Introduction

Olfaction plays a critical role in guiding insect behaviors, such as finding of mating partners, food plants and oviposition sites. The periphery process of insect olfaction is thought to involve several major steps. Firstly, external chemical volatiles enter into the chemosensilla and then are captured by odorant binding proteins (OBPs)1,2 or chemosensory proteins (CSPs)3,4,5. Secondly, the OBP or CSP bound chemical volatiles are transported to the olfactory receptors (ORs)6 located on dendrite membrane of sensory neurons, triggering the transduction of chemical signals to electric signals. In addition, sensory neuron membrane proteins (SNMPs) may also participate in the chemoreception. Two subtypes of SNMP proteins, SNMP1 and SNMP2, have been identified from different insects7. After the transduction, odorant molecules will be rapidly deactivated to resume the sensitivity of the sensory neurons, by odorant degrading enzymes (ODEs) such as carboxylesterases (CXEs)8,9 and aldehyde oxidases (AOXs)5. Antennae are the primary olfactory organs in insects, but other organs such as mouthpart appendages, legs, wings and female ovipositors also bear some olfactory sensilla and thus play some distinct roles in insect behaviors. In particular, ovipositor in moths has long been proposed to play important chemoreception roles in oviposition site selection. In female Monopis crocicapitella, multipotous sensilla on the ovipositor were observed and supposed to play role in perception of volatiles for the general assessment of the oviposition site10. In Pyralidae, sensilla on ovipositor were supposed to be involved in the perception of the oviposition-deterring pheromone secreted by the larvae and other volatiles in Homoeosoma nebulella11 and Ephestia kuehniella20. Furthermore, expression of pheromone binding protein 2 (PBP2) (a sex pheromone specific OBP) and the sex pheromone specific OR were detected in the female ovipositor of a noctuid Heliothis virescens12, suggesting a possible role of the ovipositor in feedback regulation of the female sex pheromone biosynthesis and emission from the sex pheromone gland. However, the chemoreception related proteins expressed in moth ovipositors and their functions are mostly unknown.

In most female moths, the ovipositor is anatomically in close connection onto the sex pheromone gland (PG), the site of sex pheromone biosynthesis and emission13. The ovipositor-sex pheromone glands (OV-PG) together provide an important role for the reproductive behavior of moths. Sex pheromone components in moth species are primarily C10-C18 long straight-chain, unsaturated derivatives of fatty acids, with the carbonyl carbon modified to form an oxygen-containing functional group such as alcohol, aldehyde, or acetate ester14. Their biosynthesis starts by acetyl-CoA carboxylases (ACCs) and fatty acid synthetases (FASs) catalyzing the saturated fatty acid precursor malonyl-CoA from acetyl-CoA in the first committed biosynthesis step15,16. Similar fatty-acid metabolism enzymes function in desaturation, chain-shortening by β-oxidation, functional-group modifications by reduction and finally producing the pheromone components by acetylation or oxidation17. Different combinations of these enzymes can produce unique species-specific pheromone blends in different species. So far, two classes of essential enzymes involved in moth pheromone synthesis have been functionally identified. The first is the desaturases (Dess), functioning to introduce double bonds into pheromone precursors. Dess are the most intensively studied class of enzymes involved in sex pheromone biosynthesis18. Based on the position that double bond is introduced, 5 types of desaturases (Δ9, Δ10, Δ11, Δ12 and Δ14) have been identified19,20. The second is the fatty acid reductases (FARs), responsible for reducing fatty acids to alcohols and have been functionally identified in a couple of moth species, such as pgFAR-Z/E in Ostrinia nubilalis21 and pgFAR in Bombyx mori22. Besides, some other important genes are postulated to be involved in moth pheromone production and remain to be characterized. Acyl-CoA oxidases (ACOs), are responsible for lipid metabolism by catalyzing the conversion of acyl-CoA into trans-2-enoyl-CoA during fatty acid β-oxidation; fatty acid transport proteins (FATPs) are integral membrane-bound proteins found in both the plasma membrane and endoplasmic reticulum, several of which facilitate the uptake and activation of exogenous long chain fatty acids23; acyl-CoA binding proteins (ACBPs) bind acyl-CoA esters with high specificity and affinity and thus are thought to act as intracellular transporters of acyl-CoA esters between different enzymatic systems24.

The rice stem borer, Chilo suppressalis (Walker) (Lepidoptera: Pyralidae), is a notorious rice pest in East Asian countries, causing great economic losses to rice crops25. It also oviposites on and damages variety of other crops, such as corn and cane26, suggesting an important role of chemoreception in oviposition and food plant selection. The sex pheromones of female C. suppressalis were initially identified as a binary mixture of (Z)-11-hexadecenal (Z11-16:Ald) and (Z)-13-octadecenal (Z13-18:Ald)27,28. Later on, the third component (Z)-9-hexadecenal (Z9-16:Ald) was discovered29 and the tertiary blend of Z11-16:Ald, Z13-18:Ald and Z9-16:Ald at the ratio of 48:6:5 had the maximum attraction29. However, genes involved in the pheromone biosynthesis of C. Suppressalis have not been explored. Our present study reported a genetic database of the genes expressed in the female OV-PG of C. suppressalis by using the Illumina HiSeq(TM) 2000 sequencing platform. Totally, 63 putative genes related with chemoreception and pheromone biosynthesis were identified. The tissue expression profile investigation showed that some of those genes might play important roles in oviposition behavior and sex pheromone biosynthesis of C. suppressalis.

Results

Transcriptome sequencing and sequence assembly

Sequencing of a cDNA library prepared from mRNA of the OV-PG of C. suppressalis (Fig. 1) provided about 65 million raw reads (≫4 Gb). After trimming adaptor sequences and removing low quality sequences, about 54 million clean reads remained and were assembled into 66,971 contigs (≥75 bp) with a mean length of 359 bp and the N50 length of 579 bp. After clustering and redundancy filtering, we finally acquired 37,619 longer sequences (≥150 bp), including 5,942 Clusters (15.8%) and 31,677 Singletons (84.2%) with a mean length of 628 bp and the N50 length of 884 bp. We defined these 37,619 sequences as unigenes according to some recently published papers30, although each of them may not necessarily represents a unique gene. Of the 37,619 unigenes, those ≥1000 bp accounted for 16.26% of the transcriptome assembly (Fig. 2).

Figure 1
figure 1

Dissection of C. suppressalis ovipositor-pheromone gland for RNA extraction.

(A). Ovipositor-pheromone gland was forced out by squeezing the abdomen. (B). Ovipositor-pheromone gland was cut from the 8th abdominal segment.

Full size image
Figure 2
figure 2

Distribution of Unigenes size in the C. suppressalis transcriptome assembly.

Full size image

Homology analysis and Gene Ontology (GO) annotation

Of the 37,619 unigenes, 18,037 were matched by the BLASTX homology search to the entries in NCBI non-redundant (nr) protein database with a cut-off E-value of 10−5. The highest percentage of matched sequences is to Danaus plexippus (57.97%), followed by Bombyx mori (6.49%), Papilio xuthus (4.09%), Tribolium castaneum (2.57%), Papilio polytes (1.29%), Acyrthosiphon pisum (0.91%) and Manduca sexta (0.83%). The remaining 25.8% sequences were matched to other insects (Fig. 3).

Figure 3
figure 3

Percentage of homologous hits of the C. suppressalis transcripts to other insect species.

The C. suppressalis transcripts were searched by BLASTx against the non-redundancy protein database with a cutoff E-value 10−5. Species which have more than 0.8% matching hits to the C. suppressalis transcripts are shown.

Full size image

The Gene Ontology (GO) annotation was used to classify the 37,619 unigenes into different functional groups using BLAST2GO. Based on the sequence homology, 8,506 unigenes (22.61%) could be annotated and each unigene was classified into one or more functional groups of the three biological processes (Fig. 4). Of the 8,506 annotated unigenes, more than half could align to “cellular process” (62.53%), “metabolic process” (52.92%), “binding” (51.98%), “cell” (50.48%) and “cell part” (50.48%). In total, 31,874 annotation hits (unigene-functional group) come to the biological process, 18,952 to the cellular component and 9,639 to the molecular function.

Figure 4
figure 4

Gene ontology (GO) classification of the C. suppressalis transcripts with Blast2GO program.

One unigene could be annotated into more than one GO term.

Full size image

Identification of putative genes related to chemoreception and sex pheromone biosynthesis

Moth ovipositors bear some chemoreception sensilla10,31 and the sex pheromone gland may also express some chemoreception proteins such as OBPs and CSPs that are postulated to facilitate the transportation of the sex pheromones and their precursors. By homologous searches, a total of 31 putative chemoreception genes were identified, including 9 OBPs, 10CSPs, 1 SNMP, 2 ORs and 9 ODEs (Tab. 1). In addition, 32 putative genes related to the sex pheromone biosynthesis were also obtained, which includes 1 FAS, 6 (Dess, 10 FARs, 2 ACOs, 1 ACC, 4 FATPs, 3 ACBPs and 5 elongation of very long chain fatty acids (ELOs) (Tab. 2).

Table 1 Putative chemoreception transcripts in the female ovipositor-pheromone glands of C. suppressalis. Nucleotide sequences for the identified transcripts are given in Table S1
Full size table
Table 2 Putative transcripts related to sex pheromone biosynthesis in the ovipositor-pheromone glands ofC. suppressalis. Nucleotide sequences for the identified transcripts are given in Table S1
Full size table

Among the 63 identified genes (Tab. 1 and 2), 6 genes were the same as sequences already deposited in the GenBank: 5 CsupOBPs (GenBank accession number: AGK24577.1, ACJ07120.1, AGM38609.1, AGK24580.1, ACJ07126.1) and 1 CsupSNMP (GenBank accession number: AFS50074.1), while the other 57 transcripts found in the current study were new in C. suppressalis. The abundances of the 63 genes in the transcriptome were shown in Fig. 5. Three classes of genes (Dess, FARs and CSPs) showed higher abundances than the others, with the CSPs being the highest. To validate these sequences, RT-PCR validation experiments were conducted and the results showed that all these 63 sequences were consistent with that of the PCR products.

Figure 5
figure 5

Abundances of 63 transcripts which are putatively involved in chemoreception (A) and sex pheromone biosynthesis (B) in the transcriptome dataset of C. suppressalis ovipositor-pheromone gland.

The genes expression abundance is indicated as the Reads Per Kilobase per Million mapped reads (RPKM) values.

Full size image

Expression profile of the putative genes in chemoreception and sex pheromone biosynthesis

To provide functional clues, expression profiles of the genes were investigated by RT-PCR for all 63 genes and by Quantitative Real Time RT-PCR (qPCR) measurements for selected 3 genes (Fig. 6 and 7). The results showed that only 2 chemoreception related genes (CsupCSP2 and CsupCSP10) and 3 sex pheromone biosynthesis related genes (CsupDes4, CsupDes5 and CsupFAR2) displayed higher expression in the OV-PG complex than in other tissues of C. suppressalis.

Figure 6
figure 6

Expression of C. suppressalis chemoreception and pheromone biosynthesis related transcripts in different adult tissues, determined by semi-quantitative RT-PCR.

β-actin gene was used as the positive control and NC (no cDNA template) as the negative control. OV-PG, female ovipositor-pheromone glands; A, antennae; Ab, abdomens (mixture of equivalent male and female cDNA in panel A); Le, legs (mixture of equivalent male and female cDNA in panel A); Wi, wings (mixture of equivalent male and female cDNA in panel A); Ep, epidermis; Fb, fat body; T, thoraxes. ♀, female; ♂, male. Primers used for the RT-PCR were listed in Table S3.

Full size image
Figure 7
figure 7

Relative expression levels of 3 putative pheromone biosynthesis related transcripts in different female tissues, determined by qPCR.

OV-PG, ovipositor-pheromone glands; Ep, epidermis; Fb, fat body; T, thoraxes; Le, legs; Wi, wings; ♂Ab, male abdomens. Primers used for qPCR were listed in Table S3.

Full size image

Odorant binding protein (OBP) and chemosensory protein (CSP)

The tissue expression profiles of CsupOBPs are shown in Fig. 6A. The 9 OBPs identified in this study included 2 GOBPs (OBP4: GOBP2; OBP9: GOBP1) and 7 other OBPs. Among the 9 CsupOBPs, 3 genes (CsupOBP1, CsupOBP2 and CsupOBP6) displayed a very wide range of tissue distribution in all 6 tissues, while 4 genes (CsupOBP3, CsupOBP4, CsupOBP7 and CsupOBP9) were expressed specifically or predominately in adult antennae and legs and 1 (CsupOBP8) was specifically expressed in abdomen. In addition, 4 genes (CsupOBP1, CsupOBP2, CsupOBP3 and CsupOBP6) were detected in the OV-PG.

Compared to CsupOBPs, the abundances of CsupCSPs were much higher in OV-PG of C. suppressalis. Most CsupCSPs were expressed in similar levels between olfactory and non-olfactory tissues, while CsupCSP2 and CsupCSP10 were not expressed in olfactory tissues but in OV-PG and legs (Tab. 1 and Fig. 6A).

Odorant receptor (OR) and sensory neuron membrane protein (SNMP)

2 putative CsupORs were identified from the C. suppressalis OV-PG. CsupOR2 encoding a protein with 53% identity to putative odorant receptor OR12 of Cydia pomonella (GenBank accession number: AFC91721.1) was detected in OV-PG, adult antennae and abdomens. In contrast, CsupOR1 encoding a protein with 69% identity to putative odorant receptor OR43 of C. pomonella (GenBank accession number: AFC91751.1) was expressed in all chemosensory tissues at low levels and not expressed in OV-PG and abdomens (Tab. 1 and Fig. 6A).

1 putative CsupSNMP2 was obtained from C. suppressalis. However, it was strongly expressed in adult antennae, but not detected in OV-PG (Tab. 1 and Fig. 6A).

Aldehyde oxidase (AOX) and carboxyl esterase (CXE)

AOXs catalyze the oxidation of aldehydes to carboxylic acids and they may also be involved in the degradation of sex pheromone compounds specifically in the conversion of aldehydes to carboxylic acids14,22. Only 1 AOX homolog was obtained from C. suppressalis transcriptome, which shared 66% identity to the antennae-specific AOX2 of Amyelois transitella (Tab. 1). CsupAOX was antennae-predominantly expressed, with very weak expression in OV-PG (Fig. 6A).

Esterases are hydrolases and hydrolysis of esters occurs during plant volatile and sex pheromone degradation9. We totally obtained 8 CsupCXEs from the C. suppressalis OV-PG transcriptome. Compared to the other putative chemoreception genes, all CsupCXEs were expressed at higher levels in male antennae than in female antennae, except for CXE11 that displayed the contrary. 5 CsupCXEs (CsupCXE11, 14, 17, 18 and 20) were expressed weakly in OV-PG (Fig. 6A).

Fatty acid synthase (FAS), Acetyl-CoA carboxylase (ACC) and Acyl-CoA oxidase (ACO)

FAS has been reported to catalyze the conversion of malonyl-CoA and NADPH to saturated fatty acids15. 1 putative CsupFAS cDNA in full length was identified from theOV-PG complex of C. suppressalis (Tab. 2 and Fig. 6B). It contained an ORF of 660 bp, encoding a protein with 74% amino acid identity to the FAS of Agrotis ipsilon (Gene Bank accession number: AGR49310.1). The RT-PCR revealed that CsupFAS (CL973) was expressed in all tested tissues of C. suppressalis. Similarly, the CsupACC gene was also expressed in all tissues. The gene encoded an ACC with 77% identity to ACC1 of Danaus plexippus (GenBank accession number: EHJ72299.1). As for the 2 CsupACOs, CsupACO1 and CsupACO2 shared 77% and 83% identity to ACO1 of B. mori (GenBank accession number: XP_004932400.1), respectively. Both CsupACOs were expressed in all tested tissues except for the fat body.

Desaturase (Des)

Among 6 desaturases, CsupDes4 and CsupDes5 displayed highly biased expression in OV-PG, while CsupDes2 highly fat body biased in expression, CsupDes3 and CsupDes6 were detected weakly in epidermis, thoraxes, legs and wings. All the later 3 genes were not detected to be expressed in OV-PG by the RT-PCR.

Fatty acyl-CoA reductase (FAR)

Of the 10 newly identified CsupFARs, only CsupFAR2 was predominantly expressed in OV-PG. CsupFAR10 was only detected in the thoraxes of female C. suppressalis, while 8 genes (CsupFAR1, 3, 4, 5, 6, 7, 8 and 9) displayed a wide range of tissue distribution.

Fatty acid transport protein (FATP), Acyl-CoA binding protein (ACBP) and Elongation of very long chain fatty acids (ELO)

All 4 CsupFATPs (except for CsupFATP3) and 3 CsupACBPs were expressed at very low levels in the OV-PG and were not OV-PG biased (Fig. 6B). CsupELO1-5 encoded proteins with high identity (90%, 86%, 81%, 87% and 81%) to ELOs of B. mori (GenBank accession number: XP_004931946.1, XP_004931947.1, XP_004924776.1, XP_004931951.1 and XP_004924792.1), respectively. CsupELO1, 3, and 5 were expressed both in high levels in OV-PG as well as several other tissues, the other 2 genes were very weakly expressed in the OV-PG complex (Fig. 6B).

Phylogenetic analyses

In order to assign putative functions to CsupOBPs, CsupCSPs, CsupCXEs, CsupDess and CsupFARs, phylogenetic analyses were conducted for each group of the genes. The OBP phylogenetic tree revealed that CsupOBP4 (CsupGOBP2) and CsupOBP9 (CsupGOBP1) were clustered into PBP/GOBP clade and other CsupOBPs were distributed in five different groups (Fig. 8). In addition, the 10 putative CsupCSPs were clustered with at least one lepidopteran orthologous gene (Fig. S1). The 8 CsupCXEs were distributed into three groups, with CsupCXE14, CsupCXE17, CsupCXE18 and CsupCXE20 in group (A) (mitochondrial, cytosolic and secreted esterases), CsupCXE9 and CsupCXE11 in group (B&C) (higher and lower dipteran microsomal α-esterases) and CsupCXE5 and CsupCXE16 in group (D) (integumental/antennal esterases) (Fig. S2).

Figure 8
figure 8

Phylogenetic tree of putative OBPs from C. suppressalis and some other Lepidoptera species.

The tree was constructed with MEGA5.0, using the Neighbour-joining method. Values indicated at the nodes are bootstrap values based on 1000 replicates and the bootstrap values < 50% are not shown. Amino acid sequences of the species and their accession numbers are given in Table S2.

Full size image

A phylogenetic tree using Des protein sequences from Clubiona parallela, Lampronia capitella, O. nubilalis, Ostrinia scapulalis, Spodoptera littoralis and some other lepidopteran species (Fig. 9) showed that CsupDes6 was clearly clustered into the clade of Δ14 desaturase, while CsupDes1, CsupDes2 and CsupDes4 were clustered in the clad of Δ9 (16C>18C), Δ9 (14C-26C) and Δ11 desaturase, respectively. Interestingly, the CsupDes3 and CsupDes5 were clustered in two different subclasses from other identified Dess (Fig. 9). In the FAR phylogenetic tree, CsupFAR2 was clustered with other lepidopteran pgFAR sequences21,22,32,33,34,35. Most CsupFAR sequences were clustered with at least one lepidopteran orthologous gene (Fig. S3), with only one gene (CsupFAR9) having no counterpart.

Figure 9
figure 9

Phylogenetic tree of putative Dess from C. suppressalis and some other Lepidoptera species.

The tree was constructed with MEGA5.0, using the Neighbour-joining method. Values indicated at the nodes are bootstrap values based on 1000 replicates and the bootstrap values < 50% are not shown. Amino acid sequences of the species and their accession numbers are given in Table S2.

Full size image

Discussion

It has long been reported that some olfactory sensilla are distributed on the ovipositor10,31 and thus ovipositor may play olfactory roles in detection of plant odors, ovipositor-deterring pheromones (ODEs) and sex pheromones10. To better understand the olfactory role of the moth ovipositor, we conducted investigations of chemoreception genes by analyzing the transcriptomic data of C. suppressalis OV-PG and of expression profiles of these genes by semi-quantitative RT-PCR. Our results provide direct molecular evidences for the olfactory role of moth ovipositors and also the important basis for further elucidation of the molecular mechanisms of the olfaction, as well as the sex pheromone biosynthesis in the OV-PG of C. suppressalis.

OBPs and CSPs are thought to be responsible for the binding and transport of the hydrophobic molecules including pheromones and plant volatiles1,3,4. Among the 9 CsupOBPs identified in the present study, none was OV-PG specific or biased. In contrast, CsupOBP3, 4 (GOBP1), 7 and 9 (GOBP2) were strongly antennae-biased, indicating more important olfactory roles they play in antennae. Similar as OBPs, most CsupCSPs were expressed in multiple tissues, but CsupCSP10 was weakly but more specifically detected in OV-PG (Fig. 6A) and deserves further studies with its function. Compared with 9 OBPs and 10 CSPs, only 2 CsupORs were identified in the OV-PG. It is noted that the ORco, acting as a chaperone and dimerization partner for other ORs5,36, was not found in the OV-PG, possibly due to the lower expression levels.

PBP is a subclass of OBP and is commonly thought to be involved in the reception of the sex pheromones37. Very interestingly, a study with H. virescens12 reported that the PBP2 and pheromone receptor (PR) specific to Z11-16:Ald (the major sex pheromone component) were expressed in the sensilla on the ovipositors. Considering the close vicinity between the ovipositor and the sex pheromone gland, the authors proposed a negative feedback regulation of pheromone biosynthesis, in which sensilla on the ovipositor mediate the biosynthesis and emission of the sex pheromones. However, we did not detect any CsupPBP expression in OV-PG by transcriptomic analysis. Therefore, in C. suppressalis, antennae and other tissues that express PBPs38 may play the role in the feedback regulation. In addition, the role of ovipositor in the regulation could not be excluded. The PBP expression in the ovipositor might be too low to be detected by the approach used in the present study in C. suppressalis. Alternatively, other OBPs especially the GOBP2, instead of PBPs, may function to transport the sex pheromone components in the ovipositor. GOBP2 has been suggested playing roles in sex pheromone detection in some moth species, by ligand binding assays and in site hybridization experiments39,40.

In our current study, 9 ODE candidate genes (8 CXEs and 1 AOX) were found to be expressed in the OV-PG, although in very lower level in relation to other tissues, especially the antennae. This may reflect that ovipositor is a less important olfactory organ than the antennae. In the CXEs phylogenetic tree, CsupCXEs were clustered into three insect CXE clades with other lepidopteran CXEs (Fig. S2) and at least one CXE in each of the three clades had been confirmed to act in the deactivation of ester plant volatiles and/or pheromone components8,41,42, suggesting their importance in the degradation of ester odorants.

As for the 32 pheromone biosynthesis related genes expressed in the OV-PG, 2 CsupDess (CsupDes4, CsupDes5) and 1 CsupFAR (CsupFAR2) were highly expressed in OV-PG by transcriptomic analysis (Fig. 5B) and tissue expression investigation (Fig. 6B), suggesting that these 3 genes play important roles in C. suppressalis sex pheromone biosynthesis. In particular, CsupDes4 shared an amino acid identity of 71% to a gene encoding a Δ11 desaturase in Helicoverpa zea (Genebank: AAF81787.1) (Tab. 2) and was clearly assigned to the Δ11 desaturase group in the phylogenetic tree (Fig. 9). It has been reported that the synthesis of Δ11-containing sex pheromones involves a step of Δ11-desaturation catalyzed by a Δ11 desaturase43,44. Considering that the Δ11-containing Z11-16:Ald is the major sex pheromone component in C. suppressalis, CsupDes4 is very likely involved in the desaturation step from saturated acids (16C) to unsaturated acids, with a double bond at 11th position of the carbon chain. In addition, CsupDes4 may also play a role in the biosynthesis of C. suppressalis minor pheromone component Z13-18:Ald, if this component is formed from Z11-16:Ald by carbon chain elongation. The third pheromone component in C. suppressalis is Z9-16:Ald, suggesting a existence of a Δ9 desaturase in the sex pheromone glands. However, CsupDes5, with high and biased expression in the OV-PG, was not assigned to the Δ9 desaturase group, but a group without other members. On the other hand, CsupDes1 and 2 were clustered into Δ9 desaturase group, but their expressions were not pheromone gland biased and very low in the OV-PG. Whether these 3 genes function to introduce the Δ9 double bond in Z9-16:Ald needs further investigation.

In the biosynthesis process of moth sex pheromones, once the specific unsaturated fatty acid precursors are produced, they will be converted into corresponding alcohols by FAR21,22,45. In the current study, CsupFAR2 was highly and more specifically expressed in OV-PG than other CsupFARs (Fig. 5B, 6B and 7) and the phylogenetic analysis indeed classified CsupFAR2 into the pheromone gland FAR (pgFAR) group (Fig. S3), suggesting that CsupFAR2 is responsible for the conversion of the unsaturated fatty acid precursors to corresponding alcohols in C. suppressalis.

In sex pheromone glands, aldehyde components are confirmed to be produced from alcohol oxidations by alcohol oxidase (AO)46. Unfortunately, we did not find an AO from the transcriptomic database, similar as studies with sex pheromone gland transcriptome analysis in other species using aldehyde(s) as sex pheromone component(s), such as H. virescens47, Agrotis segetum17 and A. ipsilon48. In Sesamia inferens, 1 AO was found with very low expression level in the sex pheromone gland transcriptome49. Therefore, more sensitive approaches are needed to identify the possible AO genes in moths. In addition to Des and FAR genes, we also found some genes of other classes that are thought to participate in the sex pheromone biosynthesis, such as FATP, ELO and ACBP. However, none of them were OV-PG-predominant in expression, suggesting their multiple functions in the insect physiology.

Methods

Insects rearing and collection

The rice stem borer C. suppressalis was originally collected from a rice field in the Jiangsu Provincial Academy of Agricultural Sciences, Nanjing, China (118.9°E, 32.0°N). The collected insects were reared in laboratory for several generations with rice seedlings according to the method reported by Shang et al.50 until pupation and sexed as pupae. The rearing conditions were 28°C ± 1°C, 60%–70% relative humidity and a 16 h : 8 h light dark photoperiod. Adults were provided with a cotton swab dipped in 10% honey solution and renewed daily. For transcriptomic sequencing, OV-PG of 2-day-old female moths were collected, as 2- and 3-day-old moths showed highest mating activity51, For tissue expression study, tissues including male and female abdomens (for female abdomen, OV-PG was removed), legs and wings and female OV-PG, epidermis, fat body and thoraxes were collected from 2-day-old virginal moths; while male and female antennae were collected from 2- and 3-day-old virginal moths, as more moths were needed for antennae collection. Two and three replications of tissue samples were collected for RT-PCR and qPCR measurements, respectively. All samples were collected during the 4–5 hour of the dark period and stored at −70°C until use.

cDNA library construction and Illumina sequencing

Total RNA was extracted using TRIzol regent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer's protocol. The quantity of RNA was determined using a Nanodrop spectrophotometer (Thermo Scientific, Wilmington, DE, USA) and 1.1% agarose gel electrophoresis. cDNA library construction and Illumina sequencing of the sample were performed at Beijing Genomics Institute (BGI)-Shenzhen, Shenzhen, China52. After the total RNA extraction and DNase I treatment, magnetic beads with Oligo (dT) are used to isolate mRNA (for eukaryotes) or by removing rRNAs from the total RNA (for prokaryotes). Mixed with the fragmentation buffer, the mRNA is fragmented into short fragments. Then cDNA is synthesized using the mRNA fragments as templates. Short fragments are purified and resolved with EB buffer for end reparation and single nucleotide A (adenine) addition. After that, the short fragments are connected with adapters. The suitable fragments are selected for the PCR amplification as templates. During the QC steps, Agilent 2100 Bioanaylzer and ABI StepOnePlus Real-Time PCR System are used in quantification and qualification of the sample library. At last, the library could be sequenced using Illumina HisSeq™ 2000 platform.

De novo Assembly of Short Reads and Gene Annotation

Transcriptome de novo assembly was carried out with short reads assembling program Trinity53. Trinity partitioned the sequence data into many individual de Bruijn graphs, each representing the transcriptional complexity at a given gene or locus and then processed each graph independently to extract full-length splicing isoforms and to tease apart transcripts derived from paralogous genes. The result sequences of trinity were called Unigenes. The Unigenes larger than 150 bp were first aligned by BlASTX to protein databases, including Nr, Swiss-Prot, KEGG and COG (e-value<10−5), retrieving proteins with the highest sequence similarity with the given unigenes along with their protein functional annotations. Then, we used Blast2GO program54 to get GO annotation of the unigenes and got GO functional classification by using WEGO software55.

Expression Abundance Analysis of the Unigenes

The expression abundance of these unigenes were calculated by the RPKM (Reads Per Kilobase per Million mapped reads) method56, using the formula: RPKM (A) = (1,000,000 × C × 1,000)/(N × L). In the formula, RPKM (A) is the expression abundance of gene A; C is the number of reads that uniquely aligned to gene A; N is total number of reads that uniquely aligned to all genes; and L is the number of bases on gene A. The RPKM method is able to eliminate the influence of different gene lengths and sequencing discrepancy on the calculation of expression abundance.

RNA isolation and cDNA synthesis for Reverse Transcription-PCR

Total RNA was extracted by 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, USA). The cDNAs from female OV-PG and other body parts (fat body, epidermis, thoraxes, antennae, legs, wings and abdomens (for female without the OV-PG)) were synthesized using the PrimeScriptTM RT Master Mix (TaKaRa, Dalian, China).

Reverse Transcription-PCR Analysis

Gene specific primers across ORF of predicted chemosensory genes were designed using Beacon Designer 7.6 and Primer Premier 5.0 (PREMIER Biosoft International, CA, USA). The sequences of these primers were listed in Tab. S1. PCR experiments including negative controls (no cDNA template) were carried out in a MyCyclerTM (Bio-Rad, USA) under the following conditions: 94°C for 3 min; 32 (35 for CsupACO2, CsupACBP3, CsupFAR7, CsupELO2 and CsupFATP4) cycles at 94°C for 30 sec, 60°C for 40 sec and 72°C for 50 sec and final incubation for 10 min at 72°C. The reactions were performed in 25 ul with 1 ul of single-stranded cDNA, 4.0 mM MgCl2, 0.4 mM dNTP, 0.8 mM for each primer and 2.5 U rTaq DNA polymerase (TaKaRa, Dalian, Liaoning, China). PCR products were analyzed by electrophoresis on 2.0% w/v agrose gel in TAE buffer (40 mmol/L Tris-acetate, 2 mmol/L Na2EDTA·H2O) and the resulting bands were visualized with ethidium bromide and digitized using Gel Capture (China).

In addition, all transcripts were chosen to perform a second biological replication in order to check the repeatability of the tissue expression. To validate the predicted sequences of chemoreception genes, the PCR products were purified by using the AxyPrep™ PCR Cleanup Kit (Axygen) and then were sub-cloned into a T/A plasmid using the pEASY-T3 cloning vector system (TransGene, China) following manufacturer's instructions. The plasmid DNA was used to transform into Trans1-T1 competent cells. Positive clones were checked by PCR and were sequenced by GenScript (Nanjing, China).

Quantitative Real Time-PCR Validation

The expression profiling of a total of 3 putative sex pheromone synthesis genes was carried out to validate the accuracy of the RT-PCR results using quantitative real time-PCR (qPCR) experiments. The 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, 0.4 μl each primer (10 μM), 2.5 ng of sample cDNA and 6.8 μl sterilized ultrapure H2O. The reaction programs were 30s at 95°C, 40 cycles of 95°C for 5s and 60°C for 34s. The results were analyzed using the ABI 7500 analysis software SDS 1.4. The qPCR primers (Tab. S3) were designed using Beacon Designer 7.7 (PREMIER Biosoft International, CA, USA). The mRNA levels were measured by qPCR using the SYBR Premix ExTaqTM (TaKaRa, Dalian, Liaoning, China). This was followed by the measurement of fluorescence during a 55 to 95°C melting curve in order to detect a single gene-specific peak and to check the absence of primer dimmer peaks and a single and discrete peak was detected for all primers tested. Negative controls were non-template reactions (replacing cDNA with H2O).

Expression levels of 3 genes were calculated relative to the reference gene Csupβ-action and CsupGAPDH using the Q-Gene method in Microsoft Excel-based software of Visual Basic57,58. For each sample, three biological replications were performed with each biological replication measured in three technique replications.

Phylogenetic Analyses

The phylogenetic trees were reconstructed for phylogenetic analyses of CsupOBPs, CsupCSPs, CsupCXEs, CsupDESs and CsupFARs were based on the amino sequences (the signal peptides of sequences had been removed) of the putative genes and the sequences of other Lepidoptera insects. 8 OBP sequences from C. suppressalis (amino acids>138 aa), 10 CSP sequences from C. suupressalis (amino acids>65 aa), 8 CXE sequences from C. suppressalis (amino acids>79 aa), 6 Des sequences from C. suppressalis (amino acids>90 aa) and 10 FAR sequences from C. suppressalis (amino acids>79 aa) were included in the phylogenetic tree data. The protein name and accession number of the genes used for phylogenetic tree building are listed in Tab. S2. Amino acid sequences were aligned with ClustalX 2.059 and unrooted trees were constructed by MEGA5.060 using the Neighbor-joining method, with Poisson correction of distances. Node support was assessed using a bootstrap procedure base on 1000 replicates.