The sophisticated olfactory system, particularly the peripheral chemical signal coding, is essential for insects to find mates, locate food, and avoid predators (Dweck et al., 2013; Tauxe et al., 2013; Strauch et al., 2014; Li and Liberles, 2015). Biologically important odorants are generally perceived sensitively and specifically in the multiporous sensilla hairs on insect antennae (Meijerink and van Loon, 1999; Pophof et al., 2005; Park et al., 2013; Sun L. et al., 2014). It is well established that at least three major classes of molecules are involved in this process: odorant-binding proteins (OBPs), odorant receptors (ORs), and odorant-degrading enzymes (ODEs). In brief, airborne odorants enter the hydrosoluble sensillum lymph through the sensilla pores, bind to OBPs, activate ORs and trigger signal transduction cascades and olfactory coding; odorants are then rapidly removed from the vicinity of the ORs by ODEs to restore the sensitivity of the sensory neuron (Rützler and Zwiebel, 2005; Vogt, 2005; Pelosi et al., 2006; Leal, 2013).

The highly sensitive odorant signal transduction pathway of insects represents an excellent model that researchers can use to develop new environmentally friendly pest-management strategies through targeting key molecules and screening biologically active compounds for behavioral control. Previous functional reports regarding OBPs and ORs indeed led to the rapid discovery of high-efficiency pest repellents and attractants. For example, compounds that are behaviorally active in the mirid bug Adelphocoris lineolatus were successfully screened via studies on the interaction between antenna-enriched AlinOBP10 and its putative ligands (Sun et al., 2013). In the aphid alarm pheromone EBF perception pathway, ApisOBP3 and ApisOBP7 as well as ApisOR5 were proven to be potentially crucial targets for aphid repellent screening (Sun et al., 2012; Zhang R. et al., 2017). However, compared with OBPs and ORs, similar reports on ODEs appear to be rare. Given that the rapid degradation of redundant odorants can rescue the sensitivity of odorant sensory neurons, putative genes encoding insect ODEs that are highly expressed in the chemosensory system should be identified, and their potential roles in odorant degradation deserve thorough exploration.

Carboxylesterases (CXEs) belong to the α/β-fold hydrolase superfamily and are widely distributed in insects and other organisms. CXEs commonly include a conserved catalytic triad (Ser-His-Glu) and specifically catalyze the hydrolysis of ester bonds in various substrates (Oakeshott et al., 1999, 2005). Because most insect species, including hemipteran bugs and lepidopteran moths, utilize aliphatic esters as intraspecific sex pheromones and ovipositional stimulants (Ando et al., 2004; Millar, 2005; Pan et al., 2015), many antennae-biased CXEs have been identified, and their activities associated with sex pheromone and odorant degradation have been assessed (Vogt, 2005; Jacquin-Joly and Maïbèche-Coisne, 2009). The first CXE subfamily of ODEs, known as Apol-SE (Vogt and Riddiford, 1981), or ApolPDE (Ishida and Leal, 2005), was isolated from the giant silk moth, Antheraea polyphemus. Subsequently, genes encoding putative antennal esterases were cloned and described across insect species using a polymerase chain reaction (PCR) strategy. These genes included two other CXEs, ApolODE and Apol-IE, in A. polyphemus (Ishida and Leal, 2002); Mbra-EST from the cabbage armyworm, Mamestra brassicae (Maïbèche-Coisne et al., 2004); D-AP1, a honeybee homolog of CXE in Apis mellifera L. (Kamikouchi et al., 2004); Slit-EST and Snon-EST from the Egyptian armyworm, Spodoptera littoralis, and the Mediterranean corn borer, Sesamia nonagrioides (Merlin et al., 2007); and PjapPDE, cloned from the Japanese beetle, Popillia japonica (Ishida and Leal, 2008). Furthermore, through expressed sequence tag (EST) and RNA-Seq analyses, diverse CXE genes have been identified from various insect species, such as Epiphyas postvittana (Jordan et al., 2008), Spodoptera littoralis (Durand et al., 2010b), Agrotis ipsilon (Gu et al., 2013), Sesamia inferens (Zhang Y. N. et al., 2014), Chilo suppressalis (Liu et al., 2015; Xia et al., 2015), and Spodoptera litura (Zhang et al., 2016).

Convincing evidence obtained through biochemical characterization and enzyme kinetic activity analyses showed that Apol-SE/ApolPDE displays expression specific to male antennal sensilla and exhibits rapid catalytic activity toward the acetate sex pheromone component E6Z11-16:OAc (Vogt et al., 1985; Prestwich et al., 1986; Klein, 1987; Ishida and Leal, 2005). In vitro functional analyses and potential hydrolyzed substrates of CXEs have also been documented in other insect species, particularly lepidopteran moths, whose main sex pheromone components are acetate esters (Ishida and Leal, 2008; Durand et al., 2010a, 2011; He et al., 2014a,b,c, 2015). Additionally, an extracellular carboxylesterase esterase-6 (EST-6) of Drosophila melanogaster has been demonstrated to be a potential ODE for both the sex pheromone ester cis-vaccenyl acetate (CVA) and other bioactive volatile esters, such as pentyl acetate (Chertemps et al., 2012, 2015). All of the available data support potential roles of CXEs in degrading either sex pheromones or host plant odorants containing ester functional groups.

The tea geometrid Ectropis obliqua Prout is a common pest of the tea plant, Camellia sinensis (L.), and causes serious economic damage to tea production (Ye et al., 2014; Zhang G. H. et al., 2014). Multiple electrophysiological and behavioral studies show that E. obliqua relies heavily on chemical cues to locate host plants, oviposition sites and conspecific mates. Furthermore, larval infection of tea plants strongly induces the release of several kinds of host volatiles with ester functional groups, and these ester compounds can in turn regulate the ovipositional preference of E. obliqua adult females (Sun X. L. et al., 2014). Hence, studies on the molecular mechanism of ester odorant degradation are particularly important for the identification of potential target genes mediating oviposition signal inactivation and the development of ODE-based strategies in geometrid pest management.

In this study, we identified putative genes encoding CXEs by analyzing the BLASTX annotations of transcriptomic data. The phylogenetic relationships between the candidate CXEs and homologs in other Lepidoptera species were further analyzed. Finally, the tissue expression patterns of the identified CXEs were investigated in olfactory organs (particularly in the different antennal sensilla) and non-olfactory organs, and potential functional differentiation was discussed.

In the present study, we identified and characterized 35 candidate genes encoding EoblCXEs from the chemosensory organs of the moth E. obliqua through transcriptomic analysis, including 28 full-length sequences. Phylogenetic analyses and tissue- and sex-related expression profiling showed that EoblCXEs exhibited diverse sequence structures, multiple subfamily clades and distinct expression patterns, suggesting potential differentiation of physiological functions among EoblCXEs. As expected, we demonstrated that 12 EoblCXE genes were highly expressed in E. obliqua antennae, particularly EoblCXE7 and EoblCXE13, which were strongly localized to the olfactory sensilla of both sexes. These results were consistent with the previously reported proposition that CXEs function specifically in insect olfaction and are involved in olfactory signal termination and maintenance of the sensitivity of the olfactory sensilla (Vogt and Riddiford, 1981; Vogt et al., 1985; Ishida and Leal, 2005; Vogt, 2005; Durand et al., 2010a,b, 2011; Chertemps et al., 2012, 2015).

Gene sequence identification represents the first step in elucidating the potential physiological functions of insect CXEs. Due to the lack of genomic information for E. obliqua, we identified putative CXEs through a transcriptomic approach. The number of putative CXE genes (35) identified in E. obliqua (Table 1) was comparable to those found in insect species with available genome data, including D. melanogaster (35 genes), A. gambiae (51 genes), A. aegypti (49 genes) and A. mellifera (24 genes). However, this number was significantly greater than the number of CXE genes identified in moth species that utilize ester compounds as intraspecific sex pheromones, including 20 from S. littoralis (Merlin et al., 2007; Durand et al., 2010b), 24 from S. litura (Zhang et al., 2016), 20 from S. inferens (Zhang Y. N. et al., 2014), 19 from C. suppressalis (Liu et al., 2015), 17 from A. ipsilon (Gu et al., 2013), and 30 from Cnaphalocrocis medinalis (Zhang Y. X. et al., 2017).

Insect CXEs expressed in the olfactory system are mainly related to ester odorant degradation, particularly that of lepidopteran moth sex pheromones (Vogt and Riddiford, 1981; Vogt et al., 1985; He et al., 2015); however, E. obliqua females produce sex pheromones containing unsaturated hydrocarbons and enantiomers of epoxy hydrocarbons rather than acetate esters (Yang et al., 2016a). Thus, it is particularly interesting that the number of CXE genes from E. obliqua was significantly higher than that in species that utilize ester compounds as intraspecific sex pheromones. We speculate that this situation might be attributed to either the use of a different sequencing strategy or the fact that E. obliqua adults depend on the detection of multiple odorants with ester functional groups to find host plants and egg-laying sites. Our use of a high-throughput RNA-sequencing approach in the chemosensory organs (adult antennae of both sexes, legs, wings and proboscises) enabled us to identify as many CXE genes from E. obliqua chemosensory organs as possible. The tea plant, which is the most preferred host of E. obliqua, releases large quantities of ester compounds, particularly under attack by E. obliqua caterpillars; these ester compounds, such as (Z)-3-hexenyl hexanoate and (Z)-3-hexenyl acetate, can in turn regulate the host searching and ovipositional preferences of E. obliqua female adults (Wang, 2010; Sun X. L. et al., 2014). Hence, we propose that some EoblCXEs might act as candidate ODEs and potentially exhibit crucial physiological functions in the degradation of tea plant volatiles with ester functional groups rather than the degradation of sex pheromone components produced by E. obliqua females. This inference corresponds well to the tissue- and sex-related expression patterns found in this study, in which only EoblCXE10 of the antennae-enriched EoblCXE genes displayed a male-specific expression pattern; the other EoblCXEs were either highly abundant in female antennae or were not sex-biased (Figure 3); however, even if then, this inference remains to be supported by the in vitro biochemical enzymatic experiment.

It should be noted that insect CXE genes belong to a multigene family that encodes sequence-divergent and functionally diverse proteins (Oakeshott et al., 1999; Tsubota and Shiotsuki, 2010). The 35 candidate EoblCXEs, which show average amino acid identities lower than 30%, fall into at least 10 different subclades; seven of these clades, which contain 33 EoblCXEs, possess clearly conserved functional characteristic features of the α/β-hydrolase structure, such as the Ser-Glu-His catalytic triad and the nucleophilic elbow surrounding the active-site serine residue (Gly-X-Ser-X-Gly) (Figure 1, Table 2). These features suggest that most EoblCXEs are catalytically active and participate in the degradation of diverse biologically important compounds. Other CXEs fall into the neuroligin, neuroreceptor, or gliotactin clades, including EoblCXE19 in the neuroreceptor clade and EoblCXE32 among the neuroligins, and lack the crucial residue Ser responsible for catalytic activity (Table 2); thus, these CXEs are considered to be catalytically inactive and are mainly involved in neurological and developmental functions related to sensory processing (Biswas et al., 2008; Durand et al., 2017).

Ten EoblCXEs belonging to the dipteran microsomal α-esterase or mitochondrial and cytosolic esterase clade lack a predicted signal peptide, indicating that they are intracellular esterases. These clades (particularly the α-esterases) are well-known for their involvement in the detoxification of insecticides and xenobiotics and the digestion of dietary esters (Newcomb et al., 1997; Campbell et al., 2003; Liang et al., 2007; Tang et al., 2011; Yang et al., 2016b; Gong et al., 2017). An orthologous gene of CXE10 has been functionally studied in two closely related Spodoptera species, S. littoralis (SlittCXE10) and S. exigua (SexiCXE10) (Durand et al., 2010a; He et al., 2015). Both SlittCXE10 and SexiCXE10 are reportedly expressed in the olfactory sensilla and preferentially degrade an ester plant volatile, (Z)-3-hexenyl acetate. EoblCXE10 shares high amino acid identity with SlittCXE10 and SexiCXE10 and is therefore considered an orthologous gene in E. obliqua. Its tissue expression was found to be restricted to the antennae of adults, particularly adult males, and resembled the distribution of its orthologs in Spodoptera species (Figure 3), implying a similar role in the degradation of host plant volatile compounds.

Approximately half of the candidate EoblCXEs clustered in the moth antennal esterase or ß-esterase and pheromone esterase clade and presented both a catalytically active Ser residue and a predicted signal peptide (Table 2). These CXEs represented typical secreted or extracellular esterases that could be secreted into the sensillum lymph surrounding the sensory neurons or into the hemolymph filling the antennal lumen, indicating potential roles in the degradation of odorants and the maintenance of OSN sensitivity. Indeed, functional reports regarding several members of these clades, such as ApolPDE (Vogt and Riddiford, 1981; Ishida and Leal, 2005), PjapPDE (Ishida and Leal, 2008), SexiCXE13 (He et al., 2014a), SexiCXE14 (He et al., 2014c), and SlittCXE7 (Durand et al., 2011), indicate that they play crucial roles in the degradation of insect sex pheromones and biologically important plant volatiles.

To gain further insight into the physiological roles of the EoblCXEs involved in hydrolyzing tea plant volatiles in E. obliqua olfaction, we selected EoblCXE7 and EoblCXE13 for a fluorescence in situ hybridization assay, not only because EoblCXE7 and EoblCXE13 are comparably expressed in the antennae of both E. obliqua sexes but also because EoblCXE13 clusters with ApolPDE, PjapPDE and SexiCXE13, while EoblCXE7 falls into the same clade as SlittCXE7 and SexiCXE14; therefore, these genes provide a useful model for comparative analyses of the functional evolution of CXE orthologs across lepidopteran species. In contrast to SlittCXE7 (Durand et al., 2011), the localization of EoblCXE7 and EoblCXE13 at the sensillum level is more complex. EoblCXE7 exhibits a similar cellular localization at Str and Sba between the sexes, whereas EoblCXE13 is extensively expressed at multiple sensilla of female E. obliqua but is mainly restricted to Str I in males. Interestingly, (Z)-3-hexenyl acetate, a substrate of both CXE13 and CXE7 in Spodoptera species (Durand et al., 2011; He et al., 2014a), can attract both virgin males and mated females in E. obliqua. These results suggested that EolbCXE7 and EoblCXE13 might function in the E. obliqua olfactory system; however, whether the functions of EolbCXE7 and EoblCXE13 resemble those of their orthologs SlittCXE7 and SexiCXE13 as candidate ODEs in the degradation of (Z)-3-hexenyl acetate remains to be further confirmed. Additionally, EoblCXE7 and EoblCXE13 were highly expressed at Sst, a putative gustatory sensillum in lepidopteran species (Zenker et al., 2011; Tang et al., 2015), and we therefore cannot rule out the possibility that both of these EoblCXEs might function in the gustatory system.

In summary, the present study provides the first identification and characterization of the expression patterns of candidate CXE genes in E. obliqua, a common lepidopteran insect pest of the Geometridae, which will aid the development of new pest management techniques using CXE as potential targets for the disruption of insect foraging behavior.

LS, QianW, and YoZ conceived and designed the experimental plan. LS, QianW, and QiW preformed the experiments. LS, QianW, YuZ, MT, HG, JF, QX, and YaZ analyzed the data. LS and QianW drafted the manuscript.