Protective and Detoxifying Enzyme Activity and ABCG Subfamily Gene Expression in Sogatella furcifera Under Insecticide Stress

Sogatella furcifera, an important migratory pest of rice, has substantial detrimental effects on rice production. To clarify the mechanism whereby S. furcifera responds to insecticide stress, we measured the activity of its protective [superoxide dismutase (SOD); peroxidase (POD); catalase (CAT)] and detoxifying [carboxylesterase (CarE); glutathione S-transferase (GST); mixed-function oxidase (MFO)] enzymes and the expression levels of its ATP-binding cassette subfamily G (ABCG) transporter genes in response to sublethal concentrations (LC10 and LC25) of the insecticides thiamethoxam, buprofezin, and abamectin. On the bases of the transcriptome data and the ABCG genes of Laodelphax striatellus, we obtained 14 full-length ABCG sequences for S. furcifera. RT-qPCR results showed that 13, 12, and 9 sfABCG genes were upregulated in the presence of thiamethoxam, buprofezin, and abamectin, respectively, at LC10. Moreover, 13 and 7 sfABCG genes were upregulated following treatment with thiamethoxam and abamectin, respectively, at LC25. Enzyme activity assays showed that although thiamethoxam, buprofezin, and abamectin induced GST, CarE, CAT, POD, and SOD activity, they did so at different concentrations and exposure times. The activity of MFO was generally inhibited with prolonged exposure to the three insecticides, with the inhibitory effect being most significant at 72 h. These results indicate that S. furcifera differs in its response to different types or concentrations of insecticides. Taken together, our results lay the foundations for gaining a deeper understanding of the mechanisms underlying the adaptation of S. furcifera to different types of insecticides, which would be of considerable significance for the development of effective pest management strategies.


INTRODUCTION
Sogatella furcifera, an important pest of rice, causes serious problems in rice production by sucking phloem sap from the rice plant, inflicting damage through oviposition, and transmitting viral diseases (Zhou et al., 2008). Although the use of insecticides has traditionally been an important means of control for this rice pest (Endo and Tsurumachi, 2001;Nizamani et al., 2002), recent research has shown that sublethal concentrations of insecticides can affect the reproduction, development, and chemical susceptibility of insects in such a way that it could potentially result in the resurgence of pests (Zhou et al., 2017).
In general, the detoxification process in insects can be divided into three phases: phase I, phase II (involving metabolizing enzymes), and phase III (involving transporters) (Xu et al., 2005). The main enzymes involved in the phase I and phase II detoxification processes are P450 monooxygenase, glutathione S-transferase (GST), and carboxylesterase (CarE) (Xiao et al., 2018), whereas the ATP-binding cassette (ABC) transporters are the main components of phase III (Ferreira et al., 2014). In this regard, it has previously been observed that when the nymphs of Locusta migratoria were treated with chlorantraniliprole at LC 50 , only the activities of esterase (EST) and GST increased on the first day of treatment, whereas mixed-function oxidase (MFO) activity increased only at 3 days after treatment (Cao et al., 2017). In addition, superoxide dismutase (SOD), peroxidase (POD), and catalase (CAT) are three important protective enzymes in insects that play roles in immunity, preventing free-radical-associated damage, and protecting cells from adverse environmental effects (Dubovskiy et al., 2008;Bi et al., 2010). It has previously been reported that in response to treatment with abamectin at LC 10 and LC 20 for 12 h, the activities of SOD, POD, and CAT in Harmonia axyridis were higher relative to those in the untreated control group, although these activities gradually returned to normal levels as time progressed . Furthermore, in response to treatment with imidacloprid at LC 10 and LC 20 , the SOD and POD activities in Aphidius gifuensis initially appeared to be inhibited but were subsequently stimulated, with the highest activity occurring at 36 h. Moreover, it has been found that with an increase in insecticide concentration, SOD, POD, and CAT activities show a decreasing trend (Zhu et al., 2015).
The ABC transporters comprise a large family of proteins that mediate the transport of inorganic ions, sugars, amino acids, lipids, lipopolysaccharides, peptides, metals, xenobiotics, and chemotherapeutic drugs (Higgins, 1992). In insects, this family can be subdivided into eight major subfamilies (A-H) (Dean et al., 2001). Studies of the ABC transporters in eukaryotes have revealed that they are capable of transporting structurally unrelated compounds (Dassa and Bouige, 2001;Dean et al., 2001), and researchers are thus increasingly focusing on the roles of these proteins in the transport of exogenous substances and in insecticide resistance in insects. Recent studies have shown that the expression of ABC transporters is directly related to the development of insecticide resistance (Silva et al., 2012a;Dermauw and Van Leeuwen, 2014). After treatment of Bactrocera dorsalis with malathion, abamectin, and beta-cypermethrin at an LD 50 concentration, 4, 10, and 14 bdABC genes were significantly upregulated, respectively (Xiao et al., 2018). Quantitative polymerase chain reaction (qPCR) analysis has revealed that eight ABC transporters in the ABCB/C/D/G subfamilies were upregulated in strains of Laodelphax striatellus resistant to chlorpyrifos, deltamethrin, and imidacloprid, compared with those in a susceptible strain (Sun et al., 2017). In Plutella xylostella, RNA sequencing (RNA-seq) analysis showed that ABC transporters from the ABCA/C/G/H/F subfamilies were overexpressed in chlorpyrifos-resistant strains (You et al., 2013). Nevertheless, despite the insights gained from these studies, our current understanding of the role of ABC transporters in insect resistance to insecticides remains limited.
At present, little is known regarding the effects of insecticides on the activities of the detoxifying and protective enzymes and ABC transporters of S. furcifera. Accordingly, in this study, we sought to gain insights into the roles of these enzymes and the sfABCG subfamily genes in the response of S. furcifera to insecticide-induced stress. To this end, we exposed this insect to sublethal concentrations of three insecticides (thiamethoxam, abamectin, and buprofezin) and subsequently monitored the changes in enzyme activity and gene expression levels.

Insect Treatments and Sample Collection
For the insecticide treatments, we used the rice stem dipping method (Zhou et al., 2017). Three 100 third-instar nymphs were transferred to and reared separately in glass tubes (300 mm high × 30 mm diameter) that were open at both ends and contained rice seedlings dipped in a sublethal concentration (LC 10 or LC 25 ) of thiamethoxam, abamectin, or buprofezin. Rice stems treated with distilled water were used as a control. The insects exposed to each treatment were maintained at 25 ± 1 • C and 70 ± 10% relative humidity under a 16:8 h (light:dark) photoperiod in an artificial climate box. After 48 h, 15 surviving insects from each treatment were randomly collected for extraction of RNA for a quantitative reverse-transcription PCR (RT-qPCR) assay. In addition, samples were taken at 6, 12, 24, 48, and 72 h after the treatment to determine the activity of the target enzymes. The LC 10 and LC 25 values (Supplementary  Table S1) of thiamethoxam, abamectin, and buprofezin for S. furcifera were based on previously presented results .

Gene Identification
The RNA-seq transcriptome database of S. furcifera was sequenced and annotated as described previously (Zhou et al., 2018). With the reported ABCG gene of L. striatellus as a reference, Geneious R9 software (Kearse et al., 2012) was used to assemble the transcriptome data to obtain the corresponding sequences for S. furcifera. In addition, each of the putative ABCG sequences was used as a query to search the NCBI protein database 1 to further validate their identity.

Sequence Verification
Specific primers were designed and used to amplify the internal cDNA fragments. PCRs were carried out using Sangon Biotech (Shanghai, China) Taq polymerase, under the following conditions: initial denaturation at 94 • C for 3 min; 30 cycles of denaturation at 94 • C for 30 s, annealing at 55-60 • C for 30 s, and elongation at 72 • C for 1-2 min; with a final elongation at 72 • C for 10 min. Specific primers for amplification of the 3 and 5 ends were designed using Primer Premier 6.0 (Premier Biosoft International, Palo Alto, CA, United States). Using a SMARTer R RACE 5 /3 Kit (Clontech, Mountain View, CA, United States), 3 and 5 rapid amplification of cDNA ends (RACE) were performed. Total RNA was extracted from 10 fifthinstar nymphs according to the instructions of an HP Total RNA Kit (Omega Bio-Tek, Norcross, GA, United States). Synthesis of the first-strand cDNA and PCR amplifications were carried out according to the instructions of a SMARTer R RACE 5 /3 Kit. SeqAmp DNA Polymerase (a SMARTer R RACE 5 /3 Kit component) was used for the RACE PCR, under the following conditions: 25 cycles of 94 • C for 30 s, 60-70 • C (depending on the primer) for 30 s, and 72 • C for 3 min. The overlapping PCR products were purified using an E.Z.N.A R Gel Extraction Kit, cloned into a linearized pRACE vector (a SMARTer R RACE 5 /3 Kit component), and sequenced by Sangon Biotech (Shanghai, China). The RACE sequences were assembled on the basis of the partial cDNA sequences corresponding to each fragment.

Sequence Alignment and Phylogenetic Analysis
Using ORF finder 2 , we identified the open reading frames (ORFs) of the sfABCG genes and determined the amino acid sequences of the encoded proteins. The Pfam program 3 and a search of the NCBI Conserved Domain Database 4 were used to identify the conserved domains (nucleotide-binding and transmembrane domains) of all putative ABCG genes. The ABCG gene sequences were then subjected to phylogenetic analysis, using the neighborjoining method and a bootstrap test with 1,000 replicates in the MEGA program package, v. 6.0 (Tamura et al., 2011).

Gene Expression Analysis
The mRNA levels of the ABC transporter genes under different insecticide treatments were measured by RT-qPCR using FastStart Essential DNA Green Master Mix (Roche, Indianapolis, IN, United States) in a CFX96 TM real-time quantitative PCR system (BioRad, Hercules, CA, United States). Total RNA was extracted as described above and quantified using a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific, Waltham, MA, United States) according to the manufacturer's protocols.
The RNA concentration was adjusted to 0.8 µg/µL with diethyl pyrocarbonate-treated H 2 O, and 0.8 µg of RNA was then reverse transcribed in a 20-µL reaction volume, using the PrimeScript RT Reagent Kit and gDNA Eraser (TaKaRa, Shiga, Japan), with ribosomal protein L9 (GenBank Accession No. KM885285) as an internal control. Specific primer pairs for each gene were designed using Primer Premier 6 (Supplementary Table S2). Each RT-qPCR was conducted in a 20-µL mixture containing 1 µL of sample cDNA, 1 µL of each primer (10 µM), 7 µL of diethyl pyrocarbonate-treated H 2 O, and 10 µL of FastStart Essential DNA Green Master Mix. The qPCR cycling parameters were as follows: 95 • C for 10 min, followed by 40 cycles of 95 • C for 30 s and 60 • C for 30 s. Melting curve generation was performed from 65 to 95 • C. To check the reproducibility of the assay results, the qPCR for each sample was performed using three technical replicates and three biological replicates. The comparative 2 − CT method (Livak and Schmittgen, 2001) was used to calculate the relative quantification.

Enzyme Activity Assay
In this study, we performed the following enzyme activity assays: the nitroblue tetrazolium reduction method for SOD; the guaiacol method for POD; a spectrophotometric method for CAT (based on the ultraviolet absorption of peroxide released from the activity of CAT on hydrogen peroxide); a colorimetric method for GST (based on the GST-catalyzed reaction between glutathione and 1-chloro-2,4-dinitrobenzene); and a colorimetric method for CarE (based on the CarE-catalyzed transformation of 1-naphthyl acetate to naphthyl ester, which then reacts with the Fast Blue RR salt to form an azo dye). These assays were conducted using respective commercial assay kits (Comin Biotechnology, Co., Ltd., Suzhou, China). MFO activity was measured according to the method reported by Qian et al. (2008). To check the reproducibility of the results, the enzyme activity assays for each insecticide treatment were performed using four biological replicates.

Statistical Analyses
All data were analyzed using Bonferroni corrections for multiple comparisons when the variance was homogeneous. When the variance was non-homogeneous, the Wilcoxon signed-rank test was used. In addition, the Kruskal-Wallis test was used to verify the temporal shifts within the effects of the same insecticide. All analyses were performed using SPSS version 22.0 (SPSS, Chicago, IL, United States) and the data are presented as the mean ± standard error (SE) of three or four biological replicates.

Identification and Characterization of ABC Subfamily G Transporter Genes
Using the reported ABCG gene of L. striatellus as a reference, Geneious R9 software was used to assemble the transcriptome data to obtain the corresponding sequences for S. furcifera.
We verified 14 sfABCG genes by RT-qPCR and RACE (Supplementary Table S3). The designation, accession number, length, ORF size, theoretical isoelectric point, and molecular weight of all the sfABCG genes are summarized in Table 1. The ORFs of all gene sequences ranged from 603 to 967 bp. We initially identified the characteristic nucleotide-binding domains of ABC transporters using Pfam. The nucleotide-binding and transmembrane domains of all genes were similar to those of L. striatellus (Supplementary Figure S2). As determined from the neighbor-joining tree generated from phylogenetic analysis of the ABCG genes of S. furcifera, L. striatellus, Tribolium castaneum, and B. dorsalis, the corresponding genes of each subfamily are clustered together (Supplementary Figure S1).

Effect of Insecticide Treatment on sfABCG Gene Expression
After exposing third-instar nymphs of S. furcifera to different concentrations of thiamethoxam for 48 h, we examined the relative expression levels of the 14 sfABCG genes. The results showed that the expression of only sfABCG7 was significantly downregulated (2.4-fold) after treatment with the insecticide at LC 10 (Figure 1), whereas the other 13 sfABCG genes were significantly upregulated. Among these 13 genes, sfABCG5 (766.6-fold) and sfABCG9 (5.8-fold) showed the highest and lowest upregulation, respectively. Responses to the LC 25 treatment were similar to those observed for the LC 10 treatment, with only sfABCG7 being significantly downregulated (3.5-fold) and the remaining 13 genes being significantly upregulated by 4.5-to 643.8-fold. However, we found that the expression levels of the upregulated genes showed a decreasing trend with increasing insecticide concentration, with sfABCG3 showing significantly different expression levels in response to the LC 10 and LC 25 treatments (Wilcoxon signed-rank test P < 0.05). After treatment with buprofezin at LC 10 , the relative expression levels of the 14 sfABCG genes showed trends similar to those observed for thiamethoxam at LC 10 , with only the sfABCG7 gene being significantly downregulated (3.0-fold), and the other 13 genes all being upregulated (significantly in the case of 12), Frontiers in Physiology | www.frontiersin.org from 6.0-to 924.0-fold. Although the expression of sfABCG9 was upregulated relative to that in the control, the difference was not significant (Figure 1). Buprofezin treatment at LC 25 resulted in a significant upregulation of sfABCG5 relative to the control, whereas sfABCG7 was significantly downregulated by 2.2-fold (Bonferroni-corrected P = 0.02) compared with the control level.
FIGURE 2 | Summary of the significantly upregulated genes in Sogatella furcifera under treatment with the insecticides thiamethoxam, buprofezin, and abamectin. The Venn diagram shows the putative ABC subfamily G (ABCG) transporter genes found to be significantly upregulated in the insecticide-treated insects compared with the untreated controls.
To gain a more intuitive understanding of the gene upregulation pattern in response to insecticide exposure, a Venn diagram was generated for the 13 significantly upregulated sfABCG genes after treatment with the three insecticides at LC 10 (Figure 2). Among these, sfABCG9 was upregulated by thiamethoxam only; sfABCG2, sfABCG12, and sfABCG13 were upregulated by thiamethoxam and buprofezin; and sfABCG1, sfABCG3, sfABCG4, sfABCG5, sfABCG6, sfABCG8, sfABCG10, sfABCG11, and sfABCG14 were upregulated by all Frontiers in Physiology | www.frontiersin.org three insecticides. Among the latter group, the sfABCG5 gene showed the highest upregulation responses, with expression levels 766.6-, 924.0-, and 97.4-fold higher than those of the control in response to thiamethoxam, buprofezin, and abamectin treatments, respectively.

Activity of Detoxifying Enzymes
Changes in the activity of the detoxifying enzymes in S. furcifera were examined after treatment with sublethal concentrations of the test insecticides for 6, 12, 24, 48, and 72 h (Figures 3-5). Compared with control levels, the activity of CarE was significantly increased after 6 and 12 h of treatment with thiamethoxam, buprofezin, and abamectin at the LC 10 and LC 25 levels, showing the same trend for all three insecticides and with the activity being highest at 6 h (Figure 3). It is worth noting that after treatment with the three insecticides at LC 10 and LC 25 , there was an initial increase in the overall activity of CarE with time, followed by a decrease, and then subsequently a further increase.
Glutathione S-transferase activity increased gradually and then decreased after treatment with thiamethoxam at LC 10 , peaking at 24 h (2.8-fold higher than that of the control). However, in response to treatment with thiamethoxam at LC 25 , there was no significant difference between the GST treatment and control groups after 24 h, and activity of the enzyme returned to normal levels at 72 h ( Figure 4A). After treatment with buprofezin at LC 10 and LC 25 , GST activity showed an overall increasing trend, being highest at 6 h after the LC 25 treatment, and subsequently decreasing with the prolongation of treatment time, albeit at levels significantly higher than that of the control. In contrast, we observed a significant reduction in GST activity in response to treatment with buprofezin at LC 10 for 24 h (Bonferroni-corrected P = 0.04) compared with that of the control, although again the levels had returned to normal at 72 h ( Figure 4B). In response to treatment with abamectin at LC 10 and LC 25 , the activity of GST increased significantly at 6 h, reached a maximum at 12 h, and then gradually decreased. Compared with control levels, the activity of this enzyme was significantly higher in response to the LC 25 treatment. However, similar to the response to buprofezin treatment at LC 10 , GST activity following abamectin treatment at LC 10 was not significantly different from that of the control at 24 and 72 h ( Figure 4C).
Compared with the control, the activity of MFO showed a decreasing trend in response to treatment with thiamethoxam at LC 10 , with the difference being significant at 6, 24, and 72 h. In contrast, in response to treatment with thiamethoxam at LC 25 , although the activity of MFO had decreased at 6 and 72 h, we observed a significant increase at 12 h (Bonferroni-corrected P = 0.001) relative to the control level ( Figure 5A). In response to buprofezin exposure at LC 10 , MFO activity was significantly increased at 6 and 12 h compared with that of the control, and reached a peak at 12 h (1.8-fold higher than that of the control). However, at 24, 48, and 72 h, the activity of MFO was significantly reduced. In addition, after treatment with buprofezin at LC 25 , we detected no significant difference between treatment and  control MFO activities at 6 and 12 h, whereas there was a significant increase in activity in response to treatment at 24 h, which thereafter gradually decreased ( Figure 5B). In response to treatment with abamectin at LC 10 and LC 25 , MFO activity showed a decreasing trend compared with the control levels, with the difference being significant at 48 and 72 h. However, in response to abamectin treatment at LC 10 , MFO activity was significantly higher than that of the control after 12 and 24 h ( Figure 5C).

Activity of Protective Enzymes
The activities of CAT, POD, and SOD were measured at 6, 12, 24, 48, and 72 h after exposure to sublethal concentrations of the test insecticides (Figures 6-8). Although at 6 h after treatment with thiamethoxam at LC 10 and LC 25 , we observed an inhibition of CAT activity, at 12 and 24 h the activity had increased significantly, respectively, but thereafter returned to normal levels ( Figure 6A). Following treatment with buprofezin at LC 10 and LC 25 , CAT activity had increased significantly by 1.8-and 2.1-fold at 12 and 6 h, respectively, compared with the control, and in the LC 25 treatment group thereafter gradually returned to a normal level (Figure 6B). Similarly, after treatment with abamectin at LC 10 and LC 25 , CAT activity was 2.4-and 1.9-fold higher, respectively, than that of the control at 6 h, and in the LC 25 treatment group subsequently underwent a gradual return to normal levels. However, after 48 h of LC 25 treatment, the activity of this enzyme had increased significantly to a level 1.6-fold higher than that of the control ( Figure 6C). Interestingly, in response to treatment with both buprofezin and abamectin at LC 10 , CAT activity initially increased, then decreased, and subsequently increased again with a prolongation of exposure time.
In response to treatment with thiamethoxam at both LC 10 and LC 25 , POD activity showed a tendency to initial increase and subsequently return to a normal level ( Figure 7A). In the case of the LC 10 treatment, POD activity peaked at 24 h (53.7-fold higher than that of the control) and then gradually decreased, albeit at levels still significantly higher than those of the control. In the LC 25 treatment group, POD activity peaked at 12 h (45.9-fold higher than that of the control) and then decreased gradually until reaching the normal level at 72 h ( Figure 7A). After treatment with buprofezin at LC 10 , POD activity began to increase significantly at 6 h (Bonferroni-corrected P = 0.004), peaked at 12 h (55.7-fold higher than that of the control), and then decreased gradually to a normal level after 72 h. Similarly, after treatment with buprofezin at LC 25 , POD activity showed a significant increase at 12 h (53.9-fold higher than that of the control) (Wilcoxon signed-rank test P < 0.05) and then gradually decreased to a normal level after 72 h ( Figure 7B). The responses of POD activity following exposure to abamectin at LC 10 and LC 25 showed similar patterns to those following buprofezin treatment at LC 10 and LC 25 , whereby activity peaked at 12 h (70.3-and 97.7-fold higher than that of the control, respectively) and returned to a normal level after 72 h ( Figure 7C). Compared with the control level, the SOD activity levels following thiamethoxam treatment at LC 10 and LC 25 were significantly increased at 6 h (1.4-and 2.5-fold higher than that of the control, respectively) and returned to normal levels at 12 h. Subsequently, however, the SOD activity showed a secondary significant increase at 24 h, before eventually returning to a normal level thereafter ( Figure 8A). In response to treatment with buprofezin at LC 10 , SOD activity increased significantly at 6 h (3.4-fold higher than that of the control), and then underwent a gradual decrease (Figure 8B), whereas following treatment at LC 25 , the activity of this enzyme increased significantly at 12 h to a level 1.6-fold higher than that of the control. After 48 h of exposure to buprofezin at LC 10 and LC 25 , SOD activity had decreased by 59.9 and 26.5%, respectively, compared with the control level, but had returned to a normal level at 72 h ( Figure 8B). The responses of SOD activity to treatment with abamectin at LC 10 and LC 25 showed trends similar to those following buprofezin treatment at LC 10 and LC 25 ; however, after 24 h of abamectin treatment at both sublethal concentrations, SOD activity showed a tendency to return to a normal level ( Figure 8C).

DISCUSSION
Previous studies on insects have shown that the protective enzymes SOD, POD, and CAT are related to resistance and the response to insecticide-induced stress. In this regard, it has been reported that sublethal concentrations (LC 10 and LC 25 ) of abamectin can promote upregulation of the SOD, POD, and CAT activities in Diadegma semiclausum adults, with activity increasing with increasing insecticide concentration (Jia et al., 2016). In third-instar H. axyridis nymphs exposed to LC 10 abamectin, the highest levels of SOD, POD, and CAT activity were recorded at 24, 12, and 24 h, respectively, and were significantly higher than those in the control group . In the present study, the overall levels of SOD, POD, and CAT activity in abamectin-treated (LC 10 and LC 25 ) S. furcifer tended to undergo an initial increase and thereafter gradually return to normal levels, reaching their highest levels at 12, 12, and 6 h, respectively. Interestingly, at 12 and 24 h, POD and CAT activities showed an increase in response to increasing abamectin concentration, which is consistent with the observations on D. semiclausum previously reported by Jia et al. (2016). In contrast, the levels of SOD, POD, and CAT activity in thirdinstar H. axyridis nymphs were shown to decrease with an increase in abamectin concentration . In the present study, we found that exposure to thiamethoxam initially tended to promote upregulation of the overall activities of POD and SOD and then inhibit them with an increase in the insecticide concentration from LC 10 to LC 25 , which contrasts with the observations for buprofezin (LC 10 and LC 25 ), which initially inhibited and then upregulated POD and SOD activities with increasing sublethal concentration. For CAT, buprofezin initially upregulated and then inhibited enzyme activity with increase in concentration, whereas thiamethoxam tended to initially inhibit and then upregulate CAT activity with increase in concentration. Similar observations have previously be made in Aphidius gifuensis, in which the levels of SOD, POD, and CAT activity tended to decrease with an increase in imidacloprid concentration (LC 10 , LC 20 , LC 30 , and LC 50 ) (Zhu et al., 2015). Such studies indicate that, in insects, SOD, POD, and CAT activities are related to insect resistance and the response to insecticide-induced stress, although the effects of these enzymes may be species, concentration, and time dependent.
The detoxifying enzymes CarE, GST, and MFO are also important components of insect resistance mechanisms, an increase in the activities of which is necessary during insecticide metabolism (Qi et al., 2016). Previously, it has been found that the levels of GST and MFO activity in two color morphs of the pea aphid Acyrthosiphon pisum increased in response to increasing sublethal concentrations of abamectin (LC 5 , LC 10 , and LC 20 ) following exposure for over 24 h (Wang and Liu, 2014). Similarly, the activities of CarE, GST, and MFO in Tetranychus urticae were significantly upregulated at 12 h following exposure to abamectin (LC 10 and LC 25 ) (Ru et al., 2017). In the present study, avermectin (LC 10 and LC 25 ) resulted in a similar significant induction of CarE, GST, and MFO activities in S. furcifera, at 6, 12, and 24 h, respectively. These findings indicate that insects can adapt to the stress induced by avermectin by activating their detoxifying enzymes. In addition, after treatment with thiamethoxam and buprofezin (LC 10 and LC 25 ), CarE activity showed an overall trend of initial upregulation and subsequent inhibition, with the activity being highest at 6 h. Thiamethoxam and buprofezin also significantly induced GST activity in S. furcifera, whereas these insecticides were found to have a generally inhibitory effect on the activity of MFO. Previously, it was found that GST and P450 activities in Aphis craccivora were significantly induced after treatment with cycloxaprid and imidacloprid (LC 50 ) for 48 h, whereas in contrast, the activity of the CarE activity was inhibited, although the observed difference was not significant (Wu et al., 2016). In addition, after treating Cydia pomonella with imidacloprid (LC 20 ), Shang et al. (2017) observed a significant induction of CarE and GST activity, whereas MFO activity was significantly inhibited. These findings suggest that MFO may not play a major role in the insect response to stress induced by neonicotinoid insecticides, and that the primary detoxifying enzymes are CarE and GST. The aforementioned findings indicate that detoxifying enzymes enable insects to respond to low levels of insecticide-induced stress; however, similar to protective enzymes, CarE, GST, and MFO are induced at different times in different insects. Moreover, the main enzymes involved in detoxification appear to be species dependent.
The ABC transporters are important participants in the third stage of detoxification and have been widely reported to be involved in insecticide resistance (Qi et al., 2016). In this regard, it has previously been found that the expression levels of an ABCG gene and an ABCC gene were upregulated in S. furcifera treated with a high concentration (LC 85 ) of cycloxaprid, whereas the expression levels of two ABCG genes were upregulated at a low concentration (LC 15 ) of this insecticide (Yang et al., 2016). Transcriptome sequencing has revealed that the ABCB, ABCC, and ABCG subfamily genes are expressed at high levels in a pyrethroid-resistant strain of Aedes aegypti (Bariami et al., 2012). Similarly, results of microarray experiments have shown that genes of the ABCG and ABCH subfamilies are expressed at high levels in resistant strains of Myzus persicae (Silva et al., 2012b), and that the expression levels of ABCG subfamily genes are increased in DDT-resistant strains of Anopheles arabiensis . Given that ABCG subfamily genes play a role in insecticide resistance in many insects (You et al., 2013;Yang et al., 2016;Sun et al., 2017;Xiao et al., 2018), we decided to study the expression of 14 ABCG subfamily genes in S. furcifera in response to thiamethoxam, buprofezin, and abamectin. We accordingly found that 13 of these 14 sfABCG genes were significantly upregulated after treatment with at least one sublethal concentration of insecticide. On exposure to these insecticides at the LC 10 level, 13 sfABCG genes were significantly upregulated by thiamethoxam, 12 were significantly upregulated by thiamethoxam and buprofezin, and nine were upregulated by all three insecticides. Furthermore, 13 and seven sfABCG genes were significantly upregulated after treatment with LC 25 concentrations of thiamethoxam and abamectin, respectively. These findings provide further evidence that ABC transporters probably participate in the transport of various substrates related to the resistance to different types of insecticides. Moreover, it is conceivable that, in addition to enhancing the metabolism of S. furcifera, these highly expressed sfABCG genes are associated with cross-resistance in this insect. However, these inferences need to be verified with functional experiments.
The sublethal effects of insecticides on insects are multifaceted, including their effects on insect behavior, reproduction, development, and insecticide resistance. In addition, insect adaptation to insecticide stress is a complex metabolic detoxification process involving the activity of multiple enzymes. The results of our study show that S. furcifera can eliminate insecticides in the body by activating detoxifying enzymes and ABC transporters, and also activate the protective enzyme system to prevent injury to the body. Taken together, our research results lay the foundations for gaining a deeper understanding of the mechanisms contributing to the adaptation of S. furcifera to different types of insecticides, which is of considerable significance with regards to the development of effective pest management strategies.

DATA AVAILABILITY STATEMENT
The gene sequences obtained have been submitted to the NCBI database (Accession Nos. MH481837-MH481850). Other datasets for this study are included in the manuscript and the Supplementary Files.