Correction: Chemosensory Gene Families in Adult Antennae of Anomala corpulenta Motschulsky (Coleoptera: Scarabaeidae: Rutelinae)

Background The metallic green beetle, Anomala corpulenta (Coleoptera: Scarabaeidae: Rutelinae), is a destructive pest in agriculture and horticulture throughout Asia, including China. Olfaction plays a crucial role in the survival and reproduction of A. corpulenta. As a non-model species, A. corpulenta is poorly understood, and information regarding the molecular mechanisms underlying olfaction in A. corpulenta and other scarab species is scant.

In insects, the first step in the recognition of chemical signals involves OBPs and CSPs, which are abundant in the lymph of antennal sensilla [5].OBPs are small, water-soluble proteins composed of approximately 130 amino acids that typically contain six conserved cysteine residues [6][7][8].Minus-C OBPs lack the second and fifth cysteine residues, and plus-C OBPs contain three additional conserved cysteines and one conserved proline [9].It is presumably that OBPs act as molecular carriers of hydrophobic signaling molecules that are taken into the antennae or other sensory organs from the external environment.OBPs transport these molecules through the aqueous space, and deliver them to chemoreceptor proteins located on the membranes of OSNs [10].
CSPs belong to a smaller, more conserved protein family that is characterized by four conserved cysteines that form two disulphide bonds [11,12].Although the molecular functions of CSPs remain poorly understood, it has been suggested that they are involved in semiochemical detection [13], and participate in some nonchemosensory functions [14,15].Initially identified in sex-pheromone-specific OSNs of lepidopterans, SNMPs are CD36-like proteins with two transmembrane domains [16] that are thought to act as cofactors for odorant receptors in pheromone detection [17].
Insect chemoreceptors consist of three major protein families, the ORs, GRs, and IRs, whose activation generates an electrical signal that is transmitted to the brain [18,19].The ORs are receptor proteins on the dendritic membrane surface of OSNs that are composed of approximately 400 amino acids, and contain seven transmembrane domains that form binding sites for odorant molecules [3,20,21].Insect olfactory neurons express odorant-specific ORs and a highly conserved coreceptor, known as Orco whose ortholog in Drosophila is Or83b.The Orco is required for olfactory functions and the localization of specific ORs, and the Orco and specific OR form a heterodimer that acts as an ion channel [22,23].Although GRs are members of the same receptor superfamily as ORs, GRs are primarily expressed in gustatory sensory neurons (GSNs) of gustatory organs [24].The IRs are members of the iGluR-related olfactory receptor family that was recently identified in Drosophila melanogaster, and are expressed in distinct olfactory neurons that do not express ORs or GRs, suggesting that they function through molecular mechanisms that differ from those of ORs and GRs.Previous studies have shown that IRs are expressed in coeloconic OSNs, and may function in the detection of acids and ammonia [18,25,26].
Transcriptomes based on next-generation sequencing data have been used to identify chemosensory genes in species for which a complete genomic sequence databases have been unavailable.Compared to conventional homology cloning methods, which are difficult to apply to ORs because of the low levels (20%-40%) of sequence identity shared among them, these high throughput transcriptomic approaches are more efficient for large-scale gene discovery, especially with regard to the identification of highly diverse gene families [27].Transcriptomes assembled from high-throughput sequencing data have been used to identify protein families involved in chemoreception in many lepidopterans [19,[28][29][30][31], whereas such proteins have been identified in only four species of the order Coleoptera [32][33][34].
The scarab beetle, Anomala corpulenta Motschulsky, is a major insect pest in agriculture and horticulture throughout the Asian continent, including China [35].The adults destroy the leaves of a broad range of agriculturally important host plants and the leaves and buds of various trees.The larvae of A. corpulenta damage the subterranean parts of peanuts, wheat, soybeans, and other crops, causing significant economic losses.Because of their ecological and economic importance, researchers have investigated various approaches to control A. corpulenta.However, the soil-dwelling habits of the larvae have rendered conventional chemical pesticide treatments ineffective.Chemosensory proteins in insect pests represent potential molecular targets for the development of novel population control strategies that interfere with olfaction [36].However, relatively little is known regarding the molecular mechanisms underlying odorant detection in coleopterans.
As no genome sequence information is available for A. corpulenta, we used Illumina sequencing technology to identify olfaction-related genes in A. corpulenta by assembling antennal transcriptomes of adult males and females.We identified many chemosensory genes expressing in the antennae of A. corpulenta.Our report is the first exhaustive characterization of chemosensory multigene families of a beetle in the superfamily Scarabaeidae.

Materials and Methods Insects
Live adult A. corpulenta were collected in June 2012 from fields at the Experimental Station of the Shandong Peanut Research Institute in Qingdao, China (36.75°N, 120.50°E).The insects were separated into males and females, and were reared on fresh leaves of elm tree, Ulmus pumila L. in rotating wire-mesh cages at 26±1°C temperature and 18% to 20% relative humidity in an environmental chamber.

Complementary DNA (cDNA) library construction and Illumina sequencing
The antennae were severed at the base of the pedicel using forceps, and transferred immediately to microcentrifuge tubes in liquid nitrogen.The frozen antennae were stored at −80°C.Total RNA was extracted from about five hundred female or male antennae using the Trizol reagent (Invitrogen, Carlsbad, CA, USA), according to manufacturer's protocol.The quantity and integrity of the RNA were assessed using an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA, USA).First-strand cDNA was synthesized from 10 μg of total RNA.The cDNA libraries were constructed using reagents provided in the Illumina sequencing kit, according to the manufacturer's recommendations, and the libraries were sequenced using an Illumina HiSeq 2000 platform at the Chinese National Human Genome Center (Shanghai, China).The A. corpulenta male and female antennal Illumina sequence data have been submitted to NCBI (accession numbers SRR1503480 and SRR1503444, respectively).

Read assembly and sequence annotation
All sequencing reads were filtered to remove short or low-quality reads and adapter sequences using the TGI Seqclean program (http://compbio.dfci.harvard.edu/tgi/software/).We conducted the de novo assembly using the Trinity, release 25.02.2013, program [37] using the default parameters, which did not rely on mapping the sequencing reads to a reference genome.The contigs obtained were compared with the NCBI nonredundant protein database (version 31.10.2012)using the BLASTX computational tool with a threshold E-value of 1 × 10 -5 .Gene Ontology (GO) annotations were assigned using the Blast2GO program [38] through a search of the NCBI nonredundant database using the BLASTX algorithm.The E-value cut-off was set at 1 × 10 -5 .The transcripts were categorized as cellular component, molecular function, or biological process, allowing for a meta-analyses of the gene populations.
The ORFs of the annotated olfactory gene families of A. corpulenta were identified using the Getorf utility in the EMBOSS, version 6.3.1, program (http://mobyle.pasteur.fr/cgi-bin/portal.py?#forms::getorf) [39], and verified by comparing the predicted sequences to the NCBI nonredundant protein database using the BLASTX computational tool.The cleavage sites for signal peptides were determined using the SignalP, version 4.0, program [40].The transmembrane domains of the candidate ORs, GRs, and IRs were identified using the TMHMM, version 2.0, program [41].
The multiple sequence alignment was performed using the MAFFT program [48].Alignment ambiguities and gap-containing columns were excluded from the phylogenetic analyses.The maximum-likelihood trees were constructed using the Fasttree, version 2.1.7,software [49].The tree images were created using the iTOL web-based program [50].

Reverse transcription and polymerase chain reaction (RT-PCR)
The sequences of all the identified chemosensory-related transcripts were verified using RT-PCR and direct DNA sequencing.Total RNA was extracted from antennae collected from male and female A. corpulenta.Primer pairs (S1 Table ) were designed according to the Illumina sequencing data using the Primer Premier 5 software (Premier Biosoft, Palo Alto, CA, USA).The PCR was performed using the Titanium Taq DNA polymerase (Clontech, Mountain View, CA, USA) and olfactory-gene-specific primer pairs.Thermal cycling was performed at 94°C 5 min, followed by 30 cycles of 94°C for 50 s, 48°C to 56°C for 1 min, and 72°C for 1 min, with a final extension at 72°C for 10 min.Forty cycles of PCR were used to amplify the OR, GR, and IR sequences because of their relative low levels of abundance.The amplicons were analyzed by electrophoresis on 1.5% agarose gel, and visualized using GelRed staining (Biotium, Hayward, CA, USA).The amplicons were purified using the Qiaquick Gel Extraction Kit (Qiagen, Hilden, Germany), and one amplicon for each gene was sequenced by a commercial service provider (Invitrogen, Shanghai, China).

Quantitative real time RT-PCR analysis
The expression of the putative chemosensory genes in different tissues was investigated using quantitative real time RT-PCR (qRT-PCR) experiments.The qRT-PCR was performed on the Bio-RAD CFX connect Real-Time System (Applied Biosystems, USA) using SYBR Green (Invitrogen, USA).Ten micrograms of total RNA sample was used for the synthesis of firststrand cDNA with reverse transcriptase (Takara, Japan) and the cDNA was then used as a template for qRT-PCR.The specific primers for qRT-PCR (S1 Table ) were designed using Primer Premier 5.The reaction programs were 1-30s at 95°C, 40 cycles of 95°C for 10s and 61°C for 30s.The specificity of the PCR reactions was assessed by measuring a 65-95°C melting curve.Expression levels of the genes were calculated relative to the expression of the control gene G3PDH using the 2 -ΔΔCT method [51].All reactions were performed in triplicates.

De novo assembly and gene annotation of chemosensory genes of A. corpulenta
Solexa sequencing yielded a total of 33,384,121 (99.02%) and 25,729,630 (99.09%) clean reads for the female and male transcriptomes, respectively.The clean reads from the female and male samples were assembled into 145,780 (N50 = 2,265) and 120,190 (N50 = 2,169) contigs larger than 200 bp, respectively.The largest contigs obtained from the female and male data sets were 37,531 and 41,309 bp, respectively.The assembled contigs were compared to the NCBI nonredundant protein database using BLASTX with a cut-off E-value of 1 × 10 -5 .In females, the BLASTX searches produced 92,353 hits, which comprised 63.4% of all of the contigs assembled for females.In males, the BLASTX searches produced 83,999 hits, which comprised 69.9% of all of the contigs assembled for males.The highest match percentage (44.4% in female and 44.6% in male) is to T. castaneum sequences (S2 Table ).The annotated contigs that shared > 98% amino-acid sequence identity were presumed to represent the same unigene, and only the transcript with the longest ORF was included in the analysis.Based on these criteria, bioinformatics analyses identified a total of 77 unigenes from A. corpulenta transcriptomes that belonged to gene families putatively involved in insect chemoreception, including 15 OBPs, five CSPs, one SNMP, 43 ORs, eight GRs, and five IRs (Tables 1 and 2), none of which have been reported in A. corpulenta previously.The sequences of all these transcripts assembled were experimentally validated by RT-PCR and Sanger sequencing.The nucleic acid sequences of the cDNAs shared > 99% sequence identity with the corresponding sequences in the respective transcriptomes, indicating that the transcriptome assemblies were accurate.The alignment results showed that all the sequences belonging to the same chemosensory gene family had long overlapping regions without identity, which confirmed that they represented unigenes.The amino acid sequences for all these proteins are provided in S1 File.

Gene ontology annotations
GO annotation indicated that the analyzed antennal transcriptomes of male and female adults were highly similar with respect to GO terms.In males, 16,702 A. corpulenta contigs were assigned to at least one GO term based on BLASTX matches with sequences with previously known function.In females, there were 18,624 such contigs.These contigs were annotated in three main categories: cellular component, biological process and molecular function, and 45 sub-categories.Among the biological process category, cellular and metabolic processes were the two largest groups in both male and female data sets.The molecular function category was mainly comprised of sequences involved in binding and catalytic activities.Under the cellular component category, cell and membrane were the most abundant GO terms (Fig 1).These GO annotations provide a global gene expression profile for male and female antenna of A. corpulenta, which is basically consistent with previously reported antenna of other Lepidoptera and Coleoptera insects [28,31,33].

Perireceptor olfactory gene families
OBPs.We identified a total of 15 different sequences encoding candidate OBPs in A. corpulenta, and numbered them according to the most similar homolog whenever possible (Table 1).Of these transcripts, full-length ORFs encoding 133 to 223 amino acids were identified in 14 of them, and the remaining candidate, AcorOBP2, corresponded to a partial sequence that encoded 96 amino acids.Other unigene data for these OBPs, including length, signal peptide, closest BLASTP matches, and so on, are listed in Table 1.
Sequence analysis indicated that each of the 14 candidate OBPs that contained a complete ORF also encoded a signal peptide sequence (Table 1).The OBP1 of A. corpulenta shared 86% amino-acid sequence identity with the OBP1 of Holotrichia oblita, and the OBP15 of A. corpulenta shared 84% amino-acid sequence identity with the PBP2 of Anomala schonfeldti, whereas all other candidate OBPs of A. corpulenta displayed a low level of sequence similarity with  To confirm the annotation and identify potential orthologs, a maximum-likelihood cluster analysis was performed (Fig 2).The phylogenetic tree had relatively large minus-C OBP and plus-C OBP branches.The OBPs of A. corpulenta were distributed over various branches, with all forming small clusters with the OBPs of coleopteran species.Of the 15 putative OBPs, seven clustered with other OBPs of A. corpulenta, rather than those of other insect species, suggesting that they may be the products of recent gene duplications.
CSPs.Our bioinformatic analysis identified five unigenes encoding proteins that were 108 to 261 amino acids in length, which is consistent with other insect CSPs (Table 1).Each of the putative CSP unigenes of A. corpulenta had an intact ORF with a signal peptide and four conserved cysteine residues.The maximum-likelihood tree showed that AcorCSP3 clustered confidently with TcasCSP4, DponCSP3 and DponCSP5.AcorCSP4 clustered confidently with TcasCSP6 and ItypCSP4.There were weak values supporting the other three clades: SNMPs.In general, SNMPs are classified as SNMP1 or SNMP2.Based on the transcriptome data, only one SNMP candidate was identified in the antennae of A. corpulenta.Phylogenetic analysis along with BLASTP search revealed it was classified as SNMP1 (Table 1

Receptor encoding genes families
ORs. Forty-three unigenes encoding putative ORs were identified based on the antennal transcriptome data for A. corpulenta.Transcripts for all the candidate ORs were present in the transcriptomes of both sexes.Among them, 28 contained a full-length ORF that encoded 372 to 448 amino acids.Fifteen partial sequences encoded 184-408 amino acids.One OR sequence was found that shared a high level of identity with the conserved Orco proteins of other insect species and named as AcorOrco.The amino-acid sequence of AcorOrco shared 90% identity with Or83b of Holotrichia parallela (AEG88961.1)and 81% identity with the Orco of T. castaneum (XP_008194693.1).The other ORs of A. corpulenta were highly divergent, and shared low levels of similarity with other insect ORs.Between two and eight transmembrane domains were identified in the ORs of A. corpulenta (Table 2), based on comparisons with other insect ORs [19].Thirteen (46.4%) of the 28 ORs with complete ORF of A. corpulenta contained seven transmembrane domains, which is characteristic of the OR protein family.The topology analysis indicated that the N-terminus of 21 of the 28 full-length ORs of A. corpulenta was on the cytoplasmic side of the membrane.This is the typical membrane topology of insect ORs and is inverted relative to vertebrate ORs [21].
A maximum-likelihood phylogenetic analysis was performed using a data set containing the sequences of the ORs that were longer than 200 amino acids from A. corpulenta and other four described coleopteran species [32,33,45] (Fig 4).The OR sequences were clustered into several subgroups which were numbered from 1 to 7 according to previous studies [32,33,45].Except for AcorOrco, A. corpulenta ORs were only present within previously defined coleopteran OR subgroup1, 2, and 3.The majority of the ORs of A. corpulenta clustered into species-specific expansions within receptor subgroups 2 and 3.As previously reported, receptor subgroups 4, 5, and 6 contained ORs from T. castaneum, and subgroup 7 contained ORs from M. caryae only [32,33,45], with one exception, ItypOR14, which was clustered within subgroup 5 but showed a weak supporting value.The orthologs to the functionally characterized ORs of M. caryae were identified in A. corpulenta.GRs.Bioinformatic analysis identified eight putative GR unigenes.Among them, a complete ORF was identified in AcorGR2, AcorGR3, AcorGR4, and AcorGR5, and the remaining GRs were partial sequences that encoded more than 103 amino acids.Based on the phylogenetic analysis, several potential orthologs for other known insect GRs were identified.AcorGR2 represented an ortholog with TcasGR155.AcorGR5 clustered confidently with ItypGR6.Besides, two A. corpulenta GRs clustered within the highly conserved lineage of the DmelGR21a and DmelGR63a.AcorGR3 represented an ortholog to TcasGR1, DponGR1 and DmelGR63a.AcorGR4 was the orthologous receptor of TcasGR3, ItypGR3, DponGR3 and DmelGR63a (S4 Fig) .IRs.We identified five putative IR unigenes in A. corpulenta, in three of which a complete ORF was identified.Insect IRs typically contain three transmembrane domains [25,26].We identified three transmembrane domains in four of the IRs of A. corpulenta (Table 2).According to the maximum-likelihood tree of the IRs from D. melanogaster and four coleopterans, we found each member of IR21a, IR41a and IR75q lineage in A. corpulenta (Fig 5).However, we didn't find the ortholog for co-receptor IR8a or IR25a.One A. corpulenta IR located at the base of a branch formed by 14 homologs belonging to IR75 group but its exact orthologues were not identified (Fig 5).Thus it was named AcorIR75x.Besides, one A. corpulenta IR which clustered with an IR (XP_974901.2) from T. castaneum was named AcorIRx (Fig 5).

Quantitative real time RT-PCR analysis
We investigated the expression patterns of 54 putative A. corpulenta chemosensory genes by performing a qRT-PCR analysis of the RNA extracted from adult male and female antennae, male and female legs as well as non-chemosensory tissues (mixture of male thorax and abdomen which was referred to as body tissues here after) (Fig 6).Among the 13 tested OBPs, Acor-OBP3, AcorOBP4, AcorOBP7, AcorOBP8 and AcorOBP12 displayed antenna-specific expression without sex bias except AcorOBP4, which was expressed at a higher level in male than in female.The remaining OBPs were expressed in the legs and non-chemosensory tissues as well as in the antennae.Of these widely distributed genes, AcorOBP1, AcorOBP6 and Acor-OBP15 were expressed at the highest level in male antennae compared with other body parts.The analysis on the expression of all the five A. corpulenta CSPs was performed.AcorCSP3 displayed a clear antennae-specific expression with approximately equivalent levels in both sexes.The remaining four CSPs were found to be expressed not only in antennae but also in legs and no-chemosensory tissues.The SNMP1 of A. corpulenta was expressed only in the antennae with no significant difference observed in the level between males and females.
The majority of the tested ORs were clearly antenna-specific.AcorOR10 and AcorOR15 were expressed in legs at lower levels relative to in the antennae.AcorOrco was only expressed in the antennae with no sex bias.Although we didn't find apparent sex-specific genes in these A. corpulenta olfactory receptors, AcorOR7, AcorOR9, AcorOR12, AcorOR13, AcorOR15 and AcorOR25 were more highly expressed in male antennae than in female antennae, wheras AcorOR5, AcorOR29 and AcorOR30 were expressed at higher levels in male antennae than in female antennae.All the four tested GRs were not only expressed in the antennae but also in the legs and the body tissues.AcorGR4 and AcorGR6 were more highly expressed in female antennae than in male antenna.Of the four tested IRs, AcorIR21a, AcorIR75q and AcorIR75x displayed an antenna-specific expression, among which AcorIR21a and AcorIR75q were expressed at a higher level in female than in male.AcorIRx was expressed more highly in the legs compared with in the antennae and body tissues.

Discussion
The molecular basis of chemoreception in coleopterans is relatively poorly understood compared to dipterans and lepidopterans.Major chemosensory gene families have been identified in the red flour beetle, T. castaneum, and three species of bark beetles [32][33][34].We generated antennal transcriptomes of male and female A. corpulenta using Illumina next-generation sequencing technology.We identified 77 chemosensory genes encoding members of six protein  families, including OBPs, CSPs, SNMPs, ORs, GRs, and IRs.However, transcripts present at low abundance and those that are too divergent to be identified using a BLAST search may be overlooked in transcriptome analysis.Thus, it is unlikely that the identified genes represent the total number of the related chemosensory genes in A.corpulenta.This is the first comprehensive characterization of olfactory genes in the scarab beetle.Our findings provide insight into the molecular mechanisms of olfaction in A. corpulenta, and identify potential molecular targets for controlling this pest.
A total of 15 OBPs were identified in the antennal transcriptome of A. corpulenta.This number is equal to that of OBPs previously identified in the antennae of I. typographus.By contrast, more OBPs were identified in T. castaneum based on a genomic analysis, and more were identified in D. ponderosae based on the transcriptome analysis of various tissues, including the antenna, from different life stages (S3 Table ).Our analysis of the antennal transcriptome may not have identified OBPs expressed primarily or exclusively in other tissues or life stages, such as in larve.OBPs may share a common ancestor with CSPs [42].The amino-acid sequence similarity between the OBPs and CSPs of A. corpulenta and those of other insect species indicates that OBPs are less conserved than the CSPs [42].The antennae-specific expression of OBPs has been implicated in pheromone detection, whereas expression of OBPs in other organs has been shown to be related to the other functions apart from chemosensation [36]., Some OBPs exhibited antenna-specific pattern of expression, suggesting they may play important roles in detecting odor molecules.Future functional investigation of these genes is warranted to determine their specific roles in chemical communication.
CSPs constitute another class of small binding proteins.Although most of the genomes of Drosophila species contain only four CSP genes, T. castaneum, which is more closely related to A. corpulenta, has 20 CSPs [42].We identified a modest number of CSPs in A. corpulenta, presumably because the expression of many CSP genes is limited to nonsensory tissues or other developmental stages [53].Among the five predicted CSPs of A. corpulenta, four were detected in both the antennae and non-chemosensory tissues, which is consistent with the expression patterns of CSPs in other insects [54,55].Thus, in addition to chemosensation function, CSPs in A.corpulenta may be involved in other functions such as the regeneration of legs [15] and female survival and reproduction [56].AcorCSP3 displayed apparent antenna-specific expression suggesting that it may play an important role in semiochemical recognition.
SNMPs can generally be categorized as SNMP1 or SNMP2, according to sequence similarity and intron locations [44].Although the functions of SNMPs remain largely unclear, SNMPs have been shown to be essential for pheromone responses [16].We identified only one SNMP in the antennal transcriptome of A. corpulenta, which we classified as an SNMP1, and found that it clearly displayed antenna-specific expression in both sexes, suggesting it may be important in chemoreception.The SNMPs are conserved throughout holometabolous insects [57].In certain moths, a high level of antenna-specific expression has been observed for SNMP1, whereas SNMP2 is expressed at various levels in a variety of tissues [57,58].However, studies of SNMPs in coleopterans are scant.We speculate that the tissue specificity of AcorSNMP2 expression may differ from that reported in other coleopterans [33].In addition, an SNMP2 ortholog was not found in an analysis of another scarab chafer, H. parallela (unpublished).Therefore, we speculate that the level of SNMP2 expression in the antenna of A. corpulenta may be low, and thus undetectable based on the transcriptome analysis.It is also possible that members of Scarabaeidae may have lost SNMP2 during evolution.
ORs of coleopterans from the superfamilies Tenebrionoidea, Chrysomeloidea, and Curculionidae have been described, and the findings of our study extend the knowledge base of coleopteran ORs to include the superfamily Scarabaeidae.In insects, ORs have undergone frequent gene duplications and deletion events, representing the "birth-to-death" model of evolution, which has contributed to a high level of diversity in OR genes and variability in gene number between different species.Sixty-three ORs have been identified in D. melanogaster [46], and hundreds of ORs have been identified in ants [59].Relatively few ORs were identified in the antennal transcriptomes of A. corpulenta, compared with the number of ORs in the complete genome of T. castaneum, but comparable to the number of 57 ORs identified in the antennal transcriptome of M. caryae.Maybe lower expressed ORs were not detected.However, similar to observations in T. castaneum and M. caryae, a species-specific expansion of ORs was observed in A. corpulenta.These species-specific expansions may reflect the different ecological niches of these species.ORs play a key role in odor specification.We may speculate on the possible functions of ORs in A. corpulenta through those of orthologous ORs in M. caryae, which was the first beetle in which the function of ORs was characterized.However, the effects of the rapid evolution of OR families might be inconsistent with such speculation.The AcorOR2, AcorOR12, and AcorOR13 clustered into a clade that was a sister group of McarOR3/Tca-sOR65.McarOR3 binds (S)-2-methyl-1-butanol and structurally related chemicals.Whether AcorOR2, AcorOR12, and AcorOR13 have functions that are similar to those of McarOR3 warrants further investigation [32].The expression profiling indicated that A. corpulenta ORs were highly restricted in the antennae relative to OBPs and CSPs.We also found several ORs expressed in legs and no-chemosensory tissues, similar to the findings in Sesamia inferens [60].Unlike T. castaneum olfactory co-receptor which was also expressed in legs and other tissues [61], AcorOrco showed a clear antennae-specific expression pattern.
A large number of GRs have also been identified in various insect species [24,46,62].We identified a small number of GRs in A. corpulenta, which was expected because GRs are primarily expressed in gustatory organs, such as the proboscis, rather than the antennae [24].A previous study showed that GR21a and GR63a of D. melanogaster form heterodimers, and are involved in CO 2 detection [63].Both orthologs of these two receptors were found in the antenna transcriptomes of A. corpulenta, which suggests that A. corpulenta might use CO 2 detection as a criterion for host plant selection.In contrast to AcorORs, AcorGRs were widely expressed in different tissues with relatively high leg expression, similar to that reported in T. castaneum [61].
Compared with the ORs and GRs, IRs are more highly conserved in insects, and the number of IRs is generally consistent [25].IRs can be categorized as divergent species-specific IRs and conserved antennal IRs, the latter of which are expressed in the sensory dendrites of antennae [18,25].For example, of the 66 IRs identified in drosophilids, 15 were antennal IRs [18].The orthologs of IR21a, IR41a, and IR75q of T. castaneum were identified in A. corpulenta.Acor-IR21a, AcorIR75q and AcorIR75x displayed antenna-specific expression.Our findings suggest that these IRs may function in olfaction in A. corpulenta.
This study not only described the putative A.corpulenta chemosensory genes identified from the antennal transcriptome, but also investigated the expression profile of these genes.The genes that are potentially involved in A. corpulenta chemical communication may represent novel molecular targets for controlling this beetle by regulating chemoreception.

Fig 1 .
Fig 1. Gene ontology annotations of transcripts isolated from the antennae of A. corpulenta.The X-axis shows that transcripts were assigned to the cellular component, biological process, or molecular function categories and further to 45 subcategories at GO level 2 using the Blast2GO program.The Yaxis shows percentage of transcripts.doi:10.1371/journal.pone.0121504.g001

Fig 2 .
Fig 2. Maximum-likelihood phylogenetic tree of the putative OBPs of A. corpulenta and OBPs from D. melanogaster, T. castaneum, I. typographus and D. ponderosae.The tree was constructed using the FastTree 2.1.7program.The numbers above the nodes indicate support values.Support values < 0.9 are not shown.The plus-C subfamily is marked in light yellow, and the minus-C subfamily is marked in light green.doi:10.1371/journal.pone.0121504.g002

Fig 3 .
Fig 3. Maximum-likelihood phylogenetic tree of the putative CSPs of A. corpulenta and CSPs from T. castaneum, I. typographus and D. ponderosae.The tree was constructed using the FastTree 2.1.7program.The numbers above the nodes indicate support values.Support values < 0.9 are not shown.doi:10.1371/journal.pone.0121504.g003 McarOR3 and TcasOR65 formed a strong lineage with AcorOR2, AcorOR12, and AcorOR13.AcorOR34 was clustered confidently with DponOR5 and five M. caryae olfactory receptors including McarOR5.However, McarOR20 represented no apparent orthologs to any ORs of A. corpulenta.

Fig 5 .
Fig 5. Maximum-likelihood tree of the putative IRs of A. corpulenta and IRs of D. melanogaster, T. castaneum, I. typographus and D. ponderosae.The tree was constructed using the FastTree 2.1.7program.The numbers above the nodes indicate support values.Support values < 0.9 are not shown.doi:10.1371/journal.pone.0121504.g005

Fig 6 .
Fig 6.Relative expression levels of the putative chemosensory transcripts of A. corpulenta using qRT-PCR.The X-axis shows the different body parts.The Y-axis shows the relative mRNA level.A♀, female antennae; A♂, male antennae; L♀, female legs; L♂, male, legs; B, body tissues (mixture of male thorax and abdomen).

Table 2 .
BLASTP matches for the putative ORs, GRs, and IRs of A. corpulenta.