Labeled in Thecourse of Glycolysis, Since Phosphoglycerate Kinase

THE STATE OF MAGNESIUM IN CELLS AS ESTIMATED FROM THE ADENYLATE KINASE EQUILIBRIUM* BY TRWIN A. RoSE

THE INSTITUTE FOR CANCER RESEARCH, PHILADELPHIA Communicated by Thomas F. Anderson, August 30, 1968 Magnesium functions in many enzymatic reactions as a cofactor and in complex with nucleotides acting as substrates. Numerous examples of a possible regulatory role of Mg can be cited from studies with isolated enzymes,'- and it is known that Mg affects the structural integrity of macromolecules such as trans- fer RNA" and functional elements such as ribosomes.'0 The major problem in translating this information on isolated preparations to the functioning cell is the difficulty in determining the distribution of Mg and the nucleotides among the free and complexed forms that function in the region of the cell for which this information is desired. Nanningall based an attempt to calculate the free Mg2+ and Ca2+ ion concentrations of frog muscle on the total content of these metals and of the principal known ligands (adenosine 5'-triphosphate (ATP), creatine-P, and myosin) and the dissociation constants of the complexes. However, this method suffers from the necessity of evaluating the contribution of all ligands as well as from the assumption that all the known ligands are contributing their full complexing capacity. During studies concerned with the control of glycolysis in red cells and the control of the phosphoglycerate kinase step in particular, it became important to determine the fractions of the cell's ATP and adenosine 5'-diphosphate (ADP) that were present as Mg complexes. Just as the problem of determining the distribution of protonated and dissociated forms of an acid can be solved from a knowledge of pH and pKa of the acid, so it would be possible to determine the liganded and free forms of all rapidly established Mg complexes from a knowledge of Mg2+ ion concentration and the appropriate dissociation constants. An estimate of (Mg2+) could be made from the position of equilibrium of a reaction in the cell if the equilibrium value were strongly Mg-dependent. To be useful in this regard such a reaction must be at thermodynamic equilibrium in the cell and must have all reactants in suitably measurable amounts. Preferably, the position of equilibrium should be insensitive to pH in the physiological range, since the pH within the compartment of the cell containing the components of the reaction will generally be unknown. The present study concerns the use of the adenylate kinase equilibrium for this purpose and the application of this method to the intact red cell. Theory.-When 32Pi is incubated with glycolyzing human erythrocytes, the isotope is found to be incorporated into the 9-phosphoryl position of ATP at about the same time as it is into the 7-position.12' 13 The 7-position is first to be labeled in the course of glycolysis, since phosphoglycerate kinase reacts with ADP and not with adenosine 5'-phosphate AMP.14 Thus randomization of radio- activity requires two cycles of the adenylate kinase reaction: ATP7* + AMP -- ADP + ADP,-* AMP + ATPP*, which must be very rapid relative to the initial labeling. Since, in the cell, the rate of ATP generation through glycolysis 1079 Downloaded by guest on September 27, 2021 1080 BIOCHEMISTRY: I. A. ROSE PRoc. N. A. S.

TABLE 1. Dissociation constants for Mg2+, H+, and K+ complexes of AMP, ADP, and A TP at 350 and 0.1 ionic strength. Dissociation Dissociation reaction constant Reference (AMP)H- * (AMP)2- + H+ 10^ 4 M 16 (AMP)Mg (AMP)2- + Mg2+ 10 mM 17 (ADP)H2- (ADP)3- + H+ 106.8 M 18 (ADP)Mg- (ADP)3- + Mg2+ 0.30 mM 18 (ADP)HMg (ADP)H2- + Mg2+ 7.6 mM 18 (ADP)K2- (ADP)3- + K+ 200 mM 19 (ATP)H3- (ATP)4- + H+ 10-7.05 M 18 (ATP)Mg2- (ATP)4- + Mg2+ 0.019 mM 18 (ATP)HMg-, (ATP)H3- + Mg2+ 1.35 mM 18 (ATP)K3- (ATP)4- + K+ 100 mM 19 is sufficient to maintain the ATP level, the adenylate kinase step must also be much more rapid than the degradative steps such as ATPase and adenylate deaminase. Hence, the adenylate kinase reaction must be close to equilibrium. Bowen and Kerwin15 demonstrated that the position of equilibrium of the adenylate kinase reaction is a biphasic function of magnesium concentration. At concentrations of Mg very low compared with the nucleotide concentrations, the value of [(2AMP) (ZATP) ]/(2ADP)2 = 0.37 was obtained when the reaction was started in either direction at pH 7.5, , = 0.02. (The temperature used in these determinations was unstated.) The observed equilibrium value first increased with added Mg as the ATP was preferentially titrated and then, due to the squared term in the denominator, the value decreased as the ADP became complexed. Whereas this behavior of Keq when related to total Mg (as done by Bowen and Kerwin) is a function of the adenylate concentration, it is possible to obtain an expression for the equilibrium constant that is independent of the total amounts of both the nucleotides and Mg and which depends only on the concentration of free MgS+ ion. Define the observed mass action ratio K' in terms of the significant free, protonated, and complexed forms of AMP (M), ADP (D), and ATP (T): (M2- + MH + MMg)- (T4- + TMg2- + TH3- + THMg + TK3-) K =. (D3- + DMg- + DH2- + DHMg + DK2-)2 (1) Using the values of the dissociation constants at 350 and , = 0.1 that are listed in Table 1, one obtains equation (2), which relates K' to the millimolar concentrations of K+ and Mg2+, x. (M2-) (T4-) (D3-)2

(1 + 10.4pH + +1 0.019 + 107l6pH 1 + 13) + (K+)1 -L0 + 0O8019 1+.35)+~j 100 (2) [1 + 0 3 + 106,8-pH (1 + 6 + (0) 7e0.30 2001 From the zero-Mg experimental value of Bowen and Kerwin, at pH 7.5 and K+ Downloaded by guest on September 27, 2021 VOL. 61, 1968 BIOCHEMISTRY: 1. A. ROSE 1081

= 20 mM, one can calculate, using equation (2), the pH- and K+-independent constant [(M2-) (T4-) ]/(D3-)2 which, due to compensating effects in the secon- dary pKa's of the nucleotides, remains at 0.37. Figure 1 presents the computed values K' as a function of (Mg2+) in the pH range 6.5-8.0. It will be noted that K' increases most sensitively with (Mg2+) in the range 0-0.25 mM, and that it becomes insensitive to (Mg2+) in a peak region the width of which is strongly influenced by pH. At higher (Mg2+), K' falls slowly to a limiting value of 0.17. Methods.-The nucleotides were determined by standard methods:20 ATP by yeast hexokinase and glucose-6-P dehydrogenase; ADP by pyruvate kinase and lactate dehydrogenase; and AMP by prior addition of ATP and adenylate kinase, acidification, and determination of the two equivalents of ADP formed thereby. For the determination of ATP in the presence of deoxyglucose, a tenfold excess of glucose was added to mini- mize the side reaction of ATP with deoxyglucose. At these high levels of glucose a small correction for its direct oxidation by glucose-6-P dehydrogenase was necessary. In the assay of AMP, cognizance was taken of the presence of AMP in commercial preparations of reduced diphosphopyridine nucleotide (DPNH).20 Failure to correct for this is re- sponsible for the high values previously determined21 for the adenylate kinase equilibrium. The DPNH used in these assays was purified by chromatography22 to reduce the level of AMP, although a more satisfactory method exists.21 Glycerate-2,3-diP was determined as glycerate-2-P plus glycerate-3-P after treatment with glycerate-P mutaseY20 Total Mg was kindly determined by Dr. Stuart Blum by atomic absorption. Human red cells were prepared from normal donors just prior to study.2' Experimertal.-Estimate of (Mg2+) in frozen-thawed preparations of red cells: Since human erythrocytes are known to be impermeable to the entry of Mg2+ ion,24' 25 studies were begun with disrupted preparations. One may add MgCl2 or ethylenediaminetetraacetate (EDTA) and determine the effect on the equilibrium distribution of the adenylate kinase reactants. In this way it would be possible to determine where in the K' versus (Mg2+) function the (Mg2+) of the initial hemolysate was poised. This was done with a water hemolysate (2 vol of water/vol of cells), prepared by successive freezings and thawings, that had been reconstituted by lyophilization to the original cell volume, and to which

.7-

FIG. 1.-Effect of (Mg2+) on the 2 3 4 5 6 mass action ratio K' of adenylate Mg. mm kinase according to equation (2), .1 .2 .3 .4 .5 .6 .7 .8 .9 1.0 T = 35°, , = 0.1. Mg , mM Downloaded by guest on September 27, 2021 1082 BIOCHEMISTRI': I. A. ROSE PROC. N. A. S.

AMP was added to provide greater amounts of AMP and ADP in the equilibrium for increased accuracy and freedom from artifact (see below). The results of such an experiment are given in Table 2. The value for K' of 0.77, observed for the earliest sample, could be attributed to (Mg2+) at either 134 or 1770'Al[ (Fig. 1, pH 7.2). A sample taken five minutes later gave a higher value of K'. A reasonable basis for this increased K', assuming the lower value of (Mg2+) to be the real one, would be the observed loss of 0.45 Mmole of ATP/ml due to an ATPase, thus making an almost equivalent amount of Mg available for complex formation. Part of this increased Mg2+ would react with ADP3- and displace the equilibrium to the right and to higher values of K'. For the 1770 s1AI value of (Mg2+) to be correct, one would have to postulate a decrease in (Mg2+) to explain the increase in K' and there is no a priori reason for doing so. Moreover, the higher values of uncomplexed Mg are inconsistent with reports that the total Mg of red cells is only about 2 Imoles/ml of cells.25 In experimental support of the lower value of (Mg2+) are: (1) the effect of added MgCl2, in which it is clear that K' passes through a maximum as (Mg2+) increases, and (2) the effect of EDTA which lowers K' as expected. This low value of K' in the presence of EDTA and its relative constancy with time indicates that even at very low (Mg2+), and hence low concentration of ATPMg2- and ADPMg-, the adenylate kinase equilibrium becomes established rapidly from the starting condition of (AMP) > 400 m~zmoles/ml of cell extract. If these data reflect the status of the intact cell, the level of free Mg2+ under normal glycolytic conditions should be less than 134 MM, since ATP is usually greater than 1 /mole/ml of cells. Mg2+ concentration in the whole cell: Since two concentrations of Mg2+ cor- respond to most values of the mass action ratio of adenylate kinase, a general method for deciding the correct range of (Mg2+) in the intact cell is required. The method devised makes use of the fact that when the principal ligand, ATP, is converted to weaker ligands, ADP and AMP, (Mg2+) must increase. A conse- quent rise or fall in K' would then characterize the low and high range of (Mg2+), respectively. The phosphorylation of deoxyglucose by hexokinase within the cell produces deoxyglucose-6-P, which is not an ATP-generating source. If cells

TABLE 2. Free Mg2+ concentration in the red cell extract. Additions Time AMP ADP ATP 2* Apt (Mg2 +) (mjsmoles) (min) (mtmoles/ml of extract) K' (MM) 2 116 325 700 1141 1725 0.77 134 - 7 430 350 254 1034 858 0.89 210 MgCl2 (200) 7 440 321 241 1002 803 1.03 400-600§ MgCl2 (1500) 7 457 360 248 1065 856 0.87 1270 EDTA (2000) 7 42 251 680 973 1611 0.45 20 EDTA (2000) 12 52 266 650 968 1566 0.47 30 Washed cells were hemolyzed by being frozen and thawed in 3 vol of water. The frozen hemolyzate was concentrated by lyophilization to the original cellular volume. One ml of the cold solution, pH 7.2, was added to tubes (at 350) containing the noted additions and 0.4 Emole of AMP in 0.1 ml. TCA was added at the noted times and extracts made for the analysis of the adenylates. * Z = AMP + ADP + ATP. t UP = ADP + 2-ATP. t From Fig. 1, pH 7.2. § The value K' = 1.03 is within experimental error of the peak corresponding to this range of (Mg2 +). Downloaded by guest on September 27, 2021 VOL. 61,p 1968 BIOCHEMISTRY: I. A. ROSE 1083 are briefly pretreated with lodoacetate to impair glycolysis, then deoxyglucose will cause a decrease in ATP with time. In fact, cells taken at intervals after addition of deoxyglucose have diminishing ATP, as seen in Table 3. The mass action ratio K' is observed to increase from 0.75 to 1.02, corresponding to a change from 95 MAI to 247 MAM Mg2+, assuming an intracellular pH of about pH 7.5. The alternative, a change from 2100,MM to 877 MM Mg2+ is excluded on the basis of both the direction of change and the fact that the total Mg found in the acid extract was only 1300 mumoles/ml cells. The assumption made in this calculation is that the measured values of the nucleotides in the acid extract represent their true relative concentrations in the compartment of the cell in which adeny- TABLE 3. Change in (Mg2+) with decrease of ATP of the cell. Time AMP ADP ATP (Mg2+) DMg- T4- TMg2- (min) (miumoles/ml cells) K' (AM) D3- (miumoles/ml cells) 2;Mg 0 68 285 890 0.75 95 142 45 121 605 712 5 142 390 820 0.765 100 192 63 107 567 695 15 266 450 610 0.80 114 216 82 73 437 593 25 385 472 492 0.85 132 220 97 53 368 551 55 510 350 245 1.02 247 137 114 16 207 481 Washed red blood cells were incubated in isotonic NaCl with 20 mM iodoacetate for 5 min at 370 in order to prevent glycolysis at the glyceraldehyde-P dehydrogenase step. The cells were spun down, washed with saline, and incubated as a 30% suspension in a medium of 0.043 M NaPO4, pH 7.8, 0.085 M NaCi, 4 mM KC1, and 5 mM 2-deoxyglucose. After 5 min of temperature equilibration, samples were taken at the noted times for preparation of TCA extracts for assay of total Mg, the adenylates, and glycerate-2,3-diP. The variation in the sum of the adenylates was due to incom- plete recovery. Recovery was less important than speed in manipulation. Procedures used to obtain derived values were as follows: (Mg2+) from Fig. 1, pH 7.5. D3- = ADP/[1.7 + (x/0.3)], where, from equation (2), 1.7 = 1 + 1066-sD + [(K+)/200], DMg- = (x/0.3)*D3-, T4- = ATP/ [2.36 + (x/O.019) ], where, from equation (2), 2.36 = 1 + 107.05-P + [(K+)/100], TMg2 - = (x/0.019) - T4-, and 2;Mg = TMg2- + DMg- + 0.65 (Mg2 +), with the value 0.65 taken as the fraction of cell volume that is water. late kinase operates. Since oxygenated hemoglobin does not bind ATP,26 and since the concentration of calcium in red cells is kept small by a physiological pump, the absence of mitochondria, nucleus, and endoplasmic reticula makes this assumption appear reasonable. From (Mg2+) one can calculate the amounts of ATPMg2- and ADPMg-, as shown in Table 3. It is evident that at early times about 70 per cent of the ATP and 15 per cent of the ADP are in the complex form. The sumh of the Mg that is bound to nucleotides and that which is free is only about half the amount in the cells, as determined in the extract. To account for the remaining 600-800 m~umoles of Mg, one can consider glycerate-2,3-diP and hemoglobin as possible ligands. The sample taken at zero minutes was found to contain glycerate-2,3- diP at 6.2 Amoles/ml cells, and this would remain almost constant during the incubation. The dissociation constant of the (glycerate-2,3-diP) Mg complex was determined in this laboratory by the resin competition method of Schubert27 at pH 7.5 with 0.1 M KCl present. This value, 0.9 mM, indicates that glycerate- 2,3-diP is a weak buffer for Mg2+. The large pool of this compound can account for about 600 m&moles of Mg/ml of cells at zero minutes and significantly more Downloaded by guest on September 27, 2021 1084 BIOCHEMISTRY: I. A. ROSE Peoc. N. A. S.

as the (Mg2+) rises with time. Hemoglobin was shown by equilibrium dialysis to bind 24Mg very weakly, and would not be of importance as a Mg buffer. Discussion.-The adenylate kinase equilibrium is symmetrical in form so that the ratio of products to substrates is a number without the dimensions of molar- ity. Thus, no assumption need be made concerning the volume that the components of the system occupy in the cell. On the other hand, the degree of dis- placement of the equilibrium is a function of the concentration of free Mg2+, which is the value obtained from the calculation. This cannot be converted into an amount without assuming a value for the volume of the system repre- sented by the equilibrium. However, it is the concentration (or preferably the activity) of Mg2+ that one is concerned with in matters of regulation and in the determination of liganded forms. The adenylate kinase equilibrium constant, unlike that of many other kinases, is close enough to unity to allow all components of the equilibrium to be in the measurable range, at least under certain metabolic conditions. In this respect, it is particularly valuable that the concentration of ATP is readily modified by metabolic conditions, thus allowing the system to be altered to achieve favorable concentrations for measurement. Finally, the flexibility of the system provides a test of the calculation insofar as ATP is a major ligand of Mg2+ of the cell. Thus, the adenylate kinase system is not only a test system, but also an important factor in establishing Mg2+ concentration. Unfortunately, there are few other favorable reactions or reaction se- quences that come to mind as alternatives to the adenylate kinase reaction for a measurement of (Mg2+).t The creatine kinase reaction, which is a rapid and widely distributed one, has two drawbacks: one of the reactants, creatine, is readily diffusible, making it likely that a portion of it may be extracellular. More serious, however, the creatine kinase equilibrium is such that the uncertainty in the concentration of Mg2+ is equal to the uncertainty in H+ concentration over the entire physiological range. Thus, even at pH's above the pKa's of ADP and ATP:

K (creatine) (ATP) (Cr') (T4-) + 0.019 100 (creatine-P) (ADP) (CrP2-) (D3-) +(x 0.3+ 200

since (C± Tj 1081M-2 and, (CrP2-) (D3-) (H+) 2 + 52.5x thellthen K' = 1.5 + 3.33 .1081817p17-pH, at K+ = 100 mAl. (3) Although of dubious value alone in establishing (Mg2+), the mass action ratio for creatine kinase must increase with increasing Mg2+, especially in the region 0-0.6 mM, and so might provide additional evidence about the direction of (Mg2+) changes that can be applied to the adenylate kiniase data. It may be of interest to note that, when (Mg2+) is known, the K' of creatine kinase could be used to estimate pH, according to equation (3). Downloaded by guest on September 27, 2021 VOL. 61X 1968 BIOCHEMISTRY: I. A. ROSE 1085

The literature abounds with data that may be examined in this way. The major reason for not doing so earlier is a tendency to regard the problem as too complex, since it seemed that the occurrence of multiple cellular compartments, each with its complement of nucleotides and cation, would make the interpreta- tion of data based on the total cell content of nucleotides unclear.29 Adenylate kinase is known to be present on mitochondria in a position accessible to the extramitochondrial nucleotides30 and on microsomes.31 The adenylates within the mitochondria, however, may not be equilibrated by the kinase32 and where they represent a significant proportion of the whole cell content may displace the observed ratio from equilibrium. If Mg2+ is equilibrated among kinase- containing compartments, the compartments will have the same mass action ratio K' for adenylate kinase, even if the nucleotides do not equilibrate. Since most of the adenylates of the cell are probably in the cell sap, it is expected that the observed mass action ratio will reflect the Mg2+ concentration of that region of the cell. In addition to the problem of separate compartments, one must consider the error that results if some of the nucleotides occur in a form not considered in equation (2), i.e. bound to membrane or proteins. Since the contribution of such forms may be quite small, it should be possible to limit this source of error by choosing conditions in which the amount of each nucleotide is greater than 50-100 mlimoles/gm wet tissue. AMP is often the least abundant of the three nucleotides. Clearly it would be unreasonable to consider data for this calculation from a tissue in which AMP was not greatly in excess of phosphorylase b, for example, unless a suitable adjustment was made in equation (2). Of course, the formation of AMP from such compounds as AMP-acetate, AMP-amino acids, and AMP-proteins that would hydrolyze during the preparation of the protein-free extract will be a problem when total AMP is low. The fact that the mass action ratios observed seldom are above the maximum theoretical range is reassuring. Independent of whether a precise measure of (Mg2+) can be obtained by the present method, it should be useful in determining the range of the Mg2+ ion concentration. The range has significance in establishing the approximate distribution of ADP and ATP among the free and complexed forms. This information is important for understanding the control of enzymes such as hexokinase2 3 and fumarase7 for which ATP4- is inhibitory and ATPMg2- is not, and phos- phofructokinase,5 which is tenfold more sensitive to inhibition by ATP4- than by ATP\'Ig2-. The kinases that use ADP for the generation of ATP, such as glycerate-P kinase33 and pyruvate kinase,34 behave as if ADPMg- were the sub- strate. Cells with (Mg2+) in the lower range, such as the red cell, will have only a small part of their total ADP in the active form and hence these reactions may be limited in rate by (ADPMg-). The concentration of this complex will be sensitive to changes in (Mg2+) and in total ADP, both of which change recipro- cally with changes in total ATP. Summary.-The expression relating the adenylate kinase equilibrium constant to (MNg2+), (H+), and (K+) was solved as a function of (Mg2+) at (K+) = 0.1 31 and pH 6.5, 7.0, 7.2, 7.5, and 8.0. The value [(AMP) (ATP) ]/(ADP)2 varies over a threefold range with (Mg2+) at pH 7.5 and can be a sensitive measure of Downloaded by guest on September 27, 2021 1086 BIOCHEMISTRY: I. A. ROSE PROC. N. A. S.

.MIg2+ concentration in the range 0-0.25 mM Mg2+. The adenylate kinase equilibrium was used to determine the free Mjg2 + concentration in red cell lysates and in whole cells. (Mg2+) is sufficiently low in these cells to leave substantial portions of the ATP and ADP in the uncomplexed form. The author wishes to acknowledge the excellent experimental work of Mrs. J. V. B. Warms. The programming and computing necessary for obtaining Figure 1 were due to Miss Carol Ann Casciato and Mr. Fred Soule of this Institute. * This investigation was supported by U.S. Public Health Service Research grant no. CA- 07819 from the National Cancer Institute, and by grants to this Institute: NIH grants CA- 06927 and FR-05539, and an appropriation from the Commonwealth of Pennsylvania. t Note added in proof: The influence of Mg2 + on the aconitase equilibrium and the use of the citrate/isocitrate ratio in rat heart as a measure of (Mg2 +) have been investigated by P. J. England, R. M. Denton, and P. J. Randle (Biochem. J., 105, 32c (1967)). 'Lardy, H. A., and R. E. Parks, Jr., in Enzymes: Units of Biological Structure and Function, ed. 0. Gaebler (New York: Academic Press, 1956), p. 584. 2 Raaflaub, J., and I. Leupin, Helv. Chim. Acta, 39, 832 (1956). M3Melchior, N. C., and J. B. Melchior, J. Biol. Chem., 231, 609 (1958). 4Hammes, G. G., and D. Kochavi, J. Am. Chem. Soc., 84, 2069 (1962). 6 Lowry, 0. H., and J. Passoneau, J. Biol. Chem., 241, 2268 (1966). 6 Bassham, J. A., P. Sharp, and I. Morris, Biochim. Biophys. Acta, 153, 898 (1968). 7Cohen, L. H., and P. E. Penner, Federation Proc., 24, 357 (1965). 8Allende, J. E., G. Mora, M. Gatica, and C. C. Allende, J. Biol. Chem., 240, 3229 (1965). 9 Lindahl, T., A. Adams, and J. R. Fresco, these PROCEEDINGS, 55, 941 (1966). 10 Meselson, M., M. Nomura, S. Brenner, and C. Davern, and D. Schlessinger, J. Mol. Biol., 9, 696 (1964). 11 Nanninga, L. B., Biochim. Biophys. Acta, 54, 338 (1961). 12Tatibana, M., K. Miyamoto, T. Odaka, and M. Nakao, J. Biochem., 48, 685 (1960). 13 Bartlett, G. R., Biochim. Biophys. Acta, 156, 221 (1968). 14BBucher, T., in Methods in Enzymology, ed. S. P. Colowick and N. 0. Kaplan (New York: Academic Press, 1955), vol. 1, p. 415. 16 Bowen, W. J., and T. D. Kerwin, Arch. Biochem. Biophys., 64, 278 (1956). 16 Phillips, R. C., P. George, and R. J. Rutman, Biochemistry, 2, 501 (1963). 17 Nanninga, L. B., J. Physiol. Chem., 61, 1144 (1957). 18Phillips, R. C., P. George, and R. J. Rutman, J. Am. Chem. Soc., 88, 2631 (1966). 19Melchior, N. C., J. Biol. Chem., 208, 615 (1954). 20 Lowry, 0. H., J. V. Passoneau, F. X. Hasselberger, and D. W. Schulz, J. Biol. Chem., 239, 18 (1964). 21IRose, I. A., and J. V. B. Warms, J. Biol. Chem., 241, 4848 (1966). 22Pastore, E. J., and M. Friedkin, J. Biol. Chem., 236, 2314 (1961). 23 Hofer, M., and W. Hempfling, in Methods in Enzymology, ed. R. W. Estabrook and M. E. Pullman (New York: Academic Press, 1967), vol. 10, p. 479. 24 Rogers, T. A., J. Cellular Physiol., 57, 119 (1961). 25 Walser, M., Ergeb. Physiol., Biol. Chem. Exptl. Pharmakol., 59, 185 (1967). 26 Benesch, R., R. E. Benesch, and C. I. Yu, these PROCEEDINGS, 59, 526 (1968). 27 Schubert, J., J. Physiol. Chem., 56, 113 (1952). 28 Kuby, S. A., and E. A. Noltmann, in The Enzymes (New York: Academic Press, 1962), vol. 6, p. 515. 29Control of Energy Metabolism, ed. B. Chance, R. W. Estabrook, and J. R. Williamson (New York: Academic Press, 1965), p. 369. 30 Chappell, J. B., and A. R. Crofts, Biochem. J., 95, 707 (1965). 31 Abood, L. G., and L. Romanchek, Exptl. Cell Res., 8, 459 (1955). 2 Wiliamson, J. R., M. S. Olson, B. E. Herczeg, and H. S. Coles, Biochem. Biophys. Res. Commun., 27, 595 (1967). 33 Larrson-Raznkiewicz, Ml., Biochim. Biophys. Acta, 132, 33 (1967). 34 Melchior, J. B., Biochemistry, 4, 1518 (1965). Downloaded by guest on September 27, 2021