Comprehensive detection of human terminal oligo-pyrimidine (TOP) genes and analysis of their characteristics

Abstract

Although the knowledge accumulated on the transcriptional regulations of eukaryotes is significant, the knowledge on their translational regulations remains limited. Thus, we performed a comprehensive detection of terminal oligo-pyrimidine (TOP), which is one of the well-characterized cis-regulatory motifs for translational controls located immediately downstream of the transcriptional start sites of mRNAs. Utilizing our precise 5'-end information of the full-length cDNAs, we could screen 1645 candidate TOP genes by position specific matrix search. Among them, not only 75 out of 78 ribosomal protein genes but also eight previously identified non-ribosomal-protein TOP genes were included. We further experimentally validated the translational activities of 83 TOP candidate genes. Clear translational regulations exerted on the stimulation of 12-O-tetradecanoyl-1-phorbol-13-acetate for at least 41 of them was observed, indicating that there should be a few hundreds of human genes which are subjected to regulation at translation levels via TOPs. Our result suggests that TOP genes code not only formerly characterized ribosomal proteins and translation-related proteins but also a wider variety of proteins, such as lysosome-related proteins and metabolism-related proteins, playing pivotal roles in gene expression controls in the majority of cellular mRNAs.

Figure 1.

Sequence logo of the most fixed TSSs and known TOP genes. (A) We constructed the Sequence logo of the most fixed 931 TSSs corresponding to 931 genes. 873 other genes (B) and 48 TOP genes group (C) and (D) PSWM of TOP genes. The black box at +1 position in the table shows all of the TOP genes showing same nucleotide namely ‘C’. The gray boxes in the table from −1:+4 shows the pyrimidine region. These two conditions were considered to detect TOP gene candidates.

Figure 2.

Tissue specificity of TOP genes candidates with a box-and-whisker plot. The horizontal axis shows 1645 TOP genes candidates (TOP) and 6174 not TOP candidates (not-TOP). The vertical axis shows ‘expression breadth’. This figure is drawn with ‘boxplot’ in R package.

Figure 3.

Expression profiles for TOP genes candidates. (A) HPLC fraction of cell elutions. The fractions were divided into Sub: sub-polysomal groups (1–5) and Poly: polysomal groups (7–11). Left: fraction of cells not treated with TPA, Right: fraction of cells treated with TPA. (B) Several examples of expression of detected TOP genes candidates. Relative expression levels, the highest one corresponding to 1, are shown. RPL19: ribosomal protein L19 for positive control, GAPDH for negative control, SDBCAG84: serologically defined breast cancer antigen 84, EEF1G: eukaryotic translation EF1-G, TMEM30A: transmembrane protein 30A. We showed the fraction corresponding to the potential ribosome number on ORF assuming one ribosome/150 bp. The number in each figure indicates the ratio of (TPA+:Sub/TPA+:Pol)/(TPA−:Sub/TPA−:Pol). For further details on fraction distributions, see Supplementary Research Data Figure 3. (C) Expression ratios of 81 candidates. (D) Expression ratios of 10 negative controls. Black bars correspond to the expression ratio with TPA treatment, and gray bars correspond to the expression ratio without TPA treatment. The y-axis shows the ratio of mRNA expression in polysome and sub-polysome fractions. These ratios were converted as log ratio.

Figure 4.

Correlation between translational regulation and mRNA features. Figures (A–D) show the correlation between translational regulation and the 3′-UTR, 5′-UTR, RNA and ORF length, respectively. The horizontal axis shows each mRNA length, and the vertical axis shows the ratio of mRNA level in sub-polysome and polysome fractions. The correlation coefficients were 3′-UTR: −0.56 (P < 1.1e−8), 5′-UTR: −0.42 (P < 2.9e−5), RNA: −0.61 (P < 1.1e−12), ORF: −0.53 (P < 5.1e−8).