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

Timing of early peak CAGE clusters.(A) Bar charts showing the percentage of early peak clusters associated with IEGs (ts binned in 30 min intervals), and (B) and those associated with nucleotide binding genes. The horizontal line indicates the average percentage. (C) The timing of known IEGs and transcription factors is shown for IEGs (red) and TFs (purple) assigned to the early peak signature in each MCF7 experiment. Symbols indicate the ts (plotted on the x axis) and are labelled with the gene name associated with the CAGE cluster (symbols are positioned on the y axis for legibility only).
© Copyright Policy
Related In: Results  -  Collection

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

pcbi.1004217.g002: Timing of early peak CAGE clusters.(A) Bar charts showing the percentage of early peak clusters associated with IEGs (ts binned in 30 min intervals), and (B) and those associated with nucleotide binding genes. The horizontal line indicates the average percentage. (C) The timing of known IEGs and transcription factors is shown for IEGs (red) and TFs (purple) assigned to the early peak signature in each MCF7 experiment. Symbols indicate the ts (plotted on the x axis) and are labelled with the gene name associated with the CAGE cluster (symbols are positioned on the y axis for legibility only).

Mentions: It has been proposed that nucleotide binding proteins are among the feedback regulators responsible for the attenuation of the immediate-early response to EGF [10]. This set of delayed early genes has been shown to be activated in waves following FOS, JUN and EGR1 expression in HeLa cells [10]. Following Amit et al. (2007) [10], we constructed a set of 444 genes assigned with the Gene Ontology annotation for nucleotide binding GO:0000166 (IEA assignments were excluded). Only three of these genes were transcription factors and six were IEGs therefore these sets were essentially disjoint. The set of nucleotide binding genes was over-represented in the early peak signature in all four data sets combined (p = 0.007), and for AoSMC-FGF2 and MCF7-EGF data sets individually (p = 0.018 and p = 0.003 respectively). The timing of immediate early and nucleotide binding gene expression is shown in Fig 2A and 2B where it can be seen that in AoSMC-FGF2, AoSMC-IL1b and MCF7-EGF data the largest proportion of known IEGs is found in the 30-90 min interval when ts values are binned in 30 min intervals (the proportion of clusters annotated to known IEGs is expressed as a percentage of all clusters within each 30 min period according to ts). This pattern was less apparent in the MCF7-HRG cell line where the proportion of known IEGs found in an interval exceeded the overall average towards the end of the time course. Further, in AoSMC-FGF2, AoSMC-IL1b and HRG treated MCF7 cells there was a peak in IEGs around 3 hours after stimulus (180-210min) suggesting genes currently described as IEGs may also have a role later in the immediate-early response than would be expected. Surprisingly, the proportion of nucleotide binding genes was maximal in the 60-90 min interval for the AoSMC-IL1b and MCF7-EGF data, and in all cases many nucleotide binding genes were activated concurrently with IEGs in contrast with previous reports [10]. It is also worth noting that significant downregulation did not occur until the second hour, and this may require both early induction of transcriptional repressors and the RNA degradation proteins BTG2 and ZFP36 (tristetraprolin) [29].


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)

Timing of early peak CAGE clusters.(A) Bar charts showing the percentage of early peak clusters associated with IEGs (ts binned in 30 min intervals), and (B) and those associated with nucleotide binding genes. The horizontal line indicates the average percentage. (C) The timing of known IEGs and transcription factors is shown for IEGs (red) and TFs (purple) assigned to the early peak signature in each MCF7 experiment. Symbols indicate the ts (plotted on the x axis) and are labelled with the gene name associated with the CAGE cluster (symbols are positioned on the y axis for legibility only).
© Copyright Policy
Related In: Results  -  Collection

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

pcbi.1004217.g002: Timing of early peak CAGE clusters.(A) Bar charts showing the percentage of early peak clusters associated with IEGs (ts binned in 30 min intervals), and (B) and those associated with nucleotide binding genes. The horizontal line indicates the average percentage. (C) The timing of known IEGs and transcription factors is shown for IEGs (red) and TFs (purple) assigned to the early peak signature in each MCF7 experiment. Symbols indicate the ts (plotted on the x axis) and are labelled with the gene name associated with the CAGE cluster (symbols are positioned on the y axis for legibility only).
Mentions: It has been proposed that nucleotide binding proteins are among the feedback regulators responsible for the attenuation of the immediate-early response to EGF [10]. This set of delayed early genes has been shown to be activated in waves following FOS, JUN and EGR1 expression in HeLa cells [10]. Following Amit et al. (2007) [10], we constructed a set of 444 genes assigned with the Gene Ontology annotation for nucleotide binding GO:0000166 (IEA assignments were excluded). Only three of these genes were transcription factors and six were IEGs therefore these sets were essentially disjoint. The set of nucleotide binding genes was over-represented in the early peak signature in all four data sets combined (p = 0.007), and for AoSMC-FGF2 and MCF7-EGF data sets individually (p = 0.018 and p = 0.003 respectively). The timing of immediate early and nucleotide binding gene expression is shown in Fig 2A and 2B where it can be seen that in AoSMC-FGF2, AoSMC-IL1b and MCF7-EGF data the largest proportion of known IEGs is found in the 30-90 min interval when ts values are binned in 30 min intervals (the proportion of clusters annotated to known IEGs is expressed as a percentage of all clusters within each 30 min period according to ts). This pattern was less apparent in the MCF7-HRG cell line where the proportion of known IEGs found in an interval exceeded the overall average towards the end of the time course. Further, in AoSMC-FGF2, AoSMC-IL1b and HRG treated MCF7 cells there was a peak in IEGs around 3 hours after stimulus (180-210min) suggesting genes currently described as IEGs may also have a role later in the immediate-early response than would be expected. Surprisingly, the proportion of nucleotide binding genes was maximal in the 60-90 min interval for the AoSMC-IL1b and MCF7-EGF data, and in all cases many nucleotide binding genes were activated concurrently with IEGs in contrast with previous reports [10]. It is also worth noting that significant downregulation did not occur until the second hour, and this may require both early induction of transcriptional repressors and the RNA degradation proteins BTG2 and ZFP36 (tristetraprolin) [29].

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