Cuticle protein mediates the evolution of stress resistance by generating a decoy circular RNA in spider mite

Abstract

Phytophagous mites, including Tetranychus cinnabarinus, are arthropods known for their wide infestation of host plants and pesticide resistance. We found that fenpropathrin-resistant female mites (YN-FeR, with target resistance: F1538I kdr mutation) exhibited significantly enhanced adaptability to various stress conditions, including exposure to different acaricides and high-temperature (34°C) and low-humidity environments (40% relative humidity). This evolution was attributed to cuticle thickening in resistant female mites. Cuticle protein CPR25 was identified as a critical gene mediating cuticle thickening. CPR25 regulated its own overexpression by producing a circular RNA, named circCPR25, which acted as a decoy to selectively sequester and bind to the miR-34~317 cluster. This study revealed a distinctive mechanism underlying the evolution of stress resistance in spider mites. Specifically, a cuticle protein in spider mites regulates its own overexpression by producing a decoy circRNA, thereby promoting cuticle thickening and facilitating rapid adaptation to adverse conditions.

A cuticle protein drives adaptive evolution to adversity in R-strategy pests through feedback-regulated pathways.

INTRODUCTION

The adaptation of arthropods to biological toxins and plant secondary metabolites is a slow evolutionary process (1), but the use of chemical pesticides has accelerated the adaptation process of arthropods to exogenous toxins, i.e., pesticide resistance (2, 3). Spider mites serve as an ideal model for investigating the adaptive evolutionary mechanisms of arthropods to xenobiotics due to their broad host range, which encompass more than 1100 plant species containing a wide variety of secondary metabolites (4). The issue of resistance in spider mites is particularly grave, with resistance having been observed in 96 compounds, making this a pressing challenge in arthropod management (5). The overexpression of resistance genes and genetic variations mediated by target mutations are primary drivers behind resistance evolution in spider mites (6). Notably, the mechanisms involving detoxifying enzyme genes, such as cytochrome P450, glutathione S-transferase, and carboxylesterase, which confer resistance, have been well studied in insects and mites (7–9). Furthermore, mutations affecting acetylcholinesterase (10), voltage-gated sodium ion channels (11), succinate dehydrogenase (12), and other targets have been identified in highly resistant strains (13).

Pesticide use leading to resistance is a common phenomenon in arthropods; however, a recent study has suggested that some chemicals induce developmental variations in arthropods concurrent with resistance evolution (14). Pyrethroids are extensively used in insect and mite control, comprising a substantial portion of the pesticide market (5). However, the extended history of use has exacerbated the issue of pyrethroid target resistance. Kdr mutations in sodium ion channels L1014F, M918L (15), and F1538I (16) have been identified in numerous field and indoor insect and mite strains. With the evolution of pyrethroid resistance, there are notable linked variations in arthropod cuticle development, with numerous cases observed in mosquitoes (17). For example, following the development of deltamethrin resistance in Anopheles gambiae, the cuticle experiences significant thickening (18). In addition to mosquitoes, cuticle thickening has been documented in Cimex lectularius (19), Triatoma infestans (20), and Bactrocera dorsalis (21) during the evolution of pyrethroid resistance. The cuticle, an essential component of the insect and mite body wall that is primarily composed of proteins, lipids, and hydrocarbons, plays a vital role in maintaining the body structure and water balance (22). Cuticle thickening not only enhances arthropods’ adaptation to extreme environments, such as high temperatures and drought, but also impedes the penetration efficiency of pesticides (23). This obstacle to pesticide molecule penetration is likely nonspecific and may lead to broad-spectrum resistance, posing a challenge for arthropod resistance management (17).

Cuticle proteins (CPs) constitute the primary protein constituents in the body wall of arthropods (24). Advancements in genomics have facilitated the systematic identification of CPs in various arthropod species. Although the functions of these proteins have not yet been fully validated, several studies have suggested that some CPs bind to chitin, influencing insect cuticle development (25, 26). CP overexpression is correlated with cuticle thickening in arthropods. For instance, CPLC8 overexpression is linked to cuticle thickening in Aedes aegypti (27), and overexpression of multiple CPs mediates cuticle thickening in A. gambiae (28). In insects or spider mites, gene overexpression is as a critical factor mediating physiological changes. Recent studies on the mechanisms of gene overexpression have showed on the role of key regulatory factors. These regulatory processes primarily occur at the transcriptional and posttranscriptional levels. For instance, the nuclear receptor HR96 regulates the transcription of various detoxification enzymes, mediating pesticide resistance and adaptive evolution in spider mites across different hosts (29, 30). Additionally, the regulation of detoxification enzymes and development-related genes by microRNAs (miRNAs) or long noncoding RNA at the posttranscriptional level has been elucidated in mites (31, 32). With advances in molecular biology, previously unknown classes of noncoding RNAs, such as circular RNAs (circRNAs), have also been identified. circRNA is a covalently closed-loop noncoding RNA lacking a 5′ cap structure and a 3′ poly(A) tail (33). Initially considered by-products of erroneous splicing, the regulatory functions of circRNAs on protein-coding genes in mammals have gradually been identified with advancements in molecular biology over the last decade (34). In arthropods, circRNAs from numerous insects and mites have been identified using transcriptome sequencing, and the functions of a few circRNAs have been elucidated. For example, circSfl regulates life-span extension in Drosophila (35), and circ1-3p regulates cyflumetofen resistance by interacting with miR-1-3p in Tetranychus cinnabarinus (36). However, the functions and pathways of circRNAs in regulating trait changes in arthropods remain largely unclear.

In this study, although T. cinnabarinus developed target resistance to fenpropathrin, resistant female mites displayed notable cuticle thickening, which conferred penetration resistance against diverse acaricides with different modes of action. Resistant female mites also had a higher adaptation to extreme environmental conditions characterized by high temperatures and drought. Using gene expression analysis, a distinctive CPs, CPR25, was identified, with an up-regulation over 18.6-fold in resistant strains. The substantial increase of CPR25 expression was associated with the generation of a “daughter” circRNA, referred to as circCPR25, during transcription. Similar to a “decoy” released by CPR25, circCPR25 competed for binding with the miR-34~317 cluster, which originally bound to CPR25 and negatively regulated its expression in the cytoplasm. Consequently, circCPR25 regulated CPR25 overexpression and facilitated cuticle thickening in resistant female mites. Our findings elucidated a mechanism in arthropods in which CPs produce circRNAs as decoys, subsequently regulating their own overexpression through feedback mechanisms, facilitating cuticle thickening in spider mites, and resulting in enhanced adaptability to adverse conditions.

RESULTS

Cuticle thickening mediates the stress resistance evolution in YN-FeR

In the residual coated vial (RCV) bioassay, the YN-FeR strains displayed 25.4-, 21.0-, and 10.9-fold cross-resistance to pyridaben, bifenazate, and cyflumetofen, respectively (Table 1). The cross-resistance was also found in the results of spraying method (Table 1). However, direct pesticide injection or oral into adult female mites showed no significant differences in mortality between the YN-SS and YN-FeR (Fig. 1, A to C, and Table 1), suggesting nonspecific penetration resistance when the pesticide traversed the cuticle of adult female mites of the YN-FeR strain. Compared to the YN-SS and YN-KM strain, the cuticle ultrastructure in adult female mites of the YN-FeR strain showed a significant 1.7-, 1.9-, and 1.8-fold increase in ridge, procuticle, and total cuticle thickness, respectively (Fig. 1, D to H, and fig. S1, A to C). The penetration assays for fenpropathrin, pyridaben, bifenazate, and cyflumetofen showed consistency; i.e., compared to the YN-SS strain, adult female mites of the YN-FeR strain retained more pesticide molecules on their body surface, significantly decreasing the amount of pesticide entering the body (Fig. 1I and fig. S2). Additionally, under extreme high-temperature and low-humidity conditions (34°C and relative humidity of 40%), the survival rate and body weight of adult female mites of the YN-FeR strain were significantly higher compared to those of the YN-SS strain (Fig. 1, J and K), suggesting that the thickened cuticle markedly reduced moisture loss and enhanced the adaptability of resistant adult female mites to these environmental conditions.

Table 1. Bioassay of pyridaben, bifenazate, and cyflumetofen on the YN-SS, YN-KM and YN-FeR strains.

CI, confidence interval. X², chi-square value. LC₅₀, median lethal concentration.

Acaricides	Strains	X 2	LC50 (mg/liter)	95% CI	RR*
RCV method
Fenpropathrin	YN-SS	1.2 (P = 0.8)	668.1	552.3–830.9	–
	YN-KM	1.6 (P = 0.6)	510.6	378.1–673.4	0.8
	YN-FeR	–	>80,000	–	>119.7
Pyridaben	YN-SS	4.0 (P = 0.2)	134.4	62.0–285.6	–
	YN-KM	1.0 (P = 0.8)	121.1	97.9–147.3	0.9
	YN-FeR	5.1 (P = 0.2)	3413.0	2788.6–4324.3	25.4
Bifenazate	YN-SS	2.5 (P = 0.5)	95.7	74.4–117.9	–
	YN-KM	1.7 (P = 0.6)	105.7	80.6–132.5	1.1
	YN-FeR	0.7 (P = 0.8)	2010.6	1586.8–3006.2	21.0
Cyflumetofen	YN-SS	5.4 (P = 0.1)	1.5	0.3–2.6	–
	YN-KM	3.4 (P = 0.3)	2.7	2.1–3.3	1.8
	YN-FeR	1.0 (P = 0.8)	16.4	13.9–20.8	10.9
Spraying method
Pyridaben	YN-SS	1.8 (P = 0.6)	118.3	104.3–134.3	–
	YN-FeR	2.4 (P = 0.5)	2414.0	2073.0–2838.3	20.4
Bifenazate	YN-SS	0.7 (P = 0.8)	98.3	72.7–124.0	–
	YN-FeR	1.0 (P = 0.8)	1789.2	1369.3–2180.2	18.2
Cyflumetofen	YN-SS	2.7 (P = 0.5)	2.2	1.6–2.8	–
	YN-FeR	1.6 (P = 0.5)	15.0	13.1–17.8	6.8
Oral method
Pyridaben	YN-SS	2.6 (P = 0.4)	85.3	69.0–100.7	–
	YN-FeR	1.7 (P = 0.6)	79.7	68.0–96.7	0.9
Bifenazate	YN-SS	1.0 (P = 0.8)	64.1	54.5–76.0	–
	YN-FeR	1.1 (P = 0.8)	72.7	60.4–89.7	1.1
Cyflumetofen	YN-SS	2.0 (P = 0.5)	3.9	2.9–4.9	–
	YN-FeR	3.3 (P = 0.3)	4.9	4.3–5.6	1.3

*Resistance ratio = LC₅₀ of YN-FeR/LC₅₀ of YN-SS.

Fig. 1. Cuticle thickening mediates nonspecific resistance to acaricides and enhances adaptability to high-temperature and low-humidity environments in the YN-FeR stain.

(A to C) Mortality rates of adult female mites injected with different concentrations of pyridaben, bifenazate, and cyflumetofen, respectively, after 24 hours. Twenty-five females for each concentration treatment. (D to F) Cuticle ultrastructure of YN-FeR, YN-SS, and YN-KM adult female mites, respectively. The number of the test: YN-FeR, 14 females; YN-SS, 13 females; and YN-KM, 10 females. The red dashed box indicates the area (random) undergoing progressive magnification. The magnification ratios were ×2000 (2K), ×10,000 (10K), and ×20,000 (20K). The white dashed line delineates the boundary between the cuticle layer and the underlying tissues. (G) Schematic representation of the mite cuticle structure. (H) Statistical analysis of the ridge, procuticle, and total cuticle thicknesses of the three strains. (I) Penetration ratios of fenpropathrin, pyridaben, bifenazate, and cyflumetofen in YN-SS and YN-FeR. (J) Survival rates of YN-FeR and YN-SS adult female mites in high-temperature and low-humidity environments. The detection time points were 24 and 48 hours, respectively. (K) Average weights of YN-FeR and YN-SS adult female mites at different times following high-temperature and low-humidity treatment. Thirty females for each treatment. DMSO, dimethyl sulfoxide; h, hours.

CPR25 serves as a critical factor in mediating cuticle thickening in the YN-FeR strain

A total of 44 CPs were identified in T. cinnabarinus using RNA sequencing (RNA-seq) (fig. S3 and data S1). Differential expression analysis revealed that nine CPs were overexpressed in the YN-FeR strain compared to those in the YN-SS strain, and all belonged to the CPR family. Quantitative polymerase chain reaction (qPCR) validated the RNA-seq results (Fig. 2, A and B). Although eight CPs were up-regulated, ranging from 1.8- to 4.9-fold, CPR25 exhibited a significant 18.6-fold up-regulation (Fig. 2B), suggesting its critical role in cuticle thickening in the YN-FeR strain. Compared to the dsGFP-treated group, knockdown of CPR25 expression in the YN-FeR strain (Fig. 2C and fig. S4A) resulted in a significant 1.4-, 2.1-, and 1.7-fold decrease in ridge, procuticle, and total cuticle thickness in adult female mites (Fig. 2, D to F; and figs. S1, D and E, and S5, A and B). Moreover, it significantly weakened the hindering effect on acaricide penetration (Fig. 2G and fig. S6), leading to a substantial increase in sensitivity to pyridaben, bifenazate, and cyflumetofen (Fig. 2J). However, knockdown of CPR25 expression did not significantly enhance the fenpropathrin sensitivity of YN-FeR, with statistical differences in mortality rates observed only under treatments with 80,000 and 100,000 mg/liter (Fig. 2J), suggesting that cuticle thickening was not the primary factor mediating fenpropathrin resistance in T. cinnabarinus. Under high-temperature and low-humidity conditions, CPR25 knockdown significantly decreased the survival rate and body weight of adult female mites of the YN-FeR strain compared to those of the dsGFP-treated group (Fig. 2, H and I). These findings collectively suggest that CPR25 overexpression mediates cuticle thickening in the YN-FeR strain, thereby contributing to resistance against adverse conditions.

Fig. 2. CPR25 overexpression mediates cuticle thickening and resistance against adverse environments in T. cinnabarinus.

(A) Volcano plot of differentially expressed genes between the YN-FeR and YN-SS strains, with CPR25 indicated by the arrow. A total of 206 up-regulated genes and 98 down-regulated genes. (B) Expression patterns of nine CPs in the YN-FeR and YN-SS strains. (C) The protein expression levels of CPR25 after dsRNA treated. The image of Western blotting contains three biological repetitions (dsGFP1-3 and dsCPR25-1-3) (D and E) Cuticle ultrastructure of adult female mites in the dsGFP-treated group and the dsCPR25-treated group, respectively. The number of the test: dsGFP, 14 females; and dsCPR25, 14 females. The red dashed box indicates the area (random) undergoing progressive magnification. The magnification ratios were ×2000 (2K), ×10,000 (10K), and ×20,000 (20K). The white dashed line delineates the boundary between the cuticle layer and the underlying tissues. (F) Statistical analysis of cuticle thickness in YN-FeR after silencing CPR25. (G) Penetration ratios of pyridaben, bifenazate, and cyflumetofen in YN-FeR after silencing CPR25 expression. (H) Survival rates of resistant female mites in the dsGFP-treated group and the dsCPR25-treated group under high-temperature and low-humidity conditions. The detection time points were 24 and 48 hours, respectively. (I) Average weights of resistant female mites in the dsGFP-treated group and the dsCPR25-treated group at different time points under high-temperature and low-humidity conditions. Thirty females for each treatment. (J) Bioassays with pyridaben, bifenazate, and cyflumetofen in YN-FeR after silencing CPR25 expression. The x axis represents the logarithmic scale of concentration: Log₁₀^{concentration}. Twenty-five females for each concentration treatment. FDR, false discovery rate; FC, fold change; Up, up-regulation; NS, not significant; Down, down-regulation; h, hours.

The miR-34~317 cluster inhibits CPR25 expression in the cytoplasm

miRNA plays an important role in regulating protein-coding gene expression. To elucidate the underlying reasons for the marked CPR25 up-regulation, we compared the miRNA transcriptomes between YN-FeR and YN-SS and revealed a miRNA cluster comprised of miR-34 and miR-317 in T. cinnabarinus (fig. S7, A and B). Significant down-regulation of miR-34 and miR-317 expression was observed in the YN-FeR strain (Fig. 3, A and C). Software-based prediction identified binding elements for miR-34 and miR-317 within the CPR25 coding sequence (Fig. 3B and fig. S7, C to E). A dual-luciferase reporter assay revealed significant inhibition in the fluorescence activity of the CPR25 reporter gene vector in human embryonic kidney (HEK) 293T cells following transfection with miR-34 and miR-317 mimics. This inhibitory effect was abolished with mutation of the corresponding binding elements, indicating that miR-34 and miR-317 bound to the predicted region in CPR25, thereby suppressing its expression levels (Fig. 3, D and E). Further in vivo experiments confirmed these results. Compared to the negative controls, feeding miR-34 and miR-317 mimics for 72 hours resulted in a significant decrease in CPR25 expression levels by 39.1 and 60.0%, respectively, in the YN-FeR strain (Fig. 3, F and G, and fig. S4, B and C). In addition, after feeding the miR-34 and miR-317 mimics, a comparable pattern emerged in the cuticle ultrastructure, acaricide penetration, and bioassay results for female adult mites of the YN-FeR strain, i.e., a notable reduction in cuticle thickness (Fig. 3, H to K; and figs. S1, F to H, and S5, C and D), significantly decreased impediment of cuticle penetration to pyridaben, bifenazate, and cyflumetofen (Fig. 3L and fig. S8), and significantly enhanced sensitivity of pyridaben, bifenazate, and cyflumetofen (Fig. 3M). These results suggested that the miR-34~317 cluster functions in the inverse regulation of CPR25 expression and cuticle development.

Fig. 3. The miR-34~317 cluster reverse regulates CPR25 expression.

(A) Volcano plot of differentially expressed miRNAs between the YN-FeR and YN-SS strains. A total of 108 up-regulated miRNAs and 69 down-regulated miRNAs. (B) Binding sites of the miR-34~317 cluster in the CPR25 coding sequence (CDS). The schematic diagram of CPR25 mRNA was created in BioRender: L. He (2025), https://BioRender.com/k06r452. (C) The expression patterns of miR-34 and miR-317 in the YN-SS and YN-FeR strains. (D and E) Inverse regulatory relationship of the miR-34~317 cluster on CPR25 expression using a dual-luciferase reporter system. pmCPR25, luciferase reporter vector of CPR25; pmCPR25m, mutant luciferase reporter vector of CPR25. Different letters indicate significant differences [analysis of variance (ANOVA), Tukey test, P < 0.05]. (F and G) CPR25 expression pattern in female adult mites of the YN-FeR strain after miR-34/317 overexpression. The image of Western blotting contains three biological repetitions [negative control (NC) mimic1-3 and miR-34/317 mimic1-3]. (H to J) Cuticle ultrastructure of female adult mites of the YN-FeR strain after miRNA mimic feeding. The number of the test: NC mimic, 14 females; miR-34 mimic, 14 females; and miR-317 mimic, 14 females. The red dashed box indicates the area (random) undergoing progressive magnification. The magnification ratios were ×2000 (2K), ×10,000 (10K), and ×20,000 (20K). The white dashed line delineates the boundary between the cuticle layer and the underlying tissues. (K) Statistical data on total cuticle thickness in female adult mites of the YN-FeR strain after miRNA mimic feeding. (L) Penetration ratio assays of pyridaben, bifenazate, and cyflumetofen on the cuticle of female adult mites of the YN-FeR strain after miR-34 and miR-317 overexpression, respectively. (M) Bioassay results for pyridaben, bifenazate, and cyflumetofen on female adult mites of the YN-FeR strain after miR-34 and miR-317 overexpression. The x axis represents the logarithmic scale of concentration: Log₁₀^{concentration}. Twenty-five females for each concentration treatment. FC, fold change; Up, up-regulation; NS, not significant; Down, down-regulation.

CPR25 transcription generates a daughter circRNA, known as circCPR25

CPR25 DNA transcribed mRNA, while its second and third exons underwent back-splicing, generating a circular daughter RNA, named circCPR25 (Fig. 4, A and B). A specific divergent primer was designed, and a circCPR25 product was successfully cloned from ribonuclease R (RNase R)–treated cDNA templates (Fig. 4D). No positive bands were observed in the corresponding genomic DNA (gDNA) templates (Fig. 4D). The circCPR25 back-splicing site was identified using Sanger sequencing, consistent with the RNA-seq results (Fig. 4E). Moreover, compared to the water-treated group, no significant changes were observed in the circCPR25 expression level after RNase R treatment, but the expression levels of linear CPR25 and GAPDH transcripts significantly decreased (Fig. 4F). These results strongly support the circular structure of circCPR25. qPCR revealed significant 16.0-fold up-regulation of circCPR25 in the YN-FeR strain compared with that in the YN-SS strain (Fig. 4C). Furthermore, CPR25 and circCPR25 displayed consistent expression patterns throughout the development of female adult mites aged 0 to 7 days, with a peak in expression between 0 and 3 days before a significant decrease (Fig. 4, G and H). These results suggest a robust correlation between the expression profiles of CPR25 and circCPR25 during cuticle development in female adult mites and highlight the 0- to 3-day period as pivotal for epidermal development. Subcellular localization via nuclear-cytoplasmic fractionation revealed that CPR25 predominantly localized in the cytoplasm, whereas circCPR25 was distributed in both the nucleus and cytoplasm, with a predominant presence in the cytoplasm (Fig. 4I).

Fig. 4. circCPR25 was generated by CPR25.

(A) Volcano plot of differentially expressed circRNAs between the YN-FeR and YN-SS strains, with an arrow indicating circCPR25. A total of 22 up-regulated circRNAs and 19 down-regulated circRNAs. (B) Schematic diagram depicting the generation of circCPR25 from CPR25. The arrows pointing in opposite directions indicate the convergent primers, whereas the arrows pointing outward indicate the divergent primers. (C) The expression pattern of circCPR25 in YN-SS and YN-FeR strains (D) Identification of circCPR25’s circular structure by PCR. The arrows pointing outward represent divergent primers amplification, while the arrows pointing in opposite directions indicate convergent primers amplification. The red arrow indicates the PCR product of circCPR25. (E) Identification of circCPR25’s back-splicing sites by Sanger sequencing and the red line indicates the identified splicing sites. (F) Resistance of circCPR25 to RNase R digestion. (G and H) Expression patterns of CPR25 and circCPR25 in female adult mites aged 0 to 7 days. Different letters indicate significant differences (ANOVA, Tukey test, P < 0.05) (I) Subcellular localization of both CPR25 and circCPR25. GAPDH and U6 function as control genes for cytoplasmic and nuclear localization, respectively. FC, fold change; Up, up-regulation; NS, not significant; Down, down-regulation; GAPDH, glyceraldehyde-3-phosphate dehydrogenase.

CircCPR25 functions as a decoy molecule to compete for binding with the miR-34~317 cluster

CircCPR25 predominantly localized in the cytoplasm of T. cinnabarinus cells; therefore, our investigation further determined whether it inhibited the function of the miR-34~317 cluster. Software prediction revealed the presence of binding elements within the circCPR25 sequence for the miR-34~317 cluster (fig. S7E). In YN-FeR female adult mites, streptavidin magnetic beads effectively enriched circCPR25 using biotin-labeled circCPR25 probes for RNA pulldown detection (Fig. 5A). Conversely, no significant enrichment effects were observed on the control circRNA, circ1-3p. Compared to nontarget linear genes RP18s and miR-133-5p, circCPR25 enrichment mediated significant enrichment of miR-34 and miR-317, providing evidence for the role of circCPR25 in adsorbing and binding to the miR-34~317 cluster (Fig. 5A). After the overexpression of biotin-labeled miR-34 and miR-317 in YN-FeR female adult mites, RNA pulldown assays revealed that streptavidin magnetic beads enriched miR-34 and miR-317 compared to the negative control group. They also significantly enriched circCPR25 and CPR25. Conversely, no enrichment was observed in the non-targeted control genes, RP18s and α-Tub, suggesting the competitive binding potential of circCPR25 with CPR25 for miR-34 and miR-317 in T. cinnabarinus (Fig. 5B). A dual-luciferase reporter system validated these results in vitro. In HEK293T cells, transfection of miR-34 and miR-317 mimics inhibited the fluorescence activity of the circCPR25 reporter gene vector, indicating binding (Fig. 5, C and D). When HEK293T cells were transfected with miR-34 and miR-317 mimics and circCPR25 was simultaneously overexpressed (fig. S4D), the inhibitory effect of miR-34 and miR-317 on the fluorescence activity of the CPR25 reporter vector vanished (Fig. 5E), indicating that circCPR25 competitively bound to miR-34 and miR-317 in the cytoplasm of HEK293T cells, thereby significantly reducing their ability to inhibit CPR25 expression. Additionally, fluorescence in situ hybridization (FISH) revealed the co-localization signal of circCPR25 and the miR-34~317 cluster in the cortical cells of YN-FeR female adult mites (Fig. 5, F to Q), directly indicating their binding relationship. These results collectively showed that CPR25 released circCPR25 as a decoy, enabling the latter to adsorb and bind to miR-34~317 in the cytoplasm, thereby ensuring its own overexpression.

Fig. 5. circCPR25 serves as a decoy to adsorb and bind the miR-34~317 cluster.

(A) RNA pulldown detection of circCPR25. Circ1-3p, RP18s, and miR-133-5p serve as control genes that do not exhibit any interaction with circCPR25. (B) RNA pulldown assays of the miR-34~317 cluster. RP18s and α-Tub act as control genes, as their sequences do not contain binding elements for the miR-34~317 cluster. NS, not significant. (C and D) Dual-luciferase reporter system assays determining the binding relationship between miR-34, miR-317, and circCPR25. pmCirc, luciferase reporter vector of circCPR25; pmCircm, mutant luciferase reporter vector of circCPR25. (E) Functional detection of circCPR25 competitively adsorbing the miR-34~317 cluster. pcirc25, overexpression vector of circCPR25; pmCPR25, luciferase reporter vector of CPR25. Different letters indicate significant differences (ANOVA, Tukey test, P < 0.05) (F to I) Tissue localization of circCPR25 and miR-34 in YN-FeR female adult mites, with arrows indicating co-localization fluorescence signals. (J to M) Tissue localization of circCPR25 and miR-317 in YN-FeR female adult mites, with arrows indicating co-localization fluorescence signals in the cuticle. (N to Q) Fluorescence in situ hybridization (FISH) signals of negative control probes.

CircCPR25 regulates CPR25 overexpression

A specific small interfering RNA (siCirc) was designed to target the circCPR25 back-splicing site (fig. S4E). Compared to that in the siGFP control group, the circCPR25 expression level significantly decreased by 59.3% in the siCirc-treated group. CPR25 expression was also inhibited, with a decrease of 78.3% (Fig. 6, A and D). The miR-34 and miR-317 expression levels were significantly up-regulated (Fig. 6A). Moreover, in female mites of the YN-FeR strain, ridge, procuticle, and total cuticle thickness decreased 1.5-, 2.7-, and 1.9-fold, respectively (Fig. 6, B, C, and E; and figs. S1, I and J, and S5, E and F), and their resistance effect on the penetration of pyridaben, bifenazate, and cyflumetofen significantly decreased (Fig. 6F and fig. S9). The bioassays demonstrated that silencing circCPR25 expression significantly enhanced the pyridaben, bifenazate, and cyflumetofen sensitivity of female adult mites of the YN-FeR strain, resulting in mortality rate increments ranging from 20.6 to 39.2% across concentrations (Fig. 6G). Compared to the siGFP-treated group, circCPR25 expression knockdown significantly decreased the survival rate and body weight of adult female mites of the YN-FeR strain under high-temperature and low-humidity conditions (Fig. 6, H and I). Therefore, circCPR25 overexpression in the YN-FeR strain significantly down-regulated the miR-34~317 cluster, consequently regulating CPR25 overexpression and cuticle thickening and thereby mediating resistance to adverse conditions. Silencing the expression of circCPR25 in the YN-SS strain resulted in a significant reduction in the cuticle thickness of female mites, indicating that circCPR25 plays an important role in cuticle development in T. cinnabarinus (figs. S1, K and L, and S5, G to I).

Fig. 6. circCPR25 regulates CPR25 overexpression to promote cuticle thickening in female YN-FeR mites.

(A) Expression levels of circCPR25, miR-34~317 cluster, and CPR25 after feeding on siCirc for 72 hours. GFP, green fluorescent protein. (B) and (C) Cuticle ultrastructure of female adult mites treated with siGFP and siCirc, respectively. The number of the test: dsGFP, 15 females; and dsCPR25, 14 females. The red dashed box indicates the area (random) undergoing progressive magnification. The magnification ratios were ×2000 (2K), ×10,000 (10K), and ×20,000 (20K). The white dashed line delineates the boundary between the cuticle layer and the underlying tissues. (D) The protein expression levels of CPR25 after circCPR25 was silenced. The image of Western blotting contains three biological repetitions (siGFP1-3 and siCirc1-3). (E) Statistical analysis of total cuticle thickness in YN-FeR female adult mites after silencing circCPR25 expression. (F) Impact of pyridaben, bifenazate, and cyflumetofen on the cuticle penetration ability of YN-FeR female adult mites following circCPR25 knockdown. (G) Bioassays of pyridaben, bifenazate, and cyflumetofen on YN-FeR female adult mites after silencing circCPR25. The x axis represents the logarithmic scale of concentration: Log₁₀^{concentration}. Twenty-five females for each concentration treatment. (H) Survival rates of resistant female mites in the siGFP-treated group and the siCirc-treated group under high-temperature and low-humidity conditions. The detection time points were 24 and 48 hours, respectively. (I) Average weights of resistant female mites in the siGFP- and siCirc-treated groups at different time points under high-temperature and low-humidity conditions. Thirty females for each concentration treatment.

DISCUSSION

We assessed the effectiveness of various acaricides in controlling a highly fenpropathrin-resistant strain of T. cinnabarinus (YN-FeR). The YN-FeR strain demonstrated diverse levels of cross-resistance to these acaricides. Theoretically, cross-resistance should not occur considering the disparities in the target sites and compound structures of these chemicals (37, 38). Notably, when these chemicals were directly injected or fed into the bodies of resistant female mites, no differences were observed in mortality between the YN-FeR and YN-SS strains, indicating that broad-spectrum cross-resistance may be associated with the alteration of cuticular traits. Further, silencing the esterase and P450 genes, which has been previously linked to fenpropathrin detoxification (39, 40), did not mediate a significant difference in the mortality (oral activity) of the YN-FeR strain to pyridaben, bifenazate, and cyflumetofen (fig. S4, F to J). Therefore, we hypothesized that developmental differences occur in the cuticles of resistant and susceptible female mites. This conjecture was supported by a subsequent transmission electron microscopy investigation, and we found that both the total cuticle and its different components of YN-FeR female mites were thicker than those of the YN-SS and YN-KM female mites. Additionally, significant oral activity to spider mites has been observed in bifenazate (41), and the cross-resistance was not found in the oral activity of acaricides between YN-FeR and YN-SS. However, this result does not suggest that the significant decrease of acaricide contact activity caused by cuticle thickening should be disregarded. Because, according to Insecticide Resistance Action Committee (IRAC) statistical reports, more than 90% of commercially available pesticides exhibit contact activity, with spraying being the primary method of pesticide application in the field (5). This suggests that the contact activity of acaricides plays a critical role in controlling spider mites. Consequently, the efficacy of field-applied acaricides should be assessed in terms of both contact and oral activity. The results of spraying method bioassays, which assess both oral and contact activity, show that the YN-FeR strain retained cross-resistance to three acaricides. This supports the hypothesis that cuticle thickening in spider mites significantly diminishes acaricides efficacy.

Actually, cuticle thickening is an advantageous strategy used by arthropods to resist diverse adverse conditions and facilitate adaptive evolution, including xenobiotics stress (17). Rotating the use of different chemicals in agricultural production is an important strategy for managing pesticide resistance, especially target resistance in insects and mites (42). However, the cuticle serves as a physical barrier to chemical penetration without specificity, thickening, or alterations in its composition, and it is likely to result in multiresistance development in insects or mites (17). Our results strongly support this viewpoint, as the cuticle penetration efficiency of fenpropathrin significantly decreased in female YN-FeR mites. This phenomenon was also observed for other acaricides, resulting in broad-spectrum cross-resistance. In addition, significant cuticle thickening considerably prolongs the time required for pesticides to penetrate the bodies of spider mites, potentially facilitating the increased production of detoxification metabolic enzymes. A synergistic effect between cuticle thickening and enhanced detoxification metabolism has been shown in insects (43, 44). The maintenance of moisture is a pivotal function of pests’ cuticle (45). Our results reveal that phytophagous mites with thickened cuticles had an increased capacity to retain water and adapt within hot and arid environments, indicating that cuticle thickening plays a critical role in facilitating the rapid adaptation of fenpropathrin resistant mites to the progressively warming global climate. The increased resistance to chemicals penetration is a phenotypic change associated with cuticle thickening in female YN-FeR mites. The use of radiolabeled compounds is advantageous for analyzing the penetration efficiency of pesticides, a method that has been used in the study of penetration resistance in mosquitoes (23, 46). However, this method has not yet been reported in spider mites. In this study, we applied high-performance liquid chromatography (HPLC) to evaluate the penetration efficiency of various acaricides on T. cinnabarinus, and the feasibility of this method has been validated in both spider mites and insects (47–49). Additionally, bifenazate and cyflumetofen are pro-acaricides that undergo conversion by esterases in mites. Cyflumetofen is transformed into AB-1 (38), whereas the final metabolic product of bifenazate in mites remains unidentified (50). Nevertheless, no significant expression differences have been observed in the esterases genes involved in the conversion process of cyflumetofen between the susceptible and fenpropathrin-resistant female mites (51, 52). Further, silencing four differentially expressed esterase genes in the YN-FeR strain did not lead to the significant change in the sensitivity of female adult mites to acaricides. These results suggest that the conversion efficiency of bifenazate and cyflumetofen is likely similar in both strains, thus eliminating the potential impact of conversion efficiency on the differences in acaricide content within spider mites. In addition, cuticle thickening has been frequently observed in pyrethroid resistance cases (17). Although the direct correlation between pyrethroid application and arthropod cuticle thickening requires further investigation, our results indicate that cuticle thickening presents challenges for phytophagous arthropod management.

CPs constitute the principal components of the arthropod body wall (22), and fluctuations in their expression levels frequently dictate variations in cuticle thickness. For example, BdCPCFC overexpression can increase the CP content in B. dorsalis, leading to thicker cuticle development (47). Conversely, CPs expression down-regulation results in cuticle thinning in Nilaparvata lugens, leading to aberrant body development (26). In the present study, 44 CP genes were identified in T. cinnabarinus, with 30 from the CPR family and 14 from the CPAP family. Combining information on CPs from other arthropods, we observed a significant difference in both the quantity and diversity of CP genes between spider mites (4) and model insects, such as Drosophila melanogaster (53, 54), Bombyx mori (55), and Tribolium castaneum (56). This variance likely stems from the considerable evolutionary distinctions between spider mites and insects. For example, spider mites do not develop wings or antennae, reducing the number of CPs. We found that several CPs were weakly overexpressed in the YN-FeR strain. However, CPR25 displayed 18.6-fold up-regulation, suggesting its important function in cuticle development in fenpropathrin-resistant mites. CPR25 knockdown resulted in a reduction in cuticle thickness in resistant mites. Consequently, there was a decline in the resistance phenotype, characterized by a significantly decreased cross-resistance level against various acaricides and reduced adaptability to extreme environmental conditions. These results indicate that CPR25 overexpression is pivotal in mediating genetic variation in mite cuticle development.

As an important trans-regulatory factor, miRNA-driven differential gene expression has been shown to mediate genetic variations in arthropod development (57). We identified the miR-34~317 cluster in the T. cinnabarinus genome using RNA-seq, highlighting the potential uniqueness of the miRNA cluster, as it seems to have originated from two known miRNA families. The miR-34~317 cluster showed contrasting expression patterns with CPR25 in the YN-SS and YN-FeR strains. The negative regulatory role of the miR-34~317 cluster on CPR25 expression was demonstrated in both in vitro and in vivo experiments. Increased levels of either miR-34 or miR-317 suppressed CPR25 expression, consequently hindering cuticle development in resistant female mites. MiR-34 and miR-317 demonstrated negative regulatory roles in the insect development process including neurodevelopment (58), molting (59), cuticle development (60), wing development (61), and ovarian development (62). MiR-34 and miR-317 are conserved between spider mites and insects. However, the co-regulation of gene expression by miR-34 and miR-317 through miRNA cluster formation has not been observed in insects. Given that miR-34 and miR-317 serve multiple functions in insects, the other potential target genes of miR-34~317 cluster were screened in T. cinnabarinus. We found that the expression of vitellogenin receptor could also be regulated by miR-34 (fig. S7F). The mechanism through which vitellogenin receptor overexpression enhances fertility in fenpropathrin-resistant mites has been clarified in T. cinnabarinus (63). We hypothesized that miR-34 negatively regulates the enhanced reproductive capacity of female YN-FeR mites. Although several potential target genes of the miR-34~317 cluster were found, the regulation of cuticle development was mainly attributed to the reduced repression of CPR25 expression. In YN-SS strains, feeding on either miR-34 or miR-317 mimics impeded the development of second nymphs into adults, with 16 to 18% of individuals exhibiting an eclosion failure phenotype. However, in the YN-FeR strains, feeding on miR-34 or miR-317 mimics did not affect female mite development (fig. S10), suggesting the presence of a mechanism in resistant female mites to counteract the inhibitory effects of the miR-34~317 cluster on cuticle development.

We performed a comprehensive analysis integrating RNA-seq databases with CPR25 DNA sequence information, uncovering the generation of a daughter RNA, circCPR25, during CPR25 transcription. The generation of additional circular transcripts during the transcription of protein-coding genes is not an uncommon event. Studies have shown that during the splicing of precursor mRNA, the 5′ and 3′ ends of certain elements covalently bond, forming a closed circular structure through back-splicing, referred to as circRNA (64). These additional transcripts may include exons or combinations of exons and introns derived from paternal genes. This phenomenon has been documented in both mammals and arthropods (65, 66). The regulatory roles of circRNAs in the expression of parental genes have been clarified in mammals, mainly exhibiting trans-regulatory effects, including antagonizing miRNA in the cytoplasm or recruiting transcription factors in the nucleus (67). In this study, the CPR25 and circCPR25 expression patterns showed a strong correlation. The circCPR25 sequences contained binding elements of the miR-34~317 cluster, suggesting that circCPR25 adsorbed the miRNA clusters. The interaction between circCPR25 and the miR-34~317 cluster was systematically demonstrated in this study. Specifically, in HEK293T cells, miR-34 and miR-317 transfection substantially reduced the fluorescence activity of CPR25. However, with circCPR25 overexpression, the inhibitory effect of miR-34 and miR-317 on CPR25 fluorescence activity was abolished, indicating that circCPR25 regulates CPR25 overexpression by inhibiting the function of the miR-34~317 cluster. Most circRNAs become functional soon after they are transcribed (34), and their role as upstream regulators of protein-coding genes has been established (68). Therefore, circCPR25 may serve as a decoy RNA generated during CPR25 transcription. circCPR25 interacted with the miR-34~317 cluster, up-regulating CPR25 at the posttranscriptional level. Whether considering the evolution of resistance in arthropods or aspects of their developmental processes, certain cis- or trans-regulatory mechanisms have been clarified (69–71). The occurrence of CPs generating circRNA to self-regulate its overexpression remains uncommon, suggesting that the molecular basis of trait variations in arthropods in response to environmental stress exhibits significant diversity. Knockdown of the decoy RNA (circCPR25) induced significant up-regulation in miR-34 and miR-317 levels, accompanied by a marked decrease in CPR25 expression. Consequently, the cuticle thickness of YN-FeR female mites and their penetration resistance to acaricides approached the levels of YN-SS. These results suggest that circCPR25 antagonizes miR-34~317 cluster expression, potentially serving as a critical factor in miR-34~317 cluster down-regulation and cuticle development acceleration in the YN-FeR strain. In addition, our previous study identified circRNA in T. cinnabarinus and revealed that a circRNA (circ1-3p) formed by the splicing of intergenic regions can regulate the expression of TcGSTm04 (36). Notably, most circRNAs in mites and insects are generated during the transcription of protein-coding genes (36, 65). However, it remains unclear whether their production impacts the expression of the parental genes. In this study, we found that the remarkable overexpression of CPR25 was driven by the circCPR25 produced during its transcription, which functions as a decoy, allowing CPR25 to evade attacks from the miR-34~317 cluster. This unique regulatory mechanism mediates cuticle thickening and confers rapid adaptive evolution to spider mites, an R-strategy pest, in response to adverse environmental conditions. Furthermore, despite limited functional studies on circRNA in arthropods, recent studies have indicated that circRNA plays a notable regulatory role in the resistance evolution to adversity in insects and spider mites. For example, some circRNAs showed considerable changes in expression levels in response to stress from adverse conditions, suggesting that circRNA may actively participate in the adaptive evolutionary processes of arthropods in the face of adversity (36, 72). A comprehensive analysis of the functions of circRNA in arthropods may uncover novel mechanisms of pesticide-resistance evolution and identify new control targets.

Admittedly, there were some limitations to this study. We clarified a unique posttranscriptional regulatory mechanism through which CP regulates its own overexpression via the production of a circRNA. However, this is not the entire mechanism underlying genetic-variation cuticle development in female YN-FeR-strain mites. Changes in circCPR25 expression are closely associated with variations in cuticle thickness in T. cinnabarinus. Overexpression of circCPR25 in female mites of YN-FeR promotes CPR25 mRNA translation and cuticle thickening. However, the mechanisms underlying circCPR25 overexpression remain unexplored. During CPR25 mRNA transcription, circCPR25 is simultaneously produced. Therefore, we hypothesized that significant circCPR25 overexpression in the YN-FeR strain results from enhanced CPR25 DNA transcription, thus implying the presence of a factor, such as transcription factor or nuclear receptor, which promotes CPR25 DNA transcription. The directional regulation of DNA transcription by trans-regulatory factors has been demonstrated in arthropods (30). Nevertheless, the mechanisms underlying the transcriptional enhancement of CPR25 DNA require further investigation.

Overall, using a phytophagous mite as a model, we identified a genetic link between the evolution of pyrethroid resistance and variation in cuticle development. Furthermore, we demonstrated the presence of a distinctive molecular regulatory mechanism in arthropods wherein CPs produced a decoy circRNA to regulate its own overexpression (Fig. 7). This revelation may contribute to elucidating the molecular basis of arthropod evolution toward resilience to adverse conditions.

Fig. 7. Schematic diagram of the mechanism in which circCPR25, as a decoy, regulates CPR25 overexpression, mediating cuticle thickening in spider mite.

The increased levels of mRNA, circRNA, and protein in fenpropathrin-resistant females indicated their overexpression. The dashed line signifies mechanisms that remain to be elucidated. This schematic diagram was created in BioRender: L. He (2025), https://BioRender.com/x86h794.

MATERIALS AND METHODS

Mites

The YN-SS strain was collected from fields in the Kunming, Yunnan, China, in 2015. The field strain (YN-KM) was also collected in the Kunming, Yunnan, China, in 2024 (table S1). These mites have since been reared in the laboratory under the following conditions: 26° ± 1°C, 70 to 75% relative humidity, 14:10-hour light:dark cycle, and no pesticide exposure. Following the collection of the YN-KM strain, female mites with uniform developmental stages were cultured immediately, after which biological assays and analyses of cuticle ultrastructure were performed. There was no significant difference in the sensitivity to acaricides between the YN-SS and YN-KM strains (Table 1).

The fenpropathrin-resistant strain (YN-FeR) was selected from YN-SS by continuous selection with fenpropathrin. After 16 generations of selection, the resistance ratio of YN-FeR has reached 122.3-fold. This strain has since undergone selection every two generations to maintain its resistance level, with the most recent measurement indicating a resistance level was over 119.7-fold (Table 1, table S2, and fig. S12).

Acaricide penetration assay

The detection methods for acaricides’ penetration have been reported in our previous study (48). A total of 600 5-day-old female adult mites were collected for testing, sprayed with acaricides [pyridaben (200 mg/liter), bifenazate (200 mg/liter), and cyflumetofen (10 mg/liter)], and transferred to a new centrifuge tube for 4 hours. The surface of these female mites was washed six times with acetone, and, then, they were collected and homogenized in 1 ml of phosphate-buffered saline (PBS; 0.01 M, pH 6.5). Subsequently, 10 ml of ethyl acetate was used to extract acaricides from the mites’ bodies, and the extraction process was repeated twice. The solution was evaporated and diluted to 1 ml with acetonitrile. The acaricide content entering the body and remaining on the surface was determined using HPLC (Agilent, USA). This experiment included three biological replicates. Fenpropathrin, bifenazate, pyridaben, and cyflumetofen were separated using a C18 reverse-phase analytical column (4.6 mm by 250 mm, 5 μm; Agilent Technologies, USA). The samples were analyzed under an isocratic elution program, consisting of 90% A:10% B for fenpropathrin and pyridaben, 60% C:40% B for bifenazate, and 80% A:20% B for cyflumetofen (where A means methanol, B means 0.1% acetic acid in water, and C means acetonitrile). The elution signals were monitored at 210 nm for fenpropathrin, 230 nm for bifenazate, 270 nm for pyridaben, and 230 nm for cyflumetofen. The penetration ratio was calculated as follows: penetration ratio = [A/(A + B)] × 100%. A is acaricides dose in the body and B is residual acaricides in the cuticle.

Transmission electron microscopy detection

To ensure developmental consistency, we first selected female adult mites at the same age. Briefly, 200 female adult mites were placed on each fresh leaf discs, allowed to lay eggs for 6 hours, and then removed. The eggs were cultured until they reached 5-day-old female mites. A total of 30 5-day-old female mites were collected and fixed in 4% glutaraldehyde solution for 36 hours. These mites were washed three times with 0.01 PBS (0.01 M, pH 7.4) and then fixed in 1% osmium tetroxide for 1 hour at room temperature. After three washes with PBS, the mites were dehydrated with 30-50-70-90-100% ethanol, respectively, and treated with 100% acetone for 1 hour. The treated female mites were embedded in Spurr resin (VCD:DER-736:NSA:DMAE = 25:20:65:1, mass/mass). The resin undergoes polymerization in an oven at temperatures of 40°C (12 hours), 50°C (12 hours),and 60°C (12 hours), respectively, followed by an additional polymerization period at 70°C for 24 hours. Samples were sliced using an ultramicrotome (Leica, Germany) to a thickness of 70 nm, placed on copper grids (200-mesh), stained with 2% uranyl acetate, washed three times with ultrapure water, and air-dried for 12 hours. Cuticle thickness was measured using transmission electron microscopy (HITACHI, Japan) at an operating voltage of 80 kV, and each treatment group contained at least 10 female mites. The samples were initially magnified ×2000 to assess the consistency of cuticle thickness. Subsequently, specific areas were randomly selected for further magnification (×10,000 and ×20,000) and measurement of ridge, procuticle, and total cuticle thickness.

RNA extraction, reverse transcription, and qPCR

RNA was extracted from 200 female mites using TRIzol (Thermo Fisher Scientific, USA), following the manufacturer’s instructions. Subsequently, 1 μg of RNA, cleared of genomic contamination, underwent reverse transcription using the PrimeScript RT Reagent 037a Kit. The primers for the candidate genes were designed using Primer-BLAST (www.ncbi.nlm.nih.gov/tools/primer-blast). They exhibited an amplification efficiency ranging from 90 to 110%, and a single peak in the melting curve was deemed suitable. The primer information is listed in table S3. Genes encoding the ribosomal protein, S18 (RP18s, FJ608659), and tubulin alpha-I chain (α-Tub, FJ526336) were used as reference genes. A QuantiNova SYBR Green PCR Kit (QIAGEN, Germany) was used for qPCR detection, and the reaction system was the same as in previous studies (31). The expression levels of the candidate genes were normalized using the expression levels of the two reference genes, respectively. Gene expression levels were calculated using the comparative threshold cycle (2^−∆∆Ct) method. The experiment was repeated three times.

The reverse transcription of miR-34 and miR-317 was conducted using the stem-loop method. The primers for reverse transcription comprised a 5′ stem-loop universal sequence (CTCAACTGGTGTCGTGGAGTCGGCAATTCAGTTGAG) and an 8-nucleotide (nt) miRNA-specific sequence. The primer information is listed in table S3. miRNA reverse transcription was performed using the PrimeScript RT Reagent 037a Kit (TaKaRa, Japan), with a 20-μl reaction system containing 1 μg of RNA, 2 μl of stem-loop primer, and 1 μl of downstream primers for RP18s and α-Tub. The QuantiNova SYBR Green PCR Kit (QIAGEN, Germany) was used to detect the expression levels of miRNA. RP18s and α-Tub were used as reference genes. The method for calculating miRNA expression levels is same with protein-coding genes. The experiment was repeated three times.

circRNA reverse transcription was performed using the PrimeScript RT Reagent 037a Kit (TaKaRa, Japan). Due to the absence of a poly A tail in circRNA, its reverse transcription is conducted exclusively using random primers (random 6-nt oligomer) (36). A QuantiNova SYBR Green PCR Kit (QIAGEN, Germany) was used for qPCR detection, and the reaction system was the same as in previous studies (36). To ensure specificity, the qPCR primers for circRNA were designed to be the divergent primers. RP18s and α-Tub were used as reference genes. The method for calculating miRNA expression levels is same with protein-coding genes. The experiment was repeated three times.

cDNA library and RNA-seq

Total RNA of the mite strains was extracted according to the instruction manual for the TRlzol reagent, and there were three biological replicates for each strain. The RNA integrity and concentration were checked using an Agilent 2100 Bioanalyzer (Agilent, USA). The RNA samples were treated with Epicentre Ribo-Zero to remove ribosomal RNA, followed by RNA fragmentation. First-strand cDNA was synthesized using random hexamer primers. Second-strand cDNA was synthesized using RNase H and DNA polymerase I, and the cDNA was subsequently purified using magnetic beads (AMPure XP beads). The cDNA ends were repaired, and the cDNA was PCR amplified to construct the cDNA library. RNA-seq was conducted using the Illumina NovaSeq 6000 platform. Tophat2 software was used to map the clean reads to the reference genome (4). BLASTX software was then used to compare all gene sequences against the Nucleotide, Gene Ontology, Swiss-Prot, Non-redundant, Kyoto Encyclopedia of Genes and Genomes, Clusters of Orthologous Genes, and Eukaryotic Orthologous Groups databases to obtain gene annotation information. According to the identified CP data of spider mites (4), the potential CPs obtained from sequencing were further screened and annotated.

A Small RNA Sample Prep Kit was used to construct a miRNA library. Starting with 1.5 μg of RNA, we adjusted the sample volume to a total of 6 μl with RNase-free water. T4 RNA ligase 1 and T4 RNA ligase 2 were used to ligate adapters to the 3′ and 5′ ends of small RNAs, respectively. Reverse transcription was conducted to synthesize cDNA, followed by PCR amplification. Last, gel separation was used to screen the target fragments, and gel extraction was performed to complete the library preparation. The constructed library was sequenced using the HiSeq Xten platform, with a single-end read length of 50 nt. The sample of RNA-seq contains three biological replicates.

The method of circRNA sequencing was reported in our previous study (36). Briefly, BWA software was used to align the reads to the reference genome. Fully aligned reads were removed, and the remaining reads were then anchored with a 20–base pair base sequence for genome alignment. When two anchor sequences were linearly opposite in the genome (4), the reads were amplified until a sequence convergence point was identified. When the convergence point contained GU/AG splicing sites, the reads were considered potential circRNAs. The sample of RNA-seq contains three biological replicates.

Fragments per kilobase per million mapped reads (FPKM) served as the metric for quantifying the expression levels of protein-coding genes, while transcripts per million (TPM) was used to calculate the expression levels of circRNAs and miRNAs. Genes were not further analyzed if their expression was minimal in each sample (FPKM <1 or TPM < 1). A comparative analysis was performed on the normalized sequencing data from YN-FeR and YN-SS, using |log₂ (fold change)|≥ 1 and P value < 0.05 or false discovery rate < 0.05 as the criteria for identifying differentially expressed genes.

RNA interference

We first selected female adult mites of the same age to ensure developmental consistency. In brief, 200 female adult mites were placed on every fresh leaf discs, allowed to lay eggs for 6 hours, and, then, the adult mites were removed. The eggs were cultured until the final ecdysis, at which point the female mites were considered to be 0 days old as adults and were subjected to RNAi treatment (fig. S4K). Specific double-stranded RNA (dsRNA) primers were designed on the basis of CPR25 sequence information, and the target fragment was amplified via PCR. In vitro dsRNA synthesis was performed using a TranscriptAid T7 High Yield Transcription Kit (Thermo Fisher Scientific, USA). The dsRNA was diluted to 1.5 μg/μl with RNase-free water. The dsRNA was delivered through leaf feeding (73). Fresh cowpea leaves were trimmed to 1.5 cm by 1.5 cm, dehydrated at 60°C for 90 s, and immersed in a 30-μl dsRNA solution. dsRNA targeting the GFP was used as a control. The test mites were starved for 24 hours in a 1.5-ml centrifuge tube before transferring them to the treated cowpea leaves. After 72 hours, the mites were harvested for subsequent experiments. The experiment was repeated three times.

The circRNA interference strategies was referred to the mammal’s study (74), and we designed three candidate siRNAs (siCirc) targeting the back-splicing junction of circCPR25, along with corresponding siGFP controls. To ensure specificity, we selected the most similar sequences in the CPR25 gene, based on the siCirc sequence information, to design a control siRNA (siCK). siRNAs were synthesized in vitro using a T7 RiboMAX Express RNAi System kit (Promega, USA). These siRNAs were fed to female mites for 72 hours, and the expression levels of circCPR25 and CPR25 were measured to assess the silencing efficiency and specificity of siCirc. Only siCirc with high specificity and efficient silencing (>40%) was retained for subsequent experiments. The experiment was repeated three times.

Western blotting

Specific peptide segment (YKPESSYAPKPYSAP) was selected from the amino acid sequence of CPR25 to serve as antigens for the synthesis of polyclonal antibodies, with rabbits being used as the immunization host. The synthesis of the peptide segment, immunization, and antibody purification was outsourced to PTMBIO Biotechnology Co. Ltd. (Hangzhou, China). The α-tubulin antibody and the secondary antibody were purchased from Bioworld Biotechnology Co. Ltd. (Nanjing, China). One hundred female adult mites were collected for testing, rapidly frozen in liquid nitrogen, and homogenized in 200 μl of PBS (0.01 M, pH 7.4) containing 1 mM phenylmethylsulfonyl fluoride and 1 mM EDTA. The mixture was centrifuged at 4°C for 10 min at 18,000g, and the supernatant was collected. The concentration of the supernatant was measured using a bicin-choninic acid (BCA) protein concentration assay kit (Biomed, China). Next, 10 μl of the total protein solution was mixed with loading buffer and subsequently treated in a water bath at 100°C for 10 min. Total proteins were first separated by 10% SDD–polyacrylamide gel electrophoresis and subsequently transferred onto 0.45-μm-thick polyvinylidene fluoride membranes (Millipore, USA). After that, membranes were blocked with SuperBlot Protein-free Rapid Blocking Buffer (ECOTOP Scientific, China) for 25 min and subsequently incubated with the primary antibody (dilution rate, 1:2000) overnight at 4°C. Then, membranes were washed with tris-buffered saline containing Tween 20 and incubated with a secondary antibody (dilution rate, 1:10,000) for 1 hour at room temperature. Last, specific protein bands were analyzed using enhanced chemiluminescence (Coolaber, China). The experiment was repeated three times.

Determination of adaptation under high-temperature and low-humidity conditions

At 5-day-old, female mites were allocated to 5-ml centrifuge tubes, with 30 mites per tube, and sealed with 1000-mesh gauze. These tubes were placed in an incubator with conditions set to 34°C with 40% relative humidity for the treatment group and 26°C with 75% relative humidity for the control group. Survival rates were assessed at 24 and 48 hours for both groups. At 0, 6, 12, 24, and 48 hours, the female mites were weighed using an ultraprecision electronic balance (METTLER TOLED, Switzerland). This experiment was performed with three biological replicates.

Prediction of miRNA binding sites and miRNA mimic delivery

Using RNAhybrid, RNA22, and PITA, we predicted possible binding interactions between miRNAs and CPR25/circCPR25 based on their homologous sequence regions. The minimum free energy was set to less than −20 kcal/mol. The primary focus was on known miRNAs that were significantly down-regulated in the YN-FeR strain.

MiR-34, miR-317 mimic, and negative mimic were synthesized by RiboBio Co. Ltd. (RiboBio, China). These mimics were diluted to 40 μM using RNase-free water. The leaf feeding method was used to administer the mimics to the test mites, following a procedure identical to that of dsRNA delivery. This experiment was performed with three biological replicates.

Dual-luciferase reporter system assay

A dual-luciferase reporter assay was performed to analyze the in vitro interaction between circCPR25, miR-34~317cluster, and CPR25. The pmirGLO reporter vectors of circCPR25 and CPR25 were artificially synthesized by TsingKe Biological Technology Co. Ltd. (TsingKe, China). We used a specific circRNA overexpression vector (pcircRNA-1) to express circCPR25 in the cells, and the vector was artificially synthesized by BersinBio Biological Technology Co. Ltd. (BersinBio, China). HEK293T was selected as the tool cell line to perform the dual-luciferase reporter assay. The cell culture conditions followed those reported in our previous study (31). Cells were cultured in a 96-well plate to a density of 1 × 10⁵. Transfections with miRNA mimics and plasmids were performed using TransIT-LT1 (MirusBio, USA), with various transfection combinations established. After 48 hours, 75 μl of Dual-Glo Luciferase Reagent (Promega, USA) was added to the test cells and incubated at room temperature for 15 min. The intensity of firefly luciferase was recorded using a luminometer. Subsequently, 75 μl of Dual-Glo Stop & Glo Reagent (Promega, USA) was added, and samples were incubated for another 15 min at room temperature. The intensity of Renilla luciferase was recorded. The relative luciferase activity was calculated as the ratio of Firefly/Renilla luciferase intensities. This experiment included five replicates.

PCR identification of circRNA

To confirm that circCPR25 had a circular structure, a divergent primer and convergent primer were designed and amplified in cDNA and gDNA, respectively. The detailed method has been reported in our previous study (36). Briefly, 10 μg of total RNA was treated with RNase R (Epicentre, USA) at 37°C for 30 min, and the treated RNA was purified using TRIzol reagent. Reverse transcription was performed using a PrimeScript RT Reagent Kit (TaKaRa, Japan) to synthesize cDNA. gDNA was extracted using the DNeasy Blood & Tissue Kit (QIAGEN, Germany) according to the manufacturer’s protocol. An I-5 2× High-Fidelity Master Mix (TSINGKE, China) was used to amplify the product of circCPR25 in cDNA and gDNA. PCR products were analyzed by 1% agarose gel electrophoresis and Sanger sequencing.

Subcellular localization

Subcellular localization of candidate genes was determined using a Cytoplasmic and Nuclear RNA Purification Kit (Norgen, Canada). The detailed procedures were reported in our previous study (31). Briefly, 15 mg of female mite were rapidly frozen in liquid nitrogen and homogenized in 200 μl of Separation Lysis Buffer. The supernatant and pellet were separated by cold centrifugation. The supernatant was treated with 200 μl of Buffer SK and 200 μl of ethanol and mixed thoroughly, and cytoplasmic RNA was purified using an adsorption column. The pellet was washed twice with 1× PBS and treated with 400 μl of Buffer SK and 200 μl of ethanol, and nuclear RNA was purified using an adsorption column. Reverse transcription was performed to synthesize cDNA using a PrimeScript RT Reagent Kit (TaKaRa, Japan). GAPDH and U6 were used as controls for cytoplasmic and nuclear RNA, respectively. qPCR was conducted for subcellular localization of circCPR25 in T. cinnabarinus cells. The experiment was repeated three times.

RNA pulldown

The biotin-avidin system is a common method of exploring RNA interaction. To investigate the interaction between circCPR25, the miR-34~317 cluster, and CPR25, the biotin-labeled miR-34 and miR-317 mimics were synthesized by RiboBio Co. Ltd. (China) (table S4) and delivered into female mites by feeding. After 72 hours, 50 mg of mimic-fed female mites were collected, rapidly frozen in liquid nitrogen, and homogenized in 550 μl of lysis buffer [20 mM tris-HCl (pH 7.5), 100 mM KCl, 5 mM MgCl₂, and 0.3% IGEPAL CA-630]. After cold centrifugation, 50 μl of the supernatant was used as input. Dyna beads (50 μl) were washed with solution A (0.1 M NaOH and 0.05 M NaCl), solution B (0.1 M NaCl and 70% ethanol), and lysis buffer and incubated with 500 μl of lysis buffer and 10 μl of yeast tRNA at 4°C for 2 hours. After washing the beads with lysis buffer, 500 μl of the supernatant was added and incubated overnight at 4°C. The beads were washed four times with lysis buffer, and RNA from the enrichment and input groups was purified using TRIzol reagent (Thermo Fisher Scientific, USA). Reverse transcription was performed to synthesize cDNA using a PrimeScript RT Reagent Kit (TaKaRa, Japan). RP18s and α-Tub were used as control genes, and qPCR was used to determine the enrichment rates of miR-34 and miR-317 for CPR25 and circCPR25. The experiment was repeated three times.

Biotin-labeled antisense probes were designed from the back-splicing region of circCPR25 using a scrambled sequence as a control probe (table S4), and they were synthesized by BersinBio Co. Ltd. circRNA pulldown was performed using an RNA Antisense Purification Kit (BersinBio, China). Briefly, 500 mg of female mites were collected, cross-linking with 1× PBS containing 1% formaldehyde for 20 min and then adding 4 ml of glycine (1.375 M) for an additional 5 min. The mites were homogenized in 10 ml of cold PBS, and the cell suspension was filtered using a 100-μm filter. To collect cells, samples were centrifuged, and 1.8 ml of cold lysis buffer, 18 μl of protease inhibitor, and 9 μl of RNase inhibitor were added. Then, they were lysed on ice for 10 min, and 9 μl of deoxyribonuclease (DNase) salt stock and 20 μl of DNase (20 U) were added to the cell lysate and incubated at 37°C for 10 min. Then, 18 μl of EDTA, 9 μl of EGTA, and 18 μl of dithiothreitol were added and mixed thoroughly. As the input, 200 μl of lysate was collected, and 50 μl of magnetic beads were washed with tris-HCl (10 mM, pH 7.5) and 1× hybridization buffer. Then, 800 μl of lysate was denatured at 65°C for 10 min, and the probe was denatured at 85°C for 3 min. The lysate was mixed with 100 pmol of the probe, hybridized at 37°C for 30 min, and denatured at 50°C for 5 min, and hybridization continued at 37°C for 2 hours. The mixture was transferred to a centrifuge tube containing magnetic beads and incubated at room temperature for 1 hour. The beads were washed three times with wash buffer, and RNA was purified from the probe and input groups using TRIzol reagent (Thermo Fisher Scientific, USA). Reverse transcription to synthesize cDNA was performed using the PrimeScript RT Reagent Kit (TaKaRa, Japan). Using RP18S, circ1-3p, and miR-133-5p as control genes, the enrichment rate of miR-34 and miR-317 was determined for circCPR25 by qPCR. The experiment was repeated three times.

Fluorescence in situ hybridization

To determine whether circCPR25 directly bound with the miR-34~317 cluster in T. cinnabarinus tissues, a Cy3-labeled circCPR25 probe, FAM-labeled miR-34 and miR-317 probes, and corresponding control probes were designed and synthesized by GefanBio Co. Ltd. (GefanBio, China). The probe sequences are listed in table S4. The signals of circCPR25 and miR-34/miR-317 were investigated using a fluorescent in situ hybridization kit (GefanBio, China). The detailed process was reported in our previous study (31). Briefly, the female mites were fixed in 4% paraformaldehyde at room temperature for 1 hour and embedded in optimal cutting temperature compound (OCT). A freezing microtome was used to prepare 25-μm tissue sections. The sections were treated with 0.25% hydrochloric acid at room temperature for 15 min, washed twice with diethyl pyrocarbonate water, and treated with proteinase K at 37°C for 20 min. The sections were washed with PBS and saline sodium citrate (SSC) (pH 7.5) and treated with hybridization buffer at 65°C for 1 hour. The probes were diluted in hybridization buffer (1:500, v/v), and the sections were covered, hybridized at 65°C for 48 hours, and washed sequentially with SSC (pH 7.5) and PBS. The signals were analyzed using a Zeiss LSM 780 confocal microscope (Carl Zeiss SAS, Germany).

Bioassay

The bioassay was performed using the RCV method (31). In summary, the acaricides were diluted to four to five gradient concentrations using acetone, with acetone serving only as the control. Then, 1 ml of the acaricide solution was transferred into a 2-ml centrifuge tube, thoroughly vortexed, and alternately placed upright and inverted for 15 min. The solution in the tubes was then discarded, and they were air dried for 24 hours in a fume hood. The female mites were subsequently transferred into centrifuge tubes and placed in an incubator. Mortality rates were assessed after 24 hours. Three biological replicates were performed for this experiment.

To investigate whether the differences in sensitivity to various acaricides between the YN-SS and YN-FeR strains of female mites were related to cuticle penetration, we conducted bioassays using an injection method. The acaricides were diluted to four gradient concentrations with dimethylsulfoxide, which served as the control. Five-day-old female mites were immobilized on an agarose base, and each mite was injected with 8 nl of the acaricide using a microinjection system. The treated mites were then transferred to fresh leaf discs, and mortality rates were calculated after 24 hours. Three biological replicates were performed for this experiment.

The oral activity of acaricides was assessed using a membrane device (75). In brief, a 35-mm petri dish served as the platform, with a Parafilm M placed on top, followed by a 500-mesh gauze (1.5 cm by 1.5 cm). The acaricides were diluted with a 1% Tween 80 solution to various concentrations, and 160 μl of the solution was applied to the gauze. A 1% Tween 80 solution devoid of acaricide was used as the control. The Parafilm M was then stretched six times and used to cover the solution. The liquid-filled area of the mesh sheet was surrounded with a wetted wiping paper. Fifteen female adult mites were selected and then placed in the device, and the mortality rate was assessed after 24 hours.

We performed bioassays using the spray method to further confirm the combined effects of contact and oral activity of acaricides. In brief, the acaricide was initially prepared as a stock solution using acetone and subsequently diluted with a 0.1% Tween 80 solution to establish four concentration gradients. Twenty-five 5-day-old female mites were introduced into fresh leaf discs, and 1 ml of the acaricide solution was sprayed onto them. A 0.1% Tween 80 solution devoid of acaricide was used as the control. After 24 hours, the mortality rate of the treated female mites was analyzed. Three biological replicates were performed for this experiment.

Statistical analysis

Statistical analyses were performed with SPSS 22.0. The results are presented as means ± SEM. Statistical differences were determined by independent sample t test or analysis of variance (ANOVA) followed by Tukey test. *P < 0.05; **P < 0.01; ***P < 0.001; ****P < 0.0001.

Supplementary Materials

The PDF file includes:

Figs. S1 to S12

Tables S1 to S4

Legend for data S1

Other Supplementary Material for this manuscript includes the following:

Data S1