Abstract
GH, in the presence of glucocorticoid, produces a delayed increase in lipolysis in rat adipose tissue, but the biochemical mechanisms that account for this action have not been established. Other lipolytic agents rapidly activate adenylyl cyclase (AC) and the resulting production of cAMP initiates a chain of reactions that culminates in the activation of hormone-sensitive lipase. We compared responses of segments of rat epididymal fat or isolated adipocytes to 30 ng/ml GH and 0.1 μg/ml dexamethasone (Dex) with 0.1 ng/ml isoproterenol (ISO), which evoked a similar increase in lipolysis. All measurements were made during the fourth hour after the addition of GH+Dex or immediately after the addition of ISO to cells or tissues that had been preincubated for 3 h without hormone. Although no significant increases in cAMP were discernible in homogenates of GH+Dex-treated tissues, RP-cAMPS (RP-adenosine 3′5′-phosphothioate), a competitive inhibitor of cAMP, was equally effective in decreasing lipolysis induced by GH+Dex or ISO. The proportion of PKA that was present in the active form was determined by measuring the incorporation of 32P from[γ -32P]ATP into kemptide in the absence and presence of saturating amounts of cAMP. GH+Dex and ISO produced similar increases in protein kinase A activity in tissue extracts. Treatment with GH+Dex did not change the total forskolin-stimulated AC present in either a crude membrane pellet sedimented at 16K × g or a less dense membrane pellet sedimented at 100K × g, but doubled the AC activity in the 16K pellet when assayed in the absence of forskolin. To evaluate possible effects on G proteins, pellets obtained from centrifugation of adipocyte homogenates at 16K × g and 100K × g were solubilized and subjected to PAGE and Western analysis. GH+Dex decreased Giα2 by 44% (P < 0.02) in the 16K pellets and increased it by 52% (P < 0.01) in the 100K pellets. Gsα in the 16K pellet was unaffected by GH+Dex and was decreased (P < 0.05) in the 100K pellet. Sucrose density fractionation of the 16K pellets revealed a similar GH+Dex-dependent shift of Giα2 to less dense fractions as determined by both Western analysis and[ 32P]NAD ribosylation catalyzed by pertussis toxin. No such changes were seen in the distribution of Gsα or 5′-nucleotidase. Colchicine (100 μm) blocked the GH+Dex-dependent shift of Giα2 from the 16K to the 100K pellet and blocked the lipolytic effects of GH+Dex, but not those of ISO. We conclude that by modifying the relationship between AC and Giα2, GH+Dex relieves some inhibition of cAMP production and consequently increases lipolysis.
TREATMENT with GH decreases body fat content by decreasing lipogenesis and increasing lipolysis (1). Consistent with these actions, GH increases lipolysis when added in vitro to adipocytes of many species, including rats (2), mice (3), chickens (4), rhesus monkeys (Goodman, H. M., and L.-R. Tai, unpublished), and perhaps humans (5). The lipolytic effect of GH has been studied most extensively in the rat and differs in several respects from the well characterized lipolytic effect of catecholamines. For example, epinephrine may increase glycerol production 10-fold or more within just a few minutes, and this effect is independent of ongoing RNA and protein synthesis (6). In contrast, GH seldom increases lipolysis more than about 3-fold and acts only after a delay of approximately 1–2 h, reflecting the requirement for RNA and protein synthesis (6). Glucocorticoids potentiate the lipolytic responses to both GH (6) and epinephrine (7), but stimulation of lipolysis by GH in rat adipocytes is often unmeasurable in the absence of glucocorticoid (8). However, glucocorticoids apparently are not required for GH to stimulate lipolysis in chicken (4), mouse (3), or human adipocytes (5).
Stimulation of lipolysis by epinephrine and other rapidly acting agents results from activation of adenylyl cyclase (AC), followed by activation of cAMP-dependent protein kinase (PKA), and the phosphorylation and activation of hormone-sensitive lipase (9). The biochemical pathway through which GH activates lipolysis has not been elucidated, but available data suggest that GH stimulates lipolysis by activating the same cAMP-dependent pathway as that used by catecholamines. GH activates hormone-sensitive lipase in 3T3–442A adipocytes (10) and glycogen phosphorylase in rat adipocytes (11). Both of these enzymes are substrates for PKA (12). The lipolytic and glycogenolytic actions of GH are enhanced by theophyline (13, 14), which inhibits cAMP degradation and blocks adenosine A1 receptors that provide inhibitory input to AC (15). Moskowitz and Fain (16) reported that in the presence of glucocorticoid GH caused a cycloheximide-sensitive increase in cAMP concentration in suspensions of rat adipocytes after a delay of 2 h. In these studies, however, cAMP was present largely in the incubation medium, with no detectable changes in the cells, and the cAMP concentrations did not correlate with rates of lipolysis. Subsequent studies (17, 18) showed that GH increased cAMP accumulation in adipocyte suspensions, but only when high concentrations of other agents (theophyline, catecholamines, or glucagon) that themselves increase cAMP were also present.
Receptors for hormones that produce their biological effects through changes in cAMP production belong to the superfamily of hormone receptors that have seven membrane-spanning domains and are coupled to adenylyl cyclase (19) through heterotrimeric G proteins. The GH receptor, on the other hand, belongs to the superfamily of receptors that have a single membrane-spanning region and signal through associated tyrosine kinases of the JAK family (20). Stimulation of these receptors is thought to activate transcription by way of the STAT (signal transduction and activation of transcription) proteins and the mitogen-activated protein kinases (20). No direct interaction of GH receptors with G proteins or AC has been described. Nevertheless, some observations indicate that G proteins are affected by GH. Hypophysectomy increased and treatment with GH for only 3 h decreased the apparent abundance of the inhibitory G protein (Gi) in rat adipocyte ghosts, as judged by a decrease in pertussis toxin-catalyzed NAD ribosylation (21). Roupas et al. (22) treated obese mice with a diabetogenic analog of GH and found a similar decrease in pertussis toxin-catalyzed ribosylation in adipocyte membranes. Conversely, treatment of normal rats with antiserum to GH for 2 days increased Giα2 in adipocyte plasma membranes, and this effect was reversed by simultaneous administration of GH (23). These observations coupled with the findings that 1) basal rates of lipolysis in adipocytes are under tonic inhibitory control mediated by Gi (21, 24); 2) GH increases cAMP accumulation in response to catecholamines and other agonists (19, 20); 3) GH sensitizes adipocytes to catecholamines (7, 16, 25); and 4) GH deprivation sensitizes adipocytes to antilipolytic agents (23) are consistent with the hypothesis that GH might decrease inhibitory input to AC by actions exerted on Gi and thereby increase cAMP production and lipolysis.
The present study was undertaken to evaluate the hypothesis that GH (in the presence of dexamethasone) increases lipolysis by an action on Gi that results in deinhibition of AC and the subsequent acceleration of lipolysis through cAMP-mediated events.
Materials and Methods
Animals
Normal male rats of the Charles River CD strain (Charles River Breeding Laboratories, Kingston, NY) were used in all experiments in accordance with protocols approved by the University of Massachusetts Medical School animal care and use committee. Rats were fed Purina 5008 (Ralston Purina Co., St. Louis, MO) from the time they were received until they were studied 1–2 weeks later and had attained a body weight of 180–250 μg. The rats were maintained at constant temperature (24 C) and lighting, with lights on from 0600–1800 h. Rats were killed by cervical dislocation, and epididymal fat pads were removed.
Incubations
Thin distal portions of epididymal fat pads or isolated adipocytes (26) were incubated (1:10 dilution) in Krebs-Ringer bicarbonate buffer (pH 7.4) that contained 5.5 mm glucose and 1% (wt/vol) BSA (Introgen Co., NY). Incubations were carried out in a shaking water bath (37 C) under an atmosphere of 95% O2-5% CO2 for 4 h. Human GH (Genentech, Inc., San Francisco, CA; 30 ng/ml) and 0.1μ g/ml dexamethasone (Dex; 9α-fluoro-11β,17α-trihydroxy-16α-methyl-1,4-pregnadiene-3,20-dione; Lypho-Med, Inc., Chicago, IL) were added to the incubation medium and present throughout the incubation. Because the GH-induced increase in lipolysis is not fully developed for at least 2 h, measurements of lipolysis and related biochemical events were made only during the fourth hour of incubation with GH and Dex. To do so, tissues segments were incubated for 3 h with hormone and then transferred to fresh medium of the same composition to permit collection of glycerol produced in the final hour of incubation. Because of its rapid onset of action, l-isoproterenol (ISO; Sigma Chemical Co., St. Louis, MO; 0.1 ng/ml or 1 μg/ml)) was added to tissues or cells that had preincubated for 3 h, and its effects were studied during the fourth hour. Glycerol was assayed enzymatically (27). When suspensions of isolated adipocytes were studied, the difference in glycerol concentrations in aliquots of medium sampled after 3 and 4 h of incubation was used to estimate glycerol produced in the fourth hour.
Homogenate preparations
Tissues (∼100 mg) or isolated adipocytes (20-ml packed cell volume) were disrupted in 2 vol homogenization buffer (10 mm Tris-HCl, pH 7.4; 1 mm EDTA; 8.7 ng/ml phenylmethylsulfonylfluoride; and 50 μg/ml leupeptin) at 4 C in ground glass homogenizers with 10 up/down strokes. The homogenates were allowed to settle for 10 min in an ice bath until a fat cake formed. After removal of the fat cake, the infranatants were used for preparation of membranes or assays of cAMP, AC, or PKA.
Membrane preparations
Partially purified adipocyte plasma membranes were prepared by the method of McKeel and Jarret (28). Briefly, homogenates were centrifuged at 12,000 rpm for 20 min using an SS-34 rotor in a Sorval RC2-B centrifuge (DuPont, Wilmington, DE) to obtain a 16,000 × g pellet (16K pellet) that contained about 80–90% of the plasma membranes. The resulting supernate was centrifuged for 1 h at 50,000 rpm using a Ti 60 rotor in a Beckman Coulter, Inc. (Palo Alto, CA) L8-M ultracentrifuge. The resulting 100,00 × g pellet (100K pellet) contained less than 20% of the plasma membranes as determined in assays of 5′-nucleotidase or AC (see below). In some experiments the 16K fraction was resuspended in 1 ml homogenization buffer and layered over linear sucrose density gradients (5–25%) for further purification by centrifugation using a SW41 rotor at 37,000 rpm (100,000 × g) for 90 min. Fractions were collected by carefully pipetting 0.8-ml aliquots from the tops of the tubes.
Western blot analyses
The 16K and 100K pellets or sucrose gradient fractions prepared from adipocytes that had incubated for 4 h without or with GH+Dex were dissolved in Laemmli buffer (29). Protein content was determined by the Lowry method (30) before loading samples on 10% SDS-polyacrylamide gels. Samples were resolved by electrophoresis (PAGE) with a vertical electrophoresis apparatus (Hoefer SE600, San Francisco, CA) and transferred to nitrocellulose membranes (MSI, Westborough, MA) according to the method of Towbin (31) using a Hoefer TE42 transfer apparatus. Nonspecific antibody-binding sites were blocked by soaking the membranes in 10% milk proteins in PBS for 2 h at room temperature followed by washing three times for 2 min in PBS and 1% milk proteins. Membranes were then incubated for 2 h at 4 C with a 1:1000 dilution of rabbit anti-Giα2 or Gsα in PBS that contained 1% milk proteins. The membranes were washed three times for 2 min each time and once for 20 min in PBS, 1% milk proteins, and 0.3% Tween-20. Polyclonal antibodies directed against the C-terminal decapeptide of Giα2 (KENLKDCGLF) or Gsα (QRMHLRQYELL) and characterized by Rapiejko et al. (32) were provided by Dr. Craig Malbon. Membranes were then reincubated at 4 C for 2 h with a 1:10,000 dilution of goat antirabbit antiserum conjugated to horseradish peroxidase. After washing as described above, the membranes were soaked for 1 min in enhanced chemiluminescent (ECL) reagent (Amersham Pharmacia Biotech, Arlington Heights, IL). The membranes were then exposed to x-ray film (Kodak XAR, Eastman Kodak Co., Rochester, NY) for 5 sec to 60 min. After development of the films, the membranes were stripped of antibodies by soaking in 60 mm Tris-HCl, pH 6.8, that contained 2% SDS and 0.1 m β-mercaptoethanol for 30 min at 37 C with occasional shaking and then reprobed as described above. The intensity of the chemiluminescent bands was quantitated by densitometry.
PKA was assayed according to the methods and conditions used by Roskowski (33) and Corbin (34). Homogenates of fat tissues (∼25 μl) were added to assay mixtures that consisted of homogenization buffer containing 100 μm kemptide and 100 μm[γ -32P] ATP, (SA, 5 μCi/μmol; Amersham Pharmacia Biotech) with or without 100 μm cAMP and were made up to a final volume of 50 μl. After incubation for 10 min at 37 C, 20-μl aliquots of the reaction mixtures were spotted on Whatman P81 paper (Clifton, NJ) and washed three times in 75 mm phosphoric acid and once in 95% ethanol. After drying, the papers were counted in scintillation cocktail (Optifuor, Beckman Coulter, Inc. Palo Alto, CA). Data were expressed as the ratio of enzyme activity measured in the absence and presence of cAMP and represent the degree of PKA activation by GH+Dex or ISO.
AC was assayed following the procedure of Salomon et al. (35). Briefly, 50-μl aliquots of resuspended 16K and 100K pellets were mixed with 300 μl AC assay mixture (25 mm Tris-HCl, pH 7.4; 12.5 mm MgCl2; 20 mm creatine phosphate; 250 U/ml creatine phosphokinase; 1 mm ATP; and 100 μm GTP) with and without 1 mm forskolin and incubated at 30 C for 10 min. The reaction was terminated by the addition of 50 μl of 25% ice-cold trichloroacetic acid. The mixture was centrifuged at 16K × g for 15 min, and after extraction of trichloroacetic acid with ether, the aqueous supernate was assayed for cAMP using a RIA assay kit from Amersham Pharmacia Biotech. Data were expressed as the ratio of enzyme activity measured in the absence and presence of forskolin and represent the degree of AC activation by GH+Dex.
ADP-ribosylation was carried out according to the procedure of Roupas et al. (22). After resuspension, aliquots of 16K pellets (30–50 mg protein) were incubated for 30 min at 37 C in a total volume of 100 μl 100 mm Tris-HCL (pH 8.0) that contained 2μ g/ml activated pertussis toxin (Sigma Chemical Co., St. Louis, MO), 2 mm ATP, 25 mm dithiothreitol, and 5 μCi [α-32P]NAD (SA, 5 μCi/μm; Amersham Pharmacia Biotech). Reactions were terminated by adding 50 μl 2-fold concentrated Laemmli buffer (29). After solubilization by boiling, the samples were resolved by SDS-PAGE, and ribosylation was quantified by densitometric analysis of the autoradiograms.
5′-Nucleotidase activity was used as an enzyme marker for plasma membranes. Aliquots of cell extracts were incubated for 15 min without or with α,β-methyleneadenosine diphosphate, which specifically inhibits plasma membrane-associated 5′-nucleotidase (36) in 60 mm Tris-HCl buffer (pH 7.4) that contained 100μ m [U-14C]5′-AMP (SA, 0.15 μCi/μmol; DuPont-New England Nuclear, Boston, MA). The reactions were terminated by the addition of 5% (wt/vol) ZnSO4 and 0.3 n Ba(OH)2 and centrifuged. Plasma membrane-associated 5′-nucleotidase activity was measured as the difference in the rates of appearance of [14C]adenosine in the supernatant in the absence and presence ofα ,β-methyleneadenosine diphosphate.
Statistics
The statistical significance of the data was evaluated by ANOVA for repeated measures followed by multiple pairwise t tests. All statistics were performed with the StatView computer program (Abacus Concepts, Inc., Berkeley, CA). Each experiment consisting of a separate population of adipocytes pooled from three to eight rats was considered a single observation.
Results
To evaluate the importance of cAMP for GH-induced lipolysis, 10 segments of epididymal fat from each of 16 rats were incubated without or with human GH and Dex alone or in combination for 3 h. The tissues were then transferred to fresh medium of the same composition and reincubated for an additional hour without or with 30μ m RP-cAMPS (RP-adenosine 3′5′-phosphothioate), a cell-permeant competitive inhibitor of cAMP (37). In addition, 0.1 ng/ml ISO with or without Rp-cAMPS was added to one pair of tissues that had been preincubated for 3 h without hormones (Fig. 1). GH alone, but not Dex alone, elicited a small, but significant, increase in lipolysis, and the combination of GH and Dex increased glycerol production about 4-fold. Glycerol production was increased to a similar extent by ISO, and Rp-cAMPS was equally effective in antagonizing the response to either GH+Dex or ISO. Higher concentrations of Rp-cAMPS produced no greater antagonism, presumably because Rp-cAMPS is also a competitive inhibitor of cAMP phosphodiesterase and thus permits endogenous cAMP to accumulate and partially overcome the inhibition of PKA. Although these results implicate cAMP in the lipolytic action of GH+Dex, no detectable change in the concentration of cAMP was found after 3 h of incubation with GH+Dex in parallel experiments (341 ± 35 pmol/g tissue for controls vs. 359 ± 43 for GH+Dex; n = 16). Similar concentrations of cAMP were also seen in homogenates of tissues incubated with GH alone or Dex alone and in all four treatment groups assayed at the end of the fourth hour of incubation.
Effects of RP-CAMPS on lipolysis. Segments of epididymal fat were incubated in the absence or presence of hormones for 3 h. Tissues were then transferred to fresh medium and reincubated for an additional hour, at the end of which medium was harvested for measurement of glycerol. GH (30 ng/ml), Dex (1 μg/ml), and GH+Dex were present for all 4 h. RP-CAMPS (30μ m) and ISO (0.1 ng/ml) were present only during the last hour of incubation. Each bar represents the mean and sem of 16 observations. Lipolysis was significantly increased by GH (P < 0.05), GH+Dex (P < 0.01), and ISO (P < 0.01). RP-CAMPS significantly (P < 0.01) reduced the rate of lipolysis in tissues treated with GH+Dex and ISO.
Because sustained high rates of lipolysis were observed when cAMP concentrations were at nearly basal levels in tissue (38) or cell (39, 40) homogenates, we tested the possibility that GH+Dex might increase cAMP, but only to a level that was below the sensitivity of our assay or in the cellular locale restricted to the immediate vicinity of PKA. To provide a physiologically relevant indication of cAMP concentrations, we took advantage of the observation that PKA activity correlates directly with rates of glycerol production over a wide range of lipolytic activities of a variety of cAMP-dependent agonists (41). Segments of adipose tissue were incubated for 3 h without or with GH+Dex, followed by a 1-h incubation in the absence or presence of GH+Dex or ISO. Cell-free extracts were prepared, and PKA was assayed in the absence or presence of saturating concentrations of cAMP by measuring the incorporation of 32P from[α -32P]ATP into kemptide, a specific synthetic substrate for PKA (Table 1). Each of the hormone treatments significantly increased the activity of PKA expressed either in absolute terms or as a percentage of the total activatable pool (activity percent), indicating that GH+Dex and ISO increased either the availability of cAMP or its efficacy in activating the enzyme. When the rate of glycerol production by the same tissue segments was expressed as a function of the PKA activity (Fig. 2), the data fell on a straight line with a regression coefficient (r) of 0.995, in agreement with the findings of Honnor et al. (41).
Effects of hormones on PKA activity in adipose tissue homogenates
| 32P incorporated into kemptide (cpm/μg protein·min) | |||
|---|---|---|---|
| −cAMP | +cAMP | % of maximum activity | |
| Control | 28.8 ± 8.0 | 164 ± 25.5 | 14.9 ± 1.39 |
| ISO (0.1 ng/ml) | 31.0 ± 8.2a | 156 ± 17.5 | 18.1 ± 2.11 |
| ISO (1 μg/ml) | 50.8 ± 8.3b | 160 ± 15.0 | 30.9 ± 3.19b |
| GH + Dex | 31.6 ± 7.9a | 157 ± 16.0 | 18.6 ± 1.92a |
| 32P incorporated into kemptide (cpm/μg protein·min) | |||
|---|---|---|---|
| −cAMP | +cAMP | % of maximum activity | |
| Control | 28.8 ± 8.0 | 164 ± 25.5 | 14.9 ± 1.39 |
| ISO (0.1 ng/ml) | 31.0 ± 8.2a | 156 ± 17.5 | 18.1 ± 2.11 |
| ISO (1 μg/ml) | 50.8 ± 8.3b | 160 ± 15.0 | 30.9 ± 3.19b |
| GH + Dex | 31.6 ± 7.9a | 157 ± 16.0 | 18.6 ± 1.92a |
Segments of adipose tissue (100 mg) were incubated for 4 h. GH + Dex were present throughout the 4 h, whereas ISO was present only in the final hour. PKA was assayed, as described in Materials and Methods, in the absence or presence of 100μ m cAMP. Values are the mean ± sem of 16 observations.
P < 0.05 compared to untreated control, as determined by paired analyses.
P < 0.001.
Effects of hormones on PKA activity in adipose tissue homogenates
| 32P incorporated into kemptide (cpm/μg protein·min) | |||
|---|---|---|---|
| −cAMP | +cAMP | % of maximum activity | |
| Control | 28.8 ± 8.0 | 164 ± 25.5 | 14.9 ± 1.39 |
| ISO (0.1 ng/ml) | 31.0 ± 8.2a | 156 ± 17.5 | 18.1 ± 2.11 |
| ISO (1 μg/ml) | 50.8 ± 8.3b | 160 ± 15.0 | 30.9 ± 3.19b |
| GH + Dex | 31.6 ± 7.9a | 157 ± 16.0 | 18.6 ± 1.92a |
| 32P incorporated into kemptide (cpm/μg protein·min) | |||
|---|---|---|---|
| −cAMP | +cAMP | % of maximum activity | |
| Control | 28.8 ± 8.0 | 164 ± 25.5 | 14.9 ± 1.39 |
| ISO (0.1 ng/ml) | 31.0 ± 8.2a | 156 ± 17.5 | 18.1 ± 2.11 |
| ISO (1 μg/ml) | 50.8 ± 8.3b | 160 ± 15.0 | 30.9 ± 3.19b |
| GH + Dex | 31.6 ± 7.9a | 157 ± 16.0 | 18.6 ± 1.92a |
Segments of adipose tissue (100 mg) were incubated for 4 h. GH + Dex were present throughout the 4 h, whereas ISO was present only in the final hour. PKA was assayed, as described in Materials and Methods, in the absence or presence of 100μ m cAMP. Values are the mean ± sem of 16 observations.
P < 0.05 compared to untreated control, as determined by paired analyses.
P < 0.001.
Glycerol production as a function of PKA activity. The data for PKA activity shown in Table 1 are plotted as a function of glycerol production by the same tissues. Each point represents the mean value for 16 observations. The horizontal and vertical brackets represent the sem for the PKA activity and the rates of glycerol production, respectively. The increase in glycerol production was significant (P < 0.05) in all hormone-treated groups. r, Regression coefficient.
Further support for the idea that GH+Dex act through increased cAMP production was obtained from direct assay of AC activity. Isolated adipocytes were incubated without or with GH+Dex for 4 h. The cells were harvested at the end of 4 h, and AC activity was measured in both the 16K plasma membrane fraction and in the pellets sedimented by centrifugation at 100 × g. Enzyme assays were performed in the absence and presence of a saturating concentration of forskolin to distinguish between an effect of GH+Dex on the activity and an effect on the amount of AC. The 16K pellet contained more than 4 times as much forskolin-activatable AC activity as the 100K pellet, in accord with the expected sedimentation of AC with the plasma membrane fraction. Compared with controls, rates of cAMP formation were nearly twice as high in the 16K pellets derived from GH+Dex-treated cells assayed in the absence of forskolin, but were unchanged by hormone treatment when assayed in the presence of forskolin (Table 2). GH+Dex-dependent changes in AC activity in the 100K pellet were small and statistically insignificant whether assayed in the presence or absence of forskolin, and the percent activation was unchanged by GH+Dex. GH+Dex doubled the rate of glycerol released from these cells in the fourth hour, as determined from the difference in glycerol concentration measured in aliquots of medium taken at the end of the third and fourth hours (data not shown).
The effects of GH + Dex on adenylyl cyclase activity
| cAMP production (pmol/μg membrane protein·10 min) | |||
|---|---|---|---|
| −Forskolin | +Forskolin | % of maximum activity | |
| 16K pellet | |||
| Control | 1.71 ± 0.10 | 15.64 ± 1.07 | 11.11 ± 1.08 |
| GH + Dex | 3.13 ± 0.34a | 13.27 ± 1.03 | 23.98 ± 3.12a |
| 100K pellet | |||
| Control | 0.13 ± 0.03 | 1.10 ± 0.26 | 11.64 ± 1.07 |
| GH + Dex | 0.32 ± 0.11 | 2.95 ± 0.89 | 10.84 ± 1.27b |
| cAMP production (pmol/μg membrane protein·10 min) | |||
|---|---|---|---|
| −Forskolin | +Forskolin | % of maximum activity | |
| 16K pellet | |||
| Control | 1.71 ± 0.10 | 15.64 ± 1.07 | 11.11 ± 1.08 |
| GH + Dex | 3.13 ± 0.34a | 13.27 ± 1.03 | 23.98 ± 3.12a |
| 100K pellet | |||
| Control | 0.13 ± 0.03 | 1.10 ± 0.26 | 11.64 ± 1.07 |
| GH + Dex | 0.32 ± 0.11 | 2.95 ± 0.89 | 10.84 ± 1.27b |
Isolated epididymal adipocytes were incubated for 4 h in the absence or presence of 30 ng/ml hGH and 1 μg/ml Dex. The plasma membrane-rich 16K pellet and the 100K pellet were prepared and assayed as described in Materials and Methods. The rate of cAMP production was measured as an index of adenylyl cyclase activity. Values are the mean ± sem of six observations.
P < 0.005 compared with control.
P < 0.005 compared to the percent active in the 16K pellet.
The effects of GH + Dex on adenylyl cyclase activity
| cAMP production (pmol/μg membrane protein·10 min) | |||
|---|---|---|---|
| −Forskolin | +Forskolin | % of maximum activity | |
| 16K pellet | |||
| Control | 1.71 ± 0.10 | 15.64 ± 1.07 | 11.11 ± 1.08 |
| GH + Dex | 3.13 ± 0.34a | 13.27 ± 1.03 | 23.98 ± 3.12a |
| 100K pellet | |||
| Control | 0.13 ± 0.03 | 1.10 ± 0.26 | 11.64 ± 1.07 |
| GH + Dex | 0.32 ± 0.11 | 2.95 ± 0.89 | 10.84 ± 1.27b |
| cAMP production (pmol/μg membrane protein·10 min) | |||
|---|---|---|---|
| −Forskolin | +Forskolin | % of maximum activity | |
| 16K pellet | |||
| Control | 1.71 ± 0.10 | 15.64 ± 1.07 | 11.11 ± 1.08 |
| GH + Dex | 3.13 ± 0.34a | 13.27 ± 1.03 | 23.98 ± 3.12a |
| 100K pellet | |||
| Control | 0.13 ± 0.03 | 1.10 ± 0.26 | 11.64 ± 1.07 |
| GH + Dex | 0.32 ± 0.11 | 2.95 ± 0.89 | 10.84 ± 1.27b |
Isolated epididymal adipocytes were incubated for 4 h in the absence or presence of 30 ng/ml hGH and 1 μg/ml Dex. The plasma membrane-rich 16K pellet and the 100K pellet were prepared and assayed as described in Materials and Methods. The rate of cAMP production was measured as an index of adenylyl cyclase activity. Values are the mean ± sem of six observations.
P < 0.005 compared with control.
P < 0.005 compared to the percent active in the 16K pellet.
To determine whether incubation with GH+Dex affects the abundance of G proteins in the plasma membranes, Western analyses were performed on solubilized proteins from the 16K pellet. Incubation for 4 h with GH+Dex had little or no effect on the abundance of Gs in this fraction, but severely reduced the abundance of Giα2 in the two experiments shown in Fig. 3A. Loss of Giα2 from the plasma membrane fraction might result either from accelerated degradation or from redistribution to another centrifugal fraction. Consequently, we examined both the 16K and 100K pellets and the 100K supernate. The decrease in Giα2 in the 16K pellet of GH+Dex-treated cells was accompanied by an increase in its abundance in the 100K pellet (Fig. 3B). In the seven experiments summarized in Fig. 4, treatment with GH+Dex resulted in a 44% decrease in Giα2 in the 16K pellet and a 52% increase in the 100K pellet, as determined by densitometric scanning of Western blots. It is important to note, however, that these data provide only semiquantitative estimates of G protein abundance because of differences in exposure times of the films to the ECL reactions needed for visualization of the bands in the 16K and 100K fractions and because of the nonlinearity of the film at high or low levels of chemiluminiscence. As is evident from Fig. 3B, no Giα2 was detected in the 100K supernate even after 10-fold concentration of the proteins and prolonged exposure of the Western blots. Although the abundance of Gsα in the 16K pellet was unchanged by treatment with GH+Dex, it appeared to be significantly reduced in the 100K pellet.
Effects of GH+Dex on the distribution of Giα2 and Gsα in adipocyte homogenates. Isolated adipocytes were incubated without or with GH+Dex for 4 h. The cells were harvested, and 16K and 100K pellet fractions were prepared as described in Materials and Methods. In the two experiments shown in A, the proteins in the 16K pellets were separated by PAGE, transferred to nitrocellulose membranes, and probed for Gsα or Giα2. B shows the results of two experiments in which extracts corresponding to 10% of the proteins in the 16K and 100K pellets and 100% of the supernate were examined simultaneously.
Effects of GH+Dex on the distribution of Giα and Gsα in 16K and 100K pellets obtained from adipocyte homogenates. Data obtained in the experiments shown in Fig. 3 and additional experiments of the same design were quantitated by densitometry. Because of the experimental variability in the intensity of the ECL reactions, several different exposure times were used for each experiment. Films with similar intensities were selected for densitometric analysis. Differences in intensities in adjacent lanes due to hormonal treatment are indicated by the columns designated “change.” The bars represent the mean ± sem obtained for seven experiments.
The data in Figs. 33 and 44 suggest that at least some of the decrease in Giα2 in the 16K pellet might result from an effect of GH+Dex to decrease the density of the plasma membranes in the immediate vicinity of Giα2 or to translocate Giα2 selectively to endosomic vesicles. To evaluate this possibility, we collected plasma membrane fractions on linear sucrose density gradients and analyzed the distribution of Giα2 and Gsα in Western blots probing first for Giα2 and subsequently for Gsα (Fig. 5). The exposed films are shown in the upper part of the figure. The distributions of G proteins plotted as arbitrary density units with the highest intensity assigned a value of 100 are shown below. In this and three additional experiments of identical design there was no consistent shift in the density of Gsα after treatment with GH+Dex, but the maximum intensity of the Giα2-staining bands was shifted at least one fraction to the left, i.e. toward a lower density. Similar results were obtained when aliquots of the same 16K pellets were incubated with pertussis toxin in the presence of[ 32P]NAD before application to the sucrose gradient columns. The proteins were separated by PAGE, and NAD-ribolsylation was quantified by densitometry of the resulting autoradiograms (Fig. 6). In four experiments, GH+Dex treatment shifted peak radioactivity at least one fraction leftward, consistent with the results obtained by Western blotting. In contrast, the highest activity of 5′-nucleotidase was found in the same fraction in control and hormone-treated adipocytes, indicating that GH+Dex did not substantially shift the distribution of this intrinsic plasma membrane protein to less dense fractions (Fig. 7).
Effect of incubation of adipocytes with GH+Dex for 4 h on the distribution of G proteins. Adipocytes were homogenized, and a 16K pellet was obtained as described in Materials and Methods. The pellets were resuspended in 1 ml, layered over linear 5–25% sucrose gradients, and centrifuged for 90 min at 100K × g. Aliquots were removed sequentially from the top of the gradient so that lower fraction numbers correspond to lower densities and were subjected to PAGE, followed by transfer to nitrocellulose and immunoblotting. Each membrane was probed sequentially for Giα2 and Gsα. Upper panels show Western blots obtained in a single representative experiment. Similar data were obtained in three additional experiments of identical design. To permit quantitative comparisons between films exposed to successive probes for Giα2 and Gsα, the films were analyzed densitometrically, and the intensity of each band on a film was normalized to the band with the highest intensity, which was assigned a value of 100. Although variation from experiment to experiment in the shapes of the curves and the location of the fractions with the highest intensity prevented statistical analysis, the distribution of Giα2 and Gsα always coincided in the membranes from control cells, while Giα2 was shifted to the left relative to Gsα.
![Pertussis toxin catalyzed [32P]NAD ribosylation of sucrose gradient-separated membrane proteins from control and GH+Dex-treated adipocytes. Aliquots of the same homogenates as those shown in Fig. 5 were incubated for 10 min with 2 μg activated pertussis toxin and 2 mm [32P]ATP before application to the sucrose gradient columns. Membrane proteins were resolved on SDS-PAGE, followed by autoradiography. The autoradiograms shown were prepared from the same membrane homogenates as depicted in Fig. 5. Densitometric analysis, again normalized by assigning the most intense band a value of 100, is shown in the lower portion of the figure. The data are representative of four independent experiments of identical design.](https://oup.silverchair-cdn.com/oup/backfile/Content_public/Journal/endo/140/3/10.1210_endo.140.3.6580/1/m_ee0396580006.jpeg?Expires=1675464273&Signature=4D9mq~52oEFqgiKi~Ailg019mmEWxKDbo6l~eqaqDcu-CWSAZQXzwEv15JDAedcXrAVcxOrztAdeW5eT~rQjY~oYuNgyrllEXo~6vdt3GZGmarn8oDJl14ClW5xjb84NMbpMYi9bKfUFzMNewZEzBjnvKPsbkFnovyQhI9F1Yo9CMkApIfk-r-Axf9xkm2fN6j1g-Tmw9CuX3rDA9xuc7jMNQenqSMU0bXFeINqCHdBWWmDXh3oJNPZNxZ3CdK1gRlwBW4TMKNqK2bvHOIwGEMIQMnO7v5tDqhePTmEEMVSve6uZUVnswd~XUN-xvIDp1LuLC9Agtu0arMhiDLc4sQ__&Key-Pair-Id=APKAIE5G5CRDK6RD3PGA)
Pertussis toxin catalyzed [32P]NAD ribosylation of sucrose gradient-separated membrane proteins from control and GH+Dex-treated adipocytes. Aliquots of the same homogenates as those shown in Fig. 5 were incubated for 10 min with 2 μg activated pertussis toxin and 2 mm [32P]ATP before application to the sucrose gradient columns. Membrane proteins were resolved on SDS-PAGE, followed by autoradiography. The autoradiograms shown were prepared from the same membrane homogenates as depicted in Fig. 5. Densitometric analysis, again normalized by assigning the most intense band a value of 100, is shown in the lower portion of the figure. The data are representative of four independent experiments of identical design.
Distribution of 5′-nucleotidase in sucrose density gradient-separated membrane proteins from control and GH+Dex-treated adipocytes. Similar results were obtained in three additional experiments of the same design.
To test the possibility that the migration of Giα2 with less dense plasma membrane fractions might involve membranous or submembranous proteins that associate with the cytoskeleton, adipocytes were incubated with GH+Dex in the absence and presence of 100μ m colchicine, and the distribution of Giα2 in the 16K and 100K pellets was again studied in immunoblots. In four experiments, treatment with GH+Dex decreased Giα2 by 43% in the 16K pellet (Fig. 8) and increased it in the 100K pellet by 57% in the cells that were incubated without colchicine. In the cells that were incubated with colchicine, GH+Dex produced no significant changes in the abundance of Giα2 in either fraction. GH+Dex also had no effect on lipolysis in colchicine-treated cells (Fig. 9), but produced the expected increase in lipolysis in cells incubated without colchicine. In contrast, ISO increased lipolysis to at least as great an extent in the presence of colchicine as in its absence. When colchicine was replaced with an equimolar amount of the inactive analog, lumicolchicine, no inhibition of lipolysis or redistribution of Giα2 was seen (data not shown).
Effects of colchicine on the distribution of Giα2 in adipocytes treated with GH+Dex. Adipocytes were incubated for 4 h in the absence or presence of 100μ m colchicine. GH+Dex was present from the start of the incubation. Adipocyte extracts were prepared and analyzed as described in Fig. 4. Bars represent the mean ± sem for four replicate experiments.
The effects of colchicine on the lipolytic response to GH+Dex or to 1 μg/ml ISO. Adipocytes were incubated for 4 h in the absence or presence of 100 μm colchicine. GH+Dex was present from the start of the incubation; ISO was present only in the final hour. Glycerol production in the fourth hour was obtained by subtracting the glycerol present in aliquots of medium sampled at the end of 3 h from that measured at the end of 4 h. Each bar represents the mean of four independent experiments. The vertical brackets indicate the sem. *, P < 0.05 compared with control values; **, P < 0.05 compared with GH+Dex without colchicine.
Discussion
The present findings support the hypothesis that the increase in lipolysis produced by GH+Dex or catecholamines results from activation of the same post-cAMP cascade of reactions. The cAMP antagonist, Rp-cAMPS, was as effective in blocking lipolysis induced by GH+Dex as that induced by ISO, and both GH+Dex and ISO increased the activity of PKA in adipocyte extracts to an extent commensurate with the increase in glycerol production. In addition, AC activity was increased in crude adipocyte membrane fractions prepared from GH+Dex-treated adipocytes. The absence of a detectable increase in cAMP in extracts of GH+Dex-treated tissues is not necessarily in conflict with these findings. Little correlation between lipolysis and cAMP concentrations measured in extracts of adipose tissue or adipocytes was found after prolonged or repeated stimulation with catecholamines (38, 39). Although the rate of lipolysis remained constant for 1 h in the continued presence of catecholamine, cAMP levels rose 2- to 4-fold within the initial 10 min and declined to only 15% above control levels by 60 min. From these results, the absence of a detectable increase in cAMP after 3 and 4 h of incubation with hormones that produce only a modest stimulation of lipolysis is therefore not surprising. In contrast to the lack of correlation between cAMP levels and lipolysis, glycerol production correlated well with the fraction of protein kinase that was present in the active form in adipocytes treated with various concentrations of a wide range of rapidly acting lipolytic agonists (41). These findings suggest that cAMP may not be uniformly distributed throughout adipocyte cytoplasm and that the fractional activation of PKA may provide a better index of the effective cAMP concentration than its concentration in total cell water. In the intact adipocyte, PKA may be favorably situated with respect to AC such that its activity is regulated by local changes in cAMP concentration that may be too small to have a significant influence on overall cellular levels. Earlier findings that GH+Dex did not augment lipolytic responses to membrane-permeant analogs of cAMP (14, 42) argue against the alternative possibility that GH+Dex enhances reactions distal to cAMP.
If, indeed, the accelerated rate of lipolysis produced by GH+Dex and by catecholamines results from increased production of cAMP, the differences between the rates of onset and the magnitudes of their lipolytic actions must reside in the manner in which they activate AC. Catecholamines activate AC through the agency of Gs, which couples their receptors to the enzyme. AC activity can also be increased by relieving inhibition of its basal activity, as evidenced by the profound stimulation of lipolysis seen when Gi is inactivated with pertussis toxin (21, 24). Furthermore, basal levels of cAMP increased by more than 3-fold in adipocytes of transgenic mice that were induced to express Giα2 antisense messenger RNA (43). The present findings are consistent with the possibility that GH+Dex increase the activity of AC by diminishing the inhibitory influence of Gi. These effects may represent the loss of intrinsic inhibitory input from Gi itself or from an autocrine inhibitory factor(s) that acts through the agency of Gi.
After treatment with GH+Dex, an appreciable amount of Giα2 was selectively associated with lower density membrane particles than the Giα2 in control cells or than Gsα, suggesting that GH+Dex either shifts the distribution of Giα2 within the membrane or alters the distribution of the intrinsic membrane proteins or submembranous proteins with which it forms complexes. This phenomenon was most dramatically revealed by the disappearance of some Giα2 from the 16K pellet that contains the bulk of the plasma membranes and its appearance in the 100K pellet that is enriched in endoplasmic reticulum but also contains nearly 20% of the plasma membrane fragments. The GH+Dex-related tendency of Giα2 that sedimented at 16K × g to migrate with lower density membrane fragments on sucrose gradients suggests that fragments of plasma membrane that sediment only at higher g forces may be enriched in Giα2 and thus account for these findings. Alternatively, it is possible that GH+Dex caused some transfer of Giα2 from the plasma membrane to less dense intracellular vesicles. Indeed, immunofluorescence microscopy of cultured rat and mouse adipocytes indicates that Giα2 is not confined to the plasma membrane, and that a considerable amount of Giα2 is associated with internal structures (44, 45). Clearly, the 16K fraction is not free of endoplasmic reticulum, nor is the 100K fraction free of plasma membranes, and assignment of a morphological location based upon centrifugal separation is not warranted.
The finding that Giα2 redistributes to less dense membrane fragments after treatment with GH+Dex may help explain some conflicting reports in the literature. Decreased responsiveness to adenosine, whose actions are mediated by Gi, or to the nonmetabolizable phenylisopropyl adenosine without a change in the abundance of Gi in adipocyte plasma membranes has been reported after treatment of sheep (46) or cattle (47) with GH. The membrane fractions prepared from ovine adipocytes were pelleted at 42,000 × g (46), and the fraction prepared from bovine adipose tissue was sedimented at 100,000 × g (47). Membranes prepared by both of these techniques include all of the Giα2 that we would expect to find in the 16K pellet plus most or all of the Giα2 associated with the less dense 100K pellet. That is, these preparative procedures would have missed the decrease in Giα2 in the denser membranous components when the 16K pellet is studied. Similarly, Roupas et al. (22) reported that chronic treatment of ob/ob mice with a diabetogenic analog of GH interfered with the functional relationship between Gi and activation of phospholipase C. They also found no change in the abundance of Gi present in the adipocyte membranes that they sedimented at 28K × g, although they detected a decrease in pertussis toxin-catalyzed NAD ribosylation. Conversely, Doris et al. (23) found an increase in Giα2 in a 15K pellet prepared from rat adipocytes 3 days after lowering circulating GH concentrations by injection of antiserum to rat GH. The increase was prevented by simultaneous treatment with ovine GH. These findings are in harmony with the present findings and our initial report that hypophysectomy increased and GH decreased pertussis toxin-catalyzed ADP ribosylation in rat adipocytes (21)
Haraguchi and Rodbell (48) reported a similar shift of Giα, Gsα, AC, and 5′-nucleotidase from the plasma membrane fraction to lighter fractions after stimulation of rat adipocytes with a supramaximal concentration (10 μm) of isoproterenol. They attributed this redistribution of membrane proteins to pinocytosis. Similar shifts ofα -subunits of Gq and G11 to less dense membranous components were reported as a prelude to down-regulation in response to TRH (49) or acetylcholine (50). In this regard, a high dose of TRH led first to the migration of Gq to discrete patches within the plasma membrane and subsequently to intracellular vesicles (51). The present observations differ from those of Haraguchi and Rodbell, in that only Giα2 was shifted to lower density fractions, leaving the distribution of Gsα, 5′-nucleotidase, and AC unchanged. The possibility that the effects of GH+Dex on Giα2 as reported here might also reflect a micropinocytotic process cannot be ruled out. Alternatively, G proteinα -subunits are known to associate directly with tubulin (51–54) and coimmunoprecipitate with tubulin from detergent extracts of cerebral cortical membranes (54). GH+Dex might selectively destabilize protein:protein interactions within membranes and displace Giα2 from some complexes. Involvement of tubulin in the GH+Dex-dependent redistribution of Giα2 and the stimulation of lipolysis is suggested by the findings that the microtubule disrupter, colchicine, prevented both the shift in density and the stimulation of lipolysis. Regardless of the mechanism, it appears that treatment with GH+Dex changes the physical relationship of Giα2 to both Gsα and AC. That change appears to be sufficient to decrease inhibitory input to AC and relieve some restraint on cAMP production.
This work was supported by NIDDK Grant DK-19392. Its contents are solely the responsibility of the authors and do not necessarily represent the official views of the NIH. These studies were presented to the University of Massachusetts Graduate School of Biomedical Sciences by R.G.-C.Y. in partial fulfillment of the requirements for the Ph.D. degree. Preliminary reports of these studies were presented at the Keystone Symposium, The Adipose Cell, Park City, Utah, 1994, and the 76th Meeting of The Endocrine Society, Anaheim, California, 1994.
Present address: Boston University School of Medicine, Boston Medical Center, Evans 201, 88 East Newton Street, Boston, Massachusetts 02000.
