The silkworm (Bombyx mori) is an important silk-producing domestic insect. The quality and yield of the silk produced by B. mori exceeds that of its ancestor, the wild silkworm B. mandarina. However, to date little is known about the molecular mechanisms underlying domestication-related increases in silk yield. Here, we identified a gene associated with both domestication and silk-related quantitative trait locus (QTL): Abelson tyrosine protein kinase (Abl). Population genomic data for B. mori and B. mandarina identified obvious signatures of artificial selection in the genomic region bearing the Abl gene. There were two fixed nucleotide substitutions (−244 and −1311) in the transcription factor binding motif of the upstream regulatory region of B. mori Abl. Compared to Abl in the wild silkworm B. mandarina, B. mori Abl exhibited significantly greater promoter activity and was upregulated in the silk gland from the last day of the 5^th larval instar to pupal stage (P0). Compared to the wild type, CRISPR/Cas9-generated Abl loss-of-function mutants exhibited a higher sensitivity to diseases, shorter developmental duration, and reductions in economically important silk traits, including cocoon weight, pupal weight, and cocoon layer thickness. Comparison of silk-gland transcriptomes between the wild type and the mutant indicated that genes enriched in ribosome biosynthesis, splicing, RNA transport, and the carbon metabolism were significantly downregulated in the mutants. Weighted gene co-expression network analysis (WGCNA) confirmed that genes related to ribosome biosynthesis were pivotal, driving the significant differentiation in silk yield between B. mori and B. mandarina. Here, we demonstrated that artificial selection acts on the cis-elements of silk-trait QTL gene Abl in a novel case (the domestic silkworm), and that this selection pressure increased Abl promoter activity and expression level. By promoting protein translation and synthesis, artificial selection further enhanced the robustness of silkworm larvae and improved cocoon silk synthesis.

Three duplicate samples were set for RNA-seq of slik glands of the wild type and mutant at the 6^th day of the 5^th instar larval stage (5L6D), wandering stage (W) and prepupae (PP), respectively. Silk glands were dissected, stored in dry ice and then sent to Novogene Company (Tianjin, China) for RNA extraction and RNA-seq. Sequencing libraries were generated using NEBNext® UltraTM RNA Library Prep Kit for Illumina® (NEB, USA) following manufacturer’s instructions and index codes were added to attribute sequences to each sample. Clustering of the index-coded samples was performed on a cBot Cluster Generation System using HiSeq 4000 PE Cluster Kit (Illumia) according to the manufacturer’s instructions. After cluster generation, the library preparations were sequenced on an Illumina Hiseq 4000 platform and 150bp paired-end reads were generated. Sequence raw data were qualified, filtered, built, and mapped to the silkworm reference genome (http://www.silkdb.org/silkdb/) by HISAT2 . HTSeq v0.6.0 was used to count the read numbers mapped to each gene and further normalized by DEseq2 in the R package 1.16.1 (TNLIST, Beijing, China) .Principal component analysis (PCA) was performed using genes expressed in the slikglands (Normalized reads >5) with an online software. The expression value of each gene was calculated and normalized using fragments per kilobasesof exon per million reads mapped (FPKM).