Abstract
Mammalian complement components factor B and C2 act as proteolytic subunits of the C3 convertases in the alternative and the classical activation pathways, respectively, and are believed to have diverged from a common ancestor by gene duplication. However, it is unclear when the B/C2 duplication occurred. Here, we describe two diverged B/C2-like cDNA clones (B/C2-A and B/C2-B) isolated from a bony fish, the common carp (Cyprinus carpio). B/C2-A shares the same domain structure as the factor B and C2 complement components of vertebrates reported so far and shows a close similarity to zebrafish B and medaka fish B/C2. These teleost sequences show almost the same degree of similarity to C2 and B of higher vertebrates. In contrast, B/C2-B has a novel structural feature in that it contains four short consensus repeat modules and does not have a close relative upon phylogenetic analysis. Northern blotting revealed the presence of two transcripts with different sizes for both the B/C2-A and B/C2-B in the hepatopancreas of the carp. Southern blotting suggested the presence of multiple genes for B/C2-A and a single gene for B/C2-B. Although structural features of B/C2-B are slightly more C2-like than B-like, B/C2-B has a crucial amino acid substitution in the serine protease domain, which makes it unlikely that B/C2-B functions as a C3 convertase. A possible phylogenetic relationship between the two carp sequences and mammalian C2 and B is discussed.
The complement system plays important roles as a humoral effector system in innate and acquired host defense. The activation of mammalian complement is triggered by Ag-Ab complexes through the classical pathway or by certain bacterial surfaces through the Ab-independent alternative pathway (1). Recently, a third activation mechanism, termed the lectin pathway, involving mannan-binding lectin (MBL)5 and MBL-associated serine protease, has also been identified in various animals (2, 3). Triggering of these pathways results in the proteolytic activation of the third complement component (C3), the central protein of the system. Activated C3 is involved in the mediation of various immune functions such as activation of the cytolytic cascade of complement, opsonization of foreign materials, and promotion of inflammatory responses. The activation of C3 is catalyzed by proteinase complexes called C3 convertases that are formed in each activation pathway (4).
Mammalian factor B and C2 play crucial roles as the proteolytic subunits of the C3 convertases in the alternative and the classical pathways, respectively. B and C2 have similar modular structures consisting of, from the N terminus, three short consensus repeat (SCR) modules, a von Willebrand factor domain, and a serine protease domain and are encoded by genes with the same exon-intron organization (5). It is therefore thought that the two components evolved from a common ancestor by gene duplication. An unanswered question in the evolution of B and C2 is when the B/C2-gene duplication occurred. Structurally and functionally diverged B and C2 have been well characterized only in mammals (5, 6, 7). In nonmammalian vertebrates, two Xenopus factor B genes clearly assigned as factor B have been described in addition to a partial C2-like sequence (8, 9), whereas only one B/C2-like cDNA sequence equally similar to both mammalian factor B and C2 has been isolated from the lamprey (10), a cyclostome fish, which probably lacks the classical and the lytic pathways but probably has both lectin and alternative pathways (11). Thus lamprey factor B may represents the ancestral form before the B/C2-duplication, and the duplication probably predates the divergence of amphibians. In bony fishes, functional analyses have indicated the presence of classical and alternative activation pathways comparable to those of mammals, as characterized by a divalent cation requirement, Ab dependency, and target cell specificity (12, 13, 14, 15). However, only a single type of B/C2-like sequence, which shows a similar degree of sequence similarity with mammalian factor B and C2, has been found in zebrafish and Japanese medaka fish, despite intensive efforts to find an additional B/C2-like sequences (16, 17), leaving the time of B/C2-duplication unknown. To obtain further molecular information on teleost B/C2, we have examined the presence of additional B/C2-like sequences in teleosts. Here, we describe two distinct B/C2-like cDNA clones (designated B/C2-A and B/C2-B) isolated from a teleost species, the common carp (Cyprinus carpio). B/C2-A was identified as a carp counterpart of the zebrafish and medaka fish B/C2 from the domain structure and the amino acid sequence identity. B/C2-B showed the novel structural feature of having four SCR modules and shared only 26% amino acid sequence identity with B/C2-A, suggesting the presence of two distinct B/C2-lineages in teleosts.
Materials and Methods
Materials
Carp weighing ∼30 g to 1 kg were purchased from local fish farms. Restriction enzymes were purchased from Takara (Shiga, Japan) and Nippon Gene (Toyama, Japan). EcoRI adaptor and pGEM-T vector were obtained from Promega (Madison, WI). λZAPII vector and GigaPack Gold II were purchased from Stratagene (La Jolla, CA). Dye-Primer and Dye-Terminator cycle sequencing kits were from Perkin-Elmer Japan (Chiba, Japan). PCR DIG Probe Synthesis Kit and DIG DNA Detection Kit were purchased from Boehringer Mannheim Japan (Tokyo, Japan). cDNA Synthesis Module and Hybond N and Hybond N+ nylon membranes were obtained of Amersham Japan (Tokyo, Japan). Lumiphos Plus was purchased from Wako Pure Chemical (Osaka, Japan).
RNA preparation and cDNA library construction
Total RNA was isolated from carp hepatopancreas using TriZol reagent (Life Technologies, Tokyo, Japan). Poly(A)+ RNA (5 μg) purified with oligo(dT) spin column (Pharmacia Biotech Japan, Tokyo) were subjected to synthesis of double-stranded cDNA using a cDNA Synthesis Module (Amersham Japan, Tokyo), followed by ligation with EcoRI adaptor and with λZAPII vector. After in vitro packaging with GigaPack Gold II, the library was amplified once before use.
RT-PCR amplification of mRNA segments of carp B/C2
Complementary DNA segments encoding the B/C2-like serine protease domain were amplified from the carp hepatopancreas cDNA library by PCR using the same set of primers as described for lamprey (10), medaka fish (16), and Xenopus B (8), except that the present primers lacked the artificial EcoRI sites at the 5′ ends (see Fig. 1). Thirty cycles of amplification were conducted in a Takara thermocycler TP-3000 (Shiga, Japan) using the following parameters: 95°C for 0.5 min, 55°C for 0.5 min, and 72°C for 1 min.
Probe preparation
Digoxygenin (DIG) labeling was performed using a PCR DIG Probe Synthesis Kit (Boehringer Mannheim Japan) according to the manufacturer’s instructions.
Screening of cDNA library
Plaques (1 × 106) of carp hepatopancreas cDNA library were transferred onto Hybond-N membranes. After baking at 80°C for 2 h, the membranes were prehybridized for 2 h at 65°C in a solution containing 5× SSC (1× SSC is 15 mM sodium citrate and 150 mM sodium chloride), 0.1% sodium N-lauroylsarcosinate, 0.02% SDS, and 1% DIG blocking reagent, followed by hybridization with DIG-labeled DNA probes at 65°C for 16 h. After washing with 0.1× SSC containing 0.1% SDS at 65°C for 30 min, colorimetric detection was performed using X-phosphate and NBT in the DIG DNA detection kit as described in the manufacturer’s instructions. Positive plaques were converted to plasmid using the R408 helper phage (Promega).
DNA sequencing
Nucleotide sequence analysis was performed by the dideoxy chain termination method (18) using an Applied Biosystems 377 sequencer (Foster City, CA). Each sequence was determined from both strands.
Northern blotting analysis
Total RNA was extracted from various tissues of carp using TriZol reagent according to the manufacturer’s instruction. Glyoxylated total RNA (10 μg) was separated on a 1% agarose gel and blotted onto a Hybond-N+ membrane (19), followed by baking at 80°C for 2 h. After prehybridization for 2 h at 50°C in a solution containing 50% formamide, 7% SDS, 0.1% sodium N-lauroylsarcosinate, 50 mM sodium phosphate buffer (pH 7.0), 5× SSC, 2% DIG blocking reagent, and 1 mg/ml yeast tRNA, the membranes were then hybridized at 50°C for 16 h with DIG-labeled DNA probes. The B/C2-A probe corresponded to the carp cDNA segment from the 5′ end to the HindIII site (+674), and the B/C2-B probe was a HincII fragment of B/C2-B-1 (from +1767 to +2285). Membranes were washed twice at 68°C for 30 min in 0.1× SSC containing 0.1% SDS and subjected to chemiluminescent detection with Lumiphos Plus according to the manufacturer’s instruction.
Southern blotting analysis
Genomic DNA was isolated from carp erythrocytes as describe elsewhere (19), and 10 μg of the DNA were digested to completion at 37°C for 5 h with 20 U of BamHI, EcoRI, EcoRV, HindIII, or PstI. The digests were electrophoresed on a 1% agarose gel and transferred to a Hybond-N+ membrane. After baking at 80°C for 2 h, the membrane was hybridized in the same manner as described for Northern blotting, except that the hybridization solution contained salmon sperm DNA (50 μg/ml) instead of tRNA and the hybridization was performed at 42°C. Two B/C2-A probes (A1 and A2) were used; A1 was the same as in Northern blotting, and A2 was a PCR-amplified cDNA segment spanning +1129 to +1526. B/C2-B probe was the same cDNA segments described above.
Results
RT-PCR amplification of carp B/C2
RT-PCR amplification of the carp hepatopancreas cDNA yielded three DNA bands, as shown in FigureF1 1. The doublet bands (230 and 240 bp) were separately gel purified, and each DNA was subcloned into the pGEM-T vector. Three clones containing 230-bp inserts and two clones carrying 240-bp inserts were isolated and sequenced using the T7 primer. Three clones with a 230-bp insert had an identical nucleotide sequence (designated as CP230), and two clones with a 240-bp insert had another sequence (referred to as CP240). The 690-bp band was a byproduct of PCR from the λZAP vector. The deduced amino acid sequences of the CP230 and CP240 showed 30 and 24% identity with the corresponding region of human B, and 32 and 27% with that of human C2, respectively (Fig. 1). Upon BLAST searches of the GenBank nucleotide sequence database, both CP230 and CP240 showed that the highest degree of similarity was with zebrafish B, Xenopus B, mouse B, and mouse C2. We therefore tentatively identified both sequences as B/C2 homologues of the carp.
Isolation and sequence analysis of carp B/C2-A and B/C2-B cDNA
Screening of the carp hepatopancreas cDNA library using the CP230 and CP240 probes detected 376 and 80 positive clones, respectively. Of these, five CP230-positive clones (B/C2-A) and 13 CP240-positive clones (B/C2-B) were isolated and sequenced from both termini of the insert. The complete nucleotide sequences of the longest insert, designated B/C2-A-3 (2.4 kbp) and B/C2-B-1 (3.1 kbp), were determined following the strategy shown in Figure 2.
B/C2-A-3 was 2.4-kbp long and predicted a single long open reading frame of 762 amino acids containing, from N to C terminus, three SCR modules, a von Willebrand factor domain, and a serine protease domain, as has been found in B and C2 of other animals reported to date (Fig. 3 A). Within a 115-bp long 3′-untranslated region, two overlapping sequence stretches of putative polyadenylation signals, AATAAA, were found at 24 bp and 20 bp upstream of the poly(A) tail.
B/C2-B-1 (3.1 kbp long) contained a single long open reading frame encoding an 833-amino acid sequence and a 573-bp long 3′-untranslated region in which a putative polyadenylation signal sequence, AATAAA, was found 33 bp upstream of the poly(A) tail (Fig. 3,B). The deduced amino acid sequence of B/C2-B contained a von Willebrand factor domain and a serine protease domain preceded by, not three, but four SCR modules. Alignment (Fig. 4,A) and comparison of the amino acid identities (Table I) of the SCR modules in both proteins showed that the first two modules of B/C2-B were 41% identical, suggesting that the SCR-1 of the B/C2-B was duplicated after the divergence of B/C2-A and B/C2-B (Fig. 4 B).
. | . | B/C2-A . | . | . | B/C2-B . | . | . | . | |||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
. | . | SCR-1 . | SCR-2 . | SCR-3 . | SCR-1 . | SCR-1′ . | SCR-2 . | SCR-3 . | |||||
B/C2-A | SCR-1 | 100 | 20 | 17 | 29 | 24 | 18 | 18 | |||||
SCR-2 | 100 | 34 | 18 | 14 | 33 | 23 | |||||||
SCR-3 | 100 | 16 | 18 | 32 | 46 | ||||||||
B/C2-B | SCR-1 | 100 | 41 | 16 | 14 | ||||||||
SCR-1′ | 100 | 14 | 14 | ||||||||||
SCR-2 | 100 | 21 | |||||||||||
SCR-3 | 100 |
. | . | B/C2-A . | . | . | B/C2-B . | . | . | . | |||||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
. | . | SCR-1 . | SCR-2 . | SCR-3 . | SCR-1 . | SCR-1′ . | SCR-2 . | SCR-3 . | |||||
B/C2-A | SCR-1 | 100 | 20 | 17 | 29 | 24 | 18 | 18 | |||||
SCR-2 | 100 | 34 | 18 | 14 | 33 | 23 | |||||||
SCR-3 | 100 | 16 | 18 | 32 | 46 | ||||||||
B/C2-B | SCR-1 | 100 | 41 | 16 | 14 | ||||||||
SCR-1′ | 100 | 14 | 14 | ||||||||||
SCR-2 | 100 | 21 | |||||||||||
SCR-3 | 100 |
The percentage of amino acid identity among SCR domains of carp B/C2-A and B/C2-B was calculated based on the alignment shown in Figure 4 A.
Multiple alignment and phylogenetic tree of factor B and C2 sequences from various species
The deduced amino acid sequences of carp B/C2-A and B/C2-B were aligned with zebrafish B (17), medaka fish B/C2 (16), Xenopus B (8), mouse B (5), human B (7), mouse C2 (5), human C2 (6), and lamprey B (10), primarily using the Clustal W software (20), with manual modification within the SCR domains based on the alignment shown in Figure 4,A. Carp B/C2-A and B/C2-B show significant amino acid sequence similarity to other B and C2 sequences throughout their entire length, except that B/C2-B has the additional SCR-1′ domain (Fig. 5). The calculated amino acid identities between carp B/C2-A or B/C2-B and the B and C2 of other animals, based on this alignment, are listed in Table II. B/C2-A shared 61% amino acid sequence identity with B of zebrafish, which belongs to the Cyprinidae family together with carp, and showed almost the same similarity to both mammalian B and C2 as do the other teleost B/C2 sequences. On the other hand, B/C2-B showed only 26% amino acid identity with carp B/C2-A and shared a lesser identity with the other B/C2 than did B/C2-A. Since the percentage of sequence similarities of carp B/C2-B remains lower than the other sequences even when compared after removal of the additional SCR domain (SCR-1′) of B/C2-B, the divergence of carp B/C2-B is attributable not only to the additional SCR but also to the difference throughout the entire region.
. | Carp B/C2-A . | Carp B/C2-B . |
---|---|---|
Carp B/C2-A | 25.6 (28.1) | |
Zebrafish B | 60.8 | 25.8 (29.3) |
Medaka fish B/C2 | 39.1 | 26.1 (30.0) |
Xenopus B | 32.1 | 28.1 (31.0) |
Mouse B | 34.9 | 29.3 (32.3) |
Human B | 36.1 | 29.3 (32.3) |
Mouse C2 | 34.1 | 30.3 (33.3) |
Human C2 | 34.4 | 30.5 (33.6) |
Lamprey B | 27.0 | 24.8 (27.2) |
. | Carp B/C2-A . | Carp B/C2-B . |
---|---|---|
Carp B/C2-A | 25.6 (28.1) | |
Zebrafish B | 60.8 | 25.8 (29.3) |
Medaka fish B/C2 | 39.1 | 26.1 (30.0) |
Xenopus B | 32.1 | 28.1 (31.0) |
Mouse B | 34.9 | 29.3 (32.3) |
Human B | 36.1 | 29.3 (32.3) |
Mouse C2 | 34.1 | 30.3 (33.3) |
Human C2 | 34.4 | 30.5 (33.6) |
Lamprey B | 27.0 | 24.8 (27.2) |
The value shown in parentheses is the percentage of amino acid identity when the additional SCR domain (SCR-1′ in Fig. 4) is omitted for the comparison.
A phylogenetic tree of the B and C2 was drawn using the neighbor-joining method (21) after the SCR-1′ region (Thr95–Glu173) had been removed from the carp B/C2-B sequence. As shown in Figure 6, B/C2-A formed a cluster with zebrafish B and medaka fish B/C2, which was supported by a high bootstrap value. However, the relationship of this cluster with B or C2 of other vertebrates was not clear, as shown by the low bootstrap value. B/C2-B formed a cluster with mammalian C2. This clustering is not conclusive, because the bootstrap value supporting it is low (64%).
Northern blotting analysis
Tissue distribution of the messengers for B/C2-A and B/C2-B was examined by Northern blotting using total RNA from the hepatopancreas, spleen, kidney, head kidney (pronephros), heart, and intestine. The B/C2-A probe detected a major band of 2.6 kb and a minor band of 3.2 kb from the hepatopancreas, whereas B/C2-B probe gave a major band of 3.4 kb and a minor band of 2.7 kb from the hepatopancreas (Fig. 7,A). Other organs did not show any band, indicating that both B/C2-A and B/C2-B are expressed in the hepatopancreas. No cross-hybridization between B/C2-A and B/C2-B was observed, as expected from their low sequence identity. The presence of two transcripts each for both B/C2-A and B/C2-B was further examined using hepatopancreas RNA from eight individual carp. The two transcripts of B/C2-A, seen in Figure 7,A, were detected in six of the eight carp examined (Fig. 7 B). On the other hand, the two bands of the transcripts of B/C2-B were observed in seven of eight carp with varying degrees of relative intensity. It remains to be elucidated whether these two bands represent a difference at the mRNA level, such as alternative splicing and/or alternative transcriptional initiation as reported for human C2 (22) and mouse B and C2 (5), or a difference at the genomic level.
Southern blotting analysis
As shown in Figure 8, A and B, the two B/C2-A probes (A1 and A2) hybridized with multiple bands in each digest with five different enzymes. In addition, all nine representatives showed multiple bands hybridized with the A1 probe upon HindIII digestion (Fig. 8,C). These results indicate that carp have multiple (probably two or three) B/C2-A genes, which indicates considerable polymorphism. The B/C2-B probe detected two bands in all digests (Fig. 8,D). To determine whether the two bands represent allotypes or isotypes, the same HindIII digests that were used for B/C2-A were examined. As shown in Figure 8 E, three of nine carp gave two bands and the other six gave a single band. Thus, it is probable that B/C2-B is encoded by a single-copy gene with less polymorphic alleles.
Discussion
To define when B/C2 duplication occurred, molecular cloning and genetic analysis of B/C2 have been performed on a number of teleost species. In medaka fish, RT-PCR using a primer set that allowed the amplification of both B and C2 sequences in mammals resulted in the generation of a single B/C2-like sequence. The entire medaka fish B/C2 sequence showed almost the same degree of similarity to both mammalian B and C2 (16), making it difficult to assign the medaka fish sequence to B or C2. A B/C2-like cDNA cloned from zebrafish was also equally similar to mammalian B and C2 and was reported to be a single gene product.
Here, we have demonstrated, for the first time in a teleost, the presence of two considerably diverged B/C2-like cDNA sequences (B/C2-A and B/C2-B). Carp B/C2-A is composed of three SCR modules, a von Willebrand domain, and a serine protease domain as has been found in B and C2 from other vertebrates reported so far. B/C2-A showed a 61% amino acid sequence identity with the zebrafish, which belongs to the same family (Cyprinidae) as carp, and 39% identity with the medaka fish, which is in a different order (Beloniformes), sharing a similar degree of identity with both mammalian B and C2 (Table II) as those of the two other teleost sequences. Therefore, carp B/C2-A is probably in the same lineage as the B/C2-like molecules described in medaka fish and zebrafish, which show almost the same degree of similarity to B and C2 of other vertebrates. Whereas medaka fish and zebrafish B/C2 have been described as being encoded by single-copy genes, Southern blotting analysis of carp DNA indicated the presence of multiple B/C2-A genes, which seems to be in agreement with the proposed tetraploidization event in the carp ancestor. Duplicated B genes have also been reported in Xenopus laevis, a pseudotetraploid amphibian, but both of the Xenopus genes are linked to a diploidized MHC, indicating that the Xenopus B gene has been duplicated independently of the tetraploidization (8, 9). Thus, chromosomal localization and linkage analysis are required to illustrate the mechanism that generated multiple carp B/C2-A genes.
On the other hand, an additional type of carp B/C2-like clone, B/C2-B, has a novel structural feature in that it has four SCR modules and is probably encoded by a single gene. B/C2-B also shows almost the same similarity to both mammalian B and C2, but shares only 26% amino acid sequence identity with carp B/C2-A. The high degree of sequence diversity between B/C2-A and B/C2-B suggests an ancientness of the gene duplication between these genes, most probably before speciation of teleost fish.
The presence of two diverged B/C2-like sequences led us to consider the possibility that B/C2-B corresponds to the carp counterpart of mammalian C2. This possibility is partly supported by the following observations. 1) The degree of similarity between the two carp sequences is lower than that between mammalian B and C2, even after removal of the additional SCR domain (SCR-1′) of carp B/C2-B, indicating that B/C2-B belongs to a different lineage than B/C2-A. 2) Carp B/C2-B shows a slightly higher sequence similarity to mouse and human C2 than to B from mouse and human, whereas carp B/C2-A is slightly more factor B-like than C2-like (Table II). 3) Carp B/C2-B forms a cluster with mammalian C2 in the phylogenetic tree drawn by the neighbor-joining method, although the bootstrap value is not high enough for us to be fully confident about the branching (Fig. 6). 4) The number of charged amino acids (Asp, Glu, Lys, and Arg) in the region corresponding to the mammalian exon 15-encoded sequence of B and C2 has been reported to differ greatly between human B and C2, and these residues are suggested to be involved in their functional difference (23). Carp B/C2-A and B/C2-B have 11 and 9 charged residues, respectively, whereas the number is 6 for human C2, 6 for mouse C2, 15 for human B, 12 for mouse B, 13 for Xenopus B, 12 for medaka fish B, and 13 for zebrafish B. 5) In the screening of the cDNA library, B/C2-A clones appeared about five times more frequently than B/C2-B clones. On the assumption that the frequency of cDNA clone represents the approximate level of expression, the relative abundance of B/C2-A compared with B/C2-B apparently coincides with the relationship between serum levels of B (200 μg/ml) and C2 (20 μg/ml) in mammals (4). In addition, a similar relative abundance of B/C2-A mRNA compared with B/C2-B mRNA was also observed in Northern blotting analysis. Despite these supporting data, however, the serine protease domain of B/C2-B contains a crucial amino acid substitution that could abolish its essential function as C2. Namely, Asp674 in B/C2-A is replaced by Val760 in B/C2-B at the S1 specificity crevice of trypsin-like serine proteases (24). Thus, it is unlikely that B/C2-B has a trypsin-like specificity cleaving the C-terminal site of basic amino acid residues. Inasmuch as we have determined that the C3 convertase cleavage site sequence (which corresponds to the C terminus of the C3a fragment) of carp C3 is Leu-Ala-Arg, like that of other vertebrate C3 molecules (our manuscript in preparation), B/C2-B is unlikely to act as a functional counterpart of mammalian C2. Thus, the molecular entity of the carp counterpart of mammalian C2 still remains to be elucidated. Identification of the molecules that are recognized by the SCRs of B/C2-A and B/C2-B, as well as the substrate specificities of their serine protease domains, will be of great importance in elucidating the respective functions of the two carp B/C2-like molecules. In addition, a further search for the presence of the B/C2-B-like genes in other lower vertebrates will give some insight into the evolution of the complement activation pathways.
Acknowledgements
We thank Chisato Yamada, Dr. Yoichi Kato, and Dr. Noriyuki Kuroda for technical assistance. We also thank Dr. A. W. Dodds (Medical Research Council Immunochemistry Unit, Department of Biochemistry, University of Oxford, U.K.) for helpful discussion.
Footnotes
This work was supported in part by Grant-in-Aid 07660271 to T.Y. from the Ministry of Education, Science, and Culture of Japan.
The nucleotide sequence data reported in this paper will appear in the DDBJ, EMBL, and GenBank nucleotide sequence databases with the following accession numbers: AB007004 and AB007005.
Abbreviations used in this paper: MBL, mannan-binding lectin; SCR, short consensus repeat; DIG, digoxygenin.