Mechanism of action of choleragen andE. coli heat-labile enterotoxin: activation of adenylate cyclase by ADP-ribosylation

Summary Choleragen exerts its effects on cells through the activation of adenylate cyclase. The initial event appears to be the binding of the B subunit of the toxin to ganglioside G_M1 on the cell surface, following which there is a delay prior to activation of adenylate cyclase. Patching and capping of the toxin on the cell surface, perhaps involved in the internalization of the enzymatically active subunit, may be occuring during this time. The activation of adenylate cyclase, which is catalyzed by the A₁ peptide of choleragen, does not require the B subunit or ganglioside G_M1. The A₁ peptide catalyzes the transfer of ADP-ribose from NAD to an amino acid, probably arginine, in a 42 000 dalton membrane protein. This protein appears to be the GTP-binding component (or G/F factor) of the adenylate cyclase system and is cruical to the regulation of cyclase activity by hormones such as epinephrine. ADP-ribosylation of the G/F factor is enhanced by GTP and, in some systems, by a cytosolic factor. GTP is also required for stabilization and optimal catalytic function of the choleragen-activated cyclase. Calmodulin, a calcium-binding protein, is necessary for expression of catalytic activity of the toxin-activated adenylate cyclase in brain and other tissues. The ADP-ribosyltransferase activity required for activation of the cyclase is an intrinsic property of the A₁ peptide of choleragen which is expressed only after the peptide is released from the holotoxin by reduction of a single disulfide bond. In the absence of cellular components, choleragen catalyzes the ADP-ribosylation of small guanidino compounds such as arginine as well as peptides and proteins that contain arginine. It is assumed, therefore, that the site of ADP-ribosylation in the natural acceptor protein is an arginine or similar amino acid. When guanidino compounds are not present as ADP-ribose acceptors, choleragen hydrolyzes NAD to ADP-ribose and nicotinamide at a considerably slower rate. E. coli heat-labile enterotoxin (LT) is very similar to choleragen in structure and function. It consists of two types of subunits, A and B, with sizes comparable to those of the A and B subunits of choleragen. Binding of LT to the cell surface is enhanced by prior incorporation of G_M1 but not other gangliosides; the oligosaccharide of G_M1 specifically interacts with LT and its B subunit. The A subunit of LT exhibits ADP-ribosyltransferase activity following activation by thiol to release the A₁ peptide. The A subunit of LT can be isolated in an ‘unnicked’ form and thus requires, in addition to reduction by a thiol, proteolytic cleavage to generate the active A₁ peptide. Like choleragen, LT uses guanidino compounds as model ADP-ribose acceptors and catalyzes the ADP-ribosylation of a 42 000 dalton protein in cell membrane prepatations. ADP-ribosyltransferases that use arginine as ADP-ribose acceptors are not restricted to bacterial systems; such an enzyme has been purified to apparent homogeneity (>500 000-fold) from turkey erythrocytes. Based on a subunit molecular weight of 28 000, its turnover number with arginine as the ADP-ribose acceptor is considerably higher than that of either toxin. Although with low molecular weight guanidino derivatives the substrate specificity of the enzyme is similar to that of choleragen, with protein substrates it clearly differs. The physiological role of the turkey erythrocyte transferase remains to be established.     

Membrane-bound enterotoxin of Vibrio cholerae

The mode of transport of the complex toxin molecule of Vibrio cholerae (which has a mol. wt of 84000 and consists of several subunits) across the inner and outer membranes of V. cholerae is not known. In this study we found two peptides in the outer and inner membranes of V. cholerae which may be the form in which the toxin subunits are transported across the membrane. We examined two growth conditions: aerobic growth at 37 degrees C, when most of the synthesized toxin is membrane-bound; and anaerobic growth at 37 degrees C, when little toxin remains membrane-bound, the toxin being released into the growth medium. When V. cholerae was grown aerobically at 37 degrees C, the outer and the inner membranes contained two peptides with mol. wts of approximately 22000 and 6000 which were not found in the outer or the inner membrane of anaerobically grown cells. Sodium deoxycholate, which releases membrane-bound toxin, released several peptides including the 22000 and the 6000 mol. wt peptides. Trypsin also released the 22000 and 6000 mol. wt peptides. Purified cholera toxin had three kinds of peptides, of mol. wt 21000 (A1 peptide), 11000 (B subunit) and 5000 (A2 peptide). We postulate that the membrane peptides may be precursors of the A subunit of the toxin molecule.     

Mechanism of molybdate activation of adenylate cyclase

Molybdate activation of rat liver plasma membrane adenylate cyclase has been examined and compared with the effect of glucagon, Gpp(NG)p and fluoride. Glucagon does not stimulate the detergent solubilized enzyme, though molybdate, fluoride, and Gpp(NH)p are effective in this regard. The stimulatory effects of either fluoride or molybdate are additive with those of GTP and do not require guanyl nucleotide to evoke their activation. Neither fluoride nor molybdate can substitute for GTP when glucagon is the activator of rat liver adenylate cyclase. The stimulatory effects of either ion on adenylate cyclase are additive with that produced by glucagon. Activation of adenylate cyclase by either molybdate or fluoride occurs by a mechanism distinct from that of glucagon or guanyl nucleotide. The data presented here suggest that fluoride and molybdate may act via a similar mechanism of action. Neither ion displays a lag in activation of adenylate cyclase. The pH profiles of fluoride and molybdate-stimulated adenylate cyclase activity are similar, and distinct from guanyl nucleotide-stimulated activity. Cholera toxin treatment of adenylate cyclase blocks fluoride and molybdate stimulation of the enzyme to the same extent, while enhancing the activation obtained with GTP and hormones.     

Mechanism of adenylate cyclase activation by the rat lung cytoplasmic factors

Epinephrine, histamine and prostaglandin E₁ stimulated adenylate cyclase activity in lung membranes and their stimulation of the enzyme activity was completely blocked by propranolol, metiamide and indomethacin, respectively. A partially-purified activator from the adult rat lung also enhanced adenylate cyclase activity in membranes. However, stimulation of adenylate cyclase by the rat lung activator was not abolished by the above receptor antagonists. Further, epinephrine, NaF and Gpp(NH)p stimulated adenylate cyclase activity rather readily, whereas stimulation of the enzyme activity by the lung activator was evident after an initial lag phase of 10 min. Also, the lung activator produced additive activation of adenylate cyclase with epinephrine, NaF and Gpp(NH)p. These results indicate that the lung activator potentiates adenylate cyclase activity in membranes by a mechanism independent from those known for epinephrine, NaF and Gpp(NH)p. Incubation of lung membranes for 30 min at 40°C resulted in a loss of adenylate cyclase activation by NaF and Gpp(NH)p. Addition of the released proteins to the heat-treated membranes did not restore the enzyme response to these agonists. However, heat treatment of lung membranes in the presence of 2-mercaptoethanol or dithiothreitol prevented the loss of adenylate cyclase response to NaF and Gpp (NH)p. N-ethylmaleimide abolished adenylate cyclase activation by epinephrine, NaF, Gpp(NH)p and the lung activator. These results indicate that the sulfhydryl groups are important for adenylate cyclase function in rat lung membranes.Abbreviations Gpp(NH)p 5-Guanylimidodiphosphate     

Regulation of adenylate cyclase by hormones and G-proteins

Over the past few years, it has become apparent that a large number of transmembrane signaling systems operate through heterotrimeric G-proteins [( 1] Gilman, A.G. (1984) Cell 36, 577-579; [2] Baker, P.F. (1986) Nature 320, 395). Adenylate cyclase is regulated by stimulatory hormones through Gs(alpha s beta gamma) and inhibitory hormones through Gi(alpha i beta gamma) [( 2]; Katada, T. et al. (1984) J. Biol. Chem. 259, 3586-3595), whereas the breakdown of phosphatidylinositol bisphosphate (PIP2) to inositol trisphosphate (IP3) and diacylglycerol (DG) by phospholipase C is probably also mediated by a heterotrimeric G-protein (Go or Gi) [1,2]. Similarly, the activation of cGMP phosphodiesterase by light-activated rhodopsin is mediated through the heterotrimeric G-protein transducin (Stryer, L. (1986) Rev. Neurosci. 9, 89-119). Other transmembrane signaling systems may also be found to involve G-proteins similar to those already recognized. Because of the emerging universality of G-proteins as transducers of receptor-triggered signals, it may be useful to evaluate the current models prevailing in the adenylate cyclase field, as these models seem to guide our way in evaluating the role of G-proteins in transmembrane signaling, in general.     

Determination of Vibrio cholerae enterotoxin by the passive immune hemolysis technic

To determine the antigenic determinants of cholera toxin, the passive immune hemolysis test was used. This test, proved to be highly sensitive (100-200 pg/ml) and specific, yielded results quicker than all other immunological methods for the determination of cholera toxin. The study of 36 cholera and NAG-vibrio strains revealed that V. cholerae synthesized the greatest amount of the toxin, whereas V. eltor formed a heterogenous group, comprising strains capable of synthesizing the toxin in considerable amounts, as well as strains synthetizing no toxin. Some strains of NAG-vibrio were found to produce insignificant amounts of the toxin.     

Calcium-independent activation of adenylate cyclase by calmodulin

Adenylate cyclase of Bordetella pertussis is stimulated by calmodulin by two distinct interactions. At low activator concentrations (approximately equal to 1 nM) the process is Ca2+-dependent (i.e. inhibited by EGTA added before calmodulin). High activator concentrations (approximately equal to 0.1-10 microM) stimulate adenylate cyclase also in the presence of EGTA, an effect not accounted for by residual Ca2+ or low concentrations of Ca X calmodulin, which thus appears to be due to calcium-free calmodulin. Some calmodulin dose-response curves show both phases of stimulation, separated by a plateau of activity, and half-maximal activating concentrations differ by 100-300-fold. Both effects are on the V and not the Km for ATP and are not mimicked by 10(5)-fold greater concentrations of parvalbumin or by various polyanions. In addition, adenylate cyclase stimulation at high calmodulin concentrations is greater in the presence of EGTA than in its absence. This enhancement is also produced by 1,10-phenanthroline and 8-hydroxyquinoline but not by non-chelating isomers. These compounds are poor Ca2+ chelators, stimulate at any calmodulin concentration (unlike EGTA), and suggest regulation of this adenylate cyclase by a second metal ion.     

Genetic mapping of Vibrio cholerae enterotoxin structural genes

The structural genes which constitute the cholera toxin operon, ctxAB, were genetically mapped in the Vibrio cholerae El Tor strain RV79. This strain of V. cholerae contains two copies of the ctx operon located on a 7-kilobase-pair tandemly duplicated region. We began by isolating a vibriophage VcA1 insertion mutation in one of the two ctxA genes located in this region. The mutant carrying this ctxA::VcA1 insertion, DC24, was converted to a VcA1-facilitated donor by introduction of the conjugal plasmid pSJ15, which carries an inserted copy of a defective VcA1-like prophage. The donor characteristics of DC24(pSJ15) indicated that the ctxA::VcA1 insertion mutation was near the trp region of the V. cholerae chromosome. Subsequent RV79 three-factor crosses were performed between VcA1-facilitated donors and recipient strains carrying one of two structural gene mutations in ctx, either delta ctxA23P Kmr or delta ctx-7922. The former was constructed by an in vivo marker exchange procedure and could be scored either by its kanamycin resistance phenotype or by its lack of DNA sequences homologous to the ctxA region. The delta ctx-7922 mutation is a total deletion of both ctx copies of strain RV79. The three-factor cross data strongly suggest that the two ctx loci of RV79 map between the nal and his genes of V. cholerae in the trp nal his linkage group. Physical analysis and heterologous crosses between an RV79 El Tor donor and a 569B classical recipient indicates that one of the two 569B ctx operon copies maps in the same region as the RV79 ctx loci (i.e., linked to nal). Together with previously published observations, these data show that the ctx structural genes are not closely linked to other genes known to affect toxin production in V. cholerae.     

Mechanism of action ofVibrio cholerae enterotoxin

Summary The characteristics of the cholera toxin-stimulated adenylate cyclase of toad (Bufus marinus) and rat erythrocyte plasma membranes have been examined, with special emphasis on the response to purine nucleotides, fluoride, magnesium and catecholamine hormones. Toad erythrocytes briefly exposed to low concentrations of cholera toxin (40,000 to 60,000 molecules per cell) and incubated 2 to 4 hr at 30°C exhibit dramatic alterations in the kinetic and regulatory properties of adenylate cyclase. The approximateK _m for ATP, Mg⁺⁺ increases from about 1.8 to 3.4mm in the toxinstimulated enzyme. The stimulation by cholera toxin increases with increasing ATP, Mg⁺⁺ concentrations, from 20% at low levels (0.2mm) to 500% at high concentrations (greater than 3mm). Addition of GTP, Mg⁺⁺ (0.2mm) restores normal kinetic properties to the toxin-modified enzyme, such that stimulation is most simply explained by an elevation ofV _max. GTP enhances the toxin-treated enzyme activity two-to fourfold at low ATP concentrations, but this effect disappears at high levels of the substrate. At 0.6mm ATP and 5mm MgCl₂ the apparentK _a for GTP, Mg⁺⁺ is 5 to 10m. The control (unstimulated) enzyme demonstrates a very small response to the guanyl nucleotide. 5-ITP also stimulates the toxin-treated enzyme but cGMP, guanine, and the pyrimidine nucleotides have no effect. Cholera toxin also alters the activation of adenylate cyclase by free Mg⁺⁺, decreasing the apparentK _a from about 25 to 5mm. (–)-Epinephrine sensitizes the toad erythrocyte adenylate cyclase to GTP and also decreases the apparentK _a for free metal. Sodium fluoride, which cause a 70- to 100-fold activation of enzyme activity, has little effect on sensitivity to GTP, and does not change the apparentK _a for Mg⁺⁺; moreover, it prevents modulation of these parameters by cholera toxin. Conversely, cholera toxin severely inhibits NaF activation, and in the presence of fluoride ion the usual three- to fivefold stimulation by toxin becomes a 30 to 60% inhibition of activity. The toxin-stimulated enzyme can be further activated by catecholamines; in the presence of GTP the (–)-epinephrine stimulation is enhanced by two- to threefold. The increased catecholamine stimulation of toad erythrocyte adenylate cyclase induced by cholera toxin is explained primarily by an increase in the maximal extent of activation by the hormones. Rat erythrocyte adenylate cyclase is also modified by cholera toxin. In the mammalian system the apparent affinity for the hormone appears to be increased. Cholera toxin thus induces profound and nearly permanent changes in adenylate cyclase by a unique process which mimics the stimulation by hormones in important ways, and which also accentuates the normal hormonal response. The relevance of these findings to the mechanism of action of cholera toxin is considered.Part of this work was reported at the 1974 meeting of the Federation of American Societies for Experimental Biology (Bennett & Cuatrecasas, 1974).     

ATP-dependent activation of adenylate cyclase

Incubation of rat liver plasma membranes with MgCl2, ATP, and an ATP-regenerating system at 4 degrees C provides a 4-7-fold persistent activation of adenylate cyclase. Enzyme activation is time-dependent and 48 h of incubation is usually required to achieve maximal stimulation of adenylate cyclase activity. The activation described is not affected by GTP, cAMP, or cGMP, and does not occur when ATP is replaced by a nonphosphorylating analogue, adenyl-5'-imidodiphosphate. In addition to ATP, the activation requires Mg2+ and an ATP-regenerating system. The activation described is not additive with that produced by fluoride and analysis of basal and fluoride activities following extended incubation for 48 h reveals identical activities which decay at the same rate. These results are consistent with our model (11) which invokes phosphorylation-dephosphorylation mechanisms in regulating adenylate cyclase activity.     

Charge effects in the activation of adenylate cyclase.

Polycations, including ribonuclease A, ribonuclease S protein and peptide, spermine, spermidine, and polylysines, enhance unstimulated and stimulated adenylate cyclase activity of beef thyroid membranes at low concentrations and inhibit these activities at high concentrations. Peak polylysine stimulation occurs with degrees of polymerization of 6 to 14, and for large polymers a potency limit for this maximum is reached at 4 X 10(-5) M expressed as lysine residues. Both enhancement and inhibition appear to be due to charge-charge interactions and are abolished by KC1. Polyanions are inhibitory only. The biphasic effect of polycations is seen on basal cyclase activity, occurs with prostaglandin E1- and 5'-guanylyl-imidodiphosphate-stimulated cyclase, but is most striking with thyrotropin. There is little enhancement of F--activated cyclase. The enhancement is not sensitive to changes in pH, Mg2+, or regenerating system and does not correlate with the stability constants between polycations and ATP. We suggest that the polycation effect is a general, electrostatic effect on membrane conformation and is not restricted to a particular receptor domain.     

Vasopressin-sensitive adenylate cyclase. Reversibility of hormonal activation

The reversibility of adenylate cyclase activation induced by vasopressin was studied by reducing the concentration of active peptide in contact with kidney medullo-papillary membranes. Reversibility of hormonal activation was only partial. The use of antagonists failed to demonstrate the reversibility of an adenylate cyclase activation induced by high affinity agonists. When antagonist was added after the agonist to membranes, a non-competitive inhibition was apparent. Active peptide was also eliminated from the incubation medium by treatment with agents capable of reducing the disulfide bridge of the hormonal molecule. Direct effects of reducers on adenylate cyclase activity were measured on enzyme activation induced by peptides lacking a disulfide bridge. There was no apparent correlation between the abilities of different reducers to inactivate free peptide in solution and their abilities to promote the reversibility of hormone-induced enzyme activation. Upon the addition of dithiothreitol, enzyme activity could be lowered to basal value and adenylate cyclase was again fully stimulatable. However, when dithiothreitol addition to stiumlated enzyme was combined with a 60-fold dilution of the incubation medium, no reversibility of hormonal activation occurred. These results illustrate that the processes involved in adenylate cyclase activation are only partially reversible.     

Biochemistry of Vibrio cholerae virulence: purification of cholera enterotoxin by preparative disc electrophoresis.

Procedures for cholera enterotoxin purification previously developed in this labarotory were not applicable to large-scale purification, and these methods resulted in low yields of pure toxin. An efficient scheme has been developed whereby pure cholera enterotoxin can be obtained from 6 to 8 liters of culture supernatant fluid. This method consists of concentration by membrane ultrafiltration followed by gel filtration and cation-exchange chromatography. Pure cholera enterotoxin of high biological potency was obtained after a final step of preparative acrylamide gel electrophoresis. The degree of purity of the toxin-antigen as well as its biological activity were determined at various setps of purification. This alternate technique for purification is offered because of the widespread interest in cholera enterotoxin as a specific stimulator of adenyl cyclase.     

Solubilized myocardial adenylate cyclase: activation by prostaglandins

Prostaglandins E₁, E₂, F_2α, and F_2α activated solubilized myocardial adenylate cyclase from guinea pigs and cats. The activation did not require the presence of added phospholipids in contrast to stimulation of the solubilized enzyme by catecholamines, glucagon, and histamine. The data may provide insight into the mechanism and cellular site of action of the prostaglandins.