THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 249, No. 1, Issue of January 10, pp. X32-190, 1974

Printed in U.S.A.

Cooperative Interactions in the Binding of Allosteric Effecters

to Phosphoenolpyruvate Carboxylase*

(Received for publication, April 23, 1973)

RONALD SMANDO, E. BRUCE WAYGOOD,~ AND B. D. SAN’VVAL~

From the Department of Medical Genetics, Medical Sciences Building, UwIversity of Toronto, Toronto 181, Canada

SUMMARY

Phosphoenolpyruvate carboxylase from Salmonella typhi-

murium is shown to consist of four subunits of molecular weight about 100,000. Guanidinium chloride and 0.1% sodium dodecyl sulfate only partially dissociate the carboxy- methylated subunits into monomers of molecular weight of about 50,000. Only one NH*-terminal residue, methionine, is found by the dansylation procedure, which together with peptide mapping results suggests that the enzyme is made up of identical subunits.

The enzyme binds 4 moles of L-aspartate and 4 moles of acetyl coenzyme A, i.e. each subunit, although it seems to be made up of two monomers, carries only one binding site for each one of these ligands. The binding of aspartate in the absence of allosteric activators (acetyl-CoA, guanosine tri- phosphate, and fructose diphosphate), or with the activators present singly, is hyperbolic. The dissociation constant of aspartate is 0.09 mM, but this value varies between 0.1 mM and 0.25 mM in the presence of the various activators. When two activators are present simultaneously, not only does the binding curve for aspartate become sigmoidal but the dis- sociation constant increases drastically. The dissociation constant for aspartate, for instance, in the presence of acetyl- CoA and fructose diphosphate together is about 1.5 mM. The binding of acetyl-CoA to the enzyme, in contrast to aspartate, is sigmoidal at low concentrations, but the binding curve shows “tailing” at higher concentration of this activa- tor. A likely cause of this tailing is that acetyl-CoA binds nonspecifically at higher concentrations to some other site on the enzyme surface. In the presence of a second activator, such as GTP, the binding curve for acetyl-CoA becomes hyperbolic and the dissociation constant for this ligand is considerably reduced. Neither magnesium nor manganese is required for the binding of the various ligands on the en- zyme surface. The equilibrium binding results are inter- preted on the basis of a two-state model, where the enzyme in the absence of any ligands is assumed to be present pre- dominantly in a state which binds the inhibitor, aspartate, preferentially. The cooperative interaction of activators re-

* This study was supported by an operational research grant from the Medical Research Council of Canada.

$ Recipient of a Medical R.esearch Council Studentship. $ Present address, Department of Biochemistry, University of

Western Ontario, London 72, Ontario, Canada. Reprint requests may be sent to this address.

suits in changing the conformation of the protein to a low affinity state for aspartate.

We have been interested for some time (see review, Ref. 1) in the nature and mechanism of control of the enzymes of amphi- bolic pathways in bacteria. One of the important enzymes of this pathway in the enteric bacteria is phosphoenolpyruvate carbosylase (EC 4.1.1.31)) and it has been demonstrated earlier in the case of Salmonella typhimurium that the enzyme is in- hibited by L-aspartate (2) and activated by several phosphoryl- ated compounds such as acetyl coenzyme A (3), fructose 1 ,6-di- phosphate (4), and guanosine triphosphate (5). In addition, some organic solvents, such as ethanol and dioxane, and some highly charged polyamino acids (6) powerfully activate the en- zyme and in some cases desensitize it to feedback inhibitors. P-enolpyruvate carboxylase from another enteric bacterium, Escherichia coli, has been found by several groups of workers (7-9) to behave in an almost identical manner with the enzyme from S. fyphivmrium, at least as far as its regulation is concerned. Despite the close similarity in the regulatory properties of the enzymes from E. coli and S. typhimuritm, however, the molecular weight of the S. typhimurium enzyme has been reported (10) to be exactly half of that from E. coli (7-9). Since P-enolpyruvate carboxylase from S. typhimurium shows a great propensity toward age- and concentration-dependent aggregation-deaggre- gation, the accuracy of the molecular weight estimates may be in doubt. One of the aims of the present investigations, therefore, was to reassess the molecular weight of Penolpyruvate carboxyl- ase from S. typhimurium.

The multiplicity of controls operating on P-enolpyruvate car- boxylase has raised several questions in the past (6, 10) which could not be solved by steady state kinetic measurements. Two questions of great interest, from the point of view of the mode of action of allosteric enzymes in general, were, firstly, whether each one of the effecters of the enzyme had its own site on the surface of the enzyme, and, secondly, what was the actual mecha- nism of cooperative activation of the enzyme by a set of positive effecters. It had been shown earlier (1, lo), for instance, that the percentage activation of the enzyme in the presence of acetyl-CoA and fructose 1 ,6-P2 together was much higher than the sum of the percentages of activation in the presence of each

182

by guest on October 12, 2020

http://ww

w.jbc.org/

Dow

nloaded from

183

activator alone. This was also true of the activator pairs acetyl- CoA and GTP (5) or fructose 1,6-Ps and GTP. A phenomenon analogous to this kind of modulation, viz. cooperative inhibition of enzymes by inhibitor pairs, has been known for some time (ll), but the mechanism of this inhibition is not known. We therefore decided to investigate these problems by studying the characteristics of binding of the various regulatory ligands at equilibrium to P-enolpyruvate carboxylase purified from S. typhimurium. These studies, it was expected, would also con- tribute to a clarification of the mode of action of allosteric en- zymes in general.

MATERIALS AND METHODS

Reagents-+C]Acetyl-CoA, L-[G-i%]aspartate, and sodium dodecyl [a5S]sulfate were obtained from Amersham-Searle, Des Plaines, Illinois. The reagents were purified as described earlier (12). NaH14C03 was obtained from Nuclear-Chicago. Fructose-l, 6-P2 and various nucleotides were Sigma products. Sodium salt of P-enolpyruvate and crystalline pig heart malate dehydrogenase (specific activity, 720 i.u.) were obtained from Boehringer Mannheim. Guanidinium chloride was Mann (ultra- pure grade).

Culture Media and Cell-free Extracts-Cells of S. typhimurium, strain LT2, were grown in a minimal salts medium described earlier (5). Large batches of cells were obtained by growth (at 37”) in a New Rrunswick 200.liter capacity fermentor. The cells were harvested at late log phase, washed once in 0.05 M Tris-HCl buffer, pH 8.0, and resuspended in the same buffer at a final con- centration of 50%, wet weight, of cells. These were stored frozen (at -20”) until use. When required they were thawed at 4” overnight.

Enzyme Assay-P-enolpyruvate carboxylase was assayed routinely by a coupled spectrophotometric assay described earlier (10). For a few experiments a radioisotope assay was also used. The reaction mixture for this assay consisted of 10 mM NaH14C03 (0.1 PCi), 5.0 mM P-enolpyruvate, 10 mrvr MgCl,, and 0.1 M Tris- HCl, pH 9.0, in a volume of 1.0 ml. The reaction was started at 30” by the addition of the enzyme and terminated after 5 min by adding 1.0 ml of 2.0 x H2S04 containing 10 mg of carrier oxalace- tate per ml. The mixture was then gassed for 5 min with 95y0 carbon dioxide and aliquots were counted in a scintillation counter.

Protein Concenlration-The concentration of protein in impure preparations was determined by the method of Lowry et al. (13) and in homogeneous preparations by absorbance readings at 280 nm (E::,, = 14.5).

Peptide llfups-Peptide maps of P-enolpyruvate carboxylase were obtained by tryptic digestion according to the procedure described by Kanarek et al. (14). Electrophoresis in one direc- tion was performed at 2000 volts for 100 min in pyridine-acetate buffer, pH 3.7. Descending chromatography in the second direc- tion was performed in 1-butanol-acetic acid-water (4 : 1: 15). The spots on the chromatogram were made visible by ninhydrin spray.

Reduction and Carboxymethylation-Protein at about 10 mg per ml was incubated for 4 hours at 60” in 6 M guanidinium chlo- ride, 0.01 M dithiothreitol, and 0.4 M Tris-HCl, pH 8.2. The re- duction mixture was cooled to room temperature, and iodoacetic acid was added as a freshly prepared aqueous solution (40 mg per ml) in just the amount needed to alkylate all thiol groups present, including those in the dithiothreitol. Alkylation was allowed to proceed for 1 hour in the dark, and then a few crystals of dithio- threitol were added to destroy any excess iodoacetate. The car-

boxymethylated protein was then dialyzed against deionized water.

NHz-terminal Analysis-The enzyme (5 to 10 nmoles) was allowed to react at room temperature for 30 min with a large ex- cess of dansyl chloride (5 mg per mg of protein) in 0.02 M sodium phosphate, pH 8.2, which was 4 M in urea, 25yo (by volume) dimethylformamide, and 10% (by volume) acetonitrile. The dansylatedi protein was isolated by precipitation with 10% tri- chloroacetic acid, washed with acetone, dried in vucuo, and hy- drolyzed for 4 hours at 110” in 6 N HCl. The dried hydrolysate was extracted with a few drops of water-saturated ethyl acetate and the solution was spotted in the corner of a polyamide thin layer (15). Chromatograms were developed with 1.5y0 aqueous formic acid in the first dimension, followed after drying by ben- zene-acetic acid (9: 1) in the second. The identities of the spots observed were confirmed by spotting authentic standards on the developed chromatograms, next to the unknowns, and develop- ing further with the appropriate solvent.

Centrifugation-Ultracentrifugation was done with a Reckman model E analytical ultracentrifuge equipped with interference optics and a specially designed speed control. All experiments were at 20”. High speed equilibrium data (16) were measured from the usual photographs of Rayleigh interference fringes, and were processed using the computer program of Roark and Yphantis (17).

Molecular weights of the enzyme in dodecyl sulfate were cor- rected for the binding of the detergent. The extent of binding of dodecyl sulfate to the enzyme was estimated by equilibrium dialysis using sodium dodccyl[35S]sulfate at a free sodium dodecyl sulfate concentration of 0.1% in 0.1% 2-mercaptoethanol and 0.02 M Tris-HCl, pH 8.0.

Polyacrylumide Gel Electrophoresis-Disc electrophoresis was used to obtain the molecular weight of the enzyme dissociated in 0.1% sodium dodecyl sulfate. The procedure described by Weber and Osborn (18) was followed with minor modifications. Horse heart cytochrome c or a tracking dye was used as a visible marker, and electrophoresis was performed in 57, and 10% acrylamide gels. Gels were stained with Coomassie blue.

To determine the subunit structure of I’-enolpyruvate carbos- ylase, the enzyme was cross-linked with dimethyl suberimidate as described by Davies and Stark (19) and subsequently analyzed on dodecyl sulfate gels.

Chromatography on Sephudex C-200-To obtain an independent estimate of the molecular weight of P-enolpyruvate carboxylase, gel filtration according to Andrews (20) was performed. A col- umn, 2.5 x 100 cm, was packed with Sephades G-200 and equili- brated with 0.05 M phosphate buffer, pH 7.5. The column was calibrated according to Andrews (20) for molecular weight esti- mation. The void volume was measured by blue destran, and the final elution volume by dinitrophenylalanine. The protein standards and I’-enolpyruvate carboxylase (2 mg each) were chromatographed at 4”.

Binding Sfudies-The binding of radioactive ligands to Penol- pyruvate carboxylase was studied by equilibrium dialysis. Dialysis cells holding 0.1 ml of fluid in each compartment were designed according to Myer and Schellman (21). The Visking tubing used to separate the compartments was boiled before use in 1 mM EDTA and distilled water. To achieve rapid equilib- rium the contents of the compartments were agitated by rotation at 4”. Trial runs were made before experiments at the highest

1 The abbreviation used is : dansyl, 5-dimethylaminonaph- thalene-l-sulfonyl.

by guest on October 12, 2020

http://ww

w.jbc.org/

Dow

nloaded from

184

concentration of ligand used to ascertain the time at which equilibration between the two compartments was achieved. Sampling from each compartment was done by Hamilton Sy- ringes, Hamilton Co., Whittier, California. Generally, 0.05 ml of a sample was removed from each side of the compartment and counted in a dioxane-based scintillation fluid. Quenching correc- tions, whenever needed, were made by reference to an external standard. It was ascertained that no loss of radioactive ligand or enzyme due to its possible binding or precipitation on the dialysis membrane occurred. In all the experiments reported here the enzyme remained at least 97% active at the end of the dialysis procedure (18 to 30 hours).

Enzyme PuriJication-P-enolpyruvate carboxylase was pre- pared by minor modifications of a method described earlier (10). The procedure described before (10) was scaled up to process about 1500 g (wet weight) of cells at one time. The enzyme was judged to be in a high state of purity when only one band was ob- tained by electrophoresis on dodecyl sulfate-polyacrylamide gels and only one amino acid (methionine) was obtained as the amino- terminal residue. This apparently homogeneous, freshly pre- pared enzyme was used in most of the experiments described here.

RESULTS

Chemical Properties of Enzyme

Peptide Maps-An analysis of the amount of lysine and argi- nine in P-enolpyruvate carboxylase revealed that about 52 residues are present per 50,000 daltons. This agrees well with data presented before (10). A total of about 42 peptides were located on the chromatograms. This result suggests that the subunits of the enzyme may all be equal.

NHz-terminal Residue-Chromatography of the cu-dansylated amino acids obtained by acid hydrolysis of P-enolpyruvate car- boxylase which had been treated with a large excess of dansyl chloride revealed methionine as the only dansylated residue. It would seem that, in confirmation of the conclusions drawn from peptide mapping, the subunits of the enzyme are all similar. However, since no attempts were made to find the stoichiometry between molecular size and the number of NHz-terminal methio- nine groups, the possibility cannot be excluded that some NH*- substituted amino acids may be present at the NH2 terminus of the P-enolpyruvate carboxylase peptide. Attempts made to determine the COOH-terminal group of the enzyme by using carboxypeptidases A and I%, either singly or together, all failed.

Subunit Composition and Physical Properties

Sedimentation Rates-Different concentrations of a freshly purified preparation of P-enolpyruvate carboxylase were used to determine the ~20,~ value at zero protein concentration. These experiments were performed in order to find whether the enzyme shows aggregation phenomena at high protein concentrations such as those used in equilibrium binding studies (described later). It has been reported, by Maeba and Sanwal (10) and by others (9) that aged preparations of P-enolpyruvate carboxylase do show aggregation. Under the conditions used in our experi- ments, preparations stored for more than a month at 4” in 16% ammonium sulfate turn gradually turbid, owing to the formation of insoluble aggregates. In fresh preparations of the enzyme, however, the s20,w value (12 S) is independent of protein concen- tration in the range of 1.5 to 10 mg per ml. The average ~20,~ value of about 12 remains unaltered when centrifugations are per- formed in the presence of the various effecters of the enzyme (Table I). These results mitigate against the possibility that

TABLE I Ej’ect of various efectors on sedimentation of

P-enolpyruvate carboxylase Freshly prepared enzyme (10 mg per ml) was preincubated with

the various effecters for 30 min in 0.05 M Tris-HCl, pH 8.0. Runs were performed at 44,000 rpm.

Additions Sl0.w

Acetyl-CoA (1 mM). ............................ 11.8 Acetyl-CoA (1 mM) plus GTP (10 mM). ......... 11.9 Acetyl-CoA (1 mM) plus fructose-1,6-P* (1 mM). 11.8 Fructose-1,6-Pz (20 mM). ....................... 11.8 L-Aspartate (5 mM) ........................... 11.7 (NH&S04 (10 mM). ............................ 12.1

m d

MIWI ,MlNl ,MIZl .M IiI+ll

‘: x 13 X A I

cmc. [MM.)

FIG. 1. Calcomp plots of sedimentation equilibrium results as comnuted bv the method of Roark and Ynhantis (17). Freshlv prepared enzyme was dialyzed against 0.i M Tris-HCl, pH 8.0, containing 0.1 M NaCl and 1% 2-mercaptoethanol. Concentra- tions of the protein (milligrams per ml, MM in the figure) were as indicated. The samples were centrifuged at 9173 rpm at 20".

equilibrium binding curves (presented later) obtained in the pres- ence of various effecters are due to their gross physical effects on the enzyme itself.

Sedimentation Equilibrium Studies-In buffers of moderate ionic strength containing 1.0% 2-mercaptoethanol, freshly pre- pared native P-enolpyruvate carboxylase yields a molecular weight of 400,000 + 30,000 (Fig. 1). In the absence of 2-mer- captoethanol the native enzyme seems to dissociate to smaller molecular weight species. In ultracentrifugation studies, mo- lecular weight species of about 100,000, 200,000, and 400,000 are discernible.

Carboxymethylated P-enolpyruvate carboxylase centrifuged in 6.3 M guanidinium chloride-0.1% 2-mercaptoethanol shows a concentration-dependent distribution of molecular weight species starting from about 100,000. Extrapolation to zero concentra- tion gives values of weight average molecular weights between 47,000 and 53,000 (Fig. 2). These values may be considered an upper limit, because guanidinium ion is known to bind to some proteins and thus change the partial specific volumes. For the calculation of the data in Fig. 2 we have assumed a partial spe- cific volume of 0.73.

We found by equilibrium dialysis experiments that P-enolpyru- vate carboxylase binds about 0.43 g of dodecyl sulfate anion per g of protein at a free detergent concentration of 0.1% in 0.1% 2-mercaptoethanol. Equilibrium centrifugation of preparations of P-enolpyruvate carboxylase which had been equilibrated in 0.1 y0 sodium dodecyl sulfate and 0.1 To 2-mercaptoethanol gave results comparable to the data shown in Fig. 2. Minimum mo- lecular weights (corrected for the amount of the detergent bound) in several runs varied between 46,000 and 56,000, and the largest weight average molecular weight species was about 100,000. The heterogeneity of subunit classes obtained under various de-

by guest on October 12, 2020

http://ww

w.jbc.org/

Dow

nloaded from

185

naturing conditions was analyzed further by the use of two species plots suggested by Roark and Yphantis (17)) in which the reciprocal of a given molecular weight average (n,, li;i,, etc.) is plotted as a function of the reciprocal of the next lower order average (a,, a,, etc., respectively). Associating systems which contain only two species yield straight lines in such plots, the slopes and intercepts of which are functions of the molecular weights of the two species (17). Such a two-species plot is shown in Fig. 2B for data obtained by centrifugation of carboxymethyl- ated P-enolpyruvate carboxylase in 6.3 M guanidinium chloride. The plot intersects with the hyperbola for n, = n, at two points corresponding to molecular weights of about 46,000 and 103,000, which suggests that the carboxymethylated protein is a two-component interacting system. Noncarboxymethylated enzyme samples run in 7.5 M guanidinium chloride always yielded

I A

MIW1.N (N1 ,tllZl .M (Z+ll

m El X A

x

I’. 00 0’. 20 0: I(0 0’. 60 0’. so I’. 00 CONC. (NM.)

FIG. 2. Calcomp plots of sedimentation equilibrium results as computed by the method of Roark and Yphantis (17). The native protein was carboxymethylated and dissociated in 6.3 M guanidine HCl. A, plots of l/M= versus concentration; B, a two-species plot according to Roark and Yphantis (17) for an ideal interacting two-species system. The hyperbola is calculated for M(N) = M(W) = M(Z). The protein samples were centrifuged at 29,851 rpm at 20”.

two species plots which were anomalous (concave down at lower reciprocal values of zn or a,), possibly because of charge nonideality (data not shown).

Sephadex G-fN0 Chromatography-Using a column calibrated according to Andrews (20) the molecular weight of native, freshly prepared P-enolpyruvate carboxylase was estimated to be about 400,000. The Stokes radius, calculated according to Ackers (22), was 7.3. Results of this experiment are summarized in Fig. 3. A smaller molecular weight species evident in ultracentrifugation experiments is not discernible here.

Polyacrylamide Gel Electrophoresis-Electrophoresis was per- formed in dodecyl sulfate-containing gels at 5% and 10% poly- acrylamide (Fig. 4) concentrations. The enzyme exhibited a major band with a molecular weight of 100,000 and a very faint minor band with a molecular weight of about 50,000. This sug- gests, in conformity with the results obtained in sedimentation equilibrium experiments, that the native enzyme breaks down only partially in the detergent. It may be mentioned that car- boxymethylated P-enolpyruvate kinase behaved in the same way in dodecyl sulfate-containing gels as the unmodified enzyme. The partial dissociation of P-enolpyruvate carboxylase in dodecyl sulfate is unusual, but it is not a solitary example of this kind. Several brain glycoproteins (23), rabbit muscle phospho-

- Log( Mol w t)

FIG. 3. Gel filtration of P-enolpyruvate carboxylase on Sepha- dex G-200. The arrow indicates the P-enolpyruvate carboxylase peak. Details are described in the text. The numbers over the poinis denote: I, cytochrome C; 2, ribonuclease; 9, ovalbumin; 4, bovine serum albumin; 5, alcohol dehydrogenase; 6. catalase; and 7, rabbit muscle pyruvate kinase.

‘3 RELATIVE MOBILITY

FIG. 4. Semilog plot of the monomer molecular weights (ob- tained by the dodecyl sulfate-polyacrylamide gel electrophoresis procedure) against distance of migration relative to the marker dye. The distances migrated by P-enolpyruvate are marked by arrows. The concentration of polyacrylamide in the upper line was 597., and that in the lower line, 10%. The numbers over the points denote: 1, phosphorylase; 2, bovine serum albumin; S, catalase; 4, ovalbumin; 6, aldolase; 6, glyceraldehyde 3-phosphate dehydrogenase; 7, trypsin; and 8, lysozyme.

by guest on October 12, 2020

http://ww

w.jbc.org/

Dow

nloaded from

186

glucomutase (24), and yeast phosphofructokinase (25), for in- stance, are only partially dissociated by dodecyl sulfate. This may be related to the varying amounts of the detergent bound by these proteins (26).

Results described so far suggest that P-enolpyruvate carboxyl- ase has a basic monomer with molecular weight around 50,000. Judging from the results of the NHz-terminal analysis the mono- mers seem to be identical. To confirm these conclusions, the cross-linking procedure of Davies and Stark (19) was utilized to determine the subunit structure. In 5% dodecyl sulfate-con- taining polyacrylamide gels, samples of P-enolpyruvate carboxyl- ase previously cross-linked with dimethylsuberimidate showed, after dissociation and electrophoresis, four major bands (Fig. 5). One concentration of the protein (0.33 mg per ml) and several concentrations of dimethylsuberimidate (0.025,0.25, and 1.25 mg per ml) were tried with essentially the same results as those pre- sented in Fig. 5. The approximate molecular weight of the fastest moving major band was 100,000 as judged by its mobility relative to standard protein markers (refer to Fig. 4). The molecular weights of the other three major bands could not be cal- culated from the standards used because of their high molecular weights. In general, the results obtained here are consistent with results obtained with unlinked P-enolpyruvate carboxylase. It does seem that it is hard to dissociate the dimer of 100,000 mo- lecular weight in dodecyl sulfate or other denaturing solvents.

Kinetic Studies

Earlier studies in our laboratory (5, 10) had shown that some substrates and activators of I’-enolpyruvate carboxylase have extremely low affinities for the enzyme. This fact precluded the use of these compounds in equilibrium binding studies. In order to clarify some aspects of the binding of P-enolpyruvate and Mn2+, initial velocity studies were undertaken with these compounds as variable substrates. Corwin and Fanning (27), using a radioisotopic assay for P-enolpyruvate, presented some initial velocity results which were at variance with those obtained by other workers (10) with the coupled spectrophotometric method. Since kinetic experiments have been used extensively in support of one or the other of several allosteric models proposed by diverse authors, initial velocity patterns using I’-enolpyruvate as a variable substrate were re-esamined using the isotopic assay. The results of this study are presented in Fig. 6 in the form of a Hill plot. As may be noted there are no breaks or plateaus as described by Corwin and Fanning (27). The data of Fig. 6 were obtained with a freshly prepared enzyme preparation. It has been suggested by Kirschner (28) that, under certain conditions, aged enzyme preparations may produce Hill plots suggesting negative cooperativity (29). As with the spectrophotometric assay (IO), the data in Fig. 6 yield a K, of about 12.0 mM for P-enolpyruvate, and a Hill coefficient of 1.6.

It has been demonstrated by Smith (8) that Mn2+ can replace Mg2+ in the reaction catalyzed by P-enolpyruvate in E. coli. This is also the case with the enzyme obtained in this work from S. typhimurium. When Mn2+ was used as the variable substrate, a sigmoid saturation cure was obtained (Fig. 7), suggesting that the binding of manganese to the enzyme was a cooperative process. The K, value for Mn2+ evaluated from Fig. 7 is 0.37 mM. This value may be compared with 1.0 mM, which is the K, for Mg2+ under identical assay conditions.

Binding oj Ligands at Equilibrium

It has become increasingly clear in recent years that initial velocity studies by themselves are not suitable to an understand- ing of the allosteric phenomenon. It is, for instance, well known that sigmoid rate-concentration plots can be generated if alter- nate pathways for product release are available in the reaction pathway. In order to formulate a mechanism for allosteric in-

FIG. 5. The location of the subunits of P-enolpyruvate car- boxylase in sodium dodecyl sulfate-polyacrylamide (7.5%) gels after staining with Coomassie blue. The pattern on the left,

-2.4 I 1 I 1 0

I

showing four bands, was obtained after cross-linking of the protein 0.5 1.0 1.5

with sodium suberimidate and subsequent dissociation in sodium log P-enolpyruvatd mM) dodecyl sulfate. The procedure is described in the text. The gel on the right shows the location of untreated enzyme dissociated

FIG. 6. A Hill plot of the initial velocity data with P-enolpyru- vate as the variable substrate.

in the detergent. in the text was used. The radioisotope assay described

ii refers to the slope of the line.

by guest on October 12, 2020

http://ww

w.jbc.org/

Dow

nloaded from

187

hibition and activation of P-enolpyruvate carboxylase, therefore, ligand binding to the enzyme at equilibrium was studied. Owing to technical problems it has not been possible to study this facet of the problem in any great detail. The binding of the substrate, P-enolpyruvate, and certain effecters such as fructose-l ,6-Pz and GTP could not be studied because of their low affinities. A con- sistent difficulty also has been the relatively low yields of the enzyme. For instance, on the average, only about 50 mg of the pure enzyme could be obtained from about 1500 g, wet weight, of the cells.

Binding of L-Asp&ate-Earlier kinetic studies (2, 10) had shown that the inhibition curves for n-aspartate were hyperbolic. In view of the kinetic results the equilibrium binding curve for n-aspartate was expected to yield a hyperbola. This expectation is borne out by the Scatchard plot of the binding data, presented in Fig. 8. This plot yields one binding site for aspartate per 100,000 molecular weight subunit of the enzyme. If our surmise regarding the monomer molecular weight is correct, this result suggests that a dimer may be the protomer or the actual func- tional unit of P-enolpyruvate carboxylase. This result is unex- pected but not surprising. In recent years several enzyme sys- tems have been described which have less than one ligand-binding site per monomer (30-34).

The aspartate binding data presented in Fig. 8 remain sub- stantially unaltered when this ligand is bound in the presence of 0.5 mM EDTA, 10 mM Mg2+, or 5 mM Mn2f. This suggests that

I 0 0.4 08 12 16 20

MnCI, , mM.

FIG. 7. The effect of Mn2+ on the initial velocity of the enzyme. The spectrophotometric assay described in the text was used, except that Mgz+ was omitted from the reaction mixtures.

Yl I 1 i I 0 02 04 ,- 06 08 1 .o

FIG. 8. Scatchard plot of the binding of L-aspartate to P-enol- pyruvate carboxylase. The concentration of the enzyme used was 0.2 mM (calculated on the basis of a molecular weight of 100,000). L-[i%]Aspartate was varied between 0.01 and 1.0 mM. The limiting value of r is 0.9, and the dissociation constant is 0.09 rnM.

the metal is probably not involved in the binding of aspartate. However, binding data with a metal-free enzyme would be re- quired to prove this point unequivocally. The dissociation con- stant for aspartate obtained from the data of Fig. 8 can be calcu- lated to be 0.094 mM. This value agrees very well with the Ki value obtained in earlier kinetic experiments (10).

E$ect of Allosteric Activators on Binding of L-Asparatate-To formulate a molecular model for the mode of action of P-enolpyru- vate carboxylase it was essential to know the binding behavior of aspartate in the presence of various allosteric activators. The effects of three activators of the enzyme, acetyl-CoA, fructose- 1,6-P2, and GTP, were tested separately and in various combina- tions. As shown in Fig. 9, the constant presence of a saturating amount (1 .O mM) of unlabeled acetyl-CoA during dialysis had no effect on the binding of radioactive L-aspartate to the enzyme. Neither the dissociation constant nor the number of sites binding aspartate changed in the presence of this activator. In the pres-

ence of unlabeled 10 mM fructose-l ,6-PZ (Fig. 9), again, 1 mole- cule of aspartate was found to bind to 1 molecule of 100,000 molecular weight subunit, and the KD for aspartate remained unchanged. The third activator, GTP, tested at a concentration of 10 mM (Fig. 9), however, produced a small but consistent change in the dissociation constant of aspartate. As shown in the Scatchard plot (Fig. 9) the KD in the presence of GTP is 0.25 mM, to be compared with 0.095 in its absence. The number of binding sites for aspartate is not affected by GTP. It is note- worthy that the binding curve for aspartate in the presence of various effecters does not change shape.

It had been shown earlier (1, 5, 10) that the activators of P-enolpyruvate carboxylase show a cooperative effect on the velocity of the enzyme, i.e. tested singly they cause only slight activation but when tested at the same concentration together the effect on the velocity of the reaction is more than additive. This finding suggested to us that combinations of various acti- vators might be able to affect the binding of L-aspartate on the enzyme surface. Accordingly we tested the binding of radioac- tive aspartate in the presence of 1 mM acetyl-CoA plus 10 mM fructose-l ,6-Pz and in the presence of 1 mM acetyl-CoA plus 10

FIG. 9. Scatchard plots of the binding of L-aspartate to P-enol- pyruvate carboxylase in the presence of various activators of the enzyme. The concentration of the enzyme used throughout was 0.2 mM (calculated on the basis of a molecular weight of 100,000). L-[“C]Aspartate was varied between 0.01 and 1.0 mM. The numbers above the lines denote: 1, binding in the constant pres- ence of unlabeled 10 mM GTP (line drawn with the assumption that T = 1.0 and Rn = 0.25 mM); 2, binding in the constant pres- ence of unlabeled 10 mM fructose-1,6-P* (line drawn with the assumption that T = 0.8 and KD = 0.094 mM); and 3, binding in the constant presence of unlabeled 1 mM acetyl-CoA (line drawn with the assumption that T = 0.95 and KD = 0.105).

by guest on October 12, 2020

http://ww

w.jbc.org/

Dow

nloaded from

188

mM GTP. The results are given in Fig. 10. It is clear that in both of these cases the binding of aspartate is drastically reduced. The more effective combination in this regard was found to be acetyl-CoA plus GTP (Fig. IOA). Barely perceptible binding occurred in this case at around 1.0 mM aspartate. With the second combination of activators (acetyl-CoA plus fructose-l, 6- Pz, Fig. lOB), however, only 207, of the available sites for aspar- tate were occupied at 1 .O mM aspartate. Also noteworthy here is the fact that the binding curve becomes sigmoidal or cooperative in the presence of the activators. Owing to the limited availabil- ity of the enzyme and technical difficulties attendant on the use of high concentrations of aspartate, we were unable to obtain bind- ing points beyond 1.0 mM aspartate. We could not, therefore, decide whether the interaction between the inhibitor and the activators is competitive.

Binding of Acetyl-CoA-As has been indicated before, acetyl- CoA is the most powerful activator of P-enolpyruvate carboxyl- ase. It has already been demonstrated (10) that acetyl-CoA changes the K, of P-enolpyruvate from 12.0 mM to about 1.0 mM. In view of this effect of acetyl-CoA on the kinetics of the enzyme, and also because this compound has a reasonable affinity for the enzyme (10) compared to other known activators, we decided to test its binding behavior under equilibrium conditions. A repre- sentative curve for the binding of acetyl-CoA to P-enolpyruvate carboxylase is shown in Fig. 11. The binding curve for acetyl- CoA in the absence of any other additions (lower curve, Fig. 11A) is quite complex. The initial part of the curve, at low concentra- tions of acetyl-CoA, is positively cooperative, while at higher concentrations the curve becomes negatively cooperative, i.e. it continues slowly sloping upward. Despite several attempts to obtain reliable binding data above 1 .O rnM acetyl-CoA, we did not succeed in getting consistent results. The points gave too much scatter. The behavior of the curve did not change substantially when the binding of acetyl-CoA was studied in the constant pres- ence of 10 mM Mg2+. Kirschner (28), in a recent review, sug- gested that complicated curves of the kind shown in Fig. 11 could arise if “aged” enzyme preparations are used in binding. Con- ceivably, aged enzyme preparations would contain heterogeneous molecules of protein, some of the binding sites of which may be altered or partially inactivated. This explanation is not applica- ble in our case, because the same protein preparation used to

0.16

r

ASPARTATEctJ .(mM) log ASPARTATE,(mM)

generate the acetyl-CoA binding curves yielded hyperbolic plots when binding of acetyl-CoA was studied in the presence of 10 mM fructose-l, B-P, or unlabeled 10 mM GTP (Fig. 11). A Scatchard plot of the binding data (Fig. 1lB) shows very little positive co- operativity at low concentrations of acetyl-CoA. However, it is still difficult to saturate the enzyme with acetyl-CoA. This difficulty is reflected to some degree in the value of n obtained by extrapolation of the data. This value comes to about 0.86 per 100,000 subunit molecular weight. More likely, one site for ace- tyl-CoA per dimer is present on the enzyme surface. In the pres- ence of 10 mM fructose-l ,6-PZ or 10 mM GTP the dissociation constant for acetyl-CoA is about 0.16 mM.

The work described here was largely undertaken to throw some light on one important aspect of the allosteric control of enzymes, viz. the phenomenon of cooperative regulation. This type of regulation was first described by Caskey el al. (11) in the purine biosynthetic pathway of animal tissues. Here, the first enzyme of the pathway, glutamine :phosphoribosyl amidotransferase (EC 2.4.2.14) was found to be inhibited separately by 6-hydrox- ypurine ribonucleotides, GMP and IMP, and 6-aminopurine ribonucleotides, AMP and ADP. In the presence of a mixture of these two groups of nucleotides (GMP and AMP or IMP and ADP), however, the inhibition was more than additive. This kind of control was extended in this laboratory (1, 5, 10) to cover cases where two or more effecters caused a cooperative activation. It has been demonstrated earlier (5, 10) that P-enolpyruvate car- boxylase from S. typhimurium is activated by fructose-1,6-Pz, GTP (and several other nucleotides), and acetyl-CoA. When two activator pairs are present simultaneously they cause cooper- ative activation. It seemed possible that this cooperativity is caused by a reciprocal enhancement of the affinity of one acti- vator for the enzyme in the presence of another. To understand the mode of action of the enzyme, however, we were also inter- ested in finding how the binding of the inhibitor was affected in the presence of activators separately and together.

The equilibrium binding results can be interpreted on the basis of a two-state model, where the enzyme in the absence of any ligands is assumed to be present predominantly in a state which binds the inhibitor (aspartate) preferentially. Further assump- tions of this model would be that the binding sites for all of the

FIG. 10. A, cooperative effect of two allosteric activators on the binding of L-aspartate to P-enolpyruvate carboxylase. The pro-

6- 02 /&%LP&AA.( %, 10 0 02 04 0.6 0.8 r

tein concentration throughout was 0.2 mM (calculated on the basis FIG. 11. A, binding of acetyl-CoA to P-enolpyruvate in the of a molecular weight of 100.000). The numbers above the Zirkes absence and in the Dresence of 10 mM GTP. The Drotein concen- represent: 1, no effecters; R,‘in the presence of unlabeled 1 mM tration throughout &as 0.2 mM (calculated on the dasis of a molec- acetyl-CoA and 10 mM fructose-1,6-P*; 5, in the presence of 1 mM ular weight of 100,000). Acetyl-CoA was varied between 0.01 and acetyl-CoA and 10 mM GTP. Only two points for aspartate bind- 1.0 rnM. B, Scatchard plot of the data for acetyl-CoA binding in ing are shown in Line 1 for comparison purposes. B, a Hill plot the presence of 10 mM GTP (top line in A). The limiting value of of the data from Line 2 of A. 5 refers to the slope of this plot. r is 0.86, and the dissociation constant is 0.16 mM.

by guest on October 12, 2020

http://ww

w.jbc.org/

Dow

nloaded from

189

activators and the inhibitors are separate, and that there is a direct (i.e. mediated by change in the tertiary structure of each individual subunit) reciprocal interaction of the activator sites on each subunit such that binding of one activator facilitates the binding of the other. It would have to be further assumed that the combination of a substrate and an activator on the enzyme surface or the combination of two activators (in the absence of substrate) converts the enzyme into a high affinity (for substrate) state to which the inhibitor is able to bind only with very low affinity. This model is by no means an original one, having been earlier discussed by several workers (35, 36). However, it is un- usual in the proposition that the cooperation of substrate and activator or two activator molecules is required to convert the enzyme to a low affinity state for the inhibitor. The experimen- tal data are consistent with this model. It is expected that in the absence of substrate or activators or in the presence of only one activator the binding curve for aspartate (inhibitor) would be hyperbolic. Our results clearly show that by itself aspartate binds with a dissociation constant of about 0.1 mM and this value either does not change at all or changes only slightly in the pres- ence of each of the activators singly. However, when two acti- vators are present simultaneously, and depending upon the acti- vator pair, not only does the binding curve for aspartate become sigmoidal but the dissociation constant increases drastically. Very crudely estimated, the dissociation constant for aspartate in the presence of 1 .O mM acetyl-CoA and 10 mM fructose-l, 6-PZ is more than 1.5 mM. The heterotropic interaction of aspartate in the presence of activators is reflected in the value of Hill coeffi- cient (n = 1.2) that is obtained in binding.

In keeping with the model above, the binding of acetyl-CoA to the enzyme was found to be sigmoidal at low concentrations. At higher concentrations of the ligand, however, the binding curve shows tailing (7). A very likely cause of this tailing is that acetyl-CoA binds nonspecifically at higher concentrations to some other site on the enzyme surface. Whatever the cause of this tailing may be, it is noteworthy that in the presence of a sec- ond activator, GTP, not only does the acetyl-CoA binding curve become almost hyperbolic, but the dissociation constant for acetyl-CoA is also considerably reduced (0.16 mM). A minor point that emerges from the binding studies is that neither Mg2+ nor Mn2+ is required for the binding of the inhibitor or the acti- vators to the enzyme.

From our results it appears that both aspartate and acetyl-CoA have one binding site per 100,000 g mole-i of the enzyme. The total number of sites on the enzyme oligomer depends, of course, on the molecular weight assigned to the native enzyme. Previ- ous work (10) with P-enolpyruvate carboxylase from S. typhi- murium had shown that in dilute solutions the molecular weight of the enzyme, obtained from sucrose density gradient centrifuga- tion experiments, was about 200,000. Smith (8), with E. coli enzyme, however, demonstrated that the enzyme dissociates into half-molecules on dilution. Our present ultracentrifugation re- sults obtained with freshly prepared enzyme would suggest a molecular weight of 400,000 for the native enzyme. This is sup- ported by the finding that four major bands are obtained on cross- linking and subsequent dissociation of the enzyme on dodecyl sulfate-containing polyacrylamide gels. The fast migrating major band in these studies was found to have a molecular weight of 100,000. The ultracentrifugation experiments in dodecyl sulfate and guanidine hydrochloride tend to show that the small- est subunit molecular weight for S. typhimurium enzyme is 50,000. The reason for the incomplete dissociation of the en- zyme in these denaturing solvents is not known, but this phenom-

enon, as pointed out earlier, is certainly not unique to P-enolpyru- vate carboxylase (23-25, 31). We have recently obtained data to show that, unlike the effecters, 2 moles of Mn2+ are bound per 100,000 g mole-r of the enzyme (32). This observation would also tend to suggest that the smallest polypeptide unit of P-enol- pyruvate carboxylase has a molecular weight of 50,000.

In recent years several examples have come to light where the number of polypeptide chains in an oligomer does not correspond to the number of ligand-binding sites, i.e. one to one proportion- ality does not seem to exist (30, 33, 34). In the case of rabbit muscle phosphofructokinase (33), for instance, the protomer (35) is a subunit of molecular weight 90,000 which binds 1 mole of fructose 6-P, AMP, ADP, and cyclic 3’) 5’-AMP, but 3 moles of the allosteric inhibitor ATP. This may very well be due to half- site reactivity (34). Levitzki et al. (34) have shown, as an example, that the affinity label 6-diazo-5-oxonorleucine reacts with only one-half of the glutamine sites of E. coli CTP synthe- tase, although the subunits of the enzyme all seem to be identical.

Acknowledgments-We thank Dr. Harry W. Duckworth for several suggestions and Dr. David Kells for performing the ultra- centrifuge runs.

1. 2.

3.

4.

5.

6.

7.

8. 9.

10.

11.

12.

13.

14.

15.

16. 17.

18

REFERENCES

SANWAL, B. D. (1970) Bacterial. Rev. 34, 2(r39 MAEBA, P., END S.~NW.~L, B. D. (1965) Biochem. Biophys. Res.

Commun 21, 503-508 &NOVAS, J. L., AND KORNBERG, H. L. (1966) Proc. Roy. SOC.

Ser. B Biol. Sci. 166, 189-205 S.\NW~L, B. D., AND Ma~s.4, P. (1966) Biochem. Biophys. Res.

Commun. 22, 194-199 SANDAL, B. D., AND MAEBA, P. (1966) J. Biol. Chem. 241,

4557-4562 S~NW~L, B. D., MAEB.~, P., AND COOK, R. A. (1966) J. Biol.

Chem. 241, 5177-5182 YOSHINAG~, T., IZUI, K., AND KATSUKI, H. (1970) J. Biochem.

(Tokyo) 68, 747-750 SMITH, T. E. (1971) J. Biol. Chem. 246, 4234-4241 WOHL, 1~. C., A4~~ MAI~KUS, G. (197i) J. Biol. Chem. 247,

5785-5792 MAI~BA, P., AND SANJVAL, B. D. (1969) J. Biol. Chem. 244,

2549-2557 CASKISY, C. T., ASHTON, 1). M., .IND WYNGURDEN, J. B.

(1964) J. Biol. Chem. 239, 257&2579 WRIGHT, J. A., AND SANWA;,, B. D. (1971) J. Biol. Chem. 246,

1689-1699 LOWRY, 0. H., ROSEHROUGH, N. J., FN~R, A. L., AND R.~N-

D:ILL, R. J. (1951) J. Biol. Chem. 193, 265 K~N~REK, L., MARLER, E., Baa~s~.\w, R. A., FELLOWS, R. E.,

BND HILL, It. L. (1964) J. Biol. Chem. 239.4207-4211 WOODS, K. R., AND WANG, K. (1967) Biochim. Biophys. Acla

133, 369-370 YPHI\NTIS, D. A. (1964) Biochemistry 3, 297-317 Ro.\RI<, D. E., AND YPHANTIS, D. A. (1968) A?LIL. N. Y. Acad.

Sci. 164, 245278 WEBER, K., AND OSBORN, M. (1969) J. Biol. Chem. 244, 4406-

4412 -~ 19. DAVIES, Cr. F:., .ZND STARK, G. 11. (1970) Proc. Na/. Acad. Sci.

U. S. A. 66, 651-656 20. ANDRE~S, P. (1964) Biochem. J. 91, 222-233 21. MYER, Y. P., AND SCHELLM.IN, J. A. (1962) Biochim. Biophys.

Acta 66, 361-373 22. ACKERS, G. K. (1964) Biochemislru 3, 723-730 23. KATZMAN, R. L. (1972) Biochim. k’ophys. Acta 266, 269-272 24. DUCK~ORTH, H. W., AND S~NW~L, B. D. (1972) Biochemistry

11, 3182-3187 25. KOPPERSCHLAGER, G., DIEZEL, W., PR~USCHE, M., END HOF-

MAN, E. (1972) Fed. Eur. Biochem. Sot. Le/t. 22, 133-138 26. REYNOLDS, J. A., AND TANFORD, C. (1970) Proc. Nat. Acad.

Sci. U. S. A. 66, 1002-1007 27. CORWIN, L. M., AND FANNING, G. R. (1968) J. Biol. Chem. 243,

3517-3525

by guest on October 12, 2020

http://ww

w.jbc.org/

Dow

nloaded from

190

28. KIRSCHNER, K. (1971) in Current Topics in Cellular Regulation 32. BARBER, B. H., SM‘INDO, IX., S.~N\V~.\L, B. D., AND CARVER, (HORIXKER, B. L., AND ST~DTM.~N, E., eds) pp. 167-206, J. P. (1973) Biophys. Sot. Annu. Meet. Abslr. 13, 314a Academic Press, New York 33. KEMP, R. G., AND KREBS, E. G. (1967) Biochemistry 6,423434

29. LEVITZKI, A., AND KOSHLAND, D. E., JR. (1969) Proc. Nat. 34. LEVITZKI, A., ST.\LLCUP, W. B., AND KOSHL,\ND, D. E. (1971) Acad. Sci. U. S. A. 62, 1121-1128 Biochemistry 10, 3371-3378

30. KUEHN, G. I)., BARNES, L. D., AND ATKINSON, D. E. (1971) 35. MONOD, J., WYM.~N, J., AND CHANGEUX, J. P. (1965) J. Mol. Biochemistry 10, 3945-3951 Biol. 12, 88-118

31. COFFEE, C. J., AARONSON, R. P., AND FRIEDEN, C. (1973) 36. KOSHLAND, D. E., JR., NPMETHY, G., AND FILMER, D. (1966) J. Biol. Gem. 248,1381-1387 Biochemistry 6, 365-386

by guest on October 12, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Ronald Smando, E. Bruce Waygood and B. D. SanwalPhosphoenolpyruvate Carboxylase

Cooperative Interactions in the Binding of Allosteric Effectors to

1974, 249:182-190.J. Biol. Chem. 

  http://www.jbc.org/content/249/1/182Access the most updated version of this article at

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

  http://www.jbc.org/content/249/1/182.full.html#ref-list-1

This article cites 0 references, 0 of which can be accessed free at

by guest on October 12, 2020

http://ww

w.jbc.org/

Dow

nloaded from