C1q binding and C1 activation by various isolated cellular membranes

Cellular and subcellular membranes obtained from heart, liver, and brain tissue from human, baboon, bovine, rabbit, and rat bound highly purified, radioiodinated human C1q with a high affinity (Ka = 10(8) to 10(10) M-1). The majority of these membrane preparations were able to activate fully assembled C1 as evidenced by the conversion of 125I-C1s, incorporated into C1 complexes, to 125I-C1s. C1 activation by baboon heart mitochondrial membranes required an intact C1 complex and appeared to be mediated by the binding of the C1q subcomponent in that excess C1q completely blocked C1 activation. Several experiments suggested that the heart mitochondrial membrane binding site for C1q is an integral component of the mitochondrial membrane and that C1q interacted with the membrane binding site through its globular head regions. It is suggested that the binding of C1q and the activation of C1 by cellular and subcellular membranes may be involved in the initiation and/or enhancement of the inflammatory process after acute tissue damage.     

Antibody-independent activation of C1. I. Differences in the mechanism of C1 activation by nonimmune activators and by immune complexes: C1r-independent activation of C1s by cardiolipin vesicles

C1 activation is controlled by the regulatory protein C1-inhibitor (C1-INH). In contrast to immune-complex-induced activation, which is insensitive to C1-INH, antibody-independent activation of C1 is modulated by C1-INH. The mechanisms regulating nonimmune activation were studied with two phospholipids varying in their capacity to activate C1 in the presence of C1-INH: cardiolipin (CL) and phosphatidylglycerol (PG). Whereas C1-INH consistently suppressed activation by PG vesicles, a dose-dependent increase in C1 activation was measured with CL vesicles above 40 mole %. A similar dose-response binding of C1s requiring C1q, but not C1r, was detected only on CL vesicles, but neither on PG vesicles nor on immune complexes. This binding was Ca2+-dependent, suggesting that dimeric C1s is involved and was inhibited by spermine. The C1q-bound C1s was specifically cleaved at 37 degrees C into its active 58 kDa and 28 kDa chains, in the absence of C1r. On the addition of anti-CL antibodies, the C1q-mediated cleavage of C1s by CL vesicles was specifically inhibited. The cleavage of C1r on CL vesicles was also determined. When macromolecular C1 was offered in the presence of C1-INH, C1r cleavage was detected; however, the presence of C1s was a critical factor for C1r activation, because it was required on CL vesicles, but not on immune complexes. These results show that nonimmune activation of C1 presents specific features which distinguish it from immune complex-induced activation. These characteristics varied with the capacity of antibody-independent activators to activate C1 in the presence of C1-INH.     

Inhibition of bFGF activity by complement C1s: covalent binding of C1s with bFGF

The first complement component C1s formed large aggregates with bFGF when bFGF and C1s were incubated at 37°C overnight. Under non-reducing conditions, a part of the aggregates did not penetrate into 5% polyacrylamide gel in the presence of SDS, and the rest penetrated into 5% gel but not into 12% gel. The aggregates were dissociated into monomers by reducing with 2-mercaptoethanol. Both active and inactive C1s formed aggregates with bFGF. In addition, a portion of bFGF was degraded by active C1s but not by inactive C1s. Aggregates were not formed when 2-mercaptoethanol (2 mM &base;) was added to the incubation mixture. After the incubation with C1s the growth-stimulating activity of bFGF was measured by using human umbilical vein endothelial cells (HUVEC) as indicator cells. The aggregate formation between C1s and bFGF significantly reduced the activity of bFGF. © 1998 John Wiley & Sons, Ltd.     

Measurement of macromolecular interactions between complement subcomponents C1q, C1r, C1s, and immunoglobulin IgM by sedimentation analysis using the analytical ultracentrifuge

The interactions between the complement components and with immunoglobulins are greatly enhanced by lowering the ionic strength and become readily measurable by physical techniques. Thus, the binding between C1q and IgM was previously shown to be appreciable (k = 1 x 10(6) M-1) at 0.084 M ionic strength (Poon, P.H., Phillips, M.L., and Schumaker, V.N. (1985) J. Biol. Chem. 260, 9357-9365). We have now found that, at 0.128 M ionic strength, the binding between human C1- (the activated first component of complement) and IgM was strong at physiological concentrations (k = 1 x 10(7) M-1), while under the same conditions binding between C1q and IgM was not observed. To explore the nature of the interactions responsible for this enhanced binding by C1- over C1q, mixtures of the various subcomponents of C1- were studied alone and with IgM. C1r2 did not bind to C1q, even when the ionic strength was reduced to 0.098 M, nor did the presence of C1r2 enhance the binding of C1q to IgM. In contrast, two C1s2 independently bound to C1q (k = 1 x 10(6) M-1), and caused a marked increase in its association with IgM (k = 5 x 10(6) M-1) at 0.098 M ionic strength. No detectable interaction was found between C1s2 and/or C1r2 and IgM in the absence of C1q. Moreover, there was no detectable interaction between the C1(-)-like complex formed between C1r2C1s2 and the collagenous C1q stalks (pepsin-digested C1q) and IgM. These data suggest that the binding of C1s2 to C1q, either alone or together with C1r2, induces a conformational change in C1q which results in additional C1q heads binding to complementary sites on IgM.     

Inhibitors of C1q biosynthesis suppress activation of murine macrophages for both antibody-independent and antibody-dependent tumor cytotoxicity

The inhibitors of C1q biosynthesis and secretion, 3,4-dehydro-DL-proline (DHP) and 2,2'-dipyridyl, were previously shown to suppress murine macrophage FcR-dependent phagocytosis and cytolysis of IgG-opsonized RBC targets. Inasmuch as non-antibody macrophage activators also bind C1q to initiate C1 activation, we determined the effects of these same inhibitors of C1q biosynthesis on activation of macrophages for antibody-independent, nonspecific tumor cytotoxicity by lipid A and a variety of other non-antibody activators. Preexposure of mouse inflammatory peritoneal macrophages to either DHP (0.5 to 2.5 mM) or 2,2'-dipyridyl (0.1 to 0.3 mM) for 24 h produced a dose-related suppression of their response to activation by lipid A to mediate tumor cytotoxicity of L1210 mouse leukemia targets. Inhibition of C1q secretion by DHP-treated macrophages was confirmed both by a complement hemolytic assay and by autoradiographic analysis of [35S]methionine-labeled culture supernatants. DHP-treated macrophages were inhibited in their response to direct activation and triggering of IFN-gamma-primed macrophages by lipid A, Poly I:C, and cobra venom factor for tumor cytotoxicity. DHP inhibited macrophage activation for antibody-dependent cellular cytotoxicity of L1210 tumor targets mediated by antitumor target IgG. The addition of exogenous purified C1q (2 micrograms/ml) to macrophages after DHP treatment, reconstituted their response to activation for both antibody-independent and antibody-dependent tumor cytotoxicity. Our results indicate that C1q synthesis and secretion by effector macrophages is a prerequisite for the initiation of their activation by both immune complex and by non-antibody agents that also bind C1q. It now appears that macrophage-derived C1q may act as an auxiliary amplification signal for autocrine-like modulation of the initiation of macrophage activation by both the antibody-dependent and independent pathways.     

Tumor promoter binding of the protein kinase C C1 homology domain peptides of RasGRPs, chimaerins, and Unc13s

Recent investigations discovered nonkinase-type phorbol ester receptors, RasGRPs, chimaerins, and Unc13s. Phorbol ester binding occurs at the cysteine-rich sequences of about 50 residues in the C1 domains of these receptors. Fifty-one-residue RasGRP C1 peptides except for RasGRP2 showed significant phorbol 12,13-dibutyrate (PDBu) binding, but the K(d) values of the RasGRP1 and RasGRP3 C1 peptides were about 10-fold larger than those for the corresponding whole enzymes. Addition of the C-terminal basic amino acid cluster decreased their K(d) values about 10-fold, suggesting that the positive charges of these C1 peptides play an important role in the PDBu binding in the presence of negatively-charged phosphatidylserine. The 51-mer chimaerin C1 peptides showed potent PDBu binding, while the Unc13 and Munc13-1 C1 peptides without sufficient positive charges hardly bound PDBu. By the rapid screening system using this C1 peptide library, 5-prenyl-indolactam-V was identified as a promising lead for the novel protein kinase C isozyme specific ligands.     

Limited proteolysis of beta 2-microglobulin at Lys-58 by complement component C1s

We have now demonstrated that activated complement component C1s cleaves beta 2-microglobulin at the position identical to that at which beta 2-microglobulin is cleaved in serum of patients suffering from lung cancer. The main cleavage is in the disulphide loop C-terminal to Lys-58, generating a modified form of beta 2-microglobulin with a two-chain structure. The C-terminal Lys-58 in the A chain is highly susceptible to removal by a carboxypeptidase-B-like activity causing the formation of des-Lys58-beta 2-microglobulin. This is the first demonstration of a noncomplement protein substrate for the proteolytic activity of C1s. The C1s-induced cleavage of beta 2-microglobulin can be inhibited in the presence of C1 esterase inhibitor, demonstrating a regulatory function of C1 esterase inhibitor in the C1s-induced cleavage of beta 2-microglobulin.     

The NK1 receptor localizes to the plasma membrane microdomains, and its activation is dependent on lipid raft integrity

The spatial targeting of receptors to discrete domains within the plasma membrane allows their preferential coupling to specific effectors, which is essential for rapid and accurate discrimination of signals. Efficiency of signaling is further increased by protein and lipid segregation within the plasma membrane. We have previously demonstrated the importance of raft-mediated signaling in the regulation of smooth and skeletal muscle cell contraction. Since G protein-coupled receptors (GPCRs) are key components in the regulation of smooth muscle contraction-relaxation cycles, it is important to determine whether GPCR signaling is mediated by lipid rafts and raft-associated molecules. Neurokinin 1 receptor (NK1R) is expressed in central and peripheral nervous system as well as in endothelial and smooth muscle cells and involved in mediation of pain, inflammation, exocrine secretion, and smooth muscle contraction. The NK1 receptor was transiently expressed in HEK293 and HepG2 cell lines and its localization in membrane microdomains investigated using biochemical methods and immunofluorescent labeling. We show that the NK1 receptor, similar to the earlier described beta(2)-adrenergic receptor and G proteins, localizes to lipid rafts and caveolae. Protein kinase C (PKC) is one of the downstream effectors of the NK1 activation. Its active form translocates from the cytoplasm to the plasma membrane. Upon stimulation of the NK1 receptor with Substance P, the activated PKC relocated to lipid rafts. Using cholesterol extraction and replenishment assays we show that activation of NK1 receptor is dependent on the microarchitecture of the plasma membrane: NK1R-mediated signaling was abolished after cholesterol depletion of the receptor-expressing cells with methyl-beta-cyclodextrin. Our results demonstrate that reorganization of the plasma membrane has an effect on the activation of the raft-associated NK1R and the down-stream events such as recruitment of protein kinases.     

Interplay of C1 and C2 domains of protein kinase C-alpha in its membrane binding and activation.

The regulatory domain of conventional protein kinase C (PKC) contains two membrane-targeting modules, the C2 domain that is responsible for Ca2+-dependent membrane binding of protein, and the C1 domain composed of two cysteine-rich zinc fingers (C1a and C1b) that bind diacylglycerols and phorbol esters. To understand the individual roles and the interplay of the C1 and C2 domains in the membrane binding and activation of PKC, we functionally expressed isolated C1 and C2 domains of PKC-alpha and measured their vesicle binding and monolayer penetration. Results indicate that the C2 domain of PKC-alpha is responsible for the initial Ca2+- and phosphatidylserine-dependent electrostatic membrane binding of PKC-alpha, whereas the C1 domain is involved in subsequent membrane penetration and diacylglycerol binding, which eventually lead to enzyme activation. To determine the roles of individual zinc fingers in the C1 domain, we also mutated hydrophobic residues in the C1a (Trp58 and Phe60) and C1b (Tyr123 and Leu125) domains of the native PKC-alpha molecule and measured the effects of mutations on vesicle binding, enzyme activity and monolayer penetration. Results show that the hydrophobic residues in the C1a domain are essential for the membrane penetration and activation of PKC-alpha, whereas those in the C1b domain are not directly involved in these processes. Based on these results in conjunction with our previous structure-function studies of the C2 domain (Medkova, M., and Cho, W. (1998) J. Biol. Chem. 273, 17544-17552), we propose a mechanism for the in vitro membrane binding and activation of conventional PKC that accounts for the temporal and spatial sequences of PKC activation.     

Assessment of rat brain alpha1-adrenoceptor binding and activation of inositol phospholipid turnover following chronic imipramine treatment

Chronic (21 days) treatment of rats with imipramine (10 mg/kg) did not change the density or affinity of alpha₁-adrenoceptors as measured by the specific binding of [³H]prazosin in rat cortical membranes, but produced the expected significant decrease in the density of beta-adrenoceptors labeled by [¹²⁵I]iodocyanopindolol. The functional status of brain alpha₁-adrenoceptors was also assessed by measuring the noradrenaline (NA)-induced accumulation of [³H]inositol 1-phosphate (IP₁) in brain slices from these animals. No apparent change was observed in the concentration-response relationship between NA and [³H]IP₁ accumulation in rat cerebral cortex after chronic treatment with imipramine. At concentrations higher than 1 M in vitro, imipramine and its metabolite, desipramine, produced a concentration-dependent decrease in the [³H]IP₁ accumulation elicited by NA. This inhibitory effect is likely mediated by direct blockade of alpha₁-adrenoceptors by these drugs. As the endogenous drug concentration would not reach 1 M in our preparation, the lack of changes in alpha₁-adrenoceptor response following chronic imipramine treatment are not likely attributable to residual imipramine or desipramine retained in the tissues. In conclusion, the above findings do not support previous suggestions that brain alpha₁-adrenoceptors are upregulated following chronic imipramine administration.     

Human complement subcomponent C2: purification and proteolytic cleavage in fluid phase by C1s, C1r2-C1s2 and C1

Photosynthetic membranes contain considerable regions of high surface curvature, notably at their margins, where the average radius of curvature is about 10 nm. The proportion of total membrane lipid in the outer and inner thylakoid margin monolayers is estimated at 21% and 13%, respectively. The major thylakoid lipid, monogalactosyldiacylglycerol, is roughly cone-shaped and will not form complete lamellar bilayer phases, even in combination with other thylakoid lipids. It is proposed that this galactolipid plays a role in: (a) stabilising regions of concave curvature in thylakoids; and (b) packaging hydrophobic proteins in planar bilayer regions by means of inverted micelles. This model predicts substantial asymmetries in the distribution of lipids both across and along the thylakoid bilayer plane.     

Phospholipid functional groups involved in protein kinase C activation, phorbol ester binding, and binding to mixed micelles

The specificity of the phospholipid cofactor requirement of rat brain protein kinase C was investigated using Triton X-100 mixed micellar methods. Sixteen analogues of phosphatidylserine were prepared and tested for their ability to support protein kinase C activity, [3H]phorbol 12,13-dibutyrate binding, and protein kinase C binding to mixed micelles. Phosphatidylserinol, -L-serine methyl ester, -N-acetyl-L-serine, -2-hydroxyacetate, -3-hydroxypropionate, and -4-hydroxybutyrate did not activate protein kinase C in mixed micelles containing 2 mol % of sn-1,2-dioleoylglycerol. This indicates that both the carboxyl and amino moieties are important for activation. Phosphatidyl-D-serine and -L-homoserine were incapable of supporting full activation; this demonstrates stereospecificity and the importance of the distance between the phosphate and carboxyl and amino moieties. Since 1,2-rac-phosphatidyl-L-serine and 1,3-phosphatidyl-L-serine fully supported protein kinase C activity, the stereochemistry within the glycerol backbone at the interface was not necessary for maximal activation. Neither lysophosphatidyl-L-serine nor 1-oleoyl-2-acetyl-sn-glycero-3-phospho-L-serine supported protein kinase C activity implying that the interfacial conformation is critical to the activation process. The phospholipid dependencies of [3H]phorbol 12,13-dibutyrate binding and of protein kinase C binding to mixed micelles containing sn-1,2-dioleoylglycerol did not mirror those for activation. The data demonstrate that protein kinase C possesses a high degree of specificity with respect to phospholipid activation and implicate several functional groups within the phospho-L-serine polar head group in binding and activation.     

Fluid-phase interaction of C1 inhibitor (C1 Inh) and the subcomponents C1r and C1s of the first component of complement, C1.

Interactions between proenzymic or activated complement subcomponents of C1 and C1 Inh (C1 inhibitor) were analysed by sucrose-density-gradient ultracentrifugation and sodium dodecyl sulphate/polyacrylamide-gel electrophoresis. The interaction of C1 Inh with dimeric C1r in the presence of EDTA resulted into two bimolecular complexes accounting for a disruption of C1r. The interaction of C1 Inh with the Ca2+-dependent C1r2-C1s2 complex (8.8 S) led to an 8.5 S inhibited C1r-C1s-C1 Inh complex (1:1:2), indicating a disruption of C1r2 and of C1s2 on C1 Inh binding. The 8.5 S inhibited complex was stable in the presence of EDTA; it was also formed from a mixture of C1r, C1s and C1 Inh in the presence of EDTA or from bimolecular complexes of C1r-C1 Inh and C1s-C1 Inh. C1r II, a modified C1r molecule, deprived of a Ca2+-binding site after autoproteolysis, did not lead to an inhibited tetrameric complex on incubation with C1s and C1 Inh. These findings suggest that, when C1 Inh binds to C1r2-C1s2 complex, the intermonomer links inside C1r2 or C1s2 are weakened, whereas the non-covalent Ca2+-independent interaction between C1r2 and C1s2 is strengthened. The nature of the proteinase-C1 Inh link was investigated. Hydroxylamine (1M) was able to dissociate the complexes partially (pH 7.5) or totally (pH 9.0) when the incubation was performed in denaturing conditions. An ester link between a serine residue at the active site of C1r or C1s and C1 Inh is postulated.     

Role of T-loop phosphorylation in PDK1 activation, stability, and substrate binding

3-Phosphoinositide-dependent protein kinase-1 (PDK1) phosphorylates the T-loop of several AGC (cAMP-dependent, cGMP-dependent, protein kinase C) family protein kinases, resulting in their activation. Previous structural studies have revealed that the alpha C-helix, located in the small lobe of the kinase domain of PDK1, is a key regulatory element, as it links a substrate interacting site termed the hydrophobic motif (HM) pocket with the phosphorylated Ser-241 in the T-loop. In this study we have demonstrated by mutational analysis that interactions between the phosphorylated Ser-241 and the alpha C-helix are not required for PDK1 activity or substrate binding through the HM-pocket but are necessary for PDK1 to be activated or stabilized by a peptide that binds to this site. The structure of an inactive T-loop mutant of PDK1, in which Ser-241 is changed to Ala, was also determined. This structure, together with surface plasmon resonance binding studies, demonstrates that the PDK1(S241A)-inactive mutant possesses an intact HM-pocket as well as an ordered alpha C-helix. These findings reveal that the integrity of the alpha C-helix and HM-pocket in PDK1 is not regulated by T-loop phosphorylation.     

A mutational epitope for cytochrome C binding to the apoptosis protease activation factor-1

Cytochrome c (Cc) binding to apoptosis protease activation factor-1 (Apaf-1) is a critical activation step in the execution phase of apoptosis. Here we report studies that help define the Cc:Apaf-1 binding surface. It is shown that a large number of Cc residues, including residues 7, 25, 39, 62-65, and 72, are involved in the Cc:Apaf-1 interaction. Mutation of residue 72 eliminated Cc activity whereas mutations of residues 7, 25, 39, and 62-65 showed reduced activity in an additive fashion. The implications of this binding model for both recognition and modulation of protein-protein interactions are briefly discussed.     

The neuropeptide bradykinin stimulates phosphoinositide turnover in HSDM1C1 cells: B2-antagonist-sensitive responses and receptor binding studies

Bradykinin (BK) and its analogs (1 nM-100 M) stimulated phosphoinositide (PI) turnover in murine fibrosarcoma (HSDM1C1) cells in a concentration-dependent manner. The relative potencies (EC₅₀) were: BK=48±4 nM; Lys-BK=39±3 nM; Met-Lys-BK=158±33 nM; Des-Arg⁹-BK=2617±598 nM (means±SEM, n=3–14). All these analogs were full agonists and they produced up to 5.4±0.4-fold stimulation of PI turnover at the highest concentration (10–100 M) of the peptides. In contrast, the analogs [D-Arg⁰-HYP³-Thienyl^5,8-D-Phe⁷]-BK (HYP³-antagonist), [D-Arg⁰-HYP³-Thienyl,^5,8-D-Phe⁷]-BK (Thienyl antagonist) and Des-Arg⁹-Leu⁸-BK were inactive, as agonists, at 0.1 nM-1 M in this system. These data suggested that BK-induced PI turnover in these cells was mediated via B₂-type of BK receptors. This was confirmed further by the fact that both the B₂-selective Hyp³- and Thienyl-antagonists inhibited BK-induced PI turnover with K_BS of 369±51 nM and 368±118 nM respectively while the B₁-selective antagonist, Des-Arg⁹-Leu⁸-BK, was inactive at 1 M. [³H]BK receptor binding studies revealed two binding sites, one with high affinity (K_d=0.24±0.06 nM; B_max=1.4±0.4 pmol/g tissue) and the other with low affinity (K_d=18.5±0.95 nM; B_max=25.1±0.52 pmol/g tissue), on HSDM1C1 cell homogenates. The rank order of affinity of BK analogs at inhibiting specific [³H]BK binding was similar to that found for PI turnover. Taken together, these data have provided evidence for the presence of two B₂-type BK binding sites on the HSDM1C1 cells. Based on the affinity parameters, the low-affinity component of [³H]BK binding in HSDM1C1 cells appears to be coupled to the phospholipase C-induced PI turnover mechanism. The high-affinity component has been previously shown to mediate the production of prostaglandins by activation of phospholipase A₂.     

Studies on the irreversible step of pepsinogen activation

The bond cleavage step of pepsinogen activation has been investigated in a kinetic study in which the denatured products of short-term acidifications were separated on SDS-polyacrylamide gels and the peptide products were quantitated by densitometry. Although several peptide products were observed, under the conditions of the experiments (pH values between 2.0 and 2.8, 22 degrees C), the only one that was a product of an initial bond cleavage was the 44-residue peptide, which upon removal from pepsinogen yields pepsin. The rate constant for this bond cleavage is 0.015 s-1 at pH 2.4, which is the same as that at which the alkali-stable potential activity of pepsinogen had been found to convert to the alkali-labile activity of pepsin. When the conversion of zymogen to enzyme was followed by the change in fluorescence of adsorbed 6-(p-toluidinyl)naphthalene-2-sulfonate (TNS), the rate of change in TNS fluorescence was the same as the conversion to alkali lability. However, pepstatin blocked the bond cleavage of pepsinogen to pepsin, but it permitted the fluorescence change to proceed. In fact, it accelerated the apparent rate of change of TNS fluorescence by shifting the pKa of an essential conjugate acid from 1.7 to 2.6. The conversion to alkali lability, therefore, may be considered to be a composite of a relatively slow conformational change (at the measured rate), followed immediately by a relatively fast bond cleavage.     

Structure and activity of C1r and C1s

During activation of the first component of the classical complement pathway the two zymogen subcomponents, C1r and C1s are converted to active proteolytic enzymes. Activated C1r cleaves C1s which then becomes the activator of C4 and C2. Amino acid sequence studies of the proteolytic chains of C1r and C1s, carried out in Oxford and Aberdeen respectively, have shown that they belong to the serine proteinase family. Modelling of these sequences to the three-dimensional coordinates of chymotrypsin (Birktoft & Blow 1972) reveals that both molecules have a conserved structural core, and that most of the differences lie in the external loops. Catalytically functional residues (Ile-16, His-57, Asp-102, Ser-195) are conserved, and residue 189 is aspartic acid, consistent with the known trypsin-like specificity of cleavage. Examination of the amino acid sequences of C4a, and comparison with those of the homologous molecules C3a and C5a, shows that there is a marked difference in the distribution of basic residues near the C-terminal arginine residue which is the site of action of C1s. When these amino acid sequences are modelled to the coordinates of C3a (Huber et al. 1980) and docked to the active site of C1s, the basic residues of C4a appear to interact with two glutamate residues peculiar to C1s, suggesting that this interaction may contribute to the ability of C1s to discriminate C4 from C3 and C5.     

C1 dissociation. Spontaneous generation in human serum of a trimer complex containing C1 inactivator, activated C1r, and zymogen C1s

Activation of the C1 complex in the presence of C1 inactivator (C1 IA) is known to result in the formation of tetramer C1 IA-C1r-C1s-C1 IA complexes that are dissociated from C1q. Both C1r and C1s of the tetramers are present in their activated forms. The present investigation concerned the generation of trimer complexes containing C1 IA, activated C1r, and zymogen C1s (C1 IA-C1r-C1s). C1 IA-C1r-C1s were released from C1q and were formed in high concentration during prolonged incubation (1 to 3 days) of normal serum at 37 degrees C without addition of activators. By contrast, dissociation of C1 with formation of C1 IA-C1r-C1s-C1 IA was complete within 30 min at 37 degrees C, when the serum was treated with heat-aggregated IgG (1 g/liter). On size exclusion chromatography (TSK-4000), C1 IA-C1r-C1s and C1 IA-C1r-C1s-C1 IA emerged with apparent m.w. of 320,000 and 460,000, respectively. The composition of the complexes was examined by absorption of serum with F(ab')2 anti-C1s- or anti-C1r-coated Sepharose beads. Eluates were analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis combined with immunoblotting. Under nonreducing conditions, heat-aggregated IgG-treated serum showed high concentrations of C1 IA-C1r (m.w. 202,000) and C1 IA-C1s (m.w. 194,000), while serum incubated at 37 degrees C without activators showed high concentrations of C1 IA-C1r but no C1 IA-C1s. Under reducing conditions, heat-aggregated IgG-treated serum showed m.w. 120,000 and 110,000 complexes of C1 IA and the C1r and C1s light chains, respectively. Uncleaved C1s and the m.w. 120,000 complex was found in serum that was incubated at 37 degrees C without activators. Consistent with results obtained by size exclusion chromatography, analysis by crossed immunoelectrophoresis and by electroimmunoassay showed that C1s could be released from C1 IA-C1r-C1s in the presence of EDTA.