Molecular evolution of immune genes in the malaria mosquito Anopheles gambiae.

BACKGROUND: As pathogens that circumvent the host immune response are favoured by selection, so are host alleles that reduce parasite load. Such evolutionary processes leave their signature on the genes involved. Deciphering modes of selection operating on immune genes might reveal the nature of host-pathogen interactions and factors that govern susceptibility in host populations. Such understanding would have important public health implications. METHODOLOGY/FINDINGS: We analyzed polymorphisms in four mosquito immune genes (SP14D1, GNBP, defensin, and gambicin) to decipher selection effects, presumably mediated by pathogens. Using samples of Anopheles arabiensis, An. quadriannulatus and four An. gambiae populations, as well as published sequences from other Culicidae, we contrasted patterns of polymorphisms between different functional units of the same gene within and between populations. Our results revealed selection signatures operating on different time scales. At the most recent time scale, within-population diversity revealed purifying selection. Between populations and between species variation revealed reduced differentiation (GNBP and gambicin) at coding vs. noncoding- regions, consistent with balancing selection. McDonald-Kreitman tests between An. quadriannulatus and both sibling species revealed higher fixation rate of synonymous than nonsynonymous substitutions (GNBP) in accordance with frequency dependent balancing selection. At the longest time scale (>100 my), PAML analysis using distant Culicid taxa revealed positive selection at one codon in gambicin. Patterns of genetic variation were independent of exposure to human pathogens. SIGNIFICANCE AND CONCLUSIONS: Purifying selection is the most common form of selection operating on immune genes as it was detected on a contemporary time scale on all genes. Selection for "hypervariability" was not detected, but negative balancing selection, detected at a recent evolutionary time scale between sibling species may be rather common. Detection of positive selection at the deepest evolutionary time scale suggests that it occurs infrequently, possibly in association with speciation events. Our results provided no evidence to support the hypothesis that selection was mediated by pathogens that are transmitted to humans.

Polymorphism at the defensin gene in the Anopheles gambiae complex: testing different selection hypotheses.

Genetic variation in defensin, a gene encoding a major effector molecule of insects immune response was analyzed within and between populations of three members of the Anopheles gambiae complex. The species selected included the two anthropophilic species, An. gambiae and An. arabiensis and the most zoophilic species of the complex, An. quadriannulatus. The first species was represented by four populations spanning its extreme genetic and geographical ranges, whereas each of the other two species was represented by a single population. We found (i) reduced overall polymorphism in the mature peptide region and in the total coding region, together with specific reductions in rare and moderately frequent mutations (sites) in the coding region compared with non-coding regions, (ii) markedly reduced rate of non-synonymous diversity compared with synonymous variation in the mature peptide and virtually identical mature peptide across the three species, and (iii) increased divergence between species in the mature peptide together with reduced differentiation between populations of An. gambiae in the same DNA region. These patterns suggest a strong purifying selection on the mature peptide and probably the whole coding region. Because An. quadriannulatus is not exposed to human pathogens, identical mature peptide and similar pattern of polymorphism across species implies that human pathogens played no role as selective agents on this peptide.

Spatial and temporal variation in kinship among Anopheles gambiae (Diptera: Culicidae) mosquitoes.

Genetic relatedness (kinship) among Anopheles gambiae Giles female mosquitoes was assessed using microsatellite loci in five locations across Africa and in nine samples taken between 1994 and 1999 in western Kenya. We assessed variation among samples in kinship as well as the effect of distance on kinship. Relatedness within populations was low, and differences among samples taken at various times from one locale and from different locales were minimal. Mosquitoes collected from the same compound were slightly more closely related than those collected from different compounds. Our results suggest that newly emerged female siblings move relatively short distances into a few nearby compounds for blood feeding, but that they lay eggs in a more distant location. Kinship decreased nonlinearly with increasing distance. The strongest relationship between kinship and distance was observed for mosquitoes collected 0-3 km apart (-0.014/km, P < 0.001). The effect of distance decreased with increasing distance between mosquitoes; at 7 km or more, the kinship/distance slope approached zero and the intercept became negative, suggesting that beyond this range kinship does not decline with distance. This distance may thus represent the upper limit of the diameter of the basic reproductive unit. Nevertheless, the effect of distance on kinship is weak, reflecting extensive dispersal. Because females mate within days after emergence from larval habitats, where the likelihood of mating with a sibling is presumably highest, we propose a slight inbreeding effect.