Analysis of the Antennal Transcriptome and Identification of Tissue-specific Expression of Olfactory-related Genes in Micromelalopha troglodyta (Lepidoptera: Notodontidae)

Abstract Micromelalopha troglodyta (Graeser) has been one of the most serious pests on poplars in China. We used Illumina HiSeq 2000 sequencing to construct an antennal transcriptome and identify olfactory-related genes. In total, 142 transcripts were identified, including 74 odorant receptors (ORs), 32 odorant-binding proteins (OBPs), 13 chemosensory proteins (CSPs), 20 ionotropic receptors (IRs), and 3 sensory neuron membrane proteins (SNMPs). The genetic relationships were obtained by the phylogenetic tree, and the tissue-specific expression of important olfactory-related genes was determined by quantitative real-time PCR (qRT–PCR). The results showed that most of these genes are abundantly expressed in the antennae and head. In most insects, olfaction plays a key role in foraging, host localization, and searching for mates. Our research lays the foundation for future research on the molecular mechanism of the olfactory system in M. troglodyta. In addition, this study provides a theoretical basis for exploring the relationship between M. troglodyta and their host plants, and for the biological control of M. troglodyta using olfactory receptor as targets.

Complex chemical sensory mechanisms are essential for the survival and reproduction of insects. The accurate olfactory sensory system plays a role in many insect behaviors, such as searching for food, finding mates, choosing spawning sites, and avoiding natural enemies (Brigaud et al. 2009, Zhou 2010, Lu et al. 2015. As an important part of the tissues insects use to sense and recognize odors, antennas are distributed with a large number of olfactoryrelated proteins. The proteins related to the insect olfactory system include: odorant receptors (ORs), odorant-binding proteins (OBPs), chemosensory proteins (CSPs), ionotropic receptors (IRs), odorant degrading enzymes (ODEs), and sensory neuron membrane proteins (SNMPs) (Leal, 2005).
The OR genes of insects are highly diverse, and the numbers between different species vary greatly; most of them posses between 50 and 100 (Nei et al. 2008). The first insect ORs were identified in Drosophila melanogaster Meigen by screening genomic data (Qiao et al. 2008). Insect odorant receptors (ORs) are a class of transmembrane proteins that form an ion channel on the dendritic membrane of neurons. ORs are mainly divided into two types. One is the conventional odorant receptor (specific OR), which has low homology among different insects. The other receptor is an atypical common receptor now known as Orco, which is highly conserved among different insects (Vosshall and Hansson 2011). Some studies have suggested that Orco acts as a partner of odorant receptors and works together with them (Grosse-Wilde et al. 2011, Stengl andFunk 2013).
Odorant-binding proteins (OBPs) are a class of spherical hydrophilic proteins located in the lymph fluid of insect antennae. OBPs can be divided into three categories: sex pheromone binding proteins (PBPs), general odorant binding proteins (GOBP), and antenna bind proteins (ABPx) (Vogt and Riddiford 1981, Vogt et al. 1991, Krieger et al. 1996. It has been reported that PBPs, as members of the sex pheromone identifying protein family, mainly bind to insect sex pheromones, while GOBPs are involved in identifying common odorants and sex pheromones (Grosse-Wilde et al. 2006, Wang et al. 2022. Wang et al. (2020) found that two potential PBPs (BgerOBP26 and BgerOBP40) were identified in Blattella germanica (L.). OBPs, as important participants in insect olfactory behavior, combine with lipophilic odorants in lymph fluid to form complexes, and transport them to the dendritic membrane of sensory neurons. The combination with odorants is the first step to start odorant perception. In addition, other proteins such as chemosensory proteins (CSPs), ionotropic receptors, odorant degrading enzymes(ODEs), sensory neuron membrane proteins (SNMPs), and ionic receptors (IR) are also involved in the process of odorant perception (Scott et al. 2001).
There are obvious differences between CSPs and OBP in sequence, structure, expression profile, and so on (Pieimbon et al. 2000). Unlike OBP, which is highly enriched in antennae, CSPs have a wide range of tissue expression profiles (Wang et al. 2005, Pelosi et al. 2006. ODEs are considered to play key roles in odorant inactivation to maintain the odorant receptor sensitivity of insects. Some members of carboxylesterase (CXE) are a major subfamily of ODEs.  found that a ubiquitous expression esterase SexiCXE11 may be partly involved with olfaction. In addition, SNMPs are the important olfactory functional proteins in insects. Liu et al. (2014) found that SexiSNMPs have dual functions in receptive pheromone and common odor recognition in Spodoptera exigua (Hübner). Furthermore, the ionotropic glutamate receptor (iGluR) family were named ionic receptors (IR) in D. melanogaster (Benton et al., 2009). 12 IRs genes have been identified in the antennae of Spodoptera litura (Fabricius) (Olivier et al., 2011);15 IR S gene have been found in the antenna of Cydia pomonella (L.) (Bengtsson et al., 2012).
M. troglodyta larvae feed on the leaves of poplars and are the main pests that endanger poplars. There have been outbreaks in the northeast, northwest, central plains, and south central China. Especially in the case of a single tree species, the leaves are often completely eaten during pest outbreaks, which not only affects the normal growth of poplars and causes economic losses but also destroys the ecological environment and restricts the development of forestry (Tang et al. 2009).
Olfaction plays an important role in life activities such as the feeding, mating, and positioning of M. troglodyta. As an herbivorous insect, its sensitivity and binding effect to plant volatiles can improve its selectivity and adaptability to host plants to enhance its invasion ability and lead to more serious harm. Although extensive studies have been carried out on the molecular mechanism of olfaction and the identification of chemical sensory genes in many Lepidoptera species (Elfekih et al. 2016), they may not be applicable to M. troglodyta. Therefore, the identification of olfactory genes is very important to clarify the molecular mechanism of olfaction and verify the unique olfactory receptors in M. troglodyta. However, there is no report on the olfactory repertoire and its tissue-specific expression in M. troglodyta in China or elsewhere. In this study, the transcriptome of the M. troglodyta antenna was analyzed, important olfactory-related genes (ORs, OBPs, SNMPs, IRs, and CSPs) were identified, and phylogenetic trees were used to evaluate their phylogenetic relationships with other species. In addition, the tissue-specific distribution of all olfactory-related genes was determined by qRT-PCR. These results will lay the foundation for clarifying the molecular mechanism of olfactory perception in the future. In addition, understanding olfactory behavior is essential for developing eco-friendly control strategies. For example, the combination of insect repellents or attractants is used to regulate the number of pests and natural enemies. Additionally, this study also provides a theoretical basis for exploring the relationship between M. troglodyta and host plants, which is important for the biological control of M. troglodyta using olfactory receptors as targets.

Experimental Insect Rearing and Sample Preparation
M. troglodyta population was collected from poplar in Nanjing and brought back to the laboratory for feeding. The larvae were placed in the incubator under the conditions of 26 ± 1°C, 70-80% RH, and a photoperiod of 16:8 (L:D) h, and fresh poplar leaves were collected daily as food.
The antennae, heads (without antennae), cuticles, midguts, forewings, and hindwings were dissected from adults, and then frozen in liquid nitrogen and stored in freezer at −80℃ for subsequent RNA extraction. Antennae were used for subsequent transcriptome sequencing. All tissues were used for expression profiling.

RNA Isolation and Sequencing
Around 10 mg of tissues (mixed antennae) were extracted by AG RNAex Pro Reagent (AG21102, Accurate Biotechnology, Hunan, Co., Ltd). The extracted RNA was first analyzed by gel electrophoresis on 1% (w/ V) agarose gel to detect the integrity of RNA, and then the concentration of RNA was measured by spectrophotometer (Eppendorf Bio Spectrometer). Solexa sequencing using an Illumina HiSeq 2000 was performed by Shenzhen Huada Gene Research Institute.

Sequence Assembly and Annotation
Trinity was used to perform the de novo assembly and eliminate duplicates. Then, Tgicl was used when the transcripts were aggregated into unigenes. The unigenes were blasted to compare with five databases (Nt, Nr, COG, KEGG, and SwissProt) to obtain the annotation results and the gene function was annotated from GO database using Blast2GO (Conesa et al. 2005).

Phylogenetic Analysis
Phylogenetic trees were built based on amino acid sequence alignment of the candidate genes from M. troglodyta and those of other insects' species using ClustalX2.0. The phylogenetic tree was constructed using the Neighbor joining method (NJ), and the P-distances modeling and a pairwise deletion of gaps were performed by the MEGA X software package. The reliability of the tree structure and node support was evaluated by bootstrap analysis with 1,000 replicates (Tamura et al. 2013).

Tissue-specific Expression of Olfactory-related Genes in M. troglodyta
The cDNA concentration of the tissues was adjusted to 50 μg/ml and the whole reaction preparations were carried out on ice. The primers for qRT-PCR were designed by online software (https:// sg.idtdna.com/scitools/Applications/RealTimePCR/) ( Table 1). qRT-PCR was performed with an ABI 7500 (Applied Biosystems) using SYBR Premix Ex Taq (Tli RNaseH Plus) (Takara, Japan). β-actin was used as reference gene. The melting curve was analyzed to evaluate the specificity and sensitivity of these primers, and LinReg PCR software (Version: September 2014) was used to examine the amplification efficiency of each gene. Each PCR system was set to 20 μl, including 10.0 μl SYBR premier ex Taq, 0.4 μl Rox II, 0.4 μl forward and reverse primer (10 mm), 2 μl sample cDNA (100 ng), and 6.8 μl double distilled water. qRT-PCR was divided into three stages. First, predenaturation was maintained at 95°C for 30 s, then 95°C for 5 s, 60°C for 34 s and 40 cycles at this stage, and finally 95°C for 15 s, 50°C for 1 min, 95°C for 15 s. At the end of the reaction, the amplification curve and the melting curve were confirmed. All experiments were performed four repetitions. The relative expression level of M. troglodyta mRNA was calculated by the 2 −ΔΔCt method (Giulietti et al. 2001).

Data Analysis
Data collected from these assays were subjected to analysis of variance using InStat software (GraphPad, San Diego, CA). The statistical significance of multiple sample comparisons was calculated using a one-way analysis of variance followed by Tukey's multiple comparisons. A value of P < 0.05 was considered statistically significant.

Transcriptome Sequencing and Sequence Assembly
In total, we obtained 59,076,354 clean reads from M. troglodyta antennal transcriptome. The percentages of clean reads with Q20 and Q30 quality scores were 98.65 and 95.56%, respectively. 101,597 transcripts were obtained by using Trinity, with an N50 of 1,587 bp, an average length of 740 bp, and a total length of 75,206,425 bp. 41,300 unigenes were selected from the above transcripts with a mean length of 1,092 bp and an N50 length of 2,118 bp. After eliminating repeated and short-length sequences, 41,300 Nr unigenes were annotated against five public databases using similarity searching. In total, 23,405 (56.67%) genes were annotated in the NR database; 16,677 (40.38%) genes were annotated in the Swiss-Prot database; 117,744 (42.96%) genes were annotated in the KEGG database; 8,790 (21.28%) genes were annotated in the KOG database; 2,822 (6.83%) genes were annotated in the GO database. In addition, 14,537 (35.2%) genes were not annotated with any database.

Functional Annotation of the M. troglodyta Transcriptome
The greatest number of unigene matches was for insect genomes, and Amyelois transitella (Walker) (22.98%), Bombyx mori L. (19.61%), Papilio xuthus L. (9.17%), and Papilio machaon L. (8.45%) accounted for the top 4 unigene matches based on the Nr annotations. The remaining 39.79% of the sequences showed good homology with those of other insects (Supp Fig. S1 [online only]). In total, 8,790 sequences were subjected to COG classification, and they were divided into 25 COG groups using WebMGA, with an E-value cut-off of 1e −5 . This may be related to the fact that there is still currently little data on M. troglodyta in the COG database. The unigenes were annotated to the COG database and their possible functions were predicted. From the 25 COG categories, the cluster for 'General function prediction only' (2,664) was the largest group, followed by the 'Translation, ribosomal structure and biogenesis' (1,500) group. The 'Nuclear structure' (3) and 'Extracellular structures' (9) groups were the smallest classes ( Fig. 1).

Gene Ontology Annotation
The GO database is commonly used for gene functional annotation (Ashburner et al. 2000). Using Blast2Go software, the transcriptome of M. troglodyta was successfully mapped to the three main functional processes, which comprise 52 GO terms. In other words, according to the GO gene functional classification system, 2,822 unigenes were divided into the three major functional ontologies, biological process (BP), cellular component (CC), and molecular function (MF). In the category of biological processes, the main subcategories were cellular process (1,389) and metabolic process (1,373), followed by single-organism process (904). For the cellular component category, cell (1,062), cells parts (1,042), and membrane (747) were the most highly represented. In terms of molecular function, binding (1,243) and catalytic activity (1,210) were highly enriched. Nevertheless, in all three main categories, few genes were assigned to cell killing, nucleoid, behavior, and synapse part (Fig. 2).

Identification and Tissue-specific Expression of Olfactory-related Genes
Through the analysis of the antennal transcriptome data of M. troglodyta, a total of 142 olfactory-related genes were identified, including 74 OR genes, 32 OBP genes, 13 CSP genes, 20 IR genes, and 3 SNMP genes. We downloaded the OR, OBP, CSP, IR, and SNMP gene sequences of other insects on NCBI, 5 phylogenetic trees were successfully constructed by the neighbor-joining (NJ) method using MEGA X. At the same time, we selected some genes to determine the tissue expression pattern.

Identification and Tissue Expression Profiles of Putative Odorant Receptor (OR) Genes from M. troglodyta
Through the comparative analysis of the antennal transcriptome data of M. troglodyta, 74 OR genes were identified. After obtaining the sequences of these genes, the ORFs of these genes were searched and determined by ORFfinder. The results showed that 11 of them had incomplete ORFs. In addition, 25 sequences were found to coincide with the others of the 49 sequences by amino acid sequence alignment. Therefore, only 49 sequences with differences were selected to construct the phylogenetic tree. Through the phylogenetic tree, these OR genes were scattered in different branches. We found that MtroOR38 clustered with other insect Orco genes. According to the high conservation of Orco sequences among different insect species, MtroOR38 was identified as an Orco. In addition, 8 genes, such as MtroOR48, were clustered with pheromone receptors (PRs), and these genes were identified as PR (Fig. 3A, Supp Table S1 [online only]). Eight OR genes with high expression in the transcriptome were selected from the identified OR genes, and their tissue expression profiles were measured by qRT-PCR. From the results, we found that among the 8 OR genes, 6 genes, MtroOR1 (F = 99.447, df = 5,18, P < 0.0001), MtroOR7 (F = 671.81, df = 5,18, P < 0.0001), MtroOR13 (F = 598.52, df = 5,18, P < 0.0001), MtroOR19 (F = 89.793, df = 5,18, P < 0.0001), MtroOR27 (F = 294.41, df = 5,18, P < 0.0001), and MtroOR29 (F = 832.83, df = 5,18, P < 0.0001), were highly expressed in the antennae. Among them, the expression of the MtroOR13 gene in antennae and head was very similar, and there was no significant difference. The relative expression of the other two OR genes, MtroOR10 (F = 586.3, df = 5,18, P < 0.0001) and MtroOR16 (F = 132.8, df = 5,18, P < 0.0001), in the head was significantly higher than that in the antennae (Fig. 3B).

Identification and Tissue Expression Profiles of Putative Odorant Binding Protein (OBP) Genes from M. troglodyta
OBPs have a conserved cysteine pattern of C1-X 25-30 -C2-X 3 -C3-X 36-42 -C4-X 8-14 -C5-X 8 -C6 (Xu et al. 2009). After detailed analysis of the assembled transcriptome and sequences alignment, 32 candidate OBPs were finally identified. However, during sequences alignment, we found that most candidate OBP sequences are less similar to known Lepidopteran OBPs, which may be due to the low sequences conservation among OBPs in different families. Among 32 candidate OBPs sequences, 31 candidate OBPs contained a complete ORF. In addition, 4 sequences were found to coincide with the other 28 sequences by amino acid sequence alignment. Therefore, only 28 sequences with differences were selected to construct the phylogenetic tree. According to the phylogenetic tree, these OBPs were clustered with different subfamilies, of which MtroOBP6 and  MtroOBP14 aggregated with the identified GOBP. MtroOBP28 and MtroOBP29 were classed into a subgroup in the phylogenetic tree with previously characterized PBP. In addition, MtroOBP9, MtroOBP10, and MtroOBP21 were distributed in the Minus-C subfamily, and MtroOBP31 and MtroOBP24 were distributed in the Plus-C subfamily (Fig. 4A, Supp Table S2 [online only]).

Identification and Tissue Expression Profiles of Putative Chemosensory Protein (CSP) Genes from M. troglodyta
Similar to OBPs, the CSPs have a conserved cysteine pattern of C1-X 6-8 -C2-X 16-21 -C3-X 2 -C4-X 3 (Xu et al. 2009). Through sequence alignment and bioinformatics analysis, we obtained 13 putative CSPs from the transcriptome. Except for one sequence (MtroCSP13), the remaining 12 all contained the complete ORF. The sequences with complete ORF were selected to construct the phylogenetic tree. According to the sequence alignment results, we could see that the CSPs identified had higher similarities with the CSPs identified in Lepidoptera, which also implied that the CSPs sequence was more conservative. The phylogenetic tree also showed that these CSP genes were clustered together with other insect CSPs and distributed in different branches (Fig. 5A, Supp Table S3 [online only]).

Identification and Tissue Expression Profiles of Putative Ionotropic Receptor (IR) Genes from M. troglodyta
In total, 20 IR genes were identified from the antennal transcriptome. Sequence alignment results showed that most of them had high identity with other insect IR genes. However, 7 gene sequences were short, and it was speculated that they might be partial rather than full length, so they were not included in phylogenetic trees. The results of the phylogenetic tree showed that all IR genes were clustered with IR genes of other insects, of which MtroIR1 and MtroIR2 were clustered with genes of the IR21a family, indicating that these two genes might belong to the IR21a family. MtroIR6 was clustered with the genes of the IR25a family, which means that IR6 may be a member of the IR25a family (Fig. 6A, Supp Table S4 [online only]).
We selected 8 IR genes with high expression in the transcriptome and verified the tissue expression profiles of these genes by qRT-PCR. The results showed that the tissue expression patterns of the eight IR genes were different. MtroIR8 (F = 126.18, df = 5,18, P < 0.0001) was only expressed in the antenna and midgut. The expression level of MtroIR1 (F = 219.32, df = 5,18, P < 0.0001) was the highest in the antennae, followed by the cuticle. The remaining 6 IR genes were expressed in various tissues, and the tissues with the highest relative expression levels were different. Among them, MtroIR7 (F = 5689.6, df = 5,18, P < 0.0001) was only highly expressed in antennae, and MtroIR6 (F = 718.68, df = 5,18, P < 0.0001) was highly expressed in antennae and hind wings, and MtroIR10 (F = 31.571, df = 5,18, P < 0.0001) only had the highest relative expression in the hindwings, and MtroIR12 (F = 318.67, df = 5,18, P < 0.0001) was only highly expressed in the cuticle (Fig. 6B).

Identification and Tissue Expression Profiles of Putative Sensory Neuron Membrane Protein (SNMP) Genes from M. troglodyta
In total, 3 SNMP genes were identified from the antennae transcription of M. troglodyta. After sequence alignment, it was found that MtroSNMP1 and MtroSNMP2 had complete ORFs, while MtroSNMP3 had an incomplete ORF and the sequence length was too short. The sequences with complete ORFs were selected to construct the phylogenetic tree. The phylogenetic tree results showed that MtroSNMP1 and MtroSNMP2 were divided into two subfamilies, and clustered with the SNMP genes of the corresponding subfamilies. It indicated that MtroSNMP1 belonged to the SNMP1 subfamily, and MtroSNMP2 belonged to the SNMP2 subfamily (Fig. 7A, Supp Table S5 [online only]).
Insect odorant receptors are highly differentiated members of the multigene family (Vosshall et al. 1999). At present, they are mainly divided into conventional odorant receptors and Orco (OR83b) (Pelletier et al. 2010). The insect odorant receptor family is very large. Since the first odorant receptor was identified in D. melanogaster (Pennisi, 1999) in 1999, a large number of other odorant receptors have been identified, including 79 ORs in the Anopheles gambiae Giles genome and 48 ORs in the B. mori genome (Fox et al., 2001, Sakurai et al., 2004, Nakagawa et al., 2005, Wanner et al., 2007a. By analyzing these identified odorant receptor sequences, it was found that the conventional odorant receptors have very low homology among insect species, and the pheromone receptors (PRs) have some homology in Lepidoptera moth insects (Krieger et al. 2004, Nakagawa et al. 2005, Wanner et al. 2007aMitsuno et al. 2008, Miura et al. 2009, 2010Patch et al. 2009). However, Orco (OR83b) is highly conserved among different insect species, and the amino acid sequence identity can reach more than 70% (Jones et al. 2005, Yin et al. 2013. From the antennal transcriptome we identified a total of 74 ORs, of which 49 contained a complete ORF. The NJ phylogenetic tree constructed with OR sequences from other insects showed that MtroOR38 was clustered in one branch with PxylOR83b, HarmOR83b, and HzeaOR83b. Additionally, according to the unified nomenclature  system proposed by Vosshall (Vosshall et al. 2011), we considered MtroOR38 to be a conservative Orco in M. troglodyta. We also found that 8 genes, including MtroOR48, aggregated with the genes of the PR family. These 8 genes of M. troglodyta were presumed to belong to the PR family by sequence comparison, and may play a role in detecting pheromones, especially sex pheromones. We studied the tissue expression patterns of ORs by qRT-PCR. As with most reports, a large number of ORs were mainly expressed in the antennae. In particular, the relative expression of MtroOR10 and MtroOR16 in the head was higher than that in the antennae. It was speculated that the mouthparts also had the distribution of olfactory sensors (Malpel et al. 2008). This finding suggested that MtroOR10 and MtroOR16 may be involved in other physiological activities. Among other insects, such as S. litura, SlitOR2 was slightly expressed in the head (Brigaud et al. 2009), and A. gambia (Pitts et al., 2004) and Spodoptera frugiperda (Smith) (Krieger et al., 2002) also have a small amount of OR2 expression. Therefore, further functional research on MtroOR will help to better understand the mechanism of M. troglodyta recognition of external odors at the OR level.
OBPs are important participants in insect olfactory behavior, and the combination with the odorant in the environment is also the first step to start odorant perception. Then, these odorants are transported to ORs through the lymph. (Vogt et al. 1999, Leal 2005. Based on the antennal transcriptome, we obtained 32 OBPs through homology alignment. Through the NJ phylogenetic tree constructed with other insect OBPs, we found that these OBPs were divided into several different branches. MtroOBP6 and MtroOBP14 were presumed to belong to the GOBP subfamily and were also related to GOBP1 and GOBP2 in M. troglodyta, respectively. According to the number of conserved cysteines, OBPs are divided into different subfamilies, which means that their molecular structures are different. This also implies that these OBPs may also participate in other life activities in addition to participating in olfactory mechanisms. There have been many reports that OBPs are highly expressed in insect antennae. We also studied the tissue expression patterns of OBPs by qRT-PCR. Most of these OBPs are abundantly expressed in antennae, indicating that these genes have a specific role in the olfactory process. In addition, there have been reports that OBPs are also expressed in other parts of the insect body and at different growth stages. Similar results were found in our study; e.g., although MtroOBP9 has the highest relative expression in the antennae, it also has a high expression in the head. It is also worth noting that MtroOBP7 has the highest expression in the antennae, but it also has a high expression in the forewings. Another gene, MtroOBP31, is only highly expressed in the head, which is significantly different from the relative expression in the antennae. This indicates that insect OBP exists in other tissues (wings) in addition to antennae, which may be related to adapting to complex olfactory activities. It has been reported that in S. litura (Gu et al., 2015), 3 OBPs showed body-specific expression, and 18 OBPs had similar expression levels in the antennae and other body tissues.
In total, 13 CSPs were identified, of which 12 CSPs had complete ORFs. Insect CSPs can identify nonvolatile chemicals in the environment and they are a class of acidic soluble proteins. CSP also has a structural pattern similar to OBP, but the difference is that CSP has only 4 conserved cysteines (Xu et al. 2009). It is more conservative than OBPs. Our NJ phylogenetic tree and sequence alignment have also proved that CSPs have high sequence homology. In addition, the expression profile of CSP in B. mori was different at different  developmental stages, and it was expressed in multiple tissues, which indicates that CSP bears many complex chemosensory functions (Gong et al. 2007). Previous studies have shown that antennae are not the only tissue for CSP gene expression, and the tissue expression profiles of these genes are relatively diverse. However, our results showed that MtroCSP2 and MtroCSP3 are antennae-specifically expressed, which suggests that these two genes play an important role in the olfactory process. The differential expression of other genes also precisely reflected the diversity of CSP gene function. This suggests that they may also have chemical sensing and other nonolfactory effects.
Additionally, proteins such as IRs, SNMPs, and GRs are also involved in the olfactory process of insects and have been identified in our antennal transcriptome, with the exception of GRs. The reason why GRs have not been identified may be that GRs are mostly expressed in the taste organs of insects, and the expression amount in the antennae is very small. We also identified 20 IRs and 3 SNMPs and constructed NJ phylogenetic trees using sequences containing complete ORFs. IRs are a class of conservative chemosensory receptors. The study of such genes has been relatively recent. Benton identified IRs in the D. melanogaster genome for the first time in 2009 (Benton et al. 2009). It was found that IRs of different tissues might play different functions in Drosophila melanogaster (Ai et al., 2013). Neurons expressing IR84a in antennae participated in the process of Drosophila looking for food and determining spawning sites (Dahanukar et al. 2005). Croset et al. (2010) found that DmilIR8a was mainly highly expressed in the antennae of Drosophila, suggesting that it might be involved in the function of olfactory recognition. In our experiment, MtroIR1, MtroIR 6, MtroIR 7, and MtroIR 8 were also highly expressed in antennae, so it was speculated that they might also be involved in olfactory recognition. Recent research results show that IRs play an important role in Drosophila's perception of environmental temperature changes (Ni et al. 2016). IR expressed in dorsal organ cool cells of Drosophila was involved in sensing temperature changes in the environment (Klein et al. 2015). Therefore, we speculated that the high expression of IR in the cuticle might also be related to the perception of temperature changes. SNMPs are specific proteins on olfactory neurons and dendritic membranes (Rogers et al. 2001a), with two subfamilies. The two SNMPs in our transcriptome belonged to the SNMP1 and SNMP2 subfamilies. Many studies have shown that SNMPs are related to insect odorant recognition, but there are still many speculative functions that need to be further studied (Rogers et al. 2001b). The tissue expression patterns showed that these two kinds of SNMP were expressed almost exclusively in the antennae, which just proved that they played an important role in the olfactory process of M. troglodyta and was consistent with the previous results.
Our study constructed the transcriptome of the M. troglodyta antennae, and identified 142 olfactory-related genes. The tissue-specific expression patterns of these genes were analyzed by qRT-PCR, and most of them were expressed in large numbers in the antennae. These results lay the foundation for further research on the olfactory-related proteins of M. troglodyta and provide a theoretical basis for a new approach to controlling M. troglodyta. For example, studying odorant receptors and combining them with RNA interference, could increase the sensitivity of insects to pesticides. Furthermore, it also provides a research foundation for the development of new, efficient, and environmentally friendly attractants or repellents.

Author Contributions
Conceptualization, F.T.; methodology, F.T. and W.L.; performed the experiments, K.F., Y.Z. and R.L.M.; analyzed the data and prepared the figures, Y.Z.; writing-original draft preparation, Y.Z. and K.F.; writing-review and editing, F.T. All authors have read and agreed to the published version of the manuscript.

Supplementary Data
Supplementary data are available at Journal of Insect Science online. Fig. S1. Nr classification of all M. troglodyta unigenes. Table S1. OR genes used in phylogenetic tree. Table S2. OBP genes used in phylogenetic tree. Table S3. CSP genes used in phylogenetic tree. Table S4. IR genes used in phylogenetic tree. Table S5. SNMP genes used in phylogenetic tree.