Knockout silkworms reveal a dispensable role for juvenile hormones in holometabolous life cycle

Abstract

Insect juvenile hormones (JHs) prevent precocious metamorphosis and allow larvae to undergo multiple rounds of status quo molts. However, the roles of JHs during the embryonic and very early larval stages have not been fully understood. We generated and characterized knockout silkworms (Bombyx mori) with null mutations in JH biosynthesis or JH receptor genes using genome-editing tools. We found that embryonic growth and morphogenesis are largely independent of JHs in Bombyx and that, even in the absence of JHs or JH signaling, pupal characters are not formed in first- or second-instar larvae, and precocious metamorphosis is induced after the second instar at the earliest. We also show by mosaic analysis that a pupal specifier gene broad, which is dramatically up-regulated in the late stage of the last larval instar, is essential for pupal commitment in the epidermis. Importantly, the mRNA expression level of broad, which is thought to be repressed by JHs, remained at very low basal levels during the early larval instars of JH-deficient or JH signaling-deficient knockouts. Therefore, our study suggests that the long-accepted paradigm that JHs maintain the juvenile status throughout larval life should be revised because the larval status can be maintained by a JH-independent mechanism in very early larval instars. We propose that the lack of competence for metamorphosis during the early larval stages may result from the absence of an unidentified broad-inducing factor, i.e., a competence factor.

Keywords: Bombyx; TALEN; ecdysone; juvenile hormone; metamorphosis.

Fig. 1.

TALEN-mediated gene disruption. (A) JH biosynthetic pathways in wild-type and mutant Bombyx. In the CA of wild-type Bombyx, farnesoic acid (FA) is epoxidized to JH acid by CYP15C1 and JH acid then is methylated to the JH by JHAMT (5). In the CA of mod and JHAMT mutants and JHAMT mod double mutants the presumed products are MF, JH acid, and FA, respectively (underlined). Note that the major JHs and their precursors (not illustrated in the figure) are ethyl-branched in lepidopteran insects (5, 7). (B) Schematic representations of mutant alleles used in this study. For JHAMT, Met2, and Met1 mutants, we fixed 88-, 44-, and 67-bp deletion alleles, respectively, and used them in further experiments. Gray and white bars indicate the coding regions and UTRs, respectively. Horizontal lines flanked by exons indicate introns. Red lines below the exons indicate deleted regions, and red arrows indicate the positions of premature stop codons caused by deletions. Chromosomal localizations of each gene are shown in parentheses. The nucleotide sequence of each allele is shown in Fig. S1. bHLH, basic helix–loop–helix domain; MTase, methyltransferase domain; PAC, PAC domain (24); PASA, Per-ARNT-Sim A domain; PASB, Per-ARNT-Sim B domain. (Scale bar: 1 kb.)

Fig. S1.

Knockout alleles generated in this study. TALEN-biding sites for JHAMT, Met2, and Met1 are underlined. TALEN_L and TALEN_R indicate the biding sites of the left and right TALEN monomers. Deletion regions are indicated in red, and premature stop codons are shown in blue letters. For more information, see Fig. 1 in the main text.

Fig. 2.

Effects of knockouts on embryogenesis. (A and B) Hatchability (A) and day of hatching (B) by mutant strains. Hatched larvae or unhatched embryos (stage ≥25) (67) were counted, and their individual genotypes were determined by PCR if necessary. Mutated lines are shown below the bars. Bars indicate mean + SD for hatchability (A) or the hatching day AO (B) for each genotype (n = 4–15 batches). The numbers above bars indicate the number of hatched/total embryos, and the numbers within bars indicate the number of batches used in the analysis. Asterisks indicate significant differences compared with the control. *P < 0.05, ***P < 0.001, ****P < 0.0001, Dunn’s nonparametric multiple comparisons test. (C) qRT-PCR of genes in the control strain. Data represent the mean + SD expression level of each gene (n = 4 biological replicates). The numbers below the x axis indicate hours AO. The presumed JH titer shown at the top is inferred from that in Manduca (28). Timings of the onset of blastokinesis (b) and completion of dorsal closure (DC) are based on Takami and Kitazawa (67). (D) Whole-mount in situ hybridization of JHAMT and Kr-h1 in the control strain. Arrows indicate the CA-specific signals of JHAMT. Signals in the head and thoracic legs are caused by nonspecific staining. Stages (days AO) of samples are shown at the top right of each panel. (i, ii, v, and vi) Lateral view. (iii and vi) Dorsal view. (Scale bars: 1 mm.) (E) Whole-mount in situ hybridization of Kr-h1 in JHAMT-knockout embryos. Embryos obtained from the sibling cross of JHAMT^+/− were subjected to analysis. About one-fourth of the embryos were not stained, as indicated by the white arrows. These unstained embryos were probably JHAMT homozygous mutants, as supported by the results of qRT-PCR (Fig. S6). (Scale bars: 1 mm.)

Fig. S2.

Effects of JH treatment on embryogenesis and development. (A) JH III or methoprene (1 µg per egg) was applied to the embryos of the control strain on day 3 or 4 AO. Blue, orange, and black bars indicate the percentage of hatched embryos, unhatched embryos with normal blastokinesis, and unhatched embryos with incomplete blastokinesis, respectively. Chorions of unhatched embryos were removed on day 12. Representative images are shown on the right. The numbers within bars indicate the number of individuals investigated (n = 48–50). The same volume of acetone was used as a solvent control. (B) JHAMT^−/− embryos were treated with selected amounts of JH III, MF, or methoprene on day 4 AO, 1 d before the presumed JH peak (also see Fig. 2 in the main text). Hatched larvae were reared, and their development was recorded. Note that none of the larvae metamorphosed after L1 or L2 or reached L5. All the larvae that underwent precocious metamorphosis (after L3 or L4; highlighted in red) died as larval–pupal intermediates or during the pupal stages; none became adults. The same volume of acetone was used as a solvent control.

Fig. 3.

Phenotypes of JH-deficient Bombyx. (A) Rescue experiments of JHAMT^−/− embryos. Bars indicate the hatchability of JHAMT^−/− embryos treated with selected amounts (0.001–10 µg) of JH III, methoprene, and MF on day 4 AO. Acetone was used as a solvent control. n = 7–16 individuals for each bar. (B) Time-dependent effects of JH III application. JH III (0.1 µg) was applied to JHAMT^−/− embryos at selected time points, and their hatchability was recorded. n = 7–17 individuals for each bar. (C) Dechorionation of unhatched eggs from the sibling cross of JHAMT^+/− moths. The image shows unhatched eggs on day 12 AO. After dechorionation, JHAMT^−/− larvae (individuals #1, 3, and 4) commenced locomotory behaviors, but a JHAMT^+/+ larva (individual #2) was dead. Genotypes of the larvae shown in C′ were determined by genomic PCR (C′′). (Scale bars: 1 mm.) (D) Development of JHAMT^−/− larvae treated with JH III (0.1 µg) on day 4 AO. JHAMT^−/− larvae started wandering and spinning from L3, whereas the JHAMT^+/+ and JHAMT^+/− larvae reached L4. (Scale bar: 1 cm.) (E) Dechorionated JHAMT^−/− larvae (L4) became small larval and pupal intermediates. (Scale bar: 1 cm.) (F) Development of JH-deficient mutants, which underwent precocious metamorphosis after L3 at the earliest. No larvae reached L5.

Fig. S3.

SEM analyses of neonate JHAMT^−/− larvae. Dechorionated JHAMT^−/− larvae were subjected to SEM analyses (n > 4 for each panel). A representative result is shown here. We observed no apparent morphological defects in JHAMT^−/− larvae. Scale bars are shown below each panel.

Fig. 4.

Phenotypes of JH receptor mutants. (A) Image of L2 larvae obtained from the sibling cross of Met1^+/− adults. Met1-mutant larvae exhibited retarded growth (red arrows). (Scale bar: 1 cm.) (B) Most of the Met1-mutant larvae were developmentally arrested when molting from L2 to L3. (Scale bar: 2 mm.) (C) Formation of patched pupal cuticles in L3 Met1 mutants. The old L2 cuticles of arrested Met1 larvae were removed with forceps (asterisk). Pupal patches were found around the T2 and T3 bulges (32) and the dorsolateral region of A1. (Scale bar: 1 mm.) (Inset) Magnified view of the area around the T3 bulge with pupal patches (arrowheads) (Scale bar: 0.2 mm). (D) SEM analysis of patches of pupal cuticles of L3 Met1 larvae. Cuticles with pupal characteristics are indicated by dotted lines. A magnified view is shown on the right. (Scale bars: 500 µm, Left and 50 µm, Right). (E) Image of Met2 Met1 double mutants, which were developmentally arrested during the molt from L2 to L3. The positions of the pupal patches were very similar to those in Met1 mutants. (Scale bar: 2 mm.) (F) Lethal stages in the control, Met2, Met1, and Met2 Met1 mutants.

Fig. S4.

SEM analyses of epidermis around T2 or T3 bulges in JH receptor mutants. The first to the third larval instars of the JH receptor mutants (Met2^−/−, Met1^−/−, and double mutants Met2 Met1) and control were subjected to SEM analyses. Cuticles around the T2 and T3 bulges were examined to determine the presence of pupal cuticular patches. Pupal cuticular patches were found only in Met1^−/− and Met2^−/− Met1^−/− larvae, and their formation was observed at L3 at the earliest. The white arrowheads indicate T2 or T3 bulges, and the brown arrows indicate pupal patches. More than three individuals were analyzed at each stage for each genotype, and, if necessary, specimens were genotyped by PCR after SEM analyses. Scale bars are shown below each panel.

Fig. 5.

Mosaic analysis of Met1. (A) Representative images of mosaic larvae. The larvae were reared individually and photographed at each larval instar. Individual numbers are shown on the left, and larval instars are shown at the top. Met1 mosaics (#96, 174, and 152) did not exhibit pupal characters during L1 or L2, but small pupal patches were formed in L3 (see the magnified view of #152). Many of the L3 Met1 mosaics molted to L4 and became severe larval–pupal mosaics. (Scale bars: 2 mm, L1 and L2 and 5 mm, L3–L5). (B) SEM analyses of the cuticles of Met1 mosaic larvae. Pupal patches in the A1 of an L3 larva and the dorsal region of an L4 mosaic larva are shown. Magnified views are shown on the right. Note the presence of characteristic pupal cuticles. The anterior (a) and posterior (p) axes are indicated also. (Scale bars: 500 µm, Left and 50 µm, Right). (C) The most severe example of a Met1 mosaic L4 larva. In this L4 larva, imaginal primordia grew considerably. Note the everted wing discs and the growth of imaginal compound eyes, antennae, and legs. an, antennae; eye, compound eyes; fw, forewings; hw, hindwings; leg, thoracic legs. (Scale bars: 2 mm.) (D) A mosaic larva whose abdominal prolegs were degenerated. Black and white arrowheads indicate degenerate and normal larval abdominal legs, respectively. (Scale bar: 1 mm.)

Fig. S5.

Phenotypes of JH-deficient and JH signaling-deficient Bombyx. Schematic representations of the JH-deficient and JH signaling-deficient Bombyx phenotypes. We found that the earliest precocious metamorphosis was induced after molting from L2 to L3 and was limited to small regions of the L2 epidermis (T2, T3, and A1 segments; dashed arrows). Severe metamorphic molts were induced only after L3.

Fig. 6.

qRT-PCR of Kr-h1 and broad. (A) Expression levels of Kr-h1α, Kr-h1β, and broad mRNAs in the control strain during L1–L3 (whole body), L5, and pupa (epidermis, fat body, and midgut). The relative mRNA expression levels normalized against rp49 are shown. The stages of the samples are shown below. Results represent mean + SD values (n = 4 biological replicates). HCS, head capsule slippage; P, pupation; sp, onset of spinning. (B) Expression levels of Kr-h1α, Kr-h1β, and broad mRNAs during the early larval instars of mutants. Total RNA was extracted from the whole body using individually genotyped samples (Materials and Methods). Relative mRNA expression levels normalized against rp49 are shown. As shown in a previous study (6), the Kr-h1 expression levels were lower in the mod strain than in the control. Note the very low levels of Kr-h1 mRNA in JHAMT, JHAMT mod, Met1, and Met2 Met1 mutants. In the mutants used in this study, broad expression was comparable with that in the control, but it remained at a very low level. Note that the y-axis scales differ in A and B. The data represent the mean + SD (n = 3–5 biological replicates). (C) Effects of methoprene treatment on gene expression levels. L1 larvae (24 h after hatching) were treated topically with 0.1 µg of methoprene (+) or solvent alone (acetone) (−), and total RNA was extracted from the whole body of each individual 24 h after the treatment. Values represent the mean + SD (n = 3–5) normalized against the value of the control strain treated with acetone (set to 1). The values of fold changes are indicated above the bars. Asterisks indicate significant differences compared with the control values according to the Student’s t test: *P < 0.05; **P < 0.01; ***P < 0.001. n.s., not significant.

Fig. S6.

qRT-PCR of Kr-h1 and broad during embryonic stages. (A) Expression levels of Kr-h1α, Kr-h1β, and broad mRNAs on day 5 AO, at the presumed peak of the JHs. Total RNA and gDNA were extracted from the whole bodies of individuals, and samples with genotypes defined by PCR were used in the analysis (Materials and Methods). The values represent the mean + SD (n = 3–5) normalized against that of the control strain (set to 1). Means with different letters differ significantly (Tukey–Kramer test, P < 0.05; n.s., not significant). (B) Time-course analysis of the mRNA expression levels of Kr-h1α, Kr-h1β, and broad (on days 3, 5, 7, and 9 AO). Note the very low levels of Kr-h1 mRNAs in JHAMT, JHAMT mod, Met1, and Met2 Met1 mutants. The broad mRNA expression levels were comparable among the strains used. The values represent the mean + SD (n = 3–5) normalized against that of the control strain (set to 1). (C) Effects of methoprene treatment on gene expression. At day 4 AO, eggs were treated with 0.1 µg of methoprene (+) or the solvent alone (acetone) (−). Total RNAs were extracted from individuals 24 h after the treatment. The values represent the mean + SD (n = 3–5) normalized against the value of the control strain treated with acetone (set to 1). Values of the fold changes are indicated above the bars. Asterisks indicate significant differences compared with the control values according to the Student’s t test: **P < 0.01; ***P < 0.001. n.s., not significant.

Fig. 7.

broad is essential for pupal metamorphosis in the epidermis. (A) Representative image of a mosaic pupa with patches of larval cuticles. TALEN mRNAs that targeted broad were injected into embryos, and the hatched larvae were subjected to the phenotypic analysis. (B) The animal shown in A. This pupa metamorphosed into a moth with larval–adult mosaic cuticles. Arrows in A and B indicate the positions of larval patches. Magnified views of mosaic cuticles are shown in A′ and B′. SEM analyses showed that the “white” cuticles actually exhibited the characteristic structures of larval cuticles (A′′ and B′′ ). s, scale; sc, socket cells (indicated by white arrows). (Scale bars: 2 mm in A, B, and B′, 0.2 mm in A′, and 0.1 mm in A′′ and B′′). (C) Hypothetical model of the regulatory mechanism of broad. We assume that the presence of a broad-inducing factor, or competence factor, is required for the up-regulation of broad that leads to pupal commitment (Discussion). Tissues in L1 larvae and most tissues in L2 larvae are not pupally committed, irrespective of the presence or absence of JH, because of the absence or very low levels of this competence factor. After L3, larvae can be pupally committed because of the presence or high levels of this factor, but its action is repressed by JHs, as demonstrated in the later stages in Manduca and Bombyx (–21, 45). In the last instar, the JH titer decreases to a very low level, and thus broad can be strongly induced by the putative factor and 20E, thereby leading to pupal metamorphosis.

Fig. S7.

Mosaic analysis of broad. (A) Schematic representation of broad cDNA and the TALEN target site. Note that only the A-Z1 isoform is shown here, but there are at least 14 splice variants of broad in Bombyx (22). The target site was designed in the Bric-a-brac-Tramtrack-Broad (BTB) domain because this domain is present in all the broad isoforms that have been identified. Gray and white bars indicate the coding regions and the UTRs, respectively. The TALEN target site is indicated by a red arrow. ZF, zinc finger domain. (Scale bar, 1 kb.) (B) Summary of a mosaic analysis of broad. In this experiment, the dose of TALEN mRNAs was reduced to 4 ng/µL, 100-fold lower than the dose used in standard experiments (Materials and Methods in the main text). Among 114 hatched larvae, 53 became larval–pupal mosaics, and 35 became larval–adult mosaics (Fig. 7 in the main text). (C) Mutations induced by TALENs that targeted broad. TALEN mRNAs (400 ng/µL; note that this is a standard dose) were injected into embryos, and the hatched larvae (G₀) were reared. Most of the larvae were dead by the prepupal stage, and none became pupae. gDNA was extracted from a larva that died at the prepupal stage and was used as a template for genomic PCR. PCR products were cloned into a cloning vector, and 48 clones were selected randomly and sequenced. In total, 38 clones were sequenced successfully, and all had mutations (100%). We recovered 25 mutant alleles with short insertions and deletions, which ranged from 23-bp insertions to 35-bp deletions; their nucleotide sequences are shown in the figure. The numbers in parentheses indicate the number of clones. The TALEN-binding sites are indicated by red letters.