Historical survey of the kdr mutations in the populations of Anopheles sinensis in China in 1996–2014

Anopheles sinensis has become an important malaria vector in China. The long-term extensive utilization of pyrethroids for ITNs and IRS for mosquito control in the last three decades has resulted in the occurrence of resistant An. sinensis populations in many regions. Knockdown resistance (kdr), caused by point mutations in the VGSC gene, is one of the mechanisms that confer resistance to DDT and pyrethroids. Recently, several investigations revealed the kdr occurrence in some An. sinensis populations, however, no kdr data were available earlier than 2009. A survey tracking the dynamics of the kdr mutations in past decades would provide invaluable information to understand how the kdr alleles spread in mosquito populations temporally and spatially. A survey was conducted on the kdr alleles at condon 1014 of the VGSC gene and their distributions in 733 specimens of An. sinensis and 232 specimens of the other eight member species of the Anopheles hyrcanus group that were collected from 17 provinces in China in 1996–2014. A total of three kdr alleles, TTT (F), TTG (F) and TGT (C) were detected, and TGT (C) and TTT (F) were already present in the specimens from Jiangsu and Shandong as early as 1997. The TTT (F) was the most frequent mutant allele, and largely distributed in central China, namely Shandong, Jiangsu, Anhui, Henan, Shanghai, Jiangxi and Hubei. When data were analysed in three time intervals, 1996–2001, 2005–2009, 2010–2014, the prevalence of kdr alleles increased progressively over time in the populations in central China. In contrast, the kdr alleles were less frequent in the samples from other regions, especially in Yunnan and Hainan, despite the documented presence of pyrethroid resistant populations in those regions. Interestingly, no mutant alleles were detected in all 232 specimens of eight other species in the An. hyrcanus group. The survey revealed that the kdr occurrence and accumulation in the An. sinensis populations were more frequent in central China than in the other regions, suggesting that the kdr mutations may contribute significantly to the pyrethroid resistance in the mosquitoes in central China.


Background
Mosquito control is one of the integrated programmes to prevent transmission of mosquito-borne diseases such as malaria, filariasis, and dengue fever. Chemical insecticides have been extensively used for vector management since the 1940s. Four major categories of insecticides have been utilized: organochlorines, organophosphates, carbamates and pyrethroids [1]. DDT and pyrethroids function as neurotoxins that target voltage-gated sodium channels (VGSC) and interfere electronic signaling in the nervous system, which results in paralysis and death, an effect known as knockdown [2]. One of the mechanisms that mosquitoes have developed for the resistance to DDT and pyrethroids is the target insensitivity, which is caused by mutations in the VGSC gene. A prominent mutation is the substitution of leucine at residue position 1014 in mosquitoes with the knockdown resistance (kdr) [1]. A positive correlation between the kdr mutation and the resistant phenotype to pyrethroids and DDT was well documented in various Anopheles populations [3][4][5][6][7][8][9][10].
Anopheles sinensis is an Oriental species with wide distributions in China. Anopheles sinensis is one of the principal malaria vectors in many malaria-endemic regions, especially in the central China, due to its abundant population size [11]. In China, DDT has been widely used for conventional indoor residue sprays (IRS) since 1950s, and pyrethroids have been applied for IRS and insecticide-treated nets (ITNs) since 1980s [12]. These measures have been effective in reducing malaria transmission [13]. However, the long-term applications of insecticides have resulted in the development of resistance in mosquito populations. For instance, the DDT resistance in An. sinensis was documented in Yunnan as early as 1981 [14]; the permethrin resistance was reported in Sichuan in 1989 [15]; and in Fujian in 1989-1993, resistant populations occurred two years after IRS and ITNs applications, and the resistance spread in more populations three years after applications [13].
Recent years, the kdr genotyping has been included in monitoring pyrethroid resistance in An. sinensis in China. Several investigations have been made on the distributions of kdr alleles in various An. sinensis populations, such as provinces Jiangsu, Henan, Hunan, Anhui, Jiangxi, Yunnan, and Hainan [7,[16][17][18][19][20][21]. However, these studies were conducted in the last five years, the kdr data were obtained from the mosquito specimens that were sampled in 2009 and after. No earlier kdr data were available. A historical survey tracking the dynamics of the kdr mutations in the past decades would provide invaluable information to understand how the kdr mutations occurred and spread in mosquito populations temporally and spatially. Therefore, a study was conducted to investigate the genotypes of the codon 1014 of the VGSC gene in the specimens of An. sinensis and the other eight member species of the An. hyrcanus group that were collected from 17 provinces in 1996-2014. The data revealed that the kdr alleles were already present in the specimens sampled in 1997, and kdr alleles progressively increased over decades in the An. sinensis populations in central China. The kdr were much less prevalent in the other populations, particularly in Yunnan and Hainan.

Mosquito collections and species identification
Wild mosquito adults were collected in 1996-2014 from 31 sampling sites in 17 provinces in China. Mosquitoes were caught by using light traps at livestock corrals or human landing catches, with consent of the owners and persons involved in the study. The collection information was summarized in Tables 1, 2 and Figure 1. Mosquitoes of the Anopheles hyrcanus group were sorted out in the field by morphology using the identification keys [11], and brought back to the lab. The species identity was molecularly determined by a diagnostic PCR assay based on the ribosomal DNA (rDNA) second internal transcribed spacer (ITS2) markers [22] or by the ITS2 sequencing [23].
kdr gene amplification and sequencing To identify kdr alleles, a partial sequence of S6 segment of domain II of the VGSC gene was amplified from 20-50 ng genomic DNA from single specimens using CD1 (5′-TGA TCG TGT TTC GCG TGC TG-3′) and CD2 (5′-GTC TCG TTA TCC GCC GTT GG-3′) primers [3]. The PCR kit was from Aidlab, China. The PCR reaction was carried out in Verity 96 well Thermal Cycler (Applied Biosystems, USA) included an initial step of denaturation at 94°C for 1 min, followed by 35 cycles of amplification at 94°C for 30 s, 55°C for 30 s, and 68°C for 30 s, with a final extension step at 68°C for 7 min. After electrophoresis, PCR products were purified and used for sequencing in both directions with the CD1 and CD2 primers, respectively. There were 37 specimens, of which the PCR products were cloned into plasmids (pGEMX-T Easy Vector, Aidlab, China), and then sequenced, due to the double peaks at two positions of the codon 1014.

Statistical analyses
The codon 1014 was examined by sequence analysis, and genotypes were determined. In each sample, for a particular allele, the allele frequency was calculated as: number of alleles/(sample size × 2). The mutation frequency was defined as frequency of sum of wildtype/mutant heterozygotes and mutant/mutant homozygotes, which was calculated as: (sum of wildtype/mutant and mutant/mutant individuals)/sample size. The maximum likely frequency (y) of an allele present or absent in a sample of a given size (x) was obtained from the upper of 95% confidence limit of binomial distribution, given by y = 1-0.051/x, following the example of Post and Millest [24].

Taxonomic composition of the mosquito collections
In the current study, 733 specimens of An. sinensis were used in the kdr allele survey. Some of these samples were from mosquito collections of our previous work on the molecular identification and phylogeny of the An. hyrcanus group [22,23] and the population genetics of An. sinensis [25] and Anopheles lesteri [26] since 1996. The samples included specimens from 31 sampling sites in 17 provinces in China from 1996 to 2014 ( Figure 1 and Table 1). In addition, 232 mosquitoes of eight other species in the An. hyrcanus group were identified in the samples collected in 19 locations from 11 provinces in China in 1996-2013, including 167 of Anopheles lesteri, 24 of Anopheles yatsushiroensis, 13 of Anopheles belenrae, 19 of Anopheles kleini, three of Anopheles liangshanensis, two of Anopheles peditaeniatus, two of Anopheles kunmingensis and two of Anopheles crawfordi.

Frequency and distribution of kdr mutations in the An. sinensis populations
In order to detect kdr alleles, a 343 bp fragment of the IIS6 domain of the VGSC gene was PCR amplified and sequenced directly. At codon 1014, four alleles were identified    Table S1 and Additional file 2: Table S2). In 2010-2014, 308 specimens in 13 samples from seven regions were investigated. The kdr alleles were found in Shandong, Shanghai, Anhui, Shaanxi, and Sichuan (Additional file 1: Table S1 and Additional file: 2: Table S2). In addition to the mutant genotypes mentioned above, two more mutant genotypes, TTC/TTC and TTT/TTC, were detected in Anhui and Sichuan. The mutation frequency was 41.67-100%. No mutant alleles were detected in the samples from Yunnan and Hainan (Additional file 2: Table S2).
Geographically, the kdr alleles were distributed largely in the populations from central China. Therefore, the samples from central China were further analysed. As shown in Table 3, the frequency of wildtype genotype dropped over time, from 80.91% in 1996-2001 to 0.64% in 2010-2014, whilst the frequency of mutant genotypes rose from 0 to 87.18% (Table 3, Figure 2A). Furthermore, the mutant genotypes F/F and F/C were enriched in Shandong, Shanghai and Anhui in 2012-2014 (Table 3 Table 4). Overall, the frequency of wildtype homozygotes was 100% in 1997-2001, and still relatively high as 71.71% in 2010-2014 in these regions (Figure 1, Figure 2B). The analysis revealed that the kdr alleles were enriched in the An. sinensis populations in central China with a trend of progressive increase over time. The kdr alleles were less prevalent in the other regions of China.
The allele of codon 1014 in other species of the An. hyrcanus group A total of 232 specimens from eight other species in the An. hyrcanus group were screened. No kdr alleles were  Table S2 for details).
detected. All of these specimens were the homozygotes of the wildtype allele TTG.

Discussion
In China, DDT was largely used in 1950s-1970s, and pyrethroids have been utilized since 1980s when DDT was banned [27]. The applications of pyrethroids for IRS and ITNs have greatly contributed to the success of reducing malarial transmission. The occurrence of pyrethroid resistance in An. sinensis has been documented in many regions since 1980s [28,29]. For example, in Zhejiang [30], Hubei [31], Jiangsu [32], Shandong [33], Yunnan [34], Henan [34], Fujian [35], and Hainan [36]. Recently, Wang et al. conducted a survey of pyrethroid susceptibility in the An. sinensis populations from eight malaria endemic regions including Hubei, Henan, Hunan, Jiangsu, Jianxi, Sichuan, Shanghai and Yunnan [37]. They found that the An. sinensis populations in all examined regions were resistant to deltamethrin. The mortalities were in a range of 5.96-64.54% upon exposure to the diagnostic concentration of 0.25% deltamethrin [37]. The relationship of kdr genotypes and pyrethroid resistance has been investigated in recent years. At codon 1014 of the VGSC gene, five mutant alleles, TTT(F), TTG(F), TGT(C), TCG(S) and TGG(W), have been detected. The TTT(F) allele was the most prevalent mutant, followed by the alleles TGT(C) and TCG(S); the allele TGG (W) occurred rarely [7,[16][17][18][19]21,34,38,39]. The kdr alleles were present with high frequency largely in the populations in central China, such as Jiangsu, Anhui, Shandong, Hubei, and Henan. For example, Qi and Cui detected the kdr alleles Table 3 The frequency of kdr mutations in the Anopheles sinensis samples in central China  1014F and 1014C in three populations from Henan, and the kdr allele frequency was associated with the resistant phenotype. However, no kdr alleles were detected in the two populations from Yunnan [34]. In the five populations collected from Jiangsu in 2009-2010, the kdr alleles 1014F and 1014C were detected and the frequency of allele 1014F was correlated with the resistance to beta-cypermethrin in these populations [16]. In the Guangxi collections, the kdr alleles 1014S, 1014F and 1014W were found [20]. In an investigation reported by Zhong et al. the kdr alleles 1014F and 1014C were present with high frequency (88.5-94.8%) in the resistant populations from Hunan, Hubei and Jiangsu in central China [17]. Both kdr alleles and the monooxygenase activity were significantly associated with the deltamethrin resistance, but the monooxygenase activity played a stronger role. On the other hand, no kdr alleles were detected in the two resistant populations in Yunnan, where the resistance was correlated with the monooxygenase activity [17]. A similar pattern was found in another study, in which the kdr mutation L1014F (70.0-88.9%) and L1014C (11.1-26.7%) were detected in the Anhui populations with higher frequency in the resistant mosquitoes [19]. Again, no kdr alleles were found in the mosquitoes from Yunnan. Based on a CART statistical analysis, metabolic detoxification enzymes (monooxygenases, glutathion S-transferase and carboxylesterases) played major roles in resistance to pyrethroids and DDT while kdr alleles weighed less in the context [19]. In Hainan Island, the L/F heterozygotes were present with low frequency (6.7-9.5%) in the DDT and pyrethroid-resistant individuals of the two An. sinensis populations, but no kdr alleles were detected in the sympatric An. vagus [18]. Overall, aforementioned studies demonstrated that the kdr mutations occurred largely in the An. sinensis populations in central China, where the resistance to DDT and pyrethroids was conferred primarily by the metabolic detoxification mechanisms as well as the kdr mutation. In the populations in other regions, the kdr alleles were less prevalent, and the pyrethroid resistance was conferred by the metabolic mechanisms.
In the current study, codon 1014 was examined in a large collection of the An. sinensis samples covering 17 provinces in a time span of 1996 to 2014. The kdr alleles 1014C and 1014F were found in the specimens sampled from Jiangsu and Shandong as early as 1997 (Additional file 1: Table S1). This clearly indicated that the kdr alleles already existed in the An. sinensis populations in 1990s. In line with the findings in the other reports mentioned above, the kdr alleles were more prevalent in central China. The occurrence of kdr alleles has been progressively increasing over time (Figures 1 and 2). For example, in the samples from Shandong that were collected in 1997, 2007, and 2012, the mutation frequency increased from 33.3% in 1997, 95.45% in 2007 to 100% in 2012 (Table 3). In the other regions, the kdr genotypes frequency was lower in Sichuan (0 in 1997 and 60.78% in 2014). Moreover, no kdr alleles were detected in the samples from Yunnan  Table 4). Apparently the pyrethroid resistance in these regions was conferred majorly by the mechanisms other than kdr mutations. Similar situations have been reported in An. gambiae in Africa. In south-western Nigeria, the kdr alleles were not found in the pyrethroidresistant individuals of the molecular M form of An. gambiae [40]. In a survey conducted in south-western Chad, central Africa, the allele 1014F was frequently present in the resistant S form of An. gambiae, but was not found in the M form or An. arabiensis [41]. The frequent occurrence and accumulation of the kdr alleles in An. sinensis in central China may be explained by their large population size and wide distribution range. The mutant alleles may have had better chances to be  Table 3). (B) Mosquitoes were sampled in Liaoning, Shaanxi, Guizhou, Sichuan, Chongqing, Yunnan, Hainan, Fujian, Guangdong, and Guangxi (data in Table 4). See Figure 1 for sampling sites on the map. selected and maintained in the An. sinensis populations in central China.
In this study, no kdr alleles were detected in all 232 specimens of the other eight member species of the An. hyrcanus group, including 167 specimens of An. lesteri. This resembles a finding reported by Kang et al. [42], in which the kdr alleles were only found in An. sinensis, not in the other five species (Anopheles pullus, Anopheles kleini, An. lesteri and An. belenrae) in the An. hyrcanus group from the Republic of Korea [42]. As a primary malaria vector in China, An. lesteri (previously known as An. anthropaphagus, which was rectified as An. lesteri based on the rDNA ITS2 sequences [23]), has been a major target of the vector control programmes. The extensive implementation of IRS and ITNs in 1990s-2010s has led to a remarkable reduction in An. lesteri abundance in China, which has resulted in a significant drop in malaria morbidity [43][44][45]. It is intriguing that the low kdr occurrence in some An. sinensis Table 4 The frequency of kdr mutations in the Anopheles sinensis samples in other regions  populations and absence in An. lesteri (in this study) given that both mosquitoes had been exposed to strong pyrethroid pressure. Further study is needed to investigate the resistant status and underlying mechanisms in these populations.