Directionally selected cytochrome P450 alleles are driving the spread of pyrethroid resistance in the major malaria vector Anopheles funestus

Abstract

Pyrethroid insecticides are critical for malaria control in Africa. However, resistance to this insecticide class in the malaria vector Anopheles funestus is spreading rapidly across Africa, threatening the success of ongoing and future malaria control programs. The underlying resistance mechanisms driving the spread of this resistance in wild populations remain largely unknown. Here, we show that increased expression of two tandemly duplicated P450 genes, CYP6P9a and CYP6P9b, is the main mechanism driving pyrethroid resistance in Malawi and Mozambique, two southern African countries where this insecticide class forms the mainstay of malaria control. Genome-wide transcription analysis using microarray and quantitative RT-PCR consistently revealed that CYP6P9a and CYP6P9b are the two genes most highly overexpressed (>50-fold; q < 0.01) in permethrin-resistant mosquitoes. Transgenic expression of CYP6P9a and CYP6P9b in Drosophila melanogaster demonstrated that elevated expression of either of these genes confers resistance to both type I (permethrin) and type II (deltamethrin) pyrethroids. Functional characterization of recombinant CYP6P9b confirmed that this protein metabolized both type I (permethrin and bifenthrin) and type II (deltamethrin and Lambda-cyhalothrin) pyrethroids but not DDT. Variability analysis identified that a single allele of each of these genes is predominantly associated with pyrethroid resistance in field populations from both countries, which is suggestive of a single origin of this resistance that has since spread across the region. Urgent resistance management strategies should be implemented in this region to limit a further spread of this resistance and minimize its impact on the success of ongoing malaria control programs.

Fig. 1.

Quantitative PCR results: Differential expression by qRT-PCR of 10 genes up-regulated in microarray assays in Mozambique (A) and in Malawi (B). Relative fold-change in gene copy number for CYP6P9a, CYP6P9b, CYP6P4a, and CYP6P4b in Mozambique (C) and in Malawi (D). Error bars represent SD (n = 3).

Fig. 2.

Bioassays results with transgenic strains for CYP6P9a and CYP6P9b. (A) Test with deltamethrin on transgenic Act5C-CYP6P9a flies (Exp-CYP6P9a) and three control strains [two parental (UAS-CYP6P9a and GAL4-Actin) and the progeny from the cross between the Gal4-Act5C females and the attP40 males (which do not overexpress the P450 transgene) (Cont-NO)]. (B) Test with permethrin on transgenic Act5C-CYP6P9a. (C) Test with deltamethrin on transgenic Act5C-CYP6P9b. (D) Test with permethrin on transgenic Act5C-CYP6P9b.

Fig. 3.

Metabolic activity of CYP6P9b. (A) CO-difference spectrum of E. coli membranes expressing CYP6P9b. (B) The proportion of 20 µM insecticide cleared by 0.1 µM P450 with 0.8 µM cyt b5 in the presence of NADPH is indicated by bar height. Error bars represent SD (n = 3). (C) Time course of deltamethrin and permethrin metabolism by CYP6P9b.

Fig. 4.

Schematic representation of haplotypes of CYP6P9a and CYP6P9b genes between the resistant mosquitoes from Mozambique and Malawi and the susceptible FANG strain. (A and B) The polymorphic amino acid positions for both CYP6P9a and CYP6P9b, respectively. (C) Neighbor-joining tree of CYP6P9a and (D) CYP6P9b showing two clades specific to each phenotype. A number has been given to each haplotype preceded by MAL, MOZ, or FANG if it is unique to Malawi, Mozambique, or FANG strains, respectively. The column Nb indicates the number of individuals sharing the haplotype.

Fig. 5.

A 95% parsimony network of CYP6P9a (A) and CYP6P9b (B) haplotypes between susceptible and resistant mosquitoes from both Mozambique and Malawi. Haplotypes are represented as an oval or a rectanglar shape, scaled to reflect their frequencies. Lines connecting haplotypes and each node represent a single mutation event (respective polymorphic positions are given above branches). White shapes represent haplotypes unique in susceptible mosquitoes; grey shapes represent haplotypes predominantly found in resistant mosquitoes but also in some dead mosquitoes; dark grey shapes represent haplotypes unique to resistant mosquitoes. For both genes, some susceptible haplotypes with >20 mutation differences from others could not be linked to the major network.