Characterization of Plasmodium developmental transcriptomes in Anopheles gambiae midgut reveals novel regulators of malaria transmission

Abstract

The passage through the mosquito is a major bottleneck for malaria parasite populations and a target of interventions aiming to block disease transmission. Here, we used DNA microarrays to profile the developmental transcriptomes of the rodent malaria parasite Plasmodium berghei in vivo, in the midgut of Anopheles gambiae mosquitoes, from parasite stages in the midgut blood bolus to sporulating oocysts on the basal gut wall. Data analysis identified several distinct transcriptional programmes encompassing genes putatively involved in developmental processes or in interactions with the mosquito. At least two of these programmes are associated with the ookinete development that is linked to mosquito midgut invasion and establishment of infection. Targeted disruption by homologous recombination of two of these genes resulted in mutant parasites exhibiting notable infection phenotypes. GAMER encodes a short polypeptide with granular localization in the gametocyte cytoplasm and shows a highly penetrant loss-of-function phenotype manifested as greatly reduced ookinete numbers, linked to impaired male gamete release. HADO encodes a putative magnesium phosphatase with distinctive cortical localization along the concave ookinete periphery. Disruption of HADO compromises ookinete development leading to significant reduction of oocyst numbers. Our data provide important insights into the molecular framework underpinning Plasmodium development in the mosquito and identifies two genes with important functions at initial stages of parasite development in the mosquito midgut.

Figure 1

Plasmodium berghei developmental co-expression clusters in the A. gambiae midgut. Transcriptional co-expression gene clusters that together depict the P. berghei developmental expression in the mosquito midgut from early blood bolus to late oocyst stages. In generating these clusters, genes exhibiting significant (P < 0.05) differential regulation between any two time points, assessed with one-way ANOVA, and at least 0.8-fold difference in a log₂ scale between their minimum and maximum expression after correction with the Benjamini-Hochberg hypergeometric test were grouped using a combination of SOM and K-means clustering. The Y-axis scale shows expression relative to the standard reference sample in log₂-transformed values; horizontal dashed lines indicate the signal levels of the standard reference. Black solid lines indicate the average expression of all the genes in each cluster and grey areas indicate their range of expression. Genes with known developmental expression profiles are shown with coloured solid lines. The number of genes in each cluster is indicated at the top left corner of each graph.

Figure 2

Expression and subcellular localization of P. berghei GAMER and HADO.A, B. Transcriptional expression profiles of GAMER and HADO, respectively, throughout Plasmodium development in the mosquito midgut. The DNA microarray (purple line) and qRT-PCR (blue line) profiles are shown. The vertical axis indicates the relative expression. Quantitative real-time PCR data are the mean of three biological replicates (fully independent infections) each derived from two technical replicates. Data were normalized to the level of constitutively expressed GFP transcripts. Error bars indicate standard error of the mean.C. RT-PCR analysis of GAMER and HADO in asexual blood stages using the non-gametocyte producing strain (HPE); mixed blood stages (MBS); activated (A) and non-activated (nA) gametocytes (Gc); 1 h, 3 h and 8 h zygotes; and non-purified (nP) and purified (P) in vitro produced ookinetes (Ook). The analysis was complemented with in vivo 5, 10 and 13 day (d) oocysts. P28, AMA1 and CHT1 served as stage-specific and loading controls.D. Western blot analysis of wt parasites using an anti-GAMER peptide antibody (a-GAMER). Knockout Δpbgamer parasites were used to control for non-specific signal. Anti-TAT1 (a-TAT1) antibody was used as internal control.E. Western blot analysis of wt ookinetes using an anti-HADO peptide antibody. Δpbhado ookinete protein extracts were used to control for non-specific signal.F, G. Immunofluorescence assays of purified P. berghei ANKA 2.34 wt mature, non-activated (nA) gametocytes (upper panels) labelled for P28 (red), DAPI (blue) and GAMER (green; F) or HADO (green; G) respectively. Bright field (BF) visualization is also shown. IFA of Δpbgamer and Δpbhado gametocytes, respectively, served as a negative control. Images were taken from confocal sections of fixed parasites.

Figure 3

Phenotypic analysis of Δpbgamer and Δpbhado mutant parasites.A. Asexual blood stage growth corresponding to the percentage of infected RBCs.B. Gametocytemia, indicating the percentage of infected RBCs that are at the gametocyte stage.C. Exflagellation assays, showing the percentage of male gametocytes that form exflagellation centres.D. Gametocyte-to-ookinete conversion ratio. Both the exflagellation and the gametocyte-to-ookinete conversion assays were repeated five times and statistical significance determined with a two tailed, unpaired Student's t-test.E. Distribution of P28.F, G. Oocyst loads at day 10 pi in the midguts of A. gambiae and A. stephensi mosquitoes respectively. Stars indicate statistical significance determined with Mann–Whitney U-test. The median is shown with a red line. ***P < 0.0001; **P < 0.001, *P < 0.05. n, number of independent biological replicates. Error bars indicate standard error of mean.

Figure 4

Δpbgamer exflagellation and genetic complementation assays.A. Wt and Δpbgamer exflagellation centres at 1, 10, 20 and 30 min post-male gametocyte activation. Images were obtained by time-lapse microscopy at 43× magnification. Incomplete detachment of male gametes (white arrows) is observed for Δpbgamer as opposed to complete release of male gametes (white arrowhead) in wt parasites at 30 min post-male gametocyte activation.B. Gametocyte-to-ookinete conversion ratios in genetic crosses of Δpbgamer with either male-, Δpbs48/45, or female-deficient lines, Δpbs47, showing no phenotype rescue. Genetic crosses between Δpbs48/45 and Δpbs47 were used as a positive control. The arithmetic mean and standard error of the mean (SEM) are shown. Three biological experiments were performed for each genetic cross group and P-values were calculated with the Student's t-test.C. Oocyst distribution in the midguts of A. gambiae mosquitoes at day 9 pi, following genetic crossing of Δpbgamer with either Δpbs48/45 or Δpbs47-deficient lines. Three biological experiments were performed for each of the genetic crosses group and statistical significance was determined with the Mann–Whitney U-test. The median is shown with a black line. Stars indicate statistical significance; ***P < 0.0001.