Heat shock response in yeast involves changes in both transcription rates and mRNA stabilities

Abstract

We have analyzed the heat stress response in the yeast Saccharomyces cerevisiae by determining mRNA levels and transcription rates for the whole transcriptome after a shift from 25 °C to 37 °C. Using an established mathematical algorithm, theoretical mRNA decay rates have also been calculated from the experimental data. We have verified the mathematical predictions for selected genes by determining their mRNA decay rates at different times during heat stress response using the regulatable tetO promoter. This study indicates that the yeast response to heat shock is not only due to changes in transcription rates, but also to changes in the mRNA stabilities. mRNA stability is affected in 62% of the yeast genes and it is particularly important in shaping the mRNA profile of the genes belonging to the environmental stress response. In most cases, changes in transcription rates and mRNA stabilities are homodirectional for both parameters, although some interesting cases of antagonist behavior are found. The statistical analysis of gene targets and sequence motifs within the clusters of genes with similar behaviors shows that both transcriptional and post-transcriptional regulons apparently contribute to the general heat stress response by means of transcriptional factors and RNA binding proteins.

Figure 1. Time course of the heat shock experiment.

At time 0, cells growing exponentially at 25°C were shifted to 37°C. At the indicated times, aliquots were taken to measure cell concentration, total mRNA per cell (RA) and Pol II transcription rate (TR) per cell (see “Materials and Methods”). The parameters were referred to their respective time 0 values. Bars: standard deviation (n = 3).

Figure 2. Clustering of TR and RA data.

Time course profiles for both parameters were considered for clustering. Both dataset series are normalized to time 0 value to allow a comparison between the TR and RA data. The discontinuation between the last TR point and the time 0 RA value has no real meaning and is represented as a vertical black bar. For each cluster in the tree, the number of genes and data profiles are indicated. Ordinates are on a log scale, and the horizontal line in each graph marks the time 0 level. The most significant GO categories (p value ≤10⁻⁵) are shown. The individual data for each gene and the list of genes in each cluster can be seen in Tables S1 and S2, respectively. The scale bar on the lower left side reflects the distances between the cluster profiles.

Figure 3. Sub-partition of genes in clusters 11 and 12 and Transcription Factor enrichment analysis.

The genes in clusters 11 and 12 of Figure 2 (listed in Table S2) were partitioned into subclusters using the same criteria as for the general clustering analysis. See the legend of Figure 2 for details of the representation. The most significant GO categories in each resulting subcluster (p value≤10⁻⁵) are shown on the right. Transcription factor targets enrichment in each cluster (left) or subcluster (right) is shown with its p-value. nf: not significant categories found.

Figure 4. mRNA kinetics of the genes in the 16 clusters upon heat shock.

RA values are represented in the y axis as a function of time (min) (shift from 25°C to 37°C at time 0). Experimental RA values (continuous lines) were determined as indicated in the text. Theoretical RA values (dashed lines) were determined from the experimental TR values by assuming a constant k_D identical to that of time 0. Graphics represent the mean values corresponding to all the genes in the indicated cluster in relative units, referring to the mean value at time 0. Note that different scales are employed for the y axis depending on the cluster.

Figure 5. RNA binding protein (RBP) and mRNA 3′UTR motif enrichment analyses.

Supervised and unsupervised analyses for RBP and mRNA 3′UTR motif enrichment, were carried out. In the left part of the panel, the p-values associated with the presence of the well established RBP binding motifs in the clusters are displayed. On the right, the consensus sequence motifs found in the 3′UTR of the genes belonging to the clusters are shown, as obtained by the FIRE algorithm . Only those clusters with over-representation of RBPs or significant 3′UTR motifs are shown.

Figure 6. Experimental determination of mRNA half-lives before and after heat shock.

Strains expressed HSP42 (A, strain MML980) or RRP40 (B, strain MML957) under the control of the tetO₂ promoter. Doxycycline (5 µg/ml) was added at time 0 to cultures growing exponentially at 25°C, or to cultures at 4 and 45 min after the shift to 37°C. In each case, aliquots were taken at time 0 and at successive times after adding doxycycline for total mRNA isolation and determination of the levels of the corresponding mRNA by Northern analysis. Graphics represent the evolution of experimentally determined relative RA on a log scale as a function of time for the representative experiment mean half-life values, and the standard deviations of a total of three independent experiments are also indicated. To determine these values, a linear regression of experimental data was calculated by considering only the initial points for which linearity was maintained. The upper panels show the Northern analyses of RA expressed under the respective own promoters in wild-type W303-1A cells growing at 25°C, after being shifted at time 0 to 37°C. U1 is included as the loading control. The Kruskal-Wallis statistical test was applied to show that the differences between the calculated half lives were significantly different from the others for time point 4 min (HSP42) and time 0 (RRP40) and not for the other cases.

Figure 7. TR and real and theoretical RA values of protein folding genes after heat shock.

Genes in the GO category “Protein folding” (listed in Table S3) were considered for analysis, after being subdivided between those in clusters 7–10, 11 or 12. Mean values for the three parameters were calculated and plotted as a function of time after the heat shock from 25 to 37°C. The values are represented relative to the mean value at time 0. Bars indicate the standard error for each time point. Theoretical RA values were calculated as indicated in the legend of Figure 4.