Transcriptome kinetics is governed by a genome-wide coupling of mRNA production and degradation: a role for RNA Pol II

Abstract

Transcriptome dynamics is governed by two opposing processes, mRNA production and degradation. Recent studies found that changes in these processes are frequently coordinated and that the relationship between them shapes transcriptome kinetics. Specifically, when transcription changes are counter-acted with changes in mRNA stability, transient fast-relaxing transcriptome kinetics is observed. A possible molecular mechanism underlying such coordinated regulation might lay in two RNA polymerase (Pol II) subunits, Rpb4 and Rpb7, which are recruited to mRNAs during transcription and later affect their degradation in the cytoplasm. Here we used a yeast strain carrying a mutant Pol II which poorly recruits these subunits. We show that this mutant strain is impaired in its ability to modulate mRNA stability in response to stress. The normal negative coordinated regulation is lost in the mutant, resulting in abnormal transcriptome profiles both with respect to magnitude and kinetics of responses. These results reveal an important role for Pol II, in regulation of both mRNA synthesis and degradation, and also in coordinating between them. We propose a simple model for production-degradation coupling that accounts for our observations. The model shows how a simple manipulation of the rates of co-transcriptional mRNA imprinting by Pol II may govern genome-wide transcriptome kinetics in response to environmental changes.

Figure 1. A schematic illustration of the experimental procedure.

For each strain three types of experiments were conducted: (i) a reference decay experiment where decay kinetics was measured after transcription inhibition without applying additional stress. (ii) A stress followed by transcription inhibition to measure condition specific decay kinetics, and (iii) A conventional microarray experiment where mRNA abundance was measured following the perturbation.

Figure 2. Reduced coupling in Rpb6 mutant strain.

The change in mRNA stability relative to the reference state () is plotted against the maximal fold change, defined as the maximal change in mRNA abundance for each gene across the time course. (A) Shows the wild type strain where black dots marks genes which respond to the stress. A negative correlation bywhich induced genes are destabilized is illustrated by the plotted least square line (R = −0.23, −log₁₀(p-value)>58). Fitted line y = −0.97x−0.12 with (−1.14, −0.7947) and (−0.1875, −0.05864) 95% confidence interval for each parameter. (B) The mutant measurements are plotted on top of the wild type. The negative correlation observed in the wild type, reflected by the black least straight line, is almost completely eliminated, displayed in the green least square line (R = −0.06, −log₁₀(p-value)<6). Fitted line y = 0.1852x+0.27 with (−0.1723, 0.5426) and (0.1596, 0.3805) confidence interval for each parameter. Also the width of the distribution across the x-axis is slightly narrower for the mutant strain an indication for reduced ability to modulate mRNA stability. (C) Fold enrichment for the change in stability in induced and repressed genes. The number of observed stabilized and destabilized genes within both induced and repressed groups of genes divided by the expected number, assuming no correlation. Expected number is calculated as the percentage of stabilized/destabilized genes in the genome times the induced/repressed group size.

Figure 3. Impaired response of Rbp6 mutant strain.

Hierarchical clustering of mRNA abundance temporal profiles. The three clusters marked in the dendrogram by different gray colors correspond to the three clusters in the middle and right panels. Middle panels show the fold change as a function of time for both strains, gray for wild type and green for mutant, with thick lines representing the mean of each strain. Left panels show the difference in stability in response to stress, calculated as the log2 ratio of the stress half-life by the reference half-life. The wild type value is plotted against the mutant with a blue least square line showing the general trend for each cluster. To show the slope difference between the two induced clusters fit parameters are also displayed for all clusters.

Figure 4. Wild type and mutant differences in basal conditions.

(A) correlation of steady state mRNA abundance measurements between the wild type and mutant strain. A global reduction in mRNA levels is observed in the mutant. Data is plotted in log2 scale. (B) Basal difference in stability (log 2 ratio of the reference half-lives of the mutant divided by the wild type) is plotted against the difference in mRNA abundance (log 2 ratio of basal mRNA abundance in the WT and mutant). Most changing genes show both a reduction and stabilization in the mutant strain. The black least square line shows the negative correlation between these two parameters. Fitted line y = −0.3986x−0.4557 with (−0.4183, −0.3789) and (−0.4636, −0.4477) 95% confidence intervals for each parameter. Blue line represents no change in mRNA abundance. (C) Results of a gene ontology enrichment analysis for the genes that show the largest reduction in mRNA abundance in the mutant strain.

Figure 5. A schematic illustration of the suggested model explaining the genome-wide effect of Rpb4/7.

The figure shows how an increase in the association probability following stress causes a global redistribution of the general decay machinery resulting in genome wide coupling of changes in mRNA abundance and stability. Theoretical probability for a transcript to be either imprinted or non-imprinted by Rpb4/7 during transcription in basal conditions (A) and for stress (B). A more detailed explanation is given in the main text.