Mutation of a Cuticle Protein Gene, BmCPG10, Is Responsible for Silkworm Non-Moulting in the 2nd Instar Mutant

In the silkworm, metamorphosis and moulting are regulated by ecdysone hormone and juvenile hormone. The subject in the present study is a silkworm mutant that does not moult in the 2nd instar (nm2). Genetic analysis indicated that the nm2 mutation is controlled by a recessive gene and is homozygous lethal. Based on positional cloning, nm2 was located in a region approximately 275 kb on the 5th linkage group by eleven SSR polymorphism markers. In this specific range, according to the transcriptional expression of thirteen genes and cloning, the relative expression level of the BmCPG10 gene that encodes a cuticle protein was lower than the expression level of the wild-type gene. Moreover, this gene’s structure differs from that of the wild-type gene: there is a deletion of 217 bp in its open reading frame, which resulted in a change in the protein it encoded. The BmCPG10 mRNA was detectable throughout silkworm development from the egg to the moth. This mRNA was low in the pre-moulting and moulting stages of each instar but was high in the gluttonous stage and in newly exuviated larvae. The BmCPG10 mRNA showed high expression levels in the epidermis, head and trachea, while the expression levels were low in the midgut, Malpighian tubule, prothoracic gland, haemolymph and ventral nerve cord. The ecdysone titre was determined by ELISA, and the results demonstrated that the ecdysone titre of nm2 larvae was lower than that of the wild-type larvae. The nm2 mutant could be rescued by feeding 20-hydroxyecdysone, cholesterol and 7—dehydrocholesterol (7dC), but the rescued nm2 only developed to the 4th instar and subsequently died. The moulting time of silkworms could be delayed by BmCPG10 RNAi. Thus, we speculated that the mutation of BmCPG10 was responsible for the silkworm mutant that did not moult in the 2nd instar.


Introduction
Metamorphosis and moulting is a phenomenon particular to insects, the growth and development of which includes periodic moulting. The moulting is regulated by many hormones, of which the most important are the ecdysone hormone and juvenile hormone. The exoskeleton is rebuilt by the degradation of old cuticle protein followed by the synthesis of new cuticle protein during insect moulting, and the signalling pathways of moulting hormone play an important role in this process [1][2][3]. Therefore, the cuticle protein genes are critical target genes of ecdysone, and their expression is closely related to the moulting and metamorphosis of insects. Recent studies have shown that the 1 st and 2 nd instars of silkworm are JH-independent phases in which JH does not have an important function [4].
Insect moulting involves the orderly expression of a series of genes, and the abnormalexpression of these genes might result in non-moulting. Several non-moulting silkworm mutants have been discovered [5,6], including non-moulting nm (11-13.8) [7], non-moulting dwarf nm-d (9-16.3) [8], and non-moulting Matsuno nm-m(10-27.9) [9]. The mutation gene responsible for the mutant known as non-moulting glossy nm-g (17-39.1) [10] silkworm Bombyx mori has been studied, and nm-g was identified as the key gene responsible for the mutant. This gene encodes a short chain dehydrogenase and takes part in the 'Black Box' for the synthesis of ecdysone hormone [11]. Most of these mutants were non-moulting in the 1 st instar. The silkworm mutant albino (al) is a lethal mutant with a colourless cuticle after the first ecdysis and dies without feeding on mulberry [12]. Sora Enya et al. used TALEN-mediated genome editing to generate a B. mori genetic mutant of nobo-Bm, which results in a 2 nd instar with a glossy cuticle that cannot undergo moulting and cannot develop into the 3 rd instar [13]. In addition, a mutation in GSTe7 results in the accumulation of 7-dehydrocholesterol.
We discovered a new mutant of non-moulting in the 2 nd instar (nm2) from the silkworm variety C603. The mutant develops normally in the 1 st instar and moults on time, but in the beginning of the pre-moulting stage of the 2 nd instar, the mutant larvae become lustrous, last for 6-8 days with hardly any development, cannot moult and exuviate and finally die. Genetic analysis revealed that nm2 was controlled by a single recessive gene [14]. In the present study, we mapped the candidate gene involved in the nm2 mutant by positional cloning and identified BmCPG10 as the gene most likely responsible for the nm2 mutant, BmCPG10 had a transcriptional expression level in nm2 that was lower than that of the wild-type gene. The BmCPG10 ORF of nm2 had a 217 bp deletion compared with the wild-type gene; thus, the structure of BmCPG10 and the protein encoded by BmCPG10 were altered. The titre of ecdysone in nm2 was lower than that of the wild-type, which resulted in the inability to moult in the 2 nd instar. These results might provoke further investigations into the molecular mechanisms of insect moulting and increase the understanding of the role of cuticle proteins in the moulting process of insects.

Silkworm strains
Silkworm strain p50 (standard silkworm strain), the wild-type C603 and nm2 mutant strain, were supplied by the Sericultural Research Institute (Zhenjiang, China). The larvae were reared on fresh mulberry leaves at (25 ±2)°C under a 12-h light/12-h dark photoperiod and 65 ±5% relative humidity. between p50 and P2 produced the F 1 offspring, and a positive and negative backcross (BC 1 ) between F 1 and P 2 produced the BC 1 F and BC 1 M. Eleven normal and ten mutant BC 1 F progenies from the same parents were used for the linkage analysis. The 594 mutants of BC 1 M were used for the recombination analysis.

Genomic DNA extraction
Genomic DNA of the parents (P 1 and P 2 ), F 1 individuals, and BC 1 F normal individuals were isolated from the silk gland, while the individual mutants of BC 1 F and BC 1 M were isolated from the whole 2 nd larvae without the midgut. DNA was extracted according to previously described methods [15]. The quality of the genomic DNA was determined at a 260/280 absorbance ratio, and the concentration of the genomic DNA was diluted to 100 ng/μL and stored at −20°C.

Linkage and recombination analysis
Simple sequence repeat (SSR) markers were obtained from the published SSR linkage map [16]. Polymorphisms of these SSR markers were detected by agarose gel electrophoresis. Eleven normal and ten mutant BC 1 F progenies from the same parents were used for the linkage analysis and preliminary SSR mapping. If the banding patterns of eleven normal BC 1 F progenies were inaccordance with the banding patterns of F 1 or P 2 without P1, and if the banding patterns of ten mutant BC 1 F progenies were inaccordance with the banding patterns of P 2 , then a linkage between nm2 and the SSR marker was indicated.
We downloaded the up-and down-stream sequences (http://sgp.dna.affrc.go.jp/) close to the tightly linked SSR marker of the nm2 locus and searched for new SSR markers using SSR Hunter 1.3 based on the results of the preliminary SSR mapping. Five hundred ninety four mutant BC 1 M individuals and new markers that exhibited polymorphism among the parents and F 1 individuals were used for fine mapping. The primers for the SSR markers with polymorphism are listed in S1 and S2 Tables.

Semi-quantitative and quantitative real-time PCR
Total RNAs were extracted from the pre-moulting silkworm of 2 nd instar of the wild-type and nm2 mutants, the larvaes at different developmental stages from eggs to moth, and various tissues including epidermis, fat body, midgut, silk gland, malpighian tubule, trachea, head, hemocytes, testis, ovary, muscle, prothoracic gland and ventral nerve cord from the third day in the fifth instar larvae of wild-type using RNAiso Plus (TaKaRa). The quality of the total RNAs was determined at a 260/280 absorbance ratio, as well as with the electrophoresis method, after which the RNA was stored at −80°C. After treatment with DNaseI, 1 μg of the total RNA was used to synthesize the first strand cDNA using a Primerscript Reverse Transcriptase kit (TaKaRa) according to the manufacturer's protocol. The primers of the 13 candidate genes for semi-quantitative reverse-transcription polymerase chain reaction (RT-PCR) were listed in S3 Table. The silkworm housekeeping gene Bm-actin A3 was used as an internal control for normalization of sample loading.
qRT-PCR was performed using an ABI PRISM 1 7300 Sequence Detection System (Applied Biosystems, USA). BmCPG10 mRNA and Bm-actin A3 mRNA were quantified using 2 μL of the reverse transcription reaction (equivalent to 100 ng single stranded cDNA) as a template in qRT-PCR. A 289 bp product for BmCPG10 cDNA was amplified using the following primers: forward 5-GGCGTCTATTGGTGATGGTGATAAC -3 and reverse 5-GAGTCCAAAGAAC AAGGTTCGCTTC -3. qRT-PCR was performed in a 20 μL reaction mixure using SYBR Green Supermix (Takara), according to the manufacturer's instructions. The thermal cycling profile consisted of initial denaturation at 94°C for 5 min; and 40 cycles at 94°C for 30 s, 60°C for 25 s, and 72°C for 35 s. All of the reactions were performed in triplicate, and the relative expression levels were analysed using the 2 −ΔΔCt method, where ΔΔCt = ΔCt sample − ΔC treference , and Ct refers to the cycle threshold [17]. The silkworm housekeeping gene Bm-actin A3 was used as an internal control for the normalization of sample loading.

RNAi reduction BmCPG10 mRNA experiment
The 373 bp fragment of BmCPG10 cDNA was amplified using the wild-type silkworm cDNA as a template for RNA synthesis with the following primers containing the promoter of T7: forward: 5-TAATACGACTCACTATAGGGAAGGGACATCTTGAACC-3, reverse: 5-TAATA CGACTCACTATAGGCGCAAAGTCACAGGAAAC-3. The dsRNA was synthesized using a MEGA script RNAi kit (Ambion). RNA synthesis and purification were performed according to the manufacturer's instructions, and the integrity of dsRNA was confirmed by nondenaturing agarose gel electrophoresis. The quality of the dsRNA was determined at a 260/280 absorbance ratio, and the dsRNA was diluted to a final concentration of 2000 ng/μL with an injection buffer. EGFP dsRNA was synthesized using the method of Quan et al [18].

Annotation analysis
The candidate genes in the region narrowed by the linkage analysis were annotated using the silkworm genome database(http://sgp.dna.affrc.go.jp/KAIKObase) and BLASTX from the National Center for Biotechnology Information (http://blast.ncbi.nlm.nih.gov/Blast.cgi).

Assay of ecdysteroid titres
The silkworms that were collected above were frozen and homogenized in 50% MeOH (600μL). The homogenate was centrifuged, and the supernatant was stored at -20°C. Ecdysteroid titres were assayed using an Insect Ecdysone ELISA Kit (Shanghai Meilian Bio Technology Co., Ltd.) according to the manufacturer's instructions. The quality of the protein was determined at a 280 absorbance ratio, and the concentration of the protein was diluted to 1 mg/mL and stored at −20°C.

Feeding experiments
After the separation of the non-moulting larvae during 2 nd instar, we fed mulberry leaves supplemented with 400 mg/L 20-hydroxyecdysone (20E), 8000 mg/L cholesterol and 7-dehydrocholesterol (7dC) to nm2. Supplementation with water was used as a control. The specific concentrations of 20E, cholesterol and 7dC were dependent on the preliminary feeding experiment that listed in the S4 Table. For each group, including 50 nm2 mutants, the experiment was repeated at least three times.

Results
The nm2 gene was located on the 5 th linkage group As chromatid exchanges do not occur in female silkworms and the nm2 mutation is lethal, we used the +/nm2 instead of nm2/nm2 as P 2 to establish the positional cloning group. The genotype of F 1 generation was +/nm2 and +/+, so there was no mutant separated from the F 1 generation. There was no mutant separated from the group of BC 1 F and BC 1 M when the genotype of F 1 generation was +/+, so this group of BC 1 F and BC 1 M were not used for positional cloning. And the BC 1 F and BC 1 M that had mutant separated from were used for positional cloning. Polymorphic markers on each linkage group of silkworm were found (listed in S1 Table), These polymorphic markers as well as 11 normal and 10 nm2 BC 1 F progenies that from the same parents were used to determine the linkage group on which the nm2 gene was located. According to Mendel genetic law, if a polymorphic marker was linked with the nm2 gene, then the banding patterns of 11 normal BC 1 F progenies were consistent with the F 1 or P 2 individuals, and the banding patterns of 10 nm2 BC 1 F progenies from the same parents were consistent with the P 2 individuals.
The results of marker S2529-2 indicated that 11 normal BC 1 F progenies exhibited a banding pattern similar to that of F 1 or P 2 , whereas 10 nm2 BC 1 F progenies exhibited a banding pattern similar to that of P 2 (Fig 1A). The results of other polymorphic markers that we found and located on the other linkage group indicated that11 normal and 10 nm2 BC 1 F progenies all exhibited a banding pattern similar to that of F 1 or P 2 . Thus, the nm2 gene was located on the 5 th linkage group, and the marker S2529-2 was linked to the nm2 gene. Based on the linkage analysis, eleven SSR markers were developed on the 5 th chromosome, including S2529-43, S2529-45, S2529-46, S2529-32, S2529-35, S2529-27, S2529-91, S2529-2, S2529-121, S2529-74 and S2529-104 (S2 Table), and 594 nm2 BC 1 M individuals were used for fine mapping. The results indicated that nm2 was located in a region of approximately 275 kb on the Bm_ nscaf20 from 2496529 bp to 2772188 bp compared with the silkworm genome database KAIKObase. Thirteen genes within the particular region were predicted using gene-prediction models, including BMgn002598, BMgn002690, BMgn015078, BMgn002599, BMgn002689, BMgn002600, BMgn002688, BMgn002601, BMgn002602, BMgn002603, BMgn002687, BMgn015079, and BMgn002685 (Fig 1B). BmCPG10 is the nm2 key gene Because nm2 was non-moulting in the 2 nd instar, we analysed the expression profiles of the thirteen initial candidate genes of 2 nd instar pre-moulting silkworm between nm2 and the wild-type by semi-quantitative RT-PCR (Fig 2A). Only the gene BMgn002598 was not expressed in both nm2 and wild-type. The gene BMgn002601 was not expressed in wild-type but was expressed in nm2, while the gene BMgn002602 was not expressed in nm2 but was expressed in wild-type. There were no differences in the expression of the other genes.
We obtained the open reading frame (ORF) of BMgn002601 and BMgn002602 from the nm2 mutant and wild-type strains using primers designed for the ORF of BMgn002601 and BMgn002602. We found that the BMgn002601 ORF of nm2 and of the wild-type was identical, but the BMgn002602 ORF in nm2 involved a deletion of 217 bp and resulted in a difference in the amino acid sequence of nm2 compared with the wild-type (Fig 2). BMgn002601 encodes a hypothetical protein 35 with a function that is unclear. BMgn002602 encodes a cuticle protein, BmCPG10 [19], rich in glycine according to the silkworm database (http://sgp.dna.affrc.go.jp/ KAIKObase) and GenBank of NCBI(http://blast.ncbi.nlm.nih.gov/Blast.cgi). The 217 bp deletion of BMgn002602 was located in the region that rich in glycine, which might destroy the functional domain of the BMgn002602. The deletion resulted in the backward of the termination codon, and the mutated gene was predicted to encode a new protein.
To determine whether nm2 was caused by decreasing the amount of BmCPG10 mRNA, we synthesized dsRNA corresponding to a portion of the BmCPG10 ORF and injected it into the wild-type larvae during the gluttonous stage of 2 nd instar at doses of 2000 ng per individual; we injected the same doses of EGFP dsRNA as a control. Four of the 10 individuals injected with BmCPG10 dsRNA were delayed in their moulting time during the 2 nd instar, while the controls moulted24 h after the injection (Fig 3A). The moulting time of 4 individuals injected with BmCPG10 dsRNA in the 2 nd instar could be delayed 48 h to 72 h. Using qRT-PCR, we compared the expression of BmCPG10 and BMgn002601 mRNA among individuals injected with BmCPG10 and EGFP dsRNA. The expression level of BmCPG10 mRNA in the 4 individuals injected with BmCPG10 dsRNA was significantly lower than that of the controls (Fig 3B) and the expression level of BMgn002601 mRNA in the 4 individuals injected with BmCPG10 dsRNA was significantly higher than that of the controls (Fig 3C). These results showed that the expression level of BMgn002601 mRNA could be influenced by the decreasing amount of BmCPG10 mRNA. Thus the expression of BMgn002601 mRNA in nm2 might be caused by the mutantion of BmCPG10.
We therefore concluded that the BmCPG10 gene was the most likely candidate for the nm2 mutant.

The developmental and tissue-specific transcription pattern of BmCPG10
The expression profiles of BmCPG10 at various developmental stages of wild-type were examined using the entire body, except for the midgut of the larvae by qRT-PCR. The results BmCPG10 Is Responsible for Silkworm Non-Moulting in the 2 nd Instar Mutant indicated that BmCPG10 mRNA was detectable throughout the development, from the egg to the moth. The BmCPG10 mRNA expression was low during the pre-moulting and moulting of each instar, and was highest during the gluttonous stage of 3 rd instar and was lowest during the pre-moulting of 2 nd instar (Fig 4A). To determine the tissue specificity of BmCPG10 mRNA expression in the larvae, total RNA from the third day of the fifth instar larvae of C603 was isolated from the epidermis, fat body, midgut, silk gland, Malpighian tubule, trachea, head, haemocytes, testis, ovary, muscle, prothoracic gland and ventral nerve cord and then was BmCPG10 Is Responsible for Silkworm Non-Moulting in the 2 nd Instar Mutant subjected to qRT-PCR. The results showed that the BmCPG10 gene transcription could be detected in most of the tissues. The epidermis, head and trachea had higher levels, while the midgut, Malpighian tubule, prothoracic gland and ventral nerve cord had lower levels, however the BmCPG10 gene transcript was barely detected in the haemocytes (Fig 4B).
The low titre of ecdysone in nm2 resulted in the mutant phenotype Based on the facts that the BmCPG10 encoded a cuticle protein and that the expression of most cuticle protein genes were induced by an ecdysone pulse, the expression of the cuticle protein genes required the presence and elimination of 20E [20][21][22][23][24][25]. We speculated that the mutation of BmCPG10 would have some influences on the ecdysone titre. Thus the titre of ecdysone in nm2 and wild-type were determined by ELISA. The result indicated that ecdysone titre in nm2 was significantly lower than that of the wild-type (Fig 5).
The nm2 mutant could be rescued by feeding 20E, 7dC and cholesterol Based on the difference of in the ecdysone titre between the wild-type and nm2, we speculated that the synthesis of ecdysone was affected by the mutation of BmCPG10. Cholesterol, 7dC and 20E, which are the most upstream materials, as well as the intermediate and end products for ecdysteroid biosynthesis, respectively, dissolved or suspended in water and were fed to each group, including 50 nm2; water was used as a control. As a result, 96% of the 50 nm2 that were fed 8000 mg/L cholesterol, 40% of 50 nm2 that were fed 8000 mg/L 7dC and 98% nm2 that were fed 400 mg/L 20E moulted in the 2 nd instar and grew into the 3 rd instar stage but developed slowly in the third instar, with different sizes. These were then fed fresh mulberry leaves, and the partially rescued nm2 then moulted and grew into the 4 th instar stage but could not moult in the 4 th instar and finally died (Fig 6). However, in controls that were given water, none of the all 50 nm2 individuals could not moult in the 2 nd instar and subsequently died ( Table 1). These results showed that the cholesterol that provided the raw materials for ecdysteroid biosynthesis was deficient in nm2, which was caused by the mutation of BmCPG10.

Discussion
In the present study, we fine mapped the candidate genes of the nm2 mutant by positional cloning and identified the function to illustrate the nm2 mutation at the molecular level. We focused on the two genes that had different mRNA expression between the wild type and nm2. One was BMgn002601 and the other was BmCPG10. And the key gene responsible for the nm2 was further determined by cloning of their ORF and the RNAi of BmCPG10. The ORF of BMgn002601 was identical between wild type and nm2, but the BmCPG10 ORF in nm2 involved a deletion of 217 bp in its functional domain, in addition the mRNA expression of BMgn002601 could be influenced by the expression of BmCPG10. Based on these results, we concluded that the mutation of BmCPG10 was the most likely key gene responsible for the non-moulting in the 2 nd instar silkworm mutant.
There are various types of structural cuticle protein genes in insect genomes. Currently, in all of the sequenced insects, the proportion of cuticle protein genes in an insect's total estimated protein-coding genes is more than 1% [19]. Insect cuticle, as a main component of exoskeleton [19,26,27], is very important for insect growth, reproduction, and adaptivity to the complex and changeable environment. The BmCPG10 gene encodes a cuticle protein that is rich in   [28]. The repeat sequences of glycine, including GXGX、GGXG or GGGX, can form a soft curling structure [29], which is very important for the crosslinking of cuticle protein during hardening [30]. In the present study, the 217 bp deletion of BmCPG10 was located in the functional domain that rich in glycine and include the repeat sequences of glycine, which might influence the crosslinking of cuticle protein during hardening. Okamoto et al. investigated the expression of 17 CPG genes during the last larval moult of Bombyx mori, and the result showed that all of the genes were abundantly expressed in the epidermis but were barely expressed in the fat body, haemocytes and midgut, suggesting that they were mainly involved in the construction of the newly synthesized cuticle of the epidermis [31]. The tissue expression of the BmCPG10 gene in this study was similar to that of the Okamoto et al study, the BmCPG10 gene transcription had high level in the epidermis, head and trachea, which suggested that the BmCPG10 might participate in the synthesis and construction of these organs. However, BmCPG10 mRNA was also expressed in midgut, Malpighian tubule, prothoracic gland and ventral nerve cord in the present study, which showed that BmCPG10 had other functions in addition to synthesizing the cuticle of the epidermis. The structure and composition of insect cuticle are renewed during moulting and metamorphosis, and the cuticle protein coding genes are regulated by Juvenile hormone and ecdysteroid because the cuticle proteins are the important part of the epidermis [32]. Many cuticle protein genes were reported to be up-regulated by 20E [33][34][35][36][37]. Okamoto studied the expression profiles of 17 cuticle protein genes at the 4 th moult; the results indicated that most of the cuticle protein genes tested were highly expressed when the ecdysone titre decreased or disappeared but were not expressed when the titre of ecdysone was high. Thus, the expression of cuticle protein genes was negatively correlated with the titre of ecdysone [31]. The expression of BmCPG10 mRNA showed a similar tendency to the results of Okamoto. Because of the structural variation of BmCPG10 in nm2, the mRNA expression and the encoded protein of BmCPG10 were affected, and the titre of ecdysone was lower in nm2 than that of in wild-type, which resulted in non-moulting. In the rescued experiment, the nm2 mutant could be rescued by feeding 20E, cholesterol, 7dC; because cholesterol was slightly soluble in water, but 7dC was not dissolve in water and the different solubility might have some influence on the absorption and digestion of silkworm, so the rescued rate of 7dC was lower than that of the cholesterol. All of these results gave us a hint that the BmCPG10 was closely related to the titre of ecdysone in silkworm.
Sora Enya reported a novel Halloween gene, noppera-bo (nobo), that encoded a member of the glutathione S-transferase family. The larvae that were knock-downs of the nobo gene displayed an arrested phenotype and a reduced 20E titre and could be rescued by the administration of 20E or cholesterol. The results presented the possibility that nobo played a crucial role in regulating the behaviour of cholesterol in the biosynthesis of steroids in insects [38]. The silkworm mutant nobo-Bm was non-moulting in the 2 nd instar and had a glossy cuticle phenotype, which could be rescued by the application of 20E, suggesting that the nobo family is essential for the regulation of sterol utilization in the Lepidoptera [13]. The silkworm mutant nm2 in this study was very similar to nobo-Bm and could be rescued by feeding 20E and cholesterol, which might indicate that the mutation of BmCPG10 had an influence on sterol utilization and resulted in acholesterol deficiency, which was the raw material required for the synthesis of ecdysone that were synthesized from dietary sterols. However, the specific molecular mechanism of BmCPG10 warrants further research. Supporting Information S1