MECHANISM OF HYDROLYSIS OF ADENOSINETRIPHOS- PHATE CATALYZED BY PURIFIED MUSCLE PROTEINS*

Transcription

1 MECHANISM OF HYDROLYSIS OF ADENOSINETRIPHOS- PHATE CATALYZED BY PURIFIED MUSCLE PROTEINS* By D. E. KOSHLAND, JR., ZELDA BUDENSTEIN, AND ARTHUR KOWALSKY (From the Department of Biology, Brookhaven National Laboratory, Upton, New York) (Received for publication, June 4, 1954) At the present time, the source of energy in muscular contraction and the proteins which undergo the contraction are fairly well known (2). The detailed mechanism of contraction, i.e. the manner in which the chemical energy in the adenosinetriphosphate (ATP) molecule is converted into the linear changes in the protein molecule, is still obscure. As a step in the elucidation of this interaction, the hydrolysis of ATP catalyzed by purified rabbit muscle actomyosin has been studied with the aid of isotopic tracers. The hydrolysis of ATP to adenosinediphosphate (ADP) and inorganic phosphate (Pi) has the stoichiometry of a substitution reaction; i.e., BX + Y + BY + X. This type of reaction probably occurs through a displacement mechanism in which the nucleophilic reagent, Y, makes an attack on group B (3). The first fact to be determined, therefore, was the point of hydrolytic cleavage. This was done by performing the reaction in HzOIS. Such an experiment would decide (CL) whether the ADP or the phosphate moiety was the B group of the substitution reaction and (b) whether the purified protein gave the same cleavage point as intact lobster muscle slices (4). Having identified the B part of the molecule, we performed exchange experiments with 018-labeled HzO, P32-labeled ADP, and P32-labeled phosphate to establish the existence of a B-enzyme intermediate or at least establish the kinetic limits on the lifetime of such an intermediate. EXPERIMENTAL Material and Analyses-Actomyosin was separated from rabbit muscle following the procedure of Szent-Gyorgyi (5) and then redissolved and reprecipitated five to seven times to remove traces of myokinase. Myosin was prepared by a modification of the procedures (6) of Mommaerts and Parrish and Portzehl, Schramm, and Weber and was also redissolved and reprecipitated until no detectable amount of myokinase activity was present. Myokinase activity was assayed by the deaminase method (7) or by measuring the decrease in labile phosphate on reaction with hexokinase and glucose. * A preliminary account of this work has been published (1). Research carried out at Brookhaven National Laboratory under the auspices of the United States Atomic Energy Commission. 279

2 280 MECHANISM OF HYDROLYSIS OF ATP The ATP (sodium salt) was obtained commercially (Schwarz Laboratories, Inc.) and was assayed chromatographically. A typical analysis was 1.66 moles of ATP, 0.15 mole of ADP, and 0.06 mole of Pi per mg. of material. Inorganic phosphate analyses were performed by the procedure of Fiske and Subbarow (8) or by the isobutanol extraction method (4). Adenosinediphosphate labeled with P32 was obtained by incubating adenylic acid, P3Vabeled inorganic phosphate, and oc-ketoglutarate with a suspension of rat liver mitochondria according to a procedure kindly supplied by Dr. S. Kaufman and Dr. S. Ochoa. The labeled ADP was separated chromatographically and rechromatographed until it contained less than 0.01 per cent ATP32. co.oo3r; ObIN HtI&O.IN ( I ADP 3 I ELUTRIANT VOLUME (ml) FIG. 1. Elution of ADP and ATP from Dowex 1 column The H2018 was obtained from the Stuart Oxygen Company and had an atom per cent excess of about 1.4. The O * content of the phosphate samples was obtained by a slight modification of the procedure of Cohn (9). The chromatographic procedure of Cohn and Carter (10) was modified to facilitate the rechromatographing of the adenosinephosphates. The ADP elution was performed with 0.01 M HCl and the ATP elution with 0.1 M HCI. This decreased the total salt concentration and also increased the volume necessary to obtain elution. Satisfactory separations were obtained as shown by a typical curve in Fig. 1. Hydrolysis of ATP in H2O1s in Presence of KH9320,A solution of H2018 (90 ml.) was prepared containing 0.10 M KCl, M CaC12, M tris(hydroxymethyl)aminomethane buffer, essentially carrier-free KH2P3204, and 400 mg. of solid sodium ATP. After adding 10 ml. of the purified actomyosin solution in 0.1 M KCl, and adjusting the ph to 9.0, the solution was kept at 19 for approximately 1 hour. A second addition of enzyme was made and the solution was kept at 19 for another hour until about 80 per cent of the ATP had been hydrolyzed. Alkali was added

3 KOSHLrZND, RUD~~NSTEIN, AND KOWALSKY 281 during the reaction to keep the ph at about 9.0. Trichloroacetic acid was added to stop the enzyme action and the protein was removed by centrifugation. The phosphate compounds were then precipitated as the barium salts and dissolved in HCl, the barium was removed, and the anions were separated chromatographically. Aliquots for determination of the 018 content of the medium were removed just after addition of the first enzyme solution and just before addition of the trichloroacetic acid solution. Aliquots for Pi and for P32 were taken after addition of the trichloroacetic acid and after addition of the enzyme solution, respectively. The labile phosphate in the ADP and ATP fractions was obtained by hydrolysis in hydrochloric acid, and then by adding a known amount of carrier KHzP04, precipitating the phosphorus as barium phosphate, and analyzing the solution of the precipitate as described above. Control experiments showed that no exchange of oxygen between the medium and the inorganic phosphate occurred during the acid hydrolysis. Another run was performed in a similar manner, except that the nucleotides were separated on charcoal according to the procedure of Crane and Lipmann (11). Exchange of Oxygen between KHZ04 and H ml. of the myosin (or in some cases actomyosin) solution were added to 45 ml. of an H2018 solution containing salts in such amounts that the final concentrations were 0.1 M KCl, M CaC12, and 0.02 M KH%POJ. The ph was adjusted to 8.8 with solid potassium hydroxide and the mixture was kept at 20 for 2.75 hours. The protein was removed by centrifugation and the phosphate was precipitated as barium phosphate. The barium phosphate was converted to KHzPOd and analyzed for 018 as described above. Exchange of PQ-Labeled ADP with ATP-To 7.0 ml. of a solution containing 0.32 pmole of purified ADP32 were added 9.3 mg. of sodium ATP, 44 mg. of KCl, 11 mg. of glycine, 0.75 ml. of 0.09 M CaCl?, and solid KOH to adjust the ph to 9.0. After addition of the myosin solution and dilution to 15 ml., the mixture was incubated at 20 for 20 minutes. The reaction was stopped by immersing the vessel in boiling water for 90 seconds and then cooling it rapidly in ice water. After removal of an aliquot for analysis of inorganic phosphate, the denatured myosin was removed by filtration through glass wool. The solution was then diluted to 100 ml. and made 0.5 M with respect to PUTHkOH, an aliquot was removed for counting the P32 activity, and the solution was added to the ion exchange column. After separation of the fractions, the central portions of the ATP fraction were combined and diluted to 100 ml., made 0.5 M with respect to NH40H, and rechromatographed either with or without addition of inactive ATP and ADP carriers. The control runs were similarly performed on the same ADP32 and

4 282 MECHANISM OF HYDROLYSIS OF ATP enzyme preparations. The only difference was that the enzyme in the control had been inactivated by heating for I.75 hours at 38 at ph 9 before addition to the substrate solution. Results The results of the experiment with 018 to determine the cleavage point are summarized in Table I. The inorganic phosphate produced contained an atom per cent excess equal to one-quarter that of the medium. This is clearly consistent with cleavage between the terminal phosphorus atom TABLE 0x8 Content of Compounds Isolated after Partial Hydrolysis of ATP in Hz018 Catalyzed by Actomyosin Conditions, 20, 0.1 M KCI, M CM&, 0.03 M tris(hydroxymethyl)aminomethane, ph 9. Experiment No. hydrolysis Per cent Atom per cent excess 0 8 per atom of oxygen Medium Inorganic phosphate I ADP ATP Separation method Dowex column Charcoal batch ATP-4t H20 - Ad-O-b-O-h-0 p p H i - Ad-O-~-O-~-OH+H0 8-~-0- (I) 6 il ;; O * H and its bridge oxygen (Equation 1). The single Ols atom from the medium introduced into the inorganic phosphate is diluted isotopically by the other unlabeled oxygens because of the symmetrical nature of the oxygens in the free phosphate. The deviations from the precise theoretical value are in accord with the experimental error observed by Cohn (9) and in our own laboratory and do not indicate any alternative pathway of hydrolysis. This result is further confirmed by the failure to observe any 01* in the ADP produced. The very small positive values observed are within the experimental error of the method and most probably represent small organic impurities which follow through to the final phosphate in spite of the purification steps. It should be emphasized that the quantities observed are 2 to 4 per cent of the natural abundance of O18. A trace of organic impurity which splits in the mass spectrometer to give a fragment of mass 46 can easily give a false positive value of this order of magnitude.

5 KOSHLAND, BUDENSTEIN, AND KOWALSKY 283 The 01* content of the unhydrolyzed ATP was examined to determine whether a stable free addition intermediate (12) of the type (I) is formed during the hydrolysis. No significant amount of 01* was found in the unhydrolyzed ester, the small positive values undoubtedly again representing traces of organic impurities. Ad-O-P-O-P-O-,~-O- y- y- y- 6 b It is noted that the results obtained by the charcoal elution method of Crane and Lipmann, in which all labile phosphate groups of both ATP and ADP are analyzed together, confirm those obtained by the Dowex 1 chromatographic separation. In Table II are summarized the results of experiments with P32 in which the exchange of KH2P3204 with ATP was studied. The amount of P32 activity in the ATP was essentially similar and extremely low in both the experiment with active enzyme and the control run in which no enzyme was added. The small amount, therefore, probably represents the limit of chromatographic separation. In Table III are summarized the results of the exchange of oxygen between KHlPOd and H20. Both myosin and actomyosin were used and neither showed an appreciable amount of exchange. These results are in harmony with the observed agreement with theory for the O * content of the inorganic phosphate produced in the hydrolysis, since any extensive exchange reaction would have increased this value markedly. These data are, therefore, in marked contrast to those obtained with intact lobster muscle slices (4). The results of the study of the exchange of ADP32 with ATP during the myosin-catalyzed hydrolysis are shown in Table IV. Similar data were obtained with actomyosin. The amount of activity in the ATP was found to be essentially the same in both the control experiments and those with active enzyme. The activity in the ATP was not caused by myokinase, since the enzyme preparation at a concentration 50 times that used in the exchange experiments showed no detectable myokinase action. It was also not caused by ATP32 present as an impurity in the original ADP32, for chromatographic analysis of the ADP32 preparation showed less than 0.01 per cent ATP32. Nor was it caused by some ADP32 impurity carried into the ATPs2 fraction on the column during separation after the exchange experiment, since rechromatographing of the ATP fractions on which the calculation was based showed less than a 10 per cent decrease in specific activity. Ii p O B-H (I)

6 284 MECHANISM OF HYDROLYSIS OF ATP The probable cause of this activity in the ATP fraction is the inosinetriphosphate-adenosinediphosphate transphosphorylase of Berg and Jok- TABLE Pa2 Content oj Compounds Isolated after Partial Hydrolysis of ATP Catalyzed by Actomyosin in Presence of KHzPa204 Conditions, 20, 0.1 M KCI, M CaC12, 0.03 M tris(hydroxymethyl)aminomethane, ph 9. II Experiment Per cent Total counts KHzP=Or hydrolysis added, c.p.m. Per cent of P32 activity in unhydrolyzed ATP Enzyme run... Control TABLE / III 4.4 x x Content of KHzPOa after Incubation with Myosin and Actomyosin in Hz018 Conditions, 20, 2$ hours, 0.1 M KCl, M CaClz. Enzyme Actomyosin.... Myosin... PH 0 8 atom per cent excess (,f fractions Medium K&PO. ATP hydrolyzed in 1 hr. by same amount of enzyme ~moles / TABLE IV 1 I Paz Content of ATP Fractions after Partial Hydrolysis of ATP Catalyzed by Myosin in Presence of ADPa Conditions, 20, 0.1 M KCl, M CaCh, 0.01 M glycine, 20 minutes at ph 9. Experiment state of enzyme Total Pa* activity added as ADPa Per cent hy,;irio&is Per cent of total Paz activity in ATP c.p.m. 1 Active 427,000 Control Inactive* 2,120,000 2 Active 680,000 Control Inactive* 834, * Same amount of enzyme added in control run as in experiment immediately preceding it, but the enzyme had been 98 per cent inactivated by heating for 2 hours at 38 at ph 9. lik (13). This enzyme has been shown to be present in rabbit muscle preparations, to have specificity requirements which allow uridine tri-

7 KOSHLAND, BUDENSTEIN, AND KOWALSKY 285 phosphate and inosinetriphosphate as substrates, and to be not inactivated under the mild conditions used to deactivate the myosin. It seems very likely, therefore, that it would catalyze the reaction of ADP32 with ATP to give ATP32 and ADP. Whatever the source of the 1.3 per cent P32 activity in the ATP, it is clear that it is not myosin. The actual limit of the activity in the ATP which could have been caused by it must be less than 0.1 per cent. DISCUSSION The experiments with O * show that actomyosin catalyzes cleavage between the terminal phosphorus atom and its oxygen, and it seems reasonable to assume that this occurs via a displacement on the terminal phosphorus atom. Displacement mechanisms have been or can be used successfully to explain the available facts in a number of enzyme systems, e.g. the action of inhibitors (14), stereochemistry (15, 16), correlation of FIG. 2. A single displacement mechanism in which water makes a direct primary attack on the ATP. specificity and bond cleavage (17), the failure of medium O@ to enter the kinase mechanisms (9), and the similarity of the hydrolases and transferases (3). The present study adds further evidence, since the theoretical amount of O * in the phosphate and the absence of O * from the ADP show that only one bond is broken in the ATP molecule as would be required by a displacement mechanism. Assuming that the enzyme action occurs by a displacement, two alternatives must be considered: (a) a single displacement mechanism (16) in which water makes a direct primary attack on the ATP (Fig. 2) or (b) a double displacement mechanism (16) in which the primary attack is made by the enzyme to form a phosphorylated enzyme intermediate (Equations 2 and 3). EH + ATP-4 & EPOz- + ADP-3 + H+ k-1 (2) EPOS- + Hz0 cf EH + HOPOzb (3)

8 286 MECHANISM OF HYDROLYSIS OF ATP It is clear that the single displacement mechanism of the Walden inversion type is compatible with all of the evidence presented here. No exchange of H201*, ADP32, or KH2P3204 would be expected, nor would any 01* be presented in the unhydrolyzed ATP. No covalent bond is formed with the enzyme as a result of the bond cleavage in this mechanism; the enzyme acts entirely as a catalyst providing acidic and basic groups steritally arranged to give efficient catalysis. Thus, if this is indeed the mechanism and the same mechanism is involved in the muscular contraction, the energy is transferred from ATP to enzyme without the intervention of covalent phosphoryl bonds. The double displacement mechanism is also consistent with the data, but only if certain limitations on the rates of the various steps are imposed. Thus no rapid reversible formation of a stable phosphorylated intermediate is possible or exchange with ADP32 would have been observed. Moreover, the failure to observe exchange with ADP32 cannot be ascribed to a low energy character of the phosphorylated intermediate, since, if this were so, exchange of H2018 with KHzPO, would have been measurable. If this mechanism holds, k2 (H20) must be much greater than k-1 (ADP) ; i.e., the phosphorylated enzyme intermediate is immediately hydrolyzed and is of a transient nature. The implications of the present findings in relation to muscular contraction depend in large part on how closely the purified proteins resemble the intact muscle. The extensive literature on the subject indicates many similarities and some notable differences (2). The present work establishes the similarity in the position of P-O cleavage between the purified actomyosin and the intact lobster muscle slices. The transient nature of the intermediate in the hydrolysis by the purified proteins is consistent with two major alternatives: (1) the intermediate between ATP and the muscle protein during contraction is likewise transient and (2) the ATPase action of the purified enzyme is fundamentally different from that of the intact fiber. Until further work has been completed, it appears unprofitable to speculate at length on the consequences of these alternatives, but they bear directly on the efficiency and process of energy transfer and the validity of extrapolations from purified protein to the intact system. SUMMARY 1. Adenosinetriphosphate was partially hydrolyzed in HzO * in the presence of purified rabbit muscle actomyosin. 1 atom of oxygen was introduced into each molecule of phosphate produced and no detectable Ols was found in either the ADP or unhydrolyzed ATP. 2. No exchange of KHzP3204 with ATP, of Hz018 with KHtPOI, H201S with ATP, or ADP32 with ATP was observed to be catalyzed by the muscle protein.

9 KOSHLAND, BUDENSTEIN, AND KOWALSKY The results are consistent with a displacement mechanism in which nucleophilic attack occurs on the terminal phosphorus atom and in which only a transient intermediate is formed. The implications for muscle action are discussed. Addendum-J. Gergely and W. P. Jencks have kindly informed the authors that in similar experiments with actomyosin under conditions which either activate or inhibit the ATPase action no exchange of ADPa and ATP was observed. BIBLIOGRAPHY 1. Koshland, D. E., Jr., and Budenstein, Z., Federation Proc., 13, 245 (1954). 2. Szent-Gyorgyi, A., Chemical physiology of contraction in body and heart muscle, New York (1953). Needham, D. M., Advances in Enzymol., 13, 151 (1952). Mommaerts, W. F. H. M., Muscular contraction; a topic in molecular physiology, New York (1950). Weber, H. H., and Portzehl, H., Advances in Protein Chem., 7, 161 (1952). 3. Koshland, D. E., Jr., in McElroy, W. D., and Glass, B., Mechanism of enzyme action, Baltimore, 608 (1954). 4. Koshland, D. E., Jr., and Clarke, E., J. Biol. Chem., 206, 917 (1953). 5. Szent-Gyijrgyi, A., Chemistry of muscular contraction, New York, 2nd edition, 151 (1951). 6. Mommaerts, W. F. H. M., and Parrish, R. G., J. Biol. Chem., 188, 545 (1951). Portzehl, H., Schramm, G., and Weber, H., 2. Naturforsch., 66, 61 (1950). 7. Kalckar, H. M., J. Biol. Chem., 167, 445 (1947). 8. Fiske, C. H., and Subbarow, Y., J. Biol. Chem., 66, 375 (1925). 9. Cohn, M., J. Biol. Chem., 201, 735 (1953). 10. Cohn, W. E., and Carter, C. E., J. Am. Chem. Sot., 72,4273 (1950). 11. Crane, R. K., and Lipmann, F., J. Biol. Chem., 201, 235 (1953). 12. Stein, S. S., and Koshland, D. E., Jr., Arch. Biochem. and Biophys., 39,229 (1952) ; 46, 467 (1953). 13. Berg, P., and Joklik, W. K., Federation Proc., 13, 182 (1954); Nature, 172, 1008 (1953). 14. Nachmansohn, D., and Wilson, I. B., Advances in Enzymol., 12, 259 (1951). Wilson, I. B., J. Biol. Chem., 190, 111 (1951). 15. Fitting, C., and Doudoroff, M., J. Biol. Chem., 199,153 (1952). 16. Noshland, D. E., Jr., Biol. Rev. Cambridge Phil. Xoc., 28, 416 (1953). 17. Koshland, D. E., Jr., and Stein, S. S., J. BioZ. Chem., 208, 139 (1954).

10 MECHANISM OF HYDROLYSIS OF ADENOSINETRIPHOSPHATE CATALYZED BY PURIFIED MUSCLE PROTEINS D. E. Koshland, Jr., Zelda Budenstein and Arthur Kowalsky J. Biol. Chem. 1954, 211: Access the most updated version of this article at Alerts: When this article is cited When a correction for this article is posted Click here to choose from all of JBC's alerts This article cites 0 references, 0 of which can be accessed free at ml#ref-list-1