The channel catfish genome sequence provides insights into the evolution of scale formation in teleosts

Abstract

Catfish represent 12% of teleost or 6.3% of all vertebrate species, and are of enormous economic value. Here we report a high-quality reference genome sequence of channel catfish (Ictalurus punctatus), the major aquaculture species in the US. The reference genome sequence was validated by genetic mapping of 54,000 SNPs, and annotated with 26,661 predicted protein-coding genes. Through comparative analysis of genomes and transcriptomes of scaled and scaleless fish and scale regeneration experiments, we address the genomic basis for the most striking physical characteristic of catfish, the evolutionary loss of scales and provide evidence that lack of secretory calcium-binding phosphoproteins accounts for the evolutionary loss of scales in catfish. The channel catfish reference genome sequence, along with two additional genome sequences and transcriptomes of scaled catfishes, provide crucial resources for evolutionary and biological studies. This work also demonstrates the power of comparative subtraction of candidate genes for traits of structural significance.

Figure 1. Assessment of assembly accuracy.

(a) Concordance of SNP marker positions on scaffolds of reference sequence with those on genetic linkage map, with chromosome 1 (linkage group 1=139.5 cM) being shown here. Scaffolds that span <3 cM each were omitted from the graph. (Graphs for all 29 chromosomes are shown in Supplementary Fig. 2). (b) Distribution of insert size of BAC clones on the reference genome sequence that was estimated with an average insert size of 161 Kb (ref. 21).

Figure 2. Inference of channel catfish population history.

The central bold line represents inferred population size, and the 100 thin curves surrounding the central line are PSMC estimates generated using 100 sequences randomly resampled from the original sequence. The mutation rate of 2.5e−8, adopted from medaka, was used in time scaling. Blue bars at the top of the figure represent glacial periods and orange bars, interglacial periods.

Figure 3. Characteristics of duplicated genes in the channel catfish genome.

(a) Comparison of gene copy numbers of duplicated genes among catfish (blue), zebrafish (orange) and fugu (green); (b) Interchromosomal duplicated channel catfish genes, likely derived from teleost-specific genome duplication. Numbers are chromosomes and coloured lines represent at least 25 duplicated gene pairs; (c) Comparison of duplicated genes in channel catfish, fugu and zebrafish, highlighting recent lineage-specific duplication in channel catfish as specified by the red arrow. (d) Duplicated gene clusters in zebrafish, fugu and channel catfish as a function of sequence similarities, noting the sharp splitting of duplicated clusters in channel catfish with higher e-values, suggesting a higher proportion of rapidly evolving duplicated gene clusters in the channel catfish genome.

Figure 4. Genomic hallmarks of teleosts.

Shown on the top panel are species whose genome sequences were used for the analysis. On the lower panel, comparisons with cartilaginous fish are shown on the left, and comparisons with tetrapods are shown on the right. The genome hallmarks of teleosts are shown in the middle, highlighted by genes for diverse Ig domain-containing proteins, matrix mineralization and microtubule-associated proteins.

Figure 5. Comparative subtraction analysis of transcriptomes.

(a) Interspecific comparative transcriptome analysis: the channel catfish skin transcriptome and the carp skin transcriptome was comparatively subtracted, leading to the identification of 836 genes expressed only in the carp skin but not in the channel catfish skin (left); similarly, the carp skin transcriptomes before scale regeneration and during scale regeneration were comparatively subtracted, leading to the identification of 1,173 differentially expressed genes during scale regeneration. The pool of the 836 genes and 1,173 genes shared (intersection) a total of 18 genes including 13 known genes and five uncharacterized genes. (b) A list of the 10 upregulated genes identified from the shared intersection of the genes expressed in carp but not in channel catfish skin and those differentially expressed during scale regeneration. (c) A summary of SCPP expression in the carp skin (Carp), catfish skin (Catfish) and during scale regeneration (DEG). Plus (+) indicates expression and minus (−) indicates no expression. Multiple pluses indicate degree of induced expression, and purple colour indicated that they were significantly upregulated. (d) SCPP expression during the course of scale regeneration, with their expression levels expressed as reads per kilobase of exon per million mapped reads (RPKM) on the y axis, genes indicated on the x axis, and time points expressed in bars of different colours.

Figure 6. Status of SCPP genes in channel catfish and various other fishes.

(a) Conserved syntenic regions of catfish and zebrafish containing SCPP genes were identified. In zebrafish, the SCPP genes were located on chromosomes 1, 10 and 5, whereas in channel catfish the SCPP genes were on chromosomes 4, 29 and 16. Note that SCPP7, SCPP6, SCPP9, SCPP8 and GSP37 are missing from the channel catfish genome (underlined). (b) Comparison of SCPP1 genes in channel catfish, pleco, zebrafish and fugu, noting that three exons (2, 5 and 6; light shaded) the channel catfish SCPP1 gene was mutated or deleted. (c) Comparison of SCPP5 genes in channel catfish, pleco, zebrafish and fugu, noting that only a small segment (green box) of SCPP5 homologous sequences was present as remnant, and most of its coding sequences are deleted. (d) Alignment of the sequences in the promoter regions of SCPP5 in fugu, tilapia and medaka, noting the highly conserved sequences in the transcriptional factor binding sites, which are entirely lost in the channel catfish gene (not shown). (e) Alignment of the sequences in the promoter regions of SCPP1 in fugu, tilapia, medaka, zebrafish and pleco, noting the highly conserved ETS2 transcriptional factor binding sites, which are lost in the channel catfish gene (not shown). (f) Status of SCPP genes in various teleost species in relation to their scale status (BDP, bony dermal plates; No, no scales; S, scaled; P, protuberance). Pluses (+) in blue shading indicate presence of SCPP genes, and minuses (−) indicate absence of the SCPP genes; yellow shading highlights specific absence of the SCPP genes in scaleless fishes, and red shading indicates presence of SCPP1 and SCPP5 in scaled catfish common pleco (pleco) and southern striped Raphael (doras).

Figure 7. Generation and analysis of genome and transcriptome sequences from scaled catfishes.

Scales represent the primitive condition in the Otophysi (black branches), with catfishes lacking scales (grey branches). Bony plates have evolved several times within catfishes and include all members of the family (blue branches) or just some members of a family (red branches). It is unknown whether the plated condition within callicthyids, scoloplacids and loricariids is homologous. Phylogeny simplified from Sullivan et al.. Photos by J.W. Armbruster, N.K. Lujan, M.H. Sabaj-Perez, S. Smith, K. Luckenbill, H.H. Ng, Z. Randall and L.M. Page. I. puntat, Ictaluras punctatus; P. armat, Platydoras armatulus; P. parda, Pterygoplichthys pardalis. The genes listed with ‘cross subtraction' are the number of genes that were found and expressed in the genome of scale catfishes, but not present in the genome of scaleless channel catfish.

Figure 8. Phylogenetic analysis and expression of SCPP in zebrafish.

(a) Phylogenetic analysis of SCPP genes. Acidic SCPP genes are indicated by hot pink box, and the hypermineralization ODAM is indicated by the green box, and the remaining are P/Q type SCPP genes. (b) Chromosome location and types of SCPP genes indicated with different colour (hot pink, acidic SCPP genes); black, P/Q type SCPP genes and green, hypermineralization SCPP. (c) Summary of expression patterns of the SCPP genes in scale, bone, tooth and fins. Boxes filled with solid hot pink colour represent high expression of the acidic SCPP in that tissue; boxes filled with solid black colour represent high expression of the P/Q SCPP in that tissue; boxes filled with light pink or grey colour are acidic or P/Q type, respectively, SCPP genes that are expressed at low levels in the tissue; and white boxes represent no expression of the specific SCPP genes in the tissues. Note that SPARCL1 was highly expressed only in bone and tooth. SCPP7 was highly expressed in the skin; SCPP6 was highly expressed in the bone; SCPP5 was highly expressed in bone, tooth and dorsal fin; and SCPP9 and fa93e10 was most highly expressed in the ventral fin and dorsal fin, respectively. ODAM was expressed at intermediate levels in all tested tissues.