Ribosome Profiling Uncovers the Role of uORFs in Translational Control of Gene Expression during Erythroblast Differentiation

The erythroid progenitor compartment possesses a large expansion capacity, both in vivo and in vitro, which enables a rapid restoration of peripheral erythrocytes following severe blood loss. This expansion is tightly regulated to maintain erythrocyte numbers between narrow boundaries, and to balance expansion of the erythroid compartment against the availability of iron for heme and haemoglobin production. We previously observed that control of mRNA translation is crucial for expansion of the erythroid compartment. We also showed that translation of specific transcripts is impaired in Diamond Blackfan Anemia (DBA), a severe congenital anemia due to defective ribosome biosynthesis. Transcripts can be subject to translational control through domains in the 5’- or 3’UTR, including secondary structures, protein binding sequences and upstream open reading frames (uORFs). The presence of uORFs, including those starting at non-AUG codons in the 5’UTR, may alter the level of mRNA translation, but may also result in the expression of alternative protein isoforms because translation initiation may be redirected to more downstream start codons.

The aim of our current studies is to provide a genome wide map of mRNA translation efficiency during erythropoiesis that can be used to investigate defective mRNA translation in, for instance, DBA. Ribosome profiling is a genome wide high-throughput sequencing technology for global mapping of translation initiation sites that allows translation analysis with codon resolution at the genome wide level. We first investigated translational changes occurring during differentiation of mouse erythroblasts. We used p53-deficient, growth factor dependent and differentiation competent immortalized erythroblast cultures that were expanded in presence of erythropoietin (Epo), stem cell factor (SCF) and glucocorticoids as T0, and subsequently differentiated the cells in presence of Epo for 17 and 46 hours (T17, and T46 samples). To obtain ribosome footprints, the cells were treated for 7 minutes with harringtonin or solvent, and subsequently for 5 minutes with cycloheximide, which arrests translation by stabilizing the ribosomes at translation initiation codons, or on all codons, respectively. We used optimized protocols for ribosome footprinting and data analysis, and focused the analysis on transcripts containing uORFs.

First we performed a qualitative analysis of start codon usage. The ribosome footprint data proved to be superior to previously used polyribosome recruitment. In some cases polysome recruitment appeared to represent translation of an uORFs while the protein coding ORF is hardly translated (e.g. Csf2rb2, Puma). In another set of transcripts, we found uORFs that are differentially translated during differentiation, and thereby regulate differential translation from a downstream start codon (e.g. Klf3, Use1, CD47, Kell). Finally, comparison of ribosome footprints determined in erythroblasts and in myoblasts/myotubes revealed tissue specific translation regulation of otherwise ubiquitously expressed transcripts among which transcripts encoding ribosomal proteins. Second, we will perform quantitative analysis of mRNA translation in erythropoiesis through the comparison of ribosome footprint reads in an ORF with total mRNA reads obtained from total mRNA sequencing of the same sample.

The obtained insight in transcript specific translation at codon resolution is of great value to understand many cellular processes during erythropoiesis, and will be of particular interest to understand responses to iron availability and reactive oxygen species that particularly affect translation of transcripts harboring uORFs.