RSK2 represses HSF1 activation during heat shock

Abstract

Heat shock transcription factor 1(HSF1) activation is a multistep process. The conversion of a latent cytoplasmic form to a nuclear, DNA binding state appears to be activated by nonsteroidal anti-inflammatory drugs. In previous studies, we showed that HSF 1 is phosphorylated by the protein kinase RSK2 in vitro and that this effect is inhibited by nonsteroidal anti-inflammatory drugs at the concentration that leads to the activation of HSF1 in vivo (Stevenson et al 1999). In the present study, using cells from a patient with Coffin-Lowry syndrome (deficient in RSK2), we demonstrate that RSK2 slightly represses activation of HSF1 in vivo at 37 degrees C. In Coffin-Lowry syndrome cells, HSF1-HSE DNA binding activity after treatment with sodium salicylate was slightly higher than that in untreated cells, indicating that although RSK2 is involved in HSF1 regulation, it is not the unique protein kinase that suppresses HSF1-HSE binding activity at 37 degrees C. However, heat shock treatment resulted in significantly higher HSF1-HSE binding activity in Coffin-Lowry syndrome cells as compared with normal controls, suggesting that RSK2 represses HSF1-HSE binding activity during heat shock.

Fig 1.

Level of RSK2 in normal (RSK2+/+) and CLS (RSK2−/−) cells. Whole-cell lysates from cells were analyzed by Western blot by using anti-RSK2 antibody (Santa Cruz Biotechnology, Santa Cruz, CA). RSK2 is not detectable in CLS (RSK2−/−) cells (right lane). Anti-RSK2 antibody is directed against the C terminus of the protein

Fig 2.

Transcriptional activation of HSP70B promoter by HSF1 in normal (RSK2+/+) and CLS (RSK2−/−) cells at 37°C. Wild-type HSF1 expression plasmid (pcDNA3.1HSF1; 0.2 μg/well) or empty control (pcDNA3.1; 0.2 μg/well) were cotransfected with HSP70B promoter reporter construct (HSP70B-Luc) (0.6 μg/well) along with pSV-β-galactosidase plasmids (0.6 μg/well), respectively, transfected into normal (RSK2+/+) and CLS (RSK2−/−) cells in triplicate and incubated for 20 to 24 h. Wild-type HSF-1 were force-cloned into XhoI/EcoRI sites of the pcDNA3.1 vector as described (Cahill et al 1996). HSP70B-Luc reporter plasmid was constructed as described before (Chen et al 1997). Cells were harvested for luciferase assay. Luciferase activity was normalized to β-galactosidase activity. Luciferase activity is expressed as fold activity over the empty pcDNA3.1 plasmid control

Fig 3.

Effect of sodium salicylate on HSF1-HSE binding activity in nuclear extracts from normal (RSK2+/+) and CLS (RSK2−/−) cells at 37°C. Parallel cultures of normal (RSK2+/+) and CLS (RSK2−/−) cells were pretreated for 30 minutes in presence or absence of sodium salicylate (20 mM) (lanes 5, 6, 9, and 10). Nuclear extracts (4 μg) were incubated with ³²P-labeled HSE probe and subjected to EMSA. Gel supershift assays were carried out as described before (Cahill et al 1996) (lanes 2, 4, 6, 8, and 10). Heat-shocked Hela cells (lanes 1 and 2) were positive controls. The position of HSF1-HSE complex is indicated by an arrow

Fig 4.

HSF1-HSE binding activity in nuclear extracts from normal (RSK2+/+) cells (lanes 3 and 4), CLS (RSK2−/−) cells (lanes 5 and 6), and Hela (lanes 1 and 2) cells after heat shock (43°C, 1 hour). Nuclear extracts (2 μg) were incubated with ³²P-labeled HSE and subjected to EMSA. Gel supershift assays were carried out as described before (Cahill et al 1996). Specific HSF1-HSE complex is indicated by an arrow