Ca2+ Release and Heat Production by the Endoplasmic Reticulum Ca2+-ATPase of Blood Platelets

Different sarco/endoplasmic reticulum Ca2+-ATPases isoforms are found in blood platelets and in skeletal muscle. The amount of heat produced during ATP hydrolysis by vesicles derived from the endoplasmic reticulum of blood platelets was the same in the absence and presence of a transmembrane Ca2+ gradient. Addition of platelets activating factor (PAF) to the medium promoted both a Ca2+ efflux that was arrested by thapsigargin and an increase of the yield of heat produced during ATP hydrolysis. The calorimetric enthalpy of ATP hydrolysis (ΔH cal) measured during Ca2+transport varied between −10 and −12 kcal/mol without PAF and between −20 and −24 kcal/mol with 4 μm PAF. Different from platelets, in skeletal muscle vesicles a thapsigargin-sensitive Ca2+ efflux and a high heat production during ATP hydrolysis were measured without PAF and the ΔH cal varied between −10 and −12 kcal/mol in the absence of Ca2+ and between −22 up to −32 kcal/mol after formation of a transmembrane Ca2+ gradient. PAF did not enhance the rate of thapsigargin-sensitive Ca2+ efflux nor increase the yield of heat produced during ATP hydrolysis. These findings indicate that the platelets of Ca2+-ATPase isoforms are only able to convert osmotic energy into heat in the presence of PAF.

Different sarco/endoplasmic reticulum Ca 2؉ -ATPases isoforms are found in blood platelets and in skeletal muscle. The amount of heat produced during ATP hydrolysis by vesicles derived from the endoplasmic reticulum of blood platelets was the same in the absence and presence of a transmembrane Ca 2؉ gradient. Addition of platelets activating factor (PAF) to the medium promoted both a Ca 2؉ efflux that was arrested by thapsigargin and an increase of the yield of heat produced during ATP hydrolysis. The calorimetric enthalpy of ATP hydrolysis (⌬H cal ) measured during Ca 2؉ transport varied between ؊10 and ؊12 kcal/mol without PAF and between ؊20 and ؊24 kcal/mol with 4 M PAF. Different from platelets, in skeletal muscle vesicles a thapsigargin-sensitive Ca 2؉ efflux and a high heat production during ATP hydrolysis were measured without PAF and the ⌬H cal varied between ؊10 and ؊12 kcal/mol in the absence of Ca 2؉ and between ؊22 up to ؊32 kcal/mol after formation of a transmembrane Ca 2؉ gradient. PAF did not enhance the rate of thapsigargin-sensitive Ca 2؉ efflux nor increase the yield of heat produced during ATP hydrolysis. These findings indicate that the platelets of Ca 2؉ -ATPase isoforms are only able to convert osmotic energy into heat in the presence of PAF.
Heat generation plays a key role in the regulation of the metabolic activity and energy balance of the cell. In animals lacking brown adipose tissue, the principal source of heat during nonshivering thermogenesis is derived from the hydrolysis of ATP by the sarcoplasmic reticulum Ca 2ϩ -ATPase of skeletal muscles (1)(2)(3)(4)(5)(6). This enzyme translocates Ca 2ϩ from the cytosol to the lumen of the sarcoplasmic/endoplasmic vesicles by using the chemical energy derived from ATP hydrolysis (7)(8)(9)(10). Calorimetric measurements of rat soleus muscle (2) indicate that 25 to 45% of heat produced in resting muscle is related to Ca 2ϩ circulation between sarcoplasm and sarcoplasmic reticulum. Three distinct genes encode the sarco/endoplasmic reticulum Ca 2ϩ -ATPases (SERCA) 1 isoforms, but the physiological mean-ing of isoforms diversity is not clear. The SERCA 1 gene is expressed exclusively in fast skeletal muscle (11) whereas blood platelets and lymphoid tissues express SERCA 3 and SERCA 2b genes (12)(13)(14). The catalytic cycle of the different SERCA can be reversed after a Ca 2ϩ gradient has been formed across the vesicles membrane. During this reversal, Ca 2ϩ leaves the vesicles through the ATPase in a process coupled with the synthesis of ATP from ADP and P i (7)(8)(9)(10)15). For vesicles derived from skeletal muscle, a part of the Ca 2ϩ retained during transport leaks through the Ca 2ϩ -ATPase without promoting synthesis of ATP (16 -21). This efflux is referred to as uncoupled Ca 2ϩ efflux. Recently it was shown that the SERCA 1 of skeletal muscle is able to convert osmotic energy into heat. Calorimetric measurements revealed that the amount of heat produced after the hydrolysis of each ATP molecule hydrolyzed increases 2-3-fold when a Ca 2ϩ gradient is formed across the vesicles membrane (22)(23)(24). The extra heat produced during ATP hydrolysis seems to be promoted by the uncoupled Ca 2ϩ efflux during which the energy derived from the Ca 2ϩ gradient is converted by the SERCA 1 into heat.
In the present study we have measured the heat production during ATP hydrolysis, Ca 2ϩ transport, and Ca 2ϩ efflux in vesicles derived from the sarco/endoplasmic reticulum of skeletal muscle and blood platelets, both in the absence and presence of PAF.

EXPERIMENTAL PROCEDURES
Vesicle Preparations-Vesicles derived from the dense tubules of human blood platelets and the light fraction of rabbit skeletal muscle sarcoplasmic reticulum were prepared as described previously (25,26). The muscle vesicle preparation does not contain significant amounts of ryanodine/caffeine-sensitive Ca 2ϩ -channels nor does it exhibit the phenomenon of activation of Ca 2ϩ efflux by external Ca 2ϩ , i.e. Ca 2ϩ -induced Ca 2ϩ release found in the heavy fraction of the sarcoplasmic reticulum (16). Both the muscle and blood platelet vesicles were stored in liquid nitrogen until use.
Ca 2ϩ Uptake and Ca 2ϩ Efflux-This was measured by the filtration method using 45 Ca and Millipore filters (27). After filtration, the filters were washed five times with 5 ml of 3 mM La(NO 3 ) 3 and the radioactivity remaining on the filters was counted on a liquid scintillation counter. The free Ca 2ϩ concentration was calculated using the association constants of Schwartzenbach et al. (28) in a computer program described by Fabiato and Fabiato (29) and modified by Sorenson et al. (30). For Ca 2ϩ efflux experiments, the vesicles were preloaded with 45 Ca in a medium containing 50 mM MOPS-Tris (pH 7.0), 10 mM MgCl 2 , 100 mM KCl, 20 mM P i , 0.3 mM CaCl 2 , 3 mM ATP, and 60 g of vesicles protein/ml. After 30 (muscle vesicles) or 60 (platelet vesicles) min incubation at 35°C, the vesicles were centrifuged at 40,000 ϫ g for 30 min, the supernatant was discarded, and the walls of the tubes were blotted to minimize the volume of residual loading medium. The pellet was kept on ice and resuspended in ice-cold water immediately before use. The efflux was arrested as described above for the Ca 2ϩ uptake.
ATPase Activity-This was assayed measuring the release of 32 P i from [␥-32 P]ATP. The 32 P i produced was extracted from the medium with ammonium molybdate and a mixture of isobutyl alcohol and ben-zene (31). The Mg 2ϩ dependent activity was measured in the presence of 5 mM EGTA. The Ca 2ϩ -ATPase activity was determined by subtracting the Mg 2ϩ dependent activity from the activity measured in the presence of both Mg 2ϩ and Ca 2ϩ .
Synthesis of [␥-32 P]ATP from ADP and 32 P i -This was measured as described previously (31). [ 32 P]P i was obtained from the Brazilian Institute of Atomic Energy.
Heats of Reaction-This was measured using an OMEGA Isothermal Titration Calorimeter from Microcal Inc. (Northampton, MA) (22,24). The calorimeter cell was filled with a reaction medium, and the reference cell was filled with Milli-Q water. After equilibration at the desired temperature, the reaction was started by injecting sarcoplasmic reticulum vesicles into the reaction cell and the heat change due to ATP hydrolysis was recorded starting from 2 min after the injection up to a maximum of 30 min. The calorimetric enthalpy of hydrolysis (⌬H cal ) was calculated by dividing the amount of heat released by the net amount of ATP hydrolyzed. The units used were moles for ATP hydrolyzed and kcal for the heat released. A negative value indicates that the reaction was exothermic and a positive value indicates that it was endothermic.

Ca 2ϩ Transport and Heat Production by Muscle and Platelet
Vesicles-In agreement with previous reports, we found that the vesicles derived from blood platelets ( Fig. 1A and Table I) were able to accumulate a smaller amount of Ca 2ϩ than the vesicles derived from muscle ( Fig. 2A and Table II). During transport the two vesicle preparations catalyze simultaneously the hydrolysis (Figs. 1B and 2B) and the synthesis of ATP from ADP and P i . The rate of synthesis was severalfold slower than the rate of hydrolysis. Using the same experimental conditions as those described in Tables I and II and in the presence of 1 M free Ca 2ϩ , the rates of ATP synthesis for platelets and muscle vesicles were 0.08 Ϯ 0.01 (6) and 2.57 Ϯ 0.22 (4) mol of ATP/mg protein⅐30 min Ϫ1 , respectively. These values are the average Ϯ S.E. of the number of experiments shown in parentheses. The kinetics of Ca 2ϩ transport and ATP synthesis have been analyzed in detail in previous reports (7)(8)(9)(10). In this study we focused on heat produced during transport. The overall reaction of Ca 2ϩ transport was exothermic regardless of whether muscle or platelet vesicles were used (Figs. 1C and 2C). The amount of heat released during the different incubation intervals was proportional to the amount of ATP cleaved. This could be visualized by either plotting the heat released as a function of the amount of ATP hydrolyzed (Fig. 3) or calculating the ⌬H cal using the values of heat release and P i produced at different incubation intervals (Fig. 4). Two different ATPase activities can be distinguished in both platelet and muscle vesicles. The Mg 2ϩ dependent activity requires only Mg 2ϩ for its activation and is measured in the presence of EGTA to remove contaminant Ca 2ϩ from the assay medium. The ATPase activity which is correlated with Ca 2ϩ transport requires both Ca 2ϩ and Mg 2ϩ for full activity. In both vesicle preparations, the Mg 2ϩ -dependent ATPase activity represents a small fraction of the total ATPase activity measured in presence of Mg 2ϩ and Ca 2ϩ (Figs. 1B and 2B). The amount of heat produced during the hydrolysis of ATP by the Mg 2ϩ -dependent ATPase was the same regardless of whether muscle or platelet vesicles were used (Fig. 3C) and the ⌬H cal value ( Fig. 4 and Tables I and II) calculated in the two conditions was the same as that previously measured with soluble F1 mitochondrial ATPase (23) and soluble myosin at pH 7.2 (32). For the vesicles derived from muscle (SERCA 1) the formation of a Ca 2ϩ gradient increased the yield of heat production during ATP hy-

TABLE I
Heat production during Ca 2ϩ transport and ATP hydrolysis by the platelets Ca 2ϩ -ATPase The assay medium composition and experimental conditions were as described in Fig. 1 Table I). For the muscle vesicles (Table  II), there was no difference in the ⌬H cal value of the Mg 2ϩ -dependent ATPase and the Ca 2ϩ -ATPase when the vesicles were rendered leaky (no gradient). With intact vesicles, the ⌬H cal value was more negative, i.e. more heat was produced during the hydrolysis of each ATP molecule when the free Ca 2ϩ concentration in the medium was decreased from 10 to 1 M ( Fig.  4B and Table II). During transport, the P i available in the assay medium diffuses through the membrane to form Ca 2ϩ phosphate crystals inside the vesicles. These crystals operate as a Ca 2ϩ buffer that maintains the free Ca 2ϩ concentration inside the vesicles constant (ϳ5 mM) at the level of the solubility product of calcium phosphate (8,33). The energy derived from the gradient depends on the difference between the Ca 2ϩ concentrations inside and outside the vesicles. Thus, the different values of ⌬H cal measured with the muscle vesicles with 1 and 10 M Ca 2ϩ suggest that when the free Ca 2ϩ concentration in the medium is lower, the gradient formed across the vesicles membrane is steeper; thus more heat was produced and a more negative value of the ⌬H cal for ATP hydrolysis was observed. With vesicles derived from blood platelets, there was no extra heat production during Ca 2ϩ transport regardless of the free Ca 2ϩ concentration in the medium (Figs. 3A and 4A and Table I). Similar to muscle, P i diffuses through the membrane of the platelet vesicles forming calcium phosphate crys-tal inside the vesicles that ensures the maintenance of the free Ca 2ϩ concentration in the vesicles lumen at the same level as that of the muscle (ϳ5 mM). Thus during transport, the Ca 2ϩ gradient formed across the membrane in the presence of 1 and 10 M Ca 2ϩ should be the same in muscle and platelet vesicles. These findings indicate that different from the muscle, the Ca 2ϩ -ATPase of platelets is not able to convert the osmotic energy derived from the gradient into heat. Uncoupled Ca 2ϩ Efflux-Kinetics evidence described in previous reports (22,24) indicate that the extra heat measured in muscle vesicles was related to the uncoupled Ca 2ϩ efflux mediated by the Ca 2ϩ -ATPase. This can be measured arresting the pump by the addition of an excess EGTA to the medium (Fig. 5). In this condition, the free calcium available in the medium is chelated but Mg⅐ATP and other reagents remain at the same concentration as those used in the experiments of Figs. 1 and 2. For the muscle vesicles, the efflux promoted by EGTA decreased when thapsigargin, a specific inhibitor of the Ca 2ϩ -ATPase (34,35), was added to the medium simultaneously with EGTA. The difference between the total efflux and the efflux measured in the presence of thapsigargin represents the uncoupled efflux mediated by the Ca 2ϩ -ATPase (19,36) and in muscle vesicles it represents about 70% of the total Ca 2ϩ efflux ( Fig. 5 and Table III). The uncoupled efflux can also be measured diluting vesicles previously loaded with Ca 2ϩ in a medium containing only buffer and EGTA (Fig. 6B and Tables  III and IV). For the muscle vesicles, this leakage was also decreased by thapsigargin to the same extent as that measured in the conditions of Fig. 5. The Ca 2ϩ efflux of platelet vesicles was slower than that of muscle and was not impaired by thapsigargin, regardless of the method used to measure the efflux   6A and Table IV). These data suggest that Ca 2ϩ leaks through the SERCA 1 of skeletal muscle but not through the SERCA 2B and 3 found in blood platelets. Therefore, the difference of heat production measured in muscle and platelet vesicles after formation of a transmembrane gradient ( Fig. 3 and Tables I and II) could be due to the absence of uncoupled Ca 2ϩ leakage through the Ca 2ϩ -ATPase in platelet vesicles (thapsigargin-sensitive efflux in Table III). Platelet Activating Factor-In previous reports it was shown that some of the lipids derived from the breakdown of membrane phospholipids were able to increase the uncoupled efflux mediated by the Ca 2ϩ -ATPase of skeletal muscle sarcoplasmic reticulum (37). We therefore tested different lipids in platelet vesicles in search of a compound that could promote a thapsigargin-sensitive Ca 2ϩ efflux. The reasoning was that if we could promote the leakage of Ca 2ϩ through the platelet Ca 2ϩ -ATPase then, similar to the muscle vesicles, the platelet vesicles should become able to convert osmotic energy into heat. In the course of these experiments we found that DL-␣-phosphatidylcholine ␤-acetyl-␥-O-hexadecyl could promote such an efflux in platelets but not in muscle vesicles. This phospholipid belongs to a family of acetylated phospholipids known as PAF which are produced when cells involved in inflammatory process are activated. PAF was found to inhibit the Ca 2ϩ uptake of both platelets and muscle vesicles (Figs. 7 and 8 and Table V). With the two vesicles, half-maximal inhibition was obtained with 4 to 6 M PAF. In contrast with the Ca 2ϩ uptake, the ATPase activity of the two vesicle preparations was not inhibited by PAF (Fig. 7). The discrepancy between Ca 2ϩ uptake and ATPase activity suggests that the decrease of Ca 2ϩ accumulation was promoted by an increase of Ca 2ϩ efflux and not by an inhibition of the ATPase. The amount of Ca 2ϩ retained by the vesicles is determined by the differences between the rates of Ca 2ϩ uptake and Ca 2ϩ efflux. The higher the efflux, the smaller the amount of Ca 2ϩ retained by the vesicles. The addition of PAF during the course of Ca 2ϩ uptake promoted the release of Ca 2ϩ until a new steady state level of Ca 2ϩ retention was achieved ( Fig. 8 and Table V). With both preparations, when the higher concentration of PAF was added, the lower the new steady state level of Ca 2ϩ filling (Figs. 7 and 8). The release of Ca 2ϩ promoted by PAF was not accompanied by a burst of ATP synthesis. On the contrary, PAF inhibited the synthesis of ATP driven by the coupled Ca 2ϩ efflux (Fig. 9). This indicates that Ca 2ϩ release promoted by PAF was not promoted by an increase of the reversal of the pump. A major difference between   IV Ca 2ϩ efflux from blood platelets vesicles In A the vesicles were previously loaded with 45 Ca as described under "Experimental Procedures" and then diluted in a medium containing 50 mM MOPS/Tris (pH 7.0) and 0.5 mM EGTA. The assay medium composition and experimental conditions used to measure Ca 2ϩ release after the addition of either 5 mM EGTA (2) or 6 M PAF (3) were as described in Fig the muscle and platelet vesicles was found when thapsigargin was added to the medium together with PAF. For platelet vesicles, the rate of Ca 2ϩ release measured after the addition of PAF was greatly decreased in the presence of thapsigargin ( Fig. 8A and Table IV) indicating that most of the Ca 2ϩ left the vesicles through the ATPase as an uncoupled Ca 2ϩ efflux. This could be better seen after the initial minute of incubation. In fact, the rate of release in platelet vesicles was so fast that we could not measure the initial velocity of release with the method available in our laboratory. Thus, the values with PAF in Table III differ from the other values in that it does not reflect a true rate, but only the parcel of Ca 2ϩ released during the first incubation minute. In muscle, the rate of Ca 2ϩ efflux measured after the addition of PAF was slower than that measured with platelet vesicles (compare Fig. 8, A and B) and the proportion between the Ca 2ϩ effluxes sensitive and insensitive to thapsigargin measured with PAF was practically the same as that measured when the pump was arrested with EGTA (Table IV).
Having found a compound that induces the release of Ca 2ϩ through the pump, we then measured the heat produced during ATP hydrolysis in the presence and absence of PAF. The PAF concentrations selected were sufficient to enhance the rate of efflux without completely abolishing the retention of Ca 2ϩ by the vesicles, i.e. without abolishing the formation of a Ca 2ϩ gradient through the vesicles membrane (Fig. 8). In such conditions PAF was found to enhance the amount of heat produced during the hydrolysis of ATP by blood platelets (Fig. 10 and Table V). In muscle vesicles, however, PAF was found to de-crease the amount of heat produced during ATP hydrolysis. The ⌬H cal values measured with PAF and muscle vesicles were less negative than those measured in the absence of PAF, but still more negative than the values measured in the absence of Ca 2ϩ gradient.

DISCUSSION
Heat Production-It is generally assumed that the energy released during the hydrolysis of ATP by Ca 2ϩ -ATPase can be divided in two non-interchangeable parts, one is converted into heat and the other is used to pump Ca 2ϩ across the membrane (1)(2)(3)(4)(5)(6). Here, this was observed with the platelet vesicles before the addition of PAF (Table I). The recent finding (22)(23)(24) that the SERCA 1 can convert osmotic energy into heat revealed an alternative route that increases 2-3-fold the amount of heat produced during ATP hydrolysis, therefore permitting the maintenance of the cell temperature with a smaller consumption of ATP (Table II). By this route, a part of the energy released during ATP hydrolysis is dissipated into the surrounding medium as heat. The other part is used to pump Ca 2ϩ across the membrane. During uptake, a fraction of the Ca 2ϩ accumulated flows back through the Ca 2ϩ -ATPase from the vesicles lumen to the medium driven by the Ca 2ϩ gradient. This efflux is coupled with the production of heat (thapsigargin-sensitive efflux). Thus, for the muscle vesicles, heat would be produced at least in two different steps of the energy interconversion cycle: (i) during the hydrolysis of ATP, where a part of the chemical energy released was directly converted into heat and (ii) during the leakage of Ca 2ϩ through the ATPase where part of the energy derived from ATP hydrolysis used to pump Ca 2ϩ is first converted into osmotic energy, and then converted by the enzyme into heat (22,24). The data reported show that not all SERCA isoforms are able to readily convert osmotic energy into heat. The vesicles of blood platelets, as obtained after cell fractionation, are not able to promote this conversion. These vesicles, however, can be converted by PAF into a system capable of increasing the heat production during ATP hydrolysis, suggesting that the mechanism capable of providing additional heat production can be turned on and off and this could represent a mechanism of thermoregulation specific of the cells expressing SERCA 2b and 3. Both in muscle and platelet vesicles there is a Ca 2ϩ efflux which is not inhibited by thapsigargin. We do not know through which membrane structure this Ca 2ϩ flows, but the data obtained with platelets before the addition of PAF indicate that during this efflux, osmotic energy is not converted into heat. In platelets, PAF promoted simultaneously the appearance of thapsigarginsensitive efflux and extra heat production during ATP hydrolysis (Tables IV and V) notion that the conversion of osmotic energy into heat cannot be promoted by any kind of Ca 2ϩ leakage and that a device is needed for this conversion (24). For the endoplasmic/sarcoplasmic reticulum, this device seems to be the Ca 2ϩ -ATPase itself, which in addition to interconvert chemical into osmotic energy, could also convert osmotic energy into heat. Platelets Activating Factor-Regardless of its possible physiological implication, in this study the use of PAF as an experimental tool permitted us to show that the platelet vesicles can be converted from an inactive into an active system capable of converting osmotic energy into heat. Heat generation is implicated in the regulation of several physiological processes including metabolism and energy balance of the cell. The V max of most enzymes varies significantly after a discrete temperature change leading to a substantial change of the metabolic activity of the cell (1-3). PAF was originally described as a soluble factor in blood, so it is apparent that some cells secrete it following synthesis. Subsequent experimentation revealed that the secretion of PAF varies greatly depending on the cell type. In some cells, for instance, endothelial cells, the PAF synthesized is not secreted and is used by the cell itself (for reviews, see Refs. 38 and 39). The synthesis of PAF is initiated by phospholipase A 2 . The subsequent steps of synthesis are catalyzed by enzymes located in the endoplasmic reticulum. In metabolic labeling experiments, PAF appears first in the endoplasmic reticulum and then in the plasma membrane. PAF causes an elevation of the cytosolic-free Ca 2ϩ promoted by the release of Ca 2ϩ from both the plasma membrane and from intracellular stores. This was shown in various cells that express SERCA 2B and 3 including platelets, neutrophils, macrophage, endothelial cells, and neuronal cells (38,39). PAF can therefore act in two different manners, through a receptor in the plasma membrane or as an intracellular messenger. The affinity for PAF of the cell membrane receptor is very high, and the PAF concentration needed for cell activation varies between 10 Ϫ12 and 10 Ϫ9 M, a concentration much smaller than that needed in this report to activate the thapsigargin-sensitive Ca 2ϩ efflux in platelet vesicles (38,39). The only possibility that the effect of PAF observed in this report could have some physiological implication is if the concentration of PAF reaches the micromolar range in the microenvironment surrounding the endoplasmic reticulum membrane, where the Ca 2ϩ -ATPase is located and PAF is synthesized. In this case, the local effect of PAF not only would promote the release of Ca 2ϩ from the endoplasmic reticulum (Fig. 8A) but also enhance the amount of heat produced during hydrolysis of ATP (Tables IV and V). FIG. 9. Effect of PAF on the rate of ATP synthesis by platelet (q) and muscle (E) vesicles. The assay medium composition was the same as that described in the legends to Figs. 1 and 2 except that trace amounts of [ 32 P]P i were added to the medium in order to measure the amount of [␥-32 P]ATP formed from ADP and P i after 30 min incubation at 35°C.