RNAi-mediated knockdown of juvenile hormone acid O-methyltransferase gene causes precocious metamorphosis in the red flour beetle Tribolium castaneum

Authors


T. Shinoda, National Institute of Agrobiological Sciences, 1–2 Ohwashi, Tsukuba, Ibaraki 305-8634, Japan
Fax: +81 29 838 6075
Tel: +81 29 838 6075
E-mail: shinoda@affrc.go.jp

Abstract

Juvenile hormone controls the timing of insect metamorphosis. As a final step of juvenile hormone biosynthesis, juvenile hormone acid O-methyltransferase (JHAMT) transfers the methyl group from S-adenosyl-l-methionine to the carboxyl group of farnesoic acid and juvenile hormone acid. The developmental expression profiles of JHAMT mRNA in the silkworm Bombyx mori and the fruitfly Drosophila melanogaster suggest that the suppression of JHAMT transcription is critical for the induction of larval–pupal metamorphosis, but genetic evidence for JHAMT function in vivo is missing. In this study, we identified three methyltransferase genes in the red flour beetle Tribolium castaneum (TcMT1, TcMT2 and TcMT3) that are homologous to JHAMT of Bombyx and Drosophila. Of these three methyltransferase genes, TcMT3 mRNA was present continuously from the embryonic stage to the final larval instar, became undetectable before pupation, and increased again in the adult stage. TcMT3 mRNA was localized in the larval corpora allata. Recombinant TcMT3 protein methylated farnesoic acid and juvenile hormone III acid, but TcMT1 and TcMT2 proteins did not. Furthermore, RNA interference-mediated knockdown of TcMT3 in the larval stage resulted in precocious larval–pupal metamorphosis, whereas knockdown of either TcMT1 or TcMT2 showed no visible effects on metamorphosis. Importantly, precocious metamorphosis caused by TcMT3 RNA interference was rescued by an application of a juvenile hormone mimic, methoprene. Together, these results demonstrate that TcMT3 encodes a functional JHAMT gene that is essential for juvenile hormone biosynthesis and for the maintenance of larval status.

Abbreviations
Bm

Bombyx mori

CA

corpora allata

DIG

digoxigenin

Dm

Drosophila melanogaster

EGFP

enhanced green fluorescent protein

FA

farnesoic acid

JH

juvenile hormone

JHA III

juvenile hormone III acid

JHAMT

juvenile hormone acid O-methyltransferase

JHM

juvenile hormone mimic

LA

lauric acid

MF

methyl farnesoate

PA

palmitic acid

RNAi

RNA interference

SAM

S-adenosyl-l-methionine

Tc

Tribolium castaneum

Insect juvenile hormone (JH) is a multifunctional hormone that controls a variety of physiological events, e.g. growth and development, reproduction, diapause and caste determination in social insects [1]. The most prominent role of JH is the control of insect metamorphosis, which has been studied extensively in many species [2]. In holometabolous insects, for example, larvae do not initiate larval–pupal metamorphosis until JH in the hemolymph declines at the end of the larval stage. If JH in the hemolymph is precociously eliminated by surgical removal of the corpora allata (CA), the specialized endocrine organs that secrete JH into the hemolymph, precocious metamorphic change occurs. In contrast, application of a JH mimic (JHM) at the onset of larval–pupal metamorphosis prevents metamorphosis and causes an extra larval moult in some insect species [2]. Therefore, JH has a ‘status quo’ action to prevent metamorphosis.

JH is a unique farnesoid with a methyl ester moiety at the C1 position and an epoxide group at the C10–11 position [3,4]. Natural compounds with these chemical features have been found only in insects, with one exception of JH III isolated from a Malaysian plant Cyperus iria [5]. The biosynthetic pathway of JH in CA is conventionally divided into two parts: early steps and late steps. The early steps, starting from acetyl-CoA or propionyl-CoA and leading to (homo)farnesyl diphosphate, constitute the standard mevalonate pathway and are conserved in various organisms, including vertebrates [6,7]. In contrast, the late steps, starting from (homo)farnesyl diphosphate and leading to JH, are unique to JH biosynthesis [7]. As the final step of JH biosynthesis, farnesoic acid (FA) is converted into active JH by methylation of a carboxyl group and epoxidation at the C10–11 position [8].

The identification of the genes encoding enzymes in the late steps has been hampered because of a lack of vertebrate and plant homologues. Recently, we identified and characterized the JH acid O-methyltransferase (JHAMT) gene that encodes one of the late step enzymes, first from the silkworm Bombyx mori [9], and then from the fruitfly Drosophila melanogaster [10]. In vitro enzyme assays showed that recombinant JHAMT proteins of B. mori (BmJHAMT) and D. melanogaster (DmJHAMT, CG17330) methylated carboxyl groups in JH acid and FA in the presence of S-adenosyl-l-methionine (SAM) [9,10]. JHAMT mRNA is detected primarily in CA of both B. mori and D. melanogaster, and its temporal expression profile correlates well with a change in the JH titre in the hemolymph, suggesting that the suppression of JHAMT transcription at the end of the larval stage is critical for the initiation of metamorphosis into a pupa [9,10]. However, direct evidence for the significance of JHAMT in inducing larval–pupal metamorphosis remains to be shown.

To reveal the function of JHAMT in vivo, overexpression and RNA interference (RNAi)-mediated knockdown of JHAMT were performed in D. melanogaster [10]. Overexpression of DmJHAMT caused a pharate adult lethal phenotype, as well as defects in the rotation of adult male genitalia [10], both of which are typically observed after treating wild-type insects with an excess of JHM at the end of the larval stage [11–14]. In contrast, RNAi-mediated knockdown of DmJHAMT showed no visible effect on growth and development [10]. However, whether the RNAi-mediated knockdown of DmJHAMT is effective enough to completely eliminate JH in the hemolymph needs to be examined. Functional analysis using RNAi techniques has not confirmed the significance of JHAMT in JH biosynthesis.

In this study, the red flour beetle Tribolium castaneum was chosen to analyse the in vivo function of JHAMT. In this species, RNAi-mediated knockdown of a gene of interest by injecting dsRNA into larvae is effective and easy to perform [15]. Although previous biochemical studies have disclosed the enzymatic properties of the JHAMT enzyme in intact CA of a related beetle, Tenebrio molitor [16], the JHAMT gene has not yet been identified in Coleoptera, including T. molitor. We report here the identification and functional characterization of three JHAMT-like methyltransferase genes (TcMT1, TcMT2 and TcMT3) from T. castaneum. Only TcMT3 of the three methyltransferase genes was shown by developmental and spatial expression profiles, and the enzymatic properties of the recombinant proteins, to encode a functional JHAMT gene. Furthermore, RNAi-mediated knockdown of JHAMT (TcMT3), but not TcMT1 or TcMT2, caused precocious larval–pupal metamorphosis, demonstrating that the JHAMT gene is essential for JH biosynthesis and maintenance of the larval status.

Results

Identification of three methyltransferase genes in T. castaneum

Three putative JHAMT-like methyltransferase genes were found in a genomic sequence contig (Contig4620_Contig8031) by tblastn searches of the beetle genome database with the sequences of BmJHAMT and DmJHAMT. Hereafter, these methyltransferase genes are called TcMT1, TcMT2 and TcMT3. The cDNAs containing full ORFs of TcMT1, TcMT2 and TcMT3 were amplified by RT-PCR using primers designed from the genomic sequences, and then sequenced. Comparison of the genomic sequence with the cDNA sequence revealed that TcMT2, TcMT1 and TcMT3 were located in this order (from the 5′-end to the 3′-end) with the same orientation in a ∼ 15 kb region (Fig. 1A). The deduced amino acid sequences of TcMT1, TcMT2 and TcMT3 were homologous to each other (amino acid identities, 42–50%), as well as to JHAMT of B. mori or D. melanogaster (Fig. 1B). The amino acid identities of TcMT1, TcMT2 and TcMT3 compared to BmJHAMT were 31%, 32% and 36%, respectively. The putative SAM-binding motif (motif I) is well conserved in all five methyltransferases. Each of the three TcMT genes consisted of three exons, as far as we examined (Fig. 1A), and two introns at positions 1 and 3 located in identical positions for the three TcMTs (Fig. 1B). The intron at position 3 was also conserved in DmJHAMT and BmJHAMT (Fig. 1B). Although DmJHAMT lacked an intron at position 1, BmJHAMT had an intron at position 1 and an extra intron at position 2 (Fig. 1B). The similarity in exon–intron structures of the three TcMT genes to that of the JHAMT genes in D. melanogaster and B. mori further confirmed that these are homologues of JHAMT.

Figure 1.

 Structure of the three methyltransferase genes in Tribolium castaneum. (A) Organization of the methyltransferase genes in T. castaneum. Exons are shown as boxes. (B) Alignment of TcMT1 (GenBank accession number: AB360761), TcMT2 (AB360762), TcMT3 (TcJHAMT, AB360763), Bombyx mori JH acid O-methyltransferase (BmJHAMT, BAC98835) and Drosophila melanogaster JH acid O-methyltransferase (DmJHAMT, BAC98836) sequences. Amino acids common in three or four methyltransferases are indicated by grey shadowed letters, and those common in all methyltransferases are indicated by white letters with a black background. The putative SAM-binding motif (motif I) is boxed. The positions of the introns are indicated by red lines, and the numbers above indicate the positions of the introns as described in the text.

Developmental expression profiles of TcMT1, TcMT2 and TcMT3

To examine the developmental expression profiles of TcMT1, TcMT2 and TcMT3 transcripts, quantitative RT-PCR analysis was performed (Fig. 2). The amount of TcMT1 transcript was relatively low during the embryonic and larval stages, but high in the last two days of adult development (Fig. 2A,B). The TcMT2 transcript was also weakly expressed during the embryonic and most of the larval stages, but showed a distinct peak at the beginning of the prepupal stage when the larval ocelli begin to retract and the insects become sluggish (Fig. 2C,D). The amount of TcMT3 transcript was high in the embryonic stage, decreased gradually in the second larval instar, decreased to a low level at the end of the sixth instar (day 2), and increased just before ecdysis to the seventh instar (Fig. 2E,F). The transcript level of TcMT3 gradually decreased during the final larval instar, but was still detectable in the prepupal stage (Fig. 2F). TcMT3 was undetectable in the pupal stage and during subsequent adult development, but increased again in adults by day 7 (Fig. 2F). This increase was observed in both males and females (data not shown).

Figure 2.

 Developmental expression profiles of TcMT1, TcMT2 and TcMT3 transcripts in Tribolium castaneum. Transcript levels of TcMT1 (A, B), TcMT2 (C, D) and TcMT3 (E, F) were analysed by quantitative RT-PCR, and the signal intensity was normalized to the intensity of TcRp49. In the embryonic stage and the first, second and third larval instars, RNA was isolated from a mass of eggs or larvae. From the sixth larval instar until the adult stage, RNA was isolated from individuals (three larvae in the sixth and seventh larval instars, three males and three females for pupae and adults for each time point). The means and standard deviations of expression are shown. The highest values during development (day 5 pupa for TcMT1, 84–96 h in the seventh instar for TcMT2 and the embryonic stage for TcMT3) were designated 100% for each gene.

Spatial expression profiles of TcMT1, TcMT2 and TcMT3

The tissue specificity of the TcMT1, TcMT2 and TcMT3 transcripts was examined by quantitative RT-PCR and in situ hybridization. Quantitative RT-PCR showed that TcMT1 and TcMT2 transcripts were more abundant in the posterior part of the sixth larval instar (Fig. 3A,B) and at the beginning of the prepupal stage in the seventh larval instar (Fig. 3D,E). In the sixth instar larvae, TcMT3 was specifically expressed in the anterior part, which presumably includes CA, where JH is synthesized (Fig. 3C). In contrast, the TcMT3 transcript was detected in both anterior and posterior parts of the seventh larval instar (Fig. 3F).

Figure 3.

 Spatial expression pattern of TcMT1, TcMT2 and TcMT3 transcripts in Tribolium castaneum. RNA was isolated from four larvae in the sixth instar (A–C) and from four larvae 84 h after ecdysis to the final (seventh) instar (D–F), which were cut in half between thoracic segments T2 and T3, and the transcript levels of TcMT1 (A, D), TcMT2 (B, E) and TcMT3 (C, F) in the anterior and posterior parts were examined by quantitative RT-PCR. Signal intensities relative to the highest values in the developmental expression profiles (see Fig. 2) are shown. A, anterior part; P, posterior part.

The localization of the TcMT3 transcript was further examined in the anterior part of sixth instar larvae by in situ hybridization (Fig. 4). With the antisense RNA probe, mRNA localization was found in a pair of small globular organs (Fig. 4C), but there was no obvious hybridization in these tissues with the sense RNA probe (Fig. 4B). These organs showing TcMT3 expression are the putative CA of T. castaneum. After removing the remaining head capsule with forceps, we located the putative CA on the ventral side of the brain (Fig. 4D). The size of the putative CA was approximately 15 μm in diameter.

Figure 4.

In situ hybridization of TcMT3 transcript in Tribolium castaneum. (A) Dorsal view of head and thoracic segments of a normal sixth instar larva. The area that was used for in situ hybridization and subsequent imaging (B–D) is boxed. (B–D) In situ hybridization of TcMT3. Heads of sixth instar larvae that were dissected with some part of the head capsule still attached were fixed, hybridized with sense (B) or antisense (C, D) RNA probes for the TcMT3 transcript, and detected. Pictures were taken before (B, C) and after (D) removing the head capsules with forceps. mRNA localization in the putative corpora allata is indicated by arrows, and non-specific staining in the cuticle is indicated by asterisks. A, anterior; BR, brain; D, dorsal; P, posterior; SG, sub-oesophageal ganglion; V, ventral.

Enzymatic properties of recombinant TcMT1, TcMT2 and TcMT3 proteins

The enzymatic activities of recombinant TcMT1, TcMT2 and TcMT3 proteins were examined against two potential substrates, FA and JH III acid (JHA III). Recombinant TcMT1 and TcMT2 protein did not show detectable activity to methylate these substrates. In contrast, recombinant TcMT3 protein catalysed the methylation of FA and JHA III to give methyl farnesoate (MF) and JH III, respectively (Table 1). The TcMT3 protein showed weak methyltransferase activity with normal saturated fatty acids, such as lauric acid (LA) or palmitic acid (PA), much lower than against FA and JHA III (Table 1).

Table 1.   Enzymatic activity of recombinant TcMT3 protein to FA, JHA III and saturated fatty acids. The average and standard deviation were calculated from independent enzyme assays (n = 3).
SubstrateActivity [mol·(mol enzyme)−1·min−1]
FA0.59 ± 0.04
JHA III0.56 ± 0.03
Lauric acid0.016 ± 0.002
Palmitic acid0.002 ± 0.001

JHA and JH have a chiral centre in the epoxide moiety at the C10–11 position. The stereospecificity of TcMT3 against a mixture of (10R)- and (10S)-enantiomers of JHA III was investigated by analysing the product with enantioselective HPLC. Under the conditions used in this study, (10R)- and (10S)-enantiomers of racemic JH III can be completely separated (Fig. 5A). The ratio of (10R)-JH III to (10S)-JH III in the product obtained with TcMT3 was 87 : 13 (Fig. 5B), indicating that TcMT3 catalyses the methylation of (10R)-JHA III more favourably than (10S)-JHA III.

Figure 5.

 Enantioselective HPLC profiles of racemic JH III (A) and the metabolites from racemic JHA III produced by recombinant TcMT3 protein (B). Arrows indicate (10S)-JH III and (10R)-JH III.

Effects of RNAi-mediated knockdown of TcMTs on larval–pupal metamorphosis

To examine the role of methyltransferase genes in the larval stage in vivo, RNAi-mediated knockdown of TcMT1, TcMT2 and TcMT3 was performed by injecting dsRNA at the beginning of the third instar. dsRNA for enhanced green fluorescent protein (EGFP) was injected as a control. First, the transcript levels 3–7 days after injection of dsRNA were quantified to confirm the efficiency of RNAi-mediated knockdown. As shown in Fig. 6A, injection of TcMT1 dsRNA suppressed the transcript level of TcMT1 itself compared with EGFP dsRNA-injected controls. In addition, the transcript level of TcMT2 was suppressed by injection of TcMT2 dsRNA (Fig. 6B), and the transcript level of TcMT3 was suppressed 3 days (Fig. 6C) and 6 days (Fig. 6D) after injection of TcMT3 dsRNA.

Figure 6.

 Efficiency of RNAi-mediated knockdown in Tribolium castaneum. (A) The level of TcMT1 transcript 6 days after injection with dsRNA of EGFP or TcMT1 in the third larval instar (n = 6). (B) The level of TcMT2 transcript 7 days after injection with dsRNA of EGFP or TcMT2 in the third larval instar (n = 4). (C, D) The level of TcMT3 transcript 3 days (C) and 6 days (D) after injection with dsRNA of EGFP or TcMT3 on day 0 of the fourth larval instar (day 0_4th; n = 4). (E) The level of TcMT1 transcript in pharate adults after injection with dsRNA of EGFP or TcMT1 in the prepupal stage (n = 3). (F) The level of TcMT2 transcript 3 days after injection with dsRNA of EGFP or TcMT2 on day 0 of the seventh larval instar (day 0_7th; n = 3). Means and standard deviations are shown, and the intensity in EGFP dsRNA-injected insects was set at 100% in each graph.

In the controls that received EGFP dsRNA at either 1.5–2.0 or 5.0 μg·μL−1, no significant effect on growth or metamorphosis was observed, and all of these larvae pupated at the end of the seventh or eighth larval instar and eclosed normally (Table 2). All the larvae that received TcMT1 or TcMT2 dsRNA also pupated and eclosed normally without undergoing precocious metamorphosis (Table 2). In contrast, TcMT3 RNAi caused precocious pupation, and most of the larvae pupated at the end of the sixth instar (Table 2; Fig. 7A). These pupae and adults appeared normal in their external morphology, but were much smaller than normal animals (Fig. 7). Three larvae that had been injected with TcMT3 dsRNA showed prepupal characteristics, such as larval ocellar retraction, at the end of the fifth larval instar, but only one larva of these three larvae succeeded in pupation followed by eclosion, whereas the other two arrested either as prepupa or pupa (Table 2). No significant difference in the effect of RNAi as a result of the dose of dsRNA was observed in this study.

Table 2.   Phenotypes of Tribolium larvae injected with dsRNAs on day 0 of the third instar. Numbers of animals, and the instar when they pupated, are indicated. The insects that underwent precocious metamorphosis are shown in bold.
dsRNA injectionPupationLethal phase
dsRNAConcentration (μg·μL−1)n5th6th7th8th
EGFP1.5–2.022184 
5.013121 
TcMT11.5633 
TcMT21.71046 
TcMT31.5–2.015122Pupal arrest (5th), 1
5.0191142Prepupal arrest (5th), 1; pupal arrest (7th), 1
Figure 7.

 Effects of TcMT3 RNAi-mediated knockdown in the larval stage in Tribolium castaneum. The larval instar from which each larva pupated is indicated in parentheses in each photograph. (A) Ventral view of pupae that were injected with dsRNA of EGFP or TcMT3 on day 0 of the third larval instar (day 0_3rd). Scale bar, 500 μm. (B) Dorsal and ventral views of adults injected with dsRNA of EGFP or TcMT3 in the larval stage. Scale bar, 500 μm.

As stated above, the TcMT1 transcript was expressed strongly in the last 2 days of adult development, whereas the TcMT2 transcript was expressed strongly at the beginning of the prepupal stage (Fig. 2B,D). To examine the role of TcMT1 and TcMT2 when expression levels are normally high, TcMT1 dsRNA was injected in the prepupal stage and TcMT2 dsRNA was injected at the beginning of the final larval instar. In both cases, quantitative RT-PCR confirmed that RNAi-mediated knockdown suppressed the transcript levels (Fig. 6E,F). However, all TcMT1 dsRNA-injected insects (n = 13) eclosed to form normal adults, and all TcMT2 dsRNA-injected insects (n = 4) pupated and eclosed normally.

Effects of JHM treatment on precocious metamorphosis induced by TcMT3 RNAi

To confirm that the observed precocious metamorphosis was a result of JH deficiency caused by TcMT3 knockdown, the JHM methoprene was topically applied to larvae that had been injected with TcMT3 dsRNA at the beginning of the fourth larval instar. As shown in Table 3, 84% of the larvae that received TcMT3 dsRNA and were then treated with the solvent precociously pupated at the end of the sixth instar. In contrast, 62% of the larvae (n = 21) that received TcMT3 dsRNA moulted into the seventh instar after treatment with JHM either at the fourth or fifth instar. If JHM was applied at the beginning of the sixth instar to the larvae that had received TcMT3 dsRNA, the majority (94%, n = 16) moulted into the seventh instar (Table 3). Thus, JHM application at the beginning of the sixth instar was more effective in rescuing TcMT3 RNAi-mediated precocious pupation than was JHM application in the fourth or fifth instar.

Table 3.   Phenotypes of Tribolium larvae injected with 5.0 μg·μL−1 dsRNAs on day 0 of the fourth instar, and treated with a JH mimic. Numbers and percentages of animals, and the instar when they pupated, are indicated. The insects that underwent precocious metamorphosis are shown in bold. Each larva was topically treated with 25 ng of methoprene (JHM) or the same volume of solvent as the control.
dsRNAHormonal treatmentNumber of pupae (%)
CompoundInstarn5th6th7th8th
EGFP(No treatment)4330 (70)13 (30)
Solvent4th–5th2321 (91)2 (9)
JHM4th–5th2318 (78)5 (22)
TcMT3Solvent4th–5th1916 (84)2 (11)1 (5)
JHM4th–5th218 (38)13 (62)
Solvent6th1514 (93)1 (7)
JHM6th161 (6)15 (94)

After injecting TcMT3 dsRNA and treating the larvae with JHM at the beginning of the sixth instar, 13 insects (n = 16) either arrested at eclosion or eclosed with the exuviae stuck on the elytra, whereas three eclosed successfully into adults with pupal-like urogomphi (data not shown). These phenomena may be the result of the effect of residual methoprene, as similar defects were also observed in wild-type larvae treated with JHM.

Discussion

In this study, we performed expressional and functional analyses of three methyltransferase genes (TcMT1, TcMT2 and TcMT3) identified from T. castaneum. Only TcMT3 was expressed strongly in the larval putative CA, the primary organ for JH biosynthesis. Recombinant TcMT3 protein methylated FA and JHA III, bur recombinant TcMT1 and TcMT2 proteins did not. Furthermore, RNAi-mediated knockdown of TcMT3 in the larval stage resulted in precocious metamorphosis into a pupa, presumably because of precocious shutdown of JH biosynthesis. These results demonstrate that TcMT3 encodes a functional JHAMT that is essential for JH biosynthesis. Hereafter, TcMT3 is called TcJHAMT.

TcJHAMT is expressed in a tissue-specific and stage-specific manner

In both B. mori and D. melanogaster, JHAMT mRNA was detected in large amounts in the larval CA [9,10]. In B. mori, JHAMT mRNA was detected in the third and fourth larval instars, but decreased rapidly at the beginning of the final (fifth) larval instar [9]. The JHAMT transcript of D. melanogaster was abundant in the larval stage, but was not detected in the pupal stage or during most of adult development [10]. These observations indicate that JHAMT is the key enzyme determining the timing of larval–pupal metamorphosis by controlling the rate of JH biosynthesis. In this study, we analysed the spatial and temporal expression patterns of a JHAMT orthologue in T. castaneum. The TcJHAMT transcript was expressed in the embryonic and larval stages, and decreased at the end of the final larval instar (Fig. 2E,F). In addition, the TcJHAMT transcript was detected specifically in the larval CA (Fig. 4). Although the developmental profile of JH titre has not yet been examined in T. castaneum, the temporal expression profile of TcJHAMT may correlate with JH biosynthetic activity in CA as observed in B. mori [17].

In B. mori, the BmJHAMT transcript is expressed specifically in CA until the beginning of the final larval instar [9]. In contrast, the BmJHAMT transcript is undetectable in CA at the beginning of the spinning stage, but is detected at low levels in the testis and ovary [9]. In D. melanogaster, the DmJHAMT transcript is expressed very strongly in the larval CA, and a small amount of DmJHAMT is also detected in the testis of wandering third instar larvae [10]. In this study, we found that the TcJHAMT transcript was expressed exclusively in the putative CA of the sixth instar (Figs 3C and 4), but the TcJHAMT transcript was detected in both the anterior and posterior parts of the body at the beginning of the prepupal stage (Fig. 3F). These results suggest that TcJHAMT is expressed in tissues other than CA in the prepupal stage.

Quantitative RT-PCR analysis showed that the TcJHAMT transcript exists in the prepupal stage. Recently, Parthasarathy et al. [18] have reported that the JH level in T. castaneum decreases just before entrance into the quiescent (prepupal) stage, but increases again during the prepupal stage. In the Cecropia silkworm and the tobacco hornworm M. sexta, JH reappears in the wandering stage just before pupation, and removal of CA from the final larval instar causes precocious adult differentiation of certain imaginal structures [19,20]. Whether JH in the prepupal stage of T. castaneum plays a role in preventing precocious adult development needs to be examined.

TcJHAMT methylates FA and JHA III

We have shown that recombinant TcJHAMT protein methylates (10R)-JHA III more favourably than (10S)-JHA III. JHAMT of D. melanogaster has also been reported to catalyse (10R)-JHA III preferentially over the (10S)-enantiomer [10]. To date, the absolute configuration of the chiral epoxide of natural JH III has been reported to be 10R in the lepidopteran M. sexta [21], coleopteran Tenebrio molitor [22] and orthopterans Schistocerca vaga and Locusta migratoria [23,24]. Although the chemical structure and stereochemistry of JH in T. castaneum has not yet been elucidated, it is probably the same as in other insect species.

In JH biosynthesis, FA is converted into active JH by methylation of the carboxyl group and epoxidation at the C10–11 position. Biochemical studies using CA homogenates from lepidopteran species suggest that FA is epoxidized into JH acid first, and then JH acid is methylated to JH [6,25]. In contrast, in other insect orders, such as Orthoptera and Dictyoptera, biochemical studies indicate that FA is methylated to MF, and then epoxidation occurs [6,26]. This observation is further supported by a recent study that showed that recombinant CYP15 protein of the cockroach Diploptera punctata epoxidizes MF but does not epoxidize FA [8]. In both D. melanogaster [10] and T. castaneum (Table 1), recombinant JHAMT protein methylates FA and JHA III at similar rates. Therefore, either order of reactions is possible for the late steps in JH biosynthesis in these species.

Functions of TcMT1 and TcMT2 genes

In this study, we have demonstrated that TcMT3 encodes a functional TcJHAMT gene. Although there are two more putative methyltransferase genes (TcMT1 and TcMT2) in the Tribolium genome, we conclude that they do not catalyse the methylation reaction in JH biosynthesis, because recombinant TcMT1 and TcMT2 proteins do not methylate FA or JHA III, and RNAi-mediated knockdown of TcMT1 or TcMT2 in larvae does not cause precocious larval–pupal metamorphosis. As TcMT1, TcMT2 and TcJHAMT are located in the same vicinity in the genome, and the positions of the introns are very similar in the three genes, they may have been derived through gene duplication events. In contrast with the CA-specific expression of the TcJHAMT transcript, the TcMT1 and TcMT2 transcripts are abundant in the posterior part of the sixth instar larvae (Fig. 3A–C). Interestingly, the temporal expression profiles of these three methyltransferase genes are quite different (Fig. 2), suggesting that the transcription of these genes may be regulated by hormones or other unknown factors in different ways.

At this point, the functions of TcMT1 and TcMT2 are unknown because the substrates for TcMT1 and TcMT2 have not been identified. TcMT1 and TcMT2 have putative SAM-binding motifs, and therefore it is likely that they methylate compounds with carboxyl groups, such as aliphatic or aromatic carboxylic acids. Further studies, such as in situ hybridization and enzyme assays using a variety of candidate substrates, are needed to elucidate the functions of TcMT1 and TcMT2.

Significant role of TcJHAMT in the regulation of JH biosynthesis and maintenance of the larval status

In this study, we have shown that RNAi-mediated knockdown of JHAMT in the larval stage causes precocious pupation. Importantly, this phenotype was rescued by the application of exogenous JHM, indicating that precocious metamorphosis is caused by precocious shutdown of JH biosynthesis. Therefore, we conclude that the JHAMT gene is essential for JH biosynthesis, and continuous expression in the larval stage is necessary for the maintenance of the larval status. Although the TcJHAMT transcript was suppressed significantly 3 days after dsRNA injection, i.e. day 0 of the fifth larval instar (day 0_5th; Fig. 6C), precocious metamorphosis did not occur until the end of the sixth larval instar in most cases. We assume that this time lag is caused by a long half-life for the TcJHAMT protein. Alternatively, it may take time for JH to be completely eliminated from the hemolymph because enzymes such as JH esterase and JH epoxide hydrolase are necessary for the degradation of JH in the hemolymph and tissues [27].

In some insect species, such as B. mori, it has been reported that precocious larval–pupal metamorphosis is caused by surgical removal of CA [28] or the application of chemicals with anti-JH action, such as the imidazole derivative KK-42 [29]. Recently, it has been reported that overexpression of the JH esterase gene in transgenic B. mori also results in precocious larval–pupal metamorphosis, probably as a result of precocious degradation of JH in the hemolymph [30]. As demonstrated in this study, RNAi-mediated knockdown of JH biosynthetic enzymes is a novel method to induce precocious metamorphosis. Although precocious metamorphosis can also be induced by the injection of dsRNA of the Methoprene-tolerant (Met) gene of Tribolium, probably a mediator of JH signals [31], most larvae arrest as prepupae, probably because Met function is necessary for normal pupation. In contrast, JHAMT RNAi results in miniature pupae and adults that appear normal in their external morphology.

RNAi-mediated knockdown by the injection of dsRNA into larvae or nymphs has also been reported to be effective in other insect species, such as lacewings [32], cockroaches [33–35] and milkweed bugs [36]. As demonstrated in this study, the RNAi technique is particularly useful to suppress JH biosynthesis in small insects for which it is extremely difficult to eliminate JH by traditional surgical methods. We anticipate that the RNAi technique will contribute to the elucidation of the physiological functions of JH and the molecular mode of JH action.

Materials and methods

Beetles

The wild-type strain of T. castaneum used in this study was provided by the National Food Research Institute, Tsukuba, Ibaraki, Japan. T. castaneum was raised in whole wheat flour at 30 °C. To collect eggs, adult beetles were kept in wheat flour for 1–3 days, and beetles and eggs were separated using sieves. To stage the larvae, they were individually raised in 24-well microtitre plates, and exuviae were checked every day. T. castaneum larvae do not develop synchronously: in our hands, they pupated either at the seventh or eighth larval instar. To distinguish the instar in which they pupate, the head capsule widths of early sixth and seventh instar larvae were measured using a microscope [Leica Microsystems MZ16FA/DFC500 system (Leica Microsystems, Heerbrugg, Switzerland)]. Larvae with head capsule widths of 566 ± 20 μm (mean ± SD; n = 30) in the sixth instar and 671 ± 22 μm (n = 37) in the seventh instar pupated at the end of the seventh larval instar. Larvae with head capsule widths of 529 ± 19 μm (n = 7) in the sixth instar and 633 ± 22 μm (n = 7) in the seventh instar pupated at the end of the eighth larval instar. Approximately 83% of larvae (n = 81) pupated at the end of the seventh larval instar, and 17% pupated at the end of the eighth larval instar. To investigate the developmental profile using quantitative RT-PCR, sixth instar larvae with head capsules wider than 570 μm were considered as penultimate instar larvae, and seventh instar larvae with head capsules wider than 690 μm were considered as final instar larvae, and were used for RNA isolation.

cDNA cloning of methyltransferase genes

tblastn searches were performed using the beetle genome database (http://www.bioinformatics.ksu.edu/BeetleBase/) with the sequences of B. mori and D. melanogaster JHAMT proteins, and a contig (Contig4620_Contig8031) containing three putative methyltransferase genes (TcMT1, TcMT2 and TcMT3) was identified. RT-PCR was performed to amplify the ORF of TcMT1 (828 bp) by Advantage 2 DNA Polymerase (Clontech Laboratories, Mountain View, CA, USA) with TcMT1_start and TcMT1_stop primers. Similarly, the TcMT2 ORF (846 bp) was amplified with TcMT2_start and TcMT2_stop primers, TcMT3 ORF (834 bp) with TcMT3_start and TcMT3_stop primers, and TcRp49 ORF (402 bp) with TcRp49_start and TcRp49_stop primers. It should be noted that the recognition site of the NdeI restriction enzyme was added to the 5′-end of TcMT1_start, TcMT2_start and TcMT3_start primers. The PCR products were subcloned into a pGEM-T vector (Promega Corporation, Madison, WI, USA). The DNA sequence data of TcMT1, TcMT2 and TcMT3 (TcJHAMT) were deposited in GenBank (accession numbers: AB360761 for TcMT1, AB360762 for TcMT2 and AB360763 for TcJHAMT). The sequences of the primers are listed in supplementary Table S1.

Quantitative RT-PCR analysis

The TcMT1, TcMT2 and TcMT3 transcripts were quantified using a real-time thermal cycler (LightCycler 2.0, Roche Diagnostics, Basle, Switzerland). Total RNA was isolated from the whole body of T. castaneum using an RNeasy Plus Mini Kit (Qiagen, Valencia, CA, USA). To analyse the developmental expression profile, several insects were combined for RNA isolation of the embryonic stage and the first, second and third larval instars, whereas RNA was isolated from individuals for the sixth and seventh larval instars, pupal and adult stages. To examine the tissue specificity of these genes in the sixth and seventh instars (at 84 h after ecdysis for the seventh instar), four larvae were cut in half between thoracic segments T2 and T3, and anterior and posterior parts were collected separately for RNA isolation. cDNAs were synthesized with an oligo(dT)18 primer and M-MLV reverse transcriptase (Clontech Laboratories). Quantitative RT-PCR was carried out in a 20 μL reaction volume containing SYBR Premix Ex Taq (Takara Bio, Shiga, Japan), 0.2 μm of each primer and 2–3 μL of template cDNAs or standard plasmids. PCR conditions were 95 °C for one 10 s cycle, followed by 40–50 cycles at 95 °C for 5 s and 60 °C for 20 s. The primers used for quantification are listed in supplementary Table S1. After PCR, the absence of unwanted byproducts was confirmed by melting curve analysis. For standards, serial dilutions of a plasmid containing the ORF of each gene were used. TcRp49 was used as a reference gene. Transcript levels of TcMT1, TcMT2 and TcMT3 were normalized with TcRp49 in the same samples. For each gene, the highest intensity in the developmental expression profile (Fig. 2) was set as 100%.

In situ hybridization

In situ hybridization was carried out according to a method reported for Drosophila brains [37]. The full coding region of TcMT3 was subcloned into a pGEM-T vector, and a linearized plasmid was used as the template for RNA synthesis. Digoxigenin (DIG)-labelled sense and antisense RNA probes were prepared using a DIG RNA Labelling Kit and SP6 or T7 RNA polymerase (Roche Applied Science, Mannheim, Germany), according to the manufacturer’s instructions. Heads of sixth instar larvae were dissected in NaCl/Pi, and most of the head capsules were carefully removed with forceps. Tissues were fixed in 4% paraformaldehyde at 4 °C for 40 min, and treated with 5 μg·mL−1 Proteinase K for 75 s. Re-fixation, hybridization and detection with pre-adsorbed, alkaline phosphate-conjugated anti-DIG FAB fragments and nitroblue tetrazolium/5-bromo-4-chloroindol-2-yl phosphate (Roche Applied Science) were performed as described previously [37,38]. After hybridization and detection, the remaining head capsule, fat body and muscles were carefully removed with forceps, so that the brain and CA could be seen well.

Preparation of recombinant proteins and enzyme assays

Full-length ORFs of TcMT1, TcMT2 and TcMT3 cloned into the pGEM-T vector described above were excised with NdeI and NotI restriction enzymes and subcloned into pET28a(+) expression plasmid vector (Novagen, Madison, WI, USA) that was linearized with the same restriction enzymes. The resulting constructs, TcMT1/pET28a(+), TcMT2/pET28a(+) and TcMT3/pET28a(+), were used individually to transform the Escherichia coli BL21(DE3) strain to express N-terminal 6× His-tagged recombinant TcMT1, TcMT2 and TcMT3 proteins, respectively. Expression and purification procedures were essentially the same as described previously [9]. After adding glycerol (final concentration, 25%), the purified protein aliquots were frozen in liquid nitrogen and stored at −80 °C until further analysis. Under these conditions, no obvious decrease in enzymatic activity was observed for at least 1 year.

Enzymatic analyses were performed essentially as described previously [9,10]. Briefly, purified recombinant methyltransferases were incubated individually in 500 μL of 50 mm Tris/Cl buffer (pH 7.5) containing SAM (500 μm) and one of the following acids as a substrate: FA (50 μm), racemic JHA III (50 μm), LA (100 μm) or PA (100 μm). After 5–60 min of incubation at 25 °C, the reactions were stopped by the addition of 500 μL of CH3CN and vortexing. Incubation times were adjusted so as not to consume more than 15% of the initial substrates. To analyse the generation of methylated FA (MF) and methylated JHA III (JH III), after removing the precipitate from the samples by centrifugation, the supernatants were analysed directly by RP-HPLC [10]. To analyse the methyl esters produced from LA and PA, the supernatants were extracted with hexane-containing 5 μg·mL−1 methyl tridecanoate as an internal standard and analysed by GC-MS [10]. To analyse the selectivity of recombinant TcMT3 protein between (10R)- and (10S)-enantiomers of JHA III, TcMT3 protein was incubated with racemic JHA III for 10 min, and the configuration of the metabolites extracted with hexane was analysed further by enantioselective HPLC with a ChiralPak IA column (250 mm × 4.6 mm inside diameter; Daicel, Osaka, Japan), as described previously [39].

RNAi experiments

Template DNA fragments for the synthesis of dsRNA were prepared by PCR as follows. A DNA fragment containing the 828 bp TcMT1 ORF with a T7 promoter sequence at the N-terminal end was amplified with TcMT1_start_T7 and TcMT1_stop primers to synthesize sense RNA, and a fragment containing the TcMT1 ORF with a T7 promoter sequence at the C-terminal end was amplified with TcMT1_start and TcMT1_stop_T7 primers to synthesize antisense RNA. RNA was synthesized using the T7 RiboMAX Express RNAi System (Promega Corporation). The in vitro transcription reaction was carried out at 37 °C for 1 h with T7 RNA polymerase, and the template DNA was digested by DNase I. Both strands of TcMT1 RNA were mixed and incubated at 70 °C for 10 min, and were gradually cooled to room temperature for annealing. dsRNA in the mixture was then purified by ethanol precipitation and dissolved in DNase/RNase-free water. To synthesize dsRNA for TcMT2, TcMT3 and EGFP, DNA fragments with T7 promoter sequences on both ends were amplified by PCR as follows: a fragment containing a 368 bp gene-specific region of TcMT2 (positions 1–368 in TcMT2 ORF) with T7 promoter sequences on both ends was amplified with TcMT2_start_T7 and TcMT2_3′_T7 primers, and a fragment containing a 359 bp TcMT3 sequence (positions 1–359 in TcMT3 ORF) with T7 promoter sequences on both ends was amplified with TcMT3_start_T7 and TcMT3_3′_T7 primers. The template for EGFP dsRNA (720 bp) was prepared with EGFP-F-T7 and EGFP-R-T7 primers and a plasmid containing the EGFP sequence (pEGFP, Clontech Laboratories). In vitro transcription was carried out at 37 °C for 1 h with T7 RNA polymerase, so that both strands would be synthesized simultaneously. The reaction mixture was then treated with DNase I, annealed and purified by ethanol precipitation. The primer sequences are listed in supplementary Table S1.

To knockdown genes in the larval stage, third or fourth instar larvae within 24 h after ecdysis were anaesthetized with ether for 3 min, aligned on double-sided tape, and dsRNA solution (approximately 20 nL for the third instar and 30–40 nL for the fourth instar) was injected into the abdomens at a concentration of 1.5–5.0 μg·μL−1 using a capillary tube pulled by a Narishige needle puller. To examine the efficiency of RNAi-mediated knockdown, total RNA was isolated individually from dsRNA-injected larvae several days later for quantitative RT-PCR. To knockdown TcMT1 in adult development, prepupae were injected with about 150 nL of 2.5 μg·μL−1TcMT1 dsRNA, and three pharate adults were homogenized individually for quantification of the TcMT1 transcript levels 6 days later. To knockdown TcMT2 in the final larval instar, newly moulted final instar larvae were injected with approximately 150 nL of 1.7 μg·μL−1TcMT2 dsRNA, and three larvae were homogenized individually for quantification of TcMT2 levels 3 days later. EGFP dsRNA was injected as a control. After injection, insects were reared individually in 24-well microtitre plates at 30 °C.

Hormonal treatments

Methoprene (SDS Biotech, Tokyo, Japan) was dissolved in methanol to 0.31 μg·μL−1 (1 mm), and this stock solution was diluted to 0.19 μg·μL−1 with acetone. Larvae that were injected with dsRNA on day 0 of the fourth larval instar (day 0_4th) were anaesthetized with ether for 2.5–3 min, aligned on double-sided tape, and approximately 130 nL of 0.19 μg·μL−1 methoprene solution (containing approximately 25 ng of methoprene) was topically applied on the dorsum using a 10 μL Hamilton microsyringe, on day 3_4th, day 0_5th or day 0_6th. The same volume of solvent was applied as a control. After hormonal treatment, larvae were individually reared in 24-well microtitre plates.

Acknowledgements

We thank Professor L. M. Riddiford (Janelia Farm, Howard Hughes Medical Institute, Ashburn, VA, USA), Professor D. Taylor (University of Tsukuba, Japan) and Professor H. Wojtasek (University of Opole, Poland) for critical comments on the manuscript, Dr A. Miyanoshita and Dr M. Murata (National Food Research Institute, Tsukuba, Ibaraki, Japan) for providing T. castaneum, Dr Y. Tomoyasu and Dr Y. Arakane (Kansas State University, Manhattan, KS, USA) for technical advice on RNAi experiments, and Professor S. Sakurai (Kanazawa University, Japan) for kindly providing methoprene. C. M. was supported by a research fellowship from the Japan Society for the Promotion of Science (JSPS). This work was supported by the Program for Promotion of Basic Research Activities for Innovative Biosciences (PROBRAIN).

Ancillary