Directional gene expression and antisense transcripts in sexual and asexual stages of Plasmodium falciparum

Abstract

Background: It has been shown that nearly a quarter of the initial predicted gene models in the Plasmodium falciparum genome contain errors. Although there have been efforts to obtain complete cDNA sequences to correct the errors, the coverage of cDNA sequences on the predicted genes is still incomplete, and many gene models for those expressed in sexual or mosquito stages have not been validated. Antisense transcripts have widely been reported in P. falciparum; however, the extent and pattern of antisense transcripts in different developmental stages remain largely unknown.

Results: We have sequenced seven bidirectional libraries from ring, early and late trophozoite, schizont, gametocyte II, gametocyte V, and ookinete, and four strand-specific libraries from late trophozoite, schizont, gametocyte II, and gametocyte V of the 3D7 parasites. Alignment of the cDNA sequences to the 3D7 reference genome revealed stage-specific antisense transcripts and novel intron-exon splicing junctions. Sequencing of strand-specific cDNA libraries suggested that more genes are expressed in one direction in gametocyte than in schizont. Alternatively spliced genes, antisense transcripts, and stage-specific expressed genes were also characterized.

Conclusions: It is necessary to continue to sequence cDNA from different developmental stages, particularly those of non-erythrocytic stages. The presence of antisense transcripts in some gametocyte and ookinete genes suggests that these antisense RNA may play an important role in gene expression regulation and parasite development. Future gene expression studies should make use of directional cDNA libraries. Antisense transcripts may partly explain the observed discrepancy between levels of mRNA and protein expression.

Figure 1

Density distribution of reads from bidirectional and strand-specific libraries of late trophozoite, schizont, gametocyte II, and gametocyte V. The sense fraction was calculated based on the ratio between sense and total number of reads present at each exon in the genome (total of 14,777 exons). Density represents the number of genes that have a particular ratio of forward reads over total number of reads for that gene. Strand-specific libraries (solid lines) show directionality characterized by high frequencies of exons with fraction values close to 1, whereas those from bidirectional libraries (dashed lines) have the majority of exons with sense fractions between 0.2 and 0.5.

Figure 2

Scatter plots showing exons with various numbers of antisense reads at different fractions of sense reads from the strand-specific libraries. The fraction of sense reads was plotted against the number of antisense reads from late trophozoites (LT), schizonts (Sc), gametocyte stage II (GII), and gametocyte stage V (GV) for each of the 14,777 exons in the genome. Green dots represent exons with a high number of antisense reads (> 150) and low level of sense fraction (< 0.3), and reds dots are those with high levels of sense fraction (> 0.7) and antisense reads (> 150 reads).

Figure 3

Change in transcript direction in gene PFA0245w, a putative transporter, during development from asexual to sexual stages. LT, late trophozoite; Sc, schizont; GII, gametocyte stage II; GV, gametocyte stage V. The black arrow bars indicate open reading frame direction. The blue color indicates reads in sense direction, and those in red are reads in antisense direction. Note that there are more red reads in Sc than in LT, and at the 5' end in the GV than in GII. The gap without coverage represents a repetitive region where reads were removed.

Figure 4

Distribution of antisense:sense ratio (AS:S) across the 14 parasite chromosomes and an example of 'run-through' transcript in tail-to-tail orientation. (A) Distribution of antisense:sense (AS:S) ratios across the 14 chromosomes (Ch1-14) in the GV stage. (B) A pair of genes showing transcripts possibly 'running through' into a neighboring gene in opposite orientation, possibly suppressing the expression of the neighboring gene. MAL7P1.175b is not annotated in the current Plasmodium falciparum genome.

Figure 5

Comparison of RNA-seq reads in matching intron-exon junctions of predicted gene models. (A) Relationship of the numbers of reads bridging an intron and the chance of the intron matching one intron in the predicted gene models. (B) The numbers of introns matching those in the predicted gene models decrease as the numbers of reads bridging an intron increase. (C) Plot of sensitivity against positive predictive value (PPV). Sensitivity is defined as the numbers of matched intron-exon junctions between the predicted gene models and RNA-seq/numbers of intron-exon junctions from predicted gene models, and PPV is the numbers of matched intron-exon junctions between predicted gene models and RNA-seq/total numbers of intron-exon junctions identified from RNA-seq. Red lines, from our strand-specific libraries; green, data from Bartfai et al. [19]; black, data from Otto et al. [17].

Figure 6

Comparison of expression patterns between our data and those from previous studies. Samples starting with "W" are from Le Roch et al. [24,25]. Samples starting with "O" are from Otto et al. [17], and the numbers are hours when RNA samples were isolated after erythrocyte invasion. Those with "B" are from Bartfai et al. [19], and again the numbers are hours after invasion when the RNAs were isolated. Green indicates positive correlation, and red suggests negative correlation. R, ring; ER, early ring; LR, late ring; ET, early trophozoite; LT, late trophozoite; Sc, schizont; ESc, early schizont, LSc, late schizont; M, merozoite; GII, gametocyte II; GV, gametocyte V; G1-G13, day 1 to day 13 gametocyte samples.

Figure 7

An example of stage-specific alternatively spliced genes (exon skipping) in two different developmental stages. (A) The predicted gene model of MAL13P1_42 (only part of the gene is shown); (B) reads from the bidirectional ring library; (C) reads from bidirectional late trophozoite library. The red arrowhead points to reads with a skipped exon.