Voltage-sensitive sodium channel (Vssc) mutations associated with pyrethroid insecticide resistance in Aedes aegypti (L.) from two districts of Jeddah, Kingdom of Saudi Arabia: baseline information for a Wolbachia release program

Background Dengue suppression often relies on control of the mosquito vector, Aedes aegypti, through applications of insecticides of which the pyrethroid group has played a dominant role. Insecticide resistance is prevalent in Ae. aegypti around the world, and the resulting reduction of insecticide efficacy is likely to exacerbate the impact of dengue. Dengue has been a public health problem in Saudi Arabia, particularly in Jeddah, since its discovery there in the 1990s, and insecticide use for vector control is widespread throughout the city. An alternative approach to insecticide use, based on blocking dengue transmission in mosquitoes by the endosymbiont Wolbachia, is being trialed in Jeddah following the success of this approach in Australia and Malaysia. Knowledge of insecticide resistance status of mosquito populations in Jeddah is a prerequisite for establishing a Wolbachia-based dengue control program as releases of Wolbachia mosquitoes succeed when resistance status of the release population is similar to that of the wild population. Methods WHO resistance bioassays of mosquitoes with deltamethrin, permethrin and DDT were used in conjunction with TaqMan® SNP Genotyping Assays to characterize mutation profiles of Ae. aegypti. Results Screening of the voltage-sensitive sodium channel (Vssc), the pyrethroid target site, revealed mutations at codons 989, 1016 and 1534 in Ae. aegypti from two districts of Jeddah. The triple mutant homozygote (1016G/1534C/989P) was confirmed from Al Safa and Al Rawabi. Bioassays with pyrethroids (Type I and II) and DDT showed that mosquitoes were resistant to each of these compounds based on WHO definitions. An association between Vssc mutations and resistance was established for the Type II pyrethroid, deltamethrin, with one genotype (989P/1016G/1534F) conferring a survival advantage over two others (989S/1016V/1534C and the triple heterozygote). An indication of synergism of Type I pyrethroid activity with piperonyl butoxide suggests that detoxification by cytochrome P450s accounts for some of the pyrethroid resistance response in Ae. aegypti populations from Jeddah. Conclusions The results provide a baseline for monitoring and management of resistance as well as knowledge of Vssc genotype frequencies required in Wolbachia release populations to ensure homogeneity with the target field population. Vssc mutation haplotypes observed show some similarity with those from Ae. aegypti in southeast Asia and the Indo-Pacific, but the presence of the triple mutant haplotype in three genotypes indicates that the species in this region may have a unique population history. Graphical Abstract Supplementary Information The online version contains supplementary material available at 10.1186/s13071-021-04867-3.

alternative approach to insecticide use, based on blocking dengue transmission in mosquitoes by the 23 endosymbiont Wolbachia, is being trialled in Jeddah following the success of this approach in Australia and 24 Malaysia. Knowledge of insecticide resistance status of mosquito populations in Jeddah is a prerequisite for 25 establishing a Wolbachia-based dengue control program as releases of Wolbachia mosquitoes succeed when 26 resistance status of the release population is similar to that of the wild population. 27 Methods 28 WHO resistance bioassays of mosquitoes with deltamethrin, permethrin and DDT were used in conjunction 29 with TaqMan  SNP Genotyping Assays to characterise mutation profiles of Ae. aegypti from Jeddah. 30

Results 31
Screening of the voltage sensitive sodium channel (Vssc), the pyrethroid target-site, revealed mutations at 32 codons 989, 1016 and 1534 in Ae. aegypti from two districts of Jeddah. The triple mutant homozygote 33 (1016G/1534C/989P) was confirmed from Al Safa and Al Rawabi. Bioassays with pyrethroids (Type I and II) and 34 DDT showed that mosquitoes were resistant to each of these compounds based on WHO definitions. An 35 association between Vssc mutations and resistance was established for the Type II pyrethroid, deltamethrin, 36 with one genotype (989P/1016G/1534F) conferring a survival advantage over two others (989S/1016V/1534C 37 and the triple heterozygote). An indication of synergism of Type I pyrethroid activity with piperonyl butoxide 38 4 Background 48 Target-site resistance to pyrethroids in Aedes aegypti, also known as knockdown resistance (kdr), is an 49 autosomal, incompletely recessive trait (1). Vssc mutations at codons 1016 and 1534 occur in Ae. aegypti 50 within the pyrethroid receptor sites in Domains II (S6) and III (S6) of the protein molecule (2). A third mutation, 51 S989P, which is often in perfect linkage with V1016G, is not known to reduce the sensitivity of the sodium 52 channel (2), but confers some additive pyrethroid resistance in the homozygous state in combination with 53 Dengue has been a public health problem in Saudi Arabia and particularly in Jeddah since its discovery there 61 in 1994 (9). Currently, dengue suppression relies on control of the mosquito vector, Aedes aegypti, through 62 the applications of insecticides of which the pyrethroid group has played a dominant role (5). Insecticide 63 resistance is prevalent in Ae. aegypti around the world (10, 11) and the resulting reduction of insecticide 64 efficacy is likely to be exacerbating the impact of dengue and certainly threatens the long-term utility of this 65 control method. Hence an alternative approach based on blocking dengue transmission in mosquitoes by the 66 endosymbiont Wolbachia is being trialled following the success of this approach in suppressing dengue in 67 other countries (12,13

Statistical analyses 170
Vssc mutation data from Al Safa and Al Rawabi collected over three years were analyzed for site differences 171 in genotype frequencies using contingency tables with significance tested through the chi-squared statistic. 172 Monte Carlo tests were used to determine significance because expected values in some cells were particularly 173 low even when data were combined across collection dates. These analyses were performed in IBM SPSS 174 Statistics (IBM Corp., Armonk, NY, USA;2013). 175 176 Odds ratios (with 95% confidence intervals) (29) were calculated to indicate the odds of an individual surviving 177 exposure to an insecticide if it carried one particular genotype compared with another. An odds ratio of 1 178 indicates that there is no relationship between survival and the genotype under investigation. If 95% 179 confidence intervals of the odds ratio do not span the value "1" then this suggests that the genotype is 180 associated with survival. 181 182

183
Vssc mutations in temporal field samples 184 An initial sample of ten individuals per district identified Vssc genotypes seen in samples from Asia and the 185 Indo-Pacific except for one individual from Safa which did not fit the pattern (a homozygous mutant at codon 186 1016 and 989, but a heterozygote at 1534) ( Table 1). Data from multiple studies (26, 30-33) suggest that 187 certain haplotypes of the three mutation sites predominate in a population and there is little evidence of 188 crossing over to disrupt the phase patterns found. The linkage patterns we have observed for these mutations 189 in mosquitoes from Asia and the Indo-Pacific cannot produce individuals of genotype I found in Safa ( Figure  190 1). To obtain this genotype, one parent had to contribute a triple mutant haplotype (H5) and the other 191 contributed a second haplotype (H1) that we see in Asia/ Indo-Pacific samples ( Figure 2). Saudi Arabia is one 192 of the few countries where the triple mutant genotype has been reported previously (5) and it was found as a 193 haplotype in the case here (two individuals of the same genotype) and also as four individuals made from a 194 combination of the H5 and H2 haplotypes which we have called genotype J ( Figure 2 from these sites in Saudi Arabia were those found in samples from the Indo-Pacific region (26). However, there 199 was again, one individual from Al Safa with genotype I (Table 1). 200   201 The third sample of Ae. aegypti was taken in February 2020 and again sixty samples from each site were 202 screened for Vssc mutations. Genotype I was again found in mosquitoes from Al Safa (Table 1) indicate that the triple wildtype haplotype is present in that district, but was not found as a homozygote. The 207 frequency of each of the common genotypes (A, B, C) is similar between Al Safa and Al Rawabi (Table 1) None of the bioassays conducted showed mortality in the range 98-100% which would indicate susceptibility 217 according to WHO guidelines (34). Permethrin 0.75% induced only 44.8% mortality and DDT 4% bioassays 218 showed mortality ranging from 3.2 to 10.4% confirming a high level of resistance to these compounds in Ae. 219 aegypti from Jeddah. Four bioassays with deltamethrin 0.05% showed an average mortality just under 90% 220 which indicates that the population of Ae. aegypti from Jeddah is resistant to this compound as well (Table 2). 221

Synergist bioassay 222
Odds of mosquitoes being alive if exposed to permethrin 0.75% alone are just over twice those if exposed to 223 permethrin 0.75% + PBO 4% (Table 3). PBO 4% alone caused some mortality, but the odds of being alive are 224 not greater if exposed to permethrin compared with exposure to PBO alone (Table 3). 225 226

Comparison of Vssc genotypes in dead and survivors 227
Five Vssc genotypes (A, B, C/L, J, K) were identified in the surviving pool of mosquitoes screened and four 228 genotypes (A, B, C/L, J) occurred in the dead mosquitoes exposed to 0.75% permethrin (Table 4). Two triple 229 mutants were found in the survivors. There was no significant difference (α=0.05) in the odds of being alive 230 after exposure to permethrin 0.75% if carrying one genotype over any other (Supplementary Table A). 231 232 Survivors of deltamethrin 0.05% fell into four genotypes as did the dead individuals, though the fourth and 233 least frequent genotype differed between them (I for alive and J for dead) (Table 4). Odds ratios were 234 significant (α=0.05) for two genotype comparisons indicating that the odds of surviving were higher for 235 individuals of genotype A than genotype B and also higher for genotype A than the triple heterozygote 236 (genotype C or L) (Table 5).  are not fixed for one genotype, but genotypes A, B, C are common and no homozygote wildtype individuals 253 were found in the sample. A wildtype haplotype exists, at least in Al Rawabi, so it is possible that homozygous 254 wildtype individuals (genotype F) may occur in the population, albeit rarely. There is no indication that this 255 haplotype still exists in Al Safa, but increased sampling might still find it. No instances of a mutation at codons 256 410 or 1763 were observed in the samples screened indicating lack of independent selection or contact of the 257 mosquito populations with those from Taiwan (22), Brazil (19), Mexico (35) or West and Central Africa (20). 258 The triple mutant homozygote (1016G/1534C/989P) (genotype K - Figure 2) can now be confirmed from Al 259 Safa and Al Rawabi, suggesting that Ae. aegypti from Saudi Arabia may have a unique population history which 260 could be further explored in a full genomic analysis. The implications of finding the triple mutant as haplotype 261 5 and genotypes I -L are not well understood, but provide an opportunity to study effects of these genotypes 262 in more detail in bioassays and potentially fitness experiments. Although the triple mutant genotype is rare in 263 the populations of Ae. aegypti sampled from Jeddah (2-3%), the frequency is higher than that reported from 264 Myanmar (0.98%) (32) and is high enough to facilitate collection of adequate numbers to isolate individuals 265 for crossing experiments which could allow further characterisation of the H5 haplotype and I-L genotypes in 266 bioassays. Two triple homozygous mutants were found in the live pool of our bioassay with permethrin 0.75%. 267 The Vssc genotypes found in Al Safa and Al Rawabi may all be constructed from haplotypes 1,2,3 and 5. 268 Haplotype 4, found only in Taiwan of the countries sampled in the Indo-Pacific (26) and in Indonesia (3), is not 269 found in the Saudi Arabian samples we have screened which is consistent with results of Al Nazawi, Aqili (5). 270 It is possible that the triple heterozygous genotype in Saudi Arabia could be composed of either H1 + H2 (usual 271 condition in the Indo-Pacific -genotype C) or H3 + H5 (triple wildtype plus triple mutant -genotype L) (Figure  272 4). Both H3 and H5 are rare compared with H1 and H2, but there is a possibility that some of the heterozygotes 273 have this rare configuration. It is not known whether there is any difference in insecticide resistance in 274 mosquitoes which have one or the other haplotype combinations of the triple heterozygote. 275 Potential implications of the genotypes A-E and I-K on control of Ae. aegypti with pyrethroid insecticides have 276 been surmised (Table 6) (5, 30, 36) ,but not all genotypes have been tested for effects on susceptibility. It is 277 likely that efficacy of both Type I and Type II pyrethroids is compromised by many of these modifications to 278 the Vssc, though genotypes have differential effects. 279 Although resistance to three insecticides which target the Vssc was detected in bioassays in the Ae. aegypti 280 from Jeddah in this study, the only clear association between Vssc mutations and resistance was established 281 for the Type II pyrethroid, deltamethrin, with genotype A conferring a survival advantage compared with 282 genotypes B (1534C) or C (triple heterozygote). Genotype A is known to reduce the sensitivity of the Vssc for 283 both Type I and Type II pyrethroids (2), but in our study there was no obvious effects for Type I (permethrin 284 0.75%) or DDT. It may be that this lack of effect is concentration related and a different concentration of 285 permethrin or DDT would reveal a segregation of genotypes between dead and surviving mosquitoes as 286 reported from the Jazan region of Saudi Arabia (4). In a similar study in Ae. aegypti from Malaysia (15), a 287 survival advantage over wildtype was conferred to mosquitoes by genotypes B and C when exposed to 288 permethrin 0.25%. 289 It is important to note that target-site resistance may not be the only mechanism of resistance to pyrethroids 290 that has been selected in the Jeddah populations of Ae. aegypti. A role of metabolic resistance in Ae. aegypti 291 from Al Safa and Al Rawabi is likely. Al Nazawi et al. (5) saw an increase in mortality of Ae. aegypti from Jeddah 292 with deltamethrin, using the synergist (PBO), which indicates that detoxification by oxidases (cytochrome P450 293 mixed function oxidase system) accounts for some of the pyrethroid resistance response. We also found an 294 indication of such synergism in bioassays with the Type I pyrethroid, permethrin 0.75%. 295

Conclusions 296
The continuing presence of the wildtype haplotype and minor difference in genotypes between Al Safa and Al 297 Rawabi suggests that selection for resistance may be patchy between districts. Changes in genotype 298 frequencies over time also suggest that selection for pyrethroid resistance is ongoing. The results provide a 299 baseline for ongoing monitoring of resistance particularly with implementation of resistance management 300 programs and also indicate Vssc genotypes required in Wolbachia release populations to ensure homogeneity 301 with the target field population. 302

Declarations 303
Ethics approval and consent to participate -not applicable 304 Consent for publication -not applicable 305 Availability of data and materials -All data generated or analysed during this study are included in this 306