XSD 183 may be compared with that of 1,2-propanediol. For example, at x80 = 0.1, the kcat,obs of glycerol has decreased from 1 412 sec-l to 360 sec- , a ~13 % decrease . In contrast, 1 kcat,obs of 1,2-propanediol at x80 = 0.1 is 183 sec- , a 27 % de.rease from the all-hydrogen rate. Similarly, at x5O = C.9 the rate of conversion of glycerol to s-hydroxypropionaldehyde is 80 sec -1 , a 60 % increase over the rate observed with perdeuterioglycerol alone, while at XSD = 0.9 the 1,2-propanediol kcat,obs has increased only 43 %, from -1 -1 21 s, c at XSD = 1.0 to about 30 sec . While analysis of the l,2-propanediol rate data indicates that (kDH/kHD) ~ 5 for a three-equivalent hydrogen mechanism (Table I, row 2), a manual fit to eq 3 of data obtained for glycerol near XSD = 0 and XSD = 1 indicates that in this case (kDH/kHD) ~ 1.5 + 0.3. Accordingly, we conclude that with glycerol as s1bstrate neither the first nor the second hydrogen transfer tep is strictly rate-determining in catalysis. This con- clusion is supported by the observation (Chapter II) that 5'-deoxyadenosine isolated from [l - 3H]-glycerol-inactivated dial dehydratase is enriched only about two-fold in tritium relative to substrate, while for AdoCbl isolated from holoenzyme and tritiated 1,2 - propanediol, up to a 20-fold enrichment has been observed (Essenberg et al ., 1971) . The enrichment of the intermediate hydrogen reservoir in heavy isotope is proportional to the ratio kDH/kHD ' as discussed later. 184 Relation to Microscopic Rate Constants. Although the rate constants ~H' kHD' kDH ' and kDD are overall turnover rates, their values as determined above may be used to deduce relationships between some of the microscopic rate constants. Based on previous studies, the minimum reacti.. . n scheme which may be drawn is k E + S ~n +- kl ES ➔ +- ES' kJ E + p k2 koff k 1 is the rate constant for hydrogen transfer from substrate to cofactor, while k 3 is the net rate constant for the remainder of the pathway, which includes hydrogen transfer from cofactor to product. When holoenzyme is saturated with substrate, kcat is given by (Cleland, 1975) V max ~ = Each of the rate constants k 1 , k 2 , and k 3 may manifes ~ a primary deuterium isotope effect. This expression may b :~ used to derive expressions for the observed isotope effect.; ~H/~D' kHH/kDH' and ~H/kDD in terms of the microscopic constants. For the first of these, == 185 Because the values of k 1 , k 2 j and k 3 are unknown, ~H/~D cannot be evaluated explicitly . However, limiting values of kiIH/~D may be obtained for the cases k 3 >> (k 1 + k 2 ), for which kHH/~D = 1 and k 3 << (kl+ k 2 ), for which Similar consideration of kHI-ifkDH and ~H/kDD (neglecting equilibrium isotope effects) gives the results in Table IV. In view of the values kHH/kHD = 10.8, kHH/kDD == 12, and k1rn/kDH = 2. 2 obtained experimentally, we conclude that the diol dehydratase reaction with 1, 2-propanediol as sub -· strate corresponds more closely to the left column of Table IV. In contrast, the glutamate mutase reaction (Eagar et al., 1972), with isotope effects of kHH/kDD ("kH/kn") 7. 5, ¾ml~D c~~:Dj = 7. 5/7. 0 = 1.1, and ~H/kDH (~~:D) = 7.5/1.5 = 5.0, corresponds to the right side of Table IV, with k 1 most likely rate-limiting. Methylmalonyl CoA mutase (Miller & Richards, 1969) , which exhibits value:s of 3.5, 1.1, and 1.2 respectively (and glycerol with diol dehydratase) constitutes an intermediate case in which both hydrogen abs traction from subs tr ate and hydrogen return tc, form product influence significantly the overall reaction rate. Relation to Previous Studies. The existence of an intermediate in which the cofactor exists as 186 TABLE IV Limiting Values of Overall Isotope Effects for a TwoStep Reaction k3< < (k1 + kz 2.. * ~ >> (k1 kHH/kiJD kr / kp ~HH/~.DH lt k¥ /kr k.HH/knn k~ I kp K¥ /k~ *t + ~l 1 If the hydrogens in the intermediate are equivalent with respect to k2 , the* and t labeled quantities will be somewhat reduced and increased, respectively. 187 5'-deoxyadenosine with three equivalent hydrogens has been inferred from studies of ethanolamine deaminase (Babior, 1969; Weisblat & Babior, 1971) and diol dehydratase (Fre y et al., 1967a; Essenberg et al . , 1971), using substrates and AdoCbl in which mobile hydrogens are enriched with tritium. The results and analysis described here accord with the qualitative conclusions reached earlier. More importantly, these deuterium isotope effects observed during rearrangement of 1,)2-propanediol catalyzed by diol dehydratase un .. ambigously require the intervention of a "reservoir" containing at least three equivalent hydrogens. This same point has been established for two other AdoCbl-dependent enzymes, methylmalonyl CoA mutase (Miller & Richards, 1969) and glutamate mutase (Eagar et al., 1972) . Indeed, eq 3 as developed to analyze the results with deuterium can also be informatively applied to the earlier studies employing 3 H as a trace label. These investiga- tions involved two types of experiments: of 3 H from (a) transfer [5'- 3 H]-AdoCbl to product during catalysis of unlabeled substrate and (b) catalysis with [l- 3H]-l,2propanediol or [l- 3H]-ethanolamine as substrate. In case (a), the rate of tritium transfer from cofactor to product is given by the sum of those terms in eq 3 which involve kHT(~D in eq 3). Thus: 188 dTAdoCbl ~ (n-1 x__ [n-i]XT [i - lJX SH (i-l)(nf) = - __ d__t _ _ = L. i-1) -11 n(n-1) i=l where n = 3, f = 1, and x8H = 1 for unlabeled substrate this expression reduces to: -dTAdoCbl dt which upon integration becomes: TAdoCbl = To,AdoCbl Accordingly, transfer of be first-order in the rithmic plot of the 3 H from 3 H content 3 H content AdoCbl to product should of AdoCbl; a semiloga- of AdoCbl as a function of time should be linear with a slope of -2/3 kHT. Analysis of an experiment of type bis somewhat more complex. One may monitor both the specific activity of unreacted substrate and the ratio of specific activity of substrate to that of AdoCbl. The rate of conversion of tritiated substrate to product is given by the sum of those terms of eq 3 which contain kTH and kTT' where Band C have been defined previously. tion of n = 3 and f On substitu- = 1, and neglect of terms containing 189 2 XT, XsTX-r• 2 or XST~ - . this expression simplifies to: (9) Equation 9 may be used to derive an expression for the specific activity of residual substrate as a function of the fraction of substrate converted to product x (Melander , 1960): S~ecific Activity, x [(ZkTH+kTT)/ 3kHH -lJ =s_p_e_c1.~f--1.-·c_A,_c_t_,i,_v....i_t_y_.---x---=-0... = ( l- x) ( 9 a) Thus, observation of the specific activity of residual substrate can give only a weighted average of kTH and kTT· Despite the fact that x8 T increases continuously due to depletion of unlabeled substrate, eq 2 may be adapted directly to tritium kinetics for AdoCbl~dependent rearrangements if one assumes that a steady state value of XT is attained rapidly on the time scale in XsT· of changes In fact, measurements of the specific activity of AdoCbl during catalysis with tritiated 1,2-propanediol as substrate support this assumption (Essenberg et al., 1971). Furthermore, with the approximation that x8 TkTH << . XSH~T; eq 2 becomes: (10) When X-r is given by eq 10, the amount of tritium entering 190 the intermediate is equal to the amount transferred from intermediate to product. Under these steady - state con- ditions, the specific activity of product will be determined by the overall tritium isotope effect ~IH/[(2kTH+ kTT)/3] (eq 9a) as follows: Product Su strate (11) In light of these equations, the informa tion that c.sn be derived from experiments using tritium kinetic isotope effects may be sunnnarized. mined by A value of ~T can be deter - measuring the rate of tritium transfer from [5'- 3H]-AdoCbl to product derived from unlabeled substrate . The ratio kTH/~T may be determined by measuring the specific activities of substrate and AdoCbl during catalysis with tritiated substrate . Finally, measurement of the specific activity of unreacted substrate or of product as a function of extent of reaction can yield a value for (2kTH + kTT)/3. However , in no case can one calculate th £ isotope effect for a single microscopic step of the reac tion pathway from experiments of this type in which either deuter ium or tritium isotope effects are observed. In view of these conclusions , an earlier report (Essenberg et al., 1971) merits scrutiny . Bas ed on an experiment of the type described by eq 8, a tritium isotope effect of 250·(2/3) was calculated for hydrogen 191 transfer from AdoCbl to product at 10°c. precedentedly This value is un- large and is, further, not consistent with the deuterium isotope effects observed in this work. In another experiment, Essenberg et al. measured the specific activities of AdoCbl, propionaldehyde and residual substrate as a function of time with [1- 3H]-1,2-propanediol as substrate. Their first observation, that AdoCbl be- comes enriched in tritium relative to unreacted substrate by a factor of 19, accords with the conclus i ons reached in the present work. From the present results obtained with deuterium, approximate values of kTH and kHT can be calculated (Swain et al., 1958): (1 2) Thus, ~H/kTH = 3.1 and ~H/~T = 30.9 which, though they do not represent true isotope effects for single steps in the reaction, do approximately reflect the tritium isotope effects for the first and second hydrogen transfer steps, respectively. These values, when substituted into eq 10, give a value of XrlXST ~ 10 which must be multiplied by two to obbain a specific activity ratio of 20. This result agrees well with the experimental data and predicts a tritium isotope effect for the transfer of hydrogen from AdoCbl to product of 31, a magnitude observed in other enzymatic reactions (Jencks, 1969; Knowles & Albery, 1977). 192 In contrast, their observation that the specific ac tivity of propionaldehyde product (given by eq 11) is lower by a factor of 20-47 than that of residual substra t e (given by eq 9a). indicates that hydrogen transfer from cofactor to product exhibits an apparent tritium isotope effect approximately five-fold larger than that expected on the basis of our observed deuterium isotope effects. Similar arguments also apply to the work on ethanolamine deaminase (Babior, 1969; Weisblat & Babior, 1971) who observed a tritium isotope effect of 5.8 for the conversion of [l- 3H]-ethanolamine to acetaldehyde compared to a deuterium isotope effect of 6.8-7.4. This seemingly anomalous result (a deuterium isotope effect which is larger than the apparently analogous tritium isotope effe r t) in fact follows quantitatively from the treatment of this work. The tritium isotope effect is dominated by kTH and, therefore, depends principally on the isotope effect for the first hydrogen transfer step, a conclusion also reached qualitatively by Weisblat and Babior (1971) . The overall deuterium isotope effect depends, in contrast, on knn· When the effect of isotopic substitution is kDH > ~D (and kTH > ~T), for the overall reaction one observes that ~/kD > ~/kT. When applied to the diol dehydratase react i on, our data on the deuterium isotope effects, together with the above treatment (eq 12), predict a tritium isotope effect 193 of 4 . 5, compared to a deuterium isotope effect of 12. The data in Fig. 4 confirm this prediction, and are consistent with a value of kTH ~ 81 sec- 1 (~H/kTH = 3.1). As kTT is much smaller than 2kTH, the value of kH/kT is fairly insensitiv e to the value of this rate constant. However, the value of kTT predicted from the observed value of kDD is 7 sec- 1 (~H/kTT = 36), and the observed ~/kT is not inconsistent with this prediction. Thus, the overall tritium isotope effect, as well as the data of Essenberg et al. (1971) concerning the enrichment of AdoCbl in tritium during catalysis with [l- 3HJl,2-propanediol as substrate, are consistent with the deuterium isotope effects observed in this work. The deuterium isotope effects observed in this study are within the range of those reported for free radical reactions involving cleavage of a carbon-hydnogen bond. For example, Wilen & Eliel (1958) reported isotope effects ranging from 3 to 10 for abstraction of hydrogen from adeuterated toluene and p-xylene by various peroxides . Howard et al. (1968) observed isotope effects of 8 to 20 for the oxidation of deuteriocumenes and perdeuteriotetralin by cumene hydroperoxide and t-butylhydroperoxide. Thus, our data are compatible with a mechanism involving 194 hydrogen transfer between free radicals (Babier, 1970; Finlay et al., 1972; Finlay et al., 1973). SH + S· P· + PH • IT >Co ·-< However, with the exception of the work of Walling & Cioffari (1972) and Julia (1971), and of calculations carried out by Golding & Radom (1976), there is little precedent for a 1,2-rearrangement of a radical species . In contrast, 1,2-rearnangement of a carbonium ion, such as might be generated by oxidation of the l,2-propanediol1-yl radical by cob(II)alamin, has ample precedent. In this event, the following mechanism could obtain as sug gested earlier (Eagar et al., 1972; Halpern, 1974); OHOH I I (D CH 3 CH~H + - I I CH 3CHCH • (£) OH I +I 0I + >Co < r - CH:eCHCH + >Co OH @ OHOH . >CO II< 01 < OH I 11 OHOH~ I II CH 3CHCH . II < ~CH 3~HIH + >Co 0 OH Free radical I is more stable than its precursor, ·CH 2Ad. Similarly, the rearranged radical III is less stable than 195 •CH 2Ad, and formation ,of the latter, by transfer of a C- 5 1 hydrogen to C-2 of III, would be energetically favored. The relative stability of the oxocabonium ion II could help to facilitate oxidation of the radical I by Cbl 11 , while the eventual elimination of H2o could provide a driving force for rearrangement of II and reduction of the resulting carbonium ion by CblI. CONCLUSIONS By analysis of the variation of k ca t with mole fraction deuterated substrate in the AdoCbl-dependent conversion of 1,2-propanediol to propionaldehyde, we have infer red the obligatory intervention of an intermediate hydrogen reservoir in which the hydrogen derived from substrate becomes equivalent with at least two other hydrogen atoms ; one of these three equivalent hydrogen atoms is returned to form product in the principal rate -determining step of the reaction. With glycerol as substrate, neither hydro- gen transfer reacti_on appears to be a major rate -contributing step. The magnitude of the primary deuterium isotope effects observed in these two hydrogen transfer steps are consistent with these steps' involving cleavage of a carbon-hydrogen bond. These results should, therefore, be irmnune to the criticism (Schrauzer, 1971) of experiments using tritium as a trace label. The role of the 5'-methyl group of 5 ' -deoxyadenosine as the reservoir of the three 196 equivalent hydrogens has ample precedence (for example: Babior et al., 1974; Bachovchin et al., 1978). 197 REFERENCES Abeles, R.H., & Dolphin, D. (1976) Acc . Chem. Res.~. 114 . Abeles, R.H., & Lee, H. A., Jr. (1962) Brookhaven , Symp . Biol. 15, 310. Babior, B. M. (1969) J. Biol. Chem. 244, 449. Babior, B. M. (1970) J. Biol. Chem. 245 , 6125. Babior, B. M. (1975) Acc. Chem. Res.~. 376. Babior, B. M., Carty, T. J., & Abeles, R.H. (1974) J . Biol. Chem. 249, 1689. Bachovchin, W. W., Eagar, R. G., Jr., Moore, K. W., & Richards, J. H. (1977) Biochemistry 16, 1082. Bachovchin, W.W., Moore, K. W., & Richards, J. H. (1978) Biochemistry 17, 2218. Carty, T. J., Babior, B. M., & Abeles, R.H. (1971) J. Biol. Chem. 246, 6313. Cleland, W.W. (1975) Biochemistry 14, 3220. Eagar, R. G., Jr., Baltimore, B. G., Herbst, M. M., Barker, H. A., & Richards, J. H. (1972) Biochemistry 11, 253. Eagar, R. G., Jr., Bachovchin, W.W., & Richards, J . H. (1975) Biochemistry 14, 5523. Essenberg, M. K., Frey, P.A., & Abeles, R.H. (1971) J. Am. Chem. Soc. 93, 1242. Finlay, T. H., Valinsky, J., Sato, K., & Abeles, R.H . (1972) J. Biol. Chem. 247, 4197. 198 Finlay, T. H., Valinsky, J., Mildvan, A. S., & Abeles, R. H. (1973) J. Biol. Chem. 248, 1285. Frey, P.A., Essenberg, M. K., & Abeles, R.H. (1967a) J. Biol. Chem. 242, 5369. Frey, P. A., Kerwar, S. S., & Abeles, R. H. (1967b) Biochern. Biophys. Res. Commun. 29, 873. Gold, V. (1969) Adv. Phys. Org. Chem. 2, 259. Golding, B. T., & Radom, L.T . (1976) J. Am. Chem. Soc . 98, 6331. Halpern, J. (1974) Ann. N. Y. Acad. Sci. 239, 2. Howard, J. A., Ingold, K. Y., & Symonds, M. (1968) Can . J. Chem. 46, 1017. Hull, W. E., Mauck, L., & Babior, B. M. (1975) J. Biol. Chem. 250, 8023. Hunkapiller, M. W., Forgac, M. D., & Richards, J. H. (1 976) Biochemistry 15, 5581. Jencks, W. P. (1969) Catalysis in Chemistry and Enzymolo gr , New York, N. Y., McGraw Hill, Ch. 4. Julia, M. (1971) Acc. Chem. Res. 4, 386. Knowles, J. R., & Albery, W. J. (1977) Acc. Chem. Res. 10, 105. Kollonitsch, J. (1966a) J. Chem. Soc . A, 453; (1966b) J. Chem. Soc. A, 456. Kresge, A. G. (1964) Pure Appl. Chem.~. 243. Lee, H. A., Jr., & Abeles, R.H. (1963) J. Biol. Chem. 238, 2367. _ 199 Lewis, E. S., & Robinson, J. K. (1968) J. Am . Chem. Soc . 90, 4337. Marquardt, D. W. (1963) J. Soc. Indust. Appl. Math. 11,431. Melander, L. (1960) Isotope Effects on Reaction Rates , New York, N. Y., Ronald Press, Ch. 3. Miller, W.W., & Richards, J . H. (1969) J. Am. Chem . Soc . 91, 1498. Moore, W. J. (1962) Physical Chemistry, 3rd Ed . , London, Prentice-Hall, pp. 297-298. Richards, J . H. (1970) in The Enzymes, P. D. Boyer, Ed. , I, New York, N. Y., Academic Press, pp. 321-333. Schowen, R. L. (1977) in Isotope Effects on Enzyme-Catalyzed Reactions, W.W. Cleland, M. H. O'Leary, & D. B. Northrop, Eds., Baltimore, Md., University Park Press , pp. 79 -89. Schrauzer, G. N. (1971) Advan. Chem. Ser. 100, 1. Swain, C. G., Stivers, E. C., Reuwer, J. R., Jr . , & Schaad L. H. (1958) J. Am. Chem. Soc. 80, 5885. Toraya, T., Shirakashi, T., Kosaga, T.' & Fukui , s. (1976) Biochem. Biophys. Res. Commun. 69, 475. Walling, C.' & Cioffari, A. (1972) J. Am. Chem. Soc. 94. 6064. Weisblat, D. A., & Babior, B. M . ( 19 71) J. Biol. Chem. 246, 6064. Wilen, S. H. , & Eliel, E. L. (1958) J. Am. Chem. Soc. 80, 3309. 200 Zagalak, B., Frey, P.A., Karabatsos, G. L., & Abeles, R.H. (1966) J. Biol. Chem. 241, 3028. 201 APPENDIX GROWTH OF KLEBSIELLA PNEUMONIAE AND ISOLATION OF DIOL DEHYDRATASE 202 During the past three years, modifications in the isolation procedure have significantly improved the yield and purity of diol dehydratase preparations in this laboratory. Yields have increased by up to a factor of four, and puri t y by up to 30 % over previous work. This appendix summarize s the current procedure used and makes tentative suggestions for further improvements. General References Lee, H. A., Jr., & Abeles, R.H. (1963) J. Biol. Chem. 238, 2367. Original reference . . Pozna.nskaya, A. A., Tanizawa, K., Toraya, T., Soda, K., & Fukui, S. (1977) Bioorg. Khim. I, 111. Recently published work. Some of the work described below has been carried out in collaboration with Dennis E. McGee. The Culture. Klebsiella pneumoniae (ATCC 8724) is obtained from the American Type Culture Collection as a freeze-dried culture. Nutrient Agar Stabs - for long term storage. 0 .9 g 0.5 g 0.75 g Difeo "Bacto" nutrient broth powder NaCl Difeo "Bacto" agar powder to 100 ml with distilled water Dissolve ingredients with short (2 - 3 min) heating in an autoclave (slow exhaust). Swirl to mix thoroughly, 203 and dispense 1 ml aliquots into screwcap culture tubes. Autoclave these tubes for 20 min, allow to cool in an upright position. Nutrient Agar Slopes: 5.26 g 1. 76 g 4 . 42 g Difeo "Bacto" nutrient broth powder NaCl Difeo "Bacto" agar powder to 350 ml with distilled water Dissolve ingredients with brief autoclaving, swirl to mix thoroughly, and dispense 6 - 7 ml into each culture tube. Cap the tubes and autoclave 20 min. Allow the tubes to cool at an angle and store upright at 4° C until use. Nutrient Agar Plates: 2.53 g Difeo "Bacto" nutrient broth powder 1. 9 g Difeo "Bacto" agar powder to 150 ml with distilled water Autoclave ingredients for 20 min, swirl, and dispense 20 - 25 ml into sterile petri dishes in a uv-hood under draft-free conditions. Allow plates to cool and store up- side down at 4° C until needed. The Klebsiella pneumoniae culture is maintained on aga r slopes. Every other week a single colony is isolated from an agar plate which has been grown overnight at 37 placed on a new slope. ° C and One slope is used as a master cul- ture t0 avoid contamination of the bacteria. All manipula tions are made with a sterile platinum wire loop in a uvhood in a draft-free room. 204 The Inoculation Broth: 12 g 90 ml 31.2 g 5.6 g 7.05 g 1.2 g Yeast extract glycerol K2HP0 4 •3H 2o KH Po 2 4 Autolyzed brewer's yeast fraction, SOX concentrated liquid. Amber laborator ies, Juneau, Wisc. 53039 414-386-2691 . (NH4)2S04 MgS0 4 •7H 2o to 6 1 with dist illed water Predissolve all salts. Add each ingredient in the or- der listed while diluting to volume with distilled water. Dispense altquots of approxima t ely 200 ml into 250 ml screwcap erlenmeyer flasks. These flasks are autoclaved for 30 min, allowed to cool, and the caps are tightened until use. Cells from a slope which has been grown 12 hr at 35 37° Care suspended in 1 ml of sterile distilled water and transferred to the sterile inoculation broth in a uv-hood in a draft-free room. The broths are incubated at 35 - 37° C with loosened caps for 12 hr. The Growth Medium: SO ml 1,2-propanediol 100 ml glycerol (glycerin) 3 - 5 g yeast extract 105 g KzHP04·3H20 20 g KH Po 2 4 24 g (NH 4 ) 2so 4 4 g MgS0 4 ·7H 20 to 18 1 with distilled water Predissolve all salts. Add each ingredient in the order listed while diluting to volume with di s tilled water 205 in a five-gallon carboy. The carboy is closed with a rub- ber stopper with a glass vent tube loosely plugged with nonabsorbent cotton or glass wool . The opening is covered with a double layer of aluminum foil and the medium autoclaved for 90 min. Cool and bring to 35 - 37° Cina hot . room or large bath. A fully grown inoculat£on broth is rapidly added to the center of an 18 1 medium at 35 - 37° C without stirring. The cells are allowed to grow for 24 hr (or unt i l a fair amount of froth appears), and are harvested by centrifugation over a period of 3 - 4 hr for an eight bottle batch, starting with the most frothy carboys . These large bottles should serve as the first point of variation if either growth or enzyme activity are unsatisfactory. Approximately 60 - 85 g wet cells should be ob- tained from an eight bottle harvest . correlates with poorer growth . Often better activit:1 The first variable is the yeast extract, which can be varied from 1.5 - 10 g. If low activity persists, the glycerol/propanediol ratio may be changed. The cells may be somewhat sensitive to the water source . The Harvest. A Beckman J21C centrifuge and JCF-Z ro · tor are used, at 12 KRPM and 4° C, with flow rates of 0.5 - 0.9 1/mi:n. The effluent should be examined at i ntervals for res i.dual bacteria. 206 The Cells. All procedures fr om this point on are done at 0-S 0 c, and only glass-distilled deionized water is used . After harvesting, the cell paste is weighed and suspended (a blender may be used) in 0.01 M Tris base (5 ml/lg wet cells). The suspension is stirred for 15- 30 minutes and centrifuged in a GSA rotor for one hour at 10 KRPM . The cells are suspended (may use a blender) in a minimum amount of water (maximum 2 ml/lg wet cells) and lyophilized . The Sonicate. The lyophilized cells are sonicated by the following procedure as soon as possible after the harvest. The lyophilized cells may be stored in a deep freeze (-80°C) if innnediate sonication is not possible. The dry cells are added to water (10 ml/lg dry cells) and homogenized in a blender. Dithiothreitol is added to a concentration of 15-20 mM, and the pH adjusted to 8.89.0 with 40% KOH with efficient stirring. After 30 minutes, the cells are sonicated with a Bronson sonicator (setting of "60") in an ice-methanol bath. The resulting enzyme activity is seemingly a function of both total sonication time and the duration of each burst. Therefore, the cells are sonicated for a total time of 5 min/lg dry cells in bursts which are as long as possible while still keeping the temperature below 1s 0 c. Bursts of up to 30 minutes may be attained by stirring the ice - methanol b a th . The bath cools most efficiently when it is mostly liquid. 207 After sonication. a cold suspension of Darco G60 charcoal, H2o, and distilled 1,2-propanediol (30g:115ml: 8 ml/30g dry cells) is added and efficiently stirred at 4°c for 15-20 min. The suspension is centrifuged 90 min at 11 KRPM in a GSA rotor. The supernatant is decanted, an aliquot removed for activity assay (0.05-0.l ml, diluted to 10 ml in Buffer E, below), and lyophilized for use in the enzyme preparation. The dry sonicate is stored in a deep freeze, and should contain at least 200 units /g before continuing to the end of the purification. The Assay. General reference: Bachovchin, W. W., Eagar, R. G. , Jr., Moore, K. W. , & Richards. J. H. (1977) Biochemistry 16, 1082. The assay solution should contain 60 mM 1,2-propanediol; 40-60 mM potassitnn phosphate buffer, pH 8.0; 10 unit/ml yeast alcohol dehydrogenase (ADH, Sigma); 0.2-0.25 mM nicotinamide adenine dinucleotide, reduced form (NADH, Sigma); and 20-40 µg/ml bovine serum albumin. A 2.0 ml aliquot of this solution is placed in a cuvet with a magnetic stirrer, and 0.2 ml of an enzyme-containing solution added to this. The cuvet is incubated with stirring at 37°c for 25 min., and 25 µl of 2.5 mg/ml adenosylcobalamin (AdoCbl, coenzyme B ; Sigma) solution is added to star t the reaction. The 12 absorbance at 340 nm is monitored with time to determine enzyme activity. Activity may also be assayed by the method of Eagar, 208 R. G., Jr., Bachovchin, W.W., & Richards, J. H. (1975) Biochemistry 14, 5523, via colorimetric determination of propionaldehyde-2,4-dinitrophenylhydrazone. Protein concentration is determined by the method of Lowry, 0. H., Rosebrough, N. J., Farr, A. L., & Randall, R. J. (1951) J. Biol. Chem. 193, 265. Enzyme Purification. These are not rigorous procedures, and the purification should be regarded as more of an art than a science. For those buffers containing dithiothrei t ol (DTT), best results are obtained by adding this reagent immediately before the buffer is to be used. Buffer A: 0.05 M (NH 4 ) 2so 4 (26.4 g) 0.01 M K2HP0 4 •3Hz0 (40 ml, 1 M solution) 2% 1,2-propanediol (80 ml) 0.01% NaN 3 (0.4 g) to pH 8.0 (40% KOH) to 4.0 1 with HzO 2 mM DTT (0.3 g/1) Buffer B: 0.5 M (NH 4 ) 2so 4 (66.05 g) 0.01 M K 2HP0 4 •3Hz0 (10 ml, 1 M solution) 2% 1,2-propanediol (20 ml) to pH 8.0 (40% KOH) to 1.0 lwith HzO Buffer D: 0.005 M KzHP0 4 •3HzO (5 ml, 1 M solution) 0.1 M 1,2-propanediol (7 .6 ml) 0.01% NaN 3 (0.1 g) to 1 . o 1 with H2o 299 Buffer E: 0.01 M KzHP0 4 •3H 2o (60 ml, 1 M solution) 2% 1,2-propanediol (120 ml) 0.01% NaN 3 (0.6 g) to 6 .O 1 with H2o 1 mM DTT (optional , 0.15 g/1) Buffer F: Buffer G: 10% glycerol, reagent (200 ml) 5% 1,2-propanediol (100 ml) 0.01 M potassium phosphate, pH 8.0 (20 ml, 1 M, pH 8 solution) 0.01% NaN 3 (0.2g) to 2.0 1 with HzO 0. 02% DTT (0. 2g/ 1) Assay buffer: 0.2 M potassium phosphate, pH 8.0 Assay buffer+ 2 mg/ml bovine serum albumin. The fate of Buffer C is unknown. should be distilled and not yellowish . 21 g sonicate. are noted. All 1,2-propanediol The scale is for All assay points and dilution factors All glassware should be very clean and at 25° or below (not hot'.). Crush 21 g of sonicate to a powder in a mortar and pestle and dissolve in a minimum amount of Buffer A (180220 ml), crushing lumps with a small spoon (not silver) if desired. Maintain pH at or above 8. 0 by addition of 40"~ KOH while stirring. Stir for 30 min in an ice bath. E-1: 0.2 ml/10 ml Buffer E. 210 Centrifuge 2 hr at 10 KRPM in a GSA rotor. supernatant and record volume. Decant Suspend pellet in a minimum amount of Buffer A and stir in the cold. Repeat two or three times with the supernatant, or until there is no pellet after a centrifugation. Add each pellet to the pellet (P) fraction, with more Buffer A as necessary to maintain a reasonably thin suspension. ml; P-1: E-ls: 0.2 ml/10 0.05 ml/10 ml Buffer E. At this point, P-1 may contain a substantial amount of enzyme activity, and the A280 ;A 260 ratio should be between 0.7 and 1.0. If P-1 contains 2500 units or more, it is probably worthwhile to work with it further, as discussed later. During the final spin of E-1s, prepare protamine sulfate solution by dissolving 3.0 g protamine sulfate (Calbiochem) in 147 ml Buffer A at room temperature, with pH adjusted to 8.0 with 40% KOH. Add 80 ml protamine sulfate solution dropwise with stirring to E-ls in an ice bath, keeping the pH above 8.0 with 40% KOH. Stir for 20 min, centrifuge 75 min at 11 KRPM in a GSA rotor. Discard the pellet . spin, grind (NH 4 ) 2so 4 (3 g/10 ml E-ls) . E-2. E-2: 0.2 ml/10 ml Buffer E. should be between 0.7 and 1.0. During the Supernatant is The A280 ;A 260 ratio If not, add 10 - 20 ml more protamine sulfate and treat as before. Add crushed (NH 4 ) 2so 4 (3 g/10 ml E·-2) to E- 2 slowly 211 while stirring in an ice bath and maintaining the pH above 8.0 with 40% KOH. After completing the addition, stir for 20 min and centrifuge 100 min at 11 KRPM in a GSA rotor . Supernatant is E-3s: If E- 3s contains 0. 2 ml/10 ml. substantial activity, add more crushed (NH 4 ) 2so 4 (up to 90-95% saturation) and centrifuge again . The pellets from the ammonium sulfate precipitation are suspended in 55-65 ml Buffer A and allowed to stir in the cold, overnight if desired. 0.05 ml/10 ml. The suspension is E-3: Dialyze P-1 overnight in the rocking dialyzer against Buffer E. Centrifuge E-3 for 2 hr at 15. 5 KRPM in an SS-34 rotor. Supernatant is E-4: 0.1 ml/10 ml. If E-4 contains all or a large fraction of the enzyme activity, set aside and proceed as described later. Buffer B, If not, suspend E-4's pellet in stir for 20 min. The P-1 dialysate is P-6: 0.05 ml/10 ml. Centrifuge the Buffer B suspension for 90 min at 15.5 KRPM in an SS-34 rotor. ml. Supernatant is E-Sa: 0 . 2 ml/10 Suspend the pellet in 25 ml Buffer B, stir for > 20 minutes. Centrifuge this suspension 70 min at 15 . 5 KRPM in an SS-34 rotor. Supernatant is E-5b: 0. 2 ml/10 ml. Suspend the pellet in a minimum amount of Buffer E and stir for several hours to solubilize the protein. Dialyze overnight against Buffer E in a rocking dialyaer. 212 The resulting dialysate is E-6: 0 .1 ml /10 ml. Centrifuge E-6 and/or P-6 for 2-2.5 hr at 40 KRPM in a Ti60 rotor. P-6s: Remove the supernatant· carefully (E-6s; 0.1 ml/10 ml) with a pasteur pipet , stopping when the dense, yellow suspension is reached. The pellets may be resuspended in Buffer E, solubilized, and centr ifuged again to extract more diol dehydratase. P-6s may contain a large fraction of the activity present in P-6. Transfer P-6s and E-6s to corex centr i fuge tubes. Chill the tubes in ice, and add Buffer F (0 . 03 ml/1 ml 6s supernatant) in 0.1 ml portions, while stirring with a glass rod. Allow solutions to stand for 30 min - 1 hr in the ice bath. Centrifuge at 10 KRPM for 10 min in an SS-34 rotor ; Decant supernatants, and suspend the pellets in 0.5-3 ml Buffer D, according to their apparent size . Cover and allJw to solubilize for 1-2 days before assaying . P-8a; E-8a: 0. 02 ml + 0. 4 ml Buffer E. Ass.ay O. 02 ml for enzyme activity, and 0.4 ml for protein. The 8a pellets may be combined, diluted with Buffer D to about 200 unit/ml, and dialyzed against Buffer Dor E to solubilize the protein. The supernatants from these pellets are P-7 and E- 7 . If they contain substantial activity( > 25 - 30 unit/ml), I • treat as for E-4 described below. The enzyme-containing solution (E-4, E-7, or both) is treated with ammonium sulfate as described previously , 213 and centrifuged. An effort should be made to precipitate as much enzyme activity as possible. Suspend the pellet in a minimum amount of Buffer E and dialyze against two or three 500 ml portions of Buffer E (the rocking dialyzer mav be used). This solution is then dialyzed against two 500 ml portions of Buffer G and centrifuged for one hour at 10 KRPM in an SS-34 rotor. The supernatant is passed onto a 2.6 x 8cm column of Cellex D, equilibrated with Buffer G. The column is washed with~ 10 ml Buffer G. Diol dehydratase is then eluted in 1.5- 2 ml fractions with a stepwise gradi ent of 0.08 M, 0.12 M, and 0.2 M KCl in Buffer G. A typical elution profile is shown in Fig. 1. activity is associated with two protein bands. Enzyme Most of the activity elutes at 0.12 M KCl, and is pooled as desired. We have obtained enzyme concentrations near 600 unit/ml with a specific activity of nearly 80 units/mg (see Poznanskaya et al. (1977) cited in the beginning of this appendix) in this band: E-8b: 0.015 ml+ 0 . 4 ml Buffer E. However, enzyme activity also elutes at 0.2 M KCl , although the quantity is much smaller. When enzyme from E-8a or P-8a is added to the E-4/E-7 solution prior to Cellex-D chromatography, the activity of enzyme eluting at 0.2 M KCl is enhanced. No other pr oper ty could be found which distinguishes these two forms of diol dehydratase. Their reactions with 1,2-propanediol and glycerol were identical in terms of the kinetic parameters kp, k 1 , K8 , 214 Figure 1 Elution profile of diol dehydratase during Cellex-D chromatography in Buffer G. (-o-) A280 ; (--0--) diol dehydratase activity, unit/ml. Fraction volume 2 ml. 215 0 0 -.::t 0 0 .-4 0 0 N 0 0 M 0 in .-4 rr:..d "9 I ~ 0 0 .-4 ,, / 1-l Q) ~ ~~ § s:: s::0 'M .µ CJ C'd H ~ 0 N ~ 4-00 C 0 N 0 and KG. 216 Furthermore, the kinetic isotope effects for catalysis and inactivation were the same for each fraction. Their molecular weights as measured by chromatography on Sepharose 6B were identical, as waa the four band pattern obtained by SDS-gel electrophoresis (Poznanskaya et al., 1977) . Preliminary results indicate that the relative propor tions of these two fractions (and of the amounts of enzyme appearing in E-8a/P-8a and in the E-4 supernatant) are related to the of sonication. total time and the duration of each burst For a given total sonication time, an in- crease of the "burst time" seems to correlate with decreased activity in the 8a fractions and the 0.2 M KCl fraction, and with increased enzyme activity in the E-4 supernatant and 0.12 M KCl fraction. In addition, longer bursts of sonication seem to give a higher overall yield, even when the total sonication time does not change. The reason for this behavior remains obscure. Suggested Further Studies. Alternative means of rupturing the bacteria should probably be tried, such as pulverizing with glass beads (Poznanskaya et al., 1977). The sonication procedure is susceptible to problems involving oxidation. Although the introduction of 15-20 mM DTT is a good prophylactic measure; some active enzyme is probably still destroyed by oxidation during sonication. 217 The nature of the two forms of diol dehydratase observed during Cellex D chromatography could be further probed by a systemati~ th0rough study of the effects of sonication time and time of burst on the enzyme activity profile. Finally, attempts could be made at constructing an affinity column, using a cobalamin derivative coupled to a solid support. One such attempt could be the following : CN-Cbl + ICH 2 COOH Co 2+ NaBH 4 _,.....-=-=--~~ carboxymethylcobalamin carboxymethylcobalamin + aminoalkylsepharose 0 II carbodi imide • CH 2C-NH-R-NH-sepharose. I >CO< The advantage of this method would be that the enzymecobalamin complex could be liberated by photolysis, and th~ apoenzyme resolved of OH-Cbl by the method described in Chapter I. The last four years have seen considerable improvements in the purification of diol dehydratase. With luck and effort, further ' improvements could be made, perhaps to the point where amounts of enzyme sufficient to do even NMR studies could be isolated. 218 PROPOSITIONS 219 PROPOSITION I The current body of knowledge concerning adenosyl- cobalamin (AdoCbl)-dependent rearrangements has been reviewed in this thesis (Introduction) and in the literature (Babier, 1975a,b; Abeles & Dolphin, 1976). In view of the results described in this thesis, the following investigations are proposed for further studies of the mechanism of AdoCbl-dependent rearrangements. Kinetic Isotope Effects. Rate Determining Step. In the reactions catalyzed by ethanolamine deaminase (Weisblat & Babier, 1971) and diol dehydratase (Moore et al., 1979; Chapter III), evidence has been obtained implicating the second hydrogen transfer step, from C-5' of 5'-deoxyadenosine to C-2 of product aldehyde, as the major rate-contributing step in catalysis. This conclusion differs from that reached by Essenberg et al. (1971), although the reasons for their conclusions are somewhat unclear. In view of this disagreement, and as a complement to the study described in Chapter III, catalysis with [l - 14 CJ-l,2-pro panediol may be studied. If hydrogen abstraction from C-1 of substrate makes a significant contribution to the rate of catalysis, then a 12 C/ 14 C isotope effect should be expressed, resulting in an increase of the 1 4 C-specific activity of unreacted substrate as a function of the amount of substrate consumed. 12 C/ 14 C isotope effects 220 of 1.06 - 1.17 have been observed (Schmitt & Daniels, 1953; Fry, 1970), and are about twice the magnitude of 12 C/ 13 c isotope effects (Stern & Vogel, 1971). 80 - 98 % depletion of substrate, In the range of 14 C enrichment of the residual substrate in the range 10 - 51 % relative to the initial specific activity should be observed for an isotope effect of 1. 06 - 1.12 (Melander, 1960) . Double-isotope counting methods should be useful in this regard. An appropriate standard, (3 - 3H]-l,2-propane- diol, could be synthesized from [ 3 H] - CH 3 I by the procedures outlined by Kollonitsch (1966a,b). Use of this standard, which should be converted to propionaldehyde without an isotope effect, would obviate the need for quantitative recovery of unreacted 1,2-propanediol from the reaction mixture. While a 12 C/ 14 C isotope effect of 1.06 - 1.17 would indicate that the first hydrogen transfer step is significantly rate-contributing, the observation of no or a very small (<l.03) isotope effect would support conclusions based on previous work with isotopes of hydrogen (Chapter III). Tritium Isotope Effect. Diol dehydratase and ethanol- amine deaminase exhibit large isotope effects (kH/kT ~ 160) for transfer of tritium from [S• - 3H]-AdoCbl to product derived from unlabeled substrate (Essenberg et al., 1971; Weisblat & Babior, 1971). In contrast, studies employing 221 deuterium substitution, and measurement of the overall tritium isotope effect (Chapter III) indicate kH/kT ~30- 40 for transfer of tritium from AdoCbl to product. A tri" tium isotope effect of this magnitude (30 - 40) is also indicated by measurements of the specific activity of AdoCbl and unreacted substrate during catalysis with [l - 3HJl,2-propanediol. The discrepancy between these two values of the tritium isotope effect is quite significant in relation to the catalytic mechanism. If kH/kT is in the range 30 - 40, then all known data on primary hydrogen isotope effects with diol dehydratase can be reconciled in terms of the "generally" accepted catalytic mechanism (Fig. 3, Introduction). On the other hand, if kH/kT = 160, then this mechanism represents an oversimplification and must be considerably modified to account for the large tritium isotope effect. Thus, a careful reexamination of this isotope effect is warranted. In the case of propanediol dehydratase (Essenberg et al., 1971), a determination of the number of active sites in a diol dehydratase preparation involved measurement of the amount of tritium transferred in 2 min at 10° C from [5'-3HJ-AdoCbl to propionaldehyde derived from from unlabeled 1,2-propanediol as a function of the amount of [5'-3H]-AdoCbl added to the reaction mixture. At low concentrations of [5'-3H]-AdoCbl, the amount of tritium released to product was dependent upon the [5'- 3HJ-AdoCbl 222 concentration, whereas at high concentrations it was not. Thus, the available sites could be saturated with [5'-3H] AdoCbl, leading to the important conclusion that AdoCbl does not exchange between solution and enzyme at an app reciable rate at 10° C. This result, along with the observed specific activity of the most highly purified enzyme preparation available at that time (60 unit/mg), led to an e.s tima te of a molecular weight of 2 . 5 x 10 5 dal tons. This agrees well with the molecular weight obtained from gel chromatography (Essenberg et al., 1971; Moore & Richards, unpublished results) and subunit composition (Poz nanskaya et al., 1977). However , recent resul ts (Poznan- skaya et al., 1977; Moore, McGee, & Richards, unpublished results) suggest that the actual specific activity of pure diol dehydratase lies in the range 75 - 100 unit/mg. Thus, in order to have obtained a correct molecular weight estimate from the results of the tritium washout experiment described above, Es senberg et al. (1971) may have measured 25 - 60 % more "active sites" (as defined by tritium release from [S'-3H]-AdoCbl) than were actually present . If this were in fact the case , a significant amount of exchange between free and enzyme-bound AdoCbl would have oc curred in 2 min at 10° C. In this event, an even greater degree of exchange would be expected at 37° C where control experiments were carried out fo r the isotope effect in tritium washout. Such exchange would lead to an artif a ctually 223 high estimate of the amount of tritium released to product at infinite time at 37° C, and by comparison, an apparently slow release of tritium as a func tion of time in the determination of the isotope for tritium washout into product propionaldehyde at 10° C. Several control experiments can be carried out to resolve complications in the determination of this tritium isotope effect. Exchange between free and enzyme-bound AdoCbl can be tested in two ways. First, incorporation into propionaldehyde of tritium derived from [ 5' - 3H] -Ado Cb 1 cat~ be measured by adding labeled AdoCbl to a solution containing unlabeled holoenzyme and 1,2-propanediol. The appear- ance of tritium in propionaldehyde would suggest exchange between free and enzyme-bound AdoCbl. Second, free AdoCbl may be isolated from a reaction mixture containing holoenzyme, excess AdoCbl, and [l - 3H]-l,2-propanediol. The observation of tritium associated with free AdoCbl would also suggest exchange. The tritium isotope effect can be measured by allowing enzyme and [5'-3H]-AdoCbl to react with [l -14C]-l,2propanediol. The expected half-life for tritium transfer from [5'-3H]-AdoCbl to product is less than 1 sec; this experiment is best carried out with apoenzyme in excess over [S'-3HJ-AdoCbl and with amounts of [l - 14 C]-l,2-propanediol equivalent to 20 - 150 turnovers per holoenzyme active site. The reaction can be quenched with acid, 224 neutralized, carrier aldehyde added, and propionaldehyde isolated as the semicarbazone for 3H/ 14 c assay. The 3H content of water from which AdoCbl and aldehyde have been removed should be determined to ascertain the degree of exchange of tritium from aldehyde into water. One possible interpretation of the apparently anomalc:,us large tritium isotope effect in the transfer of tritium from AdoCbl to product is the presence of a second reservoir of mobile hydrogen (in addition to C-5' of AdoCbl). Indeed Babior (1979) has presented some pre liminary evidence for such a second reservoir in the ethanolamine deaminase reaction. He found that interaction of [l - 3 HJ-ethanolamine with holoenzyme in the presence of NADH and ADH (to reduce product aldehyde to ethanol as soon as formed) followed by (i) treatment with trichloroaceti c acid to quench the reaction and denature the protein and (ii) removing ethanol by azeotropic distillation, led to tritium being recovered in water as well as in substrate, ethanol, and recovered AdoCbl. If unlabeled ethanolamine was added as a chase to holoenzyme previously exposed to [l - 3HJ-ethanolamine, little or no tritium was found in the water. These resultJ provide a basis for inferring that there exists a reservoir of hydrogen attached to such a site(s) on the protein that they are exchangeable with solvent water after denaturation of the protein; these hydrogens cannot exchange with water during the normal course of the reaction, as 225 under these conditions no tri tium from tritiated substrate or tritiated AdoCbl appears in s olvent (Brownstein Abeles, 1961; Babior, 1969). Such & a reservoir could ex- plain the anomalously large tritium isotope effects. The majority (~so%) of catalytic events would involve the reservoir as intermediate hydrogen carrier, while the 5'carbon of AdoCbl would participate the remainder (~20 %) of the time. Thus the factor of five discrepancy between the predicted (30 - 40, Chapter III) and observed (160) tritium isotope effects would be of a statistical nature. Experiments are proposed to e::icamine the possib il ity that such a second reservoir for migrating hydrogen exists in the case of diol dehydratase. A key to the proper exe- cution of these experiments is to separate any product or unreacted substrate from holoenzyme without oxygen inactivation of the holoenzyme and without exchange with solvent of any tritium acquired by the putative second able" reservoir. "exchange- To accomplish this purpose, one can carry out a reaction with [l - 3H, 14 C]-l,2-propanediol as sub .s tr ate. The resulting solution ~.a n then be passed down a BioGel P-60 column in a buffer containing a satura ting concentration (> 1 mM) of dl-2, 3-butanediol, which functior;.s as a purely competitive inhibitor (Chapter I; Moore & Richards, 1979). Labeled 1, 2-propaned:f.ol and aldehyde will be retarded by the column while the holoenzyme-dl- 2,3-butanediol complex will elute in the void volume. The 226 resulting enzyme-inhibitor compl e:K would be denatured with heat or trichloroacetic acid and lyophilize.d. To determine the amount of exchangeable hydrogen assoc.i.ated with prote!-Ln, 3H the distillate can be assayed for (a.s well as sure that no diol or aldehyde is present). 11+c to e:.1- A possible wav for aldehyde still to be present in the protein fraction is by formation of a Schiff's base with free amino groups of lysine residues. The possibility of such aldehyde -protein derivatives is suggested by the observation that [14C]-e-hydroxypropionaldehyde and [ 14 C]-acetaldehyde (Moore & Richards, unpublished results) extensively label diol dehydratase during the enzymatic reaction. Treatment of this protein with tri- chloroacetic acid would then hydrolyze these Schiff's bas<:; functionalities, releasing aldehyde to solution, where it could by enolization release its hydrogen to water. Were this the source of any tritium observed in the disttllate, 14 C should be present also. Another approach is to reduce the enzyme-dl-2,3-butanediol complex with KBH 4 to convert any imine groups to alkyl amines, which would be stable to hydrolysis by trichloroacetic acid. (The NADH/ADH reducing system does not accomplish this conversion,) the di.,tillate contains 3H and lf in the first case~ 14 C an d in the second case does not, one would know the source of radioactivity was aldehyde-enzyme Schiff 1 s base derivatives. second case 3 H is If in the still present in the distillat e (and 227 especially so if no 14 C was present :Ln the distillate in the first case) one could conclude that some exchangeable site on the enzyme has acquired tritium. An additional useful experiment would then be measurement of the rate of transfer of this "exchangeable" tritium to product aldehyde by proc edures analogous to those de scribed previously for measuring the. ·w ashout of tritium from [5'- 3 H]-AdoCbl by catalysis with unlabeled substra te. If such an alternate hydrogen reservoir does serve in a majority of turnovers to store migrati:o.g hydrogen, one should observe a nearly normal kinetic isotope effect for transfer of tritium from this r e servoir to product. The possibility that label appears in solvent via Schiff's base formation between protein and aldehyde can also be tested by -:-~arrying out the diol dehydratase reac tion in the presence of some other protein (albumin, myoglobin, etc .), isolating this protein by gel chromat o graphy , and subjecting it to analysis for exchangeable tritium as outlined above. Protein. We have obtained evidence that inacti.vatioT.1 by oxygen and glycerol involves alteration of the protein as well as of AdoCbl (Bachovchin et al . , 19 77; Chapter I), and from these results infer that onr;~ or more am:Lno a cid residues in the active site are essential for catalytic activity. To determine which residues are at or near the active 228 site , studies to identify those th a t a r e al tered dux·j_ng inactivation by glycerol or oxygen a r e propos ed . Ho lo- enzyme which has been inactivated in one o f thes e ways • d. proteo 1 yt1.ca • 11 y or • c ·1... e avec.1 w1.t • h cy anogen can b e d 1.geste bromide , and the peptide mixture s ubje c t e d t o chromate . graphy in one dimension and electrophoresi s in ·a s econd . The resulting peptide maps can b e c ompar e d to those ob tained from apoenzyme or fr om hol oen zyme which has be en exposed to 1,2- propanediol and then treat ed j ust as the :i.11activated enzyme before being cleave d into f r agments. A similar procedure could be applied to the separated subuni ts (Toraya et al., 1973; Poznanskaya et al., 1977) of inacti -• vated holoenzyme in an attempt to localize any altered amino acid residues to a particular subunit. Ultimately, isolation, followed by sequence analysis of the altered peptides should identify those residues involved in inactivation, and therefore presumably near the cataytic site and likely to play a significant role in cat alysis. While glycerol-inactivated holoenzyme present s s ome specia l problems in this regard, due to extensive label i ng of pro - tein by s-hydroxypropionaldehyde , we have reduced the de gree of this nonspecific labeling by a f a c t or of t en or more by such strat egies as t r apping t he a l dehyde produc t by KBH added to the reaction mi xtur e . I t :Ls cer tain 4 that the inactivation of holoen zyme b y glycerol doc~s not result from interaction of the pro t e i n wi t h produc t 229 s-hydroxypropionaldehyde (or its dehydrat i on product a cr olein) because of the stereospeci fi city of i nac tivat i on relative to catalysis. Furthe rmor.·e, fr ee s - hy droxyp r o- pionaldehyde added to the rea c tio n mixture does not ina ctivate holoenzyme . Another method for ident i fy ing resi due s at or n ear the active site is provided by the photo l ability o f alkyl cobalamins (Hogenkamp , 19 7 5) . Accordingl y [ 1 '+ C] - methyl-, ethyl-, and carboxyrnethyl- (et c .) c ob a lamins may serve as useful active-site directed photoaff in i.ty labels. Suer, derivatives can be readily prepared by a l kylation of cob (T) alamin with the appropriate [ 1 t1 C] - alkyliodide (Do l ph i n , 1971). Binding of these modified forms of AdoCb l, which are competitive inhibitors (Lee & Abeles, 1963) , to apoenzyme followed by photolysis of the c ar bon-cobalt bond should ·generate reactive primary alky l radi cals which may then alkylate nearby amino acid residues , or modi fy t hem in a way which would be de tee tab le b y t he t e chni ques outlined above. Appropriate control exper iments involving competitive inhibition of this photoa f fi nity labe l i ng by binding of photochemically i nert OH- Cbl or by unl abe l e d alkylcobalamin should also be carr i ed out. NMR Studies. Recent advance s in t he pur if ic at i on of diol dehydratase (Appendix) make poss ib l e s tudies of this enzyme by nuclear magnetic resonan ce. Curr entl y , concen - trations of active sites on t h e orde r of 30 - 40 µM 230 are attainable; possibly, concentrations of up to 0.1 mM may be achieved in the near future. Useful in.formati.cn could be gained by studies involving the following nucle:i : 19F, 31P, and 205Tl+. 19 (i) F. 3,3,3-Trifluoro- l,2=propanediol binds to diol dehydratase, and undergoes both catalysis and inactivation , although at greatly reduced ra.tes compared to 1,2-proparne• diol and glycerol, respectively . The three 19 F nuclei a.t C-3 of this substrate analog constitute a very sensitive probe for NMR studies of interactions of substrate with e ;.L zyme, both in the presence and absence of analogs of AdoCbl . · As the reaction of this substrate with holoenzyme is rapid on the time scale of the NMR experiment, studies of the holoenzyme-3,3,3-trifluoropropanediol complex itself will probably not be feasible (except for the inactJ:vated complex). However, binding studies such as those reported by Hull et al. (1975) with ethanolami.ne deaminase should be informative about the interact i on of enzyme with substrate when carried out in the presence of OH-Cbl , CN- Cbl, Me-Cbl, or other catalytically inactive cobalamins. ( l•. 0 l. ) 31 P . R 31 P T,lMR • • t l of . ecent 1 y, a pre 1 1.minary ~ s_ucy cobalamin has been carried out (Satter lee, 1979) . The 31 r resonance of the phosphate group in the nucleotide side chain (Introduction, Fig. 1) i.s sensitive to the ionization state of cobalt, to pH changes , and to whe ther or not the benzimidazolyl nitrogen is coordinated to cobalt . 231 Moreover, the 31 P resonance is both shifted and broadened when the cobalt atom is in the Co 2+ oxidation state and carries an unpaired electron. Thus, 31 P NMR should be a valuable tool for investigating the binding of c.o balamins to diol dehydratase. In parti.c ular, the ionization state of cobalt in the inactive holoenzyme complexes formed by glycerol and thioglycerol , which exhibit Cbl III and Cbl II uv-visible spectra, respectively, could be further probed by 31 P NMR. The importance of the phosphate group to binding of Cbl to enzyme could also be studied, since the chemical .shift of 31 P is sensitive to the involvement of phosphate oxygen in hydrogen bonds (Moon & Richards, 1973; Goetze & Richards, 1978). However, the relatively low sensitivity of 31 P (~360 times that of naturally abundant 13 c) will require accumulation of numerous free-induction decays, so study of rapid reactions such as catalysis will not be feasible. (iii) 205 Tl+. Diol dehydratase requires a monovalent cation for activity; this cation seems to play a role in binding of AdoCbl to apoenzyme (Toraya et al., 1971; Toraya & Fukui, 1972). Maximum enzyme activity is observed 0 with cations having ionic radii of about 1.4 A (K+, NH/4 Rb+, Tl+) (Toraya et al., 1971). Tl+ i s known to mimic potassium in a variety of biological processes (Britten & Blank, 1968; Manners et al., 1970), and has two spin½ . 203 ' 205 Tl. isotopes, 205 Tl has a sensitivity about twice 232 31 that of P (which results in a time saving of a factor of four in obtaining a spectrum), is 70.5 % na.turally abundant , and has a chemical shift range of over 5000 ppm (Schneid <~r & Buckingham, 1962). of The use of ZOSTl as an NMR diol dehydratase thus seems attrac t ive . probe Binding . could be studied by this method, as of 205 Tl+ to protein could binding of Cbl derivatives to the protein- 205 Tl+ complex. Conceivably, the ZOSTl+ resonance chemical shift may also be sensitive to inactivation of holoenzyme by oxygen and by glycerol and thioglycero l as de.scribed above for 31P; these results could be compared to data obtained with the competitive inhibitor dl-2,3-butanediol> which does not cause inactivation or cleavage of the carbon-cobalt bond of enzyme-bound AdoCbl (Chapter I). With ZOSTl+ in this system, chemical shift measurements will be considerably more feasible than measurements of r(·laxation times. For example, Chan & Reeves (1974) and Bangerter (1977) have determined that ZOSTl+ relaxation times are extremely sensitive to the presence of .small amounts of dissolved oxygen. 233 REFERENCES Abeles, R. H., & Dolphin, D. (1976) Acc. Chem. Res. 9, 114. Babior, B. M. (1969) J. Biol. Chem. 2l~4, 449. Babior, B. M. (1975a) in Cobalam;in, B. Babier. Ed., New York, N.Y ., Wiley and Sons, Inc,, Chapter 4. Babior, B. M. (1975b) Acc. Chem. Res. 8, 376 . Babior, B. M. (1979) Results presented at the Sixth Enzyme Mechanisms Conference in La Jo],la, CA, January 2-5. Bachovchin, W. W. , Eagar, R. G., Jr., Moore, K. W. • & Richards, J. H. (1977) Biochemistry 16, 1082. Bangerter, B. W. (1977) J . Mag. Res. 28, 141. Britten, J. R., & Blank M. (1968) Biochim. Biophys. Acta 159, 160. Brownstein, A. M., & Abeles, R.H. (1961) J. Biol. Chem . 236, 1199. Chan, S. 0., & Reeves, L. W. (1974) J . Arn. Chem. Soc . 96, 404. Dolphin, D. (1971) Methods Enzymol. 18c, 3l~. Essenberg, 11. K., Frey, P.A., & Abeles , R.H. (1971) J. Am. Chem. Soc. 93, 1242. Fry, A. (1970) in Isotope Effects in Chemical Reactions, C. J. Collins and N. S. Bowman , Eds., New York, N.Y. Van Nostrand Reinhold, Chap ter 6. Goetze , A. M., & Richards, J. H. (1978) Biochemistry !I_, 1733. 234 Hogenkamp, H. P. C. (1975) in Cobalamin, B. Babior, Ed., New York, N.Y., Wiley and Sons, Inc., Chapter 1. Hull, W. E., Mauck, L., & Babior, B. M. (1975) J . Biol. Chem. 250, 80_2 3. Kollonitsqh, J. (1966a) J. Chem. Soc . A, 453. Kollonitsch, J. (1966b) J. Chem. Soc . A, l~56. Lee, H. A., Jr., & Abeles, R.H. (1963) J. Biol. Chem. 238, 2367. Manners, J. P., Moralle e, K. G., & Williams, R. J. P. (1970) Chem. Commun. 16, 965. Melander, L. (1960)· Isotope Effects on Reaction Rates, New York, N.Y., Ronald Press, Chapter 3. Moon, R. B., & Richards, J. H. (1973) J. Biol. Chem. 248, 7276. Moore, K. W., & Richards, J. H. (1979) Biochem. Biophys. Res. Commun, in press. Moore, K. W., Bachovchin, W. W., Gunter~ J. B., & Richards, J. H. (1979) Biochemistry, i n press. Poznanskaya, A. A., Tanizawa, K., Soda, K., Toraya, T., & Fukui, S. (1977) Bioorg. Khim. 2, 111. Satterlee, J. D. (1979) J. Am. Chem. Soc . , 101, submitted . Schmitt. J .. A., & Daniels, F. (1953) J. Am. Chem. Soc. 75, 3564. Schneider, W. D., & Buckingham, A. D. (1962) Disc. Faraday Soc. 34, ll~7. 235 Stern, M. J., & Vogel, P . C. (1971) J. Chem. Phys. 55, 2007 . Toraya, T . , & Fukui, S. (1972) Biochim. Biophys. Acta_ 284 , 536. Toraya, T., Sugimoto, Y., Tamao, Y. , Shimizu, S., & Fukui » S. (1971) Biochemistry 10, 3475. Toray a, T., Uesaka, M., Kondo, M., & Fukui, S. (1973) Biochem . Biophys. Res. Commun. 52, 350 . Weisblat, D. A., & Babior, B. M. (1971) J. Biol. Chem. 246, 6064. 236 PROPOSITION II Cancer fatalities are rarely caused by a single, or primary tumor; it is the proliferat ion of mul t iple tumors throughout the body which most often causes death. This ability of tumor cells t.o colonize distant parts of the body remote from the main tumor mass is called metastasi s, and is the most deadly aspect of cancer. At present little is understood about the various processes involved in metastasis, such as tumor cell detachment, infiltration of and transport through the circulation and lymphatic systems, and invasion and colonization of the secondary sites. These mechanisms are the subject of recent reviews (Roos C::.( Dingemans, 1979; Nicholson, 1979) . Introduction into the organism of microbial stimulating agents, such as Bacillus Calrnette-Guerin_ (BCG) and Corynebacterium parvum (C. parvum) or£. granulosum, under suitable conditions appears to inhibit formation of metastases and even causes regression of the primary tumor mass. The effect is observed with either living or killed BCG or£ . parvurn preparations (Baldwin et al., 1974). In most cases, these observations have b een confined to animal cancer models, but use in immunotherapy of human malignant melanoma has been reported (Gutterman et al., 1973). Injection of a BCG suspension into a primary tumor 237 results in a decreased level of metastasis (Zban et al., 1972; Baldwin & Pirrnn, 1973a; Sparks et al.. 1974; Smith et al., 1975; Lee, 1977), and in some cases in regression of the primary tumor (Bomford & Olivotto, 1974). Appar- ently, contact of tunor cells and BCG is required for the effect to be observed; intravenous administration of BCG seems effective only against the appearance of tumors in the lung, where BCG accumulates (Baldwin & Pimm, 1973b, 1974). BCG appears to enhanc.e an immune response against tumor-specific antigens, and is relatively more effective against tumors which are antigenic (Baldwin & Pimm, 1973a). However, macrophages have also been implicated in this effect (Hanna et. al., 1972). In contrast, anti-metastatic activity by C. parvum does not seem to require cell-cell contact; both intraven•ous and intraperitoneal administration are effective (Jones & Castro, 1977; Sadler & Castro, 1976a,b; Bamford & Olivotto, 1974). Evidence has been obtained implicating a role for circulating T-cells in this effect (Sadler & Castro, 1976b; Milas et al., 1975), although an intact thymus may not be necessary. Conflicting results have been obtained on the role of macrophages (Jones & Castro, 1977 ; Bamford & Olivotto, 1974). The mechanisms responsible for these effects are poo r ly understood. BCG and C. Earvum may potentiate an immune response to tumor antigens, or cause an enhancement of 238 "nonsp~cific resistance" (Bomford & Olivotto , 1974) to t he cancer cells. Conceivably, these micr obial agents may challenge the host's irrrrn.une system with antigens which somehow mimic those on cancer c.e lls. but are significantly more antigenic, thus enhancing an irrrrnune response against the tu.~or cells. Some experiments are proposed which should help to resolve these questions . First, the "antigen mimicry" hypothesis may be tested by immunizing tumor-free organisms wi th BCG or C. parvum and then testing the resulting ant i sera for cross-reaction against the appropriate · cancer cells. The immunized animals can then be tested for the ability of these microorganisms to suppress metastasis of implanted tumors. Detection of immunologically cross-reactive antigens present in both bacteria and cancer cells would suggest a possible mode . of action for the observed suppression of metastasis . In the event of a positive result, one might predict that the immunized animals would display an overall decreased frequency of metastasis, which would probably not be further depressed by introduction of BCG or C. parvum. A negative result, that no "antigen mimicry" is observed , would suggest tha t this direct stimulation of th e immune system is not a factor in the observed antime t astat i c behavior . The possibility that administration of these mi cro organisms effects a stimulation of the immune system, thereby rendering it more competent to deal with cancer 239 cells, can be tested in the following way. Fetal mice or neonatal rats which are not yet immunologically mature may be tolerized to BCG or£. parvum by the techniques described by Billingham et al. (1953; 1955a,b). Fetal mice are best exposed to the antigen at the 16th or 17th day of fetal life (Billingham et al .. 1955b); while neonatal rats seem to yield better results than neonatal mice (Woodruff & Simpson, 1955). The tolerant organisms may then be employed to determine the effect of BCG and C. parvum on metastasis of tumors subsequently introduced. If the mechanism of metastasis suppression in any way involves the recognition by the iunnune system of BCG and C. parvum antigens as "foreign," then these microorganisms should have no effect upon proliferation of tumors in the tolerant host animals; such an effect should be observed in control animals which were injected with antigen-free solutions during the fetal or neonatal period. The observation that BCG and C. parvum could still inhibit tumor metastasis in tolerant organisms would seemingly rule out any role in this process for those components of the immune system involved in the recognition of "foreign" antigens. Several host-tumor systems are available for the pro -posed studies in which the antimetastatic activity of BCG and C. parvum has been demonstrated: (C57Bl/Rij x CBA/Rij ) Fl mice and osteosarcoma C22LR (van Putten et al., 1975); 240 CBA-T6T6 mice and T3 fibrosarcoma (Bomford & Olivotto, 1974); Sewall-Wright guinea pigs and a transplantable hepatocellular carcinoma (Smith et al., 1975) are a few examples. The effect appears to be general; and any host ·· tumor model in which it can be demonstrated should be adequate for the proposed study. The proposed experiments should clarify the mechanism of the antimetastatic effects of BCG and C. parvum; in any case, the current confusing picture should be improved. 241 REFERENCES Baldwin, R. W., & Pimm, M. V. (1973a) Br. J. Cancer 27, 48; (1973b) Int. J. Cancer 12, 420; (1974) Br. J. Cancer 30, 473. Baldwin, R. W., Cook, A. J., Hopper, D. G., & Pimm, M. V. (1974) Int. J. Cancer 13, 743. Billingham, R. E., Brent, L., & Medawar, P. B. (1953) Nature Lond. 172, 603; (1955a) Ann. N.Y. Acad. Sci. 59, 409; (1955b) Proc. Roy. Soc. Lond. 239, 357. Bomford, R., & Olivotto, M. (1974) Int. J. Cancer 14, 226 . Gutterman, J. U., Mavligit, G., & McBride, C. (1973) Lancet, 1208. Hanna, M. G., Zban, B., & Rapp, H.J. (1972) J. Natl. Can cer Inst. 48, 1441. Jones, P. D. E., & Castro, J. E. (1977) Br. J. Cancer 35, 519. Lee, Y. T. N. (1977) Cancer Res. 37, 3679. Miles, L., Kogelnik, H. D., Basic, I., Mason, K., Hunter, N., & Withers, H. R. (1975) Int. J. Cancer 16, 738. Nicholson, G. L. (1979) Sci. Amer. 240, No. 3, 66. Roos, E., & Dingemans, K. P. (1979) Biochim. Biophys. Acta 560, 135. Sadler, T. E., & Castro, J. E. (1976a.) Br. J. Surg. 63, 292; (1976b) Br. J. Cancer 34, 291. Smith, H. G., Bast, R. C., Jr., Zban, B., & Rapp, H. J. (1975) J. Natl. Cancer Inst. 55, 1345. 242 Sparks, F. C., O'Connell, T. X., Lee, Y. T. N., & Breeding , J. H. (1974) J. Natl. Cancer Inst. 53, 1825. van Putten, L. M. , Kram, L. K. J., van Dierendorck, H. H. C. , Smink, T., & Fuzy, M. (1975) Int. J. Cancer 15, 588. Woodruff, M. F. A., & Simpson, L. 0. (1955) Brit. J . Exp. Path. 36, 494. Zban, B., Bernstein, I. D., Bartlett, G. L., & Hanna, M. G. (1972) J. Natl. Cancer Inst . 49, 119. 243 PROPOSITION III The complement system of nine serum proteins (Cl-C9) is an important component of the immune system of vertebrate organisms. These proteins participate in several reactions involved in defense of the organism, such as inflammation and cytotoxicity. The complement proteins are produced as nonfunct-i.onal precursors which under suitable conditions may become ac tivated and thus enabled to participate in complement-me diated reactions . The final portion of the complement reaction sequence may be preceded by either the classical activation pathway (Muller-Eberhard, 1969), which involves interaction of the antibody-antigen complex with Cl (specifically, Clq), or the alternate (or "properdin") pathway, whose origins are less well characterized (Gotze and MullerEberhard, 1971; Muller-Eberhard & Gotze, 1972 ; Arrogave , 1972). Two recent reviews summarize current knowledge of the activation process (Porter, 1979; Reid & Porter , 1975). The classical activation pathway is initiated by interaction of the complement protein Clq with immunoglobulins IgM, and IgG subclasses IgGl, IgG2., and IgG3 (Augener et al., 1971). Although a Clq molecule can bind up to s ix Ig molecules, as few as two are required to initiate the classical activation pathway (Borsos & Rapp, 1965). Whether an antigen-induced conformational change 244 in the Ig heavy chain region (Schur & Christian, 1965; Press, 1975), or simple aggregation of antibody molecules (Hyslop et al., 1970, Augener, 1971), or both, are required for Clq bindin3 is currently unresolved. Nonetheless, evidence has been accumulated implicating the existence of specific sites of interaction, in the CH2 domain of IgG (Kehoe & Fougerean, 1969; Utsumi, 1969; Ellerson et al., 1972) and Feµ of IgM (Plaut et al., 1972). In human IgGl and rabbit IgG, amino acid residues between 271 and 291 of the CH2 region appear to be involved in complement fixation. Allen & Isliker (1974a,b) showed that modification of rabbit IgG and its Fe fragment with the tryptophan-directed reagent 2-hydroxy-5-nitrobenzyl bromide decreased complement-fixing activity, despite the observation that this modification di.cl not detectably alter the Clq binding capacity, and suggested that tryptophan at position 277 is important in complement fixation. Johnson & Thames (1976) modified IgGl with reagents directed against tryptophan (2~hydroxy-5-nitrobenzyl bromide), tyrosine (tetranitromethane), and arginine (2,3-butanedione); the complement-fixing abilities of these modified proteins were reduced by up to 77%. These workers also synthesized peptides with amino acid sequences similar to those of the Ig heavy chain region suspected to be involved in complement fixation, and determined the extent to which these peptides could 245 inhibit complement fixation by IgGl. A particularly effec- tive inhibitor was GlyTrpTyrGluArgGly, (I) which may be compared to the CH2 region sequence AsnTrpTyrValAspGly 276 281 (II) (Edelman, 1970). Interpretation of chemical modification studies such as those described above is complicated by uncertainty of the effect of a given modification. Alteration of amino acids. rriay have a direct effect on the interaction under study, or an indirect effect: a change in the conformation of the site of interaction could be caused by modification of an amino acid elsewhere in the protein. In order to resolve these ambiguities, and to investigate the nature of the antibody-Clq interaction, an investigation of the inhibitory behavior of peptides such as (I) which have been chemically modified is proposed. Of par- ticular utility should be peptides containing 3-nitrotyrosine produced by nitration with tetranitromethane. This procedure is selective, simple to execute, and resulls i n a chromophore which can be conveniently exploited for concentration assay purposes (Means & Feeney, 1971). Furthermore, the resulting nitrated tyrosine has a pK of 7.0-8.0 for deprotonation of the phenolic hydroxyl group. 246 Thus, a study of inhibition by such a modified peptide of complement fi x ation over the pH range 6-8.5 should be informqtive abou t the role played by the tyrosyl residue in binding of Clq. Moreover, a similar study of tetnanitro - methane-treated IgGl may also be valuable . The experiments described by Johnson & Thames (1976) were carried out at pH 7.4, and one or more of the seven modified tyrosine residues may have been ionized at physiological pH (manipulation of conditions of the nitration reaction could conceivably decrease the number of altered amino acids). Methods for assay of complement activity are well known (Kabat & Mayer, 1961; Lachmann & Hobart, 1978), as are those for peptide synthesis (Finn & Hofmann, 1976; Erickson & Merrifield, 1976). Peptides could perhaps also be obtained from a proteolytic or chemical digest of the ~ . portion of IgGl . Control experiments should also ~e carried out to determine the extent of complement fixation via the alternate pathway by assaying for the consumption 2+ of C 2 or by repeating the assay in the absence of Ca , which is required by the classical pathway. The pH-dependent behavior of complement fixation by IgGl and its inhibition by the peptide inhibitors which con tain nitrated tyrosine could illuminate the nature of the interraction, between Clq and the heavy chain constant region of irnmunoglobulins, which initiates the classical pathway of complement fixat ion . 247 REFERENCES Allan, R., & Isliker, H. (1974a) Immunochern. · 11, 175; (1974b) Imrnunochem. 11, 243 . Arrogave, C. M. (1972) Fed. Proc. 31, 659. Augener, W. , Grey, H. M. , Cooper , N. R. , :$c Muller-Eberhard, H.J. (1971) Irmnunochem. 8 , 1011. Borsos, T., & Rapp, H. J. (1965) Science 150, 505. Edelman, G. M. (1970) Biochemistry 9 , 3197 . Ellerson, J. R., Yasmeen, D., Painter, R.H . , & Dorrington . K. J. (1972) FEBS Lett. 24 , 318. Erickson, B. W., & Merrifield, R. B. (1976) in The Proteins, 3rd ed., H. Neurath, R. L. Hill, and C.-L. Boeder, Eds.,' New York, N. Y., Academic Press, II, p. 255. Finn, F. M., &Hofmann, K. (1976) in The Proteins, 3rd ed. ,, H. Neurath, R. L. Hill, and C.-L. Boeder, Eds., New York, N. Y. , Academic Press, II, p. 105. Gotze, 0., & Muller-Eberhard, H.J. (1971) J. Exp. Med. 134, 90s. Hyslop, N. E., Dourrnashkin, R. R., Green, N. M., & Porter, R. R. (1970) J. Exp . Med. 131, 783 . Johnson , B. J., & Thames , K. E. (1976) J. Irnrnunol. 117, 1491. Kab at, E. A., & Mayer, N. M. (1961) Experimental Im- • munochemis tE.Y., Springfield, 111. , Charles C. Thomas , p. 149 . Kehoe, J.M., & Fougereau, M. (1969) Nature 224, 1212 . 248 Lochmann, P. J. , & Holbart, M. J. ( 19 78) in Handbook of Experimental Itnmuno~ogy, D. M. Weir, Ed., Oxford, Blackwell Scientific Publications, Chapter Sa. Means, G. E., & Feeney, R. E. (1971) Chemical Modification of Proteins, San Francisco, CA, Holden-Day, Inc., pp. 183-196. Muller-Eberhard, H.J. (1969) Ann. Rev. Biochem. 38, 389. Muller-Eberhard, H. J., & Gotze, 0. (1972) J. Exp. Med. 135, 1003. Plaut, A. G., Cohen, S., & Tomasi, T. B., Jr. (1972) Science 176, 55. Porter, R.R. (1979) Int. Rev. Biochem. 23, 177. Press, E. M. (1975) Biochem. J. 149, 285. Reid, K. R. M., & Porter, R.R. (1975) in Contemporary Topics in Molecular Immunology, F. P. Inman and W. J. Mandy, Eds., New York, N. Y., Plenum Press,~, 1. Schur, P.H., & Christi.an, G.D. (1964) J. Exp. Med. 120, 531. Shimizu, A., Paul, C., Kohler, H., Chinoda, T., & Putnam, F. W. (1971) Science 182, 287. 249 PROPOSITION IV The innnune system is one of the most complex biological systems, the molecular, genetic, and cellular aspects of which have been studied intensively. The immunozlobulin molecule is composed of two light and two heavy polypeptide chains, each of which may be divided into an NH 2-terminal variable (V) and one (light chain) or three (heavy chain) constant regions (C). Three unlinked families of genes code for innnunoglobulins; one codes for the V and C regions of heavy chains, while two(>- and K) code for those of the light chains (Hood et al., 1975). The V and C regions are encoded by separate genes which become joined to form a single gene (Dreyer & Bennett, 1965). Until recently, immunogeneticists were limited to serological and protein sequence studies in their investigations of antibody genes. With the introduction of recom- binant DNA technology, the tools are now available for direct examination of the organization of antibody genes in the DNA of both antibody-producing and undifferentiated cells. Studies of immunoglobulin light chain genes with prob es derived from purified light chain rnRNA generally confirm the V-C joining hypothesis (Leder et al., 1977; Rabbits et al ., 1975; Hozumi & Tonegawa., 1976; Seidman & Leder, 1978). Hozumi & Tonegawa (1976) showed that the 250 mouse VK and CK genes lie on different DNA restriction fragments in eariy emhryonic tissues, while in the mature (differentiated) antibody-producing myeloma from which the roRNA was obtained, these genes lie on the same restriction fragment (in both chromosomes). A particularly intriguing result was that even in a differentiated myeloma cell, the VA and CA genes remain separated by a 1250 base pair se- quence; this sequence is absent from the mRNA (Brack & Tonegawa, 1977). Such intervening sequences, termed "introns" (Gilbert, 1978), have also been observed in other gene systems, such as K-light chains (Rabbit ts & Forster, 1978), S-globin (Jeffreys & Flavell, 1977), ovalbumin (Breathnach et al., 1977), and yeast tRNA (Goodman et al., 1977; Valenzuella et al., 1978). Recent evidence (Rabbitts, 1978) indicates that these introns disappear during the processing of the initial RNA transcript of the mouse light chain genes. Because of difficulties in preparation of the desired mRNA, the organization of innnunoglobulin heavy chain genes has only recently been studied (Honjo & Kataoka, 1978; Early et al., 1979; Sakano et al. 1 1979). The latter two studies , of IgA and IgGl heavy chains, respectively, indi cate that the CHl, CH2, CH3, and hinge regions of the heavy chain constant region arc each encoded by uninterrupted DNA sequences which are separated by introns of 100- 251 350 base pairs. In the IgA case (Early et al., 1979) the V and C genes are separated by 6.8 kilobases in embryonic DNA, and rearrangement apparently does occur upon differentiation . The function of the intrans remains unclear, although current evidence favors a role in splicing of the primary nuclear RNA transcript. In a large majority of intrans characterized so far, the intron-exon junction may be chosen so that the intron begins with G-T and ends with A-G (Bernard et al., 1978; Breathnach et al., 1978; Reddy et al., 1978; Van den Berg et al., 1978; Sakano et al., 1979 ). This observation suggests that splicing sites may be quite similar in vastly different species. For the purposes of both further characterization of immunoglobulin heavy chain genes, and examination of the structure of introns within a gene family, studies of the structures of the µ(IgM), o(IgD), and E(IgE) heavy chain constant region genes of the mouse are proposed. In addition to valuable information on the arrangement of these genes in the mouse genome, a comparative study of the intron sequences may provide insights into their function , in evolutionary and structural terms. For example, the observation that intron sequences between CH regions are highly conserved, while the CH µ,y,u, O,e: coding regions are considerably more diverse, might suggest that quite different s elective pressures operate on the coding 252 and intervening sequences. On the other hand, if the intron sequences in the CH gene family differ considerably, one might speculate that perhaps, with the exception of the intron-exon junction region, sequence plays a secondary role to that of size, and mutational drift could be the . source of the sequence differences. The appropriate mRNA is required as the source of a probe for the CH genes. A convenient source of this mRNA would b.e a myeloma cell line secreting antibody of the desired class. IgA, IgG, and IgM-secreting mouse myeloma cell lines are readily available. An IgD-producing rat myeloma tumor could be used; its heavy chain mRNA should cross-hybridize with mouse C0 genes. The also technique of cell hybridization to produce myeloma-like cell lines secreting monoclonal antibody of known antigen speci:_ ficity (Kohler & Milstein, 1975) can also be used; IgM and IgG-secreting mouse "hybridoma" cell lines specific for the dinitrophenyl group have been produced in this laboratory · (R. R. Hardy, personal communication) . Recently a monoclo- nal IgE-producing cell line with specificity for ovalburnin has been reported (Bottcher et al., 1978). These induced and hybrid myeloma cell lines should provide a ready source of heavy-chain mRNA for the proposed studies. 253 REFERENCES Bernard, 0., Hozumi, N., & Tonegawa, S. (1978) Cell 15, 1133. Bottcher, I., Hammerling, G., & Kapp, J.-F. (1978) Nature 275, 761. Brack, C., & Tonegawa, S. (1977) Proc. Natl. Acad. Sci. USA 7'•, 5652, Breathnach, R., Benoist, C., O'Hare, K., Gannon, G., & Chambon, P. (1978) Proc. Natl. Acad. Sci. USA 75, 4853. Breathnach, R., Mandel, J. L., & Chambron, P. (1977) Nature 270, 314. Dreyer, W. J., & Bennett, J.C. (1965) Proc. Natl. Acad. Sci. USA 54, 864. Early, P. W., Davis, M. W., Kaback, D. B., Davidson, N., & Hood, L. Gilbert, W. (1979) Proc. Natl. Acad. Sci. USA 76, 857. (1978) Nature 271, 501. Goodman, H. M., Olson, M. V., & Hall, B. D. (1977) Proc. Natl. Acad. Sci. USA 74, 5453. Hood, L. E., Campbell, J., & Elgin, S. (1975) Ann. Rev. Genetics 2_, 305. Honjo, T., & Kataoka, T. (1978) Proc. Natl. Acad. Sci. USA 75, 2140. Hozumi, N. , & Tonegawa, S. (19 76) Proc. Natl. Acad. Sci. USA J..1, 3628. Jeffreys, A. J., & Flavell, R. A . . (1977) Cell 12, 1097. 254 Kohler, G., & Milstein, C. (1975) Nature 256, l~95. Leder, P., Honjo, T., Seidman, J., & Swan, D. (1977) Cold Spring Harbor Symp. Quant. Biol. 41, 855. Rabbitts, T. H. (1978) Nature 275. 291. Rabbits, T. H., & Forster. A. (1978) Cell 13, 319. Rabbitts, T. H., Jarvis, J. M., & Milstein, D. (1975) Cell 6, 5. Reddy, V. B. , Thinnnapaya, B. , Dhar, . R., Subramanian, K. N. , Zain, B. S., Pan, J., Ghosh, P. K., Celma, M. L., & Weissman, S. M. (1978) Science 200, 494. Sakano, H. , Rogers, J. H. , Huppi, K. , Brack, C. , Trauneck~r, A., Maki, R., Wall, R., &Tonegawa, S . . (1979) Nature 277, 627. Seidman, J. G., & Leder, P. (1978) Nature 276, 790. Van den Berg, J., van Ooyen, A., Mantei, N., Schambock, A., Grosveld, G., Flavell, R. A., &Weissman, C. (1978) Nature 276, 37. Valenzuella, P., Venegas, A., Weinber.g, · F., Bishop, R., & Rutter, W. J. (1978) Proc. Natl. Acad. -Sci. USA 75, 190. 255 PROP OS IT ION V The explosive growth of the chemical industry in re_. cent decades has played a major role in technological and agricultural development. However, the accompanying re- lease into the environment of large amounts of synthetic, nonbiological compounds is of major concern, particularly in regard to their potential carcinogenicity and activity as mutagens (Mccann et al., 1975; Mccann & Ames, 1976). Recently, the Department of Health, Education and Welfare has advanced the theory that as much as 40% of all cancer may soon be caused by industrial or environmental exposure to carcinogens. Of particular potential danger are those substances which are not degraded by bacteria and fungi, or are lipophilic and thus become concentrated in the fat depots of organisms of the food chain (for . example, dioxin, DDT, and halogenated aromatics). The techniques of bacteri~l mutagenesis, especially recent developments (Bachovchin & Roberts, 1977) can be employed to "create" new strains of microogranisms whose mutant enzymes are capable of converting a target compound to less toxic or biodegradable forms. Such organisms could then be used directly to deto~ify these compounds. Techniques are available which could be used to generate mutant bacteria with a desired characteristic. In general, this procedure would involve treatment of the wi.ldtype organism (the soil bacteria Pseudomonas might be a 256 good selection) with a mutagen such as nitrosoguanidine, followed by selection of a clone of cells, using for example the technique of replica plating, of the desired phenotype (Bachovchin & Roberts, 1977). This method was used to produce a histidine auxotroph of Myxobacter 495 for a study of the mechanism of serine protease action (Bachovchin & Roberts, 1978). In the proposed study, the selection criterion could be for the ability to survive in . the presence of, or even utilize as a carbon source, the target compound. Naturally occurring strains of Pseudomonas which can metabolize halobenzoic acids and halophenols have been reported (Dorn et al., 1974; Knackmuss & Hellwig, 1978), as have conjugate strains which can utilize halobenzoic acids as a carbon source, converting them to metabolites, co 2 , and halide ion (Reineke & Knackmuss, 1979). Possible adaptive mutations which would result in the desired characteristics include gene duplication leading to hyperproduction of an enzyme (Rigby et al., 19 7L~) alteration of structural genes to give new enzymes with different specificities (Betz et al., 1974; Senior et al., 1976), or changes in regulatory processes (Wood, 1966; LeBlanc & Mortlock, 1971). Once a mutant strain capable of metabolizing or suitably altering the target compound is isolated, the fate of the compound may be studied using radioisotope 257 tracer methodology with either the whole organism or cell-free extracts. The results of this investigation could suggest which enzyme(s) had been altered by mutagenesis, and would also identify products derived from the labeled target compound. The "synthetic" organisms produced by these methods could be used for on-site treatment of wastes containing the target compound . Alternatively, but only after thorough characterization of the organism, areas known to be contaminated with the target compound could be "seeded" with the mutant bacterium, which would presumably compete effectively against native microorganisms only so long as its "subst1:.atc", the target compound, is present in sufficient quantities. This proposed study would likely involve a great dea l of labor, and the screening of prodigious numbers of colonies for the desired characteristics. However, the potential rewards are sufficiently great to justify the effort involved . 258 REFERENCES Bachovchin, W.W., & Roberts, J. D. (1977), unpublished & Roberts, J. D. (1978) J. Am. Chem. results. Bachovchin, W.W., Soc. 100, 8041. Betz, J. L., Brown, P. R., Smyth, M. J., & Clarke, P. H. (1974) Nature 247, 261. Dorn, E., Hellwig, M., Reineke, W., & Knackmuss, H. -J. (1974) Arch. Microbial. 99, 61. Knackmuss, H.-J., & Hellwig, M. (1978) Arch. Microbiol. 117, 1. LeBlanc, D. L., & Mortlock, R. P. (1971) J. Bacteriol. 106, 90. Mccann, J., & Ames, B. N. (1976) Proc. Natl. Acad. Sci. USA 73, 950. Mccann, J., Choi, E., Yamasaki, E., & Ames, B. N. (1975) Proc. Natl. Acad. Sci. USA 72, 5135. Reineke, W., & Knackmuss, H.-J. (1979) Nature 277, 385. Rigby, P. W. J., Burleigh, B. D., Jr., & Hartley, B. S. (1974) Nature 251, 200. Senior, E., Bull, A. J., & Slater, J. H. (1976) Nature 263, 476. Wood, W. A. (1966) Ann. Rev. Biochem. 35, 521.