The sera from three C8a-"y deficient patients previously reported to have a selective C8a--y defect were analyzed by SDS-PAGE and Western blot using two polyclonal antisera to C8a-y and a monoclonal antibody to C8a. All three sera exhibited C8a-'y bands that dissociated into a and y chains under reducing conditions. Quantitation of the a-'y subunit in these sera by a sensitive ELISA revealed an amount -1% of that found in normal human serum. A similar assay performed with a specific antiserum to C8,0 showed unexpectedly low levels of C8,B in these sera, which were confirmed by hemolytic titration of C8ft. The remarkable differences between C8a-'y and C8# in

The preparation of bacterial intermediates bearing complement components at various steps of the complement sequence was investigated by suspending immunoglobulin M-opsonized Escherichia coli Oll1:B4 cells in complement-deficient sera at different temperatures and ionic strengths. The optimal conditions for the formation of the intermediates at T,n were found to be an ionic strength of 0.091 , and a temperature of 37°C, except for BAC142, which could be formed equally well at room temperature. In contrast to all the other intermediates, which, once formed at Tmax, were stable in the presence of the whole serum, BAC142 decayed with a half-life of 10 min due to the lability of bound C2. Washing with a buffer of either 0.091 or 0.046 ,u did not affect the bacterial intermediates, with the exception of BAC1-3 formed either in the presence of CS-deficient serum or with purified C3 added to BAC142. All the intermediates were found to be stable after incubation in 0.091-,u buffer for 30 min at 37°C.

This study investigated the interaction between neutrophil myeloperoxidase (MPO) and the C1q component of the complement system. Using a dot-spot assay, MPO was found to bind to C1q in a dose-dependent manner. The specificity of this reaction was proved by the inhibitory effect of F(ab')2 antibodies to C1q and by the inability of MPO to bind to C1r, C1s and IgG. The interaction between MPO and C1q did not influence the enzymatic activity of the peroxidase but resulted in a more stable C1q as assessed by hemolytic assay for C1q. The protective effect of MPO on C1q did not require the presence of H2O2 in the reaction mixture nor was it inhibited by sodium azide, whereas it was abolished by heating the peroxidase. Lactoferrin and lysozyme, unlike MPO, were ineffective in protecting C1q from functional decay. Addition of H2O2 and chloride to MPO and C1q led to a complete inactivation of C1q, which could not be induced by H2O2 alone. The hypochlorite, which is known to be generated during the reaction of MPO with H2O2 and chloride, exhibited a similar inactivating effect on C1q, which was prevented by an external source of methionine.

The postnuclear supernatant of disrupted polymorphonuclear leukocytes exhibited bactericidal activity on Escherichia coli O111:B4 coated with immunoglobulin M antibodies and C5 or C8 but not on C3or C7-coated bacteria. To characterize this antimicrobial activity further, granules obtained from the postnuclear supernatant were extracted with sodium acetate (pH 4) and the soluble extract was subsequently fractionated through carboxymethyl cellulose and Sephacryl S-200. Over 90% of the activity present in the starting material was recovered in the soluble granule extract. Kinetic and dose-response analyses of the bactericidal activity of the polymorphonuclear leukocyte extract on BAC1-5 and BAC1-8 revealed different susceptibilities to killing of these two bacterial intermediates; they also differed for their susceptibilities to killing at 37°C and at room temperature. The suggestion raised by these data, that BACl-5 and BAC1-8 could be killed by different bactericidal factors, was confirmed by the findings that separate fractions of the soluble granule extract obtained by carboxymethyl cellulose and Sephacryl S-200 chromatography exhibited specific activity on either BAC1-5 or BAC1-8, whereas other fractions were active on both intermediates.

Sera from obligate heterozygotes for deficiency of the C8 beta subunit of the eighth component of human complement (C8) were analyzed for the molecular composition of C8. The C8 alpha-gamma and C8 beta subunits were separated by SDS-PAGE, visualized by immunoblotting, and the resulting bands were quantitated by laser densitometry. The laser densitometric absorption data were set to 100 arbitrary units (AU) for both subunits of pooled normal human sera. The AU values of individual normal sera ranged from 45 to 150 AU for C8 alpha-gamma (median 99 AU) and from 45 to 140 AU for C8 beta (median 101), whereas the C8 alpha-gamma/C8 beta-ratio varied from 0.7 to 1.4. Sera from C8 beta-deficient heterozygotes differed, as expected, from the normal sera for the markedly reduced levels of C8 beta (20 to 90 AU, median 55 AU) and for the higher C8 alpha-gamma/C8 beta-ratio (1.3 to 3.5). High voltage agarose gel electrophoresis was used to separate free and C8 beta-bound C8 alpha-gamma. The migration of free and C8 beta-bound C8 alpha-gamma subunit was checked by hemolytic overlay gels and by second dimension SDS-PAGE and immunoblotting. Immunochemical evaluation of C8 alpha-gamma using this system revealed about 5-14% free C8 alpha-gamma in sera with normal C8 and higher levels, from 33-71%, in the C8 beta D heterozygous sera. Functional analysis confirmed the substantial increase of free C8 alpha-gamma in the heterozygous group. We conclude that the C8 in C8 beta D heterozygous sera is characterized by increased amounts of free C8 alpha-gamma due to reduced concentrations of the C8 beta subunit. This finding may help to identify individuals heterozygous for C8 beta deficiency.