S-Adenosylmethionine Synthetase from Escherichia coZi*

Adenosylmethionine (AdoMet) synthetase has been purified to homogeneity from Escherichia coli. For this purification, a strain of E. coli which was derepressed for AdoMet synthetase and which harbors a plasmid containing the structural gene for AdoMet synthetase was constructed. This strain produces 80-fold more AdoMet synthetase than a wild type E. coli. AdoMet synthetase has a molecular weight of 180,000 and is composed of four identical subunits. In addition to the synthetase reaction, the purified enzyme catalyzes a tripolyphosphatase reaction that is stimulated by AdoMet. Both enzymatic activities require a divalent metal ion and are markedly stimulated by certain monovalent cations. AdoMet synthesis also takes place if adenyl-5’yl imidodiphosphate (AMP-PNP) is substituted for ATP. The imidotriphosphate (PPNP) formed is not hydrolyzed, permitting dissociation of AdoMet formation from tri- polyphosphate cleavage. An enzyme complex is formed which contains one equivalent (per subunit) of adenosylmethionine, monovalent cation, imidotriphosphate, and presumably divalent cation(s). The rate of product dissociation from this complex is 3 orders of magnitude slower than the rate of AdoMet formation from ATP. Studies with the phosphorothioate derivatives of ATP (ATPaS and ATPPS) in the presence of M g + , Mn”, or Co2+ indicate that a divalent ion is bound to the nucleotide

This paper reports a method for preparation of substantial quantities of homogeneous adenosylmethionine synthetase from E. coli. The amount of AdoMet synthetase in the bacteria was increased by construction of a strain of E. coli, which was derepressed for AdoMet synthetase and which harbored a plasmid containing the structural gene for this enzyme. The large amounts of homogeneous enzyme permitted detailed examination of the physical and kinetic properties of the E. coli adenosylmethionine synthetase. Particular emphasis has been placed on the role of the monovalent cation activator and the use of analogs of ATP for studies of the mechanism of the reaction.

RESULTS'
Enzyme Preparation-A strain of E. coli which produces 80-fold more AdoMet synthetase was prepared by incorporating a plasmid containing the metK+ gene3 into a metJ host. Forty-three milligrams of homogeneous AdoMet synthetase with a specific activity of 2.2 pmol/min/mg were obtained from 50 g, wet weight, of cells (see Table I of the miniprint supplement).
Molecular Weight-The molecular weight of the native AdoMet synthetase was 180,000 (determined by gel filtration). Sodium dodecyl sulfate electrophoresis showed a single band of 43,000 molecular weight, indicating that the native enzyme is a tetrameter of identical subunits. The amino acid composition of the enzyme is given in Table I1 in the miniprint.
Metal Ion Requirements of Adenosylmethionine Synthetase-Adenosylmethionine synthetase from yeast requires a divalent cation for activity and is greatly stimulated by certain monovalent cations (5). Fig. 1A in the miniprint supplement shows the effect of divalent metal ion concentration on the activity of the homogeneous E, coli enzyme. The divalent cation requirement was satisfied by Mg2' (relative maximal rate = LO), Mn2+ (0.9), Co" (l.l), Zn2' (0.35), Cd" (0.09), Nil' (0.08), or CaZ+ (0.04) but not by Cu2+ (t0.0001). In the absence of added divalent metal ion, no AdoMet formation was detected (~0.01% of the rate with Mg"). With Mg", maximal activity was not obtained until the metal ion concentration was a t least equal to the ATP concentration (measured at 3 mM, 10 m, and 20 mM ATP). Since the K , for ATP is 0.1 mM in the presence of excess Mg2' (Table 111, miniprint), the Mg2+ activation curves indicate either that free ATP is a potent inhibitor of the reaction or that free divalent cations (present in excess over those bound as MgATP) are required for full activity.
The requirement for a monovalent cation for significant activity was satisfied by K' , NH,', TI', CS' , Li' , and Na' but not by Tris' or (CH3)&N+ (Fig. lB, miniprint). Table 111 (miniprint) reports the K,,, values for ATP and methionine in the presence of various monovalent cations, the concentrations of monovalent cations required for half-maximal activation, and the relative Vmax values with the different monovalent cations. The concentration of T1' required for halfmaximal activation is substantially lower than the concentration of K' required for half-maximal activation, in common with several other enzymes in which T1' replaces K' as activator (15). Fig. 2 (miniprint) shows double reciprocal plots for ATP and methionine at several concentrations of K' . For both substrates, as the concentration of K' increases, the apparent V,,, increases, and the K,,, decreases. The data indicate substantial synergism in the binding of ATP and K' (approximately 4-fold) and in the binding of methionine and K' (approximately 12-fold) (16). In the presence of poor activators, e.g. Li' and Na', the K, values for ATP and methionine (cf. Table 111) approach the values expected for zero K' concentration (obtained by extrapolation of the data of Fig. 2 (miniprint)), showing that Na' and Li' do not appreciably synergize substrate binding. Tripolyphosphatase Activity of Purified AdoMet Synthetase-The finding of an AdoMet-stimulated tripolyphosphatase activity in preparations of yeast AdoMet synthetase supported the postulate that tripolyphosphate might be an intermediate in the AdoMet synthetase reaction (5, 17); consistent with this proposal, a small amount of tripolyphosphate was isolated from reaction mixtures containing partially purified yeast AdoMet synthetase (17).
Adenosylmethionine synthetase from E. coli co-purifies with an activity which hydrolyzes tripolyphosphate to orthophosphate and pyrophosphate. Evidence that the same enzyme catalyzes both the AdoMet synthetase reaction and the tripolyphosphatase reaction is: (i) the AdoMet synthetase and tripolyphosphatase activities are both present in the electrophoretically homogeneous enzyme; (ii) the ratio of these activities is the same in homogeneous AdoMet synthetase purified from E. coli strains which do or do not harbor the plasmid containing the metK gene although the overall purification factor differs by more than 5-fold in the two cases (Table IVC, miniprint) and different purification techniques were used (iii) identical rates of inactivation of the AdoMet synthetase and tripolyphosphatase activities were obtained by heating and by treatment with various protein modifkation reagents (Fig. 3, miniprint, and miniprint text); and (iv) the tripolyphosphatase is stimulated by AdoMet and monovalent cations (Table IVA, miniprint). The results with the homogeneous E. coli AdoMet synthetase are consistent with suggestions that a tripolyphosphatase activity is an integral part of the synthetase from various sources (7, 17).
The hydrolytic activity is relatively specific for tripolyphosphate; no activity (<0.1% of the activity with tripolyphosphate) was obtained if pyrophosphate or the cyclic trimetaphosphate was used. ATP (5 mM) is hydrolyzed to ADP at a rate approximately 0.1% of the maximal rate of tripolyphosphate cleavage.
The tripolyphosphatase activity requires a divalent cation for activity (Table IVB, miniprint) and was stimulated 10-to 20-fold by AdoMet plus monovalent cations (cf., Table IVA, miniprint). The relative ability of monovalent cations to activate the AdoMet-stimulated tripolyphosphatase differs from the pattern observed with the overall AdoMet synthetase reaction, particularly in that Li' is the best activator of the tripolyphosphatase, while it is a poor activator of the synthetase reaction. The K, for tripolyphosphate varies substantially with the monovalent cation activator; however, the concentration of adenosylmethionine required for half-maximal activation is 3 p~ for all monovalent cations tested and a t all concentrations of K' tested ( Fig. 4, miniprint). Maximal stimulation of the tripolyphosphatase activity requires both AdoMet and a monovalent cation (Table IVA, miniprint), with only a small stimulation by either AdoMet or a monovalent cation alone. With each monovalent cation examined, the rate of tripolyphosphate hydrolysis in the presence of adenosylmethionine is fast enough for it to be an integral part of the overall AdoMet synthetase reaction (Table V, miniprint).
Steady State Kinetic Mechanism of the Synthetase Reaction-Double reciprocal plots for ATP at varying concentrations of methionine and for methionine at varying concentrations of ATP intersect to the left of the 1/ V axis (Fig. 5, A and  B, miniprint). Therefore, the reaction proceeds by a sequential mechanism where no products are released until both substrates have been bound (16).
Inhibition experiments (Table  VI, miniprint) show that AdoMet, AMP-PNP, and tripolyphosphate compete with ATP and are noncompetitive with methionine. S-Carbamylcysteine, a reported inhibitor of AdoMet synthetase (18), is competitive with respect to methionine and noncompetitive with respect to ATP. Pyrophosphate and orthophosphate are both noncompetitive inhibitors for both ATP and methionine. The inhibition patterns are consistent with random binding of ATP and methionine and ordered product release, pyrophosphate and orthophosphate dissociating before AdoMet. A more detailed discussion of the kinetic mechanism is given in the miniprint section.
Selenomethionine-Selenomethionine reacts more rapidly than methionine in the synthetase reaction (Table V, miniprint), as has been reported for AdoMet synthetase from yeast (19). The activity of selenomethionine with various monovalent cations is given in Table VI1 (miniprint) as well as Table  V (miniprint). The K,,, values for selenomethionine in the presence of good activators (K' and NH,') are significantly lower than the K , values in the presence of weaker activators (Na' and Li+), paralleling the results with methionine (Table  111, miniprint). Table V (miniprint) presents the effects of different monovalent cations on the synthetase reactions with methionine and selenomethionine and on the AdoMet-stim-ulated tripolyphosphatase reaction. In the presence of K' , the rate of the AdoMet-stimulated tripolyphosphatase reaction is equal to the rate of the synthetase reaction with selenomethionine, while with other monovalent cations the tripolyphosphatase rate is faster than the synthetase rate.
Nucleotide Analogs- Table VI11 (miniprint) reports the results obtained with various nucleotides as substrates for AdoMet synthetase. 3'-Deoxy-ATP is a good substrate for the synthetase reaction, showing that the hydroxyl group at the 3' position of the ribose is not intimately involved in the reaction mechanism. The fluorescent ATP analog, formycin triphosphate (20), is also a good substrate for the AdoMet synthetase reaction.
Separation of Adenosylmethionine Formation from Tripolyphosphate Cleavage with Adenylyl Imidodiphosphate as Substrate-AMP-PNP is a substrate for AdoMet synthetase with a maximal steady-state rate of 0.04% of the maximal rate with ATP. In the reaction with AMP-PNP, AdoMet is formed (Fig. l), but no orthophosphate or pyrophosphate is detected ( t 6 % of the amount of AdoMet formed), indicating production of imidotriphosphate as the second product.

AMP-PNP
The low steady-state rate prompted investigation of the presteady-state kinetics of the reaction (Fig. 1). An initial burst of AdoMet formation occurs immediately after the reaction is initiated, followed by a slower rate of AdoMet production. The same type of curve is obtained whether the reaction is initiated by addition of enzyme or by addition of K' , M e , methionine, or AMP-PNP to enzyme which has been preincubated with other components of the reaction mixture. The burst height is proportional to the enzyme concentration but does not change when the concentration of substrates is increased to twice the concentrations used in the experiment in Fig. 1. The burst height corresponds to formation of approximately 1 mol of AdoMet/mol of enzyme subunit, indicating that on the enzyme, the equilibrium of the reaction lies far toward formation of AdoMet and imidotriphosphate. The burst of AdoMet is characterized by a first order rate constant ( t 1 / 2 -12 s, obtained from a semilog plot (not shown)) of -3.5 min". The rate is independent of enzyme and substrate concentration and, thus, reflects a slow step involving the enzyme .substrate complex and not slow substrate binding. Since a burst of AdoMet is observed, the step whose rate is being measured must either be AdoMet formation or a slow step preceding AdoMet formation. After the initial burst of AdoMet formation, the steady-state rate is proportional to the enzyme concentration, with a rate of 0.04 mol of AdoMet/ min/mol of enzyme monomer. This slower rate appears to reflect the rate of product release from the enzyme (see below). When Li' replaces K' as activating monovalent cation, a similar burst of AdoMet occurs, with tlI2 -20 s and a burst height of approximately 0.6 mol of AdoMet/mol of enzyme monomer (Fig. 6A, miniprint); the lower burst height with Li' reflects in part the closer values of the burst rate and subsequent rate, as well as perhaps incomplete saturation of enzyme with substrates or perhaps a smaller equilibrium constant for the reaction in the presence of Li+. With Li', the rate of AdoMet formation following the burst is 0.16 mol/min/mol of enzyme monomer, which is 4-fold faster than the corresponding rate in the presence of K' .
Since selenomethionine reacts more rapidly than methionine in the presence of ATP, the reaction of AMP-PNP with selenomethionine was examined in order to determine whether the rate of the burst phase is faster in the presence of selenomethionine. Fig. 6B (miniprint) shows the reactions with methionine and selenomethionine a t 5°C (to obtain better temporal resolution) in the presence of K' . The burst heights and rates are, within error, the same for the two amino acids ( tl12 -25 s, burst height approximately 1 mol of product/ mol of enzyme monomer). (With ATP as substrate, selenomethionine is a 1.6-fold better substrate than methionine a t 5°C.) An analogous result is obtained with Li' as activator (at 25"C), where no significant difference in burst rate occurs when selenomethionine replaces methionine. The steady-state rate with selenomethionine is faster than the steady-state rate with methionine with both K' and Li' as activators.
Pulse-chase experiments were performed in order to determine whether substrate binding is rapidly reversible and to investigate the rate of reversal of the AdoMet synthetase reaction, starting from enzyme-bound AdoMet and imidotriphosphate. Enzyme was incubated with [8-3H]AMP-PNP, [methyZ-14C]methionine, Mg2+, and K' for 20 s a t 5"C, a t which time a 20-fold excess of either unlabeled AMP-PNP or unlabeled methionine was added. Within the time needed to mix the sample and remove an aliquot (approximately 10 s), production of AdoMet containing the diluted isotope had stopped, while the burst of AdoMet containing label from the undiluted substrate continued (Fig. 7, miniprint). Thus, within 10 s enzyme-bound substrate had completely equilibrated with the isotopically diluted pool of free substrate. The incubation was then allowed to proceed for 8 h in order to determine whether reversal of the reaction to form dissociable substrates occurred. No change (<lo%) in the radioactivity in AdoMet from the diluted isotope occurred even if an 80-fold excess of cold methionine had been added. The rate of the reverse reaction from enzyme-bound AdoMet and imidotriphosphate to a dissociable enzyme-substrate complex was estimated as e 2 X of the rate of formation of AdoMet from AMP-PNP and methionine.
Evidence that the slow reaction rate following the initial burst of AdoMet formation results from slow product release  Fig. 1) and passed through a Sephadex G-25 column which had been equilibrated with 0.1 M Tris-chloride, pH 8.3, 0.1 M KC1, and 5 mM MgC12. Upon elution with the same buffer, both 3H and '*C co-chromatographed with the enzyme activity, with a constant ratio of 'H to I4C to enzyme activity across the peak ( Fig. 2A). The radioactivity indicated a 1 to 1 correspondence of 3H to 14C, and greater than 90% of the radioactivity from both isotopes was recovered as AdoMet after a standard assay.
When [a-"'P]AI$lP-PNP and [methyl-"H]methionine were present in the incubation mixture, .'H and ."P chromatographed in a 1 to 1 ratio with the enzyme peak, To identify the "'P-containing compound, protein was precipitated with cold 10% trichloroacetic acid followed by centrifugation. The supernatant was spotted on polyethyleneimine cellulose thin layer plates with internal standards of AMP-PNP, pyrophosphate, orthophosphate, imidodiphosphate, and tripolyphosphate. Plates were developed with 1.5 M LiC1, the standards visualized by UV absorbance or by spraying for phosphate (21), and the spots cut out and counted. The :j2P chromatographed with tripolyphosphate, consistent with the presence of ["P]imidotriphosphate. T o show the requirements for binding, parallel incubation mixtures were chromatographed in buffer which either lacked KC1 or contained 50 m~ EDTA in place of MgC12; in neither case did radioactivity co-chromatograph with the enzyme activity. No radioactivity migrated with the enzyme if [8,5'-"H]ATP was substituted for AMP-PNP, if methionine was omitted from the reaction mixture (with either [8-3H]AMP-PNP or [a-'"PIAMP-PNP present), or if after the incubation the enzyme was denatured (e.g. by addition of an equal volume of 10% sodium dodecyl sulfate followed by chromatography in 1% detergent). Thus, the radioactivity co-chromatographing with the enzyme represents AdoMet and imidotriphosphate bound to the enzyme, presumably in conjunction with Mg"+ and K'. The complex was not covalently bound to the enzyme since it was dissociated from the enzyme by sodium dodecyl sulfate and no acidinsoluble counts were found when the incubation mixture was treated with cold 10% trichloroacetic acid.
T o determine the number of monovalent cations bound per active site, a gel filtration experiment was performed in which 2"T1' was the activating monovalent cation, and the column was equilibrated with buffer containing 0. 4  ATPpS Diastereoisomers as Substrates of Adenosylmethionine Synthetase-Derivatives of ATP in which one of the nonbridge oxygens of the a-or P-phosphoryl group is replaced by sulfur exist as pairs of diastereoisomers (22). Comparison of these diastereoisomers as substrates for several enzymes has elucidated the absolute stereochemistry of the metal-ATP binding sites (22, 23). With Mg'+ (which has a strong preference for coordination to oxygen rather than sulfur ligands) as the activating divalent metal ion, AdoMet synthetase utilized the A isomer of ATP@ (as defined by Eckstein (22), S absolute configuration at the &phosphoryl group (23)) as substrate. The A isomer of ATPPS had a K,,, of 0.05 mM and a V,,, equal to 25% of the V,,, of ATP (Table VIII, miniprint). At a concentration of 1 mM, the B diastereoisomer ( R absolute configuration at the P-phosphoryl group) had no detectable activity (t0.1% the rate of ATP); a K, of 0.5 m~ was obtained in a Dixon plot.
When Mn2+ or Coz+, which coordinate either oxygen or sulfur ligands, replaced Mg2' as the activating divalent cation, AdoMet synthetase utilized both isomers of ATP@ as substrates. In the presence of Mn2+, the relative V,,, values for ATP, ATPPS(A), and ATPPS(B) were 1.0,0.5, and 0.3, while the respective K,,, values were 0.04 mM, 0.1 mM, and 0.4 mM. With Co2+ as the activating metal ion, the relative V,,,, values for ATP, ATPPS(A), and ATPPS(B) were 1.0,0.14, and 0.04, while the K , values were 0.4 mM, 0.3 mM, and 0.2 mM, respectively. The decrease in selectivity toward the diaster-eoisomers of ATPpS when metal ions which will coordinate to either oxygen or sulfur ligands replace Mg" indicates that during the AdoMet synthetase reaction a divalent cation is bound to the @-phosphoryl group of ATPPS and indicates that on the enzyme Mg2+ coordinates to the pro-S oxygen atom on the @-phosphoryl group of ATP (23).
Diastereoisomers of A T P a S as Substrates of AdoMet Synthetase-The a-sulfur-substituted analogs of ATP were also tested as substrates of AdoMet synthetase. At either 0.5 mM or 2 mM concentration, neither isomer served as substrate (<0.01% of ATP) with Mg'+, Mn", or Co2' as divalent cation. In the presence of Mg", both isomers gave K, values of 0.8 mM in a Dixon plot.
In view of recent reports that several enzymes catalyze positional interchange of oxygen atoms of the tripolyphosphate chain of ATP in the absence of an overall reaction (24) and that certain enzymes isomerize thiophosphoryl derivatives of ATP (25), it was of interest to examine whether ATPaS might be converted to either ATPyS or the C5"S-P compound A(S)PPP by AdoMet synthetase. Each isomer of ATPaS was incubated at 1 mM concentration with 0.1 m~ AdoMet synthetase, 10 m~ methionine, 50 m~ K+, and 10 mM Mg" for 4 h, and then the reaction mixtures were chromatographed on polyethyleneimine cellulose thin layer plates. ATPaS remained the sole UV-absorbing spot under conditions where ATPyS, A(S)PPP, ADP& or A(S)PP would have been readily detected. Thus, ATPaS is not isomerized or hydrolyzed at an appreciable rate. In parallel experiments using Mn2+, both isomers of ATPaS were hydrolyzed to ADPaS; after % h, ADP& remained the sole UV-absorbing spot. No attempt was made to quantify the rate of hydrolysis of ATPaS, other than to note that the rate of hydrolysis of ATPaS is of the same order of magnitude as the rate of hydrolysis of ATP.
Cr3'ATP a n d Co3+(NH&ATP--In an attempt to obtain further information regarding the metal ion coordination of ATP in the AdoMet synthetase reaction, the substitution-inert metal-nucleotide complexes Cr"ATP and C O " + ( N H~)~A T P were tested aisubstrates (Table VIII, miniprint). Neither compound showed any activity (<0.01% of Mg"ATP) in the presence or absence of 10 mM MgC12.
Rapid Hydrolysis of 5'-Mercapto-5'-Deoxyzy-ATP (A(S}PPP) by AdoMet Synthetase-A(S)PPP, an analog of ATP in which the tripolyphosphate chain is joined to the ribose by a C-S-P rather than a C-0-P linkage, was tested as a substrate for AdoMet synthetase; no formation of AdoMet was detectable (<0.01% of ATP) (Table VIII, miniprint). A(S)PPP was, however, rapidly and quantitatively hydrolyzed to the corresponding diphosphate. The rate of hydrolysis was independent of the presence of methionine but was inhibited by adenosylmethionine. The hydrolytic activity was abolished by incubation of the enzyme with 1 mM N-ethylmaleimide which inactivates AdoMet synthetase but was unaffected by incubation with 10 mM iodoacetamide which does not inhibit AdoMet synthetase (cf. miniprint). The ratio of AdoMet synthetase to A(S)PPPase activity was identical in homogeneous enzyme prepared from E. coli strains which did or did not contain the metK plasmid (Table IX, miniprint) despite a 5-fold difference in purification factor and use of different purification procedures for the enzyme from different strains, consistent with AdoMet synthetase catalyzing the A(S)PPPase reaction.
A K, of 0.6 mM was determined for A(S)PPP hydrolysis (Table VIII,  . Authentic 4',5"dehydroadenosine (0.6 mM) was tested as a substitute for ATP in the standard assay; no AdoMet was formed (<0.05% of the amount with ATP) even if 3 mM tripolyphosphate was also present in the reaction mixture. 0.9 mM 4',5'-dehydroadenosine did not inhibit the rate of AdoMet formation obtained with 0.1 mM ATP.

DISCUSSION
The current work describes the preparation of substantial quantities of homogeneous AdoMet synthetase from E. coli and investigations of the mechanism of the reaction. These studies were facilitated by the use of a metJ E. coli strain containing a rnetK+ plasmid, since this strain had an 80-fold greater amount of enzyme than a wild type E. coli.
The only other AdoMet synthetases which have been purified to homogeneity are the yeast isoenzymes ( 7 ) . AdoMet synthetases from E . coli and yeast catalyze the same overall reaction, both enzymes require divalent cations and monovalent cations for full activity, and both enzymes have a tripolyphosphatase activity which is stimulated by adenosylmethionine. However, the AdoMet synthetases from yeast and E. coli differ markedly in molecular weight and subunit composition. The two yeast enzymes have a molecular weight of 115,000, with two nonidentical subunits (55,000 and 60,000 molecular weight) ( 7 ) . The E . coli enzyme has a molecular weight of 180,000, with four apparently identical subunits. The E. coli and yeast enzymes differ strikingly in their kinetic behavior. No cooperativity is seen in the kinetics of the E.
coli enzyme, as contrasted to the negative cooperativity described for the yeast enzymes (27). As opposed to the yeast enzymes, neither a lag phase in the rate of AdoMet formation nor AdoMet stimulation of its own rate of formation has been observed with the E. coli AdoMet synthetase (27). The relatively simple kinetic behavior of the E. coli enzyme has allowed kinetic studies with fewer ambiguities than was possible with the yeast enzyme.
The kinetic experiments presented here suggest that the sequence of events in the AdoMet synthetase reaction involves (i) random addition of methionine and MgATP; (ii) formation of AdoMet and tripolyphosphate; (iii) oriented cleavage of tripolyphosphate to yield orthophosphate and pyrophosphate; and finally (iv) product release with orthophosphate and pyrophosphate dissociating before AdoMet. Comparison of the maximal rate of the AdoMet synthetase reaction and the AdoMet-stimulated tripolyphosphatase reaction indicates that the predominantly rate-determining step in the overall reaction precedes the tripolyphosphatase reaction. The ability of the diastereoisomers of ATPpS to serve as substrates for AdoMet formation depends on which divalent metal ion is present. The results indicate that a divalent cation is bound to the nucleotide during the AdoMet synthetase reaction and that on the enzyme Mg2+ is bound to the pro-S oxygen on the p-phosphoryl group of ATP.
Adenosylmethionine synthetase requires a monovalent cation, e.g. K', for maximal activity. T1' is an effective activator The A(S)PPPase reaction can be monitored continuously at 340 nm by coupling A(S)PP production to NADH oxidation with pyruvate kinase and lactate dehydrogenase ( of AdoMet synthetase and acts at lower concentrations than the other monovalent cations tested. Through use of ' "TI' , we have established that the enzyme has a single binding site per subunit for a monovalent cation. In the overall reaction, a good activator, such as K' , not only increases the maximal velocity of the reaction but substantially increases the affinity of both MgATP and methionine for the enzyme. While weak activators, such as Li' and Na', can produce 50% of the maximal activity obtained with K' , the poorer activators do not effectively potentiate substrate binding so that much higher concentrations of substrates are required for maximal activity. The influence of the monovalent cation on the rate of the reaction is shown by comparison of the reactions with methionine and selenomethionine as substrates. In the presence of K' , selenomethionine is a 2-fold better substrate than methionine, and the rate of the overall reaction with selenomethionine is equal to the rate of the tripolyphosphatase. With poorer activators, however, selenomethionine is closer to methionine in activity, and the rate of the overall reaction with both amino acids is less than the rate of the AdoMetstimulated tripolyphosphatase. The involvement of enzyme-bound tripolyphosphate as an intermediate in the AdoMet synthetase reaction, first suggested by Mudd (17), has been of particular interest in studies of the mechanism of the enzyme (7, 17, 19,27). The hydrolytic activity of the E. coli AdoMet synthetase is relatively specific for tripolyphosphate, with hydrolysis of ATP occurring a t only 0.1% the rate of tripolyphosphate cleavage. The hydrolysis of tripolyphosphate is stimulated by AdoMet and monovalent cations. The stimulation is most effective when both types of activators are present together. The ATP analog A(S)PPP is hydrolyzed to the corresponding diphosphate (A(S)PP) at a rate which is only attained with tripolyphosphate in the presence of AdoMet.
Use of the ATP analog adenylyl imidodiphosphate permits the dissociation of AdoMet formation from the subsequent hydrolytic reaction which occurs with ATP as substrate. When AMP-PNP replaces ATP in the AdoMet synthetase reaction, AdoMet is rapidly formed, but the other product of the reaction, imidotriphosphate, is not hydrolyzed. Product release from the ( E -Mg2' a AdoMet . imidotriphosphate. K') complex occurs 2500-fold more slowly than the turnover rate with ATP. The burst of AdoMet formation is stoichiometric with the concentration of enzyme subunits, indicating that the equilibrium of the reaction on the enzyme strongly favors formation of AdoMet and imidotriphosphate. The stoichiometry of one AdoMet formed per subunit further indicates that the E. coli AdoMet synthetase is composed of functionally identical subunits. The burst rate decreases when K' is replaced by Li'; however, with either of these monovalent cations, the rate of product formation is not significantly altered when selenomethionine replaces methionine. The failure of the rate of product formation to increase when selenomethionine replaces methionine in the reaction with AMP-PNP is in contrast to the reaction with ATP as substrate and suggests that with AMP-PNP as substrate the rate-determining step for the burst reaction may be a step whose rate depends on the monovalent cation present, but which precedes AdoMet formation. The possibility does remain that with AMP-PNP as substrate the reaction is unable to proceed at a faster rate than obtained with methionine, and, thus, the rate of the burst of product formation could reflect the actual AdoMet-forming reaction. Despite tests of many compounds as potential intermediates (28), no evidence for an intermediate between substrates and AdoMet has yet been obtained.
The inability to obtain significant reversal of the AdoMet synthetase reaction has been previously noted, and the rate of ATP formation from free AdoMet and tripolyphosphate was estimated as 5 X the rate of the forward reaction of yeast AdoMet synthetase (29). Pulse-chase experiments show that the rate of the reverse reaction from enzyme-bound AdoMet and imidotriphosphate is t 2 x the rate of the forward reaction, and, thus, the essentially irreversible step occurs after reversible substrate binding but before product release and may be the formation of AdoMet. In light of the results with AMP-PNP, which yields a nonhydrolyzable analog of tripolyphosphate as a product, the tripolyphosphatase reaction which normally occurs during the catalytic cycle of AdoMet synthetase may be rationalized as acting to allow the enzyme to turn over efficiently by converting the tightly bound product, tripolyphosphate, to the more weakly bound orthophosphate and pyrophosphate.  Table I for

2.8
Inhlbltlon of adenosylmethronine synthetase  c Measured from the double reciprocal plots fer selenornethmnine.
Nucleotide mbstrates for adenosylmethionine synthetase  a See text. A K, value of 0.6 mM was detemlned for the hydrolysis Of AISJPPP to AISIPP using conditions a8 described above, save for the omission Of mthionlne. The hydrolysis rate was 2.2 times the AdoMet Bynthetase activity with ATP as Substrate. ID