Analysis of human ES cell differentiation establishes that the dominant isoforms of the lncRNAs RMST and FIRRE are circular

- Analysis of human ES cell differentiation establishes that the dominant isoforms of the lncRNAs RMST and FIRRE are circular.
- Results: A transcriptome-wide increase in circRNA expression, size, and exon count was observed, with circRNA levels reaching a plateau by day 45.
- outside of the circRNA [7], is required to confirm circular- ity in individual cases..
- CircRNA biogenesis can be promoted by intronic hom- ology and knock-down of the dsRNA editing en- zyme ADAR or the RNA helicase DHX9 which can interact with ADAR1 [12], leads to upregulation of circRNA expression.
- Full list of author information is available at the end of the article.
- 2018 Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0.
- For example, a circRNA isoform of the long non-coding RNA (lncRNA) ANRIL1 has been impli- cated in atheroprotection [29, 30], while circRNAs created by an oncogenic fusion have been reported to influence tumour progression [31]..
- To identify circRNAs potentially involved in the develop- ment of a defined cell lineage, we differentiated triplicate.
- Consistent with the phenotypic changes observed, gene ontologies relevant to the eye were signifi- cantly over-represented among genes differentially expressed between time points (e.g.
- phototransduction, sensory perception of light stimulus, and structural con- stituent of the eye lens, Additional file 2), while known pluripotency genes and eye field markers were downregu- lated and upregulated respectively (Additional file 3:.
- Thus, while few genes are affected by IGF-1, they include logical candidates for involvement in the reported impact of this mitogen [42]..
- 2a), circRNA junction counts (Fig.
- 2b), or canonical exon junction counts (not shown), all identified day 0 samples as a dis- tinct group.
- The limited impact on circRNA levels is also consistent with Enuka et al..
- 2c-d), but they showed a significant ~ 4-5X increase between day 0 and day 45 (p = 0.0044 Wilcoxon rank sum test, Fig.
- Significant increases were also observed in the physical distance between circRNA donor and acceptor exons (genomic span), and the number of exons circRNAs con- tain (Fig.
- Approximately 6% of the circRNAs identified were found only in undifferentiated samples (day 0, Fig.
- To investigate the increase in circRNAs in detail, junction count distributions from all genes were then examined at day 0 and day 45.
- This established that the number of both circRNAs and circRNA junctions increased more dramatic- ally between day 0 and day 45 (Fig.
- To assess these changes at the gene level, ratios of normal- ised circRNA junction counts at day 45 relative to day 0 for each gene were plotted against the corresponding ratios for canonical junctions (Fig.
- Although QKI remained unchanged (data not shown), 2 out of the 3 human Muscleblind genes showed a significant increase by day 45 (Fig.
- 3f-g), consistent with involvement in the.
- b Phylogram based on circRNA (back-splice) junction counts.
- c-d Box and whisker plots of circRNA and canonical junction counts from all genes which generate circRNAs.
- 3 circRNA expression reaches a plateau by day 45.
- a-b Histograms comparing junction counts per transcript in day 0 and day 45.
- c-d Comparison of changes in back-splice and canonical junction count frequencies between time points.
- Junction counts were normalised to total read counts at each time- point.
- Associated p-values from t-tests comparing means in day 0 versus day 45/90 are shown.
- Fisher exact tests were first performed, using junction counts summed across repli- cates, to identify candidate differentially expressed (DE) circRNAs.
- In pairwise comparisons between the day 45 and day 90 samples (rows 5-8 Table 1) between 1 and 256 DE cir- cRNAs were identified using Fisher’s exact tests, but only 1 was significant in either of the subsequent t-tests.
- In contrast, over 2500 transcripts were identified in each pairwise comparison between day 0 samples and later time-points, and 46-496 were significant in the subse- quent t-tests (rows 1-4 Table 1).
- 4a), with more modest changes in canonical junction counts (Fig..
- Thus, these circRNAs differ from the broader popula- tion only in the magnitude/consistency of expression change, not direction.
- In an effort to increase power, this analysis was repeated with day 45 and day 90 samples combined.
- for regulation of specific circRNAs independent of the global upregulation observed prior to day 45..
- The vast majority of the 239 significant DE circRNAs have been identified previously and are derived from coding genes (Additional file 3: Figure S4).
- However, the most abundant is from RMST, a lncRNA known to be involved in the regulation of neural stem cell fate [48]..
- This established that circRNAs from non-coding genes make up a higher proportion of the total transcripts from their loci of origin than cir- cRNAs from coding genes (p <.
- Because of the known involvement of both RMST and FIRRE in developmental processes, we analysed read depth and splice junction frequency in these two genes in detail..
- RMST read and circRNA junction counts were consist- ent between replicates at each time-point, and represen- tative data from day 0 and 45 are shown in Fig.
- day 0 v day 45 IGF .
- day 45 v day 45 IGF .
- day 45 v day .
- Strikingly, when junction counts across all time-points were analysed (Fig.
- 0.5% of E12-E6 at later time-points (e.g.
- 24 and 17 v 5425 at day 45)..
- 5c), and uncovered both a high frequency of intron retention be- tween exons 1 and 2 of uc001tey1 (defined here as E5A and E5B, blue box), and a reduction in reads mapping to this region by day 45.
- a-b Expression heat maps showing relative frequency of untreated sample junction counts from 239 DE transcripts identified in both sample-level and locus-level t-tests (see methods).
- Total numbers of canonical and circRNA junctions in untreated samples at each time-point are also shown, and confirm that the increase in junction counts between days 0 and 45 affects circRNAs from both coding and non-coding genes.
- In contrast, transcription immediately upstream and downstream of the E6-E12 region, which span miRNA precursors and could potentially be involved in generation of the circRNA, showed no consistent change (Fig.
- Overall, the correlation between RNAseq junction counts and qPCR data was 0.93 (Additional file 3: Figure S8) and results are shown in Fig.
- A ~ 100 fold induction of E5- E6 (7-8 Ct) was observed between day 0 and 30, together with a similar increase in the E5-E5B junction by day 45 (both p <.
- 0.01 by day 45).
- However, there was little change in the downstream canonical E12-E13 junction.
- In silico analysis of RNAseq data from human fetal and embryonic eye tissue (Mellough et al.
- This is likely to reflect differences in activity of the single promoter region inferred from ENCODE data (Fig.
- All exon junction counts (Fig.
- By day 45 the E10-E5 junction count (362) was ~ 10 fold higher than the E1-E2 junction count (32), and >.
- This both confirmed the existence of the rearranged tran- script structures inferred from RNAseq data, and estab- lished that exons E5A, E5B, E6A, and the minor exon E10A and E10C, are integrated within multiple circRNAs up to >.
- Of the annotated FIRRE exons, only E1, E2, E4 and E13 were not identified within circRNAs by amplicon sequencing.
- a Modified UCSC browser representation of mapped read count (black) and back-splice junction count (blue) at the RMST locus in untreated replicate 1 at day 0 and day 45.
- Scales are different in the two time-points to allow visualisation of both.
- b Schematic of RMST exon structure, showing circRNA junction counts (above exons) and canonical counts (below exons) summed across untreated replicates for all time-points (0/45/90).
- Exons are numbered relative to the longest ENCODE annotation ENST ENST~ 559.6).
- The correlation between RNAseq-derived junction counts and qPCR data across all FIRRE assays was 0.75, lower than for RMST (Additional file 3: Figure S8)..
- The rela- tively low expression level of all FIRRE exons in replicate A by day 45 (Fig.
- However, relative ex- pression levels of junctions remained consistent within replicates at each time-point, and the fall in E1-E2 and E2- E3 relative to E10-E5 observed in the RNAseq data was both confirmed, and found to be significant.
- In contrast, all FIRRE structures are present at much lower levels in the adult tissues analysed relative to ES cells (Fig.
- In silico analysis of independent publicly available RNAseq datasets also confirmed the low expression of all FIRRE exons in the adult tissues analysed here, as well as skin, breast, muscle, bladder, prostate, ovary, and colon ([54], data not shown)..
- The RMST probe internal to the main E12-E6 circRNA (Fig.
- 99% of the transcriptional output from RMST in differentiating human ES cells, and a variety of adult tissues, is accounted for by this circRNA.
- 50% of the level of associated linear iso- forms in embryonic neuronal tissues [37].
- It follows that subtle changes in the activity of promoters driving both genes could have profound ef- fects upon circRNA levels and may, for instance, underpin the expression heterogeneity observed here for FIRRE at day 45.
- However, non-coding RNAs are enriched for interspersed repeats which can promote cir- cularisation so it is possible that differences in the local genomic environment will contribute to these ele- vated circRNA levels..
- Finally, although there are a growing number of reports of functional circRNAs derived from coding genes some of the clearest evidence for circRNA function involves transcripts from non-coding loci.
- Passage 34-35 of the human embryonic stem cell line H9 line (obtained from WiCell, agreement number 06- W097), were expanded and differentiated according to the protocol outlined in [41], with minor modifications [42] to generate 3D laminated retina containing the major retinal cell types including photoreceptors.
- Phase- bright structures were first observed to develop around the periphery of differentiating EBs, with optic vesicles arising as early as day 15 and optic cups developing by day 45.
- Cells were harvested at multiple time points without any physical selection for embryoid bodies, to ensure cells harvested were representative of the whole population..
- RIN values indicated some reduction in RNA quality over time, with day 0 sample RIN values ranging from 9.8-10, day 45 ranging from 7.6-9.1 and day 90 ranging from 7.1-8.9..
- Exon junction heat maps were generated using junction counts normal- ised to total read counts per sample, and mean expression value of each transcript across samples..
- To account for differences in library size and transcriptome- wide changes in circRNA expression levels between pairs of time points, contingency tables for each circRNA were constructed consisting of circRNA junction counts in both time points, versus total junction counts (circular and canonical) minus the circRNA junction counts for the circRNA being tested..
- To control for locus specific changes in total gene expression between time points, contingency tables for each circRNA were constructed using only junction counts from the locus of origin.
- These consisted of circRNA junction counts in time points A and B, versus canonical junction counts from all annotated exons from the locus being tested..
- Expression heat maps were generated from circRNA junction counts within untreated samples following removal of transcripts present in segmental duplications.
- Additional file 2: a-d.
- No significant genes were identified in the day 90 analysis..
- Additional file 3: Figure S1.
- Junction counts of major RMST and FIRRE circRNAs in human embryonic / fetal samples.
- circRNA: Circular RNA.
- lncRNA: Long non-coding RNA.
- The datasets supporting the conclusions of this article are available in the Gene Expression Omnibus repository [GEO series GSE89957]..
- Circular RNAs in the mammalian brain are highly abundant, conserved, and dynamically expressed.
- DHX9 suppresses RNA processing defects originating from the Alu invasion of the human genome.
- Exon-intron circular RNAs regulate transcription in the nucleus.
- CircRNA accumulation in the aging mouse brain.
- The circular RNA circBIRC6 participates in the molecular circuitry controlling human pluripotency.
- IGF- 1 signaling plays an important role in the formation of three-dimensional laminated neural retina and other ocular structures from human embryonic stem cells.
- Function and evolution of local repeats in the Firre locus.
- PANTHER in 2013: modeling the evolution of gene function, and other gene attributes, in the context of phylogenetic trees

Xem thử không khả dụng, vui lòng xem tại trang nguồn
hoặc xem Tóm tắt