Dengue is currently the most important arbovirus in the world. Dengue has spread widely in urban areas of tropical and subtropical regions during the last decades, including countries of Southeast Asia, Pacific and Latin America [1]. Between 2001–2011, almost 10 million dengue cases were reported in Latin America, almost 60% of these were registered in Brazil [2]. Dengue mortality can reach up to 5% of the confirmed infection cases. In addition, in tropical dengue endemic countries a loss of 1,300 disability-adjust life years (DALYs) per million people is estimated [1]. Aedes aegypti is the main dengue vector throughout the world. Control of this mosquito consists primarily of the elimination of artificial and disposable water flooded larvae breeding sites and application of insecticides. The WHO Pesticide Evaluate Scheme (WHOPES) recommends ten different compounds to eliminate larvae, including neurotoxicants (organophosphates, pyrethroids and neocotinoids), Insect Growth Regulators (chitin synthesis inhibitors and juvenile hormone analogs), and Bacillus (like B. thuringiensis var israelensis) as larvicides. However, fewer formulations are recommended for the control of adult mosquitoes, mostly five pyrethroids and one organophosphate [3].

Given their rapid mode of action and low hazardous effect to the environment, compared to organophosphate insecticides, the use of pyrethroids has increased significantly in the last two decades. Nowadays, pyrethroids are widely employed in and around households, even for pet protection and mosquito control [4]. Since Ae. aegypti is essentially an urban mosquito, it is constantly exposed to strong pyrethroid selection. As a consequence, many Ae. aegypti populations worldwide are becoming resistant to this class of insecticides [5].

Pyrethroids target the transmembrane voltage gated sodium channel (Na_V) from the insect nervous system, triggering rapid convulsions followed by death, a phenomenon known as knockdown effect [6]. The Na_V is composed of four homologous domains (I-IV), each with six hydrophobic segments (S1-S6) [7]. Because the Na_V is a very conserved protein among invertebrates, small changes are permissive without impairing its physiological role [8]. A series of mutations have been identified in different orders of insects and acarids that affect pyrethroid susceptibility, thus being referred to as ‘k nockd own r esistance’ or kdr mutations [9]. These kdr mutations may lead to conformational changes in the whole channel that maintain its physiological role but avoid insecticide action [10].

In insects, the most common kdr mutation is the substitution Leu/Phe in the 1014 site (numbered according to the Musca domestica Na_V primary sequence), followed by the Leu/Ser substitution in the same position, in Anopheles and Culex mosquitoes [11]. In the Ae. aegypti Na_V (AaNa _V ), the 1014 Leu codon is encoded by CTA, rather than TTA as in Anopheles and Culex mosquitoes. This means that two substitutions would have to be simultaneously selected in the same codon in order to change Leu to Phe (TTT) or Ser (TCA) [12]. Although several mutations have been identified in natural populations at AaNa _V [13], only the Val1016Ile and Phe1534Cys substitutions were clearly related to the loss of pyrethroid susceptibility [12, 14]. These sites are placed respectively in the IIS6 and IIIS6 regions of the channels that are known to be involved in the interaction with pyrethroids [10]. It has been previously observed that in Latin America the 1016 Ile kdr is highly disseminated [12, 15, 16] and its frequency is rapidly increasing in localities with intense pyrethroid use, such as Brazil and Mexico [15, 16]. High frequencies of 1534 Cys kdr were also observed in Grand Cayman and Martinique [14, 17].

In the current study, we demonstrate that the1534 Cys kdr mutation is present in Brazil together with the 1016 Ile allele previously found. The simultaneous occurrence of both kdr mutations at the 1016 and 1534 was found in several localities. Spatial and temporal analysis of these alleles point to a significant role of the kdr mutations in pyrethroid resistance in Brazil.

The genotyping of mutations directly related to insecticide resistance is an important surveillance tool for agricultural and sanitary purposes. Among selected mechanisms of pyrethroid resistance, kdr mutations in the voltage gated sodium channel (Na_V) are those that better correlate particular genotypes with insecticide resistance [25]. The increased efficiency of insecticide detoxification, known as metabolic resistance – involving super families of enzymes such as GST, esterases and especially the multi function oxidases P450 – may also confer resistance to pyrethroids. However, identification of these mechanisms is mainly based on enzymatic assays of low specificity [26] or on bioassays with synergist compounds [27], and are not clearly linked to particular genes. More recently, many successful transcriptome tools for metabolic resistance genes have emerged, pointing to a very complex and diverse scenario regarding insecticide selected genes and their pattern of expression among insect populations [28, 29]. Because the metabolic resistance based selection seems to have a high fitness cost, due to reallocation of energetic resources, this mechanism is expected to induce lower resistance levels, if compared to mutations in the target site molecules [30]. This was corroborated by laboratory selection with pyrethroids in an Ae. aegypti lineage: increase of the 1016 Ile ^kdr frequency was inversely proportional to the number of ‘metabolic’ genes differentially transcribed [29]. It was hypothesized that, in the presence of pyrethroid, kdr mutations are preferentially selected among other mechanisms, contributing to higher resistance levels and/or resulting in less deleterious effects.

In addition to the classical Leu1014Phe kdr mutation, several others have been associated with pyrethroid resistance [6]. Interactions of multiple Na_V mutations may modulate pyrethroid resistance levels. For instance, certain Na_V haplotypes, including synonymous substitutions, were found in two distinct field populations of Culex quinquefasciatus selected for pyrethroid resistance during 6–8 generations in the laboratory. It was suggested that some of these haplotypes were selected at an early stage of permethrin resistance and later evolved to other mutation combinations in the course of selection pressure [31]. In Ae. aegypti, a synonymous substitution at exon 20, together with an extensive polymorphism in the following intron, were linked to both Ile1011Met and Val1016Ile mutations [15, 32]. Additionally, a gene duplication event was recently described in the AaNa _V of natural populations and in a laboratory strain selected for pyrethroid resistance [33]. Although there are at least seven different mutations described in the AaNa _V , only those corresponding to the 1016 and 1534 positions are clearly related to resistance; both are placed in a domain of the sodium channel that interacts directly with the pyrethroid molecule [34].

There are two mutations described in the Aa Na_V 1016 site, Val to Ile or Gly, respectively in Latin America [12, 14, 15] and in Southeast Asia [35]. In Brazil, we found no evidence of a haplotype that contains exclusively the 1016 Ile ^kdr mutation, since it was always found together with 1534 Cys ^kdr (Na_V^R2 allele, herein). Nevertheless, we are aware that it is possible for a haplotype carrying the 1016 kdr mutation to occur in the populations examined, however, it would be present at very low frequencies. Actually, this putative allele must have occurred in two out of three Ae. aegypti populations from Grand Cayman, given that the 1016 Ile ^kdr presented a higher frequency than the 1534 Cys ^kdr substitution [14].

Differently from the 1016 position, only one substitution, Phe/Cys, was found in the 1534 site by far [14, 36]. This 1534 substitution can be linked with another one. In Thailand, the 1534 Cys ^kdr co-occurred with 1016 Gly ^kdr and 989 Pro ^kdr in the same molecule [22]. In that region an allele 1534 Cys ^kdr without mutation in 1016 site (Na_V^R1 allele, herein) seemed to be very common, since its frequency was higher than the 1016 Gly ^kdr [35].

Here we presented the distribution of the kdr variants for the AaNa_V, considering both 1016 and 1534 sites screen from several natural Brazilian populations. We considered that once these sites are very close in the genome, reporting the allele/genotypic frequencies of each site separately would not be fully informative. However, because there are still some gaps concerning the actual role of these mutations in pyrethroid resistance, regarding whether they are acting alone or synergistically, and present in cis or trans mutations, we are reporting the allele frequencies of each site rather than as an haplotype. The implication of the 1016 Ile ^kdr allele in resistance to pyrethroids was corroborated by laboratory selection, which highly increased the allelic frequency up to fixation in only five generations [29]. Accordingly, in the last decade this mutation has been rapidly spreading in natural populations from Brazil and Mexico, concomitantly with the intensification of pyrethroid usage due to the emergence of severe dengue outbreaks [15, 16]. In these cases however, the co-occurrence of the 1534 Cys ^kdr mutation has been overlooked. A recent study reported high frequencies of 1534 Cys ^kdr in Grand Cayman [14], suggesting it is not a novel mutation in Latin America. In a recent report, nine single and two double AaNa _V mutants were constructed and inserted in a Xenopus oocyte system in order to perform functional evaluations of these substitutions in the presence of type I or II pyrethroids [34]. The 1016 Ile ^kdr construct did not result in sensitivity reduction, to either pyrethroid types. On the other hand, the 1534 Cys ^kdr significantly diminished the AaNa_V sensibility to type I but not to type II pyrethroids. This same substitution in the homologous kdr site of the cockroach Na_V exhibited similar results [37].

An Ae. aegypti lineage, selected for permethrin resistance in the laboratory, exhibited high frequencies of 1016 Gly ^kdr  + 1794 Tyr^kdr substitutions in the same molecule, which suggested a synergistic effect towards pyrethroid resistance [38]. We hypothesize that mutation in the 1016 site should be important when in synergism with other specific mutations. In Brazil, the 1534 Cys ^kdr mutation is widespread throughout the territory. The Na_V^R1 allele is more frequent in North/Northeast regions whereas Na_V^R2 is more commonly present in Central/Southeast regions, generally where the highest resistance levels to pyrethroids are observed [18]. Both mutant haplotypes appear to be rapid and favorably selected in all evaluated populations. However, in the most recent samplings the Na_V^R2 double mutant was the more frequent kdr allele. The exception was MSR, in the Northeast Region, where Na_V^R2 was only recently introduced. Together these data suggest that Na_V^R2 allele would be more advantageous for pyrethroid resistance, or impose a lower fitness cost when compared to Na_V^R1. We recently demonstrated that an Na_V^R2 homozygous Ae. aegypti lineage, highly resistant to pyrethroids, exhibited a fitness cost in a series of life-trait parameters [39]. Further comparisons between Na_V^R1 and Na_V^R2 lineages will be of importance to better clarify those assumptions.

It is of note that since 2001 and up to 2009 the Brazilian Dengue Control Program employed pyrethroids in ultralow volume applications in several municipalities as part of the effort to control the dengue vector [18]. With very few exceptions, the basis for pyrethroid selection pressure derived from national campaigns is essentially the same in the whole country. Therefore, differential selection pressures would not explain the aforementioned regionalization of the kdr alleles. It is likely that the current distribution of the kdr alleles reflects distinct Ae. aegypti populations that colonized the continent. Population genetics analysis of neutral loci will help us to unravel the evolutionary routes of these resistance genes.

In conclusion, pyrethroids are the most employed insecticides worldwide and the only chemical class presently allowed in long lasting treated materials, such as nets and curtains [40]. Although novel control strategies are being tested in the field, such as those based on transgenic and on Wolbachia-infected mosquitoes [2, 41, 42], insecticides will certainly play an important role for yet a long time. Knowledge of the sodium channel diversity in natural populations together with the role of each allele regarding pyrethroid resistance as well as their fitness effects are crucial for preserving the effectiveness of this class of compounds as a viable tool against Ae. aegypti.