LSBR 1239

Control of erythropoiesis by iron-dependent expression of protein isoforms. (LSBR 1239)
Project leader: Dr. Marieke von Lindern (Dept. of Hematopoiesis, Sanquin, Amsterdam)
Ph.D. student: Nahuel Paolini (Nov. 2013 – Nov. 2017)
Research technician: Franca di Summa (May. 2015 – Sept. 2016)
Research technician: Nurcan Yagci (Sept. 2016 – May. 2017)


The number of erythrocytes in peripheral blood has to be controlled between narrow limits. Too few erythrocytes results in anemia and carries the risk of tissue damage by ischemia, whereas too many erythrocytes enhances the risk for stroke. Therefore, the process to generate erythrocytes, erythropoiesis, must be tightly controlled. Regulation of mRNA translation is an important level of control that enables rapid adjustment of gene expression in response to environmental factors. Iron is an important factor in erythropoiesis. Each erythrocyte contains approximately 250 million hemoglobin molecules that each contain 4 iron-loaded heme molecules. This iron is recycled during erythrocyte degradation and erythrocyte maturation. Lack of iron causes aggregation of globin polypeptides. Accumulation of unfolded proteins is toxic to cells, which is prevented by reducing the protein synthesis rate of erythroblasts when iron availability is low, or when unfolded proteins accumulate. These conditions activate translation initiation factor 2 (eIF2)-associated kinases 1 and 3, respectively (eIF2ak1 or heme-regulated kinase [HRI], and eIF2ak3 or protein kinase R [PKR]-like endoplasmic reticulum kinase [PERK]). eIF2 brings the amino acid methionine to the start codon of mRNA translation and provides the energy to associate the large and small subunit of the ribosome. Phosphorylation renders eIF2 inactive, which reduces translation start-site recognition. The hypothesis is that protein synthesis start sites that are preceded by a (short) translated sequence in the 5´leader of the mRNA are particularly sensitive to lower levels of active eIF2.

The aim of this project was to understand how mRNA translation affects erythropoiesis, and to identify transcripts whose translation is particularly sensitive to phosphorylation of eIF2.

Results: We investigated mRNA translation with a novel technique that identifies the position of ribosomes on transcripts. This enabled us to establish which codons are used as start codons, and to compare the density of ribosomes before and after eIF2 phosphorylation. Phosphorylation of eIF2 was induced by tunicamycin. This revealed widespread translation of sequences in the 5’leader of transcripts. In contrast to what we expected, such upstream translation did not predict the sensitivity of transcript translation to eIF2 phosphorylation. We identified 147 transcripts of which translation increased upon tunicamycin treatment, and 337 transcripts of which translation was more than average reduced. Among the downregulated transcripts was Csde1, that encodes an RNA-binding protein and whose translation is also reduced in Diamond Blackfan Anaemia (DBA). Several transcripts with increased translation encoded transcription factors, which suggests that eIF2 phosphorylation affects cellular responses on short term through mRNA translation, and subsequently by gene transcription. Transcripts with increased, or more than average decreased, translation showed different patterns of translated sequences in their 5´leader sequence: long translated sequences for enhanced translation, short translated sequences in mRNAs with exceedingly reduced translation.
One of the transcripts with enhanced translation encoded the transcription factor Tis7 (TPA-induced sequence 7). Gene expression analysis revealed that only a small group of genes was responsive to both tunicamycin treatment and Tis7 expression. The majority of stress-induced transcripts was induced independent of Tis7, and the majority of Tis7-dependent transcripts was not controlled by tunicamycin treatment. Among the transcripts that required Tis7 to be upregulated by tunicamycin was a long non-coding transcript Snhg12 that is positively associated with cell growth and cell invasiveness of cancer cells. Reduced expression of Tis7 and its targets, among which Snhg12, impaired the viability of erythroblasts.
Previously, we identified transcripts subject to control of translation by growth factors that contained start codons in a conserved context in their 5´leader sequence. Surprisingly, some strong potential start codons were not detected in our experimental data. Our preliminary data obtained with oligonucleotides that blocked scanning of the 5’leader suggested that parts of the 5’ leader may be skipped.
We also investigated how reduced expression of ribosomes, a hallmark of DBA, affects mRNA translation. We established that mutations in the ribosomal protein RPS23, found in two patients with congenital developmental impairments, affects the accuracy of mRNA translation.

In conclusion, we mapped upstream translated sequences and their role in response to eIF2 phosphorylation, and identified mechanisms that may control the cellular response to iron deficiency.