Energy-dependent Calcium Transport in Endoplasmic Reticulum of Adipocytes”

The endoplasmic reticulum from isolated rat adipocytes has the ability to actively accumulate calcium. The calcium uptake was characterized using the 20,000 x g supernatant (Sl fraction) of total cellular homogenate. Endoplasmic reticulum vesicles isolated from the Sl fraction as a 160,000 x g microsomal pellet prior to testing demonstrated little ability to accumulate calcium. The calcium uptake in the Sl fraction was localized to the endoplasmic reticulum vesicles by morphologic appearance, by the use of selective inhibitors of calcium uptake, and by high speed sedimentation of the accumulated calcium. The uptake was MgATP-and temperature-dependent and was sustained by the oxalate used as the intravesicular trapping agent. Uptake was linear with time for at least 30 min at all calcium concentrations tested (3 to 100 FM) and exhibited a pH optimum of approximately 7.0. The sulfhydryl inhibitor p-chloromercuribenzenk sulfonate produced a dose-dependent inhibition of calcium uptake with total inhibition at 0.07 pmol/mg protein. Ruthenium red and sodium azide inhibited less than 5% of the uptake at concentrations (5 PM and 10 mM, respectively) which completely blocked calcium uptake by mitochondria isolated from

The endoplasmic reticulum from isolated rat adipocytes has the ability to actively accumulate calcium. The calcium uptake was characterized using the 20,000 x g supernatant (Sl fraction) of total cellular homogenate. Endoplasmic reticulum vesicles isolated from the Sl fraction as a 160,000 x g microsomal pellet prior to testing demonstrated little ability to accumulate calcium. The calcium uptake in the Sl fraction was localized to the endoplasmic reticulum vesicles by morphologic appearance, by the use of selective inhibitors of calcium uptake, and by high speed sedimentation of the accumulated calcium. The uptake was MgATP-and temperature-dependent and was sustained by the oxalate used as the intravesicular trapping agent. Uptake was linear with time for at least 30 min at all calcium concentrations tested (3 to 100 FM) and exhibited a pH optimum of approximately 7.0. The sulfhydryl inhibitor p-chloromercuribenzenk sulfonate produced a dosedependent inhibition of calcium uptake with total inhibition at 0.07 pmol/mg protein. Ruthenium red and sodium azide inhibited less than 5% of the uptake at concentrations (5 PM and 10 mM, respectively) which completely blocked calcium uptake by mitochondria isolated from the same cells. The K,n for calcium uptake was 12 PM total calcium which corresponded to approximately 3.6 PM ionized calcium in the assay system. The maximum velocity of the uptake was 5.0 nmol (mg of microsomal protein)-' (min)-' at 24" under the assay conditions used and exhibited a Q,,, of 1.8. The uptake activity of the endoplasmic reticulum vesicles in the Sl fraction exhibited a marked time-and temperature-dependent lability which might account in part for the lack of uptake in the isolated microsomal fraction. This energy-dependent calcium uptake system would appear to be of physiologic importance to the regulation of intracellular calcium.
Calcium ions have been proposed as critical regulators of intermediary metabolism (l-8). Direct studies of the control of cellular calcium metabolism are complicated by its high degree of compartmentalization, with estimated calcium concentrations of 10m5 to lo-' M in cytosol (9) and 10m2 M in mitochondria (10) and endoplasmic reticulum (11) in a variety of cells, compared to extracellular calcium concentrations typically near lo-" M. This has made the interpretation of calcium fluxes with intact cells difficult, and has necessitated the use of subcellular fractions. Such studies have shown that mitochondria from many cell types accumulate calcium actively (12) and that sarcoplasmic reticulum rapidly accumulates calcium and plays a central role in the relaxation-contraction cycle (11). The importance of active calcium uptake by endoplasmic reticulum has been investigated only recently in other cells (13,14). Rat (15) which has been shown to contain dense deposits of calcium by the pyroantimonite technique using electronmicroscopic x-ray microprobe analysis (16). This has suggested a function in calcium homeostasis analogous to that of sarcoplasmic reticulum but there has been no demonstration of a "calcium pump" to maintain the calcium gradient suggested by morphologic investigation.
The present study is part of a series of investigations designed to elucidate the roles of the various subcellular organelles in calcium homeostasis in adipocytes and describes energy-dependent calcium uptake by endoplasmic reticulum. The uptake has been characterized using the 20,000 x g supernatant (Sl) of total cellular homogenate containing endoplasmic reticulum vesicles. Virtually no calcium uptake was present in these vesicles isolated as a microsomal fraction. The characteristics of the uptake were analogous to those in sarcoplasmic reticulum. The uptake was Mg2+-and ATP-dependent, saturable, azide-insensitive, blocked by sulfhydryl inhibi-t&, and sustained by oxalate. The K, was approximately 3.6 pM ionized calcium, which suggests a significant role at physiologic concentrations of cytosol calcium. Characterization of this uptake allows further investigation of the role of calcium as a regulator of adipocyte metabolism. 7191  (19)(20)(21). The total cell homogenate was centrifuged at 20,000 x g for 15 min to sediment mitochondria, plasma membranes, nuclei, and cellular debris. Mitochondria were isolated from the pellet as previously described using differential and gradient centrifugation (19). The supernatant, containing the microsomal fraction and cytosol, was termed Sl.' The microsomal fraction was prepared from Sl by centrifugation at 160,000 x g for 1 h and was resuspended in Trislsucrose. The microsomal fraction was a highly enriched endoplasmic reticulum preparation containing less than 3% mitochondria (19) and less than 10% plasma membranes (19,22). The contamination of endoplasmic reticulum in Sl by plasma membranes and mitochondria can be taken to be no greater than these values since the high speed centrifugation used to prepare the microsomes would certainly sediment these organelles present in Sl. The fractions were assayed for calcium uptake activity immediately or quickly frozen in dry ice-alcohol or liquid nitrogen and stored at -70". Sl fractions stored in this manner retained active calcium transport for several weeks but variations of stability among preparations necessitated careful monitoring. Assay of Calcium Uptake by Filtration -Calcium accumulation was investigated using the microsomal, Sl, and mitochondrial fractions. Except as noted, all assays were performed using a filtration technique (14,23) with minor modification. Standard incubations were performed in polystyrene tubes with constant shaking for 20 min at 24". The assay was initiated by addition of 20 to 80 pg of protein to the incubation medium containing 0.1 M KCl, 5 rnM MgCl,, 5 rnM ATP, 10 rnhf oxalate, 15 to 25 rnM Tris/HCl, pH 7.0 at 24", 1 to 100 ELM CaCl,, and 0.25 to 0.50 &i of 45CaC1, in a total volume of 300 or 500 ~1. The uptake was terminated by membrane filtration of 250-or 400-/11 aliquots, respectively, and immediate washing of the filters with three 5-ml volumes of 0.25 M sucrose. The filters were dried and 45Ca measured by liquid scintillation counting. In certain studies with the Sl fraction 5-fold higher concentrations of protein were used to allow introduction of an amount of microsomal protein comparable to that used in studies with isolated microsomes. In such experiments iso-osmolality was approximated by reducing the concentration of KC1 to 0.05 M to allow for the osmolar contribution of sucrose in Sl. Calcium uptake rates measured by this moditication were identical with those using the lower protein concentrations. In all assays of calcium uptake, appropriate control tubes without protein were included to correct for nonspecific calcium binding to the filters (routinely less than 1% of total). The specific activity of calcium was determined by counting aliquots of incubation media dried on filter discs. the Sl fraction was also determined by a centrifugation technique for comparison to the filtration method described above. Calcium uptake was terminated by sedimentation of the microsomes by centrifugation at 160,000 x g for 1 h at 4" and aliquots of both the supernatant and the solubilized pellet were counted to determine the uptake.
The uptake rates determined from either rsCaz+ accumulation in the pellet or loss of 45Caz+ from the supernatant agreed with each other within 5%. The close agreement between the filtration and centrifugation assays (see "Results") indicated that soluble protein in Sl did not affect the filtration assay (24). Electron Microscopy Technique -Samples of microsomes sedimented at 160,000 x g and of the fraction of Sl retained by 0.45-pm filters were examined by electron microscopy. Calcium uptake using Sl in the assay was compared to uptake by microsomes which had been isolated from Sl (Fig. 1). Both uptakes were calculated on the basis of microsomal protein which was 8.8 (?0.6)% of total Sl protein (n = 11). Sl demonstrated an active uptake of calcium which Calcium Transport by Adipocyte Endoplasmic Reticulum 7193 was linear for 20 to 30 min at all calcium concentrations tested tion of the structures isolated from Sl either by high speed (3 to 100 PM). The calcium uptake at 20 min using Sl was centrifugation (microsomes, Fig. 2A) or by filtration (Fig. 2B) approximately 100 times that seen using the microsomes. revealed morphologically similar, variably sized, smooth vesi-Addition of microsomal supernatant ("cytosol" fraction) to the cles without visible mitochondria.
No large vesicles with the isolated microsomes to produce a "reconstituted" Sl fraction invaginations characteristic of plasma membrane vesicles (19) did not increase the calcium uptake by the isolated micro-were seen. Examination of material recovered from membrane somes and routinely produced a slight depression of the up-filters after carrying Sl through the assay procedure revealed take. The cytosol fraction itself showed no calcium uptake. these same vesicles. In characterizing the Sl system, it was necessary to confirm that the uptake seen was attributable to endoplasmic reticulum and not to the small amount of contaminating organelles. This was investigated using three independent approaches: comparative morphologic examination of the vesicles isolated by centrifugation and filtration; measurement of calcium uptake using isolation of endoplasmic reticulum by the usual high speed centrifugation to terminate the uptake; and the use of selective inhibitors of calcium uptake.
Electron Microscopy of Material Obtained from Sl by Centrifugution and by Filtration -Electron microscopic examina-

Comparison
of Centrifugation and Filtration Systems in Assay of Calcium Uptake by Sl -Aliquots of Sl were incubated for 10 or 20 min in standard assay medium and calcium uptake was determined by isolating the vesicles either by filtration or by high speed centrifugation. The amounts of calcium uptake measured using these two separation techniques were essentially identical ( Table I).
Effects of Selected Inhibitors on Calcium Uptake-The uptake in Sl was decreased by less than 5% by ruthenium red and by less than 3% by sodium aside at concentrations which blocked over 99% of mitochondrial calcium uptake (Table II). These effects of mitochondrial inhibitors on the Sl uptake are in contrast to their effects on the isolated microsomes. The amount of uptake in Sl blocked by the mitochondrial inhibitors was consistent with the small mitochondrial contamination of the fraction and clearly indicated that the uptake was primarily non-mitochondrial.
The sulihydryl inhibitor p-chloromercuribenzene sulfonate produced a dose-dependent inhibition of the Sl calcium uptake with 50% inhibition at a concen-TIME (min) FIG (not shown) and total inhibition at 0.07 pmol/mg of protein. The residual 1% of the calcium uptake not blocked by the inhibitor (Table II) was attributable to passive binding of calcium to endoplasmic reticulum and was not affected by further increases in the concentration of the inhibitor. The inhibitory concentration observed was equivalent to the concentration of the sulfhydryl inhibitor salyrgan reported to inhibit calcium uptake by sarcoplasmic reticulum (23).
Thus the calcium uptake demonstrated in Sl was associated with the endoplasmic reticulum as judged by the morphologic appearance, sedimentation properties, and the characteristic inhibition pattern. This uptake was characterized in further studies.
Mg2+ and ATP Dependency of Calcium Uptake-Calcium uptake was dependent upon Mg2+ and ATP (Table III). The omission of either ATP or MgCl, resulted in uptake !ess than 10% of control values. Substitution of NaCl for ATP (used as the disodium salt) did not support calcium uptake. The amount of calcium associated with the endoplasmic reticulum in the absence of Mg*+ and ATP was comparable to that found in studies of passive calcium binding to isolated microsomes. The addition of MgCl* alone (Table III) lowered the binding, consistent with the previous demonstration of inhibition by Mg2+ of passive binding of calcium by isolated endoplasmic reticulum (27). The addition of ATP alone increased the calcium associated with the microsomes. The K,,, for MgATP in this system appeared to be below 0.5 mM since there was little variation in the rate of calcium uptake using equimolar concentrations of MgC& and ATP of 0.5 to 10 mM (not shown).
Dependency of Calcium Uptake on Concentration of Calcium-The rate of calcium uptake by endoplasmic reticulum in the Sl fraction was dose-dependent and saturable (Fig. 3). Double-reciprocal analysis yielded a mean K,,, of 11.8 (~6.1) PM total calcium and maximum velocity of 97 (k-6) nmol of calcium/mg of protein per 20 min in seven experiments at pH 6.8 or 7.0 using five preparations of Sl. The K,,, was 3.6 -C 1.0 PM free calcium using ionized calcium concentrations calculated at the pH of the assay according to the method of Katz et al. (26) and verified by Moore et al. (14) to account for calcium formed in a complex in the presence of 5 mM magnesium and ATP. A similar K, of 1.2 PM calcium was found using Ca/ EGTA buffers as described by Katz et al. at pH 6.8 (26). The K, values agreed quite well considering the variables and assumptions used in the calculations of free calcium. The apparent K,n was also a function of the oxalate concentration used, as described below. Calcium uptake in this system was maximal at 50 PM CaCl,. When 5 mM rather than 10 mM oxalate was used, calcium concentrations greater than 50 pM were inhibitory. Similar inhibition of calcium uptake by supraoptimal calcium concentrations has been described in detail for sarcoplasmic reticulum (28).
Effects of Oxalate on Calcium Uptake-Oxalate was necessary for the demonstration of sustained calcium uptake (Fig.  4). In the absence of oxalate, the calcium accumulating ability gradually diminished and the accumulated calcium was lost from the endoplasmic reticulum vesicles (Fig. 4). Similar findings have been reported for endoplasmic reticulum from other tissues including muscle (29). The calcium uptake at 20 min in the absence of oxalate averaged 5% of that in the presence of 10 mM oxalate under the conditions described in the legend to Fig. 4 (n = 5). Such stimulation of calcium uptake by oxalate has been interpreted as indicating active calcium translocation as contrasted to cation binding by membranes (23). When the oxalate concentration was varied between 2.5 and 20 mM, The nonlinearity seen in the presence of oxalate was due to depletion of free calcium: with the high protein concentration used in this experiment, more than 50% of total calcium was removed from the assay medium at 30 min. The kinetic properties of the uptake were investigated in studies at 5 and 10 mM oxalate (Fig. 5). These concentrations were chosen to allow comparison to calcium uptake by endoplasmic reticulum from other tissues. The maximum velocity of calcium uptake using 5 mM oxalate was half that found at 10 mM oxalate and the apparent K,,, was slightly (20%) lower at the lower oxalate concentration (Fig. 5). Temperature Dependency of Calcium Uptake-Calcium uptake was strongly temperature-dependent. At 4" the uptake was comparable to the level of passive energy-independent calcium binding to microsomes (27) and was less than 3% of the uptake at the standard incubation temperature of 24". Peak calcium uptake was observed at 37" and was twice that at 24". At this temperature, the time course of calcium uptake was essentially linear through 20 min.  6 illustrates that preincubation of Sl at 4", 24", and 37" produced a time-and temperature-dependent loss of the calciumaccumulating function.
At 24" and 37" the loss of activity was  6. Lability of the calcium uptake system in Sl. Aliquots of Sl were preincubated for the indicated times at 4", 24", or 37". Following this preincubation, the remaining calcium uptake activity was measured in the standard assay system using 30 PM calcium. Values are expressed as percentages of the uptake observed using Sl which had not been preincubated.
Each point represents the mean of duplicate determinations using a single preparation of Sl. marked within 5 min. In contrast, calcium uptake was linear during 20 to 30 min assays at both 24" and 37" (e.g. Fig. 1). This suggested a protective effect on the calcium uptake system of some component (s) in the assay medium. This was investigated by preincubating Sl in the presence of various components of the assay system and assaying the residual uptake activity (Table IV). ATP alone or with magnesium did not protect the uptake. In contrast, the inclusion of oxalate in combination with magnesium and ATP consistently produced a near-total protection of the calcium-accumulating ability.

7196
Calcium Transport by Adipocyte Endoplasmic Reticulum excellent agreement with the K,, of 1.5 to 1.9 pM ionized calcium reported for cardiac sarcoplasmic reticulum in a study employing a system similar to that used in the present report and utilizing Ca/EGTA buffers (29). In studies of other tissues, the K, reported for calcium uptake by endoplasmic reticulum has ranged from less than 1 PM ionized to approximately 25 PM total calcium (11,13,14). Allowing for differences attributable to the use of Ca/EGTA buffers, variations in the concen-tration of oxalate and ATP, etc., the K, of each of these systems approximates estimated cytosol calcium concentrations (9). In contrast, a marked difference exists between the maximum velocities (V,,,) reported for most sarcoplasmic reticulum preparations and those for endoplasmic reticulum from isolated adipocytes and from nonmuscular tissues. The V max observed in the present studies was 2 to 10 nmol (mg of protein)-' min-', similar to the values of 11 and 5.5 nmol' (mg of protein)-' min' reported for liver (14) and kidney (131, respectively. In contrast, the reported maximum velocities for sarcoplasmic reticulum range from approximately 800 to 3000 nmol (mg of protein)-' min-' (11) in comparable studies.
Strong similarities were noted between the properties of the calcium uptake and those of the passive calcium binding to the endoplasmic reticulum. The K, values for ionized calcium in the transport system both in the presence (1.2 PM) and absence (3.6 PM) of Ca/EGTA buffers were similar to the K, of 2 to 4 PM for passive binding of calcium to the high affinity sites of the adipocyte endoplasmic reticulum (27). The V,,, observed for the adipocyte endoplasmic reticulum was approximately 2 orders of magnitude less than that reported for sarcoplasmic reticulum.
This same loo-fold difference was observed between the maximum (passive) calcium-binding capacities of adipocyte microsomes and sarcoplasmic reticulum (27). Finally, the pH optimum was approximately 7.0 for both the uptake and the passive binding to the high affinity sites. These similarities suggest that the high affinity binding sites could represent the first step in the transport process and by their saturation place a limit on the rate of transport.
The data suggest that the endoplasmic reticulum of adipocytes plays an important role in cellular calcium regulation. This is supported by the previous demonstration by Hales et al. (16) that adipocyte calcium ultrastructurally appears to be concentrated in the endoplasmic reticulum network. This pool of calcium must be considered in interpreting studies of calcium fluxes using intact adipocytes with (32, 33) and without (34) hormones. In previous studies from this laboratory insulin treatment of adipocytes resulted in increased calcium binding to (27) and content of (31) the isolated endoplasmic reticulum. Preliminary data indicate an insulin-induced stimulation of the calcium uptake system (35). These findings indicate that intracellular calcium regulation by the endoplasmic reticulum is hormonally responsive. Characterization of this system will allow investigation of the role of endoplasmic reticulum in both calcium homeostasis and the mechanism of hormone action.