Sequences conserved by selection across mouse and human malaria species

Abstract

Little is known, either experimentally or computationally, about the genomic sequence features that regulate malaria genes. A sequence conservation analysis of the malaria species P. falciparum, P. berghei, P. yoelii, and P. chabaudi could significantly advance knowledge of malaria gene regulation.

Results: We computationally identify intergenic sequences conserved beyond neutral expectations, using a conservation algorithm that accounts for the strong compositional biases in malaria genomes. We first quantify the composition-specific divergence at silent positions in coding sequence. Using this as a background, we examine gene 5' regions, identifying 610 blocks conserved far beyond neutral expectations across the three mouse malariae, and 81 blocks conserved as strongly across all four species (p < 10(-6)). Detailed analysis of these blocks indicates that only a minor fraction are likely to be previously unknown coding sequences. Analogous noncoding conserved blocks have been shown to regulate adjacent genes in other phylogenies, making the predicted blocks excellent candidates for novel regulatory functions. We also find three potential transcription factor binding motifs which exhibit strong conservation and overrepresentation among the rodent malariae.

Conclusion: A broader finding of our analysis is that less malaria intergenic sequence has been conserved by selection than in yeast or vertebrate genomes. This supports the hypothesis that transcriptional regulation is simpler in malaria than other eukaryotic species. We have built a public database containing all sequence alignments and functional predictions, and we expect this to be a valuable resource to the malaria research community.

Figure 1

Phylogeny of Rodent and Human Malariae. The three mouse malaria species, P. chabaudi, P. berghei, and P. yoelii, are closely related, with 0.757 of the alignable 4-fold sites identical across all three mouse species. P. falciparum has saturated divergence from these three, with 0.359 of the alignable 4-fold sites identical across all four species. AT sites are much more commonly identical than GC sites in each grouping.

Figure 2

A 40 bp region upstream of the P. yoelii gene PY02186 conserved beyond the neutral background. The region is shown with corresponding orthologous sequences of P. chabaudi, P. berghei. The conservation pattern of the highlighted block – in addition to 609 other blocks in the P. yoelii genome – has a less than 10^-6 probability of occurring by chance, based on silent site substitution patterns. While the unhighlighted sections are mostly conserved as well, the highlighted region stands out for its high density of conserved G's and C's, which are rare (6.3%) in 4-fold degenerate positions.

Figure 3

Upstream blocks conserved beyond neutral expectations among P. falciparum and the three mouse malaria species. These blocks were identified by searching the aligned regions upstream of orthologous genes and are strong candidates for having novel functions. These are (A) at 275 bp upstream of PF14_0091, a hypothetical gene, (B) at 81 bp upstream of PF14_0682, a hypothetical gene, (C) at 185 bp upstream of PF14_0212, a hypothetical gene, (D) at 275 bp upstream of PF14_0437, a helicase truncated putative gene, and (E) at 1 bp upstream of PFA0400c, a beta3 protesome putative gene. In total, we observe 81 P. falciparum 5' regions with a conservation p-value of 10^-6or lower.

Figure 4

Downstream blocks conserved beyond neutralexpectations among P. falciparum and the three mouse malaria species. The blocks are located (A) at 172 bps downstream of PF10_0096, a hypothetical gene, (B) at 108 bps downstream of MAL8P1.157, a hypothetical gene, and (C) at 140 bps downstream of PFL2145w, a hypothetical gene. 15 conserved blocks in downstream regions had a p-value of 10^-6 or lower.

Figure 5

Variants of the AGCTAGCT motif. Each of these variants of the AGCTAGCT motif was found to have a statistically significant z-score for conservation across the rodent malaria species (Z_p) and for overoccurrence in P. yoelii (Z_n). Interestingly, this motif is also its own reverse complement. Of the 33 motifs with significant values of Z_p and Z_n, 15 of them cluster by sequence similarity into this family.