The formation of pores by streptolysin O (SLO) was analyzed in erythrocyte membranes and liposomes by immunoelectron microscopy and electron spectroscopic imaging. The binding of SLO molecules to membranes was temperature independent, while the polymerization of SLO molecules was temperature dependent. Our results also suggest that proteins in erythrocyte membranes are not involved in the formation of SLO rings.Streptolysin O (SLO) is a membrane-damaging toxic protein that is produced by many strains of group A, C, and G streptococci (2, 16). Both ring-shaped and arc-shaped structures can be observed on erythrocyte membranes that have been treated with 9,14).In this study, we evaluated a two-step theory for the formation of SLO rings and pores by SLO on membranes (7, 11) and we confirmed that SLO rings are composed exclusively of the components of SLO by immunoelectron microscopy and electron spectroscopic imaging with an imaging plate (IP).SLO was prepared from the culture supernatant of group A streptococci (Streptococcus pyogenes Richard strain) as described previously (14). Ghost membranes were prepared from rabbit erythrocytes and stored at Ϫ80ЊC in physiological saline (Ohtsuka Pharmaceutical Co., Tokyo, Japan) that contained 1% albumin and 10% sucrose. Specimens were treated with SLO in one of several ways, fixed with 2.5% glutaraldehyde for 1 min, negatively stained with 2% phosphotungstic acid, and observed under a transmission electron microscope (H-500; Hitachi Co., Tokyo, Japan).The formation of a pore begins when an SLO molecule binds to a biological membrane. To examine this step, we applied SLO to frozen and stored erythrocyte ghost membranes at 0ЊC for 3 to 5 min. No ring-shaped structures were observed (Fig. 1). However, the negative staining of membranes that had been treated with SLO at room temperature for 3 to 5 min revealed numerous ring-shaped structures with a crown (not shown), as described previously (14). The frozen ghost membranes provided a reproducible material for ultrastructural analysis of pore formation by SLO.After the reaction with SLO at 0ЊC, we fixed specimens with 1% formaldehyde in physiological saline at 0ЊC for 3 min, and then we performed immunolabeling of ghost membranes with SLO-specific goat antiserum (Nissui Pharmaceutical Co., Tokyo, Japan) and then with 5-nm colloidal gold-conjugated antibodies against goat immunoglobulin G (Zymed Laboratories Inc., San Francisco, Calif.) at room temperature for 20 min. After rinsing the membrane and fixing with glutaraldehyde as described above, we found a number of colloidal gold particles on the membranes but no ring-shaped structures or pores (Fig. 2).In an attempt to explain why rings were not seen when erythrocyte membranes were treated with SLO at 0ЊC, Duncan and Schlegel (6) suggested that adsorption of SLO to membranes at 0ЊC was insufficient for ring formation. The proposed model of an SLO molecule has a hydrophobic surface (14) which might be unstable unless some surface-protective agent is present, such as albumin or ...