Insight into the Uncoupling Mechanism of Sarcoplasmic Reticulum ATPase Using the Phosphorylating Substrate UTP*

Ca 2 1 transport and UTP hydrolysis catalyzed by sarcoplasmic reticulum Ca 2 1 -ATPase from skeletal muscle was studied. A passive Ca 2 1 load inside microsomal vesicles clearly decreased the net uptake rate and the final accumulation of Ca 2 1 but not the UTP hydrolysis rate, causing energy uncoupling. In the absence of passive leak, the Ca 2 1 /P i coupling ratio was 0.7–0.8. UTP hydrolysis did not maintain a rapid component of Ca 2 1 exchange between the cytoplasmic and lumenal compart-ments as occurs with ATP. The uncoupling process in the presence of UTP is associated with: (i) the absence of a steady state accumulation of ADP-insensitive phosphoenzyme; (ii) the cytoplasmic dissociation of Ca 2 1 bound to the ADP-sensitive phosphoenzyme; and (iii) the absence of enzyme inhibition by cyclopiazonic acid. All these characteristics confirm the lack of enzyme conformations with low Ca 2 1 affinity and point to the existence of an uncoupling mechanism mediated by a phosphorylated form of the enzyme. Suboptimal coupling values can be explained in molecular terms by the pro-posed functional model. containing 20 m M Mops, pH 7.0, 80 m M KCl, 5 m M MgCl 2 , 1 m M UTP, and sufficient 2 and/or EGTA to yield a final free Ca 2 1 in the external medium of 50 m M . This was calculated taking into account the external Ca 2 1 withdrawn with each aliquot of loaded vesicles. After dilution, the membrane protein concentration was 0.05 mg/ml. The same protocol was applied in experiments performed Ca 2 1 -un-*

The ion motive ATPases are molecular devices involved in energy coupling processes (1). A decrease in energy conservation efficiency gives rise to the phenomenon of uncoupling. A self-evident source of uncoupling is the loss of ionic gradient through ionophoric activity (2)(3)(4), although a more subtle origin may be the catalytic mechanism itself (intramolecular uncoupling). Uncoupling has been observed in different energytransducing systems and has been reported to depend on the experimental conditions tested. For instance, the sarcoplasmic reticulum (SR) 1 Ca 2ϩ -ATPase presents Ca 2ϩ transport/ATP hydrolysis coupling ratios of 2, when measured under presteady state conditions (5). However, steady state measurements have given values ranging from 1.3 to 1.8 (6 -8). The use of GTP (6), ITP (6), or other non-nucleotide substrates such as acetylphosphate (9), p-nitrophenylphosphate (10), methylumbelliferylphosphate (10), or dinitrophenylphosphate (10) provided values near to 1. Likewise, some reports have given coupling ratios of around 2 for UTP and other nonphysiological substrates (11,12).
Ca 2ϩ -ATPase intramolecular uncoupling in the presence of ATP has been related through indirect evidence to the cytoplasmic dissociation of Ca 2ϩ from the phosphorylated intermediate (7,8). However, a rapid efflux of lumenal Ca 2ϩ involving nonphosphorylated species of the enzyme has also been suggested as a possible uncoupling mechanism (13,14). The reaction cycle in the presence of ATP is difficult to analyze because the substrate has complex kinetic effects that may be complicated by the existence of rapid Ca 2ϩ exchange between the cytoplasmic and lumenal compartments. We therefore selected UTP because it behaves in a more straightforward manner when it acts as phosphorylating agent and does not support rapid Ca 2ϩ exchange. In this way, the study of the uncoupling mechanism was facilitated by the use of UTP.
Our experimental system consisted of sealed vesicles that were isolated from skeletal SR membrane. The preparation was enriched in Ca 2ϩ -ATPase protein, retained the native orientation of the membrane (15), and showed low passive permeability (16). In this study we explored different experimental conditions in the presence of UTP as an ATP surrogate. Data analysis of UTP hydrolysis, Ca 2ϩ movement, and phosphoenzyme conformations provided clear illustrations of how the uncoupling mechanism works during enzyme cycling.

EXPERIMENTAL PROCEDURES
Isolation and Quantitation of Microsomes-Fast-twitch skeletal muscle from adult female New Zealand rabbit was used as starting material. Sealed right-side vesicles mainly derived from SR longitudinal tubules were prepared by the method of Eletr and Inesi (17). The final pellet was resuspended at 10 -15 mg protein/ml, aliquoted, and stored at Ϫ80°C until use. This preparation shows no significant level of passive leakage because of the absence of Ca 2ϩ release channel (16). The SR protein was estimated by the colorimetric procedure of Lowry et al. (18) using bovine serum albumin as standard.
Free Ca 2ϩ in the Media-Concentrations of free Ca 2ϩ were adjusted by adding appropriated CaCl 2 /EGTA mixtures according to the computer program developed by Fabiato (19). Calculations were based on the Ca 2ϩ -EGTA absolute stability constant (20) and the EGTA protonation equilibria (21). Relevant Ca 2ϩ ligands and pH in the medium were also considered.
UTP Hydrolytic Activity-The release of inorganic phosphate in the min time scale was measured under stirring at 25°C, according to Lin and Morales (22). The enzymatic reaction was started by diluting aliquots 100-fold of passively Ca 2ϩ -loaded vesicles (0.085 ml) in a medium containing 20 mM Mops, pH 7.0, 80 mM KCl, 5 mM MgCl 2 , 1 mM UTP, and sufficient CaCl 2 and/or EGTA to yield a final free Ca 2ϩ in the external medium of 50 M. This was calculated taking into account the external Ca 2ϩ withdrawn with each aliquot of loaded vesicles. After dilution, the membrane protein concentration was 0.05 mg/ml. The same protocol was applied in experiments performed with Ca 2ϩ -un-* This work was supported by Spanish Ministerio de Educacion y Cultura Grant PB97-1039. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
‡ To whom correspondence should be addressed: Departamento de Bioquimica y Biologia Molecular A, Edificio de Veterinaria, Universidad de Murcia, Campus de Espinardo, 30071 Murcia, Spain. Fax: 34-968-364-147; E-mail: fbelda@fcu.um.es. 1 The abbreviations used are: SR, sarcoplasmic reticulum; Mops, 4-morpholinepropanesulfonic acid; EP, phosphorylated enzyme; TNP-ATP, 2Ј (or 3Ј)-O-(trinitrophenyl)adenosine-5Ј-triphosphate; CPA, cyclopiazonic acid; E 1 P⅐Ca 2 , phosphorylated conformation of the enzyme with high affinity Ca 2ϩ bound; E 2 and E 2 P, unphosphorylated and phosphorylated species with low affinity for Ca 2ϩ , respectively. loaded vesicles. The composition of the dilution media is given in the legend of Fig. 1. The reaction was stopped at various times by adding 1-ml samples of reaction mixture to the nitric acid/molybdovanadate reagent. The Ca 2ϩ -independent activity was evaluated in Ca 2ϩ -unloaded vesicles by including 0.5 mM EGTA in the dilution medium and omitting CaCl 2 .
UTP hydrolysis was also measured in the second time scale by a radiometric procedure. Experiments were carried out at 25°C under stirring, in the absence or presence of 5 mM potassium oxalate. SR vesicles (0.4 mg protein/ml) were initially suspended in a medium containing 20 mM Mops, pH 7.0, 80 mM KCl, 5 mM MgCl 2 , 0.5 mM EGTA, and 0.545 mM CaCl 2 , (50 M free Ca 2ϩ ). In one case, the reaction was started by adding 1 mM [␥-32 P]UTP (ϳ20,000 cpm/nmol) and in the other by adding 1 mM UTP with a pulse of [␥-32 P]UTP (ϳ1 ϫ 10 7 cpm) being delivered 2 min later. Zero time corresponded to the addition of radioactive UTP. In both cases, aliquots of the reaction mixture (0.5 ml) were quenched at different times with 0.5 ml of ice-cold 250 mM perchloric acid plus 4 mM sodium phosphate. The quenched samples were incubated in ice water bath for 5 min. The complete volume (1 ml) was filtered through a 0.45-m nitrocellulose filter, and the filtrate was collected. The subsequent treatment consisted of removing contaminant [␥-32 P]UTP by charcoal, extraction of [ 32 P]phosphomolybdate complex with isobutanol/benzene, and quantitation of 32 P i in the organic phase, according to a previously published procedure (23). The specific activity of standard 32 P i (cpm/nmol) was obtained by correlating the cpm of [␥-32 P]UTP and the chemical concentration of the substrate. When a pulse of [␥-32 P]UTP was added, the UTP concentration after 2 min of reaction (i.e. t ϭ 0) was 0.75 mM, as deduced from the activity data. A 75% yield for the organic extraction of P i and appropriate blank assays were also considered for the transformation of cpm into nmol P i . The extraction yield was experimentally measured in 0.5 ml of reaction medium containing 1 mM [␥-32 P]UTP. The procedure required chemical hydrolysis of the substrate and further processing as described for the samples.
Net Ca 2ϩ Uptake-This was measured at 25°C under stirring and with the aid of radioactive tracer. Vesicles were first passively loaded with 10, 20, or 40 mM 45 Ca 2ϩ , and the uptake was initiated by diluting aliquots 100-fold of 45 Ca 2ϩ -loaded vesicles, as described for the corresponding UTP hydrolysis experiments. After dilution, the reaction mixture consisted of 20 mM Mops, pH 7.0, 80 mM KCl, 5 mM MgCl 2 , 0.05 mg/ml of SR protein, 1 mM UTP, and 50 M free 45 Ca 2ϩ . Total CaCl 2 and/or EGTA in the dilution medium was adjusted according to the Ca 2ϩ concentration used for loading. Loading and dilution media contained 45 Ca 2ϩ at the same specific activity (ϳ10,000 cpm/nmol). Full detail is given in the legend of Fig. 3. Ca 2ϩ -unloaded vesicles were also subjected to the same protocol. In both cases, aliquots containing 1 ml of reaction mixture were filtered under vacuum at different time intervals (scale of min), and the filters were then rinsed with 10 ml of ice-cold La 3ϩ medium (20 mM Mops, pH 7.0, and 1 mM LaCl 3 ). Radioactivity was measured by liquid scintillation counting after solubilizing the filters. The unspecific Ca 2ϩ retained by the filters was subtracted by performing a blank assay in the absence of UTP.
Entry of Ca 2ϩ -UTP-dependent Ca 2ϩ entry was measured in the second time scale as follows: samples of 0.25 ml containing 20 mM Mops, pH 7.0, 80 mM KCl, 5 mM MgCl 2 , 0.545 mM 45 CaCl 2 (ϳ10,000 cpm/nmol), 0.5 mM EGTA (50 M free 45 Ca 2ϩ ), and 0.4 mg/ml of unloaded vesicles were mixed under continuous vortexing at 25°C with 1 mM UTP. The reaction was quenched by adding 5 ml of ice-cold La 3ϩ medium, and the resulting mixture was filtered under vacuum. The filters were rinsed and counted as before. Alternatively, SR vesicles were first actively loaded with 40 Ca 2ϩ at 25°C. The initial reaction medium was 20 mM Mops, pH 7.0, 80 mM KCl, 5 mM MgCl 2 , 0.545 mM CaCl 2 , 0.5 mM EGTA (50 M free Ca 2ϩ ), 0.4 mg/ml of unloaded vesicles, and 1 mM UTP. A 45 Ca 2ϩ pulse, to give ϳ10,000 cpm/nmol, was added 2 min later, zero time being taken as the moment of 45 Ca 2ϩ addition. The reaction was arrested by adding 5 ml of ice-cold La 3ϩ medium to an aliquot of 0.25 ml reaction mixture. Quenched samples were filtered and processed as described before. In some experiments, the reaction medium was supplemented with 5 mM potassium oxalate. Blank assays were performed following one of the described protocols but without the addition of UTP.
Autoradiographic Detection of EP-This was studied under two different assay conditions. Phosphorylation in the presence of Ca 2ϩ was performed at 25°C in a medium containing 20 mM Mops, pH 7.0, 80 mM KCl, 5 mM MgCl 2 , 0.545 mM CaCl 2 , 0.5 mM EGTA (50 M free Ca 2ϩ ), and 1 mg SR protein/ml. Phosphorylation in the absence of Ca 2ϩ and presence of dimethyl sulfoxide was performed at 25°C in a medium con- sisting of 20 mM Mops, pH 7.0, 80 mM KCl, 5 mM MgCl 2 , 1 mM EGTA, 20% dimethyl sulfoxide, and 1 mg of SR protein/ml. Reactions were initiated by adding 0.1 mM [␥-32 P]UTP (ϳ1 ϫ 10 6 cpm/nmol) and stopped by mixing under vortexing 1 ml of reaction mixture with 1 ml of ice-cold 7% trichloroacetic acid plus 10 mM sodium phosphate. The reaction time was 10 s for Ca 2ϩ -containing samples and 30 s for Ca 2ϩdeprived samples. Quenched samples were incubated in ice for 5 min and then sedimented at 4°C by centrifugation (10,000 rpm for 5 min). The resultant pellets were washed three times with ice-cold quenching solution and once with double distilled water and then dissolved in 0.2 ml of electrophoresis sample buffer (24). Aliquots containing 10 g of membrane protein were applied to a 7.5% polyacrylamide gel slab in the presence of 0.1% lithium dodecyl sulfate. Electrophoresis was carried out for 1 h at room temperature and 180 volts. (24). Radioactive labels in the dried gel were detected by autoradiography at Ϫ80°C. The x-ray film and the intensifying screen were Curix RP2 and Curix blue C2, respectively, from Agfa.
Phosphorylated Intermediate from [␥-32 P]UTP-The levels of radioactive EP at neutral pH and in the presence of K ϩ were measured in a medium containing 20 mM Mops, pH 7.0, 80 mM KCl, 5 mM MgCl 2 , 0.545 mM CaCl 2 , 0.5 mM EGTA (50 M free Ca 2ϩ ), 0.4 mg/ml SR vesicles, and 1 mM [␥-32 P]UTP (ϳ20,000 cpm/nmol). Potassium oxalate (5 mM) was also included when indicated. Phosphorylation was maintained for 2 min at 22°C, and the reaction was stopped by adding 5 ml of ice-cold acid medium (125 mM perchloric acid plus 2 mM sodium phosphate) to 0.5 ml of reaction mixture. The denatured protein was incubated in ice for 5 min before being filtered through a 0.45-m nitrocellulose filter. The filter, once rinsed with 25 ml of ice-cold acid medium, was solubilized and counted by liquid scintillation technique. A blank was performed by adding the acid solution before the radioactive UTP.
Phosphorylation at alkaline pH and in the absence of K ϩ was measured, as described above, in a medium containing 20 mM Tris-HCl, pH 8.0, 5 mM MgCl 2 , 0.549 mM CaCl 2 , 0.5 mM EGTA (50 M free Ca 2ϩ ), 0.4 mg/ml SR vesicles, and 1 mM [␥-32 P]UTP at ϳ20,000 cpm/nmol. Rapid Filtration Experiments on Ca 2ϩ Exchange/Dissociation-The availability of Ca 2ϩ bound to the enzyme for cytoplasmic isotopic exchange was studied at 22°C by a rapid filtration technique (25). When Ca 2ϩ -unloaded vesicles were used, 1-ml aliquots containing 20 mM Mops, pH 7.0, 80 mM KCl, 5 mM MgCl 2 , 0.149 mM 45 CaCl 2 (ϳ15,000 cpm/nmol), 0.1 mM EGTA, and 0.2 mg/ml of SR protein were added onto nitrocellulose filters placed in the filtration apparatus. The filters were immediately flushed, under electronic control, with a medium containing 20 mM Mops, pH 7.0, 80 mM KCl, 5 mM MgCl 2 , 0.149 mM CaCl 2 , 0.1 mM EGTA, and 1 mM UTP. The same procedure was applied for Ca 2ϩ dissociation in unloaded vesicles. The perfusion medium was 20 mM Mops, pH 7.0, 80 mM KCl, 5 mM MgCl 2 , and 1 mM EGTA. In other experiments, SR vesicles (0.2 mg/ml) were initially loaded with Ca 2ϩ , at 22°C for 5 min, in a medium containing 20 mM Mops, pH 7.0, 80 mM KCl, 5 mM MgCl 2 , 0.149 mM 45 CaCl 2 , 0.1 mM EGTA, and 1 mM UTP. Thereafter, 1-ml samples of actively 45 Ca 2ϩ -loaded vesicles were subjected to the rapid filtration procedure as described above. The perfusion medium was 20 mM Mops, pH 7.0, 80 mM KCl, 5 mM MgCl 2 , 0.149 mM CaCl 2 , 0.1 mM EGTA, and 1 mM UTP. Radioactivity retained by the filters was evaluated, without any further washing, by liquid scintillation counting. Blank assays were performed in the absence of SR vesicles. CPA Effect on Enzyme Activity-UTP hydrolysis as a function of time was measured at 25°C in a medium containing 20 mM Mops, pH 7.0, 80 mM KCl, 5 mM MgCl 2 , 0.149 mM CaCl 2 , 0.1 mM EGTA (50 M free Ca 2ϩ ), and 0.05 mg/ml of SR vesicles (i.e. 0.2 M Ca 2ϩ -ATPase). The reaction was started (t ϭ 0) by adding 1 mM UTP. The CPA effect was evaluated by adding 0.4 M CPA at t ϭ 9 min. The CPA effect in leaky vesicles was measured by a similar procedure, although the reaction medium was supplemented with 4 M A23187. Control assays were performed by not adding CPA. Inorganic phosphate released before and after the CPA addition was evaluated by mixing aliquots of 1 ml of reaction mixture and 1 ml of color reagent at different times (22).

TNP-ATP Fluorescence under Turnover Conditions-Changes
Materials-Radioactive tracers 45 Ca 2ϩ , and [␥-32 P]UTP were obtained from Amersham Pharmacia Biotech. TNP-ATP was from Molecular Probes Europe. CPA from Penicillium cyclopium (C 1530) and the liquid scintillation mixture (S 4023) were products of Sigma. A 5 mg/ml stock solution of CPA was prepared in dimethyl sulfoxide. The Ca 2ϩ standard solution (Titrisol) was purchased from Merck. A23187 from Streptomyces chartreusensis was a Calbiochem product. Nitrocellulose filters with a 0.45-m pore size were from Millipore. A Hoefer filtration box from Amersham Pharmacia Biotech and a rapid filtration apparatus from Bio-Logic Co. (Claix, France) were used for sample filtration.
Data Presentation-Experimental points represent the average of at least three independent determinations, each performed in duplicate. Standard errors of the mean are also included. The SigmaPlot Graph System from Jandel Scientific was used for data fitting.

RESULTS
The Effect of Lumenal Ca 2ϩ -The hydrolytic activity of the SR Ca 2ϩ -ATPase protein was evaluated by using 1 mM UTP as phosphorylating substrate. The assay conditions included a buffered medium at neutral pH, 80 mM K ϩ , 5 mM Mg 2ϩ , and 50 M free Ca 2ϩ . Fig. 1 shows that the liberation of inorganic phosphate, in the minute time scale, is characterized by an initial burst followed by a linear phase. The experiments were carried out with native (Ca 2ϩ -unloaded) SR vesicles. The effect of lumenal Ca 2ϩ on the Ca 2ϩ -UTPase activity, measured in the min time scale, was also checked. In this case, the vesicles were initially incubated for 4 h at 30°C with a Ca 2ϩ concentration ranging from 10 to 40 mM. As can be seen, the passive loading of vesicles did not significantly modify the time-dependent appearance of phosphate. Therefore, it was not possible to establish a clear dependence between the rate of UTP hydrolysis and the initial Ca 2ϩ load. Furthermore, Ca 2ϩ -independent UTP hydrolysis displayed a linear time course equivalent to 160 nmol P i /min/mg protein, a rate that represented 55% of the steady state rate measured in the presence of Ca 2ϩ .
The Ca 2ϩ -dependent and Ca 2ϩ -independent activities were further investigated by studying the accumulation of phosphorylated intermediate at 25°C after addition of 0.1 mM [␥-32 P]UTP. In one case, SR vesicles were suspended in a standard reaction medium containing 50 M free Ca 2ϩ , and in the other, the vesicles were suspended in a Ca 2ϩ -free medium and 20% dimethyl sulfoxide was added. Phosphorylation in the presence or absence of Ca 2ϩ was started by adding radioactive nucleotide and stopped by acid quenching. Once the micro-  somes had been thoroughly washed and dissolved, the membrane proteins were subjected to electrophoresis and autoradiography (Fig. 2). The electrophoretic profile of lanes 1 and 2 in Fig. 2 corresponds to samples phosphorylated in the presence or absence of Ca 2ϩ , respectively. The samples were stained with Coomassie Blue and presented the characteristic pattern of SR longitudinal tubules with Ca 2ϩ -ATPase (116-kDa) as the most abundant protein of the membrane. Moreover, autoradiographic analysis showed a radioactive band at the level of the Ca 2ϩ -ATPase protein when the vesicles were phosphorylated both in the presence (lane 3) and absence of Ca 2ϩ (lane 4). This is a clear indication that both the Ca 2ϩ -dependent and Ca 2ϩ -independent activities of UTP hydrolysis were associated with the Ca 2ϩ -ATPase protein. Lane 3 also shows the Ca 2ϩ -dependent phosphorylation of the 150-and 160-kDa calsequestrin-like proteins (26).
Experiments were also directed at measuring the UTP-dependent Ca 2ϩ uptake associated with the release of inorganic phosphate. When unloaded vesicles and 50 M free 45 Ca 2ϩ were used in the external medium, the addition of 1 mM UTP was followed by the rapid entry of 45 Ca 2ϩ , which reached asymptotic levels at approximately 80 nmol/mg protein (Fig. 3A). The effect of lumenal Ca 2ϩ on net uptake was evaluated after a preliminary passive loading of the vesicles with 10 -40 mM 45 Ca 2ϩ and their subsequent addition to a transport medium containing 50 M free 45 Ca 2ϩ . Caution was taken to ensure that the same 45 Ca 2ϩ specific activity existed in both passive loading and active transport media. The net uptake rate and the final accumulation of Ca 2ϩ decreased with increasing initial Ca 2ϩ loading (Fig. 3A). The passive 45 Ca 2ϩ load was 16, 21.5, or 30 nmol/mg of protein, when incubated with 10, 20, or 40 mM 45 Ca 2ϩ , respectively. At each time point of net uptake, we observed that the sum of the initial passive loading plus the net uptake was independent of the initial Ca 2ϩ load (Fig. 3B).
Ca 2ϩ /P i Coupling-Because Ca 2ϩ pumping activity is an energy-dependent process linked to nucleotide hydrolysis, Ca 2ϩ transport and UTP hydrolysis were evaluated in the same reaction medium, except that 45 Ca 2ϩ was included for evaluating Ca 2ϩ transport and [␥-32 P]UTP to measure UTP hydrolysis. The transport process was stopped by La 3ϩ quenching and nucleotide hydrolysis by acid/color reagent. Both parameters were measured at the beginning or after 2 min of reaction, i.e. in coupled or uncoupled vesicles. The addition of 1 mM UTP to SR vesicles incubated in the presence of 50 M free 45 Ca 2ϩ led to the nonlinear active accumulation of 45 Ca 2ϩ , tending to an asymptotic level (Fig. 4A). These measurements, in the second time scale, were carried out at the beginning of Ca 2ϩ pumping. Ca 2ϩ entry was also evaluated after 2 min of active Ca 2ϩ loading. In this case, 1 mM UTP was added to SR vesicles in the presence of 40 Ca 2ϩ , followed by a 45 Ca 2ϩ pulse 2 min later. Under these conditions there was practically no 45 Ca 2ϩ entry (Fig. 4A). Parallel experiments performed in the presence of [␥-32 P]UTP (Fig. 4B) revealed that the rate of P i release at the beginning of the reaction (430 nmol/min/mg of protein) was somewhat higher than that measured at t ϭ 2 min (320 nmol/ min/mg of protein). The effect of time on coupling (Fig. 4B,  inset) was revealed by the Ca 2ϩ /P i ratio of 0.8 at t ϭ 1 s, which showed a tendency to decrease in the following seconds. Total uncoupling of the pump was observed after 2 min of reaction.
The above coupling data prompted us to modify the assay conditions by adding 5 mM oxalate to the reaction medium and maintaining all the other conditions, including the temperature of 25°C and 50 M free 45 Ca 2ϩ . The entry of active 45 Ca 2ϩ displayed a linear time course and very similar rates (220 versus 200 nmol Ca 2ϩ /min/mg of protein) when measured at the beginning or after 2 min of active 40 Ca 2ϩ loading, respectively (Fig. 5A). The corresponding hydrolytic activities measured at the beginning or after 2 min of reaction also showed linear responses. The rates were 330 and 280 nmol P i /min/mg of protein for unloaded or Ca 2ϩ -loaded vesicles, respectively (Fig. 5B). Therefore, the coupling ratio at the beginning of pumping and in the presence of oxalate was 0.7 and did not vary with the Ca 2ϩ loading state of the vesicles (Fig. 5B, inset).
Uncoupling Features-The vectorial translocation of Ca 2ϩ involves the participation of two different EP conformations as outlined in the simplified reaction cycle (Scheme I). Therefore, the identification of accumulated EP species is a critical step for understanding the reaction cycle sustained by UTP. The experiments were performed at 22°C in a reaction medium containing 50 M free Ca 2ϩ . Total EP was measured by phosphorylating in the presence of 1 mM [␥-32 P]UTP. In parallel experiments, the ADP-sensitive E 1 P⅐Ca 2 conformation was evaluated by adding 1 mM UTP in the presence of TNP-ATP. This is an indirect procedure because the TNP-ADP or TNP-ATP fluorescence signal detects the presence of the ADP-insensitive E 2 P (27,28). Fig. 6 shows that the steady state phosphorylation level was around 3.7 nmol/mg of protein when measured at pH 7.0 and in the presence of K ϩ . Likewise, the lack of TNP-ATP fluorescence under the same conditions indicated the absence of E 2 P, and so, the accumulated species was E 1 P⅐Ca 2 . The addition of oxalate to the neutral pH reaction medium containing K ϩ was associated with a decrease in the total EP accumulated. The steady state level was ϳ2.1 nmol/mg of protein, and the phosphorylated form, as deduced from the absence of TNP-ATP fluorescence, was again E 1 P⅐Ca 2 . Some experiments were performed under conditions favoring a slow enzyme turnover and a large accumulation of E 2 P (29, 30). Our data obtained at pH 8.0 and in the absence of K ϩ and oxalate indicated that the total EP increased up to 4.8 nmol/mg of protein, and this was associated with a large increase in TNP-ATP fluorescence, i.e. a massive accumulation of E 2 P.
Because pump uncoupling in the presence of UTP induces the exclusive accumulation of E 1 P⅐Ca 2 , it was deemed of interest to examine the fate of Ca 2ϩ bound to EP. This was approached with the aid of 45 Ca 2ϩ and rapid filtration experiments. Native (Ca 2ϩ -unloaded) vesicles in a medium containing 50 M free 45 Ca 2ϩ were flushed at 22°C and in the second time scale, with a medium containing 1 mM UTP and 50 M 40 Ca 2ϩ . Thus, 45 Ca 2ϩ initially saturating the high affinity transport sites (approximately 8 nmol/mg of protein) were almost completely retained by sealed vesicles when flushed with the nonradioactive medium (Fig. 7A). It means that bound 45 Ca 2ϩ cannot be exchanged by external (cytoplasmic) 40 Ca 2ϩ during the initial seconds of the reaction when UTP is present. In fact, there was a small decrease owing to the phosphorylation/transport and the cytoplasmic isotopic exchange rates driving Ca 2ϩ inside and outside the vesicles, respectively. However, the complete removal of bound 45 Ca 2ϩ was observed in the millisecond time scale (Fig. 7A, inset) when the flushing medium contained 1 mM EGTA instead of UTP plus 40 Ca 2ϩ . In another set of experiments (Fig. 7B), SR vesicles in the presence of 50 M free 45 Ca 2ϩ were initially incubated at 22°C for 5 min with 1 mM UTP. Then, 45 Ca 2ϩ -loaded vesicles were subjected to rapid flushing/filtration with a medium containing 1 mM UTP and 50 M free 40 Ca 2ϩ . The steady state 45 Ca 2ϩ accumulated in the presence of UTP was approximately 127 conditions. The accumulation of E 1 P⅐Ca 2 at neutral pH was inferred from the absence of E 2 P. ox Ϫ , oxalate. SCHEME I. Reaction mechanism of SR Ca 2؉ -ATPase in the presence of UTP. The forward operation of the Ca 2ϩ pump produces the accumulation of phosphorylated species (E 1 P⅐Ca 2 and E 2 P). The E 1 -E 2 cycle (coupled route) involves dissociation of lumenal Ca 2ϩ from E 1 P⅐Ca 2 and further release of P i from E 2 P. The E 1 cycle (uncoupled route) consists of the release of P i from E 1 P⅐Ca 2 before Ca 2ϩ dissociation to the cytoplasmic medium and the absence of E 2 and E 2 P species. nmol/mg of protein. In this case, subsequent flushing with UTP and 40 Ca 2ϩ induced a decrease of the 45 Ca 2ϩ associated with the vesicles. The observed 45 Ca 2ϩ -40 Ca 2ϩ exchange amounted to approximately 9 nmol/mg of protein.
The mechanism of intramolecular uncoupling was confirmed by using the high affinity inhibitor CPA (31)(32)(33) and measuring UTP hydrolysis under the conditions described in Fig. 8. The incubation medium in Fig. 8A included Ca 2ϩ ionophore to make the vesicles leaky, and the reaction was started by adding 1 mM UTP. The release of P i followed a linear time course, equivalent to 700 nmol/min/mg of protein. The inhibitory effect of CPA was demonstrated in another assay by starting the reaction with UTP and then adding CPA at t ϭ 9 min. As can be seen, the initial linear accumulation of P i was clearly stopped by the presence of CPA. In this case the CPA/enzyme molar ratio was 2. However, a protecting role induced by lumenal Ca 2ϩ was revealed by performing experiments in the absence of Ca 2ϩ ionophore. Measurements of UTP hydrolysis in native vesicles (Fig. 8B) showed a linear rate equivalent to 320 nmol P i / min/mg of protein. Furthermore, when the experiments were repeated and CPA was added after 9 min of reaction, the hydrolytic activity was not inhibited. It should be noted that the same CPA/enzyme molar ratio produced complete inhibition in leaky vesicles. DISCUSSION The Ca 2ϩ -ATPase reaction cycle was studied in the presence of the nonphysiological substrate of the enzyme UTP in an attempt to understand the uncoupling phenomenon. The rate of UTP hydrolysis was barely affected by the presence of passive Ca 2ϩ load in sealed SR vesicles (Fig. 1). The Ca 2ϩ load corresponded to lumenal concentrations of 3.2, 4.3, or 6 mM, assuming 5 l/mg of protein as the volume of vesicles (34). However, the presence of the same passive load reduced the rate of UTP-dependent Ca 2ϩ uptake and the final level of accumulated Ca 2ϩ (Fig. 3A). Therefore, lumenal Ca 2ϩ plays a central role in the appearance of uncoupling as occurs in the presence of ATP (8,16). Taking together the Ca 2ϩ transport and the UTP hydrolysis data (Figs. 4 and 5), it is clear that a certain Ca 2ϩ /P i coupling can be maintained during the reaction as long as the lumenal Ca 2ϩ remains low, i.e. by including oxalate in the reaction medium. A rise in lumenal Ca 2ϩ induces complete uncoupling of the pump.
It is common practice to subtract the Ca 2ϩ -independent hydrolytic activity when calculating coupling ratios. However, this is only valid when the Ca 2ϩ -dependent and Ca 2ϩ -independent activities are operating simultaneously, as occurs when they correspond to different proteins. If the total hydrolytic activity is much higher than the Ca 2ϩ -independent activity, as occurs when dealing with ATP, subtraction does not greatly modify the coupling value. In the case of UTP, the Ca 2ϩ -dependent and Ca 2ϩ -independent activities are both related to the same protein (Fig. 2). Moreover, the Ca 2ϩ -independent activity represents about 55% of the total activity ( Fig.  1), although it would be expected to be negligible under turnover conditions because of the presence of saturating Ca 2ϩ levels in the external (cytoplasmic) medium. Therefore, subtraction of the Ca 2ϩ -independent activity overestimates the coupling, giving values close to 2 (Refs. 11 and 12 and data in this paper), so that it is concluded that the relevant activity for calculating coupling ratios is the total Ca 2ϩ -dependent activity. In the presence of oxalate or at the beginning of the reaction when oxalate is absent, a coupling ratio of 0.7-0.8 was measured (insets of Figs. 4B and 5B).
In principle, the uncoupling process can be developed through phosphorylated or unphosphorylated forms of the enzyme. Uncoupling through phosphorylated species involves the release of P i from E 1 P⅐Ca 2 before Ca 2ϩ dissociation to the cytoplasmic medium. In other words, E 1 P⅐Ca 2 can be directly hydrolyzed without interconversion into E 2 P. Moreover, this uncoupling mechanism will occur without a Ca 2ϩ flux (i.e. uptake or exchange). By contrast, uncoupling through unphosphorylated species (E 2 -E 1 ) requires the existence of a Ca 2ϩ efflux, i.e. an ADP-independent Ca 2ϩ exchange in the absence of any net entry. It should be noted that an ADP-dependent Ca 2ϩ exchange does not produce uncoupling because it occurs through the reversal of the ATP phosphorylation reaction.
The experiments on 45 Ca 2ϩ transport measured at the beginning of the reaction (Figs. 4A and 5A) corresponded to a situation of net uptake because the vesicles were initially unloaded. However, at t ϭ 2 min they corresponded to unidirectional 45 Ca 2ϩ influx, when measured in the absence of oxalate (Fig. 4A), because of the presence of lumenal 40 Ca 2ϩ or net 45 Ca 2ϩ uptake when measured in the presence of oxalate (Fig.  5A). A Ca 2ϩ influx in the absence of net uptake can be attributed to Ca 2ϩ exchange, because influx ϭ uptake ϩ exchange. In the absence of oxalate and after 2 min of reaction there was a negligible component of Ca 2ϩ influx (Fig. 4A). Therefore, rapid 40 Ca 2ϩ -45 Ca 2ϩ exchange does not occur in the presence of UTP. This lends support to the uncoupling mechanism mediated by phosphorylated species. Furthermore, uncoupling through E 2 -E 1 in the presence of UTP is unlikely, because activation of Ca 2ϩ exchange by this route was only observed when Ca 2ϩloaded vesicles were diluted in the absence of ATP, Mg 2ϩ , and Ca 2ϩ (13,14).
Further support for our model was provided by the following three observations: (i) There is an absence of accumulated E 2 P in vesicles sustaining an active Ca 2ϩ load (uncoupled) in the presence of UTP (Fig. 6), an observation that cannot be reconciled with the participation of E 2 in the uncoupling (mechanism of unphosphorylated forms). (ii) Ca 2ϩ can dissociate from EP to the cytoplasmic medium when the vesicles are Ca 2ϩ -loaded (uncoupled) but not at the beginning of the reaction (Fig. 7, A  and B). The exchange of 45 Ca 2ϩ bound to E 1 P with external 40 Ca 2ϩ when the vesicles are loaded in the presence of UTP is a clear indication that Ca 2ϩ release inside the vesicles is blocked and Ca 2ϩ dissociation occurs toward the cytoplasmic medium (35). The absence of Ca 2ϩ exchange at the beginning of the reaction supports the expected Ca 2ϩ accumulation under coupling conditions. (iii) UTP hydrolysis by Ca 2ϩ -loaded vesicles (uncoupled) was not inhibited by CPA (Fig. 8B). The uncoupled state of the enzyme was protected from CPA, an E 2directed inhibitor of the enzyme (33). Such inhibition by CPA was patent when the lumenal Ca 2ϩ load was relieved by the addition of Ca 2ϩ ionophore (Fig. 8A), which may be attributed to the existence of some enzyme turnover through the conventional cycle. The forward operation of the Ca 2ϩ pump generates the E 2 form that is the target of the high affinity inhibitor (32).
Our data indicate that E 1 P⅐Ca 2 is a keystone in the uncoupling mechanism. In this regard, the accumulation of this phosphorylated intermediate will be indistinguishable regardless of whether it is formed from ATP or UTP. The fact that the uncoupling is more easily observed in the presence of UTP suggests that the coupled route, i.e. the E 1 P⅐Ca 2 3 E 2 P interconversion is favored by ATP but not by UTP. The activating effect of ATP on the phosphoenzyme interconversion step has been described (36,37).
In experiments with ATP and Ca 2ϩ -loaded vesicles (uncoupling conditions), E 1 P⅐Ca 2 and E 2 P are accumulated, and so the uncoupling mechanism is more difficult to study owing to the existence of a rapid 40 Ca 2ϩ -45 Ca 2ϩ exchange (38,39) and the presence of E 2 and E 2 P conformations of the enzyme.
All the present observations using the phosphorylating substrate UTP can be explained by Scheme I. The coupled route (E 1 -E 2 cycle) relies on the ordered sequential breakdown of E 1 P⅐Ca 2 . The dissociation of Ca 2ϩ inside the vesicles is followed by the release of P i into the cytoplasmic medium. An alteration of this sequence, i.e. the release of P i followed by the release of Ca 2ϩ , both to the cytoplasmic medium, produces the uncoupled route (E 1 cycle). Thus, under low coupling conditions the coexistence of an E 1 -E 2 cycle and E 1 cycle is to be expected, whereas the total uncoupled state can be explained by the sole operation of the E 1 cycle. The existence of the uncoupled cycle may account for the different coupling efficiencies reported for different phosphate donor substrates and different experimental conditions.