Characterization of the Inhibition of Intracellular Ca2+ Transport ATPases by Thapsigargin”

The effects of thapsigargin (TG), a specific inhibitor of intracellular Ca2+-ATPases, were studied on vesicular fragments of sarcoplasmic reticulum (SR) mem-branes. Inhibition of Ca2+ transport and ATPase activity was observed following stoichiometric titration of the membrane bound enzyme with TG. When Ca2+ binding to the enzyme was measured in the absence of ATP, or when one cycle of Ca2+-dependent enzyme phosphorylation by ATP was measured under condi- tions preventing turnover, protection against TG by Ca2+ was observed. The protection by Ca2+ disappeared if the phosphoenzyme was allowed to undergo turnover, indicating that a state reactive to TG is produced during enzyme turnover, whereby a dead end complex with TG is formed.

The effects of thapsigargin (TG), a specific inhibitor of intracellular Ca2+-ATPases, were studied on vesicular fragments of sarcoplasmic reticulum (SR) membranes. Inhibition of Ca2+ transport and ATPase activity was observed following stoichiometric titration of the membrane bound enzyme with TG. When Ca2+ binding to the enzyme was measured in the absence of ATP, or when one cycle of Ca2+-dependent enzyme phosphorylation by ATP was measured under conditions preventing turnover, protection against TG by Ca2+ was observed. The protection by Ca2+ disappeared if the phosphoenzyme was allowed to undergo turnover, indicating that a state reactive to TG is produced during enzyme turnover, whereby a dead end complex with TG is formed.
Enzyme phosphorylation with Pi, ATP synthesis, and Ca2+ efflux by the ATPase in its reverse cycling were also inhibited by TG. However, under selected conditions (millimolar Ca2+ in the lumen of the vesicles, and 20% dimethyl sulfoxide in the medium) TG permitted very low rates of enzyme phosphorylation with Pi and ATP synthesis in the presence of ADP. It is concluded that the mechanism of ATPase inhibition by TG involves mutual exclusion of TG and high affinity binding of external Ca2+, as well as strong (but not total) inhibition of other partial reactions of the ATPase cycle. TG reacts selectively with the state acquired by the ATPase in the absence of Ca2+. This state is obtained either by enzyme exposure to EGTA, or by utilization of ATP and consequent displacement of bound Ca2+ during catalytic turnover. Thapsigargin (TG),' a plant-derived sesquiterpene lactone (Christensen et al., 1981) with tumor-promoting properties, interferes with control of the intracellular Ca2+ concentration through inhibition of sequestration mechanisms (Thastrup et al., 1990). TG inhibits Ca2+ transport by SR vesicles isolated from skeletal muscle Kijima et al., 1991), as well as all isoforms of the endo-and sarcoplasmic reticulum Ca2+ (SERCA) pumps (Campbell et al., 1991;Lytton et al., 1991). TG is a specific inhibitor of intracellular Ca2+ pumps inasmuch as neither the plasma membrane Ca2+ pump, nor other cation pumps are affected (Lytton et al., 1991). The importance of TG as an experimental tool lies in its ability to interfere specifically with intracellular Ca2+ trapping systems, and the possibility to test the effect of this interference on the large number of Ca2+-dependent intracellular functions. TG is also an important tool for the molecular characterization of the coupling mechanism of catalysis and transport in the Ca2+-ATPase.
Owing to natural abundance and high density of pump units, vesicular fragments of SR obtained from skeletal muscle constitute an ideal system for mechanistic studies of TG inhibition. Experimentation with this system revealed that TG binds to the SR ATPase stoichiometrically with very high affinity, producing inhibition of Ca2+ binding in the absence of ATP and of catalytic activity when ATP is added . Furthermore, TG reverses the intrinsic fluorescence signal produced by Ca2+ binding to the ATPase, and allows formation of bidimensional crystalline arrays of the ATPase even in the presence of Ca2+ which normally interferes with crystallization (Sagara et al., 1992).
We are here describing a series of experiments on the characterization of the TG inhibition, including its antagonism with Ca", its preferential interaction with a specific state of the ATPase, and its inhibition of various partial reactions of the catalytic and transport cycle. With these experiments we tested overall ATPase and transport activities, as well as the following partial reactions.
These reactions, which were selected for their experimental accessibility, are likely to include conformational transitions that may account for vectorial translocation of bound Ca2+ and certain kinetic features of the system (de Meis and Vianna, 1979;Inesi et al., 1980;Petithory and Jencks, 1986). Conformational transitions are also suggested by our experiments with TG, and will be considered in the discussion of the experimental results.

MATERIALS AND METHODS
Vesicular fragments of longitudinal SR were obtained from rabbit skeletal muscle according to Eletr and Inesi (1972). Thapsigargin was purchased from LC Services Corp., Woburn, MA. All other chemicals were purchased from Sigma. Cr(H20)4ATP bidentate was prepared by the method of Dunaway-Mariano and Cleland (1980). Protein concentration was measured by the method of Lowry et al. (1951), standardized with bovine serum albumin. ATP-dependent Ca2+ uptake by SR vesicles was followed by measuring fluxes of radioactive 4sCa tracer. For this purpose, the vesicles were incubated in reaction mixtures specified in the Figure legends, and serial samples were collected on 0.45-pm Millipore filters and washed three times with 5 ml of cold 2 mM LaC13 plus 10 mM MOPS, p H 7.0. The filters were then dissolved in 0.5 ml of dimethyl formamide, and processed for scintillation counting.
ATPase activity was followed by measuring production of P, according to Lin and Morales (1977), and velocity was estimated from the slope of several time points on a straight line.
Calcium binding by the SR ATPase in the absence of ATP was determined by placing SR vesicles (100 pg of protein) on a 0.65-pm Millipore filter and perfusing them with pertinent reaction mixtures (see Figure legends). The perfusion time was electronically controlled with the aid of a Bio-Logic rapid filtration apparatus. The filters were then collected for determination of radioactive (' "a) tracer associated with the vesicles.
Calcium occlusion by the ATPase in the presence of CrATP was obtained in a reaction mixture including 100 mM MOPS, pH 7.0, 80 mM KCl, 0.1 mM [4sCa]CaC12, 0.2 mM EGTA, 2 p M A23187, and 0.2 mg of SR protein/ml. The reaction was started by the addition of 0.5 mM CrATP at 25 "C, and 0.25-ml samples were taken at serial times. These samples were placed in 2 mM LaC13 plus 10 mM MOPS, pH 7, and, after 2 min, filtered through 0.45-pm Millipore filters. The filters were washed three times with 4 ml of 2 mM LaC13 plus 10 mM MOPS, pH 7, and processed for determination of 4sCa by scintillation counting. Blanks obtained in the absence of CrATP were subtracted from the values obtained with CrATP. Alternatively, small aliquots (microgram range) of quenched protein were placed on 0.45-pm Millipore filters, washed extensively with cold 0.1 N HC1 and cold HZO, and processed for determination of radioactivity on the filters.
For studies of the reversal of the Ca2+ pump, SR vesicles (35-45 pg of protein/ml) were preloaded with Ca2+ in a reaction mixture containing 50 mM MOPS-Tris, 10 mM MgC12, 20 mM Pi-Tris, 0.3 mM CaC12 (in the presence or in the absence of radioactive 4sCa tracer), and 5 mM ITP. Following 40 min incubation at 37 "C, 5.0-ml aliquots were centrifuged in a refrigerated centrifuge and each sediment was resuspended in 0.45 ml of ice-cold water. 50-pl (or larger) aliquots of this suspension were then diluted in 1.5 (or more) ml of efflux medium containing 50 mM MOPS-Tris, pH 7.0, 10 mM MgCl,, 4.0 mM Pi-Tris (in the presence or in the absence of radioactive 32P tracer), 2 mM EGTA, 10 mM glucose, and 20 pg of hexokinaselml, in the presence or in the absence of 0.2 mM ADP, a t 37 "C. Serial samples were then filtered and washed with LaC13 for determination of 4sCa remaining associated with the vesicles, or quenched with an equal volume of 1 M PCA plus 4 mM Pi, for determination of ATP synthesis ( de Meis, 1988).

RESULTS
Apparent Irreversibility of ATPase Inhibition by Thapsiargin-TG reacts with the SR ATPase with very high affinity, resulting in enzyme inhibition which is proportional to stoichiometric titration of the ATPase with TG . We found that a tight stoichiometric ratio (1:l) of added TG and inactivated ATPase is obtained when aliquots of TG dissolved in dimethyl sulfoxide are added directly to a n aqueous suspension of SR vesicles, avoiding preliminary dilution of TG in aqueous media in the absence of SR vesicles.
In the experiment shown in Fig. 1 we added to a concentrated suspension of SR vesicles an amount of TG stoichiometrically equivalent to approximately half of the Ca2+-ATPase. The resulting partial inhibition of Ca2+ transport activity corresponds to stoichiometric titration of the ATPase with TG. Under these conditions the concentration of free TG in the medium is expected to be near its Kd and, therefore, the extent added. After 5 min incubation this mixture was diluted 10,000-fold with the same medium (without protein or TG), and Caz+ transport activity was initiated by the addition of 2.5 mM ATP at 25 "C. Serial samples were taken thereafter for filtration and determination of calcium trapped in the vesicles. In other cases (A, A, 0, W) the diluted mixture was left for 2 h at 25 "C before adding ATP and starting the Ca2+ transport experiment. Samples (0 and 0) were tested without the 2-h incubation.
of protein titration should be very sensitive to changes in the concentration of free TG. Nevertheless, a 10,000-fold dilution of the inhibited protein suspension with medium containing no TG, resulted in no reversal of inhibition within a period of 2 h following dilution. This indicates that dissociation of the TG-enzyme complex is very slow, consistent with the high affinity of the inhibitor for the enzyme. On the other hand, if the ATPase treated with TG was digested extensively with trypsin or denatured with urea, TG was recovered in the washing medium of the denatured preparation, as demonstrated by inhibition of a new batch of SR vesicles. No TG was recovered by washing TG-treated vesicles with nondenaturing media (data not shown).
Inhibition of Steady State Activity in the Presence of Various Concentrations of ATP and Ca2'-When steady state ATPase velocity was measured in the presence of various ATP concentrations, we found that TG inhibition was independent of the ATP concentration ( Fig. 2 A ) . By comparison, a clearly competitive behavior (Fig. 2 A ) is displayed by cyclopiazonic acid which is another inhibitor of the SR ATPase (Seidler et al., 1989). This suggests that the TG inhibition is not related to the concentration of ATP present in the medium. We also found that, under steady state conditions, the ATPase inhibition by TG is not reversed by addition of higher Ca2+ concentrations (Fig. 2B). Nevertheless, in pre-steady state studies of partial ATPase reactions we noted that the time course of enzyme inactivation by TG was dependent on the order of reagent additions to the reaction mixture.
Inhibition of Ca2+ Binding in the Absence of ATP-High affinity binding of external Ca2+ may be considered the first partial reaction of the ATPase catalytic cycle (see reaction sequence in the Introduction), as the bound Ca2+ produces enzyme activation and in turn undergoes vectorial displacement upon enzyme phosphorylation by ATP. In the absence of ATP, Ca2+ binding to the enzyme can be measured under equilibrium conditions as a self-limited partial reaction which we found previously to be inhibited by TG (Sagara and Inesi,

1991). We now find that preincubation with
Ca2+ has a protective effect against TG.
In the experiment shown in Fig. 3, we incubated SR vesicles with TG in the presence of various concentrations of nonradioactive Ca'+ for increasing time intervals, before placing them on a rapid filtration apparatus for removal of the incubation medium. The vesicles remaining on the filters were then perfused for 0.6 s with a medium containing radioactive Ca2+ (isotopic calcium exchange is quite rapid under these conditions), to test their residual Ca2+ binding activity. We found that inhibition of Ca2+ binding was inversely proportional to the Ca2+ concentration present during the incubation with TG (Fig. 3). Therefore, in apparent contrast with inhibition of steady state ATPase activity, equilibrium measurements of Ca'+ binding in the absence of ATP show protection of the ATPase by Ca2+ against TG.  Figure. Following 5 min preincubation a t 25 "C, TG (10 nmol/ mg SR protein) was added, and 2-ml samples were filtered a t variable times (indicated in the abscissa) through 0.65-pm Millipore filters placed on a rapid filtration Bio-logic apparatus. The vesicles remaining on the filters were then perfused for 0.6 s with a medium 20 mM MOPS, pH 7.0,80 mM KC1,5 mM MgCl2, and 5 p~ ['5Ca]CaC12. The filters were then collected without any washing for determination of radioactive calcium bound to the vesicles. Blanks were obtained with filters perfused in the absence of vesicles.
Inhibition of Enzyme Phosphorylation with ATP-Addition of ATP to SR vesicles preincubated with Ca2+ leads to formation of a phosphorylated enzyme intermediate by transfer of the ATP terminal phosphate to the ATPase (Yamamoto and Tonomura, 1968). If this reaction is carried out in ice for a short time (e.g. 1 s) and in the absence of ADP, nearly stoichiometric phosphorylation of the enzyme is obtained since the rate of phosphoenzyme formation is faster (Froehlich and Taylor, 1975) and less temperature dependent than its hydrolytic cleavage.
In analogy to the previous experiment on Ca2+ binding, we then incubated SR vesicles a t 15 "C with TG for increasing time intervals in the absence (i.e. presence of EGTA) or in the presence of Ca2+, and then added ATP (plus Ca2+ to ensure its presence during phosphorylation) for 1 s at ice temperature. Even in this case we found that the inhibition was much more pronounced when SR vesicles were exposed to TG in the absence of Ca2+, while much less inhibition was obtained when 10 or 50 ~L M Ca2+ was present during the incubation with TG (Fig. 4A). Figs. 3 and 4A indicate that Ca2+ protects the ATPase from inactivation by TG. We also found (Fig. 4, B and C) that protection by Ca'+ is least effective at high temperature (37 "C) and low pH, pH 6.0, and most effective at low temperature (2 "C) and high pH, pH 8.0. It is likely that these differences are related to the temperature and pH dependence of ATPase conformational changes determining the sensitivity of the enzyme to inhibition by TG.

The experiments shown in
It is of interest that when ATP is added with TG in the presence of Ca2+, and samples are taken at serial time intervals for determination of phosphoenzyme, inactivation occurs over a period of 30 s (Fig. 5 ) , indicating that catalytic turnover produces an enzyme species which is receptive to TG thereby yielding a dead end complex. This explains the apparent discrepancy between Fig. 5 showing that Ca2+ does not protect steady state enzyme activity from TG, and Fig. 4 showing protection under conditions preventing enzyme turnover. Particularly cogent is the comparison between Fig. 5 and the upper curve of Fig. 4B, both obtained at 4 "C and in the presence of 50 PM Ca2+, the only difference being the addition of ATP with TG in the experiment shown in Fig. 5 . Occlusion of Bound Ca2+ by CrATP-ATP utilization by the SR ATPase produces occlusion of bound Ca2+ before its release in the lumen of the vesicles (for review, see Glynn and Karlish (1990)). The occluded state can be obtained in a stable form by reacting the ATPase with CrATP in the presence of Ca2+ (Serpersu et al., 1982;Vilsen and Andersen, 1987). Dissociation of radioactive occluded Ca2+ from the enzyme. CrATP complex is obtained upon addition of EGTA ( Fig. 6) or excess nonradioactive calcium isotope (not shown). Such a dissociation is very slow, and includes two distinct kinetic components likely related to sequential dissociation of the two calcium ions bound sequentially to each mole of ATPase (Inesi et al., 1980;Dupont, 1982;Inesi, 1987). We found that TG prevents Ca2+ occlusion if added to the SR ATPase before CrATP (not shown). On the other hand, addition of TG to the preformed enzyme. CrATP complex is followed by dissociation of occluded Ca2+ with kinetics quite similar to those observed following addition of EGTA (Fig.  6). This suggests that in all cases the decay kinetics of occluded state levels reflect directly the rates of Ca2+ dissociation. Therefore the role of TG (in analogy of that of EGTA) is simply that of preventing Ca2+ exchange following spontaneous dissociation of the occluded Ca'+. While EGTA exerts this effect by chelating free Ca2+ in the medium, TG prevents phosphorylation reaction was quenched after 1 s with an equal volume of 1 M PCA plus 5 mM Pi, and processed for determination of radioactive phosphoenzyme. B , as A with 50 PM Ca", but incubated a t various temperatures with TG. C, as in A with 50 pM Ca2+, but for the different pH which was maintained with a 20 mM BisTris-Tris buffer. For A-C control runs were conducted in the absence of TG. The phosphoenzyme levels obtained in the absence of TG (in the range of 2-3 nmol/mg of protein) were considered to be 100%. further binding by virtue of its own interaction with the enzyme. Since CrATP occupies the catalytic site as a stable complex either in the presence or in the absence of bound Ca'+ (Chen et al., 1991), the effect of TG cannot be attributed to TG binding within the catalytic and/or nucleotide-binding domain which is occupied by CrATP. These experiments also demonstrate that TG cannot affect the enzyme as long as Ca2+ is bound to it in its high affinity state. Inhibition of Enzyme Phosphorylation with Pi-The Ca2+-ATPase of SR can be phosphorylated with Pi instead of ATP provided that the enzyme is deprived of Ca2+ by the addition of EGTA (Masuda and de Meis, 1973;de Meis and Masuda, 1974). This reaction can be measured under equilibrium conditions independent of enzyme turnover, and can be considered the reversal of phosphoenzyme hydrolytic cleavage (see reaction sequence in the Introduction). As observed in equilibrium measurements of Ca2+ binding ; see also Fig. 3), equilibrium measurements of enzyme phosphorylation with Pi show inhibition in parallel with stoichiometric enzyme titration with TG, independent of the concentrations of Pi (Fig. 7) or Mg2+ (not shown) which are substrate and activating cation for this reaction, respectively. Contrary to analogous experiments with Ca2+ (Figs. 3 and 4), no protection is observed (Fig. 8) if enzyme is phosphorylated before adding TG, and if the Pi concentration dependence of the phosphorylation reaction is lowered by the addition of dimethyl sulfoxide (de Meis et al., 1980). It is known that ATPase phosphorylation with Pi can be obtained using either "empty" SR vesicles (as in the experiments shown in Figs. 7 and 8), or vesicles "loaded" with Ca2+ through a preliminary incubation, as long as the Ca2+ concentration in the medium outside the vesicles is lowered below 0.1 FM when the reaction with Pi is started. The Pi concentration dependence of this reaction is lower if loaded (rather than empty) vesicles are used ( de Meis, 1988;Beil et al., 1977). As long as Ca'+ concentration in the medium is very low (cO.1 PM), and depending on whether the Ca2+ concentration in the lumen of the vesicles is sufficiently high (millimolar range), the Pi reaction can occur either as with somewhat different characteristics (Beil et al., 1977;Chaloub et al., 1979). The denotation *E, as opposed to simple E, was introduced by de Meis and Vianna (1979) to distinguish the enzyme state which is promoted by the absence of Ca2+ in the high affinity outward oriented sites of the ATPase.
In the experiment shown in Figs. 7 and 9, we obtained equilibrium levels of 2.3 and 1.9 nmol of E-P/mg of protein with empty and loaded vesicles, respectively. Considering that the catalytic site stoichiometry was 5 nmol/mg in our preparation, and that saturating Pi was used, the observed E-P level suggest a Kz value near 1 in both cases.
As for the effect of TG, we found that the E-P level was reduced to 0.2 and 0.9 nmol/mg protein with empty and loaded vesicles, respectively (as long as Ca2+ in the medium was below 0.1 FM), when TG is added to match the enzyme stoichiometry (Figs. 7 and 9). Considering that the Pi concentration dependence is not changed by TG, and that saturating Pi was used, we conclude that TG lowers the levels of phosphoenzyme by reducing the equilibrium constant for the phosphorylation reaction (Kz in Equations 2 and 3). It is apparent that this reduction is somewhat greater for E.Pi than for ECazin1 Pi.
If the Ca2+ concentration in the medium is raised above micromolar, the Pi reaction is inhibited (Masuda and de Meis, 1973), since the enzyme acquires different characteristics as a consequence of Ca2+ binding to the high affinity sites on the outer surface of the vesicles, as in where *E and E are two distinct states permitting enzyme phosphorylation only by Pi or ATP, respectively. It is of interest that, while in the absence of TG the Pi reaction is totally inhibited by 35 p~ Ca2+ (Fig. 9), addition of TG interferes with the Ca2+ inhibition and brings back E-P to the same level observed with TG in the absence of Ca2+. This is due to TG inhibition of high affinity Ca2+ binding and retention of the *E state by the enzyme. Consequently, the Pi reaction is allowed to occur within the limits imposed by the K2 lowered by TG. It is of interest that the small amount of phosphoenzyme that is formed under these conditions can be utilized to form ATP upon addition of ADP (see below).
Inhibition of ATP Synthesis and Coupled Ca2+ Efflux-It was originally discovered in Hasselbach's laboratory (Barogie et al., 1971) that if SR vesicles preloaded with Ca2+ in the presence of ATP and oxalate are diluted in a medium containing Pi, ADP, and low Ca2+ (i.e. EGTA), a complete reversal of the pump is obtained, beginning with enzyme phosphorylation by Pi and resulting in ATP synthesis and coupled Ca2+ efflux. In our experiments we used vesicles preloaded in the presence of ATP and Pi (instead of oxalate). In this case, owing to a higher solubility of the calcium-phosphate complex in the lumen of the vesicles, a higher rate of Ca2+ efflux is observed upon dilution (Sande-Lemos and de Meis, 1988). A significant efflux component is observed even when the dilution medium does not contain ADP. The efflux increment obtained upon addition of ADP is then stoichiometrically (2:l under favorable conditions) related to ATP synthesis.
When we tested the effect of TG on this system, we found that the ADP-independent efflux, which is an uncoupled function of a slow ATPase channel normally involved in coupled transport (Inesi and de Meis, 1989), was not inhibited by TG. On the other hand, the ADP-dependent increment of Ca2+ efflux, as well as the coupled ATP synthesis were inhibited by stoichiometric titration of the enzyme with TG (Fig.  10, B and C). It is then apparent that TG inhibits the overall catalytic and coupled transport cycling of the SR ATPase, in both the forward and reverse directions. It can be shown, however, that under the conditions of the experiment illustrated in Fig. 9, the very small amounts of phosphoenzyme formed with Pi in the presence of TG can be utilized to form ATP upon addition of ADP (Fig. 11).

DISCUSSION
With the experiments reported above we confirmed that TG interacts stoichiometrically with the SR ATPase producing an apparently irreversible inactivation. We also found that the presence of specific ATPase ligands and substrates has a profound influence on the effect of TG, indicating that TG forms a stable inhibitory complex only with a specific state of the ATPase.
As TG inhibits Ca2+ binding to the ATPase in the absence of ATP, this reaction can be studied under equilibrium conditions and in the absence of catalytic turnover. It is then found that Ca2+, if added to the SR vesicles before TG, has a protective effect (Figs. 3 and 4). This could be explained by the following reaction sequence: where occupancy of the Ca2+ and TG sites on the enzyme is mutually exclusive, and TG reacts only with *E. Owing to the very high association constant of Ca2+ for the enzyme, most of the enzyme is in the ECa2 form in the presence of micromolar Ca2+. Therefore, even though the TG affinity for the enzyme is seemingly stronger than that of Ca2+, the rate of TG. E formation is limited by the concentration of E which is very low if Ca2+ is added before TG. It is also possible that a significant concentration of the species TG E -Ca2 is also formed, which then undergoes transformation to TG -*E and 2 Ca2+ at very low rates.
It is of interest that the protective effect of Ca2+ is less prominent at low temperature and at high pH (Fig. 4), since under these conditions the equilibrium between the two interconverting states *E and E appears to be shifted in favor of *E, as revealed by a higher reactivity to Pi (Masuda and de Meis, 1973).
If TG is added to SR vesicles in the presence of Ca2+ concentrations delaying enzyme inactivation by TG in the absence of ATP (Figs. 4 and 5), a much faster enzyme inactivation is observed upon addition of ATP (Fig. 6). This inactivation is due to production of a reactive state as a consequence of ATP utilization, and trapping of E of the enzyme in a dead end complex. The catalytic and transport cycle of the SR ATPase may be represented as follows.

71
Scheme 6, which is derived from de Meis and Vianna (1979), was augmented by branched pathways for the phosphorylation reactions with ATP or Pi, to allow external Ca2+ to bind before or after ATP (Petithory and Jencks, 1986), and internal Ca2+ to bind before or after Pi (Chaloub et al., 1979). For our present purpose, it was not necessary to include steps allowing for sequential Ca2+ binding (Inesi, 1987) or Ca2+ occlusion by enzyme phosphorylation (Glynn and Karlish, 1990). Whether the phosphorylated enzyme resides in two different conformations corresponding to E and *E, or prevalently in another conformation allowing occlusion of bound Ca2+, is a matter of controversy (Pickart and Jencks, 1982). At any rate, Scheme 6 provides an intuitive appreciation of how the enzyme would reside mostly in the E. Ca2 form in the presence of external Ca2+. Upon addition of ATP, the enzyme would then distribute itself in the various states, including *E which is reactive to TG, thereby permitting accumulation of the stable inhibitory complex TG. *E. The experiments described above indicate that *E (as opposed to E ) is characterized by reactivity to Pi and TG, relative stability in the absence of Ca2+ at low pH and high temperature, and relative instability in the presence of Ca2+ at high pH and low temperature. *E and TG. *E are distinguished from E by the low level of intrinsic fluorescence, and by their ability to form bidimensional crystalline arrays upon addition of decavanadate (Sagara et al., 1992). As opposed to *E, however, TG. *E cannot bind external Ca2+ and, therefore, is unable to utilize ATP (Figs. 4 and 5). We found that TG inhibits also enzyme phosphorylation by Pi (Figs. 7 and 8), which does not require Ca2+ (Masuda and de Meis, 1973). Contrary to Ca2+ binding, however, enzyme phosphorylation with Pi previous to addition of TG, affords no apparent protection from inactivation by TG. Furthermore, depending on the experimental conditions, TG inhibition of the Pi reaction is not complete, but is only manifested by a reduction of the phosphoenzyme level. In some cases (Fig. 9) the level of phosphorylation by Pi and the subsequent formation of ATP (Figs. 9 and 11) are actually increased by TG through interference with the inhibition by external Ca2+. This is especially true when the lumenal Ca2+ concentration is high (loaded vesicles). It is then apparent that the TG. E can bind internal Ca2+ even when binding of external Ca2+ is inhibited, and can react with Pi proceeding to ATP synthesis even though at very low rates.
In accordance with the experiments on enzyme phosphorylation by Pi, those on Ca2+ occlusion by the CrATP.ATPase complex suggest that TG does not bind within the phosphorylation and substrate-binding domain since TG interferes with Ca2+ occlusion in the presence of stable occupancy of the catalytic site by CrATP. Furthermore, these experiments suggest a direct interference of TG with binding of external Ca2+ by the ATPase, since inhibition of Ca2+ occlusion by TG occurs with the same slow kinetics observed with EGTA ( Fig.  6). This indicates that EGTA reacts with external Ca2+ and T G reacts with the enzyme only following spontaneous dissociation of occluded Ca2+, consequently preventing further association events. Therefore, TG can react with the enzyme only after dissociation of occluded Ca2+.
In conclusion, our experiments indicate that TG reacts with a conformational state acquired by the ATPase in the absence of Ca2+, and produces total inhibition of high affinity binding of external Ca". Owing to interference with activation by Ca2+, the TG.ATPase complex is unable to utilize ATP in the forward direction of the cycle.