Binding of sea anemone pore-forming toxins sticholysins I and II to interfaces--modulation of conformation and activity, and lipid-protein interaction

Abstract

Sticholysins I and II (St I and St II) are water-soluble toxins produced by the sea anemone Stichodactyla helianthus. St I and St II bind to biological and model membranes containing sphingomyelin (SM), forming oligomeric pores that lead to leakage of internal contents. Here we describe functional and structural studies of the toxins aiming at the understanding at a molecular level of their mechanism of binding, as well as their effects on membrane permeabilization. St I and St II caused potassium leakage from red blood cells and temperature-dependent hemolysis, the activation energy of the process being lower for the latter toxin. Protein intrinsic fluorescence measurements provided evidence for toxin binding to model membranes composed of 1:1 (mol:mol) egg phosphatidyl choline (ePC):SM. The fluorescence intensity increased and the maximum emission wavelength decreased as a result of binding. The changes were quantitatively different for both toxins. Circular dichroism spectra showed that both St I and St II exhibit a high content of beta-sheet structure and that binding to model membranes did not alter the toxin's conformation to a large extent. Changing the lipid composition by adding 5 mol% of negatively charged phosphatidic acid (PA) or phosphatidyl glycerol (PG) had small, but detectable, effects on protein conformation. The influence of lipid composition on toxin-induced membrane permeabilization was assessed by means of fluorescence measurements of calcein leakage. The effect was larger for ePC:SM bilayers containing 5 mol% of negative curvature-inducing lipids. Electron paramagnetic resonance (EPR) spectra of intercalated fatty acid spin probes carrying the nitroxide moiety at different carbons (5, 7, 12, and 16) evidenced the occurrence of lipid-protein interaction. Upon addition of the toxins, two-component spectra were observed for the probe labeled at C-12. The broader component, corresponding to a population of strongly immobilized spin probes, was ascribed to boundary lipid. The contribution of this component to the total spectrum was larger for St II than for St I. Moreover, it was clearly detectable for the C-12-labeled probe, but it was absent when the label was at C-16, indicating a lack of lipid-protein interaction close to the lipid terminal methyl group. This effect could be either due to the fact that the toxins do not span the whole bilayer thickness or to the formation of a toroidal pore leading to the preferential interaction with acyl chain carbons closer to the phospholipids head groups.