Let-7 microRNA is a critical regulator in controlling the growth and function of silk gland in the silkworm

ABSTRACT

The silk gland is characterized by high protein synthesis. However, the molecular mechanisms controlling silk gland growth and silk protein synthesis remain undetermined. Here we demonstrated that CRISPR/Cas9-based knockdown of let-7 or the whole cluster promoted endoreduplication and enlargement of the silk gland, accompanied by changing silk yield, whereas transgenic overexpression of let-7 led to atrophy and degeneration of the silk gland. Mechanistically, let-7 controls cell growth in the silk gland through coordinating nutrient metabolism processes and energy signalling pathways. Transgenic overexpression of pyruvate carboxylase, a novel target of let-7, resulted in enlargement of the silk glands, which is consistent with the abnormal phenotype of the let-7 knockdown. Overall, our data reveal a previously unknown miRNA-mediated regulation of silk gland growth and physiology and shed light on involvement of let-7 as a critical stabilizer and booster in carbohydrate metabolism, which may have important implications for understanding of the molecular mechanism and physiological function of specialized organs in other species.

Introduction

Silk gland development and silk protein synthesis are complex activities involving a large number of genes [1], but the molecular mechanisms in the regulation of silk gland growth and function are not fully understood. The silk gland is formed in the embryonic stage and differentiates into three divisions: anterior silk gland (ASG), middle silk gland (MSG) and posterior silk gland (PSG); the middle silk gland is further differentiated into anterior (A-MSG), middle (M-MSG) and posterior (P-MSG) parts [2,3] (Fig. 1A). Silkworm larvae dramatically increase their food intake during the final instar stage, when the silk gland cells grow most rapidly under the stimulation of nutrients and insulin signals [3,4]. The silk fibre is composed of two types of major silk proteins, sericins and fibroins, which are synthesized in the MSG and PSG, respectively, and subsequently secreted into the lumen of the silk gland for storage until spinning [5,6]. Molecular insights into regulatory mechanisms in silk gland growth and silk protein synthesis will contribute to improving silk quality and exploiting silk protein applications.

Figure 1.

Genomic disruption of let-7 and let-7C by CRISPR/Cas9. (A) Morphology and divisions of the silk gland at D7 IL5. (B) The sgRNAs used for let-7 and let-7C knockout. The sequences in red are mature miRNAs, the sequences in green are the base-pairing regions of sgRNAs, and red boxes are protospacer-adjacent motifs (PAMs). (C) Schematic representation of the transgenic vectors for let-7 and let-7C knockout. Double-sgRNA expression cassettes were driven by the U6 promoter. EGFP was used as a selection marker. (D) Positive individuals selected at the G1 embryo and adult stages by screening the EGFP marker in the eye. (E) PCR products showing the chromosomal fragment deletion between sgRNA2 and sgRNA3 in Δlet-7-MSG. (F) PCR products of the chromosomal fragment deletion between sgRNA1 and sgRNA3 in Δlet-7C-MSG. (G–I) Sequence alignment of sgRNA-targeted genomic regions. The red arrow shows a fragment deletion.

MicroRNAs (MiRNAs) are small endogenous noncoding RNAs which regulate various cell processes such as protein synthesis, DNA replication, cell proliferation, cell differentiation and apoptosis [7,8]. To date, over 600 miRNAs have been identified in the silkworm, but very few of them have been functionally determined [9,10]. The co- or differential expression of miRNAs in different divisions of the silk gland or in different silkworm strains with diverse silk yields support their important roles in this specialized organ [11–14]. Among all known miRNAs of silkworm, let-7 was the first one experimentally verified and the most comprehensively characterized for its spatial and temporal expression patterns [15]. As a highly conserved miRNA across animal phylogeny and a founding member of the miRNA community [16,17], let-7 has become a star member for mechanistic and functional studies of miRNAs in many species. We have verified that let-7 is markedly weakly expressed in individuals and silk glands from fifth-instar day-3 larvae (D3 IL5) to D5 IL5 [15]. The low expression of let-7 in specific periods and organs should have important biological significance since let-7 generally controls related phenotypes through negative regulation of its target genes. Here, using a multiplex genome editing strategy for the CRISPR/Cas9 system [18], we successfully deleted let-7 and its entire cluster (let-7C) in the middle and posterior silk gland, providing evidence that let-7 is an indispensable regulator of silk gland development and silk protein synthesis. We further confirmed these findings via transgenic overexpression of let-7. Finally, we found that let-7 controls silk gland growth and function by regulating key enzymes in glycometabolism pathways. In conclusion, this work provides insight into novel biological and molecular mechanisms coordinated by small RNAs in the specialized organ of living organisms.

Results

Knockout of let-7 and the let-7C in MSG and PSG

let-7C is a polycistronic gene spanning < 3.0 kb which is located within silkworm Chr23 and from which let-7, miR-100 and miR-2795 are transcribed (Fig. S1A). In agreement with our previous reports [12,15], further assay showed that these three miRNAs were highly expressed from the fourth moult to the adult stage (Fig. S1B), and expressed ubiquitously but weakly in the silk gland at D3 IL5 (Fig. S1C). In particular, they were weakly expressed in the silk gland during the first three days of fifth instar, and then quickly increased to a peak level at D6 IL5, and finally declined at the end of cocoon spinning (Fig. S1D). To reveal the molecular mechanisms by which let-7 and let-7C are involved in different divisions of silk gland (Fig. 1A), we performed CRISPR/Cas9-based genome editing to knock out let-7 and let-7C in the MSG and PSG. To increase the knockout efficiency, three sgRNAs were designed for chromosomal-fragment deletion of let-7 and let-7C in the genome (Fig. 1B). Especially, to readily achieve multiplex genome editing and fragment deletion of let-7C, two pairs of sgRNAs (sgRNA1/sgRNA3, sgRNA2/sgRNA3) were separately generated from a single polycistronic tRNA-gRNA (PTG) gene [18], under control of a U6 promoter (Fig. 1C). The positive G1 generation expressing double-sgRNA was screened in day 6 embryos and in adults by examining Green fluorescence protein (GFP) expressed in the eyes under the control of the neuron-specific 3 × P3 promoter (Fig. 1D). A positive G1 strain expressing sgRNA1/sgRNA3 (named let-7C-2gRNA) was used to generate a let-7C knockout. In parallel, a positive strain expressing sgRNA2/gRNA3 (named let-7-2gRNA) was used to produce a let-7 knockout.

The PSG-specific promoter FibH-p and the MSG-specific promoter Ser1-p were used to express the Cas9 protein in PSG and MSG, respectively, and positive individuals from the PSG-Cas9 and MSG-Cas9 strains were screened based on the selectable marker Red fluorescent protein (RFP) expressed in the eyes [19]. Positive G1 strains for let-7-2gRNA and let-7C-2gRNA were then separately crossed to PSG-Cas9 and MSG-Cas9 strains, generating four types of positive hybrid F1 individuals which were selected by screening of green and red fluorescence markers in the eyes of late embryos and were named Δlet-7-PSG, Δlet-7C-PSG, Δlet-7-MSG and Δlet-7C-MSG (Fig. S2A–D). To verify whether a large chromosomal fragment deletion occurred at the let-7 and let-7C loci, the fragments containing the sgRNA target sites were PCR-amplified with site-specific primers. The results showed that sgRNA2 and sgRNA3 could act simultaneously to achieve a large DNA fragment deletion for let-7 (Fig. 1E), but no truncated PCR products were clearly observed for let-7C (Fig. 1F). Reduced expression levels of miRNAs in all knockout strains were confirmed by qRT-PCR (Fig. S3). To further confirm the effect of genome editing, the PCR products were cloned and submitted to Sanger sequencing. let-7 and let-7C sequences were deleted, and there were highly polymorphic mutations at each of the three Cas9 cut sites, with deletion size ranging from 3bp to 83bp (Fig. 1G–I). A chromosomal fragment (61bp) was deleted between the sgRNA2 and sgRNA3 target sites, but the deletion of a larger DNA fragment between the sgRNA1 and sgRNA3 target sites was not detected. In conclusion, let-7 and let-7C were knocked down in the MSG and PSG, but the knockout efficacy by a single sgRNA was higher than that produced by double sgRNAs.

Deletion of let-7 and let-7C resulted in PSG enlargement and increased silk production

We dissected the larvae at different time points to compare the silk glands. The PSGs of Δlet-7-PSG were almost the same size as those of the control groups at D1 IL5 (Fig. 2A), but were significantly larger than those of the control groups from D3 IL5 (Fig. 2B–D). A similar abnormal phenotype was also observed for PSG of Δlet-7C-PSG (Fig. 2E). For more precision, we dissected the larvae at D6 IL5 and measured the length of the silk glands hanging vertically, and found that the PSGs of the knockout lines Δlet-7-PSG and Δlet-7C-PSG increased by over 60%, compared with those of control groups (Fig. 2F,G). Furthermore, the weight of PSGs from both knockout lines increased by about 150% (Fig. 2H,I), and the diameter of the PSG transverse section increased 4-fold, but there was no significant difference in the distribution of fibroin in the PSG lumen (Fig. 2J). Next, we dissected the prepupae and found no residual fibroin in the PSG lumen (Fig. 2K,L), so the increased fibroin was completely utilized in assembling the silk. Finally, we evaluated the silk yield at the pupal stage which revealed that the average cocoon weight increased by over 10% for both male and female silkworms of Δlet-7-PSG and Δlet-7C-PSG (Fig. 2M), but the average weight of pupae of each knockout strain showed no significant difference from that of the controls (Fig. 2N). Interestingly, although the expression of these miRNAs almost equally decreased in let-7 mutants compared to let-7C mutants (Fig. S3), deletion of the whole let-7 cluster could exacerbate the abnormal phenotypes since the PSGs of Δlet-7C-PSG were thicker and curlier than those of Δlet-7-PSG (Fig. 2O).

Figure 2.

Knockout of let-7 and let-7C promoted PSG enlargement and improved silk yield. (A–D) Silk glands of Δlet-7-PSG and controls at D1 IL5, D3 IL5, D5 IL5 and D7 IL5. The red line shows the division between the MSG and the PSG. (E) Silk glands of Δlet-7C-PSG and controls at D7 IL5. (F,G) The length of the PSG was significantly increased in Δlet-7-PSG (F) and Δlet-7C-PSG (G) at D7 IL5. (H,I) The weight of the PSG was significantly increased in Δlet-7-PSG (H) and Δlet-7C-PSG (I) at D7 IL5. (J) The diameter of the PSG was increased in Δlet-7-PSG and Δlet-7C-PSG at D7 IL5. (K) Silk gland of Δlet-7-PSG post cocoon spinning or at pre-pupa stage. (L) Silk gland of Δlet-7C-PSG post cocoon spinning or at pre-pupa stage. (M) Cocoon weight of female and male was significantly increased in the Δlet-7-PSG and Δlet-7C-PSG. (N) Pupa weight of female and male was not altered in Δlet-7-PSG and Δlet-7C-PSG. (O) Phenotype difference between Δlet-7-PSG and Δlet-7C-PSG at D7 IL5. The error bars indicate the mean ± SEM, *P < 0.05, ***P < 0.001.

Overexpression of let-7 led to silk gland atrophy and silk yield reduction

Next, we tried to overexpress let-7 and let-7C in the PSG in the speculation that abnormal phenotypes opposite to those of let-7 deletion might be observed. To this end, we designed a transgenic overexpression vector for let-7 and a co-expression vector for miR-100 and miR-2795 (Fig. 3A,B). Considering that a primary precursor is necessary for the generation of a mature miRNA sequence, a 190bp genomic sequence covering the let-7 precursor and a 545bp genomic sequence containing the primary precursor of [miR-100 + miR-2795] were cloned to construct their overexpression vectors (Fig. S4A–C). The transgenic vectors piggyBac [3× P3-EGFP, FibHp-let-7OE] and piggyBac [3× P3-EGFP, miR-100 + 2795OE] were injected into newly laid eggs, and then the positive overexpression individuals were screened in the G1 generation and named let-7-OE-PSG and [miR-100+ miR-2795]-OE-PSG. Expression levels of these miRNAs were all significantly increased in their own overexpression strain, respectively (Fig. 3D–F). Anatomical observation at D6 IL5 showed that the PSGs of the let-7-OE-PSG strains were only one-third the length of the wild type silk glands (Fig. 3C,G). Most interestingly, the MSGs of let-7-OE-PSG were also significantly shorter than the control MSGs (Fig. 3C,G).

Figure 3.

Transgenic overexpression of let-7 inhibited the development of PSG and the synthesis of fibroin protein. (A) Schematic diagram of the let-7 transgenic overexpression vector. (B) Schematic diagram of the miR-100+ miR-2795 transgenic overexpression vector. miRNA expression cassettes were driven by the FibH promoter. EGFP was used as a selection marker. (C) Silk gland of WT, let-7-OE-PSG and [miR-100+ miR-2795]-OE-PSG at D7 IL5. (D) Increased expression of let-7 in let-7-OE-PSG at D7 IL5. (E) Increased expression of miR-100 in [miR-100+ miR-2795]-OE-PSG at D7 IL5. (F) Increased expression of miR-2795 in [miR-100+ miR-2795]-OE-PSG at D7 IL5. (G) Length of the silk gland and its divisions decreased in the overexpression strain let-7-OE-PSG at D7 IL5. (H) The weight of cocoon with pupa increased in let-7-OE-PSG. (I) Cocoon shell weight decreased in let-7-OE-PSG. (J) Cocoon shell and pupa of let-7-OE-PSG and controls. (K) 20 mg cocoon shell of each strain was crushed and maintained in 8 M urea solution at 95 °C for one hour. (L) The weight of fibroin decreased in let-7-OE-PSG and [miR-100+ miR-2795]-OE-PSG. (M) The weight of female or male pupa increased in let-7-OE-PSG. (N) The length of female or male pupa increased in let-7-OE-PSG. The error bars indicate the mean ± SEM, *P < 0.01, **P < 0.05, ***P < 0.001.

Unlike let-7-OE-PSG, both PSG and MSG of [miR100+ miR-2795]-OE-PSG exhibited no visible abnormal phenotypes (Fig. 3C). The whole cocoon weight of let-7-OE-PSG increased by about 40%, compared with the wild type control (Fig. 3H), while the cocoon shell weight of let-7-OE-PSG decreased by about 50% (Fig. 3I). The cocoon shell of let-7-OE-PSG was very thin and fragile (Fig. 3J), similar to the phenotype of the Nd mutant which lacks a functional fibroin heavy chain gene, and has no fibroin but only sericin present in the cocoon layer [20]. Sericin, which comprises 75–80% of raw silk, is much more soluble in 8 M urea solution than the fibroin, which contributes 20–25% of raw silk [21,22]. In order to understand the nature of the transparent cocoon, 20 mg cocoon shell of each strain was crushed and maintained in 8 M urea solution at 95 °C for one hour. Solutions containing wild type or [miR-100 + miR-2795]-OE-PSG cocoon shells were turbid, while the solution from the cocoon shell of let-7-OE-PSG appeared transparent, strongly consistent with the absence of fibroin in the let-7-OE-PSG cocoon shell (Fig. 3K). Then, we dried and weighed the material that was insoluble in 8 M urea and found that silk fibroin of the let-7-OE-PSG decreased by over 80%, compared with the control, while the fibroin of [miR-100+ miR-2795]-OE-PSG decreased by about 12% (Fig. 3L). The weight per female or male pupa of let-7-OE-PSG increased by about 50%, compared with the control (Fig. 3M) and the length per female and male pupa of let-7-OE-PSG increased by about 15% (Fig. 3N). Taken together, overexpression of let-7 in the PSG led to obvious silk gland atrophy and silk yield reduction but no clear abnormal phenotypes were observed in [miR-100+ miR-2795]-OE-PSG, implying that let-7 might act as the main contributor in the let-7 cluster to normal PSG development and fibroin synthesis.

Knockout of let-7 and let-7C resulted in MSG enlargement and silk yield reduction

To verify the potential abnormal phenotypes in the MSG of Δlet-7-MSG and Δlet-7C-MSG strains, we dissected the larvae at different time points of the fifth instar. At D1 IL5, we observed almost no differences in the MSG between Δlet-7-MSG and the controls, but from D3 IL5 to the cocoon spinning stage, the P-MSG of Δlet-7-MSG larvae was longer and thicker than controls (Fig. S5A). Similarly, abnormal MSG size and appearance were found in Δlet-7C-MSG from D3 IL5 (Fig. S5B). In the wild type silkworms, the three regions of the MSG appeared smooth, and the P-MSG was shorter than M-MSG (Fig. 4A). However, in the knockout strain Δlet-7-MSG, the P-MSGs of some individuals at D6 IL5 were significantly enlarged and coarsened, and were about 30% longer than their corresponding M-MSGs (Fig. 4A). Δlet-7C-MSGs exhibited clear phenotypic differences from Δlet-7-MSGs. Abnormal phenotypes only occurred in the P-MSG of Δlet-7-MSG, but the roughly serrated surface occurred in both M-MSG and P-MSG of Δlet-7C-MSG; furthermore, the P-MSG was shorter than the M-MSG in Δlet-7C-MSG (Fig. 4A). We further examined transverse frozen sections of the MSG and found that the diameter of the MSG lumen of both Δlet-7-MSG and Δlet-7C-MSG was significantly larger than the controls, and a large amount of sericin was secreted into the lumen of mutant MSGs (Fig. 4B). All sericin proteins stored in the silk gland lumen were used for silk production in wild type silkworms, but unexpectedly, some sericin was left in the lumen of P-MSG of both Δlet-7-MSG and Δlet-7C-MSG after cocoon spinning (Fig. 4C). Consistently, the cocoon weight or silk yield was significantly decreased for both Δlet-7-MSG and Δlet-7C-MSG compared with the wild type silkworms (Fig. 4D). Together, these data indicate that let-7 or the whole let-7 cluster is required for maintaining the normal status and function of the middle silk gland in the silkworm.

Figure 4.

Abnormal phenotypes caused by deletion of let-7 and let-7C in MSG. (A) The phenotypes of Δlet-7-MSG and Δlet-7C-MSG at D7 IL5. A, anterior region of MSG; M, middle region of MSG; P, posterior region of MSG. (B) The diameter of the P-MSG was increased in the Δ let-7-MSG and Δlet-7C-MSG mutants at D7 IL5. White: microscopy under common white light illumination; DAPI: fluorescent dye 4ʹ,6-diamidino-2-phenylindole and microscopy under fluorescence. (C) Silk glands of Δlet-7-MSG and Δlet-7C-MSG showing the residual sericin in the MSG lumen at prepupal stage. (D) The cocoon weight was significantly decreased in the Δlet-7-MSG and Δlet-7C-MSG mutants. The error bars indicate the mean ± SEM, **P < 0.01, ***P < 0.001.

Knockout of let-7 and the let-7C promoted endoreduplication in silk gland cells

Silk gland growth is not determined by cell division at the embryonic stage but by cell enlargement during the larval growth periods [23]. The rapid growth of the silk gland is achieved primarily via chromosome endoreduplication in silk gland cells at larval stages [24], leading to a high level of polyploidy. To explore whether deletion of let-7 and let-7C affected cell enlargement in the silk gland, the PSGs of Δlet-7-PSG and Δlet-7C-PSG silkworms were dissected at D6 IL5 for staining cytoskeletal actin with Alexa Fluor488-labelled phalloidin. Indeed, the cell volume of the PSG increased dramatically in both knockout lines (Fig. 5A). However, a small number of silk gland cells in Δlet-7C-PSG did not expand normally (marked with a red arrow), thus resulting in the serrated surface of the silk gland (Fig. 5A). The increased DNA content in the endoreduplicated wild type silk gland cells at the fifth instar leads to a dendritic nucleus [23]. To confirm whether endoreduplication was promoted in the knockout strains, silk glands collected at D7 IL5 were stained with DAPI and examined by fluorescence microscopy. We found that DNA was distributed throughout all PSG cells of Δlet-7-PSG (Fig. 5B). Further, we found that the total DNA content in the PSG cells measured at D7 IL5 increased by 150%~200% (Fig. 5C). The P-MSG was most expanded in both Δlet-7-MSG and Δlet-7C-MSG, which was also presumably related to the enhanced endoreduplication of the silk gland cells. Indeed, the DNA content increased by 170% and 120% in the P-MSG of Δlet-7-MSG and Δlet-7C-MSG, respectively (Fig. 5D).

Figure 5.

Knockout of let-7 and let-7C promoted endoreduplication in silk gland cells. (A) The size of PSG cells increased dramatically in Δlet-7-PSG and Δlet-7C-PSG at D7 IL5. (B) DAPI staining showing DNA in the PSG of Δlet-7-PSG at D7 IL5. In (A) and (B), the red arrow shows the abnormal cell, and the dot line-enclosed area represents a single silk gland cell. (C) The DNA content increased in the P-MSG of Δlet-7-MSG and Δlet-7C-MSG at D7 IL5. (D) The DNA content increased in the M-MSG of Δlet-7-PSG and Δlet-7C-PSG at D7 IL5. (E) The ATP level increased in the M-MSG of Δlet-7-PSG and Δlet-7C-PSG at D3 IL5. The red arrow shows the malformed cell. (F) The volume of PSG cells decreased dramatically in the let-7-OE-PSG individuals at D7 IL5, but not in [miR-100+ miR-2795]-OE-PSG individuals at D7 IL5. (G) DAPI staining showing DNA in the PSG of let-7-OE-PSG at D7 IL5. In (F) and (G), the red and white arrows show the big and small cells, respectively. The error bars indicate the mean ± SEM, **P < 0.01, ***P < 0.001.

To obtain more evidence for the enhanced endoreduplication, we measured ATP levels in the P-MSG as a proxy for the amount of energy required for DNA replication, and found that the P-MSGs of Δlet-7-MSG and Δlet-7C-MSG increased by 8- and 10-fold, respectively (Fig. 5E). In summary, knockout of let-7 and let-7C remarkably accelerated silk protein synthesis and endoreduplication in silk gland cells, and as a result, contributed to the overall growth of the silk gland.

Contrary to the deletion of let-7, its overexpression in the PSG resulted in a PSG defect of up to 95%, and shortening of the MSG by at least 50% (Fig. 3G). This severe growth inhibition might be caused by decreased cell size or reduced cell number. To address this issue, PSGs were dissected out at D7 IL5 for staining with phalloidin-488 dye and examining microscopically (Fig. 5F). Cells in the PSG of let-7-OE-PSG did not grow or expand normally and were 2–5 times smaller than the control group. Even in the PSG of let-7-OE-PSG, there existed two types of cells with a size difference of more than five-fold, the large ones with a clear cell shape (marked with red arrows) and the small ones completely atrophied (marked with white arrows). Furthermore, the PSG surface became uneven with an irregular cell arrangement (Fig. 5F). However, overexpression of miR-100 and miR-2795 in the PSG did not cause significant changes in the size and shape of cells (Fig. 5F). We then stained the PSGs of let-7-OE-PSG and control strains with DAPI at D7 IL5 and found that both the large and small PSG cells of the let-7-OE-PSG contained condensed chromosomal DNA (Fig. 5G). However, it could be inferred from the degenerative PSG cells that endoreduplication was severely hindered in the PSG when let-7 was overexpressed, and that let-7-OE-PSG also lost the ability to synthesize fibroin (Fig. 3J–L). Of note, overexpression of miR-100 + miR-2975 exerted no significant effect on the shape or size of PSG (Fig. 5F), nor on the endoreduplication (Fig. 5G). Altogether, these observations show that let-7 might act as the main contributor in regulating the endoreduplication of silk gland cells, while miR-100 and miR-2795 probably only play a complementary role in this regulatory system.

Transcriptome analysis revealed differentially expressed genes and enriched pathways in response to let-7 knockout

To explore the molecular players and signalling pathways underlying the promoted enlargement of silk gland cells in response to deletion of let-7, we performed comparative transcriptome analysis of Δlet-7-PSG and Δlet-7-MSG (Supplementary Dataset S1). The three biological replicates of each group were highly consistent at the genome-wide transcriptome level (Fig. S6A). Totally, 548 differentially expressed genes (DEGs) were identified in Δlet-7-PSG, including 271 up-regulated and 277 down-regulated (Fig. S6B; Dataset S2), and 283 DEGs were identified in Δlet-7-MSG, including 184 genes up-regulated and 99 down-regulated (Fig. S6C; Dataset S3). Only 26 DEGs (4.74%) were common to Δlet-7-MSG and Δlet-7-PSG, including 6 down-regulated genes, notably mitochondrial phosphoenolpyruvate carboxykinase (PEPCK) (Fig. S6D), and 20 up-regulated genes, including lactate dehydrogenase (LDH) and glucose-6-phosphate 1-dehydrogenase (G6PDH) (Fig. S6E). The reliability of the DEGs identified in Δlet-7-PSG and Δlet-7-MSG through transcriptome sequencing analysis was verified by qRT-PCR assay (Fig. S6F,G).

Gene Ontology (GO) term analysis revealed that top-ranked biological processes enriched in the PSG of Δlet-7-PSG mostly function in the metabolic process, nucleus, carbohydrate transport, extracellular region, sulfotransferase activity and lipid binding (Fig. S7A). KEGG pathway enrichment analysis showed that the DEGs in Δlet-7-PSG were mainly enriched in metabolic pathways, especially the folate-mediated one-carbon metabolism (Fig. S7B). In detail, the KEGG analysis revealed up-regulated genes functioning in oxidation-reduction process, glycolysis/gluconeogenesis, pentose phosphate pathway, citrate cycle (TCA cycle), fatty acid metabolism, biosynthesis of amino acids and DNA replication (Dataset S2). The analysis also revealed down-regulated genes functioning in glycolysis/gluconeogenesis and TCA cycle, such as succinate-CoA ligase and mitochondrial aldehyde dehydrogenase (Dataset S2).

GO term analyses revealed that DEGs in the MSG of Δlet-7-MSG were most enriched in oxidation-reduction process, followed by fatty acid biosynthetic process, fatty acid β-oxidation, cellular metabolic process, acetyl-CoA biosynthetic process, citrate metabolic process, cellular carbohydrate metabolic process, glycolysis/gluconeogenesis, ATP citrate synthase activity and catalytic activity (Fig. S8A). KEGG pathway analysis demonstrated that DEGs in Δlet-7-MSG were also mainly enriched in metabolic pathways, followed by peroxisome, fatty acid metabolism and TCA cycle (Fig. S8B). In sum, based on transcriptome sequencing, a large number of DEGs were revealed in the PSG of Δlet-7-PSG and MSG of Δlet-7-MSG. Further functional analysis revealed that the DEGs in Δlet-7-MSG were mainly related to carbohydrate metabolism and ATP synthesis pathway, whereas the DEGs in Δlet-7-PSG were mainly related to metabolic process, folate-mediated one-carbon metabolism and carbohydrate transport. These results collectively suggest that let-7 coordinates the carbohydrate metabolic processes and the energy metabolism signalling pathways in the silk gland, providing effective mechanisms for its overall growth, the synthesis and secretion of silk proteins, and the final formation of silk fibres.

let-7 target genes

Among the DEGs identified in the knockout strains Δlet-7-MSG and Δlet-7-PSG, according to established principles of miRNA regulation, only those that were up-regulated are likely to be the targets of let-7. We predicted putative targets for let-7 using three algorithms (see Methods), and obtained 504 overlapping genes (Fig. 6A), which we then compared with the up-regulated genes identified by RNA-Seq analysis. This resulted in 22 overlapping genes (Fig. 6A; Table S1), including 10 in Δlet-7-MSG and 12 in Δlet-7-PSG. We tried to further narrow the screening range of targets based on the bivariate relationships between the temporal expression patterns of these 22 candidate targets and that of let-7 in the wild type silk gland. Only 9 genes were negatively correlated with let-7 in their temporal expression in the silk gland (Fig. S9), and were thus regarded as core candidate targets.

Figure 6.

Identification of let-7 targets and transgenic overexpression of PC in MSG. (A) Schematic diagram of let-7 target screening. Venn diagram represents the overlap of potential targets predicted by three computational programs. Twenty-two of them are up-regulated by the let-7 deletion. (B) Two putative binding sites of let-7 within the 3ʹUTR of PC mRNA. Red: let-7 seed sequences; green, mutant target sites. (C) Dual-Luciferase Reporter Assay showing that let-7 directly targets the 3ʹ-UTR of PC by binding to two target sites in vitro. (D) Dual-Luciferase Reporter Assay showing that miR-100 and miR-2795 do not target the 3ʹ-UTR of PC in vitro. (E) let-7 deletion led to an increase of the PC transcript level in the MSG at D3 IL5. (F) Schematic recombinant vector for transgenic overexpression of PC gene in the silkworm MSG. (G) Screening of positive PC-OE-MSG strains at embryonic and adult stages. (a) Embryonic stage under the blue fluorescence. (b) Adult stage under the blue fluorescence. (H) The MSG enlarged at D7 IL5 after overexpression of PC. (I) PC expression levels increased in the MSG of PC-OE-MSG strain and miRNA mutants at D7 IL5. (J) Length of the MSG increased in the PC-OE-MSG strain at D7 IL5. (K) The DNA content increased in the MSG of PC-OE-MSG at D7 IL5. (L) Weight of larvae at D5 IL5. Error bars indicate the mean ± SEM, **P < 0.01, ***P < 0.001, ns represents not significant.

The binding sites of miRNAs are generally located on the 3ʹ-untranslated region (3ʹ-UTR) of mRNAs, so we tried to determine the 3ʹ-UTRs of these candidate targets by 3ʹ RACE. Finally we successfully cloned 8 of them which we then inserted into a psiCheck-2 vector downstream of the Renilla luciferase stop codon (Fig. S10 and Table S2). We then co-transfected each recombinant vector with let-7 Mimic into HEK-293T cells. A dual-luciferase reporter gene assay showed no significant difference in the ratio of luciferase activity in 6PGD-3ʹUTR-psiCheck2, TMEM201-3ʹUTR-psiCheck2, Mmd-3ʹUTR-psiCheck2, Tret1-3ʹUTR-psiCheck2 and Chaoptin-3ʹUTR-psiCheck2 compared with the control groups Fig. S11). These results indicated that these genes might not be direct targets of let-7, and their expression change in response to let-7 deletion likely occurred downstream in the let-7 regulatory network. However, the ratio of luciferase activity in PC-3ʹUTR-psiCHECK2, Eck-3ʹUTR-psiCHECK2 and XLCOF6-3ʹUTR-psiCHECK2 was significantly lower than that of controls (Fig. S11), confirming that let-7 directly binds to the 3ʹUTR of pyruvate carboxylase (PC), ecdysteroid 22-kinase (EcK) and oocyte zinc finger protein XLCOF6 (XLCOF6), negatively regulating their expression.

Next, we tried to identify the binding sites of let-7 in these 3ʹ-UTRs. The 3ʹ-UTR of PC was predicted to contain two let-7 binding sites, at position 97 and 1375 (Fig. 6B). Mutation of either of the two sites only partially restored the activity of the reporter gene, indicating that both target sites are authentic, but they are only responsible for part of the let-7 binding effect. On the other hand, when both sites were mutated simultaneously, the activity of the reporter gene was completely restored (Fig. 6C), indicating that let-7 only has two binding sites within the 3ʹ-UTR of PC mRNA. However, PC 3ʹ-UTR was not targeted by miR-100 and miR-2795 (Fig. 6D). Further, qRT-PCR results confirmed the PC level was elevated in the MSG of Δlet-7-MSG, but remained unchanged in the PSG of Δlet-7-PSG (Fig. 6E). These results were consistent with those of transcriptome sequencing (Table S1), suggesting that the PC gene is specifically regulated in the MSG by let-7. The binding sites of let-7 within the 3ʹ-UTR of targets EcK and XLCOF6 were also clearly verified through site mutation and Dual Luciferase Reporter Assay (Fig. S12). In summary, our results confirmed that let-7 negatively regulates these three targets through the binding sites within their 3ʹUTRs.

Contributions of let-7 targets to silk gland growth revealed by CRISPR/Cas9 editing and transgenic overexpression

The three targets were all differently expressed in the tissues of silkworm, and weakly in the silk gland (Fig. S13A–C). In order to gain insight into their potential contributions to the silk gland development, we adopted the CRISPR/Cas9 system to specifically knock out these three target genes in the MSG and PSG (Figs. S14A–D, S15A–D, S16A–D). Base or fragment deletions occurred at the corresponding gRNA sites of PC, XLCOF6 and EcK with knockout efficiency of 59%, 74% and 28% respectively for mutations ranging from single bases to fragments 64 bases long (Figs. S14E, S15E and S16E). We then dissected the larvae of different knockout strains at D5 IL5, and examined their silk glands. None of the deletions of these genes in the MSG or PSG produced significant changes in the size or shape of silk glands (Figs. S14F-G, S15F-G and S16F-G).

To further elucidate the role of these targets in the silk gland, we cloned the full coding region sequence (CDS) of PC (3,549 bp) and constructed transgenic overexpression vectors containing the MSG-specific promoter Ser1p (Fig. 6F). The MSG-specific plasmids piggyBac [3× P3-EGFP, Ser1p-PC] were injected into newly laid eggs. After hatching, positive individuals of the transgenic line PC-OE-MSG were repeatedly screened at embryonic and adult stages (Fig. 6G), and the silk glands were dissected out at D7 IL5 for phenotypic investigation. The MSG of PC-OE-MSG was significantly longer than the wild type control (Fig. 6H), which was the same abnormal phenotype revealed in Δlet-7-MSG. Further, the length of the MSG increased by 25% at this stage compared with that of controls (Fig. 6I). The expression level of PC in the MSG of PC-OE-MSG line was significantly higher than that in the wild type control, and even much higher than those in the knockout lines, Δlet-7-MSG and Δlet-7C-MSG (Fig. 6J). The DNA content increased in the MSG of PC-OE-MSG at D7 IL5 (Fig. 6K), which was similar to the DNA content change in the knockout lines (Fig. 5D). There was no significant difference in the body weight of larvae between PC-OE-MSG and the wild type control (Fig. 6L). Together, no abnormal phenotype was detected after deletion of these target genes, but overexpression of PC resulted in silk gland enlargement, similar to that after deletion of let-7.

Discussion

let-7 acts as the main contributor in regulating the endoreduplication and cell growth

In many insect species, let-7 clusters with miR-100 and the lin-4 homolog miRNA-125 at a single genomic locus known as the let-7-Complex [25,26]. Although coexisting in a cluster, these three miRNAs do not function equally in controlling temporal cell fate during animal development. In Drosophila, let-7 and miR-125 but not miR-100 target the transcription factor Chimo to coordinately regulate neuronal differentiation in the mushroom body [26]. In Blattella germanica, depletion of let-7 or miR-100 causes a number of vein patterning defects while reduction of the miR-125 level induces no apparent effects [25]. In the silkworm genome, however, miR-125 is completely absent, and let-7 clusters with the conserved miR-100 and silkworm-specific miR-2795 at the same locus and they exhibit parallel spatial and temporal expression patterns [11,12,27]. Therefore, insight into the functional relationships between different miRNAs in the let-7C is crucial for understanding the molecular mechanisms in controlling the development and functions of the silk gland.

The developmental process of silk gland can be divided into morphogenesis at embryonic stages, in which the number of silk gland cells is determined, and overall growth at larval stages, in which DNA replicates for 18–19 rounds without nuclear division, resulting in gigantic silk gland cells full of nuclear material [3,28,29]. The functional relevance of several genes in the cell size and endoreduplication of the silk gland has been reported [4,29], although the underlying molecular mechanisms still remain largely unknown, as few studies have demonstrated how the overall growth of silk gland is regulated at the larval stages. The present study provides evidence that let-7 deletion enhances the overall growth of silk gland, while the whole let-7 cluster can exacerbate the abnormal phenotypes (Figs. 2 and 4). Further, the total DNA content and ATP levels significantly increased in the PSG and MSG of both let-7 and let-7C knockout strains (Fig. 5C–E). However, concurrent overexpression of miR-100 and miR-2975 exerted no significant effect on the shape and size of the PSG (Figs. 3G and 5F), nor on the endoreduplication (Fig. 5G). These data collectively support the conclusion that let-7 functions as the main contributor in regulating the endoreduplication of silk gland cells and the overall growth of silk gland, while miR-100 and miR-2795 only play a complementary role in the let-7 regulatory system. The parallel expression patterns and coordinate actions of let-7C in the silk gland might be associated with ecdysone [15]; however, additional study is required to determine the targets of miR-100 and miR-2795.

The cluster miRNAs are generally co-transcribed at a genomic locus, so when one miRNA in the cluster is knocked out, the secondary structure of their common primary precursor will be changed, and the biogenesis of other miRNAs in the cluster will thus be affected. Therefore, when let-7 was knocked down, the expression level of other miRNAs in the cluster also decreased (Fig.S3A–D). Interestingly, the expression levels of these miRNAs are almost equally reduced in let-7 mutants compared to let-7C mutants. Logically, both let-7 mutant and let-7C mutant should produce the same abnormal phenotypes. However, stronger abnormal phenotypes were observed in let-7 mutants relative to the let-7C mutants (Figs. 2O and 4A; Fig. S5). So, there arise two questions: 1. Which miRNA in the cluster is responsible for these phenotypes? 2. Why are the abnormal phenotypes of let-7 mutants different from those of let-7C mutants?

To address these issues, we overexpressed let-7 and miR-100/miR-2795 in the PSG, respectively, and found that overexpression of let-7 caused obvious silk gland atrophy and silk yield reduction but no clear abnormal phenotypes were observed in [miR-100+ miR-2795]-OE-PSG (Fig. 3C, 5F,G). We thus speculate that let-7 might act as the main contributor in the let-7 cluster to normal PSG development and fibroin synthesis, while miR-100 and miR-2795 might only play a complementary role in this regulatory system. Given the equivalent expression decrease (Fig. S3), it is difficult to explain the phenotypic differences between let-7 and let-7C mutants by the decline degree in miRNA expression. Although the knockout was confirmed by sequencing (Fig. 1G–I), we cannot rule out this possibility that let-7 and let-7C might have different knockout types produced by different sgRNAs (Fig. 1B). Therefore, although miRNA accumulation was equally affected, it is possible that some unknown functional units in the primary transcripts are also related to the silk gland growth and silk protein synthesis, and longer deletions caused by the sgRNA1/sgRNA3 might be able to exacerbate the abnormal phenotypes.

let-7 negatively regulates the accumulation of silk proteins and the overall growth of silk gland

The giant cells of PSG and MSG are efficient cellular machines for the synthesis of fibroin and sericin, respectively [2]. When let-7 and let-7C were knocked down, the increased fibroin due to the enlarged PSG was fully utilized in assembling the silk and the cocoon weight remarkably increased (Fig. 2M) but both the increased sericin synthesized by the enlarged MSG and the fibroin produced in the PSG were not fully utilized (Fig. S17), and then cocoon weight decreased (Fig. 4D). Accordingly, we postulated that the cocoon weight of silkworm probably depends mainly on the amount of fibroin synthesized in the PSG cells rather than on the sericin secreted from the MSG cells. When let-7 was overexpressed in transgenic silkworms, the growth of the PSG was significantly arrested (Fig. 3G) and the silk yield significantly decreased (Fig. 3J), further supporting a negatively regulatory role for let-7 in the whole growth of silk gland at larval stages. The MSGs of let-7-OE-PSG were significantly shorter than the control MSGs (Fig. 3H), hinting that overexpression of let-7 in the PSG cells may directly or indirectly affect the expression of genes in the MSG through signal exchanges between cells at the junction of the silk gland divisions.

The major molecular mechanisms involving let-7 might diversify in different divisions of the silk gland

A comprehensive understanding of the genetic basis of silk gland development and function is crucial for improving silk yield and silk quality through fine genetic manipulation. To this end, we performed RNA-Seq and systematically analysed the DEGs in the MSG and PSG of Δlet-7-MSG and Δlet-7-PSG at D3 IL5, and found that the DEGs were mostly involved in different biological processes (Figs. S7 and S8). DEGs in the MSG of Δlet-7-MSG were mainly enriched in the redox pathways such as tricarboxylic acid cycle and fatty acid metabolism, while the DEGs in the PSG Δlet-7-PSG were significantly enriched in the one-carbon metabolic process, which exists in the form of tetrahydrofolic acid and participates in DNA synthesis. The top-ranked enriched pathways in the PSG were different from those in the MSG, suggesting that the major molecular mechanisms involving let-7 diversify in different divisions of the silk gland. Of note, a group of genes encoding DNA polymerases were highly expressed in both PSG and MSG, and were mainly up-regulated in the PSG of Δlet-7-PSG (Table S3). These observations are strongly indicative of their contributions to the enhanced chromosomal DNA replication in the silk glands of both knockout strains as well as the strongly increased DNA contents in the PSG cells of Δlet-7-PSG (Fig. 5C,D).

let-7 coordinates the nutrient metabolism processes and energy signalling pathways

Nutrients in mulberry leaves are absorbed by silkworms and converted into carbohydrates, and the main carbohydrate in silkworm haemolymph is trehalose [30], which is then transported into the silk gland cells and hydrolysed by trehalohydrolase into glucose [31]. These processes might be promoted in the let-7 knockout strains since genes encoding the facilitated trehalose transporter Tret1 or Tretl-like were highly expressed and up-regulated in the MSG and PSG (Fig. 7A; Table S3). The genes encoding the rate-limiting enzymes of the pentose phosphate pathway (PPP) were significantly up-regulated in both let-7 knockout strains (Fig. 7B), thus facilitating the production of reductive power and pentose, and promoting the synthesis of nucleic acids and proteins. Glycolysis should be enhanced because genes functioning in the glycolytic process were highly expressed and relatively elevated (Fig. 7C). Furthermore, D-3-phosphoglycerate dehydrogenase (PGDH) and cystathionine gamma-lyase (CTH) were highly expressed and up-regulated in both knockout strains (Fig. 7D; Table S3). This suggests that glycerate 3-phosphate (3PG), an intermediate product of glycolysis, was at least partly converted into serine, glycine and cysteine, namely the amino acids materials for protein synthesis.

Figure 7.

Regulatory network of let-7 is associated with the metabolism of glucose, fatty acids and amino acids in the silk gland. (A) Trehalose enters the silk gland cells and is converted into glucose. (B) let-7 knockout led to enhanced pentose phosphate pathway. (C) Knockout of let-7 promoted Embden-meyerhof pathway. (D) 3-phosphoglyceric acid was partly converted into amino acids as raw material for protein synthesis. (E) TCA cycle was promoted. (F) Gluconeogenesis pathway was weakened. (G) OAA and acetyl-CoA were replenished by CP cycle. (H) Synthesis metabolism of fatty acids was enhanced. (I) Fatty acid beta oxidation was enhanced. The red arrow shows the up-regulated gene and the blue arrow means the down-regulated gene. The abbreviations of Fig.7 are detailed in Table S3.

Pyruvate is generally oxidized and decarboxylated into acetyl-CoA by pyruvate dehydrogenase complex (PDHC) or converted to oxaloacetic acid (OAA) by pyruvate carboxylase (PC) [32]. PDHC was highly expressed and up-regulated in both knockout strains (Table S3), ensuring a high level of acetyl-CoA in the PSG and MSG cells. The accumulated acetyl-CoA activates PC enzyme and synchronizes catalytic events within the PC tetramer [33]. PC was identified as a direct target of let-7 (Fig. 6C), but its negative response to the let-7 deletion only occurred in the MSG (Fig. 6D), thus a higher level of OAA would be produced in the MSG of Δlet-7-MSG. The genes encoding citrate synthase (CS) were up-regulated in the MSG and PSG, and consequently acetyl-CoA and OAA were normally condensed into citrate, the first product in the TCA cycle (Fig. 7E). Furthermore, the genes encoding some other key enzymes in the TCA cycle were also highly expressed and up-regulated in the PSG and MSG (Fig. 7E). Therefore, TCA cycle might be enhanced in the PSG of Δlet-7-PSG and MSG of Δlet-7-MSG, and a large amount of glucose is thoroughly decomposed to release enough ATP for the synthesis of nucleic acids and proteins required for the endoreduplication of cells as well as the overall growth of silk gland.

The transcription level of phosphoenolpyruvate carboxykinase (PEPCK) was significantly decreased in both knockout strains (Fig. 7F), showing that the accumulated OAA in the PSG and MSG cells at D3 IL5 did not promote gluconeogenesis, but likely provided the synthesis of pyrimidine nucleotides with aspartic acid. Lactate dehydrogenase (LDH) is another important enzyme involved in the redox reaction between lactate and pyruvate in glycolysis and gluconeogenesis. LDH was significantly up-regulated in the MSG of Δlet-7-MSG and the PSG of Δlet-7-PSG (Dataset S2 and S3), which strengthened the supply of pyruvate for stable glycolysis and TCA cycle and thus contributed to enough energy and raw materials required for the endoreduplication of silk gland cells and overall growth of silk glad.

Fat body is the nutrient source for the silk gland growth and silk protein synthesis. Fatty acids are synthesized in the cytoplasm, and the raw material acetyl-CoA can be supplied through the citrate pyruvate cycle (CP cycle), which might be promoted by the highly expressed cytosolic malate dehydrogenase (MDH) and NADP-dependent malic enzyme (ME) (Fig. 7G). The biosynthesis of unsaturated fatty acids should be enhanced in the silk gland of let-7 knockout strains because genes functioning in the fatty acid elongation were highly expressed and up-regulated (Fig. 7H). Consistently, the genes functioning in the fatty acid beta-oxidation were also highly expressed and up-regulated in both knockout strains, including mitochondrial Δ^3,5-Δ^2,4-Dienoyl-CoA isomerase (ECH1), acyl-CoA binding protein (ACBP), hydroxyacyl-coenzyme A dehydrogenase (HADH), peroxisomal acyl-coenzyme A oxidase 1 (ACOX1) (Fig. 7I; Table S3). Therefore, let-7 knockout leads to more vigorous synthesis and catabolism of fatty acids in silk gland cells, which further ensure the raw materials for TCA cycle and energy for the increasing protein and nucleic acid synthesis.

Regulatory network of let-7 determines the physiological state and function of the specialized organ in living organisms

Transcriptome analysis showed that hundreds of genes were significantly responsive to let-7 deletion in the MSG and PSG of silkworm. However, which of these genes were directly or indirectly regulated by let-7 remains unknown, except for the three DEGs identified as the let-7 targets. Knockout of the three target genes in both MSG and PSG caused no significant abnormal phenotypes in the silk gland (Figs. S14–S16). There might be several reasons for these outcomes. First, it is likely that these genes were only selectively deleted in certain silk gland cells, and the remaining normal cells could still maintain normal metabolic activity in the silk gland. Second, even if each target gene were completely deleted in all silk gland cells and the pathways involved were fully blocked, alternative pathways might compensate for the absence of the initial target, so the destabilizing network could be counterbalanced by compensatory mechanisms known as homoeostatic plasticity [34]. Third, hundreds of genes positively or negatively responded to the deletion of let-7 miRNA in the MSG or PSG, consistent with the likelihood that the abnormal phenotype resulting from the deletion or overexpression of let-7 is the comprehensive consequence of the derangement of the complex let-7 regulatory network, which may involve many genes and regulatory pathways. Therefore, loss of only one of its target genes could not be matched by the overexpression of let-7, and some unknown and/or undeleted target genes could compensate for the deletion of individual target genes. Collectively, our data support the conclusion that let-7 coordinates the carbohydrate metabolic processes and the energy signalling pathways in the MSG and PSG, through a regulatory network composed of its direct and/or indirect responsive genes, providing effective mechanisms for the overall growth of silk gland, and the synthesis and secretion of silk proteins.

Materials and methods

Silkworm rearing

A multivoltine, nondiapausing silkworm strain D9L, was used in this study. Larvae of silkworm were reared on fresh mulberry leaves under a 12h light/12 h dark photoperiodic regime at 25 ± 1°C.

RNA extraction and qRT-PCR validation

Experiments are described in more detail in Supplementary information, Methods. Each sample was measured in triplicate with three biological replicates.

sgRNA design, embryonic injection and screening of the positive individuals

sgRNAs were designed as described in Appendix Methods. Microinjection was performed as described in Supplementary information, Methods. The surviving G0 moths were sib-mated or crossed to wild-type individuals to generate G1 broods. Positive transgenic individuals were selected by screening the EGFP marker using fluorescence microscopy.

miRNA transgenic overexpression

The 190-bp sequence which contains pre-let-7, and the 545-bp sequence which contains pre-miR-100 and pre-miR-2795 together, were amplified by Ex Taq polymerase (Takara, Japan) using genomic DNA, and then were cloned into a pUC57-FibHp-SV40 plasmid using BamH I (Takara, Japan) and Not I (Takara, Japan) digestion. Then miRNA overexpression cassettes were digested by Asc I (NEB, USA) and inserted into a piggyBac transposon derived transgenic vector as described above. Microinjection and positive individual selection methods were as described above.

Microscopy and staining of PSG

Dissected PSG were fixed in 3.7% (vol/vol) formaldehyde overnight at 4°C and then kept at 0.1% (vol/vol) Triton X-100 in PBS for 1 h at room temperature. Samples were washed three times with PBS and kept at 3% (wt/vol) BSA in PBS for 30 min, then incubated with Alexa Fluor^TM 488 Phalloidin (Invitrogen, USA) diluted 1:200 in PBS for 1 h at 37°C followed by washing three times with PBS and incubation with DAPI (Beyotime, China) for 10 min at room temperature. Fluorescence imaging was performed using an Olympus DP80 microscope.

miRNA target prediction

Three different miRNA target prediction programmes – miRanda [35], TargetScan [36], RNAhybrid [37] were used to analyse putative targets. The 3ʹUTRs of silkworm genes were extracted using the information provided in NCBI.

Transcritome sequencing analysis

RNA-Seq experiments were carried out on Illumina HiSeq 2000 equipment (Appendix Methods). The raw data of sequencing were subjected to FastQC quality check, followed by trimming the adapters and removing the low-quality reads (Table S1). The high-quality clean reads were mapped to the reference genome of silkworm by the software HISAT2, and the totally mapped reads were assembled in expressed genes (Table S2 and Dataset S1).

Identification and knockout of the targtet genes

The targets were identified with double luciferase reporter gene assay (Appendix Methods). The three targets were knocked out with the CRISPR/Cas9 system (Appendix Methods).

DNA content assay and measure of ATP level

Total DNA in different sections of silk glands at D7 IL5 was extracted by using E.Z.N.A. Tissue DNA Kit (Omega Bio-tek, U.S.A.) and quantified by Nanodrop-2000C spectrophotometer (Thermo, U.S.A.) (Appendix Methods). ATP levels were measured by using the Enhanced ATP Assay Kit (Beyotime, China) (Appendix Methods).

Protein digestion and LC-MS/MS

The residual proteins in the P-MSGs of Δlet-7-MSG and Δlet-7C-MSG were identified by mass spectrometric analyses and the protein abundance was estimated by iBAQ algorithm (Appendix Methods).

Statistics and reproducibility

All statistical analyses were performed using GraphPad Prism 7.0 (San Diego, CA, U.S.A), and the values are presented as the means ± SEM. The mean cycle threshold (Ct) value was converted to relative expression level using the 2^−ΔΔCt method [38,39]. Statistical analyses of the expression levels were performed using a two-tailed un-paired student t-test. Significant differences are defined as * P < 0.05, **P < 0.01, ***P < 0.001.

Funding Statement

This work was supported by National Natural Science Foundation of China [31571334, 31071136, 31530071]; National Key Basic Research Project of China [2012CB114602]; and Fundamental and Frontier Research Project of Chongqing [cstc2014jcyjA00025].

Author contributions

S.L. conceived this project, performed the data analysis and wrote the manuscript. W.W. performed all expression and function experiments, and analyzed data. W.W, X.W, C.L. and Q.Y performed the target gene validation. W.W., X.W, Q.P., Q.Y, L.X. and X.P created all recombinant plasmids and bred the silkworm. S.L., S.M. and Q.X revised the manuscript. All authors revised and approved the final version of the manuscript.

Acknowledgments

We thank Professor Marian Goldsmith (University of Rhode Island Kingston, RI USA) for critical reading, kindly discussion and English use.

Data availability

The RNA-Seq data have been deposited into the National Center for Biotechnology Information Sequence Read Achieve database (accession no. PRJNA602978). https://www.ncbi.nlm.nih.gov/sra/PRJNA602978

Disclosure statement

No potential conflict of interest was reported by the authors.

Supplementary material

Supplemental data for this article can be accessed here.