Isolation and Characterization of Homogeneous Acetate Kinase from Salmonella typhimurium and Escherichia coli*

Acetate kinase from Salmonella typhimurium and Escherichia coli was purified to electrophoretic ho-mogeneity. The amino acid compositions of both proteins were similar, and the apparent molecular weights were the same, about 40,000 for the putative mono-mers. The native proteins gave higher molecular weights, suggesting that the enzymes may be oligomers, perhaps with two polypeptide subunits. Steady-state kinetic studies were performed with the enzymes isolated from both organisms and the kinetic constants were determined. The K , values were 0.07 and 7 mM for ATP and acetate, respectively. In contrast to earlier studies using less pure preparations, the homoge- neous enzymes from both strains were active only with acetate but not with propionate or butyrate. The en- zyme activity was cold-labile, and the length of reac-tivation time in the presence of Mg-ATP and acetate was dependent on protein concentration, suggesting that the monomer may not be catalytically active. The enzyme was phosphorylated with [-pa2P]ATP and the phosphoprotein was isolated. Phosphoacetate kinase was capable of transferring the phosphate group to either ADP or acetate. The accompanying paper (Fox, D. K., Meadow, N. D., and Roseman, S. shows that the phosphoryl group of phosphoacetate kinase can also be reversibly transferred to Enzyme I of the phos- phoeno1pyruvate:glycose phosphotransferase

Acetate can be converted to the key metabolic intermediate, acetyl-S-CoA, by coupling acetate kinase with the enzyme phosphotransacetylase (EC 2.3.1.8,phate acetyltransferase). The overall equilibrium constant for the combined reactions is about 1 (11). Although the major pathway for acetate utilization in bacteria appears to be via these coupled reactions, it cannot be the sole pathway. Strains deleted in the structural genes for the two enzymes, ack and pta, respectively, can grow on acetate, albeit slowly (12).
Several additional physiological functions have been proposed for acetate kinase. It may be involved in the secretion of acetate (13,14), in the synthesis of most of the ATP formed catabolically during anaerobic growth (15,16), and acetyl phosphate may be a source of energy required for the uptake of some nutrients by whole cells (17)(18)(19).
Acetate kinase can be phosphorylated by ATP or acetyl phosphate to give the phosphoenzyme, with the phosphoryl group linked to the carboxyl group of a glutamyl residue (20). Whether the phosphoprotein is an obligatory intermediate in the reaction catalyzed by the enzyme is a matter of controversy. The phosphoryl group of the phosphoenzyme is transferred to ADP or acetate (1,21,22). However, steady-state kinetic data (23)(24)(25) suggest that the mechanism for the reaction is random sequential, thus leading to the conclusion that the phosphoenzyme is not in the main reaction pathway, a contention supported by stereochemical studies (26) of the reaction products. Our interest in acetate kinase was stimulated by preliminary observations' suggesting that it interacts in some manner with IIIC'", one of the phosphocarrier proteins of the phospho-eno1pyruvate:glycose phosphotransferase system.' The latter system, designated PTS? has a variety of functions in the bacterial cell (27)(28)(29)(30), one of which is the translocation and concomitant phosphorylation of its sugar substrates. Since the rate-determining step in cell growth is frequently the rate of sugar uptake, it is not surprising to find that the PTS is stringently regulated (30), although the molecular mecha-' W. Kundig and D. K. Fox, unpublished results. The phosphoeno1pyruvate:glycose phosphotransferase system is the major system for the uptake of sugars in obligate anaerobes and is a major system in facultative anaerobes. It is widely distributed among Gram-negative and Gram-positive organisms, and is responsible for catalyzing the translocation of its sugar substrates across the cytoplasmic membranes con.comitant with their phosphorylation (27,28). Its properties and functions have been described extensively elsewhere (29,30).
Purification of Acetate Kirtase nisms underlying this regulation are not known.
Acetate kinase is a potential regulator of the PTS. The preliminary data mentioned above suggested an association between the kinase and a PTS protein. The phosphoryl groups in both the phosphorylated PTS proteins and phosphoacetate kinase are linked via "high energy" bonds (1,20,30). Thus it appeared possible that a reversible phosphotransfer reaction could occur between phosphoacetate kinase and one or more of the PTS proteins. If such transfer reactions did occur, they would link sugar transport via the PTS to the tricarboxylic acid cycle via acetate kinase (31).
The hypothesis that PTS proteins and acetate kinase interact can only be rigorously tested with homogeneous preparations of the respective proteins. Although pure PTS proteins were available (32-34), homogeneous acetate kinase has not been previously isolated from Salmonella typhimurium and E. coli. Indeed, the conflicting interpretations of the data reported in the literature on the mechanism of the enzyme could be the result of working with partially purified preparations of the enzyme.
This paper describes the isolation and properties of homogeneous acetate kinase isolated from both s. typhimurium and E. coli. The accompanying report describes studies on the interactions between acetate kinase and the PTS proteins (31). With the pure enzyme it was possible to determine for the first time whether it could be phosphorylated autocatalytically, or whether it required another factor such as a protein kinase. Finally, the enzyme preparation described here should be useful in determining whether the phosphoprotein is an obligatory intermediate in the reaction catalyzed by acetate kinase. A preliminary report has been presented (35).

RESULTS
Homogeneity-Since many reports have appeared on the purification of acetate kinase (1-3, 21, 24, 64-67), but it had not previously been obtained in homogeneous form from enteric bacteria, several methods were used to assess the purity of the preparations described here.
The purified enzyme (from Step 8, Fig. 1) was subjected to SDS-polyacrylamide gel electrophoresis with different concentrations of both protein and polyacrylamide. Typical results are shown in Fig. 2B. Under all conditions tested, only a single polypeptide band was observed and this band corresponded to M, = 40,000.
In a second set of experiments, the native protein was subjected to electrophoresis at pH 7.5 or at 8.9, and again only one protein band was detected (data not shown). Furthermore, duplicate samples of this gel were eluted (as described under "Experimental Procedures"), and the eluates were assayed for acetate kinase activity. All of the recovered activity (30% of the total loaded on the gel) corresponded to a single stainable protein band (data not shown).
Finally, electrophoresis in polyacrylamide gels of different concentrations was employed (68). In this procedure, the native protein was subjected to electrophoresis at pH 8.9 in three different concentrations of acrylamide. Only a single protein band was observed (Fig. 2 A ) .
The only impurities that would migrate identically to the "Experimental Procedures," Figs. 3-6, and Tables 1-111 are presented in miniprint at the end of this paper. Miniprint is easily read with the aid of a standard magnifying glass. Full size photocopies are available from the Journal of Biological Chemistry, 9650 Rockville Pike, Bethesda, MD 20814. Request Document No. 85M-4221, cite the authors, and include a check or money order for $10.00 per set of photocopies. Full size photocopies are also included in the microfilm edition of the Journal that is available from Waverly Press.

FIG.
1. Purification of acetate kinase. using a bioaffinity column. Affi-Gel blue resin was equilibrated with Buffer D at 4 "C and eluLed with a linear gradient of ATP. Protein was monitored by the Bradford assay (43) and acetate kinase activity was determined with 10 p1 of each fraction in the colorimetric assay. In A, acetate kinase from Step 7 (a second DEAE-Sephadex column) was adsorbed to another Affi-Gel blue column and eluted with a linear gradient of 0-3.0 mM ATP as described in the text. In B, acetate kinase from Step 7 was pooled and concentrated as described in the text and again placed on an Affi-Gel blue column. This column was then eluted with a shallower gradient of ATP (0-1.5 mM ATP). Twelve-milliliter fractions were collected and assayed as described in the text. Fractions 210-250 to 251-320 were pooled and concentrated.

FlucTDI U L R
kinase under all of these conditions were those which had the same size, shape, and net charge at two different pH values (68). The methods would have detected an impurity of 0.3% or more of the total protein. From these results, we concluded that the acetate kinase from both S. typhimurium and E. coli was at least 99.7% pure.
Amino Acid Composition-The amino acid compositions of the final preparations were determined as described under "Experimental Procedures". Slight differences in the amino acid compositions of the proteins from S. typhimurium and E. coli were found (Table 111). The largest difference was in the glycine content, with the E. coli protein containing 11 more glycine residues than that from S. typhimurium. These values were the average of at least four determinations on different hydrolysates and the difference is probably authentic. However, when contamination of a sample occurs during handling (e.g. dialysis), glycine is the most common contaminating amino acid.
The amino acid compositions of the acetate kinases isolated from Veillonella alcalescens (66) and Bacillus stearothermo-phiEus (10) are also given in Table 111 for comparison. The amino acid compositions of acetate kinases from these strains are similar to those from E. coli  measurable tryptophan and no cysteine. This lack of cysteine is characteristic of many thermophilic proteins and may be directly related to their thermostability (68, 69). Determination of the amino acid sequence of acetate kinase from S. typhimurium by Edman degradation was not possible since the NH,-terminal amino acid of acetate kinase was blocked. The nature of the blocking group was not determined.
Absorption Spectra of Acetate Kinase-Acetate kinase from E. coli and S. typhimurium were extensively dialyzed against 0.1% NaHCOs, and the protein concentration was determined by microbiuret assay (42) with the dialysate as a reference buffer. The absorption spectra were obtained with a Perkin-Elmer (Model 557) dual beam spectrophotometer.
At an absorption maximum of 277 nm, the extinction coefficient of acetate kinase was Molecular Weight of the Denatured Acetate Kina~e--'~C-Labeled acetate kinase from E. coli or S. typhimurium was subjected to gel elution chromatography on Sepharose 6B in 6 M guanidine as described under "Experimental Procedures". Acetate kinase was eluted from the column between blue dextran and soybean trypsin inhibitor (M, = 27,000). The molecular weight was determined by using the calculated distribution coefficient (0.146 for S. typhimurium and 0.161 for E. coli) and corresponded to a M , = 41,200 * 1,000 for the protein isolated from 5' . typhimurium and M , = 40,000 * 1,000 for the protein isolated from E. coli. The denatured, reduced, and alkylated acetate kinase was eluted in the linear portion of the calibration curve (Fig. 3A) and the accuracy of the column is ?7% in this region. These molecular weights of 40,000 agree well with those derived from the SDS-polyacrylamide gel electrophoresis.
Molecular Weight of the Native Protein-Gel elution chromatography of acetate kinase isolated from S. typhimurium was performed as described under "Experimental Procedures." The enzyme eluted from the Sephadex G-75 column 'Acetate Kinase 13489 either slightly ahead of or at the position of bovine serum albumin (M, = 68,000). A plot of the logarithm of the molecular weight versus Kay gave a molecular weight for acetate kinase of about 70,000 (Fig. 3B). The acetate kinase proteins isolated from E. coli or S. typhimurium or anaerobically grown S. typhimurium appear to have the same molecular weight (data not shown). Because 70,000 is very near the limit of the linear fractionation range for Sephadex G-75, the molecular weight determination was repeated with Ultrogel AcA 44, which has a broader fractionation range (10,000). The results were the same: acetate kinase in its native form behaves like a molecule that is as large or larger than bovine serum albumin.
Thus, the denatured protein exhibited an apparent molecular weight of 40,000 & 1,000 by both SDS-polyacrylamide gel electrophoresis and by chromatography in 6 M guanidine, whereas the native protein showed an apparent molecular weight of about 70,000. Acetate kinases isolated from other sources also show a difference between the molecular weight of the protein under denaturing and nondenaturing conditions (1,10,65,(70)(71)(72)(73), which suggests an associating system of apparently identical monomers.
Cold Lability of Acetate Kinase-Homogeneous acetate kinase that had been stored a t or below 0 "C exhibited a lag of 1-3 min before the maximal initial velocity was reached. This property of cold lability was found both by Anthony and Spector (1,22) and Webb et al. (58) using partially purified acetate kinase from E. coli. Prior incubation of the enzyme at room temperature for 1-2 h eliminated this lag (data not shown, and see Ref. 1). Another aspect of this behavior, not mentioned by earlier workers, is that the length of this lag decreases as the concentration of acetate kinase is increased. However, even when higher levels of protein were used, a lag period was still observed (about 1 min) unless the protein was first activated by incubation a t room temperature (data not shown). All of these results are consistent with the hypothesis that acetate kinase is an associating system and that the monomer is catalytically inactive. It is of particular interest to note that Enzyme I of the PTS exhibits the same type of behavior (82).
Metal Requirements of the Acetate Kinase Reaction-All known kinases require divalent cations, and acetate kinase is no exception. As reported by Van Campen and Matrone (74), Rose (2), Anthony and Spector (1,22), and others, acetate kinase requires Mg2' and is inhibited by Na' and Li' . For example, when NaCl or LiCl were present in the incubation mixture at final concentrations of 8.8 mM, the formation of acetyl phosphate was inhibited 50 and 62%, respectively. The homogeneous proteins from E. coli and S. typhimurium showed the same requirements and specificities: Mn'+ was as effective as Mg2+ when used at a concentration equal to that of ATP, and CaC12 could not substitute for MgC1,.
One difference between the homogeneous enzyme and the cruder preparations appears to be the ratio of Mg2+:ATP required for optimal activity. Anthony and Spector (22) found the highest activity at a Mg'+:ATP ratio of 2:l. They concluded that the actual substrate of acetate kinase was the Mg. ATP chelate, as with most kinases. The authors suggested that perhaps Mg2+ itself was needed for enzyme activity. When we did the same experiment with partially purified acetate kinase from E. coli K235, we obtained similar results: optimal activity occurred a t a Mg.ATP ratio of 1.5:1. However, when we determined this ratio using the homogeneous enzymes from both E. coli and S. typhimurium, we found the optimum ratio was 1:l. This discrepancy may be because the last step of the enzyme purification in this work used a buffer containing 8 mM MgCI,, and the enzyme was also stored in this buffer. In other words, any Mg2+ needed to activate the enzyme may have been bound to the protein.
Optimum pH-The pH optimum of the acetate kinase reaction was measured in the direction of acetyl phosphate synthesis. The shape of the pH curve and its optimum, pH 7.5 (data not shown), are in good agreement with that obtained by Rose et al. (2,3), who used a partially purified preparation from E. coli. It differs, however, from the results obtained by Anthony and Spector (1,22). They reported a pH optimum of about pH 6.4, measured in the direction of ATP synthesis and a slightly more acidic optimum for acetyl phosphate synthesis, although the shape of the curve was similar. There is no obvious explanation for these differences. They may result from differences in the conditions of the assays, in the pH optima of the reaction in each direction, in the enzymes due to genetic variation or modification during purification or growth, or impurities in the earlier preparation.
Substrate Specificity-Substrate specificities of the homogeneous enzyme were tested with propionate and butyrate, analogues of acetate previously reported to be active as phosphate acceptors (1)(2)(3)75). In contrast to the earlier results (I), neither propionic acid nor butyric acid were active at concentrations similar to or greater than those required for saturation by acetate (both the spectrophotometric and the colorimetric assays). Our results agree with those obtained by other workers using acetate kinase isolated from other bacterial strains (71) but are in contrast to the results reported with E. coli (1,3,75). However, propionic acid is a competitive inhibitor of homogeneous acetate kinase with respect to acetate (data not shown). These differences can be explained by an impurity in the older preparations that changed the substrate specificity of the kinase, or by a contaminating ki-nase@) active with propionate and butyrate.
The nucleotide specificity of homogeneous acetate kinase from S. typhimurium was studied and found to agree well with that obtained for acetate kinases from E. coli and other bacteria (10,64,70,71) and mycoplasma (72). All three of the nucleotides tested were active, although the relative apparent V,,, with ITP (77%) was less than those with ATP (100%) and GTP (125%). Since ATP and GTP are excellent substrates of the enzyme, GTP was used in studies on the interaction between acetate kinase with the PTS (31).
The apparent K, values for the nucleotides, determined at single "saturating" concentration of acetate, were 0.94 mM for Mg.ATP, 1.10 mM for Mg.GTP, and 0.78 mM for Mg.ITP. These are quite similar to those obtained by Anthony and Spector (1,22), although the relative Vmax(app) values differ in that they found that GTP was only 70% as effective as ATP.
Determination of Kinetic Constants-Initial velocities at different concentrations of the substrates were determined as described under "Experimental Procedures." Lineweaver-Burk and Hanes-Woolf (76) plots were used to analyze the initial velocity data. No evidence of nonhyperbolic behavior was seen. The kinetic constants were determined graphically from secondary plots (Lineweaver-Burk or Hanes-Wool0 of the initial velocity data. Furthermore, there was good agreement between the values for the kinetic constants calculated with all of the plotting methods used. The V,,, values for the forward and reverse reaction are 2000 and 2600 pmoI/min/mg, respectively, measured at 30 "C. The kinetic constants for the homogeneous acetate kinase from both E. coli and S. typhimurium were the same within experimental error. The K, values found were for: acetate, 7.0 mM; Mg.ATP, 0.07 mM; acetyl phosphate, 0.16 mM; and Mg-ADP, 0.50 mM. These values agree well with those of Janson and Cleland (23) and differ considerably from the apparent kinetic constants reported by some earlier workers (1-3, 10, 62, 64,65,74,75) and are much closer to the concentrations of the substrates found in bacterial cells (18,73).
Of great interest to us were the patterns observed in the primary plots of the initial velocity data. In our hands, homogeneous acetate kinase showed the same behavior observed by previous workers (23, 24, 75). Initial velocity plots for the reaction either in the direction of ATP or acetyl phosphate synthesis show patterns of lines which are characteristic of a mechanism involving a ternary complex or a mechanism that includes a ternary complex such as that suggested by Purich and co-workers (25) or Janson and Cleland (23), where a covalent intermediate can be formed but may not be in the major pathway of the reaction.
The effects of the products of the forward and reverse reactions on the initial velocities of these reactions were studied to help elucidate the mechanisms of acetate kinase and to determine whether, in some cases, product inhibition plays an important role in the kinetics of this enzyme. Acetate kinase from both E. coli and S. typhimurium was studied, and the type of inhibition and the inhibition constants were found to be the same or similar to those found by Janson and Cleland (23), and support a random sequential mechanism or a mechanism such as that suggested by these workers.
Phosphorylation of Acetate Kinase by [-y-32P]ATP-Both Anthony and Spector (22) and Webb et al. (58) succeeded in phosphorylating partially purified acetate kinase and in isolating the phosphoprotein. In our attempts to phosphorylate acetate kinase, we tested the conditions used by both groups with slight modifications. Under these conditions, homogeneous acetate kinase from both E. coli and S. typhimurium was successfully phosphorylated. An elution profile of the phosphoprotein from the Sephadex G-25 SF column is shown in Fig. 4A. The phosphoprotein was eluted in the void volume of the column, well separated from the unreacted ATP. The column buffer was 20 mM potassium pyrophosphate, pH 8.6, containing 10 mM 8-mercaptoethanol. (The use of pyrophosphate was recommended by Anthony (l), to prevent binding of ATP to the enzyme.) Two experiments were performed to verify that the 32P in the labeled protein was not simply [y-32P]ATP noncovalently bound to the enzyme. After isolation from the gel filtration column, a portion of the labeled protein fraction was applied immediately to thin layer chromatography plates (polyethyleneimine cellulose) and subjected to chromatography using System I described under "Experimental Procedures'' (Fig. 4B). None of the radioactivity was found in the ATP spot; the majority of the counts were associated with the enzyme (at the origin) or with inorganic phosphate. Also, a double-labeled experiment was performed. Acetate kinase was phosphorylated in the presence of [3H]ATP and [T-~~P]ATP of equal concentrations and specific activities (as count/min/nmol). The phosphoenzyme was isolated as described and ? sample of the fraction was subjected to thin layer chromatography on polyethyleneimine cellulose as described above. Only background 3W counts were found associated with the phosphoenzyme (Fig. 4C). Therefore, the association of 32P with the enzyme appears to be covalent and involves neither ATP, ADP, nor AMP.
One of the difficulties in isolating phosphoacetate kinase was variability in the extent of phosphorylation of the enzyme. A summary of a number of phoshorylation experiments is shown in Table IV  Acetate kinase from E. coli or S. typhimurium was incubated in the phosphorylation mixture without ATP for 1-2 h at room temperature before the reaction was started by addition of [-p3'P]ATP (specific activity, 500,000 cpm/nmol). The incubation mixture contained either 40 mM triethanolamine C1 buffer, pH 7.4, or 41.7 mM triethanolamine C1 buffer, pH 7.2, 3.3 mM dithioerythritol, and MgC12, [-y-32P]ATP, and acetate kinase in the concentrations indicated in the table. The enzyme was incubated in the complete reaction mixture for 10 min at 22 "C and transferred to a Sephadex G-25 SF column, and the phosphorylated enzyme was eluted as described in the text. The amount of 32P incorporated into acetate kinase is expressed as percentage of phosphorylation: nmol of 32P incorporated (per nmol of protein) X 100. The molecular weight of the protein used in the calculation is that of the monomer (40,000).   Table IV as well as those of Webb et al. (58) shows no definite pattern that can account for this variability. Some conditions are obviously unsatisfactory, such as performing the chromatography on the gel filtration column at room temperature (Experiments 2 and 3, Table IV) and use of pH 7.0 buffer. Increasing the molar ratio of ATP to acetate kinase or increasing the final concentration of ATP does not always result in greater incorporation of 32P (Experiments 5, 7, and 8, Table IV). However, experiments to determine the rate of phosphorylation in which a sample of the reaction mixture was treated with EDTA to stop the reaction and immediately applied to a thin layer chromatography plate and chromatographed (System 11) showed that as much as 75% of the enzyme could be phosphorylated at 5 s; therefore, the variability of the extent of phosphorylation is certainly dependent in large part on the instability of the phosphoenzyme and perhaps by slight contaminants in the media (such as trace metals) which vary from one experiment to another and which can affect the rate of spontaneous hydrolysis of the phosphoprotein.
Chemical Competence of Phosphorylated Enzyme-In order to demonstrate that the phosphorylation occurred at the active site, experiments were performed to determine whether the phosphoryl group could be transferred from the phosphoenzyme to ADP and acetate. When the phosphoenzyme was incubated with ADP, 32P counts were transferred to ADP.
The extent of transfer from the phosphoenzyme was minimally about 30%. When corrected for the hydrolysis (43-46%) of the phosphoprotein, about 50% of the 32P from the ["PI phosphoenzyme was transferred to ADP within 1 min.
Attempts to demonstrate the transfer of 32P to acetate were hindered by the lack of resolution between inorganic phosphate and acetyl phosphate on the thin layer chromatography (polyethyleneimine cellulose) plates. Although the conditions of chromatography used were those previously described (l), reproducibility of the results was not sufficient to determine stoichiometry. Semi-quantitative estimates of phosphate transfer were obtained, however, as indicated in Fig. 5. Incubation of the [32P]phosphoprotein with acetate and M&12 caused the concomitant loss of radioactivity from the phosphoenzyme at the origin and its appearance in a spot which migrated with the mobility of acetyl phosphate.
The extent of transfer of the 32P was as high as 50% and was dependent upon the presence of both MgZ+ and acetate (Fig. 5B). Incubation of the phosphoprotein in the buffer without acetate or M e resulted in 50-70% hydrolysis of the phosphate bond of the phosphoenzyme. Nature of the Phosphoryl Linkage in the Phosphoenzym-Preliminary characterization of the enzyme-phosphate bond of acetate kinase was accomplished by studying its stability as a function of pH and its susceptibility to hydrolysis by hydroxylamine. The stability of the phosphoenzyme at 37 "C in buffers of various pH values is shown in Fig. 6. The linkage is acid-labile and more stable at alkaline pH values, with the region of greatest stability between pH 9 and 13. The phosphoenzyme is very unstable in 1 N KOH (data not shown, Ref. 1) where phosphoramidates (e.g. 3-phosphohistidine) are stable (77) and it is susceptible to hydrolysis by neutral hydroxylamine (data not shown), suggesting that phosphoacetate kinase is an acylphosphate. These data agree with those of Anthony and Spector (1,21,22) and Todhunter and Purich (20), who found evidence for an acylphosphate linkage in the phosphoprotein isolated from partially purified preparations of acetate kinase.
Kinetic Competence of the Enzyme-If the phosphoenzyme is an obligatory intermediate in the reaction, the rate of phosphorylation of the protein by ATP should be at least equal to or greater than the catalytic rate of the reaction (i.e. the rate of formation of acetyl phosphate by acetate kinase in the presence of acetate and ATP). Therefore, preliminary experiments were conducted with Drs. Narlin Beatty and M. Daniel Lane (The Johns Hopkins University) with a rapid quench apparatus similar to that described by Ballou and Palmer (61). Using [y3'P]ATP and the homogeneous enzyme, incubations were conducted from 10 ms to 5 s at 22 "C, the reactions were quenched, and the 32P-phosphoprotein was measured. The apparent rate constant for protein phosphorylation was about one-third of the kc, for acetyl phosphate formation at this temperature. It must be emphasized that this is not only a preliminary value, but, because of technical problems involved in making the measurements, it is a minimal value. Nevertheless, the apparent rate constant for phosphorylation of the kinase is well within an order of magnitude of the catalytic rate constant. In view of the significance of these two values with respect to the mechanism of action of the enzyme, more extensive studies will be conducted to assess accurately the relative rates of the two reactions.

DISCUSSION
This paper reports the isolation of homogeneous acetate kinase from the enteric bacteria, S. typhimurium and E. coli.
The two preparations showed the same molecular weights (about 40,000 for the monomers) and similar amino acid compositions. These results are expected, since the two bacterial species are closely related, and other proteins from these two organisms, such as the PTS proteins, HPr and IIIGLc, are very similar or identical (30). The presence of little or no tryptophan in the two kinase preparations is not unique, since the same result has been obtained with acetate kinases isolated from mesophilic bacteria (Table 111).
As is evident from the many papers published on this subject (1-3, 21-25, 61-72), acetate kinase has been very resistant to purification when enteric bacteria such as E. coli were used as a source of the enzyme. One of the most persistent problems is the extreme lability of the kinase after partial purification, as in Step 5 of this procedure. At Step 4, the enzyme is relatively stable to storage when frozen at or below -20 "C. However, after the DEAE-Sephadex column (Step 5), we found it essential to continue as rapidly as possible, even in the presence of glycerol, which has a marked stabilizing effect on the enzyme (1, 39). The homogeneous preparations were stable to storage.
The effect of glycerol on the enzyme is important in a practical sense, but it may also have physiological implications. The mechanisms by which glycerol stabilizes enzymes, particularly those that are cold-labile, are unknown (78, 79), but it has been suggested that glycerol is effective because it lowers the dielectric constant of the medium (1) and acetate kinase may prefer a hydrophobic environment. Both acetate kinase and phosphotransacetylase can associate to some extent with bacterial membranes (80), and acetate kinase with the highly purified membrane vesicles (81) isolated in this laboratory. The membrane-associated activity did not noticeably differ from the soluble enzyme in preliminary kinetic studies (data not shown).
Acetate kinases isolated from various bacterial species have been reported to form oligomers (10, 65,70,72). The present studies suggest that the enzymes from the enteric bacteria behave similarly. Two independent methods give a molecular weight for the fully denatured monomer of 40,000, whereas the native protein elutes from a gel column as a protein fraction of M, = 68,000-70,000, which would be expected for an associating system in rapid equilibrium. More definitive evidence is obviously required concerning the putative association, and we are now attempting to collect this information. One important aspect of the current experiments will be to determine the reasons for the cold lability of the enzyme and for the lag period in the reaction when it is warmed to 36 "C. (32, 82, 83), a self-associating monomer-dimer system whose catalytically active form may be the dimer.

Both of these properties are shown by Enzyme I of the PTS
Partially purified acetate kinase (1, 21, 22, 58) was phosphorylated when incubated with ATP or acetyl phosphate.
However, these preparations were not more than 20% pure (based on specific activities relative to the homogeneous protein), and therefore it is conceivable that the phosphorylation reactions were catalyzed by contaminating protein kinases or that acetate kinase phoshorylated other proteins in the preparations. The present studies show, however, that the homogeneous proteins are phosphorylated autocatalytically. As in earlier work (Ref. 58; however, see Ref. I), the extent of phosphorylation was highly variable, depending on conditions of incubation and separation of the phosphoprotein from excess substrate, and never exceeded 0.4 mol/mol of enzyme monomer. We do not know the reason for the limited extent of phosphorylation, and the explanation may be relatively trivial (e.g. instability of the phosphoenzyme during isolation). However, more important and subtle explanations (e.g. the association-dissociation referred to above) are possible.
The phosphoryl linkage to the protein has all of the properties expected for an acylphosphate rather than a phosphoramidate, phosphoserine, phosphothreonine, or phosphotyrosine (84-88). This conclusion is based on the stability of the phosphoryl linkage as a function of pH, its reactivity with hydroxylamine, and the lack of reactivity with pyridine. The phosphoryl linkage does not show precisely the same pH stability profile as acetyl phosphate, but in at least one instance, this result has been attributed to the local environment of the acylphosphate group in phosphoproteins (89).
The homogeneous phosphoenzyme can transfer the phosphoryl group to ADP and to acetate (see "Results") and catalyzes the expected phosphate exchange reactions between substrate-product pairs (data not shown). These results are consistent with a model for the enzymatic reaction involving the phosphoenzyme as an obligatory intermediate (1,21,22) in the overall reaction. This interpretation is supported in part by our preliminary attempts to compare the rate of protein phosphorylation with the overall catalytic rate of the complete reaction (briefly described under "Results"). An extensive study of this type may show that the rate of kinase phosphorylation is sufficiently rapid so that the phosphoprotein can participate in the overall reaction.
There is considerable evidence against a model for the reaction which requires the phosphoenzyme to be an obligatory intermediate. Stereochemical experiments (26) indicate that an odd number of phosphate transfer reactions occur in the overall reaction because there is an inversion in the configuration of the transferred phosphate. If the phosphotransfer sequence was ATPkinase + acetate, then two transfers would have taken place and would have given a different result. However, the stereochemical results should be reviewed in view of the possibility that the catalytic unit may be a dimer, in which case an odd number of phosphotransfers are possible with the phosphoenzyme as an intermediate. Spector (90) has offered another explanation for the stereochemical results involving a triple displacement mechanism on the surface of the phosphoenzyme.
Kinetic data from some other laboratories using partially purified preparations of the kinase (23, 24) also argue against the phosphoenzyme as an important kinetic intermediate. Our results agree with the earlier reports. That is, the data suggest a sequential rather than the well-behaved ping-pong mechanism required by the phosphoenzyme model. Under certain conditions, however, sequential mechanisms can appear to be ping-pong and vice versa. For instance, a ping-pong mechanism, where the product of the first reaction is only slowly released from the enzyme or where it cannot be released until the second substrate binds, would yield intersecting lines in the Lineweaver-Burk plots, suggesting a sequential mechanism (a ping-pong mechanism should give parallel lines).
Complex models have been proposed for the mechanism of the reaction which attempt to assimilate all of the apparently conflicting observations into one scheme (23-25, 90). The phosphoenzyme has been incorporated into these models, but the major flux of the reaction components does not involve the phosphoprotein, and in this way the models satisfy the kinetic and stereochemical data.
The availability of the homogeneous enzyme will significantly aid in resolving the question of mechanism, first by permitting physical studies on the protein to determine whether it is, in fact, an associating system and the nature of the catalytic unit, and second, by permitting studies on the protein and substrate phosphorylation reactions using rapid quench kinetics.
It is, of course, entirely possible that the phosphoprotein plays no role in the catalytic reaction but that it has other physiological functions. The accompanying paper (31) shows, in fact, that there is phosphotransfer between phosphoacetate kinase and Enzyme I of the PTS, a reaction that may be important in regulating sugar uptake by the cell.  Butterworths Pub., Inc., Woburn, MA 77. Zetterqvist, 0. (1967) Biochim. Biophys. Acta 141,533-539 78. Jarabak, J., Seeds, A. E., Jr., and Talalay

--
The second s e t Of canditiona for phoaphorylation was adapted from the w r k of Webb pf a l . The reaction was slloved to proceed for I-20 n i n a t room tempereture (23.C) and quenched by rapid ~0 0 1 1 n g . The as described below.  -Affi-Gel Blue The resin was washed and equilibrated according to the manufacturer's directions and was regenerated after w e by washing v i t h 8 I urea.  the bulk of the protein.
The enzyme began t o be eluted at 0.5-0.6 .yl ATP. AB shovn i n P i g . l&, acetate kinase was eluted from the column with The main portion of the eluted enzyme wan pooled and concentrated am before, and p u r i t y VIIS estimeted en SDS gels. Acetate kinase la the major protein in the preparation. but contaminating protein-of higher and lower mleCUlar weight are also present. This method vas used a t t h i s s t a g e because other methods including pressure dialysle concentration In d i a l y e i s bags v i t h t h e use of Sephader 0-200 or AquacId& I1 ICalhiochcmicals) to absorb water, or use of the micon premsure c e l l r e s u l t e d In a Bubatantial 1 o~e of activity. Careful use of the Imersible Uollcular separator resulted i n a highly concentrated eample with no apparent 1 0~8 of protein or a c t i v i t y . t r s t e d and stored i n Buffer D Containing smll amounts of hTP (lese than 0.2 mu). When the concentration of Protein is lev, the enzyme loses a c t i v i t y even when frozen Md stored at -2O'C.
In the concentrated state, preparations of homogeneous acetate kinase remained a c t i v e f o r a t thaw rapidly. Repeated freezing and thaving are d e l e t e r i o u s t o the enzyme. l e a s t 1 year when stored a t -2 6 c and care ve,~ taken to freeze and t o I t should again he emphasized t h a t p u r i f i c a t i o n of the enzyme must be conducted as rapidly as possible to avoid msjor loasell and t o I s o l a t e a stable product.
Acetate kinase is mtahlc for several oonths at t h i s s t a g e when CODCIDe fpli Are aumariled i n Tables I and 11, respectively. R C S U l t 8 Of the pullficatiOn procedure for both S. Omhinvriun and