A Knockout Screen of ApiAP2 Genes Reveals Networks of Interacting Transcriptional Regulators Controlling the Plasmodium Life Cycle

Abstract

A family of apicomplexa-specific proteins containing AP2 DNA-binding domains (ApiAP2s) was identified in malaria parasites. This family includes sequence-specific transcription factors that are key regulators of development. However, functions for the majority of ApiAP2 genes remain unknown. Here, a systematic knockout screen in Plasmodium berghei identified ten ApiAP2 genes that were essential for mosquito transmission: four were critical for the formation of infectious ookinetes, and three were required for sporogony. We describe non-essential functions for AP2-O and AP2-SP proteins in blood stages, and identify AP2-G2 as a repressor active in both asexual and sexual stages. Comparative transcriptomics across mutants and developmental stages revealed clusters of co-regulated genes with shared cis promoter elements, whose expression can be controlled positively or negatively by different ApiAP2 factors. We propose that stage-specific interactions between ApiAP2 proteins on partly overlapping sets of target genes generate the complex transcriptional network that controls the Plasmodium life cycle.

Keywords: Apetala 2; ap2-l; ap2-o; ap2-o2; ap2-o3; ap2-o4; ap2-sp; ap2-sp2; ap2-sp3; oocyst.

Graphical abstract

Figure 1

Systematic Phenotyping of ApiAP2 Gene KOs in P. berghei (A) Schematic representation of P. berghei ApiAP2 genes and KO clones obtained in this study. Expression data for ring (R), trophozoite (T), schizont (S), gametocyte (G), and ookinete (O) stage are from Otto et al. (2014). (B) Summary of a high-level phenotypic screen of eleven ApiAP2 KOs. Phenotypes were called if parasite numbers were reduced >80% or increased >20% compared to a wild-type control studied in parallel. Experiments were performed as triplicates on at least two separate occasions. Previously undescribed phenotypes were confirmed with independently created KO clones. For the complete data, see Data S1.

Figure 2

STM Analysis of ApiAP2 Family and Detailed Phenotyping of Four Mutants that Fail to Produce Infectious Ookinetes (A) Competitive fitness of 16 ApiAP2 KOs (purple) as determined by barcode counting following co-transfection of barcoded vectors. Growth rates were determined from barcode counts and are expressed relative to four mutants with wild-type growth (green). Known attenuated control mutants are shown in orange. Relative growth rates are given as the arithmetic mean of measurements taken on days 5, 6, 7, and 8 post-transfection and error bars show SDs from three transfections. Asterisk (^∗) indicates significantly different from the pool of reference vectors with no growth effect (p < 0.05). (B) Oocysts per midgut 7 days after mosquitoes had fed on infected mice. At least 15 mosquitoes were dissected per group and arithmetic means ± SE are shown. (C) Male and female gametocytemia determined on Giemsa-stained blood films on day 3 post-infection. Error bars show SDs from three infected mice. (D) Distribution in 20 hr ookinete cultures of different parasite forms expressing the macrogamete/ookinete marker P28. At least 100 P28-positive cells were counted in each sample. Error bars show SDs from three cultures. (E) Sexual stage morphology 20 hr after gametocyte activation in culture as seen on Giemsa-stained blood films (top panels) and by live IFA using a Cy3-conjugated P28 antibody (bottom panels). Scale bar, 5 μm.

Figure 3

In-Depth Phenotypic Analysis of Three ApiAP2 Genes Required for Sporozoite Production (A) Midgut oocysts and salivary gland sporozoites in mosquitoes fed on infected mice. Error bars show SDs from 15 midguts (day 7 post-feeding) or 3 batches of 10 pairs of glands (day 21 post-feeding), respectively. (B) Representative fluorescence micrographs of wild-type and ap2-sp2-infected mosquito midguts showing different stages of sporogony as patterns of cytosolic GFP expression in oocysts on day 12 post-feeding (left; scale bar, 20 μm). The bar chart shows a quantitation of sporogonic stages averaged from ten midguts from two independent experiments. (C) Light and fluorescence micrographs of typical oocysts showing incomplete sporogony in ap2-sp2. Hoechst 33342 dye shows DNA. Scale bar, 5 μm. (D) TEMs showing normal sporogony in wild-type oocysts (upper panel) and incomplete budding of ap2-sp2 sporozoites from sporogonic islands (lower panel). Scale bar, 5 μm. (E) Prepatency of blood stage infections in days following administration of sporozoites by intravenous injection, or by mosquito bite. Pooled data from independent experiments using two different ap2-sp3 clones are shown. Prevalence is shown in brackets (infected mice/total).

Figure 4

Comparative Transcriptomics of ApiAP2 KOs at Different Life Cycle Stages (A) PC analysis of all transcriptome samples. Colors indicate strains, and shapes indicate life stages. (B) Number of differentially expressed genes (absolute fold change > 2 and adjusted p value < 0.01) in three ookinete mutants at different life stages. (C) Heatmap showing the expression of 4751 P. berghei genes through different strain/stage combinations. The gene transcripts are ordered according to their expression patterns using Ward hierarchical clustering. The borders of 49 gene expression clusters are marked.

Figure 5

Characterization of Selected Co-expression Clusters For each stage and mutant, the average relative expression across all genes in a cluster is given. N gives the number of genes in a cluster. GO terms and functional gene groups are shown if significantly enriched (false discovery rate [FDR]-adjusted p value < 0.05). Sequence motifs shown are significantly enriched within 2 kb upstream of the start codon, as compared with all genes. See Data S3 and S4 for a comprehensive analysis of all clusters.

Figure 6

A Double Mutant with ap2-g Identifies ap2-g2 as Repressor in Asexual Blood Stage (A) Heatmap showing expression of cluster 9 genes in blood stage cultures. Row-scaled reads per kilobase of transcript per million mapped reads (RPKM) counts for two replicates are shown for each parasite line: wild-type, both single mutants, and the ap2-g/ap2-g2 double mutant (dKO). Genes previously identified as strongly upregulated at the liver stage (Tarun et al., 2008) are marked. (B) Assessing the impact on schizont proteome and transcriptome of deleting ap2-g2 in a gametocyte-deficient ap2-g mutant. Normalized transcript and protein ratios for 2,169 genes detected as both transcript and protein are plotted against each other. See Data S5 for the underlying datasets. (C) Fluorescence micrographs showing aberrant expression and localization of the sporozoite surface protein CSP in schizonts of the ap2-g/ap2-g2 double mutant. Fixed and permeabilized infected RBCs were stained with Cy3-conjugated CSP monoclonal antibody. Schizonts from an ap2-g2 single mutant also expressed CSP, while no fluorescence was observed in ap2-g schizonts (Figure S6A). Scale bar, 5 μm.