Overexpression of Tyrosine hydroxylase and Dopa decarboxylase associated with pupal melanization in Spodoptera exigua

Melanism has been found in a wide range of species, but the molecular mechanisms involved remain largely elusive. In this study, we studied the molecular mechanisms of the pupal melanism in Spodoptera exigua. The full length cDNA sequences of tyrosine hydroxylase (TH) and dopa decarboxylase (DDC), two key enzymes in the biosynthesis pathway of melanin, were cloned, and their temporal expression patterns in the integument were compared during the larval-pupal metamorphosis process of the S. exigua wild type (SEW) and melanic mutant (SEM) strains. No amino acid change in the protein sequence of TH and DDC was found between the two strains. Both DDC and TH were significantly over-expressed in the integument of the SEM strain at late-prepupa and 0 h pupa, respectively, compared with those of the SEW strain. Feeding 5th instar larvae of SEM with diets incorporated with 1 mg/g of the DDC inhibitor L-α-Methyl-DOPA and 0.75 mg/g of the TH inhibitor 3-iodo-tyrosine (3-IT) resulted in 20% pupae with partially-rescued phenotype and 68.2% of pupae with partially- or fully-rescued phenotype, respectively. These results indicate that overexpressions of TH and DDC are involved in the pupal melanization of S. exigua.


Results
TH cDNA in the wild type and melanic strains of S. exigua. The full-length cDNA sequences of TH and DDC from S. exigua were obtained by RT-PCR cloning of a partial fragment, followed by cloning of their 5' and 3' ends by 5' and 3' RACE (random amplification of cDNA ends). The NCBI accession numbers of the cDNA sequences of TH and DDC from SEW and SEM were JF795467, JF795468, JF795469 and JF795470, respectively. The S. exigua TH (SeTH) cDNA (2259 bp in total) contains a 185-bp 5' UTR (untranslated region), an open reading frame (ORF) of 1686 bp encoding 561 amino acids, and a 386-bp 3' UTR ( Fig. 1). Sequence analysis shows that SeTH shares 97% amino acid identity with TH from Mythimna separata, 94% from Papilio xuthus, M. sexta and B. mori, 91% from Heliconius melpomene malleti and Samia cynthia ricini. Phylogenetic analysis also shows that SeTH is more closely related to TH from M. separate, which, like S. exigua, belongs to Noctuidae ( Fig. 2A). Similar to the structure of other TH proteins, SeTH contains the conserved metal binding site for iron atom (the circled histidine 389, histidine 394 and glutamate 432 in Fig. 1) and the cofactor tetrahydrobiopterin (BH 4 ) binding site (the boxed Leu251, leu252, phe257, gln267, pro385, glu390 and try429 in Fig. 1).
cDNA sequence alignment uncovers that the SeTH alleles from the wild type and melanic mutant strains only have 1 SNP (single nucleotide polymorphism) in the ORF, 1 SNP and 1 1-bp indel in the 3' UTR (underlined in Fig. 1). The only SNP in the ORF does not cause amino acid change.
DDC cDNA in the wild type and melanic strains of S. exigua. The S. exigua DDC (SeDDC) full-length cDNA is 1688 bp long, containing a 175-bp 5'UTR, an ORF of 1431 bp (from 176 to 1606) encoding 476 amino acids, and a 82-bp 3'UTR (Fig. 3). The predicted SeDDC proteins shows 95% amino acid sequence identity with Mamestra brassicae DDC, 94% with M. separate DDC, 91% with Antheraea pernyi and Papilio machaon DDC. SeDDC is grouped into the same lineage with M. separata and M. brassicae DDC, which share a common ancestor DDC with another lineage formed by the Papilionoidea DDC clades (Fig. 2B). Like other DDC, SeDDC has a cofactor binding site for pyridoxal 5'-phosphate (PLP). The PLP binding site is composed of ala148, ser149, thr152, thr242, asp271, try274, asn303 and lys306 (underlined in Fig. 3). Only one nucleotide difference was found between the DDC cDNA sequences from SEW and SEM. This substitution occurs in the 3'UTR and thus causes no amino acid change between the two DDC alleles (Fig. 3).
Overexpression of TH and DDC in the melanic SEM strain of S. exigua. Two-way ANOVA revealed significant differences in TH expression between strains and among the five developmental stages (Table 1A). There were also significant strain × stage interactions in TH expression. The normalized expression of TH in the integument of the wild type SEW strain was in a descending order of 0-h pupae (22.134 folds) >2-h pupae (4.587 folds) >late prepupa (0.936 folds) >24-h pupae (0.517 folds) > early prepuape (0.136 folds) (Fig. 4A). The temporal expression file of TH slightly differed in the melanic mutant strain SEM, which had a descending order of 0-h pupae (27.820 folds) >late prepupa (14.219 folds) >2-h pupae (3.703 folds) >24-h pupae (1.190 folds) >early puape (0.272 folds). Relative to the wild type SEW strain, the expression of TH was significantly higher in the late prepupae (15.191 folds) of the SEM strain (Fig. 4A).
Two-way ANOVA also revealed significant differences in DDC expression between strains and among the five developmental stages as well as significant strain × stage interactions in DDC expression (Table 1B). Comparison of Fig. 4A,B revealed that DDC transcripts were much less abundant than TH transcripts in both the wild and mutant strains. The normalized expression of DDC in the integument of the wild type strain SEW was ranked as 24-h pupae (0.145 folds) >0-h pupae (0.110 folds) >2-h pupae (0.070 folds) >late prepupae (0.046 folds) >early prepupae (0.044 folds) (Fig. 4B). The temporal expression profile of DDC in the integument of the mutant SEM strain was ranked as late prepupae (1.492 folds) >0-h pupae (0.339 folds) >2-h pupae (0.235 folds) >24-h pupae (0.184 folds) >early prepupae (0.075 folds). However, DDC transcripts were significantly more abundant in SEM than in SEW in late prepupae (32.435 folds) and 0-h pupae (3.082 folds), respectively (Fig. 4B).

Rescue of the mutant melanic phenotype by inhibiting TH and DDC.
To test if overexpression of TH and/or DDC is responsible for the mutant melanic phenotype in SEM, we fed 5 th instar larvae of SEM with control diets, diets incorporated with 0.25, 0.50, or 0.75 mg/g of 3-Iodo-L-tyrosine (3-IT,   Table 2), which were the same color as the SEM strain ( Fig. 5A1), whereas 6 out of the 30 pupae pupated from larvae fed on diets with 1.0 mg/g of L-α -Methyl-DOPA were dark-brown ( Fig. 5C; Table 2), i.e. partially rescued. The TH inhibitor 3-IT displayed even stronger rescue of the mutant phenotype. About half of the pupae from larvae fed on diet with 0.25 mg/g of 3-IT were partially recued and had a dark-brown color ( Fig. 5D and Table 2), while all the pupae from the PBS control diets remained black ( Fig. 5B2; Table 2). As the 3-IT dose increased, both the level of phenotype rescue (compare pupae in Fig. 5D-F) and the percentage of pupae with partially- (Fig. 5D,E) and fully-rescued (brown color pupae in Fig. 5F) simultaneously increased (Table 2), and the   5E) and even lighter ( Fig. 5F) than the wild-type morph (Fig. 5A2). The highest rescue rate was caused by 0.75 mg/g 3-IT (68.2%), which was not significantly higher than that elicited by 0.5 mg/g 3-IT (52.4%, P = 0.228) and 0.25 mg/g 3-IT (47.1%, P = 0.500). There was no significant difference between the rescue rates of 0.5 mg/g 3-IT and 0.25 mg/g 3-IT (P = 0.261).

Discussion
The great diversity of insect melanism provides rich models for studying the molecular and developmental regulation of phenotypic variations. All the melanic mutant phenotypes, whether occurred locally or globally in space/time and associated with or without a fitness cost, are concentration of one or both of the two insect melanins-Dopa melanin and Dopamine melanin 20,28,29 -on a particular body surface at a particular developmental time. This can only happen if the precursors of one or both of the two melanins, Dopa and Dopamine respectively, are overproduced, or less of them are channeled to synthesize the yellow and/or colorless pigments in the melanic insect strains. To test if one or both of the two melanin precursors are overproduced in the pupal melanic SEM strain of S. exigua, we contrasted the cDNA sequences and expression levels of TH and DDC, the two enzymes responsible for production of the Dopa and Dopamine precursors respectively, between the wild type SEW and melanic SEM strains. While we did not found any differences in the amino acid sequences of the two enzymes (Figs 1 and 3), we detected about 15-fold more TH transcripts in the early prepupae of SEM and more DDC transcripts in the late prepupae (32.435 folds) and 0-h pupae (3.082 folds) of SEM (Fig. 4). The fact that inhibiting the overproduced TH and DDC in SEM resulted in fully-and partially-rescued pupae respectively (Table 2 and Fig. 5) demonstrates that overproduction of TH and DDC is responsible for the pupal melanism in the SEM strain of S. exigua. Whether less consumption of the Dopa and Dopamine precursors for synthesis of the yellow and colorless pigments also contributes to SEM's melanic phenotype has yet to be determined.
As far as we know, our study is the first report of molecular mechanisms underlying melanism occurred only in the pupal stage of lepidoperan insects. Among the limited cases studying the mechanisms of lepidoperan melanism, 2 involved melanization in the larval stage 9,24 , 2 in the adult stage 26-28 , 1 in the larval and adult stage 20 , and 1 in the larval and pupal stage 21 . Overproduction of the Dopamine precursor due to enhanced DDC expression is linked to the larval melanism of M. sexta 24 . By contrast, less consumption of the Dopa precursor for synthesis of the yellow pigment by simultaneously-reduced expression of ebony and DDC causes partial melanization on the female adult wing of P. glaucus [25][26][27] . Similarly, less consumption of the Dopamine precursor for synthesis of the yellow or colorless pigment by deletion of the ebony ORF or the 4 th exon of AANAT is linked to the so (larval and pupal melanization 21 ) and mln (larval and adult melanization 20 ) mutants of B. mori, respectively. Thus, our finding of TH overexpression as the major contributor to the pupal melanism of S. exigua also represents the first report of the involvement of TH in insect melanism mutation. This finding is consistent with an earlier report that reduced expression of TH is linked to the light neonate color in the sex-linked chocolate (sch) mutant of B. mori 30 .
While we have discovered that overexpression of TH and DDC is responsible for the pupal melanization in the SEM strain of S. exigua, we have not identified the actual alterations that lead to increased expression of the two genes. One possibility would be an alternation in a transcription factor or small RNA that regulates expression of both TH and DDC. Another possibility would be some indel or SNP mutations in the 5' promoter regions of both genes that separately enhance expression of the two genes. The third possibility would be an alternation in the promoter region of TH or its transcription factor/ small RNA that only increases TH expression, but the resultant extra Dopa precursor induces the expression of the downstream DDC gene. Clearly, molecular cloning of the promoter regions of the two genes is the first step needed to resolve these speculations.

Methods
Experimental animals. The wild-type strain (SEW, brown color pupae) of S. exigua was established with a field collection from Jingzhou, Hubei, China in 2003 and has been maintained on artificial diet (a solid mixture of soybean flour, barley flour, yeast powder, Vitamin C, benzoic acid, sodium benzoate, acetic acid and agar) at 27 ± 1 °C, 65% relative humility and a photoperiod of 14:10 (L: D). At the 6 th generation, forty black pupae spontaneously occurred within the SEW strain. The adults emerged from the 40 black pupae were crossed to establish the pupal melanic strain (named SEM). The SEM strain has then been maintained on the same artificial diet at the same conditions with the SEW strain since then. TH and DDC cDNA Cloning. For each strain, three prepupa integuments were dissected out on a filter paper placed on ice, washed with cold 0.1% DEPC-treated water, briefly dried on a clean filter paper, transferred to a ceramic mortar with liquid nitrogen, and ground to powder in liquid nitrogen. The resultant prepupa intergument power of each strain was immediately transferred into a 1.5-mL centrifuge tube with about 500 μ L of Trizol TM reagent (Invitrogen, USA). Total RNA was isolated from the power according to the manufacturer's instructions. RNA purity was evaluated by 260/280 and 260/230 ratios measured using NanoDrop 1000 (NanoDrop, USA). DNase I (Fermentas, EU) treated RNA (1 μ g) was reverse transcribed into cDNA in a 20 μ L reaction with random primer (N9) and Oligo dT(18) primer using First-Strand cDNA Synthesis Kit (Fermentas, EU).
The primers for cloning of TH partial cDNA (Table S1) were designed based on the conserved regions of tyrosine hydroxylases of insect. The PCR conditions were 3 min denaturation at 95 °C, followed by 40 cycles of 30 sec denaturation at 94 °C, 30 sec annealing at 50 °C and 1 min extension at 72 °C, and a final 9 min extension at 72 °C. The primers for cloning of DDC partial cDNA (Table S1) were designed on the basis of S. exigua dopa decarboxylase (DDC) gene (NCBI accession no.: SEU71404). The PCR conditions were 3 min denaturation at 95 °C, followed by 40 cycles of 30 sec denaturation at 94 °C, 30 sec annealing at 56 °C and 30 sec extension at 72 °C, and a final 9 min extension at 72 °C. The full-length cDNAs of the two genes were obtained by rapid amplification of cDNA ends (RACE) technique using the 3'-Full RACE Core Set Ver. 2.0 (TaKaRa, Japan) and 5'-Full RACE Kit (TaKaRa, Japan) with the corresponding specific primers (Table S1) designed based on their partial cDNA sequences. The open reading frame (ORF) of TH and DDC were amplified from SEW and SEM strains with the corresponding flanking PCR primers (Table S1), respectively. The flanking PCR conditions for TH gene was 3 min denaturation  Table 2. Statistics of pupae color changes resulted from feeding larvae of the melanic SEM strain of S. exigua with TH and DDC inhibitors. 1 Rescue rate = (the number of dark-brown and brown pupae/the total number of pupae) × 100%.
Scientific RepoRts | 5:11273 | DOi: 10.1038/srep11273 at 95 °C, followed by 40 cycles of 30 sec denaturation at 94 °C, 30 sec annealing at 54 °C and 2 min extension at 72 °C, and a final 9 min extension at 72 °C. The flanking PCR conditions for DDC gene was 3 min denaturation at 95 °C, followed by 40 cycles of 30 sec denaturation at 94 °C, 30 sec annealing at 50 °C and 2 min extension at 72 °C, and a final 9 min extension at 72 °C. All the fragments and full ORF were cloned into pMD18-T Vector (TaKaRa, Japan) and at least three clones were sequenced for each fragment or ORF from the two strains.
Sequence alignment and phylogenetic analysis. The nucleotide and deduced amino acid sequences of TH and DDC were aligned to identify potential changes between the wild type (SEW) and mutant (SEM) strains. The conserved domains of the deduced protein sequences were localized by blast against the conversed domain database (CDD) (http://blast.ncbi.nlm.nih.gov/). The protein sequences retrieved from NCBI database and used for phylogenetic analysis were TH from the two strains of Real-time RT-PCR analysis of TH and DDC expression. Total RNA was isolated from the dissected integuments from both strains at five time points (early prepupa, late prepupa, 0 h pupa, 2 h pupa, 24 h pupa) and reverse transcribed into cDNA as described above. The cDNAs were used as the templates for quantitative RT-PCR (qRTPCR) analysis of TH and DDC expression using SYBR ® Premix Ex Taq TM II kit (TaKaRa, Japan). Real-time PCR of TH, DDC, β-actin and GAPDH (β-actin and GAPDH were internal reference genes) were performed individually in a 20 μ L reaction containing 2× SYBR ® Premix Ex Taq TM II 10 μ L, 10 μM forward primer and reverse primer (0.8 μ L each), 1 μ L template cDNA, and 7.4 μ L nuclease-free water. Specific primers (Table S1) 30,34 . The DDC inhibitor L-α -Methyl-DOPA (Sigma) water solution was mixed with artificial diets at 1.0, 0.5, and 0.1 mg/g diets according to a previous study 35 . The control diets for 3-IT and L-α -Methyl-DOPA were diets containing equal volume of PBS solution and water, respectively. For each inhibitor or control diets, 40 newly molted 5 th instar larvae of the melanic SEM strain were fed with the corresponding diets till pupation. The body colors of the SEM pupae were observed and photos were taken with a digital camera (OLYMPUS C5060) under white light. The numbers of pupae and adults obtained were recorded at the end of this experiment. The black pupae resulted from this inhibitor feeding experiment were considered as mutant melanic pupae, whereas the dark-brown and brown pupae obtained were considered as partially-and fully-rescued wild type pupae, respectively. For each concentration of inhibitor, the numbers of melanic, partially-rescued, and fully-rescued pupae were recorded and the rescue rate was calculated using the formula of the number of partially-and fully-rescued pupae/ the total number of pupae × 100%.
Statistical analysis. Two-way ANOVA was performed to examine the impacts of strain, stage and strain × stage interaction on the expressions of TH and DDC. Paired t-tests were conducted to test the significance of differences in the expression levels of TH and DDC at each of the five developmental stages between the two strains. Tukey HSD tests were performed to test the significance of differences in the expression levels of TH and DDC among the five developmental stages of each strain. The differences in the rescue rate among different concentrations of 3-IT were analyzed with one-tailed Fisher's exact test. All of the statistical tests were performed by SPSS version 16.0 software for Windows.