Sequential recruitment of SAGA and TFIID in a genomic response to DNA damage in Saccharomyces cerevisiae

Abstract

Eukaryotic genes respond to their environment by changing the expression of selected genes. The question we address here is whether distinct transcriptional responses to different environmental signals elicit distinct modes of assembly of the transcription machinery. In particular, we examine transcription complex assembly by the stress-directed SAGA complex versus the housekeeping assembly factor TFIID. We focus on genomic responses to the DNA damaging agent methyl methanesulfonate (MMS) in comparison to responses to acute heat shock, looking at changes in genome-wide factor occupancy measured by chromatin immunoprecipitation-microchip (ChIP-chip) and ChIP-sequencing analyses. Our data suggest that MMS-induced genes undergo transcription complex assembly sequentially, first involving SAGA and then involving a slower TFIID recruitment, whereas heat shock genes utilize the SAGA and TFIID pathways rapidly and in parallel. Also Crt1, the repressor of model MMS-inducible ribonucleotide reductase genes, was found not to play a wider role in repression of DNA damage-inducible genes. Taken together, our findings reveal a distinct involvement of gene and chromatin regulatory factors in response to DNA damage versus heat shock and suggest different implementations of the SAGA and TFIID assembly pathways that may depend upon whether a sustained or transient change in gene expression ensues.

FIG. 1.

Changes in factor occupancy in response to 30 min of MMS treatment in comparison to a heat shock response. (A) Venn diagram illustrating the gene membership overlap for genes in equivalently labeled clusters in panel B. (B) Cluster plots of changes in factor occupancy and gene expression (as indicated). MMS-treated and heat shock-treated data sets were clustered separately. Data were filtered to retain only those having 100% data present and >1.5-fold changes in occupancy in at least one data set. The numbers of genes meeting such criteria are indicated. Data were clustered by K means (K = 4 for MMS and K = 3 for heat shock). BY4741 represents a negative control in which the untagged parental strain was processed through the standard TAP-ChIP procedure. Data shown are for the “UAS” and “TSS” microarray probes. The Spt3 ChIP data set used the “UAS” probe, whereas the remaining factors used the “TSS” probe. (C) Median log₂ values for data sets in cluster 1.

FIG. 2.

Factor occupancy levels at RNR genes and MMS-induced genes in comparison to heat shock-induced genes. (A to E) Each panel tracks the indicated GTF and reports the median occupancy level for the indicated gene or set of genes. Values represent medians for cluster 1 data, represented as log₂ fold changes over background in the presence or absence of MMS or heat shock, as indicated. The same data used to plot fold changes in occupancy for Fig. 1 were used here. “Low” and “high” represent the bottom and top 10th percentiles of transcription frequency, as defined by Holstege et al. (17). “RP” denotes ribosomal protein genes. The “TSS” probe was used for all data except for the SAGA data, for which the “UAS” probe was used, which is where SAGA is thought to bind (23, 43). For the RNR4 gene, the “UAS” probe was used instead of the “TSS” probe because it was located closer to the known core promoter region. This places it out of range of where SAGA might be expected to bind. The absence of a significant signal from the RNR3 TSS probe (used for Sua7, Taf1, and Rpo21) may be due to a defect in the probe. Values for the RNR genes represent the averages of two measurements, whereas values for gene sets represent median values of >200 measurements (two replicates of >100 genes) and thus are more robust.

FIG. 3.

Delayed acquisition of TFIID at MMS-induced genes, as opposed to simultaneous acquisition of TFIID and SAGA at heat shock-induced genes. (A) Genome-wide Spt3 and Taf1 occupancy changes (ChIP-chip) after 0.5 and 2 h of MMS induction for MMS-induced genes (cluster 1). (B) Changes in occupancy (ChIP-seq) of Spt3 and Taf1 after 0.5 h, 1 h, and 2 h of MMS induction at the MMS-inducible genes. (C) Corresponding changes in mRNA levels are shown. *, the expression data are from the work of Gasch et al. (12). (D) Changes in occupancy (ChIP-seq) of Spt3 and Taf1 following heat shock induction for 5 and 15 min at heat shock-inducible genes. (E) Corresponding changes in mRNA are shown. *, expression data are from the work of Gasch et al. (13); a*, expression data are from the work of Zanton and Pugh (50).

FIG. 4.

Occupancy levels at genes that are repressed or unaffected by MMS or heat shock. Occupancy levels are reported for genes in cluster 2 (A; repressed or downregulated) and cluster 3 (B; genes that acquire SAGA but are not immediately activated), as described in the legend to Fig. 2.

FIG. 5.

Crt1 is not linked to MMS-induced genes. (A) Crt1 motif, obtained using MEME, for the significantly bound genes. (B) Venn diagram illustrating the overlap between Crt1- and Tup1-enriched genes genome-wide. Only the topmost 100 Crt1-occupied genes and the topmost 250 Tup1-occupied genes were used. Genes with lower measured occupancy levels of these factors (false-negative results) became less distinguishable from false-positive results and thus were not used. Such stringent filtering would therefore limit the degree of overlap. (C) Bar graph representing Crt1 and Tup1 occupancy, represented as log₂ fold changes over background in the presence or absence of MMS at RNR genes. UAS and TSS probe distributions along with Crt1 X-box sites at the RNR2, RNR3, and RNR4 genes are shown. (D) Bar graph illustrating Crt1 and Tup1 occupancies, represented as log₂ fold changes over background in the presence or absence of MMS at MMS-inducible genes. Data shown are for the “UAS” microarray probes. (E) Venn diagram representing the overlap between MMS-inducible genes (defined in the legend to Fig. 1A and B) and the Crt1- and Tup1-enriched genes.

FIG. 6.

Constitutive presence of many chromatin remodelers at DNA damage-inducible genes. Each panel tracks the indicated chromatin remodeler and reports the median occupancy levels at the “UAS” probe for the MMS- and heat shock-inducible genes along with the RNR genes. Median occupancy values are represented as log₂ fold changes over background in the presence or absence of MMS or heat shock, as indicated. The occupancy levels at the top 5th and 50th percentiles of all genes were also plotted to set the dynamic range.

FIG. 7.

Increased occupancy of Gcn4 and SAGA at MMS-induced amino acid biosynthetic genes. (A) Frequency distribution plot illustrating changes in occupancy of Gcn4 at the UAS probes in response to heat shock and MMS. (B) Bar graph representing changes in expression of the 31 amino acid biosynthetic genes in response to MMS or heat shock. Changes in expression of MMS- or heat shock-inducible genes (cluster 1 genes from Fig. 1) were used to set the dynamic range. (C) Each panel tracks the indicated GTF and reports the median changes in occupancy in response to MMS or heat shock at the amino acid biosynthetic genes as well as the MMS- or heat shock-inducible (cluster 1) genes.