Stimulation of translation by human Unr requires cold shock domains 2 and 4, and correlates with poly(A) binding protein interaction

Abstract

The RNA binding protein Unr, which contains five cold shock domains, has several specific roles in post-transcriptional control of gene expression. It can act as an activator or inhibitor of translation initiation, promote mRNA turnover, or stabilise mRNA. Its role depends on the mRNA and other proteins to which it binds, which includes cytoplasmic poly(A) binding protein 1 (PABP1). Since PABP1 binds to all polyadenylated mRNAs, and is involved in translation initiation by interaction with eukaryotic translation initiation factor 4G (eIF4G), we investigated whether Unr has a general role in translational control. We found that Unr strongly stimulates translation in vitro, and mutation of cold shock domains 2 or 4 inhibited its translation activity. The ability of Unr and its mutants to stimulate translation correlated with its ability to bind RNA, and to interact with PABP1. We found that Unr stimulated the binding of PABP1 to mRNA, and that Unr was required for the stable interaction of PABP1 and eIF4G in cells. siRNA-mediated knockdown of Unr reduced the overall level of cellular translation in cells, as well as that of cap-dependent and IRES-dependent reporters. These data describe a novel role for Unr in regulating cellular gene expression.

Figure 1. Unr stimulates translation in vitro.

(a–d) In vitro translation of uncapped non-polyadenylated (a), uncapped polyadenylated (b), capped non-polyadenylated (c) or capped polyadenylated (d) Renilla luciferase mRNA supplemented with wild type (wt) Unr or CSD mutants of Unr, relative to the buffer control (H100). The presence of a cap (O) and/or poly(A) tail (AAAA) on the mRNA (horizontal line) is indicated above each graph. Unr was in 3.7 fold molar excess over mRNA. The graphs show the mean of three independent experiments, error bars represent the standard deviation. An asterisk represents a significant difference (p < 0.05) to wt Unr. (e) Coomassie stained polyacrylamide gel showing purified recombinant wt and mutant Unr proteins, alongside BSA for quantification purposes. (f) RT-PCR of capped polyadenylated Renilla luciferase RNA isolated from in vitro translation reactions, supplemented as above. PCR products were quantified using ImageJ analysis of the agarose gel image; the values are shown beneath the gel image, relative to the buffer control. (g) UV crosslinking of wt Unr and CSD mutants of Unr to capped polyadenylated ³²P-UTP-labelled Renilla luciferase mRNA. RNA-binding of Unr and CSD mutant proteins was quantified using ImageJ analysis of the autoradiograph; the values are shown beneath the autoradiograph, relative to wt Unr. (h) Schematic of the CSD mutants of Unr. The amino acid position of F to A point mutations in each mutant is indicated. Amino acid numbering according to Unr protein isoform 2, which was the recombinant form of Unr used in this study.

Figure 2. Unr interacts with PABP1 in vitro.

(a) GST-pull down of ³⁵S-labelled wt Unr and CSD mutants with GST-PABP1, or with GST as a negative control (con). Unr pull down was quantified using ImageJ analysis of the autoradiograph, normalised against inputs; the values are shown beneath the autoradiograph, relative to wt Unr binding to GST-PABP1. (b) GST-pull down of ³⁵S-labelled wt Unr with or without GST-PABP1, in the presence or absence of RNase A. The inputs are shown to the right of each pull down and are indicated by asterisks. Unr pull down was quantified using ImageJ analysis of the autoradiograph; the values are shown beneath the autoradiograph, relative to Unr binding to GST-PABP1 in the absence of RNase A. (c) UV crosslinking of GST-PABP1, wt Unr, CSD mutant 2, alone and in combination, to uncapped, non-polyadenylated ³²P-ATP-labelled Renilla luciferase mRNA. RNA-binding of GST-PABP1 or Unr was quantified using ImageJ analysis of the autoradiograph; the values are shown beneath the autoradiograph, relative to GST-PABP1 alone. Experiments were carried out on three independent occasions and representative autoradiographs are shown.

Figure 3. Unr increases the PABP1-eIF4G interaction in cells.

(a) Immunoprecipitation (IP) from untreated HeLa cells with anti-Unr antiserum. (b) IP from untreated HeLa cells with anti-Unr antiserum or IgG control in the presence of RNase A. (c) IP from HeLa cells that had been pretreated with siRNA against Unr (siUnr) or a negative control siRNA (Scr), with anti-PABP1 antiserum. (d) IP from HeLa cells that had been pretreated with siRNA against Unr (siUnr) or a negative control siRNA (Scr), with anti-eIF4G antiserum in the presence of RNase ONE. For (a–d), input lysates and IP samples were run on the same gel, blotted onto nitrocellulose and the membrane cut to allow probing with antibodies against eIF4G (top panels), Unr (middle panels) or PABP1 (lower panels). Images separated by dotted lines represent sections of the same membrane. For (c,d), the percentage pull down of eIF4G, Unr and PABP1 from cells treated with Unr siRNA relative to control siRNA was quantified using ImageJ analysis, and is shown to the right of the western blots.

Figure 4. Unr stimulates translation in cells.

(a) Metabolic labelling (upper panel) and western blotting for Unr and GAPDH (lower panels) of HeLa and U2OS cells that had been pretreated with siRNA against Unr (siUnr) or a negative control siRNA (Scr). ³⁵S-methionine incorporation into nascent proteins was quantified using ImageJ analysis of the autoradiograph; the values are shown beneath the autoradiograph, relative to the negative control siRNA-treated samples for each cell type. Grey and black arrowheads indicate bands whose intensity increases or decreases, respectively, between negative control siRNA-treated samples and Unr siRNA-treated samples. This experiment was carried out on three independent occasions and a representative autoradiograph is shown. (b) Schematic of capped and polyadenylated dicistronic RHRVF mRNA. The upstream Renilla luciferase (Rluc) cistron is translated in a cap-dependent manner. Translation of the downstream firefly luciferase (Fluc) depends on the activity of the intercistronic HRV-2 IRES. (c) Relative translation of Rluc and Fluc from RHRVF mRNA 6 hours post-transfection of cells that had been pretreated with no siRNA, siRNA against Unr (siUnr) or a negative control siRNA (scr). The graph shows the means of three independent experiments and error bars represent the standard deviation. An asterisk represents a significant difference (p < 0.05) to the no siRNA sample. (d) Relative level of ³²P-labelled dicistronic RHRVF mRNA 6 hours post-transfection of cells that had been pretreated with no siRNA, siRNA against Unr (siUnr) or a negative control siRNA (scr), quantified by scintillation counter. The graph shows the means of two independent experiments carried out in duplicate and error bars represent the standard deviation.