Background

Insects use their olfactory systems to recognize and discriminate chemical cues in the external environment to adjust behaviors such as oviposition, host location, and predator avoidance [1,2,3]. This task is performed by two major families of small soluble proteins, odorant-binding proteins (OBPs) and chemosensory proteins (CSPs) [4,5,6,7,8]. OBPs are abundant water-soluble acidic proteins with a pattern of six conserved cysteine residues, which are paired into three interlocked disulfide bridges that bind and protect small hydrophobic ligands [9,10,11]. CSPs have four cysteines that form two disulfide bridges. CSPs are present at high concentrations in chemosensory sensilla lymphs and are broadly expressed in non-sensory tissues [5, 6, 12, 13]. In addition, some membrane bound chemosensory receptors located on the chemical sensory dendrites of olfactory neurons also participate in olfactory signal transduction. These membrane proteins, including odorant receptors (ORs), ionotropic receptors (IRs), gustatory receptors (GRs) and sensory neurons membrane proteins (SNMPs), act as the bridge between extracellular odorant signals and intracellular nerve reactions [3, 14, 15]. ORs are G-protein coupled receptors (GPCRs) with seven transmembrane domains, including a variable odor-specific protein (ORx), and a highly conserved co-receptor protein (ORco) [16, 17]. IRs are related to ionotropic glutamate receptors (iGluRs) and are essential for odor-evoked neuronal responses and for detecting environmental volatile chemicals and tastes [18,19,20,21]. Insect GRs were first identified in Drosophila melanogaster, are mainly expressed in taste organs, and are associated with contact chemoreception [22, 23]. The GR family includes many related members, and the sequence and abundance of GR family members varies greatly among different species, with the exception of carbon dioxide receptors [24]. SNMPs are membrane-bound CD36 family members located in insect olfactory neurons, which are crucial for pheromone detection [25].

Compared with groups such as Lepidoptera and Hymenoptera, the study of olfactory-related proteins in Coleoptera began relatively late. However, the olfactory genes of more than 20 species of Coleoptera, including Tribolium castaneum [26], Colaphellus bowringi [27], and Tenebrio molitor [28], have been discovered and verified. In Cerambycidae, the olfactory genes of 9 species have been identified, including Anoplophora chinensis [29, 30], Anoplophora glabripennis [31, 32], Anoplophora nobilis [33], Monochamus alternatus [34], Batocera horsfieldi [35], Xylotrechus quadripes [36], Saperda populnea [45]. In recent studies, a slow-release attractant for S. bifasciatus, mainly composed of 31 volatile substances, including 3-carene, cedrene, cedr-8(15)-ene and α-longipinene, had a good trap** effect [46]. With regard to pheromones, the antennae of S. bifasciatus were more sensitive to bark beetle pheromones in comparison with some monoterpenes and alcohols, especially the compounds 3-methy-l2-cyclohexen-1-ol and 3-methy-l2-cyclohexen-1-one [47]. In a GC–MS study of the cuticular hydrocarbons of S. bifasciatus across different 3 developmental stages, the content of methyl branched alkanes in adult cuticular hydro-carbons increased significantly compared with that of larvae, while the proportion of n-alkanes in cuticular hydrocarbons decreased, and it was speculated that 11Me-C26, 11Me-C27, and 3Me-C27 played important roles in the reproductive behaviors of adults [48].

In the process of prevention and treatment of S. bifasciatus, the identification of olfactory related genes and the study of olfactory mechanism are particularly important. An earlier study of olfactory-related genes based on antennal transcriptome analysis of S. bifasciatus in China sought to explain the reasons underlying differences in the trap** effects of a particular attractant at two locations: Bei**g and Shandong province. Olfactory-related genes were identified, and differences in the expression levels and single nucleotide polymorphisms (SNPs) of these genes between the Bei**g and Shandong populations were compared to illustrate their diversity [39]. The study demonstrated that genetic polymorphisms in olfactory-related genes and the variation in gene expression levels can explain differences in the attractiveness of traps in different locations for many insect species. However, in S. bifasciatus, some other important aspects, including genetic relationships of all chemosensory genes with other species and tissues or gender biased expression levels of some crucial genes, have not been reported. Such works can provide useful data to researcheres to comprehensively analyze phylogenetic relationship of these chemosensory genes and tentatively predict their function in chemcial cummunication in S. bifasciatus and the related species.

In this study, we identified the chemosensory genes of S. bifasciatus, after which we conducted a phylogenetic analysis of the olfactory genes of S. bifasciatus and other species. In addition, the expression profiles of OBP genes, an initial chemosensory protein for insects to specifically recognize external odorants, in different tissues and different genders of S. bifasciatus were determined by quantitative real-time PCR for the first time. Our findings may provide a foundation for future functional characterization of the chemosensory genes of S. bifasciatus and provide new information about the evolution of olfactory proteins in Cerambycidae.

Results

Transcriptome sequencing and Unigene assembly

The antennal cDNA libraries of male and female S. bifasciatus were sequenced using the Illumina HiSeq platform. A total of a total of 187,712,876 and 134,022,896 raw reads were obtained from male and female antennae. By removing low quality reads of length less than 20 nt, 185,896,722 and 132,600,212 clean reads were obtained for male and female, and the percentage of Q30 bases was more than 94.07%. Trinity was used to assemble the clean data of all samples from scratch, and the assembly results were optimized and evaluated. The results showed that the number of Unigenes assembled was 45,110, the transcript number was 75,880, and the average length of N50 was 1484 bp. The clean reads of each sample were compared with the reference sequence obtained by Trinity assembly to obtain the map** results of each sample. The comparison rate in this analysis ranged from 78.93% to 86.20%. The total numbers of Unigenes and transcripts were 43,322 and 72,790, respectively. The longest transcript of each gene was selected as the Unigene.

Homology analysis and Gene Ontology (GO) annotation

In total, 19,342 Unigenes (42.88% of all 45,110 Unigenes) were annotated to at least one of the databases using the BLASTx and BLASTn programs with an E-value cut-off of 10 e-5. A total of 18,345 (40.67%), 13,704 (30.38%), 3365 (7.46%), 13,137 (29.12%), 9367 (20.76%) and 10,424 (23.11%) Unigenes from S. bifasciatus were annotated using the Nr, Pfam, COG, Swiss-Prot, KEGG and GO databases, respectively.

Homology searches against the Nr database showed that S. bifasciatus antennal transcriptomes shared the highest homology (59.10%) with sequences from A. glabripennis, followed by Leptinotarsa decemlineata (8.73%) and T. castaneum (5.08%). The GO annotations showed that the most highly annotated levels of S. bifasciatus in the biological process branch were cellular process, single organism process, and metabolic process; among the cell component branches, cell and cell part levels were the most representative. In the molecular function branch, hierarchy of binding and catalytic activity were the most enriched terms (Fig. 1).

Fig. 1
figure 1

Functional annotation of all Unigenes based on gene ontology (GO) categorization. GO analysis was performed for cellular components, molecular functions, and biological processes

Full size image

Identification and phylogenetic analysis of candidate OBPs

A total of 32 OBPs were identified in the transcriptome of S. bifasciatus. All OBPs, with the exceptions of SbifOBP8, SbifOBP13, SbifOBP23, SbifOBP31 and SbifOBP32, had a complete ORF of at least 300 bp and were considered to be full-length OBPs (Additional file 1: Table S1). According to the FPKM value, the expression levels of male and female SbifOBPs in antennal transcriptome were divided into five groups for comparison with the results of fluorescent quantitative real-time PCR (Fig. 2). According to the multiple sequence alignment of ClustalW, analyzing the number of conserved cysteine sites contained in OBPs will contribute to its 3D structure modelling. SbifOBP9, SbifOBP10, SbifOBP12, SbifOBP13, SbifOBP14, SbifOBP15, SbifOBP17, SbifOBP18, SbifOBP20, SbifOBP21, SbifOBP22, SbifOBP24, SbifOBP27, SbifOBP28 and SbifOBP29 lacked two cysteine residues (C2 and C5) and was determined to belong to the Minus-C OBP subfamily, whereas SbifOBP19 is a member of the Plus-C OBP subfamily. Each of the remaing OBPs contains 6 conserved cysteine sites and belong to the Classic OBP subfamily.

Fig. 2
figure 2

Heat-plot of FPKM values for SbifOBPs in female antennae (FAn) and male antennae (MAn). Relative expression levels are indicated by a 8-grade color scale. The genes were divided into group1-5 according to the expression level

Full size image

A total of 233 amino acid sequences of the OBPs of S. bifasciatus, D. melanogastrodia (Diptera), Bombyx mori (Lepidoptera), Apis mellifera (Hymenoptera) and other insect species (Coleoptera) were selected to construct the phylogenetic tree. According to multiple sequence alignment and phylogenetic tree analysis, 15 OBPs belonged to the Minus-C OBP subfamily, 16 belonged to the Classic OBP subfamily and 1 belonged to the Plus-C OBP subfamily. Interestingly, the number of Minus-C OBPs was slightly less than that of the Classic OBPs in S. bifasciatus, which is unique among studied Cerambycidae species. Most of the SbifOBP clades were clustered with X. quadripes, A. chinensis, and A. glabripennis, with a bootstrap support value of more than 80. SbifOBP19 has additional cysteines and is clustered together with Plus-C OBPs from other Coleopteran insects (Fig. 3).

Fig. 3
figure 3

A maximum likelihood phylogenetic tree including candidate odor-binding proteins (OBPs) from Hymenoptera (blue), Diptera (green), Lepidoptera (purple), and Coleoptera (black). Target SbifOBPs are shown in red. Minus-C OBP, GOBP/PBP and plus-C OBP lineages are found in the beige, gray and orange regions, respectively

Full size image

Identification and phylogenetic analysis of candidate CSPs

A total of 8 chemosensory proteins were identified in the antennal transcriptome of S. bifasciatus. The FPKM values of SbifCSP1, SbifCSP3, SbifCSP4, SbifCSP6 and SbifCSP7 were much higher than 100, indicating that they were highly expressed in the antenna of S. bifasciatus. In contrast to odorant binding proteins, chemosensory proteins are widely distributed in both olfactory and non-olfactory organs, and may play a role in the growth and development of insects in addition to their roles in the olfactory system (Additional file 1: Table S2).

A phylogenetic tree was constructed to show the evolutionary relationships between the chemosensory proteins of various insects, including D. melanoglycera (Diptera), B. mori (Lepidoptera), A. mellifera (Hymenoptera) and Coleoptera species (Fig. 4). In the phylogenetic tree, most SbifCSPs were clustered with A. chinensis, B. horsfieldi and C. bowringi and the bootstrap support degree was greater than 85.

Fig. 4
figure 4

A maximum likelihood phylogenetic tree including CSP candidates for Hymenoptera (blue), Diptera (green), Lepidoptera (purple), and Coleoptera (black). Target SbifCSPs are shown in red

Full size image

Identification and phylogenetic analysis of candidate ORs

We identified 71 odorant receptors in the antennal transcriptome of S. bifasciatus, 51 of which had complete ORFs. Odorant receptor proteins are located on the neuronal membrane and are relatively low in abundance, so their overall expression levels and FPKM values are low. Of the 71 odorant receptors expressed by S. bifasciatus, only 15 had FPKM values more than 10. The FPKM values of 12 odorant receptors in males is more than that of females by two-fold, and those of SbifOR59 and SbifOR33 showed a nearly tenfold difference (Additional file 1: Table S3). The differences in the expression levels of odorant receptor proteins between males and females may indicate that they play a role in the recognition of gender-related odors.

The amino acid sequences of the odorant receptors of S. bifasciatus and other species (a total of 335 species of Hymenoptera, Lepidoptera, Diptera, and Coleoptera) were used to construct a phylogenetic tree displaying the evolutionary relationships of the odorant receptors of S. bifasciatus and hymenoptera, lepidoptera, diptera, and coleoptera insects. The evolutionary ORco lineage contains SbifORco and the ORco of other species, with a branch node that has 100 bootstrap support, confirming the identity of the odorant receptor ORco. Multiple Coleoptera odorant receptors were found to cluster into a single branch with bootstrap support greater than 70, suggesting that these receptors form a group of ORs with similar functions (Fig. 5).

Fig. 5
figure 5

A maximum likelihood phylogenetic tree including candidate odorant receptors (OR) from Hymenoptera (blue), Diptera (green), Lepidoptera (purple), and Coleoptera (black). Target SbifORs are shown in red. Specific OR and ORco lineages are found in the beige and gray regions, respectively

Full size image

Identification and phylogenetic analysis of candidate GRs

A total of 34 gustatory receptors were identified in the antennal transcriptome of S. bifasciatus. Blastx sequence alignment showed that most of the taste receptors of S. bifasciatus had high homology with those of A. chinensis. FPKM values showed that the expression levels of gustatory receptors in the transcriptome of S. bifasciatus were relatively low, and only the FPKM value of SbifGR16 exceeded 10 (Additional file 1: Table S4).

A phylogenetic tree was constructed using the gustatory receptors of 194 species, including S. bifasciatus, other Coleoptera species, and Diptera, Lepidoptera, and Hymenoptera. The phylogenetic tree (Fig. 6) has two primary branches for sugar taste receptors and bitter taste receptors. 11 gustatory receptors (SbifGR1, SbifGR7, SbifGR8, SbifGR9, SbifGR11, SbifGR14, SbifGR20, SbifGR24, SbifGR26, SbifGR29 and SbifGR33) were classified into the sweet taste receptor clade, while three gustatory receptors (SbifGR3, SbifGR23 and SbifGR30) were included in the branch of the bitter taste receptors. Most of the indentified gustatory receptors in S. bifasciatus are sweet taste receptors, which indicate that the primary function of most taste receptors in its olfactory system is carbohydrate detection.

Fig. 6
figure 6

A maximum likelihood phylogenetic tree including candidate gustatory receptors (GRs) for Hymenoptera (blue), Diptera (green), Lepidoptera (purple), and Coleoptera (black). Target SbifGRs are shown in red. Sweet and bitter receptor lineages are found in the gray and beige regions, respectively

Full size image

Identification and phylogenetic analysis of candidate IRs

A total of 18 ionotropic receptors were identified in the transcriptome of S. bifasciatus, and 4 of them (SbifIR8, SbifIR12, SbifIR15 and SbifIR17) had complete ORFs encoding ≥ 600 amino acids. The overall expression levels of ionotropic receptors were relatively low, with only SbifIR3, SbifIR12 and SbifIR18 having FPKM values greater than 10 (Additional file 1: Table S5).

Phylogenetic tree was constructed by using 107 amino acid sequences of S. bifasciatus and four model insects (D. melanogastrodia, B. mori, A. mellifera, T. castaneum) and other Coleoptera insects. It was found that SbifIR3 and SbifIR12 were in the IR8A/25A lineage and had a support degree of more than 90. However, SbifIR7 and the N-methyl-d-aspartate NMDA receptors of L. decemlineata and A. mellifera were clustered in one branch (Fig. 7).

Fig. 7
figure 7

A maximum likelihood phylogenetic tree including candidate ionotropic receptor (IRs) for Hymenoptera (blue), Diptera (green), Lepidoptera (purple), and Coleoptera (black). Target SbifIRs are shown in red. The IR8A/25A, IR NMDA, and IR21A lineages are found in the gray, beige, and orange regions, respectively

Full size image

Identification and phylogenetic analysis of candidate SNMPs

Three transcripts encoding SNMPs were identified and were used to constructa phylogenetic tree with 30 sequences from D. melanogaster (Diptera), A. mellifera (Hymenoptera) and 9 Coleoptera species, including A. glabripennis and A. chinensis. Two genes belonging to the SNMP2a lineage were of full-length genes (more than 1000 bp in length) and were clustered into one branch together with the SNMP2a protein of A. glabripennis and A. chinensis with bootstrap support of more than 85 (Fig. 8). (Additional file 1: Table S6).

Fig. 8
figure 8

A maximum likelihood phylogenetic tree including candidate sensory neuron membrane proteins (SNMPs) of the orders Hymenoptera (blue), Diptera (green), Lepidoptera (purple), and Coleoptera (black). Target SbifSNMPs are shown in red. The SNMP1 and SNMP2 lineages are found in the gray and beige regions, respectively

Full size image

Tissue expression analysis of OBPs

To verify OBP expression in the antennae and characterize the expression profiles of OBPs in 4 chemosensory tissues (antennae, legs, head and tail of the abdomen), all 32 OBPs of S. bifasciatus were selected for fluorescent quantitative real-time PCR. In the selection of internal reference genes, geNorm sorted the expression stability of seven candidate internal reference genes in different tissues of female and male S. bifasciatus by calculating the gene expression stability value M, with a smaller M value indicating a more stable internal reference gene. The optimal number of internal reference genes was determined by analyzing the variation of Vn/n + 1. In this study, the V2/3 and V3/4 values were both around 0.1 (less than the recommended value of 0.15 in the program [49, 50]). NormFinder selects the best reference genes by comprehensively calculating and comparing the expression stability values of the reference genes in a particular group; a smaller stability value indicates a more stable reference gene [51]. Based on the results of Genorm and NormFinder, the UBC gene was identified as the optimal internal reference gene (Additional file 2: Fig. S1, Fig. S2, Fig. S3).

The qPCR results showed that SbifOBP1, SbifOBP2, SbifOBP3, SbifOBP6, SbifOBP7, SbifOBP20, and SbifOBP22 were highly expressed only in the antennae (Fig. 9). The expression levels of SbifOBP2, SbifOBP3, SbifOBP6, SbifOBP7 and SbifOBP20 in females were approximately twice or more than those in males. The expression levels of SbifOBP1 was male-bias in antennae and SbifOBP22 was no significantly different in both sexes antennae (P < 0.05).

Fig. 9
figure 9

Expression patterns of candidate OBPs in S. bifasciatus. An: antennae; H: head; L: legs; AB: tail of the abdomen. The letters over the error bars (a-e) denote a significant difference (P < 0.05), and "N/a" indicates that the corresponding expression level was below the limit of detection. The genes were divided into group1-5 according to the expression level

Full size image

In addition, the expression levels of SbifOBP4, SbifOBP12, SbifOBP15, SbifOBP27 and SbifOBP29 were very poor in the antennae. The expression level of SbifOBP4 in the head and SbifOBP29 in the head and legs were higher, and both of them were male-biased expression. SbifOBP12 showed high biased expression in the male head and female abdomen. SbifOBP15 was highly expressed only at the end of the abdomen of S. bifasciatus, where its expression level in females was three times that of males. SbifOBP27 showed high biased expression in the male leg and female abdomen.

Other OBPs were expressed not only in antennae but also in other various tissues. SbifOBP14 was higher in female abdomen. SbifOBP16 was higher in male legs. SbifOBP17 and SbifOBP30 showed high male-biased expression in the head. Among the five groups divided according to the expression level, the male and female differences and expression amount shown by qPCR were the same as the FPKM value obtained by transcriptome sequencing (Additional file 3: Fig. S1).

Discussion

The olfactory genes of more than 20 species of Coleoptera have been identified, but only those of 9 species of Cerambycidae have been reported, including A. chinensis [29, 30], A. glabripennis [31, 32], A. nobilis [33], M. alternatus [34], B. horsfieldi [35], X. quadripes [36], S. populnea [39]. A total of 18 OBPs, 21 CSPs, 66 ORs, 24 GRs, 14 IRs, and 4 SNMPs in S. bifasciatus were identified by Zhang et al. (2019). In their study, sections of wood containing overwintering S. bifasciatus insects were collected from two sites, the Lingyan Forest district of Taian, Shandong province (116°59′ E, 36°21′ N), and the Mentougou district of Bei**g (115°34′ E, 39°50′ N), in March 2018; sampling was performed after eclosion in the laboratory. By comparing the expression levels of olfactory-related genes between males and females in the two populations using FPKM values, it was found that knowledge of single nucleotide polymorphisms (SNPs) may facilitate interpretation of the diversity of chemosensory genes in S. bifasciatus. In our study, we identified 32 OBPs, 8 CSPs, 71 ORs, 3 SNMPs, 18 IRs, and 34 GRs; therefore, the numbers of genes annotated in each family differed from those reported by Zhang et al. (2019). We collected S. bifasciatus at the Qianfo Hill Forest Farm in **an, Shandong Province (117°01′ E, 36°38′ N) and at the Bei**g Botanical Garden (116°28 ′ E, 40° N) in May 2019. In our study, four of the 18 OBPs (OBP6, OBP9, OBP12 and OBP16) identified by Zhang et al. (2019) were significantly expressed in both male and female S. bifasciatus (FPKM value > 1000), and 11 of the 32 OBP genes had the same expression level in both studies.

The difference in sampling strategies may be one of the reasons for the different annotation results reported in our study and that of Zhang et al. (2019). In two previous reports on A. chinensis, A. chinensis collected from Anhui Province and Fujian Province were found to possess marked differences in olfactory-related genes [29, 30]. Although we and Zhang et al. collected S. bifasciatus in the same province, the distance between the collection sites was more than 40 km. Taking Shandong Province as an example, the main peak of Lingyan forest district is 668 m above sea level, whereas that of Qianfo hill forest farm was 285 m above sea level. Zhang et al. collected S. bifasciatus at the peak of eclosion, but we collected samples at the end of eclosion. It can be inferred that differences in distance, altitude and collection time will lead to differences in gene expression levels, even within the same province. Similar polymorphisms in olfactory recognition were found in a study of the important agricultural pest Chilo suppressalis. Sex pheromone recognition by male moths among six Chinese provinces and the expression levels of 12 genes involved in sex pheromone recognition showed significant linear correlations [81]. In addition, sequencing depth, living environment, sample status and other factors also contributed to the differences. For example, in a transcriptome analysis assessing different developmental stages of B. horsfieldi, there were large differences in the number of genes identified in larvae, pupae, females and males [82].

Our study focused on the annotation and classification of several olfactory proteins. Phylogenetic trees were used to analyze the genetic relationships of olfactory proteins, and qPCR was used to analyze the expression levels of OBP genes in different tissues of S. bifasciatus, allowing us to predict the functions of SbifOBPs specifically expressed in S. bifasciatus and thus explain the mechanisms by which S. bifasciatus search for host trees and other members of the species.

Conclusions

We identified six olfactory-related genes expressed by S. bifasciatus, and we obtained the tissue expression profiles of SbifOBPs by RT-qPCR to allow us to predict their functions, providing insight into the mechanisms by which S. bifasciatus find host trees and other members of the species. Meanwhile, we compared an article on the transcriptome of the same species published in forests with our research, and analyzed the reasons for the differences at the molecular level. Our results provide a foundation for studies aimed at clarifying the mechanism by which the olfactory system responds to specific odorant molecules via experiments assessing protein expression, fluorescence binding competition, molecular docking, and behavior. Such studies will facilitate the development of attractants that will provide methods for more effective control of S. bifasciatus. Finally, our results demonstrate an ideal model for studying the olfactory behavior of other pests, and we provide new ideas and methods for controlling and monitoring other Coleoptera pests.

Methods

Insects and tissue collections

The samples of S. bifasciatus were collected from two sites, the Qianfo Hill Forest Farm of **an, Shandong Province (117°01′ E, 36°38′ N), and Bei**g Botanical Garden (116°28 ′ E, 40° N) Bei**g, in late May 2019. Damaged cypress trees with a diameter at breast height of 5–13 cm and a tree height of 2–8 m were selected. The target trees with debilitated wood were cut into 1 m long wood sections. Nine wood sections were grouped together, tied into a cube with iron wire, and then placed on an open and well-ventilated hillside. Specimens were caught when they flew out of the wood samples, and their status, sex and other information were recorded. After mating, female and male antennae were excised and then stored at − 80 °C until RNA extraction.

Total RNA extraction, cDNA library construction, and Illumina sequencing

The antennal RNA was extracted using Trizol reagent (Invitrogen, USA) and the RNeasy plus Mini Kit (No.74134; Qiagen, Hilden, Germany) according to the manufacturer’s instructions. Adult antennae of both sexes (12 antennae from males and 12 antennae from females) were cut by ophthalmic surgical scissors to allow biological repeats and then used for transcriptome sequencing. Three biological repeats were assessed for both male and female antennae.

RNA samples were tested using a Nanodrop 8000 spectrophotometer (Thermo, Waltham, MA, USA) and a 2100 Bionalyzer RNA NanoChip (Agilent, Santa Clara, CA). The RNA samples were sequenced using the Illumina Hiseq 4000 platform (Shanghai Majorbio bio Pharm Technology Co., Ltd.)

The Illumina Hiseq platform was used for short sequence sequencing. The enriched mRNA was a complete RNA sequence with an average length of several bp, so it needed to be randomly interrupted. With the addition of fragmentation buffer, the mRNA was randomly broken into small fragments of about 300 bp. Under the action of reverse transcriptase, random primers were added to perform reverse synthesis of one strand of cDNA with mRNA as the template, after which two-chain synthesis was carried out to form a stable double chain structure.

The structure of the double-stranded cDNA had a sticky end, which was supplemented by end repair mix to form a flat end. An "A" base was added to the 3′ end to connect the Y-shaped junction. The library was enriched, 15 cycles were amplified by PCR, and then the target bands were recovered with 2% agar gum. TBS380 (PicoGreen) was used to quantify the bands. The cBot was amplified by bridge PCR to generate clusters. Finally, the sequence of Illumina Hiseq was carried out (PE library, reading length 2 × 150 bp).

Assembly and functional annotation

De novo transcriptome was assembled from scratch by using Trinity (https://github.com/trinityrnaseq/trinityrnaseq/wiki) to assemble the short sequences of clean data [83], after which TransRate (http://hibberdlab.com/transrate/) and CD-HIT (http://weizhongli-lab.org/cd-hit/) were used to optimize the initial assembly sequence filter [84, 85], and BUSCO software (Benchmarking Universal Single-Copy Orthologs, http://busco.ezlab.org) was used for assessment [86]. Finally, transcripts and Unigenes were obtained. Generally, the longest sequence of each transcript was selected as the Unigene.

BLAST (Basic Local Alignment Search Tool) was used to search and compare transcript open reading frames (ORFs) with the Nr (non-redundant), String, Swiss-Prot, KEGG and other databases, to obtain the corresponding transcript annotation information [87,88,89]. The expected e-value was < 1e-5. As to the annotation results of BLAST, the Blast2GO program was used (http://www.blast2go.com/b2ghome) for Unigene processing to obtain the classification and description in the database based on the Unigene annotation results [90,91,92]. The expression level of each Unigene was represented by its FPKM value (Fragments Per Kilobase of exon model Per Million mapped reads). Higher fragment abundance indicated a higher gene expression level, and the value was calculated based on RSEM (http://www.biomedsearch.com/nih/RSEM-accurate-transcript-quantification-from/21816,040.html) [93].

Identification of chemosensory genes

With tBLASTn, the available sequences of OBP, CSP, OR, GR, IR, and SNMP proteins from Insecta species were used as queries to identify candidate Unigenes involved in olfaction in S. bifasciatus. All candidate OBPs, CSPs, ORs, GRs, IRs, and SNMPs were manually checked by evaluating the NCBI BLASTx results. ORF Finder online software (http://www.ncbi.nlm.nih.gov/gorf/gorf.html) was used to predict the ORFs of genes.

Phylogenetic analysis

The candidate OBPs and CSPs were searched for the presence of N-terminal signal peptides using SignalP5.0 (http://www.cbs.dtu.dk/services/SignalP/). TMHMM server v3.0 was used to predict the transmembrane domains of candidate ORs, IRs, and GRs (http://www.cbs.dtu.dk/services/TMHMM/). ClustalX software was used for multiple alignment of amino acid sequences [94] (Additional file 4). The maximal-likelihood (ML) method of MEGA X software was used to construct a phylogenetic evolutionary tree (with 1000 Bootstrap replicates) including S. bifasciatus and other proximal insects, whose amino acid sequences were downloaded from the protein database of NCBI [95] (Additional file 5). A JTT matrix-based approach was used to calculate the evolutionary distance [96]. Figtree 1.4.3 and ITOL (https://itol.embl.de/) software were used to edit the phylogenetic tree. Heat-plot of FPKM values plotted using TBtools v1.098652 [96, 97].The transcript levels were determined by calculating log2 (reads per kilobase per million mapped reads + 1) values.

Tissue expression analysis of OBP genes

The relative expression levels of OBP genes in both male and female tissues (antennae, legs, head and tail of the abdomen) were analyzed by fluorescence quantitative real-time PCR. The antennae, legs, heads and abdomens were collected from 12 male adult S. bifasciatus and 12 female adult S. bifasciatus for each biological replicate. Three biological replicates and three technical replicates were used with each qPCR reaction for each tissue to examine reproducibility. The total RNA of each sample was extracted by the above method, the OD260/280 values of the samples were all in the range of 1.8–2.0 by Nanodrop 8000 spectrophotometer. cDNA from various tissues of S. bifasciatus was synthesized with the PrimeScript RT kit (No. RR047A; Takara, Shiga, Japan) including gDNA-Eraser. The internal reference gene was the UBC gene of S. bifasciatus. GAPDH, RPL13, UBC, RPS3, EF1-α, AK and α-tubulin genes were selected from the antenna transcriptome of S. bifasciatus. Using real-time fluorescence quantitative pre-experiments, the genes with the most stable expression were tested as internal reference genes. Primer3plus (http://www.primer3plus.com/cgibin/dev/Primer3Plus.cgi) was used to design specific primers for odorant binding protein genes and internal reference genes of S. bifasciatus, with a primer length of 20 bp, GC content of 40–60% and a product length of 150–200 bp.

All OBP genes were subjected to RT-qPCR using the CFX Connect Real-Time System (Bio-Rad, Hercules, CA, USA). The cDNA of each tissue of male and female adults was used as the template, and the total reaction system was 25 µL, including TB Green Premix Ex Taq II (12.5 µL), DDH2O (8.5 µL), forward and reverse primers (1 µL each), and cDNA (2 µL). The RT-qPCR conditions were as follows: 95 °C for 30 s; then 40 cycles of 95 °C for 5 s and 60 °C for 30 s; followed by 65 °C to 95 °C in increments of 0.5 °C for 5 s. The reference gene was UBC. All primers used in the experiment (including the reference gene) are listed in Additional file 6: Table S1. The RT-qPCR data were analyzed by the 2-∆∆CT method [98], the relative expression levels were calculated, and GraphPad Prism version 8.0.0 for Windows (GraphPad Software, San Diego, California USA, www.graphpad.com) was used to graph the results. SPSS 23.0 was used for one-way ANOVA, and the Tukey method was used to test the significance of the difference (P < 0.05) (The version information of the software packages used for transcriptome sequencing and bioinformatics analysis was listed in Additional file 7).