http://orcid.org/0000-0001-7952-3573
Figures
Abstract
Wide and improper application of pyrethroid insecticides for mosquito control has resulted in widespread resistance in Aedes albopictus mosquitoes, an important dengue vector. Therefore, understanding the molecular regulation of insecticide resistance is urgently needed to provide a basis for developing novel resistance diagnostic methods and vector control approaches. We investigated the transcriptional profiles of deltamethrin-resistant and -susceptible Ae. albopictus by performing paired-end sequencing for RNA expression analysis. The analysis used 24 independent libraries constructed from 12 wild-caught resistant and 12 susceptible Ae. albopictus female adults. A total of 674,503,592 and 612,512,034 reads were obtained, mapped to the Ae. albopictus genome and assembled into 20,091 Ae. albopictus transcripts. A total of 1,130 significantly differentially expressed genes included 874 up-regulated genes and 256 down-regulated genes in the deltamethrin-resistant individuals. These differentially expressed genes code for cytochrome P450s, cuticle proteins, glutathione S-transferase, serine proteases, heat shock proteins, esterase, and others. We selected three highly differentially expressed candidate genes, CYP6A8 and two genes of unknown function (CCG013931 and CCG000656), to test the association between these 3 genes and deltamethrin resistance using RNAi through microinjection in adult mosquitoes and oral feeding in larval mosquitoes. We found that expression knockdown of these three genes caused significant changes in resistance. Further, we detected 1,162 single nucleotide polymorphisms (SNPs) with a frequency difference of more than 50%. Among them, 5 SNPs in 4 cytochrome P450 gene families were found to be significantly associated with resistance in a genotype-phenotype association study using independent field-collected mosquitoes of known resistance phenotypes. Altogether, a combination of novel individually based transcriptome profiling, RNAi, and genetic association study identified both differentially expressed genes and SNPs associated with pyrethroid resistance in Ae. albopictus mosquitoes, and laid a useful foundation for further studies on insecticide resistance mechanisms.
Author summary
Aedes albopictus represents a serious threat to public health as a vector for a number of arboviruses, such as dengue, Zika, and chikungunya viruses. Its ability to resist insecticides has recently been detected in several countries. However, little is known regarding molecular mechanisms that confer pyrethroid resistance. Determining such mechanisms is essential for early detection, monitoring, and management of insecticide resistance. Here, we use high-throughput RNA sequencing techniques, RNAi, and genotype-phenotype association study to characterize qualitative and quantitative differences in gene expression as well as single nucleotide polymorphisms between resistant and susceptible Ae. albopictus individuals collected from a dengue-endemic area in southern China. We identified 3 candidate genes and 5 candidate SNPs associated with deltamethrin resistance. These results enhance the understanding of insecticide resistance mechanisms among Aedes albopictus.
Citation: Xu J, Su X, Bonizzoni M, Zhong D, Li Y, Zhou G, et al. (2018) Comparative transcriptome analysis and RNA interference reveal CYP6A8 and SNPs related to pyrethroid resistance in Aedes albopictus. PLoS Negl Trop Dis 12(11): e0006828. https://doi.org/10.1371/journal.pntd.0006828
Editor: Geoffrey M. Attardo, University of California Davis, UNITED STATES
Received: March 2, 2018; Accepted: September 11, 2018; Published: November 12, 2018
Copyright: © 2018 Xu et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Data Availability: The datasets generated are available under NCBI BioProject number PRJNA475859 (Transcriptional profiles of deltamethrin resistant and susceptible Ae. albopictus).
Funding: This research was supported by a combination of funding from the: National Nature Science Foundation of China (81829004, 31830087, 81420108024) URL: http://www.nsfc.gov.cn; National Institutes of Health, USA (2R01AI136850-05A1) URL: https://www.nih.gov/grants-funding; Science Foundation of Guangdong Province (2014A030312016, 2016A020251001, 2017A030310399) URL:http://www.gdstc.gov.cn; Guangzhou Synergy Innovation Key Program for Health (201508020263) URL: www.gz.gov.cn. Postdoctoral Science Foundation (2016M602491) URL:http://jj.chinapostdoctor.org.cn; The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript.
Competing interests: The authors have declared that no competing interests exist.
Introduction
Arboviruses such as Zika, dengue, and Chikungunya by Aedes mosquitoes represent an expanding threat to global health [1–3]. Chemical control through the use of insecticides is one of the essential tools in the global strategy for mosquito-borne infectious disease control [4]. Pyrethroid is often the insecticide of choice due to its low mammalian toxicity and high insecticidal activity [4,5]. Permethrin, deltamethrin, cypermethrin, and cyfluthrin are currently the most commonly used pyrethroids for adult mosquito control in indoor residual spraying (IRS), insecticide-treated nets (ITN), and space spray treatment [6, 7]. For instance, ultra-low-volume (ULV) spray of pyrethroids has been a major measure for controlling Aedes adults in dengue-endemic areas in China [8]. However, excessive and improper application of pyrethroids for mosquito control, together with widespread agricultural usage, has resulted in a high selection pressure for pyrethroid resistance [9,10]. Insecticide resistance has become a major impediment to controlling mosquito-borne diseases worldwide. As a response to this dilemma, the World Health Organization (WHO) launched the Global Plan for Insecticide Resistance Management (GPIRM) to guide rational insecticide usage [11]. Insecticide resistance in the Asian tiger mosquito, Aedes albopictus, the major dengue vector, came to public attention in China with the dengue outbreak in 2014, when over 45,000 dengue cases were reported in Guangdong province [12]. Insecticide was applied extensively under the government-directed mosquito control program, and resistance has been reported in multiple Ae. albopictus populations in Guangdong [13]. Current resistance monitoring relies on bioassays with live mosquitoes. Understanding resistance mechanisms is essential to developing sensitive molecular monitoring tools of resistance and resistance management.
Resistance to DDT and pyrethroids has been documented in different mosquito species [10,14,15]. Two major resistance mechanisms have been previously identified, including target site insensitivity resulting from mutations in the insecticides’ target protein genes and increased metabolic detoxifications for insecticides [4]. Non-synonymous mutations in the voltage-gated sodium channel (VGSC) gene causing target-site insensitivity have been found in multiple mosquito species of public health importance, including Anopheles gambiae [16,17], Culex quinquefasciatus [15,18], Ae. aegypti [10,14], and Ae. albopictus [13]. In Ae. aegypti, V1016G, V1016I, and F1534C mutations were found to be correlated with pyrethroid resistance [19, 20], and the F1534S mutation was correlated with deltamethrin resistance in Ae albopictus [13]. Metabolic resistance involves the insecticide molecule’s bio-transformation via higher expression or presence of more efficient detoxification enzymes. Detoxification enzymes belong to four gene families: cytochrome P450 monooxygenases (P450s), carboxylesterases (COEs), glutathione S-transferases (GSTs), and UDP-glucuronosyltransferases, but a number of species-specific genes within each gene family have been found to be involved in metabolic resistance in various mosquito species, such as Anopheles sinensis, Cx. quinquefasciatus, and Ae. aegypti [10,14,15,17,21–24]. Other resistance mechanisms include mosquito physiological changes, such as cuticle thickening and digestive tract modification, which may lead to changes in insecticide absorption and penetration, and in behavioral avoidance [25, 26]. We have previously identified one mutation (F1534S) in the VGSC gene in Ae. albopictus that is significantly associated with pyrethroid resistance [13]. However, little is known about metabolic resistance in Ae. albopictus, which may result from both altered expression and mutations in the detoxification genes. For example, SNPs in GSTe2 were found to confer resistance to DDT in the malaria mosquito Anopheles funestus [27]. With this in mind, this study aims to characterize novel resistance mechanisms in Ae. albopictus.
mRNA sequencing (RNA-seq) is a powerful technique for comprehensive profiling of transcripts. To date, several studies have used RNA-seq to investigate resistance mechanisms in mosquito species, including An. gambiae [28], An. sinensis [29], Ae. aegypti [30], and Cx. quinquefasciatus [31] and Cx. pipiens pallens [32], and these studies have demonstrated that the RNA-seq technique can successfully yield useful resistance targets. In southern China, Ae. albopictus is the most important dengue vector with strong pyrethroid resistance [13]. In this context, the present study used RNA-seq on single mosquitoes and identified genes differentially expressed between resistant and susceptible Ae. albopictus mosquitoes and SNPs associated with resistance. In addition, the function of 3 candidate genes was further functionally verified by RNA interference (RNAi). Using a genotype-phenotype association study, we further identified 5 new SNPs significantly associated with resistance in natural populations.
Materials and methods
Ethics statement
This study did not involve endangered or protected species, and was not conducted on protected lands. All collections were performed on public land, and no specific permits were required for mosquito field specimen collection.
Ae. albopictus sample and transcriptome library preparation
Insecticide resistance bioassay.
The overall study design is shown in S1 Fig. RNA-seq used deltamethrin-resistant and -susceptible Ae. albopictus mosquitoes. Immature Ae. albopictus (larvae and pupae) samples were collected in September 2014 in the Longgang district, Shenzhen, China (22.7210° N, 114.2469° E), in public areas such as streets, parks, and parking lots, from more than 50 different aquatic habitats, including discarded plastic containers, flower pots, and used tires. The field-collected larvae/pupae were reared to adulthood at 28°C, 70–80% relative humidity, with 16:8h light-dark photoperiods in the insectary of Southern Medical University. After emerging, 3-5-day old female adults were identified to species by morphology before resistance bioassay using the standard WHO insecticide susceptibility tube test with 0.05% deltamethrin. The 0.05% deltamethrin and control filter papers were purchased from the WHO reference laboratory in the School of Biological Sciences, Universiti Sains Malaysia (Penang, Malaysia). Mosquitoes that were alive after the 24 h recovery period were classified as resistant (R). Susceptible ones (S) were defined as those knocked-down within the first 20 min of the standard WHO insecticide susceptibility tube test and not able to fly 3 h after being moved to a recovery tube [33, 34]. Mosquitoes were carefully examined to ensure that the inability to fly was not due to loss of legs or wings. Mosquitoes were then preserved in RNA-later (Ambion, Carlsbad, CA, USA). All bioassays were conducted under a controlled environment at 28°C and 70–80% relative humidity in the insectary of Southern Medical University. All specimens were further confirmed for species identity using PCR as previously described [13].
RNA-seq library construction.
RNA-seq library preparation and sequencing were carried out by HudsonAlpha Genomic Services Lab (Huntsville, AL, USA). A total of 24 cDNA libraries were prepared individually from 12 susceptible and 12 resistant mosquitoes. The main reason that we constructed libraries from individual mosquitoes was that this strategy enabled us to detect both transcriptome changes and SNPs associated with resistance using the RNA-seq data. This approach has advantages over the pooled mosquito approach, as the latter approach cannot detect SNPs with different frequencies between resistant and susceptible mosquitoes. Total RNA was extracted from 12 resistant and 12 susceptible female mosquito samples individually using Trizol reagent (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s protocol. Recombinant DNase I (Takara, Japan) was used to remove potential genomic DNA. RNA integrity (RNA Integrity Score 6.8) was evaluated using Nanodrop 2000 (Thermo Scientific, Delaware, ME, USA). RNA quality was analyzed on an Agilent 2100 Bio analyzer (Agilent, Palo Alto, CA, USA). Genomic DNA was extracted from mosquito legs using a Fast Tissue-to-PCR kit (Thermo fisher, Delaware, ME, USA) following manufacturer’s protocol.
The RNA-seq libraries were prepared with 500 μg of starting total RNA with an Illumina TruSeq RNA Sample Preparation Kit (Illumina, San Diego, CA), following the TruSeq protocol. The libraries were amplified with 15 cycles of PCR and contained TruSeq indexes within the adapters, specifically, indexes 1–4. Finally, amplified library yields were 30 μl of 19.8–21.4 ng/μl with an average length of ∼270 bp, indicating a concentration of 110–140 nM. KAPA Library Quantification Kits Illumina Platforms (Illumina, CA, US) were used for quantitation and dilution, and the libraries were sequenced on an Illumina HiSeq 2500 instrument with 100 bp paired end (PE) reads.
qRT-PCR validation of RNA-seq data.
We analyzed 11 randomly selected, differentially expressed genes by quantitative real-time reverse transcription PCR (qRT-PCR). Specifically, cDNA was prepared using SuperScript III (Invitrogen, MA, USA) and random primers from pooled RNA of 4 resistant or 4 susceptible mosquitoes, and three biological replicates were used. The qRT-PCR reactions were performed in a 7300 FAST Real-Time PCR System (ABI, Foster City, CA, USA) and analyzed using the rpS7 (ribosomal protein 7) gene as internal control [35]. All the qPCR primers used in this study are shown in S2 Table. Expression levels were then calculated against control mosquitoes as a calibrator using the 2-ΔΔCt method with CFX Manager Software (CFX Manager 3.1, BioRed, USA). The Pearson correlation coefficient was calculated between fold changes in transcript accumulation levels between resistant and susceptible mosquitoes, as obtained by qRT-PCR and RNA-seq, respectively.
RNAi analysis
RNAi analysis used a laboratory-selected deltamethrin-resistant population (Lab-DR strain). Briefly, larvae from a laboratory Foshan strain of Ae. albopictus were bioassayed to determine the 50% lethal concentration (LC50), following the WHO standard method [34]. This bioassay measured larval mortality after 24 hr of exposure to 5 concentrations of deltamethrin, each concentration in three replicates. The Lab-DR strain was selected from the larvae of the Foshan strain, using 0.001 mg/L of deltamethrin) as the starting concentration, and gradual increments of deltamethrin concentrations for 10 generations. The 10th generation selection used a 0.015 mg/L concentration of deltamethrin. The WHO insecticide susceptibility tube test on Lab-DR adults found an 89% mortality rate, indicating modest resistance.
RNAi analysis was conducted in both adult mosquitoes and larval mosquitoes with 3 highly differentially expressed genes. For adult mosquito RNAi, 3-day post-emergence Lab-DR mosquitoes was randomly selected and divided into control group or RNAi groups. For each of the three RNAi groups, 500ng siRNA duplex (siRNACYP6A8, siRNA01931, or siRNA00656) was microinjected into a female mosquito thorax (S2 Table). For the control group, a siRNA duplex lacking significant sequence homology to any genes in the Ae. aegypti genome, siRNA_C, was injected (S2 Table). All siRNAs were synthesized by IDT Technologies (Coralville, IA, USA). After the injection, mosquitoes were allowed to recover in a cup and maintained with 8% sucrose. An overall 16% mortality was observed 24 hr after the injection. Forty-eight hours after injection, gene knockdown efficiency was examined in 12 randomly selected individuals in each injection group using qRT-PCR, and its impact on mosquito resistance was determined using a standard WHO insecticide susceptibility tube test with 0.05% deltamethrin. For each gene, RNAi-treated female mosquitoes 48 hr after injection underwent the WHO insecticide susceptibility bioassay, with 20–25 mosquitoes per replicate for a total of 5 replicates per group [36]. Control group and RNAi groups were tested simultaneously.
For larval mosquito RNAi, we conducted oral RNAi following the method of Zhang et al. [37]. Briefly, each target gene siRNA and the siRNA_C described above (S2 Table) was mixed with Chitosan to form a Chitosan/siRNA nanoparticle. The nanoparticle was then mixed with cat food and agarose with green food dye, before feeding to 20 second-instar larvae in a petri dish with 100 ml water on a daily basis until the adult mosquitoes emerged. A total of 20 replicates with a total of 400 mosquito larvae were used for each target gene group and the control group. All larvae were successfully fed with the nanoparticles as evidenced by the green food dye in their guts. Three days after emergence, for each group 12 female adult mosquitoes were randomly selected to detect gene expression by qRT-PCR. Deltamethrin resistance bioassay was conducted using the standard WHO insecticide susceptibility tube test, with 20–25 mosquitoes per replicate for a total of 5 replicates per group.
Genotype-phenotype association study
P450-related SNP identification.
RNA-seq analysis identified 1,162 SNPs with a minimum frequency difference of 0.5 between the susceptible and resistant individuals. Because cytochrome P450 genes were considered most important to pyrethroid resistance, we aimed to identify potential SNPs associated with resistance within the P450 gene family. Cytochrome P450-related SNPs were selected using the following criteria: Mutation frequency differences between resistant and susceptible individuals that were studied by RNA-seq were at least 50%, and Fisher’s exact test P < 0.005 between the phenotyped individuals.
Genotype-phenotype association analysis.
To establish an association between SNP and phenotypic resistance in natural mosquito populations, we screened 1,350 female adults from Shenzhen, southern China, for deltamethrin resistance using the standard WHO insecticide susceptibility tube test, yielding 115 resistant mosquitoes. Randomly-selected resistant mosquitoes (n = 70) and susceptible ones (n = 70) were genotyped at 7 candidate genes to cover 9 resistance-associated SNPs identified through the above RNA-seq analysis. Primers used for SNP genotyping are shown in S2 Table. PCR products were purified with ExoSAP-IT (USB, Cleveland, Ohio, USA) according to the manufacturer’s protocol and directly sequenced by Genewiz Inc. (South Plainfield, NJ, USA).
Kdr mutation detection in Lab-DR strain adults and susceptible Foshan strain.
Randomly selected female adult mosquitoes (n = 30) from Lab-DR and from the susceptible Foshan strain (n = 30) were genotyped for kdr mutations at the L1534 locus. Portions of domains III of the VGSC gene were amplified, following protocols and primers developed by Kasai et al. [38]. PCR products were sequenced by TsingKe Inc. (Beijing, China). The sequences were analyzed using BioEdit (http://www.mbio.ncsu.edu/BioEdit/bioedit.html).
Data analysis
RNA-seq data analysis.
Sequenced reads were assigned to each sample, and adaptors were removed. Read quality for each sample was checked using FastQC. Reads with quality scores less than 20 and lengths below 30 bp were all trimmed. The resulting high-quality sequences were used in subsequent assembly. Clean reads were used as input in all the assemblers in CLC Genomics Workbench (version 8.5.1). To identify differentially expressed transcripts, reads from resistant and susceptible conditions were mapped to the Ae. albopictus reference genome, available from Vector Base database, using the transcriptomic analysis module, with the following parameters: mismatches = 2, minimum fraction length = 0.9, minimum fraction similarity = 0.8, and maximum hits per read = 10. Gene expression levels were normalized as reads per kilobase of exon model per million mapped reads (RPKM). Theoretical values were obtained from means of RPKM values for each group. Transcripts with absolute fold-change values > 2.0 and FDR corrected P-values < 0.05 were included in the gene ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways enrichment analyses. The differentially regulated genes were analyzed for gene ontology (GO) using the vector base Ae. albopictus assembly P3 database and categorized based on GO terms (level 2) for biological processes, molecular functions, and cellular components, as well as by genes up- or down-regulated in association with KEGG pathways.
SNP detection was processed for RNA-seq data obtained for individual resistant and susceptible mosquitoes with CLC Genomic Workbench, with the following parameters: Ignore positions with coverage > 100,000, ignore broken pairs, ignore non-specific match reads, with minimum read length = 20, minimum coverage = 10, minimum count = 2, minimum frequency = 35%, neighborhood radium = 5, minimum central quality = 20, minimum neighborhood quality = 15, relative read direction significance>1%, read position filter significance>1%. The site with substitution frequency above 5% was considered as a single nucleotide variant (iSNV). SNP_tool from MS-Excel was adopted to calculate the odds ratio (OR), chi-square analysis between resistant and susceptible individuals. Statistically significant SNPs were selected with OR > 2 and P < 0.005. We expect RNA-seq data to include population variations and to reflect differential expression of transcripts because the mosquitoes we used were collected from the field and our experimental design compared mosquitoes with two distinct phenotypes (resistant and susceptible).
RNAi data analysis.
A t-test was used to compare expression levels between experimental groups, and the χ2 test was used to compare 24 h mortality rates among different treatments. If mortality exceeded 20% in the control group, the replicate was excluded from the analysis. If mortality in the controls was < 20%, the mortality rate with the particular treatment was corrected using Abbott’s formula [39].
Genotype-phenotype association data analysis.
The sequences were aligned and analyzed using BioEdit (http://www.mbio.ncsu.edu/BioEdit/bioedit.html) and Codon code (http://www.codoncode.com/). The odds ratio was calculated to examine the association between SNP and resistance, and Fisher’s exact test was performed to determine the statistical significance.
Results
RNA-seq and quality control
Twenty-four cDNA libraries were constructed individually for deep sequencing on an Illumina-Solexa HiSeq 2500 platform. Raw read quality control was carried out via base composition analysis and base sequence quality analysis. A total of 674,503,592 and 612,512,034 paired-ends reads from resistant and susceptible conditions, respectively, were obtained. A summary of filtered paired-end reads is shown in Table 1. An average of 56,208,633 reads in resistant individuals and 51,042,670 in susceptible individuals were mapped to the Ae. albopictus reference genome. Gene coverage statistics for resistant and susceptible individuals showed that most of the genes (84% and 77%, respectively) reached coverage of 90–100%.
Gene ontology enrichment analysis and pathway enrichment analysis of DEGs
All unique matched reads from resistant and susceptible individuals were assembled into 20,091 transcripts in a total of 11,090 genes. From these genes, 1,130 genes were significantly differentially expressed genes (DEGs) when using the criteria of FPKM ratio of each phenotype >2 (in either direction), P < 0.0001, and FDR < 0.05. Among them, 874 genes (77.3%) were up-regulated and 256 genes (22.7%) were down-regulated in the resistant mosquitoes. These DEGs included P450s, cuticle proteins, UDP-glucuronosyltransferases, lipases, serine proteases, heat shock proteins, esterase, peptidases, ATP-binding cassette transporters, and others (Fig 1A).
Analysis was performed with Blast2GO-PRO. A) Histogram of Ae. albopictus differentially expressed sequences GO classification. Three main ontologies of GO (biological process, cellular component, and molecular function) are shown in the x-axis. The left y-axis indicates the percentage of total genes, and the right y-axis is the number of genes in each category. B) Distribution of transcripts in category of cellular components. C) Distribution of transcripts in category of molecular functions. D) Distribution of transcripts in category of biological processes.
All DEGs were classified into three main GO categories: biological processes, cellular components, and molecular functions (Fig 1). For cellular components, genes involved in cells (GO: 0005623, 249 out of 705 DEGs, or 35.3%) and cell parts (GO: 0044464, 249 out of 705 DEGs, 35.3%) were the most abundant (Fig 1B). As for molecular functions, binding (GO: 0005488, 435 out of 838 DEGs, 51.9%) and catalytic activity (GO: 0003824, 238 out of 838 DEGs, 28.4%) were the most highly represented categories (Fig 1C). For biological processes, the categories most represented were cellular processes (GO: 0009987, 294 out of 1,013 DEGs, or 29.0%) and metabolic processes (GO: 0008152, 256 out of 1,013 DEGs, 25.3%) (Fig 1D).
A total of 694 DEGs were mapped into 309 KEGG pathways. The largest category was metabolic pathways, containing 92 annotated DEGs (20.6%), followed by biosynthesis of secondary metabolites, with 36 DEGs (8.1%). In general, the most represented GO and KEGG categories were “metabolic process” and “metabolic pathway,” which were widely acknowledged to be involved in insecticide metabolism [40,41]. A total of 11 DEGs belonging to the P450 family or GST which were known to be related to insecticide detoxification are shown in Table 2. We did not find any significant DE genes related to metabolic inhibitors or cuticle protein genes.
SNP identification.
A total of 267,469 non-synonymous mutation sites were identified; among them, 1,418 polymorphisms differed significantly in frequencies between resistant and susceptible mosquitoes. These included 1,162 SNPs (81.9%), 162 multiple nucleotide polymorphisms (MNPs) (13.9%), 37 insertions (3.2%), and 39 deletions (3.4%). Among the 1,162 SNPs detected, 766 SNPs (54.1%) were mapped to the genes with annotation. Among the 766 SNPs with annotation, 9 SNPs belong to 7 P450 genes and 3 GSTs (Table 3). In the kdr gene, mutations from wildtype TTC (Phe) to TTG (Leu), or to TCC (Ser) at codon 1534 were detected, with an F1534S mutation frequency of 29.1% in resistant individuals and 16.7% in susceptible individuals.
Comparison of transcription level between RNA-seq and qRT-PCR
We used qRT-PCR to validate the expression level obtained from RNA-seq. As shown in Fig 2, a highly significant correlation was found between the two methods (R2 = 0.67), demonstrating the validity of the RNA-seq method for expression quantitation.
A) Fold change of expression value in the resistant individuals relative to the susceptible individuals, as measured by RNA-seq and qRT-PCR in 11 selected genes. B) Pearson correlation between fold changes in gene expression between resistant and susceptible mosquitoes, as determined by qRT-PCR and RNA-seq (P<0.05).
RNAi analysis indicated CYP6A8 (CCG018948), CCG013931.2, and CCG000656.1 played roles in deltamethrin resistance of Ae. albopictus
CYP6A8 (CCG018948), CCG013931.2, and CCG000656.1 were selected for RNAi analysis based on expression differences between resistant and susceptible mosquitoes. The RNA-seq analysis indicated that the expression of CYP6A8 was 3.1-fold higher in the resistant samples than in the susceptible ones, and CCG013931.2 and CCG000656.1 were 13.7 and 5.5-fold higher as evidenced from the (Fig 2). For microinjection RNAi in adult mosquitoes, the CYP6A8 expression level was reduced by 52.8% by siRNACYP6A8, compared with the control (P < 0.01) (Table 4). Similarly, CCG013931.2 expression level was reduced by 66.3% (P < 0.01), and by 71.4% for CCG000656.1 (P < 0.01). The resistance bioassay found 100% mortality in RNAi treated groups for the three genes whereas a 90.1% mortality rate was found in the control group (P < 0.01) (S3 Table). The effect of adult mosquito RNAi using the microinjection method on mosquito knockdown time is shown in S2 Fig. Mosquitoes injected with siRNA01931 exhibited significantly shorter KDT50 time (S3 Table).
Microinjection and oral delivery RNAi were used.
Oral feeding of siRNA nanoparticles in larvae indicated a similar gene expression knockdown effect compared to the microinjection group. The CYP6A8 expression level was reduced by 51.9% compared with the control (P < 0.01), by 53.5% for CCG013931.2 (P < 0.01), and by 48.0% for CCG000656.1 (P < 0.01) (Table 4). Gene expression knockdown corresponded to increased mortality rate in the RNAi-treated groups. Mortality rate was 90.5% in the control group whereas the RNAi-treated groups showed 100% mortality rate (P < 0.01) (S4 Table). The mosquito knockdown time in insecticide resistance bioassay found a significantly shorter average knockdown time in RNAi treated groups than the control group (S2 Fig).
Frequency of kdr mutations in Lab-DR resistant strain and susceptible Foshan strain
Non-synonymous mutation was detected at codon 1534 in the selected Lab-DR resistant strain adults but not in the susceptible Foshan strain. At codon 1534, a change from wildtype codon TTC (Phe) to TCC (Ser) was detected in 14 out of the 30 samples, of which all were heterozygotes. The L1534S mutation frequency was 23.3% in the selected Lab-DR resistance strain.
Association between SNPs in P450 gene family and pyrethroid resistance in natural mosquito population
We examined 9 SNPs within the 7 P450 genes identified from the RNA-seq analysis to determine SNP genotypes’ association with deltamethrin resistance using 70 randomly selected resistant mosquitoes and 70 susceptible mosquitoes (Table 5). Among the 9 SNPs examined, 5 (Arg226Ser in CYP6A8, Pro175Gln and His877Tyr in CYP9B2, Met120Ile in CYP6A8, and Cys212Ser in CYP1A1) showed significant associations with resistance.
Discussion
In the present study, we used the RNA-seq technique to study the transcriptional profile of deltamethrin-resistant and -susceptible Ae. albopictus mosquitoes. The individual mosquito-based sampling strategy used in this study is distinct from the commonly used pooled sampling strategy, and this new strategy enables us to simultaneously identify differentially expressed genes and differential SNPs that may be associated with pyrethroid resistance. We identified a total of 1,130 differentially expressed genes, including 874 up-regulated and 256 down-regulated genes. A total of 1,162 SNPs with large frequency differences between resistant and susceptible mosquitoes was identified. Using genotype-phenotype association analysis of natural mosquito populations, we identified 5 SNPs in the P450 gene family that are significantly associated with resistance. RNAi in adult mosquitoes through microinjection and larval mosquitoes through oral feeding confirmed three highly differentially expressed genes’ roles in resistance.
In our data analysis, most of the DEGs are related to metabolic pathways involving insecticide absorption, especially genes coding for cytochrome P450 monooxygenase (P450) and cytochrome C oxidase. The highly expressed DEGs in resistant individuals included specific P450 genes, such as CYP6A8, CYP9Z4, and CYP6D4, that were identified in other insects, such as Musca domestica, An. gambiae, Cx. quinquefasciatus, and An. sinensis [42–45]. We also detected one glutathione S-transferase (GST) gene (GSTE1, AGAP009195-PA) that showed increased expression in resistant mosquitoes, which was consistent with previous studies [15,34,46]. In addition to the up-regulated CYP genes, it was interesting to notice that there were several down-regulated CYP genes, including CYP301A1, CYP4AC2, and CYP4V2. Several GST genes were found to be down-regulated (GSTE1 and GSTE5), which apparently contradicts the current knowledge on GST. GST enzymes have been shown to metabolize insecticides by facilitating reductive dehydrochlorination or by conjugation reactions with reduced glutathione, to produce water-soluble metabolites that are more readily excreted. The increased GST expression in resistant mosquitoes contributes to the removal of toxic oxygen-free radical species produced through the action of pesticides [47]. Thus, the current study’s reverse finding indicates that the GST enzyme may confer insecticide resistance in a more complex way. One explanation for decreased expression in the resistant individuals is that the decrease in gene expression may be due to reactions to various endogenous and exogenous compounds, or alternatively may represent a pathophysiological signal [48,49]. Such a GST gene down-regulation observed in resistant Ae. albopictus individuals was also found in other mosquito species: Cx. quinquefasciatus and An. sinensis [50,51]. These findings suggest complex associations between metabolic genes’ expression changes and insecticide resistance.
An RNA interference assay was conducted to verify the function of candidate genes identified from RNA-seq analysis. We used two RNAi methods: the classic microinjection-based and oral feeding RNAi. We demonstrated that both methods led to >50% reduction in the expression of 3 candidate genes, and that RNAi treatment altered the resistance phenotype. Our results indicated that these 3 genes each may played a role in deltamethrin resistance regulation, however the impact of these genes on insecticide resistance at the organismal level needs further determination. One major disadvantage of the microinjection-based RNAi method is that siRNA injection led to substantial mortality in adult mosquitoes. In the control, we observed 16% mosquito mortality, likely due to mechanical injury associated with microinjection. The oral RNAi showed less than 1% mortality from the oral delivery process, consistent with oral delivery RNAi methods reported by other studies [52]. Another advantage of oral delivery RNAi is the relative ease of laboratory operation, without the need for microinjection, and subsequently a large number of RNAi-treated mosquitoes can be used for the resistance bioassay. Therefore, oral RNAi may be considered in future studies.
Using a genotype-phenotype association study with field mosquito samples from southern China, we identified 5 SNPs within the CYP gene family that were significantly associated with pyrethroid resistance. In the SNP analysis, we also found higher F1534S mutations in the kdr gene among the 12 resistant individuals than in the 12 susceptible individuals (29.1% vs. 16.7%). The L1534S mutation frequency was 23.3% in the selected resistant strain but 0 in the laboratory susceptible colony. Therefore, the impact of mutations in the kdr gene on pyrethroid resistance is contingent on the mosquitoes’ genetic background. Mutations in other genes and expression of detoxification genes need to be considered in developing predictive biomarkers for resistance.
Our study has several limitations. First, only 12 resistant and 12 susceptible individuals were subjected to RNA-seq analysis. RNA-seq using a large sample size would detect more genetic variants with small effects on resistance. Secondly, the laboratory resistant population used for RNAi functional studies was not highly resistant to deltamethrin, which lowered the ability to detect candidate genes’ phenotypic impact. Nonetheless, the present study reveals the major genetic variants associated with pyrethroid resistance in Ae. albopictus.
Conclusions
By comparing the expression profile of deltamethrin-resistant and -susceptible wild mosquitoes from China, we identified 3 candidate genes with increased expression in resistant Ae. albopictus. We determined that if interfered with, the mosquito would change phenotype status from resistant to susceptible, suggesting that these genes play a role in insecticide resistance. We also identified 5 SNPs in 4 P450 genes that were significantly associated with insecticide resistance. Overall, this study demonstrates the power of individually based transcriptome profiling, with the combination of RNAi and genotype-phenotype association analysis, in research on resistance mechanisms. That is, both differentially expressed genes and SNPs associated with pyrethroid resistance were identified in Ae. albopictus mosquitoes. Findings from this study provided candidate resistance genes and SNPs for future functional verification using the new gene editing techniques such as CRISPR/Cas9.
Supporting information
S2 Fig. Effect of RNAi on the insecticide knockdown phenotype in Aedes albopictus.
Insecticide knockdown phenotype was the time taken for a female mosquito to be knocked down in the standard WHO insecticide susceptibility tube test. a) Adult RNAi with CYP6A8 microinjection; b) adult RNAi with CCG013931.2 microinjection; c) adult RNAi with CCG000656.1 microinjection; d) larval RNAi with CYP6A8 oral delivery; e) larval RNAi with CCG013931.2 oral delivery; and f) larval RNAi with CCG000656.1 oral delivery. The mean knockdown time was compared between the specific RNAi group and control group, and appropriate statistic and P values are shown.
https://doi.org/10.1371/journal.pntd.0006828.s002
(PDF)
S1 Table. A list of gene names, functions, and primers used in qRT-PCR for RNA-seq validation and siRNA sequence for Aedes albopictus RNA interference assay.
https://doi.org/10.1371/journal.pntd.0006828.s003
(XLSX)
S2 Table. PCR primer sequences for CYP gene polymorphism analysis in natural Aedes albopictus populations.
https://doi.org/10.1371/journal.pntd.0006828.s004
(DOCX)
S3 Table. Knockdown time and mortality rate of Aedes albopictus with RNAi delivered by the microinjection method.
https://doi.org/10.1371/journal.pntd.0006828.s005
(DOCX)
S4 Table. Knockdown time and mortality rate of Aedes albopictus with orally delivered RNAi.
https://doi.org/10.1371/journal.pntd.0006828.s006
(DOCX)
Acknowledgments
The authors would like to acknowledge Songwu Cai and Shuli Xu for mosquito sample collection, Amanda Nguyen for laboratory assistance, members of the Anthony A. James lab for RNAi microinjection and mosquito rearing, and three anonymous reviewers for their critical review and suggestions.
References
- 1. Ferreira-de-Lima V.H., Lima-Camara T.N.. Natural vertical transmission of dengue virus in Aedes aegypti and Aedes albopictus: a systematic review. Parasit Vectors. 11 (2018) 77. pmid:29391071
- 2. Huang Z., Kim J., Lacruz R.S., Jr P.B., Glogauer M., Bromage T.G., et al. Snead, Epithelial-specific knockout of the Rac1 gene leads to enamel defects. Eur J Oral Sci. 119 (2011) 168–176.
- 3. Kauffman EB K.L.. Zika Virus Mosquito Vectors: Competence, Biology, and Vector Control. J Infect Dis. 16 (2017) S976–S990.
- 4. Liu N.. Insecticide resistance in mosquitoes: impact, mechanisms, and research directions. Annu Rev Entomol. 60 (2015) 537–59. pmid:25564745
- 5. Bagi J., Grisales N., Corkill R., Morgan J.C., N’Falé S., Brogdon W.G., et al. When a discriminating dose assay is not enough: measuring the intensity of insecticide resistance in malaria vectors. Malar J. 14 (2015) 210. pmid:25985896
- 6. WHO, Pesticide evaluation scheme (WHOPES), Geneva World Heal. Organ. (2011).
- 7. Michael Coleman C.L.M., Janet Hemingway, Ann Gleave Katherine, Antoinette Wiebe, Peter W. Gething. Developing global maps of insecticide resistance risk to improve vector control. Malar J. 16 (2017) 86–95.
- 8. Pei-en W.F.J.J.. Ultra-low-volume application of pyrethrins for the control of Aedes albopictus in field. Chinese J Hyg Insectic Equipments (In Chinese). 32 (2017).
- 9. Ochomo E., Bayoh N.M., Kamau L., Atieli F., Vulule J., Ouma C., M, et al. Pyrethroid susceptibility of malaria vectors in four Districts of western Kenya. Parasit Vectors. 7 (2014) 310. pmid:24996418
- 10. Kushwah R.B.S., Dykes C.L., Kapoor N., Adak T., Singh O.P.. Pyrethroid-resistance and presence of two knockdown resistance (kdr) mutations, F1534C and a novel mutation T1520I, in Indian Aedes aegypti. PLoS Negl Trop Dis. 9 (2015) e3332. pmid:25569164
- 11. WHO, Global plan for insecticide resistance management in malaria vectors, Geneva World Heal. Organ. (2012).
- 12. Ooi E.E.. The re-emergence of dengue in China. BMC Med. 13 (2015) 99. pmid:25925732
- 13. Xu J., Bonizzoni M., Zhong D., Zhou G., Cai S., Li Y., et al. Multi-country Survey Revealed Prevalent and Novel F1534S Mutation in Voltage-Gated Sodium Channel (VGSC) Gene in Aedes albopictus. PLoS Negl Trop Dis. 10 (2016) e0004696. pmid:27144981
- 14. Li C.-X., Kaufman P.E., Xue R.-D., Zhao M.-H., Wang G., Yan T., et al. Relationship between insecticide resistance and kdr mutations in the dengue vector Aedes aegypti in Southern China. Parasit Vectors. 8 (2015) 325. pmid:26068925
- 15. Yanola J., Chamnanya S., Lumjuan N., Somboon P..Insecticides resistance in the Culex quinquefasciatus populations from northern Thailand and possible resistance mechanisms. Acta Trop. (2015). pmid:26091622
- 16. Chouaïbou M D.L., Kouadio FB, Tia E. First report of the East African kdr mutation in an Anopheles gambiae mosquito in Côte d’Ivoire. Wellcome Open Res. 2 (2017) pmid:28317032
- 17. Silva A.P.B., Santos J.M.M., Martins A.J.. Mutations in the voltage-gated sodium channel gene of anophelines and their association with resistance to pyrethroids—a review. Parasit Vectors. 7 (2014) 450. pmid:25292318
- 18. Low V.L., Chen C.D., Lim P.E., Lee H.L., Tan T.K., a L Y. Lim , et al. First molecular genotyping of voltage gated sodium channel alleles in Culex quinquefasciatus populations in Malaysia. Pestic Biochem Physiol. 107 (2013) 127–31. pmid:25149246
- 19. Alvarez L.C., Ponce G., Saavedra-Rodriguez K., Lopez B., Flores A.E.. Frequency of V1016I and F1534C mutations in the voltage-gated sodium channel gene in Aedes aegypti in Venezuela. Pest Manag Sci 71 (2015) 863–869. pmid:24935645
- 20. a Stenhouse S., Plernsub S., Yanola J., Lumjuan N., Dantrakool A., Choochote W., et al. Detection of the V1016G mutation in the voltage-gated sodium channel gene of Aedes aegypti (Diptera: Culicidae) by allele-specific PCR assay, and its distribution and effect on deltamethrin resistance in Thailand. Parasit Vectors. 6 (2013) 253. pmid:24059267
- 21. Ponce G., Sanchez I.P., Garcia S.M., Torrado J.M., Lozano-Fuentes S., Lopez-Monroy B., at al. First report of L1014F kdr mutation in Culex quinquefasciatus in Mexico. Insect Sci. (2015). pmid:25765734
- 22. Chang X., Zhong D., Fang Q., Hartsel J., Zhou G., Shi L., et al. Multiple resistances and complex mechanisms of Anopheles sinensis mosquito: a major obstacle to mosquito-borne diseases control and elimination in China. PLoS Negl Trop Dis. 8 (2014) e2889. pmid:24852174
- 23. Hu Z., Du Y., Nomura Y., Dong K.. A sodium channel mutation identified in Aedes aegypti selectively reduces cockroach sodium channel sensitivity to type I, but not type II pyrethroids. Insect Biochem Mol Biol. 41 (2011) 9–13. pmid:20869441
- 24. Ishak I.H., Riveron J.M., Ibrahim S.S., Stott R.. The Cytochrome P450 gene CYP6P12 confers pyrethroid resistance in kdr -free Malaysian populations of the dengue vector Aedes albopictus. Nat Publ Gr. (2016) 1–13.
- 25. Liu H., Yang P., Cheng P., Wang H., Liu L., Huang X., et al. Resistance Level of Mosquito Species (Diptera: Culicidae) from Shandong Province, Chin. Int J Insect Sci. (2015) 47–52.
- 26. Thomsen E.K., Koimbu G., Pulford J., Jamea-Maiasa S., Ura Y., Keven J.B., et al. Mosquito Behavior Change After Distribution of Bednets Results in Decreased Protection Against Malaria Exposure. J Infect Dis. 215 (2017) 790–797. pmid:28007921
- 27. Riveron J.M., Yunta C., Ibrahim S.S., Djouaka R., Irving H., Menze B.D., et al. A single mutation in the GSTe2 gene allows tracking of metabolically based insecticide resistance in a major malaria vector. Genome Biol. 15 (2014) R27. pmid:24565444
- 28. Bonizzoni M., Afrane Y., Dunn W.A., Atieli F.K., Zhou G., Zhong D., et al. Comparative Transcriptome Analyses of Deltamethrin-Resistant and -Susceptible Anopheles gambiae Mosquitoes from Kenya by RNA-Seq. PLoS One. 7 (2012). pmid:22970263
- 29. Zhu G., Zhong D., Cao J., Zhou H., Li J., Liu Y., et al. Transcriptome profiling of pyrethroid resistant and susceptible mosquitoes in the malaria vector, Anopheles sinensis. BMC Genomics. 15 (2014) 448. pmid:24909924
- 30. Frederic Faucon J.-P.D., Thierry Gaude, Isabelle Dusfour, Vincent Navratil, Vincent Corbel,Waraporn Juntarajumnong, et al. In the hunt for genomic markers of metabolic resistance to pyrethroids in the mosquito Aedes aegypti: An integrated next-generation sequencing approach. PLoS Negl Trop Dis. 11 (2017).
- 31. Reid W.R., Zhang L., Liu F., Liu N.. The Transcriptome Profile of the Mosquito Culex quinquefasciatus following Permethrin Selection. PLoS One. 7 (2012). pmid:23071746
- 32. Lv Y., Wang W., Hong S., Lei Z., Fang F., Guo Q.. Comparative transcriptome analyses of deltamethrin ‑ susceptible and ‑ resistant Culex pipiens pallen s by RNA‑seq. Mol Genet Genomics. 291 (2016) 309–321. pmid:26377942
- 33.
World Health Organization, Test procedures for insecticide resistance monitoring in malaria vector mosquitoes. 2013. doi:ISBN 978 92 4 150515 4.
- 34. Zhong D., Chang X., Zhou G., He Z., Fu F., Yan Z., et al. Relationship between knockdown resistance, metabolic detoxification and organismal resistance to pyrethroids in Anopheles sinensis. PLoS One. 8 (2013) e55475. pmid:23405157
- 35. Zhou Y., Liu Y., Yan H., Li Y., Zhang H., Xu J., et al. miR-281, an abundant midgut-specific miRNA of the vector mosquito Aedes albopictus enhances dengue virus replication. Parasit Vectors. 7 (2014) 488. pmid:25331963
- 36. Aïzoun N., Aïkpon R., Padonou G.G., Oussou O., Oké-Agbo F., Gnanguenon V., et al. Mixed-function oxidases and esterases associated with permethrin, deltamethrin and bendiocarb resistance in Anopheles gambiae s.l. in the south-north transect Benin, West Africa. Parasit Vectors. 6 (2013) 223. pmid:23919515
- 37. Zhang X., Zhang J., Zhu Y.K.. Chitosan/double‐stranded RNA nanoparticle‐mediated RNA interference to silence chitin synthase genes through larval feeding in the African malaria mosquito (Anopheles gambiae). Insect Mol Biol. 19 (n.d.) 683–693. pmid:20629775
- 38. Kasai S., Ng L.C., Lam-Phua S.G., Tang C.S., Itokawa K., Komagata O., et al. First detection of a putative knockdown resistance gene in major mosquito vector, Aedes albopictus. Jpn J Infect Dis. 64 (2011) 217–221.
- 39.
World Health Organizatio, Test procedures for insecticide resistance monitoring in malaria vector mosquitoes-2nd ed. 2016.
- 40. Bonizzoni M., Ochomo E., Dunn W.A., Britton M., Afrane Y., Zhou G., et al. RNA-seq analyses of changes in the Anopheles gambiae transcriptome associated with resistance to pyrethroids in Kenya: identification of candidate-resistance genes and candidate-resistance SNPs. Parasit Vectors. 8 (2015) 474. pmid:26381877
- 41. Safi N.H.Z., Ahmadi A.A., Nahzat S., Ziapour S.P., Nikookar S.H., Fazeli-Dinan M., et al. Evidence of metabolic mechanisms playing a role in multiple insecticides resistance in Anopheles stephensi populations from Afghanistan. Malar J 16 (2017) 100. pmid:28253925
- 42. Hardstone M.C., Komagata O., Kasai S., Tomita T., Scott J.G.. Use of isogenic strains indicates CYP9M10 is linked to permethrin resistance in Culex pipiens quinquefasciatus. Insect Mol Biol. 19 (2010) 717–726. pmid:20629774
- 43. Liu N., Li T., Reid W.R., Yang T., Zhang L.. Multiple cytochrome P450 genes: Their constitutive overexpression and permethrin induction in insecticide resistant mosquitoes, Culex quinquefasciatus. PLoS One. 6 (2011). pmid:21858101
- 44. Zhu F., Li T., Zhang L., Liu N.. Co-up-regulation of three P450 genes in response to permethrin exposure in permethrin resistant house flies, Musca domestica. BMC Physiol. 8 (2008) 18. pmid:18817570
- 45. Zhu F., Liu N.. Differential expression of CYP6A5 and CYP6A5v2 in pyrethroid-resistant house flies, Musca domestica. Arch Insect Biochem. Physiol. 67 (2008) 107–119. pmid:18163524
- 46. Marcombe S., Farajollahi A., Healy S.P., Clark G.G., Fonseca D.M.. Insecticide resistance status of United States populations of Aedes albopictus and mechanisms involved. PLoS One. 9 (2014) e101992. pmid:25013910
- 47.
I.I. A. Rami Horowitz. Advances in Insect Control and Resistance Management. 2016. https://doi.org/10.1007/978-3-319-31800-4
- 48. Yang T., Liu N.. Genome analysis of cytochrome p450s and their expression profiles in insecticide resistant mosquitoes, culex quinquefasciatus. PLoS One. 6 (2011). pmid:22242119
- 49. a Carvalho R., Azeredo-Espin A.M.L., Torres T.T.. Deep sequencing of New World screw-worm transcripts to discover genes involved in insecticide resistance. BMC Genomics. 11 (2010) 695. pmid:21143848
- 50. Cui P.H., Lee A.C., Zhou F., Murray M.. Impaired transactivation of the human CYP2J2 arachidonic acid epoxygenase gene in HepG2 cells subjected to nitrative stress: Research paper. Br J Pharmacol. 159 (2010) 1440–1449. pmid:20180943
- 51. Marinotti O., Nguyen Q.K., Calvo E., James A.A., Ribeiro J.M.C.. Microarray analysis of genes showing variable expression following a blood meal in Anopheles gambiae. Insect Mol Biol. 14 (2005) 365–373. pmid:16033430
- 52. Mysore K., Sun L., Tomchaney M., Sullivan G., Adams H., Piscoya A.S., et al. siRNA-Mediated Silencing of doublesex during Female Development of the Dengue Vector Mosquito Aedes aegypti. PLoS Negl Trop Dis. 9 (2015). pmid:26544686