Analysis of phosphorylation of human heat shock factor 1 in cells experiencing a stress

Abstract

Background: Heat shock factor (HSF/HSF1) not only is the transcription factor primarily responsible for the transcriptional response of cells to physical and chemical stress but also coregulates other important signaling pathways. The factor mediates the stress-induced expression of heat shock or stress proteins (HSPs). HSF/HSF1 is inactive in unstressed cells and is activated during stress. Activation is accompanied by hyperphosphorylation of the factor. The regulatory importance of this phosphorylation has remained incompletely understood. Several previous studies on human HSF1 were concerned with phosphorylation on Ser303, Ser307 and Ser363, which phosphorylation appears to be related to factor deactivation subsequent to stress, and one study reported stress-induced phosphorylation of Ser230 contributing to factor activation. However, no previous study attempted to fully describe the phosphorylation status of an HSF/HSF1 in stressed cells and to systematically identify phosphoresidues involved in factor activation. The present study reports such an analysis for human HSF1 in heat-stressed cells.

Results: An alanine scan of all Ser, Thr and Tyr residues of human HSF1 was carried out using a validated transactivation assay, and residues phosphorylated in HSF1 were identified by mass spectrometry and sequencing. HSF1 activated by heat treatment was phosphorylated on Ser121, Ser230, Ser292, Ser303, Ser307, Ser314, Ser319, Ser326, Ser344, Ser363, Ser419, and Ser444. Phosphorylation of Ser326 but none of the other Ser residues was found to contribute significantly to activation of the factor by heat stress. Phosphorylation on Ser326 increased rapidly during heat stress as shown by experiments using a pSer326 phosphopeptide antibody. Heat stress-induced DNA binding and nuclear translocation of a S326A substitution mutant was not impaired in HSF1-negative cells, but the mutant stimulated HSP70 expression several times less well than wild type factor.

Conclusion: Twelve Ser residues but no Thr or Tyr residues were identified that were phosphorylated in heat-activated HSF1. Mutagenesis experiments and functional studies suggested that phosphorylation of HSF1 residue Ser326 plays a critical role in the induction of the factor's transcriptional competence by heat stress. PhosphoSer326 also contributes to activation of HSF1 by chemical stress. To date, no functional role could be ascribed to any of the other newly identified phosphoSer residues.

Figure 1

Validation of the transactivation assay used, transactivation analysis of mutant S326A, and isolation of FLAG-HSF1 for phosphorylation analyses. (A) Transactivation assay of cells in 96-well dishes co-transfected with reporter gene mixture (LEXA-fLUC and pRL-CMV) and 0–40 ng LEXA-HSF1. C: Control unheated cells; HS: Cells subjected to standard heat-treatment for 30 min at 44°C. (B) Relative amounts of HSF1-immunoreactive protein in cells transfected with 0.5 ng β-galactosidase expression construct (B-GAL) or 0.5 ng LEXA-HSF1. Extracts prepared one day after transfection were analyzed by anti-HSF1 western blot. A quantitation of the HSF1 signals, shown in the inserts on top, is presented. Note that the insert only shows the relevant (scanned) portion of the blot, i.e., polypeptide sizes from about 60–100 kD. HSF1 forms appear as several closely spaced bands above a single band representing a nonspecific signal. (C) Coomassie Blue-stained gel containing immune-isolated FLAG-HSF1. M: markers having molecular weights in kD as indicated on the left. Note that the cluster of bands appearing at and below the 98 kD marker was identified as HSF1 forms by a parallel anti-HSF1 western blot (not shown). (D) Transactivation assay comparing activities of HSF1 forms in cells co-transfected with luciferase reporter gene mixture and expression constructs (0.5 ng) for FLAG-LEXA-HSF1 substitution mutant S326A, parent FLAG-LEXA-HSF1 or B-GAL. One day after transfection, cells were exposed either to 300 μM CdCl₂ for 2 h (Cd) or to standard heat treatment (HS), or were left untreated (C). Relative firefly luciferase activities assayed 6 h after treatments are shown. (E) FLAG western blot of parallel (heat-treated) cultures from the experiment analyzed under C that compares expression levels of FLAG-LEXA-HSF1 mutant S326A and parent FLAG-LEXA-HSF1 one day after transfection. The tubulin signal was used as loading control. The numbers below the blots represent a quantitative comparison of the mutant S326A and parent FLAG-LEXA-HSF1 signals in the FLAG blot.

Figure 2

Heat-induced oligomerization and phosphorylation of HSF1. (A) Native anti-HSF1 blot showing heat-induced oligomerization of endogenous HSF1 in response to heat treatment at 44°C (HS) for the times indicated on top of the blot. T: HSF1 trimers; D: heterodimers; M: monomers. (B) Parallel cultures to those used in A for an analysis of HSF1 oligomerization were employed here for an examination of global or Ser³²⁶-specific phosphorylation of endogenous HSF1. The anti-HSF1 western blot shown reports the distribution of HSF1 forms of different apparent size (reflecting different levels of phosphorylation) in extract samples (lysate) or in protein immunoprecipitated from the same extracts by pSer³²⁶ antibody (ip). The portion of the blot depicted only shows protein signals larger than about 70 kD. (C) Detection of heat-induced phosphorylation of Ser³²⁶ by western blot using anti-pSer³²⁶ antibody. Parallel cultures were transfected with small amounts of FLAG-HSF1. One day later, the cultures were either left untreated (C) or were heat-treated for 30 min at 44°C (HS) and were processed for western blot immediately following the heat treatment. The anti-pSer³²⁶ blot reports on induction of Ser³²⁶ phosphorylation by heat, and the parallel anti-FLAG blot shows that similar amounts of FLAG-HSF1 were compared in the anti-pSer³²⁶ blot. Data from densitometry are shown below the blots. Brackets shown on the side of blots indicate lengths of regions scanned. In A, the monomer signal was quantitated.

Figure 3

Comparative analyses in HSF1-negative mouse embryo fibroblasts. (A) Comparison of DNA-binding abilities of HSF1 and HSF1 mutant S326A. An electrophoretic mobility shift assay was carried out using an HSE DNA probe and extracts from heat-treated or not-heat-treated cells that had been transfected one day earlier with the constructs indicated on top of the gel. An anti-HSF1 western blot of the same samples reporting on the relative levels of expression of HSF1 forms from the different constructs is shown below the gel. HSF1/HSE: HSF1-DNA complex; NS: nonspecific signal, serving as a loading control for the group of samples from not-heated cells, and, independently, for the group of samples from heat-treated cells; HS: heat-treated for 30 min at 43°C; C: not heat-treated. (B) Heat-induced transactivation of endogenous hsp70 gene(s) in cells transfected with small amounts of the constructs indicated above the blots and with luciferase reporters HSP70-fLUC and pRL-CMV. One day after transfection, cultures were either left untreated (C) or were heat-treated for 30 min at 43°C (HS). Extracts were prepared after 6 h of further incubation at 37°C and were used for western blots. The top blot was probed with an antibody recognizing HSP70, whereas the bottom blot was probed with HSF1 antibody. The bottom blot demonstrates that mutant and wildtype HSF1 forms accumulated to similar levels. Note the presence of a weak nonspecific signal present in all lanes that happened to co-migrate with the transfected HSF1 forms. (C) Quantitative comparisons of data of B. (D) Relative luciferase reporter activities in the same extracts that were analyzed for HSP70 in B. Representative results from one of several independent experiments are shown in this Figure.