0161-5890/92 35.00+0.00 Pergamon Press Ltd
Molecular ~~rnu~ology, Vol. 29,No.7/S, pp. 911-916, I992 Printed in Great Britain.
ANALYSIS OF C5b-8 BINDING SITES IN THE C9 MOLECULE USING MONOCLONAL ANTIBODIES: PARTICIPATION OF TWO SEPARATE EPITOPES OF C9 IN C5b-8 BINDING MICHIYO HATANAKA,* TSUKASA SEYAJ.ATSUSHI YODEN,~ KYOKO FUKAMOTO,? TOSHIHIKO SEMBA11and SHINYAINAI~ TDepartment of Clinical Pathology, Osaka Medical College, Takatsuki 569, Japan; SDepartment of Immunology, Center for Adult Diseases Osaka, Higashina~-ku Osaka 53’7, Japan; §Department of Pediatrics, Hirakata Municipal Hospital, Hirakata 573, Japan; and /iApplied Biosystems Japan Inc., Tokyo 134, Japan (First received 25 October 1991; accepted in revisedform
26 December 1991)
Abstract-C5b-8 binding sites in C9 were examined using mAbs raised against C9. Among 16 mAbs, two, designated P40 and X197, blocked C9-mediated EACI-8 lysis. C9 pretreated with the mAbs failed to bind to EACl-8 at 4°C. In addition, the mAbs became inaccessible to the C9 that had been incorporated into EACl-8 at 4°C. These findings suggest that C9 binding to EACI-8, but not its membrane spanning or polymerization, is blocked by mAbs. By immunoblotting analysis using a-thrombin proteolytic fragments derived from C9 [a N-terminal fragment of mol. wt 25,000 (C9a) and a C-terminal one of mol. wt 37,000 (C9b)] and tryptic fragments of C9 (mol. wts 53,000 (C9a’) and 20,000 (C9b’)), the epitopes of P40 and Xl97 were mapped to the N-terminal and C-terminal regions of C9b, respectively. Both P40 and Xl97 bound to the C9 polymerized with Zn*+ in the fluid phase, whereas Xl97 but not P40 reacted with the membrane attack complex (MAC) formed on membranes. The results suggest that two distinct epitopes are involved in C9 binding to EACl-8, and behave in a different manner for globular C9 bound to EACl-8 at 4”C, C9 assembled in MAC, or poiy-C9 induced by Znr+. These mAbs may be useful in clarifying the conformational states of C9 and in analyzing the molecular interaction between C9 and its inhibitors.
INTRODUCTION The ninth component of complement (C9) is a soluble glycoprotein that is the last component in the assembly of the membrane attack complex (MAC). When the complement is activated on target membranes, C9 binds to C5b-8 and its conformation changes from a globular to a tubular form (Silversmith and Nelsestuen, 1986). This triggers self-polymerization of C9 molecules to form transmembrane channels on the membrane leading to cell lysis. Thus, the binding of C9 to C5b-8 plays a key role in the function of C9. However, little is known about the molecular mechanism whereby C9 interacts with C5b-8 concomitant with induction of its unfolding and membrane insertion. Several studies have revealed the molecular architecture and properties of monomeric globular C9 (DiScipio et al., 1984; Stanley er al., 1985; Stanley and Herz, 1987). Analysis of the cDNA-derived amino acid sequence of C9 showed that the C-terminal fragment produced
*Author to whom correspondence should be addressed. Abbreviations: Complement
components and intermediate cells are expressed follo~ng the WHO r~ommendations. C9DS, C9 deficient serum; SFU, site-forming unit; HRF, homologous restriction factor; membrane MAC, attack complex; NHS, normal human serum; PBS, phosphate buffered saline; SDS-PAGE, sodium dodecylsulfate-polyacrylamide gel electrophoresis.
by a-thrombin (C9b) is more hydrophobic than the N-terminal fragment (C9a) (DiScipio er al., 1984). However, unlike many integral membrane proteins, C9 has no continuous stretches of hydrophobic amino acids relevant to its interaction with membranes and to self-polymerization. A putative structural unit with amphipathi~ properties, therefore, may be involved in the function of C9. Ishida et al. (1982), reported that the segments reacting with a membrane-restricted photoaffinity probe have been located in C9b. Peitsch et al. (1990), also identified two amphipathic rx-helices (292-308 and 313-333) in C9b. Since these segments interact with membrane lipids, they are presumed to represent the membrane spanning domain of C9. However, it remains unknown whether these domains are associated with the binding of C9 to C5b-8. In the previous study the domains of C9 responsible for the binding of C9 to C5b-8 and for cytolysis (Yoden et al., 1988) have been analysed, and 16 mAbs raised against monomeric C9. Two of these Abs, P40 and X197, inhibited C9 activities presumably by preventing C9 binding to C5b-8. These two Abs are directed against C9b, suggesting that the binding sites of C9 for CSb-8 are located in C9b (Yoden et al., 1988). These studies have been extended and herein show that the epitopes of P40 and X197 are mapped to different tryptic fragments, C9a’ and C9b’, respectively. The results imply that there are two distinct binding sites for C5b-8 in the C9b region: one is defined by P40 and the other by X197.
911
MICHIYO HATANAKA et
912 MATERIALS
AND METHODS
C9, serum and antibodies C9 was isolated from normal human serum (NHS) by the method of Biesecker and Miiller-Eberhard (1980), and further purified to homogeneity with Mono Q and Superdex 200 FPLC chromatography (Pharmacia LKB Biotechnology, Uppsala, Sweden). Mouse mAbs against human C9, P40, X195 and X197, were produced as previously described (Yoden er at., 1988). These mAbs were affinity purified using a Protein A column (Bio-Rad Laboratories, Richmond, CA, U.S.A.). Mouse IgGl (kappa) was purchased from Cooper Biomedical (Malvern, PA, U.S.A.). ‘*‘I-labeled sheep anti-mouse IgG F(ab’), was purchased from Amersham International (Bucks, U.K.). C9-deficient serum (C9DS) was prepared by passing NHS through an anti-C9 column. Preparation of intermediate ceils Sensitized sheep erythrocytes (EA) and intermediate cells, EACI-7, were prepared as described by Kitamura and Nagaki (198 1). EACl-8 was prepared by incubating EACl-7 with excess C8gp for 60 min at 4°C. Throughout the experiments, EA and intermediate cells were used at a concentration of 1.5 x 10’ cells/ml.
One hundred microliters of C9 (40 SFU/ml) were incubated with various concentrations of mAbs for 1.5min at 37°C. EACI-8 (0.1 ml) and GGVB (1.2 ml), 2.5 mM veronal buffer containing 0.1% gelatin, 2.5% glucose, 71 mM NaCl, I .OmM MgCl, and 0. I5 mM CaCI, (pH 7.4) were added and the mixture was incubated for 30 min at 4°C. The cells were then washed with GGVB to remove unbound C9. The EACl-9 was resuspended in 1.5 ml of GGVB then incubated for an additional 2 hr at 37°C. The released hemoglobin, reflecting the amounts of C9 bound to EACl-8, was measured at 4 15 nm.
qf poly-C9
C9 was incubated with ZnSO, at a final concentration of 100 PM in tris-buffered saline for 2 hr at 37°C. ~~nd~~~gqf mAbs to EACl-9
3ind~~g of mAbs to MAC EA (1.5 x 109) was incubated with C9DS (0.5 ml) and excess purified C9 (32 pg) in 50 ml GGVB for 25 min at 37°C. The lyzed erythrocyte ghosts were sedimented by centrifugation at 15,OOOg for 30 min and washed four times in PBS. The ghosts were resuspended in 1.1 ml of PBS containing 0.5% BSA. Aliquots (100 ~1) were incubated with 10 pg of each mAb for 45 min at 37°C. After two washes in PBS-O.S% BSA, the ghosts were incubated with 220,000 cpm of “*I-1abeled sheep anti-mouse IgG F(ab’), for 1 hr. The ghosts were washed three times in PBS-BSA, then counted in a y counter. Proteoiysis of C9 Human cl-thrombin was a gift from Dr S. Iwanaga (Kyushu University). Immobilized TPCK-trypsin was purchased from the Pierce Co. (Rockford, IL, U.S.A.) and washed before use in TBS (pH 7.5). C9 (40 ,ug) was incubated with a-thrombin (4gg) for 30 min at 37°C. The reaction was terminated by the addition of SDS at a final concentration of 0.2%. C9 (6Opg) was treated with 0.2 ml of immobilized TPCK-trypsin (13 U/ml) for 15 min at 37°C. The reaction was terminated by the removal of the immobilized trypsin by centrifugation at 10,OOOgfor 5 min. Gel electrophoresis and immunoblotting
Binding of C9 to EAC 1-8
Preparation
al.
af 4°C
EACl-8 (5.6 x lo*) was incubated with purified C9 (20 pg) in 7.5 ml of GGVB for 30min at 4°C. After three washes with GGVB, the cells were suspended in 1.1 ml of GGVB. The sample (100 ~1) was incubated with 10 pg of mAbs for 45 min. The cells were then washed twice with GGVB and incubated with ‘*‘Ilabeled sheep anti-mouse IgG F(ab’), (220,000 cpm) for 1 hr. Samples were layered over 225 ~1 of a mixture of 8 vol of dibutylphthalate and 2 vol of dinonylphthalate in microtest tubes. The tubes were centrifuged at 8O~g for 2 min. The tips of the tubes were cut and counted in a 7 counter. All procedures were performed at 4°C.
SDS-PAGE was performed by the method of Laemmli (1970). Proteins were electrophoretically transferred onto PVDF membranes (Millipore, Bedford, MA, U.S.A.) as described by Towbin ei al. (1979). A semi-dry discontinuous system (Pharmacia LKB Biotechnology) was used according to the manufa~turer’s instructions. Membrane strips were incubated with each mAb and binding was visualized using peroxidase-conjugated goat anti-mouse IgG (Cooper Biomedical, Malvern, PA, U.S.A.) followed by 4-chloronaphthol and H201. N-terminal sequence analysis C9 fragments produced by a-thrombin and immobilized trypsin were separated by SDS-PAGE and transferred to PVDF membrane. Proteins were stained with Coomassie Brilliant Blue R-250, and the stained bands were excised for analysis by the automated gas phase protein sequencer, Model 470A (Applied Biosystems, Foster City, CA, U.S.A.), equipped with an on-line HPLC, Model 120A. RESULTS
EfSect of mAbs to C9 on C9 binding to EACl-8 The effect of anti-C9 mAbs on C9-mediated EACI-8 cell lysis was tested at two temperatures (Fig. 1). C9 binds to EACI-8 at 4°C maintaining its globular conformation and is not polymerized (Boyle et al., 1978). EACI-8 incubated with C9 at 4°C followed by removal of the unbound C9 was lyzed by an additional incubation at 37°C. Pretreating C9 with P40 or Xl97
913
CSb-8 binding sites in C9
(4 120 -I
I
mAb ( Wml )
.l
10
1 mAb
100
( jJg/ml )
Fig. 1. (A) Effect of mAbs on C9 binding to EACI-8. C9 was incubated with mAbs for 15 min at 37°C. EACl-8 and GGVB were added and the mixture was incubated for 30 min at 4°C. Unbound C9 was removed by washing with GGVB. Formed EACl-9 was resuspended in GGVB and incubated for an additional 2 hr at 37°C. (B) Effect of mAbs on EACl-8 lysis by C9. C9 was incubated with mAbs for 15 min at 37°C. EACl-8 and GGVB were added and the mixture was incubated for 3 hr at 37°C. 0, P40; A, X197; n , X195.
by C9 to a similar extent and in a dose-dependent manner [Fig. l(A)]. Both P40 and X197, therefore, block C9 binding to EACI-8. By incubation of C9 with EACl-8 at 37°C both the binding of C9 to EACI-8 and C9 polymerization occur (Boyle er al., 1978). C9 pretreatment with P40 or Xl97 also blocked EACl-8 lysis by C9 as reported previously (Yoden et al., 1988). P40 was more efficient than X197, since the ID,, of P40 and Xl97 were 7 and 40 pgg/ml, respectively [Fig. l(B)]. X195, a mAb recognizing C9a, blocked EACl-8 lysis marginally under both conditions, with an ID,, of > 125 pg/ml. The difference in ID,, was not due to that of the binding affinity of the mAbs for intact C9, since P40 and Xl97 had K,, of the same order for C9, and Xl95 possessed the highest affinity among the three mAbs for intact C9, although its inhibitory effect on the C9-mediated EACl-8 lysis was minimal (Yoden et al., 1988). inhibited
the lysis of EACl-8
Epitope mapping of mAbs
cr-Thrombin cleaved C9 at a single site into C9a (mol. wt 25,000) and C9b (mol. wt 37,000), as reported
previously (Biesecker et al., 1982). TPCK-trypsin digested C9 into many fragments as previously reported (Biesecker et al., 1982). More specific digestion reported by several groups (Stanley et al., 1985; Laine and Esser, 1989) still resulted in the generation of fragments of mol. wt 34,000 and small fragments other than major 53,000 and 20,000 fragments. We obtained limited tryptic 53,000 and 20,000 mol. wt fragments using immobilized TPCK-trypsin. The 53,000 and 20,000 mol. wt tryptic fragments were designated as C9a’ and C9b’, respectively. Results of the N-terminal sequence analyses of these fragments are shown in Table 1. Since the N-terminus of C9 is blocked (Biesecker er al., 1982) C9a and the 53,000 mol. wt
fragment derived by trypsin digestion (C9a’) were Nterminal fragments, whereas C9b and the 20,000 mol. wt fragment (C9b’) were from C-terminus. Based upon the sequences of C9b and C9b’, a-thrombin cleaved 244His-‘45Gly, whereas trypsin cleaved at 391Arg-392Ala. 394Asnwas not detected, so this site is actually N-glycosylated. Using blotting analysis (Fig. 2), attempts were made to map the epitopes of P40, Xl95 and X197. The epitopes for the two antibodies capable of inhibiting C9-mediated EACl-8 lysis were mapped to different regions of the C9 molecule: P40 recognized C9b and C9a’, and Xl97 recognized C9b and C9b’. The Xl97 epitope is located in the C-terminal region of C9b, whereas that of P40 is in the N-terminal region. Xl95 reacted with C9a and C9a’, suggesting that its epitope resides in the N-terminal portion of C9. In the previous study, it was reported that P40 did not recognize any of the tryptic fragments of C9 digested with TPCK-trypsin (Yoden et al., 1988). Since several TPCK-trypsin fragmentation
minor fragments were yielded with even after 1.5 min digestion, further of C9a’ might have caused a loss
in immuno-reactivity with P40. The results are summarized in Fig. 3. These mAbs were used in the following experiments to analyze C5b-8 binding sites of c9.
Table 1. N-terminal sequence analyses of proteolytic fragments of C9 fragments produced by cc-thrombin (C9a and C9b) or immobilized TPCK-trypsin (C9a’ and C9b’) were electrophoresed on a 12% polyacrylamide gel and transferred to a PVDF membrane. The areas of each fragment were excised and subjected to N-terminal sequence analysis as described in Material and Methods Fragments c9 C9a C9b C9a’ C9b’ -,
1
2
3
4
Cycle No. 5 6 7
8
9
10
G K G S nd nd nd nd nd nd --_------AV-ITSENLI
Not detected, nd, not determined.
MICHIYO HATANAKAet ul.
914
~-thrombin
trypsin
244 1GfqGSF. ..2%!wN
N
1
1
C9a’
1 -X195~C
C
CQb
I
CQa
i
IT w
CQb’
1
~4o+*X197-
Fig. 3. Location of the epitopes of the three mAbs on the C9 molecule. This C9 model was constructed from the results of the N-terminal (Table 1) and the blotting analyses (Fig. 2).
C9 binding to EACl-8 did not accumufate as much on the EACI-9 as on the control TgGl. The EACl-9, on which MAC was formed, was prepared by incubating EA cells with C9DS then with purified C9 for 25 min at 37°C. The EACl-9 ghosts were washed, reacted with mAbs, and the levels of bound mAbs estimated with labeled second Ab (Table 2). Xl97 in addition to Xl95 markedly bound to the MAC-formed EACl-9 under these conditions, although the amount of Xl97 bound was less than that of X195. The Xl97 epitope in C9 appeared to become temporally inaccessible to the mAb through binding to EACl-8 and again accessible after the formation of MAC. On the other hand, the P40 epitope remained unexpressed throughout the C9 activation/ polymerization.
123
4
6
Reactiuities
Fig. 2. (A) Epitope mapping of mAbs on fragments produced by a-thrombin. Proteolytic fragments of C9 produced by cr-thrombin were electrophoresed on a 12% gel and transferred to a PVDF membrane. Lane 1, Coomassie stained C9; tane 2. Coomassie stained ~-thrombin-cleaved C9; lanes 3-5 show immunoblots of ~-thrombin-cleaved C4 (lane 3. P40; lane 4, X197; lane 5, X195). (B) Epitope mapping of mAbs on fragments produced by immobilized TPCK--trypsin. Proteolytic fragments of C9 produced by immobilized TPCK-trypsin were electrophoresed on a 12% polyacrylamide gel and transferred to a PVDF membrane. Lane 1, Coomassie stained C9; lane 2, Coomassie stained immobilized TPCK-trypsin cleaved C9; lanes 3-5 show immunoblots of immobilized TPCK-trypsin cleaved C9 lane 3, P40; lane 4, X197; lane 5, X195). An unidentified fragment was detected by Xl97 just above C9a’ (lane 4). This fragment is also slightly visible in
Coomassie stained trypsin-cleaved
Reactivities cells
of the mAbs
C9 (lane 2).
to C9 with preformed
EACl-9
Further investigation into whether the mAbs were still capable of binding to the C9 incorporated into EACl-8 was carried out (Table 2). C9 was incubated with EACl-8 for 30 min at 4°C which allows one C9 molecule (with a globular form) to bind to one EACl-8. After the cells were washed, mAbs were added and the levels of the bound mAbs estimated using ‘2sI-labeled second Ab. Only Xl95 markedly bound to the EACl-9 formed at 4°C. The two mAbs capable of bIocking the
of the mAbs
with polymerized
C9
C9 was polymerized in the fluid phase by incubation with 100,uM ZnSO,. Reactivities of P40, Xl97 and Xl95 with the polymerized C9 were next tested (Fig. 4). The polymerized C9 were heterogenous, with mol. wt ranging from I80,OOO to 200,000 to more polymerized forms which hardly move into polyacrylamide gels. All three mAbs reacted in immunoblots with the various forms of C9 including poly-C9, though P40 tended to react weakly with both C9 and poly-c9.
Table 2. Binding
of mAbs to EACl-9
Bound
to EACl-9
(4°C)
(at 4°C) or to MAC Bound to MAC?
Antibody
@pm)
@pm)
No Ab IgGI P40 x197 x195
2138 5411 4710 1974 13,255
1414 583.5 5808 10,467 28,744
“EACI-8 was incubated with purified C9 in GGVB for 30 min at 4°C. After unbound C9 was removed, the cells were incubated with 10~8 of mAbs for 45min at 4°C. Bound mAbs were detected by ‘251-labeled anti-mouse IgG F(ab’)2. ‘EA were incubated with C9DS and purified C9 for 25 min at 37°C. The stroma of the lyzed EA was washed in PBS to remove unbound C9, then incubated with 10~8 of mAbs for 30min at 37’C. Bound mAbs were detected with 1251-labeled sheep anti-mouse IgG F(ab’),
C5b-8 binding sites in C9
Fig. 4. Immune-reactivity of mAbs to potymerized C9. C9 was incubated with ZnSO, at a final concentration of 100 FM for 2 hr at 37°C. The samples were electrophoresed on an 8-18% linear gradient gel and transferred to a PVDF membrane. Lane 1, Coomassie stained poly-C9; lanes 2-4 show immunobiots of polyC9 (lane 2, P40; lane 3, X197; lane 4, X195). DISCUSSION
In the present study, at least two epitopes in C9 associated with the binding to EACl-8 were defined with mAbs. It is accepted that, during complement activation at 37°C C9 binds first to the C5b-8 complex, which triggers its unfolding and induces membrane insertion concomitant with self-polymerization (Silversmith and Nelsestuen, 1986). When C9 is incubated with EACl-8 at 4”C, one C9 molecule binds to one C5b-8, maintaining its globular form. Thus, only the initial step, C9 binding to EACl-8, is permitted to proceed at 4°C (Boyle et al., 1978). The lysis of the 4”C-formed EACl-9 was abrogated when C9 was pretreated with mAb, P40 or X197. The result suggests that these epitopes are involved in EACI-8 binding. Evidence of the two mAbs recognizing EACI-8 binding sites was further verified by the fact that the mAbs failed to interact with the C9 that bound to EACI-8 at 4°C. Although both P40 and Xl97 inhibited C9 binding to EACl-8 to a similar extent, P40 inhibited the lysis of EACl-8 by C9 at 37°C more efficiently than X197. The augmented C9 inhibitory activity of P40 in the lysis, therefore, can be attributed to the blocking of some steps after binding to EACl-8 cells. Using immunoblotting analysis of proteolytic fragments of C9, the epitopes of P40 are mapped within 245G1y- 39’Arg (N-terminal region of C9b) which contains two amphipathic a-helices, the putative membrane-spaning region of C9. This may reinforce the idea that P40 blocks C9 unfolding, membrane-spanning or polymerization besides the EAC l-8 binding steps.
915
A number of mAbs to C9 were obtained and their epitopes mapped by several groups (Stanley et al., 1985; Kontermann et al., 1990). Most of these mAbs are anti-C9a, suggesting that epitopes on native C9 are predominantly exposed on C9a. The C9b region is probably hidden in native C9 and thus has weak immunogenisity. As far as we know, no mAbs recognizing the membrane-spanning region, similar to P40, have been reported. In this context, P40 is unique and may be useful for clarifying the roles of the C9b portion in C9. The P40 epitope is distinct from that of Xl97 which Lys. Thus, the two separate is localized within 3g2AlaN 538 epitopes present in C9b are related to C9 binding to EACl-8 with a similar efficiency. Meri et af. (1990) proposed that there are two binding sites to C5b-8 in C9, and that CD59 inhibits the second binding. Stanley and Herz (1987) also suggested that C9 domain 4 (containing the P40 epitope) and probably domain 5 (containing the Xl97 epitope) unfolds secondary to the binding of C9 to C5b-8, resulting in an extended form of C9 and insertion into the membrane. Our data support their concepts at the molecular level, i.e. the two separate epitopes present in domains 4 and 5 are involved in C9 binding to EACI-8. Kontermann et al. (1990) on the other hand reported that a number of mAbs recognizing C9a inhibit C5b-8 binding. If this is the case, C9a assists in CSb-8 binding. Although we obtained mAbs to C9a, they, including X195, barely influenced C5b-8 binding. The discrepancy between their results and ours may be in part caused by the different strategy for screening the inhibitory mAbs and assay systems. Furthermore, it has been documented that treating C9 with glycopeptidase F reduced its hemolytic activity presumably by inhibiting C9 binding to EACl-8 (Konte~ann and Rauterberg, 1989). Glycosyl residues on C9 are likely to be important for C9 hemolytic activity. Both P40 and X197, however, bind well to the C9 pretreated with glycopeptidase F (data not shown). In this respect, some portions of the C9 molecule other than those presented in this paper may additionally modulate EAC 1-8-C9 interaction. The mAbs were raised against native C9. The three mAbs used in this study bound with high afhnity to native C9 as well as to the 2-mercaptoethanol-treated C9 (meaning that the mAbs recognize the primary sequence) as summarized in Fig. 4, but reacted in a different manner with the EACl-9 formed at 4°C MAC, or poly-C9. Xl97 binds both poly-C9 and MAC while P40 binds poly-C9 but not MAC. The reactivity of P40 with poly-C9 is not in agreement with a previous report (Takata et al., 1989) for reasons as yet unknown. The difference between P40 and Xl97 in C9 reactivity would therefore reflect a structural difference among these types of C9. These mAbs may be useful for clarifying the conformational states of C9. The roles of the two CSb-8 binding sites in C9b is unknown. It is of great interest to address the relationship between the C9 epitopes related to C5b-8 binding and C9 inhibitors, because HRF and CD59 are both anchored to membranes via glycosyl
MICHIYO HATANAKA
916
phosphatidylinositol and functions as inhibitors through blocking C9 binding to EACl-8 on the same cell membranes resulting in a disturbance of MAC formation (Meri et al., 1990; Rollins
and Sims,
1990).
Acknowledgements-The authors thank Dr Kuramitsu (Osaka University) for N-terminal analysis of C9 fragments, and Dr Iwanaga (Kyushu University) for providing cc-thrombin. Thanks are also due to Drs Matsumoto and Kitamura (Center for Adult Diseases Osaka) for valuable discussions.
REFERENCES Biesecker G., Gerard C. and Hugli T. E. (1982) An amphiphilic structure of the ninth component of human complement. Evidence from analysis of fragments produced by rthrombin. J. biol. Gem. 257, 2584-2590. Biesecker G. and Miiller-Eberhard H. J. (1980) The ninth component of human complement: purification and physicochemical characterization. J. Immun. 124, 1291-1296. Boyle M. D. P., Langone J. J. and Borsos T. (1978) Studies on the terminal stages of immune hemolysis-III. Distinction between the insertion of C9 and the formation of a transmembrane channel. J. Immun. 120, 1721-1725. DiScipio R. G., Gehring M. R., Podack E. R., Kan C. C., Hugh T. E. and Fey G. H. (1984) Nucleotide sequence of cDNA and derived amino acid sequence of human complement C9. Pmt. mm. Acad. Sci. U.S.A. 81, 7298-7302. Ishida B., Wisnieski B. J., Lavine C. H. and Esser A. F. (1982) Photolabeling of a hydrophobic domain of the ninth component of human complement. J. biol. C’henr. 257, 10,551-10,553. Kitamura H. and Nagaki K. (1981) Inhibitory effect of sugars on the spontaneous lysis of EACl-8. Moiec. rmmun. 18, 985-990. Kontermann R., Deppisch R. and Rauterberg E. W. (1990) Several epitopes on native human complement C9 are involved in interaction with the C5b-8 complex and other C9 molecules. Eur. J. fmmun. 20, 623-628. Kontermann R. and Rauterberg E. W. (1989) N-deglycosylation of human complement component C9 reduces its hemolytic activity. Molec. Immun. 26, 1125-I 132.
et al.
Laemmli U. K. (1970) Cleavage of structural proteins during the assembly of the head of the bacteriophage T4. Nature (Lund.) 227, 680-685. Laine R. 0. and Esser A. F. (1989) Detection of refolding conformers of complement protein C9 during insertion into membranes. Nature (Land.) 341, 63-65. Meri S., Morgan B. P., Davies A., Daniels R. H., Olavesen M. G., Waldmann H. and Lachmann P. J. (1990) Human protectin (CD59), an 18,OOt&20,000 MW complement lysis restricting factor, inhibits C5b-8 catalysed insertion of C9 into lipid bilayers. immunology 71, I-9. Peitsch M. C., Amiguet P., Guy R., Brunner J., Maize1 J. V. Jr and Tschopp J. (1990) Location and molecular modelling of the membrane-inserted domain of the ninth component of human complement and perforin. Mo/ec. fmmun. 27, 589-602. Rollins S. A. and Sims P. J. (1990) The complement-inhibitory activity of CD59 resides in its capacity to block incorporation of C9 into membrane CSb-9. J. Immun. 144, 3478-3483. Silversmith R. E. and Nelsestuen G. L. (1986) Assembly of the membrane attack complex of complement on small unilamellar phospholipid vesicles. Biochemistry 25, 852-860. Stanley K. K. and Herz J. (1987) Topological mapping of complement C9 by recombinant DNA techniques suggests a novel mechanism for its insertion into target membranes. EMBO J. 6, 1951-1957. Stanley K. K., Kocher H.-P., Luzio J. P.. Jackson P. and Tschopp J. (1985) The sequence and topology of human complement component C9. EMBO J. 4, 375-382. Takata Y., Moriyama T., Fukumori Y., Yoden A.. Shima M. and Inai S. (1989) A biotin-avidin sandwich ELISA for quantification of intact complement component C9. The sera from hereditary C9 deficient individuals completely lack C9. J. Immun. ‘fete. 117, 107-113. Towbin Ii., Staehelin T. and Gordon J. (1979) Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc. natn. Acad. Sci. U.S.A. 76, 4350-4354. Yoden A.. Moriyama T., Inoue K. and Inai S. (1988) The role of the C9b domain in the binding of C9 molecules to EAC l-8 defined by monoclonal antibodies to C9. .J. fmmun. 140, 2317-2321.
Molecular ~~rnu~ology, Vol. 29,No.7/S, pp. 911-916, I992 Printed in Great Britain.
ANALYSIS OF C5b-8 BINDING SITES IN THE C9 MOLECULE USING MONOCLONAL ANTIBODIES: PARTICIPATION OF TWO SEPARATE EPITOPES OF C9 IN C5b-8 BINDING MICHIYO HATANAKA,* TSUKASA SEYAJ.ATSUSHI YODEN,~ KYOKO FUKAMOTO,? TOSHIHIKO SEMBA11and SHINYAINAI~ TDepartment of Clinical Pathology, Osaka Medical College, Takatsuki 569, Japan; SDepartment of Immunology, Center for Adult Diseases Osaka, Higashina~-ku Osaka 53’7, Japan; §Department of Pediatrics, Hirakata Municipal Hospital, Hirakata 573, Japan; and /iApplied Biosystems Japan Inc., Tokyo 134, Japan (First received 25 October 1991; accepted in revisedform
26 December 1991)
Abstract-C5b-8 binding sites in C9 were examined using mAbs raised against C9. Among 16 mAbs, two, designated P40 and X197, blocked C9-mediated EACI-8 lysis. C9 pretreated with the mAbs failed to bind to EACl-8 at 4°C. In addition, the mAbs became inaccessible to the C9 that had been incorporated into EACl-8 at 4°C. These findings suggest that C9 binding to EACI-8, but not its membrane spanning or polymerization, is blocked by mAbs. By immunoblotting analysis using a-thrombin proteolytic fragments derived from C9 [a N-terminal fragment of mol. wt 25,000 (C9a) and a C-terminal one of mol. wt 37,000 (C9b)] and tryptic fragments of C9 (mol. wts 53,000 (C9a’) and 20,000 (C9b’)), the epitopes of P40 and Xl97 were mapped to the N-terminal and C-terminal regions of C9b, respectively. Both P40 and Xl97 bound to the C9 polymerized with Zn*+ in the fluid phase, whereas Xl97 but not P40 reacted with the membrane attack complex (MAC) formed on membranes. The results suggest that two distinct epitopes are involved in C9 binding to EACl-8, and behave in a different manner for globular C9 bound to EACl-8 at 4”C, C9 assembled in MAC, or poiy-C9 induced by Znr+. These mAbs may be useful in clarifying the conformational states of C9 and in analyzing the molecular interaction between C9 and its inhibitors.
INTRODUCTION The ninth component of complement (C9) is a soluble glycoprotein that is the last component in the assembly of the membrane attack complex (MAC). When the complement is activated on target membranes, C9 binds to C5b-8 and its conformation changes from a globular to a tubular form (Silversmith and Nelsestuen, 1986). This triggers self-polymerization of C9 molecules to form transmembrane channels on the membrane leading to cell lysis. Thus, the binding of C9 to C5b-8 plays a key role in the function of C9. However, little is known about the molecular mechanism whereby C9 interacts with C5b-8 concomitant with induction of its unfolding and membrane insertion. Several studies have revealed the molecular architecture and properties of monomeric globular C9 (DiScipio et al., 1984; Stanley er al., 1985; Stanley and Herz, 1987). Analysis of the cDNA-derived amino acid sequence of C9 showed that the C-terminal fragment produced
*Author to whom correspondence should be addressed. Abbreviations: Complement
components and intermediate cells are expressed follo~ng the WHO r~ommendations. C9DS, C9 deficient serum; SFU, site-forming unit; HRF, homologous restriction factor; membrane MAC, attack complex; NHS, normal human serum; PBS, phosphate buffered saline; SDS-PAGE, sodium dodecylsulfate-polyacrylamide gel electrophoresis.
by a-thrombin (C9b) is more hydrophobic than the N-terminal fragment (C9a) (DiScipio er al., 1984). However, unlike many integral membrane proteins, C9 has no continuous stretches of hydrophobic amino acids relevant to its interaction with membranes and to self-polymerization. A putative structural unit with amphipathi~ properties, therefore, may be involved in the function of C9. Ishida et al. (1982), reported that the segments reacting with a membrane-restricted photoaffinity probe have been located in C9b. Peitsch et al. (1990), also identified two amphipathic rx-helices (292-308 and 313-333) in C9b. Since these segments interact with membrane lipids, they are presumed to represent the membrane spanning domain of C9. However, it remains unknown whether these domains are associated with the binding of C9 to C5b-8. In the previous study the domains of C9 responsible for the binding of C9 to C5b-8 and for cytolysis (Yoden et al., 1988) have been analysed, and 16 mAbs raised against monomeric C9. Two of these Abs, P40 and X197, inhibited C9 activities presumably by preventing C9 binding to C5b-8. These two Abs are directed against C9b, suggesting that the binding sites of C9 for CSb-8 are located in C9b (Yoden et al., 1988). These studies have been extended and herein show that the epitopes of P40 and X197 are mapped to different tryptic fragments, C9a’ and C9b’, respectively. The results imply that there are two distinct binding sites for C5b-8 in the C9b region: one is defined by P40 and the other by X197.
911
MICHIYO HATANAKA et
912 MATERIALS
AND METHODS
C9, serum and antibodies C9 was isolated from normal human serum (NHS) by the method of Biesecker and Miiller-Eberhard (1980), and further purified to homogeneity with Mono Q and Superdex 200 FPLC chromatography (Pharmacia LKB Biotechnology, Uppsala, Sweden). Mouse mAbs against human C9, P40, X195 and X197, were produced as previously described (Yoden er at., 1988). These mAbs were affinity purified using a Protein A column (Bio-Rad Laboratories, Richmond, CA, U.S.A.). Mouse IgGl (kappa) was purchased from Cooper Biomedical (Malvern, PA, U.S.A.). ‘*‘I-labeled sheep anti-mouse IgG F(ab’), was purchased from Amersham International (Bucks, U.K.). C9-deficient serum (C9DS) was prepared by passing NHS through an anti-C9 column. Preparation of intermediate ceils Sensitized sheep erythrocytes (EA) and intermediate cells, EACI-7, were prepared as described by Kitamura and Nagaki (198 1). EACl-8 was prepared by incubating EACl-7 with excess C8gp for 60 min at 4°C. Throughout the experiments, EA and intermediate cells were used at a concentration of 1.5 x 10’ cells/ml.
One hundred microliters of C9 (40 SFU/ml) were incubated with various concentrations of mAbs for 1.5min at 37°C. EACI-8 (0.1 ml) and GGVB (1.2 ml), 2.5 mM veronal buffer containing 0.1% gelatin, 2.5% glucose, 71 mM NaCl, I .OmM MgCl, and 0. I5 mM CaCI, (pH 7.4) were added and the mixture was incubated for 30 min at 4°C. The cells were then washed with GGVB to remove unbound C9. The EACl-9 was resuspended in 1.5 ml of GGVB then incubated for an additional 2 hr at 37°C. The released hemoglobin, reflecting the amounts of C9 bound to EACl-8, was measured at 4 15 nm.
qf poly-C9
C9 was incubated with ZnSO, at a final concentration of 100 PM in tris-buffered saline for 2 hr at 37°C. ~~nd~~~gqf mAbs to EACl-9
3ind~~g of mAbs to MAC EA (1.5 x 109) was incubated with C9DS (0.5 ml) and excess purified C9 (32 pg) in 50 ml GGVB for 25 min at 37°C. The lyzed erythrocyte ghosts were sedimented by centrifugation at 15,OOOg for 30 min and washed four times in PBS. The ghosts were resuspended in 1.1 ml of PBS containing 0.5% BSA. Aliquots (100 ~1) were incubated with 10 pg of each mAb for 45 min at 37°C. After two washes in PBS-O.S% BSA, the ghosts were incubated with 220,000 cpm of “*I-1abeled sheep anti-mouse IgG F(ab’), for 1 hr. The ghosts were washed three times in PBS-BSA, then counted in a y counter. Proteoiysis of C9 Human cl-thrombin was a gift from Dr S. Iwanaga (Kyushu University). Immobilized TPCK-trypsin was purchased from the Pierce Co. (Rockford, IL, U.S.A.) and washed before use in TBS (pH 7.5). C9 (40 ,ug) was incubated with a-thrombin (4gg) for 30 min at 37°C. The reaction was terminated by the addition of SDS at a final concentration of 0.2%. C9 (6Opg) was treated with 0.2 ml of immobilized TPCK-trypsin (13 U/ml) for 15 min at 37°C. The reaction was terminated by the removal of the immobilized trypsin by centrifugation at 10,OOOgfor 5 min. Gel electrophoresis and immunoblotting
Binding of C9 to EAC 1-8
Preparation
al.
af 4°C
EACl-8 (5.6 x lo*) was incubated with purified C9 (20 pg) in 7.5 ml of GGVB for 30min at 4°C. After three washes with GGVB, the cells were suspended in 1.1 ml of GGVB. The sample (100 ~1) was incubated with 10 pg of mAbs for 45 min. The cells were then washed twice with GGVB and incubated with ‘*‘Ilabeled sheep anti-mouse IgG F(ab’), (220,000 cpm) for 1 hr. Samples were layered over 225 ~1 of a mixture of 8 vol of dibutylphthalate and 2 vol of dinonylphthalate in microtest tubes. The tubes were centrifuged at 8O~g for 2 min. The tips of the tubes were cut and counted in a 7 counter. All procedures were performed at 4°C.
SDS-PAGE was performed by the method of Laemmli (1970). Proteins were electrophoretically transferred onto PVDF membranes (Millipore, Bedford, MA, U.S.A.) as described by Towbin ei al. (1979). A semi-dry discontinuous system (Pharmacia LKB Biotechnology) was used according to the manufa~turer’s instructions. Membrane strips were incubated with each mAb and binding was visualized using peroxidase-conjugated goat anti-mouse IgG (Cooper Biomedical, Malvern, PA, U.S.A.) followed by 4-chloronaphthol and H201. N-terminal sequence analysis C9 fragments produced by a-thrombin and immobilized trypsin were separated by SDS-PAGE and transferred to PVDF membrane. Proteins were stained with Coomassie Brilliant Blue R-250, and the stained bands were excised for analysis by the automated gas phase protein sequencer, Model 470A (Applied Biosystems, Foster City, CA, U.S.A.), equipped with an on-line HPLC, Model 120A. RESULTS
EfSect of mAbs to C9 on C9 binding to EACl-8 The effect of anti-C9 mAbs on C9-mediated EACI-8 cell lysis was tested at two temperatures (Fig. 1). C9 binds to EACI-8 at 4°C maintaining its globular conformation and is not polymerized (Boyle et al., 1978). EACI-8 incubated with C9 at 4°C followed by removal of the unbound C9 was lyzed by an additional incubation at 37°C. Pretreating C9 with P40 or Xl97
913
CSb-8 binding sites in C9
(4 120 -I
I
mAb ( Wml )
.l
10
1 mAb
100
( jJg/ml )
Fig. 1. (A) Effect of mAbs on C9 binding to EACI-8. C9 was incubated with mAbs for 15 min at 37°C. EACl-8 and GGVB were added and the mixture was incubated for 30 min at 4°C. Unbound C9 was removed by washing with GGVB. Formed EACl-9 was resuspended in GGVB and incubated for an additional 2 hr at 37°C. (B) Effect of mAbs on EACl-8 lysis by C9. C9 was incubated with mAbs for 15 min at 37°C. EACl-8 and GGVB were added and the mixture was incubated for 3 hr at 37°C. 0, P40; A, X197; n , X195.
by C9 to a similar extent and in a dose-dependent manner [Fig. l(A)]. Both P40 and X197, therefore, block C9 binding to EACI-8. By incubation of C9 with EACl-8 at 37°C both the binding of C9 to EACI-8 and C9 polymerization occur (Boyle er al., 1978). C9 pretreatment with P40 or Xl97 also blocked EACl-8 lysis by C9 as reported previously (Yoden et al., 1988). P40 was more efficient than X197, since the ID,, of P40 and Xl97 were 7 and 40 pgg/ml, respectively [Fig. l(B)]. X195, a mAb recognizing C9a, blocked EACl-8 lysis marginally under both conditions, with an ID,, of > 125 pg/ml. The difference in ID,, was not due to that of the binding affinity of the mAbs for intact C9, since P40 and Xl97 had K,, of the same order for C9, and Xl95 possessed the highest affinity among the three mAbs for intact C9, although its inhibitory effect on the C9-mediated EACl-8 lysis was minimal (Yoden et al., 1988). inhibited
the lysis of EACl-8
Epitope mapping of mAbs
cr-Thrombin cleaved C9 at a single site into C9a (mol. wt 25,000) and C9b (mol. wt 37,000), as reported
previously (Biesecker et al., 1982). TPCK-trypsin digested C9 into many fragments as previously reported (Biesecker et al., 1982). More specific digestion reported by several groups (Stanley et al., 1985; Laine and Esser, 1989) still resulted in the generation of fragments of mol. wt 34,000 and small fragments other than major 53,000 and 20,000 fragments. We obtained limited tryptic 53,000 and 20,000 mol. wt fragments using immobilized TPCK-trypsin. The 53,000 and 20,000 mol. wt tryptic fragments were designated as C9a’ and C9b’, respectively. Results of the N-terminal sequence analyses of these fragments are shown in Table 1. Since the N-terminus of C9 is blocked (Biesecker er al., 1982) C9a and the 53,000 mol. wt
fragment derived by trypsin digestion (C9a’) were Nterminal fragments, whereas C9b and the 20,000 mol. wt fragment (C9b’) were from C-terminus. Based upon the sequences of C9b and C9b’, a-thrombin cleaved 244His-‘45Gly, whereas trypsin cleaved at 391Arg-392Ala. 394Asnwas not detected, so this site is actually N-glycosylated. Using blotting analysis (Fig. 2), attempts were made to map the epitopes of P40, Xl95 and X197. The epitopes for the two antibodies capable of inhibiting C9-mediated EACl-8 lysis were mapped to different regions of the C9 molecule: P40 recognized C9b and C9a’, and Xl97 recognized C9b and C9b’. The Xl97 epitope is located in the C-terminal region of C9b, whereas that of P40 is in the N-terminal region. Xl95 reacted with C9a and C9a’, suggesting that its epitope resides in the N-terminal portion of C9. In the previous study, it was reported that P40 did not recognize any of the tryptic fragments of C9 digested with TPCK-trypsin (Yoden et al., 1988). Since several TPCK-trypsin fragmentation
minor fragments were yielded with even after 1.5 min digestion, further of C9a’ might have caused a loss
in immuno-reactivity with P40. The results are summarized in Fig. 3. These mAbs were used in the following experiments to analyze C5b-8 binding sites of c9.
Table 1. N-terminal sequence analyses of proteolytic fragments of C9 fragments produced by cc-thrombin (C9a and C9b) or immobilized TPCK-trypsin (C9a’ and C9b’) were electrophoresed on a 12% polyacrylamide gel and transferred to a PVDF membrane. The areas of each fragment were excised and subjected to N-terminal sequence analysis as described in Material and Methods Fragments c9 C9a C9b C9a’ C9b’ -,
1
2
3
4
Cycle No. 5 6 7
8
9
10
G K G S nd nd nd nd nd nd --_------AV-ITSENLI
Not detected, nd, not determined.
MICHIYO HATANAKAet ul.
914
~-thrombin
trypsin
244 1GfqGSF. ..2%!wN
N
1
1
C9a’
1 -X195~C
C
CQb
I
CQa
i
IT w
CQb’
1
~4o+*X197-
Fig. 3. Location of the epitopes of the three mAbs on the C9 molecule. This C9 model was constructed from the results of the N-terminal (Table 1) and the blotting analyses (Fig. 2).
C9 binding to EACl-8 did not accumufate as much on the EACI-9 as on the control TgGl. The EACl-9, on which MAC was formed, was prepared by incubating EA cells with C9DS then with purified C9 for 25 min at 37°C. The EACl-9 ghosts were washed, reacted with mAbs, and the levels of bound mAbs estimated with labeled second Ab (Table 2). Xl97 in addition to Xl95 markedly bound to the MAC-formed EACl-9 under these conditions, although the amount of Xl97 bound was less than that of X195. The Xl97 epitope in C9 appeared to become temporally inaccessible to the mAb through binding to EACl-8 and again accessible after the formation of MAC. On the other hand, the P40 epitope remained unexpressed throughout the C9 activation/ polymerization.
123
4
6
Reactiuities
Fig. 2. (A) Epitope mapping of mAbs on fragments produced by a-thrombin. Proteolytic fragments of C9 produced by cr-thrombin were electrophoresed on a 12% gel and transferred to a PVDF membrane. Lane 1, Coomassie stained C9; tane 2. Coomassie stained ~-thrombin-cleaved C9; lanes 3-5 show immunoblots of ~-thrombin-cleaved C4 (lane 3. P40; lane 4, X197; lane 5, X195). (B) Epitope mapping of mAbs on fragments produced by immobilized TPCK--trypsin. Proteolytic fragments of C9 produced by immobilized TPCK-trypsin were electrophoresed on a 12% polyacrylamide gel and transferred to a PVDF membrane. Lane 1, Coomassie stained C9; lane 2, Coomassie stained immobilized TPCK-trypsin cleaved C9; lanes 3-5 show immunoblots of immobilized TPCK-trypsin cleaved C9 lane 3, P40; lane 4, X197; lane 5, X195). An unidentified fragment was detected by Xl97 just above C9a’ (lane 4). This fragment is also slightly visible in
Coomassie stained trypsin-cleaved
Reactivities cells
of the mAbs
C9 (lane 2).
to C9 with preformed
EACl-9
Further investigation into whether the mAbs were still capable of binding to the C9 incorporated into EACl-8 was carried out (Table 2). C9 was incubated with EACl-8 for 30 min at 4°C which allows one C9 molecule (with a globular form) to bind to one EACl-8. After the cells were washed, mAbs were added and the levels of the bound mAbs estimated using ‘2sI-labeled second Ab. Only Xl95 markedly bound to the EACl-9 formed at 4°C. The two mAbs capable of bIocking the
of the mAbs
with polymerized
C9
C9 was polymerized in the fluid phase by incubation with 100,uM ZnSO,. Reactivities of P40, Xl97 and Xl95 with the polymerized C9 were next tested (Fig. 4). The polymerized C9 were heterogenous, with mol. wt ranging from I80,OOO to 200,000 to more polymerized forms which hardly move into polyacrylamide gels. All three mAbs reacted in immunoblots with the various forms of C9 including poly-C9, though P40 tended to react weakly with both C9 and poly-c9.
Table 2. Binding
of mAbs to EACl-9
Bound
to EACl-9
(4°C)
(at 4°C) or to MAC Bound to MAC?
Antibody
@pm)
@pm)
No Ab IgGI P40 x197 x195
2138 5411 4710 1974 13,255
1414 583.5 5808 10,467 28,744
“EACI-8 was incubated with purified C9 in GGVB for 30 min at 4°C. After unbound C9 was removed, the cells were incubated with 10~8 of mAbs for 45min at 4°C. Bound mAbs were detected by ‘251-labeled anti-mouse IgG F(ab’)2. ‘EA were incubated with C9DS and purified C9 for 25 min at 37°C. The stroma of the lyzed EA was washed in PBS to remove unbound C9, then incubated with 10~8 of mAbs for 30min at 37’C. Bound mAbs were detected with 1251-labeled sheep anti-mouse IgG F(ab’),
C5b-8 binding sites in C9
Fig. 4. Immune-reactivity of mAbs to potymerized C9. C9 was incubated with ZnSO, at a final concentration of 100 FM for 2 hr at 37°C. The samples were electrophoresed on an 8-18% linear gradient gel and transferred to a PVDF membrane. Lane 1, Coomassie stained poly-C9; lanes 2-4 show immunobiots of polyC9 (lane 2, P40; lane 3, X197; lane 4, X195). DISCUSSION
In the present study, at least two epitopes in C9 associated with the binding to EACl-8 were defined with mAbs. It is accepted that, during complement activation at 37°C C9 binds first to the C5b-8 complex, which triggers its unfolding and induces membrane insertion concomitant with self-polymerization (Silversmith and Nelsestuen, 1986). When C9 is incubated with EACl-8 at 4”C, one C9 molecule binds to one C5b-8, maintaining its globular form. Thus, only the initial step, C9 binding to EACl-8, is permitted to proceed at 4°C (Boyle et al., 1978). The lysis of the 4”C-formed EACl-9 was abrogated when C9 was pretreated with mAb, P40 or X197. The result suggests that these epitopes are involved in EACI-8 binding. Evidence of the two mAbs recognizing EACI-8 binding sites was further verified by the fact that the mAbs failed to interact with the C9 that bound to EACI-8 at 4°C. Although both P40 and Xl97 inhibited C9 binding to EACl-8 to a similar extent, P40 inhibited the lysis of EACl-8 by C9 at 37°C more efficiently than X197. The augmented C9 inhibitory activity of P40 in the lysis, therefore, can be attributed to the blocking of some steps after binding to EACl-8 cells. Using immunoblotting analysis of proteolytic fragments of C9, the epitopes of P40 are mapped within 245G1y- 39’Arg (N-terminal region of C9b) which contains two amphipathic a-helices, the putative membrane-spaning region of C9. This may reinforce the idea that P40 blocks C9 unfolding, membrane-spanning or polymerization besides the EAC l-8 binding steps.
915
A number of mAbs to C9 were obtained and their epitopes mapped by several groups (Stanley et al., 1985; Kontermann et al., 1990). Most of these mAbs are anti-C9a, suggesting that epitopes on native C9 are predominantly exposed on C9a. The C9b region is probably hidden in native C9 and thus has weak immunogenisity. As far as we know, no mAbs recognizing the membrane-spanning region, similar to P40, have been reported. In this context, P40 is unique and may be useful for clarifying the roles of the C9b portion in C9. The P40 epitope is distinct from that of Xl97 which Lys. Thus, the two separate is localized within 3g2AlaN 538 epitopes present in C9b are related to C9 binding to EACl-8 with a similar efficiency. Meri et af. (1990) proposed that there are two binding sites to C5b-8 in C9, and that CD59 inhibits the second binding. Stanley and Herz (1987) also suggested that C9 domain 4 (containing the P40 epitope) and probably domain 5 (containing the Xl97 epitope) unfolds secondary to the binding of C9 to C5b-8, resulting in an extended form of C9 and insertion into the membrane. Our data support their concepts at the molecular level, i.e. the two separate epitopes present in domains 4 and 5 are involved in C9 binding to EACI-8. Kontermann et al. (1990) on the other hand reported that a number of mAbs recognizing C9a inhibit C5b-8 binding. If this is the case, C9a assists in CSb-8 binding. Although we obtained mAbs to C9a, they, including X195, barely influenced C5b-8 binding. The discrepancy between their results and ours may be in part caused by the different strategy for screening the inhibitory mAbs and assay systems. Furthermore, it has been documented that treating C9 with glycopeptidase F reduced its hemolytic activity presumably by inhibiting C9 binding to EACl-8 (Konte~ann and Rauterberg, 1989). Glycosyl residues on C9 are likely to be important for C9 hemolytic activity. Both P40 and X197, however, bind well to the C9 pretreated with glycopeptidase F (data not shown). In this respect, some portions of the C9 molecule other than those presented in this paper may additionally modulate EAC 1-8-C9 interaction. The mAbs were raised against native C9. The three mAbs used in this study bound with high afhnity to native C9 as well as to the 2-mercaptoethanol-treated C9 (meaning that the mAbs recognize the primary sequence) as summarized in Fig. 4, but reacted in a different manner with the EACl-9 formed at 4°C MAC, or poly-C9. Xl97 binds both poly-C9 and MAC while P40 binds poly-C9 but not MAC. The reactivity of P40 with poly-C9 is not in agreement with a previous report (Takata et al., 1989) for reasons as yet unknown. The difference between P40 and Xl97 in C9 reactivity would therefore reflect a structural difference among these types of C9. These mAbs may be useful for clarifying the conformational states of C9. The roles of the two CSb-8 binding sites in C9b is unknown. It is of great interest to address the relationship between the C9 epitopes related to C5b-8 binding and C9 inhibitors, because HRF and CD59 are both anchored to membranes via glycosyl
MICHIYO HATANAKA
916
phosphatidylinositol and functions as inhibitors through blocking C9 binding to EACl-8 on the same cell membranes resulting in a disturbance of MAC formation (Meri et al., 1990; Rollins
and Sims,
1990).
Acknowledgements-The authors thank Dr Kuramitsu (Osaka University) for N-terminal analysis of C9 fragments, and Dr Iwanaga (Kyushu University) for providing cc-thrombin. Thanks are also due to Drs Matsumoto and Kitamura (Center for Adult Diseases Osaka) for valuable discussions.
REFERENCES Biesecker G., Gerard C. and Hugli T. E. (1982) An amphiphilic structure of the ninth component of human complement. Evidence from analysis of fragments produced by rthrombin. J. biol. Gem. 257, 2584-2590. Biesecker G. and Miiller-Eberhard H. J. (1980) The ninth component of human complement: purification and physicochemical characterization. J. Immun. 124, 1291-1296. Boyle M. D. P., Langone J. J. and Borsos T. (1978) Studies on the terminal stages of immune hemolysis-III. Distinction between the insertion of C9 and the formation of a transmembrane channel. J. Immun. 120, 1721-1725. DiScipio R. G., Gehring M. R., Podack E. R., Kan C. C., Hugh T. E. and Fey G. H. (1984) Nucleotide sequence of cDNA and derived amino acid sequence of human complement C9. Pmt. mm. Acad. Sci. U.S.A. 81, 7298-7302. Ishida B., Wisnieski B. J., Lavine C. H. and Esser A. F. (1982) Photolabeling of a hydrophobic domain of the ninth component of human complement. J. biol. C’henr. 257, 10,551-10,553. Kitamura H. and Nagaki K. (1981) Inhibitory effect of sugars on the spontaneous lysis of EACl-8. Moiec. rmmun. 18, 985-990. Kontermann R., Deppisch R. and Rauterberg E. W. (1990) Several epitopes on native human complement C9 are involved in interaction with the C5b-8 complex and other C9 molecules. Eur. J. fmmun. 20, 623-628. Kontermann R. and Rauterberg E. W. (1989) N-deglycosylation of human complement component C9 reduces its hemolytic activity. Molec. Immun. 26, 1125-I 132.
et al.
Laemmli U. K. (1970) Cleavage of structural proteins during the assembly of the head of the bacteriophage T4. Nature (Lund.) 227, 680-685. Laine R. 0. and Esser A. F. (1989) Detection of refolding conformers of complement protein C9 during insertion into membranes. Nature (Land.) 341, 63-65. Meri S., Morgan B. P., Davies A., Daniels R. H., Olavesen M. G., Waldmann H. and Lachmann P. J. (1990) Human protectin (CD59), an 18,OOt&20,000 MW complement lysis restricting factor, inhibits C5b-8 catalysed insertion of C9 into lipid bilayers. immunology 71, I-9. Peitsch M. C., Amiguet P., Guy R., Brunner J., Maize1 J. V. Jr and Tschopp J. (1990) Location and molecular modelling of the membrane-inserted domain of the ninth component of human complement and perforin. Mo/ec. fmmun. 27, 589-602. Rollins S. A. and Sims P. J. (1990) The complement-inhibitory activity of CD59 resides in its capacity to block incorporation of C9 into membrane CSb-9. J. Immun. 144, 3478-3483. Silversmith R. E. and Nelsestuen G. L. (1986) Assembly of the membrane attack complex of complement on small unilamellar phospholipid vesicles. Biochemistry 25, 852-860. Stanley K. K. and Herz J. (1987) Topological mapping of complement C9 by recombinant DNA techniques suggests a novel mechanism for its insertion into target membranes. EMBO J. 6, 1951-1957. Stanley K. K., Kocher H.-P., Luzio J. P.. Jackson P. and Tschopp J. (1985) The sequence and topology of human complement component C9. EMBO J. 4, 375-382. Takata Y., Moriyama T., Fukumori Y., Yoden A.. Shima M. and Inai S. (1989) A biotin-avidin sandwich ELISA for quantification of intact complement component C9. The sera from hereditary C9 deficient individuals completely lack C9. J. Immun. ‘fete. 117, 107-113. Towbin Ii., Staehelin T. and Gordon J. (1979) Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc. natn. Acad. Sci. U.S.A. 76, 4350-4354. Yoden A.. Moriyama T., Inoue K. and Inai S. (1988) The role of the C9b domain in the binding of C9 molecules to EAC l-8 defined by monoclonal antibodies to C9. .J. fmmun. 140, 2317-2321.
