Heat Shock Protein 25 or Inducible Heat Shock Protein 70 Activates Heat Shock Factor 1

The expression of heat shock proteins (HSPs) is known to be increased via activation of heat shock factor 1 (HSF1), and excess expression of HSPs exerts feedback inhibition of HSF1. However, the molecular mechanism to modulate such relationships between HSPs and HSF1 is not clear. In the present study, we show that stable transfection of either Hsp25 or inducible Hsp70 (Hsp70i) increased expression of endogenous HSPs such as HSP25 and HSP70i through HSF1 activation. However, these phenomena were abolished when the dominant negative Hsf1 mutant was transfected to HSP25 or HSP70i overexpressed cells. Moreover, the increased HSF1 activity by either HSP25 or HSP70i was found to result from dephosphorylation of HSF1 on serine 307 that increased the stability of HSF1. Either HSP25 or HSP70i inhibited ERK1/2 phosphorylation because of increased MKP1 phosphorylation by direct interaction of these HSPs with MKP1. Treatment of HOS and NCI-H358 cells, which showed high expressions of endogenous HSF1, with small interfering RNA (siRNA) of either HSP27 (siHSP27)or HSP70i (siHSP70i) inhibited both HSP27 and HSP70i proteins; this was because of increased ERK1/2 phosphorylation and serine phosphorylation of HSF1. The results, therefore, suggested that when the HSF1 protein level was high in cancer cells, excess expression of HSP27 or HSP70i strongly facilitates the expression of HSP proteins through HSF1 activation, resulting in severe radio- or chemoresistance.

The expressions of small HSPs, especially HSP25 (HSP27), 3 and the inducible HSP70 (HSP70i) have been shown to enhance the survival of mammalian cells exposed to various types of stimuli that induce stress and apoptosis (1,2). Considering the key role of HSPs played in protec-tion against stress-induced damage, it is of great importance to elucidate the regulatory mechanisms responsible for HSP expression.
The inducible HSP expression is regulated by the heat shock transcription factors (HSFs). In response to various inducers such as elevated temperatures, oxidants, heavy metals, and bacterial and viral infections, most HSFs acquire DNA binding activity to the heat shock element (HSE), thereby mediating transcription of the heat shock protein genes, which results in accumulation of HSPs (3,4). Although several genes encoding multiple HSF isoforms have been identified in vertebrates (5,6), only HSF1 is shown to be essential for the transcriptional activation of HSPs in mammalian organisms (4,7). Under normal conditions, cellular HSF1 exists in a transcriptionally repressed state (8,9), however, it is activated by stress in a multistep process involving trimerization, acquisition of HSE binding activity, novel phosphorylation, and trans-activation of HSP genes (3,10,11).
In recent years, a conceptual framework for stress-induced activation and feedback repression of HSF1 has emerged (3,12,13). However, our recent study indicated that, in RIF cells that have relatively low HSP25 and inducible HSP70 (HSP70i) expressions, HSP25 or HSP70i mutually coregulated each other and increased their own protein expression through HSF1 activation (14).
Recently, it was shown that ERK1/2 activation by mitogenic stimulation leads to phosphorylation of serines 303 and 307 of HSF1 and consequent association of HSF1 with 14-3-3. 14-3-3 binding inhibits both the transcriptional activity and nuclear accumulation of HSF1 (15). HSF1 seems to exert a reciprocal influence on cell proliferation in that HSP70, a major product of HSF1 transcriptional activity, inhibits the ERK pathway. It is intriguing to note that ERK1/2 is known to be a repressor of HSP synthesis (16), although HSP induction and ERK1/2 activation result in cytoprotection. HSF1 is subjected to additional layers of regulation, including up-regulation through binding the apoptosis modulator DAXX and modulation by ERK1/2 (17,18). Hierarchical phosphorylation of HSF1 within its transcriptional regulatory domain by ERK1 (on serine 307) and glycogen synthase kinase-3 (GSK3) (on serine 303) was also reported (19,20).
In this study, we attempted to further elucidate molecular mechanisms governing the relationship between HSPs and HSF1, and found that excess expression of HSP25/27 or HSP70i increased endogenous HSPs such as HSP25/27 or HSP70i through HSF1 activation, which resulted from its dephosphorylation of serine 307. Dephosphorylation of HSF1 on serine 307 was mediated by inhibition of ERK1/2 phosphorylation by direct interaction of MKP1 with HSP25 or HSP70i. Interac-tion between MKP1 and HSP25 or HSP70i increased MKP1 phosphorylation, which resulted in ERK1/2 dephosphorylation.
Plasmids-Wild type mouse Hsp25 and inducible Hsp70 were cloned into pcDNA3. The mutant construct, Mkp1 (C258S), was constructed by PCR by using the overlap extension primer. Cysteine was replaced by serine to generate a phosphatase activity-deficient mutant. The mutants Hsf1-S307A and Hsf1-S307D were constructed using the method as described above.
Cell Transfection-Pre-designed siRNA for human HSP25 and inducible HSP70 (Ambion catalog numbers 19706, 51187, and 16706) and negative control siRNA were purchased from Ambion, Inc. (Austin, TX). The cells were transfected with the siRNAs for 48 h with the use of Lipofectamin TM 2000 (Invitrogen). Transient transfection of all cell types was carried out using Plus Lipofectamine TM reagent (Invitrogen) and Lipofectamine TM reagent.
Immunofluorescence Analysis-For immunofluorescence analysis, cells were fixed with 4% paraformaldehyde, permeabilized with 0.1% Triton X-100 in PBS, and then washed three times with PBS. Cells were then incubated with anti-HSP25/HSP70 and anti-HA diluted 1:200 in PBS with 5% fetal bovine serum for 1 h at room temperature in a humidified chamber. Excess antibody was removed by washing coverslips three times with PBS. Cells were then incubated with fluorescein isothiocyanate-conjugated secondary antibody (Dako, Produktionsvej, Denmark) at 1:200 dilution in PBS with 5% fetal bovine serum for 4 h, and then incubated with 0.5 g/ml propidium iodide (Molecular Probes, Inc., Eugene, OR) for 5 min at room temperature. After washing three times with PBS, coverslips were mounted onto microscope slides using ProLong antifade mounting reagent (Molecular Probes). The slides were examined by a confocal laser scanning microscope (Leica Microsystems). A, representative Western blot analysis showing the relative levels of HSP25 and HSP70i in clones derived from RIF cells transfected with HSP25 or inducible HSP70 and TR cells, which are thermoresistance clones of RIF cells. B, activity of HSP25 or HSP70i promoter driving expression of the luciferase reporter gene in RIF cells, which were transfected with HSP25 or inducible HSP70 and grown in phenol red-less medium containing stripped serum. Luciferase activity was normalized to ␤-galactosidase transfection, which was included in all cotransfections to control for variation in transfection efficiency. Percentage of the empty pGL3 vector control is set at 100%. Mean Ϯ S.D. of at least three independent experiments. Polyacrylamide Gel Electrophoresis and Western Blot-For PAGE and Western blots, cells were solubilized with lysis buffer (120 mM NaCl, 40 mM Tris (pH 8.0), 0.1% Nonidet P-40), the samples were boiled for 5 min, and an equal amount of protein (40 g/well) was analyzed on 10% SDS-PAGE. After electrophoresis, proteins were transferred onto a nitrocellulose membrane and processed for immunoblotting. Blots were further incubated with horseradish peroxidase-conjugated secondary antibody diluted at 1:5,000, and specific bands were visualized by chemiluminescence (ECL, Amersham Biosciences). Autoradiographs were recorded onto X-Omat AR films (Eastman Kodak Co.).
Reporter Assay-Cells were plated at about 8 ϫ 10 4 cells per 35-mm dish and cultured for 24 h before transfection. After incubation, cells were transfected with 200 ng of ␤-galactosidase expression vector pSV110 and 200 ng of reporter gene, along with HSP25, inducible HSP70, and AP-1 expression vector. Cells were harvested 48 h later, luciferase activity was assayed, and results were normalized to the ␤-galactosidase expression (Promega). All the results represent the mean of three independent experiments.
Electrophoretic Mobility Shift Assay-Two complementary singlestrand DNA oligonucleotide, each 36 bases long and containing the HSE oligonucleotide, were annealed and used in the electrophoretic mobility shift DNA binding assay. The DAN sequences were 5Ј-GATCCTC-GAAGGTTCGAGGATCCTCGAAGGTTCGAG-3Ј and 3Ј-GAGCT-TCCAAGCTCCTAGGAGCTTCCAAGCTCCTAG-5Ј. The HSE was labeled with [␥-32 P]ATP using T4 polynucleotide kinase, and 5 g of nuclear extract was incubated with the [ 32 P]HSE at room temperature for 30 min. Sequentially the HSF1-HSE complex was separated from unbound HSE on 6% Tris-borate-EDTA gel. The gel was dried, and the autoradiogram was taken.

RESULTS
Overexpression of HSP25 and HSP70i Facilitated Expressions of Endogenous HSPs Such as HSP25 and HSP70i in RIF Cells-When pcDNA3 vectors carrying Hsp25 or Hsp70i were transfected to RIF cells and stable cell lines were obtained, either HSP25 or HSP70i overexpressed cells were found to facilitate mutual expressions of the corresponding HSPs as well as their own protein expressions comparable with that of TR cells that are thermoresistant RIF cells, and showed high expressions of HSP25 and HSP70i (14) (Fig. 1A). We examined 2 or 3 clones and similar phenomena were obtained (supplemental Fig. 1). The reporter assay for Hsp25 and Hsp70i also indicated that the promoter activity for both Hsp25 and Hsp70i were significantly activated by HSP25 or HSP70i overexpression (Fig. 1B), suggesting that excess expressions of HSP25 and HSP70i increased promoter activity of both Hsp25 and Hsp70i.
HSF1 Activation Was Observed in HSP25 or HSP70i Overexpressed Cells-Because HSF1 is known to bind to HSE to promote HSP transcription (3), nuclear fractions were incubated with 32 P-labeled HSE, and the HSF1-HSE complex was subjected to polyacrylamide gel electrophoresis. As seen in Fig. 2A, the amount of HSF1-HSE complex was found to have significantly increased in both Hsp25-and Hsp70i-transfected cells. TR cells showed strong activation of HSF1, and stable transfection of Hsp25 or Hsp70i also showed increased HSF1 activation. Moreover, binding of HSF1-HSP70i in the cytosol of RIF cells was abundant, however, no binding activity was shown in NIH3T3 cells (supplemental Fig. 2). Reverse transcriptase-PCR analysis for HSF1 and HSF2 revealed that the amounts of these genes were not significantly different by HSP25 or HSP70i overexpression (Fig. 2B). However, a greater amount of HSF1 and HSF2 proteins was found in HSP25 or HSP70i overexpressed cells, suggesting post-translational regulation of HSF1 by HSP25 or HSP70i (Fig. 2C). Determination of the half-life of HSF1 also indicated that HSP25 or HSP70i overexpression increased HSF1 stability (Fig. 2D).
Mutant Forms of Hsf1 Abolished the Enhanced Expression of HSPs by HSP25 or HSP70i-To examine whether HSF1 activation by ectopic expression of HSP25 or HSP70i was responsible for the increased expression of endogenous HSPs, C and N terminus deletion mutants of Hsf1 were transiently transfected to the stably transfected cells of Hsp25 or Hsp70i. Fig. 3A shows that increased promoter activity of both Hsp25 and Hsp70i was dramatically inhibited by additional transfection of deletion mutants of Hsf1. Western blot analysis also indicated that transient transfection of C terminus and N terminus deletion mutants of Hsf1 to either HSP25 or HSP70i overexpressed cells did not induce any endogenous HSP25 and HSP70i at all. As shown in Fig. 3B, overexpression of HSP25 and HSP70i facilitated endogenous expression of HSPs through HSF1 activation. To elucidate whether ectopic expression of HSP25 or HSP70i can affect endogenous expression of these proteins, transient transfection of HA-tagged Hsp25 was performed, and HSP25 protein expression was studied at 24, 48, and 72 h after transfection. As shown in Fig. 3C, ectopic expression of HA-tagged HSP25 was decreased at 48 h of transfection and could not be detected at 72 h, however, endogenous Hsp25 expression was increased from 48 h of transfection, which coincided with the decrease of extrinsic HSP25. Therefore, the results indicate that HSP25 or HSP70i affect expression of the corresponding endogenous HSPs through HSF1 activation.
HSP25 and HSP70i Inhibited ERK1/2 Phosphorylation-Because ERK1/2 activity is involved in HSF1 phosphorylation and activation   (Fig. 4A). When the reporter assay for Hsp25 or Hsp70i was performed, transient transfection of Erk1 inhibited promoter activities of Hsp25 or Hsp70i in HSP25-or HSP70i-overexpressed cells. However, inhibited transcriptional activity of AP-1 of one ERK1/2 downstream pathway in overexpressed cells of Hsp25 or HSP70i was restored or more augmented by Erk1 transfection (Fig. 4B). Therefore, the data strongly indicate that HSP25 or HSP70i overexpression induced dephosphorylation of ERK1/2, which might affect HSF1 phosphorylation.
Serine Phosphorylation of HSF1 Was Reduced by HSP25 or HSP70i Transfection-Because ERK1/2 activity affects serine phosphorylation of HSF1 (21), immunoprecipitation of HSF1 and immunoblotting of phosphoserine were performed. Serine phosphorylation of HSF1 disappeared in Hsp25-or Hsp70i-transfected RIF cells (Fig.  5A). Because ERK1/2 activation and consequent phosphorylation of serine 307 of HSF1 lead to inactivation of HSF1 (19), wild type or phosphorylation-deficient mutants of serine 307 (S307A) of Hsf1 were transiently transfected to stably overexpressed RIF cells of HSP25 or HSP70i. Expressions of HSP25 and HSP70i also disappeared in the transfectant of the S307D mutant, whereas transfectant of wild type Hsf1 and S307A expressed these proteins. Therefore, the results indicate that phosphorylation of HSF1 on serine 307 inactivated HSF1, and that HSP25 and HSP70i dephosphorylated HSF1 on serine 307 (Fig. 5B). To clarify whether down-regulation of ERK1/2 by HSP25 or HS70i is related to HSF1 phosphorylation, Erk1 was transiently transfected to HSP25 or HSP70i overexpressed cells. Fig.  5C shows that HSF1 phosphorylation was restored and expression of HSP25 or HSP70i was reduced by Erk1 transfection. In addition, S307A had a longer half-life than that of wild type Hsf1 and S307D when detected using HA-tagged mutants (Fig. 5D). Confocal microscopy also revealed that nuclear expression of HSF1 was more abundant in HSP25 or HSP70i overexpressed cells than that of RIF control cells. Moreover, HA-tagged Hsf1-S307A mutant as well as wild type Hsf1 were located in the nucleus, however, the majority of Hsf1-S307D mutant was not present in nucleus; Hsf1-S307D was present in the nuclear membrane and did not enter the nucleus (Fig. 5E).
Therefore, these results indicate that HSP25 or HSP70i inhibited the nuclear translocation of HSF1 via decreased phosphorylation on serine 307, which in turn led to transcriptional activation of HSF1.
Interaction between HSP25 or HSP70i and MKP1 Dephosphorylated ERK1/2-Because heat shock increased expressions of the phosphorylated form of MKP1, which was found to be associated with ERK1/2 and JNK1/2 inactivation (22), we determined MKP1 phosphorylation in HSP25-or HSP70i-overexpressed cells. As seen in Fig. 6A, serine-phosphorylated MKP1 expression was increased in HSP25 or HSP70i overexpressed cells, and interaction between HSP25 or HSP70i and MKP1 was also increased. Transient transfection of site-directed mutants of Mkp1 on cysteine 258 to serine (Mkp1-C258S), which abolished phosphatase activity of MKP1 (23), revealed that expressions of HSP25, HSP70i, and HSF1 were reduced by Mkp1-C258S transfection, and phosphorylations of both HSF1 and ERK1/2 by HSP25 or HSP70i were restored by this mutant (Fig.  6B), even though mutation of Mkp1 on cysteine 258 did not affect the binding activity with HSP25 or HSP70i (data not shown). The data indicate that the site of cysteine 258 on MKP1 is important for phosphorylation of ERK1/2.
siRNA of HSP27 or HSP70i Treatment Inhibited HSF1 Expression-To unravel the physiological relevance of regulation existing between HSP27 or HSP70i and HSF1, HOS and NCI-H358 cells, which showed increased expressions of HSP27, HSP70i, as well as HSF1 (Fig. 7), were transfected with siRNA of HSP27 (siHSP27) or HSP70i (siHSP70i). Endogenous expression of HSPs as well as HSF1 expression was found to be dramatically inhibited in siHSP27 or siHSP70i-transfected HOS and NCI-H358 cells (Fig. 7). In addition, inhibited ERK1/2 phosphorylation was restored by these siRNAs, suggesting that HSP27 or HSP70i facilitated HSF1 activation by ERK1/2 dephosphorylation.

DISCUSSION
HSPs are known as the molecular chaperone because of its major role in protein folding (24 -27). Because HSF1 induces Hsp70i and Hsp25 gene expression in response to heat stress, it has been suggested that HSPs may function in a negative feedback mechanism to return HSF1 to its inactive monomeric state (12,13). However, in the present study, ectopic transfection of HSP25 or HSP70i in RIF cells increased the expression of endogenous HSPs, such as HSP25 and HSP70i, through activation of HSF1. NIH3T3 cells or L929 cells, which showed relatively low HSF1 expression did not exhibit such phenomena (supplemental Fig. 3). Therefore, when HSF1 expression was high, such as in RIF, HOS, and NCI-H358 cells, overexpression of HSP25 or HSP70i facilitated more expression of endogenous HSPs, which may result in more severe radio-or chemoresistance.
Overexpression of HSP25 or HSP70i in RIF cells induced transcriptional activation of endogenous HSP25 and HSP70i, which was mediated by HSF1 activation (Figs. 1-3). Moreover, treatment of HOS and NCI-H358 cells, which showed high expression of HSF1 protein, with Hsp27 or Hsp70i siRNA inhibited endogenous expression of HSPs and HSF1 (Fig. 7), suggesting that HSP25 or HSP70i overexpression activated HSF1. A recent study, which derived a similar conclusion to ours, also suggested that cognate HSP70 was required for regulation of HSF1 during heat stress and subsequent target gene expression, and also showed greatly decreased HSF1 activation when transfected with siR-NAs targeted to cognate Hsp70 (28).
Upon activation, HSF1 undergoes several modifications such as trimerization and localization to specific nuclear structures. Independent of monomer to trimer transition, phosphorylation of HSF1 is strongly induced during heat shock. Although the bulk phophorylation level of HSF1 correlates well with its transcriptional activity, phosphorylation of Ser-303, Ser-307, and Ser-363 has been shown to repress the transactivation capacity of HSF1 (19,20,29) and phosphorylation of Ser-307 is affected by ERK1/2 activation (21). The result in the present study showed that HSP25 or HSP70i overexpression dephosphorylated HSF1 on serine 307 of HSF1 (Fig. 4). The phosphorylation mimicked mutant form of serine 307 inhibited expression of endogenous HSPs in HSP25 or HSP70i overexpressed cells, suggesting that HSP25 or HSP70i affected the phosphorylation status of HSF1, especially on serine 307. Moreover, the increased half-life and nuclear translocation of the dephophorylated mimicked mutant of serine 307 (Fig. 4) indicated that dephosphorylated HSF1 on serine 307 was the transcriptionally active form of HSF1. ERK1/2 phosphorylation was inhibited in Hsp25-or Hsp70i-transfected cells (Fig. 5) and ERK1/2 phosphorylation has been reported to be involved in HSF1 phosphorylation on serine 307 (16,21). ERK1/2 was found to be overexpressed in Hsp25-or Hsp70i-transfected cells, and transcriptional activation of endogenous HSP25 or HSP70i was abolished, suggesting that HSP25 or HSP70i-mediated ERK1/2 down-regulation may be involved in dephosphorylation of HSF1 on serine 307. Phosphatase involved in ERK1/2 signaling is MKP1.
MKP1 is a phosphatase with dual specificity, induced by various stresses, and inactivates ERKs (30). ERK1/2 phosphorylates serine 307 of HSF1 (31). HSP70 has been shown to suppress ERK phosphorylation by direct interaction with MKP1 (32). In the present study, ectopic expression of HSP25 or HSP70i increased MKP1 phosphorylation on the serine residue, resulting in increased binding activity with MKP1 and ERK1/2 dephosphorylation. Our results (Fig. 6) that HSP25 also bound MKP1 is a new finding, suggesting that HSP25 also has a potentiality of MKP1 activation. In addition, the phosphatase activity defective mutant of Mkp1 (MKP1-C258S) restored ERK1/2 phosphorylation and HSF1 expression, suggesting that interaction of MKP1 with HSP25 or HSP70i activated MKP1 phosphorylation, which was correlated with ERK1/2 dephosphorylation. Based on the data obtained, we suggest that overexpression of HSP25 or HSP70i activates HSF1, which is mediated by interaction with MKP1. Interaction of MKP1 with HSP25 or HSP70i activates MKP1 phosphorylation and then ERK1/2 inactivation, resulting in HSF1 dephosphorylation on serine 307. Because HSF1 dephosphorylation facilitated translocation of HSF1 to the nucleus (Fig. 4) (3,12,33), it is highly likely that this activated HSF1 enhanced more expression of endogenous HSPs such as HSP25 and HSP70i.
In conclusion, the present data suggest the possibility that overexpression of HSPs does not always show feedback inhibition of HSF1. In certain cancer cells that have HSF1 excessively expressed, overexpression of HSPs facilitated more endogenous HSPs production through HSF1 activation.