Limits...
Transcriptional dynamics reveal critical roles for non-coding RNAs in the immediate-early response.

Aitken S, Magi S, Alhendi AM, Itoh M, Kawaji H, Lassmann T, Daub CO, Arner E, Carninci P, Forrest AR, Hayashizaki Y, FANTOM ConsortiumKhachigian LM, Okada-Hatakeyama M, Semple CA - PLoS Comput. Biol. (2015)

Bottom Line: Surprisingly, these data suggest that the earliest transcriptional responses often involve promoters generating non-coding RNAs, many of which are produced in advance of canonical protein-coding IEGs.Consistent with this, we find that the response of both protein-coding and non-coding RNA IEGs can be explained by their transcriptionally poised, permissive chromatin state prior to stimulation.Our computational statistical method is well suited to meta-analyses as there is no requirement for transcripts to pass thresholds for significant differential expression between time points, and it is agnostic to the number of time points per dataset.

View Article: PubMed Central - PubMed

Affiliation: MRC Institute of Genetics and Molecular Medicine, University of Edinburgh, Edinburgh, United Kingdom.

ABSTRACT
The immediate-early response mediates cell fate in response to a variety of extracellular stimuli and is dysregulated in many cancers. However, the specificity of the response across stimuli and cell types, and the roles of non-coding RNAs are not well understood. Using a large collection of densely-sampled time series expression data we have examined the induction of the immediate-early response in unparalleled detail, across cell types and stimuli. We exploit cap analysis of gene expression (CAGE) time series datasets to directly measure promoter activities over time. Using a novel analysis method for time series data we identify transcripts with expression patterns that closely resemble the dynamics of known immediate-early genes (IEGs) and this enables a comprehensive comparative study of these genes and their chromatin state. Surprisingly, these data suggest that the earliest transcriptional responses often involve promoters generating non-coding RNAs, many of which are produced in advance of canonical protein-coding IEGs. IEGs are known to be capable of induction without de novo protein synthesis. Consistent with this, we find that the response of both protein-coding and non-coding RNA IEGs can be explained by their transcriptionally poised, permissive chromatin state prior to stimulation. We also explore the function of non-coding RNAs in the attenuation of the immediate early response in a small RNA sequencing dataset matched to the CAGE data: We identify a novel set of microRNAs responsible for the attenuation of the IEG response in an estrogen receptor positive cancer cell line. Our computational statistical method is well suited to meta-analyses as there is no requirement for transcripts to pass thresholds for significant differential expression between time points, and it is agnostic to the number of time points per dataset.

No MeSH data available.


Related in: MedlinePlus

Non-coding RNA gene activation.(A) Histograms of ts for lncRNA, snoRNA, snRNA and miRNA precursors show these genes are activated rapidly. (B) Density plot of early peak gene length against ts for all RNA biotypes (grey symbols), lncRNA (red symbols) and miRNA precursors (blue symbols). LncRNA and miRNA form distinct clusters of RNAs activated with a wide range of kinetics.
© Copyright Policy
Related In: Results  -  Collection

License
getmorefigures.php?uid=PMC4401570&req=5

pcbi.1004217.g004: Non-coding RNA gene activation.(A) Histograms of ts for lncRNA, snoRNA, snRNA and miRNA precursors show these genes are activated rapidly. (B) Density plot of early peak gene length against ts for all RNA biotypes (grey symbols), lncRNA (red symbols) and miRNA precursors (blue symbols). LncRNA and miRNA form distinct clusters of RNAs activated with a wide range of kinetics.

Mentions: The role of ncRNA in the immediate-early response is not well understood. A small number of mature miRNA that respond to EGF signalling have been identified [12, 35], and, in yeast, lncRNA have been shown to poise GAL genes for rapid activation [36]. Clusters assigned to the early peak signature were over-represented relative to other signatures for lncRNA and miRNA precursors (p = 0.013 and 3.8e-3, respectively, by hypergeometric test), and late peak and decay signatures were over-represented for snRNA (p = 1.9e-4 and 0.022 respectively). Thus lncRNA and miRNA precursors showed an analogous kinetic response to IEGs. The distributions of ts for lncRNA, snoRNA, snRNA and miRNA assigned to the early peak category are shown in Fig 4A. These distributions can be compared with those for protein coding genes, including known IEGs (S7 Fig). These RNA biotypes had more rapid kinetics as shown by the number with ts of less than 30 min.


Transcriptional dynamics reveal critical roles for non-coding RNAs in the immediate-early response.

Aitken S, Magi S, Alhendi AM, Itoh M, Kawaji H, Lassmann T, Daub CO, Arner E, Carninci P, Forrest AR, Hayashizaki Y, FANTOM ConsortiumKhachigian LM, Okada-Hatakeyama M, Semple CA - PLoS Comput. Biol. (2015)

Non-coding RNA gene activation.(A) Histograms of ts for lncRNA, snoRNA, snRNA and miRNA precursors show these genes are activated rapidly. (B) Density plot of early peak gene length against ts for all RNA biotypes (grey symbols), lncRNA (red symbols) and miRNA precursors (blue symbols). LncRNA and miRNA form distinct clusters of RNAs activated with a wide range of kinetics.
© Copyright Policy
Related In: Results  -  Collection

License
Show All Figures
getmorefigures.php?uid=PMC4401570&req=5

pcbi.1004217.g004: Non-coding RNA gene activation.(A) Histograms of ts for lncRNA, snoRNA, snRNA and miRNA precursors show these genes are activated rapidly. (B) Density plot of early peak gene length against ts for all RNA biotypes (grey symbols), lncRNA (red symbols) and miRNA precursors (blue symbols). LncRNA and miRNA form distinct clusters of RNAs activated with a wide range of kinetics.
Mentions: The role of ncRNA in the immediate-early response is not well understood. A small number of mature miRNA that respond to EGF signalling have been identified [12, 35], and, in yeast, lncRNA have been shown to poise GAL genes for rapid activation [36]. Clusters assigned to the early peak signature were over-represented relative to other signatures for lncRNA and miRNA precursors (p = 0.013 and 3.8e-3, respectively, by hypergeometric test), and late peak and decay signatures were over-represented for snRNA (p = 1.9e-4 and 0.022 respectively). Thus lncRNA and miRNA precursors showed an analogous kinetic response to IEGs. The distributions of ts for lncRNA, snoRNA, snRNA and miRNA assigned to the early peak category are shown in Fig 4A. These distributions can be compared with those for protein coding genes, including known IEGs (S7 Fig). These RNA biotypes had more rapid kinetics as shown by the number with ts of less than 30 min.

Bottom Line: Surprisingly, these data suggest that the earliest transcriptional responses often involve promoters generating non-coding RNAs, many of which are produced in advance of canonical protein-coding IEGs.Consistent with this, we find that the response of both protein-coding and non-coding RNA IEGs can be explained by their transcriptionally poised, permissive chromatin state prior to stimulation.Our computational statistical method is well suited to meta-analyses as there is no requirement for transcripts to pass thresholds for significant differential expression between time points, and it is agnostic to the number of time points per dataset.

View Article: PubMed Central - PubMed

Affiliation: MRC Institute of Genetics and Molecular Medicine, University of Edinburgh, Edinburgh, United Kingdom.

ABSTRACT
The immediate-early response mediates cell fate in response to a variety of extracellular stimuli and is dysregulated in many cancers. However, the specificity of the response across stimuli and cell types, and the roles of non-coding RNAs are not well understood. Using a large collection of densely-sampled time series expression data we have examined the induction of the immediate-early response in unparalleled detail, across cell types and stimuli. We exploit cap analysis of gene expression (CAGE) time series datasets to directly measure promoter activities over time. Using a novel analysis method for time series data we identify transcripts with expression patterns that closely resemble the dynamics of known immediate-early genes (IEGs) and this enables a comprehensive comparative study of these genes and their chromatin state. Surprisingly, these data suggest that the earliest transcriptional responses often involve promoters generating non-coding RNAs, many of which are produced in advance of canonical protein-coding IEGs. IEGs are known to be capable of induction without de novo protein synthesis. Consistent with this, we find that the response of both protein-coding and non-coding RNA IEGs can be explained by their transcriptionally poised, permissive chromatin state prior to stimulation. We also explore the function of non-coding RNAs in the attenuation of the immediate early response in a small RNA sequencing dataset matched to the CAGE data: We identify a novel set of microRNAs responsible for the attenuation of the IEG response in an estrogen receptor positive cancer cell line. Our computational statistical method is well suited to meta-analyses as there is no requirement for transcripts to pass thresholds for significant differential expression between time points, and it is agnostic to the number of time points per dataset.

No MeSH data available.


Related in: MedlinePlus