Species-Specific Functions For Long Noncoding RNAs During Erythropoeisis

Abstract Mammalian genomes encode thousands of long noncoding (lnc) RNAs, although the extent to which lncRNAs regulate biological functions is not known. Indeed, most lncRNAs show low levels of conservation among mammals, which could suggest limited or species-restricted functions. We used RNA-seq to identify 683 and 1189 lncRNAs expressed during mouse and human erythropoiesis respectively, of which ∼25% are erythroid-specific. Primary cell chromatin immunoprecipitation (ChIP) studies indicate that most mouse erythroid lncRNAs arise from classical promoters, and their genomic loci are bound by the hematopoietic transcription factors GATA1 and SCL/TAL1. Using the G1E-ER4 cell line, which has an inducible form of GATA1, we found that lncRNAs whose genomic loci are bound by GATA1 in primary cells are upregulated on GATA1 induction in the cell line, similar to coding genes. We compared RNA-Seq transcriptomes of erythroblasts from 8 different mice strains and showed that lncRNA expression is highly conserved within strains of the same species. A comparison of the human and mouse lncRNA transcriptomes showed that only 15% of mouse erythroblast lncRNAs have detectable syntenic transcripts in human erythroblasts, and that most lncRNAs are not detectable in humans. To identify functional lncRNAs, we used RNA interference in primary erythroblasts to analyze the functions of 21 abundantly expressed murine erythroid lncRNAs, 15 of which are mouse specific and 6 of which have orthologous transcripts in humans. Suppression of seven different lncRNAs inhibited various aspects of erythroid maturation including loss of cell volume, down-regulation of the surface marker CD44, and enucleation. The suppression of enucleation ranged from 25% reduction to 75% reduction of enucleation demonstrated with at least 2 separate shRNAs for each of the 7 lncRNAs (p<0.0005) (Figure 1). Of the seven lncRNAs that have a demonstrable function in murine erythropoiesis, none are expressed in human erythroblasts, indicating that lack of conservation between different species does not necessarily predict a lack of function. These results reflect marked evolutionary differences between protein-coding genes and lncRNAs, and indicate that the latter exert tissue- and species-specific roles in mammalian development.Figure 1Figure 1. Disclosures: No relevant conflicts of interest to declare.