Selective mRNA translation controls erythropoiesis; regulation and function of the RNA-binding protein Csde1.
Dr M. von Lindern
Duration:
Name researcher:
4 years
Amount granted:
€220.000
Year:
2011
Project number:
1140
Project leader:
Dr Marieke von Lindern (Dept. of Hematopoiesis, Sanquin, Amsterdam)
PhD student: Kat Moore (July 2012 – July 2016)
Research technician: Nurcan Yagci (Sept. 2014 – Sept. 2016)
About the project
The Csde1 protein contains seven RNA-binding domains, called cold-shock domains. Hence its name Cold shock domain E1. The Csde1 protein is widely expressed, but erythroblasts express 100-fold more Csde1 compared to precursors of other hematopoietic lineages. This suggests that Csde1 fulfils a specific role in erythropoiesis. We also observed very low Csde1 expression in erythroblasts of patients with Diamond Blackfan Anemia (DBA), compared to erythroblasts of healthy controls. This low expression of Csde1 was a direct consequence of the reduced number of ribosomes that characterizes DBA, and lowering Csde1 expression impaired the proliferation and maturation of erythroblasts. We hypothesised that Csde1 has an important role in erythropoiesis.
The aim of the project was to understand the function of Csde1 in erythropoiesis, and to investigate whether and how reduced Csde1 expression contributes to the severe anemia of DBA.
We first identified the transcripts that associate with Csde1 in the ´mouse eryhroblast leukemia´ cell line MEL. These transcripts mainly encoded proteins involved in mRNA translation and protein stability, and in mitochondrial function. To evaluate the effect of Csde1 we attempted to reduce Csde1 expression. Lentiviral expression of shRNA resulted in induction of Csde1 expression as a consequence of virus transduction. The shRNA was effective, but reduced Csde1 level to parental levels. Next, we used Crisp/Cas9-mediated gene editing to abrogate Csde1 expression or to delete the first cold shock domain. We obtained clones with low expression of N-terminally truncated Csde1, but never clones without Csde1. Thus, we confirmed our previous findings that Csde1 is essential for erythroblast survival. Analysis of Csde1 deletion mutants showed changes in mRNA stability and translation. Due to feedback mechanisms these changes were modest. For instance, cells homozygous for truncated Csde1 contained less Pabpc1 mRNA but increased levels of Pabpc1 protein, suggesting decreased mRNA stability and increased translation.
Analysis of Csde1-containing protein complexes revealed Strap (serine- and threonine kinase receptor associated protein) as a strong partner of Csde1. Pabpc1 and Pabpc4 were also identified as Csde1-associated proteins. Strap expression did not alter the binding of Csde1 to target transcripts, but it affected the expression level and/or translation of Csde1-bound transcripts. Of particular interest is the binding to and regulation of Pabpc1 and Pabpc4, together with the regulation of eIF4G3. Binding of Pabp to the poly(A) tail and to the eIF4G scaffold in the cap/binding complex is crucial for efficient translation. The interaction between Pabp en eIF4G creates a link between the start and the end of a mRNA and enables an efficient restart of the protein synthesis machinery after it finishes synthesis at the end of the mRNA. Pabpc4 and eIFG3 are isoforms that may be important during hypoxia, which prompts for future experiments to investigate proliferation, differentiation and survival of erythroblasts under hypoxic conditions.
Concurrently we characterized where, and how efficient, ribosomes bind to specific transcripts. This method is indicated as ribosome footprinting and included an experimental identification of translation start codons. Translation of Csde1 appeared sensitive to tunicamycin-induced phosphorylation of translation initiation factor 2 (eIF2). Phosphorylation of eIF2 inhibits protein synthesis to protect cells from the toxicity of protein aggregates. Translation of some transcripts, such as Csde1, is reduced at a more than average rate. The ribosome footprints indicated which sequences in the 5´leader of the Csde1 transcript are translated. It also indicated that Csde1 is generated as two protein isoforms, one of which is a 13-amino acid N-terminal extension of the annotated protein. This protein is generated from a CUG start codon, which explains why this protein was not previously predicted. Whether the two Csde1 isoforms have distinct functions needs to be investigated.
Conclusion: Csde1 is an essential RNA-binding protein in erythroblasts. Csde1 expression is reduced when the protein synthesis capacity is reduced, as is the case in DBA, and when the cell is at risk for formation of toxic protein aggregates. Csde1 acts in concert with Strap to control protein homeostasis.
Thesis 2018: The spiders at the center of the web. Csde1 and strap control translation in erythropoiesis, Kat Moore