Multivalent Feedback Inhibition of Aspartokinase in Bacillus polymyxa KINETIC STUDIES*

Aspartokinase, partially purified from extracts of Bacillus polymyxa, is inhibited by combinations of L-threonine and L-lysine at concentrations below 1 mu and by the amino acids separately at much higher concentrations. At 25”, the presence of the feedback inhibitors leads only to a decrease in the maximal velocity of the aspartokinase reaction. However, at 3’7” the atRnity for r,-aspartate is also markedly reduced by the inhibitors. Sigmoid substrate kinetics are not observed under any conditions. At 37”, but not at 25’, the apparent K, for L-aspartate is highly dependent on enzyme concentration, increasing from 0.4 mu to about 50 mu as the enzyme concentration is decreased from 13.4 to 0.17 units per ml. At 25’, the apparent K, for L-aspartate remains constant at about 1.5 mrvr over the same range of enzyme concentration. The presence of dioxane also leads to an increase in the apparent K, for L-aspartate, but at 25’, higher concentrations of the solvent are required to exert an effect than at 37’. It is suggested that at 37”, but not at 25’, the active form of aspartokinase dissociates into lower molecular weight units which have a markedly lower affinity for L-aspartate than the native enzyme.

In bacteria, the amino acids n-lysine, n-threonine, n-isoleucine, and n-methionine are derived from n-aspartate by a series of reactions, the first of which is the phosphorylation of L-aspartate, catalyzed by the enzyme aspartokinase (ATP : n-aspartate 4phosphotransferase, EC 2.7.2.4). Considerable diversity has been observed in different bacterial species in the feedback * This investigation was supported by Grant GB-3157 from the National Science Foundation.
A preliminary report was presented at the 50th meeting of the Federation of American Societies of Experimental Biology,Atlantic City, April 1966. i Recinient of Career Develoament Award l-K3-GM-9848 from the National Institute of General Medical Sciences, United States Public Health Service. regulation of aspartokinase.
For example, in Escherichiu coli there are three different aspartokinases, one of which is inhibited by n-lysine, another by n-threonine (1, 2). Many other bacteria have only a single aspartokinase. In Rhodopseudomonas spheroides this is inhibited by aspartic semialdehyde (3) and in Rhodospirillum rubrum by n-threonine and specifically reversed by n-isoleucine (4). In Rhodopseudomonas capsulatus (5) and in some Bacillus species (6), aspartokinase is inhibited by the concerted action of n-threonine and n-lysine.
Such multivalent feedback inhibition by two products of a biosynthetic pathway also occurs in several other systems. These include the inhibition of glutamine phosphoribosylpyrophosphate amidotransferase of pigeon liver (7) and of Aerobacter czerogenes (8) by AMP and GMP; the inhibition of one of the aspartokinases of E. coli by n-lysine and n-leucine, n-phenylalanine, n-isoleucine, or n-methionine (9) ; and the inhibition of 0-succinylhomoserine synthetase of E. coli by n-methionine and S-adenosylmethionine (10). Multivalent feedback inhibition is thus a regulatory mechanism of general biological significance, and is of special interest in relation to postulated mechanisms of allosteric regulation of enzyme activity. This paper describes the kinetics of interaction of the inhibitors n-tbreonine and n-lysine and the substrate n-aspartate with the aspartokinase of B. polymyxa and the effects of temperature and dioxane on the kinetic properties of the enzyme.
EXPERIMENTAL PROCEDURE i%faterials-y-32P-ATP was prepared by the method of Glynn and Chappell (11) and was further purified by adsorption and elution from Norit A charcoal.
Amino acids were obtained from the following sources: nn-allothreonine, K and K Laboratories; N'-formyl-n-lysine, n-threonine methyl ester, and n-threonine amide, Cycle; L-threonine, Sigma; all other amino acids were obtained from Calbiochem.
Streptomycin sulfate and ammonium sulfate (enzyme grade) were products of Mann. Dioxane (reagent grade) was refluxed over KOH and redistilled immediately before use.
Enzyme Assay-Aspartokinase activity during enzyme purification was assayed by a procedure essentially like that of Black and Wright (12 I  1 unit of creatine phosphokinase, 10 mM MgC&, 800 mu NHzOH . KCl, pH 8.0, 10 mM L-aspartate, and enzyme fraction. After 30 min at 37", the reaction was terminated by the addition of 0.5 ml of 50% (by weight) trichloracetic acid, adjusted to pH 0.9 with NaOH.
FeCla, 2 M (0.2 ml), was added and the precipitate was removed by filtration.
The absorbance at 520 rnp was read against a control mixture from which L-aspartate had been omitted.
A unit of aspartokinase was defined as the amount of enzyme which catalyzed the formation of 1 pmole of aspartyl hydroxamate ( specific and is subject to interference by compounds such as guanine nucleotides and aspartic semialdehyde. The kinetic studies described in this paper were conducted with enzyme preparations devoid of ATPase and inorganic pyrophosphatase activity and could therefore utilize (except for ATP kinetics) an assay procedure that measured the L-aspartate-dependent conversion of T-~~P-ATP to 32P-acyl phosphate (6). Since the equilibrium of the aspartokinase reaction does not favor the formation of aspartyl phosphate, hydroxylamine was included in the incubation mixtures to permit more extensive reaction. The following components were incubated at 37" for 30 min (unless otherwise indicated) in a final volume of 1.0 ml: 1 mM Y-~~P-ATP, 1 mM MgC12, 50 mM triethanolamine.HCl, pH 8.0, 400 mM NH2OH+KCl, pH 8.0, 10 mM P-mercaptoethanol, 1 mM potassium phosphate, L-aspartate (omitted from controls), and appropriate amounts of enzyme. Incubation was terminated by the addition of 2% ammonium molybdate in 2 N H2S04. After 5 min at room temperature, 5 ml of isobutanolbenzene (1 :I) were added and the tubes were stoppered and shaken. A sample of the organic phase, containing the 3'Pi formed from aspartyl phosphate, was transferred to a planchet, dried, and counted in a Nuclear-Chicago gas flow counter. The results were corrected for a small amount of 32Pi (about 30 mpmoles), either present initially in the ATP preparation or liberated from ATP nonenzymatically during the assay, by the use of control mixtures.
Aspartyl phosphate formation by this procedure agreed well with estimates based on the formation of aspartyl hydroxamate.
Growth of Organism-Bacillus polymyxa, strain Pfizer 2459, was grown at 37" with aeration as described previously (15) in the synthetic medium of Katznelson and Lochhead (16) supplemented with 4.2y0 ammonium sulfate, 1 pg per liter of biotin, and 0.1 mM L-methionine' The cells were harvested at the end of the exponential phase of growth, washed with 0.02 M potassium phosphate, pH 7.5, containing 0.03 M fimercaptoethanol, and stored at -15". Enzyme PuriJication-About 50 g of thawed cells were suspended in 3 volumes of 0.02 M potassium phosphate, pH 7.5, containing 0.03 M P-mercaptoethanol, and were sonically disrupted in a Branson sonifier (model S-75) for 10 min. All procedures were carried out at 0 to 4". After centrifugation at 100,000 x g for 1 hour, the supernatant solution (crude extract) was treated with 2.4 mg of streptomycin sulfate per mg of protein (added as an about 25% aqueous solution) and The fraction which precipitated between 25 and 30 g of ammonium sulfate per 100 ml of Fraction II was redissolved in 5 to 10 ml of 0.03 M potassium phosphate, pH 7.5, containing 1 mM P-mercaptoethanol, and passed over a column (2.5 x 40 cm) of Sephadex G-200 in the same buffer. The fractions with significant amounts of aspartokinase, which emerged after about 0.6 column volume, were combined, diluted with 0.67 volume of glycerol, and stored at -15".
The combined Sephadex fractions derived from about 500 g of frozen cells were pooled, diluted with an equal volume of 1 mM ,&mercaptoethanol, and passed over a column (1.6 x 40 cm) of DEAE-Sephadex A-50 (Cl-), equilibrated with 10 mM potassium phosphate, pH 7.5, and 1 IRM /3-mercaptoethanol in 20% glycerol.
The column was developed with 2 liters of a linear gradient of 0 to 0.4 M KC1 in this buffer. The fractions which contained aspartokinase activity were combined, diluted with 2 volumes of the above phosphate-mercaptoethanolglycerol buffer, and again passed through a column (1.6 x 40 cm) of DEAE-Sephadex A-50 (Cl-) equilibrated with the same buffer containing 0.1 M KCl. The column was washed with 250 ml of the latter and developed with 2 liters of a linear gradient of 0.1 to 0.4 M KC1 in the phosphate-mercaptoethanol-glycerol buffer. The fractions of highest specific activity, which contained about one-half of the recovered enzyme activity, were combined, diluted with glycerol to a final concentration of 40%, and stored at -15".
Under these conditions, no loss of enzyme activity was observed over a period of more than a year. The steps involved in the purification procedure are summarized in Table I ions. At 37", the requirement for monovalent cations was nearly absolute, and half-maximal stimulation of enzyme activity occurred with 30 mM NH&l or 90 mu KC1 (Fig. 1). The degree of stimulation by ammonium ion was greater than that produced by potassium ion. In contrast, at 2.5" the requirement for monovalent cations was less stringent than at the higher temperature, and potassium provided more effective stimulation than ammonium ion (Table II). Divalent cations were also required; at 1 mM ATP, the apparent K, for Mg++ was 1 mM.
Aspartokinase (0.33 unit; 16 units per mg of protein) was assayed under standard conditions (see "Experimental Procedure") except that hydroxylamine was omitted. The reaction velocities are expressed as the amount of aspartyl hydroxamate formed in 30 min.
Substrate XpeciJcity-ATP was three times more effective as a phosphate donor than dATP and could not be replaced by GTP, ITP, CTP, or UTP. The apparent K, of aspartokinase for ATP was 1 mM (Fig. 2). ADP inhibited the reaction competitively with respect to ATP and noncompetitively with respect to n-aspartate. At 1 mM ATP and 1 mM MgC12, 1 mM ADP inhibited 50%.
The aspartate analogues, succinate, maleate, n-glutamate, and nn-2-amino-3phosphonopropionate, had no iniluence on the reaction. The kinetic parameters for the interaction of aspartokinase with n-aspartate will be discussed in a later section.
Inhibition by Amino Acids-Aspartokinase was strongly inhibited by combinations of n-threonine and n-lysine at concentrations below 1 mrvr and also by higher concentrations of n-threonine or n-lysine alone. The specificity of this inhibition was tested by the use of structural analogues of n-lysine and Lthreonine. N'-formyl-n-lysine, n&sine, n-ornithine, and Lcr, y-diaminobutyrate (2 mM each) were not inhibitory in the presence of n-threonine (2 rnhf). Similarly, 4 mM nn-allothreonine, 2 mM nn-homoserine, 2 mM n-threonine, and 2 mu n-serine had no effect in the presence of 2 mM n-lysine. On the other hand, 2 mM n-threonine amide, 2 mM n-threonine methyl ester, and ar-amino-/?-hydroxyvalerate (8 mM mixed isomers) inhibited aspartokinase in the presence of 2 mM L-lysine to an extent corresponding to 65, 80, and lOO%, respectively, of the inhibition produced by 2 mM L-threonine. Combinations of naturally occurring amino acids other than L-threonine and L-lysine had no inhibitory effect on aspartokinase.
Thus, the other amino acids derived from L-aspartate, namely L-methionine, L-isoleucine, L-a , y-diaminobutyrate, and L-and mesodiaminopimelate, did not inhibit aspartokinase when tested either alone (2 mM each) or in combination with 2 mM L-threonine or 2 mM L-lysine. Moreover, a mixture of 17 amino acids (all amino acids occurring naturally in protein except L-threonine, L-lysine, and L-aspartate, at a concentration of 2 mM each) was not inhibitory either with or without 2 mM L-threonine or L-lysine.
No combination of amino acids led to a reduction of the inhibition produced by L-threonine and Llysine.
A slight and somewhat variable stimulation of aspartokinase by L-methionine alone was observed (20 to 30% stimulation at 2 MM).
The sensitivity of aspartokinase to feedback inhibition was not altered during the course of purification and storage of the enzyme.
In fact, it was not possible by a large variety of treatments to dissociate enaymic activity and susceptibility to inhibition.
When aspartokinase was heated, treated with p-hydroxymercuribenzoate, iodoacetate, 3-methylcncoxindole, acetic anhydride, pronase, or carboxypeptidase, or assayed in the presence of urea, organic solvents, or extremes of pH, under conditions in which only between 10 to 30% of the enzymic activity remained, the residual activity was as sensitive to inhibition by L-threonine and L-lysine as the untreated enzyme. Time Course of Reaction-The rate of the aspartokinase reaction was not constant (Fig. 3), even when measured over short initial time intervals, for reasons which are not clear. Nevertheless, the reaction rates at various times were in a constant ratio when the enzyme was assayed under different conditions.
For example, the extent of inhibition of aspartokinase by L-threonine plus L-lysine was constant over a period of 2 hours (Fig. 3). Thus it is not likely that the standard assay procedure, which measured the amount of product formed in 30 min rather than initial reaction rates, led to appreciable errors.
Inhibition Kinetics-The effect of L-threonine and L-lysine on aspartokinase activity is shown in Fig. 4. For convenience, the inhibitor concentration has been plotted on a logarithmic scale. Inhibition by the amino acids separately occurred only at much higher concentrations than those required for inhibition by combinations of L-threonine and L-lysine. When the same data were plotted on a linear scale (not shown), a high degree of sigmoidicity was noted when L-threonine and L-lysine were varied simultaneously.
No (or only slight) sigmoidicity was noted when only one inhibitor was varied, either alone or in the presence of a constant concentration of the coinhibitor. Under standard conditions at 37" (Fig. 5A), L-threonine and L-lysine inhibited in a mixed competitive-noncompetitive manner with respect to L-aspartate, since in the presence of the inhibitors the apparent K, for L-aspartate increased from 1.5 mM to 20 mM, while the V,,, decreased from 2.7 to 1.5 pmoles per mg of protein.
The presence of the inhibitors had no effect on the shape of the substrate saturation curves, which were frequently characterized by a slight downward curvature when plotted in the double reciprocal form or by upward concavity when V/S was plotted against V (see insert, Fig. 58).
At 25", the effect of the feedback inhibitors was noncompetitive, and the Procedure") at 10 mM L-aspartate. The curves, from left to righl, represent experiment.s with varying L-lysine at 2 mM L-threonine, varying L-threonine at 2 mM Llysine, varying L-threonine and L-lysine at equimolar concentrations, varying L-lysine alone, and varying L-threonine alone. apparent K, for L-aspartate (1 mM) was unaffected while the V max decreased from 4.2 to 0.74 pmoles per mg of protein.
E$ects of Temperature and Dilution on Aspartate Kinetics-As shown in Table III, when the enzyme was diluted and assayed at 37" at low aspartate concentrations, a much more than proportional decrease in activity was observed, whereas at high aspartate concentrations or at 25", enzyme activity was nearly proportional to enzyme concentration. This had the striking consequence that at low, but not at high substrate concentration, the activities measured at 25" were greater than the values obtained at 37". These effects were not changed by the addition of serum albumen or other proteins and suggested that at 37", but not at the lower temperature, the affinity of aspartokinase for L-aspartate might be a function of enzyme concentration.
The experiments shown in Fig. 6 confirmed this prediction: as the aspartokinase concentration was increased from 30 pg to 60 pg per ml, the apparent K, for Laspartate, determined at 37", decreased from 11 mM to 3.2 mM, whereas at 25' it remained constant at 1 mM. In Table  IV, the kinetic parameters of aspartokinase as determined under a variety of conditions are summarized. The values obtained in duplicate determinations indicate that the estimation of apparent K, and V,,, was subject to some error; but, nevertheless, it is clear that when the enzyme concentration was varied over an 80-fold range at 37", the apparent K, for L- The enzyme was assayed as described under "Experimental Procedure" in absence (0) and presence (0) of 1 mM each Lthreonine and n-lysine.
The incubations were carried out with 150 pg of Fraction VI at 37" (Fig. 5A) or with 60 pg of Fraction VI at 25" (Fig. 5B). The reaction velocities are expressed as the amount of product formed in 30 min. aspartate changed about lOO-fold, while at 25", the K, for Laspartate remained virtually the same. The V,,, of the reaction also seemed to be dependent on enzyme concentration, but to a smaller extent than substrate affinity. Moreover, at 37" the presence of n-threonine and n-lysine invariably led to an increase in the apparent K, for L-aspartate as well as to a reduction in the V,,,, whereas at 25" the feedback inhibitors led only to a reduction of the V,,,. E$ect of Dioxane on Aspartate Kinetics-Like dilution, the addition of dioxane to the incubation medium resulted in a decrease in the apparent affinity of aspartokinase for L-aspartate. At 37", dioxane had a greater effect than at lower temperatures, and the addition of 5% dioxane led to an increase of the apparent K, for n-aspartate from 7 mM to 70 mM (Fig.  7A). At 25", 5% dioxane had little effect on the apparent Km for n-aspartate, which remained unaltered at about 3 mM, but higher levels of dioxane also affected this parameter, which

If multivalent
feedback inhibition is to provide an effective mechanism for the regulation of the aspartate biosynthetic pathway, it is necessary that neither n-threonine nor L-lysine alone at physiological concentrations should inhibit aspartokinase, but that combinations of these amino acids at concentrations sufficient to support protein synthesis should substantially reduce aspartokinase activity. The data given in Fig. 4 show that the aspartokinase of B. polymyxa meets these requirements.
While n-threonine and n-lysine alone produce 50% inhibition of the enzyme only at relatively high concentrations (10 to 20 mM), combinations of n-threonine and Llysine inhibit aspartokinase by 50% at a concentration of 0.4 mM each. It is of interest to note that a third product of the aspartate biosynthetic pathway, n-methionine, does not participate in the feedback regulation of aspartokinase.  (17), combinations of n-threonine and n-lysine inhibit the growth of B. polymyxa, and this inhibition can be overcome by the addition of n-methionineS3 Multivalent feedback inhibition of aspartokinase could be effected by several different mechanisms.
Perhaps the simplest of these is based on the two-state allosteric model of Monod, Wyman,and Changeux (18) and assumes that n-threonine and n-lysine bind, independently, to the catalytically inactive state of the enzyme in preference to the catalytically active form. Alternatively, one might assume, in terms of the induced fit model proposed by Koshland (19,20), that binding of one of the inhibitors to the enzyme produces a conformational change which facilitates the binding of the second inhibitor. Both types of models predict cooperative interactions between the coinhibitors and weaker, if any, cooperativity in the action of each inhibitor alone. These predictions are in accord with the results shown in Fig. 4. From inhibition data of this kind it is difficult to rule in favor of either one of the two models, and in order to distinguish between them it will be necessary to analyze directly the binding of the inhibitors to highly purified preparations of the enzyme. The study of the effect of the feedback inhibitors on the substrate kinetics of aspartokinase revealed an interesting phenomenon.
While at 25" the feedback inhibitors led only to a reduction in the V max of aspartokinase, at 37" the inhibitors also promoted a substantial reduction in the affinity of the enzyme for n-aspartate (Fig. 5). Thus, in the terms of Monod, Wyman,and Changeux (18), at 25' aspartokinase was best described as a V system, but at 37" as a K system of inhibition. However, unlike a K system, aspartokinase did not present sigmoid substrate kinetics either in the absence or the presence of the feedback inhibitors.
This suggested that the effect of the inhibitors on the affinity for n-aspartate at 37" was perhaps not directly related to enzyme inhibition but rather was due to unusual properties of aspartokinase at 37". We had observed earlier that the dependence of aspartokinase activity on the presence of monovalent cations was quite different at 25" and Multivalent Feedback Inhibition.

I
Vol . 242,No. 21 at 37" (Table II), and therefore it appeared likely that other properties of aspartokinase were also highly temperaturedependent.
This was found to be indeed the case. As shown in Fig. 6 and and summarized in Table IV, at 37", but not at 25", the apparent K, of aspartokinase for L-aspartate was highly dependent on enzyme concentration, increasing markedly as the enzyme was diluted.
This recalled a similar, although less striking, observation with glutamic dehydrogenase, in which the affinity for NADH of the alanine dehydrogenase activity was found to increase at high enzyme concentrations (21). In the light of our current understanding of the glutamic dehydrogenase system (22), we may ascribe this effect to association of the enzyme to high molecular weight forms at high protein concentrations and to a higher affinity of NADH for the aggregated form of the enzyme. A similar reaction might occur in the aspartokinase system.4 Thus, in order to account for the behavior of aspartokinase, we might assume that at 37", but not at low temperatures, the active form of aspartokinase is subject to dissociation to a lower molecular weight species which is also enzymically active but has a markedly lower affinity for L-aspartate.
Since the equilibrium position of a dissociating system shifts towards the dissociated state upon dilution, the dilution of aspartokinase at 37" would lead to a relative increase in the low molecular weight form and hence to an increase in the apparent K, for r,-aspartate.
But also, since addition of feedback inhibitors will result in a reduction of the amount of active enzyme, at 37" these would also promote an increase in the proportion of the dissociated form of active aspartokinase, with a concomitant increase in the apparent K, for n-aspartate.
In other words, even though the feedback inhibitors would directly affect only the VI,,,, of aspartokinase, by causing a relative increase in the dissociated form of the enzyme, they could indirectly also affect the affinity for substrate. This hypothesis is supported by the experiments with dioxane. The addition of dioxane to the aspartokinase incubation mixtures results in a marked decrease in the affinity for n-aspartate, both at 37" and, at sufficiently high solvent concentrations, also at 25' (Fig. 7). Thus, the effect of dioxane on aspartokinase resembles that of dilution.
Since dioxane has been shown to promote the dissociation of glutamic dehydrogenase to smaller units (23), it is not unlikely that the effects of both dilution and dioxane on the apparent K, of aspartokinase for L-aspartate are due to a dissociation of the enzyme. In a subsequent