Measurement of the association constants of the complexes formed between intact C1q or pepsin-treated C1q stalks and the unactivated or activated C1r2C1s2 tetramers

The association constants between C1q and C1r2C1s2 and between C1q and C1r2C1s2 were measured in solution using a new technique which employs sucrose gradient ultracentrifugation to estimate thermodynamic association constants. In this technique, zones of dilute, radioiodine-labeled C1q were sedimented through uniform concentrations of either C1r2C1s2 or C1r2C1s2. The zones remained intact, indicating that the dynamic equilibrium was rapid compared with the time of centrifugation. The observed increases in the sedimentation coefficients of the C1q zones were assumed to be directly proportional to the fraction of C1q bound in the dynamic equilibrium. Binding curves were constructed by performing the measurements at many C1r2C1s2 and C1r2C1s2 concentrations. The association constants were estimated from the midpoints of the binding curves and found to be 6.7 X 10(7)M-1 for C1r2C1s2 binding to 125I-C1q. After activation of the C1r2C1s2 the association constant decreased 10-fold to 7.1 X 10(6)M-1. These association constants refer to solvent conditions of pH 7.35, 1 mM Tris, 5 mM Ca2+ and 150 mM NaC1, pH 7.35. Similar measurements were performed with the collagenous peptic fragment of C1q and both 125I-C1r2C1s2 and 125I-C1r2C1s2. The association constants were independent of the state of activation and both found to be about 2 X 10(7) M-1, suggesting that most if not all of the interactions between C1q and C1r2C1s2 were confined to the collagenous portion of C1q.     

The reaction between the complement subcomponent C1q, IgG complexes and polyionic molecules.

The strength of the bond between 125I-labelled C1q and immune complexes, Fc piece, dextran sulphate, polyglutamic acid and polylysine has been investigated. The binding of C1q to Fc piece, small molecular weight (less than 10,000) dextran sulphate, polyglutamic acid and polylysine have value; for the functional affinity constant (Ko) in the range of 0.2-1.5 X 10(4) M-1. In contrast the binding of C1q to immune complexes and large molecular weight polyions (greater than 100,000 is much greater and lies in the range 3 X 10(7)--4 X 10(8) M-1. The differences in the binding constants between the two groups can be explained if the Fc piece and small molecular weight compounds bind to only 1 head of the C1q molecule but the immune complexes and large molecules bind to 2 heads. There are probably 6 binding sites on the C1q molecule for dextran sulphate. The enhancement of the binding affinity of C1q by reduction in ionic strength and the reaction with polyions, indicate that ionic groups are present near or within the binding sites.     

Quantification of antibodies to the C3d subcomponent of human C3.

Purified soluble C3d has been employed to measure the concentration of anti-C3d antibodies in immune rabbit sera. Multiple batches of C3d, prepared from C3-C3b substrate by treatment with C3b-inactivator (KAF), after labelling with 125I, retained 80% immunoreactivity, and were stable on storage at -50 degrees and +4 degrees. Concentrations of anti-C3d were determined by Scatchard analysis of equilibrium concentrations of bound and free C3d in a mixture of 125I-labelled C3d and anti-C3d. Separation of bound from free C3d was by G-75 Sephadex filtration. Assuming a 1:1 molar ratio in the antibody-C3d complex, anti-C3d antibody concentrations for four rabbit whole antisera and four IgG preparations fell in the range 288-2433 microgram/ml, with Ko values of 6-2 X 10(8)-2-9 X 10(9) litres/mol. One commercial antiglobulin-serum contained 3-6 microgram anti C3d/ml and had a Ko value of 1-7 X 10(8) litres/mol. Values for anti-C3d concentrations measured independently by an indirect method employing 125I-labelled sheep anti-rabbit IgG averaged 20% lower than those obtained with 125I-labelled C3d. Antibody concentrations were correlated with antiglobulin agglutination titres against C3d-coated red cells; a titre of 1 was given by an anti-C3d concentration of 0-5 microgram/ml.     

The interaction of 1-anilino-8-naphthalene sulphonate with human C1q

C1q has 12 binding sites for 1-anilino-8-naphthalene sulphonate (ANS), two per peripheral subunit. This number increases to 18 upon weak-acid-induced conformational transition in the globular heads. One ANS binding site is present in each C gamma 2 domain of human IgG. ANS is bound by C1q with a higher affinity (Ka = 2.07 X 10(6) M-1) than by the Fc fragment (Ka = 9.07 X 10(4) M-1) of human IgGl. Hence the inhibitory capacity of C1q binding to IgG immune complexes of ANS probably reflects its preferential binding to the globular heads of C1q. The characteristics of ANS-C1q binding may in part explain the hydrophobic component of the C1q-IgG interaction. It is suggested that an ionic-hydrophobic two-step process is involved in the contact between C1q and IgG.     

In vivo degradation of rat C1q induced by intravenous injection of soluble IgG aggregates.

Immune complexes are able to bind and activate the first component of complement, C1. Upon activation of C1, C1r and C1s are rapidly inactivated by C1-In which also forms a complex with these two subcomponents, resulting in their release from C1-immune aggregate complexes. The fate of C1q after the binding C1 to immune complexes in vivo is not clear and, therefore the clearance of radiolabelled rat C1q was investigated in normal rats and in rats receiving soluble aggregated human IgG. 125I-labelled rat C1q was cleared with a half-life (T 1/2) of 12.4 hr in normal rats. Injection of AIgG into rats that had previously received 125I-C1q accelerated the clearance of 125I-C1q, resulting, finally, in a T 1/2 of 53 min. The levels of circulating endogenous C1q were also followed using haemolytic titrations and immunochemical measurements. Directly after injection of AIgG into rats, there was a rapid decrease in C1q haemolytic activity to less than 25% of the initial value after 10 min. The rate of disappearance of C1q antigen, was, however, much slower, the lowest concentration being 30% at 2 hr. C1q haemolytic activity and the C1q antigen level returned to virtually normal values after 24 hr. Plasma samples were taken at different time intervals after the injection of AIgG and subjected to gel filtration on Sephacryl S-400 columns. It was found that, in the 10 min samples, C1q antigen and C1q haemolytic activity, each with an estimated molecular weight (MW) of 400,000, were detected together. In addition, there was C1q antigen with a MW of less than 69,000 without C1q haemolytic activity. SDS-PAGE analysis of the various serum samples indicated that the low MW C1q antigen had an apparent MW of 25,000. Measurement of uptake of 125I-C1q in various organs indicated that the main site of clearance of 125I-C1q is the liver.     

C1q binding by a high affinity anti-fluorescein murine monoclonal IgM antibody and monomeric subunits

Comparative interactions of purified rabbit C1q with 18-2-3, a high affinity (2-3 X 10(10) M-1) anti-fluorescein (anti-F1) murine monoclonal IgM antibody (pentamer) and constitutive monomeric subunits (IgMs) were studied. Using a solid phase radioimmunoassay (SPRIA), based on immobilized polyvalent antigen, it was shown that the mechanism of C1q binding to IgM was characteristically multiphasic while IgMs yielded monophasic binding curves. The latter compared qualitatively and quantitatively with a monoclonal IgG2a anti-fluorescein antibody with the same intrinsic affinity of 2-3 X 10(10) M-1. C1q binding efficiency to antibodies was significantly enhanced when the immunoglobulins interacted with immobilized multivalent antigen. Monoclonal IgM antibody bind identically to six F1-carrier protein conjugates independent of epitope (F1) density. In contrast, the C1q-antibody interaction binding was dependent upon epitope density. An average distance between F1 epitopes of 80 A was optimal for C1q binding by IgM. At low concn of IgM, when fluorescein was bound by antigen-binding sites on adjacent subunits of an intact pentamer, C1q appeared to bind IgM intramolecularly.     

Enhancement of the binding of C1q to immune complexes by polyethylene glycol

The binding of 125I-labelled human C1q to insoluble rabbit IgG:ovalbumin immune complexes was enhanced by polyethylene glycol (PEG, Mr 8 x 10(3)) in the concn range 0-2.5% (w/v). C1q with native immunoglobulin bindings sites rendered inactive by diethylpyrocarbonate treatment did not bind to immune complexes in the presence of PEG. The ionic strength dependence of the binding was independent of the presence of PEG. There was a linear relationship between the logarithm of the apparent affinity constant of the C1q:immune complex interaction and PEG concn.     

An analysis of the fluid phase C1q binding assay. The effect of endogenous C1q on the precipitation and detection of an immune complex model

We examined the effect of endogenous C1q on the sensitivity of the fluid-phase C1q binding assay (C1qBA) in detecting an immune complex (IC) model, heat-aggregated IgG (HAIgG), at concentrations of 10-10,000 micrograms/ml sample. Results in normal human serum (NHS) or plasma (NHP) were compared with those in heat-inactivated NHS (NHS/56) in which most endogenous C1q was depleted by heat denaturation. Higher HAIgG concentrations were required in NHP and NHS to produce the same 125I-C1q precipitation seen in NHS/56. This decreased sensitivity varied from 70% at low HAIgG concentrations to 0% at high concentrations, as predicted for a large pool of endogenous C1q, in equilibrium with 125I-C1q, but in excess of that which could bind to all but the highest concentrations of IC model. In serum depleted of functional C1q on an immunoadsorbant of HAIgG, the precipitation of radiolabeled HAIgG under C1qBA conditions was concentration dependent and generated a saturation curve, showing that only a fraction of IC are usually precipitated in this assay. HAIgG precipitation was enhanced 1.4-fold in NHS/56 (8 micrograms C1q/ml) and three-fold in NHS (67 micrograms C1q/ml) suggesting that IC size is increased by endogenous C1q. In dual label experiments using 131I-HAIgG, the precipitation of 125I-C1q in NHS/56 was directly proportional to IC model precipitation, but markedly discordant in NHP, showing the measurement of IC in heat-inactivated sera superior to that in native serum. A comparison of the C1q:HAIgG ratio in PEG precipitates with that in samples, indicated that equilibrium was established between C1q and IC model. Thus the precipitation of 125I-C1q in the C1qBA represents (1) the fraction of total C1q bound to IC, and (2) the fraction of IC precipitated by PEG.     

Antibody-independent interactions of fibronectin, C1q, and human neutrophils with Treponema pallidum.

Although recent evidence suggests that fibronectin may be involved in the attachment of treponemes to mammalian cells, its possible role in promoting phagocytosis of Treponema pallidum has not been investigated. In the present study, we examined the antibody-independent interactions of fibronectin, C1q, and human polymorphonuclear leukocytes with T. pallidum. Binding of [125I]fibronectin was specific and saturable with an affinity constant of approximately 2 X 10(7) M-1. The number of binding sites per treponeme at 37 degrees C, irrespective of the mammalian source of fibronectin, was between 2,500 and 7,500, with a mean of approximately 4,700. Binding of [125I]C1q to T. pallidum, in the absence of antibodies to the organism, also was saturable and specific. Pretreatment of treponemes with C1q enhanced binding of soluble [125I]fibronectin two- to threefold and also increased attachment of 125I-surface-labeled treponemes to fibronectin-coated surfaces. Treatment of 125I-labeled T. pallidum with fibronectin alone, or together with C1q, however, did not enhance surface phagocytosis by neutrophils.     

Nature of the interaction between the C1q and C1r2S2 subunits of the first component of human complement

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.     

C1q binding substances in pemphigus and bullous pemphigoid. Detection with a [131I] C1q binding assay.

A modification of the [125I]C1q binding assay was developed to allow the estimation of C1q binding activity (C1q BA) in pemphigus and bullous pemphigoid sera. The modifications include lower final concentration of PEG 6000 (1-5%) which permitted the use of sera that had been stored at -20 degrees C for extended periods of time; use of 131I instead of 125I and an [131I] C1q concentration of 5 microng/ml rather than 1 microng/ml. EDTA was used at a final concentration of 0-13 M to obviate the need for heat inactivation of sera. Sera from seventy-one patients with pemphigus and from 142 patients with bullous pemphigoid were tested for C1q BA. Of these 40% of the pemphigus and 20% of the bullous pemphigoid patients showed elevated C1q BA. A relationship between elevated C1q BA in serum and active disease was noted. Sequential samples from forty patients with pemphigus and thirty-seven patients with bullous pemphigoid demonstrated two different types of relationship between serum antibody titres to cutaneous antigens and C1a BA. In some patients serum antibody titres and C1q BA increased and decreased simultaneously; in others, increase of C1q BA followed increase of antibody titre and coincided with its decrease. The latter relationship supports the hypothesis that C1q BA may represent at least in part antigen-antibody complexes containing cutaneous antigens.     

Smooth muscle and epithelial cells express specific binding sites for the C1q component of complement.

In injury and inflammation, interactions of complement C1q with C1q receptors may provide attachment sites for cell localization and tissue regeneration. Cultured smooth muscle cells (58%), epithelial cells (26%), and endothelial cells (25%) attach to C1q-coated surfaces, while only 6% of cultured B cells (Raji) attach. Endothelial and Raji cells express C1q receptors, but C1q receptors (C1qR) on smooth muscle cells and epithelial cells have not previously been demonstrated. Evidence is provided that smooth muscle cells express an average of 1.5 x 10(6) C1qR/cell (K alpha = 10(8) M-1) and that epithelial cells express an average of 0.7 x 10(6) C1qR/cell (K alpha = 1.4 x 10(8) M-1). Binding properties of C1qR, and immunoreactivity to anti-C1qR antibodies, are characterized. The antibodies specifically recognize a 67-kDa component of smooth muscle cell lysates and inhibit cell attachment to C1q substrates. We conclude that distribution of C1qR may be ubiquitous; binding properties, size, and antigenicity of various C1qR may be related, but adhesive function may be tissue specific.     

Anti-C1q column: ligand specific purification of immune complexes from human serum or plasma. Analysis of the interaction between C1q and immune complexes

An efficient and reproducible procedure has been developed for the specific isolation of immune complexes. PEG precipitation of EDTA serum or plasma was an essential preliminary step to separate complex-bound from free C1q. PEG had no discernible effect on the molecular weight size of the extracted complexes. Redissolved complexes were incubated with a Sepharose-4B column coated with anti-human C1q antibodies and following removal of unbound material the bound complexes were sequentially eluted with 0.02 M EDTA, 0.5 M NaCl and 1 M propionic acid. Characteristics of the affinity column were established by the purification of 125I-labelled BSA-anti-BSA complexes and heat-aggregated IgG (HAGG) incubated in normal human serum (NHS). EDTA and NaCl eluted complexes were of similar molecular size and contained antigen, specific antibody, as well as human IgM, IgG, albumin, C3, C3c, C3d and C1q. Acid eluted complexes contained the highest yield of specific antigen and antibody and comprised in addition human C1q and C3d. Activation of complement components after C1q made the bond between C1q and immune complexes resistant to 0.5 M NaCl and interfered with the binding between solid phase anti-C1q and complex bound C1q. Using BSA-anti-BSA complexes and HAGG activated in NHS it was apparent that only a minority of the complexed material was isolated via the C1q ligand and this probably applies to the C1q binding assay. Most complexed material could be isolated using an anti-C3 affinity column.     

C1q binding and C1q deviation assays using enzyme labelled C1q

Methods for the enzyme labelling of C1q are described. The C1q conjugates have been incorporated into suitably modified C1q deviation and C1q binding assays for circulating immune complexes. These assays show the expected high positivity rate in sera from patients with systemic lupus erythematosus. The assays using enzyme labelled C1q confer considerable advantages over their counterparts using 125I.     

Short amino acid sequences derived from C1q receptor (C1q-R) show homology with the alpha chains of fibronectin and vitronectin receptors and collagen type IV.

The human C1q receptor (C1q-R) is a 65-70-kd, highly acidic, hydrophobic glycoprotein that is expressed on a wide variety of cell surfaces. Although the C1q-R itself appears to bind preferentially to C1q, the region of the ligand to which C1q-R binds is the primary binding site for several other molecules, including fibronectin, laminin, and C1q inhibitor (chondroitin 4-sulfate proteoglycan) as well as the complement C1r2C1s2 tetramer. In order to further characterize the C1q-R molecule with regard to its structure and function, highly purified C1q-R was obtained from Raji cells using DEAE-Sephacel and C1q-Sepharose CL-4B chromatography. Studies performed with 125I-labeled C1q-R demonstrated that whereas the C1q-R molecule binds poorly to a variety of human collagens including types II, III, and V, markedly enhanced binding is observed with type IV collagen and moderately enhanced binding with type I collagen. Amino acid composition studies show that the C1q-R molecule contains approximately 44% hydrophobic and 12.6% hydrophilic residues with a ratio of negatively charged to positively charged residues of about 2:1. Treatment of 125I-labeled C1q-R with endoglycosidase F lowers the apparent molecular size from 70 to 58 kd, whereas endoglycosidase H lowered the size to 64 kd. Treatment with neuraminidase, on the other hand, shifted the size of C1q-R to 60 kd. These results suggest the presence of several highly sialylated complex-type or high mannose-type N-linked oligosaccharide side chains. Because purified C1q-R has a blocked amino terminus, amino acid sequences representing internal fragments of the molecule were generated by electroblotting and in situ enzymatic digestion. When these short sequences were searched against the National Biomedical Research Foundation computer data base, a seven-amino-acid sequence, VSWQGQI, showed significant homology (100% and 80% in a five-amino-acid overlap, respectively) with the alpha chains of the human fibronectin (alpha 5 beta 1) and vitronectin (alpha v beta 3) receptors, and to a lesser degree with epidermal growth factor receptor and T cell receptor. A second sequence, ISEDNIR, showed homology with mouse collagen type IV (86% in a six-amino-acid overlap), calmodulin (60% in a seven-amino-acid overlap), and a Leishmania major surface antigen, gp63. These observations seem to predict that C1q-R has pockets of conserved sequences that are similar to those not only present in its ligand(s) but also in other cell surface receptors that may, in part, fulfill similar functions.     

A microplate adaptation of the solid-phase C1q immune complex assay

A method has been developed for the detection of C1q binding immune complexes in serum in which microculture plates are used as the solid-phase matrix for adsorption of C1q. This micromethod used only one-tenth of the amount of both C1q and [¹²⁵I]anti-human immunoglobulin per test and enabled 7 times as many samples to be tested in triplicate in comparison with the number performed in duplicate by the standard tube assay.¹²⁵I-labelled C1q studies showed that adsorption varied with the brand of microplate used, some types of plate binding up to 66% of the labelled material. This is considerably more than that bound by the polystyrene tubes generally used for this assay.The increased capacity of the method allowed the binding to plates coated with the same batch of C1q to be assessed at various times after storage for 8–10 weeks at both 4°C and ?70°C. A marked decrease in binding to C1q of aggregates of IgG was observed on storage for longer periods. Results with different batches of aggregated IgG which had been ultracentrifuged suggested that this may be used as an effective standard, stable on storage at ?70°C.A comparison with the standard tube method of aggregated IgG, normal control sera and sera from patients with SLE showed that the micromethod adaptation of the C1q solid-phase binding radioimmunoassay is more economical and easier to perform and does not impair accuracy or standardization.     

Dissociation of C1 and concentration dependence of its activation kinetics

The activation of the zymogen C1s to the enzyme C1s in the human C1 complex [C1q(c1rC1s)2] was studied as a function of the concentrations of (C1rC1s)2 and C1q which were saturated with oligomers of rabbit IgG. A large concentration dependence of the sigmoidal kinetics was observed in the 2-180 nM concentration range. This was explained by association-dissociation equilibria between the antibody-saturated C1q and various forms of the (C1rC1s)2 complex (unactivated to activated). The establishment of these equilibria (binding constant 2 x 10(7) M-1) was assumed to be fast as compared to the rates of the activation steps (rate constants 10(-3) and 10(-2) sec-1 at 30 degrees C). The fast re-equilibration of the C1 complex explains the finding that small amounts of antibody-saturated C1q catalysed the activation of large amounts of C1s. The interpretation of the kinetic results was supported by a direct demonstration of the dissociation of C1 into C1q and (C1rC1s)2 by analytical and density gradient centrifugation. No difference was found between the rates of activation and the dissociation properties of reconstituted C1 and C1 isolated from serum.     

Co-operation between the pair of C gamma 2 domains in Clq-binding by rabbit IgG

The single site binding constants of rabbit IgG and its plasmin-derived fragments F(acb)2, Facb and F(ab)2 for human C1q were measured by the sedimentation velocity method. The intact IgG and F(acb)2 having the paired C gamma 2 domains gave an identical association constant at 20 degrees C (Ka) of 3.02 X 10(4) M-1 in the presence of a physiological concn of salt and on the basis of six sites per C1q. The C1q-binding affinity was found to be decreased to 1.04 X 10(4) M-1 in the reduced, monomerized fragment Facb. Under the same conditions F(ab)2, which is completely unable to activate the classical complement cascade, gave an apparent C1q-affinity of 0.36 X 10(4) M-1. The results, together with previous observations, led us to the conclusion that the C1q-binding site of rabbit IgG is constituted associatively by the pair of C gamma 2 domains, each of which providing a limited, complementary part of the binding free energy between IgG and C1q.     

On the Interaction of the First Complement Component C1 and its Subunit Clq with Solid-Phase IgM Immune Complexes

The interaction of C1 and C1q with solid-phase anti-dextran MOPC-104E IgM was studied. An enzyme-linked immunosorbent assay (ELISA) was used to detect bound C1q. The results revealed that immobilized IgM is converted to the functionally active 'staple' conformation by the specific polyvalent ligand dextran (B 1355/S). C1q is fixed to IgM dependent on the antigen concentration, and its binding might be explained by assuming a functional binding constant (K) of approximately 10(9) M-1. Molecules bound with a K in the range of 10(7) M-1 cannot be detected by this ELISA procedure. The fixation of C1q saturated with an excess of the C1r2S2-tetramer differs from that of free C1q. C1q incorporated in the C1 complex rapidly dissociates independently of the antigen concentration. Since the complement binding sites are located at definite positions on the IgM molecule because of its pentameric structure, it is suggested that the distinguishable association properties of C1 and C1q are brought about from the altered flexibility of the C1q molecule complexed with C1r2S2.     

Reversible biotinylation of C1q with a cleavable biotinyl derivative Application in C1q receptor (C1qR) purification

Reversible biotinylation of human C1q without impairment of its physiologic functions has allowed us to develop a simple and rapid purification method for C1q receptor (C1qR). The biotinylating reagent, NHS-SS-biotin (M_r 606.7) contains an extended connector or cross-linker arm which limits steric hindrance and is bridged by a cleavable disulfide bond to the biotin component. Biotinylation was achieved by mixing C1q (in PBS, pH 7.4) with NHS-SS-biotin (dissolved in dimethyl formamide) in a 50:1 v/v and 1:25 mol/mol ratio and allowing the reaction to continue at room temperature for 4 h. The mixture was then dialyzed against PBS pH 7.4 (2 × 1 liter) and analyzed by SDS-PAGE and hemolytic assay using C1q depleted serum. Under these conditions neither denaturation of the protein nor loss of hemolytic activity was evident. Such biotinylated C1q (Bio-C1q) was used to pull out the C1qR from detergent-solubilized (1% NP-40 in PBS, pH 7.4 plus inhibitors) ¹²⁵I-surface labeled membrane solution that had been first centrifuged (1 h, 45 000 × g, 4°C) and then sequentially precleared with immobilized protein A, protein A-IgG and gelatin. The mixture of Bio-C1q and membrane solution was then incubated (20 h, 4°C), applied to immobilized avidin (equilibrated with PBS, pH 7.4, 0.1% NP-40) and after washing, the bound C1qR was eluted with equilibrating buffer containing 1 M NaCl, and the C1q by same buffer containing 100 mM DTT. The eluted C1qR contained a major M_r 70 000 molecule which upon reduction electrophoresed with an apparent M_r of 85 000–90 000 as assessed by SDS-PAGE analysis. In addition, a faint single chain band of 30–40 kDa was eluted with the major band and may represent a non-covalently associated part of the C1qR molecule.