Rapid isolation and biochemical characterization of rat C1 and C1q

Molecular Immunology, Vol. 30, No. 5, pp. 43340, Printed in Great Britain.

RAPID

ISOLATION

1993

0161-5890/93$6.00 + 0.00 Pergamon Press Ltd

AND BIOCHEMICAL CHARACTERIZATION RAT Cl AND Clq

OF

M. G. WING,* D. J. SEILLY, D. J. BRIDGMAN and R. A. HARRISON Molecular Immunopathology

Unit, MRC Centre, Hills Road, Cambridge CB2 2QH, U.K.

(First received 10 August 1992, accepted in revised form 17 October 1992)

Abstract-Using a human IgG-Sepharose column to which rabbit anti-human IgG was bound (rabbit anti-human/human IgG-Sepharose), human and rat Cl or Clq were isolated from serum in a single step, and the Clq further purified to homogeneity by FPLC. This procedure allowed the rapid isolation of haemolytically active Cl or Clq, with a yield equal to or greater than published methods. The availability of human and rat Clq allowed comparison of the two molecules, revealing differences in their mobility on SDS-PAGE as well as on agarose gel electrophoresis. Amino terminal sequence analysis demonstrated greater than 78% residue identity between rat Clq A, B and C chains and the published human and mouse sequences. Similar amino acid compositions suggest that the homology extends throughout the molecules. In addition to the major A:B and C:C dimer bands, rat, unlike human Clq, contained minor dimer species. These may reflect heterogeneity in glycosylation and or lysine and proline hydroxylation.

INTRODUCTION Cl is a calcium-dependent trimolecular complex of Clq, r and s in the molar ratio of 1:2:2 in man. Activation

of the classical pathway of complement follows the interaction of the globular head of Clq with multiple Fc regions on immunoglobulins or with polyanionic structures (see Lachmann and Hughes-Jones, 1984 for review). This interaction results in a conformational change in the Clq molecule (Heinz, 1989) leading to activation of the proenzyme Clr, which in turn activates Cls. Cls then cleaves C4 and C2 resulting in the generation of a C3 convertase which leads to the deposition of C3b on the cell membrane and initiation of the membrane attack complex assembly. Purification of active human Cl and Clq is made difficult by the low serum concn of Cl and its susceptibility to activation during purification. In an attempt to solve these problems, a number of purification protocols have been devised for human C 1. These include euglobulin precipitation followed by gel filtration (Gigli et al., 1976) or alternatively, binding Cl to aggregated IgG (Miiller-Eberhard and Kunkel, 1961), immune complexes (Arlaud, 1979) or IgG-Sepharose (Bing, 1971; *Author to whom correspondence should be addressed. Abbreviations: FPLC, fast protein liquid chromatography; Cl, first component of complement; Clq, r, s, subcomponents of Cl; IgG, immunoglobulin G; VBS, Verona1 buffered saline; NPGB, p-nitrophenyl p-guanidinobenzoate; PBSAZ, phosphate buffered saline plus 10mM sodium azide; EDTA, ethylendiaminetetra-acetic acid disodium salt; SDS-PAGE, sodium dodecyl sulphate-polyacrylamide gel electrophoresis; HEPES, N-2-hydroxyethl piperazine-N-2ethane sulphonic acid; CAPS, 3-(cyclohexylamino)I-propanesulfonic acid; E, sheep erythrocytes; EA, antibody coated sheep erythrocytes; CFD, complement fixing diluent.

Medicus and Chapuis, 1980). Early attempts were made to utilize these protocols for rat Cl purification, however the yield and purity of these preparations was poor. In the procedure described here, modification of a human IgG-Sepharose column, by allowing a rabbit polyclonal anti-human IgG antiserum to bind to it, resulted in the efficient purification of rat Cl or Clq in a single step from serum. The Clq was further purified to homogeneity by FPLC on a Mono S cation exchange column, as described by Stemmer and Loos (1984) allowing comparative biochemical analysis of human and rat Clq to be made. MATERIALS

AND METHODS

Serum Rat blood was obtained by cardiac puncture, clotted for 2 hr at 37°C then the serum harvested following centrifugation at 4500g. Human blood was obtained from healthy volunteers and the serum prepared as above. Bufirs Verona1 buffered saline (VBS) comprised 0.14 M sodium chloride, 5 mM sodium barbitone, 4 mM sodium azide, pH 7.4. A stock solution of 30 mM p -nitrophenyl p-guanidinobenzoate (NPGB) was made up in formamide. Rabbit anti-human IgG-Sepharose

column

A human IgG preparation was made from fresh serum by caprylic acid precipitation (Steinbuch and Audran, 1969). The serum pH was adjusted to 5.0 with 14% acetic acid, then l/20 volume of caprylic acid was added with constant stirring at room temp. After 30 min, the precipitate was centrifuged at 50,000 g. The supernatant was then reprecipitated with 2/3 volume of saturated 433

434

M. G.

WI~\IGet al.

ammonium sulphate. After 30 min on ice, the precipitate was centrifuged at 5O,OOOg,redissolved in H@, then extensively dialysed against 0.1 M borate buffer, 0.15 M NaCl (pH 8.0). Approximately 250 mg of IgG was offered to 70ml of cyanogen bromide activated Sepharose in 0.1 M Na,BO,, 0.15 M NaCl (pH 8.0), resulting in 3 mg of IgG being coupled per ml of Sepharose. Fifty-five ml of the above human IgG-Sepharose column was incubated with 55 ml of rabbit anti-human IgG serum with constant rotation for 1.5 hr at 4°C. The gel was then poured into a 30 x 230mm column, washed with three column volumes of phosphate buffered saline plus 10 mM sodium azide (PBS-AZ), followed by three column volumes of I M NaCl and finally equilibrated with VBS buffer with or without 10 mM EDTA. C 1 and C 1q preparation The buffers and flow rates used for preparation of the human and rat Cl were essentially those reported by Medicus and Chapuis (1980) with minor modifications which are described below. All procedures were carried out at 4°C. For Cl preparation, human or rat serum was half diluted with VBS plus 0.15 mM CaCl,, t mM MgCI, and NPGB to a final concn of 3 x IO-’ M, then loaded on to the rabbit anti-human/human IgG-Sepharose column. All subsequent steps were carried out in this buffer. The flow rate of the column was two column volumes/hr for sample application, Unbound protein was washed off, then the Cl eluted with VBS plus 0.15 mM CaCl,, 1 mM MgC12 with 0.4 or 1.OM NaCl for human and rat Cl, respectively. Clq was prepared as for Cl, except that that the serum was made 10 mM EDTA to dissociate the Clr/s from the Clq and VBS plus 10 mM EDTA was used to dilute the serum and wash the column. Cl or Clq containing fractions were identified by SDS-PAGE and the relevant fractions pooled and concentrated by Amicon ultrafiltration for C 1. To reverse any dissociation of C 1r/s and C Iq from the Cl complex induced by the 1 M NaCl used to elute it from the column, the final ionic strength of the buffers used in the activation assays was adjusted to 0.15 M NaCl. The protein-containing fractions which did not bind to the column during the Clq preparation were pooled, then concentrated back to the original volume of serum and used as a Cl q-depleted reagent. C Iq was further purified by FPLC according to the method of Stemmer and Loos (1984). Clq-containing fractions were dialysed against 20mM HEPES, 60 mM NaCl, 10mM EDTA (pH 7.8), then applied to an FPLC Mono S cation exchange column at room temp, washed, then the Clq was eluted with a NaCl gradient to 1 M.

SDS-PAGE

and electroblotting

Proteins were run reduced or non-reduced on IO--20% SDS-PAGE gradient gels in 0.375M Tris-HCI (pH8.8) according to Laemmli (1970). Iodoacetamide at a final concn of 5 mM was required in the sample preparation buffer of non-reduced samples to prevent dissociation of Cl q dimers during electrophoresis. For two-dimensional electrophoresis, the samples were reduced, electrophoresed on IO-20% SDS-PAGE gels in the first dimension, then the unfixed gel was soaked for 30 min in 0.142 M Tris-HCI, pH 6.8; 0.1% SDS; I mM EDTA. The region of the gel which contained the Ciq was cut out with reference to mol. wt markers which had been run on the gel, then separately fixed and stained with Coomassie blue R. The first dimension slice was then placed horizontally on top of a second lo--20% SDS--PAGE gel and a stack poured above it. V8 protease at 6pg/ml in 0.142 M Tris-HCl pH 6.8; 0.1% SDS; 1 mM EDTA; 10% glycerol; and 0.05% Bromophenol blue was loaded on top of the stack, then the gel electrophoresed (see Cleveland et al., 1977). For amino-terminal sequence analysis, the gels were electroblotted onto Problott (Applied Biosystems) in 10 mM CAPS plus 10% methanol(pH 1I), and stained with Coomassie blue R. Bands of interest were cut out and sequenced on a Applied Biosystems 477A protein sequencer equiped with a 120A analyser and 610A data analysis software. Electrobiotted samples were sequenced on a Blott; TM cartridge. Amino acid compositions were determined using a Applied Biosystems 420H hydrolysis derivitizer equiped with a 130A separation system and 920A data analysis. Agurose gel electro~~ores~.~ A 1% agarose gel was prepared in 23 mM KH,PO,; 10 mM Na,HPO,, pH 6.6; and 5 pg of protein run in the same buffer. The gel was fixed in 12% TCA, and stained with Coomassie blue R. Haemolytic assay A 10% suspension of washed sheep erythrocytes (E) was incubated with a complement-fixing rat anti-sheep erythrocyte antibody for 30 min at 4°C in complement fixing diluent plus 0.1% gelatin (CFD) (Oxoid, Basingstoke, U.K.). The antibody coated sheep erythrocytes (EA) were washed, then resuspended to a 10% suspension. The E and EA were then incubated with serum or the purified complement components for 30 min at 37°C and haemolysis measured by the release of haemoglobin into the supernatant monitored at 412 nm; 100% haemolysis was determined by the lysis of E or EA with water.

Radiolabeling C 1 and C 1q

Functional integrity of C 1

Cl and Clq were iodinated using the iodogen method (Fraker and Speck 1978) and the labeled protein and free ‘*‘I separated on a Sephadex G25 column. Bovine serum albumin was added to a final concn of 0.1% and the iabelled proteins stored at -70°C until required.

To determine the integ~ty of the rat Cl, it was incubated at 37°C for 3 hr to allow autoactivation to proceed, then analysed by SDS-PAGE under reducing conditions to allow the cleavage of the Clrjs molecule into its heavy and light chains to be observed. Biological

Characterization

and isolation of rat Cl and Clq

435

activity of the Cl was examined by incubating “‘I-Cl with a 10% suspension of E or EA for 30 min on ice followed by a further 30 min at 37°C. The cells were washed, then the supernatant and pellet run reduced on SDS-PAGE and autoradiographed. Cl activation was accessed by the cleavage of the Clr/s molecule into their heavy and light chains. RESULTS Yields The yield of rat Cl from the rabbit anti-human IgG-Sepharose column was in the order of 7.4 mg per 100 ml rat serum, which was approximately 85% pure. Known contaminating proteins include IgG and albumin from their mol. wt and amino terminal sequence (data not shown). The same column yielded 5 mg of partially pure rat Clq from 100 ml serum, which following FPLC resulted in a yield of 1.5-2.6 mg of highly purified Clq. The Clq was free of contaminating rabbit, rat or human IgG measured using radial immunodiffusion, which was able to detect antibody at 1 pg/ml (data not shown). Haemolytic

activity

of the C lq preparation

A 10% suspension of EA was incubated with a 1:40 dilution of Clq deficient serum in CFD (Oxoid, U.K.) with or without purified Cl or Clq for 30 min at 37°C. The cells were centrifuged and the percentage lysis determined by the release of haemoglobin. Whilst Clq deficient serum, Cl and Clq failed to lyse EA, Cl and Clq were able to restore the haemolytic activity of the Clq depleted serum to a level comparable to that of whole NRS (Fig. 1). Functional

integrity

of C 1r/s

SDS-PAGE analysis of rat Cl following autoactivation revealed the complete cleavage of the intact 93kDa Clr/s molecule into a 60 kDa heavy chain band, the Clr and s heavy chains resolving poorly under these conditions, and the 36 and 34 kDa light chains (Fig. 2). The absence of activated Cl in the starting material confirms the integrity of the purified protein. To demonstrate

Fig. 2. SDS-PAGE alysis of purified rat Cl following autoactivation. Unactivated (lane 1) or autoactivated Cl (lane 2) was run under reducing conditions on a l&20% gradient SDS-PAGE. The position of intact Clr/s and its heavy and light chains are marked together with mol wt markers. specific activation of Cl, radiolabeled Cl was incubated with a 10% suspension of E or EA in CFD for 30 min at 4°C to facilitate binding, then at 37’C for another 30 min to allow activation to proceed. The cells were pelleted, washed twice, then the cells and supernatant separated by SDSPAGE under reducing conditions and autoradiographed. Incubation of Cl with the EA but not E resulted in the cleavage of the intact Clr/s molecule into the heavy and light chains. This was most clearly seen in the supernatant, but was also visible on the cells (Fig. 3). Comparison of rat and human C lq by polyacrylamide and agarose gel electrophoresis

CCNT

ClqDef Clq NRS

Cl

Clq

Cl

Nxs

+ Clq Del NRS

Fig. 1. Reconstitution of lytic activity of Clq-depleted serum with purified Cl or Clq. Lysis was determined by the release of haemoglobin, which was monitored at 412 nm.

SDS-

When the proteins were analyzed by SDS-PAGE under non-reducing conditions, there was evidence of some A:B and C:C dimer breakdown. This was largely resolved by including iodoacetamide in the sample preparation buffer, to scavenge any free thiol groups (Fig. 4). Comparison of human and rat Clq revealed some differences in the apparent mol. wt of human Clq A, B and C chains which were 35, 32 and 25 kDa compared to 30,28 and 26 kDa for the rat. These accounted for the differences seen in the apparent mol. wt of the human and rat A:B and C:C dimers. In contrast to human Clq, an additional dimer was visible in the rat Clq sample. As well as exhibiting differences in mol. wt, rat Clq displayed a higher isoelectric point compared to human,

M. G. WING et al.

436

amino-terminal sequences of the fragments were consistent with their being derived from a V8 protease digestion, the fragments resulting from cleavage between a predicted glutamic acid and either a lysine or leucine residue. Sequencing of the lower mol. wt dimer identified a C:C homodimer, whilst the higher mol. wt dimer gave comparable levels of A and C chain sequences. The failure to detect a B chain sequence was expected in view of its blocked amino terminus, however the origin of the C chain signal is unclear, and may represent contamination of the A:B heterodimer. The band migrating between the two dimers was identified as containing a C chain also.

c

Amino

acid composition

of human and rat Clq

Amino acid compositions are expressed in moles percent. As seen in Table 1, there is good agreement between the experimentally determined composition of human Clq, and that derived from the published cDNA sequence of the A, B and C chains (Sellar et al., 1991). Comparison of the amino acid composition of human and rat Clq revealed few differences between the two species. A small decrease in the number of isoleucine and tyrosine residues in the rat protein was detected whilst the number of serine residues was increased. DISCUSSION

Fig. 3. Specific activation of Cl. Radiolabeled Cl was incubated with EA or E, then the washed cells (lanes 2-3) and supernatants (lanes 4-5) run on a 10% SDS-PAGE under reduced conditions and autoradiographed. A sample of radiolabeled Cl was run in lane 1.

which was reflected in its higher mobility (Fig. 5). Amino-terminal

sequence

analysis

on agarose

gels

of rat C lq

Amino-terminal sequence analysis of the bands which, though run under non-reducing conditions, had a comparable mobility and pattern to monomeric Clq chains revealed that they had intact N-termini, indistinguishable from the sequences derived from the bands which remained as dimers under identical conditions. Whilst sequences with greater than 78% identity to human or mouse A and C chains were found, no B chain sequences were detected (Fig. 6a). Since the amino terminus of the human B chain is blocked, rat Clq was subjected to V8 protease digestion on a two-dimensional SDS-PAGE gel in order to generate fragments suitable for sequence analysis. Following electroblotting. two fragments were successfully sequenced which readily aligned with the human and mouse Clq B chain. The larger fragment was identified as containing part of the collagen-like region, whilst the smaller fragment was derived from the globular head (Fig. 6b). The size and

Despite efficiently coupling human IgG to CNBractivated Sepharose as described by Medicus and Chapuis (1980) human, as well as rat Cl, failed to bind to this column. This finding, together with the experiences of others, who have also reported the unreliable Clq binding property of IgG-Sepharose columns, resulted in our modifying the protocol. We speculated that whilst IgG density would be an important factor for Clq binding, perhaps the most important variable was the position and number of the covalent bonds between the IgG molecule and the Sepharose. The cyanogen bromide method of coupling proteins to Sepharose potentially utilises the c1amino group of the amino-terminal residue and the el amino group of lysine residues, and it has been shown that the binding site for Clq on the Fc region of IgG involves the species conserved sequence Gly,,,, Lys,,, and LYS,~*(Duncan and Winter. 1988). Consequently, highly efficient protein coupling, could well utilise either or both of these lysine residues, and thus adversely affect Clq binding. Hence we reasoned that a column which had the Fc regions of the IgG molecules accessible to Clq, rather than being potentially involved in the linkage to the Sepharose, would bind Clq more efficiently. To this end a polyclonal rabbit anti-human IgG was added to the column, which by binding through its Fab had its Fc region free to interact with the Clq. This resulted in a column capable of purifying Cl or Clq in a single step from serum, with yields equal to. or greater than other published methods. The Cl was intact as judged by the

Characterization

and isolation of rat Cl and Clq

437

-17 -12

Fig. 4. Comparison of the the relative mobility of rat and human Clq on SDS-PAGE under reduced and non-reduced conditions. Rat or human Clq were run in duplicate on lO_20% gradient SDS-PAGE either reduced (lanes 1417), or non-reduced with (lanes 14) or without (lanes 67) iodoacetamide. Molecular wt markers are included on lanes 11-12. absence of Clr/s cleavage prior to autoactivation 37”C, and its specific activation following exposure antibody-coated sheep erythrocytes demonstrated integrity.

at to its

Fig. 5. Comparison of the relative mobility of native rat and human Clq on agrose gels. Rat or human Clq were run in triplicate MIMM

on a 1% agarose gel in a phosphate to determine their relative isoelectric

3015-B

buffer at pH 6.6 mobility.

The Clq preparations could be further purified to homogeneity by passage over a Mono S cation exchange column. Purification of human Clq using the same protocol as that for rat Clq provided the opportunity for comparative biochemical analysis of Clq from both species. Examination of human and rat Clq by SDS-PAGE showed evidence of dimer breakdown under non-reducing conditions; this could largely be prevented by the addition of iodoacetamide to the loading buffer, indicating that the monomers arise as a consequence of the sensitivity of the interchain disulphide bond to thiol attack. Since the A:B and C:C dimers are held in covalent linkage by the interchain disulphide bonds between the cys-4 residues of each chain, an alternative hypothesis would be that limited proteolysis immediately carboxy-terminal to the cys-4 residues, in one or more of the individual chains had occured, releasing the essentially monomer sized carboxyterminal fragments. This was however ruled out by analysis of the monomer and dimer bands of nonreduced Clq, which showed no difference in their aminoterminal sequences. Whilst only A and C chain sequences were detected in the monomer bands of non-reduced Clq, this was clearly a consequence of the blocked amino terminus of the B chain rather than any difference in sensitivity of the dimers to breakdown, since iodoacetamide protected all dimer species. In vitro breakdown of guinea pig Clq dimers to monomers in the absence of thiol reduction has been reported previously (Hitschold et al., 1983) and whilst human Clq is

M. G. WING et al.

438

15

1

(a)

HUMAN RAT

Clq

Clq

A

EDVCRAPDGKKGEAG ******* ** * ** EDVXRAPNGKDGVAG ************ *

A

MOUSE

Clq

A

EDVCRAPNGKDGAPG

HUMAN

Clq

C

NTGCYGIPGMPGLPGAPGKDGYDGLPGP * ************* ***** ** NAGXYGIPGMPGLPGTPGKDGHXGLQXP

28

1

RAT

Clq

C

*

121 116 55 E FGEKGDPG.........D HVITN **** ***** * * (E)LGEKGDAG........(E)KVITN * ***** ***** * *

41

(b)

HUMAN RAT

Clq

Clq

MOUSE

B

B

Clq

B

E FGEKGDPG.........E

KVITN

The human A,B amd C chain sequence was derived from Sellar et the mouse B chain from Wood et al (1988) and Petry et al al.(1991), (19891, and the mouse A chain from Petry et al (1991).

Fig. 6. Comparison of the amino acid sequence derived from purified rat Clq with the published human and mouse sequences. Amino-terminal sequence of the rat A and C chain (a), and rat B chain derrived following V8 protease digestion (b).

generally regarded as being stable, we have observed its breakdown to monomers also. The phenomenon of dimer instability is therefore not species related, although susceptibility to breakdown may be. A physiological role for the monomers is at present unclear. However there are reports of in uivo degradation of rat Clq to low mol. wt forms following interaction with aggregated IgG (Veerhuis et al., 1985). The possibility that the monomers may bind preferentially to the Clq receptors on cells of the macrophage series, resulting in their removal from the circulation, appears to be ruled out by preliminary studies which demonstrated that macrophages bind the monomers and intact molecules in the same proportions as seen in the offered material (data not shown). Variation in the relative electrophoretic mobility of the three chains of Clq has been reported for mice, where it has been shown that the A chain migrates in the C chain position under reduced conditions (Petry et al., 1991). In this regard rat Clq is like more human Clq, with the C chain being the fastest dimer and monomer. An additional dimer band of intermediate mobility was identified which gave both A and C chain sequence. This is reminiscent of the findings of Stemmer and Loos (1984), who also observed an additional dimer in guinea pig and mouse Clq, suggesting this might represent different stages of glycosylation of the A:B dimer. Human Clq contains one N-linked glycosylation site on the A chain (Reid et al., 1982), the remainder of the 8% (w/w) carbohydrate consisting of glucosylgalactosyl disaccharides or galactose groups associated with certain hydroxylysine residues (Shinkai and Yonemasu, 1979; Reid, 1979).

Variable glycosylation of such sites on rat Clq may provide an explanation for the heterogeneity in mobility of the dimers on SDS-PAGE. Alternatively, heterogeneity may arise as a consequence of variable lysine and proline hydroxylation. As well as forming a homodimer, there was a suggestion from sequence data obtained from electroblotted

Table 1. Amino

Amino

acid

Alanine Cysteine Aspartic acid Glutamic acid Phenylalanine Glycine Histidine Isoleucine Lysine Leucine Methionine Proline Arginine Serine Threonine Valine Tryptophan Tyrosine

Acid composition rat Clq.

of human

Human 4.3” 1.9 8.4 8.8 4.3 17.6 1.6 4.3 5.6 5.9 2.0 10.2 4.3 5.4 5.3 6.3 0.5 3.4

4.2’ N.D. 7.6 8.8 4.3 17.4 1.2 4.3 3.4 5.9 5.6 10.9 4.3 4.8 5.2 6.3 N.D. 3.2

and

Rat 5.2’ N.D. 9.9 9.6 3.7 18.5 0.9 2.9 3.4 5.3 3.3 11.5 3.2 7.6 5.1 5.1 N.D. 1.8

The results shown are derived from the “published cDNA sequence (Petry et al., 1989) or ’ experimentally and expressed in moles percent. N.D.: not done.

Characterization Reduced

Non-Reduced Rat AB+CC AB+CC c,c

Hu -

AB

-

CC

Rat

HU

--

-$

= -

f!L

439

and isolation of rat Cl and Clq

-C

Fig. 7. A schematic representation of rat Clq as visualized on SDS-PAGE. The identity of the rat Clq chains seen following SDS-PAGE is indicated based on the most likely interpretation of the sequencing data.

homogeneity of its chains. The novel Cl and Clq isolation procedure described here allows rapid protein isolation from a number of species. This will enable a broader comparative analysis, which will allow the universal properties of Cl and Clq to be more fully defined. Acknowledgements-The authors would like to acknowledge the invaluable intellectual contribution made by Dr M. J. Hobart during this study. This work was funded by the Wellcome Trust and The Medical Research Council.

REFERENCES bands, that the C chain may form a heterodimer with either the A or B chain or both in the rat. As seen in Fig.

3, rat Cl q dimers, unlike human, are closely spaced, and resolve into three distinct bands. The dimer with the highest mobility gave C chain sequences only, and was accordingly proposed to be the C:C dimer. The middle band gave predominantly C chain sequence with some A sequence, and the band with the lowest mobility gave both A and C chain sequences with similar initial yields. One interpretation of this data is that A:C, and possibly B:C, heterodimers are formed. If this were the case, one would predict an excess of C chains over the other chains in intact Clq. In an attempt to resolve this issue, the A:C sequence ratios in a liquid sample of rat and human Clq were determined. These revealed an equal signal from both chains, rather than the higher initial yield of C chain sequence, which would be predicted if the C chain was forming both homo- and heterodimers (data not shown). Contamination of the A:B dimers with C:C dimers trailing during migration through the gel would also provide a possible explanation for the observed sequence data, though this seems unlikely since there was no contamination of the A chain with C chain sequence when the reduced sample was analysed. A more probable explanation for the apparent discrepancy in dimer sequence data is that A and C chain heterogeneity gives rise to multiple A:B and C:C dimers on SDS-PAGE, and that the C:C dimers are transferred in higher yield than the A-containing dimers to the PVDF membrane. Co-migration of these with the A:B dimer would result in a single band giving both A and C chain sequence, the B chain being blocked. This is schematically drawn on Fig 7, which takes into account the most likely explanation of the sequencing data. Comparison of the limited rat Clq sequence with those of man and mouse show it to be highly conserved, 78-83% with human and 89% with mouse. This is not unexpected as most of the sequence data is derived from the ordered (Gly-X-Y) collagen-like region, which has severe constraints on acceptable mutations. The dimer heterogeneity noted in this study of rat Clq is not unique to the rat, as they have been reported in other species (Stemmer and Loos, 1984). Human Clq may therefore be unusual in the apparent chemical

Arlaud G. J., Sim R. B., Duplaa A.-M. and Colomb M. G. (1979) Differential elution of Clq, Clr and Cls from human Cl bound to immune aggregates. Use in the rapid purification of Cl subcomponents. Molec. Immun. 16, 445-450. Bing D. H. (1971) Purification of the first component of human complement by affinity chromatography on human y globulin linked to Sepharose. J. Immun. 107, 1243-1249. Cleveland D. W., Fischer S. G., Kirschner M. W. and Laemmli U. K. (1977) Peptide mapping by limited proteolysis in sodium dodecyl sulphate and analysis by gel electrophoresis. J. biol. Chem. 252, 1102-l 106. Duncan A. R. and Winter G. (1988) The binding site for Clq on IgG. Nature 322, 738-740. Fraker P. J. and Speck J. C. (1978) Protein and cell membrane iodinations with a sparingly soluble chloroamide, 1,3,4,6tetrachloro-3a,6a-diphenylglycoluril. Biochem. biophys. Res. Comm. 80, 849-857.

Gigli I., Porter R. R. and Sim R. B. (1976) The unactivated form of the first component of human complement, Cl. Biochem. J. 157, 541-548.

Heinz H. P. (1989) Biological functions of Clq expressed by conformational changes. Behring Inst. Mitt. 84, 20-31.

Hitschold T., Golan M. D., Rabs U. and Loos M. (1983) Purification and physiochemical properties of Clq from guinea pig serum. Molec. Immun. 20, 213-221. Lachmann P. J. and Hughs-Jones N. C. (1984) Initiation of complement activation. Springer Semin. Immunopathol. 7, 143-162.

Laemmli U. K. (1970) Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227, 68&685.

Medicus R. G. and Chapuis R. M. (1980) The first component of complement. 1. Purification and properties of native Cl. J. Immun.

125, 39c-395.

Miiller-Eberhard H. J. and Kunkel H. G. (1961) Isolation of a thermolabile serum protein which precipitates y-globulin aggregates and participates in immune hemolysis. Proc. Sot. exp. biol. Med. 106, 291-295.

Petry F., Reid K. B. M. and Loos M. (1989) Molecular cloning and characterisation of the complementary DNA coding for the B-chain of murine Clq. FEBS Lett. 258, 89-93.

Petry F., Reid K. B. M. and Loos M. (1991) Gene expression of the A- and B-chain of mouse Clq in different tissues and the characterisation of the recombinant A-chain. J. Immun. 147, 3988-3993.

440

M. G. WING et al.

Reid K. B. M. (1979) Complete amino acid sequences of the three collagen-like regions present in the subcomponent Clq of the first component of human complement. Biochem. J. 179, 367-37 1. Reid K. B. M., Gagnon J., Frampton J. (1982) Completion of the amino acid sequences of the A and B chains of subcomponent Clq of the first component of human complement. Biochem. J. 2@3, 559-569.

Sellar G. C., Blake D. J. and Reid K. B. M. (1991) Characterisation and organisation of the genes encoding the A-, Band C-chains of human complement subcomponent Clq. Biochem. J. 274, 481490.

Shinkai H. and Yonemasu K. (1979) Hydroxylysine-linked glycosides of human complement subcomponent C Iq and of various collagens. Biochem. J. 177, 847-852.

Steinbuch N. and Audran R. (1969) The isolation of IgG from mammalian sera with the aid of caprylic acid. Archs. biothem. Biophys. 134, 279-284.

Stemmer F. and Loos M. (1984) Purification and characterisation of human, guinea pig and mouse Clq by Fast Protein Liquid Chromatography (FPLC). J. Immun. Meth. 74, 9---16.

Veerhuis R., Van Es L. A. and Daha M. R. (1985) In ~tiun degradation of rat Clq induced by intravenous injection of soluble IgG aggregates. Immunology 54, 801-810. Wood L., Pulaski S. and Vogeli G. (1988) cDNA clones coding for the complete murine B chain of complement Clq: nucleotide and derived amino acid sequence. Immun. Lett. 17, 59-62.