Prediction of the general transcription factors associated with RNA polymerase II in Plasmodium falciparum: conserved features and differences relative to other eukaryotes

Abstract

Background: To date, only a few transcription factors have been identified in the genome of the parasite Plasmodium falciparum, the causative agent of malaria. Moreover, no detailed molecular analysis of its basal transcription machinery, which is otherwise well-conserved in the crown group of eukaryotes, has yet been reported. In this study, we have used a combination of sensitive sequence analysis methods to predict the existence of several parasite encoded general transcription factors associated with RNA polymerase II.

Results: Several orthologs of general transcription factors associated with RNA polymerase II can be predicted among the hypothetical proteins of the P. falciparum genome using the two-dimensional Hydrophobic Cluster Analysis (HCA) together with profile-based search methods (PSI-BLAST). These predicted orthologous genes encoding putative transcription factors include the large subunit of TFIIA and two candidates for its small subunit, the TFIIE beta-subunit, which would associate with the previously known TFIIE alpha-subunit, the TFIIF beta-subunit, as well as the p62/TFB1 subunit of the TFIIH core. Within TFIID, the putative orthologs of TAF1, TAF2, TAF7 and TAF10 were also predicted. However, no candidates for TAFs with classical histone fold domain (HFD) were found, suggesting an unusual architecture of TFIID complex of RNA polymerase II in the parasite.

Conclusion: Taken together, these results suggest that more general transcription factors may be present in the P. falciparum proteome than initially thought. The prediction of these orthologous general transcription factors opens the way for further studies dealing with transcriptional regulation in P. falciparum. These alternative and sensitive sequence analysis methods can help to identify candidates for other transcriptional regulatory factors in P. falciparum. They will also facilitate the prediction of biological functions for several orphan proteins from other apicomplexan parasites such as Toxoplasma gondii, Cryptosporidium parvum and Eimeria.

Figure 1

Schematic view of the predicted general transcription factors associated with RNA polymerase II in Plasmodium falciparum. Components which have been predicted in previous studies and in the present analysis, are displayed respectively in blue and in red. The components which have not been predicted from sequence analyses are shown in grey and white. Grey boxes indicate components for which potential candidates exist, but which cannot be discriminated from sequence analysis alone, due to the absence of specific domains. Green boxes indicate the HFD-containing TAF pairs which have not been identified in Plasmodium falciparum.

Figure 2

Comparison of the HCA plots of TFIIA large subunit (TOA1/TFIIA α/β panel A) and small subunit (TOA2/TFIIA γ panel B) subunits from different species, highlighting the conservation of the hydrophobic core of the domains constituting the two proteins in the hypothetical proteins from Plasmodium falciparum. The sequences are shown on a duplicated α-helical net, in which strong hydrophobic amino acids (VILFMYW) are circled. These form clusters, which mainly correspond to the internal faces of regular secondary structures (α-helices and β-strands). The way to read the sequence and special symbols is indicated in the inset. The regions initially detected by PSI-BLAST (either with significant (TOA1/TFIIA α) or marginal E-value (TOA2/TFIIA γ) are indicated with a dotted line. Cluster similarities are shaded in grey, identities are shown in white on a black background. Despite low level of sequence identity, hydrophobic clusters are well conserved, supporting the presence of a common fold. The deduced 1D alignment is shown at the bottom and at right. The positions of regular secondary structures, as observed from experimental data (pdb 1nh2) are shown up to the HCA plot. Two hypothetical proteins from Plasmodium falciparum share significant similarities with the small subunit of TFIIA (bottom panel) and are thus _PfTFIIA small subunit candidates. The similarity with PFI1630c was revealed using its homolog sequence in Plasmodium yoelii (see text). This sequence was missed during our first PSI-BLAST search because of an error in the intron prediction. This novel ortholog has been found through HCA sequence comparison of the translated DNA sequences (the boxed and underlined sequences in the HCA plot and 1D alignment, respectively, correspond to the sequence which was first included in an intron, as predicted automatically from genome data).

Figure 3

1D alignment of different TAF subunits (TAF1, TAF2 and TAF7) with hypothetical proteins from P. falciparum. Identical and similar amino acids are boxed in black and in grey, respectively. Although restricted to a limited length, the similarity regions highlighted here match the inter-species conserved regions (see text). These similarities are supported at the 2D level using HCA (data not shown).

Figure 4

Comparison of the HCA plots of TAF10 from human (TAF30) and yeast (TAF25) and the P. falciparum hypothetical protein PFE1110w (see figure Fig. 2 for explanation) Cluster similarities are shaded in grey, identities are shown in white on a black background. Vertical bars indicate cluster links. The deduced 1D alignment is shown at the bottom.

Figure 5

Comparison of the HCA plots of the TAF14 subunit in humans (AF9) and in yeast (TFG3) with that of the Plasmodium falciparum MAL8P1.131 hypothetical protein (see Fig. 2 for explanation). Cluster similarities are shaded in grey and identities are shown in white on a black background. A N-terminal YEATS domain [57] is present in the three sequences, whereas HCA detects a common domain in the C-terminal end of only human AF9 and yeast TFG3. The conserved clusters of this C-terminal domain are not detected in the MAL8P1.131 sequence, suggesting that this protein may not correspond to _PfTAF14.

Figure 6

A) Identification of a small globular domain common to the C-termini of TFIIE α subunits of higher eukaryotes and P. falciparum. Top panel: Comparison of the corresponding HCA plots (see figure Fig. 2 for explanation). Cluster similarities are shaded in grey and identities are shown in white on a black background. The position of the globular domain is boxed. Despite the low level of sequence identity, hydrophobic clusters are well conserved, supporting the presence of a common fold. Bottom panel: HCA-deduced 1D alignment. B) Comparison of the TFIIE β subunits of S. cerevisiae, Cryptococcus neoformans and Plasmodia species (P. yoelii and P. falciparum). Cluster similarities are shaded in grey and identities are shown in white on a black background. The deduced 1D alignment is shown at the bottom.

Figure 7

Comparison of the human TFIIF β subunit (RAP30) with the hypothetical protein PF11_0458 from P. falciparum. Top panel: Comparison of the corresponding HCA plots (see Fig. 2 for explanation). The N- and C-terminal structured domains are boxed, according to the limits defined on the basis of experimental data (pdb 1f3u (chain A) and 2bby, respectively). Cluster similarities are shaded in grey and identities are shown in white on a black background. Putative correspondences in the C-terminus of the first domain are reported with dashed lines. Secondary structures, as observed in the experimental structures, are reported up to the RAP30 sequence. The corresponding 1D alignment is shown in the bottom panel.

Figure 8

Comparison of the human TFIIH P62 subunit with the hypothetical MAL3P7.42 from P. falciparum. Significant similarities were detected for the region including two BSD domains (boxed, panel B). These were supported at the 2D level by comparison of the HCA plots (see Fig. 2 for explanation), where cluster similarities are shaded in grey and identities are shown in white on a black background. Marginal similarities observed in PSI-BLAST are located within the C-terminal parts of the proteins (ranging from ~250 to the C-terminus (human) and from ~aa 370 to the C-terminus (P. falciparum)). These were also supported at the 2D level, as illustrated here on a segment in the most distal C-terminal part of the two proteins (panel C). Upstream of the BSD domain, cluster similarities together with sequence identities can be observed in the most proximal N-terminal part of the protein sequences, corresponding to a pH domain in human p62 (panel A). Secondary structures, as observed in the experimental structure of human p62 (pdb 1pfj; [80]), are reported up to its sequence. The corresponding 1D alignment is shown in panel D.

Figure 9

Comparison of the amino acid distribution in the proteomes (predicted proteins) of P. falciparum, Homo sapiens, Saccharomyces cerevisiae and Arabidopsis thaliana (5334, 32035, 6699 and 27857 sequences, respectively). Amino acids are grouped with respect to the structural classes defined previously in [30]. The first class defines strong hydrophobic amino acids (V, I, L, F, M, Y, W), that display a similar propensity for yielding regular secondary structures (α-helices and β-strands. The third class includes coil-forming amino acids (G, P, D, N, S), whereas the intermediate class (A, R, C, Q, T, E, K) contains amino acids for which coil and secondary structure forming propensities are similar. The total class I amino acid content is similar in Plasmodium falciparum with respect to other proteomes (see comments in the discussion section). The frequency of cysteine, which is also a frequent component of hydrophobic cores, does not differ in Plasmodium falciparum relatively to other organisms. One can also observe a stable frequency for histidine, which has always a remarkably neutral behavior in the secondary structure propensities.