Isolation, Expression Profiling, and Regulation via Host Allelochemicals of 16 Glutathione S-Transferases in the Chinese White Pine Beetle, Dendroctonus armandi

The Chinese white pine beetle (Dendroctonus armandi) is undoubtedly one of the most important pests causing ecological damage in the Qinling Mountains. When bark beetles invade conifers, they must overcome host tree defenses, including primary resistance and induced resistance responses. Moreover, this induced resistance occurs following herbivory by bark beetles. Bark beetles have a corresponding defense mechanism for degrading toxic compounds, and glutathione S-transferases (GSTs) can catalyze the binding of endogenous substances that reduce glutathione (GSH) to various harmful electrophilic substrates, increasing their solubility and facilitating their excretion from cells. In this experiment, we successfully obtained sixteen full-length sequences of D. armandi, which belonged to four GST categories (delta, epsilon, sigma, and theta). The transcript levels of sixteen GSTs in D. armandi were compared at four developmental stages (larvae, pupae, teneral adults, and adults), three different tissues (antennae, gut, and reproductive organs), and under various levels of terpenoid stimuli during feeding on phloem tissue to evaluate the various relevant modes of action. This study aids in the understanding of the interaction between monoterpenes and beetles, and beetles’ detoxification through GSTs.

The Chinese white pine beetle, Dendroctonus armandi (Coleoptera: Curculionidae; Scolytinae) is one of the most harmful pests in the Qinling Mountain ecosystem (Yin et al., 1984;Chen and Yuan, 2000). Various native insects, such as D. armandi, can undergo population increases that reach epidemic levels, leading to widespread tree death in the forest ecosystem south of the Qinling Mountains. D. armandi has mainly damaged healthy Chinese white pine (Pinus armandi) over the past 30 years or more (Chen and Tang, 2007). The resin contains volatile monoterpenes and sesquiterpenes include α-pinene, β-pinene, β-caryophyllene, and β-myrcene, which, produced by the damaged tree, are harmful for beetles (Reid and Purcell, 2011;Dai et al., 2015). Moreover, the terpenoid substances can also be used by beetles for the colonization and selection of hosts (Delorme and Lieutier, 1990;Raffa and Smalley, 1995;Klepzig et al., 1996). Furthermore, monoterpenes are indispensable substances in pheromone precursor synthesis for beetles (Lindström et al., 1989;Seybold, 1993;Byers, 1995) or serve as the main attractant of other herbivores and even fungal symbionts (DiGuistini et al., 2011).
The beetles must first survive exposure to the toxic substances and then convert these terpenoids into other substances. Bark beetles have a corresponding defense mechanism for degrading toxic compounds, composed primarily of cytochromes P450 (CYP), esterases, and glutathione S-transferases (GSTs), which are the three major multi-gene family enzymes involved in this defense (López et al., 2013).
Various insect life stages are constantly exposed to exogenous and endogenous toxic substances such as pesticides and plant defense allelochemicals, which are needed to degrade toxins via GSTs in insects (Ketterman et al., 2011). The cytoplasmic GSTs in insects are mainly divided into six categories: delta, epsilon, omega, sigma, theta, and zeta. And GSTs in epsilon and delta are insect-specific which have the largest number of Coleoptera species like Tribolium castaneum (Shi et al., 2012). Further study also shows that delta and epsilon class GSTs are specifically involved in insecticide resistance in Meligethes aeneus to the most commonly used pyrethroids (Erban and Stara, 2014).
GSTs function mainly in the cell, inform the active electrophilic group attack, through the catalytic reduction of GSH, and participate in the detoxification process along with other detoxification enzymes (Sheehan et al., 2001). The GST enzymes activity and GSH content have been studied for a long time, which were considered as an important element in the insecticide resistance of beetles (Reidy et al., 1990). But their functions in the detoxification of the host chemical defense were concerning. Previous study has reported that GSTs have been cloned from the spruce budworm Choristoneura fumiferana, which mainly feeds on conifer tissues (Feng et al., 2001). Moreover, studies on the GST gene of mountain pine beetles Dendroctonus ponderosae have included the functional annotation of the expressed sequence tags and the resulting genome (Keeling et al., 2012. Robert et al. (2013) performed a transcriptome sequencing analysis (RNA-seq) on males and females after starvation or feeding for 24 h and showed a significant increase in some GST transcripts after feeding on lodgepole pine Pinus contorta tree tissue. And the insect physiology conditions like life-stage and sex should been considered in studying GST enzymes activity under toxic stress (Wilczek et al., 2003).
In this study, we isolated sixteen GST genes of D. armandi, and detected their expressions at different life history stages, and in different organs and tissues. The main goal of this study was to determine the response of GSTs in Chinese white pine beetles when exposed to terpenoids or feeding on host phloem.

Collection of Insect Samples
Samples were collected from infested P. armandi at the southern slope of the Qinling Mountain, Shaanxi Province, China (33 • 18 -33 • 28 N, 108 • 21 -108 • 29 E). At the end of April 2018, 1.0-1.5 m logs were cut from infested trees, transported to the laboratory, and placed in a pot with moist sand. The outer layer was covered with gauze. Hundreds of larvae, pupae, and adults were collected from the logs. The adults were divided into three substages: teneral adults (brightly colored body), emerged adults, and feeding adults (those who invaded a new host) according to their different behaviors (Dai et al., 2014). Only the latter two sub-stages: emerged and feeding adults were separated by the sex according to the back view of the abdomen (Lyon, 1958). Then, beetles were placed in a plastic container filled with wet paper and transported at 4 • C.

Isolation of GST Genes
Total RNA was isolated from 9 larvae, 6 pupae, 3 teneral adults (indicated by a brightly colored body), 6 emerged adults (1:1 ratio), and 6 feeding adults (1:1 ratio) (those feeding on new conifers) following the manufacturer's instructions of the UNIQ-10 Column TRIzol Total RNA Isolation Kit (Sangon Biotech, Shanghai, China). The obtained RNA was checked for integrity using 1% agarose gel electrophoresis, and its quality was examined with a NanoDrop 2000 (Thermo Scientific, Pittsburgh, Pennsylvania, United States). The cDNA template strand was synthesized with the FastKing RT Kit (with gDNase) (Tiangen, Beijing, China) according to the instruction manual.
Specific primers (Supplementary Table S1A) were designed for each gene of interest using the Primer Premier 5.0 software based on an existing transcriptome database (Dai et al., 2015). PCR amplification of the cDNA was performed with a C1000 thermocycler (Bio-Rad Laboratories, Hercules, CA, United States). The reaction system was 20 µL in volume, including 1 µL of cDNA, 0.5 µL of each primer (10 µM), and 10 µL of 2 × EasyTaq PCR SuperMix (TransGen Biotech, Beijing, China). The reaction conditions were 30 s at 94 • C, 30 cycles of 30 s at 94 • C, 30 s at 55-62 • C, 1 min at 72 • C, and a last extension step of 10 min at 72 • C.
The PCR products were aligned with a 2K Plus DNA Marker (TransGen Biotech) in 1% agarose gel electrophoresis, where the agarose gel was stained with 1 µL GelStain 10,000× (TransGen Biotech). The resulting PCR product was further purified using a Gel Purification Kit (Omega, GA, United States), and the resulting DNA was ligated into the pMD TM 18-T Vector (TaKaRa Bio Inc., Dalian, China) and then transformed into DH5α chemically competent cells of Escherichia coli (TransGen Biotech). After the bacterial solution was applied to the plate and cultured overnight in an incubator, white colonies were transferred to Amp/X-Gal/IPTG/LB plates, and then three replicates per plate were sent to Sangon Biotech (Shanghai, China) for sequencing. Partial fragments of the sequence were returned by sequencing, and the partial sequences were compared and analyzed with the National Center Biotechnology Information database 1 by manual editing with DNAMAN (v.7.0).
The partial sequences identified above were used for the primer design to amplify the complete sequences of the GST genes with a SMARTer RACE cDNA Amplification Kit (Clontech Laboratories Inc., Mountain, CA, United States). The amplicons were purified, cloned, and sequenced as described previously. To avoid chimera sequences, specific primers (Supplementary Table S1C) were designed based on the obtained complete sequences to amplify the complete DNA of each GST gene. Amplification reactions were carried out in 20 µL volumes containing: 1 µL cDNA, 0.25 µM of each primer, and 10 µL 2 × EasyTaq PCR SuperMix (TransGen Biotech, Beijing, China). The PCR reaction was performed as follows: 94 • C for 5 min; 30 cycles of 94 • C for 30 s, Tm of each pair of primers for 30 s and 72 • C for 1 min; and a final extension for 10 min at 72 • C. PCR products were visualized on 1% agarose gels, purified, cloned, and sequenced for both strands as described above.
The full-length sequence of the GST genes were translated into an amino acid using the Open Reading Frame Finder 2 , searched against the NCBI database using BLASTp (Altschul et al., 1990).

Full-Length Sequence Analysis of GST Genes
The DNA sequences and the translated amino acid sequences were searched against the NCBI nr database using blastx and analyzed in the GenBank database on the NCBI website 1 . The amino acid sequence was BLAST-searched with the existing insect GST sequences in the GenBank database, and the GST gene sequences that were fully verified were uploaded to GenBank (accession numbers: MK796851-MK796867). 1 https://www.ncbi.nlm.nih.gov 2 https://www.ncbi.nlm.nih.gov/orffinder A total of 44 GST sequences of other insects were selected for the phylogenetic analysis according to the BLASTp searching used by deduced amino acid sequences of the obtained DaGSTs. The phylogenetic tree of DaGSTs was constructed with the maximum likelihood method using the Mega 5.0 (Tamura et al., 2013) software. The Whelan and Goldman (WAG) model was supported by best fitting model testing in Mega 5.0 (−ln L = 5008.27) with a gamma parameter value G = 2.70. The bootstrap value was calculated after 500 pseudoreplicates to estimate the support for each node.
The GST amino acid sequences were subjected to multiple sequence alignment analysis with CLUSTALX v.2.0 (Thompson et al., 1997), using the default software settings. The secondary structure templates of the GST proteins were obtained from the Brookhaven Protein Data Bank website 3 , and crystal structures (Delta and Epsilon: 4YH2; Sigma: 5H5L; Theta: 5ZFG) were used for alignment and analysis. The results of the alignment were visualized with ESPript (Robert and Gouet, 2014), and the secondary structure was assigned to the corresponding sequence.
The molecular weight (kDa) and isoelectric point of the obtained GST gene sequences were analyzed using ProtParam (Gasteiger et al., 2005). The resulting GST amino acid sequence was then subcellularly localized using the TargetP 4 (Emanuelsson et al., 2000) website, and its function was predicated with the default parameters.

Development Stages
Samples of five developmental stages of D. armandi were collected: larvae, pupae, and adult which divided into three substages: teneral adults (brightly colored body), emerged adults, and feeding adults (those who invaded a new host). Only emerged and feeding adults were separated by sex.

Feeding Treatments
A total of 180 emerged adults (90 each ♀/ ♂) were selected and separated according to sex with 30 individuals of each sex placed into one group. One group was placed in Petri dishes (90 × 15 mm) containing new filter paper and stored in a dark environment for 2 h as a control. The other two beetle groups were placed in 2 mL centrifuge tubes contains 2 × 2 cm sections of fresh host phloem one by one, and placed in a dark environment for 12 and 24 h. After feeding treatment, viability was observed under a dissecting microscope. As no beetles died after feeding treatment, for every group, 12 beetles of each sex were randomly selected for real-time qPCR detection.

Terpenoid Treatment
Thirty beetles of different stages: larvae, pupae, teneral, and emerged adults (only emerged adults divided by sex) were placed in a 90 mL Petri dish (90 × 15 mm) containing moist filter paper, respectively. A dry filter paper was placed inside the top of the dish. The terpenoids used in this experiment were identical to those in the fumigant toxicity tests previously performed by the team (Dai et al., 2014). The five terpenoids used were as follows: (−)-α-pinene (98%), (−)-β-pinene (99%), (+)-3-carene (90%), (±)-limonene (95%), and turpentine oil (Aladdin Industrial Corporation, Shanghai, China). The terpenoids were added to the top filter paper at a concentration of 1,000 ppm, and the Petri dish was immediately sealed with Parafilm. Each terpenoid treated three Petri dishes of each stage. The treated dishes were placed in an incubator in a dark environment at room temperature for 8 h and 24 h. Insect viability in each dish was examined under a dissecting microscope after fumigation with the terpenoids. With 20% mortality, 12 living beetles were selected for real-time qPCR detection. Untreated beetles of each stage were used as controls.
Furthermore, other groups with emerged adults were treated with terpenoids at 1,000 ppm for 8 h, performed as above. And antennal, gut, and reproductive tissues of emerged adult beetles were dissected, and used for real-time qPCR detection.

Real-Time qPCR
RNA extraction and first-strand cDNA synthesis were performed according to the cloning of GST genes method. Specific primers for each sequence were designed using Primer Premier 5.0 based on the full-length gene sequences obtained. The RT-PCR system had a total reaction volume of 20 µL, including 0.4 µL per primer (10 µmoL·L −1 ), 1 µg of cDNA template, and 10 µL of Roche SYBR Green mix (Roche Diagnostics Mannheim, Germany). All reactions were performed with a CFX96 TM Real-Time PCR Detection System (Bio-Rad). The reaction conditions were as follows: 95 • C for 30 s, 95 • C for 5 s for 40 cycles, Tm (melting temperature) for each pair of primers (Supplementary Table S1D) for 15 s, and 72 • C for 20 s. Three biological replicates plus three technical replicates were used.
The amplification efficiency and effectiveness of the RT-qPCR for different genes were analyzed by linear regression between the dilutions of the different concentration gradients (1.0, 10 −1 , 10 −2 , 10 −3 , 10 −4 , 10 −5 , 10 −6 , 10 −7 , 10 −8 ) of the cDNA for quantification. The processed cDNA template was diluted into a concentration gradient, and 5 µL of each RT-qPCR template from the previous concentration gradient dilution was added. PCR was performed three times for each gene, and its efficiency was estimated as follows: efficiency = (10 −1/slope −1) × 100, where the efficiency value was 100% ± 10% (Supplementary Table S1D). The estimated PCR verification was verified by R 2 -values, which were > 0.90. In addition, it should be noted that amplification primers specificity was to be determined by melting curves.
The transcriptional expression levels of D. armandi GSTs were normalized to the transcript levels of β-actin (KJ507199) and CYP4G55 (KR012821), as these two genes have stable expression levels in different stages, tissues, and feeding or terpenoids treatment (Dai et al., 2019). By calculating the geometric mean of these two genes, the accuracy of the relative expression level results can be improved. The 2 − Ct method was used to determine the relative expression of D. armandi GSTs (Schmittgen and Livak, 2008). To evaluate the significant difference between different treatments, all of the 2 − Ct values were log 2 transformed.
One-way ANOVA analyzed the transcriptional expression levels of D. armandi GSTs in different development stages and different time points of feeding/fumigation treatments of each stage, sex, and stimulus with a Tukey test. And the difference between the sexes of adult beetles was performed with independent-samples T-tests. The expression profiles of the GSTs in different tissues of each sex were analyzed with two-way ANOVA including interaction. Statistical analysis was performed using SPSS 18.0 (IBM SPSS Statistics, Chicago, IL, United States), and the results were plotted with GraphPad Prism 6.0 (GraphPad Software, CA, United States).

Sequence Analysis of Full-Length D. armandi GST Genes
A total of 16 full-length GST codon sequences were obtained from D. armandi which belonged to the delta, epsilon, sigma, and theta families. They were named as DaGSTs according to their homology with the GSTs of other insects (Table 1 and Figure 1) According to the location of the cells, the GSTs we cloned belonged to the cytosolic class ( Table 1).
The DaGSTs gene of D. armandi and its GSTs have high homology with those of other insects according to multiple sequence alignment analysis. Multiple comparisons of DaGSTs with other insect GST amino acid sequences also showed that the N-terminal domain and the C-terminal domain in the secondary structure are β −α −β −α −β −β −α and 5-6 α-helices, respectively (Figure 2). The DaGSTs in D. armandi had high similarity with four types of amino acids of other insects ( Table 2). The maximum likelihood phylogenetic tree showed that most of GSTs from D. armandi were gathered with GSTs from beetles in Curculionidae like Lissorhoptrus oryxophilus and Sitophilus oryzae (Figure 1). But some GSTs in the sigma family from the Chinese white pine beetle had low homology with known GSTs in the NCBI databases (Figure 1). Based on comparison analysis, DaGSTs have greater than 34% identity with GSTs in other insects ( Table 2), which also agrees with the viewpoint of Chelvanayagam's proposed GST classification standard (Chelvanayagam et al., 2001).

Different Developmental Stages
Significant differences in the GST expression levels among the different developmental stages were identified with one-way ANOVA (Supplementary Table S2). The transcription levels of most of the GST genes had significant differences among different stages (Figure 3). Between sexes, only DaGSTd2, DaGSTe8, DaGSTe10, and DaGSTs3 had significant differences in emerged adults (Figure 3). Moreover, in feeding adults, the expression levels of all GSTs significantly differed between sexes (Figure 3).
In comparison with the expression levels in the larval stage, many DaGSTs were upregulated in the pupal stage, and most of them were upregulated in teneral adults (Figure 3). All DaGSTs were upregulated in feeding females, but DaGSTd1, DaGSTd2, and DaGSTs4, were downregulated in males (Figure 3).

Feeding Treatment
Significant differences were found in the expression levels of all GST genes after feeding on host phloem for 12 and 24 h in FIGURE 1 | Phylogenetic analysis of sequences of GST genes from Dendroctonus armandi was performed using the amino acid substitution model WAG+ G (-lnL = 5008.27, G = 2.7011). The matched genes are listed by the accession numbers of the corresponding sequence in GenBank. Bootstrap values after 500 pseudoreplicates are shown as nodes. The evolutionary branch is divided into four parts: epsilon is orange, theta is green, delta is blue, and sigma is red.
females, but only DaGSTt2 had obvious downregulation in males (Figure 4 and Supplementary Table S3).
Females over expressed all GSTs in comparison to starvation adults after feeding on host phloem. But the transcription levels of DaGSTe8 and DaGSTt3 were higher at 12 h after feeding on host phloem than 24 h (Figure 4 and Supplementary Table S3).

Terpenoid Treatments
One-way ANOVA of 16 GST gene expression levels were performed after terpenoid treatment for 8 and 24 h in different stages of Dendroctonus armandi with a Tukey test (Supplementary Tables S4, S5).
The response of GSTs in males to terpenoids treatment was unlike females. After exposure with (+)-3-carene almost every Frontiers in Physiology | www.frontiersin.org FIGURE 3 | Quantitative expression of GST genes (mean ± SE, n = 3) in different developmental stages of Dendroctonus armandi. L, larvae; P, pupae; TA, teneral adults; EA, emerged adults; FA, feeding adults, females and males of the two latter stages of Dendroctonus armandi. GST expression was normalized to the transcription levels geometric mean of β-actin and CYP4G55. Different letters indicate significant differences from the different development stages (P < 0.05) by Duncan's test. * Significant differences between sexes of the same stage as measured by independent Student's t-tests ( * P < 0.05, * * P < 0.01, * * * P < 0.001). The 2 − Ct and SE values were log 2 transformed for plotting (DaGSTs expression is upregulated if the − Ct has a positive value, and downregulated for negative values). Statistically significant differences in gene expression are shown in Supplementary Table S2. GST was significantly overexpressed in males at 24 h. But only (−)-α-pinene, (−)-β-pinene, (±)-limonene, and turpentine caused significant downregulation in some GSTs like DaGSTe2, DaGSTe8, DaGSTe11, and DaGSTt2 after an exposure of 8 h ( Figure 5E and Supplementary Table S5).

Different Tissues Without Any Treatment
The significant differences between sexes were found in the expression level of six GSTs (DaGSTe2, DaGSTe8, DaGSTe10, DaGSTs4, DaGSTs6, and DaGSTt2), but significant differences were found in 14 GSTs among different tissues. The significant differences were only found in the expression level of DaGSTe2, DaGSTe8, DaGSTe10, and DaGSTe12 for sex * tissues interaction (Figure 6 and Supplementary Table S6).

Different Tissues After Terpenoid Treatments
After treatment with (−)-α-pinene, more than half of the GSTs were significantly overexpressed in the gut of females and reproduction organ of males (Supplementary Table S7). In contrast, most of them were significantly downregulated in the gut of males and reproduction organ of females after treatment with (+)-3-carene. The treatment with (−)-β-pinene only caused two GST genes (DaGSTe8 and DaGSTe11) to be overexpressed in the antennae of females, and two genes (DaGSTd2 and DaGSTs7) to be downregulated in the gut of males. The (±)-limonene and turpentine induced a few GSTs of epsilon in the antennae, gut, or reproduction organ of females (Table 3).

DISCUSSION
We successfully obtained sixteen cDNA gene sequences belonging to GSTs from the transcriptome of D. armandi. The gene sequences of the obtained GSTs were divided into four families (delta, epsilon, sigma, and theta) by phylogenetic analysis. This is consistent with the GST classification of other insects (Liu et al., 2017). Delta and epsilon are two different GST families in insects, and the GSTs that belong to these two families have the same metabolic function as pesticides (Ketterman et al., 2011). The GSTs in epsilon and delta are insect-specific which FIGURE 4 | Quantitative transcript levels of GST genes (mean ± SE, n = 3) in feeding adults (12 and 24 h) of Dendroctonus armandi. GST expression was normalized to the transcription level geometric mean of β-actin and CYP4G55. Different letters indicate significant differences at P < 0.05 (with Tukey test, lowercase for females and uppercase for males, no letter means no significant difference among times). Statistically significant differences in gene expression are shown in Supplementary Table S3. have large species-specific members in Coleoptera species like Tribolium castaneum (Shi et al., 2012). Further study also shows that delta and epsilon class GSTs are specifically involved in insecticide resistance in Meligethes aeneus to the most commonly used pyrethroids (Erban and Stara, 2014). Thus, the GST genes of these two families are also closely related to insect resistance to pesticides because the expression levels of the GSTs in the delta and epsilon families are upregulated in insecticide-resistant pests (Li et al., 2007). D. armandi has two delta-family GSTs and eight epsilon-family DaGSTs; these DaGSTs may also metabolize pesticides and other harmful compounds. The resin present in conifers is the most important component of the chemical defense against the threat of foreign organisms. The resin secreted by a tree can inhibit beetle invasion, slow the invasion speed, and even repel the beetles with its detection (Erbilgin et al., 2003). Moreover, the monoterpene in the resin has toxic effects on the pest and causes debility or death through the respiratory system (Smith, 1965;López et al., 2011). The dispersal flight of the Dendroctonus pseudotsugae is oriented toward fresh wind-thrown and cut trees when these are present, while the beetle is attracted by oleoresin, particularly its α-pinene, camphene, and limonene fractions. The beetles are found to be repelled by Pseudotsuga menziesii resin and its fraction, including α-pinene and β-pinene, camphene, limonene, terpineol, and geraniol, when tested at close range in the laboratory (Rudinsky, 1966). And the toxicities of monoterpenes to bark beetles like D. ponderosae and D. armandi were performed with fumigation experiments (Reid and Purcell, 2011;Dai et al., 2015).
The experimental design and the theories behind the experiment were similar to those used in studies of detoxification enzymes (P450 enzymes, carboxylesterases, and GSTs) in D. armandi (Dai et al., 2014(Dai et al., , 2015(Dai et al., , 2016. The phloem of damaged conifers excretes volatile monoterpenoids, which can then enter the bodies of bark beetles through different routes. The chemical compounds enter the bark beetle body by penetrating the epidermis, accompanied by the invasion of its respiratory and digestive systems (Prates et al., 1998). Insect physiology conditions like life-stage and sex should be considered in studying GST enzymes activity under toxic stress (Wilczek et al., 2003). The strategy of this experiment is in consideration of not only how beetles intake terpenoids (feeding or stimulation), but also the different development stages of bark beetles. Therefore, the feeding experiments reflect GST gene function in the insect digestive and respiratory system, but the fumigation treatments only show GST gene function in the insect respiratory system. And the expression of GSTs in the different sexes of adult beetles was compared as a different behavior in host invasion.
The female adults had a higher expression of GSTs especially after feeding on host trees than males and other development stages under field conditions. And feeding with phloem of the host also induced more GSTs in females compared with males. These all indicated that GST genes were involved in the host selection of female adults and the multiple chemical compounds during host tree colonization (Zhang et al., 2010). The significantly increased transcript and protein accumulation of the DpGSTs gene with RNA-seq and proteomics after D. ponderosae fed on conifer phloem tissue (Robert et al., 2013;Pitt et al., 2014). Furthermore, when different developmental D. armandi were exposed to terpenoids, the expression levels FIGURE 6 | Quantitative expression of GST genes in different tissues of different sexes in Dendroctonus armandi. Antennae (A), gut (G), and reproductive organs (RO). GST expression was normalized to the transcription level geometric mean of β-actin and CYP4G55. Statistically significant differences in gene expression between sexes, tissues and sex*tissue are indicated with asterisks ( * P < 0.05, * * P < 0.01, * * * P < 0.001). The 2 -Ct and SE values were log 2 transformed for plotting. Statistically significant differences in gene expression are shown in Supplementary Table S6. TABLE 3 | Significant difference of GST genes express trend (mean ± SE, n = 3) in different adult tissues [antennae (A), gut (G), and reproductive organs (RO)] of Dendroctonus armandi after terpenoids treatment for 8 h.
Considering that different GST families have been reported to have different responses to monoterpenes in beetles, epsilon and sigma DaGSTs were differentially regulated in the different feeding and terpenoids treatment (Dai et al., 2016). This time, more differences were seen between sex in the DaGST transcript levels, and the regulation of some DaGSTs showed stimuli specifics with terpenoids treatment. GST gene transcription was more sensitive in females to monoterpenes than that in males, but 3-carene was a special stimulus to males which caused extensive overexpression of GSTs. (+)-3-carene lead to higher male beetle mortality in fumigant toxicity tests than other terpenoids (Dai et al., 2015), and also had a greater impact on the transcription of most esterase genes (Dai et al., 2019). The higher expression levels of GSTs in females had a relationship with the resistance to host chemicals which might be also involved in female oviposition.
As the expression profiles of GSTs in different tissues of Chinese white pine beetle shows some sex or tissue specifics. The overexpression of some GSTs in female antennae after being treated with monoterpenes indicates odor-degrading enzyme (ODE) functions (Dickens et al., 1992;Wojtasek and Leal, 1999;Maïbèche-Coisne et al., 2004). The results indicate that female insects first drill into conifers and their antennae need to metabolize various chemicals released by the host trees to avoid persistent excitation. Exotic pests can also use these chemicals to synthesize hormones or pheromones for population regulation (Byers, 1995). α-pinene has the highest proportion of volatile monoterpenes obtained from the resin of P. armandi (Huo and Chen, 2010). This is why many DaGSTs were overexpressed after exposure to α-pinene in the female gut, as the female gut participates in the detoxification process (Kasai et al., 2000;Tartar et al., 2009) and then transforms α-pinene to corresponding pheromones via intestinal bacteria present in some beetles (Kleinhentz et al., 1999;Seybold et al., 2000). The significant overexpression of many DaGSTs in male reproductive organs after being treated with α-pinene also demonstrates that the beetles convert tree monoterpenes into corresponding substances for population regulations (Zhao et al., 2017).
In conclusion, the DaGST genes belonging to different families participate in different metabolic processes in beetles and have different functional characteristics. It was confirmed that four families of GSTs in D. armandi were involved in the detoxification of terpenoids produced by host trees. The involvement in the metabolism of foreign substances was through different response patterns in both the feeding and stimulation treatments. The gene expression levels of GSTs were overexpressed in female beetles after being fed on host phloem which indicates that females have higher resistance to host chemical defenses because of their behaviors such as building galleries for oviposition. Moreover, in combination with the study of oxidative stress caused by conifer steroids, the current study aids in the understanding of the overcoming strategy of D. armandi for host chemical defense.

DATA AVAILABILITY STATEMENT
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Supplementary Material.

AUTHOR CONTRIBUTIONS
HG, LD, and HC contributed to conceptualization. HG, DF, and YS performed the experiments and investigation. HG analyzed the data and wrote the manuscript. All authors have read and agreed to the published version of the manuscript.