Localized recruitment of a chromatin-remodeling activity by an activator in vivo drives transcriptional elongation

Abstract

To understand the role of chromatin-remodeling activities in transcription, it is necessary to understand how they interact with transcriptional activators in vivo to regulate the different steps of transcription. Human heat shock factor 1 (HSF1) stimulates both transcriptional initiation and elongation. We replaced mouse HSF1 in fibroblasts with wild-type and mutant human HSF1 constructs and characterized regulation of an endogenous mouse hsp70 gene. A mutation that diminished transcriptional initiation led to twofold reductions in hsp70 mRNA induction and recruitment of a SWI/SNF remodeling complex. In contrast, a mutation that diminished transcriptional elongation abolished induction of full-length mRNA, SWI/SNF recruitment, and chromatin remodeling, but minimally impaired initiation from the hsp70 promoter. Another remodeling factor, SNF2h, is constitutively present at the promoter irrespective of the genotype of HSF1. These data suggest that localized recruitment of SWI/SNF drives a specialized remodeling reaction necessary for the production of full-length hsp70 mRNA.

Figure 1.

Wild-type and mutant human HSF1 are functional when stably expressed in mouse Hsf1^-/- fibroblasts. (A) Schematic of wild-type and mutant human HSF1 constructs showing the DNA-binding domain (DBD), trimerization domain (TD), regulatory domain (RD), and activation domains (AD1 and AD2) of HSF1. (B) Mobility of wild-type and mutant versions of human HSF1. Nuclear extracts from resting and heat-shocked cells were resolved by SDS-PAGE and probed with a polyclonal antibody specific for human HSF1; the mobility shift (HSF1-P) has previously been shown to be caused by phosphorylation. (C) Wild-type and 3A hHSF1 restore thermotolerance to Hsf1^-/- cells; 3F hHSF1 does not. The percentage of viable cells was determined by treating cells with Trypan Blue immediately following exposure to different temperature conditions. Standard errors are indicated (n > 3).

Figure 2.

Expression of wild-type or acidic-mutant hHSF1 restores hsp70 mRNA induction; expression of phenylalanine-mutant hHSF1 does not. (A) S1 analysis of the total level of hsp70 mRNA following heat shock. (B) Quantitation of S1 analysis normalized to signal from the actin probe. Standard errors are indicated (n = 3). (C) Nuclear run-on analysis. RNA was hybridized to four probes across the hsp70 gene and one probe to the mouse skeletal β-actin gene. (D) Quantitation of nuclear run-on data (n = 3). For each cell line, the fold increase in signal following heat shock for each probe is presented, corrected for background signal from the UTR probe as well as from the filters themselves, and is normalized to the number of uridines in each probe.

Figure 3.

The chromatin structure of the hsp70 gene is altered following heat shock induction. (A) Restriction enzyme accessibility. Nuclei were harvested from Hsf1^+/+ cells under normal growth conditions, following treatment with α-amanitin for 1 h, following incubation at 43°C for 1 h, or after treatment with amanitin followed by heat shock. Harvested nuclei were digested with one of the indicated enzymes (cut), and DNA was isolated and digested with a second, reference enzyme (uncut). BanI cleavage produced a set of bands 63–67 bp in size, most likely because of heterogeneity introduced during LMPCR. (B) Quantification from at least two independent preparations of nuclei are expressed as percentage of the signal from the in vivo restriction enzyme site relative to the total signal in each lane.

Figure 4.

Heat-shock-induced chromatin remodeling depends on HSF1 activation domains. Nuclei were harvested from Hsf1^-/-, 3A, and 3F cells under normal growth conditions or following incubation at 43°C for 1 h. Restriction enzyme accessibility was assayed and quantitated as in Figure 3 from three independent preparations of nuclei.

Figure 5.

SNF2h is present at the hsp70 promoter in resting cells, whereas SWI/SNF is recruited to the promoter following heat shock and associates with downstream chromatin in a polymerase-dependent manner. (A) Chromatin immunoprecipitations (ChIPs) were carried out with an anti-BRG1 antibody, an anti-SNF2h antibody, or no antibody and PCR primers that centered on -608, +42, or +2340 relative to the start of transcription (n = 3). All reactions in this figure contain 0.5% of the input material and 2.5% of the no-antibody or antibody samples recovered. (B) BRG1 recruitment in Hsf1^-/- cells is rescued by WT or 3A hHSF1. ChIPs were carried out using an anti-BRG1 antibody, an anti-SNF2h antibody, or no antibody, and isolated DNA samples were amplified using primers that centered on +42 (n = 3). (C) HSF1 is recruited to the hsp70 promoter of all three transgenic cell lines following heat shock. ChIPs were carried out using an anti-BRG1 antibody, an anti-HSF1 antibody, or no antibody, and isolated DNA samples were amplified using primers that centered on +42 (n = 3). (D) BRG1 association with downstream chromatin following heat shock depends on polymerase movement. ChIPs were carried out on resting and heat-shocked Hsf1^-/- (n = 2) and Hsf1^+/+ cells (n = 4), and Hsf1^+/+ cells that had been treated with α-amanitin (n = 3) using an anti-BRG1 antibody, an anti-PolII antibody, or no antibody, and isolated DNA samples were amplified using primers that centered on -608, +42, +612, +1378, and +2340 relative to the start of transcription.