INTRODUCTION

The existence of multiple somatostatin receptors was first suggested by radioligand binding studies showing differential affinities in various tissues for the endogenous ligands, somatostatin-14 and somatostatin-28, and for octapeptide and hexapeptide somatostatin analogs (1, 2, 3, 4). Five distinct somatostatin receptors have since been cloned from the human and rat, designated sstr1, sstr2, sstr3, sstr4, and sstr5; sstr2 exists as two splice variants, sstr2a and sstr2b (3, 4, 5, 6, 7). All five somatostatin receptors are expressed in the brain, pituitary gland, and pancreatic islets; expression of a specific somatostatin receptor in other tissues is variable and overlaps with expression of other somatostatin receptors (2, 4, 8). All five somatostatin receptors belong to the family of G protein-coupled receptors and exhibit motifs for regulatory phosphorylation by protein kinase A, protein kinase C, and calmodulin kinase 2 (4). Several signaling pathways are known to be activated by one or more of these receptors via pertussis toxin-sensitive inhibitory G proteins. These include inhibition of adenylyl cyclase (2, 3, 4), activation of several types of K⁺ channels (inward rectifying), delayed rectifying, ATP-sensitive (4, 9), and large conductance Ca²⁺-activated K⁺ channels (10), inhibition of voltage-dependent Ca²⁺ channels (11), activation of several enzymes including phospholipase C (PLC)¹ (12, 13, 14), phospholipase A₂ (PLA₂) (15, 16), mitogen-activated protein kinase (15, 16), and serine/threonine and phosphotyrosine phosphatases (10, 17).

Various inhibitory G proteins couple somatostatin receptors to specific responses in native cells and transformed cell lines (1, 2, 3, 4, 18). The coupling of somatostatin receptors to inhibition of adenylyl cyclase has been extensively studied (19, 20, 21). Apparently discordant results have been obtained that mainly reflect the extent or stability of expression of cloned receptors and/or the absence of a full complement of inhibitory G proteins (4, 22). When expressed in COS-7 cells, which possess a full complement of G proteins, all five human somatostatin receptors are coupled to inhibition of adenylyl cyclase via unidentified pertussis toxin-sensitive G protein(s) (20). When expressed in CHO cells, which lack one or more G proteins, sstr1, sstr2, or sstr3 are variously coupled to inhibition of adenylyl cyclase via G_i1, G_i2, G_i3, or G_o (2, 3, 4, 13, 23). In pituitary GH₄C₁ cells, which mainly express ssrt2 receptors inhibition of adenylyl cyclase is mediated by G_i2 and G_i3, while in pituitary AtT-20 cells, it is mediated by G_i1 (21, 24). Recent studies suggest that all five human somatostatin receptors expressed in COS-7 cells are coupled to activation of phosphoinositide (PI)-specific PLC- and Ca²⁺ mobilization via pertussis toxin-sensitive G protein(s) with an order of potency of sstr5 > sstr2 > sstr3 > sstr4  sstr1 (12). The human sstr1 expressed in monoclonal CHO cells mediates stimulation of IP₃ formation (13). Neither the PLC- isozyme nor the type of inhibitory G protein mediating PI hydrolysis has been identified.

In the present study, we have used dispersed intestinal smooth muscle cells and plasma membranes to characterize the signaling pathways initiated by somatostatin receptors. Specific G protein and PLC- antibodies were used to identify the G protein subunits and PLC- isozymes mediating phosphoinositide hydrolysis, Ca²⁺ mobilization, and contraction, and the G protein subunits mediating inhibition of adenylyl cyclase.

EXPERIMENTAL PROCEDURES

Muscle cells were isolated separately from the circular and longitudinal muscle layers of guinea pig intestine by sequential enzymatic digestion, filtration, and centrifugation as described previously (25). Briefly, muscle strips were incubated at 31 °C for 30 min in HEPES medium with Type II collagenase (0.1%) and soybean trypsin inhibitor (0.1%). The muscle strips were then washed and the cells allowed to disperse spontaneously for 30 min. The cells were harvested by filtration through 500-µm Nitex and centrifuged twice at 350 × g for 10 min.

In experiments with blocking antibodies, the cells were permeabilized as described previously (25, 26) by incubation for 10 min with saponin (35 µg/ml) in a medium containing 20 m NaCl, 100 m KCl, 5 m MgSO₄, 1 m NaH₂PO₄, 25 m NaHCO₃, 0.34 m CaCl₂, 1 m EGTA, and 1% bovine serum albumin. The cells were centrifuged at 350 × g for 5 min, washed free of saponin, and resuspended in the same medium with 1.5 m ATP and ATP-regenerating system (5 m creatine phosphate and 10 units/ml creatine phosphokinase).

Muscle cells were suspended in HEPES medium containing 1% bovine serum albumin, 10 µ amastatin, 1 µ phosphoramidon, and 0.7 m bacitracin. Triplicate aliquots (0.3 ml) of cell suspension (10⁶ cells/ml) were incubated with 50 p ¹²⁵I-somatostatin-14 alone or in the presence of different concentrations (1 p to 10 µ) unlabeled ligand. Bound and free radioligand were separated by rapid filtration under reduced pressure through 5 µ polycarbonate Nucleopore filters, followed by repeated washing (three times) with 3 ml of ice-cold HEPES medium containing 0.2% bovine serum albumin (27). Nonspecific binding was measured as the amount of radioactivity associated with the muscle cells in the presence of 10 µ unlabeled ligand. Specific binding was calculated as the difference between total and nonspecific binding. Nonspecific binding was 21.5 ± 4.8% of total binding.

cAMP was measured in intact and permeabilized circular muscle cells by radioimmunoassay as described previously (25). Lyophilized samples were reconstituted in 500 µl of 50 m sodium acetate (pH 6.2) and acetylated with triethylamine/acetic anhydride (3:1, v/v) for 30 min before radioimmunoassay. cAMP was measured in duplicate and the results expressed as pmol/10⁶ cells.

1,4,5-IP₃ mass was measured in intact and permeabilized muscle cells as described previously (25, 28) using Amersham's assay system. One ml of muscle cell suspension (10⁶ cells/ml) was incubated with Li⁺ for 10 min at 31 °C, after which somatostatin was added for 30 s and the reaction terminated with ice-cold 10% perchloric acid. After centrifugation for 10 min at 750 × g, the supernatant was extracted and IP₃ content in the aqueous phase was measured. Results were expressed as pmol/10⁶ cells.

Ca²⁺ release was measured in dispersed intestinal muscle cells as described previously (25, 26). Muscle cells (10⁶ cells/ml) were incubated with ⁴⁵Ca²⁺ (10 µCi/ml) and antimycin (10 µ); an ATP regenerating system (5 m creatine phosphate and 10 units/ml creatine phosphokinase) was present when measurements were made in permeabilized cells. Ca²⁺ uptake was initiated with 1.5 m ATP and measured at intervals for 90 min when a steady state was attained. Somatostatin was then added and ⁴⁵Ca²⁺ cell content determined after 30 s. Ca²⁺ efflux was expressed as percent decrease in steady-state ⁴⁵Ca²⁺ cell content. In other experiments, G protein or PLC- antibodies were added 30 min after ATP and the incubation maintained for another 60 min. Addition of antibodies had no effect on steady-state Ca²⁺ uptake.

Contraction was measured in intact and permeabilized muscle cells by scanning micrometry as described previously (25, 26). A cell aliquot containing 10⁴ muscle cells/ml was added to 0.1 ml of medium containing somatostatin and the reaction terminated after 30 s with 1% acrolein. The effect of PLC- and G protein antibodies was determined in permeabilized muscle cells after preincubation for 1 h with 10 µg/ml amounts of each antibody separately. The lengths of muscle cells treated with somatostatin were compared with the length of untreated cells and contraction expressed as percent decrease in mean cell length.

PLC- Activity in Plasma Membranes

PLC- activity was determined in plasma membranes prelabeled with myo-[³H]inositol by a modification of the method of Uhing et al. (29) as described previously (25, 30, 31). A plasma membrane fraction was obtained from cell homogenates by centrifugation at 200,000 × g for 2 h and resuspended in 50 m Tris-HCl (pH 7.5), 1 m EGTA, 100 µg/ml leupeptin, and 100 µg/ml antipain to yield a final concentration of 2 mg of protein/ml (10,000-20,000 cpm/mg of protein). PLC assay was initiated by addition of 0.4 mg of membrane protein to 25 m Tris-HCl (pH 7.5), 0.5 m EGTA, 10 m MgCl₂, 300 n free Ca²⁺, 10 µ GTP, 5 m phosphocreatine, and 50 units/ml creatine phosphokinase in a total volume of 0.4 ml. After incubation at 31 °C for 60 s, the reaction was terminated with 25% trichloroacetic acid (w/v) and the supernatant extracted with diethyl ether; the amount of labeled inositol phosphates in the aqueous phase was counted. The trichloroacetic acid-soluble radioactivity at zero time (~150 cpm) was subtracted from all values. PLC activity was expressed as cpm/mg of protein × min.

Cell homogenates were prepared from dispersed intestinal circular muscle cells and solubilized on ice for 1 h in 20 m Tris (pH 8.0), 1 m EDTA, 1 m dithiothreitol, 100 m NaCl, and 0.5% sodium cholate (25). The suspension was centrifuged at 13,000 × g for 5 min. Solubilized membrane proteins (60-70 µg) were resolved by 12% SDS-PAGE and electrophoretically transferred to nitrocellulose membranes. After incubation in 5% nonfat dry milk to block nonspecific antibody binding, the blots were incubated for 12 h at 4 °C with G protein antibodies and for 1 h with anti-rabbit IgG conjugated with horseradish peroxidase. The bands were identified by enhanced chemiluminescence.

G proteins selectively activated by somatostatin were identified by the method of Okamoto et al. (32). Ten ml of muscle cell suspension (2 × 10⁶ cells/ml) were homogenized in 20 m HEPES medium (pH 7.4) containing 2 m MgCl₂, 1 m EDTA, and 2 m dithiothreitol. After centrifugation at 27,000 × g for 15 min, the crude membranes were solubilized for 60 min at 4 °C in 20 m HEPES medium (pH 7.4) containing 2 m EDTA, 240 m NaCl, and 1% CHAPS. The membranes were incubated for various periods at 37 °C with 60 n [³⁵S]GTPS in a solution containing 10 m HEPES (pH 7.4), 100 µ EDTA and 10 m MgCl₂. The reaction was stopped with 10 volumes of 100 m Tris-HCl medium (pH 8.0) containing 10 m MgCl₂, 100 m NaCl, and 20 µ GTP, and the mixture placed in wells precoated with specific G protein antibodies. After incubation for 2 h on ice, the wells were washed three times with phosphate buffer solution containing 0.05% Tween 20 and the radioactivity from each well was counted. Coating with G protein antibodies (1:1000) was done after the wells were first coated with anti-rabbit IgG (1:1000) for 2 h on ice.

Results were calculated as means ± S.E. of n experiments using cell suspensions obtained from different animals. Statistical significance was evaluated using Student's t test for paired or unpaired values.

Monoclonal antibody to PLC-1 was obtained from Upstate Biotechnology Inc., Lake Placid, NY, and polyclonal antibodies to PLC-2, PLC-3, PLC-4, G_o, and G_q/11 were obtained from Santa Cruz Biotechnology Inc., Santa Cruz, CA. Pertussis toxin (PTx) and polyclonal antibodies to G_i1, G_i1-2, G_i3, G_s, and G were obtained from Calbiochem, and monoclonal antibody to G_i2 was obtained from Chemicon, Temecula, CA. ¹²⁵I-cAMP, ¹²⁵I-somatostatin-14, myo-[³H] inositol, [³⁵S]GTPS, and ⁴⁵CaCl₂ were obtained from DuPont NEN; nitrocellulose membranes from Bio-Rad; collagenase type II and soybean trypsin inhibitor from Worthington; and -myo-inositol 1,4,5-trisphosphate assay system from Amersham Corp. All other chemicals were obtained from Sigma. Somatostatin analogs, SMS 201-995 (octreotide), -Phe-Phe-Tyr--Trp-Lys-Val-Phe--Nal-NH₂ (SS-939), -Phe-Phe-Phe--Trp-Lys-Thr-Phe-Thr-NH₂ (SS-938), and -Phe-Cys-Tyr--Trp-Lys-Abu-Cys-Nal-NH₂ (SS-931) were synthesized by one of us (D. H. C.) (33).

RESULTS

Western blot analysis of solubilized membrane fractions derived from dispersed intestinal muscle cells demonstrated the presence of a full complement of G proteins: G_q/11, G_s, G_o, G_i1, G_i2, and G_i3; a band elicited with G_i1-2 antibody reflects the presence of G_i1 and G_i2 (Fig. 1). Western blot analysis of PLC- isozymes disclosed the presence of PLC-1, PLC-2, PLC-3, and PLC-4 isozymes (34).

The G protein antibodies used in Western blot analysis were used to identify the G proteins activated by somatostatin in plasma membranes. Incubation of muscle cell membranes with somatostatin (1 µ) and [³⁵S]GTPS (60 n) for 20 min caused a significant, time-dependent increase in the binding of [³⁵S]GTPS-G complexes to wells precoated with specific antibody to G_i1, G_i1-2, and G_o, but not to wells precoated with antibodies to G_i2, G_i3, G_s, and G_q/11 (Fig. 2 and Table I). The increase in bound radioactivity reflected somatostatin-dependent activation of the dissociated  subunits of G_i1 and G_o by [³⁵S]GTPS; the increase observed with G_i1-2 antibody reflected the activation of G_i1.

Time course of binding of somatostatin-stimulated GTPS-G

i1

i1

Table I. The binding of somatostatin-stimulated GTPS-G complexes in smooth muscle membranes to specific G protein antibodies (Ab) Crude membranes from dispersed circular muscle cells were incubated for 20 min with [35S]GTPS in the absence (basal) or presence of 1 µ somatostatin and then added to wells precoated with specific G protein antibodies. Increase in binding of [35S]GTPS-G complexes was obtained in wells precoated with Gi1 and Go antibodies. Values are means ± S.E. of four experiments. **, p < 0.01.  [35S]GTPS-G protein complex bound Basal + Somatostatin % increase cpm Gi1 Ab 3444  ± 288 9237  ± 1042 168  ± 30** Gi2 Ab 4010  ± 201 4357  ± 267 8  ± 7 Gi3 Ab 3643  ± 1377 3695  ± 538 1  ± 12 Go Ab 3643  ± 377 6654  ± 566 82  ± 15** Gq/11 3544  ± 381 3852  ± 333 9  ± 9

Somatostatin-14 bound specifically to intestinal circular but not longitudinal muscle cells. ¹²⁵I-Somatostatin-14 binding was inhibited in a concentration-dependent fashion by unlabeled somatostatin with an IC₅₀ of 2.3 ± 0.8 n (Fig. 3).

Somatostatin-14 caused concentration-dependent contraction of dispersed intestinal circular but not longitudinal muscle cells with an EC₅₀ of 2 ± 1 n (Fig. 3). Three somatostatin analogs (SS-931, SS-939, and SMS201-995), which shared high affinity for sstr3 but exhibited different patterns of selectivity for other somatostatin receptors caused concentration-dependent contraction. The EC₅₀ values for SS-931 (low affinity, i.e. >1 µ, for sstr1 and sstr5), SMS201-995 (low affinity for sstr1 and sstr4) and SS-939 (low affinity for sstr1 and sstr2) were 50, 100, and 250 p, respectively (Fig. 3). The potency of these analogs implied interaction with sstr3 on intestinal circular smooth muscle cells.

Somatostatin-14 increased Ca²⁺ release and IP₃ formation in a concentration-dependent fashion with EC₅₀s of 10 ± 4 and 45 ± 6 n, respectively (Fig. 4). Maximal Ca²⁺ release (27 ± 2% decrease in steady-state ⁴⁵Ca²⁺ cell content) and IP₃ formation (5.0 ± 0.5 pmol/10⁶ cells above a basal level of 3.5 ± 0.3 pmol/10⁶ cells) were elicited by 1 µ somatostatin. Treatment of the cells for 1 h with PTx (800 ng/ml) inhibited maximal somatostatin-induced IP₃ formation, Ca²⁺ release, and contraction by 84 ± 2%, 77 ± 3% and 86 ± 6%, respectively (Table II).

Table II. Effect of pertussis toxin (PTx; 800 ng/ml) on somatostatin-induced PLC activity in plasma membranes, and IP3 formation, Ca2+ release, muscle contraction, and forskolin-stimulated cAMP in dispersed intestinal muscle cells PLC activity was expressed as increase in inositol phosphate formation above basal levels (598 ± 49 cpm/mg protein × min). 1,4,5-IP3 mass was measured by radioreceptor assay and expressed as increase in pmol/106 cells above basal levels (3.50 ± 0.3 pmol/106 cells). Ca2+ release was expressed as % decrease in steady-state Ca2+ content (steady-state levels: 2.65 ± 0.43 nmol/106 cells) in cells preloaded with 45Ca2+. Contraction was expressed as % decrease in control cell length (115 ± 3 µm). cAMP was measured in the presence of 10 µ forskolin and the results expressed as pmol/106 cells above basal levels (basal: 4.3 ± 0.2 pmol/106 cells: forskolin:  20.7 ± 1.6 pmol/106 cells; forskolin plus somatostatin:  3.6 ± 0.6 pmol/106 cells or 82 ± 3% inhibition in the absence of PTx). Values are means ± S.E. of four to six experiments. **, p < 0.01.  Somatostatin Somatostatin + PTx % inhibition PLC (cpm/mg protein × min) 2250  ± 237 329  ± 117** 85  ± 5 IP3 (pmol/106 cells) 5.0  ± 0.5 0.8  ± 0.1** 84  ± 2 Ca2+ release (%) 23.8  ± 3.1 5.5  ± 0.8** 77  ± 3 Contraction (%) 26.1  ± 1.6 3.6  ± 1.6** 86  ± 6 cAMP (pmol/106 cells) 3.6  ± 0.6** 17.7  ± 1.4 14  ± 7

Somatostatin inhibited forskolin-stimulated cAMP formation in a concentration-dependent fashion with an EC₅₀ of 0.23 ± 0.14 n (Fig. 4). Maximal inhibition of cAMP formation was elicited by 0.1 µ somatostatin (forskolin alone: 20.7 ± 2.0 pmol/10⁶ cells; forskolin plus somatostatin: 3.4 ± 0.6 pmol/10⁶ cells or 83 ± 4% inhibition). Treatment of the cells for 1 h with PTx reversed the inhibition of forskolin-stimulated cAMP formation (forskolin: 19.8 ± 1.2 pmol/10⁶ cells; forskolin plus somatostatin: 3.6 ± 0.6 pmol/10⁶ cells; forskolin plus somatostatin and PTx: 17.7 ± 1.4 pmol/10⁶ cells) (Table II).

Identification of PLC- Isozymes and G Proteins Functionally Coupled to Somatostatin Receptors

PI-specific PLC activity (inositol phosphate formation) measured in plasma membranes derived from dispersed intestinal circular muscle cells increased by 514 ± 54% above basal level upon addition of 10 µ GTP and 1 µ somatostatin. Pretreatment of the cells for 1 h with 800 ng/ml PTx before membrane isolation inhibited PLC activity in plasma membranes by 85 ± 5% (Table II). Pretreatment of plasma membranes for 1 h with 10 µg/ml PLC-3 antibody inhibited PLC activity by 67 ± 3% (Table III), whereas pretreatment with 10 µg/ml PLC-1, PLC-2, or PLC-4 antibody had no significant effect (range of inhibition: 6 ± 4% to 15 ± 8%; not significant).

Table III. Effect of antibodies to PLC- isozymes (10 µg/ml) and G protein subunits (10 µg/ml) on 1 µ somatostatin-stimulated PLC activity in plasma membranes, and IP3 formation, Ca2+ release, and contraction in dispersed permeabilized intestinal muscle cells PLC activity was expressed as increase in inositol phosphate formation above basal levels (437 ± 56 cpm/mg protein × min). IP3 mass was measured by radioreceptor assay and expressed as increase above basal levels (3.03 ± 0.2 pmol/106 cells). Ca2+ release was expressed as percent decrease in steady-state Ca2+ content (2.52 ± 0.46 nmol/106 cells) in cells preloaded with 45Ca2+. Cell contraction was measured by scanning micrometry and expressed as percent decrease in control cell length (102 ± 2 µm). Values are means ± S.E. of four to five experiments. **, p < 0.01.  PLC Activity 1,4,5-IP3 Ca2+ Release Contraction  cpm/mg-min  pmol/106 cells  % in cell Ca2+  % in length Somatostatin 2250  ± 237 4.7  ± 0.2 24.3  ± 2.2 25.0  ± 1.5 + PLC-1 Ab 1912  ± 134 4.2  ± 0.5 26.1  ± 1.9 23.3  ± 1.5 + PLC-2 Ab 2114  ± 198 4.1  ± 0.6 25.4  ± 2.5 25.7  ± 0.8 + PLC-3 Ab 747  ± 135** 1.4  ± 0.4** 5.5  ± 1.1** 4.2  ± 0.4** + PLC-4 Ab 2051  ± 134 4.9  ± 0.6 24.6  ± 3.2 26.3  ± 1.6 + Gi1 Ab 2083  ± 112 4.4  ± 0.4 24.8  ± 2.9 23.9  ± 2.5 + Gi1-2 Ab 2094  ± 146 4.7  ± 0.3 24.0  ± 2.1 23.0  ± 1.3 + Go Ab 2067  ± 139 5.1  ± 0.7 25.3  ± 4.4 26.9  ± 2.9 + Gi3 Ab 2107  ± 149 4.4  ± 0.3 23.3  ± 1.7 24.6  ± 2.5 + G Ab 828  ± 108** 1.5  ± 0.3** 5.0  ± 0.7** 3.4  ± 1.0** + Gq/11 Ab 2111  ± 135 4.2  ± 0.6 22.9  ± 1.2 23.8  ± 0.8

The G protein antibodies used in Western blot analysis were used to identify the G proteins coupled to activation of PLC-3. Pretreatment of plasma membranes with 10 µg/ml common G antibody inhibited somatostatin-stimulated PLC activity by 63 ± 5%, while pretreatment with 10 µg/ml G_i1-2, G_i1, G_i3, G_o, G_s, or G_q/11 antibody had no significant effect (range: 6 ± 6% to 8 ± 6%; not significant) (Table III).

The effect of PLC- and G protein antibodies on contraction, Ca²⁺ release, and IP₃ formation was examined in permeabilized muscle cells. Permeabilization had no significant effect on somatostatin-induced contraction (26.1 ± 2.7% decrease in cell length in intact cells versus 25.0 ± 1.5% in permeabilized cells). Pretreatment of the cells with 10 µg/ml G antibody inhibited maximal somatostatin-induced IP₃ formation, Ca²⁺ release, and contraction by 68 ± 6%, 79 ± 3%, and 86 ± 4%, respectively (Table III). Pretreatment of the cells with 10 µg/ml PLC-3 antibody had a similar effect, inhibiting somatostatin-induced IP₃ formation, Ca²⁺ release, and contraction by 71 ± 8%, 77 ± 4%, and 83 ± 2%, respectively (Table III). No other G protein or PLC- antibody had any significant effect on contraction, IP₃ formation, or Ca²⁺ release (range: 1 ± 9% to 12 ± 12%; not significant). Contraction and Ca²⁺ release induced by exogenous IP₃ (1 µ) were not affected by pretreatment of the cells with either G or PLC-3 antibody (inhibition: 3 ± 5% and 2 ± 4%; not significant).

G protein antibodies were also used to explore the coupling of somatostatin receptors to inhibition of adenylyl cyclase in permeabilized muscle cells. Forskolin stimulated the formation of cAMP to the same extent in intact (20.7 ± 2.0 pmol/10⁶ cells) and permeabilized (20.6 ± 1.2 pmol/10⁶ cells) muscle cells, and somatostatin (1 µ) inhibited forskolin-stimulated cAMP to the same extent in intact (82 ± 3%) and permeabilized muscle cells (78 ± 4%). Pretreatment of permeabilized muscle cells with 10 µg/ml G_i1 and G_o antibody reversed somatostatin-induced inhibition of forskolin-stimulated cAMP to 48 ± 2% and 63 ± 5%, respectively (Fig. 5). The effects of 10 µg/ml G_i1 or G_o antibody were optimal since they were not exceeded when a higher concentration (20 µg/ml) was used (47 ± 3% and 59 ± 3% inhibition, respectively). The effect of a combination of G_i1 and G_o antibodies, however, was additive reversing inhibition to 26 ± 6% when 10 µg/ml amounts of each antibody were used, and 23 ± 7% when 20 µg/ml amounts of each antibody were used (Fig. 5).

Pretreatment of permeabilized muscle cells with 10 µg/ml G_i1-2 antibody reversed somatostatin-induced inhibition of forskolin-stimulated cAMP formation to 52 ± 6%. The effect was not exceeded when G_i1-2 antibody was combined with G_i1 antibody (50 ± 10%), implying that the effect of G_i1-2 antibody reflected blockade of G_i1. Pretreatment with 10 µg/ml G_i2, G_i3, G_q/11, or G had no significant effect on somatostatin-induced inhibition of forskolin-stimulated cAMP (range: 75 ± 3% to 76 ± 4%).

DISCUSSION

The present study was performed in native cells (the smooth muscle cells of the circular muscle layer of the intestine) that express one somatostatin receptor type, sstr3, and a full complement of G proteins (G_q/11, G_s, G_i1, G_i2, G_i3, and G_o) and PLC- isozymes (PLC-1, PLC-2, PLC-3, and PLC-4). The combination of properties rendered these cells a suitable preparation with which to examine two major signaling pathways linked to somatostatin receptors: inhibition of adenylyl cyclase and activation of PLC-. Most recent studies on G protein-dependent coupling of somatostatin receptors to signaling pathways have used cloned receptors expressed in a variety of cell lines (2, 3, 4, 12, 13, 14). However, the absence of a full complement of G proteins in some cell lines and variable degrees of receptor expression and/or receptor-G protein coupling have complicated analysis (4, 22). The somatostatin receptors in intestinal smooth muscle were coupled to two pertussis toxin-sensitive G proteins, G_i1 and G_o, the  subunits of which inhibited adenylyl cyclase, while their subunits activated a specific PLC- isozyme, PLC-3, to stimulate PI hydrolysis, IP₃-dependent Ca²⁺ release, and muscle contraction.

The exclusive presence of sstr3 in intestinal smooth muscle was demonstrated using three octapeptide somatostatin analogs, SMS201-995, SS-931, and SS-939, known to possess high affinity for sstr3 and low affinity (>1 µ) for various combinations of cloned sstr1, sstr2, sstr4, and sstr5. All three analogs caused muscle contraction with EC₅₀ values in the 50-250 p range consistent with interaction with sstr3. Radioligand binding studies using these and other somatostatin analogs also showed expression of sstr3 in gastric smooth muscle cells of the same species (35).

The G proteins activated by sstr3 were identified using specific G protein antibodies. Two approaches were used that yielded complementary results. In the first, the G protein antibodies were used to identify the receptor-activated  subunits bound to GTPS. In the second approach, the same G protein antibodies were used to block receptor-stimulated responses in plasma membranes (PI hydrolysis) and permeabilized muscle cells (IP₃ formation, Ca²⁺ release, muscle contraction, and inhibition of cAMP formation). The results may be summarized as follows.

1) Somatostatin stimulated the binding of [³⁵S]GTPS to the  subunits of G_i1 and G_o but not to the  subunits of other G proteins.

2) Somatostatin stimulated IP₃ formation, Ca²⁺ release, and contraction; inhibited forskolin-stimulated cAMP formation in dispersed muscle cells; and stimulated PI hydrolysis in smooth muscle cell membranes. All the effects were blocked by pretreatment of the cells with pertussis toxin.

3) Somatostatin-stimulated IP₃ formation, Ca²⁺ release and contraction in permeabilized muscle cells, and PI hydrolysis in plasma membranes were selectively blocked by G antibody and PLC-3 antibody.

4) Inhibition of forskolin-stimulated cAMP formation by somatostatin was additively blocked by G_i1 and G_o antibodies, implying the participation of both G proteins in this response and confirming the results showing selective binding of [³⁵S]GTPS to G_i1 and G_o in plasma membranes.

Somatostatin receptors are known to share several properties with other receptors coupled to inhibitory G proteins (e.g. opioid µ-, -, and -receptors, adenosine-A₁ receptors, and muscarinic-M₂ receptors). All these receptors activate various K⁺ channels, inhibit voltage-dependent Ca²⁺ channels, and inhibit adenylyl cyclase activity (1, 36, 37). In intestinal smooth muscle cells, these receptors activate also PLC-3 (25, 34). However, the G proteins involved in inhibition of adenylyl cyclase and activation of PLC-3 differ for each receptor. Somatostatin receptors (sstr3 in this study) are coupled to inhibition of adenylyl cyclase via G_i1 and G_o and to activation of PLC-3 via the G subunits of both G proteins, while opioid µ-, -, and -receptors are coupled to inhibition of adenylyl cyclase via G_i2 and G_o and to activation of PLC-3 via the G subunits of both G proteins (34). Adenosine-A₁ receptors, on the other hand, are coupled to inhibition of adenylyl cyclase via G_i3 only and to activation of PLC-3 via both the  and subunits of G_i3 (25). The activation of PLC-3 by G conforms to a pattern of preferential activation of this PLC- isozyme by receptors coupled to inhibitory G proteins (38, 39, 40). In contrast to somatostatin, adenosine, and opioid receptors, cholecystokinin receptors are coupled via G_q to activation predominantly of PLC-1 (80%), and to a lesser extent, PLC-3 (20%) (31). It is noteworthy that the distinctive patterns of signaling by somatostatin, adenosine, opioid, and cholecystokinin receptors were identified in intestinal smooth muscle cells using the same panels of PLC- and G protein antibodies providing support for the effectiveness and selectivity of these antibodies in blocking function (25, 30, 31, 34).

Somatostatin-14, like opioid µ-, -, and -receptor agonists, did not bind to or cause contraction of muscle cells from the adjacent longitudinal muscle layer (Ref. 27 and present study). However, adenosine acting via A₁ receptors caused pertussis toxin-sensitive Ca²⁺ mobilization and contraction (41). Although longitudinal muscle expresses a full complement of PLC- isozymes, neither PLC-1 nor PLC-3 appears to be involved in Ca²⁺ mobilization and muscle contraction (31, 34). The main PI substrate in longitudinal muscle was shown to be phosphatidylinositol 4-monophosphate, and PI hydrolysis generated only minimal amounts of IP₃ (28). Agonist-induced contraction is mediated by G protein-dependent activation of PLA₂, which initiates a cascade involving arachidonic acid-induced Ca²⁺ influx and Ca²⁺-induced Ca²⁺ release via ryanodine-sensitive, IP₃-insensitive Ca²⁺ channels (42, 43, 44). Preliminary studies suggest that activation of PLA₂ may be mediated by subunits of inhibitory G proteins.²

Evidence for stimulation of PI hydrolysis by somatostatin has been obtained for all five cloned receptors expressed in COS-7 cells, and for sstr1 expressed in CHO cells (12, 13). The PLC- isozyme mediating this effect has not been identified; the involvement of one or more inhibitory G protein suggests preferential activation of PLC-2 or PLC-3 by abundant subunits (13, 38, 39, 40).

In summary, this study identified two signaling pathways initiated by interaction of somatostatin with sstr3 receptors in intestinal smooth muscle. The pathways involve activation of two G proteins, G_i1 and G_o: the  subunits of both G proteins mediate inhibition of adenylyl cyclase while the subunits mediate activation of PLC-3 resulting in IP₃-dependent Ca²⁺ release and muscle contraction.