Permethrin and DDT resistance in Anopheles gambiae s.s. associated with a leucine-serine knockdown resistance (kdr) mutation in the voltage-gated sodium channel gene was discovered recently in western Kenya where a large scale permethrin-impregnated bed net (ITN) program has been implemented. Collections of An. gambiae s.l. were made from intervention and control villages prior to and after onset of the program. The kdr genotypes were determined using allele-specific polymerase chain reaction diagnostic tests. In An. gambiae s.s., the frequency of the kdr mutation prior to ITN introduction was approximately 3-4% in western Kenya and zero in samples from the coast. After ITN introduction, the kdr mutation increased in ITN and neighboring villages from approximately 4% to approximately 8%, but remained unchanged in villages at least 20 km distant and was not detected in coastal Kenya. The identical leucine-serine mutation was found in a single An. arabiensis individual among 658 tested. The leucine-phenylalanine kdr mutation common in west African An. gambiae populations was not detected in An. gambiae s.l. from Kenya. Implications for the population structure and control of An. gambiae are discussed.

The M and S molecular forms of Anopheles gambiae are undergoing speciation as they adapt to heterogeneities in the environment, spreading malaria in the process. We hypothesized that their divergence despite gene flow is facilitated by reduced recombination at the centromeric (proximal) end of the X chromosome. We sequenced introns from 22 X chromosome genes in M and S from two locations of West Africa where the forms are sympatric. Generally, in both forms nucleotide diversity was high distally, lower proximally, and very low nearest the centromere. Conversely, differentiation between the forms was virtually zero distally and very high proximally. Pairwise comparisons to a close relative, the sibling species Anopheles arabiensis, demonstrated uniformly high divergence regardless of position along the X chromosome, suggesting that this pattern is not purely mechanical. Instead, the pattern observed for M and S suggests the action of divergent natural selection countering gene flow only at the proximal end of the X chromosome, where recombination is reduced. Comparison of sites with fixed differences between M and S to the corresponding sites in A. arabiensis revealed that derived substitutions had been fixed in both forms, further supporting the hypothesis that both have been under selection. These derived substitutions are fixed in the two West African samples and in samples of S from western and coastal Kenya, suggesting that selection occurred before the forms expanded to their current ranges. Our findings are consistent with a role for suppressed genetic recombination in speciation of A. gambiae.R educed recombination contributes to the persistence of species or emerging species in the face of gene flow (1, 2). Despite some interbreeding and hybrid formation, species-specific regions of the genome can be preserved from introgression and homogenization of different genetic backgrounds if they are not subject to crossing over. Recent elaborations of this concept in organisms as diverse as humans, sunflowers, fruit flies, and mosquitoes invoke chromosomal inversions because recombination is effectively suppressed between the break points of chromosomal inversion heterozygotes (3-6). If captured by inversions, genes involved in assortative mating and species-specific ecological adaptations would remain associated longer relative to the case of free recombination. For pairs of species that have diverged in parapatry or sympatry, these models predict (i) that they are likely to differ by fixed inversions, (ii) that genes involved in reproductive isolation and species-specific adaptations should preferentially map to these inversions, and (iii) that significantly greater genetic divergence will accumulate in rearranged versus colinear regions. These predictions seem to be upheld for the group of mosquito sibling species known as the Anopheles gambiae complex (7, 8) but not for A. gambiae itself.A. gambiae, the primary vector of malaria in subSaharan Africa, is undergoing speciation. The incipient species (designate...

Emerging species within the primary malaria vector Anopheles gambiae show different ecological preferences and significant prezygotic reproductive isolation. They are defined by fixed sequence differences in X-linked rDNA, but most previous studies have failed to detect large and significant differentiation between these taxa elsewhere in the genome, except at two other loci on the X chromosome near the rDNA locus. Hypothesizing that this pericentromeric region of the X chromosome may be accumulating differences faster than other regions of the genome, we explored the pattern and extent of differentiation between A. gambiae incipient species and a sibling species, A. arabiensis, from Burkina Faso, West Africa, at 17 microsatellite loci spanning the X chromosome. Interspecific differentiation was large and significant across the entire X chromosome. Among A. gambiae incipient species, we found some of the highest levels of differentiation recorded in a large region including eight independent loci near the centromere of the X chromosome. Outside of this region, no significant differentiation was detected. This pattern suggests that selection is playing a role in the emergence of A. gambiae incipient species. This process, associated with efficient exploitation of anthropogenic modifications to the environment, has public health implications as it fosters the spread of malaria transmission both spatially and temporally.

In the malaria vector Anopheles gambiae, alternative arrangements of chromosome 2 (2La and 2L+(a)) vary in relative frequency along clines of aridity, suggesting the action of natural selection on targets within the inversion. Our long term goal of detecting such targets depends in part on the level of genetic exchange between arrangements. Accordingly, we estimated recombination rates on 2L from the backcross progeny of 2La/+(a) heterokaryotypes and as a control, from 2L+(a) homokaryotypes. In homokaryotypes, the recombination rate was uniform at ~2.0 centimorgans per megabase (cM/Mb). In heterokaryotypes, recombination within the rearranged region was reduced to < 0.5 cM/Mb, with slightly higher but nevertheless reduced levels (< 1.0 cM/Mb) flanking the rearrangement. Yet, gene exchange was recorded between nearly all markers, including those very near the distal inversion breakpoint. These results suggest that reduced recombination is a necessary but not sufficient mechanism for genetic isolation between alternative arrangements, and that the targets of natural selection can be identified against the different chromosomal backgrounds.