Quantitative Trait Loci for Refractoriness of Anopheles gambiae to Plasmodium cynomolgi B

See allHide authors and affiliations

Science  18 Apr 1997:
Vol. 276, Issue 5311, pp. 425-428
DOI: 10.1126/science.276.5311.425


The severity of the malaria pandemic in the tropics is aggravated by the ongoing spread of parasite resistance to antimalarial drugs and mosquito resistance to insecticides. A strain of Anopheles gambiae, normally a major vector for human malaria in Africa, can encapsulate and kill the malaria parasites within a melanin-rich capsule in the mosquito midgut. Genetic mapping revealed one major and two minor quantitative trait loci (QTLs) for this encapsulation reaction. Understanding such antiparasite mechanisms in mosquitoes may lead to new strategies for malaria control.

Melanotic encapsulation, an immune reaction in which invading parasites are enclosed and destroyed within a melanin-rich capsule, is widespread among insects. Malaria parasites, which must develop into oocysts in the mosquito midgut, can also be encapsulated in some refractory vector strains, resulting in a block to disease transmission (1). The mechanism of parasite rejection is a key to the biology of interaction betweenPlasmodium and its vector, and an understanding of this mechanism may ultimately be useful in malaria control strategies such as mosquito population replacement using robust refractory strains.

Fully refractory and susceptible strains of A. gambiae have been selected for the ability to encapsulate or tolerate, respectively, oocysts of Plasmodium cynomolgi, a simian parasite. These strains respond similarly to most Plasmodium species, including the human pathogen P. falciparum (1). Many different paracentric chromosomal inversions are present inA. gambiae from both natural and laboratory-adapted strains (2). The previously described strains differ by a large paracentric inversion (2La) covering polytene divisions 23 through 26 of the left arm of chromosome 2. Refractoriness and susceptibility to P. cynomolgi B have been associated with the 2L+ a/+ a and2La/a karyotypes, respectively (3). To avoid the expected suppression of recombination by inversion polymorphism and to facilitate genetic analysis, we selected new refractory and susceptible strains bearing the same2L+ a/+ a arrangement (4). Their reciprocal crosses yielded generally refractory F1 female progeny; backcrosses (BCs) of reciprocal F1 progeny revealed no sex-linked genetic component to parasite encapsulation (5). The BC of F1 to the refractory strain produced progeny that were all highly refractory (Fig. 1A), indicating a dominant effect of the refractory allele (or alleles) of the loci involved (6).

Figure 1

Encapsulation of P. cynomolgi B is dominant and is controlled by a major locus. Distributions of traits are shown for BC female progeny from one family backcrossed to the refractory line (R/S × R/R) (A) and from the seven families backcrossed to the susceptible line (R/S × S/S) (B and C). Mosquitoes showing no infection could not be represented in (A) and (B) (7 and 18, respectively); they are shown as a white bar in (C).

BCs to the susceptible strain generated seven families (E1 through E7), with 19, 34, 36, 36, 25, 29, and 31 BC female progeny, that were blood-fed with infected P. cynomolgi B and then scored individually for intensity of infection and for encapsulation phenotype. Most of the BC progeny were either refractory or susceptible, which suggested that a single genetic locus is the major determinant of the encapsulation response (Fig. 1B). By contrast, the intensity of infection (normal plus encapsulated parasites) in the BC progeny varied broadly, which suggested that there is no simple genetic component of inheritance (Fig. 1C).

The microsatellite markers shown in Fig. 2 were heterozygous in the F1 female parent and thus genetically informative. The marker genotypes in families E1 to E5 were correlated with the encapsulation phenotypes (7, 8) by means of Mapmaker/QTL (9, 10). When these five families were analyzed together, two QTLs were identified (Fig. 2). The major one (Pen1, for Plasmodium encapsulation 1) was identified near marker H175, with a combined LOD score of 22.7 that explains ∼54% of the trait. A minor QTL (Pen2) was also identified near marker H758, with a LOD score of 4.1 that explains ∼13% of the trait. By correlation of the genetic and cytogenetic maps with a subset of markers mapped by in situ hybridization to the polytene chromosomes (2, 5, 7),Pen1 is probably located in division 8 of the right arm of chromosome 2 (2R), whereas Pen2 is in or near division 43 of the left arm of chromosome 3 (3L). Progressively decreasing LOD scores in regions of the same chromosome away from each QTL are expected because of linkage. Secondary peaks were also observed, which, if significant, might indicate additional QTLs. Therefore, the whole genome was rescanned, with Pen1 and Pen2 both fixed (that is, assuming the presence of these two QTLs and mapping the residual variation). This revealed only one additional significant QTL,Pen3, which is near marker H135 about 45 centimorgans (cM) from marker H175 and probably near division 14 on 2R. The total contribution from these three QTLs amounted to 70% of the trait. By contrast, the X chromosome showed no influence on the refractoriness phenotype (Fig. 2).

Figure 2

QTLs controlling the encapsulation response toP. cynomolgi B. The genetic map was generated by genotyping 150 progeny from families E1 to E5. A few closely linked markers appeared in an order different from that in the standard map (7). Asterisks indicate locations of the encapsulation QTLs; ▪ indicates a marginal peak in the infection LOD score (see Table 1). The approximate locations of the centromeres (C) and the arms of each chromosome (R, right; L, left) are labeled. Noninformative microsatellite markers are not shown.

Both the refractory and susceptible strains are polymorphic for the 2Rbc inversion, which encompasses polytene divisions 11B through 14A, and could affect the data for Pen3. We controlled for this variable by monitoring a nearby visible marker,collarless (4, 6). All F1 females were chosen to be c/c. Moreover, we reanalyzed data pooled from families derived from the same type of susceptible BC father, C/C (families E2, E3, and E6) orC/c (families E1, E4, and E5). In both cases,Pen3 was identified, linked most closely with eitherH135 or the next available marker, H770. Similarly, Pen3 was identified near H135 in family E7, which was derived from a c/c BC father and was analyzed using selected microsatellite markers. Thus, it is unlikely that the 2Rbc inversion polymorphism explains any of the QTLs. Analysis of all seven families combined yielded similar results (Table 1). Pen1 was located 1.5 cM from H175 (LOD = 36.0) and contributed 60% of the trait; Pen2 was mapped 8.0 cM from H758 and contributed 19%. Pen3 was identified 4.0 cM fromH135 when either Pen1 alone or Pen1and Pen2 were fixed. The combined actions ofPen1, Pen2, and Pen3 appeared to control 76% of the trait.

Table 1

Pen1, Pen2, and Pen3control the ability to encapsulate P. cynomolgi B oocysts. The underlined markers map closest to the three encapsulation QTLs; the boldface marker indicates a marginal peak in the infection LOD score (see Fig. 2). Numbers of progeny (out of 210 total from families E1 to E7) successfully genotyped for each marker and of progeny homozygous for the susceptible-derived alleles were subjected to a χ2 test; significant deviations from the expected 1:1 ratio are indicated. Some genotypings failed for technical reasons; the E7 maternal parent was uninformative for markerH788.

View this table:

These results were further validated by examining the distribution of infection and encapsulation phenotypes of genotypically sorted progeny (Fig. 3). Heterozygosity (the presence at marker H175 of an allele derived from the refractory parent) conferred a certain degree of phenotypic refractoriness; the combined effect of refractory markers at all three loci was much stronger. Conversely, homozygosity (the absence of a refractory allele at these loci) resulted in mosquitoes that were completely, or in a few individuals almost completely, susceptible.

Figure 3

In combination, the three QTLs control virtually completely the refractoriness to P. cynomolgi B. The numbers of normal and encapsulated oocysts in each mosquito are represented. Symbols • and ⧫ indicate homozygosity and heterozygosity, respectively, for the susceptible-derived allele only at marker H175 (upper panels) or at markersH175, H135, and H758 (lower panels). Note the substantial effect of the H175 genotype alone (upper left versus upper right panel) and the additive effects ofH135 and H758 (upper versus lower panels).

In contrast, no QTL was identified for the intensity of infection, except for a small peak (LOD = 2.38; 6% variance explained) near marker H769 on chromosome 2 (Table 1). None of the refractoriness QTLs controlled the intensity of infection by P. cynomolgi B. Moreover, only a marginal deviation from the expected 1:1 ratio of homozygosity to heterozygosity in the BC progeny was observed, between but not at Pen1 and Pen3(Table 1). This observation suggested that the QTLs were detrimental only to the parasites, not to the mosquito itself.

A genetic region that includes Pen1 has been shown (11) to be involved in the melanotic coating of abiotic Sephadex beads injected into the mosquito thorax (12). Interestingly, neither parasite nor bead encapsulation mapping experiments identified any locus within the 2La region. The reported association of 2L+ a with refractoriness (3) may be an artifact of the previously available strains (for example, a result of suppression of recombination). It is also possible that a locus within the 2La region is required for the expression of Pen1, Pen2, and Pen3but is not directly involved in encapsulation. The new strains are both2L+ a/+ a and may already carry the same permissive allele at this locus.

Melanotic encapsulation is only one of several types of refractoriness of anopheline mosquitoes to Plasmodium parasites (13). Another common type is manifested earlier, before or during parasite invasion of the midgut epithelium (14). Two QTLs each have been identified for the susceptibility of Aedes aegypti mosquitoes to P. gallinaceum (15) and Brugian worms (16). However, these Ae. aegypti QTLs control the intensity of parasite infection and thus differ from the Pen loci of A. gambiae. Hence, our results establish that the development of malaria parasites can be blocked by two independent refractory mechanisms that are both temporally and functionally different.

The nature of Pen1, Pen2, andPen3 is not known, although Pen3 maps in the general area where the prophenoloxidase gene is also located (17). In any case, the detailed localization of genes involved in the A. gambiae encapsulation response offers the opportunity to clone these genes positionally and to characterize the antiparasitic immune response of this vector at both the genetic and molecular levels.

  • * To whom correspondence should be addressed. E-mail: fhc1{at}


View Abstract

Navigate This Article