Complement-induced ultrastructural membrane lesions: requirement for terminal components

The step in the complement (C) sequence at which 8- to 11-nm ring-shaped lesions are formed on antibody-coated erythrocytes (EA) has remained controversial. Some workers have concluded that these lesions appear at the C5 step and are not ultrastructural correlates of lysis; others hold that these lesions are formed only after the action of C8 and C9 in association with lysis. We have re-examined this problem by using sheep EA and human sera genetically lacking C5, C6, C7, or C8. Electron micrographs of negatively stained membranes (x 220,000) were read in blind fashion and the results correlated with 125I-C5 binding. Rare structures resemblind C-induced ring lesions were found on EA exposed to C5-deficient (C5D), C6D, C7D and C8D sera or to heated normal serum, with no significant differences among these sera (lesion density 0 to 0.26/mum2). Fresh normal serum (NHS) produced 140 to 220 ring lesions/mum2. C5 binding to EA in C8D serum was 60% of that observed in an NHS control; in C6D and C7D sera C5 binding was 4 to 11% of the normal value. Iodine treatment of sera (to enhance C5 uptake by C2 oxidation) increased C5 binding in C6D serum to 40 to 65% of that seen in native NHS; in iodine-treated C7D and C8D sera C5 binding was 250 and 440%, respectively, of the native NHS value. No increase in ring lesions was observed, however, except in the iodine-treated NHS. Thus, in whole serum, C5 binding is not sufficient to produce ultrastructural membrane rings in the absence of later-acting C components, at least through C8. The formation of ring lesions appears to have C requirements similar to those necessary for lysis.