Mannose-binding lectin-associated serine proteases (MASPs) are involved in complement activation through the lectin pathway. To elucidate the phylogenetic origin of MASP and a primordial complement system, we cloned two MASP cDNAs from amphioxus (Branchiostoma belcheri) of the cephalochordates, considered to be the closest relative of vertebrates. The two sequences, orthologues of mammalian MASP-1 and MASP-3, were produced by alternative processing of RNA from a single gene consisting of a common H chain-encoding region and two L chain-encoding regions, a structure which is similar to that of the human MASP1/3 gene. We also isolated two MASP genes from the ascidian Halocynthia roretzi (urochordates) and found that each of them consists simply of an H chain-encoding region and a single L chain-encoding region. The difference in structure between the ascidian MASP genes and the amphioxus/mammalian MASP genes suggests that a prototype gene was converted to the MASP1/3-type gene possessing two L chain-encoding regions at an early stage of evolution before the divergence of amphioxus. This conclusion is supported by the presence of MASP-1 and MASP-3 homologues in almost all vertebrates, as demonstrated by the cloning of novel cDNA sequences representing lamprey (cyclostomes) MASP-1 and Xenopus MASP-3. The ancient origin of MASP-1 and MASP-3 suggests that they have crucial functions common to all species which emerged after cephalochordates.
Mannose-binding lectin (MBL)4-associated serine proteases (MASPs) play crucial roles in activation of the lectin pathway of the complement system (1, 2, 3, 4, 5). The lectin pathway is thought to be pivotal in innate immunity during the lag period before the onset of adaptive immunity and to play a central role in host defense in lower animals lacking adaptive immunity, such as jawless fish (cyclostomes) and ascidians (urochordates) (6). In humans, lectins such as MBL (7) and ficolins (8, 9, 10) bind to carbohydrate moieties and the proenzyme forms of MASPs, which form complexes with these lectins, are converted to their active forms, resulting in the proteolytic activation of the early complement components, C4, C2, and C3 (11, 12, 13, 14). To date, five members of the MASP/C1r/C1s family have been identified in humans, MASP-1 (1, 2, 3), MASP-2 (5), the recently identified MASP-3 (15), and two proteolytic components of the classical pathway, C1r and C1s (16, 17, 18, 19). All of them consist of six domains, two C1r/C1s/Uegf/bone morphogenetic protein 1 (CUB) domains, an epidermal growth factor (EGF)-like domain, two complement control protein (CCP) domains, and a C-terminal serine protease domain. Proteolytic cleavage of the proenzyme forms of MASPs between the second CCP and the C-terminal protease domains creates the active forms comprised of two polypeptides, the H and L chains bridged by a disulfide bond. Human MASP-1 and MASP-3 are produced from a single gene (MASP1/3 gene) by alternative polyadenylation and splicing of the transcripts between the common H chain-encoding region and two l chain-encoding regions (15), which is unique among members of the serine protease superfamily in possessing the double l chain-encoding regions.
There are several important unanswered questions, including the precise stoichiometry and function of lectin-MASP complexes and the phylogenetic origins of MASPs. MASP proteins and/or cDNAs have been isolated from various species of vertebrates (20) and from a species of invertebrates, the ascidian Halocynthia roretzi (urochordates) (21). Based on the primary structures and exon organization of the genes, we identified two lineages in this family (20). The TCN-type, including MASP-1 and ascidian MASPs, has a histidine-loop disulfide bridge (22) in the protease domain, a TCN codon at the active site serine, and split exons for the protease domain. The AGY type, including MASP-2, MASP-3, C1r/C1s and lower vertebrate MASPs, is characterized by the absence of the histidine-loop disulfide bridge, an AGY codon at the active site serine and a single exon for the protease domain. However, more information is needed to establish the phylogeny of the MASP/C1r/C1s family and its relationship to the evolution of the lectin pathway. There is evidence indicating that a lectin-MASP complex, C3 and its receptor may have been the minimal ancestral components of a primitive complement system which developed before the divergence of urochordates (6, 23).
In the present study, we cloned six novel MASP cDNA sequences, two each from amphioxus, lamprey, and Xenopus, which are located at critical positions of the chordate phylogenetic tree. We report on the origin of the MASP1/3 gene and present a plausible mechanism for the evolution of the MASP family of genes. We also discuss the concurrent emergence of constituents of the lectin pathway during evolution.
Materials and Methods
The solitary ascidian H. roretzi and amphioxus Branchiostoma belcheri were harvested in Mutsu Bay and Kitakyushyu in Japan, respectively. Lamprey Lampetra japonica and Xenopus laevis were obtained from local dealers in Fukushima and Aomori in Japan, respectively.
Cloning amphioxus, lamprey, and Xenopus MASP cDNAs
RNA was isolated from lamprey and Xenopus liver and the poly(A)+ fraction was purified by passage through an oligo(dT)-cellulose column (Clontech Laboratories, Palo Alto, CA). The cDNA libraries were constructed in λZAPII according to the manufacturer’s instruction (Stratagene, La Jolla, CA). The amphioxus notocord cDNA library was a gift from Prof. N. Satoh (Kyoto University, Kyoto, Japan). The cDNAs of the MASP/C1r/C1s family in Xenopus, lamprey, and amphioxus were amplified by PCR using the cDNAs as templates and degenerated primers which were designed based on the conserved sequences of the MASP/C1r/C1s family: xxCxYD (MAHF1, 5′-TAISDMTGYKWDTAYGA-3′), CGxxxP (MAHR1, 5′-GGBVIYKKHTYHCCACA-3′), CxPxCG (MAHLF, 5′-TGYIWDCCWRHITGYGG-3′), TAAHxx (MALR2, 5′-ACVIMRTGDGCIGCYGT-3′), MxCAGx (MALR1, 5′-TMBCCRGCRCARAWCAT-3′), and YR2 (see Ref. 20). PCR was nested using MAHF1 and MAHR1 for the first and MAHF1 and YR2 for the second, or using MAHLF and MALR1 for the first and MAHLF and MALR2 for the second, and the products were cloned into pGEM-T easy vector (Clontech Laboratories) and sequenced. To complete the cDNA sequences, cDNAs from the respective libraries were cloned and 5′- and 3′-rapid amplification of cDNA ends was performed using a kit (Marathon; Clontech Laboratories).
PCR-based cloning of amphioxus genomic MASP DNA
To amplify five regions of the amphioxus MASP gene, PCRs were performed with the genomic DNA as a template. The 5.4, 0.5, 0.7, 0.8, and 2.1 kb PCR products for the region encoding the entire common H chain, the intronic region between the H chain and the L chain of amphioxus MASP-3, the region encoding the entire L chain of MASP-3, the intronic region between the L chains of MASP-1 and MASP-3, and the region encoding the entire L chain of MASP-1, respectively, which overlap and cover the entire amphioxus MASP1/3 gene, were cloned in pGEM-T easy and sequenced.
Cloning ascidian MASPa and b genomic DNA
Genomic DNA was prepared from the hepatopancreas of the ascidian H. roretzi, and the library was constructed in λDASH as described (Stratagene). The genomic DNA clones of the ascidian MASPa and MASPb genes were obtained by screening the library with ascidian MASPa and MASPb cDNAs (21) as probes. Two clones, As096 and As171, were selected for nucleotide sequencing. The uncloned region between the two clones (0.9 kb) was amplified by PCR with ascidian genomic DNA as a template.
Cloning Xenopus MASP1/3a and lamprey MASP-A genes
To determine the partial structure of the Xenopus MASP1/3a gene, which encodes MASP-1a (identical to the previous MASP; Ref. 20) and MASP-3a (one of two MASP-3 identified in the present study), PCR was performed using cDNAs corresponding to the H chain of MASP-1a and the L chain of MASP-3a as primers and Xenopus genomic DNA as a template. The 3.8-kb PCR product was cloned in pGEM-T easy and sequenced. Another PCR was performed using cDNAs corresponding to the L chain of MASP-3a and the L chain of MASP-1a as primers. The PCR product of ∼14 kb was cloned in pGEM-T easy and sequenced.
Clones of the lamprey MASP-A (identical to the previously reported lamprey MASP; Ref. 20) gene were obtained by screening the library constructed in λFIXII. One of two positive clones was selected for nucleotide sequencing.
Nucleotide sequence analysis
DNA sequences were determined by the dideoxy chain termination method using a DNA sequencer (Model 4000; LI-COR, Lincoln, NE). The labeling reaction was conducted using the SequiTherm Excell II sequencing kit (Epicentre Technologies, Madison, WI).
Construction of phylogenetic tree
Twenty-three members of the MASP/C1r/C1s family including six new sequences identified in the present study were aligned at the amino acid sequence level using Clustal W software (24). A pairwise distance matrix was obtained by calculating the proportion of the different amino acids. The matrix was then used to construct trees by the neighbor-joining method (25). Bootstrap analysis was used to assess the reliability of branching patterns.
Primary structures of amphioxus MASP-1 and MASP-3 deduced from their cDNAs
The deduced amino acid sequences from the two amphioxus MASP cDNAs revealed that they have identical H chains and distinct L chains. One of the cDNAs is 2695-bp long coding for a 680 amino acid protein, termed amphioxus MASP-1 (Fig. 1⇓). The L chain of this sequence is 41.8% identical to that of human MASP-1, slightly higher than its identities to the L chains of human MASP-2, MASP-3, C1r and C1s (<37.1%) (Table I⇓). Interestingly, the active site serine of amphioxus MASP-1 is encoded by an AGY (AGC) codon and the two cysteines which form the histidine-loop disulfide bridge are present in the protease domain. Because all MASP/C1r/C1s sequences analyzed thus far have either an AGY codon without the histidine-loop or a TCN codon with the histidine-loop, this is a novel combination.
The other amphioxus cDNA is 2782-bp long coding for a 687 amino acid protein, termed amphioxus MASP-3. The amino acid sequence of amphioxus MASP-3 showed no special relationship to the five members of the human MASP/C1r/C1s family (Table I⇑). Like its MASP-1 counterpart, the active site serine of amphioxus MASP-3 is encoded by an AGC codon and the histidine-loop disulfide bridge is present. Both MASP-1 and MASP-3 consist of putative signal peptides followed by the proenzyme forms of MASP, which consist of the conserved six domains. In both amphioxus MASP-1 and MASP-3, the triad of catalytic residues (His, Asp, and Ser) and the arginine and isoleucine which form the proteolytic cleavage site essential for generating the active form of MASP, are also conserved. The two MASPs have an aspartic acid residue at a position −6 relative to the active site serine, suggesting trypsin-type substrate specificities (26).
Primary structures of novel MASPs from lamprey and Xenopus
We cloned two novel cDNA sequences from lamprey liver. One is 3254-bp long and codes for a 663 amino acid protein, termed lamprey MASP-1 (Fig. 1⇑). Lamprey MASP-1 has 45.0% amino acid sequence identity to human MASP-1, which is much higher than to other members of this family in humans (Table I⇑). Like the other vertebrate TCN-type MASPs, the active site serine is encoded by a TCN (TCG) codon and the histidine-loop disulfide bridge is present. These structural features suggest that this sequence is the orthologue of mammalian MASP-1. Lamprey MASP-1 is the first TCN-type MASP identified in lower vertebrates other than tetrapods. The other cDNA is 2991-bp long and codes for a 704 amino acid protein, termed lamprey MASP-B (Fig. 1⇑). The L chain is <37% identical to those of five members of the human MASP/C1r/C1s family (Table I⇑). The L chain of lamprey MASP-B is 55.7% identical to that of the previously identified lamprey MASP (20) which we renamed MASP-A in the present study. Like lamprey MASP-A, MASP-B is a typical AGY-type MASP. Northern blots with lamprey MASP-1 H chain and L chain cDNAs as probes showed very similar patterns which consisted of 6.8- and 3.6-kb bands (data not shown).
By PCR-based cloning, we identified two cDNA clones from X. laevis, termed MASP-3a and MASP-3b. Xenopus MASP-3a and MASP-3b cDNAs are 4457-bp and 4696-bp long, respectively, and each codes for a 717 amino acid protein (Fig. 1⇑). MASP-3a and MASP-3b share 91.8% amino acid sequence identity and have 68.4 and 67.3% identity, respectively, to human MASP-3. The H chain of Xenopus MASP-3a is identical to that of Xenopus MASP-1 (20), which we renamed MASP-1a in the present study. Northern blots showed that Xenopus liver expresses a 4.8-kb transcript for MASP-3a and a 3.6-kb transcript for MASP-1a (data not shown).
Consensus sequences of MASP-1 and MASP-3
We found a few differences in the consensus sequences of the L chains of MASP-1 and MASP-3 (Fig. 1⇑B). An important difference was observed at a position −5 relative to the active site serine, being alanine in MASP-1 and serine or threonine in MASP-3. Recently, the residue at this position is reported to be crucial for substrate diversity in the evolution of serine proteases (27). The aspartic acid residue at a position −6 indicates that both MASP-1 and MASP-3 have trypsin-type specificities (26). These suggest that MASP-1 and MASP-3 cleave different substrates through their trypsin-type specificities. The substrate of MASP-3 is still unknown.
Phylogenetic trees of the MASP/C1r/C1s family
A phylogenetic tree constructed based on the alignment of the amino acid sequences of the H chains of MASPs shows that amphioxus MASP-1/3 forms a branch with ascidian MASPa and MASPb (Fig. 2⇓A), suggesting its ancient origin. Interestingly, in another tree constructed using the sequences of the L chains (protease domains), amphioxus MASP-1 formed a tight branch with the vertebrate MASP-1 group, with a high bootstrap percentage of 92.1% (Fig. 2⇓B). This suggests that amphioxus MASP-1 is the orthologue of vertebrate MASP-1, although the codon for the active site serine (AGY) is different from that of vertebrate MASP-1 (TCN).
The trees clearly indicate that lamprey MASP-1 is the orthologue of mammalian/amphibian MASP-1. In addition, lamprey MASP-A and MASP-B, Xenopus MASP-3a and MASP-3b, the previously cloned carp MASP-A (identical to our previous carp MASP) and shark MASP occupy positions homologous to respective MASP-3, suggesting that they are all orthologous or closely related to mammalian/amphibian MASP-3.
Structure of the amphioxus MASP gene
The amphioxus MASP gene spans ∼9 kb, which includes a region encoding an H chain and two regions encoding L chains, one for MASP-1 and the other for MASP-3 (Fig. 3⇓). This structure clearly suggests that the two amphioxus MASPs are generated from a single gene. The signal peptide and H chain are encoded by eight exons, the MASP-3 L chain by a single exon and the MASP-1 L chain by five split exons. The overall structure of the amphioxus MASP gene is similar to that of the human MASP1/3 gene (15), suggesting that amphioxus MASP-3 may be the orthologue of mammalian/amphibian MASP-3, although the amphioxus MASP-3 L chain has no specific identity to human MASP-3 L chain as described above. The view that amphioxus MASP-1 is the orthologue of mammalian/amphibian MASP-1 is also supported by the fact that four of six positions of introns inserted into the protease domain are conserved between the amphioxus and human MASP-1 genes (data not shown).
Structures of the ascidian, lamprey, and Xenopus MASP genes
The ascidian MASPa and MASPb sequences are the most divergent among the members of this family. Each gene consists simply of a region encoding an H chain and a region encoding an L chain (Fig. 3⇑). The MASPa gene spans ∼7 kb and contains 17 exons. The protease domain of the MASPa gene is encoded by seven exons. In addition, the MASPb gene is located adjacent to the MASPa gene in a tail-to-tail arrangement at a distance of 3.5 kb. The MASPb gene spans ∼7 kb and contains 15 exons. The exon-intron organization of the MASPb H chain is very similar to that of MASPa. In contrast, the exon-intron organization of the MASPb L chain is considerably different from that of MASPa, consisting of five exons. The positions of introns inserted into the protease domain are conserved at 2 of 10 individual positions in MASPa and MASPb, and only at one position in the two ascidian MASPs and human MASP-1.
Lamprey MASP-A, one of the most divergent MASP-3-like sequences in the vertebrate MASP/C1r/C1s family, spans ∼13 kb and contains 11 exons (Fig. 3⇑). The entire organization of exons is the same as that of human MASP2 and C1s genes (20, 28). In addition, we found that the L chains of lamprey MASP-1 and MASP-B are encoded by a split and single exon, respectively, as summarized in Table II⇓. It was found that a single exon encodes the Xenopus MASP-3a L chain, and that it is located between the most downstream exon encoding the common MASP-1/3a H chain and the most upstream exon encoding the MASP-1a L chain (Fig. 3⇑). This organization is similar to that of the human MASP1/3 gene (15), suggesting that the Xenopus MASP1/3a gene produces both MASP-1a and MASP-3a by alternative RNA processing. Another PCR-based analysis showed that the Xenopus MASP-3b L chain is also encoded by a single exon (Table II⇓).
Thus, a clear difference was observed in the structures of the ascidian and the amphioxus/vertebrate genes, there being a single TCN-type L chain-encoding region in the former and two L chain (AGY-type and TCN-type)-encoding regions or a single AGY-type L chain-encoding region in the latter.
All of the vertebrate and amphioxus MASP/C1r/C1s genes have an intronless exon encoding an AGY-type L chain, which is located downstream of the H chain-encoding region (Fig. 3⇑). In contrast, two ascidian MASP genes have no such exon in their corresponding regions. This difference suggests that a processed intronless region might have been inserted between the regions encoding the H chain and TCN-type L chain of a prototype gene. The insertion of the intronless region occurred in the vertebrate lineage after the divergence of urochordates, but before the divergence of cephalochordates. From the structures of two amphioxus MASPs, it can be speculated that an ancestral sequence inserted had the structural features of a TCN-type L chain, and that the change from the TCN-type to the AGY-type occurred in at least two steps, the base change from TCN to AGY and the loss of an intron in the early step and the loss of a histidine-loop disulfide bridge in the late step. The AGY codon for the active site serine of amphioxus MASP-1 may have been generated exceptionally in the amphioxus lineage after the branching of cephalochordates. Although the codon usage of the active site serine in amphioxus MASP-1 is different from those in vertebrate MASP-1, amphioxus MASP-1 was grouped into the TCN-type in this report (Table II⇑), owing to common structural features such as the presence of a histidine-loop disulfide bridge and split exons for protease domain. Similarly, amphioxus MASP-3 was grouped into the AGY-type although it is different from vertebrate AGY-type MASPs in having a histidine-loop disulfide bridge.
It is likely that all of the amphioxus/vertebrate MASP/C1r/C1s genes evolved from an ancestral MASP1/3-type gene by gene duplication, because all of the vertebrate/amphioxus MASP/C1r/C1s genes have an intronless exon encoding an AGY-type L chain in common. It is known that the human MASP2 gene lacks the TCN-type L chain-encoding region, which is replaced by an unrelated gene (28). Thus, the absence of the TCN-type L chain-encoding region in some genes, such as the human MASP2 and the lamprey MASP-A genes, might be explained by the loss of a TCN-type-encoding region during evolution. At present, it is unclear whether the loss of the TCN-type L chain-encoding region happened only once, or more than once, during evolution.
The present study established that, unlike the MASP2 and C1r/C1s genes, the MASP1/3 gene has an ancient origin which can be traced back at least to the amphioxus (cephalochordate) lineage. The origin of MASP-1 may be traced further back to the ascidian (urochordate) lineage, although the view that two ascidian MASPs are the orthologues of vertebrate/amphioxus MASP-1 is still controversial. The lectin pathway seems to have developed step-by-step into a sophisticated system involving retrotransposition (or partial gene duplication) to generate the MASP1/3 gene, gene duplication to generate the MASP2 gene before the amphibian lineage, and by alternative processing of MASP-2 mRNA to produce the truncated form.
Like mammals, such as humans and mice, Xenopus has three types of MASP: MASP-1, MASP-2, and MASP-3. The lectin pathway in Xenopus seems to be similar to that in mammals, although MBL has not yet been identified in Xenopus. Our preliminary study showed that a homologue of human ficolins, which are other lectins complexed with MASPs in human (32, 33), is present in Xenopus serum. Lamprey also has at least three types of MASP: MASP-1, MASP-A, and MASP-B which are encoded by three distinct genes. A problem arises from the phylogenetic trees as to whether either (or both) lamprey MASP-A or MASP-B is an orthologue of mammalian/amphibian MASP-3, because these two sequences occupy a position corresponding to a putative lamprey MASP-3, and because we failed to clone a mammalian/amphibian-type MASP-3 from lamprey, which should have an H chain common to its MASP-1 counterpart. In addition, Northern blots with two distinct cDNA probes for lamprey MASP-1 H chain and L chain showed very similar patterns, suggesting that the lamprey MASP-3 in question is absent or expressed less in liver. It is possible that unlike in mammals, amphibias, and amphioxus, in lamprey MASP-1 and MASP-3 are produced from distinct genes. A similar question should be asked with respect to MASP-1/3 in shark (cartilaginous fish) and carp (bony fish), because the homologue of MASP-1 should be present in these species. Recently, another carp MASP cDNA, designated carp MASP-B (34), and two carp C1r/s cDNAs (31) were isolated. Carp MASP-A and MASP-B have 70% identity, suggesting, as is the case with Xenopus MASP-3, that they are structurally and probably functionally similar isoforms. These four carp sequences are all of the typical AGY-type (Table II⇑). The cloning of orthologues of mammalian/amphibian MASP-1 is needed to establish a clearer phylogenetic relationship of this family in shark and carp.
The evolution of the MASP/C1r/C1s family seems to be synchronized with the evolution of their substrates, such as C3 and C4. To date, C3 has been isolated from invertebrate deuterostomes such as sea urchin (ethinoderma), an ascidian and amphioxus, as well as from vertebrates (see reviews, Refs. 6 and 23). MASP-1 is known to cleave C3 into active C3b (11, 13, 14). As shown in the present study, the origin of MASP-1 can be traced back to amphioxus and possibly back to ascidians. In contrast, MASP-2 seems to have a recent origin going back only to amphibia. MASP-2 is known to specifically cleave C4 into active C4b (12, 13, 14). To date, C4 has been isolated only from vertebrates, such as bony fish, Xenopus, chicken, and mammals, suggesting that it emerged concurrently with adaptive immunity.
MBL, which forms complexes with MASPs, has been isolated from various vertebrates, such as carp, chicken, and mammals. Previously, we reported an MBL-like lectin in an ascidian, termed GBL, which specifically recognizes glucose and forms complexes with ascidian MASPa and MASPb (35). Ficolins have also been isolated from ascidians (36). Thus, MBL and ficolins have ancient origins comparable to those of MASP-1 and MASP-3.
The present study elucidated another important aspect of the evolution of these genes. The positions and phases of the introns inserted into both ends of all six domains are completely conserved in all the MASP/C1r/C1s genes analyzed (detailed data not shown). This suggests their early origin and conservation throughout evolution. In contrast, the positions and phases of the introns inserted into the internal area of each domain are varied, possibly suggesting their late origin. For example, the exon-intron structures for the H chains among the MASP genes have three differences in the second CUB domain and in two CCP domains. The first CCP domain of ascidian/amphioxus MASPs is coded by a single exon, whereas the homologous domain of vertebrate MASPs is coded by two split exons. The second CCP domain of ascidian and vertebrate MASPs is encoded by two exons, although their positions and phases are different. The homologous domain of amphioxus MASP is encoded by a single exon. It is known that mouse factor B and mouse/human complement C2 each contain three CCP modules, and that each CCP is encoded by a single exon (37, 38). These facts suggest that an ancestral gene for the CCP module might have been a single continuous sequence, and that the introns in the MASP/C1r/C1s family might have been inserted independently, an intron into the second CCP domain of ascidian MASPs at phase 2 and two introns into two CCP domains of vertebrate MASPs at phases 0 and 1.
In conclusion, by cloning six novel MASP cDNAs, we demonstrated that MASP-1 and MASP-3 have ancient origins going back at least to the amphioxus lineage. We speculate that a prototype gene was converted to the MASP1/3-type gene having two L chain-encoding regions at an early stage of evolution before the divergence of amphioxus, and that all members of the MASP/C1r/C1s family in amphioxus/vertebrates evolved from this ancestral gene by gene duplication. The ancient origin of MASP-1 and MASP-3 suggests that they have crucial functions common to all species which emerged after cephalochordates.
We thank Prof. N. Satoh, Department of Zoology, Graduate School of Science, Kyoto University and Prof. P. J. Zhang, Institute of Oceanology, Chinese Academy of Sciences for kindly giving us the amphioxus notochord cDNA library and amphioxus genomic DNA, respectively. We also thank K. Kanno for excellent technical assistance.
↵1 This work was supported in part by grants from the Ministry of Education, Science, Sports, and Technology of Japan.
↵2 The nucleotide sequence data reported in this paper will appear in the DDBJ, EMBL and GenBank nucleotide sequence databases with the following accession numbers: AB078636 and AB078637 for Xenopus MASP-3a and MASP-3b cDNAs, respectively; AB089265 and AB089266 for lamprey MASP-1 and MASP-B cDNAs, respectively; AB089267 and AB089268 for amphioxus MASP-1 and MASP-3 cDNAs, respectively; AB078907-AB078909 for the Xenopus MASP1/3a gene; AB078887-AB078894 for the lamprey MASP-A gene; AB089507 for the amphioxus MASP1/3 gene; AB078885 and AB078886 for ascidian MASPa and MASPb genes, respectively.
↵3 Address correspondence and reprint requests to Dr. Yuichi Endo, Department of Biochemistry, Fukushima Medical University School of Medicine, 1-Hikariga-oka, Fukushima 960-1295, Japan. E-mail address:
↵4 Abbreviations used in this paper: MBL, mannose-binding lectin; MASP, MBL-associated serine protease; EGF, epidermal growth factor; CUB, C1r/C1s/Uegf/bone morphogenetic protein 1; CCP, complement control protein.
- Received December 9, 2002.
- Accepted February 24, 2003.
- Copyright © 2003 by The American Association of Immunologists