Introduction

The two most widely applied vector control tools, insecticide treated nets (ITNs) and indoor residual spraying (IRS) have contributed greatly to the decline in global malaria rates [1, 2]. Pyrethroids are the most commonly used insecticides in control programs due to their low human toxicity and high efficacy against vectors [3, 4]. Previously, DDT, an organochlorine, was the most widely used insecticide for vector control with its use spread out over multiple countries for malaria control [5, 6]. The widespread use of these insecticides has likely contributed to the selection of resistance across sub-Saharan Africa [7] (http://www.irmapper.com/).

Increased resistance to pyrethroids is particularly troubling since this is the only class of insecticides approved by WHO for use on ITNs [3]. If ITNs are rendered ineffective, a surge in malaria transmission could follow [8]. Resistance to pyrethroids has been reported from multiple sites in western Kenya [9, 10] with both target site and metabolic resistance mechanisms implicated [9–15]. DDT and pyrethroids function by binding to the voltage gated sodium channels (Vgsc) on the mosquito’s neurons delaying the closing of the sodium channel; prolonging the action potential and causing repetitive neuron firing, ultimately resulting in paralysis and death [8, 16].

In Anopheles gambiae s.l., knock down resistance (kdr) is commonly caused by mutations in the Vgsc- either from leucine (TTA) to phenylalanine (TTT) or leucine to serine (TCA) [11, 17] at codon 1014. Vgsc-1014S (formerly kdr-east) was first reported in Kenya in 2000 and has been observed to occur at high frequencies in the local An. gambiae populations [10, 11]. Thus far, there has been no report of the existence of Vgsc-1014F (formerly kdr-west) in Kenya but has been reported in Uganda and Tanzania in the recent past [18, 19]. Our work demonstrates the emergence of Vgsc-1014F in western Kenya in two principal malaria vectors, An. gambiae and An. arabiensis.

Findings

Material and methods

This study was conducted in four malaria endemic districts in western Kenya with two distinct Vector control interventions: Rachuonyo and Nyando where IRS (Deltamethrin in 2011 and lambdacyhalothrin in 2012) was combined with ITNs (treated with permethrin or deltamethrin); and in Bondo and Teso where only ITNs are deployed [9]. Mosquito collections were performed annually between June and September in 2011, 2012 and 2013. Mosquito sampling, rearing and bioassays of emergent adults were conducted as described in Ochomo et al. [9].

Species identification & Vgsc genotyping

DNA was extracted from whole specimens and a PCR assay [20] was used to distinguish between An. gambiae and An. arabiensis. DNA samples were genotyped to identify the kdr genotype at amino acid position 1014 of the Vgsc using a modification of the protocol by Bass et al., [21] as described in Mathias et al., [10].

Exon sequencing of Vgsc

Previous studies in western Kenya have only reported the presence of Vgsc-1014S mutation. Therefore, in order to confirm the presence of Vgsc-1014F and to determine if was a de novo origin, a subsample of the Vgsc-1014F carriers were Sanger sequenced. Prior to sequencing, conventional PCR was used to amplify the exon 20 [22] which contains the 1014 locus]. Samples were sequenced at Centre for Genomic Research, University of Liverpool, UK and resulting sequences aligned using CodonCode aligner (http://www.codoncode.com/aligner/).

Analysis for the origin of Vgsc-1014F Mutation

Gene sequences obtained from the sequencing exons 20 and 27 were aligned using Codon Code aligner (http://www.codoncode.com/) and the contigs transferred to DnaSP (http://www.ub.edu/dnasp/) as FASTA files. The files were concatenated and then run using the PHASE algorithm in DnaSP [23]. The phased files were exported as a Phylip file to TCS, a statistical parsimony software for phylogenetic network estimation (http://darwin.uvigo.es/software/tcs.html).

Results

Frequency of Vgsc-1014S and Vgsc-1014F in the study sites from 2011 to 2013

We observed low frequencies of Vgsc-1014S in An. arabiensis,even though we had high frequencies of the same allele in An. gambiae, they were much lower than has been reported previously [10]. We saw a simultaneous appearance of Vgsc-1014F inboth An. gambiae and An. arabiensis in 2012 in all four study sites (Table 1) and thereafter compared the mean frequencies of the genes among the three years (Table 2). Of these, 19 samples (12 An. gambiae and 7 An. arabiensis) were sequenced. 3 An. gambiae and 3 An. arabiensis were confirmed to be homozygous for Vgsc-1014F with one An. arabiensis heterozygote detected. The sequences were deposited in GenBank (Accession numbers: KR867642- KR867651, KT758295-KT758303).

Table 1 Frequency of Vgsc-1014F and Vgsc-1014S mutations in An. gambiae and An. arabiensis populations of western Kenya from 2011 to 2013
Full size table
Table 2 Comparison of mean frequencies of Vgsc-1014F and Vgsc-1014S mutations in An. arabiensis using ANOVA and Tukey’s test. A similar analysis could not be done for An. gambiae as only one site (Teso) had sufficient numbers of An. gambiae
Full size table

Only a single 1014F haplotype was observed (Fig. 1), suggesting a common origin in the species and subsequent interspecific transfer. However it should be noted that our ability to resolve different haplotypes was constrained by the low levels of diversity at this locus and our small amplicon length (478bp).

Fig. 1
figure 1

A TCS plot of the three haplotypes present in the populations assayed. White colour represents An. gambiae while black colour represents An. arabiensis

Full size image

Discussion

This is the first report of Vgsc-1014F in Kenya, which appears to have emerged in both An. gambiae and An. arabiensis around 2012 and is confirmed via DNA sequencing in multiple samples. The gene has previously been observed in Uganda [18], then much later in Tanzania [19] and now in Kenya. We have developed this report to alert researchers and programmatic staff to the presence of Vgsc-1014F mutation in these two important Anopheles vectors so that they can modify their resistance marker screening procedures. It is important therefore that we start monitoring allele and genotype frequencies so that we can assess their impact in an area of high bednet ownership.