Midgut microbiota of the malaria mosquito vector Anopheles gambiae and interactions with Plasmodium falciparum infection

Abstract

The susceptibility of Anopheles mosquitoes to Plasmodium infections relies on complex interactions between the insect vector and the malaria parasite. A number of studies have shown that the mosquito innate immune responses play an important role in controlling the malaria infection and that the strength of parasite clearance is under genetic control, but little is known about the influence of environmental factors on the transmission success. We present here evidence that the composition of the vector gut microbiota is one of the major components that determine the outcome of mosquito infections. A. gambiae mosquitoes collected in natural breeding sites from Cameroon were experimentally challenged with a wild P. falciparum isolate, and their gut bacterial content was submitted for pyrosequencing analysis. The meta-taxogenomic approach revealed a broader richness of the midgut bacterial flora than previously described. Unexpectedly, the majority of bacterial species were found in only a small proportion of mosquitoes, and only 20 genera were shared by 80% of individuals. We show that observed differences in gut bacterial flora of adult mosquitoes is a result of breeding in distinct sites, suggesting that the native aquatic source where larvae were grown determines the composition of the midgut microbiota. Importantly, the abundance of Enterobacteriaceae in the mosquito midgut correlates significantly with the Plasmodium infection status. This striking relationship highlights the role of natural gut environment in parasite transmission. Deciphering microbe-pathogen interactions offers new perspectives to control disease transmission.

Figure 1. Comparison of bacterial diversity for the three 16S libraries at the genus level.

Tag abundance was compared for three 16S libraries, and the graph shows the bacterial flora of six mosquito midguts. The three 16S libraries were obtained using primer sets targeting different 16S domains, as described in the Materials and Methods section. Only the most abundant categories (>2%) were considered. The S1 library only reached 95%, showing this domain allowed identifications for a greater number of minor clades. For mosquito NKD97, S2 and S3 primer sets only allowed the identification at the Enterobacteriaceae family level, whereas S1 reached the assignment at the genus level, Serratia.

Figure 2. Relative abundance of the different bacterial classes within each mosquito midgut sample.

Mvan and NKD (Nkolondon) indicate the geographical origin of mosquitoes, NG being mosquitoes of the laboratory colony Ngousso. Pf+ and Pf− designate the P. falciparum infection status of the challenged mosquitoes, positive and negative, respectively. Only class rank groups that represented >1% of the total reads, and identified in at least 30% of mosquitoes, are shown. “Unclass” represents tags that could not be assigned to the class level, and were grouped into a higher taxonomical rank.

Figure 3. Redundancy analysis for gut bacterial communities (taxonomic rank = class) in field and laboratory mosquitoes.

The length of arrows indicates the strength of correlation between the variable and the ordination scores. Blue arrow: bacterial classes, green arrow: environmental variables. The Monte Carlo permutation test was used to test the statistical significance of the relationship between environmental variables and the bacterial classes. The “Flavo” (Flavobacteriaceae) segregates with “labo” environmental variable, “Alpha” (Alphabacteriaceae) with the “NKD” environmental variable (P<0.05). All other bacterial classes segregate along the second axis, with the “Mvan” environmental variable.

Figure 4. Relative abundance of Enterobacteriaceae in P. falciparum non-infected (Pf−) or P. falciparum infected (Pf+) mosquitoes.

The Enterobacteriaceae loads are significantly higher in P. falciparum infected mosquitoes (P = 0.004; Mann–Whitney test). The boxes represent the interquartile range (25–75^th percentile), and the line within each box corresponds to the median value.