Introduction

Species-specific sex pheromones play a key role in sex communication and subsequent mating of most moth species (Vogt 2005). To date, sex pheromones of more than 640 lepidopteran species have been reported (Ando et al. 2004). Moth pheromones usually are a blend of two or more components, and are generally divided into two types based on the presence (Type-I) or absence (Type-II) of a terminal functional group in the components (Ando et al. 2004; Millar 2000). Type-I pheromones are used by most moths, and composed of C10–C18 unsaturated acyclic aliphatic compounds with a functional group such as a formyl, hydroxyl, or acyloxyl group (Ando et al. 2004; Witzgall et al. 2004). Type-II pheromones mainly are composed of C17-23 hydrocarbons with 2 or 3 double bounds at the 3, 6, or 9 positions, and their corresponding epoxy derivatives (Arn et al. 1997). Moth pheromones especially the Type-I pheromones are produced and released by specialized pheromone glands (PG) located along the inter-segmental membrane between abdominal segments 8th and 9th of females (Raina et al. 2000; Tillman et al. 1999).

A general pathway of biosynthesis for Type-I pheromones has been established starting with a palmitic or stearic acid synthesized de novo in PG through modifications of fatty acid biosynthetic pathway (Ando et al. 2004). By combinations of several enzymatic reactions (i.e., desaturation, chain-shorting reaction, reduction, acetylation, and oxidation), the palmitic or stearic acids are converted step-wise to the final pheromone components (Moto et al. 2004; Park et al. 2008; Tillman et al. 1999). So far, classes of essential enzymes involved in moth pheromone synthesis have been identified biochemically or/and molecularly. They are: 1) Desaturases, to introduce double bonds into pheromone precursors, are the most intensively studied class of enzymes involved in sex pheromone biosynthesis. Many desaturase genes acting on certain carbon chain positions have been functionally characterized, including Δ5 (Foster and Roelofs 1996; Hagstrom et al. 2014), Δ6 (Wang et al. 2010), Δ9 (Liu et al. 1999; Park et al. 2008), Δ11 (Fujii et al. 2011; Jeong et al. 2003), Δ10-12 (Moto et al. 2004), and Δ14 desaturases (Roelofs and Rooney 2003). 2) Fatty acid reductases (FAR), responsible for reducing fatty acids to alcohols, also have been functionally identified in some moth species, including pgFAR-Z/E of Ostrinia nubilalis (Lassance et al. 2010) and pgFAR of Bombyx mori (Moto et al. 2003). 3) Alcohol oxidase (AO) and acetyltransferase (ACT) have been suggested by biochemical studies to oxidize alcohols into the corresponding aldehyde components (Teal and Tumlinson 1987; Wang et al. 2010) and to convert alcohols into acetate esters (Jurenka and Roelofs 1989), respectively. However, none of the genes encoding these two enzyme classes has been functionally characterized. In addition, fatty acid transport proteins (FATP) and acyl-CoA binding proteins (ACBP) have been found to play roles in the production of bombykol in B. mori by RNA interference (Ohnishi et al. 2006, 2009). Although the subject has been studied intensely, unknown classes of genes that might be involved in pheromone production are still likely to be elucidated, for example, acetyl-CoA carboxylases (ACC), acyl-CoA oxidases (ACO), aldehyde reductases (ALR), elongation of fatty acids proteins (ELO), and short-chain dehydrogenases (SCD) (Vogel et al. 2010). These enzymes are potentially important in moth sex pheromone biosynthesis. Exhaustively identifying candidate genes from different species is important to better understand pheromone biosynthesis and speciation in moths.

In the males, three major classes of protein are involved in the perception of female sex pheromones. They are pheromone binding proteins (PBPs), odorant receptors (ORs), and pheromone degrading enzymes (PDEs). PBPs are thought to bind hydrophobic pheromone molecules and transport them through the aqueous sensillum lymph to reach the membrane surface of the sensory neuron where ORs are situated (Field et al. 2000; Leal 2013; Pelosi et al. 2006; Vogt 2005). The binding of odorant molecule alone or in complex with PBPs activates ORs and initiates a signal transduction cascade leading to an electrical signal. After OR activation, the pheromone molecules are thought to be rapidly inactivated to restore the sensitivity of ORs for receiving new chemical signals (Vogt 2003, 2005; Vogt and Riddiford 1981). Pheromone inactivation is crucial for maintaining high sensitivity of pheromone reception, and the process has been demonstrated to be accomplished by enzymatic degradation in the sensillar lymph (Ferkovich et al. 1980; Kasang 1971; Leal 2013; Mayer 1975; Vogt and Riddiford 1981). However, only a few PDE genes have been functionally characterized (Chertemps et al. 2012; Choo et al. 2013; Durand et al. 2010, 2011; Ishida and Leal 2005, 2008), and the inactivation mechanisms are not well understood.

The purple stem borer, Sesamia inferens (Lepidoptera: Noctuidae), is a polyphagous insect pest found in many Asian countries (Chai and Du 2012). It damages various crops, including rice, corn, and sugarcane, and since the 1990s has become one of the major rice pests in China (Gao et al. 2010; Xu et al. 2011). The female sex pheromone of S. inferens is a blend of (Z)-11-hexadecenyl acetate (Z11-16:OAc), (Z)-11-hexadecenol (Z11-16:OH), and (Z)-11-hexadecenal (Z11-16:Ald) (Zhu et al. 1987), which are molecules containing three distinct terminal functional groups typically found in Type-I sex pheromones. Thus S. inferens serves as a good model to study enzymes involved in the biosynthesis, as well as the degradation of sex pheromones.

In the present study, we took advantage of previously obtained transcriptomic data from adult antenna and female sex pheromone gland of S. inferens (Zhang et al. 2013) and identified 73 putative genes that might be involved in sex pheromone biosynthesis and degradation. Tissue expression patterns of these genes were investigated using qPCR, and phylogenetic analyses were performed in an effort to predict gene function. The analysis revealed that some identified genes are specifically or highly expressed in antennae or in the female sex pheromone glands. Based on the results, we proposed a putative pathway of sex pheromones biosynthesis and degradation in S. inferens and highlighted gene candidates for further functional studies.

Methods and Materials

Insects and Tissue Collection

Larvae of the purple stem borer, S. inferens, originally were collected from a rice field at the Jiangsu Provincial Academy of Agricultural Sciences, Nanjing, China, and reared on fresh wild rice stems in glass bottles (diam = 7 cm, height = 11 cm) until pupation and sexing (Zhang et al. 2012). Rearing conditions were 28 ± 1 °C, 70–80 % RH and a 14:10, L, D photoperiod. Adults were provided with a cotton swab dipped in 10 % honey solution, and renewed daily. All tissues were collected from 3-d-old virgin male and female adults at 5–7 h into the scotophase, and immediately stored at−70 °C until use. Sex pheromone glands were extracted from female abdomens.

RNA Isolation and cDNA Synthesis

RNA was isolated from antennae, legs, and wings of 20–30 individuals; abdomens and PGs of 15–20 individuals; and fat bodies, thoraxes, and epidermises of 10–15 individuals. Total RNA was extracted using the SV 96 Total RNA Isolation System (Promega, Madison, WI, USA) following the manufacturer’s instructions. RNA quality was checked with a spectrophotometer (NanoDropTM 1000, Thermo Fisher Scientific, Waltham, MA, USA). Single-stranded cDNA templates were synthesized from 1 μg of total RNA from various tissue samples using the PrimeScript™ RT Master Mix (TaKaRa, Dalian, China).

Sequence Retrieval, RACE Amplification and Sequence Analysis

All putative genes were retrieved from previously obtained transcriptomic data (Zhang et al. 2013) that was reassembled using Trinity (v2012-10-05) (Novogene, Beijing, China). The SMART™ RACE cDNA Amplification Kit (Clontech, Mountain View, CA, USA) was used to amplify the 5′ and 3′ regions of target genes following the manufacturer’s instructions. The RACE PCR products were subcloned into the pEASY-T3 cloning vector (TransGene, Beijing, China) and positive clones were sequenced by GenScript (Nanjing, China). Full-length sequences were determined by assembling the cDNA fragments and the sequences obtained from the 5′ and 3′ RACE PCR. The RACE primers (Table S1) were designed using Primer Premier 5.0 (PREMIER Biosoft International, CA, USA).

Open reading frames (ORFs) of the putative genes were predicted using ORF finder (http://www.ncbi.nlm.nih.gov/gorf/gorf.html). The similarity searches were performed using the NCBI-BLAST network server (http://blast.ncbi.nlm.nih.gov/).

Phylogenetic Analysis

Phylogenetic trees were constructed by phylogenetic analyses of SinfDess, SinfFARs, SinfCXEs and SinfAOXs, based on the amino acid sequences of these genes and sequences of other insects. The desaturase (Des) data set contained six sequences from S. inferens, and 53 from other insects. The FAR data set contained three sequences from S. inferens, and 54 from other insects. The CXE data set contained 15 sequences from S. inferens (amino acids >80 aa), and 69 from other insects. The AOXs data set contained three sequences from S. inferens, and 11 from other insects. The protein names and accession numbers of the genes used for phylogenetic tree construction are listed in Table S2. Amino acid sequences were aligned using ClustalX 2.0, and unrooted trees were constructed by MEGA5.0 using the Neighbor-joining method, with Poisson correction of distances. Node support was assessed by bootstrap using 1,000 bootstrap replicates.

Reverse Transcription-PCR Analysis

Gene-specific primers across ORFs of predicted chemosensory genes were designed using Primer Premier 5.0 (PREMIER Biosoft International, CA, USA) and were listed in Table S1. PCR experiments including negative controls (no cDNA template) were carried out in a MyCycler™ (Bio-Rad, USA) under the following conditions: 94 °C for 4 min; 28–35 cycles at 94 °C for 30 sec, 60 °C for 30 sec, and 72 °C for 40 sec, with a final incubation for 10 min at 72 °C. The reactions were performed in 25 μl with 15 ng of single-stranded cDNA, 2.0 mM MgCl2, 0.2 mM dNTP, 0.4 μM for each primer and 1.25 U rTaq DNA polymerase (TaKaRa, Dalian, Liaoning, China). PCR products were analyzed by electrophoresis on 1.5 % w/v agrose gel in TAE buffer (40 mM Tris-acetate, 2 mM EDTA), and the resulting bands were visualized with ethidium bromide. The gene encoding S. inferens glyceraldehyde-3-phosphate dehydrogenase (SinfGAPDH) was used as a reference gene for checking the integrity of the cDNA template and expression quantification of the target genes. To check the repeatability of the tissue expression measurements, 15 genes were randomly chosen to perform a second biological replicate.

Quantitative Real-Time PCR

Quantitative Real-Time PCR (qPCR) was performed on an ABI 7500 (Applied Biosystems, Foster City, CA, USA) using a mixture of 10 μl 2× SYBR Green PCR Master Mix (TaKaRa, Dalian, Liaoning, China), 0.4 μl of each primer (10 μM), 2.5 ng of sample cDNA, and 6.8 μl sterilized ultrapure H2O. The reaction programs were 30 sec at 95 °C, 40 cycles of 95 °C for 5 sec, and 60 °C for 34 sec. The results were analyzed using the ABI 7500 analysis software SDS 1.4. The qPCR primers (Table S1) were designed using Beacon Designer 7.7 (PREMIER Biosoft International, CA, USA). Amplification was followed by the measurement of fluorescence during a 55 to 95 °C melting curve in order to detect whether a single gene-specific peak was observed and to check for the absence of primer dimer peaks. A single and discrete peak was detected for all primers tested. Negative controls were reactions without templates (replacing cDNA with H2O).

Expression levels of 46 putative sex pheromone biosynthesis genes were calculated relatively to the reference gene using the Q-Gene method in the Microsoft Excel-based software of Visual Basic (Muller et al. 2002; Simon 2003). To ensure the reliability of data analysis, we chose SinfGAPDH and Sinf28SrRNA as reference genes. Each sample had three biological replicates each with three technique replicates.

Statistical Analysis

Data (mean ± SE) from various samples were subjected to one-way nested analysis of variance (ANOVA) followed by a least significant difference test (LSD) for mean comparison. Two-sample analysis was performed by Student t-test using SPSS Statistics 17.0 (SPSS Inc., Chicago, IL, USA).

Results

Identification of Putative Genes Involved in Sex Pheromone Biosynthesis and Degradation

By blast analysis of the transcriptome data of S. inferens (Zhang et al. 2013) using the reported genes of other moth species in NCBI as queries (Gu et al. 2013; Strandh et al. 2008; Vogel et al. 2010), we identified a total of 46 genes encoding proteins putatively involved in sex pheromone biosynthesis. These included six Dess, three FARs, five FATPs, three ACBPs, five ACTs, seven ACCs, four ACOs, six ALRs, four ELOs, and two SCDs (Table 1). A total of 27 genes encoding putative odorant-degrading enzymes were identified, including 18 carboxylesterases (CXEs), three aldehyde oxidases (AOXs), and six alcohol dehydrogenases (ADs) (Table 2).

Table 1 The best blastx matches of putative sex pheromone biosynthesis genes of Sesamia inferens
Full size table
Table 2 The best blastx matches of putative sex pheromone degradation genes of Sesamia inferens
Full size table

Among the 73 newly identified genes, three Dess, one FAR, two FATPs, one ACBP, ten CXEs, three AOXs, and four ADs contained complete open reading frames (ORF) (Tables 1 and 2), which were obtained either by transcriptome analysis (17 genes) or by RACE methodology (seven CXEs).

Phylogenetic Analysis

In order to assign putative functions to the 73 genes, phylogenetic analyses were conducted for each group of the enzymes. A phylogenetic tree of the desaturases (Fig. S1) showed that four S. inferens desaturases clearly clustered in four different groups of insect desaturases; including the Δ11-desaturases (SinfDes4), Δ9-desaturases (18C > 16C) (SinfDes2), Δ9-desaturases (16C > 18C) (SinfDes3), and Δ9-desaturases (14C-26C) (SinfDes5). In the FAR phylogenetic tree, all three S. inferens FARs clustered within the lepidopteran pgFAR group, which contains previously identified FARs involved in moth sex pheromone biosynthesis (Hagstrom et al. 2012) (Fig. S2).

According to Durand et al. (2010b), a set of currently available insect carboxylesterase sequences was used to construct a CXE phylogenetic tree. The tree revealed that the 15 S. inferens CXEs divided into three different groups; an intracellular clade (A–C) (five genes), an extracellular clade (D–G) (nine genes) and a neurosignalling clade (K–M) (one gene) (Fig. S3). All three S. inferens AOXs clustered within the lepidopteran antennal AOX group. Within this group, two aldehyde oxidase, MbraAOX and AtraAOX2, have been reported to be involved in the degradation of aldehyde pheromone components (Choo et al. 2013; Merlin et al. 2005) (Fig. S4).

Tissue Expression Profile of the Putative Sex Pheromone Biosynthesis Genes

qPCR analyses of the expression patterns of 46 putative sex pheromone biosynthesis genes showed that only three genes, SinfDes4, SinfFAR2, and SinfFATP2, displayed PG-predominant or biased expression (Fig. 1). The others were expressed at relatively higher levels in fat body, leg or thorax.

Fig. 1
figure 1figure 1

Expression patterns of putative sex pheromone biosynthesis genes in different tissues of female Sesamia inferens. PG, female pheromone glands; FB, fat bodies; L, legs; W, wings; T, thoraxes and E, epidermis

Full size image

Tissue Expression Profile of the Putative Sex Pheromone Degradation Genes

The expression patterns of all 27 putative ODE genes characterized by RT-PCR showed ten genes with antennae-specific or biased expression (Fig. 2 and Fig. S5). SinfCXE10 was expressed specifically in the antennae of both sexes. SinfCXE13, 14, 18, and 20 also had highly antennae-biased expression. Further qPCR tests indicated that these five S. inferens CXE genes all had similar expression levels between male and female antennae.

Fig. 2
figure 2

Expression patterns of putative odorant-degrading enzyme (ODE) genes in Sesamia inferens. (a) Expression of all ODE genes by RT-PCR. GAPDH gene was used as a positive control and NC (no cDNA template) as a negative control. PG, female pheromone glands; A, antennae; T, thoraxes; Ab, abdomens (female without PG); L, legs; W, wings. (b–d) Relative expression levels of ODE genes in adult antennae of both sexs. ♀, female; ♂, male. An asterisk indicates a significant difference between male and female expression levels (P < 0.05, Student t-test), whereas “NS” indicates no significant difference (P > 0.05, Student t-test)

Full size image

Among the three S. inferens AOX genes, SinfAOX1, and SinfAOX2 were specifically expressed in antennae and male-biased (significantly higher in males than in females), whereas SinfAOX3 was expressed equally in the antennae of both sexes and male abdomens (Figs. 2a, c, and Fig. S5). Of the six S. inferens AD genes, SinfAD1 and SinfAD3 were expressed at higher levels in antennae than in other tissues (antenna bias), but the expression levels were similar between sexes (Figs. 2a, d, and Fig. S5).

To check the repeatability of the tissue expression, 15 genes were randomly selected to do a second measurement using different cDNA templates, and the results were consistent with those from the first measurement (Fig. S5).

Discussion

The present study has identified 73 new genes, including 46 putative genes involved in sex pheromone biosynthesis and 27 in pheromone degradation, through analysis of the transcriptomic data of adult antennae and female sex pheromone glands of S. inferens (Zhang et al. 2013). Phylogenetic analyses and tissue expression profiling highlighted two pheromone biosynthesis genes and ten pheromone degradation genes that are good candidates for further functional studies and could potentially be used as target genes for pest control purposes.

The three sex pheromone components (Z11-16:OAc, Z11-16:OH, and Z11-16:Ald) of S. inferens (Zhu et al. 1987) all contain an Δ11-double bond. The biosynthesis pathway of Δ11-containing sex pheromone has been studied in several moth species. The defined pathway involves a step of Δ11-desaturation catalyzed by a Δ11-desaturase (Fujii et al. 2011; Hao et al. 2002; Liu et al. 2002b; Roelofs et al. 2002; Wu et al. 1998). It is likely that a Δ11-desaturase is responsible for the introduction of the Δ11-double bond in the pheromone components of S. inferens. It is also possible, although rarely reported (Liu et al. 2002a), that a Δ9-desaturase can participate in the production of Δ11-containing pheromone component by introducing a Δ9-double bond at 14:CoA, followed by a carbon chain elongation to Δ11-16:CoA. In the present study, six desaturase genes were obtained, but only one (SinfDes4) displayed a PG-predominant expression pattern. Further phylogenetic analysis showed that SinfDes4 is clearly assigned to the Δ11-desaturase group, closely related to MbraZ11 of Mamestra brassicae (GenBank accession no. ABX90049). SinfDes5 is assigned to the Δ9-(14-26C)-desaturase group, while the remaining three genes belong to other desaturase groups. Therefore, SinfDes4 is likely involved in the desaturation of 16C saturated acids to the unsaturated acids, with a double bond introduced at 11th position of the carbon chain. However, the involvement of SinfDes5 in this process can not be ruled out, with further studies required.

Previous research revealed that in the process of sex pheromone biosynthesis, once the specific unsaturated fatty acid precursors are produced, they will be converted into corresponding alcohols by FARs (Hagstrom et al. 2012; Lassance et al. 2010; Lienard et al. 2010; Moto et al. 2003). In S. inferens, we identified three FARs, but only SinfFAR2 is expressed at significantly higher levels in PG than in other tissues, suggesting a role in the biosynthesis of Z11-16:OH. The other two FAR-encoding genes that were identified might not be involved in this process, as their expressions in PGs were not higher than in other tissues, even though they were clustered in the moth pgFAR group in the phylogenetic analysis. The third gene class showing high levels of expression in PG are the SinfFATP genes, of which SinfFATP1 was the most highly expressed. FATPs have been functionally confirmed to bind and transport fatty acids across insect hemolymph into PG cells for pheromone biosynthesis in B. mori (Ohnishi et al. 2009) and Eilema japonica (Qian et al. 2011). SinfFATP1 may play a similar role in pheromone biosynthesis of S. inferens.

Alcohol oxidases (AOs) and acetyltransferases (ACTs) are essential in the biosynthesis of formyl and acyloxyl components, respectively, although no such gene encoding such an activity has been identified. In our study, we found one AO and five ACT genes, but none of them showed PG-predominant expression. However, expression patterns alone cannot exclude the possibility that these genes might be involved in sex pheromone biosynthesis, as they may also play roles in other physiological processes. Another possibility is that genes with PG-specific or biased expression patterns might not be revealed by transcriptome sequencing due to very low expression levels. In addition, mRNA levels are not always consistent with protein levels (Lee et al. 1999; Newman et al. 2006). There might be genes being PG-specific or biased at protein level that do not show such patterns at the mRNA level. Alternative approaches are needed to detect candidates of AO and ACT genes that are specifically expressed in PGs.

In general, insect ODEs have the ability to attack specific functional groups, such as acetate esters, aldehydes, and alcohols (Vogt 2005). S. inferens should have ODEs of all three subclasses in the antennae as the sex pheromone contains components of all three functional groups. In the present study, 18 CXEs were identified, of which SinfCXE10 was antennae-specific and another four SinfCXEs were antennae-biased (SinfCXE13, 14, 18, and 20), suggesting that they may play a role in the degradation of Z11-16:OAc or plant volatile esters. To date, only a single AOX gene has been functionally characterized in insects from Amyelois transitella (Choo et al. 2013). Of the three SinfAOX candidates identified in the current study, SinfAOX1 and SinfAOX2 were antennae-specific and SinfAOX3 was predominantly expressed in antennae. SinfAOX1 and SinfAOX2 also were significantly male biased in their expression, which strongly suggests a role in the degradation of the Z11-16:Ald component. As for ADs acting in the degradation of alcohol pheromone components, no gene has been reported, although the existence of AD genes was suggested by a biochemical study two decades ago (Kasang et al. 1989). We obtained six SinfADs from S. inferens antennae, and among those, SinfAD1 and SinfAD3 were predominantly expressed in antennae, but not male biased in their expression. It is possible that those two genes have functions in degradation of plant volatiles or of both plant volatiles and sex pheromones. The ability of the same enzyme to degrade both sex pheromones and plant volatiles may constitute an efficient system that could participate to maintain the high sensitivity of the sex pheromone detection system, as illustrated by SlitCXE7 from S. littoralis (Durand et al. 2011) and AtraAOX2 from A. transitella (Choo et al. 2013).

Finally, based on the general moth sex pheromone biosynthesis pathway suggested in previous studies (Albre et al. 2012; Choi et al. 2005; Jurenka 2004; Ohnishi et al. 2006; Roelofs et al. 2002; Tillman et al. 1999), and the results obtained in the current study, we propose a putative biosynthesis and degradation pathway of the sex pheromone in S. inferens (Fig. 3). In the diagram, several candidate genes for some key steps are suggested and in our opinion they should be considered first for further functional studies in the future. For other steps where no candidate gene is suggested, more studies with alternative approaches are required.

Fig. 3
figure 3

A proposed biosynthesis and degradation pathway for the sex pheromone components of Sesamia inferens [adapted from (Jurenka 2004; Moto et al. 2003; Tillman et al. 1999; Vogt 2005)]

Full size image