WHY IS THE TRANSLATION SELECTIVITY OF LOW RIBOSOME AVAILABILITY
DELETERIOUS FOR ERYTHROPOIESIS?
1,2,3
Ismael Boussaid
1,2,3
,Michaela Fontenay
(1)Institute Cochin, Department Development, Reproduction, Cancer, Paris, France
(2) INSERM U1016, Paris, France����
(3) CNRS UMR8104, Paris, France
(3) Assistance Publique-Hôpitaux de Paris, Hôpitaux Universitaires Paris Centre – Cochin, Service
d’hématologie biologique, Paris, France Lead contact : Michaela Fontenay
(michaela.fontenay@inserm.fr)
(4) Université Paris Descartes, Sorbonne-Paris Cité, Paris, France
Ribosomopathies like 5q- syndrome or Diamond Blackfan Anemia (DBA) are linked to
haploinsufficiency or mutation in genes encoding ribosomal proteins (RP) leading to a
predominant erythroid phenotype. In DBA models, the decrease of ribosome content has been
associated to a defective GATA1 translation. In acquired 5q- myelodysplastic syndrome (MDS), the
monoallelic deletion of RPS14 gene results in altered ribosome biogenesis. We previously reported
that the erythroid phenotype can be partly explained by a caspase-dependent cleavage of GATA1
due to an absence of its chaperone HSP70 in the nucleus of MDS erythroblasts. Because of the
indispensable role of RPS14 protein for the maturation of the small 40S ribosome subunit, we
further investigated the impact of ribosome biogenesis defect on GATA1 and global translation.
Here we confirmed that GATA1 protein expression is decreased in 5q- primary erythroblasts and
in RPS14 shRNA-expressing normal erythroblasts. To get further insights into this defect, we
developed an inducible shRNA to RPS14 in the UT7/Epo cell line. Polysome profiling confirmed the
decrease of 40S subunit and the absolute quantification of RP by deep proteomics demonstrated
a 2-fold reduction of RPS consistent with the reduction of ribosome content by half in these cells.
We then tested the hypothesis of a decrease in GATA1 translation, by comparing the global
transcriptome and the translatome. We observed a decoupling between transcriptome and
translatome suggesting a translation selectivity.
Parameters known to regulate translation initiation or elongation like the thermodynamic
characteristics (i.e. fold energy, and length of untranslated regions UTRs) and the codon
composition of the transcripts may be involved in their selection to translation. Indeed, the
shortness of transcripts, high structuring of 3’UTR and enrichment in optimal codons were the
parameters that change translation rules and predict decreased translation of GATA1 and other
erythroid transcripts in limiting ribosome availability. We confirmed that translation changes were
driven by 3’UTR characteristics and codon composition using reporter assays and proteomics. We
also confirmed that those features governing translation selectivity in shRPS14 human cells also
governed translation selectivity in shRPS19 or shRPL5 human primary erythroid cells. Finally, at
each step of normal erythropoiesis, the poor correlation between the transcriptome and the
proteome was explained by translation regulation leading to enrichment in proteins, which
transcripts harbored the characteristics of UTRs structuring and codon composition shared by
those selectively less translated in the context of ribosomopathies.
In summary, our results show that limiting ribosome availability leads to an altered translation of
a selected subset of transcripts involved in the impairment of erythroid differentiation. These
observations reveal a regulatory role for ribosome content in the control of normal erythroid
differentiation.