The Glyceraldehyde 3-Phosphate Dehydrogenases of Liver and Muscle COOPERATIVE INTERACTIONS AND CONDITIONS FOR FUNCTIONAL REVERSIBILITY

SUMMARY A method is described for the isolation of glyceraldehyde 3.phosphate dehydrogenase from rabbit liver. The enzyme has been crystallized as the NAD complex and its chemical and physical properties have been compared with those of the muscle enzyme. The kinetics of the reversible reaction catalyzed by the dehydrogenases is sensitive to temperature and ionic strength and has been examined at 37, pH 7.4, in a solvent of ionic strength 0.1. Under these conditions the outstanding new feature of the reverse reaction, the reductive dephosphorylation of 3-phosphoglyceroyl phosphate, is the positive cooperativity of the response of both enzymes to concentration increments of the acyl phosphate and the increase in this cooperativity as a function of concentration of the cosubstrate, NADH. At low concentrations of the acyl phosphate, the reduced coenzyme exerts an inhibitory function which is abolished by high concentrations of acyl phosphate or by low concentrations of NAD acting as a heterotropic effector. The same concentrations of NAD which activate at low acyl phosphate concentration inhibit at high concentrations of the 3-carbon substrate by competing with NADH. Activation of the reaction by NAD involves a transition of the acyl phosphate saturation function from a sigmoidal to the hyperbolic form and hence sensitizes the enzyme to lower concentrations of acyl phosphate. The substrate inhibition by NADH and the activation by NAD may be accounted for by the binding of the nucleotides to catalytic sites of the tetrameric protein that are not occupied by the acyl enzyme intermediate. Although the competitive advantage in binding to the apoprotein strongly favors NAD over NADH, the competition under reaction conditions tends to be equalized at nonacylated sites of the acylated protein and is drastically reversed at the acylated sites. Thus at kinetic saturation levels of the acyl phosphate KNADH remains in the IO-PM range, but half-maximal inhibition by NAD occurs at concentrations of the latter that are in excess of 1000 PM. These properties involve both ligandinduced conformational transitions of the protein and a local steric effect of the acyl group that interferes with the binding * This work was supported in part by Research Grant AM-13115 from the National Institutes of Health.

A method is described for the isolation of glyceraldehyde 3.phosphate dehydrogenase from rabbit liver. The enzyme has been crystallized as the NAD complex and its chemical and physical properties have been compared with those of the muscle enzyme. The kinetics of the reversible reaction catalyzed by the dehydrogenases is sensitive to temperature and ionic strength and has been examined at 37, pH 7.4, in a solvent of ionic strength 0.1. Under these conditions the outstanding new feature of the reverse reaction, the reductive dephosphorylation of 3-phosphoglyceroyl phosphate, is the positive cooperativity of the response of both enzymes to concentration increments of the acyl phosphate and the increase in this cooperativity as a function of concentration of the cosubstrate, NADH. At low concentrations of the acyl phosphate, the reduced coenzyme exerts an inhibitory function which is abolished by high concentrations of acyl phosphate or by low concentrations of NAD acting as a heterotropic effector. The same concentrations of NAD which activate at low acyl phosphate concentration inhibit at high concentrations of the 3-carbon substrate by competing with NADH. Activation of the reaction by NAD involves a transition of the acyl phosphate saturation function from a sigmoidal to the hyperbolic form and hence sensitizes the enzyme to lower concentrations of acyl phosphate.
The substrate inhibition by NADH and the activation by NAD may be accounted for by the binding of the nucleotides to catalytic sites of the tetrameric protein that are not occupied by the acyl enzyme intermediate. Although the competitive advantage in binding to the apoprotein strongly favors NAD over NADH, the competition under reaction conditions tends to be equalized at nonacylated sites of the acylated protein and is drastically reversed at the acylated sites. Thus at kinetic saturation levels of the acyl phosphate KNADH remains in the IO-PM range, but half-maximal inhibition by NAD occurs at concentrations of the latter that are in excess of 1000 PM. These properties involve both ligandinduced conformational transitions of the protein and a local steric effect of the acyl group that interferes with the binding * This work was supported in part by Research Grant AM-13115 from the National Institutes of Health.
$ Predoctoral trainee 1965 to 1969, supported by Training Grant GM-00152 from the National Institutes of Health. of the oxidized but not the reduced form of the pyridine nucleotide.
In the forward or oxidative phosphorylation reaction the reciprocal kinetic plots are entirely linear except under conditions of product inhibition by NADH. The preferential binding of NADH at an acylated site is not strongly expressed in the inhibition by added NADH because such sites are already occupied by product NADH that is released subse. quent to the rate-limiting acyl group transfer from enzyme to orthophosphate ion. The acyl phosphate is a strong inhibitor of the forward reaction, noncompetitive with NAD and strictly competitive with aldehyde.
Intracellularly, the enzyme always operates in the presence of high concentrations of NAD which are required to provide the driving force for the energy-conserving reaction. Functional reversibility, under the adverse substrate concentration distributions in the aerobic hepatocyte, is maintained by regulatory effects of NAD binding which promote the reaction of extremely low concentrations of acyl phosphate, and by the differential effects of acylation of the protein which allow access of substrate NADH and at the same time eliminate a prohibitive inhibition by the high concentrations of NAD. The catalytic properties of the liver and muscle enzyme are qualitatively similar but exhibit some quantitative differences that would favor the more diversified metabolic requirements of liver. However, no definitive structural differences have been identified. tions catalyzed by the dehydrogenase and 3-phosphoglycerate kinase remain in equilibrium over a wide range of metabolic conditions (2).
To examine the possibility that the dehydrogenase is synthesized in tissue-specific forms the rabbit liver enzyme has been crystallized and analyzed.
Although the occurrence of one or a few conservative ammo acid replacements has not been excluded, no definitive structural differences have been identified.
The previous kinetic analysis had been done at 26O and low ionic strength.
Since the enzyme has a strong affinity for anions (3) and has several properties which exhibit a marked temperature dependence (4), the reaction kinetics has been re-examined at a physiological temperature and in a solvent of an ionic strength that more closely approximates natural levels. Under these conditions, the reciprocal interactions between NAD and the acyl enzyme intermediate which had been assigned a prominent role in the kinetics of glyceraldehyde 3-phosphate oxidation, but not in its reversal, are now found also to dominate the steady state kinetic behavior of the enzyme in the reduction of 3-phosphoglyceroyl phosphate.
Under the new experimental conditions, which simulate critical features of the natural reaction environment, the dehydrogenase exhibits a combination of homotropic and heterotropic cooperative responses that operate in a way to keep the reaction freely reversible under conditions that would be highly adverse to the reversible action of an enzyme monomer. This is in contrast with the behavior of numerous other allosteric enzymes, the kinetic behaviors of which serve to regulate the activity of a metabolic pathway in a single direction.

Materials
Pyridine nucleotides, supplied in previously assayed vials, ATP, and the barium salt of m-glyceraldehyde a-phosphate diethyl acetal were products of Sigma Chemical Co. There was less than 0.03% NADH in the NAD.
The cyclohexyl ammonium salt of n-glyceraldehyde 3-phosphate dimethyl acetal, synthesized by the method of Ballou and Fischer (5) was the gift of Dr. C. D. Gutsche.
Phosphoglycerate kinase, crystallized from yeast, and the sodium salt of n-3-phosphoglycerate were obtained from Boehringer and Sons. n-3-Phosphoglyceroyl phosphate, made enzymically from the m-aldehyde and purified by ion exchange chromatography, was prepared by a previously described method (1). The glyceraldehyde 3-phosphate dehydrogenase of rabbit skeletal muscle was prepared by minor modifications of the method of Cori, Slein, and Cori (6) and maintained throughout the preparation and in storage in the presence of dithiothreitol and EDTA. The dehydrogenase of yeast was prepared essentially by the method of Krebs (7) and handled as described elsewhere (4).
Trypsin, five times recrystallized, was a product of Worthington Chemical Co. lo-Methyl-3(3-dimethylaminopropyl)carbodiimide was obtained from Ott Chemical Co. Reagent grade urea was deionized and recrystallized before use. Imidazole (Aldrich) was treated with charcoal and recrystallized from benzene to remove fluorescent impurities.

Methods
Chemical Analysis-Amino acid analyses of 16-to 72-hour hydrolysates of the X-carboxymethylated protein (8) were made on the long columns of an amino acid analyzer by standard methods. Free carboxyl groups were estimated by the carbodiimide activation of glycine methyl ester addition and the subsequent measurement of the increase in protein-bound glycine, following the procedure of Hoare and Koshland (9). The reaction was run in 8 M guanidine hydrochloride at pH 4 for 2 hours at 25" in a pH stat. For tryptic peptide mapping, hydrolyses were carried out at pH 8.5 in 0.1 M ammonium bicarbonate for 12 hours at 24" with additions of 0.02 mg of trypsin to 4.1 mg of S-carboxymethylated protein at 0, 4, and 7) hours. Chromatography in I-butanol-acetic acid-water (4: 1: 5) and electrophoresis at pH 3.7 in pyridine-acetic acid-water (1: 10 : 289) were performed as described by Katz et al. (10). In addition to ninhydrin staining, replicate maps were also stained with Ehrlich's reagent for tryptophan and with 8-hydroxyquinoline and bromine for arginine (11,12). Free thiol groups of the native enzyme and of enzyme denatured in 8 M guanidine hydrochloride were measured spectrophotometrically by reaction with sodium dithio-bis-orthonitrobenzoate (13). Amino end groups were determined as the 2,4-dinitrophenyl derivates (14) and by Edman degradation (15). Ebctrophoresis-Zone electrophoresis at pH 6.5 and 8.5 was done on cellulose acetate (Microzone) and at the alkaline pH in polyacrylamide gel. Electrofocusing measurements were made in a llO-ml column containing Ampholine (LKB) in a sucrose gradient.
The column was operated at 1000 volts, 10" for 3 to 6 days. No losses of enzyme activity were observed under these conditions.
The columns were emptied through an ultraviolet absorbance monitor, and the absorbance and specific enzyme activities of individual fractions were measured on a Cary 15 spectrophotometer.
Sedimentation-Molecular weight measurements by equilibrium sedimentation employing a filled Epon capillary-type double sector centerpiece, column heights of 2.3 to 2.7 mm, and by interference optics were made essentially as described in the Spinco model E manual with the modification that an initial overspeed was used to shorten the equilibration time (16). Fringe displacements were measured in a Gaertner microcomparator.
Sedimentation velocity runs in a 4" single sector aluminum centerpiece were monitored by schlieren optics. Diffusion measurements were made in the free boundary cell of the Aminco, model H, electrophoresis apparatus and were monitored by interference optics.
We are indebted to Dr. Robert Yue for the diffusion measurements.
Kinetic Measurements-Initial velocity measurements of the forward reaction, that of aldehyde oxidation, were measured fluorometrically.
For the purpose, a simple filter fluorimeter was used with a low pressure mercury arc. The signal was taken directly from the photomultiplier to an 11-inch adjustable zero, adjustable range strip chart recorder (Leeds and Northrup) giving a full scale deflection with 0.6 PM NADH.
When still higher sensitivity was required, the recorder was replaced by a Tektronix model 568 storage tube oscilloscope sweeping at 10 set per scale division and allowing the measurement of linear initial rates during the production of as little as 2 x lO+ M NADH.
The high sensitivities were required to distinguish between some of the strong primary and secondary actions of substrates and products.
The same apparatus could be used for the reverse reaction at initial NADH concen-Orations of 4 I.~M or less. However, at higher NADH concentrations, the large zero suppression at high sensitivity resulted in a drifting signal  addition, contains 0.1 M potassium chloride. n , l , glyceraldehyde-3-P dehydrogenase of muscle; 0, 0, the liver enzyme. due to phototube fatigue.
A Cary model 15 spectrophotometer and the 0 to 0.1 absorbance slide wire with cells of lo-cm light path was therefore also used for the reverse reaction.
Product inhibition presented no technical problem in this reaction direction, but the filling and mixing time of the lo-cm absorption cells was longer than desirable.
In some instances, an open rectangular cell of 5-cm light path was used allowing a dead time after manual enzyme addition of a few seconds.
Stock dilutions of enzyme were kept at the highest allowable concentration, and additions to the reaction mixtures were made with 5-or lo-p1 pipettes.
Small normalizing corrections were required for activity decay of stock solutions over a period of several hours. Since the enzyme concentrations in the test solutions were in the extremely dilute range of 10-3 pg ml-1 and hence susceptible to accelerated decay, activity controls covering the duration of a single initial velocity measurement were also made by sampling this concentration of enzyme incubated for various periods of time in several incomplete reaction mixtures.
Linear initial velocities lasting up to 2 min were acceptable without correction.

Purijication
and Analysis of Liver Enzyme Initial E&act-The livers of six adult rabbits were promptly removed, diced, and homogenized in a blendor at low speed for 20 set in 2 volumes of cold 0.4 M sucrose containing 1% sodium chloride, 5 mM EDTA, and dithiothreitol.
The pH of the suspension was adjusted to 5.8 with hydrochloric acid and the suspension was centrifuged at 23,000 x g for 20 min at 0".
Precipitations with Ethanol-After adjusting the pH of the extract to 6.8, 1 volume of 95% ethanol previously chilled to -5" was added at a rate that maintained the solution temperature below +5".
The suspension was then spun at 20,000 x g for 15 min at -10'. Approximately 65% of the original activity was precipitated in this step, together with a large amount of denatured protein.
The packed precipitate was resuspended by a brief homogenization in 100 to 150 ml of 0.05 M potassium phosphate, pH 7.0, containing 5 mM EDTA, 5 mM dithiothreitol, and 1 mM NAD.
Insoluble material was removed by ultracentrifugation at 78,000 x g for 3 hours at 0". The enzyme in the supernatant solution was reprecipitatecl exactly as described above, extracted from the precipitate with 5 to 15 ml of the same extracting solution, and clarified by ultracentrifugation as before. Ammonium Sulfate Fractionation-The clarified enzyme solution was dialyzed in a l-cm diameter casing against 150 ml of the phosphate potassium chloride buffer to remove residual ethanol. Weighed portions of ammonium sulfate were added, basing the concentrations upon the saturation table of Green and Hughes (17). Precipitates obtained at 0.55 and 0.59 saturation were allowed to form for 1 hour in the cold and were removed by centrifugation and discarded. The ammonium sulfate was then increased to 0.64 saturation, and the pH was raised to 8.3 with ammonium hydroxide. Crystals appeared overnight at 4". They were usually contaminated with a smaller molecular weight protein that was retained on a DEAE-Sephadex column equilibrated with 0.01 M potassium phosphate, pH 7.0. On such a column the dehydrogenase emerges with the solvent front.
Second crystals obtained under the above conditions contain approximately 3 moles of NAD per 144,000 g of protein.
The purification is summarized in Table I. If NAD is omitted from the extraction solutions in the above procedure apoenzyme in small yield may be obtained after arduous refractionation with ammonium sulfate.
Crystals of the holoenzyme may be obtained either as thin plates resembling those of the muscle enzyme or as stout rods.
Molecular Weight-Sedimentation constants of the liver and muscle enzyme as a function of protein concentration are plotted in Fig. 1. The reported protein concentrations are averages over the time of the experiment, corrected for radial dilution. Within small limits of error, the sets of points extrapolate to the same value of l/80 and exhibit the same slope. ~20,~ for both enzymes is 7.60. Converted to standard conditions, the diffusion constant is Dzo,u, = (5.12 f 0.041) x lo+ cm2 set-I, compared with the value of (4.97 f 0.03) x lo+ for the muscle enzyme reported by Fox and Dandliier (18). With the calculated anhydrous partial specific volume of 0.740, the molecular weight of the liver enzyme by this method is 139,000. Sedimentation  b The 24-hour muscle analysis is the average of three determinations from separate hydrolysates.
The other columns are single hydrolysates.
All have been normalized to 32.0 alanine residues per subunit. 0 Reference 20. d Reference 21. 0 Cysteine was determined as the carboxymethylated derivaequilibrium runs on two preparations give linear plots of In J against r2. At protein concentrations of 1.5 and 4.6 mg ml-l, molecular weights of 143,000 and 139,000, respectively, were obtained.
Schlieren patterns of sedimentation velocity runs in 5 M guanidine hydrochloride showed no sign of asymmetry. The sedimentation constant of the dissociated liver enzyme is ~20,~ = 1.9 f 0.09 compared with a value of 1.82 reported by Harrington and Karr (19) for the muscle enzyme and corresponding to a subunit molecular weight of 36,000.
Amino Acid Analyses-The results of amino acid analysis are listed in Table II. Although a few single amino acid replacements cannot be excluded, there are no definitive differences in the composition of the rabbit liver and muscle enzymes. Moreover, the results are in close agreement with those tabulated from the total sequence analysis of the pig muscle enzyme by Harris and Perham (20). The NHt-terminal groups were identified as valine by paper chromatography of the dinitrophenyl derivatives and were quantitated as the phenyl thiohydantoins, 3.9 f 0.1 moles per mole of tetramer for both proteins.
In the free carboxyl group analysis, 30 2. Electrofocusing of the apoglyceraldehyde-3-P dehydrogenase of rabbit muscle after 135 hours at 10" in a pH 8 to 10 ampholine sucrose gradient.
Activity is expressed in units of pAtno min-1 under the standard assay conditions of Table I. --s 22 c! -1.08 3. Electrofocusing of the apoglyceraldehyde-3-P dehydrogenase of rabbit liver for 65 hours under the conditions of Fig. 2. and muscle enzymes. If the value of 39 aspartyl plus glutamyl residues per subunit of the pig muscle enzyme applies to the rabbit enzymes, the observed 31 to 32 free carboxyl groups in the latter correspond to an 80% recovery and the difference may not be significant.
Yields of 80 to 90% had been reported for proteins of known structure (9). Our own control analysis of lysozyme by this method appeared to be quantitative.
Within the limits of variation of independent runs on hydrolysates of the same protein, the tryptic peptide maps of the liver and muscle enzymes showed no reproducible differences. As required for four identical subunits, 10 of the resolved spots stained for arginine.
However, in both cases only two of the expected three stained for tryptophan.
The free thiol group content of the liver enzyme, as routinely isolated and measured by titration with dithio-bis-orthonitrobenzoate, tends to be low, but the expected 8 mole equivalents per mole of native tetramer are obtained after incubation with dithiothreitol and dialysis; 16 mole equivalents (hO.5) are measured after similar treatment in 4 M urea. A low thiol content has been reported for the calf liver enzyme by Heinz and Kulbe (22) who isolated the apoprotein.
Several marked differences between the calf liver and rabbit muscle enzymes were reported by these workers, but because of some inconsistencies in the analyses the calf protein requires further examination.
Electrophuresis-The liver and muscle enzymes migrate as single slightly diiuse bands in polyacrylamide gel or cellulose  23) for the muscle enzyme is readily confirmed with holoenzyme preparations from both tissues. However, we have obtained similar results with the pig muscle enzyme which has exhibited no structural variation in a total sequence analysis (22). The heterogeneity is greatly decreased when enzyme, stripped of its bound NAD by charcoal treatment in dithiothreitol, is examined. Electrofocusing curves obtained with the muscle and liver apoenzymes are shown in Figs. 2 and 3. The isoionic point of the major component in both cases is 8.52 f 0.02 as expected from the amino acid compositions and as predicted from the ionic strength dependence of the isoelectric points and the anion-binding properties (24). The isoionic point of the minor component, constituting about 15% of the total is 8.32 f 0.03. The major band exhibits maximal or near-maximal enzyme activity even after 3 to 5 days of electrofocusing.
The specific activity of minor band is only slightly lower after an approximate corection for the somewhat variable base-line absorbance.
When the bands are isolated and refocused, they concentrate at their original positions.
A small p1 (isoelectric point) difference in the direction observed for the minor band could arise from a component that still carried bound nucleotide, from differential ampholme binding, from amide group variation, or from a conformational isomerism. A thiol to disulfide conversion would shift the p1 in the opposite direction since it is likely that the exposed thiol groups are ionized at pH 8.5 (25).

KINETIC RESULTS
Glyceraldehyde 3-phosphate dehydrogenases isolated from widely diverse species in the evolutionary scale constitute a family of homologous tetramers each composed of a single type of subunit (26, 27). The reversible catalytic reaction has been studied by a variety of independent methods (24) and proceeds through a substituted enzyme intermediate.
In the direction of aldehyde oxidation, an acyl enzyme thioester is formed by the oxidation by NAD of a presumed thiohemiacetal adduct of aldehyde and enzyme.
The acyl group is then transferred to orthophosphate ion. The acyl enzyme intermediate in the reverse reaction is formed by phosphate displacement from the acyl phosphate and is then reduced by NADH.
In the absence of NADH there is a powerful activation by NAD of phosphate exchange between acyl phosphates and orthophosphate (28). The interaction between NAD and the acyl enzyme intermediate is a reciprocal one since acylation of the enzyme promotes the release of bound NAD (1) and high NAD concentrations shift the acyl group transfer equilibrium in the direction of acyl phosphate formation (29). An activation by NAD of the reductive dephosphorylation of acyl phosphate, implied by the above results, was not observed in steady state kinetics at pH 7.5 and 25" at low ionic strength (1) but was found qualitatively by de Vijlder et al. (30) to occur under special conditions. This effect has been studied in greater detail by Trentham (31) who employed the lobster and sturgeon muscle enzymes and measured kinetic transients by the stopped flow method.
Trentham's work has provided independent evidence that bound NAD promotes the acyl group transfer in both reaction directions and also accelerates the formation and dissociation of the enzyme complex with glyceraldehyde 3-phosphate. Kinetic transients provide a powerful approach to certain properties of the catalysis but have been restricted to high enzyme and substrate concentrations.
However, the dehydrogenase acts on extremely low concentrations of acyl phosphate and NADH.
These concentrations may be studied by the steady state kinetic method, and under the appropriate conditions the multiple functions of NAD and new properties of the enzyme interactions with the acyl phosphate and NADH are observable.
Our initial objective was to search for functionally significant differences between the liver and muscle enzymes. As it turns out, the two enzymes are qualitatively quite similar but share properties that had not been fully recognized and which are pertinent both to mechanism and intracellular function.

Reductive Dephosphorylation of 3-Phosphoglyceroyl Phosphate
Salt E$ects-Initial velocities at sets of fixed substrate concentrations were measured in 0.01 M imidazole at pH 7.4 and 37" as a function of the concentrations of a series of neutral univalent electrolytes.
The results were similar with the salts examined and involve some minor ion specificities superimposed upon a general ion strength effect. The largest effects were given by sodium and potassium chlorides and are illustrated for potassium chloride in Fig. 4. Initial velocities increase with the concentration of potassium chloride to a maximum at about 0.06 M and then decline with increasing concentrations of the salt. The form of the kinetics as well as the absolute rates change as one passes from the rising to the descending portions of the salt concentration curve. Linear reciprocal plots are obtained only at low ionic strength in the absence of potassium chloride.
In the measurements to be described 0.1 M potassium chloride was selected as the supporting electrolyte.
At this salt concentration, the enzyme, although not at the activity maximum, is still in the salt-activated state and also exhibits optimal stability. The initial velocities are also more reproducible where the ionic strength functions are less steep.
Cooperative E$ects-Reciprocal plots of initial velocities at sets of fixed concentrations of the cosubstrates are shown in reciprocal NADH lines are also curved but diverge with increasing concentrations of NADH.
Such behavior is indicative of a strong substrate inhibition by NADH that is opposed by increasing concentrations of the acyl phosphate.
Thus, at 20 PM acyl phosphate, the reciprocal NADH plot is linear, corresponding to a hyperbolic saturation function and an apparent KNADH of about 7 PM. Direct plots of vi against [Acyl-Pj are sigmoidal, and the apparent cooperativity increases with NADH concentration.
Each subunit of the dehydrogenase binds 1 molecule of NAD or NADH competitively (32), and the subunits are believed to be equivalent in the apoprotein.
Whether or not a given site exhibits a catalytic or an inhibitory function when occupied by NADH should depend upon its occupancy by the cosubstrate. Thus, at low acyl phosphate concentration NADH may bind to an otherwise vacant site and act as an inhibitor.
At high acyl phosphate concentrations, the inhibitory sites disappear and only the substrate function of NADH is expressed. In such a model This interpretation is supported by the following effects of NAD upon the rates of acyl phosphate reduction.
Activation and Inhibition by NAD-The effects of NAD concentration upon the reductive dephosphorylation of acyl phosphate by NADH are shown for the liver and muscle enzymes in Figs. 6 and 7. At subsaturating concentrations of acyl phosphate, the enzyme activity increases and passes through a maximum as a function of the concentration of NAD. The broad activity maximum occurs in the range of 80 to 150 PM NAD, is not identical for the two enzymes, and is followed by an inhibition that increases slowly with further increments of NAD concentration.
The kinetic basis of these effects is indicated by the reciprocal plots of Fig. 8. In the absence of NAD the Issue of January 10, 1972 C. M. Smith and S. F. Velick 279 curve is typically steep and concave upward. NAD addition drastically reduces the slope and abolishes the curvature. A change of this type is equivalent in a direct plot to the transformation of a sigmoidal to a hyperbolic saturation function for acyl phosphate.
Activation occurs only at low acyl phosphate concentrations and results from a diminution in the effective Micha,elis constant of the acyl phosphate.
The maximal velocity is diminished to an extent that depends upon the concentration of NAD.
In the inhibitory region NAD is competitive with NADH (Fig. 9).
These results follow the pattern exhibited by the NADH inhibition.
Activation of the reaction by NAD at low acyl phosphate concentration should result from NAD binding to a site that is unoccupied by the 3-carbon substrate.
In so doing, NAD competitively opposes any NADH inhibition at that site and also promotes the intramolecular acylation step on another subunit.
A conformational transition of the protein associated with NAD binding is implied.
Independent evidence for the occurrence of opposing conformational transitions associated with the binding of the conjugate forms of the pyridine nucleotide is provided by the fact that NAD stabilizes (1) and NADH destabilizes (33) the apoprotein.
The activation of the reaction by NAD not only eliminates the cooperative enzyme response to acyl phosphat,e but also blocks a site which at high acyl phosphate concentration would act catalytically. Hence, at high acyl phosphate concentrations, which require no heterotropic activation, NAD limits the maximal velocity.
There is a great disparity, at low acyl phosphate concentrations, between the NAD concentrations that activate and those that inhibit.
Half-maximal activation occurs in the range of 5 to 10 PM NAD depending in part upon the concentration of NADH.
Half-maximal inhibition is not clearly defined by the data but occurs at NAD concentrations in excess of 1000 PM. The difference may be interpreted as a measure of the relative binding affinities of NAD at nonacylated and acylated sites. In the absence of other ligands, the dissociation consta,nt of the high affinity enzyme-NAD complexes at 37" is about 3 X lO+ M (4). The somewhat higher concentration of NAD required for halfmaximal activation may be attributed to competition with inhibitory NADH and to indirect conformational effects of ncylation. The much weaker binding of NAD at an acylated site, where it would form an inhibitory dead end complex, should arise from direct steric interference by the enzyme-bound acyl group. This behavior is consistent with the reciprocal interactions between NAD and acyl phosphate that have been cited. An important additional property of the steric interference by the acyl group is that it is specific for the oxidized form of the pyridine nucleotide, since the kinetic K,,,, at high acyl phosphate concentration remains in the low micromolar range. Geometric and possible conformational differences of the bound conjugate forms of the pyridine nucleotides (34) are sufficient in principle to account for this degree of specificity.
All of the above reactions were initiated by the addition of enzyme-NAD complex at a final concentration of about 10-n M. Holo-rather than apoenzyme was used to improve the stability of the dilute stock solutions.
Initial velocities were measured during the production of as little as 2 X IO-* M NAD and could be extrapolated back linearly to the origin at zero time. Although it is unlikely, at 37" in the presence of NADH and acyl phosphate, that any significant fraction of the enzyme remained complexed with NAD, the enzyme was added in the active form. It may be presumed to have been held in that state by acylation and the effects of NAD as a leaving group during the several thousand reaction cycles per min that were measured.
Inhibitions by Aldehyde and Orthophosphate-The main conclusion to be drawn from Fig. 10 is that glyceraldehyde 3-phosphate is an inhibitory product but that, unlike NAD, it does not relieve the substrate inhibition by NADH nor does it eliminate the cooperative response of the enzyme to acyl phosphate.
In plotting these results, as well as those that involve the aldehyde as a substrate of the forward reaction, the aldehyde concentration was taken as the total concentration of n-aldehyde that was present in the DL mixture.
Calculated on this basis, the rates with an authentic sample of the n-aldehyde, synthesized by a stereospecific method (5) and the same concentration of the n-aldehyde present in the racemate were identical.
The unnatural L-aldehyde was therefore neither a substrate nor an inhibitor under our conditions. Recent findings of Trentham et al. (35) indicate that in aqueous solution at 20" the aldehyde is actually 970/, in the hydrated gem-diol form which does not interact with the enzyme. Assuming this conclusion to be correct and taking the absorbance of the aldehyde solution at 280 nm to be a measure of the concentration of the free or active aldehyde,  I  I  I  I  I  I  I  I  I   60 we calculate that at 37" the active aldehyde concentration is 10% of the total D form present. On this basis the aldehyde concentrations in Fig. 10 should be divided by 10, and the aldehyde should be considered to be a relatively strong inhibitor but not anomalously strong as was reported for low ionic strength and 26" (1). The altered behavior of the aldehyde as an inhibitor under the present conditions is primarily a temperature and not an ionic strength effect.
An orthophosphate concentration in the physiological range of 3 mm has no effect on the kinetic parameters of NADH and exerts an approximate 10% inhibition directed against the acyl phosphate at concentrations of the latter from 2 to 8 PM.

Oxidative Phosphorylation of Glyceraldehyde S-phosphate
Potassium Chloride E$ects--As shown in Fig. 11, the effects of potassium chloride concentration upon the initial velocities of the forward reaction are similar to those on the reverse reaction, maximal activation occurring at about 0.06 M followed by a progressive inhibition with increasing salt concentration. be noted that at 15 mM phosphate the curve for the muscle but not the liver enzyme is shifted to the left. This is one of several minor differences between the two enzymes that are expressed primarily at low ionic strengths. The kinetic properties of the two enzymes, except for some quantitative differences, are formally quite similar and only the more extensive results obtained with the muscle enzyme will be presented.
Kinetics of Reaction-The initial velocities in Figs The outstanding feature of the results, compared with those of the reverse reaction, is the linearity of the plots.
Thus, the NAD concentration function exhibits no significant deviation from a straight line over the range of 20 to 2000 PM. Over the concentration range tested the only substrate cross reaction expressed in the K, values is between aldehyde and orthophosphate ion. At high concentrations of the latter, the K,ldehyde is increased.
In a direct plot at fixed aldehyde and N-4D concentration, this would be expressed as a substrate inhibition by phosphate and is probably the converse of the substrate inhibition by aldehyde previously reported (1). In the latter work, arsenate had been used in place of phosphate and was used at concentrations two to three orders of magnitude lower than the phosphate concentrations employed here. Under the present conditions there is no significant substrate inhibition by aldehyde.
Insofar as can be determined under the present conditons, the enzyme at any fixed concentration of two of its three substrates in the forward reaction functions in a single conformational form. This may be attributed to the combined effects of NAD as a substrate and activator and to the occurrence of high steady state concentration levels of the acyl enzyme intermediate. As in the earlier work, the acyl group transfer from enzyme to orthophosphate ion is considered to be rate limiting at pH 7.4 The KNAD of about 80 pM satisfies kinetic criteria for a dissociation constant of the complex of NAD with the acylated enzyme. Its magnitude relative to the smaller dissociation constant of the complex in the absence of aldehyde is attributed to the indirect or conformational effect of acylation.
The results are consistent with the conclusions of Smith (37) that NADH in the forward reaction is released subsequent to the acyl group transfer.
If NAD promoted acyl group transfer to orthophosphate in the forward reaction by displacing product NADH from an acylated site, we would expect nonlinear kinetics and a range of activation by NAD comparable to the extended range of NAD inhibition of the reverse reaction, and this is not observed.
However, as shown in the following section, a nonlinear function of NAD concentration is obtained in the presence of NADH added as an inhibitory product.  with respect to both substrates. One of these plots is shown in Fig. 15. Similar results at a fixed concentration of NADH with NAD and aldehyde as independent variables are shown in the double reciprocal plots of Fig. 16. At a low concentration of one substrate and high concentrations of the other two, 6 PM NADH is a strong inhibitor.
The concentration dependence of the rates of the inhibited reactions are nonlinear with respect to both of the independently varied substrates and resemble the response in the reverse reaction to acyl phosphate in the absence of NAD.
In terms of the above arguments, the primary action of NAD in overcoming the NADH inhibition should occur in the competition with NADH for a nonacylated site. Although there is an extremely strong competitive advantage of NADH over NAD at an acylated site, such sites are already occupied by product NADH which is released subsequent to the acyl group transfer.
Hence, competition at these sites is not expressed in the forward reaction. The opposition to NADH inhibition by increasing concentrations of glyceraldehyde 3-phosphate should be a consequence of an increased steady state concentration of the acyl enzyme intermediate and a corresponding decrease in the number of nonacylated sites available for the inhibitory action of the reduced coenzyme.  Tables III and IV are  practical quantities, the substrate concentrations which give half-maximal velocities under the defined sets of conditions. In some cases they correspond to apparent Michaelis constants derived from linear reciprocal plots, and in others they are the estimated midpoints of sigmoidal kinetic saturation curves. The important features of the reverse reaction are the small magnitudes of the half-saturation values and their dependence upon the positive or negative effector functions of the pyridine nucleotides. -4pparent maximal velocities with respect to a given substrate at the designated concentrations of the others are given only in relative units to indicate the magnitudes of the effects.
Observed rates in absolute units of enzyme turnover are in the range of 3000 to 8000 nun-* and are in excess of 20,000 mine1 response of the enzyme to concentration increments of the acyl phosphate, readily observable in the test tube, is probably never when extrapolated to the limit for both substrates with maxi-expressed intracellularly. mally active protein.
The parameters of the forward reaction The intracellular conditions under which the dehydrogenase require little additional comment other than to mention that must maintain functional reversibility are imposed in part by KILldehyde may be as low as 2 to 5 plf if the enzyme sees only the the mechanism and energetics of the reaction itself and in part unhydrated form at 37". by the requirements and properties of t.he numerous other osi-The reductive dephosphorylation of acyl phosphate catalyzed dation-reduction and phosphorylation reactions with which it is by the dehydrogenase exhibits many of the kinetic properties coordinated.
Without the violation of any fundamental prin described by the allosteric transition model of an oligomeric pro-ciples, except economy, the problems imposed upon the tein proposed by Monod ef al. (38). However, complica.tions dehydrogenase might have been solved in nature by the evolution are introduced by the occurrence and properties of the substituted of two different monomeric enzymes, each catalyzing the same enzyme intermediate, the multiple functions of the pyridine over-all equilibrium but with different kinetic parameters. By nucleotides, and the ability of the four active centers of the pro-taking advantage of the interaction capabilities of the appropritein to play either a catalytic or a regulatory role depending ate oligomer, a single genetic message has s&iced to provide the upon the partition among them of the five substrates of information for the formation of a tetramer, proportionate parts the reversible reaction.
Conformational transitions between of which can act alternately and in combination as an NAD enzyme states of different catalytic activity are clearly associated enzyme and an NADH enzyme. with pyridine nucleotide binding and some of the general relationships that prevail under reaction conditions have been REFERENCES indicated.
These properties do not correlate in detail with the against the reverse reaction.