Studies on the Enzymatic Utilization of Amino Acyl Adenylates: the Formation of Adenosine Triphosphate* PAUL BERG Department of dlicrobioloyy, Washington University School of Medicine, St. Louis, JIissouri (Received for publication, March 3, 1958) The enzymatic formation of acetyl coenzyme A from adeno- sine triphosphatc (ATP),' aretate, and CoA was recently shown to occur (1, 2) by a two step reaction mechanism. This involves first, the reaction of ATP and awtate with an elimination of inorganic pyrophosphate and the formation of the 5'-phospho acetyl derivative of ASP. In the presence of CoA, this acetyl deiivative is cleaved to forin acetyl CoA and A5P (Equations 1 to 3). Mg++ BTP + scctate acetyl adenylate + PP (1) (2) hcetyl adenylate + CoA-acetyl CoA + A5P Mg++ ATP + acetate + CoA -' acetyl CoA + A5P + PP (3) The widespread occurrence of this type of reaction in the activation of other fatty acids (3-6), inorganic acids (7-9), and other acyl compounds (10, 11) has been established by a number of investigators. Of particular interest was the discovery by I-loagland et al. (12, 13) and others (14-18) of enzymes which catalyze an amino acid-dependent exchange of ATP and PP3. Purification of several of these enzymes and further studies of the mechanism of the reaction (14-17) revealed that they mere relatively specific for a single amino acid and furthermore, that the reaction very likely involved the formation of a carboxyl phosphate anhydride derivative of A5P and the amino acid with the liberation of inorganic pyrophosphate (Equation 4). Mg++ ATP + amino acid,'amino acyl adenylate + PP (4) The evidence supporting this hypothesis is as follows: (a) There is essentially no exchange of PP32 with ATP in the absence of the amino acid (15, 16); (b) in the presence of hydrosylamine there is an accumulation of equivalent amounts of the amino acid hydroxamate, A5P, and PP (15, 16); (c) free h5P is not fornied as evidenced by the failure to find significant exchange of C1*-A5P with ATP (13, 15), and (d) 018 from thc carboxyl group of the amino acid is transferred to the phosphate group of A5P in the presence of hydroxylamine (19, 20). Various at- tcnipts to isolate the postulated amino acyl adenylates have to date been unsuccessful and these failures have led to the notion that these compounds do not exist in the free state but rather in n complex with the enzyme (13). * This work was supported by a grant from the U. S. Public Health Service. 1 The abbreviations used are: ATP, adenosine triphosphate; co,4, coenzyme A; PP, inorganic pyrophosphate; A5P, adenosine- 5'-monophosphate; Tris, tris(hydroxymethy1)aminomethnne. A-2, alcohol fraction; AS-1, -2, ammonium sulfate fractions. More direct evidence supporting the proposal of amino acid adenylate formation was obtdinecl by using the chemically synthesized derivatives. DeMoss et al. (21) demonstrated that the cnzyme from Escherichia coli which catalyzes the L-leucine- dependent exchange of PP3 with ATP also effects the conversion of bleucgl adenylate to ATP. Subsequently it was reported (22) that L-methionyl adenylate is convwted to ATP by an analogous enzyme isolated from yeast. In the present paper several aspects of the conversion of methionyl adenylate and other amino acyl adenylates to ATP are reported and in a second study (23) methods for the chemical synthesis and purification of a number of amino acyl adcnylatcs are described. MATERIALS AND METHODS was prepared by heating NazHP320a at 400" for 1 hour and isolated by chromat.ography on a Dowes 1 Cl- column. ATP, triphosphopyridine nucleotide, hexokinase, and glucose-fj- phosphate dehydrogenase were purchased from the Sigma Chemical Company. All the amino acids used were Cfp grade as provided by the California Foundation for Hiochemic,al Research. The various amino acid adenylates mere prepared, purified, and determined as described in a second report (23). Assay Procedures-Generally two met'hods were used to measure the formation of ATP from the amino acyl adenylates. The first relied on the conversion of PP32 to a Norit-adsorbable forin (15, 24). The incubation mixture ( 1 nil.) contained 160 pmoles of potassium succinate buffer, pH 5.6, 5 pmoles of MgCl?, 2 pmoles of PP32 containing 2.5 to 10 x 10* e.p.ni. per pniole, between 1 and 1.5 /*moles of the amino acid derivative and the enzyme. All the components except the amino acyl adenylate were mixed, equilibrated at 37" for 3 minutes and then the amino acyl adenylate was added. After 3 minutes the reaction was terminated by the addition of 0.5 ml. of 0.7 hi perchloric acid and the ATP formed was adsorbed on Norit and treated as previously described (15). 1 unit of activity is equal to the formation of 1 /*mole of ATP in 15 minutes under these conditions. In the second procedure the same incubation mixture was used except for the substitution of unlabeled PP for the PP32 and the ATP formed was determined spectro- photometrically by TPN reduction in the coupled hexokinase, glucose-6-phosphate dehydrogenase system (25). The values obtained with these two procedures differed by less than 10 per cent. Using PP3* incorporation as a measure of the reaction, the amount of ATP formed was proportional to the amount of enzyme protein added. Thus with 3.3, 6.5, 13, 26, and 52 fig. of protein, there were 23, 24, 25, 25, and 26 units of activit'y per mg. of protein, respectively. GO1 602 Conversion of Amino Acyl Adenylates to A1'P Vol. 233, No. 3 The purification of the enzyme was followed by measuring the tmethionine-dependent exchange of PP32 and ATP as previously described (15). The unit is defined as the incorporation of 1 pmole of PP3* into ATP per 15 minutes. Enzyme Preparation-In a previous report (L5), a procedure for the purification of an enzyme from brewers' yeast, which catalyzes the exchange of PP32 with ATP in the presence of L-mcthionine, was described. 'This involved fractionation of the yeast extract with alcohol followed by ammonium sulfate precipitation. These procedures yielded a preparation with a specific activity of about 15 (units per mg. of protein) corrlpilred to the initial extract which had a specific activity of approxi- mately 0.2. In the present work, the purification procedure was &ended with a view towards achieving more purified enzyme and to obtaining fractions with varying specific ac- tivitics for studies of the spccificity of thc rcaction. The following is a description of the preparative procediires for the enzymc fractions uscd in the current work. To 25 ml. of the previously described fraction (15) AS-1 (specific activity 8.1, 4.9 mg. of protein per ml) was added 0.5 ml. of a solution of crystalline pancreatic ribonuclease (10 mg. per ml.) and the mixture was incubated at room temperature for 15 minutes. After cooling to 3", 6.5 gm. of ammonium sulfate were added and the pH was adjusted to pH 4.6 by the addition of 0.6 ml. of 1 N sulfuric acid. ilfter 5 minutes the precipkate was centrifuged and discarded. 2.1 gm. of am- monium sulfate were added to the supernatant fluid and after 5 minutes the precipitate was centrifuged and dissolved in 10 ml. of 0.1 M potassium succinate buffer, pH 6.0 (AS-2; specific activity 16.9; 2.5 mg. of protein per ml.). To 4 ml. of fraction AS-2 were added 8 ml. of cold water and 1 ml. of aged alumina Cr gel (26) (containing 15 mg. dry weight per ml.). The gel was centrifuged and washed once with 10 nil. of water, once with 10 ml. of 0.05 M potassium phosphate buffer, pH 6.5, and then with two 5 ml. portions of 0.1 M potas- sium phosphate buffer, pH 7.5. The major portion of the enzymatic activity appeared in the first pH 7.5 eluate (48 per cent) and an additional 10 per cent was present in the second pH 7.5 eluate (first Cr gel eluate; specific activity 32; 0.5 mg. of protein per ml.). Fraction AS-2 was also further purified by zone electrophoresis on a cellulose column (27). 2 ml. of AS-2 (specific activity 17.5; 2.7 mg. of protein per ml.), which originally had been dissolved in 0.02 M Tris buffer, pH 8.0, containing 0.02 hi KC1 and then dialyzed for 3 hours at 3" against the same buffer, was layered on top of a cellulose column (50 x 1 cm.) and subjected to a potential of 420 v. for 15 hours at 3". The enzyme was recovered by displacement elution from the column with 0.02 M Tris buffer, pH 8.0, containing 0.02 M KC1. 50 per cent of the enzyme activity was eluted in 4 fractions (Fractions 11 to 14). The specific activities of the Fractions 11 to 14 wcrc 33.5, 42.5, 32.6, and 20.0, respectively. RESULTS Upon examination of Equation 4, we may predict that an enzyme preparation catalyzing the L-methionine-dependcnt exchange of PP32 with ATP should convert Lmethionyl adenyl- ate to ATP in the presence of PP at a rate not slower than the rate of the exchange. Fig. 1 shows the rate of ATP synthesis from the chemically prepared L-methionyl adenylate. Calcula- t,ion of the initial rate gives a figure of 15.9 pmoles of ATP formed per 15 minutes per mg. of protein. The rate for the exchange reaction under the identical conditions was 4.6 pmoles of PP32 incorporated into ATP per 15 minutes per mg. of pro- tein. The conversion of tmethionyl adenylate to ATP did not go to completion but stopped when approximately 40 to 50 per cent of it appeared as ATP. This limited conversion was due to the destruction of the metliionyl adrnylate during the period of the reaction as indicated by the following cx- periment. An incubation of 0.91 fimolr of methionyl adrnylate and excess PP for 10 minutcs resulted in the formation of 0.39 pmole of ,4TP, but a second 10 minute incubation with a fresh addition of enzyme or of PP resulted in no further increase in ATP (0.39 pmole of ATP). However, a second addition of the amino acyl adenylate after the first incubation, followed by a second 10 minute incubation period, yielded a total of 0.79 pmole of ATP as espected. Further discussion of the lability of methionyl adenylate is presented later. The requirements for the conversion of methionyl adenylate to ATP are shown in Table I. In the absence of enzyme, PP, Mg++, or methionyl adenylate, there was no ,4TP formed. Fur- thermore, if the methionyl adenylate was treated with 0.01 N KOH for 5 minutes at 25' to form free A5P and tmethionine, there was no ATP formation. Stoichwnictry of the Reaction-Attempts to study the stoichiom- etry of the reaction and obtain a balance of the various com- ponents were complicatcd by the marked instability of the substrate under the conditions used. Measurements of the tmethionyl adenylate disappearance were therefore made in the presence and absence of PP to determine the amount of methionyl adenylate utilized for ATP formation. The results, shown in Table 11, indicate that the increment in mcthionyl adenylate disappearance resulting from the presence of PP is in agreement with the amount of ATP synthesized and with the amount of PP32 incorporated into ATP. It is not yet clear whether the methionyl adenylate breakdown in the absence of PP is enzymatically catalyzed. We have found that the half-life of methionyl adenylate under the condi- 0.5~ .c Oa3t i Oe2t I 0 cn aJ 1 I I I I 4 8 12 16 20 Minutes FIG. 1. Kinetics of conversion of L-methionyl adenylate to ATP: The reaction mixture contained, in 1 ml., 160 *moles of potassium succinate buffer, pH 5.0, 5 pmoles MgClz, 2 pmoles PP32 containing 4.6 X lo4 c.p.m. per pmole, 1.05 pmoles L-meth- ionyl adenylate, 85 pg. of enzyme AS-1, spccific activity 7.8. Temperature, 37". September 1958 Components Complete Minus PP Minus L-methionyl adenylate Minus MgClz Minus enzyme Complete, but with hydrolyzed L-methionyl ade- nylate P. Berg ATP formed* pnioles 0.40 0.00 0.00 0.04 0.00 0.00 603 minutes 5 10 15 ~ pmole -0.12 -0.23 -0.26 * ATP was dctermined in a 0.1 ml. aliquot of the heated reaction mixture with the coupled herokinase, glucose-6-phosphate dehy- drogenase assay (25). The reaction mixture contained, in 1 ml., 150 pmoles potassium succinate buffer, pH 5.5, 5 pmoles of MgC12, 3 pmoles of PP, 1.6 pmoles of methionyl adenylate, 165 pg. of .4S-1, specific activity 5.0. Temperature 37", time 15 minutes. The L-methionyl adeny- late was hydrolyzed in 0.01 N KOH at 25" for 5 minutcs and then the solution was neutralized with HC1. TABLE I1 The disappearunce of L-methionyl adenylate and formation of ATP Time I L-Mrthionyl adenylate. ATP formed$ &mole 0.14 0.22 0.25 PP= incor orated into A&P~ pmole 0.13 0.21 0.24 * L-Methionyl adenylate was determined as the hydroxamate with acid FeCL (23). The values for L-methionyl adcnylate dis- appearance in the presence of PP were 0.60, 0.86, and 0.96 pmole, while in the absence of PP the values were 0.48, 0.83, and 0.70 pmole. t ATP was measured in a heated aliquot of the reaction mix- ture with hexokinase and glucose-6-phosphate dehydrogenase (25). $ PP3* incorporation was determined as described in "Meth- ods." The reaction mixture contained, in 0.5 ml., 75 pmoles of potas- sium succinate buffer, pH 5.6, 2.5 pmoles of MgC12, 2 pmoles of PP or PP32 containing 5.4 X lo4 c.p.m. per pmole, 1.15 pmoles of L-methionyl adenylate and 28 pg. of AS-1, specific activity, 7.3. Temperature, 37". tions shown in Table I1 (absence of PP) is 7.5 minutes, whereas with a heat-inactivated sample of the enzyme it was about 14 minutes. This is to be contrasted with a value of greater than 25 minutes observed in the absence of the enzyme preparation. This 2-fold increase in the rate of breakdown of amino acyl adenylate may be due to enzymatic activity or to some other heat-labile component present in the enzyme preparation. Further studies are needed to clarify this point. In order to equate the conversion of methionyl adenylate to ATP with the methionine-dependent L4TP-PP32 exchange re- action, it seemed pertinent to determine if both reactions are catalyzed by the same enzyme. A number of fractions at various stages of purification have been examined for both activities and within the limits of the accuracy of the measure- ments made, the ratios of thc two activities were essentially constant (Table 111). It should be pointed out that these ratios comparc the two reactions under somewhat different conditions. The exchange reaction is carried out at pH 8.0 whereas the amino acyl adenylate conversion to ATP is per- formed at pH 5.6. The eschange reaction activity at 5.6 is about 60 per cent that of pH 8.0 and therefore the ratio is actually between 0.25 and 0.33. No preparation has been found which catalyzes one reaction and not the other. As another line of evidence, experiments which will be discussed later showed that free L-methionine competitively inhibits the conversion of L-methionyl adenylate to ATP and that the Kr value for methionine is in close agreement with the observed K, value for methionine in the ATP-PP32 exchange assay. Thcse findings suggest that Lmethionine acts at the same site and presumably with the same enzyme when functioning as a substrate or as an inhibitor. Specificity-It was previously reported (15) that the L methioninc-activating enzyme was relatively specific for L- methionine since other naturally occurring amino acids did not promote a significant ATP-PP32 exchangc reaction. It was there- fore of interest to examine the specificity with respect to the reverse reaction, namcly, the conversion of other amino acyl adenylates to ATP (Table IV). Although ATP formation from L-methionyl adcnylate is most rapid, there is significant ATP formation from L-seryl adenylate, L-phenylalanyl adenylate, and cvcn D-methionyl adenylate, but little or no utilization of L-trypt,ophanyl adenylate. ATP formation was confirmed by the spectrophometric assay and was not observed when the amino acyl adenylates had been hydrolyzed with dilute alkali. Under the identical conditions however, there was very little or no significant exchange of PP32 with ATP in the presence of L-serine, L-phenylalanine, L-tryptophan, or u-methionine (Table V). These measurements of the exchange reaction, now made at pH 5.6, are in agreement with the previous findings carried out at pH 8.0. The finding of ATP formation from D-methionyl adenylate posed the question of whether there was contamination of the material with the L-methionine derivative. The L-methionine TABLE I11 Comparison of the rate of the ATP-PP3z exchange reaction with the rate of conversion of L-methionyl adenylate to ATP in various enzyme preparations Fraction Alcohol (A-2) Ammonium sulfate (AS-1) Ammonium sulfate (AS-2) Alumina Cy gel eluate Cellulose column electropho- resis of AS-2, Fraction: 10 11 12 13 14 15 ~~ pecific activity ATP-PPJZ ex- change reac- tion A units/mR. of prolein 2.4 7.3 13.1 32.0 10.4 33.5 42.5 32.6 20.0 9.7 peci6c activity, L-methionyl adenylate conversion to ATP B unilslmg. of prolein 5.1 16.0 25.8 65.1 20.8 60.5 86.7 65.0 40.7 20.6 - Ratio, A:B __ 0.47 0.46 0.51 0.49 0.50 0.55 0.49 0.50 0.49 0.47 __ 604 Conversion of Amino Acyl Adenylates to ATP Vol. 233, No. 3 TABLE IV The conversion of amino acyl adenylates to ATP by the methionine-activating enzyme Amino acyl adcnylatc ATP formation I None r,-Methionyl adenylate L-Seryl adenylate L-Phenglalunyl adenylate I,-Tryptophanyl adenylate D-Methionyl adenylate rmolesjl5 min./mg. of protein 0.05 42.0 10.0 5.0 0.10 6.0 The reaction mixture contained, in 1 ml., 150 pmoles of potas- sium succinate buffer, pH 5.6, 5 pmoles of MgCl?, 2 pmoles of PP32 containing lo5 c.p.m. per pmole, 1.1 pmoles of L-methionyl ade- nylate, 1.2 pmoles of L-seryl adenylate, 1.2 pnioles of L-phenyl- alanyl adenylate, 1.2 pmoles of L-tryptophanyl adenylate, and 1.1 pmoles of u-methionyl adenylate. In all cases, except with L-tryptophanyl adenylate, enough enzyme was used so that 0.25 to 0.50 pmole of ATP was formed during the incubation period. TABLE V The exchange of PP2 and -4TP with ~-rnethionine, L-serine, L-phenylalanine, L-tryplophan, and D-methionine Amino acid I PPa2 incorporated into ATP None L-Methionine L-Serine L-Phenylalanine L-Tryptophan D-Methionine unilsjmg. o/ prolein <0.10 10.4 <0.10 <0.10 <0.10 <0.10 The reaction mixture was the samc as described in Table IV except that 2 pmoles of ATP and 2 rmoles of each of the amino acids replaced the amino acid adenylates. could have been present as an original contaminant in the D- mcthionine; it could have also been formed during the synthesis of the adenylate or during the incubation with the enzyme. All of thcse possibilities were eliminated by the following experi- ments. Polarimetric examination of the starting D-methionine showed an [a], in 5 N HC1 of +23.9" compared to the value +23.4" given in the literature (28). Moreover, the sample of D-methionine did not catalyze the exchange of ATP and PP32 at a level which would have detected a 1 per cent contamination with kmethionine. Hydrolysis of D-methionyl adenylate with 0.01 N KOH and testing of the hydrolysate for its ability to replace L-methionine in the exchange reaction was likewise negative. For example, in one experiment, 1 pmole of L- methionine and a hydrolysate of L-methionyl agenylate con- taining 1 pmole of L-methionine gave an incorporation of 0.2 and 0.19 pmole of PP2, respectively, into ATP. On the other hand, 1 pmole of D-methionine, free or in the form of the D- methionyl adenylate hydrolysate, gave less than 1 per cent of this value. Furthermore, preincubation of the ~-methionyl adenylate with the enzyme and then followed by hydrolysis and testing in the exchange reaction likewise gave no evidence for the presence of L-methionine. This apparently anomalous finding posed the question whether additional mechanisms exist for utilizing the amino acid adenyl- ates other than that shown in Equation 4. We investigated this question first by determining if the conversion of L-methionyl adenylate and other amino acid adenylates to ATP was catalyzed by the same enzyme. Various enzyme preparations were compared for their ability to convert L-methionyl and L-scryl adenylates to ATP (Table VI). The results showed that the ratios of the two activities were essentially constant over an approximately 17-fold range of enzyme purification. Moreover, attempts to effect a preferential heat inactivation of one activity with respect to the other demonstrated that the kinetics of inactivation were essentially first order with both substrates and that the ratio of the two activities remained the same throughout (Table VII). The same result was obtained under somewhat varied conditions for the heat inactivation (Esperi- inent 2). Additional evidence which supports the idea that a single enzyme catalyzes the conversion of both Gmethiongl and L-seryl adenylates to ATP mas obtained by kinetic experiments. L-Methionine competitively inhibits the conversion of both L-methionyl and L-seryl adenylates to ATP (Table VIII). IrSerine (2 to 48 x M), on the other hand, does not inhibit ATP formation from either of these substrates. Dixon (29) has described a graphic method for determining the k'l and K, values which involves a plot of 1/V against the inhibitor concentration (I) at two different substrate concentrations (Fig. 2A and B). The point at which the two lines intercept one another is equal to -Kx and the intercept of each curve with the abscissa is equal to -ICl(S/K + 1). From these data, the Zix values for kmethionine acting as an inhibitor of ATP formation from L-methionyl and r,-seryl adenylates were 3 TABLE VI The activity of various enzyme preparations in the conversion of L-methionyl and L-seryl adenylutes to ATP Preparation A-2 AS-1 AS-2 AS-2a -4S-2~ -4S-2d Alumina Cy gel eluate Cellulose column electropho- resis of AS-2, Fraction : BS-2b 10 11 12 13 14 15 Specilk activity L-Methionyl adenylate conversion to ATP A unitslnrg. o/ prolein 5.1 16.0 25.8 33.4 27.2 25.9 19.8 65.1 20.8 60.5 86.7 65.0 40.7 20.6 L-Seryl conversion adenylate to ATP B unitslmg. o/ pinlein 1.2 4.0 6.1 8.1 6.6 6.5 4.8 15.9 4.1 12.6 19.9 15.4 8.4 4.0 Ratio, A:B 4.3 4.0 4.2 4.1 4.1 4.0 4.1 4.1 5.1 4.8 4.4 4.2 4.9 5.1 ~ AS-% to 2d refers to ammonium sulfate fractions derived from AS-2 and which were prepared as follows: 0 to 0.44 saturation (AS-2a), 0.44 to 0.48 saturation (AS-2b), 0.48 to 0.51 saturation (AS-2c), 0.51 to 0.62 saturation (AS-2d). September 1958 P. Berg GO5 and 3.5 X lo-* M, respectively. The K, for L-methionine in the exliange reaction measured under the same conditions (pH 5.6) was 2.6 x lo-* M. The agreement between the apparent dis- sociation constants of free methionine acting both as a sub- strate and inhibitor suggests that it is acting at the same site in both reactions. The K, for L-methionine at pH 5.6 lyas as TABLE VI1 Heat inactivation of enzyme preparation catalyzing conversion of L-mcthionyl and L-seryl adenylates to ATP Experi- ment I I1 rime at 45" min. 0 1 2 3 5 7 0 2.5 5 7.5 L-Methionyl adenylate conversion to ATP A unils/ml. 7.6 6.4 3.6 2.3 1.2 0.68 97 92 80 68 L-Seryl adenylate conversion to ATP B units/ntl. 1.8 1.6 0.88 0.52 0.28 0.16 23 18 18 16 Ratio, A:B 4.2 4.0 4.1 4.4 4.3 4.3 4.2 5.1 4.4 '4.2 In Experiment I the enzyme (AS-2,280 fig. per ml.) was kept in a water bath at 45" and aliquots were removed at various times, cooled and the activity with both L-methionyl and L-seryl ade- nylates determined. In Experiment 11, the enzyme (AS-2, 3.1 mg. per ml.) containing 1.8 pmoles of ATP and 2 pmoles of MgC12 per ml. was treated as above. The cooled aliquots were diluted 50- to 200-fold in the final reaction mixture so that the amount of ATP present was negligible. TABLE VI11 The effect of methionine on the conversion of L-melhionyl and L-seryl udenylates to ATP Concentration of amino acyl adenylate X loa Y L-Methionyl adenylate: 1.1 1.1 1.1 1.1 1.1 0.55 2.2 2.2 L-Seryl adenylate: 1.2 1.2 1.2 2.4 2.4 4.8 4.8 Concentration if L-methionine ~~ x IOZ .w 0.0 2.4 5.4 10 20 10 11 21 0.6 1.2 0.6 1.2 0.6 1.2 Amino acyl adenylate con- iersion to AT1 unils/mg. a/ profern 31.5 24.0 19.5 13.6 8.3 9.2 18.0 11.8 7.4 3.5 2.5 4.4 3.5 5.3 4.3 The assay conditions are described in "Methods." zyme used was AS-2 with a specific activity of 16. ihibition % 24 38 57 74 71 43 62 53 66 40 53 28 42 Che en- 15i / I I 1 1 I I I I 3'2 I I 23456 [I], ,u Moles per ml. 7 I I I I 2 I/' 12345 [I],,u Moles per ml. FIG. 2. Analysis of the inhibition of ATP formation from I,- methionyl and L-seryl adenylates by L-methionine: Curve A, inhibition of L-methionyl adenylate conversion to ATP; Curve B, inhibition of L-seryl adenylate conversion to ATP. 1/V is the reciprocal of the rate of ATP formation in pmoles per minute per mg. of protein. The reactions were carried out as described in "Methods. " mentioned above 2.6 x M but calculation of previous data obtained at pH 8.0 gave a value of 1.0 x N or about 25 times lower. The factors contributing to this higher affinity at pH 8.0 are not known and require further work. The K, for &me- thionyl adenylate determined graphically (29) was 3.7 f 0.8 x 10-5 M and the K. for L-seryl adenylate was 1.2 f 0.2 x 10-3 M. Thus there seems to be about a 30-fold difference in the apparent affinity of the two amino acid adenylates at pH 5.6. During the course of these studies Dr. W. P. Jencks suggested? that amino acids other than L-methionine might be active in the ATP-PP32 reaction if tested at high concentrations. When this was carried out it was found that L-serine, o-methionine, and 2 Private communication. 606 Conversion of Amino Acyl Adenylates to ATP Vol. 233, No. 3 L-threonine, had low but significant activity in promoting the eschange, while Gtryptophan (0.03 hx) and L-isoleucine (0.14 M) lid no detectable activity. Fig. 3 shows that at 0.13 M, L-serine does promote a slow incorporation of PP32 into ATP. Higher concentrations of L-serine were inhibitory. D-Methionine (0.03 hi) and L-threonine (0.25 ra) were one-fifth and two-thirds as ac- tive as L-serine while higher concentrations were also inhibitory. Since essentially all the studies of the amino acid-activating en- zymes (15-17) have revealed a very high affinity of the enzyme for the amino acid in question, its seems unlikely that this slow rate of exchange found nt high concentrations of amino acid is chic to the presence of small amounts of enzymes specific for these amino acids. Rather, it secms more likcly that the enzyme which is "specific" for L-met,liionine has a relatively low affinity and activity with certain other amino acids. These findings point to t,he likelihood that although the enzyme is relatively specific for a single amino acid and its adenylate, it also utilizes other amino acids and their adenyl derivatives with a lower effi- ciency. Similar findings have been reported with a fatty acid-activat- ing enzyme by Jencks and Lipmann (6). They showed that with the enzyme specific for fatty acids of intermediate chain length (4 to 12 carbon atoms), and under their usual assay condi- tions, acetate (5 x 10-3 M) was not converted to acetyl Coil. However, acetyl adenylate was utilized for acyl CoA synthesis at about half the rate of hexanoyl adenylate. FurtKr studies re- vealed that with higher conccntrations of acetate, acetyl CoA 50- c Q, .- .I- - 2 Q 40- d, E 01 a Q i- L- - a 30- c .- N - ro a - 20- a 0 Ln Q, 0 - 5 ?--x X- None 30 60 90 120 150 180 Minutes FIG. 3. The exchange of PP3* and ATP in the presence of high concentrations of L-serine: The reaction mixture (1 ml.) contained 100 pmoles of Tris buffer, pH 8.0, 5 pmoles of MgCL, 2 pmoles of ATP, 2 pmoles of PP3* containing 6 X lo4 c.p.m. per pmole, 1 pmole of L-methionine or 125 pmoles of L-serine. Temperature, 37". synthesis was demonstrable; this being maximal at an acetate concentration of 0.4 M. DISCUSSION The conversion of L-methionyl adenylate to ATP by the puri- fied methionine-activating enzyme supports the mechanism pro- posed for the bmethionine dependent exchange of ATP and PPr2 shown in Equation 4. The surprising finding in these ex- periments was that, although the enzyme is relatively specific for Irmethionine in the exchange reaction, other amino acid adenyl- ates are also converted to ATP. Purificat,ion of the enzyme and kinetic studies have indicated however, that at least in the ca.se of the utilization of L-methionyl and L-seryl adenylates, the same enzyme is responsible for both reactions. Further studies re- vcaled that the specificity for methionine was relative since at very high concentrations of L-serine, L-threonine, and D-methi- onine there was a detectable ATP-PPS2 exchange. Similar find- ings to those described here have also been reported for experi- ments with crude extracts of Escherichia coli by DeMoss et al. (21). These workers observed that Icalanyl adenylate was con- verted to ATP while L-alanine was inactive in promoting the exchange of PPS2 and ATP. We have made similar findings in E. coli extracts with L-seryl and L-threonyl adenylate^.^ Exactly which enzyme (or enzymes) is rcsponsiblc for the utilization of these compounds remains to be determined. The failure to detect the accuniulation of L-methionyl adenyl- ate from ATP and L-methionine remains a puzzling feat'ure of this reaction. Because of the rapid destruction of methionyl adenylate at pH 8 it does riot seem likely that the enzymatically formed compound exists to any significant extent in the free state. If this were so, one might expect to observe a significant increase in the breakdown of ATP to A5P and PP in the presence of L-methionine. This, however, was not observed in our earlier studies (15). The question of whether the enzymatically formed amino acid adenylates are bound to the enzyme and thereby stabilized remains to be resolved. The mechanism of niet,liionyl aclenylate formation resembles ` the first step in the formation of the fatty acid-CoA derivatives (2-6). More recent studies (30) indicate that the similarity ex- tends to the second step as well since the methionine-activating , enzyme has been shown to transfer the amino acid moiety to a polyribonucleotide. Thus the amino acid-activating enzymes like the fatty acid-activating enzymes appear to catalyze both the transfer of an adenyl group from ATP to an acyl group and the transfer of the acyl group to an appropriate acceptor. SUMNARY A purified cmet~iiunlrie-activating enzyme from yeast has been shown to convert L-methionyl adenylate to adenosine triphos- phate (ATP) in the presence of inorganic pyrophosphate (PP). The rate of this conversion is consistent with its role as an inter- mediate in the L-methionine dependent exchange of PPS3 and ATP. ATP + L-methionine L-methionyl adenylate + PP The same preparation also catalyzes a significant conversion of L-seryl, bphenylalanyl, and D-methionyl adenylates to ATP even though it does not utilize the free amino acids in the exchange re- s Unpublished observations of the author. September 1958 P. Berg 607 :tetion. Purification studies have indic:tted that the cnzyine (2.6 x which utilizes kseryl adenylate is the same one which converts L-methiongl adenylate to ATP. This was further supported by the finding that the K, for L-methionine in the eschmge rextion M) is the same as the 1'1 value for L-methionine acting as an inhibitor of L-methionyl adenylate conversion to ATP (3 x M) and for L-sergl adenylate conversion to ATP (3.5 x RRFRII.13N CES M). 1. BERG, P., J. Am. Chena. SOC., 77, 3163 (1955). 2. BERG, I>., J. Bioi. Chem., 222, 991 (1956). 3. TALBERT, P. T., AND HUENNEKINS, F. M., J. Ani. Cheiii. soc., 4. YENG, C. €1. L., Hiochini. et Bioph~ys. Acta, 22, 42 (1956). 78, 4671 (1956). 5. i\lOYED, H. S., .4ND LIPMdNN, F., J. Bacterid., 73, 117, (1957). 6. JENCKH, w. P., AND IAPhIANN, F., J. BiOl. ChC7tl., 225, 207 (1957). 6409 (1956); J. Biol. Chem., 229, 837 (1957). Cheni. Snc., 78, 6408 (1956). Federution Proc., 15, 214 (1956). Reports, 1955. Federation Proc., 16, Y36 (1957). 7. ROBBIXS, P. W., AND LIPIMANN, F., J. Ani. Chenz. Soc., 78, 8. BANDURSKI, R. S., WILSON, L. G., .~ND SQUIRES, C. I,., J. Ana. '3. BACHIIAWAT, 13. IC., WOESSNER, J. l?., JR., AND COON, M. J., 10. RIAAS, W., 3rd International Congress oi Biochemistry, Brussels, 11. REED, L. J., LEACH, F. It., KOIKE, M., AXD LEVITCH, AI. E., 12. HOAGLAND, NI. B., BiochittL. et Biophys. Acla, 16, 288 (1955). 13. HOAGLAND, M. B., KELLER, E. B., AND ZARIECNIK, P. C., .J. Riol. Chem., 218, 345 (1956). Acla, 22, 49 (1956). 14. D~Moss, J. A,, AND NOVELLI, G. D., Bzochim. et Biophys. 15. BERG, P., J. Biol. Chevi., 222, 1025 (1056). 16. DAVIE, E. w., I