Pertussis Toxin Blocks Both Cyclic AMP-mediated and Cyclic AMP ...

THE JOURNAL OF BIOLOGICAL CHEMISTRY 0 1985 by The American Society of Biological Chemists, Inc.

Vol. 260, No. 24, Issue of October 25, pp. 13138-13145.1985 Printed in U.S.A.

Pertussis Toxin Blocks Both Cyclic AMP-mediated and Cyclic AMP-independent Actions of Somatostatin EVIDENCE FOR COUPLING OF Ni TO DECREASES IN INTRACELLULAR FREE CALCIUM*

(Received for publication, June 17, 1985)

Bruce D. KochSgli, Laneta J. Dorflingerell , and Agnes SchonbrunnSg From the $Cell and Developmental Biology Program, Haruard Medical School, and the $%.uboratofy of Toxicology, Harvard School of Public Health, Boston, Massachusetts 02115

The neuropeptide somatostatin inhibits hormone release from G&C1 pituitary cells via two mechanisms: inhibition of stimulated adenylate cyclase and a CAMP- independent process. To determine whether both mechanisms involve the guanyl nucleotide-binding protein Ni, we used pertussis toxin, which ADP-ribosylates Ni and thereby blocks its function.

Pertussis toxin treatment of G&C1 cells blocked somatostatin inhibition of both vasoactive intestinal peptide (VIP)-stimulated CAMP accumulation and prolactin secretion. In membranes prepared from toxin- treated cells, somatostatin inhibition of VIP-stimulated adenylate cyclase activity was reduced and lzsI-Tyrl- somatostatin binding was decreased more than 95%. In contrast, pertussis toxin did not affect the biological actions or the membrane binding of thyrotropin-releasing hormone. These results indicate that ADP-ribosylated Ni cannot interact with occupied somatostatin receptors and that somatostatin inhibits VIP- stimulated adenylate cyclase via Ni.

To investigate somatostatin’s CAMP-independent mechanism, we used depolarizing concentrations of K+ to stimulate prolactin release without altering intracellular CAMP levels. Measurement of Quin-2 fluorescence showed that 11 mM K+ increased intracellular [Ca”] within 5 s. Somatostatin caused an immediate, but transient, decrease in both basal and K+-elevated [Ca2+]. Consistent with these findings, somatostatin inhibited K+-stimulated prolactin release, also without affecting intracellular CAMP concentrations. Pertussis toxin blocked the somatostatin-induced reduction of [Ca”’]. Furthermore, the toxin antagonized somatostatin inhibition of K+-stimulated and VIP-stimulated secretion with the same potency (EDSO = 0.3 ng/ml). These results indicate that pertussis toxin acts at a common site to prevent somatostatin inhibition of both Ca2+- and CAMP-stimulated hormone release. Thus, Ni appears to be required for somatostatin to decrease both CAMP production and [Ca2+] and to inhibit the actions of secretagogues using either of these intracellular messengers.

* This work was supported by National Institutes of Health Re- search Grant AM 32234 (to A. s.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “aduertisement” in accord- ance with 18 U.S.C. Section 1734 solely to indicate this fact.

7 Recipient of Institutional National Research Service Award 5T32GM07258 and a grant from the Harvard Medical Science Fund.

11 Recipient of National Research Service Awards AM 07082 (in- dividual) and AM 07337 (institutional). Present address: Science, Engineering, and Diplomacy Fellow, American Association for the Advancement of Science, 1776 Massachusetts Avenue, Washington, D. C. 20523.

The hypothalamic peptide somatostatin is a physiological regulator of growth hormone release and also inhibits the secretion of many other hormones and neurotransmitters by specific target cells of the pancreas, gut, and nervous system (1). The mechanisms by which somatostatin elicits these inhibitory effects are still poorly understood. The GH4C1 clonal rat pituitary cell strain has been a particularly useful model system for studying the biochemical events involved in somatostatin inhibition of hormone release (2, 3). In these cells, we have shown that somatostatin inhibits growth hormone and prolactin secretion by at least two different mechanisms: one mediated by changes in cAMP levels and the other independent of changes in cAMP concentrations (4,5). Two lines of evidence indicate that the CAMP-mediated effects of somatostatin result from the interaction of the somatostatin receptor with membrane adenylate cyclase. First, somatostatin inhibits VIP1-stimulated adenylate cyclase activity in GH,Cl cell membranes in a manner which parallels its effect on cAMP levels in cells (6). Second, the phospho- diesterase inhibitor IBMX does not reduce somatostatin inhibition of VIP-stimulated cAMP production in intact cells (4).

Less is known about the mechanism by which somatostatin inhibits hormone secretion independently of changes in intracellular cAMP levels (5). However, the recent observation that somatostatin reduces [Ca2+]i in GH cells (7) suggests that changes in cellular Ca2+ metabolism may be involved.

[1251-Tyr1]~~mato~tatin binding to pituitary membranes is modulated by guanyl nucleotides (6, 8), indicating that the somatostatin receptor is likely to regulate adenylate cyclase via a guanyl nucleotide-binding protein. Such a regulatory component (Ni) has recently been identified and shown to couple the enzyme to other inhibitory receptors (9, 10). A 41,000-Da subunit of this protein is ADP-ribosylated by pertussis toxin, a protein produced by Bordetellupertussis (9,lO). Following pertussis toxin treatment of cells, receptor-mediated inhibition of adenylate cyclase is blocked and the specific

The abbreviations used are: VIP, vasoactive intestinal peptide; 8- Br-CAMP, 8-bromoadenosine 3’:5’-monophosphate, [Ca2+]i, the intracellular free calcium ion concentration; F-lOlh, Ham’s F-10 medium supplemented with 5 mg/ml lactalbumin hydrolysate; Gpp(NH)p, guanyl-5’-yl imidodiphosphate; HBS, 4-(2-hydroxy- ethyl)-lrpiperazineethanesulfonic acid (Hepes)-buffered salt solution containing 20 mM Hepes, 118 mM NaC1, 4.6 mM KCI, 1 mM CaC12, 10 mM D-Glucose (pH 7.2); HBS/BSA, HBS supplemented with 0.1% bovine serum albumin; IBMX, 3-isobutyl-1-methylxanthine; Ni, the inhibitory guanyl nucleotide-binding regulatory subunit of adenylate cyclase; PtdIns-4,5-Pz, 1-(3-sn-phosphatidyl)-~-myoinositol 4,5-bis- phosphate; TRH, thyrotropin-releasing hormone; Me-TRH, [3- methyl-histidine]thyrotropin-releasing hormone; EGTA, ethylene glycol his(@-aminoethyl ether)-N,N,N’,N’-tetraacetic acid.

13138

Role of Ni i n Somatostatin Action 13139

binding of inhibitory agonists to membrane preparations is decreased (11-14). Both of these actions of pertussis toxin are believed to result from the inability of ADP-ribosylated Ni to properly couple inhibitory receptors to adenylate cyclase.

Since pertussis toxin treatment of GH cell membranes results in the ADP-ribosylation of a single 41,000-Da protein (E), we have used this toxin to address the role of Ni in somatostatin action. First, we examined whether Ni was the guanyl nucleotide-binding protein which coupled the somatostatin receptor to adenylate cyclase. Second, we investigated the role of Ni in the CAMP-independent actions of somatostatin, including its effects on [Ca2+]i. Our results indicate that both the CAMP-dependent and CAMP-independent actions of somatostatin require a functional Ni. Preliminary reports of some of these data has been presented?

EXPERIMENTAL PROCEDURES

Materials Pertussis toxin was the kind gift of Dr. Erik Hewlett, University

of Virginia, or was purchased from List Biologicals (Campbell, CA). VIP, somatostatin, and bombesin were obtained from Bachem (Torr- ance, CA). TRH was obtained from Beckman Instruments. [L-histidyl-4-3H,~-prolyl-3,4,-3H]Me-TRH was purchased from New England Nuclear. 8-Br-cAMP, Quin-2 acetoxymethyl ester, adenosine deaminase (Type VII), and IBMX were obtained from Sigma. Reagents for the prolactin radioimmunoassay were provided by the National Hor- mone and Pituitary Program of the National Institute of Arthritis, Diabetes, and Digestive and Kidney Diseases, courtesy of Dr. S. Raiti.

Methods Cell Culture-GH4C1 cells were grown in Ham's F-10 medium

supplemented with 15% horse serum and 2.5% fetal calf serum as previously described (5, 16). Medium was changed every 2-3 days. For Quin-2 experiments, cells were plated at a density of 3 X lo5 cells/ml and used after three medium changes. For all other experiments, cells were plated at 0.75-1 X lo5 cells/ml and used after five medium changes.

Cells were pretreated with the indicated dose of pertussis toxin in Ham's F-10 medium supplemented with 15% horse serum and 2.5% fetal calf serum containing 100 units/ml penicillin and 100 pg/ml streptomycin (GIBCO, Grand Island, NY) for 2 h at 37 "C, diluted with an equal volume of culture medium, and incubated overnight (-22 h).

Measurement of Prolactin Release-The effect of various com- pounds on the release of presynthesized, stored prolactin (17) was determined as previously described (5). In brief, cells were preincubated for two 30-min periods at 37 "C in release medium and then fresh medium containing the appropriate drugs or hormones and 0.25 units/ml adenosine deaminase was added. Adenosine deaminase was included to inactivate adenosine produced by GH& cells, since this nucleoside inhibits hormone ~ecret ion.~ After a 30-min incubation at 37 "C, the medium was collected, centrifuged to remove any floating cells (1000 x g for 10 min at 4 "C), and assayed for prolactin content by radioimmunoassay.

Two different media were used in the prolactin release experiments, as specified in each figure legend. The initial experiments were done using F-lOlh, as previously described (5). In this medium, somatostatin inhibits basal and VIP-stimulated hormone secretion by -50%. Since the intrinsic fluorescence of F-lOlh interfered with the measurement of [Ca'+Ii using Quin-2, later experiments were done in HBS/BSA. In HBS/BSA, basal hormone secretion is lower than in F-lOlh and is only slightly inhibited by somatostatin. However, the response to secretagogues was larger in HBS/BSA, and somatostatin inhibited VIP-stimulated prolactin secretion to a similar extent (-50%) as in F-lOlh. The cells were incubated in a 5% carbon dioxide

'B. D. Koch, L. J. Dorflinger, E. Hewlett, and A. Schonbrunn (1984) ZVth International Congress on Prolactin, Charlottesville, VA, Abstr. 67; B. D. Koch and A. Schonbrunn (1985) Proceedings of the 67th Meeting of the Endocrine Society, Baltimore, MD, June 19-21,

3Dorflinger, L. J., and Schonbrunn, A. (1985) Endocrinology, in p. 228.

press.

incubator with F-lOlh and in ambient atmosphere with HBS/BSA. When K+ was used as a secretagogue, the indicated concentration is the final, not the added, K+ concentration. Since in Quin-2 experiments the sodium ion concentration could not be decreased to com- pensate for the added K+, this adjustment was also omitted in most hormone secretion and cAMP experiments. However, control experiments have shown that the additional sodium has no effect on secretory responses.

scribed (5). Rat prolactin (NIADDKD-rPRL-1-5) was radioiodinated Prolactin radioimmunoassay was performed as previously de-

by the chloramine-T method (18). The prolactin standard was NIADDKD-rPRL-RP-2. Antibody-bound antigen was precipitated by the addition of inactivated Staphylococcus aureus (IgGsorb, The Enzyme Center, Malden, MA) (19).

Measurement of Cyclic AMP Levels-The effect of hormones on extracellular cAMP accumulation was determined as previously described (4). Briefly, after two 30-min incubations in F-lOlh, the first without and the second with 100 pM IBMX, cells were incubated in fresh medium containing 100 p~ IBMX and the indicated drugs and hormones. After 20 min, the cells were centrifuged (1000 X g for 10 min at 4 "C) and the cAMP concentration in the medium was measured by radioimmunoassay (20). Under the conditions of these experiments, changes in extracellular cAMP concentrations parallel effects on intracellular cAMP levels (4).

The effect of drugs on intracellular cAMP concentrations was determined in HBS/BSA without IBMX. Cells were preincubated for two periods of 30 min in HBS/BSA, and then fresh HBS/BSA containing chemicals or hormones was added. At the indicated time, the medium was aspirated and the cells were extracted with 1 ml of ice-cold 95% ethanol, 0.1% trifluoroacetic acid. The extracts were dried in a Speed-Vac concentrator (Savant Instruments, Inc., Hicks- ville, NY), resuspended in 50 mM acetate buffer (pH 6.2), and centrifuged to remove precipitated protein (3000 X g for 10 min at 4 "C). The cAMP content was measured by radioimmunoassay (20).

Measurement of /Ca'+]i-Changes in [Ca'+]i were monitored by measuring Quin-2 fluorescence, using a slight modification of published procedures (21). In brief, cells were removed from 100-mm plates by a short exposure to HBS lacking Ca'+ and containing 0.02% EDTA, washed, and incubated with 100 p~ Quin-2 acetoxymethyl ester in HBS for 30 min at 37 "C. Cells were then washed twice with HBS and resuspended in HBS/BSA at a concentration of 2-4 X lo6 cells/ml, and their fluorescence (excitation 339 nm, emission 492 nm) was monitored in a well stirred cuvette with a Perkin-Elmer 650-10s fluorescence spectrophotometer. Hormones were added to the cuvette from -200 X stock solutions. Tracings were digitized on a bitpad, normalized by converting the fluorescence readings to per cent satu- ration of the Quin-2, and plotted on a laser printer.

Measurement of Adenylate Cyclase Actiuity-GH4Cl cell membranes were prepared as previously described (6), except that the buffers contained the following protease inhibitors: 0.25 mM EGTA, 3 mM benzamidine, 5 pg/ml leupeptin, 7 pg/ml pepstatin, 5 pg/ml soybean trypsin inhibitor, and 5 pg/ml lima bean trypsin inhibitor. Adenylate cyclase activity was determined by the method of Salomon et al. (Ref. 22; as modified in Ref. 6).

Ligand Binding-[1251-Tyr1]Somatostatin (2200 Ci/mmol) binding to GH&, cells was determined during a 60-min incubation at 37 'C in F-lOlh as previously described (23). Nonsaturable binding was determined in the presence of 100 nM unlabeled somatostatin and was subtracted from total binding.

['251-Tyr1]S~mato~tatin binding to GHICl membranes was measured as previously described (6). Membranes were incubated with 0.1 pCi/ml [1251-Tyr1]~~mato~tatin for 90 min at 37 "C in 50 mM Tris, 2 mM EDTA, 7 mM MgClp, 1% bovine serum albumin, 75 units/ml bacitracin (pH 7.6). Following the incubation, the membranes were diluted and centrifuged at 4 "C for 6 min at 18,000 rpm, and the membrane-associated radioactivity was measured. Nonsaturable binding was determined in the presence of 30 nM somatostatin and subtracted from total binding. [~-histidyl-4-~H,L-prolyl-3,4-~H]Me-TRH (77 Ci/mmol) binding to

membranes was determined by same method, using 0.16 pCi/ml [3H] Me-TRH as the radioligand. Nonsaturable binding was measured in the presence of 2.5 p~ TRH.

RESULTS

Effect of Pertussis Toxin on Somatostatin Receptor Function in Membranes-Two lines of evidence have indicated that the

13140 Role of Ni in Somatostatin Action

somatostatin receptor is coupled to adenylate cyclase in GH4C1 cell membranes. First, somatostatin inhibits stimulated adenylate cyclase activity (6). Second, guanyl nucleotides decrease the binding affinity of somatostatin receptors in membrane preparations (6). Therefore, we examined the effect of pertussis toxin treatment on these two indicators of somatostatin receptor function.

In membranes prepared from control cells, VIP stimulated adenylate cyclase activity 3.6-fold and somatostatin inhibited this stimulation by 15%-(Table I). In membranes from pertussis toxin-treated cells, VIP caused a similar stimulation of adenylate cyclase activity (4.5-fold), but somatostatin only inhibited this stimulation by 3%. Thus, somatostatin inhibition of stimulated adenylate cyclase activity was blocked by pertussis toxin treatment.

The data in Fig. 1 show the effects of pertussis toxin treatment on the membrane binding of [1251-Tyr1]somatostatin and [3H]Me-TRH. Although TRH has been shown to stimulate PtdIns-4,5-P2 hydrolysis (24-26), its receptor also appears to be coupled to a guanyl nucleotide-binding protein (27). In control membranes, Gpp(NH)p, a nonhydrolyzable analog of GTP, reduced [1251-Tyr1]~~mato~tatin binding by 78% (Fig. 1, left panel) and [3H]Me-TRH binding by 63%

TABLE I Effect of pertussis toxin treatment on somatostatin inhibition of VIP-

stimulated adenylate cyclase Cells were pretreated with or without 75 ng/ml pertussis toxin,

membranes were prepared, and adenylate cyclase activity was determined in the absence of hormones or in the presence of VIP (30 nM), somatostatin (SRIF) (100 nM), or both peptides. Per cent inhibition was calculated by the following formula: (VIP - VIP and SRIF)/ (VIP - control) X 100.

Peptide Adenylate cyclase activitf

Control Toxin-treated

nmollminlmg Control 0.51 f 0.013 (3) 0.33 f 0.010 (4) Somatostatin 0.48 f 0.012 (4) 0.30 f 0.007 (4) VIP 1.85 f 0.021 (4) 1.47 f 0.046 (4) VIP + somatostatin 1.65 +. 0.033 (4) 1.44 f 0.040 (3) % inhibition 15 f 3 3 + 5

a Mean + S.E.; the number of observations is shown in parentheses.

Control Toxin Rx FIG. 1. Effect of pertussis toxin treatment on the binding of

[1261-Tyr1]somatostatin and ['HIMe-TRH to membranes. GH,C1 cells were preincubated with (Toxin Rx) or without (Control) 50 ng/ml pertussis toxin. Saturable binding of [1251-Tyr1]somatostatin (SRIF) ( le f t pawl) or [3H]Me-TRH (right panel) was determined in membranes prepared from the pretreated cells. The binding reaction was carried out either in the absence (0) or presence (8) of 100 p M Gpp(NH)p. Each bar represents the mean f S.E. ( n = 3) of the saturable binding.

(Fig. 1, right panel). These observations confirm published results (6, 27). In membranes-prepared from pertussis toxin- treated cells, [lZ5I-Tyr1 ]somatostatin binding in the absence of Gpp(NH)p was reduced by more than 95%, whereas [3H] Me-TRH binding was unaffected. As in control membranes, Gpp(NH)p reduced [3H]Me-TRH binding by 65% in pertussis toxin-treated membranes. In contrast, this guanyl nucleotide had no detectable effect on the low level of saturable [lZ5I- Tyrl]somatostatin binding measured in pertussis toxin- treated membranes.

These results indicate that pertussis toxin treatment spe- cifically blocked the interaction of the somatostatin receptor with a guanyl nucleotide-binding protein, thereby disrupting the functional coupling of this receptor to adenylate cyclase. In the same membrane preparations, the apparent interaction of the TRH receptor with a guanyl nucleotide-binding protein was unaffected.

Effect of Pertussis Toxin on the CAMP-mediated Actions of Somatostatin in Cells-Since pertussis toxin treatment blocked somatostatin inhibition of adenylate cyclase, we examined the toxin's effects on those cellular actions of somatostatin which appear to result from this inhibition (3-6). In control cells, VIP stimulated cAMP accumulation 11-fold, and somatostatin inhibited the VIP stimulation by 52% (Fig. 2, lower panel). In pertussis toxin-treated cells, VIP stimulation of cAMP accumulation was unchanged, but somatostatin no longer inhibited this stimulation. The effect of pertussis toxin treatment on prolactin release paralleled the effects on CAMP accumulation (Fig. 2, upper panel): toxin pretreatment did not alter VIP stimulation, but blocked somatostatin inhibition.

Control Toxin Rx FIG. 2. Effect of pertussis toxin treatment on somatostatin

inhibition of VIP-stimulated prolactin release and cAMP accumulation. Cells were pretreated with (Toxin Rx) or without (Con- trol) 70 ng/ml pertussis toxin. The effect of peptides on prolactin release (upper panel) or extracellular cAMP accumulation (lower panel) was then determined in F-lOlh. Each bar represents the mean f S.E. ( n = 6). 0, control; 0, 100 nM VIP; 100 nM VIP plus 100 nM somatostatin.

Role of Ni in Somatostatin Action

In intact cells, pertussis toxin blockage of somatostatin action was not simply due to inhibition of somatostatin binding. The saturable binding of [1251-Tyr1]somatostatin to GH4C1 cells was unaffected by pertussis toxin treatment, whereas somatostatin inhibition of hormone secretion was abolished (Table 11). The apparently different effects of pertussis toxin treatment on ['251-Tyr']somatostatin binding to cells and membranes (Fig. 1 and Table 11) probably result from the different forms of the receptor measured under the two conditions. In membranes incubated in the absence of guanyl nucleotides, ['251-Tyr']somatostatin binding is primar- ily a measure of the high affinity state of the somatostatin receptor which exists in the absence of GTP (6). In contrast, [1251-Tyr1]~~mato~tatin binding to intact cells probably reflects binding to the low affinity form of the somatostatin receptor induced by GTP.

In summary, pertussis toxin has parallel effects on somatostatin inhibition of VIP-stimulated adenylate cyclase activity, cAMP accumulation, and prolactin secretion. This indicates that somatostatin inhibition of stimulated cAMP levels and prolactin secretion results from its inhibition of adenylate cyclase via Ni.

Effect of Pertussis Toxin on CAMP-independent Actions of Somatostatin-Somatostatin inhibition of basal hormone secretion in F-lOlh occurs without a concomitant reduction of basal cAMP concentrations (5). Furthermore, this inhibition is not blunted when cellular cAMP concentrations are phar- macologically raised to levels many times those required to maximally stimulate hormone release (5). To determine whether Ni might also be involved in this CAMP-independent action of somatostatin, we first examined the effect of pertussis toxin treatment on somatostatin inhibition of basal hormone secretion in F-lOlh. The results in Table I1 demonstrate that somatostatin inhibition of basal hormone secretion was totally blocked in pertussis toxin-treated cells.

To further characterize the effect of pertussis toxin treatment on the CAMP-independent actions of somatostatin, we used four secretagogues which do not regulate adenylate cyclase activity (Fig. 3). 8-Br-CAMP directly stimulates CAMP- dependent protein kinases and therefore elicits CAMP-in- ducible effects a t a site distal to the activation of adenylate cyclase. High K+ concentrations are believed to stimulate hormone release by depolarizing the plasma membrane, open- ing voltage-sensitive Ca'+ channels, and increasing [Ca2+Ii (28). TRH and bombesin do not reproducibly affect cAMP levels in GH4Cl cells and have been proposed to stimulate hormone secretion by increasing PtdIns-4,5-Pz hydrolysis and subsequently [Ca2+Ii (4,21,24-26). Pertussis toxin treatment

TABLE I1 Effect of pertussis toxin treatment on j'Z5Z-Tyr']somatostatin binding

Cells were pretreated with or without 75 ng/ml pertussis toxin. The effect of 100 nM somatostatin on prolactin release was subsequently determined during a 30-min incubation in F-lOlh as described under "Methods." After the release incubation, the same cells were incubated for 60 min at 37 "C in fresh F-lOlh containing ['251-Tyr'] somatostatin (115,330 cpm/ml) with or without 100 nM somatostatin to measure both total and nonsaturable '251-Tyr'-somatostatin binding.

Pretreatment Peptide added Prolactin ['"I-Tyrl]Somatostatin released" bound"

nglml cpmjdish Control Control 128 -C 2.7 10,750 f 150

Pertussis toxin Control 119 -t 2.0 11,730 f 610 Somatostatin 98 f 2.7 1,670 f 90

Somatostatin 114 f 3.3 1,720 f 90 @Mean k S.E. ( n = 3).

Control V I P BBr-CAMP K'

13141

7

Control TRH Bombesin

FIG. 3. Effect of pertussis toxin treatment on somatostatin inhibition of secretagogue-stimulated prolactin release. Cells were pretreated without (upper panels) or with (lower panels) 50 ng/ ml pertussis toxin. Prolactin release was then determined in HBS/ BSA, either in the absence (a) or presence (Ed) of 100 nM somatostatin and the secretagogues shown. The concentrations of secretagogues used were: VIP, 100 nM; g-Br-cAMP, 5 mM; K+, 11 mM; TRH, 1 pM; bombesin, 100 nM. Each bar represents the mean f S.E. ( n = 6).

did not affect the degree to which 8-Br-cAMP, K+, TRH, or bombesin stimulated prolactin release (Fig. 3), although the stimulatory effect of VIP was slightly potentiated in some experiments. The lack of inhibition of TRH action is consistent with our observation (Fig. 1) that the guanyl nucleotide- sensitive binding of [3H]Me-TRH to membrane preparations was unaffected by pertussis toxin treatment. In control cells, somatostatin inhibited both basal hormone secretion and the stimulatory effects of all secretagogues (Fig. 3, top panels). Pertussis toxin treatment prevented these actions of somatostatin (Fig. 3, bottom panels). Thus, pertussis toxin blocks somatostatin inhibition of hormone secretion stimulated by secretagogues which do not affect adenylate cyclase, suggest- ing that the CAMP-independent mechanisms by which somatostatin inhibits hormone secretion may also require Ni.

Effect of K+ and Somatostatin on Intracellular CAMP Leu- els-Since the mechanisms by which TRH and bombesin stimulate secretion are complex, we chose to use 11 mM K+ as a secretagogue with which to further investigate the CAMP- independent mechanism of somatostatin action. As the results in Fig. 4 demonstrate, high concentrations of K+ (50 mM) did not affect intracellular cAMP levels at any time from 30 s to 30 min. In the same experiment, VIP produced a rapid rise in intracellular CAMP, which then dropped to a sustained elevated plateau level. Somatostatin, either alone or in combi- nation with 50 mM K+, had a negligible effect on intracellular cAMP levels (-8% decrease; data not shown), while it reduced the VIP-stimulated increase by approximately 50% at all time points (Fig. 4). When extracellular cAMP accumulation was measured in the presence of 100 p~ IBMX, 11 mM K+ was also without effect (data not shown). In accord with these results, K+ depolarization produces a different pattern of protein phosphorylations in GH cells than do agents which elevate intracellular cAMP levels (29). Thus, both the stimulatory action of K+ and its inhibition by somatostatin seem to be independent of changes in intracellular cAMP concentrations.

Effect of Pertussis Toxin on Somatostatin-induced Decreases in [Ca2+Ji-Although the biochemical basis of the CAMP- independent mechanism of somatostatin action is unknown, several studies have suggested that regulation of [Ca2+Ji might


be involved (7, 30, 31). We therefore used the fluorescent chelating dye Quin-2 to determine the effect of somatostatin on basal and K+-elevated [Ca2+]; with and without pertussis toxin treatment (Fig. 5). In control GH& cells, somatostatin caused a transient drop in resting [Ca2+]i (Fig. 5A). [Ca2+Ii reached a minimum approximately 30 s after the addition of somatostatin, and then slowly returned to the original resting level. Pretreatment of the cells with pertussis toxin resulted in an almost complete block of the somatostatin reduction in [Ca2+Ii (Fig. 5B). Table I11 summarizes the results of a number of such experiments. In control cells, somatostatin reduced [Ca2+Ii below resting level, whereas in pertussis toxin-treated cells. somatostatin no longer had a significant effect on basal

I I

10 20 30

Minutes FIG. 4. Effects of VIP and K+ on intracellular cAMP levels.

Cells were incubated at 37 "C in HBS/BSA (0) or HBSPSA containing 50 mM K+ (.), 100 nM VIP (M), or 100 nM VIP plus 100 nM somatostatin (A) for the times shown. Intracellular cAMP levels were measured at the indicated times after addition of medium. Eachpoint represents the mean k S.E. (n = 3).

I I

400 - A 600

400

200 - r 600

2 400

.I - 1 1 ' ' '

"0 400 - B

300 -

1 SRIF t

200 I 2oo

We next explored the effect of somatostatin on K+-stimulated [Ca2+Ii (Fig. 5, C and D). Addition of 11 mM K+ caused a very rapid rise in [Caz+]i, which decayed to a nadir before rising to a relatively sustained plateau level. Addition of 100 nM somatostatin during the plateau phase caused a transient drop in [Ca2+Ii in control cells (Fig. 5C). However, in pertussis toxin-treated cells, somatostatin had no significant effect on K+-stimulated [Ca2+Ii (Fig. 5D and Table 11). The effect of somatostatin on [Ca2+Ii is likely to be receptor-mediated, since 100 nM dicarboxymethyl somatostatin caused a much smaller decrease in [Ca2+Ii than the same concentration of somatostatin (Fig. 5, B and C). Dicarboxymethyl somatostatin, like other reduced analogs (32), is at least 100-fold less potent than somatostatin at inhibiting hormone secretion and adenylate cyclase activity (data not shown). Thus, somatostatin reduces both basal and potassium-elevated [Ca2+]i, and these effects are blocked by pertussis toxin treatment. Therefore somatostatin's effects on [Ca2+Ii appear to be mediated via Ni.

Concentration Dependence for Pertussis Toxin Inhibition of Somatostatin Action-The results in Fig. 3 show that pertussis toxin treatment prevented somatostatin inhibition of both VIP-stimulated and K+-stimulated hormone secretion. Since these secretagogues are believed to stimulate prolactin release

TABLE I11 Effect of pertussis toxin treatment on somatostatin-induced reduction

Cells were pretreated with or without a maximal concentration (10 or 100 ng/ml) of pertussis toxin. The effect of 100 nM somatostatin on basal or K+-stimulated [Ca"]; was determined by monitoring the fluorescence of Quin-8-loaded cells in experiments analogous to those shown in Fig, 5. In each experiment, the per cent inhibition produced by somatostatin was calculated as the change in [Ca2+]i divided by the [Ca2+Ii before adding somatostatin.

of [Ca2+1;

Decrease in [Ca"]; by somatostatin" Pretreatment

Basal 11 mM K+ %

Control 20.7 f 2.60 (3) 31.0 f 2.38 (4) Pertussis toxin 3.7 +. 2.03 (3) 1.0 f 1.00 (4)

Mean & S.E.; the number of observations is shown inparentheses.

Minutes Minutes

FIG. 5. Effect of pertussis toxin treatment on somatostatin regulation of [Caa+]i. Cells were preincubated with either medium alone (top panels) or with pertussis toxin (lower panels: 10 ng/ml in left panel; 100 ng/ml in right panel). Changes in. [Ca2+Ii were determined in suspended cells by monitoring the fluorescence of intracellular Quin-2. Arrows mark the addition of ,100 nM somatostatin (SRIF), 100 nM dicarboxymethyl somatostatin (CMSRIF), or vehicle. Neither 11 mM K+ nor somatostatin had any effect on the intrinsic fluorescence of GH& cells (data not shown). The left two panels show the effects of somatostatin on basal [Ca"Ii and the right two panels show the effects of somatostatin on [Ca*+]i stimulated by 11 mM K+.

Role of Ni in Somatostatin Action 13143

FIG. 6. Concentration dependence for pertussis toxin inhibition of somatostatin action. Cells were pretreated with the indicated concentrations of pertussis toxin for 24 h without dilution. Prolactin release was then determined in HBS/BSA in the absence of any agents (0) or in the presence of either 100 nM VIP (left panel, W and 0) or 11 mM K+ (right panel, A and A). Each secretagogue was tested both in the absence (solid symbols) and presence (open syrnbok) of 100 nM somatostatin. Each point represents the mean f S.E. (n = 6).

40 I 1

1 I I 8 0.1 1 .o IO 4 1

by a CAMP-mediated and a Ca2+-mediated mechanism, re- spectively, this result suggested that Ni was involved in both the CAMP-dependent and the CAMP-independent actions of somatostatin. Although pertussis toxin has been shown to ADP-ribosylate only a single 41,000-Da protein in GH cell membranes (15), it remained possible that the toxin had multiple targets in intact cells. To address this possibility, we determined the concentration dependence for pertussis toxin blockage of somatostatin inhibition of both VIP-stimulated (Fig. 6, left panel) and K+-stimulated (Fig. 6, right panel) prolactin secretion. Both effects of somatostatin were completely blocked by pertussis toxin with the same ED5,, (0.3 ng/ ml), indicating that the toxin acts at a single, common site to prevent somatostatin inhibition of VIP-stimulated and K+- stimulated prolactin secretion,

DISCUSSION

We have previously shown that somatostatin inhibits hormone secretion from GH4Cl cells by at least two mechanisms: inhibition of stimulated adenylate cyclase and a second CAMP-independent process (3-6). The results presented here demonstrate that pertussis toxin blocks somatostatin inhibition of VIP-stimulated adenylate cyclase activity, cAMP accumulation, and prolactin secretion. Since pertussis toxin ADP-ribosylates only a 41,000-Da polypeptide, presumably Ni, in GH cell membranes ( E ) , these results indicate that Ni mediates somatostatin inhibition of VIP-stimulated prolactin secretion by coupling the somatostatin receptor to adenylate cyclase. However, pertussis toxin treatment also blocked the CAMP-independent inhibition of hormone secretion by somatostatin. Furthermore, pertussis toxin prevented the reduction of basal and K+-elevated [Caz+]i by somatostatin, effects which were also produced without any changes in intracellular cAMP levels. These results indicate that Ni not only mediates somatostatin inhibition of adenylate cyclase, but also the CAMP-independent components of somatostatin action, including the reduction in [Ca2+]i. To our knowledge, this is the first evidence for the coupling of Ni to inhibition of [Ca2+]i.

In parallel with our results in GH4Cl pituitary cells, pertussis toxin treatment has been shown to reduce somatostatin inhibition of stimulated adenylate cyclase, cAMP accumulation, or hormone secretion in other systems. In S49 lymphoma cell membranes, pertussis toxin treatment blocks somatostatin inhibition of forskolin-stimulated adenylate cyclase activity (13). Pertussis toxin treatment also blocks somato-

statin inhibition of stimulated CAMP accumulation and hormone secretion from several types of endocrine cells: pan- creatic &cells (33), pituitary somatotrophs (34), and AtT2O/ Dl6 pituitary tumor cells (35). The general observation that pertussis toxin blocks somatostatin inhibition of hormone secretion, in addition to stimulated adenylate cyclase activity and cAMP accumulation, has been taken as evidence that somatostatin elicits its inhibitory effects on hormone secretion by regulating adenylate cyclase.

Two lines of evidence support the hypothesis that somatostatin can inhibit hormone secretion from GH4Cl cells by a CAMP-independent mechanism. First, somatostatin can inhibit hormone secretion under a variety of conditions in which there are no changes in intracellular cAMP concentrations. These include inhibition of basal hormone secretion (4, 5 ) , inhibition of the stimulatory effect of 11 mM K+, and inhibition of the stimulatory effects of TRH and bombesin (36). Second, somatostatin can inhibit secretion stimulated by 8- Br-CAMP, which raises intracellular CAMP levels without affecting adenylate cyclase, and by forskolin, which produces supramaximal intracellular cAMP levels even in the presence of somatostatin (5). The observation that pertussis toxin blocked a CAMP-independent effect of somatostatin with the same potency as it blocked its CAMP-mediated action suggested that both mechanisms required the same pertussis toxin-sensitive site, namely Ni. Thus, our results provide a new function for Ni in addition to mediating inhibition of adenylate cyclase.

Somatostatin has been shown to decrease basal [Ca2+Ii in GH cells (7), and this reduction in [Ca2+]i provides a possible mechanism for its CAMP-independent action. Consistent with this possibility, we have observed that somatostatin decreases [Ca2+Ii in the presence of a maximal concentration of the cAMP analog 8-(4-chlorophenylthio)-cAMP.4 The result that pertussis toxin prevented both somatostatin reduction of [CaZ+li and its inhibition of K+-stimulated secretion suggests that Ni may activate the CAMP-independent mechanism of somatostatin action by triggering decreases in [Ca2+Ii. The biochemical mechanism by which Ni may reduce [Ca2+Ii clearly provides an interesting problem for future research.

Examination of guanyl nucleotide effects on ['251-Tyr1] somatostatin binding to GH4Cl cell membrane preparations (6) indicated that the somatostatin receptor interacts with a guanyl nucleotide-binding protein by a process similar to that

* B. D. Koch and A. Schonbrunn, unpublished observations.


originally proposed for the @-adrenergic receptor (37). Namely, guanyl nucleotides appear to destabilize a high affinity ternary complex of the somatostatin receptor, bound hormone, and a guanyl nucleotide-binding protein, thereby pro- ducing a decrease in [1251-Tyr']s~mato~tatin binding. [lZ5I- Tyr'] Somatostatin binding to membrane preparations pri- marily reflects binding to this high affinity ternary complex (6); thus, the observed decrease in membrane binding and loss of guanyl nucleotide sensitivity upon pertussis toxin treatment are consistent with inactivation of Ni. However, it is not clear why pertussis toxin treatment reduced ['"I-Tyr'] somatostatin binding to membrane preparations more than did maximal concentrations of Gpp(NH)p. Perhaps ADP- ribosylation completely prevents Ni interaction with the somatostatin receptor, whereas the binding of Gpp(NH)p to Ni only reduces its affinity for the receptor. Unfortunately, the binding of [1251-Tyr']somatostatin to pertussis toxin-treated membranes was too low to permit experimental examination of this hypothesis. Nevertheless, the ability of pertussis toxin to reduce [1251-Tyr*]~~mato~tatin binding to GH4Cl cell membranes suggests that the somatostatin receptor interacts with Ni and that this interaction is disrupted by pertussis toxin treatment.

Unexpectedly, we found that although ['251-Tyr']somatostatin binding to membrane preparations was reduced by pertussis toxin treatment, ['251-Tyr1]~~mato~tatin binding to cells was unaffected. The presence of GTP in cells, but its absence in the membrane incubations, may be responsible for this difference. In contrast to membrane binding, ['z51-Tyr'] somatostatin binding to intact cells is likely to reflect binding to the low affinity form of the somatostatin receptor induced by GTP. Therefore, the different effects of pertussis toxin on [1251-Tyr1]somatostatin binding to cells and membranes may be due to the different forms of the receptor measured in the two assays.

The action of pertussis toxin to ADP-ribosylate a 41,000- Da polypeptide component of the inhibitory guanyl nucleotide-binding subunit of adenylate cyclase, and thereby reduce hormonal inhibition of the enzyme, has been demonstrated in many target cells (11-14). Furthermore, pertussis toxin treatment has been shown to decrease the affinity of specific membrane receptors for a number of neurotransmitters which cause inhibition of adenylate cyclase (14). Thus, pertussis toxin treatment appears to generally disrupt both Ni-dependent inhibition of adenylate cyclase activity and the interaction of inhibitory receptors with Ni. Recently, pertussis toxin has also been shown to block the action of a stimulatory agent, the chemotactic peptide fMet-Leu-Phe (38-41). This effect of pertussis toxin also correlates with ADP-ribosylation of a 41,000-Da polypeptide (38-40). Since Wet-Leu-Phe has been shown to increase PtdIns-4,5-P2 hydrolysis and [CaZ+], in neutrophils (41,42), the ability of pertussis toxin to block its actions indicates that Ni may also be involved in adenylate cyclase-independent stimulatory mechanisms. Our results indicate that Ni does not play such a role in GH4Cl cells: pertussis toxin treatment did not affect the guanyl nucleotide- sensitive binding of TRH to GH4C1 cell membranes, nor did it reduce TRH stimulation of hormone secretion. Since TRH also stimulates PtdIns-4,5-Pz hydrolysis and increases [Ca2+Ii in GH cells (21, 24-26), our results suggest that not all secretagogues which utilize these intracellular messengers act via receptors which are coupled to a pertussis toxin substrate.

We have previously demonstrated that a series of somatostatin analogs show ihe same potency for inhibition of VIP- stimulated CAMP accumulation and inhibition of basal prolactin secretion in GH4C1 cells (43). These results indicated

that the CAMP-mediated and CAMP-independent actions of somatostatin were triggered by the same receptor. The results presented here provide evidence that both somatostatin inhibition of adenylate cyclase and its CAMP-independent inhibitory actions require a pertussis toxin substrate, Ni. Fur- thermore, the ability of pertussis toxin to block somatostatin reduction of [Ca2+]; suggests a possible biochemical mechanism for somatostatin's CAMP-independent actions and dem- onstrates a new function for Ni.

Acknowledgments-We thank Dr. Erik Hewlett for the gift of pertussis toxin and for helpful comments on this manuscript, Dr. William Toscano, Jr. and Diane Toscano for help with the adenylate cyclase assays, Dr. Paul Albert for advice on the [Ca2+]i measure- ments, Dr. David Presky for his determination of [1251-Tyr1]somatostatin binding to GH,Cl cells, and Helen Bikkal for expert technical assistance. We would also like to thank the J o s h Diabetes Center for use of the bitpad.

REFERENCES

1. Reichlin, S. (1983) N. Engl. J. Med. 309, 1495-1501,1556-1563 2. Tashjian, A. H., Jr. (1979) Methods Enzymol. 5 8 , 527-535 3. Schonbrunn, A., Dorflinger, L. J., and Koch, B. D. (1985) So-

matostatin (Pate1 y., and Tannenbaum G., eds) pp. 305-324, Plenum Press, New York

4. Dorflinger, L. J., and Schonbrunn, A. (1983) ErzdocrinoZogy 113, 1541-1550

5. Dorflinger, L. J., and Schonbrunn, A. (1983) Endocrinology 113 , 1551-1558

6. Koch, B. D., and Schonbrunn, A. (1984) Endocrinology 114 ,

7. Schlegel, W., Wuarin, F., Wollheim, C. B., and Zahnd, G. R.

8. Enjalbert, A., Rasolonjanahary, R., Moyse, E., Kordon, C., and

9. Bokoch, G. M., Katada, T., Northup, J. K., Ui, M., and Gilman,

10. Neer, E. J., Lok, J. M., and Wolf, L. G. (1984) J. Biol. Chem.

11. Murayama, T., and Ui, M. (1983) J. Biol. Chem. 258,3319-3326 12. Hildebrandt, J. D., Sekura, R. D., Codina, J., Iyengar, R., Man-

Clark, C. R., and Birnbaumer, L. (1983) Nature 302 , 706-709 13. Aktories, K., Schultz, G., and Jakobs, K. H. (1983) FEBS Lett.

14. Kurose, H., Katada, T., Amano, T., and Ui, M. (1983) J. Biol. Chem. 258,4870-4875

15. Wojcikiewicz, R. J. H., Dobson, P. R. M., Irons, L. I., Robinson, A., and Brown, B. L. (1984) Biochem. J. 2 2 4 , 339-342

16. Tashjian, A. H., Jr., Yasumura, Y., Levine, L., Sato, G. H., and Parker, M. L. (1968) Endocrinology 8 2 , 342-352

17. Dannies, P. S., and Tashjian, A. H., Jr. (1973) J. Biol. Chem. 248,6174-6179

18. Peake, G. T., Morris, J., and Buckman, M. T. (1979) in Methods of Hormone Radioimmunoassay (Jaffe, B. M., and Behrman, H. R., eds) pp. 223-244, Academic Press, New York

19. Goding, J. W. (1978) J. Immunol. Methods 2 0 , 241-253 20. Steiner, A. L. (1979) in Methods of Hormone Radioimmunoassay

(Jaffe, B. M., and Behrman, H. R., eds) pp. 3-17, Academic Press, New York

21. Albert, P. R., and Tashjian, A. H., Jr. (1984) J. Biol. Chern. 2 5 9 , 5827-5832

22. Salomon, Y., Londos, C., and Rodbell, M. (1974) Anal. Biochem.

23. Schonbrunn. A.. and Tashiian, A. H.. Jr. (1978) J. Biol. Chem.

1784-1790

(1984) Cell Calcium 5 , 223-236

Epelbaum, J. (1983) Endocrinology 113,822-824

A. G. (1984) J. Biol. Chem. 259,3560-3567

259,14222-14229

158,169-173

58,541-548

253,6473-6483 " .

24. Sutton, C. A., and Martin, T. F. J. (1982) Endocrinology 110, 1273-1280

25. Martin, T. F. J. (1983) J. Bid. Chem. 2 5 8 , 14816-14822 26. Gershengorn, M. C., Geras, E., Purrello, V. S., and Rebecchi, M.

27. Hinkle. P. M.. and Kinsella. P. A. (1984) J. Biol. Chem. 259, J . (1984) J. Biol. Chem. 259, 10675-10681

344513449 '

. .

28. Tan, K.-N., and Tashjian, A. H., Jr. (1984) J. Biol. Chem. 2 5 9 , 418-426

29. Drust, D. S., and Martin, T. F. J. (1982) J. Biol. Chern. 2 5 7 ,

Role of Ni in Somatostatin Action 13145

7566-7573

1180

J. Physwl. 233, C164-Cl71

Chem. 18,123-126

30. Kraicer, J., and Chow, A. E. H. (1982) Endocrinology 111,1173-

31. Pace, C. S., Murphy, M., Conant, S., and Lacy, P. E. (1977) Am.

32. Rivier, J., Brazeau, P., Vale, W., and Guillemin, R. (1975) J. Med.

33. Katada, T., and Ui, M. (1979) J. Biol. Chem. 254, 469-479 34. Cronin, M. J., Rogol, A. D., Myers, G. A., and Hewlett, E. L.

35. Reisine, T., Zhang, Y.-L., and Sekura, R. (1985) J. Pharmmol.

36. Westendorf, J. M., and Schonbrunn, A. (1982) Endocrinology

(1983) Endocrinology 113, 209-215

Exp. Ther. 232, 275-282

110,352-358

37. DeLean, A., Stadel, J. M., and Lefiowitz, R. J. (1980) J. Biol.

38. Okajima, F., and Ui, M. (1984) J. Biol. Chem. 259, 13863-13871 39. Bokoch, G. M., and Gilman, A. G. (1984) Cell 39,301-308 40. Lad, P. M., Olson, C. V., and Smiley, P. A. (1985) Proc. Nutl.

Acad. Sci. U. S. A. 82, 869-873 41. Molski, T. F. P., Naccache, P. H., Marsh, M. L., Kermode, J.,

Becker, E. L., and Sha’afi, R. I. (1984) Biochem. Biophys. Res. Commun. 124,644-650

42. Volpi, M., Yassin, R., Tao, W., Molski, T. F. P., Naccache, P. H., and Sha’afi, R. I. (1984) Proc. Nutl. Acad. Sci. U. S.A. 81, 5966-5969

43. Schonbrunn, A., Rorstad, 0. P., Westendorf, J. M., and Martin, J. B. (1983) Endocrinology 113, 1559-1567

Chem. 255,7108-7117

Pertussis Toxin Blocks Both Cyclic AMP-mediated and Cyclic AMP ...

Documents

Transcript of Pertussis Toxin Blocks Both Cyclic AMP-mediated and Cyclic AMP ...