Sustained signaling leading to T cell activation results from prolonged T cell receptor occupancy. Role of T cell actin cytoskeleton.

Using antigen-specific T cell clones and peptide-pulsed antigen-presenting cells (APCs) we investigated the mechanisms that lead to sustained signaling, known to be required for activation of effector function. Four lines of evidence indicate that the T cell actin cytoskeleton plays a crucial role in T cell activation by antigen-pulsed APCs, but is not required when T cell receptor (TCR) is cross-linked by soluble antibodies. First, addition of antibodies to the major histocompatibility complex molecules recognized by the TCR aborts the ongoing intracellular calcium concentration ([Ca2+]i) increase in performed T-APC conjugates, indicating that the sustained signaling requires the continuous occupancy of TCR. Second, time-lapse image recording shows that T lymphocytes conjugated to peptide-pulsed APCs undergo a sustained [Ca2+]i increase, which is accompanied by the formation of a large and changing area of contact between the two opposing membranes. Third, drugs that disrupt the actin cytoskeleton, Cytochalasin D and and C2 Clostridium botulinum toxin induce a rapid block of [Ca2+]i rise, coincident with a block of the cyclic changes in T cell shape. Finally, the addition of Cytochalasin D or of anti-MHC antibodies to preformed conjugates inhibits interferon gamma production in an 1-antigen dose- and time-dependent fashion. These results identify T cell actin cytoskeleton as a major motor for sustaining signal transduction and possibly for driving TCR cross-linking and offer an explanation for how T cells equipped with low affinity TCR can be triggered by a small number of complexes on APCs.

Nerve growth factor triggers microfilament assembly and paxillin phosphorylation in human B lymphocytes.

Increasing evidence suggests that the nervous system is involved in allergic inflammation. One of the potential regulatory molecules of the neuroimmune system is nerve growth factor (NGF). Recent studies from our group demonstrated the presence of a functional NGF receptor (NGFR) on human B lymphocytes. Moreover, we showed that gp140trk tyrosine kinase, which serves as an NGFR, was involved in transduction of early signaling events in human B lymphocytes. The mechanisms by which NGF initiates the signaling cascade and the link between the neuroimmune systems are unknown. We have focused on the role of the cytoskeleton as a possible mediator for transduction of signals induced by NGF. Polymerized actin (F-actin) content was determined by fluorescent staining and immunoblotting with antiactin antibody. Addition of NGF caused a time- and concentration-dependent increase in F-actin content, and maximum effects were noted after 1 min. These increases in F-actin content and NGF-induced thymidine incorporation could be blocked by incubating the cells with cytochalasin D and botulinum C2 toxin before the addition of NGF. Incubation of human B lymphocytes with 10 nM K252a, an inhibitor of Trk kinase, decreased NGF-induced microfilament assembly by 75%. In immunoprecipitation experiments, addition of NGF to B cells induced a rapid increase in the tyrosine phosphorylation of paxillin, one of a group of focal adhesion proteins involved in linking actin filaments to the plasma membrane. Coimmunoprecipitation studies demonstrated the association between gp140trk kinase and paxillin. Together, these observations suggest that actin assembly is involved in NGF signaling in human B cells, and that paxillin may be essential in this pathway after phosphorylation by gp140trk kinase.

Monoglucosylation of low-molecular-mass GTP-binding Rho proteins by clostridial cytotoxins.

Rho proteins, which are involved in receptor-mediated regulation of the actin cytoskeleton, are substrates for ADP-ribosylation by Clostridium botulinum C3 toxins. Recently, it was shown that Rho and other members of the Rho subfamily of low-molecular-mass GTP-binding proteins are glucosylated by C. difficile toxins A and B. Glucosylation occurs at threonine-37, which is a crucial amino acid residue for the regulatory functions of the small GTP-binding proteins. These toxins should prove useful as tools for studying the functions of Rho proteins.

Influence of botulinum C2 toxin on F-actin and N-formyl peptide receptor dynamics in human neutrophils.

Stimulation of human neutrophils with the chemotactic N-formyl peptide causes production of oxygen radicals and conversion of monomeric actin (G-actin) to polymeric actin (F-actin). The effects of the binary botulinum C2 toxin on the amount of F-actin and on neutrophil cell responses were studied. Two different methods for analyzing the actin response were used in formyl peptide-stimulated cells: staining of F-actin with rhodamine-phalloidin and a transient right angle light scatter. Preincubation of neutrophils with 400 ng/ml component I and 1,600 ng/ml component II of botulinum C2 toxin for 30 min almost completely inhibited the formyl peptide-stimulated polymerization of G-actin and at the same time decreased the amount of F-actin in unstimulated neutrophils by an average of approximately 30%. Botulinum C2 toxin preincubation for 60 min destroyed approximately 75% of the F-actin in unstimulated neutrophils. Right angle light scatter analysis showed that control neutrophils exhibited the transient response characteristic of actin polymerization; however, after botulinum C2 toxin treatment, degranulation was detected. Single components of the binary botulinum C2 toxin were without effect on the actin polymerization response. Fluorescence flow cytometry and fluorospectrometric binding studies showed little alteration in N-formyl peptide binding or dissociation dynamics in the toxin-treated cells. However, endocytosis of the fluorescent N-formyl peptide ligand-receptor complex was slower but still possible in degranulating neutrophils treated with botulinum C2 toxin for 60 min. The half-time of endocytosis, estimated from initial rates, was 4 and 8 min in control and botulinum C2 toxin-treated neutrophils, respectively.

Ephrin-A5 induces collapse of growth cones by activating Rho and Rho kinase.

The ephrins, ligands of Eph receptor tyrosine kinases, have been shown to act as repulsive guidance molecules and to induce collapse of neuronal growth cones. For the first time, we show that the ephrin-A5 collapse is mediated by activation of the small GTPase Rho and its downstream effector Rho kinase. In ephrin-A5-treated retinal ganglion cell cultures, Rho was activated and Rac was downregulated. Pretreatment of ganglion cell axons with C3-transferase, a specific inhibitor of the Rho GTPase, or with Y-27632, a specific inhibitor of the Rho kinase, strongly reduced the collapse rate of retinal growth cones. These results suggest that activation of Rho and its downstream effector Rho kinase are important elements of the ephrin-A5 signal transduction pathway.

Involvement of the GTP binding protein Rho in constitutive endocytosis in Xenopus laevis oocytes.

To study an endocytotic role of the GTP-binding protein RhoA in Xenopus oocytes, we have monitored changes in the surface expression of sodium pumps, the surface area of the oocyte and the uptake of the fluid-phase marker inulin. Xenopus oocytes possess intracellular sodium pumps that are continuously exchanged for surface sodium pumps by constitutive endo- and exocytosis. Injection of Clostridium botulinum C3 exoenzyme, which inactivates Rho by ADP-ribosylation, induced a redistribution of virtually all intracellular sodium pumps to the plasma membrane and increased the surface area of the oocytes. The identical effects were caused by injection of ADP-ribosylated recombinant RhoA into oocytes. The C3 exoenzyme acts by blocking constitutive endocytosis in oocytes, as determined using a mAb to the beta 1 subunit of the mouse sodium pump as a reporter molecule and oocytes expressing heterologous sodium pumps. In contrast, an increase in endocytosis and a decrease in the surface area was induced by injection of recombinant Val14-RhoA protein or Val14-rhoA cRNA. PMA stimulated sodium pump endocytosis, an effect that was blocked by a specific inhibitor of protein kinase C (Gö 16) or by ADP-ribosylation of Rho by C3. Similarly, the phorbol ester-induced increase in fluid-phase endocytosis in oocytes was inhibited by Gö 16, C3 transferase, or by injection of ADP-ribosylated RhoA. In contrast to C3 transferase, C. botulinum C2 transferase, which ADP-ribosylates actin, had no effect on sodium pump endocytosis or PMA-stimulated fluid-phase endocytosis. The data suggests that RhoA is an essential component of a presumably clathrin-independent endocytic pathway in Xenopus oocytes which can be regulated by protein kinase C.

Microinjection of recombinant p21rho induces rapid changes in cell morphology.

The rho proteins, p21rho, are ubiquitously expressed guanine nucleotide binding proteins with approximately 30% amino acid homology to p21ras, but their biochemical function is unknown. We show here that microinjection of constitutively activated recombinant rho protein (Val14rho) into subconfluent cells induces dramatic changes in cell morphology: 15-30 min after injection cells adopt a distinct and novel phenotype with a contracted cell body and finger-like processes still adherent to the substratum. Ribosylation of Val14rho with the ADP-ribosyltransferase C3 from clostridium botulinum, before microinjection, renders the protein biologically inactive, but it has no effect on either its intrinsic biochemical properties or on its interaction with the GTPase activating protein, rho GAP. Micro-injection of ribosylated normal rho, on the other hand, has a similar effect of injection of C3 transferase and induces complete rounding up of cells. We also report striking biochemical changes in actin filament organization when contact-inhibited quiescent 3T3 cells are injected with Val14rho protein. The effects induced by activation or inactivation of p21rho described here, suggest that the biological function of this protein is to control some aspect of cytoskeletal organization.

Botulinum C2 toxin ADP-ribosylates actin and enhances O2- production and secretion but inhibits migration of activated human neutrophils.

The binary botulinum C2 toxin ADP-ribosylated the actin of human neutrophils. Treatment of human neutrophils with botulinum C2 toxin for 45 min increased FMLP-stimulated superoxide anion (O2-) production 1.5-5-fold, whereas only a minor fraction of the cellular actin pool (approximately 20%) was ADP-ribosylated. Effects of botulinum C2 toxin depended on toxin concentrations, presence of both components of the toxin, and incubation time. Cytochalasin B similarly enhanced O2- production. The effects of botulinum C2 toxin and cytochalasin B were additive at submaximally, but not maximally effective concentrations and incubation time of either toxin. Botulinum C2 toxin also enhanced stimulation of O2- production by Con A and platelet-activating factor, but not by phorbol 12-myristate 13-acetate (PMA). Botulinum C2 toxin increased FMLP-induced release of N-acetyl-glucosaminidase by 100-250%; release of vitamin B12-binding protein induced by FMLP and PMA was enhanced by approximately 150 and 50%, respectively. Botulinum C2 toxin blocked both random migration of neutrophils and migration induced by FMLP, complement C5a, leukotriene B4, and a novel monocyte-derived chemotactic agent. The data suggest that botulinum C2 toxin-catalyzed ADP-ribosylation of a minor actin pool has a pronounced effect on the activation of human neutrophils by various stimulants.

Human fat cell lipolysis is primarily regulated by inhibitory modulators acting through distinct mechanisms.

The effects of adenosine deaminase and of pertussis toxin on hormonal regulation of lipolysis were investigated in isolated human fat cells. Adenosine deaminase (1.6 micrograms/ml) caused a two-to threefold increase in cyclic AMP, which was associated with an increase in glycerol release averaging 150-200% above basal levels. Clonidine, N6-phenylisopropyladenosine, prostaglandin E2, and insulin caused a dose-dependent inhibition of glycerol release in the presence of adenosine deaminase. Pretreatment of adipocytes with pertussis toxin (5 micrograms/ml) for 180 min resulted in a five- to sevenfold increase in cyclic AMP. Glycerol release was almost maximal and isoproterenol caused either no further increase or only a marginal additional increase of lipolysis after pretreatment with pertussis toxin, whereas cyclic AMP levels were 500 times higher than in controls. The effects of antilipolytic agents known to affect lipolysis by inhibition of adenylate cyclase activity, i.e., clonidine, N6-phenylisopropyladenosine, and prostaglandin E2, were impaired. In contrast, the antilipolytic action of insulin was preserved in adipocytes pretreated with pertussis toxin. As in controls, the peptide hormone had no detectable effect on cyclic AMP after pertussis toxin treatment. The findings support the view that the antilipolytic effect of insulin does not require adenylate cyclase or phosphodiesterase action. In addition, the results demonstrate that, upon relief of endogenous inhibition, human fat cell lipolysis proceeds at considerable (adenosine deaminase) or almost maximal (pertussis toxin) rates. A certain degree of inhibition, therefore, appears to be necessary for human fat cell lipolysis to be susceptible for hormonal activation.

The low molecular mass GTP-binding protein Rho is affected by toxin A from Clostridium difficile.

Enterotoxin A is one of the major virulence factors of Clostridium difficile, and the causative agent of antibiotic-associated pseudomembranous colitis. In cell culture (NIH-3T3, rat basophilic leukemia cells) toxin A inhibits Clostridium botulinum ADP-ribosyltransferase C3 (C3)-catalyzed ADP-ribosylation of the low molecular mass GTP-binding Rho proteins. Rho participates in the regulation of the microfilament cytoskeleton. Decrease in ADP-ribosylation of Rho occurs in a time- and concentration-dependent manner and precedes the toxin A-induced destruction of the actin cytoskeleton. Action of toxin A is not due to proteolytical degradation of Rho or to an inherent ADP-ribosyltransferase activity of toxin A. Toxin A-induced decrease in ADP-ribosylation is observed also in cell lysates and with recombinant RhoA protein. A heat stable low molecular mass cytosolic factor is essential for the toxin effect on Rho. Thus, the enterotoxin (toxin A) resembles the effects of the C. difficile cytotoxin (toxin B) on Rho proteins (Just, I., G. Fritz, K. Aktories, M. Giry, M. R. Popoff, P. Boquet, S. Hegenbath, and C. Von Eichel-Streiber. 1994. J. Biol. Chem. 269:10706-10712). The data indicate that despite different in vivo effects, toxin A and toxin B act on the same cellular target protein Rho to elicit their toxic effects.

Adenosine diphosphate-ribosylation of G-actin by botulinum C2 toxin increases endothelial permeability in vitro.

The endothelial cytoskeleton is believed to play an important role in the regulation of endothelial permeability. We used botulinum C2 toxin to perturb cellular actin and determined its effect on the permeability of endothelial cell monolayers derived from porcine pulmonary arteries. The substrate for botulinum C2 toxin is nonmuscle monomeric actin which becomes ADP-ribosylated. This modified actin cannot participate in actin polymerization and, in addition, acts as a capping protein. Exposure of endothelial cell monolayers to botulinum C2 toxin resulted in a dose- (3-100 ng/ml) and time-dependent (30-120 min) increase in the hydraulic conductivity and decrease in the selectivity of the cell monolayers. The effects of C2 toxin were accompanied by a time- and dose-dependent increase in ADP-ribosylatin of G-actin. G-Actin content increased and F-actin content decreased time- and dose-dependently in C2 toxin-treated endothelial cells. Phalloidin which stabilizes filamentous actin prevented the effects of botulinum C2 toxin on endothelial permeability. Botulinum C2 toxin induced interendothelial gaps. The effects occurred in the absence of overt cell damage and were not reversible within 2 h. The data suggest that the endothelial microfilament system is important for the regulation of endothelial permeability.

Regulation of mitogen-activated protein kinases in cardiac myocytes through the small G protein Rac1.

Small guanine nucleotide-binding proteins of the Ras and Rho (Rac, Cdc42, and Rho) families have been implicated in cardiac myocyte hypertrophy, and this may involve the extracellular signal-related kinase (ERK), c-Jun N-terminal kinase (JNK), and/or p38 mitogen-activated protein kinase (MAPK) cascades. In other systems, Rac and Cdc42 have been particularly implicated in the activation of JNKs and p38-MAPKs. We examined the activation of Rho family small G proteins and the regulation of MAPKs through Rac1 in cardiac myocytes. Endothelin 1 and phenylephrine (both hypertrophic agonists) induced rapid activation of endogenous Rac1, and endothelin 1 also promoted significant activation of RhoA. Toxin B (which inactivates Rho family proteins) attenuated the activation of JNKs by hyperosmotic shock or endothelin 1 but had no effect on p38-MAPK activation. Toxin B also inhibited the activation of the ERK cascade by these stimuli. In transfection experiments, dominant-negative N17Rac1 inhibited activation of ERK by endothelin 1, whereas activated V12Rac1 cooperated with c-Raf to activate ERK. Rac1 may stimulate the ERK cascade either by promoting the phosphorylation of c-Raf or by increasing MEK1 and/or -2 association with c-Raf to facilitate MEK1 and/or -2 activation. In cardiac myocytes, toxin B attenuated c-Raf(Ser-338) phosphorylation (50 to 70% inhibition), but this had no effect on c-Raf activity. However, toxin B decreased both the association of MEK1 and/or -2 with c-Raf and c-Raf-associated ERK-activating activity. V12Rac1 cooperated with c-Raf to increase expression of atrial natriuretic factor (ANF), whereas N17Rac1 inhibited endothelin 1-stimulated ANF expression, indicating that the synergy between Rac1 and c-Raf is potentially physiologically important. We conclude that activation of Rac1 by hypertrophic stimuli contributes to the hypertrophic response by modulating the ERK and/or possibly the JNK (but not the p38-MAPK) cascades.

Microbial toxins and the glycosylation of rho family GTPases.

Large clostridial cytotoxins act on cells by glycosylating low molecular mass GTPases using nucleotide-sugars as the sugar donor. These toxins are important virulence factors in human and animal diseases, but are also valuable cell biology tools. Recent findings shed some light on their mode of action and provide new insights into the structure/activity relationship of these bacterial toxins.

Effect of disruption of actin filaments by Clostridium botulinum C2 toxin on insulin secretion in HIT-T15 cells and pancreatic islets.

To examine their role in insulin secretion, actin filaments (AFs) were disrupted by Clostridium botulinum C2 toxin that ADP-ribosylates G-actin. Ribosylation also prevents polymerization of G-actin to F-actin and inhibits AF assembly by capping the fast-growing end of F-actin. Pretreatment of HIT-T15 cells with the toxin inhibited stimulated insulin secretion in a time- and dose-dependent manner. The toxin did not affect cellular insulin content or nonstimulated secretion. In static incubation, toxin treatment caused 45-50% inhibition of secretion induced by nutrients alone (10 mM glucose + 5 mM glutamine + 5 mM leucine) or combined with bombesin (phospholipase C-activator) and 20% reduction of that potentiated by forskolin (stimulator of adenylyl cyclase). In perifusion, the stimulated secretion during the first phase was marginally diminished, whereas the second phase was inhibited by approximately 80%. Pretreatment of HIT cells with wartmannin, a myosin light chain kinase inhibitor, caused a similar pattern of inhibition of the biphasic insulin release as C2 toxin. Nutrient metabolism and bombesin-evoked rise in cytosolic free Ca2+ were not affected by C2 toxin, indicating that nutrient recognition and the coupling between receptor activation and second messenger generation was not changed. In the toxin-treated cells, the AF web beneath the plasma membrane and the diffuse cytoplasmic F-actin fibers disappeared, as shown both by staining with an antibody against G- and F-actin and by staining F-actin with fluorescent phallacidin. C2 toxin dose-dependently reduced cellular F-actin content. Stimulation of insulin secretion was not associated with changes in F-actin content and organization. Treatment of cells with cytochalasin E and B, which shorten AFs, inhibited the stimulated insulin release by 30-50% although differing in their effects on F-actin content. In contrast to HIT-T15 cells, insulin secretion was potentiated in isolated rat islets after disruption of microfilaments with C2 toxin, most notably during the first phase. This effect was, however, diminished, and the second phase became slightly inhibited when the islets were degranulated. These results indicate an important role for AFs in insulin secretion. In the poorly granulated HIT-T15 cells actin-myosin interactions may participate in the recruitment of secretory granules to the releasable pool. In native islet beta-cells the predominant function of AFs appears to be the limitation of the access of granules to the plasma membrane.

Targeted delivery of an ADP-ribosylating bacterial toxin into cancer cells.

The actin cytoskeleton is an attractive target for bacterial toxins. The ADP-ribosyltransferase TccC3 from the insect bacterial pathogen Photorhabdus luminescence modifies actin to force its aggregation. We intended to transport the catalytic part of this toxin preferentially into cancer cells using a toxin transporter (Protective antigen, PA) which was redirected to Epidermal Growth Factor Receptors (EGFR) or to human EGF receptors 2 (HER2), which are overexpressed in several cancer cells. Protective antigen of anthrax toxin forms a pore through which the two catalytic parts (lethal factor and edema factor) or other proteins can be transported into mammalian cells. Here, we used PA as a double mutant (N682A, D683A; mPA) which cannot bind to the two natural anthrax receptors. Each mutated monomer is fused either to EGF or to an affibody directed against the human EGF receptor 2 (HER2). We established a cellular model system composed of two cell lines representing HER2 overexpressing esophageal adenocarcinomas (EACs) and EGFR overexpressing esophageal squamous cell carcinomas (ESCCs). We studied the specificity and efficiency of the re-directed anthrax pore for transport of TccC3 toxin and established Photorhabdus luminescence TccC3 as a toxin suitable for the development of a targeted toxin selectively killing cancer cells.

Rho-modifying C3-like ADP-ribosyltransferases.

C3-like exoenzymes comprise a family of seven bacterial ADP-ribosyltransferases, which selectively modify RhoA, B, and C at asparagine-41. Crystal structures of C3 exoenzymes are available, allowing novel insights into the structure-function relationships of these exoenzymes. Because ADP-ribosylation specifically inhibits the biological functions of the low-molecular mass GTPases, C3 exoenzymes are established pharmacological tools to study the cellular functions of Rho GTPases. Recent studies, however, indicate that the functional consequences of C3-induced ADP-ribosylation are more complex than previously suggested. In the present review the basic properties of C3 exoenzymes are briefly summarized and new findings are reviewed.

Clostridial Rho-inhibiting protein toxins.

Rho proteins are master regulators of a large array of cellular functions, including control of cell morphology, cell migration and polarity, transcriptional activation, and cell cycle progression. They are the eukaryotic targets of various bacterial protein toxins and effectors, which activate or inactivate the GTPases. Here Rho-inactivating toxins and effectors are reviewed, including the families of large clostridial cytotoxins and C3-like transferases, which inactivate Rho GTPases by glucosylation and ADP-ribosylation, respectively.

Rho proteins: targets for bacterial toxins.

GTP-binding proteins of the Rho family are regulators of the actin cytoskeleton and molecular switches in various signal transduction pathways. The Rho proteins are targets for bacterial protein toxins that either inactivate GTPases by ADP-ribosylation or glucosylation, or activate them by deamidation. Rho proteins play essential roles in host cell invasion by bacteria.

Synergistic inhibition of human platelet adenylate cyclase by stable GTP analogs and epinephrine.

Adenylate cyclase inhibition by stable GTP analogs and their interaction with epinephrine were studied in human platelet membranes. Whereas basal enzyme activity was increased by these nucleotides, the stable GTP analogs decreased the adenylate cyclase activity stimulated by fluoride or forskolin by maximally 60 to 70%, with the potency order, guanosine 5'-O-(3-thiotriphosphate) (GTP gamma S) greater than guanyl-5'-ylimidodiphosphate greater than guanyl-5'-ylmethylenediphosphate. The inhibition of the forskolin-stimulated enzyme by GTP gamma S was half-maximal at about 4 nM, occurred after a time lag period, which was inversely related to the GTP gamma S concentration, and was resistant to washing of the membranes. Prostaglandin E1-stimulated activity exhibited a biphasic response towards GTP gamma S, with activation occurring at low (1 nM) and inhibition at higher GTP gamma S concentrations. The inhibitory effect of GTP gamma S was competitively antagonized by GTP. This antagonism was prevented by epinephrine, which inhibited the stimulated platelet adenylate cyclase in the presence of GTP to the same degree as observed with GTP gamma S alone. In the absence of GTP, epinephrine largely diminished the time lag required for the inhibitory action of GTP gamma S. Furthermore, the decrease in final activity induced by GTP gamma S was amplified by epinephrine. Whereas the acceleration of the inhibitory action of GTP gamma S was observed at low and high GTP gamma S concentrations, the amplification by epinephrine was observed only at submaximally effective concentrations of GTP gamma S.

Cholera toxin does not impair hormonal inhibition of adenylate cyclase and concomitant stimulation of a GTPase in adipocyte membranes.

The influence of cholera toxin on hormonal inhibition of adenylate cyclase and concomitant stimulation of low Km GTPase was studied in adipocyte membrane preparations. In hamster adipocyte ghosts, cholera toxin caused an about 8-fold activation of the adenylate cyclase. The antilipolytic hormonal factors, prostaglandin E1 (1 micro M), N6-phenylisopropyladenosine (1 micro M) and nicotinic acid (30 micro M) reduced both basal and cholera toxin-stimulated enzyme activities. Similar data with regard to inhibition of cholera toxin-stimulated adenylate cyclase were obtained in mouse and rat adipocyte ghosts. As studied in hamster adipocyte ghosts, prostaglandin E1 (1 micro M) increased GTP hydrolysis by a low Km GTPase by about 3-4 fold. Pretreatment of the membrane preparation with cholera toxin did not impair prostaglandin E1-induced GTPase stimulation. The data suggest that cholera toxin does not directly affect the GTPase enzyme stimulated by adenylate cyclase inhibitory hormones.