Comparison between ATP-Supported and GTP-Supported Phosphate Turnover of the Calcium-Transporting Sarcoplasmic Reticulum Membranes

Transcription

1 Eur. J. Biochem. 11, (1979) Comparison between ATP-Supported and GTP-Supported Phosphate Turnover of the Calcium-Transporting Sarcoplasmic Reticulum Membranes Nelly RONZANI, Andrea MIGALA, and Wilhelm HASSELBACH Max-Planck-Institut fur Medizinische Forschung, Abteilung Physiologie, Heidelberg (Received June 6, 1979) The study deals with the interrelationship of the phosphate-transferring activities of the calciumtransporting sarcoplasmic reticulum membrane vesicles : the phosphate exchange between nucleoside triphosphate (NTP) and nucleoside diphosphate (NDP) (NTP-NDP exchange), the calcium-dependent NTPase, and the phosphorylation of NDP by inorganic phosphate in the presence of NTP (NTP-Pi exchange). Different nucleotides were used as phosphate donors and acceptors. It is demonstrated for the phosphate transfer from ITP to GDP that the NTP-NDP exchange exhibits ping-pong kinetics with Mg-ITP and unliganded GDP as substrates. The apparent affinities of the enzyme for the nucleoside diphosphate and triphosphate species are deduced according to this mechanism. The enzyme's affinity for the nucleoside triphosphates and diphosphates depends on its functional state being considerably lower under conditions of NTP-NDP exchange than during NTP splitting or NTP synthesis. ATP and GTP are split with the same low rates when calcium-activated NTPase is inhibited by high internal calcium concentrations after calcium transport has reached steady state. The rates of the NTP-NDP exchange reactions, however, differ by a factor of about 1 being E 3 pmol. mg-'. min-' for ATP-ADP and only E.3 pmol. mg-'. min-' (22 C) for GTP- GDP. When the sarcoplasmic reticulum vesicles are made calcium-permeable, the calcium transport ATPase is turned on and the rates of GTP and ATP splitting increase about tenfold. Yet, while the rate of ATP-ADP exchange is little reduced, the rate of GTP-GDP exchange drops by approximately 5 %. The persisting exchange activity of calcium-permeable vesicles demonstrates that high internal calcium concentrations are not required for the transfer of the protein-bound phosphoryl group to NDP during NTP-NDP exchange. The calcium-transporting membranes of the sarcoplasmic reticulum are especially suited for studying energy-dependent ion translocation due to their exceptional simplicity in structure and function. The energydependent calcium movement is causally coupled with the phosphorylation of the transport protein which is the main structural component of the membranes (cf. [1,2]). In other ion transport systems the interrelationship between ion translocation and phosphorylation mainly relies on the demonstration of the Abbreviations. EGTA, ethylene glycol bis(j-aminoethyl ether)- N,N,N',N'-tetraacetic acid; NTP, nucleoside triphosphate; NDP, nucleoside diphosphate; NTP-NDP exchange, phosphate exchange between NTP and NDP or nucleosidediphosphate kindse activity; NTP-P, exchange, incorporation of inorganic phosphate into NTP during NTP-NDP exchange; NTPase, nucleoside triphosphatase activity. Enzymes. NTPase (EC ); NTP-NDP exchange or nucleosidediphosphate kinase (EC ); pyruvate kinase (EC ); phospholipase AZ (EC ); adenylate kinase (EC ). occurrence of a phosphoprotein after acid denaturation. Therefore its direct involvement as a transport intermediate of the ion translocation reaction sequence has repeatedly been questioned [3]. For the calcium transport system of the sarcoplasmic reticulum, in contrast, the involvement of a phosphorylated intermediate in ion translocation could be substantiated in the native protein. Calcium uptake and the calciumdependent ATP splitting are paralleled by a rapidly proceeding phosphate exchange between ATP and ADP. This calcium-dependent nucleoside diphosphokinase activity is difficult to reconcile without assuming the formation of a phosphorylated intermediate [4-61. The just-mentioned parallelism had led to the assumption that the exchange reaction might be directly coupled with the inward movement of calcium ions across the membranes and, on account of the principle of microreversibility, also to the efflux of calcium during ATP synthesis [2]. Thus the nucleoside diphos-

2 594 Phosphate Turnover of Sarcoplasmic Reticulum Membranes Supported by ATP and GTP phokinase activity does represent an essential sequence in the reaction chain of the sarcoplasmic calcium transport which in a simple version is illustrated by Scheme 1 [7]. Scheme 1 Caz Ca2 ATP + E L-?E-P b E. E* -E-P + P + ADP This report deals with the analysis of the nucleoside diphosphokinase activity of the sarcoplasmic-reticulum calcium-transport system in different functional states by comparing different nucleotides as substrates. It will be shown that at pseudo equilibrium when calcium transport has levelled off the nucleotide specificity of the calcium-dependent NTPase is lost, while the nucleoside diphosphokinase activity displays a pronounced specificity for its substrates. In this state the enzyme exhibits lower apparent affinities for different nucleotides than during unidirectional activity. It will further be shown that the nucleoside diphosphokinase activity is only indirectly connected with calcium translocation in contrast to the hydrolysis of the phosphorylated intermediates formed by phosphoryl transfer from the energy-yielding substrates when calcium ions are accumulated and in contrast to phosphoprotein formation by the incorporation of inorganic phosphate when ATP is synthesized during calcium release. MATERIALS AND METHODS Preparation of Sarcoplasmic Reticulum Vesicles The vesicular fragments of the sarcoplasmic reticulum were prepared at 4 "C from rabbit skeletal muscle as described by Hasselbach and Makinose [8] and modified according to de Meis and Hasselbach [9]. Calcium-Dependent NTPase Activity Nucleoside triphosphatase activity was determined by measuring the rate of inorganic phosphate liberation at 22 C. The incubation medium contained 2 mm histidine buffer ph 7., 4 mm KC1, 5 mm MgCI2 and.1 mm CaC12. The concentrations of NTP and NDP are given in the legends. The reaction was started by the addition of the sarcoplasmic reticulum vesicles at.1 mg protein/ml. Aliquots were taken at different times and the reaction was stopped with trichloroacetic acid (final concentration 3 %), + filtered, and inorganic phosphate in the filtrate was determined according to Rockstein and Herron [lo]. Calcium-independent splitting of the nucleoside triphosphates was measured in the presence of 2 mm EGTA. Phosphate Exchange between NTP and NDP NTP-NDP exchange was carried out according to Makinose [Ill using ['4C]NDP as phosphate acceptor. The incubation medium contained 2 mm histidine buffer ph 7., 4 mm KCI,.1 mm CaC12, NTP, ['4C]NDP and MgClz as indicated in the figures or tables. ATP-ADP exchange was measured in the presence of.1 mm P1,P5-bis(5'-adenosyl) pentaphosphate to inhibit adenylate kinase aclivity present as a contamination in the preparation [12]. The reaction was started by the addition of the sarcoplasmic reticulum vesicles at.2 mg protein/ml for ATP-ADP exchange or.2mg protein/ml for GTP-GDP exchange. Aliquots of 2 ml were taken at different times and the reaction stopped with.5 ml of 15 o/, cold trichloroacetic acid. After removal of the protein by centrifugation trichloroacetic acid was extracted twice from the supernatant with 5 ml of diethyl ether. When low concentrations of nucleotides were used, cold nucleotides were added to the samples. The separation of the nucleotides was performed by thin-layer chromatography carried out on polyethylenimine-cellulose (TCL plastic sheets PEI-cellulose F; Merck, Darmstadt) with.5 M KH2P4 for ATP and ADP or 2. M formic acid and.5 M LiCl for GTP and GDP. The spots of the nucleotides were localized by ultraviolet light and cut out for liquid scintillation counting. In the experiment with open vesicles the calcium ionophore X537A dissolved in ethanol was added to the assay medium at a final concentration of 2 FM. NTP-Pi Exchange NTP-Pi exchange was assayed by measuring the formation of [Y-~~P]NTP from NDP and [32P]phosphate. The incubation medium used was the same as the medium described for NTP-NDP exchange except that potassium [32P]phosphate(pH 7.) was added. Incubations were performed at 22 "C and the reaction stopped with trichloroacetic acid (final conc. 3 "/,). Inorganic phosphate was removed by the method of Avron [13]. To 1 ml of the filtrate after trichloroacetic acid precipitation, 1.2 ml of acetone,.3 ml of distilled water saturated with a 1 : 1 mixture of isobutyl alcohol and benzene and 7. ml of a 1 : 1 mixture of isobutyl alcohol and benzene saturated with distilled water were added. The mixture was vigorously stirred and after phase separation.8 ml of a 2 M H2S4 solution

3 N. Ronzani, A. Migala, and W. Hasselbach 595 containing 5 % (w/v) ammonium molybdate was added. After gentle mixing and incubation for 5 min at room temperature the tube was vigorously stirred for 3 s. After phase separation the isobutyl alcohol/benzene layer was removed and.1 ml of 2 mm KHzPO4 was added to the aqueous layer. The extraction with 7 ml of isobutyl alcohol/benzene was repeated and after removal of the organic phase 1. ml aliquot of the water phase was counted in a liquid scintillation counter. Phosphoprotein Formation The sarcoplasmic reticulum protein was phosphorylated by [y3'p]gtp or [p3'p]itp in assay media of the following composition: 4 mm KCl, 2 mm histidine, 5 mm MgC12,.1 mm CaC12,.12 mm EGTA. The concentrations of GTP and GDP are given in the legends. The reaction was quenched at appropriate time intervals with ice-cold trichloroacetic acid (5 % final concentration and 25 mm phosphate). After centrifugation at "C the supernatant was discarded and the pellet resuspended in ice-cold washing solution of 1 % trichloroacetic acid plus 5 mm phosphate. The protein was precipitated on a glass fiber filter and washed repeatedly. The protein precipitate with the filter was resuspended in 3 ml.1 M NaOH. After heating in boiling water for 5 min, 1 ml.1 M H3P4 was added. Aliquots were taken to determine 32P activity by liquid scintillation counting. The concentrations of the various ion species were calculated employing the stability constants given by Schwarzenbach [14] for EGTA and collected by Martell [15] for the various nucleotides. Under the conditions of our experiments, values of 1 x lo4 M-' and 1.6 x lo3 M-l were adopted as stability constants for all NTP and NDP species respectively, at 22 "C [16,17]. The apparent stability constant of the magnesium nucleoside diphosphate complexes in solutions containing 4 mm KC1 and 2 mm histidine ph 7. were checked using a model 91 ionalyzer (Orion) equipped with a divalent cation electrode. Values of (1.6 k.2). lo3 M-' were found for Mg- ADP and Mg-GDP which is in good agreement with the values for Mg-ADP and Mg-IDP [18] reported in the literature. For the experiments performed at "C, 1.13 M-I was used as stability constant for GDP [15]. [14C]ADP, [14C]GDP, [32P]phosphate and [Y-~~PIGTP were obtained from Amersham Buchler (Braunschweig, F.R.G.). [y-32p]1tp was prepared as described for [p3'p]atp by Glynn and Chapel1 [19]. The calcium ionophore X537A was a generous gift of Hoffmann-La Roche Ltd. (Basel, Switzerland). P',P5-bis(5'-adenosyl)pentaphosphate, GTP and GDP were purchased from Boehringer (Mannheim, F.R.G). All other chemicals were P.A. grade. RESULTS Calcium-dependent NTP-NDP Exchange, Substrate Specificity and Aflinity The analysis of the mechanism underlying the nucleoside diphosphokinase activity catalyzed by the calcium transport enzyme of the sarcoplasmic reticulum membranes is made difficult, especially for the adenine nucleotides, by two properties of the preparation. a) The sarcoplasmic membranes are contaminated by the ubiquitous adenylate kinase [5] which can falsify the exchange reaction by ATP formation from ADP. b) On account of the high apparent affinities of the calcium-transport enzyme for ATP and ADP [2] the evaluation of the dependence of the exchange reaction on the concentration of the nucleotides requires measurements at low concentrations of both substrates. Under such conditions, however, the hydrolysis of ATP can change the low substrate concentrations significantly leading to an apparent reduction of the exchange activity. While the effect of the adenylate kinase can be abolished by adding the effective inhibitor P',P5-bis- (5'-adenosyl) pentaphosphate the action of the ATPase would remain disturbing if the ATP-ADP exchange activity exhibits the same high affinity for the nucleotides as reported for phosphoprotein formation and ATPase activity. The difficulties discussed for the ATP-ADP exchange reaction can be supposed to be much less aggravating for the guanosine and inosine phosphates which are not interconverted by the adenylate kinase and for which much lower affinities for the enzyme have been reported [21]. For GTP as well as for ITP, GDP was adopted as phosphate acceptor because of its commercial availability as I4C-labelled compound. The exchange reaction was studied with closed sarcoplasmic reticulum vesicles after calcium uptake has reached pseudo equilibrium in order to find out if under these conditions other kinetic parameters prevail than far from equilibrium. At pseudo equilibrium no net calcium uptake occurs and the calcium-dependent NTPase is strongly suppressed by the high calcium concentration inside the vesicles. The interference of the NTP hydrolysis with the exchange reaction is thus largely excluded. The ratio of calcium to vesicles was chosen so that the free calcium concentration in the medium remained above 1 pm. For the nucleotide couples ATP-ADP, GTP-GDP and ITP-GDP the dependence of the rate of phosphate exchange on the concentration of the respective nucleoside diphosphates and triphosphates is illustrated in Fig. 1-4 and 7a. Concentrations of ionized magnesium and calcium were chosen in order to obtain optimal activity [5,22] (compare Table 4). The exchange rate ob-

4 596 Phosphate Turnover of Sarcoplasmic Reticulum Membranes Supported by ATP and GTP 31-4 a, m c m c u x? a t 4.4 Fig Fig. 2 rf,, s l/[gtp] (mm-') l/ [GDP](mM-') n [GTPI, (mm) [GDP], (mm) Fig. 1. Dqendence on ATP concentration ofthe ATP-ADPphosphute exchange. The assay media contained 2 mm histidine, ph 7.,O.l mm CaC12, 4 mm KC1,.1 mm P1,P5-bis(5'-adenosyl) pentaphosphate, 2 mm ADP; (-) average and standard error of four experiments. The total magnesium concentration exceeds the concentration of the nucleotide by 1 mm. The reaction was started by the addition of.2 mgvesicular protein. ml-' of closed sarcoplasmic reticulum vesicles at 22 "C. The concentration oftotal ATP is plotted on the abscissa. Inset: reciprocal plot of the data with the concentration of Mg-ATP as substrate on the abscissa Fig. 2. Dependence on ADP concentration of the ATP-ADP phosphate exchange, The assay medium was composed as described in Fig. 1 except that the total ATP concentration was 2 mm and the total ADP concentration changed as indicated: (-) average and standard error of four experiments. Inset: reciprocal plot with the concentration of unliganded ADP on the abscissa Fig. 3. TJw dependence oii GTP concentration of the GTP-GDP exchange reaction. The assay medium contained 2 mm histidine, ph 7.,.1 mm CaC12, 4 mm KCI, 2 mm GDP and 1 mm of magnesium chloride in excess of the nucleotides. The total concentration of GTP is given on the abscissa, protein.1 mg. ml of closed sarcoplasmic reticulum vesicles, t = 22 C. Inset: the data are plotted doubly reciprocally with the concentration of Mg GTP on the abscissa Fig.4. The dependence on GDP concentration of the GTP-GDP exchange reaction. The concentration of GTP was 2 mm. The total GDP concentration is plotted on the abscissa. Other conditions are as described in Fig. 3. Inset: data are plotted doubly reciprocally taking into account unliganded GDP as substrate of the reaction Fig. 4 served at 22 "C at near saturating concentrations of proximately the same interval of total nucleotide con- GTP, ITP and GDP respectively, are more than centration between.2 and 2 mm. 1 times lower than the corresponding exchange rate In contrast, clearly separate concentrations of of the adenosine phosphates being % 3 pmol. mg-i the NTP species are required for the activation of the. min-'. The comparison of Fig. 1-4 and 7a further exchange reaction at near saturating concentrations reveals that the activation of the exchange reaction of the nucleoside diphosphates. This result seem's to by the three nucleoside diphosphates occurs in ap- be in accordance with the differences in the affinities

5 N. Ronzani, A. Migala, and W. Hasselbach I 597,,,/I F 2 5 k z :8 % - 7, , 8 I I I 8 3, I I I I I I 1, : * which the calcium dependent NTPase activity of the enzyme exhibits for its different substrates as shown by Fig. 5. GTP, and especially ITP, must be regarded as low affinity substrates as compared to the much higher apparent affinity which ATP has for the enzyme (Table 1). The table shows that the values of the affinities for the three NTP Species estimated by NTP cleavage and phosphate exchange decline in the same order ATP > GTP > ITP. Exchange Reaction Mechanism An essential prerequisite for the kinetic analysis of the exchange reaction is the ascertainment of the nucleotide species which the enzyme uses as substrates. Under the prevailing conditions the nucleotides either react with unliganded ions or as their magnesium complexes [2,23,24]. Since it can be taken as proven that the exchange reaction proceeds via the formation of a phosphoenzyme, the determination of the dependence of the phosphoprotein level on the magnesium concentration in the media containing y-phosphatelabelled nucleoside triphosphates and unlabelled diphosphates is the most convincing experiment that can be performed to decide which are the two of the four nucleotide couples (MgNTP-MgNDP, MgNTP- NDP, NTP-MgNDP, NTP-NDP) used by the enzyme (M. Makinose, unpublished results). Fig. 6 demonstrates that at 2 "C and at "C the phosphoprotein level increases when the magnesium concentration is raised in media containing [Y-~~PIGTP and unlabelled GDP. A similar phosphoprotein profile was obtained when [y-32p]itp was used as phosphate donor and GDP as phosphate acceptor. The addition of increasing concentration of magnesium to the phosphorylation assay can give rise to an elevation of the phosphoprotein level only, when magnesium nucleoside triphosphate acts as phosphate donor and unliganded nucleoside diphosphate as phosphate acceptor. This substrate specificity is taken into account for the evaluation of the affinities of the enzyme for its substrates as well as for the evaluation of rate constants of the reaction. Among the investigated nucleoside triphosphates ITP exhibits the lowest apparent affinity for the enzyme. It seems, therefore, especially suited for analyzing the exchange reaction because its concomitant hydrolysis by the enzyme does not significantly change the relatively high substrate level. Fig. 7 shows the exchange rates obtained by experiments in which the concentration of ITP and GDP were varied systematically. The graphs obtained when the rates are plotted doubly reciprocally for different constant magnesium ITPjGDP ratio can be represented fairly well by straight lines intersecting on the ordinate. The results support a reaction sequence according to which phosphorylation of the protein by the nucleoside triphosphates must be followed by the release of the corresponding diphosphates before the enzyme can bind another nucleoside diphosphate molecule as phosphate acceptor [17]. This ping-pong Bi-Bi mechanism can be described by the reaction sequence of E + MgT AEMgT AMgE k-, - P. D Scheme 2 K &MgE - P + D

6 598 Phosphate Turnover of Sarcoplasmic Reticulum Membranes Supported by ATP and GTP [Mg-ITP] (rnm) l/[mg-itp] (rnm-') Fig.7. Dependence of the ITP-GDP phosphate exchange reaction on the concentration of ITP at different GDP concentrations. (A) Total GDP concentrations () 1 mm; (v) 2 mm; (A) 4 mm; (M) 6 mm. The corresponding concentrations of unliganded GDP are.32 mm,.57 mm, 1. mm and 1.3 mm. The measurements were performed with closed vesicles at 22 C. Other conditions are described in the legend of Fig.3. (B) Double reciprocal plot obtained from A for different constant ratios of Mg-ITP and unliganded GDP ().5 mm; (A) 1. mm; (V) 1.3 mm and (M) 2. mm in which the substrates magnesium nucleoside triphosphates (MgT) and nucleoside diphosphates (D), are in rapid equilibrium with the corresponding enzyme substrate complex EMgT and MgE - P. D. The phosphoryl transfer reaction between these intermediates is assumed to proceed slowly. Under these assumptions and with the conservation of the enzyme one obtains the following expressions : mation concerning the existence of such silent intermediates can be obtained by evaluating the overall equilibrium constant of the exchange reaction deduced either from the exchange data or from phosphoprotein determination at different NTP and NDP ratios. The exchange data should yield the true equilibrium constant Keq = [El. [T]/[E - PI. [D] because the exchange reaction monitors only those intermediates which appear in the expression for the equilibrium constant. In contrast, other intermediates possibly present in the reaction chain are also registered when the equilibrium constant is evaluated from phosphoprotein (EP) determinations. That is because the analytically determined equilibrium constant For constant ratios, p = nucleoside triphosphate/ nucleoside diphosphate the latter equation can be rearranged to ~ which yields straight lines of different slopes intersecting on the ordinate. This functional relationship is consistent with the results depicted in Fig. 7B. The apparent affinity constants Kfl ; Kbl obtained from the exchange experiments are compiled in Table 1 and compared with apparent affinity constants deduced from measurements of NTP binding, NTP phosphorylation, NTP hydrolysis or NTP formation. It is evident that the exchange experiments yield considerably lower affinities than all other estimates. Such differences in the apparent affinities can be expected if in the reaction chain intermediates are formed which do not participate in the exchange reaction. Infor- ([El + P*I) [MgTl - ([E- PI + [E-3m (compare Scheme 1) includes possible phosphoprotein species (E-P) as well as unphosphorylated intermediates (E*) which do not participate directly in the exchange reaction. Fig. 8 illustrates the phosphoprotein levels at different GTP/GDP (Fig. 8 A), ITPjGDP (Fig. 8 B) and ATP/ADP (Fig. 8 C) ratios. The overall equilibrium constant K, for the substrate couple GTP-GDP is very similar to that deduced from the exchange experiment (Table 1). For the substrate couple GTP-GDP the following alternative arises from the coinciding results of two different approaches. a) Under exchange conditions during pseudo equilibrium only those intermediates exist which are directly involved in the exchange reaction. b) The results are also compatible with the existence of silent intermediates like E* and E-P provided that they exist in the same proportion as the active

7 ~ -- ~ P e. 9 wl W W Table 1. Apparent affinities and rate constants of NTP-NDP interaction with surcophnic reticulum membranes The compiled data compare substrate affinities, equilibrium constants and rate constants obtained by NTP-NDP exchange measurements at pseudo equilibrium of calcium transport with those deduced from initial rate measurements and binding studies, respectively. The experiments were performed at 2-22 "C. For the apparent equilibrium constants K ' [MgTl KT [Eplmax- ([EP1' eq - - ~ a n Ks d = [MgT1). KT and KO are the reciprocal affinities of the enzyme from the nucleoside tri- and diphosphates as used in Eqn (Ic). [EPI. [Dl KD [EPI ' [Dl The second-order rate constants are obtained from the exchange rates at low substrate concentrations (Fig. 1-4) according to kl = u,,/([e]. [MgT]) and k-l = u,,/([ep]. [D]) (Scheme 3). The corresponding concentrations of E and EP are taken from phosphoprotein determinations at different NTP/NDP ratios (Fig. 8). The first-order rate constants k; and k13 (Scheme 2) are calculated using either the relation k' = k/k between the second-order rate constant k and the affinities K for the respective nucleoside triphosphates and diphosphates or from the measured rates and the corresponding concentrations of E or EP Nucleo- k-1 kl3 Affinity Apparent equilibrium Rate of phosphorylation of closed vesicles Rate of dephosphorylation of closed vesicles tides constants A" Bb _ exchange E -P _ ~ initial exchange initial exchange Ke, formation max rate ki k; ki k; max rate k-1 kl3 KS iz 66' 5 P E i r z v1 IE ' i5 MgATP 5x15 8x13 PO, 261 ADP 3x16 3x13 [I1 MgGTP 3x lo4 2x lo3 GDP 3x lo4 1.3~ lo3 o'6 111 MgITP 2x lo3 1.1 x lo3 1'2 [I, 261 J } [2] 35 [22,26] 4x x1~ 3 > 35 [26] >4x17 >85 lo ' 4x ~1~ &.3 6 [26] 2x ' 1.3~ o a Obtained from NTP binding, hydrolysis and synthesis experiments. Obtained from NTP-NDP exchange (KF1 KG'). Fassold, E., Waas, W., and Hasselbach, W.. unpublished results.

8 6 Phosphate Turnover of Sarcoplasmic Reticulum Membranes Supported by ATP and GTP - 5 Q l e 1-2 lo-' 1 [ATPI imm) Fig.8. Phosphoprotein level at diffcrent GTP-GDP, ITP-GDP und A TP-ADP rutios. Closed vesicles were phosphorylated as described in Materials and Methods with [g-"p]gtp (A), with [y-32p]1tp(b) and with [p3'p]atp (C) in the presence of different concentrations of GDP (A, B) and ADP (C). Abscissa: concentration of total [;,-3ZP]NTP; concentration of total GDP () mm, (W).5 mm, (v) 1 mm, (A) 2 mm. (x) 4 mm: ( ) concentrations of total ADP. () mm. (W).1 mm, (A).5 mm. The assay of experiment A contained 5 mm MgC12. In experiment B the magnesium concentration was 1 mm in excess of the total nucleotide concentration. In experiment C 1.5 mm MgClz were present. Phosphorylation was terminated by acid quenching after 1 s and 2 s at 22 'C. The values of the equilibrium constants Ks = { [EP],,, - ([EP] [MgT]))i [EP]. [D] were obtained from the phosphoprotein levels at the corresponding Mg-NTP (MgT) and NDP (D) concentrations. The substrate concentrations MgT and D were calculated using the stability constants given in Materials and Methods. EP must be considered to be the sum of the intermediates E - P and E-P of Scheme 1 ones due to the following relationship [E]/[E*] = [E - P]/[E-PI; Keq/Ks 2: 1. Consequently the existence of silent intermediates cannot be proven by this approach when the enzyme catalyzes GTP-GDP exchange. If, however, the ratio Ke,,/K, differs essentially from one, as it is the case for the substrate couple ITP-GDP and the adenine nucleotides, the ratio of the unphosphorylated intermediates [El/( [El + [E*]) becomes different from that of the phosphorylated ones [E - P]/([E - PI + [E-PI) which is in evidence for silent intermediates. The distribution of the intermediate apparently depends on the nature of the used substrates. An additional finding which is in support of silent intermediates is the dependence of the K, value for the adenine nucleotides on their concentrations as it emerges from the results of Fig. 8 C yielding K, values between 1 and 1. Kinetics of the Exchange Reaction Deduced from Steady State and Initial Rate Measurenient The evaluation of the rate constants of the exchange reaction should yield further information concerning the existence of silent intermediates. The estimate of the kinetic constants based on the observed exchange rates, the determined phosphoprotein levels and the computed equilibrium constants registered only those intermediates which in fact participate in the exchange reaction. For the evaluation of the second-order rate constants at nonsaturating substrate concentrations the reaction pathway simplifies to Scheme 3 : Scheme 3 i t E + MgT MgE - P + D. The second order rate constants kl and k-1 as well as the first-order rate constants ki and k13 which result from Scheme 2 are collected in Table 1. The values are much smaller than those reported for the transfer of the terminal phosphate group of the triphosphates to the enzyme and for the dephosphorylation of the phosphorylated enzyme by the nucleoside diphosphate as determined by initial rate measurements [22,26]. An estimate of the individual rate constants not based on measurements under steady-state conditions has been performed at "C where the rate of phosphoprotein formation and its decay can be measured by manual quenching. This approach, likewise, relies on assumptions concerning concentrations of the phosphate-accepting and the phosphate-donating enzyme species. It must be considered to be very unlikely that the occupation of the different enzyme states is identical under steady state and initial rate conditions. The basic assumptions for the initial rate measurements are given by Scheme 4: E + M ~T+M~E - P + D I". E+P Scheme 4 which yields the following expressions.69/ti 1 (ki. [MgT]) kz; [D] = ;.69/t-i= (k-i. [D]) + kz; [TI =;.69jtz = kz

9 N. Ronzani, A. Migala, and W. Hasselbach 61 c Time (s) Time (s) Fig. 9. Formation and decay of phosphoprotein. (A) Closed sarcoplasmic vesicles were phosphorylated at "C in media containing 5 mm MgC12, 4 mm KCI, 2 mm histidine, ph 7.,.1 mm CaClz and ().3 mm, (W).5 mm, (v).1 mm [y-32p]gtp. The reaction was started by the addition of the protein and terminated by acid quenching at the times given on the abscissa. (B) For measuring the decay of phosphoprotein the vesicles were phosphorylated for 2 min with.1 mm [Y-~'P]GTP. Phosphorylation was interrupted by the addition of 5 mm EGTA. At the indicated time intervals, phosphoprotein decay was stopped by acid quenching. () 5 mm EGTA alone; (M) 5 mm EGTA, 2 mm GDP; (V) 5 mm EGTA, 5 mm GDP. (C) Phosphoprotein level at different GTP/GDP ratios; t = "C. The assay contained () no GDP; (m).5 mm GDP; (v) 1 mm GDP and 2 mm GDP. Concentrations of GTP as given on the abscissa. Other conditions as described in A. The apparent equilibrium constant KS was calculated as described in Fig.8 for the rate constants of phosphoprotein formation kl, the rate constants of ATP formation k- and the rate constant for the spontaneous decay of phosphoprotein they can be deduced from the observed half times (Fig. 9). The rate of phosphoprotein formation and its decay induced by the addition of EGTA alone as well as by the addition of EGTA and GDP simultaneously, is illustrated by Fig.9. The decay of the phosphoprotein following the addition of EGTA alone which is applied to prevent rephosphorylation of the enzyme by the nucleoside triphosphate cannot be neglected as compared to the decay and induced by simultaneous addition of EGTA and GDP. For the spontaneous decay at C a rate constant of.34 sc1 was obtained which is in perfect agreement with the value deduced from the calcium-dependent GTPase activity at "C being 8 nmol. mg-'. min-' which yields with an assumed phosphoprotein level of 4nmol.rng-',.26 s-l for kz. According to Scheme 4, the rate constants kl and k-1 computed from the observed half times depicted in Fig.9 are 6 M-' sc1 and 45 M-' s-' respectively. The ratio of the rate constants k-l/kl = 1 agrees quite well with the equilibrium constant Ks obtained from phosphoprotein measurement at "C (Fig. 9C). Because of the various uncertainties inherent in this kind of esti- Table 2. Dependence on temperature of GTP-GDP exchange The assays contained 2 mm GTP, 2 mm GDP, 5 mm MgC12, 2 mm histidine adjusted to ph 7. at the respective temperature, 4 mm KCI,.1 mg vesicular protein/ml Temperature "C Rate of exchange pmol. mg-'. min-' mate this agreement seems remarkable. Yet, surprisingly, the values of kl and k-1 are quite high as compared to those computed from steady state measurements at 2 C (Table l). This would indicate a relatively small temperature coefficient of the exchange reaction which, however, contradicts the experimental findings. Table 2 shows that the exchange reaction has approximately the same high temperature coefficient as calcium-dependent NTP hydrolysis and calcium transport [27]. If a temperature coefficient of 2.5, which is a minimum value, is applied to transform the rate constants found under pseudo equilibrium conditions at 2"C, values are obtained which are distinctly smaller than the values determined by the

10 ~~ ~~~ ~~~~ ~ 62 Phosphate Turnover of Sarcoplasmic Reticulum Membranes Supported by ATP and GTP Table 3. Independence of NTP-NDP exchange of inorganic phosphate The assay medium contained 4 mm KCI, 2 mm histidine ph 7. and MgClz as described in Materials and Methods. The ATP-ADP exchange assay contained.2 mg protein. ml-', The protein concentration in the GTP-GDP exchange assay was.2 mg. ml-', temperature = 22 C Exchange Nucleo- Nucleo- Inorganic Velocity side side phosphate diphos- triphosphate phate Table 4. Dependence on calcium concentration of NTP-NDP exchange of closed and open sarcoplasmic reticulum vesicles The assay media contained 4 mm KCI, 2 mm histidine, ph 7. and MgClz as described in Materials and Methods. The media contained [ATP] = [ADP] = 2 mm and GTP = I mm, GDP = 4 mm respectively. The assay with 'zero' calcium contained.2 mm EGTA. The low calcium concentrations were adjusted by calcium- EGTA buffers containing.2 mm total EGTA. The vesicles were opened by the addition of 5 pmol ionophore X537A to the assay medium before incubation. Protein concentration: ATP-ADP exchange.2 mg/ml, GTP-GDP exchange.2 mg/ml mm ~- ATP-ADP.5. I GTP-GDP 4. 1.o 4. 1.o 4. 1.o o o o pmol. mg protein-'. min-' Ca2+ mm NTP-NDP exchange with closed vesicles open vesicles ATP-ADP GTP-GDP ATP-ADP GTP-GDP pmol. mg protein-'. min-' 'Zero' initial rate measurements at "C. Phosphorylation and dephosphorylation experiments performed at 2 "C with a rapid mixing device confirm the disparity of the two approaches for estimating the rate constant (Table 1) (E. Fassold, W. Waas, and W. Hasselbach, unpublished results). Dependence on Phosphate and Calcium Ions of the NTP-NDP Exchange Reaction Besides the concentrations of nucleoside diphosphates and triphosphates, the activities of the calcium transport system depend on the concentration of the other reaction participants, calcium and phosphate ions. Concentrations of phosphate as they are required for the activation of NTP synthesis do not markedly interfere with ATP [29] or with GTP hydrolysis (not shown). Likewise, the results collected in Table 3 demonstrate that neither the ATP-ADP nor the GTP- GDP exchange are influenced by phosphate when present at such concentrations. This finding shows that inorganic phosphate does not significantly affect the interaction of these nucleotides with the enzyme. The requirement of ionized calcium for the ATP- ADP exchange is well documented [5,3]. Maximum rates were observed at a pca of 6.. As shown in Table 4, the GTP-GDP exchange has a similar calcium requirement. The exchange activity is low atpca above 7, it reaches its maximum value between 6 and 5. If the concentration of ionized calcium in the assay exceeds 1 pm, the NTP-NDP exchange reactions start to decline and inhibition becomes complete when millimolar concentrations of calcium ions are present. Hence, inhibition occurs in approximately the same concentration range of ionized calcium in which the calcium-dependent phosphate liberation and the calcium transport are inhibited. Yet, while the activity of the calcium-dependent NTPase of closed vesicles becomes depressed, because the concentration of ionized calcium rises in their internal space [31-341, the exchange measured under the same conditions seems to be fully activated. Various observations indicate that high concentrations of calcium ions in the external solution do not interfere with the phosphorylation of the enzyme by ATP [35,36]. Interrelationship between Calcium-Dependent NTP Hydrolysis and NTP-NDP Exchange Table 5 shows that the rates of calcium-dependent ATP and GTP cleavage ofclosed vesicles are in the same low range of E.5 pmol. mg-'. min-l. The total calcium concentration in the assay media was sufficiently high to prevent the solution being depleted of calcium by the vesicles. Such low calcium-dependent activities of closed vesicles are observed only when freshly prepared vesicles (24 h) were used. When the vesicles are made permeable for calcium ions by the calcium ionophore X537A (2 pm), the rate of cleavage rises at 22 C for ATP to approximately 1 pmol.mg-' ~min-'andforgtpto.6pmol~mg-'~min-' when the triphosphates were regenerated by phosphoenolpyruvate and pyruvate kinase. Addition of 2 mm

11 ~~ ~~~ N. Ronzani, A. Migala, and W. Hasselbach 63 Table 5. Nucleoside triphosphatase activity of' closed and opened sarcoplasmic reticulum vesicles Nucleotides NTPase activity basal splitting extra splitting closed and open closed open vesicles v e s i c 1 e s vesicles pmol P,. mg protein-' min-' ATP " ATPIADP GTP " GTP/GDP a The NTPase activity was measured in the presence of 2 mm phosphoenolpyruvate and 2 pg. ml-' of ATP pyruvate phosphotransferase at 22 "C. The assay media contained histidine 2 mm, CaC12.1 mm, MgClz 5.mM, KC1 4mM, and NTP 2mM. ADP or GDP were added at a final concentration of 2 mm Table 6. Relationship between NTPuse activity, NTP-NDP and NTP-Pi exchange reactions of sarcoplasmic reticulum vesicles NTP-Pi exchange was measured as described in Materials and Methods at.2 pm ionized calcium in media containing 2 mm NTP, 2 mm NDP, 5 mm 32P-labelled inorganic phosphate, 5 mm MgClz, 2 mm histidine ph 7., 4 mm KCI, t = 22 "C. NTPase and NTP-NDP exchange were measured as described in Table 3 and Fig. 1 Activities measured Adenine Guanine nucleotides nucleotides pmol mg protein-'. min-' Closed vesicles NTPase activity NTP-NDP exchange NTP-Pi exchange.6.35 Opened vesicles NTPase activity NTP-NDP exchange NTP-Pi exchange.2.4 NDP considerably diminishes the rate of NTP hydrolysis of opened vesicles whereby ATP hydrolysis is more strongly reduced than GTP hydrolysis. In contrast, NTP cleavage of closed vesicles is not affected by the respective NDP species. The increase of the NTPase activity after relieving calcium inhibition by making the vesicles calcium-permeable is accompanied by a decrease of the exchange rates. The ATP- ADP exchange declines by approximately 15 % from 3.8 pmol. mg-'. min-' to 3 pmol. mg-'. min-' while GTP-GDP exchange declines by 6 % which, however, only corresponds to a decline of %.15 pmol. mg-'-. min-' (Table 6). The opening of the vesicles should lead to a decline of the phosphoprotein level when the acceleration of phosphate liberation catches up with the rate of phosphorylation at reduced NTP concentrations. This decline is more easily to verify for GTP than for ATP. The phosphoprotein level starts to fall when the GTP concentration becomes lower than.1 mm and reaches 5 % of its maximum value at 4 pm GTP. According to Scheme 4, this concentration reflects the ratio of the rate constant for the formation of phosphoprotein kl and its decay by hydrolysis k2: [GTP][EPI~~~/z = kz/kl. The decay rate kz can be calculated from the phosphoprotein level (2 nmol. mg-') and the rate of phosphate liberation (.3 pmol. mg-'. min-') both measured at 3 pm GTP (Fig. 5 and 8). The obtained value 2.5 s-' yields when introduced in the above relation, an approximate second-order rate constant of kl = 6. lo4 M-' SKI. The same experiment performed in the presence of GDP makes it possible to compute k-' = 3. lo4 M-'s-'. Thus, kl and k-1 increase approximately 5-fold when GTP hydrolysis is turned on, reaching values which are similar to those obtained by initial rate measurements (Table 1). Table 7. Dependence on calcium concentration of NTP-Pi exchange reaction The incubation medium contained 2mM GTP, 2mM GDP, 5 mm 32P-labelled phosphate, 5 mm MgC12, 2 mm histidine ph 7., 4 mm KCI,.2 pm Ca2+ was adjusted with EGTA. t = 22 C + Ca2 Closed vesicles Opened vesicles ATP-P, GTP-P, ATP-P, GTP-P, mm nmol NTP-P, mg protein-' min-' Phosphorylation by Inorganic Phosphate of NDP during NTP-NDP Exchange - NTP-P Exchange In contrast to the exchange of phosphate between NTP and NDP the incorporation of inorganic phosphate into the ATP fraction only takes place with appreciable rates as long as the vesicles are closed. Phosphate incorporation has been observed either as NTP net formation when calcium is released from calcium-loaded vesicles in the presence of NDP, or as NTP-Pi exchange when in the simultaneous presence of NTP and NDP net calcium accumulation has ceased, that means under quasi-equilibrium conditions [l, 37,381. Under the latter conditions the incorporation of inorganic phosphate into GTP and ATP was measured and the results are compared with the rates of NTP hydrolysis and NTP-NDP exchange. The sum of the latter activity represents the total phosphate turnover of the enzyme. Tables 6 and 7 show that phosphate incorporation into both nucleoside diphos-

12 64 Phosphate Turnover of Sarcoplasmic Reticulum Membranes Supported by ATP and GTP a, m a, x +- a [Phosphate] (mm) Fig. 1. Dependence of NTP-supported phosphate incorporation into NDP on the concentration of inorganic phosphate. Sarcoplasmic reticulum vesicles.1 mg. ml-' were incubated in media containing 1 mm MgCIz and the following nucleotide concentrations: () 5 mm ATP, 2 mm ADP; (A).5 mm ATP, 2 mm ADP; (m) 5 mm GTI' 2 tnv GDP; (A).5 mm GTP, 2 mm GDP; t = 22 "C phates measured in the presence of 5 mm inorganic phosphate is a much slower process than the exchange of the terminal phosphate between nucleoside diphosphate and triphosphate. GTP is labelled much more rapidly than ATP. For closed vesicles the rate of incorporation of inorganic phosphate into ATP is 6 times smaller than the rate of phosphate turnover. In the case of GTP this ratio is only 8. As compared to GTP, ATP seems to counteract incorporation of inorganic phosphate. This kind of competition is more clearly revealed when the dependence of phosphate incorporation on the concentration of phosphate is studied (Fig.1). In the presence of.5-5 mm ATP the apparent phosphate affinity seems to be so low that the affinity constant cannot be evaluated. Phosphate incorporation rises increasingly with the phosphate concentration. In contrast the phosphate incorporation into GTP saturates with rising phosphate concentrations and an apparent affinity of z 1 M-' is obtained. The same affinity of the enzyme for phosphate was observed during ATP and GTP net synthesis (not shown, cf. [l]). Since neither the splitting of ATP or GTP by open vesicles nor the NTP-NDP exchange by closed vesicles is inhibited by inorganic phosphate the competition between ATP and inorganic phosphate indicates that the enzyme form which accepts inorganic phosphate does not exist in appreciable quantities during NTP splitting and does not interfere with NTP-NDP exchange. When the vesicles are opened the rate of incorporation of inorganic phosphate into the NDP species drops considerably, but is much more pronounced for GDP than for ADP. Therefore, the ratio of NTP turnover over/phosphate incorporation increases from 6 to 15 for ATP and from 8 to 15 for GTP (Table 6). This change cannot be reversed by the addi- tion of high calcium concentrations to open vesicles (Table 7) in contrast to observations of de Meis and Carvalho [39] who used IDP as phosphate acceptor. This behaviour must be attributed to the inhibition exerted by high concentrations of calcium ions on the NTP-NDP exchange resulting from the reduction of the concentrations of unliganded NDP by calcium NDP formation. DISCUSSION The most remarkable results of this study on the calcium-transporting membranes of the sarcoplasmic reticulum concern the characteristics of the phosphate turnover catalyzed by the membranes under different functional conditions. The total phosphate turnover has been measured as a sum of the rate of phosphate transfer between nucleoside diphosphate and triphosphate (NTP-NDP exchange) and the rate of phosphate liberation. To manipulate the rate of phosphate liberation, use was made of its inhibition by high intravesicular concentrations of ionized calcium. This kind of inhibition always occurs when closed sarcoplasmic vesicles actively accumulate calciurn ions [ As demonstrated, closed vesicles liberated inorganic phosphate in the presence of calcium ions from ATP and GTP with approximately the same rate, yet they transfer phosphate between the respective nucleoside triphosphates and diphosphates under the same conditions with very different velocities, being tenfold higher for the adenine than for the guanine nucleotides. The nearly identical low values found for the rate of ATP and GTP splitting by closed vesicles strongly indicate that nucleotide-specific steps in the hydrolytic pathway (Scheme 1) have lost their modifying function [ It is therefore natural to assume that nucleotide specificity must reside in the reaction transferring the phosphate residue from the nucleoside triphosphate to the nucleoside diphosphate itself. When the vesicles were made permeable for calcium ions the rates of nucleoside triphosphate hydrolysis increase 7-2-fold and concomitantly the rate of phosphate exchange declines (Table 6). Owing to the high ATP-ADP exchange rate of approximately 3 pmol. mg protein-'. min-' its relative reduction does not exceed 15 %. The increase of the rate of GTP hydrolysis by.3-.5 pmol. mg-.'. min-' leads, however, to a decline in the relatively slow exchange rate of 6% from.25 to.1 pmol. mg-'. min-'. When calcium-dependent GTP hydrolysis is turned on by phospholipase A2 digestion the GTP-GDP exchange becomes unmeasurably small. The fact that ATP-ADP exchange is only little reduced when the membrane vesicles are made permeable strongly argues against the idea that high internal calcium concentrations are necessary for the transfer of the protein-bound phosphoryl group to NDP during NTP-

13 N. Ronzani, A. Migala, and W. Hasselbach 65 NDP exchange [2,1]. When the total phosphate turnover is considered we find that it remains nearly constant for the adenine nucleotides but it increases for the guanine nucleotides when NTP hydrolysis can proceed with full speed. This behaviour of the phosphate turnover can be reconciled only under certain conditions with the reaction Scheme 1 in which the transition between unphosphorylated states (E" + E) of the enzyme has been assumed to be rate-limiting [37,38]. The increase of the phosphate turnover observed when the hydrolysis of GTP is turned on requires that the concentration of the GTP accepting intermediate 'E' increases at the expense of the other silent intermediate. When on the other hand, the phosphate turnover does not rise when the enzyme starts splitting ATP the rate of formation and decomposition of E must remain equal. This can only be the case when the concentration of the intermediate E - P declines whilst the concentration of the precursors of E increases. The connotation that intermediates other than those directly involved in the NTP-NDP exchange exist, gets further support from the remarkable difference between the rate of phosphate turnover during steady state and the initial rate of phosphoprotein formation. The trivial explanation that the exchange rate is low solely due to the slowness of the reverse reaction does not apply. The values of the rate constants k-1 respectively k-3 (Schemes 2, 3), describing E - P dephosphorylation by GDP or ADP determined by initial rate measurement are not much different from those for E - P formation (Table 1). The disappearance of the great differences between the initial rate of phosphoprotein formation and the rate of phosphate turnover at pseudo equilibrium indicates that the accumulated calcium ions are presumably responsible for the low rates near equilibrium. The great shift of the enzyme's affinities for the nucleoside phosphates as well as inorganic phosphate [29,41,42] observed after calcium accumulation might therefore be directly related to the discussed activity constraint. The assumed silent intermediates might be participants of the exchange reaction itself like E. T and E - P. D [Eqn (l)]. An accumulation of these intermediates is indicated if the apparent equilibrium constant K, depends on the substrate concentration. This is not the case when the enzyme reacts with GTP-GDP or ITP-GDP but becomes apparent for the adenine nucleotides. The observation that the calcium-dependent NTPase activity and the concomitant calcium transport are very similar for GTP and ATP, although the corresponding nucleoside diphosphokinase activites differ by a factor of 1, raises doubts as to a direct coupling between nucleoside diphosphokinase activity and calcium translocation [2,43]. These doubts are strengthened by the result of simultaneous measure- ments of the rate with which calcium ions are exchanged between the medium and the vesicles and the rate of NTP-NDP exchange at pseudo equilibrium. Although the NTP-NDP exchange rates differ considerably, the rates of calcium exchange are of the same magnitude [28]. In contrast to the rate of phosphate turnover the rate of NTP synthesis during calcium efflux seems not to depend on the nature of the phosphate accepting NDP. Under comparable conditions at 2 C and at optimal concentrations of ADP, IDP and GDP the same low rates of z.1 pmol. mg-'. min-' have been reported for the calcium-efflux-driven NTP synthesis. Evidently a reaction step prior to the transfer of phosphate to the nucleoside diphosphates limits the rate of nucleoside triphosphate synthesis. From the optimal rate of NTP synthesis and the corresponding phosphoprotein level of only.5 nmol. mg-' a turnover rate of z 3-4 s-' (2 "C) results. This rate is 1 times slower than the rate of phosphate exchange between ATP and ADP. The step which determines the rate of NTP synthesis is most likely the phosphorylation of the enzyme by inorganic phosphate energized by the existing calcium gradient. The importance of the special separation of intravesicular and extravesicular calcium for the phosphorylation of NDP species is stressed by the fact that calcium-loaded vesicles present in media containing low concentrations of ionized calcium catalyze a much faster NTP-Pi exchange than open vesicles [39,44]. The saturation of the internal and the external calcium binding sites of the sarcoplasmic reticulum membranes by high concentrations of calcium does not substantially activate incorporation of inorganic phosphate neither into GTP nor into ATP. The fact that the phosphate-ntp exchange is not stimulated is due to the severe inhibition of phosphate transfer to NDP because in the presence of high concentrations of calcium in the medium Ca-NDP is formed which cannot accept phosphate. REFERENCES 1. Hasselbach, W. (1978) Biochim. Biophys. Acta, 515, Tada, M., Yamamoto, T. & Tonomura, Y. (1978) Physiol. Rev. 58, Whittam, R. & Chipperfield, A. R. (1975) Biochim. Biophys. Acta, 415, Hasselbach, W. & Makinose, M. (1962) Biochem. Biophys. Res. Commun. 7, Makinose, M. (1966) Biochem. Z. 345, Inesi, G., Maring, E., Murphy, A. J. & MacFarland, B. H. (197) Arch. Biochem. Biophys. 138, , 7. Makinose, M. (1973) FEBS Lett. 37, Hasselbach, W. & Makinose, M. (1963) Biochem , de Meis, L. 8 Hasselbach, W. (1971) J. Bid. Chem. 246, Rockstein, M. &Herron, W. (1951) Anal. Chem. 23,

14 66 N. Ronzani et al. : Phosphate Turnover of Sarcoplasmic Reticulum Membranes Supported by ATP and GTP 11. Makinose, M. (1969) Eur. J. Biochem. 1, Beil, F. U., von Chak, D. & Hasselbach, W. (1977) Eur. J. Biochem. 81, Avron, M. (196) Biochim. Biophys. Acta, 4, Schwarzenbach, G. (196) Die Komplexometrische Titration, F. Enke, Stuttgart. 15. Martell, A. E. (1971) in Stability Constants ojmeta1-ion Complexes, pp , The Chem. Soc., London. 16. Adolfsen, R. & Moudrianakis, E. N. (1978) J. Biol. Chem. 253, Fromm, H. J. (1975) Mol. Biol. Biochem. Biophys Hotta, K., Brahms, J. & Morales, M. (1961) J. Am. Chem. Soc. 83_, Glynn, K. M. & Chapell, J. B. (1964) Biochem. J. 9, Meissner, G. (1973) Biochim. Biophys. Acta, 298, de Meis, L. & de Nullo, M. C. F. (1973) J. Biol. Chem. 248, Froehlich, J. P. & Taylor, E. W. (1975) J. Biol. Chem. 25, , 23. Makinose, M. & Boll, W. (1979) Cation Flux across Membranes (Mukohata, Y., ed.) pp. 89-1, Acad. Press, New York. 24. Kanazawa, T., Yamada, S., Yamamoto, T. & Tonomura, Y. (1971) J. Biochem. (Tokyo) 7, Hasselbach, W. (1972) in Molecular Bioenergetics and Macromolecular Biochemistry (Weber, H. H., ed.) pp , Springer-Verlag Berlin, New York. 26. Vergorski-Almeida, S., Kurzmack, M. & Inesi, G. (1978) Biochemistry, 17, Inesi, G. & Watanabe, S. (1976) Arch. Biochem. Biophys. 121, Reference deleted. Beil, F. U., von Chak, D., Hasselbach, W. & Weber, H. H. (1977) Z. Naturforsch. 32c, Suko, J. & Hasselbach, W. (1976) Eur. J. Biochem. 64, Makinose, M. & Hasselbach, W. (1965) Biochem. Z. 343, Weber, W. (1971) J. Gen. Physiol. 57, Hasselbach, W. (1966) Ann. New York. Acad. Sci. 137, tkemoto, N. (1974) J. Biol. Chem. 249, Froehlich, J. &Taylor, E. W. (1976) J. Biol. Chem. 251, Souza, D.. G. & de Meis, L. (1976) J. Biol. Chem. 251, Carvalho, M. G. C., de Souza, D. G. & de Meis, L. (1976) J. Biol. Chem. 251, Shigekawa, M. & Dougherty, J. P. (1978) J. Biol. Chem. 253, de Meis, L. & Carvalho, M. G. C. (1974) Biochemistry, 13, Sumida, M., Wang, T., Mandel, T., Froehlich, J. B. & Schwartz, A. (1978) J. Biol. Chem. 253, , de Meis, L. (1976) J. Biol. Chem. 251, Punzengruber, C., Prager, R., Kolossa, M., Winkler, F. & Suko, J. (1978) Eur. J. Biochem. 92, Makinose, M. (1975) Calcium Transport in Contraction and Secretion (Carafoli, E., ed.) pp , North Holland, Amsterdam. Makinose, M. (1971) FEBS Lett. 12, N. Ronzani, A. Migala, and W. Hasselbach*, Abteilung Physiologie, Max-Planck-Institut fur Medizinische Forschung, JahnstraIje 29, D-69 Heidelberg, Federal Republic of Germany * To whom correspondence should be addressed