Reallocation of RNA Pol II upon osmotic shock. (A) Dynamics of depletion and recruitment of RNA Pol II in stress were measured by a time course of ChIP-QPCR experiments. Cells in log growth were subjected to osmotic shock by addition of 0.4 M KCl to the growth medium, and samples were collected at the indicated time points. Housekeeping genes with high RNA Pol II occupancy pre-stress are quickly depleted of Pol II upon stress and begin recovering within 10 min (average behavior of PDC1, TDH3, ILV5). Stress-induced genes recruit polymerase upon stress treatment and then rapidly return to low RNA Pol II occupancy (average behavior of RTC3 and STL1). Data shown are the averages of biological replicate experiments. (B) RNA Pol II occupancy was measured by ChIP-sequencing of Rpb3, the third largest subunit of RNA Pol II, in control (mock-treated with YEPD) cells and in cells subjected to 5 min of osmotic shock induced by addition of 0.4 M KCl to the growth medium for 5 min. Histograms show the distribution of RNA Pol II occupancy across all yeast ORFs in each condition. RNA Pol II occupancy is measured at each ORF in reads per kilobase per million (RPKM), after subtracting the input value. Regions of overlap between two or more ORFs were excluded from analysis. Inset plots show the 95^th percentile for RNA Pol II occupancy in each condition; the 95^th percentile is indicated by a dashed line in the main plot. (C) Scatter plot showing RNA Pol II occupancy during rapid growth in YEPD (x-axis) vs. osmotic shock in 0.4 M KCl for 5 min (y-axis). The number of reads that align to each ORF is normalized to RPKM. Each point represents one gene, and points are color-coded to show ORFs that are occupied by Hog1 in response to stress (red points). (D) Distinct sets of genes are top RNA Pol II targets in the presence and absence of stress. RNA Pol II occupancy in rapid growth (gray plots) and after osmotic shock for 5 min in 0.4M KCl (black plots). Values plotted are the number of reads that align at each point along the chromosome, and each ChIP-seq dataset is scaled to one million reads. Arrows indicate the position and direction of each open reading frame. The five highest occupancy ORFs during osmotic shock are shown. (E) The five highest occupancy ORFs during rapid growth in YEPD are plotted, as in (D).

Reallocation of RNA Pol II upon osmotic shock is Hog1-dependent. (A) RNA Pol II is depleted from a heterologous gene upon stress. The TET-ON system was used to drive expression of P_TETO7-LACZ, encoded on a plasmid, to a high level in wild-type cells or in a strain lacking Hog1. Induced cells were then subjected to osmotic shock (as described in (A). RNA Pol II enrichment over a featureless (subtelomeric) region was measured at LACZ and at two control genes: ADH1, which shows high RNA Pol II occupancy in the absence of stress, and RTC3, which shows stress-induced RNA Pol II occupancy. Values plotted are relative RNA Pol II enrichment values for each gene. Error bars show the standard deviation of three biological replicates. (B) Scatter plot showing RNA Pol II occupancy in a Δhog1 strain during rapid growth in YEPD (x-axis) vs. osmotic shock in 0.4 M KCl for 5 min (y-axis). Each point represents one gene. The number of reads that align to each ORF are normalized to reads per kilobases per million (RPKM).

The Hog1 cognate TFs Sko1 and Hot1 target stress-induced genes for transcription by a Hog1–RNA Pol II complex. (A) Colocalization of Hog1, Sko1, and Hot1 at example genes occupied by Hog1 during stress. Lighter colored overlay for Pol II and Hog1 ChIP-seq shows ChIP signal in the Δsko1Δhot1 strain. Black arrows and dashed gray guidelines mark the position of the GPD1, RHR2, STL1, and SCM4 ORFs, and white triangles show predicted transcription factor binding sites. Values plotted are the number of reads that align at each point along the chromosome, and each ChIP-seq dataset is scaled to one million reads. Highlighted regions are those shown to be sufficient for Hog1 ORF occupancy. (B) RNA Pol II and Hog1 presence in the LACZ ORF driven by three different promoters (defined as 1000 base pairs upstream of each gene): the stress gene promoters P_GPD1 and P_RHR2, and P_CYC1, an inactive promoter that serves as a negative control. These promoter-LACZ constructs are carried on a plasmid. Hog1 and RNA Pol II ChIP enrichment was measured by qPCR during rapid growth in SD medium lacking tryptophan to maintain the plasmid (bars labeled −) and in response to osmotic shock induced by addition of 0.4 M KCl to the medium (bars labeled +). Enrichment values are normalized by ChIP to POL1. For Hog1 enrichment, no-stress values were normalized to one, and stress values show the relative change in enrichment. Error bars show the standard deviation of three biological replicates.

Analysis of binding determinants of Sko1 and Hot1. (A) Binding behavior (left panel) and discovered motifs (right panel) for Hot1 and two classes of Sko1 binding sites observed by ChIP-seq. Sko1 and Hot1 binding is shown in control conditions and in osmotic shock (treatment with 0.4 M KCl for 5 min. Sko1 no-stress peaks show decreased enrichment upon osmotic shock (example peaks from FSH1 and PHD1 promoters, top left panel.) Stress-induced Sko1 (center left panel) and Hot1 (bottom panel) binding peaks increase in ChIP enrichment upon stress (example peaks in STL1 and RTC3 promoters). A bioinformatics search of 43 pre-stress Sko1 binding sequences reproduces a previously reported Sko1 binding motif, 5′-TGACGT-3′ (). A bioinformatics search of 17 stress-induced Sko1 binding peak sequences reveals a low-information variant of the Sko1 consensus site. A search of 11 high-confidence Hot1 binding peaks allows de novo discovery of a candidate motif for Hot1. (B) Hog1 enrichment during stress at ORFs, promoters, and 3′ regions decreases in the absence of Sko1 and Hot1. Box and whisker plot shows Hog1 enrichment in the ORFs, promoters (1000 bp upstream) and 3′ regions (1000 bp downstream) for the set of 28 ORFs occupied by Hog1 during stress in wild-type cells. Enrichment is defined as Hog1 ChIP-seq reads minus mock IP, divided by the input signal. Enrichment values in wild-type (wt) cells are shown in dark blue; and enrichment values for Δsko1Δhot1 cells are shown in light blue. The edges of each box show the 25^th and 75^th percentiles, and the line within each box indicates the median. Whiskers extend to the highest and lowest values that are not considered outliers (three standard deviations above or below the median). Outliers are plotted as individual points. (C) Stress-induced Sko1 binding requires Hog1, but not Hog1 phosphorylation, of Sko1. Sko1 binding was measured by ChIP against the HA epitope in wild-type, Δhog1, and sko1- S108A, T113A, S126A (which cannot be phosphorylated by Hog1) strains, all tagged with three copies of the HA tag at the N′ of the endogenous copy of Sko1. ChIP experiments were performed on cells in log-phase growth in rich medium, in the presence or absence of osmotic shock (induced by 5 min in 0.4 M KCl). The data points on the bar graph are the averages of at least three biological replicates, and error bars represent the standard deviation of the measurements.

A stress-specific binding mode of Sko1 targets Hog1 to ORFs. (A) In stress, genes bound by a Sko1/Hot1/Hog1 complex (see Table S3) show highest levels of RNA Pol II and Hog1 in ORFs. Scatter plot shows ChIP-seq of Hog1-HA compared with ChIP-seq of Rpb3 after 5 min of osmotic shock with 0.4M KCl. The number of reads that align to each ORF are normalized to reads per kilobases per million (RPKM). Genes that exhibit stress-induced colocalization of Hog1 with Sko1 or with both Sko1 and Hot1 in their regulatory region are shown in red; all other genes are shown in black. (B) Scatter plot shows ChIP-seq of Hog1-HA compared with ChIP-seq of Rpb3 during rapid growth in rich medium.