0161-5890/92 $5.00 + 0.00 0 1992 Pergamon Press plc
Molecular Immunology, Vol. 29, No. 2, pp. 251-256, 1992 Printed in Great Britain.
ACTIVATION OF THE FIFTH COMPONENT OF HUMAN COMPLEMENT, C5, WITHOUT CLEAVAGE, BY METHIONINE OXIDIZING AGENTS W. VOGT,*B. ZIMMERMANN, D. HESSEand R. NOLTE Max Planck Institute for Experimental Medicine, Hermann D-3400 Giittingen, Germany
Rein-Strasse 3,
(First received 17 March 1991; accepted in revised form 6 h4ay 1991)
Abstract-Purified human CS was incubated with chloramine T (Cl-T) or N-chloro-succinimide (N-Cl-S) in barbital buffer, pH 7.2. The treatment led to C5 activation: Cl-T- and N-Cl-S-treated C5 acquired a binding site for C6; upon incubation with C6 and subsequent addition of C7, C8 and C9 a membrane attack complex formed which lysed non-sensitized guinea pig red cells (reactive lysis). While the physiological activation of C5 follows its specific cleavage, the resulting fragment C5b representing the activated C5 and expressing the C6 binding site, the treatment with the mentioned chemicals does not lead to fragmentation of the C5 protein. So, functionally, the product of the chemical treatment is CSb-like, but chemically, it comprises the whole protein; no C5a is released. Cl-T and N-Cl-S are known to more or less selectively oxidize methionine residues in proteins, dependent on the conditions. Other sensitive amino acid residues are tryptophan and cysteine. Conditions were chosen for treatment of C5 with Cl-T which exclude attack on tryptophan, and we have ensured that human C5 does not contain free cysteine residues. Further, oxidation of about 60% of the methionine residues of C5 by Cl-T was demonstrated by amino acid analysis. So, all evidence points to methionine residue(s) as the site of attack of Cl-T and probably also of N-Cl-S. The oxidation product of methionine, its sulphoxide, may cause a change in structural conformation of C5 which involves expression of the C6 binding site. Earlier it was found that oxidation of C5 by hydroxyl radicals leads to its activation without cleavage. Since the properties of this CSb-like product resemble those of the product of treatment with Cl-T and N-Cl-S, it is suggested that the formerly found activation of human C5 by hydroxyl radicals is also mediated by oxidation of methionine residue(s) in the C5 protein.
INTRODUCTION In its native state the fifth component of complement is not active. To be utilized in complement reactions it has to be activated. Physiological activation of C5 is a specific enzymic process: the protein is cleaved into two fragments which express different functional activities: C5a, the chemotactic peptide significant in inflammatory processes, and C5b, the larger protein fragment that has the transient capability to bind C6 and
in this way to form the nucleus for the assembly of the membrane damaging complex, C5b-9. The cleavage is effected by specific enzymes of the complement system, the C5 convertases of the classical or alternative pathway. In recent years, non-enzymic procedures leading to C5 activation have become known: freezing and thawing or short exposure of C5 solutions to pH 6.4 (Rother et al., 1978; Dessauer et al., 1984). Furthermore, oxygen radicals, in particular hydroxyl radicals, are capable of activating C5 (Vogt et al., 1989, 1991). The non-enzymic activation by hydroxyl radicals is not accompanied by peptide cleavage of the C5 protein chains.
*Author to whom correspondence
should be addressed.
Hence, no C5a is released and no chemotactic activity is generated. Thus activation by oxygen free radicals means conversion of the whole C5 protein to a CSb-like conformation allowing it to bind C6 and to generate the membrane attack complex in co-operation with C7, C8 and C9. The fact that hydroxyl radicals convert C5 to an activated species suggested that the change was induced by oxidation of some amino acid(s) leading to conformational changes of the protein. There is one other protein known to be rather sensitive to oxidative damage in which the locus of damage has been elucidated: u 1 proteinase inhibitor. Here, oxidation of two out of eight methionine residues is responsible for the loss of activity (Johnson and Travis, 1979). Methionine is particularly sensitive to oxidation, and there are two reagents which more or less selectively oxidize methionine residues in proteins: chloramine T (Cl-T) and Nchloro-succinimide (N-Cl-S) (Shechter et al., 1975). We have tested these agents and found that, indeed, they both convert C5 to a CSb-like species without peptide cleavage and thus mimic the effect of hydroxyl radicals, with even better yield. The results suggest that the activation of C5 by the mentioned oxidants and also by oxygen radials is induced by oxidation of methionine residues. 251
252
w. VOW et al. MATERIALS
AND METHODS
Buffers The buffer (VBSGG) used in the classical immune haemolysis assay for estimation of native C5, contained 2.5 mM barbital buffer pH 7.4, 80 mM NaCl, 0.15 mM CaCl,, 1 mM MgCl,, 2.5% glucose and 0.1% gelatin. For estimation of CSb-like C5 activity in reactive lysis assays, the buffer contained 10 mM EDTA (EDTA-VBSGG). Oxidations of C5 were carried out in barbitalsaline (5 mM barbital buffer, pH 7.2; 0.15 M NaCl). Complement
components
The purification of human C3, C5, C6, C7, C8 and C9 has been described earlier (Vogt et al., 1984, 1989). A stock solution of purified C5 in barbital-saline (0.8 mg/ml) was kept in small portions over liquid nitrogen. Before use, it was incubated for 16 hr at 37°C in order to inactivate any preformed CSb-like C5 (Vogt et al., 1991). Other complement components, used for C5 estimations were bought from Cordis Corporation. Chemicals Chloramine T (Cl-T), N-chloro-succinimide (N-Cl-S), L-methionine, its N-acetyl derivative and OH analogue were purchased from Sigma. Activated Thiol-Sepharose 4B was a preparation from Pharmacia. Before use it was swollen and washed as recommended by the manufacturer. Oxidation
of C5
Solutions of human C5 in barbital-saline of pH 7.2 were incubated with Cl-T or N-Cl-S at varying concns for varying times, at room temp. The reaction was stopped by addition of excess L-methionine (for functional assays) or its hydroxy analogue (for amino acid analysis). The calculations of molar ratios of oxidant to C5 protein were based on a molecular mass of C5 of 200 kDa.
procedure developed by Shechter et al. (1975) was used. In brief, the samples of C5 were treated with cyanogen bromide to convert non-oxidized methionine to homoserine (lactone). Then they were hydrolyzed with 6 N HCl in the presence of dithiothreitol to reduce methionine sulphoxide back to methionine. Thus any methionine determined in the subsequent amino acid analysis referred to the sulphoxide. Amino acid determination of C5 hydrolysates was carried out after precolumn derivatization with (dimethylamino)azobenzene-sulphonyl chloride according to the method of Knecht and Chang (1986). Gradient polyacrylamide
gel electrophoresis
Reduced and non-reduced samples gradient gels as described (Laemmli, et al., 1985). RESULTS
Oxidation
of C5 with Cl-T
C5 was treated with Cl-T solutions in barbital-saline buffer, pH 7.2, for 10 min at room temp. Then methionine was added to inactivate unreacted Cl-T. After further 10 min the product was mixed with C6, incubated for 3 hr at 37°C to allow the development of an active C56 complex from any CSb-like C5 formed (Vogt et al., 1989, 1991), and tested for C56 activity in the reactive lysis assay. Further, residual functional C5 activity was tested in the classical immune haemolysis assay. The tests showed that the treatment with Cl-T resulted in partial loss of native C5 function and in the generation of CSb-like activity. Figure 1 shows the dependence of the yield on the concn of Cl-T. Highest yields (12,00&25,000 CHSO/mg C5) were obtained when the incubation mixture contained 80 mol Cl-T per mol of C5. In Fig. 2 the time kinetics are shown. Regularly, the CSb-like activity Moles Chloramine T/ Mole C5 LO
Estimation
of C5 were run on 1970; von Zabern
of native C5
60 I
80
100 I
100
C5 was assayed by its function in the classical immune haemolysis assay, using sheep red cells in the state EAC14, varying dilutions of the C5 preparation to be tested, excess C2gp, C3h”, and C6C9gp. Estimation
of CSb-like
C5 activity
The products of C5 oxidation were mixed with C6h”, diluted with EDTA-VBSGG, incubated for 3 hr at 37°C and then tested in a reactive lysis assay system, using non-sensitized guinea pig red cells, and human C7, C8 and C9 (Vogt et al., 1989, 1991). Activities are expressed in CHSO/mg C5. Amino acid sulphoxide
analysis;
determination
of
methionine
Free amino acids in model mixtures were quantified in an amino acid analyser, with automatic ninhydrin staining of the effluent. For the determination of methionine sulphoxide in (oxidized) C5 preparations the
”
I
0
$0
160
liO,,M-
[ Chloramine T]
Fig. 1. Conversion of native C5 to CSb-like C5 by chloramine T. Concentration dependence. Eight microgrammes CS (0.04 nmol) were incubated with varying concns of Cl-T, as indicated in the lower abscissa. Total volume 30~1 barbitalsaline, pH 7.2. After 10 min incubation at room temp. the reaction was stopped by addition of 250pM methionine. CbS-like activity was measured by reactive lysis assay; loss of C5 refers to native C5 which was assayed by its function in the classical immune haemolysis test (see Materials and Methods).
C5 activation by methionine G 0
253
oxidation
20000-100
F 2 Y 2 % c g
__s
10000-
-50
m aB Y 7 iz 0
%
z = : 7 x
0 0
I
I L
I 12
a
, 0 16 min
Time
Fig. 2. Conversion of native C5 to CSb-like C5 by chloramine T. Time kinetics. Thirty-two microgrammes C5 (0.16 nmol) were incubated with 12.8 nmol of Cl-T in 80 ~1 barbital-saline buffer, at room temp. At varying times, portions were withdrawn and injected into methionine solution to stop the reaction. Assays as in Fig. 1 legend.
the loss of functional native C5 proceeded. The peak of CSb-like activity was reached in about l&12 min. Thereafter, the activity slowly declined. In Fig. 3 a gradient polyacrylamide gel electrophoresis of Cl-T-treated and untreated C5 is presented. Neither the non-reduced C5 oxidation products nor the reduced samples show any cleavage of the C5 protein by Cl-T.
developed
more
slowly
than
Specljicity of Cl-T for methionine oxidation
Shechter et al. (1975) found that at pH 8.5 of all natural amino acids only methionine and cysteine were oxidized by Cl-T, while at pH 2.2 Cl-T attacked tryptophan, in addition. N-Cl-S was more aggressive, oxidizing tryptophan not only under acidic conditions but at pH 8.5, too. When C5 was treated with either N-Cl-S or Cl-T at pH 8.5, no CSb-like activity was seen. Control experiments showed that CSb-like C5, generated at Oxidation with N-Cl-S pH 7.2, became rather unstable when brought to pH 8.5 and was unable to form an active complex with C6 after Treatment of human C5 with N-Cl-S in barbitalreneutralization. In order to ensure that the specificity saline also led to the generation of CSb-like activity, of Cl-T for methionine held also for pH 7.2 where the accompanied by partial loss of native C5 function. Optimal yields of CSb-like activity (lO,OOO-20,000 experiments with C5 were conducted, mixtures of leucine and methionine CHSO/mg C5) were obtained when the molar ratio of (as internal standard), tryptophan oxidant to C5 protein was around 25 : 1.) i.e. about 1 mol (1 mM each) were incubated with 3 mM Cl-T in barbitalN-Cl-S per mol methionine. Higher doses of N-Cl-S saline, pH 7.2, under the same conditions as in the experiments with C5. After 10min at room temp. caused considerable loss of CSb-like activity indicating excessive damage to the C5 protein (Fig. 4). Like Cl-T, excess Cl-T was inactivated with N-acetyl-methionine N-Cl-S did not cause cleavage of any peptide chain in C5 (Shechter et al., 1975). Subsequent amino acid analysis revealed that methionine was completely lost (Fig. 3).
-c5 cm,
c5a-
cq3-c5p-
w 1234567 Fig. 3. Gradient polyacrylamide gel electrophoresis of native and oxidized C5. Lanes 14, reduced samples; lanes 5-7, non-reduced samples. 1, C4 marker; 2 and 5, C5 oxidized with N-Cl-S (molar ratio 1:25); 3 and 6, C5 oxidized with Cl-T (molar ratio 1:60); 4 and 7, native C5.
254 Moles N-Cl-S / Mole C5 12.5 I
37.5
25
G; 020000-
50 I -100
0
I 25
I 50
, 0 75pM
[N-Cl-S]
Fig. 4. Conversion of native C5 to CSb-like C5 by oxidation with N-Cl-S. Concentration dependence. Same conditions as in Fig. 1 legend, except that Nl-Cl-S was used as oxidant, in concns as indicated.
while 96% f 10 S.D. of the tryptophan were recovered (N = 3). Cl-T would oxidize also the free SH-groups of cysteine residues in proteins. To our knowledge there is no report on the existence of such groups in human C5. We checked for the presence or absence of free cysteine residues by attempting to bind C5 to activated ThiolSepharaose. Fifty microlitres packed activated ThiolSepharose (containing about 50 nmol activated thiol groups) were mixed with 20 ~1 C5 (about 0.1 nmol) in 0.05 M phosphate buffer, pH 7.0, containing 0.3 M NaCl and 1 mM EDTA. The mixture was gently shaken at room temperature for lOmin, then centrifuged. C5 estimation in the supernatant showed 90% recovery, i.e. the C5 had not been bound and, hence, did not contain free SH-groups of cysteine residues. Demonstration
of methionine
oxidation
in C5 by Cl-T
Forty microgrammes human C5 (0.2 nmol) were treated with 16 nmol Cl-T (0.29 mM) in 55 ~1 barbital saline for 10 min at room temp. i.e. under conditions found to be optimal for the yield of CSb-like C5. Then excess oxidant was inactivated with the OH analogue of methionine. In the products and in nonoxidized control samples methionine sulphoxide was determined by amino acid analysis (Shechter et al., 1975; see Materials and Methods). The original C5 preparation contained 1.61% (molar) methionine, of which 0.13% were already present as methionine sulfoxide; in the Cl-T-treated sample 1.09% methionine sulphoxide were found. Accordingly, 0.96% methionine had been oxidized to the sulphoxide by the treatment with Cl-T. In a second set of experiments a C5 preparation was used which had been stored at 0°C for 2 weeks before use. Here, 1.55% methionine were found in the preparation of which 0.25% had been spontaneously oxidized; the treatment with Cl-T led to additional oxidation of 0.89%. Thus 57759% of the methionine residues present in the original C5 preparation were oxidized by Cl-T to the sulphoxide under the prevailing conditions.
DISCUSSION In the course of complement activation C5 is cleaved by a C5 convertase fixed to the target. The larger of the two cleavage products, C5b, expresses a binding site for C6, and with this new property, represents the activated C5 (the resulting C5b6 complex combines with the late complement components, C7C9; the product, C5b-9, inserts in the target membrane, disturbs its continuity and causes cell lesions). The oxidation of C5 by either Cl-T or N-Cl-S proceeds without cleavage of a peptide bond. Hence, no C5a fragment is released but the whole C5 protein acquires a conformation which allows it to bind C6 and in this way to start the formation of the membrane attack complex. Insofar, the oxidation of C5 by the oxidants used means activation to a CSb-like form. The main functional difference between C5b generated by a complement convertase and CSb-like C5 produced by oxidation is the higher stability of the latter: while the physiologically released C5b loses its C6 binding site if this is not immediately occupied by the ligand, and becomes inactivated (DiScipio et al., 1983) the CSb-like product of C5 oxidation keeps its C6 binding capacity much longer. C6 was regularly added only 10 min after stopping the oxidation with methionine, i.e. around 20 min after beginning of the oxidation, and yet high yields of CSb-like activity were obtained. Even when the addition of C6 was delayed by 1 hr around 90% of the CSb-like activity was preserved. Much more CSb-like activity-12,00&25,000 CH50/ mg C5was obtained by oxidation with Cl-T than earlier by exposure of C5 to oxygen free radicals200&5000 CHSO/mg (Vogt et al., 1991). Like radical treatment, chemically induced activation is accompanied by loss of native C5 function. This is partially due to the higher efficiency in complement lysis of native C5 compared with preformed C5b6; native C5 is activated at the target site, and can there be inserted directly after the CSb6 complex formation. Preformed C56 complexes diffuse freely in the medium, and may be inactivated by forming S-C5-9 complexes in the
255
C5 activation by methionine oxidation fluid phase which can no longer be inserted in a membrane. The difference in efficiency has been discussed earlier (Vogt et al., 1989). On the other hand, some loss of native C5 function seems to be due to nonspecific damage to the C5 protein induced by the oxidant. This is clear from the different slopes of the time courses of loss of native C5 and of generation of CSb-like C5. While N-Cl-S oxidizes tryptophan in addition to methionine, Cl-T has been found to be selective for the latter at pH 8.5 (Shechter et al., 1975). At pH 2.2 it also oxidizes tryptophan. Because of the largely increased lability at pH 8.5 of the CSb-like C5 product of oxidation, treatment of C5 with Cl-T at this pH was not possible. It had to be confirmed, therefore, that at pH 7.2 which was used for the oxidation experiments, Cl-T did not attack tryptophan but only methionine. The oxidation of model mixtures consisting of leucine, methionine and tryptophan proved that Cl-T was specific for methionine at this pH, too. Cysteine would be oxidized by either oxidant used. However, in agreement with lacking positive reports we excluded the presence of cysteine residues in C5 by showing that it is not bound to activated ThiolSepharose. This finding confirms own earlier results (Vogt et al., 1984), in which it was further shown that a related protein, namely inactivated human C3 which has one free cysteine residue, is bound to activated ThiolSepharose. So, all available evidence suggests that at least Cl-T, probably also N-Cl-S, activate C5 by oxidation of methionine residue(s) in C5. The oxidation by Cl-T of about 60% of the methionine residues of C5 to the sulfoxide, under conditions which allow optimal generation of CSb-like C5, was demonstrated by amino acid analysis. It is unlikely that oxidation of all these residues is necesary for the activation. Rather the conversion of some of them may be adverse, since excess treatment with oxidants led to reduction of CSb-like activity. N-Cl-S was more aggressive than Cl-T, and 25 mol N-Cl-S per mol C5 was found optimal for production of CSb-like C5 activity. Higher amounts reduced the yield considerably. The C5 protein has been reported to contain 23 methionine residues (DiScipio et al., 1983). This would mean that about 1 mol N-Cl-S per methionine residue was sufficient for optimal oxidation of C5. For Cl-T 34 mol per mol methionine (80 mol per mol C5) were needed for optimal results. When it was first shown that OH radicals convert C5 to an activated, CSb-like, species the radicals were generated by hydrogen peroxide plus traces of iron (Vogt et al., 1989). Methionine is-besides cysteinethe most sensitive amino acid in proteins to oxidation by H,O,, especially when metal ions are present (Toennies and Callan, 1939; Caldwell and Tappel, 1964; Wolff et al., 1986). In these reactions it is likely that OH radicals are involved (Wolff et al., 1986). (For the generation of hydroxyl radicals from hydrogen peroxide metal ions are absolutely essential; whether the oxidation
of methionine by hydrogen peroxide would proceed at all in the strict absence of metal ions appears not to have been studied.) These observations suggest that the earlier discovered activation of C5 by hydroxyl radicals is also based on oxidation of methionine residues in C5. It is proposed that the conversion of certain methionine residues in C5 to the sulphoxides by radicals or chemicals leads to changes in intra-molecular binding forces of the protein with subsequent conformational rearrangements which express a binding site for C6. Acknowledgement-We thank discussions about determination sulphoxide.
Dr H. Kratzin for valuable of methionine and methionine
REFERENCES Caldwell K. A. and Tappel A. L. (1964) Reactions of selenoand sulfoamino acids with hydroperoxide. Biochemistry 3,
1643-1647. Dessauer A., Rother U. and Rother K. (1984) Freezethaw activation of the complement attack phase: I. Separation of two steps in the formation of the active C56 complex. Acta Path. Microbial. Immun. &and. C92, Suppl. 284, 75-81. DiScipio R. G., Smith C. A., Miiller-Eberhard H. J. and Hugli T. E. (1983) The activation of human complement component C5 by a fluid phase C5 convertase. J. biol. Chem. 258, 10,629-10,636. Johnson D. and Travis J. (1979) The oxidative inactivation of human a-1-proteinase inhibitor. Further evidence for methionine at the reactive center. J. biol. Chem. 254,
40224026. Knecht R. and Chang J.-Y. (1986) Liquid chromatographic determination of amino acids after gas-phase hydrolysis and derivatization with (dimethylamino)azobenzenesulfonyl chloride. Analyt. Chem. 58, 2375-2379. Laemmli U. K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature
(Lend.) 227, 680-685. Rother U., Hansch G., Rauterberg F. W., Jungfer H. and Rother K. (1978) Deviated lysis: lysis of unsensitized cells by complement. V. Generation of the activity by low pH or low ionic strength. Z. Zmmun.-Forsch. 155, 118-129. Shechter Y., Burstein Y. and Patchornik A. (1975) Selective oxidation of methionine residues in proteins. Biochemistry
14, 44974503. Toennies G. and Callan T. P. (1939) Methionine studies. III. A comparison of oxidative reactions of methionine, cysteine, and cystine. Determination of methionine by hydrogen peroxide oxidation. J. biol. Chem. 129, 481490. Vogt W., Damerau B., von Zabern I., Nolte R. and Brunahl D. (1989) Non-enzymic activation of the fifth component of human complement, by oxygen radicals. Some properties of the activation product, CSb-like C5. Molec. Immun. 26, 1133-l 142. Vogt W., Ltihmann B. and Hesse D. (1984) “Inactivated” third component of complement (C3b-like C3; C3i) acquires C5 binding capacity and supports C5 activation upon covalent fixation to a solid surface. Complement 1, 87-96. Vogt W., Nolte R. and Brunahl D. (1991) Binding of iron to the 5th component of human complement directs
256
W. Veer
oxygen radical-mediated conversion to specific sites and causes non-enzymic activation. Complement Injlammat. (in press). von Zabern I., Damerau B., Nolte R., Griinefeld E. and Vogt W. (1985) Generation of anaphylatoxin activity related
et al. to C3a, by treatment of human serum with the nitrogen nucleophile N,H, or the chaotrope KSCN. Stand. J. Immun. 22, 639-65 1. Wolff S. P., Garner A. and Dean R. T. (1986) Free radicals, lipids, and protein degradation. TIBS 11, 27-31.
Molecular Immunology, Vol. 29, No. 2, pp. 251-256, 1992 Printed in Great Britain.
ACTIVATION OF THE FIFTH COMPONENT OF HUMAN COMPLEMENT, C5, WITHOUT CLEAVAGE, BY METHIONINE OXIDIZING AGENTS W. VOGT,*B. ZIMMERMANN, D. HESSEand R. NOLTE Max Planck Institute for Experimental Medicine, Hermann D-3400 Giittingen, Germany
Rein-Strasse 3,
(First received 17 March 1991; accepted in revised form 6 h4ay 1991)
Abstract-Purified human CS was incubated with chloramine T (Cl-T) or N-chloro-succinimide (N-Cl-S) in barbital buffer, pH 7.2. The treatment led to C5 activation: Cl-T- and N-Cl-S-treated C5 acquired a binding site for C6; upon incubation with C6 and subsequent addition of C7, C8 and C9 a membrane attack complex formed which lysed non-sensitized guinea pig red cells (reactive lysis). While the physiological activation of C5 follows its specific cleavage, the resulting fragment C5b representing the activated C5 and expressing the C6 binding site, the treatment with the mentioned chemicals does not lead to fragmentation of the C5 protein. So, functionally, the product of the chemical treatment is CSb-like, but chemically, it comprises the whole protein; no C5a is released. Cl-T and N-Cl-S are known to more or less selectively oxidize methionine residues in proteins, dependent on the conditions. Other sensitive amino acid residues are tryptophan and cysteine. Conditions were chosen for treatment of C5 with Cl-T which exclude attack on tryptophan, and we have ensured that human C5 does not contain free cysteine residues. Further, oxidation of about 60% of the methionine residues of C5 by Cl-T was demonstrated by amino acid analysis. So, all evidence points to methionine residue(s) as the site of attack of Cl-T and probably also of N-Cl-S. The oxidation product of methionine, its sulphoxide, may cause a change in structural conformation of C5 which involves expression of the C6 binding site. Earlier it was found that oxidation of C5 by hydroxyl radicals leads to its activation without cleavage. Since the properties of this CSb-like product resemble those of the product of treatment with Cl-T and N-Cl-S, it is suggested that the formerly found activation of human C5 by hydroxyl radicals is also mediated by oxidation of methionine residue(s) in the C5 protein.
INTRODUCTION In its native state the fifth component of complement is not active. To be utilized in complement reactions it has to be activated. Physiological activation of C5 is a specific enzymic process: the protein is cleaved into two fragments which express different functional activities: C5a, the chemotactic peptide significant in inflammatory processes, and C5b, the larger protein fragment that has the transient capability to bind C6 and
in this way to form the nucleus for the assembly of the membrane damaging complex, C5b-9. The cleavage is effected by specific enzymes of the complement system, the C5 convertases of the classical or alternative pathway. In recent years, non-enzymic procedures leading to C5 activation have become known: freezing and thawing or short exposure of C5 solutions to pH 6.4 (Rother et al., 1978; Dessauer et al., 1984). Furthermore, oxygen radicals, in particular hydroxyl radicals, are capable of activating C5 (Vogt et al., 1989, 1991). The non-enzymic activation by hydroxyl radicals is not accompanied by peptide cleavage of the C5 protein chains.
*Author to whom correspondence
should be addressed.
Hence, no C5a is released and no chemotactic activity is generated. Thus activation by oxygen free radicals means conversion of the whole C5 protein to a CSb-like conformation allowing it to bind C6 and to generate the membrane attack complex in co-operation with C7, C8 and C9. The fact that hydroxyl radicals convert C5 to an activated species suggested that the change was induced by oxidation of some amino acid(s) leading to conformational changes of the protein. There is one other protein known to be rather sensitive to oxidative damage in which the locus of damage has been elucidated: u 1 proteinase inhibitor. Here, oxidation of two out of eight methionine residues is responsible for the loss of activity (Johnson and Travis, 1979). Methionine is particularly sensitive to oxidation, and there are two reagents which more or less selectively oxidize methionine residues in proteins: chloramine T (Cl-T) and Nchloro-succinimide (N-Cl-S) (Shechter et al., 1975). We have tested these agents and found that, indeed, they both convert C5 to a CSb-like species without peptide cleavage and thus mimic the effect of hydroxyl radicals, with even better yield. The results suggest that the activation of C5 by the mentioned oxidants and also by oxygen radials is induced by oxidation of methionine residues. 251
252
w. VOW et al. MATERIALS
AND METHODS
Buffers The buffer (VBSGG) used in the classical immune haemolysis assay for estimation of native C5, contained 2.5 mM barbital buffer pH 7.4, 80 mM NaCl, 0.15 mM CaCl,, 1 mM MgCl,, 2.5% glucose and 0.1% gelatin. For estimation of CSb-like C5 activity in reactive lysis assays, the buffer contained 10 mM EDTA (EDTA-VBSGG). Oxidations of C5 were carried out in barbitalsaline (5 mM barbital buffer, pH 7.2; 0.15 M NaCl). Complement
components
The purification of human C3, C5, C6, C7, C8 and C9 has been described earlier (Vogt et al., 1984, 1989). A stock solution of purified C5 in barbital-saline (0.8 mg/ml) was kept in small portions over liquid nitrogen. Before use, it was incubated for 16 hr at 37°C in order to inactivate any preformed CSb-like C5 (Vogt et al., 1991). Other complement components, used for C5 estimations were bought from Cordis Corporation. Chemicals Chloramine T (Cl-T), N-chloro-succinimide (N-Cl-S), L-methionine, its N-acetyl derivative and OH analogue were purchased from Sigma. Activated Thiol-Sepharose 4B was a preparation from Pharmacia. Before use it was swollen and washed as recommended by the manufacturer. Oxidation
of C5
Solutions of human C5 in barbital-saline of pH 7.2 were incubated with Cl-T or N-Cl-S at varying concns for varying times, at room temp. The reaction was stopped by addition of excess L-methionine (for functional assays) or its hydroxy analogue (for amino acid analysis). The calculations of molar ratios of oxidant to C5 protein were based on a molecular mass of C5 of 200 kDa.
procedure developed by Shechter et al. (1975) was used. In brief, the samples of C5 were treated with cyanogen bromide to convert non-oxidized methionine to homoserine (lactone). Then they were hydrolyzed with 6 N HCl in the presence of dithiothreitol to reduce methionine sulphoxide back to methionine. Thus any methionine determined in the subsequent amino acid analysis referred to the sulphoxide. Amino acid determination of C5 hydrolysates was carried out after precolumn derivatization with (dimethylamino)azobenzene-sulphonyl chloride according to the method of Knecht and Chang (1986). Gradient polyacrylamide
gel electrophoresis
Reduced and non-reduced samples gradient gels as described (Laemmli, et al., 1985). RESULTS
Oxidation
of C5 with Cl-T
C5 was treated with Cl-T solutions in barbital-saline buffer, pH 7.2, for 10 min at room temp. Then methionine was added to inactivate unreacted Cl-T. After further 10 min the product was mixed with C6, incubated for 3 hr at 37°C to allow the development of an active C56 complex from any CSb-like C5 formed (Vogt et al., 1989, 1991), and tested for C56 activity in the reactive lysis assay. Further, residual functional C5 activity was tested in the classical immune haemolysis assay. The tests showed that the treatment with Cl-T resulted in partial loss of native C5 function and in the generation of CSb-like activity. Figure 1 shows the dependence of the yield on the concn of Cl-T. Highest yields (12,00&25,000 CHSO/mg C5) were obtained when the incubation mixture contained 80 mol Cl-T per mol of C5. In Fig. 2 the time kinetics are shown. Regularly, the CSb-like activity Moles Chloramine T/ Mole C5 LO
Estimation
of C5 were run on 1970; von Zabern
of native C5
60 I
80
100 I
100
C5 was assayed by its function in the classical immune haemolysis assay, using sheep red cells in the state EAC14, varying dilutions of the C5 preparation to be tested, excess C2gp, C3h”, and C6C9gp. Estimation
of CSb-like
C5 activity
The products of C5 oxidation were mixed with C6h”, diluted with EDTA-VBSGG, incubated for 3 hr at 37°C and then tested in a reactive lysis assay system, using non-sensitized guinea pig red cells, and human C7, C8 and C9 (Vogt et al., 1989, 1991). Activities are expressed in CHSO/mg C5. Amino acid sulphoxide
analysis;
determination
of
methionine
Free amino acids in model mixtures were quantified in an amino acid analyser, with automatic ninhydrin staining of the effluent. For the determination of methionine sulphoxide in (oxidized) C5 preparations the
”
I
0
$0
160
liO,,M-
[ Chloramine T]
Fig. 1. Conversion of native C5 to CSb-like C5 by chloramine T. Concentration dependence. Eight microgrammes CS (0.04 nmol) were incubated with varying concns of Cl-T, as indicated in the lower abscissa. Total volume 30~1 barbitalsaline, pH 7.2. After 10 min incubation at room temp. the reaction was stopped by addition of 250pM methionine. CbS-like activity was measured by reactive lysis assay; loss of C5 refers to native C5 which was assayed by its function in the classical immune haemolysis test (see Materials and Methods).
C5 activation by methionine G 0
253
oxidation
20000-100
F 2 Y 2 % c g
__s
10000-
-50
m aB Y 7 iz 0
%
z = : 7 x
0 0
I
I L
I 12
a
, 0 16 min
Time
Fig. 2. Conversion of native C5 to CSb-like C5 by chloramine T. Time kinetics. Thirty-two microgrammes C5 (0.16 nmol) were incubated with 12.8 nmol of Cl-T in 80 ~1 barbital-saline buffer, at room temp. At varying times, portions were withdrawn and injected into methionine solution to stop the reaction. Assays as in Fig. 1 legend.
the loss of functional native C5 proceeded. The peak of CSb-like activity was reached in about l&12 min. Thereafter, the activity slowly declined. In Fig. 3 a gradient polyacrylamide gel electrophoresis of Cl-T-treated and untreated C5 is presented. Neither the non-reduced C5 oxidation products nor the reduced samples show any cleavage of the C5 protein by Cl-T.
developed
more
slowly
than
Specljicity of Cl-T for methionine oxidation
Shechter et al. (1975) found that at pH 8.5 of all natural amino acids only methionine and cysteine were oxidized by Cl-T, while at pH 2.2 Cl-T attacked tryptophan, in addition. N-Cl-S was more aggressive, oxidizing tryptophan not only under acidic conditions but at pH 8.5, too. When C5 was treated with either N-Cl-S or Cl-T at pH 8.5, no CSb-like activity was seen. Control experiments showed that CSb-like C5, generated at Oxidation with N-Cl-S pH 7.2, became rather unstable when brought to pH 8.5 and was unable to form an active complex with C6 after Treatment of human C5 with N-Cl-S in barbitalreneutralization. In order to ensure that the specificity saline also led to the generation of CSb-like activity, of Cl-T for methionine held also for pH 7.2 where the accompanied by partial loss of native C5 function. Optimal yields of CSb-like activity (lO,OOO-20,000 experiments with C5 were conducted, mixtures of leucine and methionine CHSO/mg C5) were obtained when the molar ratio of (as internal standard), tryptophan oxidant to C5 protein was around 25 : 1.) i.e. about 1 mol (1 mM each) were incubated with 3 mM Cl-T in barbitalN-Cl-S per mol methionine. Higher doses of N-Cl-S saline, pH 7.2, under the same conditions as in the experiments with C5. After 10min at room temp. caused considerable loss of CSb-like activity indicating excessive damage to the C5 protein (Fig. 4). Like Cl-T, excess Cl-T was inactivated with N-acetyl-methionine N-Cl-S did not cause cleavage of any peptide chain in C5 (Shechter et al., 1975). Subsequent amino acid analysis revealed that methionine was completely lost (Fig. 3).
-c5 cm,
c5a-
cq3-c5p-
w 1234567 Fig. 3. Gradient polyacrylamide gel electrophoresis of native and oxidized C5. Lanes 14, reduced samples; lanes 5-7, non-reduced samples. 1, C4 marker; 2 and 5, C5 oxidized with N-Cl-S (molar ratio 1:25); 3 and 6, C5 oxidized with Cl-T (molar ratio 1:60); 4 and 7, native C5.
254 Moles N-Cl-S / Mole C5 12.5 I
37.5
25
G; 020000-
50 I -100
0
I 25
I 50
, 0 75pM
[N-Cl-S]
Fig. 4. Conversion of native C5 to CSb-like C5 by oxidation with N-Cl-S. Concentration dependence. Same conditions as in Fig. 1 legend, except that Nl-Cl-S was used as oxidant, in concns as indicated.
while 96% f 10 S.D. of the tryptophan were recovered (N = 3). Cl-T would oxidize also the free SH-groups of cysteine residues in proteins. To our knowledge there is no report on the existence of such groups in human C5. We checked for the presence or absence of free cysteine residues by attempting to bind C5 to activated ThiolSepharaose. Fifty microlitres packed activated ThiolSepharose (containing about 50 nmol activated thiol groups) were mixed with 20 ~1 C5 (about 0.1 nmol) in 0.05 M phosphate buffer, pH 7.0, containing 0.3 M NaCl and 1 mM EDTA. The mixture was gently shaken at room temperature for lOmin, then centrifuged. C5 estimation in the supernatant showed 90% recovery, i.e. the C5 had not been bound and, hence, did not contain free SH-groups of cysteine residues. Demonstration
of methionine
oxidation
in C5 by Cl-T
Forty microgrammes human C5 (0.2 nmol) were treated with 16 nmol Cl-T (0.29 mM) in 55 ~1 barbital saline for 10 min at room temp. i.e. under conditions found to be optimal for the yield of CSb-like C5. Then excess oxidant was inactivated with the OH analogue of methionine. In the products and in nonoxidized control samples methionine sulphoxide was determined by amino acid analysis (Shechter et al., 1975; see Materials and Methods). The original C5 preparation contained 1.61% (molar) methionine, of which 0.13% were already present as methionine sulfoxide; in the Cl-T-treated sample 1.09% methionine sulphoxide were found. Accordingly, 0.96% methionine had been oxidized to the sulphoxide by the treatment with Cl-T. In a second set of experiments a C5 preparation was used which had been stored at 0°C for 2 weeks before use. Here, 1.55% methionine were found in the preparation of which 0.25% had been spontaneously oxidized; the treatment with Cl-T led to additional oxidation of 0.89%. Thus 57759% of the methionine residues present in the original C5 preparation were oxidized by Cl-T to the sulphoxide under the prevailing conditions.
DISCUSSION In the course of complement activation C5 is cleaved by a C5 convertase fixed to the target. The larger of the two cleavage products, C5b, expresses a binding site for C6, and with this new property, represents the activated C5 (the resulting C5b6 complex combines with the late complement components, C7C9; the product, C5b-9, inserts in the target membrane, disturbs its continuity and causes cell lesions). The oxidation of C5 by either Cl-T or N-Cl-S proceeds without cleavage of a peptide bond. Hence, no C5a fragment is released but the whole C5 protein acquires a conformation which allows it to bind C6 and in this way to start the formation of the membrane attack complex. Insofar, the oxidation of C5 by the oxidants used means activation to a CSb-like form. The main functional difference between C5b generated by a complement convertase and CSb-like C5 produced by oxidation is the higher stability of the latter: while the physiologically released C5b loses its C6 binding site if this is not immediately occupied by the ligand, and becomes inactivated (DiScipio et al., 1983) the CSb-like product of C5 oxidation keeps its C6 binding capacity much longer. C6 was regularly added only 10 min after stopping the oxidation with methionine, i.e. around 20 min after beginning of the oxidation, and yet high yields of CSb-like activity were obtained. Even when the addition of C6 was delayed by 1 hr around 90% of the CSb-like activity was preserved. Much more CSb-like activity-12,00&25,000 CH50/ mg C5was obtained by oxidation with Cl-T than earlier by exposure of C5 to oxygen free radicals200&5000 CHSO/mg (Vogt et al., 1991). Like radical treatment, chemically induced activation is accompanied by loss of native C5 function. This is partially due to the higher efficiency in complement lysis of native C5 compared with preformed C5b6; native C5 is activated at the target site, and can there be inserted directly after the CSb6 complex formation. Preformed C56 complexes diffuse freely in the medium, and may be inactivated by forming S-C5-9 complexes in the
255
C5 activation by methionine oxidation fluid phase which can no longer be inserted in a membrane. The difference in efficiency has been discussed earlier (Vogt et al., 1989). On the other hand, some loss of native C5 function seems to be due to nonspecific damage to the C5 protein induced by the oxidant. This is clear from the different slopes of the time courses of loss of native C5 and of generation of CSb-like C5. While N-Cl-S oxidizes tryptophan in addition to methionine, Cl-T has been found to be selective for the latter at pH 8.5 (Shechter et al., 1975). At pH 2.2 it also oxidizes tryptophan. Because of the largely increased lability at pH 8.5 of the CSb-like C5 product of oxidation, treatment of C5 with Cl-T at this pH was not possible. It had to be confirmed, therefore, that at pH 7.2 which was used for the oxidation experiments, Cl-T did not attack tryptophan but only methionine. The oxidation of model mixtures consisting of leucine, methionine and tryptophan proved that Cl-T was specific for methionine at this pH, too. Cysteine would be oxidized by either oxidant used. However, in agreement with lacking positive reports we excluded the presence of cysteine residues in C5 by showing that it is not bound to activated ThiolSepharose. This finding confirms own earlier results (Vogt et al., 1984), in which it was further shown that a related protein, namely inactivated human C3 which has one free cysteine residue, is bound to activated ThiolSepharose. So, all available evidence suggests that at least Cl-T, probably also N-Cl-S, activate C5 by oxidation of methionine residue(s) in C5. The oxidation by Cl-T of about 60% of the methionine residues of C5 to the sulfoxide, under conditions which allow optimal generation of CSb-like C5, was demonstrated by amino acid analysis. It is unlikely that oxidation of all these residues is necesary for the activation. Rather the conversion of some of them may be adverse, since excess treatment with oxidants led to reduction of CSb-like activity. N-Cl-S was more aggressive than Cl-T, and 25 mol N-Cl-S per mol C5 was found optimal for production of CSb-like C5 activity. Higher amounts reduced the yield considerably. The C5 protein has been reported to contain 23 methionine residues (DiScipio et al., 1983). This would mean that about 1 mol N-Cl-S per methionine residue was sufficient for optimal oxidation of C5. For Cl-T 34 mol per mol methionine (80 mol per mol C5) were needed for optimal results. When it was first shown that OH radicals convert C5 to an activated, CSb-like, species the radicals were generated by hydrogen peroxide plus traces of iron (Vogt et al., 1989). Methionine is-besides cysteinethe most sensitive amino acid in proteins to oxidation by H,O,, especially when metal ions are present (Toennies and Callan, 1939; Caldwell and Tappel, 1964; Wolff et al., 1986). In these reactions it is likely that OH radicals are involved (Wolff et al., 1986). (For the generation of hydroxyl radicals from hydrogen peroxide metal ions are absolutely essential; whether the oxidation
of methionine by hydrogen peroxide would proceed at all in the strict absence of metal ions appears not to have been studied.) These observations suggest that the earlier discovered activation of C5 by hydroxyl radicals is also based on oxidation of methionine residues in C5. It is proposed that the conversion of certain methionine residues in C5 to the sulphoxides by radicals or chemicals leads to changes in intra-molecular binding forces of the protein with subsequent conformational rearrangements which express a binding site for C6. Acknowledgement-We thank discussions about determination sulphoxide.
Dr H. Kratzin for valuable of methionine and methionine
REFERENCES Caldwell K. A. and Tappel A. L. (1964) Reactions of selenoand sulfoamino acids with hydroperoxide. Biochemistry 3,
1643-1647. Dessauer A., Rother U. and Rother K. (1984) Freezethaw activation of the complement attack phase: I. Separation of two steps in the formation of the active C56 complex. Acta Path. Microbial. Immun. &and. C92, Suppl. 284, 75-81. DiScipio R. G., Smith C. A., Miiller-Eberhard H. J. and Hugli T. E. (1983) The activation of human complement component C5 by a fluid phase C5 convertase. J. biol. Chem. 258, 10,629-10,636. Johnson D. and Travis J. (1979) The oxidative inactivation of human a-1-proteinase inhibitor. Further evidence for methionine at the reactive center. J. biol. Chem. 254,
40224026. Knecht R. and Chang J.-Y. (1986) Liquid chromatographic determination of amino acids after gas-phase hydrolysis and derivatization with (dimethylamino)azobenzenesulfonyl chloride. Analyt. Chem. 58, 2375-2379. Laemmli U. K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature
(Lend.) 227, 680-685. Rother U., Hansch G., Rauterberg F. W., Jungfer H. and Rother K. (1978) Deviated lysis: lysis of unsensitized cells by complement. V. Generation of the activity by low pH or low ionic strength. Z. Zmmun.-Forsch. 155, 118-129. Shechter Y., Burstein Y. and Patchornik A. (1975) Selective oxidation of methionine residues in proteins. Biochemistry
14, 44974503. Toennies G. and Callan T. P. (1939) Methionine studies. III. A comparison of oxidative reactions of methionine, cysteine, and cystine. Determination of methionine by hydrogen peroxide oxidation. J. biol. Chem. 129, 481490. Vogt W., Damerau B., von Zabern I., Nolte R. and Brunahl D. (1989) Non-enzymic activation of the fifth component of human complement, by oxygen radicals. Some properties of the activation product, CSb-like C5. Molec. Immun. 26, 1133-l 142. Vogt W., Ltihmann B. and Hesse D. (1984) “Inactivated” third component of complement (C3b-like C3; C3i) acquires C5 binding capacity and supports C5 activation upon covalent fixation to a solid surface. Complement 1, 87-96. Vogt W., Nolte R. and Brunahl D. (1991) Binding of iron to the 5th component of human complement directs
256
W. Veer
oxygen radical-mediated conversion to specific sites and causes non-enzymic activation. Complement Injlammat. (in press). von Zabern I., Damerau B., Nolte R., Griinefeld E. and Vogt W. (1985) Generation of anaphylatoxin activity related
et al. to C3a, by treatment of human serum with the nitrogen nucleophile N,H, or the chaotrope KSCN. Stand. J. Immun. 22, 639-65 1. Wolff S. P., Garner A. and Dean R. T. (1986) Free radicals, lipids, and protein degradation. TIBS 11, 27-31.
