C-type lectin-like domains in Fugu rubripes

Abstract

Background: Members of the C-type lectin domain (CTLD) superfamily are metazoan proteins functionally important in glycoprotein metabolism, mechanisms of multicellular integration and immunity. Three genome-level studies on human, C. elegans and D. melanogaster reported previously demonstrated almost complete divergence among invertebrate and mammalian families of CTLD-containing proteins (CTLDcps).

Results: We have performed an analysis of CTLD family composition in Fugu rubripes using the draft genome sequence. The results show that all but two groups of CTLDcps identified in mammals are also found in fish, and that most of the groups have the same members as in mammals. We failed to detect representatives for CTLD groups V (NK cell receptors) and VII (lithostathine), while the DC-SIGN subgroup of group II is overrepresented in Fugu. Several new CTLD-containing genes, highly conserved between Fugu and human, were discovered using the Fugu genome sequence as a reference, including a CSPG family member and an SCP-domain-containing soluble protein. A distinct group of soluble dual-CTLD proteins has been identified, which may be the first reported CTLDcp group shared by invertebrates and vertebrates. We show that CTLDcp-encoding genes are selectively duplicated in Fugu, in a manner that suggests an ancient large-scale duplication event. We have verified 32 gene structures and predicted 63 new ones, and make our annotations available through a distributed annotation system (DAS) server http://anz.anu.edu.au:8080/Fugu_rubripes/ and their sequences as additional files with this paper.

Conclusions: The vertebrate CTLDcp family was essentially formed early in vertebrate evolution and is completely different from the invertebrate families. Comparison of fish and mammalian genomes revealed three groups of CTLDcps and several new members of the known groups, which are highly conserved between fish and mammals, but were not identified in the study using only mammalian genomes. Despite limitations of the draft sequence, the Fugu rubripes genome is a powerful instrument for gene discovery and vertebrate evolutionary analysis. The composition of the CTLDcp superfamily in fish and mammals suggests that large-scale duplication events played an important role in the evolution of vertebrates.

Figure 1

Fugu genome sequence and annotation. A. Fugu selectin gene cluster annotation in the EnsEMBL database (v.2 annotation is shown, v.3 annotation is almost identical to v.2). Gene models predicted by us based on comparison with human selectins are shown in the grey box. As shown, the CTLD is encoded by the 5' exon in fSELP, fSELL and fSELE; the TM segment is encoded by the 3' exon. EnsEMBL predicted transcripts, GenScan predictions and similarity features are shown on the tracks below. Stable IDs for EnsEMBL transcripts are given. The TMHMM track shows ORFs encoding TransMembrane regions predicted by the TMHMM program (see Methods). B. Fragments of group VI genes found on various scaffolds. CTLD numbers indicate sequential number of CTLD in full-length MManR, while numbers for the CTLD in the partial sequences indicate the MManR CTLD sequence they are most closely homologous to.

Figure 2

CTLDcps with novel domain architectures. Fugu CTLD-containing proteins, which do not fit into the existing CTLDcp classification are shown. Domain abbreviations are explained in the text. Roman numbers near names indicate suggested new group names for the new Fugu sequences, which also have new predicted human homologues. C-terminal CTLDs of DEC205-FUSE that are not present in the v.3 assembly are shown in light pink.

Figure 3

Phylogenetic relationships between fish and human CTLDs. A phylogenetic tree built on a ClustalW alignment of a 95% non-redundant collection of predicted Fugu CTLDs and known human and fish CTLDs. Link domains and group VI CTLDs were excluded from the alignment. The tree was built by the neighbor-joining method with 100 bootstrap trials using the ClustalW program. PhyloDraw was used to draw the radial cladogram shown. Branches containing CTLDs from CTLDcps belonging to the same group are shaded; group numbers are marked. Lower case prefixes in the identifiers indicate taxonomic origin: h – Homo sapiens, f – Fugu rubripes, zbrfs – Danio rerio (zebrafish), g – Gillichthys mirabilis, gldhs – Carassius auratus (goldfish), carp – Cyprinus carpio (common carp), rsmlt – Osmerus mordax (rainbow smelt), slmn – Salmo salar (Atlantic salmon), wfldr – Pseudopleuronectes americanus (winter flounder), ahrng – Clupea harengus (Atlantic herring), servn – Hemitripterus americanus (sea raven), jpeel – Anguilla japonica (Japanese eel), medak – Oryzias latipes (Japanese medaka), c – Paralabidochromis chilotes (cichlid fish).

Figure 4

CTLDcp duplication dates. A. Average number of synonymous substitutions per synonymous site (Ks) for CTLDcp paralogue pairs based on full-sequence (triangle) and CTLD-only (diamond) alignments, measured with four different methods (see Methods). Error bars show one standard deviation in the CTLD-only measurements. All possible pairwise alignments between the MManR fragments and between the three CRTL1 paralogues were analyzed. Only homologous regions were used for MManR fragment alignments. B. A linearized phylogenetic tree built by the neighbor-joining method from Poisson-corrected distances between ClustalW-aligned sequences of CTLDs 3–5 from Fugu, mouse and human MManRs. Sequence of the human PLA2R region containing CTLDs 3–5 was used as an outgroup. T_hm – time of separation between human and mouse [96 Myr; 60], T_fish – time of separation between ray-finned and lobe-finned fishes [430 Myr; 32]. Time of duplication (T_dupl) was calculated using average between molecular clock calibrated with T_hm and with T_fish.

Figure 5

Relationships between fish, mouse and human group V and II CTLDs. Non-redundant set of CTLD sequences from known human and mouse CTLDcps classified as groups II and V, Fugu CTLDcps classified as group II, and putative killer cell receptor from Paralabidochromis chilotes (cKLR) were aligned with ClustalW. A consensus phylogenetic tree was built from 100 bootstrap trials using the protdist (with PAM distance matrix) and neighbor programs from the PHYLIP package. Black triangle shows position of cKLR. Bootstrap values higher than 40 are indicated.