B Subunits of Shiga Toxin 2 Induce VWF Secretion by a cAMP-Independent Protein Kinase A Signaling Pathway.

Abstract 3046

Poster Board II-1022

Shiga toxin (Stx) is composed of an A subunit that inactivates ribosomes and 5 B subunits that bind globotriaosylceramice (Gb3) on cell surfaces to facilitate internalization and retrograde transport of the toxin to the cytosol. Stx-producing E. coli cause hemolytic uremic syndrome (HUS) by damaging renal microvascular endothelium. In animal experiments, Stx causes fatal thrombotic microangiopathy in ADAMTS13-deficient mice by inducing the secretion of von Willebrand factor (VWF). Using cultured human umbilical vein endothelial cells (HUVECs), we have shown that the B subunits of either Stx1 or Stx2 are sufficient to stimulate VWF secretion, independent of the catalytically-active A subunit. HUVECs were perfused in a parallel plate flow chamber with Stx preparations and fluorescently-labeled anti-VWF, and secreted VWF strings were visualized in real time by immunofluorescence microscopy. Either Stx1B or Stx2B rapidly induced the secretion of long strings of VWF that attached to the cell surface and bound platelets with high affinity. VWF string formation was maximal within 5 min and was blocked by soluble analogs of Gb3 or anti-Stx B subunit antibodies. Stx1B subunits activated phospholipase C (PLC), increased intracellular Ca²⁺, and activated protein kinase C (PKC). Stx2B is homologous to Stx1B, but their amino acid sequences differ by 34% and Stx2B activates distinct signaling pathways. For example, pretreatment of HUVECs with a chelator of intracellular Ca²⁺ (0.1 mM BAPTA-AM, 30 min), Ca²⁺-free medium (pH 7.2 PBS, 30 min), a phospholipase C (PLC) inhibitor (5 μM U73122, 15 min) or a protein kinase C inhibitor (5 μM 19-31, 1 h) blocked VWF secretion in response to Stx1B, but not Stx2B. The role of intracellular Ca²⁺ was also assessed using a fluorescent BD Calcium Assay. Stx1B (5 nM), histamine (0.1 mM) or thrombin (1 U/ml) induced a similar transient rise in intracellular Ca²⁺ level that peaked by 30 sec and declined to baseline over 5 min, whereas Stx2B did not affect the intracellular Ca²⁺ level. Conversely, pretreatment with a protein kinase A inhibitor (5 μM H89, 30 min) had no effect on Stx1B-induced VWF string formation, but decreased Stx2B-induced VWF string formation more than 90%. Neither Stx1B nor Stx2B increased intracellular cAMP in cell extracts within 5 min. To more directly detect cAMP production, HUVECs were transfected with a CFP-Epac-YFP construct that exhibits decreased FRET upon binding cAMP, which is detected by confocal microscopy. Forskolin caused the expected increase in CFP/YPF ratio, indicating a rise in cAMP, whereas Stx1B and Stx2B did not. Despite the absence of a change in cAMP, Stx2B caused an increase in PKA activity comparable to that induced by forskolin, when assayed in cell extracts with a synthetic peptide substrate (Assay Designs) kit. In addition, treatment of HUVECs with Stx2B increased the phosphorylation of the endogenous PKA substrate, 14-3-3ζ. These results indicate that Stx2B subunits stimulate VWF secretion through a previously unsuspected signaling pathway that involves binding to cell surface Gb3 and cAMP-independent activation of PKA. Therefore, Stx2 may contribute to the pathogenesis of HUS by at least two mechanisms that affect microvascular endothelium: Stx2A subunit-dependent inhibition of protein synthesis, and Stx2B subunit-dependent induction of VWF secretion.