INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

GROWTH HORMONE (GH) regulates not only the metabolism of carbohydrate and lipid in adipocytes (12) but also the concentration of intracellular free Ca²⁺ (9, 10, 32, 33). Earlier studies in this laboratory demonstrated at least two apparently independent effects of GH on Ca²⁺ in rat epididymal adipocytes (33). At ~2 min after addition of GH to freshly isolated adipocytes, intracellular Ca²⁺ concentrations ([Ca²⁺]_i) increased about twofold (10, 32). Similar responses to GH were subsequently reported in human IM9 lymphocytes (16), in rat insulinoma cells (4), and in Chinese hamster ovary cells that express transfected GH receptors (34). In adipocytes this response to GH was abolished by exclusion of Ca²⁺ from the extracellular fluid, addition of 100 nM nimodipine, or treatment with protein kinase C inhibitors (10). Conversely, the effect of GH was mimicked by the dihydropyridine Ca²⁺ channel activator BAY K 5552, partial depolarization of the adipocyte membrane with 30 mM KCl, or treatment with 1,2-dioctanoyl-sn-glycerol (DOG), suggesting strongly that GH activates voltage-sensitive L-type Ca²⁺ channels secondary to activation of protein kinase C (10).

In adipocytes that were maintained in vitro for 3 h in the absence of hormones (GH-deprived cells), resting intracellular Ca²⁺ concentrations ([Ca²⁺]_i) declined from ~200 to ~100 nM (9, 33). The decline in [Ca²⁺]_i was prevented by brief exposure of the cells to GH (GH-treated cells) during the 1st h of a 3- or 4-h incubation period (33). The ability of GH to sustain resting levels of Ca²⁺ appeared to require ongoing transcription and was blocked by actinomycin D (33). The decline in [Ca²⁺]_i in GH-deprived cells was accompanied by a decline in thapsigargin-insensitive Ca²⁺-ATPase activity and Mn²⁺ permeability through dihydropyridine-sensitive channels (9) compared with freshly isolated or GH-treated cells, suggesting that maintenance of resting [Ca²⁺]_i depends on maintenance of Ca²⁺ influx rather than interference with Ca²⁺ extrusion. Furthermore, the decline in [Ca²⁺]_i in GH-treated cells to levels seen in GH-deprived cells after treatment with nimodipine suggested that activity of L-type Ca²⁺ channels might account for the effect of GH on basal levels of Ca²⁺ (9).

Although voltage-sensitive L-type Ca²⁺ channels probably have little influence on resting [Ca²⁺]_i in most cells in which the resting membrane potential is more negative than 60 mV, they are likely to play a crucial role in this regard in adipocytes in which the resting membrane potential has been measured with microelectrodes to be in the 20- to 40-mV range (1, 3, 17, 29, 37). This membrane potential is in the range of the "window" in which there is sustained Ca²⁺ channel activity (8). L-type channels in heart, skeletal muscle, and nerve are activated only after the plasma membrane has been depolarized to about 40 to 30 mV (27). The increase in [Ca²⁺]_i that results from partial membrane depolarization, treatment with BAY K 5552, or DOG was also attenuated in GH-deprived cells (10). In addition, no acute change in [Ca²⁺]_i was seen for at least 30 min after GH was added to GH-deprived adipocytes (10, 32). Together with the decrease in resting [Ca²⁺]_i, these observations suggest that GH-deprived adipocytes might have fewer functional L-type Ca²⁺ channels in their plasma membranes. The experiments described here were undertaken to test the hypothesis that GH maintains the abundance of functional L-type Ca²⁺ channels in adipocyte plasma membranes.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Animals. Male rats of the Charles River CD strain were obtained from the breeding facilities at Kingston, NY, and maintained in the vivarium under conditions of constant lighting (lights on from 0600 to 1800) and temperature (23°C). Rats were fed Purina 5008 laboratory chow (Ralston Purina, St. Louis, MO) and water ad libitum and studied when they attained weights of 150-200 g. The animals were killed by cervical dislocation, and the epididymal and perirenal fat bodies were immediately harvested. All procedures were in accord with protocols approved by the University of Massachusetts Medical School Institutional Animal Care and Use Committee.

Isolation and handling of cells. Tissues pooled from 2-30 rats were finely minced, suspended in Krebs-Ringer bicarbonate incubation medium enriched with 5.5 mM glucose and 1% (wt/vol) albumin (Intergen, Purchase, NY), and incubated for 20 min with 1 mg/ml collagenase (lot 143710, type A, Boehringer Mannheim Biochemicals, Indianapolis, IN) according to the procedure of Rodbell (31) as modified in this laboratory (13). After filtration through silk, the fat cells were washed three times and resuspended at a dilution of 1:3 in incubation medium. Some cells were studied without further processing and hence are referred to as freshly isolated. Other fat cells from the same pools were incubated at 37°C in a shaking water bath under an atmosphere of 95% O₂-5% CO₂ in the absence or presence of 100 ng/ml recombinant human GH (generously provided by Genentech, S. San Francisco, CA). After 1 h the fat cells were again washed, resuspended in fresh medium, and incubated for an additional 2 h in the absence of GH. Adipocytes that were incubated with GH in the 1st h are referred to as GH treated; those that were incubated without GH are referred to as GH deprived. In a few experiments, adipocytes were incubated for 5 h at a dilution of 1:10.

Preparation of adipocyte membranes. Adipocyte plasma membranes were prepared according to the protocol of McKeel and Jarett (23) and microsomal membrane fractions according to Simpson et al. (35). Briefly, 3-5 ml (packed cell volume) of adipocytes were homogenized with a Dounce homogenizer in three volumes of 0.25 M sucrose containing 50 mM Tris buffer, 1 mM phenylmethylsulfonyl fluoride, and 5 mM CaCl₂, pH 7.4. All steps were performed at 4°C. The homogenate was centrifuged at 16,000 g for 20 min. The fat cake was removed and discarded, and the supernate and pellet were processed further. The supernate was centrifuged at 48,000 g for 20 min to obtain a "heavy" microsomal pellet and again at 212,000 g for 1 h to obtain a "light" microsomal pellet. The 16,000-g pellet was resuspended in 10 ml of homogenization buffer, layered over a 9-45% sucrose gradient, and centrifuged for 1 h at 100,000 g. The plasma membrane fraction was visible as a white band near the top of the gradient. To obtain a "combined" membrane fraction, cell homogenates were centrifuged at 16,000 g to remove the fat cake and then applied over a 38% (wt/vol) sucrose cushion. After centrifugation at 100,000 g for 1 h, the mitochondria and nuclei formed a pellet. The supernate was then centrifuged at 212,000 g for 1 h to sediment a combined plasma membrane and microsomal pellet. Suspensions of plasma membranes and heavy and light microsomes containing ~2.5 mg protein/ml as measured by the procedure of Lowry et al. (22) were stored at 95°C for later use.

Ligand binding studies. Specific binding of (+)-[³H]PN-200-110 (Amersham Life Science, Arlington Heights, IL), a dihydropyridine that binds specifically to the -subunit of L-type Ca²⁺ channels (11), was used to quantify Ca²⁺ channels. Binding studies were carried out in triplicate by a modification of the procedure of Glossman and Ferry (11). Briefly, cell membranes (400 µg protein/tube) were incubated with 8.4 nM [³H]PN-200-110 (specific activity 86 Ci/mmol) in the dark for 2 h at 37°C in the absence or presence of 1,000-fold molar excess of unlabeled PN-200-110 (generously donated by Sandoz, East Hanover, NJ) or nimodipine (Research Biochemicals International, Natick, MA). Membrane-bound and free ligand were separated by filtration through 24-mm Whatman (GF/C) glass microfiber filters, as described by Morton and Froehner (25), and the dried filters were counted by liquid scintillation spectroscopy. Specific binding is defined as the difference between total ³H-ligand binding and the binding obtained in the presence of unlabeled ligand.

Immunoassays. Adipocyte membranes were extracted with 1% Triton X-100 plus 0.1% (final concentration) SDS. After clarification by centrifugation at 16,000 g for 10 min, aliquots containing 200-300 µg of protein were incubated for 18 h at room temperature in 96-well microtiter plates that were coated with 1 µg/ml of the capture antibody, monoclonal antibody (MAb) 2B, raised against the ₁-subunit of the L-type Ca²⁺ channel purified from rabbit skeletal muscle (24). The plates were washed to remove unbound antigen and further incubated for 2 h with 1 µg/ml of biotinylated MAb 1A, which recognizes an epitope of the ₁-subunit that is distinct from that recognized by MAb 2B (24). Biotinylation was carried out by incubating the antibody with sulfosuccinimidyl 6-(biotinamide)hexanoate (Pierce, Rockford, IL) for 30 min at room temperature according to the vendor's recommended protocol. After they were washed to remove excess secondary antibody, the plates were incubated for 2 h with 0.1 µg/ml streptavidin-alkaline phosphatase conjugate (Pierce), washed again, and incubated for 2-16 h with 100 µl of p-nitrophenyl phosphate substrate (0.5 mg/ml). Substrate hydrolysis was estimated from light absorbance at 405 nm. The optical density values were compared with values obtained for standards prepared from crude t-tubule extracts of rat skeletal muscle in which the specific activity of Ca²⁺ channels had been determined by [³H]PN-200-110 binding. In some experiments the -subunit was also quantified using a monoclonal antibody obtained from SWant (Bellizona, Switzerland), and the ₂/-subunit was measured using MAb 20A (25).

Measurement of [Ca²⁺]_i. Immediately after cellular isolation or after a designated incubation period, freshly isolated, GH-deprived, and GH-treated adipocytes were incubated for 30 min with the Ca²⁺-sensitive fluorescent dye (13) fura 2-AM (20 µM; Molecular Probes, Eugene, OR). [Ca²⁺]_i was measured according to procedures described previously (32, 33). Briefly, 100 µl of a 1:30 (vol/vol) suspension of fura 2-loaded cells were pipetted into a temperature-controlled Plexiglas perifusion chamber mounted on the stage of an inverted microscope (Nikon Diaphot). The buoyant cells were held in place by adhesion to a coverslip coated with Cell-Tak (Collaborative Research, Bedford, MA) and perifused at a flow rate of 1 ml/min with Krebs-Ringer bicarbonate medium enriched with glucose and containing 0.1 mg/ml BSA. The adipocytes were imaged with a ×10 ultraviolet lens (Nikon, glycerin immersion) and sequentially illuminated with 3-ns pulses of light (337 or 380 nm) delivered every 33 ms through a bifurcated quartz fiber extending from a nitrogen laser and a tunable dye laser (Laser Science, Cambridge, MA) to the epiport of the inverted microscope. Images were recorded with a charge-coupled device camera (CCTV, New York, NY) and resolved into 512 × 240 pixels with an image processor (Recognition Technology, Westborough, MA). Data processing was performed using a computer (NEC Power Mate). Data were selected for processing, and background fluorescence was subtracted from an area of interest (11 × 7 pixels) usually positioned over the perinuclear region of chosen adipocytes. In a typical experiment, cells were observed for 10-15 s so that 30-40 cells could be examined during a 9-min scanning period. [Ca²⁺]_i was estimated by comparison with an in situ calibration curve constructed from the fluorescence ratios measured in the cytosol of GH-treated and GH-deprived adipocytes that were made permeable to Ca²⁺ by treatment with 1 µM ionomycin and equilibrated with 10⁸-10⁴ M extracellular Ca²⁺ (39).

Electrophoresis. Membrane fractions were solubilized in Laemmli sample buffer (19), and 10 µg of protein from each were applied to each lane in 7.5% polyacrylamide slab gels. The proteins were resolved by electrophoresis and visualized with Coomassie blue.

Enzyme assays. ,-Methylene adenosine diphosphate-inhibitable 5'-nucleotidase was used as an enzyme marker for the plasma membranes (5). Assays were performed as described by Burger and Lowenstein (6) in the absence or presence of ,-methylene adenosine diphosphate. Briefly, aliquots of membrane fractions (1-15 µg protein) were incubated in 1 ml of 60 mM Tris · HCl (pH 7.4) containing 100 µM ATP and 0.015 µCi of [³H]AMP (Dupont-NEN, Boston, MA; specific activity 0.2 µCi/pmol) for 15 min at 37°C. The reaction was terminated by addition of 200 µl of 5% (wt/vol) zinc sulfate and 200 µl of 0.3 N barium hydroxide. After centrifugation the supernate was removed and counted by liquid scintillation.

Statistics. Data were analyzed by Student's t-test for paired or unpaired values as appropriate with the Bonferroni adjustment (36) to compensate for type I errors when multiple comparisons were made.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Measurement of specific binding of ³H-labeled dihydropyridines to Ca²⁺ channels in intact adipocytes was not feasible because of low channel abundance and a partition coefficient that favors sequestration of the ligand in the fat droplet. However, specific binding of [³H]PN-200-110 was obtained with a sucrose gradient-purified plasma membrane fraction prepared from disrupted adipocytes (23). Scatchard analysis of [³H]PN-200-110 binding to plasma membranes derived from freshly isolated adipocytes (Fig. 1) indicated ~1.2 × 10¹¹ binding sites/µg membrane protein and half-saturation at a concentration of ~3.5 nM PN-200-110. This affinity is similar to reported dissociation constants of 3.87 nM for skeletal muscle L-type Ca²⁺ channels (11) and 2.15 nM for clonal pituitary cells (20). If it is assumed that 250-300 µg of membrane protein can be recovered from 1 ml of packed cells (~2 × 10⁷ cells) and that 50-100% of the plasma membranes is recovered in our purification procedure, each adipocyte contains ~2,000-3,000 specific binding sites.

View larger version (12K): [in this window] [in a new window]   Fig. 1.   Concentration-dependent binding of (+)-[3H]PN-200-110 to plasma membranes prepared from freshly isolated adipocytes from 30 rats. Triplicate aliquots of membranes (400 µg protein/tube) were incubated with 1 nM [3H]PN-200-110 and 0-10 nM unlabeled (+)-PN-200-110 in darkness for 2 h at 37°C. Linear regression analysis gives a slope (r2 = 0.966) corresponding to dissociation constant of 3.5 nM and an x-intercept corresponding to a Bmax of 76 fmol. Similar data were obtained in 2 experiments.

Because of the relatively low specific activity of the available tritiated ligand and the low abundance of membrane binding sites, it was not practical to perform multiple Scatchard analyses for each experimental condition. Ca²⁺ channel abundance was therefore estimated in plasma membranes prepared from freshly isolated, GH-deprived, and GH-treated adipocytes using ligand at the near-saturating concentration of 8.4 nM (Fig. 2). Deprivation of GH for 3 h resulted in an ~80% decrease in specific binding compared with freshly isolated or GH-treated cells. The time course for the decrease in PN-200-110 binding and for the decline in [Ca²⁺]_i is shown in Fig. 3. After 2 h of GH deprivation, binding was reduced by about one-half, and by 4 h it declined even further to ~20% of that seen in freshly isolated cells (Fig. 3A). By comparison, [Ca²⁺]_i as measured with fura 2 was unchanged at 2 h of GH deprivation but fell precipitously to ~25% of that seen in freshly isolated cells in the succeeding 2 h (Fig. 3B), suggesting that changes in channel abundance may precede changes in [Ca²⁺]_i.

View larger version (18K): [in this window] [in a new window]   Fig. 2.   L-type Ca2+ channels in adipocyte plasma membranes as determined by [3H]PN-200-110 binding. Plasma membranes (400 µg protein) prepared from freshly isolated, growth hormone (GH)-deprived, or GH-treated adipocytes were incubated in triplicate with 8.4 nM [3H]PN-200-110 in presence or absence of 8.4 µM unlabeled PN-200-110. Bars, means ± SE for binding of [3H]PN-200-110; n, number of independent measurements. Adipocytes obtained from 10 rats were used for each experiment. Values correspond to ~2,250, 560, and 2,100 channels/cell in freshly isolated, GH-deprived, and GH-treated adipocytes, respectively, with assumptions that 1 ml of packed cells contains 250 µg of plasma membrane protein and 2 × 107 cells and that 70% of binding sites are occupied at this concentration of ligand.

View larger version (26K): [in this window] [in a new window]   Fig. 3.   Time course for development of effects of GH deprivation. A: specific binding of [3H]PN-200-110 to adipocyte membranes prepared 2 and 4 h after incubation without GH shown as percentage of values obtained in freshly isolated cells. Bars, means ± SE for 3 populations of cells. P values are for comparison with freshly isolated cells. B: intracellular Ca2+ concentration ([Ca2+]i) in adipocytes after 2 and 4 h of incubation without GH expressed as percentage of value obtained for freshly isolated cells. Bars, means ± SE for 8 populations of adipocytes, each including 20 cells.

Decreased binding to membranes prepared from GH-deprived adipocytes might reflect decreased affinity for the ligand as a result of channel inactivation or a decrease in channel abundance due to altered channel turnover or cellular distribution. To resolve this issue, using two monoclonal antibodies (MAb 1A and MAb 2B) that recognize distinct epitopes of the purified ₁-subunit of skeletal muscle L-type Ca²⁺ channels, we developed a sandwich immunoassay to estimate the abundance of channel proteins. Ca²⁺ channels in a crude t-tubule preparation of rat skeletal muscle and quantified by binding and sandwich assays were used as standards (Fig. 4). In agreement with the results of the binding studies shown in Fig. 2, immunoassays of plasma membrane extracts revealed a significant reduction in Ca²⁺ channel protein in membranes prepared from GH-deprived adipocytes compared with freshly isolated or GH-treated adipocytes (Fig. 5). Estimates of immunoassayable channel abundance in freshly isolated and GH-treated cell membranes were in close agreement with the results of binding experiments, but the decline observed in GH-deprived cells appeared less severe by immunoassay than by ligand binding.

View larger version (18K): [in this window] [in a new window]   Fig. 4.   Quantification of Ca2+ channels by specific ligand binding and optical density readings obtained in immunoassays. Pooled rat quadriceps muscles (10 g) were homogenized, and a membrane preparation was obtained by differential centrifugation (6, 23). Specific binding of 8.4 nM [3H]PN-200-110 was assayed as described for adipocyte membranes. Solubilized extracts of muscle membranes (0-25 µg) were incubated in microtiter plates with monoclonal antibody (MAb) 2B as capture antibody and biotinylated MAb 1A as detecting antibody.

View larger version (20K): [in this window] [in a new window]   Fig. 5.   L-type Ca2+ channels in adipocyte plasma membranes as determined by immunoassay for 1-subunit. Ca2+ channel proteins were extracted from plasma membranes prepared from freshly isolated, GH-deprived, or GH-treated adipocytes and assayed in triplicate in 96-well microtiter plates with MAb 2B as capture antibody and biotinylated MAb 1A for detection. Bars, means ± SE; n, number of independent measurements. Values correspond to ~2,900, 1,600, and 2,500 channels/cell in freshly isolated, GH-deprived, and GH-treated adipocytes, respectively. P value is for comparison with freshly isolated or GH-treated cells.

In contrast with findings of fewer Ca²⁺ channels in membranes purified from GH-deprived adipocytes, binding of [³H]PN-200-110 in a combined plasma membrane-microsomal membrane fraction failed to detect any differences in abundance in GH-deprived or freshly isolated adipocytes (Fig. 6). This finding suggested that GH might regulate channel distribution rather than channel synthesis or degradation. To assess this possibility, the presence of Ca²⁺ channels in microsomal and plasma membrane fractions obtained from freshly isolated and GH-deprived adipocytes was estimated by ligand binding assays. With the assumption of 70% occupancy of binding sites at the ligand concentration of 8.4 nM, these data correspond to 3,800 ± 640 binding sites in the combined membranes of freshly isolated cells and 3,300 ± 570 binding sites in the combined membranes of GH-deprived adipocytes. Nearly 80% of those binding sites (3,000 ± 300) were recovered in the plasma membrane fraction of freshly isolated cells compared with only ~24% (780 ± 120) in the plasma membrane fraction of GH-deprived cells. Conversely, <20% of the binding sites were in the microsomal fraction of freshly isolated cells (720 ± 170) compared with 42% (1,400 ± 210) in the microsomes of GH-deprived cells.

View larger version (30K): [in this window] [in a new window]   Fig. 6.   Binding of [3H]PN-200-110 by membrane fractions prepared from freshly isolated and GH-deprived adipocytes. Combined membranes, pellet obtained when "postmitochondrial" supernate was centrifuged for 1 h at 212,000 g; microsomes, combined heavy and light microsomal fractions. Data were corrected for different amounts of protein recovered in each fraction and normalized to packed cell volume to provide insight into actual cellular distribution of channels. In terms of specific activity, plasma membrane fraction of freshly isolated adipocytes bound 280 ± 27 fmol of [3H]PN-200-110 per mg protein compared with 73 ± 12 for GH-deprived cells; microsomal fraction prepared from freshly isolated cells bound 88 ± 21 fmol of PN-210-110 compared with 170 ± 26 for GH-deprived cells. Bars, means ± SE of results of 5 independent experiments, each including adipocytes pooled from 30 rats.

Immunoassay of channel distribution in freshly isolated, GH-treated, and GH-deprived adipocytes yielded similar conclusions (Fig. 7). GH deprivation significantly decreased the abundance of Ca²⁺ channel proteins in the plasma membrane fraction and increased their recovery in the light microsomes. These changes were prevented by GH treatment. Once again the decrease in Ca²⁺ channels in the plasma membrane measured immunologically (Fig. 7) was not as profound as that estimated from binding studies (Fig. 6) and corresponds to ~600 channels per cell, ~75% of which could be accounted for by the increase in channel proteins in the light microsomes. GH had little if any effect on the abundance of channel proteins found in the heavy microsomes, which probably include some plasma membranes and some light microsomes.

View larger version (28K): [in this window] [in a new window]   Fig. 7.   Immunoassay of Ca2+ channels in membrane fractions of freshly isolated, GH-deprived, and GH-treated adipocytes. Bars, means ± SE of 7 experiments with freshly isolated cells and 14 experiments each with GH-deprived and GH-treated cells. Each experiment included adipocytes pooled from 10 rats. * P < 0.05 compared with freshly isolated or GH-treated adipocytes.

Modifications of the immunoassay procedure to detect the ₂/-subunit or the -subunit provided similar data on channel redistribution (Fig. 8). To assay for the ₂/-subunit, MAb 2B was again used as the capture antibody, and biotinylated antibody to the ₂/-subunit (MAb 20A) was used for detection. To assay for the -subunit, an MAb directed against an epitope on the -subunit was used as the capture antibody, and biotinylated MAb 1A was used for detection. Like the ₁-subunit, the - and ₂/-subunits of the Ca²⁺ channels were decreased in the plasma membrane fractions prepared from GH-deprived adipocytes and increased in the light microsomal fraction. Although recoveries of partially purified membrane fractions varied, the data shown in Figs. 7 and 8 indicate that most (>50%) of the Ca²⁺ channels that disappeared from the plasma membranes of GH-deprived cells could be accounted for in the light microsomal fraction. Table 1 shows the nearly identical distribution ratios of the three channel proteins in the plasma membrane and light microsomal fractions of freshly isolated and GH-deprived adipocytes. These data are the means of three independent experiments, each including adipocytes from 15 rats. Thus it appears that the channels remain intact and may be translocated as functional complexes.

View larger version (26K): [in this window] [in a new window]   Fig. 8.   Effects of GH treatment and deprivation on abundance of L-type Ca2+ channel subunits in plasma membranes (A) and light microsomes (B). Bars, means ± SE of 3 experiments, each including adipocytes from 15 rats. All 3 subunits were significantly higher (P < 0.05) in plasma membranes of GH-treated than GH-deprived adipocytes (A) and significantly higher (P < 0.05) in light microsomes of GH-deprived adipocytes (B).

Clearly, the fractionation scheme used did not yield pure plasma membranes and a pure light vesicular membrane fraction. The specific activity of the plasma membrane marker enzymes 5'-nucleotidase and thapsigargin-insensitive Ca²⁺-ATPase was at least three times higher in the plasma membrane fraction than in the microsomal fraction, whereas the heavy microsomal fraction contained the bulk of the thapsigargin-sensitive Ca²⁺-ATPase characteristic of the endoplasmic reticulum (Fig. 9). The relatively low abundance of plasma membrane and endoplasmic reticulum marker enzymes is consistent with the findings of Simpson et al. (35), who reported that this fraction consisted largely of Golgi membranes. GH deprivation did not affect the distribution of any of the three enzymes measured, although the amount of the Ca²⁺-ATPase recovered in the plasma membrane fraction was significantly lower than that recovered in GH-treated cells in accord with earlier findings (9). Qualitative differences among the three fractions were also evident from the array of the most abundant proteins, as seen in each fraction after electrophoretic separation on polyacrylamide gels and staining with Coomassie G200. Each membrane fraction showed a distinct pattern of darkly staining bands, but there were no obvious differences between GH-deprived and GH-treated cells, indicating that GH deprivation did not produce any global rearrangement of cellular proteins (data not shown).

View larger version (19K): [in this window] [in a new window]   Fig. 9.   Distribution of enzyme activities in plasma membranes, heavy microsomes, and light microsomes prepared from GH-treated and GH-deprived adipocytes. Bars, means ± SE of 4 experiments each involving 8 rats. Plasma membrane fractions contained significantly greater (P < 0.01) activities of 5'-nucleotidase (A) and thapsigargin-insensitive Ca2+-ATPase (B) than light or heavy microsomes. Activity of thapsigargin-sensitive Ca2+-ATPase (C) was significantly greater (P < 0.05) than zero only in heavy microsomal fraction. Thapsigargin-insensitive Ca2+-ATPase activity was significantly (P < 0.05) higher in plasma membranes of GH-treated adipocytes, but GH treatment did not significantly affect activity of either enzyme in any fraction.

To determine whether the translocation of Ca²⁺ channels can be reversed by treatment with GH, adipocytes were incubated for a total of 5 h in the absence of GH or with GH present for various times (Fig. 10), and [Ca²⁺]_i or the ₁-subunit was measured. [Ca²⁺]_i declined by a factor of ~2 during the first 3 h of incubation without GH and remained at that low level for the subsequent 2 h. Addition of GH at 3 h restored [Ca²⁺]_i by the 5th h (Fig. 10A). When GH was present during the first 3 h of the 5-h incubation, [Ca²⁺]_i remained constant or rose slightly (data not shown). Similarly, Ca²⁺ channels measured by immunoassay declined by ~40% in the plasma membrane fraction after 5 h without GH and increased by nearly twofold in the light microsomes. Exposure of adipocytes to GH for 1 h at the end of the 3rd h resulted in a nearly complete reversal of these changes by the end of the 5th h. The distribution of Ca²⁺ channel protein measured at the end of the 5-h incubation period was unchanged compared with freshly isolated cells when GH was present for the first 3 h of incubation (data not shown). The data suggest that GH not only prevents the migration of Ca²⁺ channels out of the plasma membrane but also promotes their reinsertion.

View larger version (17K): [in this window] [in a new window]   Fig. 10.   Reversibility of effects of GH deprivation on [Ca2+]i and channel protein distribution. A: [Ca2+]i was measured at start of 5-h incubation and 3 and 5 h later. All cells were deprived of GH for first 3 h. Some cells were then incubated with 100 ng/ml GH for 1 h (GH @ 3 h). Each point represents mean ± SE for number of cells indicated in parentheses. [Ca2+]i was decreased significantly at 3 h (P < 0.01) and declined further (P < 0.05) at 5 h. Treatment with GH increased [Ca2+]i significantly (P < 0.01) above 3- and 5-h values in untreated cells. Cells from same populations that were treated with 100 ng/ml GH for first 3 h showed no decline in [Ca2+]i (data not shown). B: adipocytes were incubated for 5 h without GH, with 100 ng/ml GH present only for first 3 h or only during 4th h. Bars, means ± SE for 5 independent experiments each involving adipocytes from 30 rats. 1-Subunit was significantly diminished in plasma membrane fraction of cells incubated without GH for 5 h compared with GH treatment for first 3 h (P < 0.01) or GH treatment for only 4th h (P < 0.05) and significantly increased in light microsomes compared with cells treated with GH for first 3 h (P < 0.01) or only 4th h (P < 0.05).

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Results of the foregoing experiments are consistent with the hypothesis that GH maintains functional L-type Ca²⁺ channels in the adipocyte plasma membrane. Channel abundance in partially purified plasma membrane preparations was markedly reduced after GH deprivation as determined by ligand binding assays and by immunoassays for three different channel proteins. Although the particular isoforms of Ca²⁺ channel subunits that are present in adipocytes are unknown, the very fact that they could be measured by immunoassay indicates that adipocytes express L-type Ca²⁺ channels that must share at least some immunologic determinants with L-type channels that are expressed in skeletal muscle. The decrease in plasma membrane channels was accompanied by an increase in channel abundance in the light microsomal fraction in GH-deficient adipocytes. This shift in channel distribution was prevented by including GH in the initial incubation buffer and was reversed by treating GH-deprived cells with GH.

In these experiments the abundance of Ca²⁺ channels in different membrane fractions was estimated by ligand binding and by immunoassay. Estimates of channel abundance in ligand binding assays as carried out in these studies assume that at least two important conditions are met: 1) all channels are accessible to the ligand and 2) affinity for the ligand is constant. Immunoassay provides an estimate of the amount of channel proteins that are present in the sample and, because the proteins are extracted from their native membrane environment, neither accessibility nor configuration as related to ligand binding affinity is a relevant concern. Agreement between the two assay procedures was quite good for channel abundance in membranes of freshly isolated or GH-treated cells, as evident from comparison of Figs. 2 and 5. However, the abundance of Ca²⁺ channels in the plasma membranes of GH-deprived adipocytes was two- to threefold greater when estimated by immunoassay than by ligand binding. Additionally, ~75% of the immunoassayable channel proteins lost from the plasma membrane in GH-deprived cells could be accounted for by the gain in the light microsomes (Fig. 7), but less than one-half can be accounted for when measured by ligand binding assay (Fig. 6). Several factors may contribute to these discrepancies. It is possible that the discrepancies are artifactual because of the imprecision inherent in measuring low levels of specific binding or that the immunoassay overestimates channel protein abundance at the low end of the detectable range. Alternatively, it is possible that GH deprivation decreases affinity for the ligand as well as channel abundance. The affinity of the ₁-subunit for dihydropyridines is not constant and varies with the activity state (20, 21). Binding assays were performed at a single ligand concentration of 8.4 nM, at which theoretically 70% of the ligand binding sites in freshly isolated or GH-treated cells are occupied. A fivefold decrease in affinity, for example, would produce an apparent twofold decrease in Ca²⁺ channel abundance. The possibility that GH maintains the activity as well as the abundance of Ca²⁺ channels will probably be resolved only by application of electrophysiological studies.

Another explanation for the discordant assay results in the GH-deprived adipocytes derives from the finding that dihydropyridine binding sites are accessible to the ligand only from the extracellular side of the membrane (2). The binding sites of any L-type channels present in intracellular (endosomic) vesicles that escaped rupture by the homogenization procedure would face inward and, hence, be inaccessible to [³H]PN-200-110. These channels would escape detection in a ligand binding assay, but not in an immunoassay. The observed higher specific activity of ligand binding sites in microsomal vesicles than in plasma membranes of GH-deprived cells (Fig. 6) indicates that contamination with plasma membranes cannot account for all the binding detected in the microsomes and that at least some of the channels in these vesicles were accessible to [³H]PN-200-110. Nevertheless, the combined heavy and light microsomal fractions of GH-deprived cells contained ~30% more channels when estimated by immunoassay than by ligand binding, raising the possibility that as many as one-third of the channels in the light microsomes may be inaccessible to the ligand. If so, any intact light microsomal vesicles that might contaminate the plasma membrane fraction would contribute more to the immunoassay estimates of channel abundance than to the ligand binding estimates. Such contamination would be quantitatively less important for the plasma membranes of freshly isolated or GH-treated cells, because the light microsomes of these cells contain fewer channels than those of GH-deprived cells. Our data provide no estimate of light microsomal contamination of the plasma membrane fraction, but assay of specific marker enzymes by Simpson et al. (35) in comparable experiments suggests that as much as 20-25% of the protein recovered in the plasma membrane fraction was derived from the light microsomal fraction. This degree of contamination of the plasma membrane fraction with intact microsomal vesicles that were impermeable to the ligand could account for as much as 40% of the discrepancy between the two assays.

The findings that the time and magnitude of the reversible decline in [Ca²⁺]_i roughly paralleled the decline in Ca²⁺ channel abundance in the partially purified membranes lends further credence to the idea that GH regulates the distribution of these channels and that resting [Ca²⁺]_i in adipocytes are at least partly determined by rates of Ca²⁺ influx through these channels. The earlier decline in plasma membrane channel abundance than in [Ca²⁺]_i seen in Fig. 3 appears to indicate that the loss of plasma membrane channels preceded the fall in [Ca²⁺]_i, but it is important to note that [Ca²⁺]_i is also partly determined by the rate of Ca²⁺ extrusion by the Ca²⁺-ATPase, which also declined in GH-deficient adipocytes (8) (Fig. 9). The time course of that decline is unknown. In this regard, it is noteworthy that the Ca²⁺-ATPase did not translocate to the light microsomes with the L-type Ca²⁺ channels in GH-deficient cells (Fig. 9).

The findings of reciprocal and reversible changes in the abundance of channels in the plasma membrane and the light microsomal fraction strongly support an argument for hormone-dependent shuttling of channels between the surface membrane and some less dense intracellular membranous structures. Similar reversible internalization of L-type Ca²⁺ channels in response to depolarization was reported by Liu et al. (20) in clonal pituitary cells. The reversible downregulation of these channels in rat GH₄C₁ pituitary cells after prolonged treatment with epidermal growth factor (15) may represent a similar phenomenon. GH may now be added to the growing list of hormones, including insulin (30), antidiuretic hormone (ADH) (18), and parathyroid hormone (26), the actions of which include regulation of the insertion and removal of intrinsic membrane proteins that determine membrane permeability characteristics. Likewise, L-type Ca²⁺ channels may join GLUT-4, aquaporin, and the sodium-phosphate cotransporter as proteins that are shuttled into and out of the plasma membrane in a hormone-dependent manner. Many questions remain to be answered, however. The successful immunologic assays that measure the expression of ₂/- and -subunits along with the ₁-subunit indicate that the channel complex remained intact in the translocation process and was not disrupted by our solubilization procedure. The internalization of Ca²⁺ channels in GH-deprived cells differs in this regard from the internalization of Ca²⁺ channels brought about by prolonged depolarization of GH₄C₁ cells (21). In the latter case, the ₁- and -subunits appeared to dissociate from the ₂/-subunit, which remained associated with the plasma membrane. Failure to account for a relatively small amount of channel proteins lost from the plasma membrane and not recovered in microsomes may be ascribed to methodological limitations, but some dissociation and loss to the soluble cell extract cannot be ruled out. In the adipocyte, unlike the GH₄C₁ cell, there was no evidence of selective retention of the ₂/-subunit in the plasma membrane. Translocation of intact or partially dissociated channel complexes might best be envisioned as an endocytic process akin to that described for GLUT-4 (30). However, the nature, composition, and cellular location of the putative vesicles that harbor the Ca²⁺ channels in the GH-deficient state remain to be determined.

There are some notable differences in the GH-dependent translocation of Ca²⁺ channels and in the translocation of GLUT-4 in response to insulin or translocation of aquaporin in response to ADH. The actions of insulin on GLUT-4 translocation and of ADH on aquaporin are apparent within minutes (18, 30) and are independent of protein synthesis, whereas the actions of GH on Ca²⁺ channels require hours and may depend on transcription (33). The nature of any induced protein(s) that may be required for translocation and their role in the putative translocation process are unknown. It is possible, however, that the delay in the GH effect reflects time needed for synthesis of some short-lived protein that signals a translocation process similar to that involved in GLUT-4 or aquaporin translocation.

Although the precise role of L-type Ca²⁺ channels in adipocyte physiology has yet to be established, it is clear that cellular Ca²⁺ levels are associated with, and probably govern, responsiveness to the insulin-like actions of GH (33). Because changes in intracellular Ca²⁺ are involved in signaling and executing many different cellular functions, adjustment of the abundance of Ca²⁺ channels may provide a means for adjusting the sensitivity of many cellular processes to hormones and other agonists. For example, if the abundance of L-type Ca²⁺ channels in the plasma membranes of secretory cells were increased by GH, a given concentration of agonist might be expected to produce a greater secretory response. Further studies are required to ascertain whether GH or other hormones enhance the sensitivity of target cells to agonists in this manner.