Conservation and divergence of known apicomplexan transcriptional regulons

Abstract

Background: The apicomplexans are a diverse phylum of parasites causing an assortment of diseases including malaria in a wide variety of animals and lymphoproliferation in cattle. Little is known about how these varied parasites regulate their transcriptional regulons. Even less is known about how regulon systems, consisting of transcription factors and target genes together with their associated biological process, evolve in these diverse parasites.

Results: In order to obtain insights into the differences in transcriptional regulation between these parasites we compared the orthology profiles of putative malaria transcription factors across species and examined the enrichment patterns of four binding sites across eleven apicomplexans. About three-fifths of the factors are broadly conserved in several phylogenetic orders of sequenced apicomplexans. This observation suggests the existence of regulons whose regulation is conserved across this ancient phylum. Transcription factors not broadly conserved across the phylum are possibly involved in regulon systems that have diverged between species. Examining binding site enrichment patterns in light of transcription factor conservation patterns suggests a second mode via which regulon systems may diverge - rewiring of existing transcription factors and their associated binding sites in specific ways. Integrating binding sites with transcription factor conservation patterns also facilitated prediction of putative regulators for one of the binding sites.

Conclusions: Even though transcription factors are underrepresented in apicomplexans, the distribution of these factors and their associated regulons reflect common and family-specific transcriptional regulatory processes.

Figure 1

Conservation patterns of putative Plasmodium falciparum transcription factors. Orthologs predicted via reciprocal best BLAST hits (RBH) are indicated with light blue and those predicted by OrthoMCL are indicated with purple. 42.24% of the transcription factors do not have orthologs in non-Plasmodium apicomplexans demonstrating order-level specificity of transcriptional networks. ** The figure does not incorporate RBH information for non-apicomplexans. It does however include the associated orthology information for OrthoMCL.

Figure 2

Assessing conservation (enrichment) of binding sites across species. a. The best instance of the binding site being studied (one of the four in Figure 3) is identified upstream of each gene in the genome. b. The binding site is then permuted 100 times to generate 100 random sites with the same base composition as the real binding site. c. The best instance of each permuted site in the regions upstream of each gene is located in the same genome. d. The distribution of scores for the real site is compared to that of the permuted sites and score at the top one percentile of permuted sites is used as a cutoff for identifying high confidence hits for the real site. e. A p-value is computed representing the statistical significance of the overlap between genes whose upstream regions have hits for the real site and the list of genes in the process expected to be regulated by the site. f. The enrichment analysis (steps a-e) is repeated in all genomes and enrichment p-values are compared across species.

Figure 3

Transcription factor binding sites considered in this study. The functional contexts considered in the study are highlighted in blue. The transcription factor for the site, if known, and key publications are also presented.

Figure 4

Conservation of the four regulons. Regulons have generally diverged across the apicomplexan phylum but genera-related conservation patterns are evident. This table presents enrichment p-values for each of the four regulons (rows) in each of the eleven species studied (columns). Species names (column headings) are abbreviated as presented in Table 1. Significant enrichment p-values are given in boldface. If the known factor for a particular regulon is conserved in a species (as determined in earlier analysis of conservation of patterns of putative transcription factors) we highlight the cell for that entry in light blue.

Figure 5

Comparison of the expression of two putative G-box regulators and average expression of putative targets. The highest-ranking candidate positive regulator, PFC0690c, has a correlation of 0.651 with the set of ribosomal genes with high confidence G-box hits (n = 11). The highest-ranking negative regulator, PF11_0442, has a correlation of -0.838 with these same genes. In both cases, the expression of these putative targets lags that of the candidate regulators.

Figure 6

The ribosomal regulon system in apicomplexans. a) Enrichment p-values of three binding sites upstream of ribosomal genes. Stars are color-coded to represent the three transcription factor binding sites at the bottom of the image. The G-box, sporozoite and TRP-1 sites are 'starred' with green, purple and yellow respectively. b) Schematic phylogeny of apicomplexan species considered. Stars below a particular species indicate that the corresponding binding site likely regulates ribosomal genes in that particular species.