**3. Discussion**

Although rice planthoppers in China are still sensitive to fluorodinitrile, our results showed that the toxicity of sulfoxaflor to *S. furcifera* decreased when the Sus-Lab strain was successively screened with the sublethal dose of sulfoxaflor, which indicated that *S. furcifera* presented some resistance to sulfoxaflor. Liao et al. [24] monitored the resistance levels of *Nilaparvata lugens* to sulfoxaflor from 2013 to 2016 in China and found that all field-collected populations were still sensitive, with LC50 values ranging from 1.63 to 13.20 mg/L (resistance ratio from 0.8 to 6.8-fold). Liao et al. [25] continuously screened *N. lugens* with a sublethal dose of sulfoxaflor in approximately 39 intervals and finally obtained an extremely sulfoxaflor-resistant strain with a resistance ratio of 183.6-fold. Ma et al. [26] performed continuous selection of *Aphis gossypii* gradually increasing LC50 concentrations of sulfoxaflor based on bioassays of the parental generations for a total of 27 generations in the laboratory, finally resulting in a 366.4-fold resistance ratio compared with the susceptible strain. Therefore, there is an extremely high risk of insects developing resistance to sulfoxaflor, and it is necessary to perform resistance monitoring in field populations, along with the investigation of resistance mechanisms and cross-resistance to design integrated pest management strategies.

Insecticide resistance is inevitable after the application of insecticides, and the main reasons include a reduced penetration rate, increased detoxification and metabolism of insecticides (MFOs [27], CarEs [28], GSTs [29]) and decreased sensitivity at the target site. Wei et al. [30] found that PBO and

TPP could increase bifenthrin toxicity in resistant *A. gossypii* Glover strains by 2.38- and 4.55-fold, respectively. Liao et al. [25] showed that the toxicity of sulfoxaflor to sulfoxaflor- resistant *N. lugens* (Stål) showed a synergistic effect with PBO resulting in a 2.69-fold relative synergistic ratio, and the P450 enzyme activity of SFX-SEL was increased 3.50-fold compared with that in the unselected strain (UNSEL). Mao et al. [31] also reported that a resistant strain (NR) with a high nitenpyram resistance level (164.18-fold) and cross-resistance to sulfoxaflor (47.24-fold) showed a 3.21-fold increase in P450 activity compared to that in NS, and resistance also showed a synergistic effect (4.03-fold) with the inhibitor PBO, suggesting a role of P450. Our results further demonstrated that the inhibitors PBO, TPP, and DEM showed some synergism with sulfoxaflor regarding the toxicity and inhibition of the activities of three types of metabolic detoxification enzymes in the Sus-Lab and SF-Sel strains; this effect was especially strong for the inhibitor PBO in the SF-Sel strain.

Normally, the contribution of the overexpression of detoxification metabolism genes to an increased detoxification ability, especially which of the P450 genes related to insecticide detoxification metabolism, is the main reason for insect resistance to insecticides [32–34]. Jones et al. [35] reported that the resistance of the ALM07 strain of B-biotype populations of *Bemisia tabaci* adults to imidacloprid reached 180-fold, and the relative normalized expression of the resistance gene *CYP6CM1* in adults and nymphs reached 4.2- and 200-fold in the resistant strain, respectively. The overexpression of *CYP6AY1* contributes to the development of resistance to imidacloprid in *N. lugens* [36]. Liao et al. [25] and Mao et al. [31] also demonstrated that the reducing expression of *CYP6ER1* in sulfoxaflor-resistant strain through RNAi could significantly increase its' susceptibility to sulfoxaflor. Our transcriptome data and qRT-PCR results also indicated that two P450 genes, *CYP6FD1* and *CYP4FD2,* and three transcription factors, *NlE78sf*, *C2H2ZF1* and *C2H2ZF3*, were clearly upregulated in the SF-Sel strain. The RNAi results also showed that when 3rd-instar nymphs were injected with the *CYP6FD1* and *CYP4FD2* dsRNA, the relative expression of *CYP6FD1* and *CYP4FD2* was decreased, causing the insects to be more sensitive and ultimately to show higher mortality compared with negative dsGFP control treatment. However, it is still uncertain which transcription factors are mainly responsible for regulating the overexpression of *CYP6FD1* and *dsCYP4FD2*, and require further study in the future.

The P450s area multi-enzyme complex, and the first step in the metabolism of exogenous toxic substances is recognition by a CYP protein, which binds the toxin; then, electrons are transferred by electron donors to exogenous REDOX substances [37]. At present, the examination of P450 structure in insects generally concentrates on assessing highly conserved sequence motifs, such as the residue pairs WxxxR in helix C, CxxT in helix I, ExxR in helix K, RxxF in the meander region, and FxxGxRxCxG in the canonical heme-binding domain [38]. Our research showed that the molecular structure of sulfoxaflor was surrounded by the active pocket of *CYP6FD1*, while the active pocket was located near the heme-binding region. This protein exhibits a predicted active site structure with an oval shape [39], a large volume, and large substrate channels, allowing sulfoxaflor to fit the active site cavity. The spacious cavity of P450 enzymes enables larger molecules to access the heme-bound oxygen of the reaction center; therefore, we hypothesize that *CYP6FD1* could present a greater metabolic ability than *CYP4FD2* [40,41].

On the basis of our results, the main findings show that it is likely that *S. furcifera* will develop resistance to sulfoxaflor and that upregulation of detoxification enzymes such as P450s is a likely mechanism. However, the authors also show that the toxicity of sulfoxaflor is increased by using synergistic agents, so perhaps this is one possible approach that could be used in the field to prevent rapid development of resistance to this compound. Meanwhile, we also find that two main P450 genes (*CYP6FD1* and *CYP4FD2*) could be related to the development of resistance of *S. furcifera* to sulfoxaflor. Our results should provide a foundation for subsequent efforts to investigate the expression of *CYP6FD1* and *CYP4FD2* in heterologous expression systems, such as baculovirus- infected Sf9 cells, and metabolic processes in vitro and transcriptional regulation of the two genes in the further investigations.
