It has previously been shown that thrombin effects on endothelial cells can be mediated via G-proteins, which couple the thrombin receptor to several key physiological responses. As G-proteins are known targets of bacterial toxins, specific toxins were used to further characterize G-protein involvement in thrombin activation of bovine pulmonary arterial endothelial cells (6PAEC) and human umbilical vein endothelial cells (HUVEC). Homogenates were exposed to several bacterial toxins in the presence of 32P-N AD and ADP ribosylation of proteins determined by autoradiography of SDS-PAGE gels. Major substrates were a 40 kDa protein for pertussis toxin, 39, 45 and 52 kDa proteins (G5) for cholera toxin, a 21 kDa protein for botulinum toxin C, and a 43 kDa protein (actin) for botulinum toxin C2α. The increase in either HUVEC or BPAEC PGI2 release induced by thrombin was not altered by pretreatment with any toxin. However, 1 h treatment of BPAEC monolayers with 1 μg/ml pertussis toxin resulted in dramatic barrier dysfunction, which was synergistic with the albumin permeability induced by 1 μM thrombin. In contrast, pretreatment with 1 μg/ml cholera toxin completely prevented the thrombin-induced barrier dysfunction. Moreover, contraction and gap formation due to thrombin challenge, observed by phase contrast microscopy, was greatly augmented by pertussis toxin and prevented by cholera toxin. Whereas 5 μg/ml botulinum toxin C did not affect either basal or thrombin-induced barrier dysfunction, botulinum toxin C2a increased basal BPAEC permeability over four-fold. Thus, bacterial toxins have specific and divergent effects on thrombin-induced endothelial cell responses. Botulinum toxin C2a appears to interact directly with actin to produce barrier dysfunction. In contrast, cholera toxin promotes barrier function via its known effects on Gs, stimulating adenylate cyclase and increasing cAMP. Because cholera toxin and pertussis toxin (via inhibition of Gi) both increase cAMP, yet have opposing effects on barrier function, the present results suggest that pertussis toxin produces barrier dysfunction via ADP ribosylation of a novel G-protein other than Gi or via a novel action of Gi.