A genome-wide tethering screen reveals novel potential post-transcriptional regulators in Trypanosoma brucei

Abstract

In trypanosomatids, gene expression is regulated mainly by post-transcriptional mechanisms, which affect mRNA processing, translation and degradation. Currently, our understanding of factors that regulate either mRNA stability or translation is rather limited. We know that often, the regulators are proteins that bind to the 3'-untranslated region; they presumably interact with ribonucleases and translation factors. However, very few such proteins have been characterized in any detail. Here we describe a genome-wide screen to find proteins implicated in post-transcriptional regulation in Trypanosoma brucei. We made a library of random genomic fragments in a plasmid that was designed for expression of proteins fused to an RNA-binding domain, the lambda-N peptide. This was transfected into cells expressing mRNAs encoding a positive or negative selectable marker, and bearing the "boxB" lambda-N recognition element in the 3'-untranslated region. The screen identified about 300 proteins that could be implicated in post-transcriptional mRNA regulation. These included known regulators, degradative enzymes and translation factors, many canonical RNA-binding proteins, and proteins that act via multi-protein complexes. However there were also nearly 150 potential regulators with no previously annotated function, or functions unrelated to mRNA metabolism. Almost 50 novel regulators were shown to bind RNA using a targeted proteome array. The screen also provided fine structure mapping of the hit candidates' functional domains. Our findings not only confirm the key role that RNA-binding proteins play in the regulation of gene expression in trypanosomatids, but also suggest new roles for previously uncharacterized proteins.

Figure 1. Proof of concept of the tethering approach.

(A) Schematic diagram of reporter mRNAs and tethered proteins, not to scale. Above: BLA and GFP reporter mRNA with five boxB recognition sites in the 3′-UTR. Middle: Control BLA 3′-UTR mRNA without boxB. Below: PGKB mRNA with boxB recognition sites at the 3′-UTR. A tetracycline-inducible EP1 procyclin promoter (flag) drives PGKB expression. The downstream promoter, P_VSG, gives constitutive expression of the resistance marker. Bottom: Fusion protein with N-terminal lambda-N peptide and C-terminal myc tag. (B) Cumulative growth curves of cells expressing the BLA reporters with tetracycline-inducible lambda-N-PABP1, in the presence of 5x blasticidin (25 µg/ml) with or without tetracycline. For comparison, cells expressing lambda-N-PABP1 and non-transfected cells (without N-protein, parental cell line) grown in 50x blasticidin (250 µg/ml) are shown. The +boxB cells with lambda-N-PABP express a little of the lambda-N fusion even in the absence of tetracycline (panel D), which may result in slightly better growth in the presence of 25 µg/ml blasticidin. (C) Cumulative growth of cells containing the BLA-boxB reporter expressing N-CAF1-myc in the presence or absence of tetracycline and 5x blasticidin. (D) Western blot analysis of N-PABP1-myc, N-CAF1-myc and N-GFP proteins. Expression of PABP1-myc and CAF1-myc was detected with anti-myc, using aldolase as loading control. (E) Effect of tethered PABP1, GFP and CAF1 on growth after tetracycline-mediated induction of both PGKB and tethered protein expression.

Figure 2. Schematic of the overexpression library and the growth conditions analyzed.

After transfection, the library was grown under non-inducing (tet −) and inducing conditions (tet +) with increasing concentration of blasticidin and then, genomic DNA was isolated from survival populations. Adaptor-ligated sequencing libraries were prepared from specific PCR reactions and Illumina sequenced. Illumina sequencing reads containing a backbone vector junction sequence were then mapped to the T. brucei reference.

Figure 3. Results for ZC3H11.

(A) Sequence reads that were linked to the lambda-N peptide were extracted, mapped onto the trypanosome genome and visualised using the Integrative Genomics Viewer (IGV) tool. The Figure shows the sequence coverage plots for a ∼12 kbp region of chromosome 5. Open reading frames are in purple and mRNA untranslated regions in dark grey. Reads were derived from tetracycline-induced cells grown without blasticidin (upper panel), or with the 10x blasticidin concentration (bottom panel). For better visualization, the plotted Y scales display different range values. In the 10x blasticidin plot, note the cluster over ZC3H11; the isolated “pile” downstream of Tb927.5.820 would be removed later in the filtering process. (B) These are the same results as in (A), but in close-up focusing on the ZC3H11 (Tb927.5.810) open reading frame. Each “stack” represents a unique target fragment, mapped to forward (plus) and reverse (minus) strands, for induced cells grown without (upper panel) and with (bottom panel) blasticidin pressure. Light blue boxes indicate the ZC3H11 coding sequence (transcribed left to right) while zinc finger domain and HNPY motif (residues 186–189) are indicated in green and orange respectively. The minimum region that was required for activity in tethering assays was the C-terminal part, starting at amino acid 186.

Figure 4. Enriched protein categories and Pfam domains.

(A) Proteins for which fragments increased BLA resistance are classified by category. Different classes of gene function are shown in different colors. For highly enriched groups (Fisher's exact test, P<0.05) the E-values are shown in parentheses. On the right, all Pfam domains that were significantly enriched are shown (P<0.01, Fisher's exact test); lower p-values are represented in red, higher p-values are represented in blue. The numbers of each domain identified are shown in the corresponding boxed. (B) Proteins for which fragments suppressed PGKB expression. Details are as in (A).

Figure 5. Validation of mRNA-fate regulators.

Expression of CAT in cells expressing different myc-tagged lambda-N-fusion proteins was assayed after 24 h tetracycline induction and activities were expressed relative to a control with no lambda-N protein (green). Results are arithmetic mean ± standard deviation of at least 3 independent clones. As comparison, the fold per CDS (log2 values) are shown (grey). (left) Domain structures of analysed T. brucei proteins as detected by SMART (http://smart.embl-heidelberg.de). Different domains are specific colours as shown on the blocks. P5CR, pyrroline-5-carboxylate reductase; CD, cytidine deaminase; LCR, low-complexity region; RRM, RNA recognition motif; RBD, RNA-binding domain. RBP11 and Tb927.10.15760 are negative controls, which did not have reproducible effects in the tethering screen.