ZNF9 activation of IRES-mediated translation of the human ODC mRNA is decreased in myotonic dystrophy type 2

Abstract

Myotonic dystrophy types 1 and 2 (DM1 and DM2) are forms of muscular dystrophy that share similar clinical and molecular manifestations, such as myotonia, muscle weakness, cardiac anomalies, cataracts, and the presence of defined RNA-containing foci in muscle nuclei. DM2 is caused by an expansion of the tetranucleotide CCTG repeat within the first intron of ZNF9, although the mechanism by which the expanded nucleotide repeat causes the debilitating symptoms of DM2 is unclear. Conflicting studies have led to two models for the mechanisms leading to the problems associated with DM2. First, a gain-of-function disease model hypothesizes that the repeat expansions in the transcribed RNA do not directly affect ZNF9 function. Instead repeat-containing RNAs are thought to sequester proteins in the nucleus, causing misregulation of normal cellular processes. In the alternative model, the repeat expansions impair ZNF9 function and lead to a decrease in the level of translation. Here we examine the normal in vivo function of ZNF9. We report that ZNF9 associates with actively translating ribosomes and functions as an activator of cap-independent translation of the human ODC mRNA. This activity is mediated by direct binding of ZNF9 to the internal ribosome entry site sequence (IRES) within the 5'UTR of ODC mRNA. ZNF9 can activate IRES-mediated translation of ODC within primary human myoblasts, and this activity is reduced in myoblasts derived from a DM2 patient. These data identify ZNF9 as a regulator of cap-independent translation and indicate that ZNF9 activity may contribute mechanistically to the myotonic dystrophy type 2 phenotype.

Figure 1. ZNF9 specifically binds to a human cellular mRNA IRES element.

A. Increasing amounts of recombinant, full-length ZNF9 were incubated with ³²P-end-labeled ODC RNA probe and cross-linked. Reactions were separated by polyacrylamide gel electrophoresis, transferred to nitrocellulose and exposed to X-ray film. Membranes were stained for ZNF9 protein using Ponceau S. B. Recombinant ZNF9 was incubated with ³²P-end-labeled ODC RNA probe and either a 50-fold molar excess of bovine liver tRNA or a 50-fold molar excess of unlabeled ODC RNA competitor (comp) probe. Samples were processed as in A.

Figure 2. ZNF9 interacts with the translating human eukaryotic ribosome.

A. Western blots of sucrose gradient fractions of HeLa cells extracts transfected with either ZNF9-myc or LacZ-myc and probed with the indicated antibodies. 10% of each fraction was analyzed by western blotting. WCL: whole cell extract, C: cytoplasmic, N: nuclear B. ZNF9 and GAPDH western blotting of supernatant and polysome fractions of HeLa cell lysates isolated in the absence or presence of EDTA. Quantification was normalized to the levels of ZNF9 in the supernatant fraction in the absence of EDTA C. ZNF9 western blotting of sucrose gradient fractions of HeLa cell lysates in the absence or presence of EDTA. Equal cell lysate amounts (10 O.D. units) were applied to each gradient (−/+50 mM EDTA), and each fraction was TCA precipitated. 50% of the total precipitated protein was analyzed by western blotting. D. ZNF9 western blotting analysis of supernatant and pellet fractions from ribosome salt wash experiments. Western blots were quantified and the data are reported as ratios of ZNF9 in the pellet fraction to the amount of ZNF9 in the supernatant, normalized to the values for 100 mM potassium acetate.

Figure 3. ZNF9 activates cap independent-translation of the human ODC mRNA.

A. Diagram of the ODC bicistronic vector for measuring cap-independent translation of human ODC. 5′meG: 5′-methyl-G cap structure. B. HeLa cells expressing the ODC bicistronic vector and either V5-ZNF9 or V5-LacZ were analyzed for luciferase activity. The level of cap-independent translation is reported as a ratio of cap-independent translation (firefly) to cap-dependent translation (Renilla). Values are averages of 6 independent experiments. The error bars are the standard error of the mean (S.E.M). The insert figure is a representative western blot analysis showing the V5-LacZ and V5-ZNF9 levels in the HeLa cells lysates. C. ZNF9 mRNA and protein levels after RNAi of ZNF9 in HeLa cells. HeLa cells expressing either a control shRNA or a ZNF9-specific shRNA were analyzed using quantitative RT-PCR and western blotting to quantify expression of ZNF9 RNA and protein, respectively. The RT-PCR values were computed using the ΔΔCt method and are normalized to the levels of GAPDH mRNA. Error bars represent the standard error. The insert figure is a western blot showing the ZNF9 proteins levels. Recombinant ZNF9 (rZNF9) was used as a positive control, and equal amounts of protein were loaded in each lane as assessed by BCA protein assays. D. HeLa cells expressing the ODC bicistronic vector and either a control or ZNF9-specific shRNA were assessed for luciferase activity. The level of cap-independent translation is reported as a ratio of cap-independent translation (firefly) to cap-dependent translation (Renilla). Values are the averages of 3 independent experiments and are normalized to the control shRNA. The error bars are the S.E.M.

Figure 4. Cap-independent translation of ODC is defective in DM2 myoblasts.

A. IRES activity in wild-type and DM2 myoblast. Equal numbers of human wild-type or DM2 myoblasts were infected with virions expressing the bicistronic ODC reporter vector and were analyzed for luciferase activity. The level of cap-independent translation is reported as a ratio of cap-independent translation (firefly) to cap-dependent translation (Renilla). Values are averages of 6 independent experiments and are normalized to IRES activity in wild-type myoblasts. B. Differences in ZNF9 mRNA expression in wild-type and DM2 myoblasts. Quantitative real-time PCR was performed on equal amounts of total mRNA from wild-type and DM2 myoblasts. Values were computed using the ΔΔCt method and are normalized to levels of GAPDH mRNA. Error bars represent the S.E.M. C. IRES activity in wild-type and DM2 myoblast overexpressing ZNF9. Equal numbers of wild-type or DM2 myoblasts were infected with the bicistronic ODC reporter vector and vectors for either the overexpression of LacZ or ZNF9. Cell lysates were analyzed for luciferase activity as described above. Values are reported as a ratio of cap-independent translation to cap-dependent translation and are representative of 4 independent experiments. The values are normalized to the value of IRES activation in wild-type myoblasts overexpressing LacZ. Error bars represent the S.E.M.