ADP-ribosylation Of Membrane Proteins Research Articles

Evidence for the presence of G-proteins, adenylyl cyclase and phospholipase C activities in lymphatic smooth muscle cell membranes

In plasma membranes derived from bovine mesenteric lymphatic smooth muscle cells, guanine nucleotide and forskolin stimulated adenylyl cyclase (AC) activity in a concentration-dependent manner, indicative of the presence of the stimulatory G-protein G, linked to AC. There was no significant enzyme inhibition by low concentrations of guanine nucleotide and no effect on basal or guanine nucleotide-stimulated activity following pertussis toxin treatment of cells, suggesting the absence of G, linked to inhibition of AC. Furthermore, there was no effect of adrenaline, isoprenaline or clonidine on basal or forskolin-stimulated activities, nor was there any specific binding of the β-adrenoceptor ligand [ 125I]cyanopindolol to membranes, suggesting that catecholamine receptors do not modulate AC activity in these membranes. Pertussis toxin-mediated ADP ribosylation of membrane proteins and Western immunoblotting analysis revealed the presence of G-protein subunits of G αi2, G αq, G α11 and G β1. In experiments designed to identify a possible effector enzyme for these G-proteins, membranes were screened with a range of antibodies raised against phospholipase C (PLC) β, γ and δ isozymes. Though no evidence was obtained by Western blotting for any of these proteins, PLC activity was concentration-dependently stimulated by Ca 2+, but not by AlF 4-, GTP[S], or purified G βγ subunits. Finally, no specific binding to membranes of the α 1-adrenoceptor ligand [ 3H]prazosin or the α 2-adrenoceptor ligand [ 3H]yohimbine was obtained. In conclusion, this study provides evidence for a G s-dependent stimulation of AC, and for the presence of G i2 and G q/11, which do not appear to regulate a PLC activity also identified in lymphatic smooth muscle cell membranes. Furthermore, neither AC nor PLC appear to be associated with catecholamine receptors.

Molecular Characterization of a Glycosylphosphatidylinositol-Linked ADP-Ribosyltransferase From Lymphocytes

Mono ADP-ribosyltransferases catalyze the transfer of the ADP-ribose moiety of nicotinamide adenine dinucleotide (NAD) to proteins. It was reported by Wang et al (J Immunol 153:4048, 1994) that incubation of mouse cytotoxic T lymphocytes (CTL) with NAD resulted in the ADP-ribosylation of membrane proteins and inhibition of cell proliferation and cytotoxicity. Treatment of CTL with phosphatidylinositol-specific phospholipase C (PI-PLC) before incubation with NAD prevented the inhibitory effects of NAD on the cells, consistent with the removal of a glycosylphosphatidylinositol (GPI)-anchored ADP-ribosyltransferase on the lymphocyte surface. We have identified and cloned a GPI-linked ADP-ribosyltransferase from Yac-1 mouse T-cell lymphoma cells. The deduced amino acid sequence of the Yac-1 transferase was 70% and 41% identical to those of the rabbit skeletal muscle and chicken heterophil, respectively. It contained three noncontiguous sequences similar to those found in several of the bacterial toxin and vertebrate ADP-ribosyltransferases. Based on crystallography of the bacterial toxins, these regions are believed to form, in part, the catalytic site consistent with a common mechanism for the ADP-ribose transfer reaction. In rat mammary adenocarcinoma (NMU) cells transformed with the Yac-1 transferase cDNA, transferase activity was present on the cell surface and was released into the medium by treatment of cells with PI-PLC. Thus, we have cloned a novel gene that has properties identical to the transferase detected in CTL, and may be involved in the NAD-dependent regulation of proliferation and cytotoxicity.

Ontogeny of guanine-nucleotide-binding regulatory proteins in rabbit liver

Ontogeny of trimeric GTP-binding regulatory proteins (G-proteins) and their subunits in rabbit liver during neonatal development was studied, by using bacterial-toxin-catalysed ADP-ribosylation of membrane proteins, immunoblot analysis to quantify the alpha-subunit (alpha s and alpha i) of stimulatory (Gs) and inhibitory (Gi) G-protein and the beta-subunit, and reconstitution assay with cyc- membranes (from Gs-deficient variant of S49 lymphoma cell) to measure Gs activity. Under optimal conditions of ADP-ribosylation, little cholera-toxin substrate (alpha s) was detected in membranes from liver of neonatal animals up to 24 h of age. Thereafter ribosylatable alpha s proteins, i.e. 45 kDa (alpha s-1) and 52 kDa (alpha s-2) proteins, were increasingly evident, reaching maximal levels in membranes from animals aged 4-6 weeks. The concentrations of alpha s-1 and alpha s-2, as determined by immunoblotting, were 6.1 +/- 0.8 and 2.7 +/- 0.4 pmol/mg of protein respectively at birth, and did not change during 0-24 h after birth. Thereafter they gradually increased to maximal levels of 22.1 +/- 1.3 and 10.5 +/- 0.7 pmol/mg of protein for alpha s-1 and alpha s-2 respectively, within 6 weeks. The beta-subunit also showed a similar 3-4-fold increase during the same age span. In contrast, the pertussis-toxin substrate (alpha i) was clearly evident even in membranes from term animals and in all age groups studied. Its developmental pattern, as assessed by ADP-ribosylation, was the same as that determined by immunoblot analysis. The functional activity of Gs in cholate extracts of membranes exhibited similar developmental pattern to that of cholera-toxin-mediated labelling. This activity also paralleled the concentrations of alpha s as measured by immunoblotting. These results suggest differential expression of G-protein subunits in liver during neonatal development.

Pertussis toxin abolishes the effect of neuropeptide Y on rat resistance arteriole contraction

Pertussis toxin (PTX) was used to investigate the possible involvement of a G protein in the mechanism of action of neuropeptide Y (NPY) on rat mesenteric arterioles. ADP ribosylation of membrane protein was assessed by gel electrophoresis followed by autoradiography. The effect of NPY was studied on contractions elicited either by addition of calcium to depolarized vessels or by the calcium agonist BAY K 8644. Under conditions in which PTX ADP-ribosylated a 40-41 kDa protein, the potentiation of contractions induced by NPY in both protocols was abolished. In contrast, the contraction induced by norepinephrine was unaffected by PTX. It is concluded that a G protein mediates the potentiating effect of NPY on external calcium-dependent contractions in rat mesenteric arterioles.

Low-Molecular-Weight GTP-Binding Proteins Serving as ADP-Ribosylation Substrate for ADP-Ribosyltransferase from Clostridium botulinum and Their Relation to Phosphoinositides Metabolism in Thymocytes1

ADP-ribosyltransferase from Clostridium botulinum type C strain was found to induce an increase of inositol phosphates (IPs) formation in murine thymocytes membranes. Incubation of electropermeabilized murine thymocytes with the enzyme also caused an increase of IPs formation in the cells. This increase of IPs formation in the enzyme-treated membranes and electropermeabilized cells was dependent on the amount of both NAD and the enzyme, suggesting that the stimulation of phosphoinositide-specific phospholipase C (PLC) was related to ADP-ribosylation of membrane proteins by the enzyme. On the other hand, in calf and murine thymocytes two proteins with the same molecular weight of 21,000 were found to be ADP-ribosylated by the botulinum ADP-ribosyltransferase. A minor ADP-ribosylation substrate was shown by two-dimensional polyacrylamide gel electrophoresis to be G21k, a low-molecular-weight GTP-binding protein (G protein) suggested previously by us to be involved in PLC regulation [Wang, P. et al. (1987) J. Biochem. 102, 1275-1287; (1988) 103, 137-142; and (1989) 105, 461-466], and the other major ADP-ribosylation substrate was identified as a rho A protein. Under the experimental conditions of the IPs formation study, ADP-ribosylation of both G21k and rho A proteins by botulinum ADP-ribosyltransferase in membranes and permeabilized cells was observed. These results suggest that botulinum ADP-ribosyltransferase-induced PLC stimulation in thymocytes is closely correlated with ADP-ribosylation of the low-molecular-weight G proteins.

ADP-ribosylation of cell membrane proteins by staphylococcal α-toxin and leukocidin in rabbit erythrocytes and polymorphonuclear leukocytes

Staphylococcal α-toxin resulted in ADP-ribosylation of the 37 and 41 kDa proteins of a membrane preparation from rabbit erythrocytes. In the presence of 100 μM GTP, the toxin ADP-ribosylated proteins of 54 and 59 kDa and potentiated ADP-ribosylation of the 37 and 41 kDa forms. GTP had no effect on ADP-ribosylation of membrane proteins in the absence of α-toxin. Incubation of a membrane preparation of rabbit polymorphonuclear leukocytes with the S and F components of staphylococcal leukocidin resulted in ADP-ribosylation of the 37 and 41 kDa proteins, respectively. Furthermore, the 37, 41, 54 and 59 kDa proteins were ADP-ribosylated by leukocidin in the presence of GTP. The ADP-ribosylation of these proteins was observed to be dependent on the incubation time and toxin dose and was abolished by prior boiling. Addition of agmatine did not attenuate ADP-ribosylation of these proteins. These results demonstrate that staphylococcal α-toxin and leukocidin possess ADP-ribosyltransferase activities which are potentiated by GTP and suggest that ADP-ribosylation reactions are responsible for development of the cytolytic activities of these staphylococcal toxins.

ADP-ribosylation of membrane proteins by bacterial toxins in the presence of NAD glycohydrolase

The ADP-ribosylation of membrane G proteins is difficult to achieve in tissues that are rich in membranebound NAD glycohydrolase (NAD + glycohydrolase, EC 3.2.2.5). For many animal species this problem can be surmounted by inhibiting NAD hydrolysis with a combination of the anti-tuberculous drug, isonicotinic acid hydrazide, and the NAD analog, 3-acetylpyridine adenine dinucleotide, which act synergistically. In their presence, the ADP-ribosylation of cholera and pertussis toxin substrates reach plateau levels even with only 10 μM NAD. Although 3-acetylpyridine adenine dinucleotide acts as a weak substrate for the toxins, it is simple to estimate its contribution to the ADP-ribosylation and thus to determine the total amount of ADP-ribosylation substrate present in a tissue sample. NAD glycohydrolases that are insensitive to isonicotinic acid hydrazide are also less sensitive to 3-acetylpyridine adenine dinucleotide, but may be inactivated by dithiothreitol. Isonicotinic acid hydrazide adenine dinucleotide, the product of an exchange reaction eatalysed by NAD glycohydrolase, runs with NAD in most thin-layer chromatographic systems. It can be separated from NAD, and quantitated if the chromatographic solvent contains benzaldehyde. Isonicotinic acid hydrazide itself inhibits NAD glycohydrolase. It need not first be converted into isonicotinic acid hydrazide adenine dinucleotide.

ADP-ribosylation of a 21-24 kDa eukaryotic protein(s) by C3, a novel botulinum ADP-ribosyltransferase, is regulated by guanine nucleotide.

Besides botulinum C2 toxin, Clostridium botulinum type C produces another ADP-ribosyltransferase, which we termed 'C3'. ADP-ribosyltransferase C3 has a molecular mass of 25 kDa and modifies 21-24 kDa protein(s) in platelet and brain membranes. C3 was about 1000 times more potent than botulinum C1 toxin in ADP-ribosylation of membrane proteins. C3-catalysed ADP-ribosylation of the 21-24 kDa protein(s) was decreased by stable guanosine triphosphates, with the potency order GTP[S] much greater than p[NH]ppG greater than p[CH2]ppG. GTP[S] inhibited the ADP-ribosylation caused by C3 by maximally 70-80%, with half-maximal and maximal effects occurring at 0.3 and 10 microM-GTP[S] respectively. The concomitant addition of GTP decreased the inhibitory effect of GTP[S]. GTP[S]-induced inhibition of ADP-ribosylation was resistant to washing of pretreated platelet membranes. The data suggest that the novel botulinum ADP-ribosyltransferase C3 modifies eukaryotic 21-24 kDa guanine nucleotide-binding protein(s).

Type II heat-labile enterotoxin of Escherichia coli activates adenylate cyclase in human fibroblasts by ADP ribosylation.

Type II heat-labile enterotoxin (LT-II) from Escherichia coli causes characteristic morphological changes and accumulation of cyclic AMP in Y-1 adrenal cells, but it is not neutralized by antisera against choleragen (CT) or the classical type I heat-labile enterotoxin (LT-1) from E. coli. The action of purified LT-II on CT- and LT-I-responsive human fibroblasts was investigated and compared with that of CT. Fibroblasts incubated with LT-II or CT had an increased cyclic AMP content as well as a fourfold elevation of membrane adenylate cyclase activity. In membranes, activation of cyclase by toxin was enhanced by NAD, GTP, and dithiothreitol. The effect of LT-II on intact fibroblasts or membranes was increased by trypsin treatment of toxin. Since activation of adenylate cyclase by LT-II was stimulated by NAD, the ability of LT-II to catalyze the [32P]ADP-ribosylation of membrane proteins in the presence of [32P]NAD from control and LT-II- and CT-treated fibroblasts was investigated. Similar proteins were [32P]ADP-ribosylated in membranes exposed to LT-II or CT; LT-II- and CT-specific labeling was significantly decreased in membranes prepared from cells preincubated with either LT-II or CT. These studies are consistent with the hypothesis that LT-II, similar to CT and LT-I, increases cyclic AMP by activating adenylate cyclase through the GTP-dependent ADP-ribosylation of specific membrane proteins.

Incubation of bovine thyroid slices with thyrotropin is associated with a decrease in the ability of pertussis toxin to adenosine diphosphate-ribosylate guanine nucleotide regulatory component(s).

Pretreatment of bovine thyroid slices with TSH resulted in desensitization of TSH-sensitive adenylyl cyclase activity but no change in stimulatory nucleotide binding regulatory component of adenylyl cyclase (Gs) activity assessed by reconstitution of the Gs-defective cyc-S49 adenylyl cyclase system. Possible changes in substrates for pertussis toxin (PT)-induced ADP ribosylation due to TSH treatment and/or in endogenous ADP ribosylation of membrane proteins were explored. Using 10 microM [32P]NAD+ as substrate, endogenous ADP ribosylation was not observed in membranes from control or TSH-treated slices. ADP ribosylation of alpha-subunits of Gs by cholera toxin was also unaffected by incubation of thyroid slices with TSH. In contrast, ADP ribosylation of 40 kilodalton (kDa) substrates for PT was decreased between 40% and 60% by TSH treatment. This effect of TSH was dependent on its concentration and the time of incubation of the slices and was specific for labeling of the 40 kDa PT substrate. Prostaglandin E1 treatment of thyroid slices, which results in a much smaller homologous desensitizing effect, did not result in changes in ADP ribosylation by PT. The effect of incubation of slices with TSH was abolished by pretreatment of the membranes with 0.3-1.0% Lubrol PX, which increased the labeling of the 40 kDa polypeptides. The data suggests that TSH induces in thyroid tissue a redistribution of 40 kDa polypeptides changing their availability to PT.

The activation of adenylate cyclase from small intestinal epithelium by cholera toxin.

ADP-ribosylation of membrane proteins from rabbit small intestinal epithelium was investigated following incubation of membranes with [32P]NAD and cholera toxin. Cholera toxin catalyzes incorporation of 32P into three proteins of 40 kDA, 45 kDa and 47 kDa located in the brush-border membrane. In contrast, basal lateral membranes do not contain any protein which becomes labeled in a toxin-dependent manner when incubated with cholera toxin and [32P]NAD. The modification of membrane proteins from brush border occurred in spite of the virtual absence in these membranes of adenylate cyclase activatable either by cholera toxin, vasoactive intestinal peptide (VIP) or fluoride. The three agents activated adenylate cyclase when crude plasma membrane were used. Cholera toxin activated fivefold at 10 micrograms/ml. Vasoactive intestinal peptide activated at concentrations from 10-300 nM, the maximal stimulation being sixfold. Fluoride activated 10-fold at 10 mM. When basal lateral membranes were assayed for adenylate cyclase it was found that, with respect to the crude membranes, the specific activity of fluoride-activated enzyme was 3.3-fold higher, VIP stimulated enzyme was maintained while cholera-toxin-stimulated enzyme showed half specific activity. Moreover, while fluoride stimulated ninefold and VIP stimulated fivefold, cholera toxin only stimulated twofold at the highest concentration. The results suggest that the activation by cholera toxin of adenylate cyclase located at the basal lateral membrane requires ADPribosylation of proteins in the brush border membrane.

Effect of Bordetella pertussis toxin on ADP-ribosylation of membrane proteins, adenylate cyclase activity and insulin release in rat pancreatic islets

Exposure of rat pancreatic islet membranes to [α- 32P]-NAD + in the presence of Bordetella Pertussis toxin (islet-activating protein) reveals the ADP-ribosylation of a peptide with a Mr close to 41 kDa, which corresponds to the α-subunit of the guanine nucleotide regulatory protein Ni. Islets removed from rats pretreated with the Bordetella Pertussis toxin display a specific increase in adenylate cyclase responsiveness to GTP and are characterized by a resistance to the inhibitory action of α 2-adrenergic agonists upon either adenylate cyclase activity or glucose-induced insulin release.

Intravesicular NAD has no effect on sodium-dependent phosphate transport in isolated renal brush border membrane vesicles.

The effects of intravesicular NAD on Na+-dependent 32Pi uptake were investigated in isolated rat kidney brush border membrane vesicles (BBMV). NAD was introduced into the vesicles by osmotic shock, and extravesicular NAD was removed by passing the vesicles through a anion exchange column. The effectiveness of the osmotic shock procedure and the hydrolysis of extra- and intravesicular NAD were controlled by enzymatic analysis and thin layer chromatography. ADP-ribosylation of the membrane proteins was analyzed in vesicles osmotically shocked in the presence of either [adenylate-32P]-NAD or [adenine-2,8-3H]-NAD by SDS-polyacrylamide gel electrophoresis. It was found that the Na+-dependent Pi uptake was inhibited when the BBMV were incubated with NAD at alkaline pH, which resulted in rapid NAD hydrolysis. When NAD was present in the intravesicular space only, the Na+-dependent Pi uptake was not inhibited. 32P from NAD was rapidly incorporated into a number of brush border membrane proteins, but no incorporation of 3H-adenine could be detected. The results provide evidence that NAD does not inhibit Pi transport by a direct interaction with the cytoplasmic side of the brush border membrane. No evidence of ADP-ribosylation of the brush border membrane protein(s) was found.

NAD +-induced inhibition of phosphate transport in canine renal brush-border membranes Mediation through a process other than or in addition to NAD + hydrolysis

We previously demonstrated inhibition of Na +-dependent 32P i transport in canine renal brush-border membranes in association with NAD +-induced ADP ribosylation of membrane protein(s) and postulated that NAD + inhibits P i transport across the brush-border membrane via ADP ribosylation. Recently it was shown that incubation of rat brush-border membrane with NAD + resulted in release of P i which was prevented by EDTA. It was proposed that NAD +-mediated inhibition of 32P i transport might occur through this mechanism. To determine whether NAD + inhibited 32P i transport by a mechanism other than or in addition to release of P i, we compared Na +-dependent 32P i counterflow in brush-border membrane equilibrated with P i or with P i generated from NAD +. Release of P i from NAD + incubated with brush-border membrane was confirmed. The increased uptake of 32P i which was demonstrated in brush-border membrane equilibrated with P i was not measured when intravesicular P i was generated from a concentration of NAD + which effected ADP-ribosylation of brush border membranes (100 μM NAD +). In contrast, increased uptake of 32P i was demonstrated when intravesicular P i was generated from 1 μM NAD + which did not effect ADP ribosylation. Mg 2+-dependent ADP ribosylation of brush-border membrane incubated with NAD + was demonstrated which persisted during the time interval of 32P i uptake measurements. Our findings are compatible with the hypothesis that NAD +-induced ADP ribosylation of brush-border membrane protein(s) results in inhibition of P i transport across the membrane in vivo. EDTA may act to prevent this inhibition in brush-border membrane by chelation of Mg 2+ and decreased ADP ribosylation.

Mechanism of action of glycopeptide hormones and cholera toxin: what is the role of ADP-ribosylation?

The cultured murine Leydig tumor cell line MLTC-1 and the normal rat thyroid strain FRTL have adenylate cyclase activities that are stimulated by human chorionic gonadotropin (hCG) and thyrotropin, respectively. Both cell types also respond to choleragen. Activation of adenylate cyclase in membranes by choleragen required NAD whereas stimulation of the enzyme by hormones did not. With [alpha-32P]NAD as a donor, ADP-ribosylation of membrane proteins was determined under the same conditions used to assay adenylate cyclase activity. Under these conditions, choleragen, but not the hormones, caused the ADP-ribosylation of subunits of the regulatory component (G/F) of adenylate cyclase in both FRTL and MLTC-1 membranes. In the absence of any effectors, several membrane proteins became labeled but the hormones did not cause the specific labeling of these or any other membrane proteins. Pretreatment of intact MLTC-1 cells with hCG did not block the ability of choleragen to ADP-ribosylate G/F in isolated membranes; labeling was actually enhanced in a manner related to the length of exposure to hCG. In contrast, pretreatment of the cells with choleragen inhibited ADP-ribosylation of G/F by the toxin in isolated membranes. Extracts of membranes from untreated, hCG-treated, and choleragen-treated MLTC-1 cells were used to reconstitute adenylate cyclase activity in membranes from the cyc- variant of S49 lymphoma cells which lacks a functional G/F. Toxin but not hormone treatment caused an increase in the basal activity of adenylate cyclase in the reconstituted system. Our results indicate that ADP-ribosylation of the regulatory component of adenylate cyclase is required for choleragen action but not for hormone action.

ADP ribosylation of membrane proteins from human fibroblasts. Effect of prior exposure of cells to choleragen.

Major labeled bands of Mr = 42,000 and 47,000 were observed on sodium dodecyl sulfate polyacrylamide gel electrophoresis of human skin fibroblast membranes incubated with choleragen and [32P]NAD. Prior incubation of intact fibroblasts with choleragen blocked specifically the subsequent in vitro labeling of these two proteins. The effect of choleragen was dependent on time, temperature, and toxin concentration. Neither the choleragen A subunit nor the B subunit nor the A1 peptide could replace the holotoxin in cell incubations. Inhibition of subsequent in vitro labeling by prior exposure of cells to choleragen was correlated with increased cellular cAMP. Incubation of fibroblasts with prostaglandin E1 and isoproterenol, which activate adenylate cyclase by different mechanisms, did not block subsequent labeling with choleragen and [32P]NAD. The results suggest that proteins of Mr = 42,000 and 47,000 may be in vivo substrates for choleragen in human fibroblasts.

ADP-ribosylation of membrane proteins and activation of adenylate cyclase by cholera toxin in fat cell ghosts from euthyroid and hypothyroid rats

Incubation of fat cell ghosts with activated cholera toxin, nucleoside triphosphate, cytosol, and NAD results in increased adenylate cyclase activity and the transfer of ADP-ribose to membrane proteins. The major ADP-ribose protein comigrates on sodium dodecyl sulfate-polyacrylamide gels with the putative GTP-binding protein of pigeon erythrocyte membranes ( M r 42 000), which is also ADP-ribosylated by cholera toxin. The treatment with cholera toxin enhances the stimulation of the fat cell membrane adenylate cyclase by GTP, but the stimulation by guanyl-5′-yl imidodiphosphate is unaltered. Subsequent stimulation of fat cell adenylate cyclase by 10 μM epinephrine is not particularly affected. These changes were qualitatively the same for membranes isolated from fat cells of hypothyroid rats. Although the cyclase of these membranes has a reduced response to epinephrine, guanyl-5′-yl imidodiphosphate or GTP, as compared to euthyroid rat fat cell membranes, the defect is not rectified by toxin treatment and cannot be explained by a deficiency in the cholera toxin target.