Proc. Nat. Acad. Sci. USA Vol. 70, No. 7, pp. 2077-2081, July 1973

Activation of a Covalent Enzyme-Substrate Bond by Noncovalent Interaction with an Effector (NAD binding/structural asymmetry/acyl glyceraldehyde 3-phosphate dehydrogenase/ negative cooperativity) 0. P. MALHOTRA* AND SIDNEY A. BERNHARDt Institute of Molecular Biology and Department of Chemistry, University of Oregon, Eugene, Ore. 97403 Communicated by V. Boekelheide, April 9, 1973

ABSTRACT The absorption spectrum of an active- reaction pathway (4). The subsequent chemical transforma- site specific chromophoric acyl enzyme, sturgeon 3-(2- tion of such intermediates is often the rate-limiting step in furyl) - acryloyl- glyceraldehyde - 3-phosphate dehydroge- nase, is reported. This acyl enzyme undergoes all of the catalysis. An example, par excellence, of the formation of a catalyzed reactions characteristic of the intermediate of stoichiometrically significant covalent intermediate is for the physiological acyl enzyme, 3-phospho-D-glyceroyl- muscle glyceraldehyde-3-phosphate dehydrogenase (GPDH; glyceraldehyde-3-phosphate dehydrogenease. The rates of EC 1.2.1.12), a tetrameric molecule composed of identical sub- reactions of both these acyl enzymes depend strongly on the extent of interaction of the acyl enzyme with the units. In this case an acyl (3-phosphoglycerol-) enzyme can be oxidized coenzyme, NAD+, even where the "redox" isolated in stoichiometrically significant yield (5, 6). Con- properties of the coenzyme are not required. Likewise, sistent with this finding are kinetic inferences that the rate- the spectral properties of chromophoric acyl enzyme limiting step in the oxidative phosphorylation catalyzed by are affected by the extent of bound NAD. Under the pseu- muscle GPDH (Eq. 1) near neutral pH is the of dophysiological conditions reported herein, there is a phosphorolysis stoichiometric limitation of two furylacryloyl-acyl groups the acyl enzyme (7). For such oligomeric enzymes, particularly per enzyme molecule containing four covalently-equiva- when rate-control occurs subsequent to intermediate forma- lent subunits. The binding of NAD both to the apoenzyme tion, we should consider the effect of allosteric effectors not and to the diacyl enzyme is heterogeneous: at low ex- only on the unsubstituted enzyme, but on the stoichiometri- tents of NAD occupancy, NAD binding is stronger. The binding to acyl enzyme can be quantitatively described by cally significant enzyme-substrate covalent intermediates. an enzyme model involving a tetramer with 2-fold sym- 0 0 0 metry, and consequently containing equal numbers of two 1I NADI 11 HPOje V classes of sites. NAD binding to difurylacryloyl-enzyme R-C-H + EH _ R-C-E _ ' R-COP32 + EH occurs virtually discretely, first to the two unmodified n-Glyceraldehyde NADH Acyl 3-Phospho- (tight-binding) sites, followed by looser binding to the two enzyme D-glyceroyl acyl-sites. NAD occupancy at these latter sites transforms 3-phosphate HAsO}@ phosphate the chromophoric acyl spectrum from that characteristic H20 of a model furylacryloyl-thiol ester in H20 to a highly per- RCOO0 + EH turbed furylacryloyl spectrum characteristic ofmonomeric [1] native "active-thiol" furylacryloyl-enzymes. Likewise the acyl reactivity towards arsenolysis depends on the extent We have reported the isolation and spectrophotometric of NAD bound to the loose sites. Elimination of the tight identification of a chromophoric acyl enzyme with properties binding of NAD to the difurylacryloyl enzyme tetramer by analogous to that of the physiologically isolable 3-phospho- alkylation of the remaining two free SH groups with iodo- acetate has no apparent influence on the NAD-dependent glyceroyl GPDH (8). This j#(2-furyl)-acryloyl enzyme (FA- furylacryloyl-spectral perturbation at the "two equivalent enzyme), like the physiologically important acyl enzyme, acyl sites," even though it eliminates the apparent "nega- undergoes phosphorolysis, arsenolysis, and reduction by tive cooperativity" in NAD binding. NADH. Catalytic reactions involving acyl enzyme exhibit an absolute requirement for bound NAD even where the oxidative Models for allosteric activation and inhibition have centered Under primarily on the premise that this type of regulation involves potential of the coenzyme is not required (8-11). variation in the effectiveness of ligand binding (1-3). Accord- appropriate conditions, acyl enzymes can be obtained with ing to these models, allosteric effectors function by regulating stoichiometric Emits of acylation of 2 or of 4 acyl groups per tetramer (8, 12, 13). Due to concentration limitations in acyl the distribution among quaternary conformational states limitation (and consequently the fraction of states with high affinity for phosphate and to other factors, the stoichiometric substrates). In various enzyme-catalyzed reactions, formation of two acyls per tetramer is more stringent with FA-enzyme of covalent enzyme-substrate intermediates is essential to the than with the 3-phosphoglyceroyl-enzyme. Isolation of FA-enzyme from a large excess of acyl phosphate Abbreviations: GPDH, glyceraldehyde 3-phosphate dehydro- results in a diacyl tetramer containing two equivalents of genase; FA, j0(2-furyl)-acryloyl; FAP, jl(2-furyl)-acryloyl phos- bound NAD (8). The UV-visible absorption spectrum of this phate. acyl enzyme species is consistent with that of a model com- * On leave from Chemistry Department, Banaras Hindu Uni- pound FA-thiol ester in water, and exhibits only a small versity, Varanasi, India. change in spectrum when the enzyme protein is denatured t Address reprint requests to this author. (8, 14). In contrast, a monomeric enzyme that, like GPDH, 2077 Downloaded by guest on September 27, 2021 2078 Biochemistry: Malhotra and Bernhard Proc. Nat. Acad. Sci. USA 70 (1973) Since the native FA-GPDH is catalytically convertible to all of the products characteristic of the catalytic true substrate reactions (Eq. 1), it was surprising to us, on the basis of our past experience with chromophoric acyl enzymes, that FA- GPDH was spectrally unperturbed. A potential explanation lies in the known effect of NAD on the thermodynamic sta- bility of this acyl enzyme (8). As seen by chemical equilibrium measurements, the free energy of formation of the acyl enzyme bond in the FA-enzyme is strongly affected by the presence of bound NAD. The unperturbed acyl enzyme spectrum re- ported previously was for a species containing two or less bound NAD per tetramer. In the experiments that follow, we report on the effect of higher concentrations of NAD (concen- trations in the physiological range) on the spectral and chem- ical properties of the FA-enzyme. There is a profound effect of noncovalently bound NAD on the bonding properties of the acyl enzyme thiol ester, as indicated by changes in the FA absorption spectrum and the reactivity of the acyl-enzyme bond. MATERIALS AND METHODS The procedure of Allison and Kaplan (19) for purification of sturgeon muscle GPDH was modified, yielding a higher .02 specific activity. The DEAE-cellulose chromatography step was replaced by a DEAE-Sephadex chromatography step and was followed by an additional step involving carboxy methyl-

.0 cellulose chromatography. The activity was estimated accord- (D>a0 ing to Ferdinand (20) and protein concentration was obtained .02 from absorbance at 280 and 260 nm. The specific activity of our enzyme preparation was 310-320 Ferdinand units/mg protein, the highest muscle GPDH activity reported. The .041 reaction of FA-phosphate (FAP) with enzyme was done as described (8). The molecular weight of the enzyme was as- sumed to be 1.4 X 105. FA-apoenzyme was prepared by passing a solution of the FA-enzyme through a small column of 310 340 370 400 Wavelength, A(nm) charcoal [washed according to Krimsky and Racker (12) ]. The 280/260 ratio of the FA-apoenzyme was in the range 2.02-2.10 FIG. 1. (a) FA-Absorption spectrum of FA-GPDH at different and was invariant to further charcoal treatments. Carboxy- concentrations of NAD. GPDH concentration is 4.2 MM and methylated FA-enzyme was prepared by treatment of the contains 1.4 FA-groups per tetramer. The NAD content varies FA-enzyme (without charcoal treatment) with a 10- to 20-fold as follows: (D) No NAD; (i) 2.4 mol of NAD per mol of GPDH; 82 ,uM; (G) 1.09 mM. (b) Difference spectra: [(FA-enzyme + excess of iodoacetate solution (pH about 7), and isolated by NAD) - (FA-enzyme) - (NAD)]. FA-enzyme in (-(b is that passage through a Bio-gel (P-30) column. The 280/260 ratio at described above. NAD concentrations are (D 41 MuM; (G 82 MM; this stage was about 1.8 and increased to 2.06 on charcoal 1.09 mM. Curve (G) is the difference spectrum (1.09 mM NAD treatment. The isolated carboxymethylated FA-enzyme was -no NAD) for carboxymethylated FA-enzyme (4.2 MAM) contain- deacylated (1-2 mM arsenate, pH 7.0) and free reactive SH ing 1.2 FA-groups and 2.8 carboxymethyl groups per tetramer. groups were estimated with Ellman's reagent [5,5'-dithio-bis- (2-nitrobenzoic acid)]. The number of such groups was in- variably equal to the number of FA groups that were initially contains a cysteine sulfhydryl at the active site and functions present, indicating that the rest of the reactive SH groups (4 by formation of a native enzyme thiol ester intermediate, - number of FA groups per tetramer) had reacted with exhibits highly perturbed ("red-shifted") chromophoric acyl iodoacetate. For rate studies, the time-dependent disappear- enzyme thiol ester spectra (14, 15). The denatured chromo- ance of the FA-GPDH absorption band was monitored at phoric acyl enzyme spectra correspond to those for the chro- 360 nm after addition of 10 Al of 40 mM arsenate solution (pH mophoric thiol esters in water. Likewise, chromophoric acyl 7.0) to a solution of FA-enzyme (total volume 0.8 ml). enzyme intermediates formed with ser;ne oxygen at the active The buffered solvent used in all the experiments contained site are highly "red-shifted" relative to the corresponding 0.01 M ethylenediamine chloride-0.1 M KCl-1 mM EDTA model esters and denatured acyl enzymes (16, 17). For "active at pH 7.0. The temperature was 25° unless stated otherwise. serine" acyl-enzymes, where a wide variety of enzyme species The spectra were recorded with a Cary model 14 spectro- have been studied, the Xm. for any particular chromophoric photometer interfaced to a Varian 620i computer. They con- acyl group in the native acyl enzyme is enzyme-species in- sist of measurements at 2.0-nm intervals, the value at each variant, hence suggesting that the chemical-bond structure of wavelength having been obtained by averaging of multiple the oxygen ester (and presumably the thiol ester) are altered absorbance measurements. For precise low absorbance from the usual transplanar configuration (16-18). measurements, a Cary 16 spectrophotometer was used. Downloaded by guest on September 27, 2021 Proc. Nat. Acad. Sci. USA 70 (1978) Effector Activated Enzyme-Substrate Bond 2079 RESULTS The absorption spectra of a native FA-enzyme in the presence of varying amounts of NAD are shown in Fig. la. The diacyl enzyme isolated by Sephadex chromatography contains two tightly bound NAD, which can be removed by charcoal treatment. The spectra in the presence of NAD are potentially complicated to a small extent by formation of a new broad electronic absorption band of low extinction (the "Racker band") due to interaction of NAD with unmodified active-site cysteine SH. Indeed, the loss of absorbance on treatment of the diacyl-di NAD-enzyme with charcoal (Fig. la) can be quantitatively accounted for by the loss of two equivalents of "Racker band" per enzyme tetramer; hence no further Free [NADD' (uM)' 01 "Racker band" absorption is anticipated on further NAD ° .04 *08 *12 binding to the di NAD-di FA-enzyme. NAD binds to sites Toal W[] (mM) where the active-site SH groups are modified. Large spectral FIG. 3. Titration of FA-GPDH with NAD at 420 nm. GPDH changes characteristic of perturbation in the chromophoric concentration is 10 /AM and contains 1.8 FA-groups per tetramer. acyl (FA) group are observed at higher NAD concentrations Note the biphasic nature of this curve in contrast to the simple where more than two NADs per enzyme are bound (Fig. la). hyperbola in Fig. 2. (Insert) At 420 nm the change in absorbance Note the "red-shift" of Xmax. This NAD-induced perturbation due to Racker-band formation is equal to that due to the FA- of the FA-enzyme chromophore is readily distinguished from GPDH spectral perturbation, so that the ratio (observed A A): (A A at NAD-saturation) is equal to the fraction of total spectral changes due to formation of NAD-enzyme "Racker- NAD sites occupied. A total stoichiometry of four NAD sites band" by the spectral peculiarities of the latter. The spectral per tetramer has been presumed for calculation of the concentra- properties of the diffuse "Racker band" have been well tion of free NAD for the double-reciprocal plot. characterized. The T-7r* transition of the FA-chromophore is a narrow single electronic absorption band. Thus, the NAD- binding sites responsible for the "Racker band" can be dis- The Xmaxc of acyl enzyme at high concentrations of NAD is at tinguished from the NAD-binding sites that perturb the about 360 nm (Fig. la). This is also the Xma. of the FA chromo- acyl spectrum, if such discrete sites exist. From the NAD phore when it is attached to the active-site cysteine of native concentration dependence of the absorbance at 390 nm papain, a monomeric, nonallosteric enzyme (14). Spectra of (Fig. 2) compared to that at 420 nm (Fig. 3), the sites respon- FA-enzymes containing different amounts of bound NAD, sible for the "Racker band" are occupied at concentrations greater than two NAD per tetramer, are shown in Fig. la. As much lower than that required to perturb the FA-thiol ester can be seen in Fig. la, and more clearly in Fig. lb, there is no enzyme spectrum. discernible intermediate acyl spectrum differing from both the FA-apoenzyme spectrum and that obtained under conditions of nearly complete NAD saturation. Rather, the distribution 100 KI between the two spectral species is affected by the extent of NAD binding. Also, appreciable perturbation of the acyl enzyme spectrum, from that anticipated for a model thiol ester, occurs only when there are more than two NAD bound per tetrameric enzyme. Distinctions among NAD binding sites become clear when 60 the remaining two active-site SH groups of the di FA-tetramer are carboxymethylated with iodoacetate. In this way (Eq. 2) a di FA-dicarboxymethyl-enzyme can be prepared in which there is no possibility for "Racker band" formation upon NAD binding. (FA-S)2E(SH)2 + 2 I-CH2COO'9 (FA-S)2E(SCH2COOe)2 + 2H9 + 219 [21 0 0-2 0 4 0-6 08B In the presence of higher concentrations of NAD, the FA-di- [NADJ (mM) carboxymethyl-enzyme (Eq. 2) shows a perturbation in the FIG. 2. Effect of NAD on the catalytic and spectral properties FA-enzyme spectrum identical to that observed with the non- of FA-GPDH. Spectral perturbation titration of apo FA-GPDH alkylated FA-enzyme, provided that due account of "Racker (4.1 ,uM protein containing 1.9 FA-groups per tetramer) with band" formation is taken in the nonalkylated case (Fig. lb). NAD at 390 nm (0); spectral perturbation titration of carboxy- This is substantiated by an identical NAD-concentration methylated FA-GPDH (2.251sM protein containing 1.5 FA-groups and 2.5 carboxymethyl groups per tetramer) with NAD at 390 nm dependence for the spectral perturbation in the alkylated and (0); initial rates of arsenate-catalyzed deacylation of FA-GPDH nonalkylated FA-enzymes (Fig. 2). Carboxymethylation (1.65 ,uM protein containing 2.1 FA-groups per tetramer) in the merely destroys the tight binding to two sites, and conse- presence of 0.5 mM arsenate (X). The NAD-saturated concen- quently the apparent "negative cooperativity" in NAD bind- tration-independent maximal rate of deacylation is 4.8 acyl groups ing. These results illustrate the inherent intramolecular struc- per acyl site per min. tural asymmetry of the diacyl enzyme. Downloaded by guest on September 27, 2021 2080 Biochemistry: Malhotra and Bernard Proc. Nat. Acad. Sci. USA 70 (1973) It is possible to terminate the acylation reaction at stoi- solution. There are essentially only two spectra for FA- chiometries less than two acyl groups per tetramer and to enzyme; the spectrum obtainable for native FA-papain and alkylate the remaining free SH groups with iodoacetate. native FA-GPDH at high [NAD] and the spectrum obtain- Acyl-alkyl enzymes were prepared as indicated in Table 1, able for denatured FA-enzymes. This result might be con- and the extent of perturbation of the acyl-spectrum by a trasted with that obtained by measuring the spectra of model nearly saturating NAD concentration was measured. At a FA chromophores in various solvents, a process that leads to a fixed protein concentration, the NAD-induced absorbance continuum of spectra dependent on the changing solvent changes are proportional to the acyl content (Table 1), fur- polarity (17). Previously one of us has proposed that the ther confirming that the acyl spectral perturbations are in- characteristic perturbed acyl-enzyme spectrum results from a dependent of alkylation at the remaining SH sites.t higher-energy configurational ground state of the acyl ester In contrast to acylation of GPDH with FAP, which follows bond in native acyl enzymes. This proposal was based on the first-order kinetics, the removal of acyl groups with specific spectra of acyl chromophores at the active-sites of enzymes acceptors (phosphate, arsenate, or NADH) exhibits more com- from both "active-serine (oxygen)" and "active-cysteine plex kinetics (8). The initial slope of time-dependent deacyla- (thiol)" esters. For "active-serine" acyl enzymes, we have tion, however, was first order in FA-enzyme concentration suggested a discrete configuration for the perturbed acyl (unpublished data). We have now measured these initial rates enzyme bond (17) based on the known three-dimensional of deacylation with 0.5 mM arsenate (pH 7.0), at different structure of indole-3-acryloyl chymotrypsin (21). This sug- NAD concentrations. The NAD concentration dependence of gestion is consistent with the spectrum of the acyl enzyme, the acyl-enzyme bond reactivity, as measured by the rate of and, furthermore, presents a plausible route for catalysis by deacylation, is the same as that obtained in the spectral per- facilitation of nucleophilic attack at a more electrophilic car- turbation experiments (Fig. 2). bonyl-carbon. Our knowledge of the electronic properties and electronic spectra of sulfur compounds is not so extensive as to DISCUSSION make such definitive mechanistic proposals. Nevertheless, we We believe that these results are significant in the following note the similar, and even larger, perturbation to higher Xmax ways: (i) an intermediate in enzyme catalysis has been demon- of the acyl enzyme spectrum for "active-cysteine" acyl strated to be "activatable" by the noncovalent binding of a enzymes. An oligomeric enzyme-substrate intermediate re- ligand that does not participate in the subsequent catalyzed quires bound effector for activation, whereas the correspond- reaction process. (ii) The acyl enzyme so activated exists in ing monomeric intermediate does not require any bound ligand an alternate chemical configurational state, as indicated by for its presumably identical activation. the unique, activated-acyl absorption spectrum. It is important to note the reactive chemical properties of During our work, several interesting conclusions regarding FA-papain as contrasted with an ordinary thiol ester (14). acyl-enzyme stability and enzyme-NAD affinity were un- Likewise the conditions of NAD concentration required for covered: Diacyl apoenzyme contains two classes of NAD bind- optimal reactivity of FA-GPDH, according to Eq. 1, are those ing sites. The sites of high NAD affinity are those in which the that lead to the same perturbed FA-enzyme spectrum (Xmax SH groups are unmodified. Activation of the acyl bond is at 360 nm) (Fig. 2). Hence the term "activated acyl enzyme" mediated by NAD binding to the subunit containing the acyl is equally applicable to the spectral and the chemical proper- group. This is the NAD binding site of lower affinity in the ties. diacyl enzyme.§ Chemical modification of the tight binding A charge transfer complex has been proposed to give rise to NAD sites by carboxymethylation of the SH groups has little a unique NAD-enzyme electronic absorption band (the or no effect on the NAD-induced activation of the acyl-enzyme "Racker band") (22). Our results show that those NAD bind- bond. Rather, alkylation results in the loss of the "ineffective" ing sites that give rise to "Racker band" do not perturb the tight NAD binding. FA-enzyme spectrum. It is also of interest to note the lower Activation of Chemical Bonds by The affinity for NAD at the acyl-perturbing sites. Such apparently Ligand Binding. be due to an fit be- spectrum of FA-GPDH in the presence of higher concentra- weaker interactions might "imperfect" tions of NAD (Fig. la) is identical to the spectrum of the FA- enzyme formed by specific acylation of the active-site cysteine TABLE 1. Effect of NAD on the acyl spectrum of FA-GPDH at of papain. This enzyme is monomeric. Denaturation of the different extents of acylation FA-enzymes, including FA-GPDH, leads to a spectrum iden- for model FA-thiol esters in Number of acyl tical to that obtained aqueous groups per mole of Initial A A A It is possible that acylation to any mean extent between 0 and 2 enzyme acyl per tetramer actually represents a statistical mixture of 0.5 0.0121 0.016 diacyl- and nonacylated tetramers. Such a mixture would lead to 0.8 0.0185 0.026 results identical to those listed in Table 1. This alternative seems 1.1 0.0218 0.037 unlikely on the basis of reported pseudo-first-order kinetics for 1.5 0.0307 0.044 the acylation reaction (8), although such kinetic data are not unambiguous as regards site heterogeneity. § We have not eliminated the possibility that there are two NAD- The reaction of GPDH with FAP was terminated at different binding sites per subunit, one tighter than the other. Conse- submaximal extents of acylation, and the remaining free SH quently our conclusion depends on the unlikelihood of this possi- groups of the enzyme were exhaustively alkylated with iodoace- bility. On the basis of various NAD-GPDH binding isotherms, a tate (see Methods). NAD (1.1 mM) was added to a fixed con- stoichiometry of four NAD binding sites per 140,000 molecular centration of protein (5.2 ,uM) and the absorbance changes weight appears to be well substantiated (24, 28). were noted at 390 nm. Downloaded by guest on September 27, 2021 Proc. Nat. Acad. Sci. USA 70 (1978) Effector Activated Enzyme-Substrate Bond 2081

tween the coenzyme and the enzyme site. Alternatively, it of weak ligand-site interactions to a chemical-bond perturba- might reflect a coupling of a relatively strong NAD-site inter- tion in chemoreceptors. action with an energetically unfavorable quarternary struc- This work was supported by Public Health Service Grant tural change, thus using part of the ligand-binding free energy GM10451-09 and National Science Foundation Grant GB313- to stabilize that conformation that results in chemical-bond 75X1. activation. 1. Monod, J., Wyman, J. & Changeux, J. P. (1965) J. Mol. On the Existence of 2-Fold Symmetry in Tetrameric Muscle Biol. 12, 88-118. 2. Koshland, D. E., Jr., Nemethy, G. & Filmer, D. (1966) GPDH. We (8, 23, 24) and others (25) have suggested that Biochemistry 5, 365-385. the subunit structure of muscle GPDH has maximal 2-fold 3. Atkinson, D. E., Hathaway, J. A. & Smith, E. C. (1965) symmetry (C2) and hence functions as a pair of dimers (a J. Biol. Chem. 240, 2682-2690. dimer of dimers). This is apparent for the stoichiometrically- 4. See for example Jencks, W. E. (1969) in Catalysis in Chem- limited di FA-enzyme tetramer. However, we have argued istry and Enzymology (McGraw-Hill Book Co., New York), pp. 42-59. strongly in the past that unmodified apoenzyme also has a 5. Velick, S. F. & Furfine, C. (1963) in The Enzymes, eds. maximal 2-fold symmetry (23, 26). The results presented Boyer, P. D., Lardy, H. & MyrbAck, K. (Academic Press, herein add further substance to the argument. A 2-fold sym- New York), 2nd ed., Vol. VII, pp. 243-273. metry tetramer might a priori be anticipated to have two dis- 6. Bloch, W., MacQuarrie, R. A. & Bernhard, S. A. (1971) in J. Biol. Chem. 246, 780-790. tinct kinds of binding sites (27): the extreme, only two of 7. Furfine, C. S. & Velick, S. F. (1965) J. Biol. Chem. 240, the four potential enzyme sites might be catalytically reactive, 844-855. thus leading to "half-site reactivity." The fact that the two 8. Malhotra, 0. P. & Bernhard, S. A. (1968) J. Biol. Chem. 243, tight NAD-binding sites in the diacyl enzyme can be 1243-1252. blocked by alkylation without changing the effector role of 9. Racker, E. & Krimsky, I. (1952) J. Biol. Chem. 198, 731- 743. NAD at the two acyl sites suggests that the ligands bound to 10. Hilvers, A. G. & Weenen, J. H. M. (1962) Biochim. Biophys. the tight binding sites do not affect NAD binding at the Acta 58, 380-383. "weaker" sites. Previously, the existence of "negative co- 11. Trentham, D. R. (1971) Biochem. J. 122, 59-69. operativity," particularly for NAD binding to muscle GPDH 12. Krimsky, I. & Racker, E. (1963) Biochemistry 2, 512-518. (28), has been cited in refutation of the allosteric model of 13. Trentham, D. R. (1971) Biochem. J. 122, 71-77. 14. Hinkle, P. M. & Kirsch, J. F. (1970) Biochemistry 9, 4633- Monod et al. (1). This allosteric model in its simplest formula- 4643. tion did not consider the possibility of 2-fold symmetry in 15. Bender, M. L. & Brubacher, L. J. (1964) J. Amer. Chem. tetramers, or of any other subunit arrangements in oligomers Soc. 86, 5333-5334. in which the symmetry was suboptimal. Phenomenologically, 16. Bernhard, S. A., Lau, S. J. & Noller, H. (1965) Biochemistry if a quaternary 4, 1108-1118. "negative cooperativity" is possible conforma- 17. Bernhard, S. A. & Lau, S. J. (1971) Cold Spring Harbor tional state has more than one kind of binding site for ligand, Symp. Quant. Biol. 36, 75-83. as for example, a tetramer with maximal 2-fold symmetry. 18. Charney, E. & Bernhard, S. A. (1967) J. Amer. Chem. Soc. "Negative homotropic" interactions are not allowed by the 89,2726-2733. allosteric model. The present results give no evidence for 19. Allison, W. S. & Kaplan, N. 0. (1964) J. Biol. Chem. 239, 2140-2152. "negative homotropic" interactions in sturgeon muscle 20. Ferdinand, W. (1964) Biochem. J. 92, 578-585. GPDH; the "negative cooperativity" is a result of site hetero- 21. Henderson, R. (1970) J. Mol. Biol. 54, 341-354. geneity consequent to the maximal 2-fold symmetry of the 22. Kosower, E. M. (1956) J. Amer. Chem. Soc. 78, 3497-3501. tetramer. 23. Bernhard, S. A. & MacQuarrie, R. A. (1973) J. Mol. Biol., The results presented herein do not give the molecular de- 74, 73-78. 24. Matthews, B. W. & Bernhard, S. A. (1973) Annu. Rev. tails by which the noncovalent interaction of an effector with Biophys. Bioeng. 2, in press. an oligomeric enzyme results in the activation of chemical 25. Watson, H. C., Duee, E. & Mercer, W. D. (1972) Nature bonds. Nevertheless, it is abundantly clear from our results New Biol. 240, 130. that a set of weak interactions between ligand and enzyme site 26. MacQuarrie, R. A. & Bernhard, S. A. (1971) J. Mol. Biol. can to 55, 181-192. lead strong perturbation of chemical bonds, a mecha- 27. Hanson, K. R. (1968) J. Mol. Biol. 38, 133-136. nism that has often been proposed without direct evidence. 28. Conway, A. & Koshland, D. E., Jr. (1968) Biochemistry 7, The present results also provide a basis for invoking a coupling 4011-4023. Downloaded by guest on September 27, 2021