The mechanism of action of Escherichia coli heat-stable toxin

Microbial toxins and diarrhoeal diseases: introduction and overview

Without question, diarrhoeal diseases constitute one of the greatest causes of morbidity and death on a global scale. To an increasingly recognized extent, they are caused by an expanding array of microbial products or "toxins'. The symposium focuses on microbial products that alter normal bowel function either by augmenting secretory pathways or by selectively destroying mucosal cells or pathways, thus leading to an imbalance in the concert of normal absorptive function that results in diarrhoea. An understanding of normal intestinal physiology is thus the key to unraveling the specific actions of microbial toxins. In many instances, the microbial toxins are themselves providing unique pharmacological tools with which to dissect normal intestinal function. Specifically, families of enterotoxins are reviewed that appear to cause secretion through the recognized second messengers of cyclic AMP and cyclic GMP as well as cyclic nucleotide-independent and calcium-dependent pathways. Potential "third messengers' such as the protein kinases, through which one or more of the second messengers may act, are also considered. We examine cytotoxins that alter the orchestrated function of specialized regions of intact intestinal mucosa by selectively impeding or killing certain cells, so leading to small intestinal or colonic pathology and contributing to diarrhoea. We also consider a wide range of recognized bacterial and parasitic agents and their enterotoxic products. In some instances, these toxins may strikingly resemble our own endogenous humoral regulators or hormones. At this point, the possible roles of viruses or other transmissible genome products in this area await further clarification. Finally, we examine pharmacological and immunological approaches to attacking the toxins themselves or the deranged physiology they cause, in order to approach the control of the potentially devastating diseases of diarrhoea.

Mechanisms by which cyclic nucleotides and other intracellular mediators regulate secretion

The regulation of Na+ and Cl- transport across surface membranes and tight junctions of intestinal epithelium is mediated through at least three intracellular messengers: (i) 3',5'-cyclic AMP, activating two types of cyclic AMP-dependent protein kinase, (ii) 3',5'-cyclic GMP, binding to a unique isoenzyme of cyclic GMP-dependent protein kinase, enriched in the intestinal brush border, and (iii) Ca2+ ions, partially acting through calmodulin and a Ca2+/phospholipid-dependent protein kinase (C kinase). Recent data on the subcellular distribution and molecular properties of the high affinity cyclic nucleotide and Ca2+ receptors, their influence on the phosphorylation state of specific membrane proteins, and the possible role of these target proteins in ion transport regulation, are reviewed. The following aspects are accentuated: (1) the asymmetrical compartmentation of cyclic AMP-dependent protein kinase isoenzymes in the enterocyte and its functional implications; (2) the structure and function of microvillous cyclic GMP-dependent protein kinase; (3) the integration of cyclic AMP and cyclic GMP signals through co-phosphorylation of a 25 000 Mr protein in the intestinal-microvilli; (4) the identification of C kinase in villous and crypt cells; (5) various levels of interaction between cyclic nucleotide and Ca2+ signals; and (6) priority areas for future studies on stimulus-secretion coupling.

Toxins which activate guanylate cyclase: heat-stable enterotoxins

Certain enteropathogenic bacteria, including strains of Escherichia coli, Klebsiella pneumoniae and Yersinia enterocolitica, elicit their diarrhoeagenic effects by elaborating small molecular weight, heat-stable enterotoxins (STs). Structural and functional characteristics indicate that ST peptides are heterogeneous and two major subtypes, STa and STb, have been identified. Molecules of STa, unlike those of STb, are methanol-soluble and elicit their pathogenic effects by activating host cell guanylate cyclase activity and thereby increasing tissue cyclic GMP content: this increase in cyclic GMP causes fluid secretion. STa binds to specific proteinaceous receptors on intestinal cells but the nature of STa-receptor coupling to guanylate cyclase is poorly understood. The actions of STa, including binding to its receptor, activation of guanylate cyclase and stimulation of electrolyte transport, are rapid, reversible and tissue-specific. STa activates only particulate and not soluble guanylate cyclase. It alters the Vmax but not the apparent Km of this enzyme for Mg-GTP or Mn-GTP. At concentrations above 0.5-1 mM, calcium inhibits the STa activation of guanylate cyclase. The effects of calmodulin antagonists such as chlorpromazine on the activation of guanylate cyclase by STa are less clear. Inhibitors of phospholipid and arachidonic acid cascade pathways interfere with both basal and STa-stimulated guanylate cyclase. Membrane integrity is essential for STa activation of guanylate cyclase and the STa-receptor complex may activate the enzyme by intramembrane protein-protein interactions and/or perturbations. Interference with membrane phospholipid could alter such coupling.

Expression of putative Escherichia coli heat-labile enterotoxin (LT) receptors on intestinal brush borders from pigs of different ages

Many strains of enterotoxigenic Escherichia coli (ETEC) that cause diarrhoea in young piglets secrete a heat-labile enterotoxin (LTp) that binds to specific glycoconjugates on porcine intestinal epithelial cells. Binding of LTp to an appropriate glycoconjugate facilitates the uptake and trafficking of the toxin into the cell, where it stimulates intracellular changes that promote fluid secretion and diarrhoea. The objective of the current study was to identify the LTp-binding glycoconjugates on porcine intestinal epithelial cells, the natural target cells for LTp. We found that LTp binds, in an age-correlated manner, to an acidic glycosphingolipid (GSL) that co-migrated with GM1 on thin-layer chromatography (TLC), a small acidic GSL that appears to be a sulphatide, a neutral GSL that co-migrated with neolactotetraglycosylceramide (nLc4) on TLC, and two glycoproteins (36 and 205 kDa). Of these potential LTp receptors, the GM1-co-migrating GSL was detected most intensely in young animals, while the other four LTp-binding glycoconjugates were detected most intensely in older pigs (> or= 4 weeks). Since ETEC primarily cause disease in young piglets, the GM1-co-migrating GSL is the most likely candidate for a functional LTp receptor.

Guanylin-, heat-stable enterotoxin of Escherichia coli- and vasoactive intestinal peptide-induced water and ion secretion in the rat intestine in vivo

The heat-stable enterotoxin of Escherichia coli binds to an intestinal receptor, guanylyl cyclase-C, and produces cGMP to induce diarrhea. Guanylin is an endogenous ligand of this receptor. In the present in vivo study, the intestinal water and ion secretion induced by mucosal application of 2 nmol/ml guanylin or 5 or 10 units/ml heat-stable enterotoxin into closed loops was compared in the rat. The characteristics of secretion induced by cAMP following intravenous perfusion of 1.2 nmol/100 g per h vasoactive intestinal peptide were compared to those induced by cGMP. Unidirectional Na+ and Cl- fluxes were estimated by addition of 22Na into the loop and i.v. injection of 36Cl. Guanylin induced less water and ion secretion than that produced by heat-stable enterotoxin in the colon, confirming the results of in vitro studies, and also in duodenum and ileum. The cAMP- or cGMP-mediated response had a similar pattern, i.e., an inhibition of Na+ absorption and an increase in anion secretion.

Involvement of K+ channel modulation in the proabsorptive effect of nitric oxide in the rat jejunum in vivo

The role of K+ channels in the mediation of the nitric oxide(NO)-induced proabsorptive effect in intestinal fluid transport was investigated in a functional study, using a model of ligated jejunal loops of anaesthetized rats in vivo. The K+ channel opener cromakalim and the K+ channel blocker glibenclamide were administered under basal conditions as well as under conditions, when fluid secretion was influenced by N omega-nitro-L-arginine methyl ester (L-NAME), prostaglandin E2, Escherichia coli heat stable enterotoxin a (E. coli STa) or L-arginine. Intravenous infusion of cromakalim (63.5 micrograms/kg per min) significantly enhanced net fluid absorption compared to controls, totally abolished net fluid secretion induced by L-NAME (0.55 mg/kg per min), reversed net fluid secretion induced by intraluminal instillation of E. coli STa (10 units/ml) to absorption, but did not influence fluid secretion elicited by close i.a. infusion of prostaglandin E2 (79 ng/min). Close i.a. infusion of glibenclamide (0.16 mg/kg per min) reversed net fluid absorption to net secretion, blocked the absorptive effect of L-arginine (8.88 mg/kg per min) and reduced the proabsorptive effect of cromakalim. The secretory effect of L-NAME was not further enhanced by glibenclamide. These results suggest that modulation of basolateral K+ channels by NO is involved in the mediation of its proabsorptive effect, since opening and closure of K+ channels mimicked, respectively counteracted, the action of NO-donors and inhibitors of NO-synthesis on intestinal fluid transport. The role of prostaglandins in the proabsorptive effect of NO remains to be elucidated. These results furthermore support the role of K+ channel openers as potential new antidiarrheal drugs.

N-terminal myristoylation is required for membrane localization of cGMP-dependent protein kinase type II

The apical membrane of intestinal epithelial cells harbors a unique isozyme of cGMP-dependent protein kinase (cGK type II) which acts as a key regulator of ion transport systems, including the cystic fibrosis transmembrane conductance regulator (CFTR)-chloride channel. To explore the mechanism of cGK II membrane-anchoring, recombinant cGK II was expressed stably in HEK 293 cells or transiently in COS-1 cells. In both cell lines, cGK II was found predominantly in the particulate fraction. Immunoprecipitation of solubilized cGK II did not reveal any other tightly associated proteins, suggesting a membrane binding motif within cGK II itself. The primary structure of cGK II is devoid of hydrophobic transmembrane domains; cGK II does, however, contain a penultimate glycine, a potential acceptor for a myristoyl moiety. Metabolic labeling showed that cGK II was indeed able to incorporate [3H]myristate. Moreover, incubation of cGK II-expressing 293 cells with the myristoylation inhibitor 2-hydroxymyristic acid (1 mM) significantly increased the proportion of cGK II in the cytosol from 10 +/- 5 to 35 +/- 4%. Furthermore, a nonmyristoylated cGK II Gly2 --> Ala mutant was localized predominantly in the cytosol after transient expression in COS-1 cells. The absence of the myristoyl group did not affect the specific enzyme activity or the Ka for cGMP and only slightly enhanced the thermal stability of cGK II. These results indicate that N-terminal myristoylation fulfills a crucial role in directing cGK II to the membrane.

Significance of nitric oxide in the stimulation of intestinal fluid absorption in the rat jejunum in vivo

1. The effects of inhibiting nitric oxide (NO)-synthase on fluid transport, mucosal cyclic GMP and cyclic AMP levels and intraluminal prostaglandin E2 (PGE2)-release were studied in a model of ligated jejunal loops of anaesthetized rats in vivo. Experiments were performed under basal conditions as well as under conditions, when net fluid secretion was induced by Escherichia coli heat stable enterotoxin a (E. coli STa) or PGE2. 2. Intravenous infusion of the NO-synthase inhibitor N omega-nitro-L-arginine methyl ester (L-NAME, 0.25-50 mg kg-1, 45 min) dose-dependently reversed net fluid absorption to net secretion, whereas infusion of D-NAME, the inactive enantiomer of L-NAME, in corresponding doses did not influence net fluid transport. N omega-nitro-L-arginine (L-NOARG, 25 mg kg-1), another NO-synthase inhibitor, also elicited net secretion of fluid. 3. L-NAME (25 mg kg-1)-induced net fluid secretion was reversed to net absorption by infusion of L-arginine (400 mg kg-1) or sodium nitroprusside (1 mg kg-1) and s.c. administration of indomethacin (10 mg kg-1). Hexamethonium (1 mg kg-1, s.c.), a ganglionic blocker and granisetron (100 micrograms kg-1, s.c.), a 5-HT3-receptor antagonist, did not influence L-NAME-induced net secretion. 4. Net fluid secretion induced by intraluminal instillation of E. coli STa (10 units ml-1) was enhanced by infusion of L-NAME (25 mg kg-1) and was inhibited by infusion of L-arginine (400 mg kg-1) and sodium nitroprusside (1 mg kg-1). D-Arginine (400 mg kg-1) did not influence E. coli STa-induced fluid secretion. Likewise, net fluid secretion induced by i.a. infusion of PGE2 (79 ng ml-1, 30 min) was enhanced by infusion of L-NAME and was inhibited by L-arginine and sodium nitroprusside. D-Arginine(400 mg kg-1) did not influence PGE2-induced fluid secretion.5. PGE2 levels in intraluminal fluid were not elevated after infusion of L-NAME (25mgkg-1) compared to controls.6. Mucosal cyclic GMP and cyclic AMP levels after L-NAME-treatment were not different from control values.7. These results indicate that nitric oxide plays an important role in the regulation of intestinal fluid transport. The data suggest a nitric oxide-dependent proabsorptive tone in the intestine, which possibly involves the enteric nervous system and suppression of prostaglandin formation. This proabsorptive tone also may downregulate fluid secretion induced by E. coli STa or PGE2.

Structure and function of cholera toxin and the related Escherichia coli heat-labile enterotoxin

Cholera and the related Escherichia coli-associated diarrheal disease are important problems confronting Third World nations and any area where water supplies can become contaminated. The disease is extremely debilitating and may be fatal in the absence of treatment. Symptoms are caused by the action of cholera toxin, secreted by the bacterium Vibrio cholerae, or by a closely related heat-labile enterotoxin, produced by Escherichia coli, that causes a milder, more common traveler's diarrhea. Both toxins bind receptors in intestinal epithelial cells and insert an enzymatic subunit that modifies a G protein associated with the adenylate cyclase complex. The consequent stimulated production of cyclic AMP, or other factors such as increased synthesis of prostaglandins by intoxicated cells, initiates a metabolic cascade that results in the excessive secretion of fluid and electrolytes characteristic of the disease. The toxins have a very high degree of structural and functional homology and may be evolutionarily related. Several effective new vaccine formulations have been developed and tested, and a growing family of endogenous cofactors is being discovered in eukaryotic cells. The recent elucidation of the three-dimensional structure of the heat-labile enterotoxin has provided an opportunity to examine and compare the correlations between structure and function of the two toxins. This information may improve our understanding of the disease process itself, as well as illuminate the role of the toxin in studies of signal transduction and G-protein function.

Pharmacologic action of Escherichia coli heat-stable (STa) enterotoxin

Escherichia coli produces a heat-stable (STa) enterotoxin that belongs to a family of peptides that mediate several diarrheal diseases, including traveler's diarrhea and epidemic diarrhea in infants and newborns. The STa enterotoxin consists of 18 or 19 amino acids and is encoded by genes specified on a transposon. Intestinal secretion is induced by specific binding to high affinity receptors that reside on the brush border cell membrane of the small intestine. Receptor activation by STa enterotoxin induces a sequence of events that culminate in the release of fluid and electrolytes into the intestinal lumen. These events include the stimulation of particulate guanylate cyclae and subsequent increase of intracellular cyclic GMP, involvement of particulate protein kinase, elevation of intracellular calcium, and activation of the phosphatidylinositol pathway. The release of archidonic acid and production of prostaglandins and/or leukotrienes have also been implicated in the action of STa. Evidence indicates that the STa enterotoxin receptor may be a single multifunctional membrane protein.

Presence of functional receptors for the Escherichia coli heat-stable enterotoxin in the gastrointestinal tract of the chicken

Receptors for the Escherichia coli heat-stable enterotoxin (STa) were shown to be present throughout the digestive tract of the chicken, with binding activity present not only in the intestinal epithelium but also in the intestinal smooth muscle. Brush border membrane vesicles (BBMV) purified from chicken enterocyte homogenates and plasma membranes (SMPM) purified from intestinal smooth muscle homogenates were compared with pig enterocyte BBMV. All had similar 125I-STa binding affinities, but the 50% effective concentration for STa activation of guanylate cyclase was higher in SMPM than in BBMV. Maximal STa-stimulated guanylate cyclase activities were similar in chicken and pig BBMV and were seven- to eightfold higher than in SMPM, and the STa receptor density was five- to sixfold higher. Patterns unique to each membrane were demonstrated after affinity labelling of STa receptors with 125I-STa, sodium dodecyl sulfate-polyacrylamide gel electrophoresis, and autoradiography. The results demonstrated STa-stimulated guanylate cyclase activity in birds as well as mammals and suggested that there are different functional STa receptors in chicken BBMV and SMPM.

Nucleotide regulation of heat-stable enterotoxin receptor binding and of guanylate cyclase activation

Certain nucleotides were found to regulate the binding of the Escherichia coli heat-stable enterotoxin (STa) to its receptor in pig intestinal brush border membranes. ATP and adenine nucleotide analogues inhibited 125I-STa binding, while guanine nucleotide analogues stimulated binding, with maximal effects at 0.5-1.0 mM. The strongest inhibitors were adenosine 5'-[beta gamma-imido]triphosphate (App[NH]p) (36%) and adenosine 5'-[beta-thio]diphosphate (ADP[S]) (41%). Inhibition did not require Mg2+, and was blocked by p-chloromercuribenzenesulphonate (PCMBS). Stimulation of binding required Mg2+, was not prevented by PCMBS and was maximal with GDP[S] (41%). While App[NH]p and MgGDP[S] appeared to be acting at different sites, they also interfered with each other. These nucleotides exerted only inhibitory effects on STa-stimulated guanylate cyclase activity, in contrast with the stimulatory effects of adenine nucleotides on atrial natriuretic peptide (ANP)-stimulated guanylate cyclase. Inhibition by low concentrations of MgApp[NH]p and MgATP was weaker above 0.1 mM, while MgGDP[S] and magnesium guanosine 5'-[gamma-thio]triphosphate (MgGTP[S]) inhibited in a single phase. Inhibition by MgApp[NH]p, at all concentrations, was competitive with the substrate (MgGTP), as was that by MgGDP[S] and MgGTP[S]. Whereas membrane guanylate cyclases usually show positively co-operative kinetics with respect to the substrate, STa-stimulated activity exhibited Michaelis-Menten kinetics with respect to MgGTP. This changed to positive co-operativity when Lubrol PX was the activator, or when the substrate was MnGTP. These results suggest the presence of both a regulatory and a catalytic nucleotide-binding site, which do not interact co-operatively with STa activation.

Age-dependent changes in affinity-labeled receptors for Escherichia coli heat-stable enterotoxin in the swine intestine

Intestinal brush border membranes from 1-day-old and 4-week-old (day of weaning) pigs were affinity labeled with an Escherichia coli heat-stable enterotoxin (STa) by cross-linking 125I-STa to receptor proteins with disuccinimidyl suberate. Analysis by sodium dodecyl sulfate-polyacrylamide gel electrophoresis and autoradiography revealed that a radioactive protein with a relative molecular weight of 137,000 to 145,000 was present in both age groups. A strongly radioactive protein with an apparent Mr of 90,000 was present in the 1-day-old animals but not in those that were 4 weeks old. The major radioactive protein present in the older pigs had an Mr of 64,000 to 67,000, but this protein was missing or very weakly radioactive in the younger pigs. There was no significant difference between the groups in receptor affinity for STa, although the receptor density in the older animals was marginally significantly greater. STa-stimulated guanylate cyclase activity in membranes from 1-day-old pigs was only one-sixth that in 4-week-old pigs, although the basal and Lubrol PX-stimulated activities were similar.

Treatment of diarrhea of neonatal calves

Therapeutic strategies for the treatment of diarrhea of neonatal calves should be logical and should be targeted at correction of physiologic dysfunction. Appropriate, specific antimicrobial or antiprotozoal therapy should be instituted when colibacillosis, salmonellosis, or giardiasis is confirmed or suspected. All calves with diarrhea should be rehydrated if necessary, and proper nutritional support should be provided. Antisecretory agents such as flunixin meglumine and bismuth subsalicylate may be beneficial for treatment of calves with colibacillosis and salmonellosis. Adsorbants, such as attapulgite and bismuth subsalicylate, also may reduce loss of fluids. Perhaps loperamide or a similar drug will be proven effective in calves in the future. Potentially harmful drugs include several antimicrobial agents when they are administered orally, because they result in malabsorption; kaolin and pectin, which increase loss of ions during diarrhea; and motility modifiers that cause a decrease in all types of intestinal motor function. Finally, success should be measured by indicators of production such as survivability, days treated, weight gained, and net profit. Our goal should be to restore and maintain the health of the calf, not simply to alter the volume and consistency of the feces.

Solubilization and reprecipitation from intestinal brush border membranes of a complex containing guanylate cyclase activatable by the heat-stable enterotoxin

Extraction of pig intestinal brush border membranes with the zwitterionic detergent 3-[(3-cholamidopropyl)-dimethylammonio]-1-propanesulfonate (Chaps) in the presence of 0.5 M KCl yielded a solution which contained 60-70% of the receptor for the Escherichia coli heat-stable enterotoxin (STa) and of the Lubrol PX-activated guanylate cyclase activity present in the membrane. When the supernatant solution was diluted fivefold with 10 mM Hepes buffer (pH 7.4) and kept at 4 degrees C overnight, a precipitate formed. Centrifugation yielded a pellet (P2) which contained 25-30% of both the cyclase and the receptor in the original membranes, with a 2.5- to 3-fold enrichment of both. The process could be repeated for further enrichment (P4). The addition of MgCl2 to the diluted extract affected both basal and STa-stimulated activity of P2; 1 mM was optimal. P2 resembled membranes with respect to competitive inhibition of 125I-STa binding by STa, and the concentration-dependent activation of cyclase by STa. Guanylate cyclase in resolubilized P2 was also activated by STa. Most of the enzymes interfering with guanylate cyclase determinations were removed, as were the brush border marker enzymes sucrase and gamma-glutamyltransferase, and a GTP-binding protein that is a pertussis toxin substrate. Specific cross-linking of 125I-STa to receptors in the membrane was preserved in P2 and P4, the three proteins showing the strongest radioactivity having relative molecular masses of 55,000-60,000, 70,000-80,000, and 135,000-140,000. P2 and P4 appear to contain a complex of membrane proteins with certain functional properties intact.

Segment-specific effects of the heat-stable enterotoxin of E. coli on electrolyte transport in the rat colon

The heat-stable enterotoxin of E. coli (STa) induced an increase in short-circuit current (Isc) in the rat colon. The maximal increase in Isc was about three times larger in the proximal than the distal colon. The action of STa was mimicked by 8-Br-cyclic GMP. Unidirectional flux measurements revealed that STa decreased Na+ and Cl- absorption in the distal colon, while it decreased Na+ absorption and activated Cl- secretion in the proximal colon. In the distal, but not in the proximal colon, indomethacin inhibited the action of STa and of 8-Br-cyclic GMP. Inhibition by indomethacin could be overcome by addition of prostaglandin E2 or forskolin, but not by addition of a non-hydrolysable analogue of cyclic AMP, suggesting an action of STa on cyclic AMP hydrolysis. Amrinone and trequinsin, two inhibitors of cyclic GMP-inhibited phosphodiesterases, mimicked the action of STa on Isc and inhibited the response to a subsequent administration of the toxin indicating the modulation of a cyclic GMP-inhibited phosphodiesterase by STa in the distal colon. The results give evidence for different intracellular action sites of STa in the two parts of the rat colon.

Activation of intestinal brush border guanylate cyclase by aromatic disulphide compounds

Guanylate cyclase in pig intestinal brush border membranes was stimulated by certain aromatic disulphides. The most effective were 6-thioguanine disulphide [(TGS)2], 6-mercaptopurine disulphide, 6,6'-dithiodinicotinic acid, 5,5'-dithiobis-(2-nitrobenzoic acid) and 5-carboxy-2-thiouracil disulphide. (TGS)2 stimulated activity 15-fold when present at 0.1 mM. The optimum concentration for each disulphide was different, and higher concentrations were inhibitory. There was no activation by alkyl disulphides or by N-ethylmaleimide. Activation by 50 microM-(TGS)2 was partially reversed by later addition of 0.1 mM-dithiothreitol, whereas activation by the Escherichia coli heat-stable enterotoxin STa was relatively unaffected. Pretreatment of the membranes with (TGS)2 produced a concentration-dependent inhibition of STa-stimulated activity, while stimulating basal activity, until the activities were equal at 50 microM. Activity was [Mg2+]-dependent, the optimal [Mg2+] progressively increasing as the enzyme was stimulated by (TGS)2, STa and Lubrol PX respectively. However, (TGS)2 pretreatment prevented the shift to higher [Mg2+]optima induced by STa or Lubrol alone. Substitution of Mn2+ for Mg2+ in the reaction elevated basal activity and eliminated by activation (TGS)2. (TGS)2 only inhibited Mn2(+)-dependent activity (both basal and stimulated). The affinity of 125I-STa for its receptor was slightly increased by (TGS)2. We propose that (TGS)2 undergoes thiol-disulphide exchange with at least three different protein thiols of decreasing reactivity. The first is associated with Mg2(+)-dependent activation, the second is associated with a tonic inhibition of activity and the third is associated with the catalytic activity, although probably not at the active site.

Effect of antisecretory factor on Escherichia coli STa enterotoxin-induced alkalinisation of pig jejunal acid microclimate

The effect of challenge by Escherichia coli STa enterotoxin on pig jejunal mucosal surface pH was investigated in vivo. Exposure to STa resulted in a rapid and reversible alkalinisation (P less than 0.001) of the jejunal mucosa from 6.27 +/- 0.11 (5) to 6.89 +/- 0.03 (5). This action of STa is probably mediated through cyclic 3'5'-guanosine monophosphate (cGMP) since the 8-bromo analogue of cGMP induced the same effect as that observed after STa challenge. The action of STa on mucosal pH was partially inhibited by pre-administration of an antisecretory factor (ASF) preparation. The action of 8-bromo cGMP was unchanged by the presence of ASF. This implies that ASF inhibition occurs during the early stages of STa action prior to stimulation of guanylate cyclase. This effect of STa on the pig jejunal mucosal surface pH, or acid microclimate, may explain why weak acid supplementation of oral rehydration solutions can be ineffective in certain types of diarrhoeal disease.

Expression of the cloned gene for enterotoxin STb of Escherichia coli

This study involved the construction of hybrid plasmids to produce heat-stable enterotoxin type II of Escherichia coli (STb). The translation of the open reading frame for the STb gene estA was demonstrated in several ways. Studies using in vivo labeling with [35S]cysteine demonstrated a radiolabeled protein band on sodium dodecyl sulfate-polyacrylamide gel electrophoresis of the expected molecular weight of 5,000 for toxin STb. Insertion of translational or transcriptional termination signals into the BglII site of the estA gene blocked the expression of estA. The estA gene was cloned into high-expression vector pKC30 downstream from the strong pL promoter. Northern (RNA) blotting assays revealed a 10- to 20-fold increase in mRNA produced by strain C600F(pKC30STb) over other STb-producing strains, compared with little or no increase in enterotoxin activity demonstrated by bioassay. The estA gene, with its own promoter and Shine-Delgarno region and a portion of the sequence for the signal peptide deleted, was also inserted under the control of the tac promoter. Even after induction of the tac promoter by addition of isopropyl-beta-D-thiogalactopyranoside, no biologic enterotoxin activity could be identified. Neutralizing antibodies to STb were produced in rabbits by using either a purified OmpF-STb-beta-galactosidase fusion protein or a 19-amino-acid synthetic STb peptide coupled to keyhole limpet hemocyanin.

The effect of Escherichia coli heat-stable enterotoxin on protein kinase activity

Isolated rat enterocytes were incubated with E. coli heat-stable enterotoxin or buffer alone and the protein kinase activity and cyclic GMP level determined on the particulate fraction or cytosol, respectively. In the control cells, particulate protein kinase activity and cyclic GMP concentration were at a maximum after 20 sec and 1 min of incubation, respectively. In heat-stable enterotoxin-treated cells the particulate protein kinase activity was significantly increased (P less than 0.05) after 20 sec of incubation, but decreased (P less than 0.05) after 30 sec, 1 min and 2 min, when compared to the control reaction. During this time period the concentration of intracellular cyclic GMP increased 10-fold. The effect of heat-stable enterotoxin on particulate protein kinase activity and cyclic GMP concentration was dose-dependent. Analysis of radioactive membrane phosphorylation products indicate a role for phosphoproteins with a mol. wt of 25,000 and 120,000. These results suggest that the action of heat-stable enterotoxin may involve an effect on protein kinase.

Effects of heat-stable Escherichia coli enterotoxin on intestinal alkaline secretion and transepithelial potential difference in the rat intestines in vivo

The effects of the heat-stable enterotoxin of Escherichia coli (STa) on intestinal alkaline secretion and transepithelial electric potential difference (PD) were investigated in vivo in denervated segments of rat jejunum, ileum, and proximal colon. STa caused a significant increase in alkaline secretion in the jejunum but not in the ileum or colon. The jejunal effect of STa may be ascribed to a stimulation of bicarbonate secretion and/or an inhibition of Na+/H+ exchange. With regard to PD, STa caused a marked rise in colonic PD, whereas only a small response was found in the jejunum. No effect on PD was seen in the ileum. Hexamethonium (10 mg/kg intravenously) significantly diminished the effects of STa on PD, whereas only a small inhibition of the STa-induced alkaline secretion was observed. The effect of lidocaine on PD and alkaline secretion was found to be similar to that of hexamethonium. Atropine had no effect on any of the studied variables. These findings suggest that STa exerts, via nerves, a profound influence on the jejunal transport mechanisms responsible for the changes in PD, whereas the influence on alkaline secretion is to a large extent not mediated via enteric nerves. Thus, the extent of enteric nervous control of epithelial function differs for different transport functions. The findings also indicate that the bicarbonate ion is not the anion mainly responsible for the fluid secretion elicited by STa.

The molecular basis of chloride channel dysregulation in cystic fibrosis

The opening and closing of chloride (Cl-) channels in the apical membrane of epithelial cells is regulated by hormones, neurotransmitters and enterotoxins (intestine) acting through a variety of intracellular messengers, including cyclic nucleotides (cAMP, cGMP), calcium (Ca) and diacylglycerol (DAG). The chloride impermeability of epithelial membranes observed in cystic fibrosis (CF) patients does not result from a defect in the Cl- conducting properties of the channel or in channel recruitment but stems either from a defect in a key regulator of the channel, presumably a phosphoprotein, or from the hyperactivation of a channel closing mechanism, presumably a protein phosphatase or a down-regulating protein kinase (i.e. protein kinase C). In vitro phosphorylation of isolated intestinal brush border membranes has revealed the existence of a 25,000 molecular weight proteolipid (p25) acting as cosubstrate for both cGMP- and cAMP-dependent protein kinases and cross-reacting with antibodies directed against the cytoplasmic tail of the band 3 anion exchanger from erythrocytes. The putative role of p25 in Cl- channel regulation and its relationship to an unidentified GTP-binding protein recently implicated in Cl- channel activation is discussed on the basis of a regulatory model indicating potential sites of the CF defect at a molecular level.

Effect of Escherichia coli heat-stable enterotoxin on cell volume and intracellular inorganic ions of rat intestinal cells

1.--Electron micrographs of rat jejunum mucosa incubated for 1 h in the presence of Escheria coli heat-stable enterotoxin (STa) in the lumen shows alterations of villous cells as well as of crypt cells. The brush border of mature enterocytes is partially desintegrated and covered with a thick mucus. Crypts are occupied on half of their height by cells very similar to Paneth cells, loaded with numerous large dark inclusions. 2.--Cell volume and intracellular inorganic ion concentrations have been estimated in mucosal scrapings of jejunum sacs, incubated in vitro for 1 or 3 h. The quick action (1 h of incubation) of STa is a swelling of the intestinal calls accompanied by an increase in Na+, Cl- and Ca2+ intracellular concentrations and a decrease in the K+ and Mg2+ ones. The delayed action (3 h of incubation) is an increase of extracellular space and a decrease in cell volume; and at the same time the intracellular concentration of Na+, Cl-, K+, Ca2+ and Mg2+ is augmented. 3.--After 3 h of incubation intestinal cells from the other levels of intestine (duodenum, ileum and colon) show the same variations in cell volume and intracellular inorganic ion concentrations under the influence of STa, as those recorded in the jejunum. 4.--The present work favours the hypothesis that all intestinal cells, villous or cryptic, are involved in the alteration of fluid ion transport ending in diarrhea.

Demonstration of cGMP-dependent protein kinase and cGMP-dependent phosphorylation in cell-free extracts of platelets

Homogenates, membranes and cytosol of rat and human platelets were found to contain cGMP-dependent protein kinase immunoreactivity. Specific cGMP-dependent protein kinase immunoreactivity was about 1.7 pmol protein kinase/mg protein for homogenates of human platelets and 0.7 pmol/mg for homogenates of rat platelets; the majority appeared to be associated with the membrane fraction. In membranes of platelets low concentrations of cAMP (0.5-2 microM) stimulated the phosphorylation of five major proteins with apparent relative molecular masses, Mr, of 240 000, 130 000, 50 000, 42 000 and 22 000 while low concentrations of cGMP (0.5-2 microM) stimulated the phosphorylation of three major proteins with apparent Mr of 130 000, 50 000 and 46 000. An affinity-purified antibody against the cGMP-dependent protein kinase was prepared which specifically inhibited the activity of cGMP-dependent protein kinase. In membranes of human platelets this affinity-purified antibody inhibited the cGMP-stimulated phosphorylation of the three proteins with Mr of 130 000, 50 000 and 46 000 while it had no effect on the cAMP-dependent and cyclic-nucleotide-independent protein phosphorylation. The results demonstrate that platelets contain a cGMP-dependent protein kinase and at least three specific substrates for this enzyme. Two of these substrates, the proteins with apparent molecular Mr of 130 000 and 50 000, are substrates for both cAMP- and cGMP-dependent protein kinase. The protein with apparent Mr of 130 000 appears to be closely related to an intrinsic plasma membrane protein of vascular smooth muscle cells which is a substrate for a membrane-associated cGMP-dependent protein kinase. Therefore, cGMP-dependent protein kinase and cGMP-regulated phosphoproteins may mediate in platelets the intracellular effects of those hormones, vasodilators and drugs which elevate the level of cGMP and inhibit platelet aggregation.

Modulation of salt permeabilities of intestinal brush-border membrane vesicles by micromolar levels of internal calcium

A possible modulation of ion permeabilities of rat intestinal brush-border membrane vesicles by Ca2+, a putative second messenger of salt secretion, was explored by three independent methods: (1) measurements of [3H]glucose accumulation driven by a Na+ gradient; (2) stopped-flow spectrophotometry of salt-induced osmotic swelling; (3) 86Rb+, 22Na+ and 36Cl- flux measurements. Cytoskeleton-deprived membrane vesicles were prepared from isolated brushborders by thiocyanate treatment. Intravescicular Ca2+ levels were varied by preincubating vesicles in Ca-EGTA buffers in the presence of the Ca2+-ionophore A23187. At Ca2+free greater than 10(-5) M, initial Na+-dependent glucose uptake in the presence of a 0.1 M NaSCN gradient (but not in its absence) was inhibited by about 50 per cent as compared to EGTA alone (ED50 approximately equal to 10(-6) M Ca2+). By contrast, initial rates of 22Na+ uptake and reswelling rates of vesicles exposed to a NaSCN gradient were increased at least 2-fold by 10(-5) M Ca2+free. Both observations are compatible with a Ca2+-induced increase of the Na+-permeability of the vesicle membrane. The modulation of ion transport was fully reversible and critically dependent on internal Ca2+, suggesting a localization of Ca2+-sensor sites at the inner surface of the microvillous membrane. As shown by radiotracer and osmotic swelling measurements, micromolar Ca2+ additionally increased the flux rate of K+, Rb+, Cl- and NO-3 but did not change the membrane permeability for small uncharged molecules, including glucose and mannitol. The effect of Ca2+ on ion permeabilities could be blocked by Ba2+ (10(-3) M) or Mg2+ (10(-2) M), but not by amiloride (10(-3) M), apamin (2 X 10(-7) M), trifluoperazine (10(-4) M) or quinine (5 X 10(-4) M). At present it is unclear whether Ca2+ activates a nonselective cation and anion channel or multiple highly selective channels in the vesicle membrane.

Local changes in fractional saturation of cGMP- and cAMP-receptors in intestinal microvilli in response to cholera toxin and heat-stable Escherichia coli toxin

Cyclic nucleotide modulation of electrolyte transport across intestinal brushborder membranes is initiated by binding of cGMP and cAMP to high-affinity receptors at the interior of the microvilli. Previously these receptors have been identified by photoaffinity-labelling techniques as regulatory domains of cGMP- and cAMP-dependent protein kinases. In the present study, the receptor concentration in isolated brushborder membrane vesicles and their fractional saturation in absorptive and secretory states of the tissue were estimated. In microvillous membrane vesicles isolated from rat small intestine in the absorptive state, about 10% of the total number of cGMP receptors (25.5 pmol/mg protein) and 40% of all cAMP receptors (28.7 pmol/mg protein) were occupied by endogenous cyclic nucleotides. Luminal exposure of the intestinal segments in vivo to heat-stable Escherichia coli toxin for 3-5 min increased the occupancy of cGMP receptors by about 5-fold without affecting receptor-bound cAMP levels. In contrast, incubation with cholera toxin for 2 h increased the fractional saturation solely of cAMP receptors by 2-fold. Addition of heat-stable E. coli toxin to cholera toxin-pretreated segments, again raising the cGMP levels by 5-fold, did not reduce the amount of receptor-bound cAMP. This finding argues against the concept that increased levels of cAMP during cholera would mimick cGMP effects on ion transport by low-affinity binding to microvillar cGMP receptors. This analysis of local changes in cyclic nucleotide levels at the microvillous level might help to explore the mechanism of action of other secretagogues or antidiarrhoeal agents and to delineate a possible compartmentation of cGMP and cAMP pools within the intestinal mucosa responding differently to external signals.

Characterization of intestinal brush border guanylate cyclase activation by Escherichia coli heat-stable enterotoxin

Intestinal brush border guanylate cyclase was previously reported to be activated by the Escherichia coli enterotoxin (STa). This system was reexamined in order to develop a hypothesis for the mechanism of activation. The extent of activation was previously underestimated, since by using sodium azide to inhibit competing reactions and ethylene glycol bis(beta-aminoethyl ether) N,N-tetraacetic acid to chelate Ca2+, which is inhibitory, maximal activations of 30- to 50-fold were obtained. Ca2+ inhibition was only partially relieved by the calmodulin inhibitor calmidazolium. Inhibitors of the O2-dependent activation of soluble guanylate cyclase had no effect on STa activation; hence, it was concluded that STa activation did not involve arachidonate release and oxidation. STa was able to further increase activity already elevated by the nonionic detergent Lubrol PX. The membrane-active agent filipin, which was previously reported to inhibit both basal and agonist-stimulated adenylate cyclase, did not inhibit STa activation of guanylate cyclase. Digitonin, another cholesterol binder, inhibited STa activation at low concentrations, which disappeared at higher concentrations. Both of these agents stimulated basal activity. Dimethyl sulfoxide produced a concentration-dependent inhibition of STa activation, while increasing basal activity 7-fold. Ethanol inhibited both basal and STa-stimulated activity, with the former being more affected. Benzyl alcohol, like ethanol, a "fluidizer" of cell membranes, also inhibited both basal and activated enzymes. We concluded that STa directly activates this guanylate cyclase and, because of the differential effects of inhibitors on basal and STa-stimulated activity, propose a receptor-mediated mechanism.

Reversal of the biological activity of Escherichia coli heat-stable enterotoxin by disulfide-reducing agents

Various disulfide-reducing agents, mostly thiols and thiol precursors, were examined for their ability to reduce the disulfide bonds in the Escherichia coli heat-stable enterotoxin STa; reduction of the bonds results in loss of biological activity. The biological activity measured was the stimulation of guanylate cyclase in pig intestinal brush border membranes by STa. Nearly all of the compounds inactivated STa, although at different rates; a smaller number appreciably decreased guanylate cyclase activity when they were introduced into the reaction mixture after STa bound to its receptor. With dithiothreitol, the decrease in reaction rate was both time and concentration dependent and resulted in a reversal to basal activity. The anionic thiols were relatively ineffective in reversing activation, the neutral monothiols were moderately effective, and the aminothiols and neutral dithiols were the most effective. The order of effectiveness of the compounds was S-2-(3-aminopropylamino)ethanethiol greater than 2,3-dimercaptopropanol = 2-aminoethylisothiuronium bromide greater than dithiothreitol greater than 2-mercaptoethylamine greater than alpha-thioglycerol. These compounds were used in weanling pig ligated-intestinal-loop bioassays to determine if STa-induced secretion was reduced when they were injected 20 min after the STa. Instead of S-2-(3-aminopropylamino)ethanethiol we used the phosphorylated derivative S-2-(3-aminopropylamino)ethylphosphorothioic acid; this compound and 2,3-dimercaptopropanol were the only compounds that reduced STa-induced secretion and had no direct secretory or pathological effects.

Intestinal receptor for heat-stable enterotoxin of Escherichia coli is tightly coupled to a novel form of particulate guanylate cyclase

A novel form of particulate guanylate cyclase tightly coupled by cytoskeletal components to receptors for heat-stable enterotoxin (ST) produced by Escherichia coli can be found in membranes from rat intestinal mucosa. Intestinal particulate guanylate cyclase was resistant to solubilization with detergent alone, with only 30% of the total enzyme activity being extracted with Lubrol-PX. Under similar conditions, 70% of this enzyme was solubilized from rat lung membranes. The addition of high concentrations of sodium chloride to the extraction buffer resulted in greater solubilization of particulate guanylate cyclase from intestinal membranes. Although extraction of intestinal membranes with detergent and salt resulted in greater solubilization of guanylate cyclase, a small fraction of the enzyme activity remained associated with the particulate fraction. This activity was completely resistant to solubilization with a variety of detergents and chaotropes. Particulate guanylate cyclase and the ST receptor solubilized by detergent retained their abilities to produce cyclic GMP and bind ST, respectively. However, ST failed to activate particulate guanylate cyclase in detergent extracts. In contrast, guanylate cyclase resistant to solubilization remained functional and coupled to the ST receptor since enzyme activation by ST was unaffected by various extraction procedures. The possibility that the ST receptor and particulate guanylate cyclase were the same molecule was explored. ST binding and cyclic GMP production were separated by affinity chromatography on GTP-agarose. Similarly, guanylate cyclase migrated as a 300,000-dalton protein, while the ST receptor migrated as a 240,000-dalton protein on gel filtration chromatography. Also, thiol-reactive agents such as cystamine and N-ethylmaleimide inhibited guanylate cyclase activation by ST, with no effect on receptor binding of ST. These data suggest that guanylate cyclase and the ST receptor are independent proteins coupled by cytoskeletal components in membranes of intestinal mucosa.