Identification of the C3b Receptor-binding Domain ... - John D. Lambris

5 downloads 55 Views 3MB Size Report
immunosorbent assay, CR1 bound to C3b, iC3b, and. C3c but not to C3d, and this binding was inhibited by soluble C3b and C3c. Further attempts to generate a.
THEJOURNALOF BIOLOGICAL CHEMISTRY

Vol. 263, No. 28, Issue of October 5, pp. 14586-14591, 1988 Printed in U.S.A.

0 1988 by The American Society for Biochemistry and Molecular Biology, Inc.

Identification of the C3b Receptor-binding Domain in Third Component of Complement* (Received for publication, April 14, 1988)

J. David Becherer and John D. LambrisS From the Basel Institute for Immunology, Grenzacherstrasse 487, 4005-Basel, Switzerland

We report here that complement receptor type one surface receptors (for reviews see Refs. 1-4). Human C3 is a (CR1) binds to a region of C3b that is contained within 195,000-Da glycoprotein (1.2 mg/ml in serum) composed of the NHz terminus of the a’ chain. In an enzyme-linked an a and @ chain which are covalently linked via a disulfide immunosorbent assay, CR1 bound to C3b, iC3b, and bond. Upon complement activation, C3 is cleaved byC3 C3c but not to C3d, and this binding was inhibited by convertase into C3a ( M , = 9,000) and C3b (MI = 185,000), soluble C3b and C3c. Further attempts to generate a both of which are capable of binding specific receptors. This small C3 fragment capable of binding CR1 were un- cleavage results in a transiently expressed membrane-attachsuccessful. However, elastase degradation of C3 gen- ing site in additionto conformational changes in the molecule erated four species of C3c (C3c I-IV), two of which that expose binding sites for factors B (5, 6), I (7), and H (8, bound CR1. NH2-terminal sequence analysis and sodium dodecyl sulfate-gel electrophoresis of the C3cs 9), properdin (10, l l ) , C5 (12),and CR1 (13, 14). C3bis susceptible to furthercleavage whichproceeds in several steps. indicated that the 0 chainsandthe40,000-dalton COOH-terminal a’ chain fragments were identical; the The first cleavage, between the Arg-Ser bond at residues NHz-terminal a’ chain fragments of C3c I-IV varied 1281-1282 of the C3 sequence, is made by factor I with factor from 21,000 to 27,000 daltons and accounted for the H serving as a cofactor. A second factor I cleavage occurs differential binding toCR1. C3c-I and 11,which do not rapidly between another Arg-Ser bond (residues 1298-1299) bind CR1, were missing 8 and 9 residues from the NH2 leading to the formation of iC3b (15, 16), which can then be further cleaved to C3c and C3dg. In addition, factor I-meterminus of the a’chain when compared with the intact diated degradation of C3b can also occur in the presence of a’ chain of C3b. C3c-I11 and IV, which bind CR1, had NHz termini identical to the intact NHz-terminal a’ different cofactor molecules such as factor H (16), CR2 (17), chain of C3b. Using iodinated concanavalin A and en- CR1 (16, 18) and gp 45-70 (19). The activation and subsedoglycosidase H, we showed that the NHz-terminala’ quentinactivation of C3 resultsin an array of cleavage chains of C3e-I and I11 were glycosylated, while C3c- fragments which bind specifically to membrane receptors, and I1 and IV were not. Therefore, these data indicated these receptors, in turn,mediate many of the biological effects that the amino terminus of the NHz-terminala’ chain of the complement system (1-3,20). fragment of C3c was responsible for binding CR1 while The C3b receptor, CR1, is a polymorphic glycoprotein with the COOH terminus of this fragment wasnot involved M , values ranging from 160,000-250,000 that binds the C3b since thepresence or absence of this region in C3c did and iC3b fragments of C3 andthe C4b fragment of C4. not affect CR1 binding C3c. to Subsequently, two pep- Initially purified by Fearon (18)from erythrocyte membranes, tides were synthesized from the NHz-terminala’ chain CR1 is also found on polymorphonuclear leukocytes, monofragment of C3c: X42, 42 residues in length from the cytes and macrophages, podocytes, dendritic cells, B cells, and NHz terminus and C30,30 residues in length from thea small population of T cells (21). In addition, CR1 has been COOH terminus. X42 inhibited bindingof CR1 to C3b, and this effect was also observed with antipeptide an- found in human plasma (22). Molecular genetic studies have tibodies against the X42 peptide. The C30 and other shown that CR1 contains long homologous repeating units C3-derived peptides and antipeptide antibodies had no which are comprised of short consensus repeats(23) that resemble the shortconsensus repeats of other C3b/C4b-bindeffect on the bindingof CR1 toC3b. ing proteins such as CR2 (24, 25), factor H (26, 27), and C4binding protein (28). CR1 has been shown to exhibit various functional properties depending on the cell type expressing the receptor. On phagocytic cells such as neutrophils and The complement system is composed of 20 plasma proteins monocytes, CR1 appears to enhance the phagocytosis of C3bthat function in host defense against foreign pathogens. Parbearing particles (29). The role of CR1 on non-phagocytic ticularly, the thirdcomponent of complement C3l has received much attention due to its central role in both the classical cells is less understood, but recent studies indicate that stimand alternative pathways of the complement cascade and due ulation of CRl on B cells enhances B cell differentiation (30, to its dynamic ability to bind numerous ligands and cell 31). As mentioned above, CR1 can serve as a cofactor for the factor I-mediated cleavage of C3b and for the processing of *This workwas supported by F. Hoffman-La Roche Ltd. Co., immune complexes (32). The affinity of C3b for its receptor varies; monomeric C3b has an apparent K, from 0.5 X lo6 Basel, Switzerland. $ To whom correspondence should be addressed. “l-2 X IO6M” (33, 34) while dimeric C3b binds with higher The abbreviations used are: C3, the third component of comple- affinity (KO= 5 X IO7 M-’) (34), as is the case when C3b is ment; CR1, complement receptor type one; ELISA, enzyme-linked covalently bound to surfaces. Additionally, it has been shown immunosorbent assay; ConA, concavalin A; mAb, monoclonal antibody; PBS, phosphate-buffered saline; FPLC,fastprotein liquid that iC3b (KO= 2 x lo5 M-‘) (35) and C3c (36) bind to CR1 chromatography; SDS, sodium dodecyl sulfate; endo-H, endoglycosi- but with lower affinity. In this study, in order to identify the structural requiredase H.

14586

14587

The CR1 Binding Domainin C3 ments of C3 involved in the binding of CR1, different fragments of C3 were generated and characterized with respect to their primary structure and by their ability to bind CR1. Based on this, we localized the binding site on C3b for CR1 to a domain in the NH2terminus of the a' chain of C3b. This was confirmed by the use of synthetic peptides from this region, as well as with the corresponding antipeptide antibodies, both of which inhibited the binding of CR1 to C3b and c3c. EXPERIMENTALPROCEDURES

Materials-Trypsin and soybean trypsin inhibitor were purchased from Worthington. Porcine pancreatic elastase was obtained from Serva and concanavalin A from Vector Laboratories. Polyvinylidene difluoride membranes were from Millipore. Antibodies-mAb To5 was purchased from Dakopatts, and peroxidase-conjugated goat anti-mouse Ig was purchased from Bio-Rad. Antibodies against synthetic peptides were made in rabbits by subcutaneous injections of the synthetic peptides coupled to keyhole limpet hemocyanin by glutaraldehyde (37). The rabbit immunoglobulin was fractionated from other serum proteins by ammonium sulfate precipitation. The pellet was resuspended, dialyzed against PBS, pH 7.2, and the volume was adjusted to equal the starting volume of serum. Cells-Tonsils were obtained from Bezirkspital, Rheinfelden, Switzerland. Cells were isolated from homogenized tissue and washed five times with PBS and thenlysed with lysing buffer (1%Nonidet P-40 in PBS containing 5 mM EDTA, 2 mM phenylmethanesulfonyl fluoride, 1 p~ leupeptin, 1 p M pepstatin, and 2 mM diisopropyl fluorophosphate) for 1 h at 4 "C at a concentration of 5 X 10' cells/ml. Insoluble material was removed by centrifugation, and the supernatants were stored at -70 "C prior to use. Synthetic Peptides-Three different peptides, representing residues 727-768 (X42), 894-923 (C30), and 1402-1435 (M34) of the C3 sequence (see Fig. 8) were synthesized using an Applied Biosystems 430A synthesizer by the standardsolid-phase procedure of Merrifield et al. (38, 39). The synthesis was carried out on a 4-methylbenzhydrylamine resin with tert-butyloxycarbonyl-protectedamino acids, dicyclohexylcarbodiimide couplings, and trifluoroacetic acid (40%) for deprotection. The peptides were cleaved from the resin with anhydrous hydrogen fluoride in the presence of anisole as a scavenger (40). Theywere then washed with cold ether andextracted with 10% acetic acid and lyophilized. crude peptides were purified bygel filtration on a Sephadex G-15 column equilibrated with 4% acetic acid followed by reverse-phase high performance liquid chromatography on aC-18 column (Vydac) using a 10-80% acetonitrile gradient containing 0.1% trifluoroacetic acid. Purity and composition of the peptides were confirmed by thin layer chromatography and amino acid analysis using an Applied Biosystems 420A derivatizer with an on-line 130A analyzer. All peptides were resuspended and dialyzed in M, 1000 cut-off dialysis membranes (Spectrum) against PBS at pH 7.2 before use. Preparation of C3 Fragments-C3 was isolated from EDTA-plasma as previously described (41) except that the C3was passed over Mono-Q HRlO as a finalstep of purification. C3b wasgenerated from C3 by limited digestion with trypsin (90 s at 37 "C with a 1%enzyme/ substrate (w/w)), and thereaction was stopped by the addition of 3% soybean trypsin inhibitor. C3c was generated by incubating C3 with 5% elastase (w/w) (Serva) for 6 h at 37 "C. Both C3b and C3c were purified immediately after cleavage by passage over a Mono-Q HRlO column (Pharmacia LKB Biotechnology Inc.) attached to an FPLC system (Pharmacia LKB Biotechnology Inc.). Separation was achieved by equilibrating the Mono-Q column in 20 mM Tris-HC1, pH 7.5, and eluting with 200 ml of a 0-500 mM NaCl gradient at a flow rate of 4 ml/min. Assay for Binding of CRI to C3 Fragments-The binding of CR1 to C3 fragments was tested by ELISA. Briefly, microtiter wells were coated with C3 or its fragments (8 pg/ml, 50 pl/well) overnight at 4 "C. After saturation with 1%bovine serum albumin, 1%ovalbumin in PBS, pH 7.2, 50 pl of tonsil lysates, serially diluted, were allowed to react (30 min at 23 "C). The bound CR1 was detected with mAb To5 whichis specific forCR1. Bound To5 was detected with a peroxidase-conjugated goat anti-mouse Ig. Between every step in the ELISA the wells werewashed three times with 0.05% Tween in PBS, pH 7.2 Inhibition of CRI Binding to C3b or C3c by Fluid Phase C3 Frag-

ments, Synthetic Peptides, or Antipeptide Antibodies-Inhibition of CR1 binding to C3c fixed to microtiter wellswas performed as described above for the binding of CR1 to C3 fragments except that various dilutions of C3b or C3c (0.2-12 p M ) or the synthetic peptides ) preincubated with 50 pl of tonsil lysates (1 X lo7 (10-700 p ~ were cells/ml) for 1 h prior to the incubation with C3c. In the case of inhibition of CR1 binding to C3c by antipeptide antibodies, serially diluted antipeptide antibodies were preincubated with the C3c fixed to microtiter wells before the addition of 50 p1 of tonsil lysates (1 X IO7cells/ml). After the preincubation period with the various inhibitors, the assays were performed as described above for the binding of CR1 to C3 fragments. Maximum binding was that binding observed in the presence of control proteins, peptides, or antibodies (preimmune serum in the case of inhibition of antipeptide antibodies). Polyacrylamide Gel Electrophoresis-Electrophoresis of the purified proteinfragments was performed in the presence of sodium dodecyl sulfate (SDS) as described by Laemmli (42). The samples were reduced with 2% mercaptoethanol. Protein bands were detected by staining with 0.1% Coomassie Blue R-250 (Bio-Rad) in isopropyl alcohol/methanol/acetic acid/HzO (2.82.01.04.2, by volume). The molecular weights were estimated using reference proteins. Digestion of C3c Fragments by Endoglycosidase-H-C3c was treated with endoglycosidase-H (endo-H) (Boehringer Mannheim) at a ratio of 0.04 units/mg of protein in 50 mM acetate buffer at pH 5.5 for 24 h at 37 "C as previously described (43). Detection of Carbohydrate Moieties in C3c-C3c fragments were electrophoresed on 12% SDS-polyacrylamide gels under reducing conditions, and the proteins were electroblotted onto nitrocellulose filters (Bio-Rad)as described by Towbin et al. (44). The nitrocellulose sheets were blocked with 2% polyvinylpyrrolidone (Sigma) in PBS, pH 7.2. The filters were incubated with '251-ConA(1 X lo6 cpm/ml), that had been labeled by the Iodogen method (45), for 2 h at room temperature. The sheets were washed extensively with 2% polyvinylpyrrolidone and then 0.05% Tween in PBS prior to autoradiography. Protein Sequence Analysis-NHz-terminal sequences of the different chains of C3c were obtained by electroblotting the protein onto polyvinylidene difluoride filters and sequencing the excised bands directly from the filters by modifying the method described by Matsudaira (46).Briefly, SDS-polyacrylamide gels werepolymerized overnight and prerun for 1-2 h a t 5 mA/gel before the reduced C3 samples were loaded. The protein was then transferred to polyvinylidene difluoride membranes at 0.8 mA/cm2 ofgel for 1 h in a semidry electroblotter (JKA Biotech) using, as the transfer buffer, 0.025 mM N-ethylmorpholine (Pierce) titrated with formic acid to pH 8.3. The transferred protein was visualized as previously described (46) and the protein bands were excised and stored in Eppendorf tubes at -20 "C prior to sequencing. Sequence determinations were performed by Edman degradation using an Applied Biosystems 470A gas-phase sequenator with an on-line 120A phenylthiohydantoin analyzer and a 900A data handling system. RESULTS

Binding Specificity of CRl for C3 Fragments-To study the binding of CR1 to C3 fragments, we developed an ELISA where mAb To5, specific for CR1, was used to detect CR1 bound to C3 fragments fixed to microtiter wells. Fig. 1 shows that CR1 binds to C3b,iC3b, and C3c butnot to C3d. Furthermore, the binding to C3c was inhibited by fluid-phase C3b and, toa lesser extent, fluid-phase C3c (Fig. 2). Additional attemptsto generate a small fragment ofC3c capable of binding CR1 were made utilizing both chemical and enzymatic methods. The isolated chains of C ~ Cobtained , by electroelution from SDS-polyacrylamide gels, failed to bind CR1 while C3c that had been cleaved by CNBr also failed to bind CR1. Also, extensive cleavage of C3bby trypsin, chymotrypsin, clostropain or Staphylococcus aureus V8 protease failed to generate a fragment smaller than C3c that could bind CR1. Degradation, Isolation, and Reactivity of Elatae-generated C3 Fragments-Cleavage of C3 by elastase generates two major fragments known as C3c and C3d. We incubated C3 with elastase for 6 h at 37 "C and, after separationon a MonoQ anion exchange column, we obtained C3d and a heterogeneous mixture of C3c (Fig. 3). Thesefractions were then

The CRl Domain Binding

14588

I

I

sb

0.78 1156 3.12 61251215 2k

I

Lysed cells added (x105)

FIG. 1. Binding of CR1 to C 3 fragments. Two pmol of C3 (A), C3b (0).iC3b (O),C3c (O), and C3d (*) were fixed to ELISA plates. After saturation with 1%bovine serum albumin, 1%ovalbumin in PBS, pH 7.2, 50 ~1 of tonsil lysates (1 X 10’ cells/ml) were serially diluted, and bound CR1 was detected using the CR1-specific mAb To5 followed by peroxidase-labeled goat anti-mouse Ig.

I

1

0.0850.19

I

0.37

0175

115

4

E3 fragmend , p~

6

I

1

12

FIG. 2. Inhibition of CR1 binding toC3c in the presence of fluid-phase C 3 fragments. Binding of CR1 to C3c fixed on ELISA plates was carried out by preincubating tonsil lysates with various amounts (0.2-12 p ~ of) fluid-phase C3b (D), C3c I11 (O),and C3c I1 (0)as described under “Experimental Procedures.” Bound CR1 was detected using mAb To5 followed by goat anti-mouse Ig. See Fig. 7 for the differences between C3c-I1 and 111.

tested for CR1 binding and pooled according to thisreactivity and to their similarity in SDS-polyacrylamide gels. In our ELISA assay, it was found that C3c-I11 and IV bound CR1 while C3c-I and I1 did not (Fig. 4). These results led us to further investigate the difference between these C3c fragments. SDSStructural Analysis of the C3c-I-IV-Analysisby polyacrylamide gel electrophoresis showed that the p chains and the COOH-terminal a’ chain fragments of the four C3c species had identical M , of 70,800 and 40,000, respectively, while the NHz-terminala’ chain fragmentshad M, of 25,700, 21,100,27,400, and 23,600 for C3c fragments I-IV, respec-

in C3

tively. Thus thedifference in molecular weight between these fourfragments is localized totheNHz-terminal a’ chain fragment of C3c (Fig. 5 ) . In order to further characterize the different NHz-terminal a’ chainfragments of C3c I-IV, lZ5ConAwas used to detect whether these fragments contained the carbohydrate moiety present at the COOH-terminal end of this fragment (residue 917 of the C3 sequence). As seen in Fig. 5 , lZ5I-ConAbound only to C3c-I and I11 while C3c fragments I1 and IV were negative for ConA binding, indicating that, in the case of C3c-I1 and IV, elastase had cleaved internally to the carbohydrate moiety. Furthermore, treatment of C3c-I and I11 with endo-H resulted in a shift in M , of the NHz-terminal a’ chain fragment of C3c-I from 25,700 to 21,700 and C3c-I11from 27,400 to 23,800 while C3c-I1 and IVwere unaffected by endo-H (Fig. 6). Thus, between the species ofC3c that bound CR1 (C3c-I11 and IV) and those which did not (C3c-I and 11),a 2,000-2,600-dalton difference exists that is due solely to the absence of approximately 1723 residues. Further characterization of these fragments was provided by NHz-terminal sequence analysis of the individual chains. The NHz-terminal sequence of the p chain and of the 40,000dalton COOH-terminal a’ chain of C3c I-IV were identical, starting at residues 1and 1329 of the C3 sequence respectively (Fig. 7). This is identical with the sequence published by Tack et al. (47) for the same elastase-generated chains of C3c. C3cI11 and IV, which bound CR1, had amino termini beginning at residues 728 of the C3 sequence while the 21,100-dalton 8 band ofC3c-11, which does notbind CR1,wascleaved residues further in the C3 sequence, at residue 736. C3c-I, which possesses the carbohydrate moiety on its a’ chain but does not bindCR1, had an NH2 terminusthat began at residue 735 of the C3 sequence. Thus, the sequencing data coupled with the data from the carbohydrate studies mapping the COOH termini suggest that the NHz-terminal portion of the a’ chain of C3c is involved in CR1 recognition. Inhibition of CRl Binding to C3bby Synthetic Peptides or Antipeptide Antibodies-To confirm that CR1 binds to the NHz terminus of the a’ chain, we synthesized peptides from the NH, terminus (X42), and the COOH-terminus (C30) of the NHz-terminal a’ chain ofC3c (Fig. 8). Also, a third peptide, M34, which blocks properdin binding to C3b, (48) was used as a control. As seen in Fig. 9,peptide X42 from the NH, terminus of the a’ chain (residues 726-768) inhibited CR1 binding to C3c while both C30 and M34 had no effect. In a similar ELISA where CR2 binding to C3d was detected using mAb HB-5, peptide X42 had no inhibitory effect (data not shown). In addition, the X42 peptide showed noinhibitory effect when tested in the same properdin assay where M34 has been shown to be inhibitory (48) (datanot shown). Finally, antipeptide antibodies against the X42 peptide blocked CR1 binding to C ~ Cwhile , antipeptide antibodies against C30 did not inhibit (Fig. 10). Similarly, both X42 and antipeptide antibodies against X42 inhibited CR1 binding to C3b (data not shown). DISCUSSION

The aim of this study was to localize the binding site for CR1 in the C3b fragment ofC3. Because ofC3’s ability to bind numerous other ligands and cell surface receptors, there is considerable interest in localizing these binding sites within C3. Currently, binding domains of less than 40 residues have been found in C3 for factor H (491, properdin (481, CR2 (50), CR3 (51),and theC3a receptor (52).Localization of the CR1binding site has proven troublesome since the different fragments of C3can interact simultaneously with different recep-

The CR1 Domain Binding

in C3

14589

FIG. 3. PurificationofC3c-elastase by FPLC. 50 mg of C3 were incubated with 2.5 mg of elastase for 6 h a t 37 "C before injection onto a Mono-Q HR 10/10 anion exchange column. The column was equilibrated with 20 mM Tris-HC1, pH 7.5, and the C3 fragments were eluted with a 200-ml gradient from 0 to 500 mM NaCl a t a flow rate of 4 ml/ min.

A

B

2.0. 1.8.

1.6. 1.4. CU

1.2.

G 1.0. 0.8.

0.6

1

2

3

4

Coomassie Stain

Lysed cells added(xlO5) FIG.4. Binding of CR1 to C3c-elastasefragments I-IV. Binding of CR1 to C3c-I (O),C3c-II(O), C3c-111 (O), and C3c-IV (W) was performed as described in Fig. 1.

tors and since small peptide fragments of C3, generated enzymatically or chemically, losetheir ability to bind CR1. This report describes the cleavage of C3 by elastase, characterization of the resulting fragments, and the localization of a domain in C3b required forCR1 recognition. Employing an ELISA assay with tonsil lysates as a source of CR1, we showed that C3, C3b, iC3b, and C3c but not C3d bind CR1. The fact that C3 binds CR1 suggests that attachment tothe ELISA plate alters C3's conformation, thus causing it tobehave "C3b-like."The binding in this assay was specific since fluid-phase C3b and C3c could inhibit CR1 binding, although 30 times more C3c than C3b was neededto see 50% inhibition. These results suggest that theconformation of the binding site is altered, but not abolished, upon cleavage of C3b to C3c or that there is a second binding site which is present on C3b but not on C3c that increases C3b's affinity for CR1. It has beensuggested that C3b has two binding sites for factor H(49), a moleculewhich shares functional and structural similarities with CR1. Earlier studies have shown that C ~ Cgenerated , by elastase treatment of C3, is able to bind CR1 (36). Accordingly, when

1

2

3

4

Binding of '251-ConA

FIG. 5. SDS-polyacrylamide gel electrophoresis and binding of '261-ConA to C3c fragments. A, C3c fragments generated by elastase and purified on a Mono-Q column were electrophoresed ona 12% SDS-polyacrylamide gel under reducing conditions and proteins were stained with Coomassie Blue. B , A duplicate gel was run, and the proteins were electroblotted onto nitrocellulose, incubated with '251-ConA,washed, and subjected to autoradiography. Lane I , C3c-I; lane 2, C3c-11; lane 3, C3c-111; lane 4, C3c-IV. The M,35,000 protein in lane 1 and, to a lesser extent in [am2 is C3d.

we degraded C3with elastase, in addition to C3d, we obtained, after purification on a Mono-Q column, C3c fragments that differed with respect to their molecular weights and their reactivity toward CR1. The elution of C3c in four different peaks indicates that charge differences exist between the different species of C3c generated by elastase. It is interesting to note that only the NH2-terminal fragment of the a' chain of C3c wasaffected by this differential cleavage. This suggests that this fragment is responsible for binding CR1 since only C3c-I11 and IV bound CR1 (see Fig. 3 and4). NH2-terminal sequence data were usedto characterize these C3c fragments. The @ chains and the 40,000-dalton COOHterminal a' chain fragments all had identical NH2 termini, corresponding to residues 1 and 1329 of the predicted C3 sequence (53) and agreeing with the previously published sequence for the @ and 40,000-dalton chains of C3c prepared with elastase (47). This fact precluded the possibility that these chains, despite their identical apparent molecular weights from SDS-polyacrylamide gels, were responsible for CR1 binding due to differential cleavage at opposite ends of the fragment. The NH2-terminal CY'chains had molecular weights of 27,500, 25,500, 23,500, and 21,000 for C3c-111, I,

The CRl Binding in Domain

14590

C3

40.0kDa 27.4kDa 25.7kDa 23.8kDa 21.7kDa -

0 0

200

d l o

300

460

5 6 0 ' 660

'

7&3

Peptide. pM

FIG.9. Inhibition of CR1 binding toC3c by synthetic peptides. 50 pl of tonsil lysates (2 X lo' cells/ml) were preincubated for 1 h a t room temperature with 50 p1 of synthetic peptides X42 (O), C30 (A),and M34 (0)a t various concentrations (10-700 p ~ ) 50 . pl of the incubation mixture was then transferred to microtiter wells precoated with C3c. Bound CR1 was detected using mAb To5.

1 2 3 4 5 6 7 8 FIG. 6. SDS-polyacrylamide gel electrophoresis of C3c IIV after treatment with endo-H. C3c fragments I-IV, after treatment with endo-H, were run under reducing conditions on a 12% SDS-polyacrylamide gel. Lanes I , 3, 5, and 7, C3cI-IV, respectively; lanes 2,4,6, and 8, C3cI-IV, respectively, after treatmentwith endoH. The M , listed are only for thosefragments which had their carbohydrate moiety removed by endo-H digestion. CI

N-IERYIHAL SEOUENCE

FPLGYfNlS

c lb CIC-I CIC-II CIC-Ill CIC-IV

E CHAIN

SPYYSIITP fPWYSlITP ,PYYSI SPYYSI

SPYYSI

4OkDa

rArDOL1 KArDOLl KArDOL KArDOLl

OF

CQI

27kDa

CHO

BINDING

SNLDEDIIAEEN(a'-CHAIN) .-..---. AEfHIYSPS

-..--..-rw1v:ns

-NLDEDIIAEfNIVS -NLDEDIlAEfNIV

FIG. 7. Comparison of the NHz-terminal sequences, CR1 binding, and presence of the u' chain carbohydratemoiety for C3c-I-IV. Comparisons of the NHp-terminal sequence obtained for the different chains of the elastase-generated C3c by Edman degradation. CHO represents the carbohydrate moiety present on the a' chain ofC3c. The numbering of the residues for these peptides is taken from Ref. 53 after subtracting the signal peptide sequence.

PtE!E

CI RESlDUf

X42

727-768

C 30

894-923

Y34

1402-1435

0

I"

I

'

,

vl0

I

'130

I 1im Antibody dilution I

'

5 jl0

'

lj60

FIG. 10. Inhibition of CR1 binding to C3c by antipeptide antibodies. Serial dilutions of antipeptide antibodies X42 (0)and C30 (0)were incubated with C3c fixed to microtiter wells for 30 min a t room temperature. Plates were washed and theELISA was carried out asdescribed under Experimental Procedures.

SfOUfNCE

SNLDEDlIAffNIVSRSEIPESUlUNV~DlKEPPKNGlSlKL VYHHFISDGVRKSLKVVPfGlRMNKlVAVR

GVDRYISKYfLDKAFSDRNlLlIYLDKVSHSEDD

rn 'X42

FIG. 8. Amino acid sequences of the syntheticpeptides used in this study. The numbering of the residues for these peptides is as described in Fig. 7.

IV, and 11, respectively. The NH2-terminal a' chain fragments C3c-I11 and IV hadidentical NH2 termini with cleavage occurring between serine and asparagine at residues 727 and 728. Also, these fragments bound CR1. C3c-I and 11, which do not bind CR1, were cleaved 7 and 8 residues further into the C3 sequence, COOH-terminal to isoleucine and alanine a t residues 734 and 735 of the C3 sequence, respectively. These 8 residues were not enough to account for the discrepancies in the M , of these chains, thus indicating that differential cleavage also occurred at the COOH terminus. The presence of the carbohydrate on the COOH terminus of this chain allowed approximation of the COOH termini of C3c-IIV (43, 54). Iodinated ConA indicated that C3c-I and I11 possessed carbohydrate and endo-H treatment of these fragments revealed anapparent size of approximately 3,500daltons for the carbohydrate moiety. C3c-I1 and IV lacked carbohydrate onthis chainof C ~ Chaving , been cleaved internally to this moiety. These data eliminated the COOH ter-

\

,a

k I 2 N M

' Y

mJ51 vcm

181

-,++

s 14 ,1003

A ,

-003

~COOH

a-chain

c3

s-s-s $-cham

FIG. 11. Aschematic representation of the C3 molecule. Sites of cleavage by factor I ( I ) , trypsin (T), and elastase ( E ) are indicated. The location of peptides X42 and C30 and of the carbohydrate (CHO) moieties are also shown. All the assignments are based on direct structural analysis except for the elastase cleavage sites at theCOOH terminus of M , 27,000 fragment.

minus as being necessary for CR1 binding, since C3c-I contained thea' chain carbohydrate but failed to bindCR1, while C3c-IV was cleaved internally to the carbohydrate moiety but still bound CR1 (see Fig. 11).These results, coupled with the NH2-terminal sequence data and the differential binding of CR1, indicated that an intact a' chain of C3b was essential for CR1 binding to C3b. This would explain why trypsin-

The CR1 Binding Domain in C3 cleaved C3c still bound CR1, since trypsin fails to cleave the NHz terminus of the a' chain of C3b ( 5 5 ) , and why C3c-11, which is missing 8 residues from the NH2terminus of the a' chain when compared with C3c-111, does notinhibitCR1 binding to C3c (see Fig. 2). The requirement of an intact a' chain for CR1 binding results from conformational constraints or from the necessity of an intact linear sequence led us to synthesize two peptides from the C3 sequence (see Fig. 11).Peptide X42, of 42 residues corresponding to the NHzterminal region of the a' chain of C3c-111, was synthesized to insure that the peptide would be long enough to span the entire CR1-binding domain. Peptide C30, from the COOHterminal end of this same fragment, was synthesized to conclusively show that this region ofC3c was not involved in CR1 binding. The inhibition of CR1 binding to C3c by X42 and not C30 further indicated that the CR1-binding site in C3c is localized in the NH2-terminala' chain fragment. This was substantiated by the fact that antipeptideantibodies against X42 also inhibited binding of CR1 to C3c. Therefore, the CR1-binding site in C3b is localized to a domain that is contained within the NHz-terminal42 residues of the a' chain of C3b. Further elucidation of the CR1-binding site using overlapping peptides will allowus to determine the necessary elements in the interaction of C3b with its receptor. Interestingly, the first 36 residues of the a' chain of C3b, which are spanned by the synthetic peptide X42, are encoded for by a separate exon in the C3 gene (56) and the fact that the CR1binding site resides in this domain supports the supposition that functional units of proteins are encoded by individual exons (57, 58). Similarly, other mapped binding sites in C3 for factor H, properdin, and CR2 are encoded by separate exons. Acknowledgments-We thank Drs. J. Alsenz and C. Servis for critical reading of the manuscript and D. Avila for excellent technical help. REFERENCES 1. Lambris, J. D., and Muller-Eberhard, H. J. (1986)Mol. Immunol. 23,12371242 2. Lambris, J. D. (1988)Immunol. Today, in press 3. Ross, G. D., and Medof, M. E. (1985)Adu. Immunol. 37,217-267 4. Sim, R. B., Malhotra, V., Day, J. A,, and Erdei, A. (1986)Immunol. Lett. 14.183-190 5. Fear& D. T., Austen, K. F., and Ruddy, S. (1973)J. Exp. Med. 138,1305I

~~~

~~

~

1313

6. Brad;; V., Lee, G. D., Nicholson, A,, Shin, H., and Mayer, M. M. (1973)J. Immunol. 111, 1389-1400 7. Halbwachs, L., and Lachmann, P. J. (1976)S c u d . J. Immunol. 6,697-704 8. Whaley, K., and Ruddy, S. (1976)J. Exp. Med. 144,1147-1163 9. Weiler, J. M., Daha, M. R., Austen, K. F., and Fearon, D. T. (1976)Proc. Nutl. Acud. Sci. U. S.A. 73.3268-3272 10. Chapitis, J., and Lepow, I. H. (1976)J. Exp. Med. 143, 241-257 11. Schreiber, R. D., Medicus, R. G., Gotze, O., and Muller-Eberhard, H. J. (1975)J. Exp. Med. 142,760-772 12. Vogt, W., Schmidt,G., von Buttlar, B., and Diemingr, L. (1978)Immunology 34.29-40 13. Nelson, R. A,, Jr. (1953)Science 118,733-737 14. Gigli, I., and Nelson, R. A. Jr. (1968)Exp. Cell Res. 51,45-67

14591

15. Harrison, R. A,, and Lachmann, P. J. (1980)Mol. Immunol. 17,9-20 16. Ross. G. D., Lambris, J. D., Cain, J. A,, and Newman, S. L. (1982)J. Immunol. 129,2051-2060 17. Mitomo, K., Fujita, T., and Iida, K. (1987)J. Exp. Med. 166,1424-1429 18. Fearon, D. T. (1979)Pmc. Nutl. Acud. Sci. U. S.A. 76,5867-5871 19. Seya, T., Turner, J. R., and Atkinson, J. P. (1986)J. Exp. Med. 163,837855 20. Schreiber, R. D. (1984)Springer Semin. Immunoputhol. 7,221-249 21. Fearon, D. T., and Wong, W. W. (1983)Annu. Reu. Immunol. 1,243-271 22. Yoon, S. H., and Fearon, D. T. (1985)J. Immunol. 134,3332-3338 23. Klickstein, L. B., Wong, W. W., Smith, J. A., Weis, J. H., Wilson, J. G., and Fearon, D. T. (1987)J. Exp. Med. 166,1095-1112 24. Weis, J. J., Fearon, D. T., Klickstein, L. B., Wong, W. W., Richards, S. A., de BNyn Kops, A,, Smith, J. A., and Weis, J. H. (1986)Proc. Nutl. Acud. Sci. U. S. A. 83,5639-5643 25. Moore, M. D., Cooper, N. R., Tack, B. F., and Nemerow, G. N. (1987)Proc. Nutl. Acud. Sci. U. S. A. 84,9194-9198 26. Ripoche, J., Day, A. J., Harris, T. J. R., and Sim, R. B. (1988)Biochem. J. 249,593402 27. Kristensen, T., and Tack, B. F. (1986)Proc. Nutl. Acud. Sci. U. S.A. 83, 3963-3967 28. Chung, L. P., Bentley, D. R., and Reid, K. B. M. (1985)Biochem. J. 230, 133-141 29. Ehlenberger, A. G., and Nussenzweig, V. (1977)J. Exp. Med. 146, 357371 30. Weiss, L., Delfraissy, J. F.,Vazquez,A,, Wallon, C., Galanaud, P., and Kazatcbkine, M. D. (1987)J. Immunol. 138,2988-2993 31. Daha, M. R., Bloem, A. C., and Ballieux, R. E. (1984)J. Immunol. 132, 1197-1201 32. Cornacoff, J. B., Hebert, L. A., Smead, W. L., VanAman, M. E., Birmingham, D. J., and Waxman, F. J. (1983)J. Clin. Inuest. 71, 236-247 33. Pangburn, M. K. (1986)J. Immunol. 136,2216-2221 34. Arnaout, M. A., Melamed, J., Tack, B. F., and Colten, H.R. (1981)J. I ~ 127,1348-1354 ~ ~ ~ L 35. Gordon, D. L., Johnson, G. M., and Hostetter, M. K. (1987)Immunology 60,553-558 36. Ross, G. D., Newman, S. L., Lambris, J. D., Devery Pocius, J. E., Cain, J. A,, and Lachmann, P. J. (1983)J. Exp. Med. 158,334-352 37. Briand, J. P., Muller, S., and Van Regenmortel, M. H. V.(1985)J. Immunol. Methods 78,59-69 38. Merrifield, R. B. (1963)J. Am. Chem. SOC.85,2149-2154 39. Stewart, J. C., Pensa, C., Matsueda, G., and Harris, K. (1976)in Pe tides (Loffer, A., ed) pp. 285-290,Editions de 1'Universitede Bruxelles, S m x elles 40. Barany, G., and Merrifield, R. B. (1980)in The Peptides (Groos, E., and Meienhofer, J., eds) Vol. 2,pp. 100-250,Academic Press, New York 41. Lambris, J. D., Dobson, N. J., and Ross, G. D. (1980)J. Exp. Med. 162, 1625-1644 42. Laemmli, U.K. (1970)Nature 227,680-685 43. Hirani, S., Lambris, J. D., and Miiller-Eberhard, H. J. (1985)J. ImmunoL 134,1105-1109 44. Towbin, H., Staehelin, T., and Gordon, J. (1979)Proc. Nutl. Acud. Sci. U. S.A. 76,4350-4354 45. Fraker, P. J., and Speck, J. C. (1978)Biochem. Biophys. Res. Commun. 80, 849-857 46. Matsudaira, P. (1987)J. Biol. Chem. 262, 10035-10038 47. Tack, B. F., Janatova, J., Thomas,M. L., Harrison, R. A., and Hammer, C. H. (1981)Methods Enzymol. 80 (part C),64-101 48. Daoudaki, M. E., Becherer, J. D., and Lambris, J. D. (1988)J. Immunol. 140,1577-1580 49. Lambris, J. D., Avila, D.,Becherer, J. D., and Muller-Eberhard, H. J. (1988) J. Biol. Chem., 263,12147-12150 50. Lambris, J. D., Ganu, V. S., Hirani, S., and Muller-Eberhard, H. J. (1985) Proc. Nutl. Acad. Sci. U. S. A. 82,4235-4239 51. Wright, S. D., Redd P A Jong, M. T. C., and Erickson, B.W. (1987) Proc. Natl. Acud. &i. U. i. A. 84,1965-1968 52. Hugli, T. E. (1984)Springer Semin. Immunoputhol. 7 , 73-99 53. De Bruijn, M. H. L., and Fey, G. H. (1985)Proc. Nutl. Acud. Sci. U. S. A. 82,708-712 54. Hirani, S., Lambris, J. D., and Muller-Eberhard, H.J. (1986)Biochem. J. 233,613-616 55. Eggertsen, G., Hellman, U., Lundwall, A,, Folkersen, J., and Sjoquist, J. (1985)Mol. Immunol. 22,833-841 56. Barnum, S. R., and Wetsel, R. A. (1988)in Biochemistry und Molecular Biolopy of Complement (Sim, R. B.. ed) Kluwer Academic Press, Lancaster,) Ln press 57. Gilbert, W. (1978)Nature 271,501 58. Traut, T. W. (1988)Proc. Nutl. Acud. Sci. U. S. A . 86,2944-2948