The complement requirements for killing of a rough, serum susceptible strain of E. coli (J5) were examined. Experiments showed that all five components of the terminal complex (C5b, C6, C7, C8 and C9) were required for killing. Furthermore, multimeric C9 within C5b-9 (C9:C5b-8 ratio of 3.3:1) was required for killing. This was shown in direct binding experiments with purified, radiolabeled C7 and C9, by measuring binding of an antibody specific for multimeric C9 to J5 bearing C5b-9, and by transmission electron microscopy of J5 outer membranes bearing C5b-9. Preliminary experiments measuring release of markers for J5 have shown that multimeric C9 within C5b-9 is required for release of the large periplasmic marker beta lactamase and the small cytoplasmic marker Rb86. These experiments suggest that either C5b-9 bearing low C9 multiplicities does not have access to the cytoplasmic space or that the E. coli K+ transport systems compensate for C5b-9 channels until C9:C7 ratios of greater than 3.3:1 are achieved. Experiments were continued on the mechanism of action of bactericidal antibody for E. coli 0111. Bactericidal antibody did not change the distribution of C3 on the capsule and outer membranes. However, nearly 1/5 of C3 deposited in the presence of IgG bound covalently to the antibody molecule. We therefore prepared covalent complexes of C3b-IgG. These complexes were 3-5-fold more efficient than IgG in presensitizing 0111 for direct complement killing, suggesting that formation of C3b-IgG complexes may be critical for the serum bactericidal reaction. The mechanism of action of blocking IgG for Neisseria gonorrhoeae was tested. Results showed that blocking IgG enhanced rather than blocked complement consumption and deposition on GC. Furthermore, blocking IgG competed with bactericidal IgG for binding to GC. Finally, we showed that blocking IgG led to deposition of C3 at new sites on the outer membrane, sites which do not support formation of a bactericidal C5b-9 complex.
