67. Jg.,Hefl 9, 1970

435

Stereochemistry of Enzyme Reactions at Prochiral Centers H. G. FLoss* Department of Medicinal Chemistry and Pharmacognosy School of Pharmacy and Pharmacal Sciences, Purdue University, West Lafayette, Indiana 47907, U.S.A.

Herrn Pro/. Dr. Kurt Mothes anl~fllich seines 70. Geburtstages in Dankbarkeit und Verehrung gewidmet. Stereospecific isotopic labeling sometimes enables distinctions to be made between several possible metabolic pathways, as was done, for instance, with the ergot alkaloids. This method can provide clues to the mechanism of enzymatic reactions, as illustrated by chorismate synthetase and other enzymes. It is even possible in some cases to study a whole complex of questions relating to the evolution of enzymes, the stereostrueture of their active centers, or cooperation between two enzymes.

Zusammen[assung: Sehr viele biochemische Reaktionen spielen sich an Kohlenstoffatomen ab, die zwei gleiche (a) und zwei verschiedene ( b u n d d) Substituenten tragen, z.13. Methylen-Gruppen mit z w e i verschiedenen Liganden (R-CH2-R'). Solche Caabd-Systeme bezeichnet man Ms prochirale Zentren. Die beiden Substituenten a sind nieht identisch, sondern unterscheiden sich in ihrer r~umlichen Anordnung relativ zu den anderen drei Gruppen. Enzymreaktionen an prochiralen Zentren sind im allgemeinen stereospezifisch, d.h. ein gegebenes Enzym reagiert bei einer Substitutions- oder ]Eliminierungsreaktion nur mit einem der beiden rgumlich verschiedenen Substitueuten a. Durch stereospezifische Isotopenmarkierung eines yon beiden lassen sich die zwei Substituenten a unterscheiden, und man kann somit den stereoehemischen Verlauf enzymatischer Reaktionen an prochiralen Zentren bestimmen. In dem folgenden Artikel wird die Anweudung dieser Methodik zur Nl~rung verschiedener biochemischer Probleme diskutiert und an Beispielen illustriert.

enzymatic replacement of one H by another group could proceed with either complete loss or complete retention of the isotope. Subsequent experimentation has, of course, fully confirmed the correctness of this predicition [4] and it is today generally accepted that as a rule enzymatic reactions at a prochiral center are stereospecific, although, as to every rule there may be a number of exceptions. It must be emphasized that the ability to distinguish between heterotopic substituents is not a unique property of enzymes. Rather, the two heterotopic groups, although chemically identical, differ in their environment. If one looks from one group a at the other three substituents, the sequence

a -+ b -->d is clockwise, whereas looking from the other one

1. Introduction In 1948, Ogston introduced a new concept into biochemistry. It had been well established b y that time that enzymatic reactions of asymmetric (chiral) molecules, particularly if they involve the chiral center itself, are usually stereospecific, i.e., the enzyme usually reacts one enantiomer to the exclusion of the other. Ogston [1 ] pointed out that enzymes might be expected to exhibit stereospecificity also in reactions at a carbon atom which has two like and two different substituents. Such a carbon atom, a Caabd system, is called prochiral [2] and the two heterotopic substituents a are named the ibro-R and the pro-S group, depending on whether elevation of the group to be named over the other one gives rise to a chiral center of R or S configuration according to the Cahn-IngoldPrelog rules [3]. An enzyme, according to Ogston, could be expected to distinguish between the two heterotopic substitnents at a prochiral center, if the substrate binds to the enzyme in a three-point attachment. Thus, if at a carbon atom C H I R R ' only one of the hydrogens (e.g., H~) is labeled with tritium, * This article is based on lectures given at the USDA Northern Regional Laboratory, Peoria, Illinois (May 1969) and at the 2rid Natural Products Symposium, University of Connecticut, StolTs (August 1969). - - Recipient of a Research Career Development Award from the National Institutes of Health.

a-+b-+d appear in counterclockwise order (Fig. 1). Therefore, in any interaction with a chiral center, the two heterotopic groups can potentially behave differently. For example, the reaction of an asymmetric reagent with the heterotopic groups will proceed through two diastereomeric transition states, which m a y differ in their relative stability or in the probability of their formation, and m a y therefore lead to formation of products in an unequal ratio. A more thorough consideration of these relationships can be found in an b

Fig. 1. Caabd system

436

H.G. Floss: Stereochemistry of Enzyme Reactions at Prochiral Centers

article by Hirschmann [5]. Whereas in the majority of chemical reactions of this type the degree of stereoselectivity is modest, enzymatic reactions are usually completely stereospecific, because enzymes m a y be regarded as" near perfect asymmetric agents". It will be clear from tile above that a "three-point a t t a c h m e n t " of the substrate to the enzyme (in the sense of three binding sites) is not a necessary requirement for stereospecfficity towards a prochiral center. The enzyme may, for example, merely b y its three-dimensional structure allow binding of tile substrate to the active site in only one particular orientation. In a very general way, Popjak and Cornforth [6] state that "...dissymmetric treatment of a substrate b y an enzyme can occur whenever tile enzyme imposes, whether actively b y binding or passively b y obstruction, a particular orientation on the substrate at tile site of reaction" and they predict that "according to this general hypothesis tile attachment of the substrate to the enzyme b y one group alone would suffice for a completely stereospecific reaction". Stereospecificity of a reaction can only be detected if the substrate or the product or both are chiral. However, the chirality may only be due to tile presence of an isotopic label. In a number of cases, as in the phosphorylation of glycerol with glycerol kinase [7], where one of the heterotopic groups is replaced by another substituent not yet present at this center, a stereospecific reaction can be detected without the help of an isotope by virture of the fact that the product is optically active. More frequently, however, it is necessary to make tile two heterotopic substituents distinguishible by labeling one of them with an isotope, in order to detect a stereospecific reaction. Isotopic labeling of one heterotopic group b y an isotope, of course, transforms tile prochirai into a chiral center, because by labeling the other group one can obtain an enantiomeric compound, even though the chirality m a y not be detectable b y most of tile classical methods. It can then be determined whether in a reaction the labeled or the nonlabeled group has been replaced or eliminated or whether after replacement of another group at this carbon atom the configuration of the product is the same as that of the substrate or opposite. It is inferred then that tile unlabeled substrate reacts in the same way as the labeled one, i.e., that if (R)-ethanol-l-T gives rise to non-labeled acetaldehyde and tritiated NADH in the alcohol dehydrogenase reaction, it is tile pro-R hydrogen from C-t of unlabeled ethanol, which is transferred to the pyridine nucleotide. B y the use of substrates labeled stereospecifically with an isotope and/or b y configurational analysis of stereospecifically labeled products one can determine the steric course of stereospecific reactions even if the non~lal substrates and products of the reactions are devoid of chirality. The isotopes used most frequently in such studies are deuterium and tritium, tile isotopes of hydrogen, mainly because the prochiral centers most frequently encountered in biological molecules are methylene groups. As indicated above, investigations of the stereochemistry of enzyme reactions at prochiral centers frequently require the preparation of compounds labeled stereospecifically with an isotope and/or the

Naturwissenscha/ten

determination of tile configuration of molecules which owe their chirality to the presence of an isotope. It is beyond the scope of this article to discuss the methods used in such studies. Likewise, it is not the purpose of this paper to review all the enzyme reactions which have been investigated so far. Both these topics are covered in the reviews quoted earlier [4]. Instead, it will be attempted to outline in which ways stereochemical information of this kind can be used to answer biochemical questions. It is obvious that, at least at this time, merely showing that an enzymatic reaction is stereospecific or even determining its steric course per se is not of much merit unless it can tell us something about the system, e.g., the mechanism of the reaction, tile nature of tile enzyme, etc.

2. Distinction between Di[[erent Metabolic Pathways In a number of instances stereospecific labehng of a substrate or stereospecific degradation of a reaction product have been used to distinguish between alternate possible metabolic pathways. This approach is illustrated b y some work in which we made use of mevalonic acids stereospecifically tritiated at carbon atom 4 to clarify a question in the biosynthesis of ergot alkaloids. These alkaloids (e.g., agroclavine, Fig. 2) characteristically contain the tetracyclic ergoline ring system and are derived in nature from the amino acid tryptophan and an isoprene unit provided by mevalonic acid (Fig. 2) [8]. Work from Arigoni's laboratory and by our group has shown that one of tile last reactions leading to the ergoline ring system, tile ring closure of the tricyclic precursor alkaloid chanoclavine-I to give agroclavine and then elymoclavine,, involves a cis-trans isomerization at the allylic double bond [ 9 - - t i ] . Tile labeled carbon atom arising from C-2 of mevalonate was found to be C-7 in chanoclavine-I, which occupies the cis position relative to tile benzyhc carbon atom at the z] 8, 9 double bond, but upon cyclization this ctlanoclavine-I gave rise to agroclavine and elymoclavine labeled in C-17, i.e., in the trans position. This raised the new question as to whether in the original dimethylallyl substituent the cis or tile trans methyl group was derived from C-2 of mevaionate. The consequences of these two possibilities are outlined as pathways A and B in Fig. 2. Obviously, if pathway A is correct tile biosynthetic sequence involves not only one but two cis-trans isomerizations. To resolve this question, we made use of information on the stereochemistry of isoprenoid biosynthesis available mainly from the work of Cornforth and Popjak [6]. These investigators have shown that in the isopentenyl pyrophosphate isomerase reaction, which produces dimethylallyl pyrophosphate carrying tile label from C-2 of mevalonate in the trans methyl group [t2, t3], it is the pro-R hydrogen at C-2 of isopentenyl pyrophosphate (corresponding to the pro-4 S hydrogen of mevalonate) which is eliminated, whereas the pro-2S hydrogen is retained [14]. Likewise, tile condensation of dimethylallyl pyrophosphate and isopelltenyl pyrophosphate units to give geranyl and farnesyl pyrophosphate, i.e., compounds with a tmns geometry of the newly formed double bond, proceeds with loss of tile 1~ro-4S and retention of the pro-4R hydrogen of mevalonate [141. On the other hand, the pro-4R hydrogen is eli-

H.G. Floss: Stereochemistry of Enzyme Reactions at Prochiral Centers

57. Jg., Heft 9, 1970

437

OH Q

~ / ,

A.

NHk

NH 2

|

f

isopentenyl

dimethylallyl

pyrophosphate

pyrophosphate

H dimethylallyltryptophan

H

/ / OH

I pathway A 1COOH

R

NH2

a,~NCH3

H,,, 4 5

HO

"~/

"N H tryptophan

(3R, 4R)-mevalonic acid-4-T

v

"N

H

chanoclavine-I OH

i pathway B 4 H

|

H~NtnI2

9 9

[

>

|

9isopentenyl

dimethylallyl

pyrophosphate

pyrophosphate

H dimethylallyltryptophan

\

NH,

H

--

H agroclavine (R'= It) elymoclavine(R'=OH)

~'x

9

Z g

R = H or COOH

"N

H

Fig. 2. Biosynthesis of ergot alkaloids

minated and the pro-4S hydrogen retained in the conversion of mevalollate into rubber, which involves the generation of cis double bonds [t5]. In order to distinguish between pathways A and B, we therefore fed (4R)- and (4S)-mevalonic acids-4-T to alkaloid-producing cultures of the ergot fungus. These commercially available preparations [t4, 161 were actually racemic mixtures with respect to the configuration at both C-3 and C-4, i.e., (4R)-mevalonate-4-T was an equimolar mixture of (3 R, 4R)- and (3S, 4S)-mevalonate-4-T. However, in a separate experiment it was confirmed that as in mammalian liver [t7] only the 3 R enantiomer is metabolized by the ergot fungus [1t]. The experimental finding [t8] that (4 S)-mevalonate-4-T lost all its tritium (relative to an internal standard of mevalonate-2-14C) and (4R)-mevalonate-4-T retained the tritium during the conversion into elymoclavine, therefore, showed that the original geometry of the allylic double bond must be trans, i.e., the methyl group originating from C-2 of mevalonate occupies the trans position relative to the benzylic carbon atom. This result obviously favors the pathway A shown in Fig. 2.

8. Study o/Enzyme Reaction Mechanisms B y far the most frequent use of stereochemical probes of this kind is in the study of enzyme reaction mechanisms. Since its stereochemistry is one integral part of an enzyme reaction, stereochemical informa-

tion can often aid in the elucidation of the reaction mechanisnl. It is, however, important to be aware of the limitations of this approach. First of all, because an enzymatic conversion is brought about by interaction between the substrate(s) and certain groups of the enzyme, the mechanism of a given enzyme reaction is determined by the particular arrangement of these groups relative to each other and the substrate(s), i.e., by the three-dimensional structure of the enzyme. The latter is the result of the evolution of this particular enzyme. It follows that ultimately the mechanism of an enzyme reaction is determined b y the evolution of the enzyme catalyzing the reaction. If in the analogous nonenzymatic system a reaction Call proceed by two mechanisms, one of which is energetically favored (e.g., elimination of H + and X from adjacent carbon atoms by a concerted or by a non-concerted mechanism), the corresponding enzyme reaction will proceed by only one discrete mechanism. Experience shows that in most cases this is the one which is also energetically favored in the chemical system (e.g., concerted (tram)elimination of HX). However, this is by no means a strict rule. All it says is that most enzymes have evolved in such a way as to increase the rate of reaction via the mechanism which is the chemically most reasonable. But just as there are enzymes which can selectively catalyze one of a number of possible reactions of a substrate, even if that reaction should be only a minor one in the chemical system, an enzyme can in principle also catalyze a given reaction via a mechanism which in

438

H.G. Floss: Stereochemistry of Enzyme Reactions at Prochiral Centers COOH 0

COOH

|176

HO/~ ,

H T (Z)-phosphoenolpyruvate-3-T

T (2R, 3S)-phospho,glycerate-3-T

glucose COOtt T~H H OH 9

COOH (2R, S, 3R)-malate-3-T 1

~

--y

COOH " ~ O ~' I~ ,,,::H

HO"

H

H

COOH

more thorough discussion of this subject, the reader is referred to an article by Rose [t9]. Based on these considerations one can use guidelines to evaluate the mechanistic significance of stereochemical information on enzyme reactions, which are illustrated in the following b y some examples. a) A mechanism is rendered unlikely if the observed steric course of the reaction disagrees with that required b y the mechanism. As an example let us consider some recent work on the chorismate synthetase reaction. Chorismate synthetase, an enzyme of the shildmic acid pathway of aromatic amino acid biosynthesis [20], catalyzes a 1,4-conjugate elimination of phosphoric acid from enolpyruvyl-shikimate 3-phosphate to give chorismate (Fig. 3). It has been suggested that this elimination proceeds by a concerted (E2') mechanism [2t], which would require [22, 23~ that the phosphate group and the proton are eliminated on the same side of the ring or cis (loss of H s from C-6) (cf. Fig. 4). In order to examine this question experimentally, we prepared shikimic acids

""OH

H

erythrose-4phosphate

COOo~,',H i <

(6R)-shikimate-6-T 1

COOH L C H

.CH20(~

COOH ~,,,T H CH2 H'--[ [~H [[ .,....'x.. . c . (~O'" H..~,..,OHO . COOH

" (~OH2C H HO""

H

,,'::H H

(3S)-3-deoxy-arabinoheptulosonate-7-phosphate (DAHP)-3-T COO~ H > ~

I,,"H

; H2 "coo.

+ HTO fumarate

+ TO(~ (6R)-enolpyruvylshikimatechorismate 3-phosphate-6-T Fig. 3. Preparation of shikimic acid tritiated stereospecifically at C-6, its degradation and conversion into chorismate the non-enzymatic system is the less favored one (e.g., cis elimination of H X even if the chemical reaction proceeds predominantly with a trans stereochemistry). Strictly speaking, the analogy with the mechanism of the corresponding nonenzymatic reaction can therefore not be used to predict the mechanism and steric course of an enzyme reaction, although experience shows that in a good number of cases the prediction turns out to be correct. Another difference between chemical and enzymatic reactions determines largely to what extent conclusions on the mechanism can be drawn from knowledge of the steric course of an enzyme reaction. Some reaction mechanisms imply a certain steric course. For example, a bimolecnlar nucleophilic displacement (Sx2) involves inversion of the configuration, a concerted bimolecular elimination (E2) will result in removal of trans-diaxial substituents. In other words, the various events of the overall reaction are stereochemically interdependent. In other mechanisms, however, the individual steps are stereochemically independent of each other. For example, in reactions proceeding through free carbonium ions (e.g., S~I processes) the stereochemical result in a purely chemical system is more or less complete racemization, because the fate of the carbonium ion is, apart from possible solvent or neighbouring group effects, independent of its previous history. The same reaction catalyzed b y an enzyme will nevertheless be stereospecific, but the overall steric course is not dictated b y the reaction mechanism. Thus, if an enzyme would catalyze an S~t process, the overall result could be either inversion or retention of configuration. For a

Naturwissenschaflen

H I ,~OR

, -~)

/X-. ,.

@Or

f

\ HO-- i H "\

HR E 2' mechanism

Enz'Xe H

HO---'|

L

H

CH2 II R: -C I

COOH Enz.X~ H

i/oR

I)/~ 6

@

/

--s

@/

cooH'-"

HO --~'~ H

X-groupmechanism Fig. 4, Two possible mechanisms of the chorismate synthetase reaction and their stereochemical consequences

57. Jg., Heft 9, 1970

H . G . Floss: Stereochemistry of Enzyme Reactions at Prochiral Centers

439

tritiated stereospecifica/ly at C-6 by the route shown considerable numer of other enzymes, e.g., maleate in Fig. 3. Ghicose-l-T was phosphorylated with hexo- hydratase, aconitase, citraconase and mesaconase also kinase, followed by isomerization to fructose-6-phos- catalyze reversible trans additions of water to carbonphate using phosphoglucose isomerase. In the latter carbon double bonds and that in the analogous elireaction a second hydrogen is introduced stereo- minations of R - N H 2 catalyzed by aspartase, fl-methylspecifically at C-I to give the (1 S)-t-T specimen. The aspartase, adenylosuccinase and argininosuccinase diastereomeric sample (1 R)-fructose-6-phosphate-t-T the steric course is also trans [291. These findings was obtained in the same way from mannose-t-T strongly suggest that the trans stereochemistry of using phosphomannose isomerase. The two samples these reactions is not accidental but rather inherent of fructose-6-phosphate were converted into phos- in their mechanism and thus support a concerted (E 2) phoglycerates-3-T by the glycolytic enzymes in the mechanism for these reactions [321. This argument presence of arsenate. The phosphoglycerates together has been used to support a certain common mechanism with nonlabeled erythrose-4-phosphate on incubation for a series of 7 a/dose-ketose isomerase reactions inwith phosphoglyceromutase, enolase and a cell-free vestigated for their stereochemistry. All these enzymes extract of E. coli mutant 83-24, which is blocked in catalyze the reversible tautomerization of an a/dose the further conversions of shikimate, gave two samples or aldose phosphate into the corresponding 2-ketose of shikimate tritiated at C-6. Aliquots of these shiki- or its phosphate. Five of these enzymes, triosephosmate samples were chemically degraded to malates, phate isomerase [331, phosphoglucose isomerase [341, representing carbon atoms 7, t, 6 and 5, which were phosptloribose isomerase [341, D-xylose isomerase [351, analyzed for the position of tritium at C-3 (represent- and L-arabinose isomerase [351, utilize an a/dose ing C-6 of shikimate) by the fumarase reaction. From which at C-2 has R configuration and produce a the data and the known steric course of the fumarase ketose in which the newly introduced hydrogen at C-1 reaction E24, 251 it followed that the material ob- occupies the pro-R position (Fig. 5 a). Tile other two tained from glucose-l-T was predominantly (6R)-shikimate-6-T and H"-c/O'... H'-C--'OH that from mannose-l-T was preH\C~~ H% - O H X~ B:H* "'X---+ B:H* "][ X B: X dominantly (6 S)-shikimate-6-T. Enl: H*~ zymatic conversion of these shikiR/c--oH mate samples into chorismates showed that (6R)-shikimate-6-T lost a) ~ Enzyme ~ Enzyme ~--Enzyme--~ most of its tritium in this transEnzyme formation, whereas (6S)-shikimateH~ H~c/O'. H--.c/OH 6-T gave chorismate with retention H\c~O H* ,~,.C--OH of most of the tritium. In most of B: H. . . . . . X~ B:H* ""X ~ B:H* { .X B: " X these reactions the change in tritium RfiC_OH content was measured against an internal standard of 14C-labeled b) L Enzyrne ~ ~ - Enzyme ~ L---Enzyme----~ material. This study proved that Enzyme the elimination of phosphate and Fig. 5- Stereochemical course and proposed mechanism of aldose-ketose isohydrogen in the chorismate syn- merase reactions thetase reaction occurs on opposite sides of the ring or trans [261. The same conclusion was drawn independently by Hill and enzymes, phosphomannose isomerase [36] and I~Newkome [271, who chemically synthesized (6R)- and arabinose isomerase (assayed with L-fucose) [35] pro(6 S)-shikimate-6-D [281 and determined the deuterium duce a ketose in which the newly introduced hydrogen incorporation from these samples into phenyla/anine occupies tile pro-S position at C-1 from an a/dose and tyrosine by E. coll. These results do not support with 2S configuration (Fig. 5 b). Thus, the configurathe suggested E2' mechanism. We have proposed a tions at the two centers involved are always the 2-step X-group mechanism for the reaction (Fig. 4), same relative to each other in these reactions. It has which fits the observed steric course [261. It must been proposed [37] that all these reactions proceed be emphasized, however, that the stereochemica/results, through an endiol intermediate. Since for at least although they are compatible with it, do not prove some of them tile proton transfer has been shown such a mechanism. [35, 38, 391 to be partly or completely intramolecular, b) Stereochemical constancy of a series of analogous it should be mediated by only one group of the reactions m a y support a certain mechanism, usually enzyme. Abstraction and addition of the proton are a concerted one. In the light of the previous discussion, therefore expected to occur at the same side of the the fact that the stereochemistry of a reaction agrees plane of the intermediate [371. As a consequence of with that required by a certain mechanism cannot this and the observed stereochemistry, the endiol be taken as evidence that the reaction proceeds by intermediate must have cis geometry (i.e., the two this mechanism. For example, the finding that oxygens are cis). In view of the stereochemica/confumarase catalyzes a trans elimination of water from stancy of tile series, tile cis-endiol intermediate is (S)-malate [24, 25] does not prove a concerted (E2) likely to be an essential requirement of the reaction mechanism for this reaction, because if the reaction mechanism [351. As shown in Fig. 5 the arrangement were non-concerted there is a 50% probability that of the groups at the active sites of the two types the overall stereochemical result would also be trans of isomerases may be quite similar, the only signifelimination. However, it is known (cf. [4cJ) that a icant difference being the position of the basic

/

f

31

Naturvfissenschaftent970

l

I

440

H . G . Floss: Stereochemistry of Enzyme Reactions at ProchiralCenters

Naturwissenschaflen

group of the enzyme relative to 6H 6H the plane of the endiol intermediate. c) Stereochemical inconstancy of a a) series of analogous reactions may / .,~ I H R n I ~H s H* support a non-concerted mechanism. B12 ~ Bl2 If related reactions proceed with different stereochemistry, this may * binding sites be taken as evidence that the overall steric course of these reactions is not HO~"-'-~ /OH ~ O H CH3 OH dictated by their mechanism, i.e., that the individual steps involved are stereochemically independent of each other. Lienhard and Rose have applied this argument to two oxidative fl-decarboxylation reactions. It b) Coenzyme H-Coenzyme was found that in the conversion Fig. 7. Steric course of the propanediol dehydrase reaction (from [45] and [46]) of isocitrate to ~-ketoglutarate by isocitrate dehydrogenase the replacement of the carboxyl group by a proton occurs with retention of configuration [40, 4t] migrates from C-I to become HI~ at C-2 of propion(Fig. 6 a), whereas in the analogous decarboxylation of 6- aldehyde and in (R)-propanediol HI~ migrates from phosphogluconate to ribulose-5-phosphate by phospho- C-1 to become H s at C-2 of the product [44, 45]. The gluconate dehydrogenase it involves inversion of config- reaction thus always involves inversion of conuration [42] (Fig. 6b). Their tentative conclusion that figuration at C-2. Rdtey et al. [46] also showed that the elimination of the carboxyl group and the addition in the reaction with both substrates the hydroxyl of the proton occur in two stereochemically in- group from C-2 migrates stereospecifically to C-l, dependent steps, presumably via an enzyme-bound followed by a dehydration of the resulting propaneenol intermediate, is supported by the finding that l,l-diol to the free aldehyde, which again is stereoboth enzymes also catalyze a COs-independent stereo- specific. As a result, in the propionaldehyde obtained specific exchange of a proton in the decarboxylation from (R)-propanediol the carbonyl oxygen originates product. The hydrogen which is exchanged occupies from the secondary hydroxyl group and in the product the same steric position as that introduced in the from (S)-propanediol it originates from the primary overall reaction. hydroxyl group. The reactions of the two enantiomeric substrates except the final dehydration therefore take place in an entirely "mirror image fashion". This COOH F?OOH] COOH] has been explained by the assumption that the two [ ~--OH[, C-=O H--CI OH substrates are bound by the same three binding sites I HOOC~C--H* 9 H~C--H* > of the enzyme, resulting in an identical steric relationship between the leaving group and the migrating CH2 CH2 atoms E45]. This is illustrated in Fig. 7a. Although a) ctoo L OO j 1 COOH this has not been demonstrated, it seems likely that the migration of the hydroxyl group also proceeds COOH H with inversion of the configuration at C-1 as is shown H*--C--OH V H*--C--OH-1 H*--C]--OH in Fig. 7b. Finally, it has been found that the hydrogen I > > C=O transferred in the reaction is transiently bound to the BI~ coenzyme [47, 48]. In this process, both hydrogens at C-5' of the coenzyme equilibrate with the substrate hydrogen which undergoes transfer [49]. The studies b) on the stereochemistry of the propanediol dehydrase Fig. 6. Stereochemistry of the isocitrate dehydrogenase reaction have thus provided some detailed information reaction (a) and the 6-phosphogluconate dehydrogenase on the mechanism of this reaction and of vitamin B12reaction (b) catalyzed reactions in general which would have been almost impossible to obtain without the use of stereospecific labeling. Stereochemical studies can also often shed light on the steps involved in more complex enzyme reactions. One of the intriguing examples of this kind is the pro- 4. Enzyme Evolution panediol dehydrase reaction, a coenzyme Bib-requiring dismutation of 1,2-propandiol to propion- In certain cases stereochemical information may tell aldehyde, which has been studied mainly by the us something about the evolution of a series of engroups of Abeles and Arigoni. First of all it is sur- zymes. Rose and his collaborators [50] have recently prising that both(R)- and (S)-propanediol are utilized by determined the stereochemistry of the carboxylation the enzyme [43]. The reaction involves the transfer of phosphoenolpyruvate (PEP) by three enzymes, of a hydrogen from C-t to C-2, which with both sub- PEP-carboxylase (EC4.tA.31), PEP-carboxykinase strates is stereospecific [44]. In (S)-propanediol H s (EC 4.1.t.32) and PEP-carboxytransphosphorylase

o.,,

.o_,.

b4! /


L

I

/

57. Jg., Heir 9, 1970

H . G . Floss: Stereochemistry of Enzyme Reactions at Prochiral Centers

/

441

/

/

I

/~J~

v

famesyl pyrophosphate

~

v

~

xH

squalene

H

and/or

- - ~

and/or

H.. .H,H %

squalene epoxide

T

lanosterol

Fig. 8. Formation of asymmetrically labeled squalene and its conversion into lanosterol

(EC4AA.38). Starting from (3R)- and (3S)-phosphoglycerate-3-T, they prepared the two isomers of phosphoenolpyruvate-3-T, the configuration of which followed from the knowledge [5t] that enolase catalyzes a trans elimination of the elements of water. These samples of tritiated phosphoenolpyruvate were used as substrates in the carboxylation reactions, the resulting oxalacetate samples were immediately converted into malate with an excess of malic dehydrogenase and the malates were analyzed for the configuration of the labeled hydrogen atoms by the fumarase reaction. The result of this study was that in all three reactions the attack at C-3 of phosphoenolpyruvate is at the side of the double bond plane viewed at which the substituents phosphate, carboxyl, methylene appear in counterclockwise order or the si face [2]. Our study on the stereochemistry of chorismic acid formation also allowed us to deduce the steric course of the DAHP synthetase reaction (Fig. 3), the first specific reaction of the shikimic acid pathway in which C-3 of phosphoenolpyruvate condenses with C-1 of erythrose-4-phosphate. It was found [26] that this reaction also involves si attack at C-3 of phosphoenolpyruvate. Thus, all four addition reactions at C-3 of phosphoenolpyruvate so far studied proceed by attack at the same side of the double bond plane, although the stereochemistry of these reactions is not determined by their mechanism. If other addition reactions at C-3 of phosphoenolpyruvate should be found also to involve si attack without exceptions this would strongly suggest a common evolutionary origin of these enzymes. One could then assume that all these enzymes are made up of a common phosphoenolpyruvate-binding protein, which holds the substrate in such a way that only one face is exposed, and a second protein which is different 31"

in each of these enzymes. The latter portion would specify what particular reaction at C-3 of phosphoenolpyruvate the enzyme catalyzes.

Exploration o/Enzyme Active Sites Occasionally, studies on the stereochemistry of an enzyme reaction may provide information about the active site of the enzyme involved. For example, the studies on propanediol dehydrase mentioned above suggest a certain arrangement of the substrate bound to the enzyme (and therefore also the binding sites) relative to the catalytically active portion of the Bl~ coenzyme (Fig. 7a).

Cooperation between Enzymes An interesting application of stereospecific labeling was recently explored by Etemadi et al. [52]. These workers asked the question whether particle-bound enzymes of a biosynthetic chain may be organized in such a way that the substrate molecules pass from one to the next in an ordered or oriented fashion without dissociating from the structure. To investigate this question they made use of their previous finding that squalene, a perfectly symmetrical molecule, is biosynthesized from two molecules of farnesyl pyrophosphate in an unsymmetrical fashion [6]. Thus, in the squalene synthetase reaction,the pro-S hydrogen from C-i of one of the two farnesyl pyrophosphate moieties is replaced by a hydrogen from C-4 of NADPH, giving rise to squalene which is labeled in only one half of the molecule if tritiated NADPH is used. The next enzyme in the pathway, squalene oxidase, which like the synthetase is particulate,

442

H.G. Floss: Stereochemistry of Enzyme Reactions at Procbiral Centers

would n o t be able to distinguish b e t w e e n the two ends of squalene if the l a t t e r h a d completely dissociated from the s y n t h e t a s e a n d m o v e d a r o u n d i n the s y s t e m a t r a n d o m before b e i n g a t t a c h e d to the oxidase. I n this case, 50% each of the squalene molecules w o u l d be epoxidized i n the labeled a n d i n t h e u n l a b e l e d p o r t i o n a n d lanosterol, resulting from the cyclization of t h e squalene expoxide, w o u l d carry 50% each of the t r i t i u m at C - t t a n d C-t2, the two c a r b o n a t o m s o r i g i n a t i n g from the center carbons of squalene. O n the other h a n d , if the squalene w o u l d be " h a n d e d o v e r " b y the s y n t h e t a s e to the oxidase i n a n ordered fashion, the e p o x i d a t i o n would occur o n l y a t the labeled or o n l y a t the u n l a b e l e d half of the molecule. As a result, the t r i t i u m w o u l d be p r e s e n t o n l y at C - I I or o n l y at C-12 of t h e derived lanosterol (Fig. 8). B y this reasoning, a n y d e v i a t i o n from a t : t ratio of the t r i t i u m at C-II a n d C-12 of lanosterol w o u l d i n d i c a t e t h a t t h e process of the transfer of squalene from t h e s y n t h e t a s e to t h e oxidase is n o t completely r a n d o m . The e x p e r i m e n t was carried out with a r a t liver h o m o g e n a t e a n d d e g r a d a t i o n of the labeled lanosterol revealed t h a t the ratio of ~Hn/SHI~ was 1.28. However, there was some residual t r i t i u m i n the molecule u n a c c o u n t e d for a n d the a u t h o r s therefore felt t h a t the evidence was n o t sufficient yet to allow the conclusion t h a t squalene h a d been oxidized preferentially from one end. This s t u d y is b e i n g c o n t i n u e d u s i n g more suitable conditions [531.

Conclusion

Since t h e original p u b l i c a t i o n b y Ogston o n l y two decades ago, a great n u m b e r of e n z y m e reactions at prochiral centers h a v e b e e n investigated. I n this article, it has been a t t e m p t e d to i n d i c a t e a n d illust r a t e some of the ways i n which such studies on the s t e r e o c h e m i s t r y of e n z y m e reactions can help to answer biochemical questions. A t the same time, these examples also serve to illustrate l i m i t a t i o n s to this approach. Work from the author's laboratory mentioned in this paper was supported by U.S. Public I-Iealth Service Grants No. AM 11662 and FR 05586 and by Eli Lilly and Company. The author wishes to thank Drs. J. Retey, Ziirich, J. E. Robbers, Lafayette, and I. A. Rose, Philadelphia, for reading the manuscript and for their helpful comments. Eli Ogston, A. G.: Nature 162, 963 (1948). - - [2] Hanson, K . R . : J. Amer. Chem. Soe. 88, 2731 (t966). - - [3] Cahn, R. S., Ingold, C., Prelog, V.: Angew. Chem. 78, 413 (t966); Angew. Chem., Intern. Ed. 5, 385 (t966). - - [4] For some reviews see: a) Levy, H. R., Talalay, P., Vennesland, t3., in: Progress in Stereochemistry. 13y P. t3. D. de laMare and W. Klyne. London: Butterworths 1962. Vol. 3, p. 299. b) Cornforth, J.W., Ryback, G.: Ann. Rep. Progr. Chem.

Naturwissemchaflen

42, 428 (1965). c) Arigoni, D., Eliel, E.L., in: Topics in Stereochemistry. By E. L. Eliel and N. L. Allinger. New York: John Wiley and Sons. VOI. 4, in print. - - [5] Hirschmann, H., in: Essays in 13iochemistry. 13y S. Graft. New York: John Wiley and Sons t965, p. 156. - - [6] Popjak, G., Cornforth, J . W . : 13iochem. J. 101, 553 ( t 9 6 6 ) . - [7] 13ublitz, C., I~ennedy, E. P.: J. Biol. Chem. 211, 951 (1954). - - [8] cf.: Weygand, F., Floss, H . G . : Angew. Chem. 75, 783 (1962); Angew. Chem., Intern. Ed. 2, 243 ( t 9 6 3 ) . - [9] Fehr, T., Acldin, W., Arigoni, D.: Chem. Commun. 1966, 80t. - [t0] Floss, H. G., et al.: ibid. 1967, t05. - - [11] Floss, H. G., et al.: J. Amer. Chem. Soq. 90, 6500 (1968). - - [12] A15goni, D.: Experientia 14, 153 (1958). - - [13] Birch, A. J., et al.: J. Chem. Soc. 1962, t 5 0 2 . - [14] Cornforth, J.W., et al.: Proc. Roy. Soc. Ser. 13 163, 492 (1966). - - [t5] Archer, 33. L., et al.: ibid. 163, 5t9 ( 1 9 6 6 ) . - [161 Radiochemical Center, Amersham. - - [17] Cornforth, R. H., Cornforth, J.W., Popjak, G.: Tetrahedron 18, 1351 (1962). - - [18] Floss, H. G.: Chem. Commun. 1967, 8 0 4 . - [19] Rose, I.A.: Ann. Rev. 13iochem. 35, 23 (1966). - - [20] cf.: Lingens, F.: Angew. Chem. 80, 384 (t968); Angew. Chem., Intern. Ed. 7, 350 (1968). - - [2t] Morell, H., etal.: J. Biol. Chem. 242, 82 ( 1 9 6 7 ) . - [22] Fukui, If. : Tetrahedron Letters 1965, 2 4 2 7 . [23] Anh, N. T.: Chem. Commun. 1968, 1089. - - [24] Gawron, O., Fondy, T. P.: J. Amer. Chem. Soc. 81, 6333 ( 1 9 5 9 ) . [25] Anet, F. A.L.: ibid. 82, 994 (1960). - - [26] Onderka, D. K., Floss, H. G.: ibid. 91, 5894 (1969). - - [27] Hill, R. K., Newkome, G. R.: ibid. 91, 5893 (1969). - - [28] Hill, R. IZ., Newkome, G. R. : Tetrahedron Letters 1968, 1 8 5 1 . - [29] The author is aware of only one exception: the hydration of dehydroshikimate to dehydroquinate involves a ais addition of water [30]. It has been pointed out that this reaction might be initiated by OH- rather than by a proton. Interestingly, the internal cyclization of cis, cis-muconic acid to muconolactone, which presumably is initiated by the carboxylate anion, also proceeds by cis addition [31]. - - [30] Hanson, K.R., Rose, I.A.: Proc. Nat. Acad. Set. (U. S.) 50, 98t ( 1 9 6 3 ) . - [3t] Avigad, G., Englard, S.: Fed. Proc. 28, 345 (1969). - - [32] However, as pointed out beIow, stereochemical constancy of a series of related enzyme reactions could conceivably also be the result of a common evolutionary origin of these enzymes (cf. Section 4). - - [33] Rose, I. A. : J. Amer. Chem. Soc. 80, 5835 ( 1 9 5 8 ) . - [34] Rose, I.A., O'ConneU, E. L,: 13iochem. 13iophys. Acta 42, 159 (1960). - - [35] Rose, I. A., O'Connell, E. L., MortIock, R. P. : ibid. 178, 376 (t969).-[36] Topper, Y. J.: J. 13iol. Chem. 225, 419 (1957). - - [37] Rose, I.A.: 13rookhaven Symp. Biol. 15, 293 ( t 9 6 2 ) . [38] Rose, I. A., O'Connell, E.L.: J. 13iol. Chem. 236, 3086 (1961). - - [39] Simon, H., Medina, R.: Z. Naturforschg. 21b, 496 ( 1 9 6 6 ) . - [40] Englard, S., Listowsky, I.: 13iochem. Biophys. Res. Commun. 12, 356 (1963). - - [4t] Lienhard, G. E., Rose, I. A.: 13iochemistry 3, 185 (1964). - - [42] Lienhard, G.E., Rose, I.A.: ibid. 3, 190 ( 1 9 6 4 ) . - [43] 13rownstein, A . M . , Abeles, R. H.: J. 13ioL Chem. 236, t199 (t96t). - [44] Zagalak, 13., et al. : ibid. 241, 3028 (t966). - - [45] R~tey, J., Umani-Roncbi, A., Arigoni, D. : Experientia 22, 72 (t966). - - [46] R~tey, J., et al.: ibid. 22, 502 (1966). - [47] R6tey, J., Arigoni, D.: ibid. 22, 783 (t966). - - [48] Frey, P.A., Kottke Essenberg, M., Abeles, R. H. : J. Biol. Chem. 242, 5369 (1967). - - [49] Frey, P. A., Kerwar, S. S., Abeles, R. H.: 13iochem. 13iophys. Res. Commun. 29, 873 ( t 9 6 7 ) . [501 Rose, I.A., et al.: J. Biol. Chem. 244, 6t30 ( 1 9 6 9 ) . [5t] Cohn, M., et al.: J. Amer. Chem. Soc. 92 (in press). - [52] Etemadi, A. H., Popjak, G., Cornforth, J . w . : 13iochem. J. 111, 445 (1969). - - [53] Cornforth, J . W . : Quart. Rev. 23, t25 (1969). Received January 7, and February tl, t970