Diversity in copy number and structure of a silkworm morphogenetic gene as a result of domestication

Abstract

The carotenoid-binding protein (CBP) of the domesticated silkworm, Bombyx mori, a major determinant of cocoon color, is likely to have been substantially influenced by domestication of this species. We analyzed the structure of the CBP gene in multiple strains of B. mori, in multiple individuals of the wild silkworm, B. mandarina (the putative wild ancestor of B. mori), and in a number of other lepidopterans. We found the CBP gene copy number in genomic DNA to vary widely among B. mori strains, ranging from 1 to 20. The copies of CBP are of several types, based on the presence of a retrotransposon or partial deletion of the coding sequence. In contrast to B. mori, B. mandarina was found to possess a single copy of CBP without the retrotransposon insertion, regardless of habitat. Several other lepidopterans were found to contain sequences homologous to CBP, revealing that this gene is evolutionarily conserved in the lepidopteran lineage. Thus, domestication can generate significant diversity of gene copy number and structure over a relatively short evolutionary time.

Figure 1.—

Presence of CBP in lepidopteran insects, including B. mandarina, the putative wild ancestor of B. mori. (A) Western blot analysis of larval midguts from lepidopterans using an anti-B. mori CBP antibody. Weak bands were observed in the lepidopteran species other than B. mori. (B) Alignment of the partial predicted amino acid sequences of CBP from lepidopterans. (C) A neighbor-joining tree generated using the amino acid sequences of CBP from lepidopterans. Human StAR, one of the closest homologs to CBP in humans (Alpy and Tomasetto 2006; Bhosale et al. 2009), was chosen as the outgroup. B. mori and B. mandarina share an identical position in this tree. Bootstrap values were based on 1000 replicates.

Figure 2.—

Amplification of the copy number of CBP in B. mori strains containing the Y allele. (A) Schematic exon/intron structure of CBP, including its alternatively spliced isoform BmStart1. Authentic CBP is composed of seven exons, namely exons B1–2 and C1–5, while BmStart1 is composed of 12 exons, including exons A1–7 and C1–5. Although this structure was derived from the strain Kinshu X Showa, which contains the +^Y allele (Sakudoh et al. 2005), the sequence of the CATS retrotransposon was intentionally left out of this figure for the purpose of clarity. Furthermore, although this structure was generated by connecting partial nucleotide sequences, Southern blot analysis confirmed the overall structural integrity (Sakudoh et al. 2005). (B) Geographical distribution of B. mandarina individuals used in this study: 1–3, Sakado, Saitama, Japan; 4, Fuchu, Tokyo, Japan; 5 and 6, Huzhou, Zhejiang, China; 7, Hangzhou, Zhejiang, China; 8 and 9, Taichung, Taiwan; 10, Taipei, Taiwan; and 11–13, Daegu, South Korea. (C–E) Copy numbers of each exon or gene in multiple B. mori strains and B. mandarina individuals calculated by quantitative PCR. Error bars represent mean ± standard deviation (SD), n = 5 for strain p50, while n = 3 for all other strains. dib and Cameo2 were used as internal standards. For B. mori, the genotype of the Y gene is shown in parentheses. Two or more individuals of strains N4, w1-pnd, UAS-CBP, Nistari, and 925 were analyzed in this study.

Figure 3.—

Genotyping of CBP in multiple B. mori strains and B. mandarina individuals. (A) Schematic genomic structures of several versions of the CBP gene, identified either in a previous study (Y-a, Y-b, and +^Y-a) (Sakudoh et al. 2007) or from the silkworm genome database generated from B. mori strain p50 (+^Y-b) (International Silkworm Genome Consortium 2008). The annealing sites for the primers used in C and Figure 4 are indicated. (B) Western blot analysis showing that CBP protein was absent from strain p50, indicating that the protein was not produced from the +^Y-b sequence, likely due to the complete absence of exon B2. (C) Genotyping of CBP by genomic PCR. The lower band of ∼200 bp for Y-a in strain UAS-CBP is likely from the UAS-CBP transgene (Sakudoh et al. 2007). The structure of CBP in B. mandarina individual number 13 remains undetermined. (D) Provisional summary of the presence/absence (+/−) of each CBP sequence in B. mori and B. mandarina. (E) Schematic genomic structure of CBP, as assessed by analysis of one clone from a Japanese B. mandarina genomic fosmid library. A single target site for the CATS retrotransposon was observed in the B. mandarina sequence between exons B2 and C1, suggesting that the insertion of CATS had not occurred.

Figure 4.—

Copy numbers of types Y-b and +^Y-a in B. mori strains. (A and B) The copy number of each CBP sequence in diverse B. mori strains was calculated by quantitative PCR (mean, SD; n = 4 for primer-145 and -146 and primer-1 and -40; n = 3 for primer-149 and -152 and primer-151 and -154). The primers used were the same as those used to generate the data in Figure 3C and Figure S2B. Strains for which no bands for Y-b and +^Y-a were detected in Figure 3C were also confirmed by quantitative PCR to have a value of zero. (C) Genotyping of the +^Y-a sequence by genomic PCR in several individuals from strain 925. The 925 individuals in which bands corresponding to +^Y-a were not detected were found to be zero by quantitative PCR.

Figure 5.—

Confirmation of copy number variation and structural differences in Bombyx CBP by Southern blot analysis. Genomic DNA was fully digested with EcoRV. For the exon B1 probe, the lengths of the bands for Y-a, Y-b, and +^Y-a were predicted to be 8200, 5890, and 10,320 bp, respectively, on the basis of available sequences (Sakudoh et al. 2007). A dib probe that hybridized to dib was used as an internal standard.

Figure 6.—

Phylogenetic analysis based on SNPs in B. mori and B. mandarina Y-a and Y-b. (A) Schematic representation of the region of genomic sequence used in our phylogenetic analysis. The double underline is for C and the double wavy underline is for D. The sequences of Y-a and Y-b were determined after amplification by PCR using the primer pairs, primer-1 and primer-18 and primer-1 and primer-146, respectively, except for the Y-a sequence of Japanese B. mandarina from a genomic fosmid library. (B) Representative portions of the alignment of the genomic sequences near the 3′ end of exon B2 (i) and the target site of CATS (ii). Positions in which the sequences were not identical to Japanese B. mandarina Y-a are boxed. The Y-a and Y-b sequences of B. mori were nearly or completely identical among the strains analyzed. The sequences of the CATS target site were not included in this phylogenetic analysis. (C and D) Neighbor-joining trees showing each version of CBP from B. mori and B. mandarina, based on an analysis of SNPs at the third codon positions in exon B2 (C) and in the intron located next to the 3′ end of exon B2 (D). Bootstrap values were based on 1000 replicates.

Figure 7.—

Potential contributions of both Y-a and Y-b to mRNA transcription of CBP. A representative alignment of the CBP genomic DNA and cDNA sequences from strain N4. SNPs in Y-a are boxed.