- Email: [email protected]
EBV receptor (CR2) of human B lymphocytes
Molecular
/nnww/o~J~.
Vol. 23.
No.
I I. pp. 1249-1253,
1986
Olhl-5890186 Pergamon
Printed in Great Britain.
$3.00 + 0.00 Journals Ltd
STRUCTURE AND FUNCTIONS OF GP 140, THE C3d/EBV RECEPTOR (CR2) OF HUMAN B LYMPHOCYTES RAYMOND FRADE*
Laboratoire
de Biochimie des Anti&es
de Membrane, INSERM U. 139. H6pital Henri Mondor. 94 010 Creteil, France
(Received
12 June 1986)
Abstract-Analysis of the interaction of human C3 fragments with human B lymphoma cell line, led us to isolate gp 140. the C3 receptor of Raji cells. Rabbit anti-gp 140 was prepared against this highly purified receptor. Using these polyclonal antibodies, it was found that: (a) gp 140 is the C3d receptor (CR2) which reacts with the C3d site expressed on C3d. C3dg. C3bi and at a less extent on C3b. Gp 140 is a specific marker of human B lymphocytes; (b) gp 140 is also the Epstein-Barr virus receptor (EBVR): (c) CR2 is a membrane site involved in B-cell regulation; and (d) the C3d/C3dg receptor (CR2) of human B lymphocytes is distinct to the C3dg receptor (CR4) of human neutrophils.
INTRODUCTION Among all complement components, C3, the third component. appears as one of the main plasma proteins by its pivotal place in the two complement activation pathways, i.e. the alternative and the classical pathway, and by its elevated concn, in the range of 1.2-1.5 mg/ml (Fearon et al., 1976; MiillerEberhard and Schreiber, 1980). C3 is a 185,000 mol. wt glycoprotein, consisting of two chains: the 115,000 alpha chain and the 70,000 beta chain, linked by disulfide bridges (Zemel et al., 1980). During activation of human C3, through classical or alternative pathways, different fragments were successively generated: C3a and C3b, then C3bi and finally C3c plus C3dg; C3dg is cleaved into C3d and C3g (Minta et al., 1977; Fontaine and Rivat, 1979; Harrisson and Lachmann, 1980). C3 is under the control of serine proteinases as factor I, whose reaction with C3b is mediated by factor H (Lambris et al., 1980). Activated C3 fragments are able to react with the cell surface through either an acceptor binding site or receptor binding sites. The acceptor interaction involves the thiol-ester bond. carried by the alpha chain, which is hydrolyzed during the first step of activation of C3 into C3a and C3b. Such hydrolysis generates an unstable acyl group which reacts with hydroxyl groups present either in the water, to lead to soluble C3b, or on carbohydrate groups, carried by immunocomplexes, bacteria, yeast walls or cell surfaces, to generate particle-bound C3b (Law and Levine, 1977; Tack et al.. 1980: Mi.iller-Eberhard and Schreiber, 1980). Such acceptor interaction is a covalent binding distinct from the receptor interactions. *Present address: Laboratoire de Biochimie des Antigenes de Membrane. INSERM U.23. H6pital Saint Antoine. 75012 Paris. France.
Non-covalent binding depends on the polymerization degree of C3 fragments and on the nature of C3 receptors. Molecular analysis of C3 receptors have been conducted by following two complementary studies: (a) by kinetics studies of C3 interactions with cell surface; and (b) isolation and/or characterization of C3 receptors molecules. Since 1980, five types of C3 receptors were identified: C3a receptor; C3b receptor or CR]; C3bi receptor or CR3; C3d/C3dg receptor or CR2; and C3dg receptor or CR4. Each of these receptors reacts with a specific site expressed mainly, but not exclusively, on each type of C3 fragments mentioned above. In this paper data on the C3djC3dg receptor (CR2) of human B lymphocytes and on the C3dg receptor (CR4) of human neutrophils is reported on. RESULTS AND DISCUSSION Interactions of C3 fragments with human cell surface need polymeric conformations
Interaction of “‘I-C3b, was originally analyzed by Frade and Strominger (1980) on the human B lymphoma cell Raji. This interaction is characterized by an affinity constant of lo6 M-’ for soluble C3b and of 10’ M-’ for particle-bound C3b (Frade and Strominger, 1980, 1981). Similar data were found by others. Berger er al. (1981) calculated on affinity constant of 1.5 x 10’ M-’ for particle-bound C3b on human neutrophils, while Arnaout ec al. (198 1) found affinity constants of lo6 M-l for monomeric C3b and of 2 x 10’ M-’ for polymerized C3b binding on human erythrocytes and B lymphocytes. The physiological meaning of such affinity constants is supported by the data that polymeric C3b, but not soluble C3b, could induce histaminase release on human granulocytes (Melamed et al., 1982). Moreover, soluble C3b form is more likely physiologically irrelevant since. under the control of factor H and factor I, it is instantaneously cleaved in C3bi then successively in C3dg and in C3d.
1249
1250
RAYMOND FRADE
Thus. it appears from these data. that C3b binding with the ceil surface needs polymerized C3b, or C3b associated to others proteins or particles, rather than monomeric soluble C3b. Such polymeric conformation is also needed for the interaction of C3d with the cell surface (Melchers et al., 1985). GP 140, a C3blC3dglC3d binding membrane protein, is the C3d receptor (CR,?)
Identification of CR2 was undertaken by following two different methods: one was to isolate by afhnity chromatography on sepharose bound C3b the C3 receptor of Raji cells, to characterize it and to prepare successively polyclonal and monoclonal antibodies against this highly purified receptor; the other involved using monoclonal antibodies prepared randomly against the whole B-lymphocyte cells. The first method was undertaken by us in 1981. Cell-surface antigens were labeled by 12%odine using the lactoperoxidase method, then membranes were prepared. Membrane components were solubilized in presence of NP-40 detergent. We have shown that a l~,O~ mol. wt membrane component could be isolated by affinity chromatography on sepharose bound C3b. This membrane component is constituted by one single chain which, under non-reducing or reducing conditions, carries a carbohydrate moiety and conserves its binding properties on sepharose bound C3b after solubilization. We called it gp 140 (Bare1 et uf., 1981). This membrane component reacts with particle-bound C3b (Bare1 et al., 1981) and iZSI-labeled soluble C3b (Charriaut et ul., 1982), supporting its C3b receptor like properties, while the binding site of the C3b molecule involved- in its interaction with gp 140 was not determinated. Cole er al. (1983) confirmed later that gp 140 acts as a C3b binding membrane protein. It should be noted that gp 140 is distinct from the C3b/C4b receptor, CRl, which was isolated from human erythrocyte membranes and associated to a 190,000-250,000 membrane glycoprotein (Fearon, 1980: Dobson et al., 1981; Dykman et al., 1984). Moreover, the binding of C3b on Raji celfs is highly controversial. For instance. while we (Frade and Strominger, 1981) and others (Gaither et al., 1983; Dierich er al., 1983) have described particles-C3b rosettes formation with Raji cells, Dobson et al. (1981) and Tedder et al. (19836) have shown, using a monoclonal antibody against the C3b receptor extracted from human erythrocytes, that CRI is not expressed on the Raji cell surface. However, Weis et al. (1984) also found later EC3b rosettes formation with Raji cells. Dierich et u!. (1983) have mentioned that Raji cells could secrete factor H activity, a 150,000 serum co-factor of factor I proteolysis of C3b. These data and the close mol. wts of H and gp 140, prompted us to analyze the relationship between both molecules. Polyclonal antibodies were prepared by injecting
highly purified gp 140 in a rabbit. For instance. 20 jig of gp 140 were purified from 20 g of Raji cells (Frade et al., 1984).
Up to this point we have shown that gp 140 is not related to factor H and that C3b binding to gp 140 is not mediated by H. These results were supported by: (a) cytofluorimetry analysis of rabbit anti-gp 140 and rabbit anti-H binding on Raji cell surface: (b) comparing the inhibition of specific cytotoxic assays by each antigen; (c) measuring the inhibition by each antibody of the specific binding of soluble or particle bound C3b to the cell surface; and (d) analyzing the binding of each antibody on solubiiized antigens using immunoblotting techniques, after polyacrylamide gel electrophoresis in absence or in presence of sodium dodecyl suIfate (Frade er al.. 1984). Thus, gp 140 appeared as a peculiar C3b binding protein and its relationship with others C3 receptors was analyzed. It should be noted that as early as in 1983, we had evidence that gp 140, while reacting as a C3b receptor as mentioned above, acted as a C3d receptor and that gp 140 is a specific marker of human peripheral B l~phocytes. We reported these preliminary data at the Xth International Complement Workshop in Mainz, in Mai 1983 (Frade et al., 1984). We have shown that polyclonal anti-gp 140 inhibited Raji cell rosettes with EC3b, EC3bi. EC3dg and EC3d and also blocked normal lymphocyte rosettes with EC3dg and ECM, without affecting CR1 and CR3 activity. Moreover, a monoclonal anti-C3 (C3b/130) described by others as reacting with the d region highly expressed on EC3bi, EC3dg and EC3d and poorly exposed on EC3b, completely inhibited EC3bi, EC3dg and EC3d rosettes with Raji cells, but had not effect on EC3b rosettes (Frade er al., 198%). Thus, gp 140, originally isolated by us as a C3b-like receptor, is the C3d receptor (CR2) of human B cells. Gp 140 reacts with the C3d site highly expressed on C3d, C3dg and C3dg fragments and to a lesser extent with sepharose-bound C3b or “*I-labeled C3b, supporting our previous data. Cellular distribution of gp 140 and of C3d-receptor activity restricted to different human B-cell lines, and to normal B lymphocytes, prompted others, in 1983, to analyze whether monoclona1 antibodies prepared randomly against whole human B cells, and selected as specific marker of B lymphocytes, could recognize the C3d receptor. Thus, three monoclonal antibodies were successively identified as monocfonal anti-C3d receptors: B2 (Nadler et al., 1981), HB-5 (Tedder et al., 19830) and OKB-7 (Nemerow CI al., 1986b). Iida er al. (1983) and Weis et al. (1984) identified a 140,000 and a 145,000 glycoprotein, as a C3d receptor. using the monoclonal antibodies (mAb) anti-B2 and HB-5. respectively. However, anti-B2 and HB-5 inhibited ECM rosettes poorly, unless cross-linked with goat anti-mouse IgG. Nemerow et crl. (19856) have shown that the monoclonal OKB-7. which immu-
CR2. the C3d/EBV receptor
noprecipitates a I~.000 membrane antigen. inhibits directly CXdg-microsphere rosettes. It was found that treatment of Raji cells by rabbit anti-gp 140 blocked the upvdke of the three
125l
EBV binding. Polyclonat anti-gp 140 IgCi directly inhibits the binding of EBV to Raji cells at the same concn that the binding of EC3d on cells is inhibited, whereas a 35 times higher concn of anti-gp 72 IgG or ‘Z51-labeled monoclonal antibodies anti-B2, HIS.5 and preimmune serum IgG does not. Anti-gp 140 IgG GKB-7, indicating that each of these monoclonal treatment also inhibits the induction of EBVantibodies recognized epitopes present on gp 140 determined nuclear antigen in normal tonsil B lym(Frade et af., 198%). phocytes or in EBV-negative Ramos cells, whereas high concns of anti-gp 72 IgG or prcimmune serum The C3dlC3dg receptor qf ltui~~u~B [~~~pi~~)cstes IgG have no effect (Frade ef ul., 198%). (CR2) is distinct to the C3dg receptor of’ hunwn Fingeroth er al. (1984) have shown that solubilized neutrophils (CR4) membrane antigen recognized by the monoclonal HB-5, which reacts with CR2. binds EBV particles. The neutrophil C3dg receptor was examined to More recently, Nemerow et al. (19856) have found determine its relationship to lymphocyte CR2. While that monoclonal OKB-7 inhibits both ECM and neutrophil rosettes with EC3d were undetectable, a specificity for C3d was suggested by the inhibition of EBV binding on CR2. EC3dg rosettes by fluid phase C3d-complexes bearing These data support strongly that gp 140, the C3d no detectable C3dg. However, ‘such neutrophil receptor (CR2) of human B lymphocytes, is also the EC3dg and EC3bi rosettes were not inhibited by EBVR. Meanwhile, the effect of rabbit anti-gp 140 IgG on rabbit anti-gp 140 nor an excess of anti-CRI, antithe B cell proliferative response was tested. Purified CR2 and anti-CR3 added together. In addition, B cells from human blood were induced to proliferate neutrophils did not bind “51-labeled anti-gp 140, antiby partially purified B-cell growth factor (BCGF)B2, HB-5 or OKB-7 (Frade et al., 198%). Recently, containing supernatant from activated T cells. The Vik and Fearon (1985) have shown that the binding anti-C3d receptor Ffab’)? enhanced the BCGFof Iabeled C3dg on human neutrophils is different to dependent B-cell proliferation. The effect was dosethe one on human B lymphocytes. dependent and was observed in the presence of Thus, the neutrophil C3dg receptor is distinct from different concns of BCGF, it did not correspond to gp 140, the B lymphocyte CR2, and should be desiga change in the time course of the response. The nated CR4. anti-C3d receptor F(ab’)2 had no mitogenic effect in Functions of GP 140, the C3d receptor (CR2) the absence of T-cell supernatant. In contrast, the Jondal et al. (1976) suggested. early in 1976, that undigested antiC3d receptor IgG suppressed the a structural relationship existed between the receptor BCGF-dependent B-cell proliferation. Our results for the Epstein-Barr (EBVR) and the receptors for emphasize the involvement of the C3d receptor in the the third component of complement (C3) expressed regulation of B-cell response to T-cell products on the human B-lymphocyte surface. For instance, (Frade er al., 1985~). erythrocyte antibody complement, but not erythroRecently, Nemerow et aI. (1985~) have also found cyte antibody, rosette formation was inhibited by that monoclonal anti-CR2 induces activation and EBV preparations (Jondal et al., 1976). In a large differentiation of human B lymphocytes. Moreover. it panel of human-cell lines tested, C3 receptors and was found by others that interaction of particleEBVR were either co-expressed or simultaneously bound C3d (Melchers et al., 1985) or U.V.inactivated absent (Jondal et al., 1976). C3 receptors and EBVR Epstein-Barr virus particles (Masuchi e? al., 1985) co-capped with each other but not with other memwith CR2 leads to the enhancement of B lymphocytes brane antigens, as IgM, B,-microglobulin and Fc proliferation. receptor (Yefenof et al., 1976). The association beThese data support strongly that cross-linking of tween EBVR and C3 receptors on the surface of gp 140, the C3d/EBV receptor (CR2), at the cell human B-lymphoma lines was also supported by surface by probes as F(ab’), anti-gp 140, particlemembrane-receptor stripping (Yefenof et al., 1977). bound C3d or inactivated EBV particles induces More recently, among C3 receptors, an association post-membrane signals which are involved in the has been suggested between EBVR and C3d receptor regulation of B-cell activation. Among the molecular (CR2) rather than C3b receptor (CRl) (Jonsson et events associated to “in vitro*’ activation of human al., 1982). peripheral B lymphocytes, we have recently found We have analyzed the relationship between gp 140, that gp 140, the CR2, is phosphorylated (Bare1 er al., the C3d receptor, and the EBVR using rabbit anti1986). gp 140 IgG. Polyclonal anti-gp-72, a C3-binding membrane component not related to the EBVR but REFERENCES also expressed on the Raji cell surface. was used as a Arnaout M. A., Melamed J., Tack B. F. and Colten H. R. control. Binding of rabbit IgG and EBV on cells was (1981) Characterization of the human complement fC3b) assessed by using flow cytofluorimetry. A semireceptor with a fluid phase dimer. 1. ~~i~~7~~7. 127, 1348-1354. quantitative bioassay was also used to measure the
1252
RAYMOND
Bare1 M., Charriaut C. and Frade R. (1981) Isolation and characterization of a C3b receptor-like molecule extracted from membranes of human B lymphoblastoid cell (Raji). FEBS Lelf. 136, 11l-1 14. Bare1 M., Vazquez A., Charriaut C., Aufredou M. T., Galanaud P. and Frade R. (1986) Gp140. the C3d/EBV receptor (CR2) is phosphorylated upon in oirro activation of human peripheral B lymphocytes. FEBS Letr. 197, 353-356. Berger M.. Gaither T. A., Hammer C. H. and Frank M. M. (1981) High affinity interaction of fluid phase C3b but not C3 with the neutrophil C3b receptor. Fedn Proc. 40, 44614469. Charriaut C., Bare1 M. and Frade R. (1982) Behaviour of soluble human “‘1 C3b, the third component of complement. after binding to human cells. Eur. J. Immun. 12, 289-294. Cole J. L., Dykman T. L., McDermott R. P. and Atkinson J. P. (1983) Isolation of 3 distinct C3b binding proteins (C3b-BP) from human leukocytes. Immunobiology 164, 223.
Dierich M. P., Scheiner 0.. Mussel H. H., Hammer K. P., Schopf R. E. and Schultz T. (1983) Characterization of complement receptors. Molec. Immun. 19, 1255-1265. Dobson N. J., Lambris J. D. and Ross G. D. (1981) Characteristics of isolated erythrocyte complement receptor type one (CRlC4bC3b receptor) and CR1 specific antibodies. J. Immun. 126, 693-698. Dykman T. R., Hatca T. A. and Atkinson J. P. (1984) Polymorphism of the human C3b/C4b receptor. J. exp. Med. 159, 691-703. Fearon D. T. (1980) Identification of the membrane glycoprotein that is the C3b receptor of the human erythrocyte, polymorphonuclear leukocyte. B lymphocyte and monocyte. J. exp. Med. 152, 20-30. Fearon D. T., Daha J., Weiler J. M. and Austen K. F. (1976) The natural modulation of the amplification phase of complement activation. Transplant Rev. 32, 12-25. Fingeroth J. D., Weis J. J., Tedder T. F., Strominger J. L., Biro P. A. and Fearon D. T. (1984) Epstein-Barr virus receptor of human B lymphocytes is the C3d receptor CR2. Proc. nam. Acad. Sri. U.S.A. 81, 451&4514. Fontaine M. and Rivat C. (1979) A study of the breakdown of the third component of human complement. Analyt. Immun. 13OC, 349-366.
Frade R.. Bare1 M.. Ehlin-Henriksson B. and Klein G. (1985a) Gp 140, the C3d receptor of human B lymphocytes, is also the Epstein-Barr virus receptor. Proc. narn. Acad. Sci. U.S.A. 82, 1490-1493.
Frade R., Bare1 M.. Krikorian L. and Charriaut C. (1984) Analysis of gp 140, a C3b binding membrane component present on Raji cells: a comparison with H. Eur. J. Immun. 14, 542-548. Frade R., Crevon M. C., Bare1 M., Vasquez A., Krikorian L., Charriaut C. and Galanaud P. (19856) Enhancement of human B cell proliferation by an antibody to the C3d receptor, the gp 140 molecule. Eur. J. Immun. 15, 73-76. Frade R., Myones B., Bare1 M.. Krikorian L., Charriaut C. and Ross G. D. (1985~) Gp 140, a C3b binding membrane component of B lymphocytes is the C3djC3dg receptor (CR2) and is distinct from the neutrophils C3dg receptor (CR4). Eur. J. Immun. 15, 1192-1197. Frade R. and Strominger J. (1980) Interaction of soluble “‘1 human C3b. the third component of complement, to specific receptors in human cultured B lymphoblastoid cells: characterization of a low affinity interaction. J. Immun. 125, 1332-1339.
Frade R.
and Strominger J. (1981) Interaction of particlebound ‘1SI-C3b. the third component of complement with specific receptors on human B lymphoblastoid cells (Raji). Immun. Lerl. 3, 173 178. Gaither T. A.. Magrata I. T., Berger M.. Hammer C. H.. Novikos L., Santaclla M. and Frank M. M. J. (1983)
FRADE
Complement receptor expression by neoplastic and normal human cel1s.J. Imrkn. 131, 899-905. Harrison R. A. and Lachmann P. J. (1980) The physiological breakdown of the third component of human complement. Molec. Immun. 17, 9-20. Hutt-Fletcher L. W., Fowler E., Lambris J. D.. Simmons J. G. and Ross G. D. (1983) Studies of the Eostein-Barr virus receptor found on Raji cells. II. A comparison of lymphocyte binding sites for Epstein-Barr virus and C3d. J. Immun. 130, 1309-1312. Iida K., Nadler L. and Nussenzweig V. (1983) Identification of the membrane receptor for the complement fragment C3d by means of a monoclonal antibody. J. exp. Med. 158, 1021-1033.
Jondal M., Klein G., Oldstone M. B. A., Bokish V. and Yefenof E. (1976) Surface markers on human B and T lymphocytes. VIII: association between complement and Epstein-Barr virus receptors on human lymphoid cells. Stand. J. Immun. 5, 401410.
Jonsson V.. Wells A. and Klein G. (1982) Receptors for the complement C3d component and the Epstein-Barr virus are quantitatively coexpressed on a series of B-cell lines and their derived somatic cell hybrids. Cell. Immun. 72, 263-276. Lambris J. D., Dobson N. J. and Ross G. (1980) Release of endogeneous C3b inactivator from lymphocyte in response to triggering membrane receptors for BIH globulin. J. exp. Med. 152, 1625-1644. Law S. K. and Levine R. P. (1977) Interaction between the third complement protein and cell surface macromolecules. Proc. natn. Acad. Sri. U.S.A. 74, 2701-2705. Masuchi M. G., Szigeti R., Ernberg I. and Klein E. (1985) B lymphocyte activation induced by non-transforming Epstein-Barr virus. 7th Europ. Immun. Meeting, Israel, p. 158. Melamed J., Arnaout M. A. and Colten H. R. (1982) Complement (C3b) interaction with the human granulocyte receptor: correlation of binding of fluid-phase radiolabeled ligand with histaminase release. J. Immun. 128, 2313-2318. Melchers F., Erdei A., Schultz T. and Dierich M. P. (1985) Growth control of activated, synchronized murine B cells by the C3d fragment of human complement. Nature, Lond. 317, 264-267.
Minta J. O., Man D. and Morat H. R. (1977) Kinetics studies on the fragmentation of the third component of complement by trypsin. J. Immun. 118, 2192-2196. Miiller-Eberhard H. J. and Schreiber R. D. (1980) Molecular biology and chemistry of the alternative pathway of complement. Adv. Immun. 29, l-53. Nadler L. M., Stashenko P., Hardy R., Van Agthoven A., Terhorst C. and Schlossman S. F. (1981) Characterization of a human B cell specific antigen B2 distinct from BI. J. Immun. 126, 1941-1943. Nemerow G. R., McNaughton M. E. and Cooper N. R. (19850) Binding of monocional antibody to the Epstein-Barr virus (EBV)/CR2 receptor induces activation and differentiation of human B lymphocytes. J. Immun. 135, 3068-3073. Nemerow G. R., Wolfert R., McNaughton M. and Cooper N. R. (1985b) Identification of the Epstein-Barr virus (EBV) receptor and its relationship to the complement C3d receptor (CR2). Fedn Proc. Fedn Am. Sots exp. Biol. 44, 3376.
Tack B. F., Harrison R., Janatova J., Thomas M. L. and Prahl J. W. (I 980) Evidence for the presence of an internal thioester in the third component of complement. Proc. nom. Acad. Sci. U.S.A. 77, 5764-5768.
Tedder T. F., Clement L. T. and Cooper M. D. (1983a) Use of monoclonal antibodies to examine differentiation antigens on human B cells. Fedn Proc. Fedn Am. Sots exp. hiol. 42, 718.
Tedder T. F.. Fearon D. T.. Gartland G. L. and Cooper
CR2,
the C3d/EBV
M. D. (19836) Expression of C3b receptor on human B cells and myelomocytic cells but not natural killer cells. J. /nmun. 130, 166881671. Vik D. P. and Fearon D. T. (1985) Neutrophils express a receptor for iC3b. C3dg and C3d that is distinct from CRI. CR2 and CR3. J. Immure. 134, 2571 -2579. Weis J. J., Tedder T. F. and Fearon D. T. (1984) Identification of a 145.000 Mr membrane protein as the C3d receptor (CR2) of human B lymphocytes. Proc. natn. Acad. Sri. U.S.A. 81, 881-885. Yefenof E.. Klein G.. Jondal M. and Oldstone M. B. A., (I 976) Surface markers on human B and T-lymphocytes.
receptor
1253
IX. Two-color immunofluorescence studies on the association between EBV receptors and complement receptors on the surface of lymphoid cell lines. hf. J. Cancer 17, 693-700. Yefenof E., Klein G. and K. Kvarnung (I 977) Relationship between complement activation, complement binding and EBV absorption by human hematopoetic cell lines. Cell. Immun. 31, 225-233. Zemel E. S., Cartwright-Harwick C. M. and Nilsson U. R. (1980) Effect of reduction on the structural, biologic and immunologic properties of the third component of human complement. J. Immun. 125, 1099-I 103.
/nnww/o~J~.
Vol. 23.
No.
I I. pp. 1249-1253,
1986
Olhl-5890186 Pergamon
Printed in Great Britain.
$3.00 + 0.00 Journals Ltd
STRUCTURE AND FUNCTIONS OF GP 140, THE C3d/EBV RECEPTOR (CR2) OF HUMAN B LYMPHOCYTES RAYMOND FRADE*
Laboratoire
de Biochimie des Anti&es
de Membrane, INSERM U. 139. H6pital Henri Mondor. 94 010 Creteil, France
(Received
12 June 1986)
Abstract-Analysis of the interaction of human C3 fragments with human B lymphoma cell line, led us to isolate gp 140. the C3 receptor of Raji cells. Rabbit anti-gp 140 was prepared against this highly purified receptor. Using these polyclonal antibodies, it was found that: (a) gp 140 is the C3d receptor (CR2) which reacts with the C3d site expressed on C3d. C3dg. C3bi and at a less extent on C3b. Gp 140 is a specific marker of human B lymphocytes; (b) gp 140 is also the Epstein-Barr virus receptor (EBVR): (c) CR2 is a membrane site involved in B-cell regulation; and (d) the C3d/C3dg receptor (CR2) of human B lymphocytes is distinct to the C3dg receptor (CR4) of human neutrophils.
INTRODUCTION Among all complement components, C3, the third component. appears as one of the main plasma proteins by its pivotal place in the two complement activation pathways, i.e. the alternative and the classical pathway, and by its elevated concn, in the range of 1.2-1.5 mg/ml (Fearon et al., 1976; MiillerEberhard and Schreiber, 1980). C3 is a 185,000 mol. wt glycoprotein, consisting of two chains: the 115,000 alpha chain and the 70,000 beta chain, linked by disulfide bridges (Zemel et al., 1980). During activation of human C3, through classical or alternative pathways, different fragments were successively generated: C3a and C3b, then C3bi and finally C3c plus C3dg; C3dg is cleaved into C3d and C3g (Minta et al., 1977; Fontaine and Rivat, 1979; Harrisson and Lachmann, 1980). C3 is under the control of serine proteinases as factor I, whose reaction with C3b is mediated by factor H (Lambris et al., 1980). Activated C3 fragments are able to react with the cell surface through either an acceptor binding site or receptor binding sites. The acceptor interaction involves the thiol-ester bond. carried by the alpha chain, which is hydrolyzed during the first step of activation of C3 into C3a and C3b. Such hydrolysis generates an unstable acyl group which reacts with hydroxyl groups present either in the water, to lead to soluble C3b, or on carbohydrate groups, carried by immunocomplexes, bacteria, yeast walls or cell surfaces, to generate particle-bound C3b (Law and Levine, 1977; Tack et al.. 1980: Mi.iller-Eberhard and Schreiber, 1980). Such acceptor interaction is a covalent binding distinct from the receptor interactions. *Present address: Laboratoire de Biochimie des Antigenes de Membrane. INSERM U.23. H6pital Saint Antoine. 75012 Paris. France.
Non-covalent binding depends on the polymerization degree of C3 fragments and on the nature of C3 receptors. Molecular analysis of C3 receptors have been conducted by following two complementary studies: (a) by kinetics studies of C3 interactions with cell surface; and (b) isolation and/or characterization of C3 receptors molecules. Since 1980, five types of C3 receptors were identified: C3a receptor; C3b receptor or CR]; C3bi receptor or CR3; C3d/C3dg receptor or CR2; and C3dg receptor or CR4. Each of these receptors reacts with a specific site expressed mainly, but not exclusively, on each type of C3 fragments mentioned above. In this paper data on the C3djC3dg receptor (CR2) of human B lymphocytes and on the C3dg receptor (CR4) of human neutrophils is reported on. RESULTS AND DISCUSSION Interactions of C3 fragments with human cell surface need polymeric conformations
Interaction of “‘I-C3b, was originally analyzed by Frade and Strominger (1980) on the human B lymphoma cell Raji. This interaction is characterized by an affinity constant of lo6 M-’ for soluble C3b and of 10’ M-’ for particle-bound C3b (Frade and Strominger, 1980, 1981). Similar data were found by others. Berger er al. (1981) calculated on affinity constant of 1.5 x 10’ M-’ for particle-bound C3b on human neutrophils, while Arnaout ec al. (198 1) found affinity constants of lo6 M-l for monomeric C3b and of 2 x 10’ M-’ for polymerized C3b binding on human erythrocytes and B lymphocytes. The physiological meaning of such affinity constants is supported by the data that polymeric C3b, but not soluble C3b, could induce histaminase release on human granulocytes (Melamed et al., 1982). Moreover, soluble C3b form is more likely physiologically irrelevant since. under the control of factor H and factor I, it is instantaneously cleaved in C3bi then successively in C3dg and in C3d.
1249
1250
RAYMOND FRADE
Thus. it appears from these data. that C3b binding with the ceil surface needs polymerized C3b, or C3b associated to others proteins or particles, rather than monomeric soluble C3b. Such polymeric conformation is also needed for the interaction of C3d with the cell surface (Melchers et al., 1985). GP 140, a C3blC3dglC3d binding membrane protein, is the C3d receptor (CR,?)
Identification of CR2 was undertaken by following two different methods: one was to isolate by afhnity chromatography on sepharose bound C3b the C3 receptor of Raji cells, to characterize it and to prepare successively polyclonal and monoclonal antibodies against this highly purified receptor; the other involved using monoclonal antibodies prepared randomly against the whole B-lymphocyte cells. The first method was undertaken by us in 1981. Cell-surface antigens were labeled by 12%odine using the lactoperoxidase method, then membranes were prepared. Membrane components were solubilized in presence of NP-40 detergent. We have shown that a l~,O~ mol. wt membrane component could be isolated by affinity chromatography on sepharose bound C3b. This membrane component is constituted by one single chain which, under non-reducing or reducing conditions, carries a carbohydrate moiety and conserves its binding properties on sepharose bound C3b after solubilization. We called it gp 140 (Bare1 et uf., 1981). This membrane component reacts with particle-bound C3b (Bare1 et al., 1981) and iZSI-labeled soluble C3b (Charriaut et ul., 1982), supporting its C3b receptor like properties, while the binding site of the C3b molecule involved- in its interaction with gp 140 was not determinated. Cole er al. (1983) confirmed later that gp 140 acts as a C3b binding membrane protein. It should be noted that gp 140 is distinct from the C3b/C4b receptor, CRl, which was isolated from human erythrocyte membranes and associated to a 190,000-250,000 membrane glycoprotein (Fearon, 1980: Dobson et al., 1981; Dykman et al., 1984). Moreover, the binding of C3b on Raji celfs is highly controversial. For instance. while we (Frade and Strominger, 1981) and others (Gaither et al., 1983; Dierich er al., 1983) have described particles-C3b rosettes formation with Raji cells, Dobson et al. (1981) and Tedder et al. (19836) have shown, using a monoclonal antibody against the C3b receptor extracted from human erythrocytes, that CRI is not expressed on the Raji cell surface. However, Weis et al. (1984) also found later EC3b rosettes formation with Raji cells. Dierich et u!. (1983) have mentioned that Raji cells could secrete factor H activity, a 150,000 serum co-factor of factor I proteolysis of C3b. These data and the close mol. wts of H and gp 140, prompted us to analyze the relationship between both molecules. Polyclonal antibodies were prepared by injecting
highly purified gp 140 in a rabbit. For instance. 20 jig of gp 140 were purified from 20 g of Raji cells (Frade et al., 1984).
Up to this point we have shown that gp 140 is not related to factor H and that C3b binding to gp 140 is not mediated by H. These results were supported by: (a) cytofluorimetry analysis of rabbit anti-gp 140 and rabbit anti-H binding on Raji cell surface: (b) comparing the inhibition of specific cytotoxic assays by each antigen; (c) measuring the inhibition by each antibody of the specific binding of soluble or particle bound C3b to the cell surface; and (d) analyzing the binding of each antibody on solubiiized antigens using immunoblotting techniques, after polyacrylamide gel electrophoresis in absence or in presence of sodium dodecyl suIfate (Frade er al.. 1984). Thus, gp 140 appeared as a peculiar C3b binding protein and its relationship with others C3 receptors was analyzed. It should be noted that as early as in 1983, we had evidence that gp 140, while reacting as a C3b receptor as mentioned above, acted as a C3d receptor and that gp 140 is a specific marker of human peripheral B l~phocytes. We reported these preliminary data at the Xth International Complement Workshop in Mainz, in Mai 1983 (Frade et al., 1984). We have shown that polyclonal anti-gp 140 inhibited Raji cell rosettes with EC3b, EC3bi. EC3dg and EC3d and also blocked normal lymphocyte rosettes with EC3dg and ECM, without affecting CR1 and CR3 activity. Moreover, a monoclonal anti-C3 (C3b/130) described by others as reacting with the d region highly expressed on EC3bi, EC3dg and EC3d and poorly exposed on EC3b, completely inhibited EC3bi, EC3dg and EC3d rosettes with Raji cells, but had not effect on EC3b rosettes (Frade er al., 198%). Thus, gp 140, originally isolated by us as a C3b-like receptor, is the C3d receptor (CR2) of human B cells. Gp 140 reacts with the C3d site highly expressed on C3d, C3dg and C3dg fragments and to a lesser extent with sepharose-bound C3b or “*I-labeled C3b, supporting our previous data. Cellular distribution of gp 140 and of C3d-receptor activity restricted to different human B-cell lines, and to normal B lymphocytes, prompted others, in 1983, to analyze whether monoclona1 antibodies prepared randomly against whole human B cells, and selected as specific marker of B lymphocytes, could recognize the C3d receptor. Thus, three monoclonal antibodies were successively identified as monocfonal anti-C3d receptors: B2 (Nadler et al., 1981), HB-5 (Tedder et al., 19830) and OKB-7 (Nemerow CI al., 1986b). Iida er al. (1983) and Weis et al. (1984) identified a 140,000 and a 145,000 glycoprotein, as a C3d receptor. using the monoclonal antibodies (mAb) anti-B2 and HB-5. respectively. However, anti-B2 and HB-5 inhibited ECM rosettes poorly, unless cross-linked with goat anti-mouse IgG. Nemerow et crl. (19856) have shown that the monoclonal OKB-7. which immu-
CR2. the C3d/EBV receptor
noprecipitates a I~.000 membrane antigen. inhibits directly CXdg-microsphere rosettes. It was found that treatment of Raji cells by rabbit anti-gp 140 blocked the upvdke of the three
125l
EBV binding. Polyclonat anti-gp 140 IgCi directly inhibits the binding of EBV to Raji cells at the same concn that the binding of EC3d on cells is inhibited, whereas a 35 times higher concn of anti-gp 72 IgG or ‘Z51-labeled monoclonal antibodies anti-B2, HIS.5 and preimmune serum IgG does not. Anti-gp 140 IgG GKB-7, indicating that each of these monoclonal treatment also inhibits the induction of EBVantibodies recognized epitopes present on gp 140 determined nuclear antigen in normal tonsil B lym(Frade et af., 198%). phocytes or in EBV-negative Ramos cells, whereas high concns of anti-gp 72 IgG or prcimmune serum The C3dlC3dg receptor qf ltui~~u~B [~~~pi~~)cstes IgG have no effect (Frade ef ul., 198%). (CR2) is distinct to the C3dg receptor of’ hunwn Fingeroth er al. (1984) have shown that solubilized neutrophils (CR4) membrane antigen recognized by the monoclonal HB-5, which reacts with CR2. binds EBV particles. The neutrophil C3dg receptor was examined to More recently, Nemerow et al. (19856) have found determine its relationship to lymphocyte CR2. While that monoclonal OKB-7 inhibits both ECM and neutrophil rosettes with EC3d were undetectable, a specificity for C3d was suggested by the inhibition of EBV binding on CR2. EC3dg rosettes by fluid phase C3d-complexes bearing These data support strongly that gp 140, the C3d no detectable C3dg. However, ‘such neutrophil receptor (CR2) of human B lymphocytes, is also the EC3dg and EC3bi rosettes were not inhibited by EBVR. Meanwhile, the effect of rabbit anti-gp 140 IgG on rabbit anti-gp 140 nor an excess of anti-CRI, antithe B cell proliferative response was tested. Purified CR2 and anti-CR3 added together. In addition, B cells from human blood were induced to proliferate neutrophils did not bind “51-labeled anti-gp 140, antiby partially purified B-cell growth factor (BCGF)B2, HB-5 or OKB-7 (Frade et al., 198%). Recently, containing supernatant from activated T cells. The Vik and Fearon (1985) have shown that the binding anti-C3d receptor Ffab’)? enhanced the BCGFof Iabeled C3dg on human neutrophils is different to dependent B-cell proliferation. The effect was dosethe one on human B lymphocytes. dependent and was observed in the presence of Thus, the neutrophil C3dg receptor is distinct from different concns of BCGF, it did not correspond to gp 140, the B lymphocyte CR2, and should be desiga change in the time course of the response. The nated CR4. anti-C3d receptor F(ab’)2 had no mitogenic effect in Functions of GP 140, the C3d receptor (CR2) the absence of T-cell supernatant. In contrast, the Jondal et al. (1976) suggested. early in 1976, that undigested antiC3d receptor IgG suppressed the a structural relationship existed between the receptor BCGF-dependent B-cell proliferation. Our results for the Epstein-Barr (EBVR) and the receptors for emphasize the involvement of the C3d receptor in the the third component of complement (C3) expressed regulation of B-cell response to T-cell products on the human B-lymphocyte surface. For instance, (Frade er al., 1985~). erythrocyte antibody complement, but not erythroRecently, Nemerow et aI. (1985~) have also found cyte antibody, rosette formation was inhibited by that monoclonal anti-CR2 induces activation and EBV preparations (Jondal et al., 1976). In a large differentiation of human B lymphocytes. Moreover. it panel of human-cell lines tested, C3 receptors and was found by others that interaction of particleEBVR were either co-expressed or simultaneously bound C3d (Melchers et al., 1985) or U.V.inactivated absent (Jondal et al., 1976). C3 receptors and EBVR Epstein-Barr virus particles (Masuchi e? al., 1985) co-capped with each other but not with other memwith CR2 leads to the enhancement of B lymphocytes brane antigens, as IgM, B,-microglobulin and Fc proliferation. receptor (Yefenof et al., 1976). The association beThese data support strongly that cross-linking of tween EBVR and C3 receptors on the surface of gp 140, the C3d/EBV receptor (CR2), at the cell human B-lymphoma lines was also supported by surface by probes as F(ab’), anti-gp 140, particlemembrane-receptor stripping (Yefenof et al., 1977). bound C3d or inactivated EBV particles induces More recently, among C3 receptors, an association post-membrane signals which are involved in the has been suggested between EBVR and C3d receptor regulation of B-cell activation. Among the molecular (CR2) rather than C3b receptor (CRl) (Jonsson et events associated to “in vitro*’ activation of human al., 1982). peripheral B lymphocytes, we have recently found We have analyzed the relationship between gp 140, that gp 140, the CR2, is phosphorylated (Bare1 er al., the C3d receptor, and the EBVR using rabbit anti1986). gp 140 IgG. Polyclonal anti-gp-72, a C3-binding membrane component not related to the EBVR but REFERENCES also expressed on the Raji cell surface. was used as a Arnaout M. A., Melamed J., Tack B. F. and Colten H. R. control. Binding of rabbit IgG and EBV on cells was (1981) Characterization of the human complement fC3b) assessed by using flow cytofluorimetry. A semireceptor with a fluid phase dimer. 1. ~~i~~7~~7. 127, 1348-1354. quantitative bioassay was also used to measure the
1252
RAYMOND
Bare1 M., Charriaut C. and Frade R. (1981) Isolation and characterization of a C3b receptor-like molecule extracted from membranes of human B lymphoblastoid cell (Raji). FEBS Lelf. 136, 11l-1 14. Bare1 M., Vazquez A., Charriaut C., Aufredou M. T., Galanaud P. and Frade R. (1986) Gp140. the C3d/EBV receptor (CR2) is phosphorylated upon in oirro activation of human peripheral B lymphocytes. FEBS Letr. 197, 353-356. Berger M.. Gaither T. A., Hammer C. H. and Frank M. M. (1981) High affinity interaction of fluid phase C3b but not C3 with the neutrophil C3b receptor. Fedn Proc. 40, 44614469. Charriaut C., Bare1 M. and Frade R. (1982) Behaviour of soluble human “‘1 C3b, the third component of complement. after binding to human cells. Eur. J. Immun. 12, 289-294. Cole J. L., Dykman T. L., McDermott R. P. and Atkinson J. P. (1983) Isolation of 3 distinct C3b binding proteins (C3b-BP) from human leukocytes. Immunobiology 164, 223.
Dierich M. P., Scheiner 0.. Mussel H. H., Hammer K. P., Schopf R. E. and Schultz T. (1983) Characterization of complement receptors. Molec. Immun. 19, 1255-1265. Dobson N. J., Lambris J. D. and Ross G. D. (1981) Characteristics of isolated erythrocyte complement receptor type one (CRlC4bC3b receptor) and CR1 specific antibodies. J. Immun. 126, 693-698. Dykman T. R., Hatca T. A. and Atkinson J. P. (1984) Polymorphism of the human C3b/C4b receptor. J. exp. Med. 159, 691-703. Fearon D. T. (1980) Identification of the membrane glycoprotein that is the C3b receptor of the human erythrocyte, polymorphonuclear leukocyte. B lymphocyte and monocyte. J. exp. Med. 152, 20-30. Fearon D. T., Daha J., Weiler J. M. and Austen K. F. (1976) The natural modulation of the amplification phase of complement activation. Transplant Rev. 32, 12-25. Fingeroth J. D., Weis J. J., Tedder T. F., Strominger J. L., Biro P. A. and Fearon D. T. (1984) Epstein-Barr virus receptor of human B lymphocytes is the C3d receptor CR2. Proc. nam. Acad. Sri. U.S.A. 81, 451&4514. Fontaine M. and Rivat C. (1979) A study of the breakdown of the third component of human complement. Analyt. Immun. 13OC, 349-366.
Frade R.. Bare1 M.. Ehlin-Henriksson B. and Klein G. (1985a) Gp 140, the C3d receptor of human B lymphocytes, is also the Epstein-Barr virus receptor. Proc. narn. Acad. Sci. U.S.A. 82, 1490-1493.
Frade R., Bare1 M.. Krikorian L. and Charriaut C. (1984) Analysis of gp 140, a C3b binding membrane component present on Raji cells: a comparison with H. Eur. J. Immun. 14, 542-548. Frade R., Crevon M. C., Bare1 M., Vasquez A., Krikorian L., Charriaut C. and Galanaud P. (19856) Enhancement of human B cell proliferation by an antibody to the C3d receptor, the gp 140 molecule. Eur. J. Immun. 15, 73-76. Frade R., Myones B., Bare1 M.. Krikorian L., Charriaut C. and Ross G. D. (1985~) Gp 140, a C3b binding membrane component of B lymphocytes is the C3djC3dg receptor (CR2) and is distinct from the neutrophils C3dg receptor (CR4). Eur. J. Immun. 15, 1192-1197. Frade R. and Strominger J. (1980) Interaction of soluble “‘1 human C3b. the third component of complement, to specific receptors in human cultured B lymphoblastoid cells: characterization of a low affinity interaction. J. Immun. 125, 1332-1339.
Frade R.
and Strominger J. (1981) Interaction of particlebound ‘1SI-C3b. the third component of complement with specific receptors on human B lymphoblastoid cells (Raji). Immun. Lerl. 3, 173 178. Gaither T. A.. Magrata I. T., Berger M.. Hammer C. H.. Novikos L., Santaclla M. and Frank M. M. J. (1983)
FRADE
Complement receptor expression by neoplastic and normal human cel1s.J. Imrkn. 131, 899-905. Harrison R. A. and Lachmann P. J. (1980) The physiological breakdown of the third component of human complement. Molec. Immun. 17, 9-20. Hutt-Fletcher L. W., Fowler E., Lambris J. D.. Simmons J. G. and Ross G. D. (1983) Studies of the Eostein-Barr virus receptor found on Raji cells. II. A comparison of lymphocyte binding sites for Epstein-Barr virus and C3d. J. Immun. 130, 1309-1312. Iida K., Nadler L. and Nussenzweig V. (1983) Identification of the membrane receptor for the complement fragment C3d by means of a monoclonal antibody. J. exp. Med. 158, 1021-1033.
Jondal M., Klein G., Oldstone M. B. A., Bokish V. and Yefenof E. (1976) Surface markers on human B and T lymphocytes. VIII: association between complement and Epstein-Barr virus receptors on human lymphoid cells. Stand. J. Immun. 5, 401410.
Jonsson V.. Wells A. and Klein G. (1982) Receptors for the complement C3d component and the Epstein-Barr virus are quantitatively coexpressed on a series of B-cell lines and their derived somatic cell hybrids. Cell. Immun. 72, 263-276. Lambris J. D., Dobson N. J. and Ross G. (1980) Release of endogeneous C3b inactivator from lymphocyte in response to triggering membrane receptors for BIH globulin. J. exp. Med. 152, 1625-1644. Law S. K. and Levine R. P. (1977) Interaction between the third complement protein and cell surface macromolecules. Proc. natn. Acad. Sri. U.S.A. 74, 2701-2705. Masuchi M. G., Szigeti R., Ernberg I. and Klein E. (1985) B lymphocyte activation induced by non-transforming Epstein-Barr virus. 7th Europ. Immun. Meeting, Israel, p. 158. Melamed J., Arnaout M. A. and Colten H. R. (1982) Complement (C3b) interaction with the human granulocyte receptor: correlation of binding of fluid-phase radiolabeled ligand with histaminase release. J. Immun. 128, 2313-2318. Melchers F., Erdei A., Schultz T. and Dierich M. P. (1985) Growth control of activated, synchronized murine B cells by the C3d fragment of human complement. Nature, Lond. 317, 264-267.
Minta J. O., Man D. and Morat H. R. (1977) Kinetics studies on the fragmentation of the third component of complement by trypsin. J. Immun. 118, 2192-2196. Miiller-Eberhard H. J. and Schreiber R. D. (1980) Molecular biology and chemistry of the alternative pathway of complement. Adv. Immun. 29, l-53. Nadler L. M., Stashenko P., Hardy R., Van Agthoven A., Terhorst C. and Schlossman S. F. (1981) Characterization of a human B cell specific antigen B2 distinct from BI. J. Immun. 126, 1941-1943. Nemerow G. R., McNaughton M. E. and Cooper N. R. (19850) Binding of monocional antibody to the Epstein-Barr virus (EBV)/CR2 receptor induces activation and differentiation of human B lymphocytes. J. Immun. 135, 3068-3073. Nemerow G. R., Wolfert R., McNaughton M. and Cooper N. R. (1985b) Identification of the Epstein-Barr virus (EBV) receptor and its relationship to the complement C3d receptor (CR2). Fedn Proc. Fedn Am. Sots exp. Biol. 44, 3376.
Tack B. F., Harrison R., Janatova J., Thomas M. L. and Prahl J. W. (I 980) Evidence for the presence of an internal thioester in the third component of complement. Proc. nom. Acad. Sci. U.S.A. 77, 5764-5768.
Tedder T. F., Clement L. T. and Cooper M. D. (1983a) Use of monoclonal antibodies to examine differentiation antigens on human B cells. Fedn Proc. Fedn Am. Sots exp. hiol. 42, 718.
Tedder T. F.. Fearon D. T.. Gartland G. L. and Cooper
CR2,
the C3d/EBV
M. D. (19836) Expression of C3b receptor on human B cells and myelomocytic cells but not natural killer cells. J. /nmun. 130, 166881671. Vik D. P. and Fearon D. T. (1985) Neutrophils express a receptor for iC3b. C3dg and C3d that is distinct from CRI. CR2 and CR3. J. Immure. 134, 2571 -2579. Weis J. J., Tedder T. F. and Fearon D. T. (1984) Identification of a 145.000 Mr membrane protein as the C3d receptor (CR2) of human B lymphocytes. Proc. natn. Acad. Sri. U.S.A. 81, 881-885. Yefenof E.. Klein G.. Jondal M. and Oldstone M. B. A., (I 976) Surface markers on human B and T-lymphocytes.
receptor
1253
IX. Two-color immunofluorescence studies on the association between EBV receptors and complement receptors on the surface of lymphoid cell lines. hf. J. Cancer 17, 693-700. Yefenof E., Klein G. and K. Kvarnung (I 977) Relationship between complement activation, complement binding and EBV absorption by human hematopoetic cell lines. Cell. Immun. 31, 225-233. Zemel E. S., Cartwright-Harwick C. M. and Nilsson U. R. (1980) Effect of reduction on the structural, biologic and immunologic properties of the third component of human complement. J. Immun. 125, 1099-I 103.