Mechanism of action of Vibrio cholerae enterotoxin. Effects on adenylate cyclase of toad and rat erythrocyte plasma membranes

Cholera toxin: A paradigm for multi-functional engagement of cellular mechanisms (Review)

Cholera toxin (Ctx) from Vibrio cholerae and its closely related homologue, heat-labile enterotoxin (Etx) from Escherichia coli have become superb tools for illuminating pathways of cellular trafficking and immune cell function. These bacterial protein toxins should be viewed as conglomerates of highly evolved, multi-functional elements equipped to engage the trafficking and signalling machineries of cells. Ctx and Etx are members of a larger family of A-B toxins of bacterial (and plant) origin that are comprised of structurally and functionally distinct enzymatically active A and receptor-binding B sub-units or domains. Intoxication of mammalian cells by Ctx and Etx involves B pentamer-mediated receptor binding and entry into a vesicular pathway, followed by translocation of the enzymatic A1 domain of the A sub-unit into the target cell cytosol, where covalent modification of intracellular targets leads to activation of adenylate cyclase and a sequence of events culminating in life-threatening diarrhoeal disease. Importantly, Ctx and Etx also have the capacity to induce a wide spectrum of remarkable immunological processes. With respect to the latter, it has been found that these toxins activate signalling pathways that modulate the immune system. This review explores the complexities of the cellular interactions that are engaged by these bacterial protein toxins, and highlights some of the new insights to have recently emerged.

Membrane traffic and the cellular uptake of cholera toxin

In nature, cholera toxin (CT) and the structurally related E. coli heat labile toxin type I (LTI) must breech the epithelial barrier of the intestine to cause the massive diarrhea seen in cholera. This requires endocytosis of toxin-receptor complexes into the apical endosome, retrograde transport into Golgi cisternae or endoplasmic reticulum (ER), and finally transport of toxin across the cell to its site of action on the basolateral membrane. Targeting into this pathway depends on toxin binding ganglioside GM1 and association with caveolae-like membrane domains. Thus to cause disease, both CT and LTI co-opt the molecular machinery used by the host cell to sort, move, and organize their cellular membranes and substituent components.

The activation of rabbit intestinal adenylate cyclase by cholera toxin

Brush-border and basal-lateral membranes were prepared from rabbit intestinal epithelial cells by differential centrifugation and MgCl2 precipitation. The ADP-ribosylation of proteins in these fractions when incubated with [adenylate-32P]NAD+ and cholera toxin was investigated. Three proteins of molecular mass 45, 40 and 37 kDa were labelled in a toxin-dependent manner in each membrane fraction. The incorporation of 32P-labelled ADP-ribose was 18-fold greater in brush-border membranes than in basal-lateral membranes, comparable to the enrichment of sucrase (marker enzyme for the brush border) in these membranes. There was a 20% release of the 40 and 45 kDa proteins from the brush-border membrane following this ADP-ribosylation. Activation of adenylate cyclase by both cholera toxin and sodium fluoride was 2.7- and 2.3-fold greater, respectively, in basal-lateral membranes than in brush-border membranes, comparable to the enrichment of Na+/K+-ATPase (marker enzyme for the basal-lateral membrane) in these membranes. The effect of sodium fluoride on membranes pretreated with cholera toxin revealed no increase in adenylate cyclase activity above that due to the toxin. This presumably means that both toxin and fluoride activate adenylate cyclase by the same regulatory protein. The results show that cholera toxin catalyzes the ADP-ribosylation of regulatory proteins in the brush-border membrane, and these proteins then migrate to the basal-lateral membrane where they activate the catalytic component of adenylate cyclase.

Reduced hormone-stimulated adenylate cyclase activity in NIH-3T3 cells expressing the EJ human bladder ras oncogene

Recent studies have shown that the 21-kilodalton protein (p21) Ha-ras gene product shares sequence homology with and may exhibit biochemical properties similar to the mammalian guanine nucleotide-binding proteins. These data suggested that one of the biochemical functions of p21 in the vertebrate cell may be to regulate adenylate cyclase [ATP pyrophosphate-lyase (cyclizing), EC 4.6.1.1]. We determined both in intact NIH-3T3 murine cells and in membranes isolated from these cells that the hormone-stimulated adenylate cyclase activity of cells expressing the EJ human bladder carcinoma oncogene (EJ-ras) is significantly reduced compared with control cells. Thus, the levels of cAMP measured in the EJ-ras-transformed cells by radioimmunoassay are reduced 78% and 93% after prostaglandin and isoproterenol stimulation, respectively, compared with the levels in control cells. Treatment of the EJ-ras-transformed cells with pertussis toxin or cholera toxin did not correct the alterations in adenylate cyclase activity. Cells expressing the normal human Ha-ras gene displayed intermediate levels of adenylate cyclase hormone sensitivity; these levels of adenylate cyclase activity were greater than those in the EJ-ras-transformed cells but lower than in control cells. Hormone-stimulated adenylate cyclase activities in cells transfected with Rous sarcoma virus DNA were similar to those in control cells. These data support the hypothesis that both the normal and mutated Ha-ras p21s are related to guanine nucleotide-binding proteins.

Sensitivity to injected cholera toxin of the sodium efflux in single barnacle muscle fibers

A study has been made of the effect of microinjected cholera toxin (CT) on the efflux in single barnacle muscle fibers. Characteristically, injected CT causes sustained stimulation of the ouabain-insensitive Na efflux but only after a lag phase. An effect is seen with as little as a 10(-7) M-solution of CT. Sustained stimulation after a lag phase is also seen following injection of subunit A fragment. Enrichment of fibers with NAD+ fails to enhance the response to CT. Prior injection of GTP or its non-hydrolyzeable analogue, Gpp(NH)p, markedly reduces the response to CT, whilst prior injection of CT reduces the response to guanine nucleotides. Evidence is also brought forward that omission of external Ca2+ reversibly reduces the response to CT and that pre- or postinjection of EGTA markedly reduces the response to CT. In addition, fibers preinjected with CT show increased aequorin light emission. Whereas verapamil and Cd2+ are ineffective, both Mg2+ and trace metals, e.g. Fe and Zn, reverse the response to CT following injection. Prior injection of protein kinase inhibitor reduces the response to CT. As for calmodulin inhibitors, e.g. chlorpromazine, imipramine and mepacrine, they are effective in reducing the response to CT but not calmodulin antibody (IgG). Collectively, the above results are compatible with the view that sustained stimulation of the ouabain-insensitive Na efflux by injected CT is due to persistent activation of adenylate cyclase by the toxin and that a fall in myoplasmic pCa facilitates or augments this activation mechanism.

Cell-free extracts of Salmonella inhibit protein synthesis and cause cytotoxicity in eukaryotic cells

Cell-free, unconcentrated sonic extracts of several serotypes of Salmonella caused extensive detachment of intact Vero cells. Undiluted sonic extracts of these strains exhibited cell detachment in the range 20-50%. Upon dilution, the extract preparations caused a linear, dose-related cytotoxic effect on Vero cell monolayers. The heat-lability (100 degrees C for 30 min) of much of the cytotoxic activity in the extract ruled out the possible involvement of endotoxin in this toxic effect for eukaryotic cells and suggested that this toxic factor is probably a protein. It was demonstrated that these sonic extracts also inhibited the incorporation of 3H-leucine by Vero-cells and that the inhibitory events occurred 1-2 hr after exposure. When subjected to heating at 100 degrees C for 30 min, the ability of Salmonella extracts to inhibit protein synthesis of Vero cells was significantly but only partially destroyed. Because of Salmonella extract-treated Vero cells did not release 3H-uridine until 24-48 hr after addition of sonic extracts, cell lysis was considered to be a secondary event resulting from the early shutdown of protein synthesis, rather than a direct effect of the toxic factor on membrane integrity. Further studies are needed to determine if these two biological activities of Salmonella sonic extracts are due to a single toxic molecule or result from two distinct toxin molecules.

Characterization of Salmonella toxin released by mitomycin C-treated cells

The enzyme-linked immunosorbent assay and the Chinese hamster ovary floating cell assay for cholera toxin have proven to be sensitive and reliable tests for determining the antigenic and biological characteristics of Salmonella toxin, respectively. The addition of mitomycin C to the culture media 3 h after inoculation resulted in increased amounts of Salmonella toxin in culture filtrates but had the reverse effect on cell sonic extracts. Our data suggested that the increased amounts of Salmonella toxin culture filtrates caused by mitomycin C were due to cell lysis, resulting in the release of intracellular toxin, rather than to an increase in the synthesis of Salmonella toxin. The biological activity of Salmonella toxin was heat labile at 100 degrees C. The antigenic structure of the toxin appeared to remain intact after exposure to temperatures as high as 100 degrees C but was altered somewhat when the toxin was subjected to autoclaving. The toxin had an isoelectric point in the pH range from 4.3 to 4.8 and an estimated molecular weight which appeared to be more than 110,000. With the exception of the range for its isoelectric point, its molecular weight, and its low concentration in filtrates and sonic extracts, Salmonella toxin was very similar in biological and antigenic characteristics to cholera toxin. The antigenic and biological assays described here provide an effective basis for extending our study of Salmonella toxin.

Activation of pigeon erythrocyte adenylate cyclase by cholera toxin. Partial purification of an essential macromolecular factor from horse erythrocyte cytosol

A cytosolic, macromolecular factor required for the cholera toxin-dependent activation of pigeon erythrocyte adenylate cyclase and cholera toxin-dependent ADP-ribosylation of a membrane-bound 43,000 dalton polypeptide has been purified 1100-fold from horse erythrocyte cytosol using organic solvent precipitation and heat treatment. This factor, 13,000 daltons, does not absorb to anionic or cationic exchange resins, is sensitive to trypsin or 10% trichloroacetic acid and is not extractable by diethyl ether. Activation of adenylate cyclase by cholera toxin requires the simultaneous presence of ATP (including possible trace GTP), NAD+, dithiothreitol, cholera toxin, membranes and the cytosolic macromolecular factor. Reversal of cholera toxin activation of adenylate cyclase, and of the toxin-dependent ADP-ribosylation, requires the presence of the cytosolic factor. The ability of the purified cytosolic factor to influence the hormonal sensitivity of liver membrane adenylate cyclase may provide clues to its physiological functions.

Synergistic protection against experimental cholera by immunization with cholera toxoid and vaccine

Rabbits were immunized with two parenteral injections of Wellcome toxoid PX389A, Wyeth toxoid 20101, or Merck bivalent vaccine. Other groups of rabbits were immunized with combinations of the Merck vaccine and each of the two toxoids. Antitoxin responses were monitored in each group of rabbits before livecell challenge of each animal by the ligated intestinal loop assay. Inaba and Ogawa strains of Vibrio cholerae were used for challenge experiments. Basically, the data indicate that the toxoids were equivalent in antigenic potency and antitoxin responses were unaffected by combination of the toxoids with the whole-cell vaccine. The 50 microgram doses of each toxoid as well as the 4 X 10(9) cells of the bivalent vaccine provided the same magnitude of protection against live-cell challenge with either Inaba or Ogawa vibrios. Immunization with either toxoid in combination with the bivalent vaccine resulted in a synergistic protective response against live-cell challenge of intestinal loops with V. cholerae. Synergistic protection was observed when toxoid and vaccine were administered together by the oral and parenteral routes. Maximum protection was obtained when rabbits were immunized with the combined toxoid-whole-cell vaccine administered by both oral and parenteral routes.

The regulation of adenylate cyclase of the adrenal cortex

This short review summarizes some of the data concerning the regulation of adrenocortical adenylate cyclase by ACTH and other putative effectors, such as guanosine and nucleotides, divalent cations and adenosine. The available information on ACTH-sensitive adenylate cyclase of the adrenal cortex is discussed in comparison to other cyclase systems and the possible biochemical mechanisms of action of ACTH on the adrenal cortex.

Mechanism of action of choleragen

Choleragen exerts its effect on cells through activation of adenylate cyclase. Choleragen initially interacts with cells through binding of the B subunit of the toxin to the ganglioside GM1 on the cell surface. Subsequent events are less clear. Patching or capping of toxin on the cell surface may be an obligatory step in choleragen action. Studies in cell-free systems have demonstrated that activation of adenylate cyclase by choleragen requires NAD. In addition to NAD, requirements have been observed for ATP, GTP, and calcium-dependent regulatory protein. GTP also is required for the expression of choleragen-activated adenylate cyclase. In preparations from turkey erythrocytes, choleragen appears to inhibit an isoproterenol-stimulated GTPase. It has been postulated that by decreasing the activity of a specific GTPase, choleragen would stabilize a GTP-adenylate cyclase complex and maintain the cyclase in an activated state. Although the holotoxin is most effective in intact cells, with the A subunit having 1/20th of its activity and the B subunit (choleragenoid) being inactive, in cell-free systems the A subunit, specifically the A1 fragment, is required for adenylate cyclase activation. The B protomer is inactive. Choleragen, the A subunit, or A1 fragment under suitable conditions hydrolyzes NAD to ADP-ribose and nicotinamide (NAD glycohydrolase activity) and catalyzes the transfer of the ADP-ribose moiety of NAD to the guandino group of arginine (ADP-ribosyltransferase activity). The NAD glycohydrolase activity is similar to that exhibited by other NAD-dependent bacterial toxins (diphtheria toxin, Pseudomonas exotoxin A), which act by catalyzing the ADP-ribosylation of a specific acceptor protein. If the ADP-ribosylation of arginine is a model for the reaction catalyzed by choleragen in vivo, then arginine is presumably an analog of the amino acid which is ADP-ribosylated in the acceptor protein. It is postulated that choleragen exerts its effects on cells through the NAD-dependent ADP-ribosylation of an arginine or similar amino acid in either the cyclase itself or a regulatory protein of the cyclase system.

Irreversible activation of adenylate cyclase of toad erythrocyte plasma membrane by 5'-guanylylimidodiphosphate

The irreversible activation of adenylate-cyclase by 5'guanylylimidodiphosphate, a phosphoramidate analog of 5'GTP, has been examined in toad (Bufus marinus) plasma membranes using the technique of preincubating the membranes with the nucleotide under various controlled conditions followed by washing and subsequent assay of enzyme activity. Activation of adenylate cyclase by Gpp(NH)p, but not GTP, is essentially permanent and persists following extensive washing, prolonged incubation at 30 degrees C in the absence of the nucleotide, and after dissolution of the membranes with Lubrol PX. (-)-Isoproterenol increases the activation observed with maximal concentrations of Gpp(NH)p from eight- to 10-fold (in the absence of hormone) to 50- to 100-fold; final activities as high as 10-15 nmoles of cyclic AMP per min per mg protein are achieved. The activated state obtained with isoproterenol and Gpp(NH)p is also permanent and is not inhibited by propranolol. The synergism between Gpp(NH)p and hormone requires the simultaneous presence of these compounds, and the time-dependent enhancement of activation with (-)-isoproterenol may be interrupted by addition of propranolol. The stimulation is slow, and may proceed for as long as 45 min at 30 degrees C in the presence of maximal concentrations of Gpp(NH)p and (-)-isoproterenol. Very little activation occurs at 0 degrees C. The time course of activation at 30 degrees C exhibits an accelerating phase lasting from 5 to 30 min when Gpp(NH)p is added directly during assay of cyclase activity or when the membranes are preincubated for various times and washed prior to assay for a fixed time. The lag period occurs in the presence and absence of (-)-isoproterenol, although the rate of increase in velocity is greater with hormone. The length of the accelerating phase decreases with increasing concentrations of Gpp(NH)p, although it is still evident with maximal levels of Gpp(NH)p and hormone. However, prewarming the membranes at 30 degrees C for 10 min in the absence of Gpp(NH)p or (-)-isoproterenol results in an immediate onset of linear activation at a rate which is achieved in untreated membranes only after about 10 min. The events occurring during prewarming at 30 degrees C are readily reversible since chilling the warmed membranes to 0 degrees C results in a time course of activation identical to that of membranes maintained at 0 degrees C until addition of Gpp(NH)p. Activation is proportional to the concentration of Gpp(NH)p within the range of 10(-8) to 10(-4) mM. The apparent affinity for Gpp(NH)p increases with increasing time of incubation. The primary effect of increasing the concentration of Gpp(NH)p is to decrease the time required to obtain a maximal rate of activation. The possible relevance of these findings to the mechanism of action of Gpp(NH)p, adenylate cyclase and hormones is discussed within the context of current views of biological membranes which recognize the lateral mobility of membrane molecules.

Immunological probes into the mechanism of cholera toxin action

The use of antibodies to specific cell surface proteins or to ligands which interact with cell surface receptors is a powerful tool for analyzing the properties of membrane proteins and the consequences of specific cell surface ligand-receptor interactions. Two central observations concerning membrane structure and function, - the diffusibility of membrane proteins (1) and ligand-triggered modulation of specific receptors (2), have derived from the use of antibodies to analyze the properties of membrane proteins. In our study of the mechanism of action of cholera toxin, a protein which binds to a specific cell surface receptor and results in the activation of adenyl cyclase, considerable information has been gained through the use of immunological techniques. This review will briefly summarize the data underlying our current concept of cholera toxin action at the cell membrane and will emphasize those observations made through the use of immunological approaches.

Structure and function of cholera toxin and hormone receptors

The enterotoxin from Vibrio cholerae is a protein of 100,000 mol wt which stimulates adenylate cyclase activity ubiquitously. The binding of biologically active 125I-labeled choleragen to cell membranes is of extraordinary affinity and specificity. The binding may be restricted to membrane-bound ganglioside GM1. This ganglioside can be inserted into membranes from exogenous sources, and the increased toxin binding in such cells can be reflected by an increased sensitivity to the biological effects of the toxin. Features of the toxin-activated adenylate cyclase, including conversion of the enzyne to a GTP-sensitive state, and the increased sensitivity of activation by hormones, suggest analogies between the basic mechanism of action of choleragen and the events following binding of hormones to their receptors. The action of the toxin is probably not mediated through intermediary cytoplasmic events, suggesting that its effects are entirely due to processes involving the plasma membrane. The kinetics of activation of adenylate cyclase in erythrocytes from various species as well as in rat adipocytes suggest a direct interaction between toxin and the cyclase enzyme which is difficult to reconcile with catalytic mechanisms of adenylate cyclase activation. Direct evidence for this can be obtained from the comigration of toxin radioactivity with adenylate cyclase activity when toxin-activated membranes are dissolved in detergents and chromatographed on gel filtration columns. Agarose derivatives containing the "active" subunit of the toxin can specifically absorb adenylate cyclase activity, and specific antibodies against the choleragen can be used for selective immunoprecipitation of adenylate cyclase activity from detergent-solubilized preparations of activated membranes. It is proposed that toxin action involves the initial formation of an inactive toxin-ganglioside complex which subsequently migrates and is somehow transformed into an active species which involves relocation within the two-dimensional structure of the membrane with direct perturbation of adenylate cyclase molecules (virtually irreversibly). These studies suggest new insights into the normal mechanisms by which hormone receptors modify membrane functions.

Irreversible stimulation of adenylate cyclase activity of fat cell membranes of phosphoramidate and phosphonate analogs of GTP

The ability of 5'-guanylylimidodiphosphate (Gpp(NH)p) to stimulate irreversibly the adenylate cyclease activity of fat cell membranes has been studied by preincubating the membranes with this or related analogs followed by assaying after thoroughly washing the membranes. Activation can occur in a simple Tris-HCl buffer, in the absence of added divalent cations and in the presence of EDTA. Dithiothreitol enhances the apparent degree of activation, perhaps by stabilization. The importance of utilizing optimal conditions for stabilizing enzyme activity, and of measuring the simultaneous changes in the control enzyme, is illustrated. The organomercurial, p-aminophenylmercuric acetate, inhibits profoundly the activity of the native as well as the Gpp(NH)p-stimulated adenylate cyclase, but in both cases subsequent exposure to dithiothreitol restores fully the original enzyme activity. However, the mercurial-inactivated enzyme does not react with Gpp(NP)p, as evidenced by the subsequent restoration of only the control enzyme activity upon exposure to dithiothreitol. Thus, reaction with Gpp(NH)p requires intact sulfhydryl groups, but the activated state is not irreversibly destroyed by the inactivation caused by sulfhydryl blockade. GTP and, less effectively, GDP and ATP inhibit activation by Gpp(NH)p, but interpretations are complicated by the facts that this inhibition is overcome with time and that GTP and ATP can protect potently from spontaneous inactivation. These two nucleotides can be used in the Gpp(NH)p preincubation to stabilize the enzyme. The Gpp(NH)p-activated enzyme cannot be reversed spontaneously during prolonged incubation at 30 degrees C in the absence or presence of GTP, ATP, MgCl2, glycine, dithiothreitol, NaF or EDTA. The strong nucleophile, neutral hydroxylamine, decreases the Gpp(NH)p-activated enzyme activity and no subsequent activation is detected upon re-exposure to the nucleotide.

Mechanism of activation of adenylate cyclase by Vibrio cholerae enterotoxin. Relations to the mode of activation by hormones

The influence of Vibrio cholerae enterotoxin (choleragen) on the response of adenylate cyclase to hormones and GTP, and on the binding of 125I-labeled glucagon to membranes, has been examined primarily in rat adipocytes, but also in guinea pig ileal mucosa and rat liver. Incubation of fat cells with choleragen converts adenylate cyclase to a GTP-responsive state; (-)-isoproterenol has a similar effect when added directly to membranes. Choleragen also increases by two- to fivefold the apparent affinity of (-)-isoproterenol, ACTH, glucagon, and vasoactive intestinal polypeptide for the activation of adenylate cyclase. This effect on vasoactive intestinal polypeptide action is also seen with the enzyme of guinea pig ileal mucosa; the toxin-induced sensitivity to VIP may be relevant in the pathogenesis of cholera diarrhea. The apparent affinity of binding of 125I-labeled glucagon is increased about 1.5- to twofold in choleragen-treated liver and fat cell membranes. The effects of choleragen on the response of adenylate cyclase to hormones are independent of protein synthesis, and they are not simply a consequence to protracted stimulation of the enzyme in vivo or during preparation of the membranes. Activation of cyclase in rat erythrocytes by choleragen is not impaired by agents which disrupt microtubules or microfilaments, and it is still observed in cultured fibroblasts after completely suppressing protein synthesis with diphtheria toxin. Choleragen does not interact directly with hormone receptor sites. Simple occupation of the choleragen binding sites with the analog, choleragenoid, does not lead to any of the biological effects of the toxin.

Mechanism of activation of adenylate cyclase by cholera toxin

Cholera toxin (choleragen) can stimulate adenylate cyclase [EC 4.6.1.1; ATP pyrophosphate-lyase (cyclizing)] activity in whole particulate fractions or purified plasma membranes of homogenates of isolated fat cells provided special precautions are taken to stabilize the enzyme during the required preincubation period. As observed with intact cells, the activation exhibits a protracted (about 25 min) lag phase, and it is blocked by ganglioside GM1 and choleragenoid ("binding" subunit of toxin). The 36,000 molecular weight subunit ("active" subunit), a hydrophobic polypeptide which does not block choleragen binding or action, can directly activate the enzyme in intact cells without a lag phase. Its effects are not blocked by ganglioside GM1 or choleragenoid, yet the stimulated activity exhibits reduced fluoride and enhanced isoproterenol sensitivity, properties characteristic of the choleragen-activated enzyme. Binding of the 125I-labeled 36,000 molecular weight subunit to cells is not saturable and is unaffected by gangliosides, choleragen, or choleragenoid, and the bound material behaves as an integral membrane protein; this protein may simply partition into the membrane matrix. With increasing time of incubation cell-bound choleragen may dissociate into its component subunits, but these remain in the membrane. Using a double antibody immunoprecipitin system, substantial precipitation of cyclase activity occurs with antisera against the 36,000 molecular weight subunit provided toxin activation has occurred. The normal process of activation may involve an initially inactive toxin--ganglioside complex which, as a result of lateral mobility and multivalent binding (lag phase), results in destabilization of the molecule with release of the "active" subunit into the membrane core where it can spontaneously associate with and perturb the cyclase complex.

Mechanism of activation of adenylate cyclase by Vibrio cholerae enterotoxin

The kinetics and properties of the activation of adenylate cyclase by cholera enterotoxin have been examined primarily in toad erythrocytes, but also in avian erythrocytes, rat fat cells and cultured melanoma cells. When cholera toxin is incubated with intact cells it stimulates adenylate cyclase activity, as measured in the subsequently isolated plasma membranes, according to a triphasic time course. This consists of a true lag period of about 30 min, followed by a stage of exponentially increasing adenylate cyclase activity which continues for 110 to 130 min, and finally a period of slow activation which may extend as long as 30 hr in cultured melanoma cells. The progressive activation of adenylate cyclase activity by cholera toxin is interrupted by cell lysis; continued incubation of the isolated membranes under nearly identical conditions does not lead to further activation of the enzyme. The delay in the action of the toxin is not grossly dependent of the number of toxin-receptor (GM1 ganglioside) complexes, and is still seen upon adding a second dose of toxin to partially stimulated cells. Direct measurements indicate negligible intracellular levels of biologically active radioiodinated toxin in either a soluble or a nuclear-bound form. The effects are not prevented by Actinomycin D (20 mug/ml), uromycin (30 mug/ml), cycloheximide (30 mug/ml), sodium fluoride (10 mM) or sodium azide (1 mM); KCN, however, almost completely prevents the action of cholera toxin. The action of the toxin is temperature dependent, occurring at very slow or negligible rates below certain critical temperatures, the values of which depend on the specific animal species. Thetransition for toad erythrocytes occurs at 15 to 17 degrees C, while rat adipocytes and turkey erythrocytes demonstrate a discontinuity at 26 to 30 degrees C. The temperature effects are evident during the lag period as well as during the exponential phase of activation. The rate of decay of the stimulated adenylate cyclase activity of cultured melanoma cells indicates a half-time of about 36 hr. The rate of exponentially increasing activity and extent of enzyme activation are related to the number of bound toxin molecules according to a Langmuir adsorption isotherm and are half-maximal when about 2000 molecules of toxin are bound per cell. It is proposed that initially cholera toxin binds ineffectively, but that it is converted to an active form during the lag phase. This process may involve lateral motion of toxin-GM1 ganglioside complex within the plane of the membrane. The kinetics of adenylate cyclase activation are consistent with the possibility that during the exponential phase a bimolecular association is proceeding between the active form of the cholera toxin and some other membrane component. The possibility is considered that the cholera toxin molecule may bind directly to adenylate cyclase. These considerations may prove useful in understanding the possible interactions of active hormone-receptor complexes with adenylate cyclase in cell membranes.