- Split View
-
Views
-
Cite
Cite
Gao Xiong, Xiaoling Tong, Tingting Gai, Chunlin Li, Liang Qiao, Antónia Monteiro, Hai Hu, Minjin Han, Xin Ding, Songyuan Wu, Zhonghuai Xiang, Cheng Lu, Fangyin Dai, Body Shape and Coloration of Silkworm Larvae Are Influenced by a Novel Cuticular Protein, Genetics, Volume 207, Issue 3, 1 November 2017, Pages 1053–1066, https://doi.org/10.1534/genetics.117.300300
- Share Icon Share
Abstract
Body shape and color patterns of insect larvae are fundamental traits for survival. Typically, transcription factors or members of signaling pathways...
The genetic basis of body shape and coloration patterns on caterpillars is often assumed to be regulated separately, but it is possible that common molecules affect both types of trait simultaneously. Here we examine the genetic basis of a spontaneous cuticle defect in silkworm, where larvae exhibit a bamboo-like body shape and decreased pigmentation. We performed linkage mapping and mutation screening to determine the gene product that affects body shape and coloration simultaneously. In these mutant larvae we identified a null mutation in BmorCPH24, a gene encoding a cuticular protein with low complexity sequence. Spatiotemporal expression analyses showed that BmorCPH24 is expressed in the larval epidermis postecdysis. RNAi-mediated knockdown and CRISPR/Cas9-mediated knockout of BmorCPH24 produced the abnormal body shape and the inhibited pigment typical of the mutant phenotype. In addition, our results showed that BmorCPH24 may be involved in the synthesis of endocuticle and its disruption-induced apoptosis of epidermal cells that accompanied the reduced expression of R&R-type larval cuticle protein genes and pigmentation gene Wnt1. Strikingly, BmorCPH24, a fast-evolving gene, has evolved a new function responsible for the assembly of silkworm larval cuticle and has evolved to be an indispensable factor maintaining the larval body shape and its coloration pattern. This is the first study to identify a molecule whose pleiotropic function affects the development of body shape and color patterns in insect larvae.
INSECTS exhibit body shape and color patterns in the adult and larval stages to either disguise them against the background, and allow them to approach unsuspecting prey (O’Hanlon et al. 2014), or to protect themselves from their own predators (Joron et al. 2011; Dasmahapatra et al. 2012; Hossie et al. 2015; Prudic et al. 2015). Larval adaptations can be especially intriguing. For example, the dorsally bulging segments of the caterpillar of the mother of pearl moth, Patania ruralis, facilitate its high-speed escape method, which involves the larva rolling away from predators in a doughnut shape, where the dorsal side of each segment stretches to produce a perfectly smooth curved shape (Brackenbury 1997). On the other hand, the eyespot color markings of some large lepidopteran larvae make them less susceptible to be attacked (Hossie et al. 2015). Thus, the high diversity of both body shape and color patterns make insect larvae outstanding models for studies of adaptation and of its underlying molecular basis (Ronshaugen et al. 2002; Suzuki and Nijhout 2006; Futahashi and Fujiwara 2008; Shirataki et al. 2010).
One of the key structures that determines body shape and color patterns in insect larvae is the cuticle. The insect cuticle is an extracellular matrix (ECM) covering the entire body. It generally consists of chitin fibers, structural cuticular proteins (CPs), and some catecholamines and lipids (Moussian 2010). Previous reports showed that disruption of chitin fiber synthesis, decomposition, or modification in Drosophila melanogaster (Moussian et al. 2005, 2006) and Tribolium castaneum (Zhu et al. 2008; Chaudhari et al. 2011, 2013), and mutations of CPs in Bombyx mori (Qiao et al. 2014), led to defects in cuticle assembly and ultimately caused abnormal body shapes. On the other hand, the diversity of larval color patterns depends on the content and distribution of pigments in the cuticle, which are regulated by transcription factors and pigment biosynthesis pathway genes expressed in the underlying epidermis (Shirataki et al. 2010; Osanai-Futahashi et al. 2012; Yoda et al. 2014). Although these studies have advanced our understanding of the diversity of body shapes and color patterns, they focused on the formation of each trait independently. Since the two larval traits coexist in the ECM, it is possible that a common factor may cocontrol body shape and color pattern simultaneously. The molecular characterization of such factors, however, remains elusive.
In this work, we characterize a mutation in the silkworm, B. mori, named Bamboo (Bo, chromosome 11–28.8 cM) that causes variation in both body shape and pigmentation (Kanekatsu et al. 1988). Using positional cloning followed by functional validation, we identified a low complexity CP gene, BmorCPH24, that is responsible for the Bo phenotype. In addition, our findings indicate that novel or fast-evolving genes can have key functions in development of larval body shape and color patterns.
Materials and Methods
Insects
The wild-type domesticated silkworm strains (Dazao, +Bo/+Bo from Bo/+Bo inbreeding), Bo mutant strains (Bo/+Bo, Bo/Bo) and other strains without Bo phenotype, were obtained from the Silkworm Gene Bank of Southwest University, China. B. mandarina, a wild relative of B. mori, was collected in a field in Chongqing, China. All domesticated silkworm strains were reared on mulberry leaves under a 12 hr/12 hr light/dark photoperiod at 24°.
Positional cloning
The Bo mutant and Dazao strains were used as the parental strains for genetic mapping. F1 heterozygous males were produced from a cross between a female Bo and a male Dazao. One thousand and forty-five BC1M progeny from the backcross Dazao ♀× (Bo × Dazao) ♂ were used for recombination analysis. For fine mapping, we used nine polymorphic PCR markers identified among the parents, and these were assessed in BC1M individuals. Primers used for genotyping are listed in Supplemental Material, Table S1 in File S1.
Complementary DNA (cDNA) synthesis from the larval epidermis and other organs
Various silkworm tissues were dissected from larvae into cold phosphate-buffered saline (PBS). They included the ovary, testis, head, anterior silk gland, middle silk gland, posterior silk gland, epidermis, midgut, fat body, malpighian tubule, hemocyte, trachea, ventral nerve cord, and wing discs. The tissues were powdered in liquid nitrogen, and total RNA was extracted and purified using TRIzol (Invitrogen, Carlsbad, CA) according to the manufacturer’s protocol. First cDNA was generated by reverse transcription using the PrimeScript RT reagent Kit with gDNA Eraser according to the supplier’s instructions and was used for quantitative and qualitative RT-PCR analyses.
Quantitative RT-PCR
To quantify the expression of BmorCPH24 and other genes, we used the Bio-rad CFX96 sequence detection system with an iTaqSYBRGreen (Bio-rad) according to the manufacturer’s protocol. The primers designed for amplification are listed in Table S2 in File S1. The B. mori ribosomal protein L3 (rpL3, GenBank: AY769270.1) was used as the internal control. All assessments were performed on three or four biological replicates per sample. Gene expression levels were normalized by those of a random sample, which were set to 1.0. Statistical comparisons were made by Student’s t-test or paired Student’s t-test.
Identifying the Bo mutation site
Primers were designed based on the sequence of BGIBMGA011765 and BGIBMGA011766 mRNA and are shown in Table S1 in File S1. Total RNA was isolated from the wild-type and Bo strains using TRIzol reagent (Invitrogen), and cDNA libraries were established following the instructions of the Reverse Transcript Kit (No. RR037A, Takara). PCR products were cloned into a PMD19-T vector (Takara) and sequenced. For multi-strain analyses, PCR was performed on genomic DNA, and the amplified PCR product was sequenced.
In situ hybridization
The digoxigenin-labeled RNA probes for antisense and sense strands of BmorCPH24 were synthesized using the T7 RiboMAXTM Express RNAi System (Promega, Madison, WI). A whole dorsal epidermis from the sixth to the ninth segment of day 2, fifth instar larvae were carefully dissected to protect the epidermal cells from falling off. The muscles and fat body were retained on the epidermis until the staining was removed. The materials were then fixed immediately in 4% paraformaldehyde in PBS. In situ hybridization was performed according to a previously described protocol (Futahashi and Fujiwara 2005).
RNAi experiment
The double-stranded RNA (dsRNA) targeting BmorCPH24 and red fluorescence protein (dsRed, the negative control) were synthesized using the RiboMAX Large Scale RNA Production System T7 kit (Promega). Length of the two fragments was 173 and 495 bp, respectively. The synthesized dsRNA was injected into the left side of U strain larvae at the beginning and the second day of fourth instar larval stage at the dose of 80 µg per individual. Then we placed PBS, the electric buffer, on the injected side (left) and the control side (right), and shocked the injected region immediately with a 20-V electric voltage. The phenotype was observed in fifth instar larvae. Thereafter, we determined the expression level of BmorCPH24 in the larvae by qRT-PCR. The electroporation technique was performed according to a previously described protocol (Ando and Fujiwara 2013).
CRISPR/Cas9-mediated gene knockout
The small guide RNA (sgRNA) for knocking out BmorCPH24 was synthesized using the RiboMAXTM Large Scale RNA Production System T7 kit (Promega) and Cas9 mRNA was synthesized using the mMESSAGE mMACHINE T7 kit (Ambion). The sgRNA and Cas9-encoding mRNA were mixed at a dose of 400 ng/µl, and the mixture was injected into newly laid eggs. The injected individuals were reared to moths and mated with each other randomly to produce the G1 generation. Knockout phenotypes were examined in the resulting larvae. To identify the mutation site and type of larvae edited by CRISPR/Cas9, we used the appropriate primer sets to clone and sequence the genomic DNA.
Construction of piggyBac-based transgenic vectors
We first designed a set of primers to amplify BmorCPH24 with a BamHI recognition site in the sense and a NotI recognition site in the antisense primers, respectively. Using the cDNA of wild-type as template, the ORF of BmorCPH24 was amplified and cloned into the pMD19-T vector. Then BamHI and NotI were used to cleave pMD19- BmorCPH24 and pSL [IE1-GFP-SV40] to recover BmorCPH24 and pSL [IE1-SV40], which were ligated to yield the resulting pSL [IE1- BmorCPH24-SV40]. Subsequently, both pSL [IE1- BmorCPH24-SV40] and piggyBac target vector (pPIGA3EGFP) were cleaved with AscI and BglII, respectively. Then the IE1- BmorCPH24-SV40 cassette was cloned into pPIGA3EGFP to generate piggyBac [A3EGFP, IE1-BmorCPH24-SV40] (Figure S6Aa in File S1).
Transgene expression by piggyBac
The procedure reported by Ando and Fujiwara (2013) was performed to inject the transgenic vector into larvae. First, the overexpression vector, piggyBac [A3EGFP, IE1- BmorCPH24-SV40], and the helper plasmid with piggyBac transposase were mixed in a 1:1 molar ratio to make a final concentration of 1.4 µg/µl. Then, we injected 5 µl of the mixture into each N4 individual on day 2, fourth instar by capillary injection. After that, we immediately placed PBS, the electric buffer, on the injected side (left) and the uninjected side (right), and shocked the injected region with 20 V electric voltage. Fluorescence of enhanced green fluorescent protein (EGFP) was screened at the fifth instar larval stage.
Staining and light microscopy
The dorsal epidermis of fifth instar larvae was carefully dissected in cold PBS and then fixed in 4% formaldehyde/PBS for 30–60 min at room temperature. Cells were stained by DAPI (Beyotime; 10 μg/ml), and chitin was stained by Fluostain (wheat germ agglutinin-FITC labeled, Sigma, St. Louis, MO; 10 μg/ml). The samples were visualized under a fluorescence microscope (Olympus BX63).
Measurement of caspase activity
The dorsal epidermis of fourth and fifth instar larvae was carefully dissected in cold PBS and then ground in a mortar. The powders were resuspended in 500 μl lysis buffer (20 mM Tris pH 7.5,150 mM NaCl, 1% Triton X-100, 0.1 mM PMSF, Beyotime) and placed on ice for 60 min. Then they were centrifuged at 10,000 × g for 10 min at 4° and the supernatant was saved. Caspase activity was determined using the Caspase-Glo 9, 3/7 assay systems (Promega). The samples were incubated with 100 μl Caspase-Glo 9 (3/7) reagent for 3 hr at room temperature in the dark and the fluorescence was measured using a microplate reader.
Phylogenetic analysis
To understand the evolutionary history of BmorCPH24, we performed phylogenetic analysis using the genes encoding cuticular protein with low complexity sequence from four different insect species, B. mori (Futahashi et al. 2008), Dendrolimus punctatus (Yang et al. 2017), Manduca sexta (Dittmer et al. 2015), and Anopheles gambiae (Cornman and Willis 2009). All proteins were aligned using Muscle. The phylogenetic tree was constructed using maximum likelihood with the MEGA6 software. The confidence levels for various phylogenetic clades were assessed by bootstrap analysis (1000 replicates). To determine whether BmorCPH24 orthologous existed in other species, we additionally used BmorCPH24’s sequence as a query to perform BLASTP and TBLASN searches against the National Center for Biotechnology Information (NCBI) (http://www.ncbi.nlm.nih.gov/protein), Manduca base (https://i5k.nal.usda.gov/Manduca_sexta), Spodobase (http://bioweb.ensam.inra.fr/Spodobase/), ChiloDB (http://ento.njau.edu.cn/ChiloDB/), Heliconius Genome Project (http://www.butterflygenome.org/), and DBM-DB (http://iae.fafu.edu.cn/DBM/).
Data availability
Strains are available upon request. The authors state that all data necessary for confirming the conclusions presented in the article are represented fully within the article.
Results
Bo mutant has epidermal defects
Bo is a dominant mutation, which was first identified by Kanekatsu in 1988 (Kanekatsu et al. 1988). Larvae of Bo exhibit a bamboo-like body shape with a dilated thorax and a slim but hard abdomen (Figure 1). As in the plant, the segmental boundaries bulge outward from the body, and the main area of each segment is narrow and constricted (Figure 1, A and B). Expansion of the dorsal cuticle is limited upon feeding and the last abdominal segments of the fifth instar larvae usually curl up after the larvae are satiated with mulberry leaves (Figure 1B). Bo mutants also display reduced markings and pigmentation on their cuticle (Figure 1C. orange ovals). The introduction of the Bo mutation, via genetic crosses, into two other strains, L and U, with large amounts of pigmentation on their larval body, led to reduced pigmentation in these strains (Figure 1D). This suggests that the Bo mutation has a general repressive effect on pigmentation.
To examine how Bo affected the shape of each larval body segment, we analyzed the length/width (L-W) ratios of the cuticle of each segment in wild-type and Bo mutants. In the nomenclature of silkworm larvae, the internode is the main portion of each segment and is the region with pigmentation, while the intersegmental fold is the minor, posterior portion of each segment, which is soft and has no pigmentation (Figure 1E). In wild-type larvae, the L-W ratio of the internode is significantly higher than the L-W ratio of the intersegmental fold (Figure 1F). In contrast, the L-W ratio of Bo mutants was variable, with the L-W ratio of the internode being comparable, or even lower, than that of the intersegmental fold (Figure 1F). Together, the reduction of the L-W ratio in the internode and its increase in the intersegmental fold may lead to a constricted segment and a bulge in the posterior segmental boundaries, simultaneously accompanied by the reduced coloration (Figure 1, B–F). These results indicate that the phenotype of the Bo mutant could result from defects in the larval cuticle during development.
The Bo locus is a low-complexity cuticular protein gene named BmorCPH24
To identify the gene responsible for the Bo phenotype, we performed positional cloning using the BC1M progeny. Previous linkage analyses mapped the Bo locus to the 28.8 cM in the 11th chromosome (Oota and Kanekatsu 1993). This position is 1.4 cM away from the locus of another mutation, dimolting (mod, chromosome 11–27.4 cM), which is a mutation in the cytochrome P450 gene, CYP15C1 (BGIBMGA011708, Bm_nscaf3031: 935905 bp..927663 bp) (Daimon et al. 2012). In light of this, we developed nine polymorphic markers (Table S1 in File S1) on Bm_nscaf3031 and Bm_nscaf3034 for finer scale mapping of Bo. By genotyping these markers in 1045 BC1M individuals, we pinpointed the Bo locus to a 143-kb genomic region between marker “d” and “h,” which contains 15 predicted genes (Figure 2A). We designed 13 specific primer pairs for qRT-PCR (Table S2 in File S1) to perform gene expression analyses in the fifth instar Bo and wild-type larvae (Dazao, +Bo/+Bo). Three of the predicted genes, BGIBMGA011715, BGIBMGA011716, and BGIBMGA011717, shared the same cDNA sequence and could not be differentiated. The results showed that a single gene, BGIBMGA011765, was significantly down-regulated in Bo (Figure 2B and Figure S1 in File S1). Meanwhile, cloning and sequence comparisons of 15 genes revealed that a deletion and an insertion were identified in the coding region of BGIBMGA011765 and BGIBMGA011766 in the Bo mutant, respectively. In the Bo mutant, a 48-bp nucleotide sequence was inserted into BGIBMGA011766, and BGIBMGA011765 carried a 5-bp deletion that resulted in a frame shift mutation indicating the loss-of-function of BGIBMGA011765 in Bo mutants (Figure 2C and Figure S3A in File S1). Further, multi-strain comparisons showed that the 5-bp deletion in the BGIBMGA011765 gene was unique to Bo and was not found in 33 non-Bo phenotype silkworm strains (Figure S2 in File S1). However, no Bo mutant-specific mutation was found in the BGIBMGA011766 gene (Figure S3B in File S1). These results collectively suggest that BGIBMGA011765 is the most likely gene responsible for the Bo mutant phenotype. BGIBMGA011765 is a low-complexity CP gene without the typical CP domain, previously named BmorCPH24 (Okamoto et al. 2008; Liang et al. 2010).
BmorCPH24 is mainly expressed in the larval epidermis
To better understand the spatiotemporal expression pattern of BmorCPH24 in wild-type individuals, qRT-PCR was conducted across 14 different tissues in larvae. Expression profiles from day 3, fifth instar larvae revealed high expression of BmorCPH24 in the epidermis but not in the other tissues (Figure 3A). BmorCPH24 epidermal expression peaked during the feeding phase of each instar, then decreased rapidly in the molting phase, and declined to almost zero during the larval wandering, pupal, and adult stages (Figure 3B). This expression pattern suggests an important role for BmorCPH24 in larval cuticle assembly.
BmorCPH24 expression is associated with changes in body shape and coloration
To further elucidate the molecular mechanism of Bo, we tested whether the expression of BmorCPH24 correlated with finer scale variation in body shape and pigmentation markings in Bo and wild-type (+Bo/+Bo) larvae. We first investigated the expression of BmorCPH24 in the larval markings using qRT-PCR. BmorCPH24 was highly expressed in the pigmented areas of the fifth dorsal segment of wild-type larvae, both at the fourth and fifth instar larval stages (Figure 4A). These results suggest that BmorCPH24 may be required for the development of larval markings. Furthermore, in situ hybridizations showed that BmorCPH24 was mainly expressed at the internode (in) of each segment and was below detection level at the posterior margin of each segment, in the intersegmental fold (in-f) (Figure 4B). The overall expression level of BmorCPH24 in Bo was also lower than in +Bo/+Bo (Figure 4B). This expression pattern perfectly coincided with the shorter internodes and longer intersegmental folds of Bo mutants. qRT-PCR also revealed that BmorCPH24 was expressed at 11-fold higher levels in the dorsal vs. ventral part of the epidermis (Figure 4C), and at gradually higher levels toward the posterior end of the larval body (Figure 4C). In all cases, levels of BmorCPH24 were significantly lower in Bo than in +Bo/+Bo (Figure 4C). Thus, these results indicate a strong association between the expression pattern of BmorCPH24 and the Bo phenotype.
BmorCPH24 produces the Bo phenotype in silkworm
To test whether BmorCPH24 caused the Bo phenotype, we used a previously established functional tool, electroporation-mediated RNAi (Ando and Fujiwara 2013; Yamaguchi et al. 2013; Yoda et al. 2014; Osanai-Futahashi et al. 2016). We synthesized dsRNA targeting BmorCPH24 and introduced it into the left larval epidermis of the fourth instar U strain using electroporation. From the 15 treated individuals, five showed reduced pigmentation in the treated region while no significant color variation was observed in the nontreated regions or in controls injected with dsRed (Figure S4, A and B in File S1). As larvae developed, the extension of the cuticle in the pigmentation-inhibited region became limited and constricted in the RNAi group at the third day of the fifth instar, while this was not observed in the opposite control region of the same animal and in the dsRed injected control larvae (Figure S4A in File S1). Furthermore, mRNA levels of BmorCPH24 were reduced in the lighter treated regions vs. untreated regions (Figure S4C in File S1).
To evaluate how a more extreme BmorCPH24 loss-of-function mutation would affect the body shape and coloration of the wild-type Dazao strain, we disrupted the gene using CRISPR/Cas9-mediated gene editing. We designed a guide RNA targeting a specific site in BmorCPH24 exon2 using CRISPRdirect (http://crispr.dbcls.jp/) (Figure 5A), and then simultaneously injected the sgRNA and Cas9 mRNA into newly laid embryos (G0 generation). Complete Bo-like morphology was observed in multiple individuals (∼200 larvae) of the G1 generation (Figure 5B). Genomic sequencing of five mutant individuals identified four kinds of frame shift mutations all disrupting the BmorCPH24 coding region (Figure 5C). These results show that BmorCPH24 is necessary for the development of a normal larval body shape and pigmentation during silkworm development.
BmorCPH24 is involved in formation of larval endocuticle
The insect cuticle generally is composed of multiple layers, the envelope and the epicuticle (outer), the exocuticle (medial), and the endocuticle (inner). The exocuticle and endocuticle are composed of chitin and protein complexes (Moussian 2010). The exocuticle is deposited before the pharate period preceding ecdysis, while the deposition of endocuticle, which is rich in chitin, occurs postecdysis (Andersen et al. 1995). Thus, the expression pattern of BmorCPH24 indicates that it may be involved in the synthesis of endocuticle. To investigate the structure of the BmorCPH24-dependent cuticle in more detail, we observed cuticle sections of the wild-type and Bo mutant larvae from the first day to the third day of the fifth instar. The observation showed that the endocuticle was not deposited both in wild-type and Bo on the first day of the fifth instar, the period shortly after molting, while the outer two layers, the epicuticle and the exocuticle, had comparable morphology between the wild-type and the mutant (Figure 6). On the second day of the fifth instar, the endocuticle was massively deposited in wild-type larvae, as observed by direct visualization of the chitin in the layer using labeled wheat germ agglutinin-FITC (Figure 6). Note that the second day of the fifth instar corresponds to the expression peak of BmorCPH24 in wild-type larvae (Figure 3B). In contrast, endocuticle formation in Bo was extremely inhibited on the second and third days of the fifth instar (Figure 6). These results suggest an indispensable role of BmorCPH24 in the development of larval endocuticle after ecdysis.
In our previous study, we identified a larval cuticular protein (LCP) gene, BmLCP17, which is responsible for the silkworm stony (st) mutant phenotype (Qiao et al. 2014). Note that the Bo mutant has a similar but stronger phenotype compared to st (Figure S5A in File S1). BmLCP17, BmLCP18, BmLCP22, and BmLCP30 are evolutionarily conserved and abundantly expressed LCPs with typical R&R consensus motifs (Nakato et al. 1994, 1997; Shofuda et al. 1999). A microarray performed on different larval developmental stages revealed a consistent expression pattern of BmorCPH24, BmLCP17, BmLCP18, BmLCP22, and BmLCP30; high expression in feeding phases vs. low expression in molting phases (Figure S5B in File S1). This consistent spatiotemporal expression suggests that these LCPs are probably involved in building the endocuticle after the molt. We investigated the expression of these genes in the epidermis of the wild-type and the Bo mutants in fifth instar larvae. In Bo mutants, the four LCPs were significantly down-regulated when compared to +Bo/+Bo (Figure S5C in File S1). Expression of the four LCPs was also down-regulated in the BmorCPH24KO individuals. Thus, these results suggested that BmorCPH24 is able to affect the expression of other LCPs responsible for the assembly of endocuticle.
The relationship between BmorCPH24 and larval pigmentation
The positive correlation of BmorCPH24 expression with larval marking pigmentation suggests this gene is associated with pigment synthesis. To test whether ectopic expression of BmorCPH24 was sufficient to induce pigmentation, we constructed a BmorCPH24 overexpression vector driven by the IE1 constitutive promoter that also contained the EGFP gene driven by the A3 promoter (Figure S6Aa in File S1). The vector [IE1-BmorCPH24/A3-EGFP] and piggyBac helper plasmid were injected simultaneously into day 2, fourth instar wild-type larvae (+P) and then electroporated into cells. BmorCPH24-transformed cell lineages that expressed EGFP were visualized microscopically and dissected in the subsequent larval fifth instars (Figure S6Ab in File S1). The uninjected right side without a EGFP signal was used as the control. qRT-PCR confirmed that the expression of BmorCPH24 was up-regulated in the EGFP positive region (Figure S6Ac in File S1). However, no pigmentation was observed in the BmorCPH24 overexpressed region, which indicates that presence of BmorCPH24 protein is not sufficient to induce pigmentation (Figure S6Ab in File S1).
The absence of BmorCPH24, however, is able to inhibit pigmentation, especially in the region of the twin-spot markings on the dorsal side of epidermis in Bo and L mutant larvae, a strain where Wnt1 is ectopically expressed in those markings (Yamaguchi et al. 2013) (Figure 1, C and D). Wnt1 regulates genes related with pigment synthesis and plays an important role in the wing pigmentation of D. guttifera (Werner et al. 2010) and in the wing eyespots of Bicyclus anynana (Ozsu et al. 2017). To test for possible epistasis of BmorCPH24 on Wnt1 we surveyed the expression of Wnt1 in the marking region of wild-type and Bo on the third day of the fourth instar (the end of the feeding phase). Wnt1 mRNA expression was significantly lower in Bo mutants relative to wild-type indicating epistasis of BmorCPH24 on Wnt1 (Figure S6B in File S1). Taken together, these results show that BmorCPH24 is required for the synthesis of cuticle pigments, and its deficiency leads to the down-regulation of genes also required for pigment synthesis, such as Wnt1.
Disruption of BmorCPH24 induces apoptosis in epidermal cells
As shown above, deficiency of BmorCPH24 caused the down-regulation of several LCPs as well as of Wnt1. However, since BmorCPH24 is a cuticle structural protein, it is unable to directly regulate the expression of these genes. We, thus, investigated whether deficiency of BmorCPH24 led to epidermal cell death and, thus, to lower levels of these other genes. A previous report showed variation in epidermal cell morphology in Bo, with the cell nuclei being irregularly distributed (Oota and Kanekatsu 1993), and our epidermal sections showed a significantly reduced number of nuclei in Bo epidermal cells compared with wild type (Figure 6). Nuclear staining of epidermal cells from day 3 of the fourth instar (the end of the feeding phase) and day 3 of the fifth instar showed that wild-type nuclei have a uniform distribution and size, whereas they are uneven in Bo mutants, and are accompanied by apoptotic bodies (Figure 7A), suggesting that cell death was being promoted by a BmorCPH24 deficiency.
To confirm the presence of apoptosis in Bo mutants, we surveyed the expression of several apoptosis-related genes in these mutants and in wild type. Proapoptotic genes surveyed at day 3 of the fifth instar, including BmICE and BmICE2, were up-regulated, while the antiapoptotic gene BmBuffy was down-regulated in Bo mutants compared with wild type (Figure 7B). Furthermore, protease activity of two important apoptotic signaling molecules, the initiator caspase-9 and effector caspase-3/7, was enhanced in the epidermis of Bo mutants compared with wild-type, both in the fourth and fifth instar larvae (Figure 7C). Together, these results suggest that deficiency in BmorCPH24 induces apoptosis in larval epidermis, likely leading to lowered expression of LCPs and Wnt1.
BmorCPH24 likely arose via gene duplication after the divergence of B. mandarina and M. sexta
To better understand the origin of BmorCPH24, we investigated its evolutionary history using a phylogenetic approach. First, we performed a phylogenetic analysis of genes encoding cuticular proteins with low complexity (CPLC) sequence from four different insect species: B. mori, D. punctatus, M. sexta, and A. gambiae. Most CPs belonging to the same species clustered together, indicating recent lineage-specific duplications, but BmorCPH24 clustered with the CPLCA family of D. punctatus (Figure 8A), a family whose proteins are rich in alanine residue. No close orthologous gene was found in B. mori, M. sexta, and A. gambiae, but one paralogous gene, BmorCPH26 (BGIBMGA011767), with 77% amino acid sequence similarity, was obtained from silkworm (Figure 8A). We additionally used BmorCPH24’s sequence as a query to perform BLASTP and TBLASN searches against the nonredundant (nr) protein database in NCBI, and against the genomes or transcriptomes of six additional lepidopteran insects: M. sexta, Spodoptera frugiperda, Chilo suppressalis, Danaus plexippus, Papilio polytes, and Plutella xylostella. No orthologs of BmorCPH24 were found, but one ortholog of BmorCPH26, the paralog of BmorCPH24, was found in M. sexta (E-value <10−3), and we named it MsexCPH26 (Figure S7A in File S1). These observations indicate that BmorCPH24 was likely produced by gene duplication from BmorCPH26 after the split from its sister clade M. sexta (Kawahara and Breinholt 2014) (Figure 8B).
Gene duplication is an important source of novel genes because, due to redundancy, one of the copies is free to evolve new expression domains or new functions (He and Zhang 2005; Innan and Kondrashov 2010). To test whether BmorCPH24 evolved new expression domains, we investigated the spatial expression pattern of BmorCPH26, the paralog of BmorCPH24, in B. mori. We detected ubiquitous expression of BmorCPH26 in multiple tissues at low levels (Figure S7B in File S1). Furthermore, MsexCPH26 was not part of the transcriptome of M. sexta larval epidermis (Dittmer et al. 2015). This suggests that the ancestral function of MsexCPH26/BmorCPH24 proteins may be unrelated to the assembly of larval cuticle. BmorCPH24 appears, thus, to have evolved a new function in larval cuticle assembly. Based on these findings, we propose that BmorCPH24 is a novel gene that has evolved a new biological function over a short evolutionary time scale, after undergoing gene duplication.
Discussion
CPs were previously shown to be important in the assembly of an insect’s exoskeleton and ultimately its body shape (Guan et al. 2006; Arakane et al. 2012; Jasrapuria et al. 2012; Soares et al. 2013; Qiao et al. 2014; Noh et al. 2015; Tajiri et al. 2017). However, prior to this study, it was not known that CPs could also affect an insect’s pigmentation pattern. In this study, we demonstrated by fine mapping and gene knockout analysis that a novel gene, BmorCPH24, encoding a cuticular protein with a low complexity sequence, is responsible for the silkworm Bo mutant phenotype, which exhibits bamboo-like body shape and reduced pigmentation. Additional experiments, such as genetic crosses and RNAi knockdown, also showed that loss of BmorCPH24 inhibits pigmentation in two other silkworm strains, L and U, and causes changes in their body shape. These results further validate that BmorCPH24 can simultaneously affect body shape and color patterns in silkworm larvae.
BmorCPH24 is one of the most abundantly expressed genes in the epidermal cDNA library of B. mori (Okamoto et al. 2008). This indicates that BmorCPH24 may be an abundant cuticle structural protein that is essential for larval cuticle morphology in B. mori. In this study, qRT-PCR showed that BmorCPH24 was specifically and highly expressed in the epidermis (Figure 3, A and B). In general, deficiency of highly abundant CPs can limit cuticle expansion (Arakane et al. 2012; Qiao et al. 2014; Noh et al. 2015). In Bo mutants, disruption of BmorCPH24 inhibits the synthesis of endocuticle, which likely limits the expansion of cuticle and leads to the Bo abnormal body shape. Deficiency of BmorCPH24 is accompanied by the down-regulation of LCP17, the gene responsible for st mutant, and another three LCPs, which may explain why phenotypic effects of Bo are more extreme than those of st.
Because BmorCPH24 is a cuticular protein, instead of a transcription factor, the manner in which it regulates the expression of other LCPs is likely indirect. In T. castanenum, changes in expression of one CP gene have no effect on the expression of other CP genes (Noh et al. 2015). Thus, the down-regulation of LCPs in B. mori Bo mutants may be the result of epidermal cell apoptosis caused by loss of BmorCPH24 proteins. Bo mutants have abnormal epidermal cell development and we speculate that this abnormal development affects the expression of the other genes related to the synthesis of endocuticle in postecdysial larvae. Why depletion of a highly abundant CP leads to cell apoptosis, however, is still unclear.
The color pattern of insect larvae is determined by two sequential biological processes: the synthesis of pigment precursors by a series of related genes in the epidermis, and the deposition of these precursors in the epicuticle and exocuticle after a series of oxidation processes that turns them into pigments (Futahashi et al. 2010). Upstream of this process is the expression of Wnt1 (Yamaguchi et al. 2013), which regulates the gene yellow, a pigment synthesis gene that affects wing pigmentation in D. guttifera (Werner et al. 2010). Mutations in BmorCPH24 appear to induce cell apoptosis, which inhibits the differentiation and development of epidermal cells, which disrupts the expression of Wnt1 and leads to decreased pigmentation. Epidermal cell apoptosis in Bo may also affect the expression of genes responsible for the color patterns of other strains, such as U, causing depigmentation in them as well. On the other hand, transcription levels of BmorCPH24 were significantly up-regulated in the marking region of wild-type larvae (Figure 4A). And we recently found that multiple CP genes, including BmorCPH24, were also up-regulated in silkworm larva with excess melanin accumulation (Wu et al. 2016). In addition, research in silkworm and butterflies has shown that cuticular nanostructure is finer in the black region compared with less pigmented regions, which is indicative of a tight association between pigmentation and the surface nanostructure that is under the charge of CPs (Janssen et al. 2001; Futahashi et al. 2012; Tan et al. 2016). Consequently, we speculate that BmorCPH24 may function as a melanin-related CP that cross-links with the cuticular pigments.
A surprising finding of this study is that deficiency of a cuticle structural protein, BmorCPH24, induced apoptosis of larval epidermal cells. Previous studies, however, had already pointed to the tight connection between epidermal cells and their ECM, which provides structural and biochemical support to the surrounding cells (Moussian and Cohen 2016). Cells are able to sense the mechanical properties of the ECM, and changes in the ECM are able to regulate cellular processes (Discher et al. 2005; Plotnikov et al. 2012), such as apoptosis (Wang et al. 2000). An attractive hypothesis to explain how mutations in BmorCPH24 led to pigmentation and body shape defects in B. mori is that lower levels of BmorCPH24 changed the biological microenvironment of cells leading to the production of apoptosis signals, which induced the apoptosis process. These changes in the microenvironment may have been mediated by mitochondria. Mitochondria are indispensable for the apoptosis pathway and dysfunctions in mitochondria induces apoptotic cell death (Green and Kroemer 2004). The mitochondria apoptosis pathway is mainly regulated by Bcl-2 and Caspase family proteins (Wang and Youle 2009). In silkworm, Bmbuffy was previously identified as one antiapoptotic gene belonging to the Bcl-2 family (Pan et al. 2014), while BmICE and BmICE2 function as proapoptotic caspases (Yi et al. 2014; Chen et al. 2015). Our qRT-PCR results showed that BmorCPH24 deficiency led to the down-regulation of Bmbuffy, and to the up-regulation of BmICE and BmICE2 (Figure 7B). Furthermore, the activity of caspase9 and caspase3/7 were also enhanced significantly in the Bo mutant (Figure 7C). Collectively, these results indicate that BmorCPH24 deficiency may induce apoptosis through the caspase-dependent apoptosis pathway.
In summary, we identified a fast-evolving gene, BmorCPH24, that is responsible for the silkworm Bo mutant. This gene has undergone neofunctionalization after having duplicated and evolved a new function in constructing and pigmenting larval cuticle. These findings not only contribute to our understanding of the regulation and evolution of insect larval cuticle morphology, but also highlight a new and specific target gene for Bombycidae pest control.
Acknowledgments
We thank Haruhiko Fujiwara and Junichi Yamaguchi for the advice on in situ hybridization, and Zhengwen Yan and Chenxing Peng on cuticle dissection. We also thank Duan Tan, Songzhen He, and Jiangbo Song for helpful comments. This work was supported by the Hi-Tech Research and Development 863 Program of China (grant no. 2013AA102507), the National Natural Science Foundation of China (no. 31372379, no. 31472153), China Agriculture Research System (CARS-18-ZJ0102), Chongqing Youth Science and Technology Talent Training Project (cstc2014kjrc-qnrc80001), and the municipal graduate student research innovation project of Chongqing (no. CYB2015066).
Footnotes
Supplemental material is available online at www.genetics.org/lookup/suppl/doi:10.1534/genetics.117.300300/-/DC1.
Communicating editor: M. Wolfner
Literature Cited
Author notes
These authors contributed equally to this work.