Robert J. Ziccardi’s research while affiliated with The Scripps Research Institute and other places

We had previously demonstrated that in normal human serum (NHS) nascent C3b inhibited C1 activation by immune complexes (IC). We have now investigated the mechanism of this feedback inhibition. For these studies, EA-IgG were added to solutions containing physiological concns of purified C1, C1-In, C2, C3 and C4. Mixtures were then incubated at 37 degrees C for 30 min. Western blot and autoradiographic analyses revealed that almost half of the IgG molecules had become covalently linked to C3b in a 1:1 complex with the C3 alpha' chain of C3b being bound to the heavy chain of IgG. IgG-C3b and free IgG were separated by ion exchange chromatography and immune complexes were formed with each. The consumption of complement in NHS by EA-IgG and EA-(IgG-C3b) were then compared. The results indicate that binding of C3b to IgG did not significantly inhibit the C1 activating potential of the IgG. Thus feedback inhibition is not due to the binding of C3b to IgG. An alternative mechanism was next explored. After incubation of EA-IgG with C1 through C3, EA were separated from supernatant fluid by centrifugation. It was determined that one-third to one-half of the IgG had been released from the erythrocytes. Release appears not to have been due to C3b binding to IgG, since the released IgG-C3b readily bind to fresh sheep erythrocyte (E), and since IgG that was free of C3b was also released from EA by complement, it is more likely that C3b binding to the E caused the dissociation of antibody. These results indicate that under physiological conditions, the C1 activating potential of an immune complex is greatly reduced as the result of the binding of nascent C3b to the antigen moiety of the IC, thereby causing the displacement of complement activating antibody. In addition to IgG, IgG-C3b is also released from the IC.

We have demonstrated that immune complexes turn over C1, i.e., limiting quantities of immune complexes activate an excess of C1. This was readily apparent in a system of purified C1 and C1-inhibitor (C1-In) but not in normal human serum (NHS). The following results indicate that C3 and C4 are the serum factors responsible for the inhibition of C1 turnover by immune complexes. 1) In a purified protein system composed of C1 and C1-In at pH 7.5, ionic strength 0.14 M, doses of immune complexes that activated all the C1 in 60 min at 37 degrees C yielded no detectable C1 activation when C2, C3, and C4 were also present. All proteins were at their physiologic concentrations. Activation was quantified by SDS-PAGE analysis and hemolytic titration 2) In order to inactivate C3 and C4, NHS was treated with 50 mM methylamine (MeAm) for 15 min at 37 degrees C, after which the MeAm was removed by dialysis. The activities of C1, C2, and C1-In were unaffected by this treatment. Doses of immune complexes that consumed no C1 in NHS, consumed all the C1 in MeAm-treated NHS (MeAm-NHS). 3) Reconstitution of MeAm-NHS with physiologic concentrations of C3 and C4 rendered the serum again resistant to excessive C1 consumption by immune complexes. Immune complexes used in these studies included EA-IgG, EA-IgM, tetanus-human anti-tetanus, and aggregated human IgG. There appeared to be specificity to the inhibition reaction since C4 by itself could inhibit C1 consumption by EA-IgM, whereas the presence of C3 was also required to control EA-IgG. Finally, N-acetyl-L-tyrosine was added to NHS at a final concentration of 30 mM. This nucleophile did not interact with native C3 or C4, nor did it directly activate C1. However, upon the addition of low doses of immune complexes, acetyl tyrosine did yield uncontrolled C1 activation, presumably by binding nascent C3b and C4b and thereby blocking their attachment to the immune complexes. We conclude that in NHS there is a mechanism of feedback inhibition by which nascent C3b and C4b inhibit C1 turnover by immune complexes. This mechanism of control might be physiologically important in that it prevents excessive complement activation by low concentrations of immune complexes.

The association of native C1 with physiologically relevant proteins was studied by ultracentrifugation. 125I-C1 was centrifuged through numerous sucrose density gradients, each of which contained a different concentration of monomeric (19S) IgM throughout the gradient. The s-rate of C1 (16S) increased with increasing IgM input to a maximum of 32S. In the absence of C1q, the C1r2s2 subunit did not bind to the Ig. In gradients containing physiologic concentrations of IgM (1.3 mg/ml) at 0.14 M ionic strength, the observed s-rate of C1 was 21S. In the presence of 13 mg/ml IgG, C1 sedimented with an s-rate of 19S. Thus, under physiologic conditions, a significant fraction of native C1 is reversibly bound to monomeric Ig. SDS-PAGE analyses show that this interaction does not lead to C1 activation. The interaction of native C1 with C1 inhibitor (C1-In) was studied by ultracentrifugation at physiologic ionic strength. Purified 125I-C1-In alone sedimented with an s-rate of 4S. However in the presence of excess native C1, one-third of the C1-In co-sedimented with C1 at a 16S position. For these studies, 100 microM nitrophenylguanidinobenzoate (NPGB) was present throughout the sucrose density gradient to prevent C1 activation during centrifugation. As the concentration of NPGB was increased, the percent of 125I-C1-In at 16S decreased, indicating that C1-In was binding (reversibly) to the C1 active site region(s), which is at least partially accessible in uncleaved C1. In controls, when NPGB was omitted or activated C1 was used, the s-rate of 125I-C1-In was only 12S due to the release of C1rC1s(C1-In)2 from activated C1. Thus, under physiologic conditions native C1 is reversibly bound to C1-In.

The strength of interaction between the C1q and C1r2S2 subunits of C1 was studied as a function of temp. During centrifugation through sucrose density gradients at 4 degrees C, macromolecular C1 readily dissociated as it sedimented away from its free subunits. In contrast, at 20 degrees C, C1 remained associated as the 16S complex throughout centrifugation, thus indicating a stronger interaction between C1q and C1r2S2 at the higher temp. C1-inhibitor (C1-In) or nitrophenylguanidinobenzoate was present during centrifugation to prevent C1 activation. That native C1 was in fact the species being studied was confirmed by SDS-PAGE analysis. To investigate this temp dependence without using inhibitors, an alternative approach was used. Trace amounts of 125I-C1q were centrifuged through numerous sucrose density gradients, each of which contained a different concn of native C1r2S2 throughout the gradient. The s-rate of 125I-C1q increased with increasing C1r2S2 input. An association constant of 4.9 X 10(7) M-1 was calculated for this reversible interaction at 4 degrees C. However, at 20 degrees C, the data indicated a much higher affinity reaction since the addition of far less C1r2S2 was required for the s-rate of 125I-C1q to reach the 16S plateau. The presence of Cl-In did not affect these results. We have demonstrated that the association of C1q with C1r2S2 increases with increasing temp, a finding suggestive of a hydrophobic interaction. However, since we also show that C1 readily dissociates with increasing NaCl concn, the C1q-C1r2S2 interaction must, in fact, be ionic in nature. We therefore conclude that the temp dependence of the inter-subunit interaction is the result of a conformational change(s) within one of the subunits, and propose that this change may be similar to that occurring during Cl activation.

The binding of C4b to C4b-binding protein (C4BP) was demonstrated at physiological ionic strength by analytical ultracentrifugation. The sedimentation rate of C4BP gradually increased from 9.4 S to a maximum of 18.5 S with increasing C4b concentration. The stoichiometry of different C4BP X C4b complexes was calculated from the sedimentation-velocity data. A linear relationship was established between the number of C4b bound per C4BP and the sedimentation rate of the complex. In order to define further the C4BP-C4b interaction, sucrose density gradient ultracentrifugation was also used. Trace amounts of 125I-C4BP were centrifuged through 12 sucrose density gradients, each of which contained a different concentration of C4b throughout the gradient. The sedimentation rate of the C4BP increased with increasing C4b input to a maximum of 19.5 S. These binding data, in conjunction with the stoichiometry measurements determined in the analytical ultracentrifuge, were analyzed by the methods of Scatchard and Hill. At physiological ionic strength, C4BP exhibited four binding sites for C4b, each having an association constant of 1.2 X 10(7) M-1. A Hill coefficient of 1.1 was calculated, indicating that the four binding sites were independent. At reduced ionic strength, two additional sites were detected. The sedimentation coefficient of C4BP(C4b)6 was 24 S. The hydrodynamic data suggest that after four C4b molecules have bound to C4BP, the binding of additional C4b is sterically hindered. This interpretation implies that all six binding sites on C4BP are identical. C4BP also bound C4(H2O) (the product resulting from spontaneous hydrolysis of the thiol ester bond in native C4) and weakly bound C4c, but had no measurable affinity for native C4 or C4d at physiological ionic strength. A low-affinity interaction between C3b and C4BP was also demonstrated in the analytical ultracentrifuge. The C4BP X C3b complex was specific because C4BP mediated the cleavage of C3b by Factor I to C3bi with concomitant dissociation of the complex.

Antibody-independent interactions of C1 with several E. coli strains were examined. Purified C1 was directly activated by the semi-rough mutant E. coli J-5, its parental wild-type strain, E. coli 0111:B4, and two clinical isolates, E. coli (P) and E. coli (A), in the absence of C1 inhibitor. E. coli J-5 activated C1 about 10-fold more rapidly and bound approximately threefold more C1 than the other strains. E. coli J-5, but not the other strains, also bound C1s2, provided that the subcomponent was offered to the bacteria in the presence of C1q and calcium; such binding was thus independent of the presence or absence of C1r2. After C1 activation in the absence of C1 inhibitor, activated C1s spontaneously dissociated from E. coli 0111:B4, (P), and (A), but remained associated with E. coli J-5. The regulatory protein C1 inhibitor prevented C1 activation by the weaker activators, E. coli strains 0111:B4, (P), and (A), but had no effect on C1 activation by E. coli J-5. Although C1 inhibitor thus failed to modulate C1 activation by E. coli J-5, it did block the enzymatic activity of activated C1 bound to this strain. Analyses of the molecular processes involved revealed differences with other systems. In the presence of C1 inhibitor, the C1s subunit of C1 activated by E. coli J-5 underwent further cleavage with the release into the supernatant of C1s fragments and complexes of C1 inhibitor with light chain fragments. Such fragments were not disulfide-linked to the remainder of the C1s molecule. The bulk of the heavy chain remained adherent to the surface of E. coli J-5. This finding documents the presence of a binding site for activated C1s on the surface of E. coli J-5 and localizes this site to the heavy chain. These studies thus indicate that several E. coli strains are direct C1 activators. Furthermore, E. coli J-5 provides another example of a direct C1 activator having binding sites not only for C1q but also for dimeric C1s. The studies also show that there are multiple properties of particles which determine the ability to activate C1, the rate of activation, the possibility of regulation of the activation process by C1 inhibitor, and the fate of activated C1.

Immune complex-induced C1 activation and fluid phase C1 autoactivation have been compared in order to elucidate the immune complex role in the C1 activation process. Kinetic analyses revealed that immune complex-bound C1 activates seven times faster than fluid phase C1 spontaneously activates. The rate of spontaneous C1 activation increased after decreasing the solution ionic strength. In fact at one-half physiologic ionic strength (i.e., 0.08 M), the kinetics of spontaneous C1 activation were indistinguishable from the kinetics of activation of immune complex-bound C1 at physiologic ionic strength. The enhanced fluid phase C1 activation at low ionic strength resulted neither from C1 nor C1q aggregation, nor from selective effects on the C1r2S2 subunit; however, at the reduced ionic strength, the C1 association constant (defined for C1q + C1r2S2 in equilibrium C1qr2S2) did increase to 2.3 X 10(8) M-1, which is equal to that for C1 bound to an immune complex at physiologic ionic strength. Therefore, C1 can spontaneously activate in the fluid phase as rapidly as C1 on an immune complex when the strength of interaction between C1q and C1r2S2 is the same in both systems. In conclusion, under physiologic conditions, C1q and C1r2S2 are two weakly interacting proteins. Immune complexes provide a site for the assembly of a stable C1 complex, in which C1q and C1r2S2 remain associated long enough for C1q to activate C1r2S2. Thus, immune complexes enhance the intrinsic C1 autoactivation process by strengthening the association of C1q with C1r2S2.

The first component of human complement (C1) was reconstituted from equimolar concentrations of its purified subunits C1q, C1r, and C1s, in the presence of each of nine different metal ions for the purpose of studying the qualitative and quantitative nature of the metal ion requirement for C1 assembly and function. For C1 reconstituted with each metal ion, three assays characteristic of C1 were performed as follows: (1) spontaneous C1 activation in the absence of the regulatory protein C1-inhibitor was assessed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis analysis by simultaneously quantifying the specific proteolysis of the C1r and C1s subunits; (2) C1 activation induced by aggregated IgG in the presence of C1 inhibitor was similarly analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis; and (3) formation of 16 S macromolecular C1 was determined in the analytical ultracentrifuge. For all experiments, trace metal contaminants were removed from buffers and proteins. All divalent cations tested from the first transition period of the periodic table (i.e. Ca2+, Mn2+, Co2+, Ni2+, and Zn2+) effectively mediated the formation of functional macromolecular C1. Dose curves showed maximal C1 assembly and activation at ion concentrations of 30 to 50 microM for each of the above metal ions. However, when ion concentrations were increased above 50 microM, C1 assembly and activation became inhibited. The further to the right in the periodic table, the better inhibitor was the metal ion. Competition experiments indicated that the ion binding sites mediating inhibition are distinct from those promoting activation. Other metal ions that also effectively mediated C1 assembly and function were Cd2+ and Tb3+; however, Mg2+ and Ba2+ were ineffective. All metal ions that mediated C1 assembly and activation also promoted C2 consumption by C1 in normal human serum treated with aggregated IgG. In conclusion, the assembly and function of C1 can be mediated by numerous metal ions. In direct opposition to accepted theory, there is no specific requirement for calcium.

Complement is a group of serum proteins that together play a vital role in the host defense against infection. Stimulation of the complement system triggers sequential biochemical reactions, which are accompanied by the generation of numerous biologically active mediators of inflammation, ultimately leading to the destruction and clearance of invading organisms. The classical pathway of complement activation is initiated by the first complement component (C1). After an activating substance, such as an immune complex, binds and activates C1, this C1 then activates the second (C2) and fourth (C4) complement components thereby triggering the complement cascade. The purpose of this report is to summarize the known molecular principles involved in C1 activation and its physiologic control, while emphasizing contemporary aspects. Cl is of interest not only because of its importance as the initiator of the classical complement pathway, but also since it is the most readily defined and easily studied mediator of immunoglobulin function. Furthermore, C1 in itself is an intriguing biochemical model involving specific protein-protein interactions, induced conformational changes, and activation by limited proteolysis.

The first component of human complement (C1) readily dissociates under physiologic conditions into two subunits - C1q and C1r2C1s2. The equilibrium constant for this reaction has been determined for native C1 in fresh normal human serum by hemolytic titration. Standard technology was modified to simulate physiologic conditions. Furthermore, assays were carried out at numerous poncentrations of sensitized erythrocytes, thereby allowing the calculation of the percent of associated C1 at different total C1 concentration. Increased C1 dissociation was observed with dilution. From these data, an association constant of 4.5 × 108 M−1 was calculated for native C1. Thus in normal human serum approximately ten percent of the C1 is present as free C1q and C1r2C1s2.