Nucleotide specificity of cardiac sarcoplasmic reticulum. GTP-induced calcium accumulation and GTPase activity

We previously demonstrated that the hydrolysis of GTP by canine cardiac sarcoplasmic reticulum is not sensitive to calcium and does not support the translocation of calcium and oxalate into the vesicular space. In response to GTP, however, calcium is accumulated into a compartment which is sensitive to pH and ionophore. In the present paper, we further explored the relationship between GTP hydrolysis and GTP-induced calcium accumulation. Both ATP- and GTP-induced calcium accumulation were prevented by the sulfhydryl reagent, N-ethylmaleimide (NEM; I50 = 0.2 mM). In contrast, the sensitivity of NTP hydrolysis to NEM differed markedly; GTPase activity was not affected by NEM, whereas ATPase activity was markedly inhibited. Conversely, although the GTPase was noncompetitively inhibited by the ATP analogue, adenylyl imidodiphosphate (Ki = 8 microM), and was competitively inhibited by the GTP analogue, guanylyl imidodiphosphate (Ki = 60 microM), GTP-induced calcium accumulation was not affected by the NTP analogues at any concentration. Therefore, the GTP-dependent accumulation of calcium into the pH- and ionophore-sensitive compartment of cardiac SR may not require GTP hydrolysis but may be dependent on GTP binding. The previously reported noncompetitive inhibition of the GTPase by ATP was also observed when the calcium-dependent hydrolysis of ATP was prevented by NEM (Ki = 1.2 microM). Along with the noncompetitive inhibition of the GTPase by adenylyl imidodiphosphate, the inhibition of the GTP by ATP in the presence of NEM suggests that ATP binding may be involved in the observed inhibition. The Ki for the noncompetitive inhibition of GTPase activity is compatible with ATP binding to the high affinity catalytic site of the ATPase. Thus, although GTP-induced calcium accumulation differs somewhat from ATP-dependent calcium translocation, the similarities between the two processes (i.e. similar time courses and sensitivity to pH, ionophore, and sulfhydryl modification) suggest that they may be related in some manner.     

Pathway for ATP synthesis by sarcoplasmic reticulum ATPase

Millisecond mixing and quenching experiments were performed in order to study the rate of phosphorylation by P_i of the Ca²⁺-dependent ATPase of sarcoplasmic reticulum vesicles. A rapid phosphoenzyme formation was observed when the vesicles were preincubated in the absence of Ca²⁺ prior to the addition of P_i and Mg²⁺ to the medium, the half-time being in the range of 6 to 10 ms. A lag phase and a 5- to 10-fold slower rate of phosphoenzyme formation were observed when the enzyme was preincubated with Ca²⁺ prior to the addition to the reaction mixture of P_i, Mg²⁺, and an excess of ethylene glycol bis(β-aminoethyl ether)N,N′-tetraacetic acid. The rate of phosphoenzyme hydrolysis was measured either by the addition of Ca²⁺ or, in the absence of Ca²⁺, by tracing the hydrolysis of radioactive phosphoenzyme upon the addition of nonradioactive P_i. In the presence of Ca²⁺, the rate of phosphoenzyme hydrolysis was found to be one order of magnitude slower than the rate of hydrolysis measured in the absence of Ca²⁺. Different rates of phosphoenzyme formation and cleavage were found depending on whether sarcoplasmic reticulum vesicles or purified Ca²⁺-dependent ATPase were used. A transient phosphorylation by P_i was observed when the enzyme was preincubated in the absence of Ca²⁺ and then added to a medium containing P_i, Mg²⁺, and excess of Ca²⁺. The enzyme was phosphorylated during the initial 100 ms, the phosphoenzyme formed being slowly hydrolyzed in the subsequent incubation intervals. In these conditions ATP synthesis was observed if ADP was added to the mixture 100 ms after starting the reaction. No transient phosphorylation by P_i was observed when the enzyme was preincubated with Ca²⁺. Synthesis of a small but significant amount of ATP was observed when the enzyme was preincubated in the absence of Ca²⁺ and then added to a medium containing P_i, ADP, Mg²⁺, and 20 mm CaCl₂. This was not observed when the enzyme was preincubated in the presence of Ca²⁺.     

Accelerating effect of ATP on calcium binding to sarcoplasmic reticulum ATPase

The rise of intrinsic fluorescence due to calcium binding to sarcoplasmic reticulum ATPase occurs with a kobs of approximately 2 s-1 at pH 6.0, which is much lower than that observed at neutral pH. This is consistent with a H+-Ca2+ competition for the high-affinity sites. An accelerating effect of ATP on the calcium-induced transition can be clearly demonstrated at that pH. Nonhydrolyzable nucleotides, such as AMP-PNP, do not elicit the same response. Acetylphosphate also accelerates the calcium-induced fluorescence rise, demonstrating that this effect is limited to substrates that are able to form the phosphorylated enzyme intermediate. This effect, which is attributed to occupancy of the phosphorylation domain of the catalytic site, is distinct from the known secondary activation of enzyme turnover which is produced by ATP and by inactive nucleotide analogs, but not by acetylphosphate.     

Nucleotide specificity of canine cardiac sarcoplasmic reticulum. Differential alteration of enzyme properties by detergent treatment

We previously demonstrated that, in contrast to the hydrolysis of ATP, the hydrolysis of GTP by canine cardiac sarcoplasmic reticulum is not sensitive to calcium. Based on a variety of qualitative and quantitative considerations (cf. Tate, C. A., Bick, R. J., Chu, A., Van Winkle, W. B., and Entman, M. L. (1985) J. Biol. Chem. 260, 9618-9623), we suggested that the hydrolysis of ATP and GTP appears to be effected by the same enzyme. In the present paper, we examined the sensitivity of both enzymatic activities to low concentrations of detergent. With nonsolubilizing concentrations of the nonionic detergent, octaethylene glycol monododecyl ether, the hydrolysis of GTP was rendered partially calcium-sensitive resulting from a slightly increased total (Ca2+ + Mg2+)-GTPase activity and a markedly inhibited calcium-independent (Mg2+-dependent) GTPase activity. Calcium-dependent ATPase activity was increased with octaethylene glycol monododecyl ether, mimicking the effect of the ionophore, A23187. Calcium-dependent ATPase activity and detergent-induced calcium-dependent GTPase activity were similar in (a) calcium sensitivity, (b) sensitivity to mersalyl, and (c) pressure inactivation through dilution and centrifugation, all of which differed from the untreated calcium-independent GTPase activity. Calcium-dependent ATPase activity differed from calcium-dependent GTPase activity with (a) a higher nucleotide affinity, (b) a lower vanadate sensitivity, and (c) a calcium sensitivity for phosphoenzyme formation. Thus, the detergent-induced perturbation of the GTPase resulted in an enzyme with many characteristics qualitatively and quantitatively similar to the calcium ATPase.     

Oxalate, calcium uptake and ATPase activity of sarcoplasmic reticulum vesicles.

Ca++-uptake and Mg++-Ca++-dependent ATPase activity of skeletal muscle sarcoplasmic reticulum vesicles were reciprocally affected by increasing the oxalate concentration from 0 to 4 mM. At 0-0.1 mM oxalate approximately 17% of the calcium was removed by the vesicles from the medium while the ATPase activity was maximal (approximately 0.66 mumoles Pi mg-1 protein min-1). Between 0.1 to 0.2 mM oxalate the ATPase activity was reduced to one-fifth but the uptake rose sharply and 100% of the 45Ca++ was removed from the medium. The uptake was maintained at this level at oxalate concentrations greater than 0.4 mM but the ATPase activity remained inhibited. The kinetics of Ca++-uptake and ATPase activity were also differentially affected by oxalate. In the presence of oxalate, ruthenium red had only a very slight inhibitory effect on the calcium uptake. Addition of 0.1 mM EGTA removed 80% of the Ca++ from preloaded vesicles within 10 min. The formation of insoluble Ca-oxalate salt on the surface of the vesicle is suggested by these results. Calculations based on the Ksp of the calcium oxalate salt are presented to show its formation and the possible speciation of a Ca-oxalate complex which may affect the Ca++-uptake and ATPase activity.     

Cooperative proton and calcium binding by sarcoplasmic reticulum ATPase.

The classic work on binding of calcium to CaATPase is analyzed by an objective non-linear least squares procedure of 74 data points over six pH values. Binding of two calciums to the basic form of the sites occurs with an equilibrium stability constant product of log K1K2 = 13.2. Owing to competition from protons, this value drops in acidic and neutral solutions, becoming, for example, 11.9 at pH 6.8. Binding of the two calciums is so strongly cooperative that its extent is difficult to estimate reliably; there is very little of the one calcium species. Two protons are also bound cooperatively to the calcium sites. In solutions of calcium free protein, at pH less than 7.6 the predominant species holds two protons at the calcium sites, while at greater pH the dominant species bears no protons; there is very little of the intermediate one proton species. The analysis also reveals the likely presence of a small, less than statistical, amount of a ternary complex bearing one calcium and one proton.     

The MgATPase activity of rat cardiac sarcoplasmic reticulum is a function of the calcium ATPase protein.

Magnesium-dependent ATPase (MgATPase) activity is associated with many E1-E2 or P-type transport ATPases including the sarcoplasmic reticulum (SR) calcium ATPase. The SR isolated from rat heart has a MgATPase activity which is 6-12 times faster than the MgATPase activity of the SR isolated from dog heart. To determine the origin of the high MgATPase activity of rat heart SR, we compared and contrasted cardiac SR isolated from both species. The preparations were similar in the following ways: (i) contamination by other organelles; (ii) the comigration of MgATPase activity with calcium-dependent ATPase (CaATPase) activity through a sucrose gradient; (iii) a similar ATPase activity sensitivity to pH and ATP concentration; (iv) the high and similar of sensitivity of ATPase activity to detergent; and (v) a similar protein profile. In both preparations, a single protein in the 105,000-Da region of polyacrylamide gels was phosphorylated by ATP, and the phosphorylated species was an acylphosphate formed in the presence and absence of calcium. Dimethyl sulfoxide, which slows acylphosphoenzyme breakdown, markedly inhibited both CaATPase and MgATPase activities of both preparations but not other enzyme activities. Importantly, the specific inhibitor of the SR calcium pump, thapsigargin, completely inhibited the CaATPase activity with an I50 of 6-7 nM; however, a higher concentration (I50 of 2 microM) was required to inhibit the MgATPase activity of the rat cardiac SR. These results provide evidence that the MgATPase activity of rat cardiac SR is part of the enzyme cycle of the calcium ATPase protein.     

Inhibition of dicarboxylic anion transport by fluorescein isothiocyanate in skeletal sarcoplasmic reticulum.

It has been demonstrated previously that dicarboxylic anions are cotransported during ATP-dependent Ca2+ transport by skeletal muscle sarcoplasmic reticulum (SR) membranes, and that anion cotransport stimulates Ca2+ transport. In the current study, we present evidence that dicarboxylic anion cotransport and Ca2+ transport are kinetically distinct in SR, but both functions are mediated by the CaATPase protein. Preincubation of SR with 40 microM fluorescein isothiocyanate (FITC) (pH 7.0) inhibited essentially all of the Ca2+ ATPase activity, as well as active oxalate-supported and oxalate-independent 45Ca2+ accumulation. The addition of 1 mM beta, gamma-methyleneadenosine 5'-triphosphate (AMP-PCP) to the preincubation media fully protected the dicarboxylic anion-independent Ca2+ ATPase activity and the oxalate-independent active 45Ca2+ accumulation from the inhibitory effects of FITC; however, the ATP-associated [14C]oxalate accumulation, the oxalate-dependent 45Ca2+ accumulation, and the oxalate- and maleate-dependent stimulation of Ca2+ ATPase activity were not protected by AMP-PCP. Thus, the dicarboxylic anion accumulation and the stimulation of Ca2+ uptake by dicarboxylic anions could be functionally separated from the ATP-dependent, anion-independent Ca2+ translocation. FITC bound exclusively to the 100-kDa (CaATPase) and 92-kDa (phosphorylase) proteins in the SR membranes and to purified CaATPase in sodium dodecyl sulfate-polyacrylamide gel electrophoresis; 1 mM AMP-PCP inhibited 50-55% of the FITC fluorescence on the 100-kDa protein, but did not significantly alter fluorescence on the 92-kDa protein. Two-dimensional gel analysis demonstrated a single 100-kDa protein in longitudinal SR membranes. FITC appears to inhibit ATP-dependent Ca2+ transport, and dicarboxylic anion translocation through interaction at separate domains of the CaATPase protein.     

Inhibition of mitochondrial F1 ATPase and sarcoplasmic reticulum ATPase by hydrophobic molecules

The hydrophobic nature of the active site of two energy-transducing ATPases was explored by comparing interactions between Pi and each of three hydrophobic drugs in the absence and presence of organic solvents. The drugs tested were the Fe . bathophenanthroline complex and the anticalmodulin drugs, calmidazolium and trifluoperazine. All inhibit the Pi in equilibrium with ATP exchange reaction catalyzed by submitochondrial particles and the ATPase activity of both submitochondrial particles and soluble F1 ATPase. The inhibition by the three drugs is reversed by either raising the Pi concentration or by adding organic solvent (dimethylsulfoxide, ethyleneglycol or methanol) to the medium. The inhibition of the Pi in equilibrium with ATP exchange by trifluoperazine becomes more pronounced when the electrochemical proton gradient formed across the membrane of the submitochondrial particles is decreased by the addition to the medium of the proton ionophore carbonylcyanide p-trifluoromethoxyphenylhydrazone. The ATPase activity and the Ca2+ uptake by sarcoplasmic reticulum vesicles are inhibited by the Fe . bathophenanthroline complex, calmidazolium and trifluoperazine. Phosphorylation of the ATPases by Pi, synthesis of ATP from ADP and Pi and the fast efflux of Ca2+ observed during reversal of the Ca2+ pump are inhibited by the three drugs. The inhibition is reversed by raising the concentration of Pi or dimethylsulfoxide. The three drugs tested appear to compete with Pi for a common binding site on the Ca2+-ATPase. The data presented are interpreted according to the proposal that the catalytic site of an enzyme involved in energy transduction undergoes a hydrophobic-hydrophilic transition during the catalytic cycle.     

Inhibition of beta-adrenergic stimulated calcium pump of rat cardiac sarcoplasmic reticulum by tricyclohexyltin hydroxide

The effects of tricyclohexyltin hydroxide (Plictran), an organotin acaricide, on 45Ca2+ uptake and Ca2+ ATPase were studied in vitro and in vivo in rat heart ventricular membrane vesicles, primarily sarcoplasmic reticulum. There was a concentration dependent inhibition of both 45Ca2+ uptake and Ca2+ ATPase in vivo as well as in vitro. Isoproterenol, a beta-adrenergic agonist, stimulated 45Ca2+ uptake and Ca2+ ATPase of sarcoplasmic reticulum and this was also inhibited by Plictran. Since cardiac relaxation is mediated by beta-adrenergic stimulation via Ca+ uptake by sarcoplasmic reticulum, the inhibition of calcium pump activity by Plictran may result in alterations in cardiac Ca2+ fluxes leading to cardiac dysfunction.     

Oligovanadate binding to sarcoplasmic reticulum ATPase. Evidence for substrate analogue behavior

Solutions of vanadate were controlled through concentration and pH adjustment to give specific compositions of mono- and oligovanadates. By monitoring the EPR spectrum of iodoacetamide spin-labeled ATPase, it is shown that decavanadate and the oligovanadate species present at neutral pH exhibit behavior typical of a substrate analogue. This is seen in terms of Ca2+ binding site affinity (microM), outward Ca2+ site orientation, and conformational effects on the enzyme normally associated with enzyme activation. In contrast, monovanadates exhibit behavior identical to that observed with Pi, with one exception: the vanadoenzyme is stable to Ca2+ in the concentration range of high affinity binding at the vanadate concentrations used here (200 microM). It is further demonstrated that Ca2+ binding in the 100 microM range directly induces enzyme devanadation of the monovanadate enzyme complex through Ca2+ binding to internal sites. Extensive array formation of dimeric ATPase units is found only with decavanadate in the absence of Ca2+, and then stoichiometric amounts are sufficient. Electron micrographs of dimeric arrays show evidence of increased penetration into the lipid bilayer, including freeze-fracture replicas which show evidence of corresponding "pits" in the inner leaflet of the bilayer. In turn, EPR spectra provide a means of following vanadate binding to the ATPase per se, as well as monitoring Ca2+-induced changes in the vanadoenzyme conformation, as only binding to specific sites on the enzyme affect the EPR spectrum.     

Regulation of calcium transport by protein phosphatase activity associated with cardiac sarcoplasmic reticulum

Canine cardiac sarcoplasmic reticulum is phosphorylated by an endogenous calcium-calmodulin-dependent protein kinase on a 22,000 proteolipid, called phospholamban. Phosphorylation by the calcium-calmodulin-dependent protein kinase is associated with stimulation of the initial rates of calcium transport (Davis, B. A., Schwartz, A., Samaha, F. J., and Kranias, E. G. (1983) J. Biol. Chem. 258, 13587-13591). The present study shows that protein phosphatase activity, associated with canine cardiac sarcoplasmic reticulum vesicles, can catalyze dephosphorylation of the calcium-calmodulin-dependent sites on phospholamban. The activity was maximally stimulated by manganese; fluoride was inhibitory, but its effect was reversible. Dephosphorylation of phospholamban, which was prephosphorylated by calcium-calmodulin-dependent protein kinase, resulted in a reduction of the stimulation on calcium transport rates, particularly at submaximal calcium concentrations. The decrease in calcium transport was associated with a statistically significant decrease in the apparent affinity (EC50) for calcium. Rephosphorylation of phospholamban by the endogenous calcium-calmodulin-dependent protein kinase caused full recovery of the stimulation on calcium transport rates and reversal of the effects mediated by the protein phosphatase. Thus, the calcium pump in cardiac sarcoplasmic reticulum appears to be under reversible regulation mediated by endogenous calcium-calmodulin-dependent protein kinase and protein phosphatase. Such regulation may represent an important control mechanism for the myocardium.     

Utilization of arylazido-ATP by sarcoplasmic reticulum ATPase in the absence of calcium

The ATP analog arylazido-ATP 5'-triphosphate) (3'-O-(3-[N-(4-azido-2-nitrophenyl)amino]propionyl)adenosine 5'-triphosphate) was shown to phosphorylate the calcium-ATPase from sarcoplasmic reticulum in the absence of calcium. Levels of 0.6 nmol of phosphoenzyme/mg of protein were attained. Calcium either at micromolar or millimolar concentrations did not affect the level of phosphoenzyme. A non-Michaelian dependence of the hydrolytic activity as a function of analog concentration was obtained in the absence of calcium. Calcium addition did not modify either the analog concentration dependence for the activation of hydrolysis or the maximal rate of hydrolysis. In the presence of micromolar calcium, arylazido-ATP promoted calcium accumulation inside the vesicles, and a steady-state level of 100 nmol of calcium/mg of protein was maintained. ESR spectra of spin-labeled ATPase showed that addition of the analog in the absence of calcium caused a spectral change, and the resulting spectral parameters were different from those obtained for ATP under similar conditions. Calcium addition did not cause any further modification of the spectra, which was clearly distinct from the change when ATP was used. The partition coefficient of the analog from a water medium into an organic phase was found to be 1 order of magnitude higher than that of ATP. It is suggested that it might be the hydrophobic nature of the analog which makes it bypass the calcium requirement for utilization of the substrate by the ATPase.     

Fluoride-inhibited calcium ATPase of sarcoplasmic reticulum. Magnesium and fluoride stoichiometry.

The sarcoplasmic reticulum (SR) CaATPase is inactivated by fluoride in the presence of magnesium (Murphy, A. J., and Coll, R. J. (1992) J. Biol. Chem. 267, 5229-5235). The inactive complex is very stable and can be isolated free of other components by 48 h of dialysis at 4 degrees C (Murphy, A. J., and Coll, R. J. (1992) J. Biol. Chem. 267, 16990-16994). In this study, we used a fluoride-specific electrode to determine that the amount of tightly bound fluoride in the complex was 9.4 +/- 2 nmol mg-1 SR protein. The rate constant of inactivation was very similar to the rate constant of fluoride incorporation and varied directly as the square of the fluoride concentration. Luminal Ca2+ accelerated reactivation of the inhibited enzyme, and the rate constants of activity regain and fluoride release were very similar. Although required for inhibition, added magnesium did not accelerate reactivation. Analysis for magnesium using antipyrylazo III of the inhibited enzyme showed 4.1 +/- 0.4 nmol mg-1 SR protein. As there is much evidence in the literature supportive of an estimate of calcium pumps equal to approximately 4-5 nmol mg-1 SR protein, our results indicate that each inhibited enzyme contains two tightly bound fluorides and one tightly bound magnesium.     

Steady state kinetics of ATP synthesis and hydrolysis coupled calcium transport catalyzed by the reconstituted sarcoplasmic reticulum ATPase

The steady state kinetics of calcium transport driven by ATP hydrolysis and ATP synthesis catalyzed by purified, reconstituted calcium ATPase has been investigated as a function of the transmembrane calcium gradient. Purified calcium ATPase was reconstituted into phospholipid vesicles enabling control of the transmembrane calcium gradient. Calcium transport was monitored spectrophotometrically by the calcium indicator, Arsenazo III. Thus, only the enzymatic activity of coupled transport was measured. It was shown under conditions of low external calcium that ATP hydrolysis and synthesis follow simple Michaelis-Menten kinetics and that Michaelis constants obtained for both processes appear independent of the calcium gradient. The maximum velocities for both hydrolysis and synthesis strongly depend on the transmembrane calcium gradient. Based on these results, a mechanism is proposed in which a random addition of substrates for ATP synthesis is followed by random release of ATP and calcium. By measuring the ATP hydrolysis and synthesis under identical conditions, determination of the equilibrium constant for ATP hydrolysis as a function of the transmembrane calcium gradient was possible. Our results indicate that the thermodynamics of the catalytic cycle can be totally accounted for by the energetics of transport of 2.2 +/- 0.3 calciums and the hydrolysis of 1 ATP. An equilibrium constant for ATP hydrolysis in the absence of a calcium gradient was determined to be 4.0 X 10(4).     

Comparison of ATPase activity of cardiac sarcoplasmic reticulum fraction from rat and dog

The purpose of the present study was to compare the ATPase activities of cardiac SR in two species in which the different intrinsic myocardial contractility can only partially be explained by the different properties of cardiac myosins. In cardiac SR isolated from rat heart, the total ATPase activity was 1512.5 +/- 23.3 nmol Pi/mg protein/min, nearly four times as high as in dog cardiac SR (408.8 +/- 28.9 nmol Pi/mg protein/min). The Ca2+-activated ATPase in rat cardiac SR represented only 23.8% of the total ATPase activity, while in dog cardiac SR it was approximately 50% of the total. Thus, the specific Ca2+-activated ATPase was nearly two times higher in the cardiac SR of the rat than in that of the dog. This higher rate of ATP hydrolysis in rat cardiac SR may be, at least in part, responsible for the increased intensity and shorter duration of the active state in the rat myocardium. Polyacrylamide gel electrophoresis of SR showed that the relative amount of Ca2+-pump protein was two times higher in dog heart, similar to the percentage of Ca2+-activated ATPase activity. At the same time, the specific Ca2+-activated ATPase activity and the relative amount of Ca2+ pump protein in both the rat and dog cardiac SR were inversely related.     

Reaction of a carbodiimide adduct of ATP at the active site of sarcoplasmic reticulum calcium ATPase

An adduct of a carbodiimide and ATP was synthesized from 1-ethyl-3-[3-(dimethylamino)propyl]carbodiimide (EDC) and the nucleotide. Despite its limited stability (t1/2 for hydrolysis of about 5 min at 25 degrees C), it was shown to react with and inactivate the calcium ATPase of sarcoplasmic reticulum in its vesicular, nonionic detergent-solubilized and purified forms. Saturation kinetics, with an ATP-EDC concentration dependence midpoint in the 10 microM range, were observed, suggesting an active-site affinity which is similar to ATP. The reaction was specific in that inactivation required reaction of about one adduct per ATPase. The modified enzyme could no longer be phosphorylated by ATP or Pi or hydrolyze p-nitrophenyl phosphate, but retained the ability to undergo the high-affinity calcium-dependent fluorescence change. It also bound trinitrophenyl-ADP and other nucleotides at least 10-fold more weakly than the unmodified ATPase. The inactivation reaction required the presence of Mg2+ and Ca2+ and was prevented by nucleotides such as ATP and ADP. For magnesium, the inactivation-enabling effect occurred with a midpoint of 3 mM. In the case of calcium, the transition resembled high-affinity binding in that it occurred cooperatively with a midpoint in the micromolar range. Higher [Mg2+] shifted this transition to higher [Ca2+]. Polyacrylamide gel electrophoresis (PAGE) demonstrated that the reaction converted the ATPase (Mr = 1.1 x 10(5)) to a species with an apparent Mr = (1.7-1.8) x 10(5). Since nonionic detergent-solubilized ATPase and purified ATPase gave similar results, intramolecular cross-linking is implicated.(ABSTRACT TRUNCATED AT 250 WORDS)