Artificial Selection on Storage Protein 1 Contributes to Increase of Hatchability during Silkworm Domestication

Like other domesticates, efficient utilization of nitrogen resource is also important for the domestic insect, the silkworm. Deciphering how artificial selection act on silkworm genome for improved utilization of nitrogen resource and further human-favored domestication traits will provide unique cues from the insect scenario for understanding general rules of Darwin’s evolutionary theory on domestication. Storage proteins (SP), which belong to a hemocyanin superfamily, basically serve as a source of amino acids and nitrogen during metamorphosis and reproduction in insects. Here through genomic search and further screening of artificial selection signature on silkworm SPs, we discovered a candidate domestication gene, i.e. the methionine-rich storage protein1 (SP1), which is uniquely diverged from the others and showed increased expression in the ova of domestic silkworms. Knockout of SP1 via CRISPR/Cas9 approach resulted in dramatic decrease in egg hatchability, without obvious impact on egg production, which was similar to the case in the wild silkworm compared with domestic one. Larval development or metamorphosis were not affected by SP1 knockout. Comprehensive ova comparative transcriptomes indicated a general repression of gene expression, specifically vitellogenin, chorion proteins and structural component proteins in the extracellular matrix (ECM)-interaction pathway, as well as enzymes in folate biosynthesis, in both the mutant and the wild silkworm with the mutated allele, compared to the wild type domestic silkworm. Wild silkworms with the wild allele also showed generally down-regulated expression of genes enriched in structural constituent of ribosome and amide and peptide biosynthesis. This study exemplified a novel case that artificial selection could directly act on nitrogen resource protein to affect egg nutrient and eggshell formation, and activate ribosome for improved biosynthesis and increased hatchability during domestication. The findings shed new light on both understanding of artificial selection and silkworm breeding from the angle of nitrogen and amino acid resource. Author summary Like other domesticates, nitrogen resource is also important for the domestic insect, the silkworm. Deciphering how artificial selection act on silkworm genome for improved utilization of nitrogen resource and further human-favored domestication traits, will provide unique cues from insect scenario, for understanding general rules of Darwin’s evolutionary theory. However, mechanism of domestication in the silkworm is largely unknown to date. Here we focused on one important nitrogen resource, i.e, the storage proteins (SP). We discovered that the methionine-rich storage protein1 (SP1) which is divergent from the other SPs are the only target of the artificial selection. We proposed based on functional evidence together with the key findings of comprehensive comparative transcriptome, that artificial selection, on one hand favored higher expression of SP1 in the domestic silkworm, which would subsequently up-regulate the genes or pathways vital for egg development and eggshell formation. On the other hand, artificial selection consistently favored activated ribosome activities and improved amide and peptide biosynthesis and in the ova, as it might act in the silk gland for increased silk-cocoon yield. We here exemplified a novel case that artificial selection could directly act on nitrogen resource protein for human desired domestication trait.


65
The silkworm Bombyx mori is the only fully domestic insect species, which 66 originated from its wild ancestor B. mandarina about 5000 years ago. During this 67 process, the domestic silkworm evolved rapidly under human-preferred selection. 68 Deciphering how artificial selection act on silkworm genome for human-favored 69 domestication traits, will provide unique cues from insect scenario, for understanding affect the metamorphosis and cocoon yield [1]. These findings suggest that, like 75 domestic plants and animals, domestic silkworms also tend to have efficient 76 utilization of nitrogen resources to adapt to human-preference [1][2][3]. Besides the 77 glutamate and aspartate metabolism which is an ammonia re-assimilated system[4], 78 we further wonder whether other kind of nitrogen resources are also affected by 79 artificial selection. If this is the case, how they contributed to silkworm phenotypic 80 changes during domestication. 81 Insect storage proteins (SP) are another important source of amino acids and 82 nitrogen, which belong to a special conserved arthropod hemocyanin superfamily [5]. 83 Most insects have at least two main kinds of storage proteins, i.e, arylphorin and 84 methionine-rich storage proteins and some species have other non-typical SPs [6]. SPs 85 have been cloned or predicted in many insect species, including Lepidoptera moths Running title: SP1 increased hatchability during silkworm domestication 6 / 34 86 and butterflies [7][8][9][10]. Insect SPs are suspected to serves as a source of amino acids 87 and nitrogen the for pupae and adults during metamorphosis and reproduction [11], 88 however solid functional evidence on its biological significance is rather few [9,12]. 89 In plant, storage proteins are mainly reserved in seeds, together with other nutrient 90 such as oil and starch, to supply energy for seed germination, growth [13,14]. 91 Especially in crops, seed SPs function in providing energy for humans and animals 92 and they are of great interest and target for breeding and improvement [13][14][15]. 93 In the domestic insect, the silkworm, previous studies preliminarily characterized 94 three SPs mainly based on cues of gene or protein expression pattern [7, [16][17][18]. 95 Especially the methionine-rich SP1 was implied to contribute to adult female  [19,20]. Here in this study, we conducted a genome-wild identification of silkworm 105 SPs and taking advantages of the genomic data resource of a batch of representative 106 domestic and wild silkworms [1], we conducted selection signature screening of these amino acid supply will be also required for improved silkworm reproduction.

SP1 is diverged from the other SPs and is the only one targeted by artificial
119 selection 120 Totally, we identified 7 SPs in the silkworm genome by blast search. SP1 showed the 121 dramatically highest methionine content (10.98%) ( Table 1). Phylogenetic analysis 122 showed that SP1 was one distinct clade whereas the others were in another one, 123 indicating an obvious divergence between SP1 and the others (Fig 1A). SP1 is located 124 in Chromosome 23 and the others are clustered in Chromosome 3, suggesting possible 125 tandem duplication events during evolution. Interestingly, by screening of artificial 126 selection signatures on the genomic region bearing SP1 and the other SPs 127 respectively, we detected strong signatures in the SP1 region (see material and 128 methods) (Fig 1B and 1C). Furthermore, we detected strong differentiation in allelic 129 frequency in the upstream of SP1 (Fig 1D). Correspondingly, SP1 were differentially 8 / 34 130 expressed in the ova between domestic and wild silkworms, with higher expression in 131 the domestic one ( Fig 1E). We also detected 11 SNPs that caused amino acid changes 132 in the coding sequence of the gene (S1 Fig

Knockout of SP1 by CRISPR/Cas9
137 To explore the functional impact of BmSp1 in silkworm domestication, we firstly 138 investigated the biological role of this gene in the silkworm through CRISPR/Cas9 139 knocking out system. For single guide RNA (sgRNA) design, we selected highly 140 specific targets in the first exon close to the translation starting site, namely, S1 and 141 S2 (Fig 2A). We choose another site S3 close to the end of the first exon, more than 142 60 bp downstream of S1 and S2 (Fig 2A and Table 2), to obtain a potentially large 143 fragment deletion by injecting the pool of three gRNAs. After mutation screening of 144 the injected eggs (G0 generation), the gRNAs targeting the above three sites 145 successfully guided DNA editing and generated a variety of mutation types, including 146 4-9 bp deletions or small insertions followed by a large deletion (Fig 2B).

147
By mutation screening of the exuviae of the fifth instar larva in the G0 cocoon, we 148 successfully identified 26 mosaic mutant G0 moths. We then generated pairwise 149 crosses of these G0 mutants with similar mutant genotypes for the G1 populations.

150
After mutation screening of the G1 eggs, we selected two populations with large 151 deletions for further feeding and mutation screening (see Material and methods).

/ 34
152 Finally, in the G2 generation we obtained two types of homozygous mutants, i.e., 153 MU1 and MU2 ( Fig 2C). As to MU1, there was an 8 bp insertion followed by a 63 bp 154 deletion in the BmSP1 coding sequences. As to MU2, there was a 4 bp insertion 155 followed by a 65 bp deletion. The mutations occurred at +29 bp of the first SP1 exon 156 in MU1 and MU2, respectively (Fig 2C), resulting in reading frame shift mutations 157 and severe premature termination close to the translation starting site, with stop 158 signals at +10 aa and +37 aa of theSP1 protein (Fig 2D).

159
Both SP1 mutants and the wild silkworm had decreased hatching rate and Vg 160 expression 161 We selected and maintained MU1 population for assay on phenotypes related to 162 reproduction and metamorphosis, such as number of eggs, hatching rate, pupa weight, 163 and cocoon weight. Compared with the wild-type, which showed hatching rates of 164 about 90%, the hatching rates of SP1 mutants were dramatically decreased, with a 165 mean egg hatching rate of about 40% (Fig 2E), whereas the number of eggs produced 166 was not obviously affected ( Fig 2E) , neither did the whole pupa weight or cocoon 167 shell weight ( Fig 2E). Given that the data were obtained from large replicates (126 168 replicates for hatchability assay and 320 replicates for pupa and cocoon weight), the 169 results are convincible. Loss-of-function mutation resulted in significant decreased 170 expression of SP1 and Vg in the ova, based on the RNA-seq data ( Fig 2F).

171
Given that knockout of SP1 caused reduced hatching rate ( Fig 2E) and that 172 expression of SP1 in the ova of wild silkworm were significantly lower than that of 173 domestic one, we further suspected that artificial selection on SP1 might improve 10 / 34 174 silkworm hatching rate during domestication. Supporting this hypothesis, we found 175 that hatching rates of the wild silkworm were generally lower compared with that of 176 the domestic one and, similar to SP1 mutant, no obvious differences was detected in 177 egg production between the wild and the domestic silkworm ( Fig 3A). The lower 178 hatching rate of wild silkworm was also reported in other studies [21,22]. Meantime, 179 we also found that similar to SP1 mutants, expression of Vg in the ova of the wild 180 silkworm was drastically lower compared with domestic one (Fig 3B). These results   RNA-seq data for each sample (S1 Table). Totally, there were 561 genes identified as  Table). As expected, there are much more differential 11 / 34 196 expressed genes (2882) between the wild and domestic silkworm, since wild 197 silkworm are much more genetically and phenotypically different from the domestic 198 one, compared with the silkworm mutant. It is interesting that, down-regulated genes 199 (1761) were also significantly more than up-regulated (1121) ones (p=2.2e-16, 200 Chi-squared test with Yates' continuity correction) (Fig 3A and S3 Table). The results 201 suggested that transcriptome repression in ova might be an output of SP1 depletion, in 202 both the SP1-KO mutant (Fig 2F and Fig 4A) and in the wild silkworm ( Fig 1E and 203 Fig 4A). 204 We identified 302 common genes shared in the two sets of DEGs. KEGG 205 enrichment analysis indicated that these common differential expressed genes were genes, it is notable that they were mostly down-regulated, in both the SP1 mutant and 221 the wild silkworm ( Fig 4B). This results suggested that the common influence of 222 repression of SP1 in both the SP1 mutant and the wild silkworm at transcriptome level 223 is on genes or pathways involved in reproduction, such as ovarian follicle cell 224 development or proliferation, and eggshell formation. 225 We further generated enrichment analyses on DEGs in the two sets of comparison 226 independently and observed consistent pattern (Table 3 and (Table 3) Table 4). The involved genes in these KEGG or GO terms were nearly all 234 down-regulated in the wild silkworm (Table 4) Table 4). The related genes were also mostly down-regulated in the wild silkworm.   Methionine-rich SP1 seems to be of special interest, since methionine is reported to be functions in reproduction process but not obviously affect growth.

271
Given that in those cocoon-producing silk moths, another nitrogen utilization 272 system such as gluminate/glutamine cycle system were reported to be vital in   Table 2.