HSF1-TPR Interaction Facilitates Export of Stress-induced HSP70 mRNA*♦

Stress conditions inhibit mRNA export, but mRNAs encoding heat shock proteins continue to be efficiently exported from the nucleus during stress. How HSP mRNAs bypass this stress-associated export inhibition was not known. Here, we show that HSF1, the transcription factor that binds HSP promoters after stress to induce their transcription, interacts with the nuclear pore-associating TPR protein in a stress-responsive manner. TPR is brought into proximity of the HSP70 promoter after stress and preferentially associates with mRNAs transcribed from this promoter. Disruption of the HSF1-TPR interaction inhibits the export of mRNAs expressed from the HSP70 promoter, both endogenous HSP70 mRNA and a luciferase reporter mRNA. These results suggest that HSP mRNA export escapes stress inhibition via HSF1-mediated recruitment of the nuclear pore-associating protein TPR to HSP genes, thereby functionally connecting the first and last nuclear steps of the gene expression pathway, transcription and mRNA export.

The up-regulation of heat shock proteins such as HSP70 that occurs in response to exposure to elevated temperature and many other stress conditions is vital for the ability of cells to survive these stresses. Because of their crucial cytoprotective function, it is very important that up-regulation of HSP expression after stress occur as rapidly and as efficiently as possible. An intriguing finding of past studies is that stress conditions inhibit the export of many mRNAs from the nucleus, but mRNAs encoding heat shock proteins continue to be efficiently exported during stress (1)(2)(3)(4)(5)(6)(7). However, the mechanism by which HSP mRNAs bypass this stress-associated export inhibition was not known. HSF1 is the transcription factor responsible for up-regulating transcription of HSP70 and other genes in response to elevated temperature and other stress conditions. HSF1 performs this function by undergoing stress-induced trimerization to the DNA-binding form and then interacting with heat shock elements in the promoters of these genes to increase their transcription (8,9). TPR is a 270-kDa polypeptide that is associated with the nuclear basket on the nucleoplasmic face of the nuclear pore complex (10 -17). Previous data suggest that TPR is involved in the export of both mRNAs and proteins from the nucleus (13, 16, 18 -20). During the course of yeast two-hybrid experiments in our laboratory, we identified the existence of an interaction between HSF1 and the TPR protein. The results presented here suggest that the HSF1-TPR interaction could be important for the export of HSP mRNAs during stress conditions.
In Vitro Binding Assay-In vitro translated 35 S-labeled TPR- (14 -117) or TPR-(1218 -1320) was incubated with GST 5 -HSF1 and GST bound to glutathione-agarose. After washing, bound proteins were analyzed by SDS-PAGE and autoradiography to detect the 35 S-labeled TPR proteins. The amounts of GST-HSF1 and GST proteins bound to the beads were determined by SDS-PAGE followed by Western blotting using goat anti-GST polyclonal antibody (Amersham Biosciences).
Immunoprecipitation Analysis-HeLa cells (American Type Culture Collection) were heat-shocked at 42°C for 1 h, and extracts of these cells were then subjected to immunoprecipitation using rabbit anti-HSF1 polyclonal antibody (21) or rabbit nonspecific IgG followed by Western blotting using mouse anti-TPR monoclonal antibody (Oncogene Research Products).
Chromatin Immunoprecipitation Assay-ChIP assays were performed with Jurkat cells as described previously (22) using mouse anti-TPR monoclonal antibody and mouse nonspecific IgG (as a negative control). The primers used to amplify the promoter regions of the stress-inducible HSP70i gene and histone H4 gene were as follows: HSP70i, 5Ј-ctcagggtccctgtccc-3Ј and 5Ј-tgagccaatcaccgagc-3Ј; and histone H4, 5Ј-gagagggcggggacaattga-3Ј and 5Ј-ttggcgtgctcggtgtaggt-3Ј. PCR products were then separated on polyacrylamide gels and stained with ethidium bromide. ChIP samples were also analyzed by quantitative real-time PCR as described below.
HSP mRNA Export Analysis-pEGFP-TPR-(14 -117) and pEGFP-TPR-(1218 -1320) were generated using a PCR-based strategy to amplify the relevant regions while adding restriction sites to each end to allow ligation into the pEGFP-C2 vector. Cloning junctions were checked by sequencing. One g of each plasmid or empty vector was cotransfected with either the HSP70i-luciferase (luciferase expressed from the stress-inducible human HSP70i promoter) (21) or RSV-luciferase plasmid using Effectene (Qiagen Inc.) following the manufacturer's instructions. The RSV-luciferase plasmid was a kind gift of Dr. Dan Noonan and was generated by replacing the chloramphenicol acetyltransferase coding sequence from the RSV-chloramphenicol acetyltransferase plasmid (23) with the luciferase coding sequence. Cells were heat-shocked at 42°C for 1 h, and cytoplasmic and nuclear fractions were then prepared by hypotonic lysis as follows. Cells were swollen in 5 packed cell volumes of 10 mM HEPES (pH 7.9), 1.5 mM MgCl 2 , 10 mM KCl, 1 mM dithiothreitol, and 1 mM phenylmethylsulfonyl fluoride (Buffer A) for 10 min on ice. Cells were then centrifuged at 2000 rpm for 10 min, resuspended in 2 packed cell volumes of Buffer A, and lysed by 20 strokes of a Dounce homogenizer (type B pestle). The nuclei and cytoplasm were separated by centrifugation at 2000 rpm for 10 min. Separation was verified by viewing fractions with a microscope. mRNA was extracted from each fraction using TRIzol reagent (Invitrogen) following the manufacturer's instructions. To analyze mRNA concentrations, each pool was subjected to an RNase protection assay using SuperSignal RPA III (Ambion, Inc.) following the manufacturer's instructions with a probe for either luciferase or L32 ribosomal protein mRNA. The probe for luciferase mRNA was constructed via in vitro transcription using MAXIscript (Ambion, Inc.) and biotinylated UTP (Roche Applied Science). The template for in vitro transcription was created by PCR with the HSP70-luciferase plasmid and primers 5Ј-cacggaaagacgatgacg-3Ј and 5Ј-taatacgactcactataggttgggtaacgccaggg-3Ј. This PCR product contained the 3Ј-end of the luciferase mRNA, ending with the polyadenylation signal (yielding a protected fragment of 325 bp) and untranscribed vector sequence (resulting in an unprotected fragment of 438 bb). The construction of the L32 probe was described previously (24).
To determine whether expression of the two HSF1-binding regions of TPR affects endogenous HSP70 mRNA levels, GFP-TPR-(14 -117) and GFP-TPR-(1218 -1320) or empty vector were transfected into HeLa cells according to the manufacturer's protocol. Cells were heat-shocked and fractionated as described above. Total RNA was extracted with TRIzol according to the manufacturer's instruction. RNA was resuspended in 10 mM Tris-HCl (pH 7.5), 2.5 mM MgCl 2 , and 0.5 mM CaCl 2 and incubated with RNase-free DNase I to remove possible genomic contamination. cDNA was prepared from samples using ImProm-II reverse transcriptase (Promega Corp.) and oligo(dT) 16 primers. cDNA samples were assayed by quantitative real-time PCR as described below.
RNA Immunoprecipitation Analysis-HeLa cells (1.5 ϫ 10 6 ) were transfected with 2 g of either the HSP70i-luciferase or RSV-luciferase plasmid using Effectene following the manufacturer's instructions. Cells were then heat-shocked for 1 h at 42°C, washed once with ice-cold phosphate-buffered saline, and cross-linked with 2% paraformaldehyde for 12 min while rotating. Cross-linking was quenched with 125 mM glycine for 5 min; and cells were washed twice with ice-cold phosphate-buffered saline, harvested by scraping, and snap-frozen in liquid nitrogen. Cells were resuspended in 2 ml of low stringency radioimmune precipitation assay (RIPA) buffer (50 mM Tris-HCl (pH 7.5), 1% Nonidet P-40, 0.5% sodium deoxycholate, 0.05% SDS, 1 mM EDTA, 150 mM NaCl, 1ϫ Complete protease inhibitor mixture (Roche Applied Science), and 80 units of RNaseOUT (Invitrogen)), pipetted 20 times, and incubated on ice for 10 min. Cells were then sonicated three times (80 -90% output) for 20 s and centrifuged at 16,000 ϫ g for 10 min at 4°C. The supernatant was precleared with 20 l of protein G-Sepharose (GE Healthcare) and washed with low stringency RIPA buffer and 100 g/ml yeast tRNA (Ambion, Inc.) for 2 h at 4°C. During the preclearing, the low stringency RIPA buffer-washed protein G-Sepharose beads were coated with 5 g of either anti-TPR antibody or mouse IgG (Sigma) in low stringency RIPA buffer. The precleared supernatant was then incubated with the antibody-coated beads for 90 min at 4°C with rotation. Complexes were washed five times for 10 min each at room temperature with 1 ml of high stringency RIPA buffer (50 mM Tris-HCl (pH 7.5), 1% Nonidet P-40, 1% sodium deoxycholate, 0.1% SDS, 1 mM EDTA, 1 M NaCl, 1 M urea, and 0.2 M phenylmethylsulfonyl fluoride). Beads were then resuspended in 100 l of 50 mM Tris-HCl (pH 7.5), 5 mM EDTA, 10 mM dithiothreitol, and 1% SDS, and cross-links were reversed for 1 h at 70°C. mRNA was isolated as described above, and samples were analyzed by quantitative PCR as described below.
Quantitative Real-time PCR-This was performed with the Mx4000 system (Stratagene) using Brilliant SYBR Green QPCR Master Mix (Stratagene). Samples were checked for specific amplification using dissociation curve analysis included with the software. PCR products were also assayed on polyacrylamide gels with ethidium bromide staining to ensure that they were of the expected size. For quantitative PCR analysis of the TPR-HSP70 promoter ChIP assay, the primers used were as follows: HSP70, 5Ј-caacacccttcccaccgccactc-3Ј and 5Ј-ctgattggtccaaggaaggctgg-3Ј; and histone H4, 5Ј-gagagggcggggacaattga-3Ј and 5Ј-ggtcatgtccggctgtggaaag-3Ј. The C t values were normalized to input DNA (DNA before the immunoprecipitation step) and IgG controls. Data are represented as -fold differences above mouse control IgG relative to input DNA using the formula 2 ((Ct(IgG)ϪCt(Input))Ϫ(Ct(TPR)ϪCt(Input))) . For the experiments analyzing luciferase cDNA, primers used for quantitative PCR analysis were 5Ј-gtctgaattccagtcgatgtacacgttcg-3Ј and 5Ј-cacgaagcttgcatgcgagaactccacgc-3Ј. The C t values were normalized to input cDNA (cDNA made from NOVEMBER 23, 2007 • VOLUME 282 • NUMBER 47 JOURNAL OF BIOLOGICAL CHEMISTRY 33903 total RNA before the immunoprecipitation step) and IgG controls, which were set as 1 unit. Data are represented as -fold differences relative to these two values using the formula 2 ((Ct(IgG)ϪCt(Input))Ϫ(Ct(TPR)ϪCt(Input))) .

HSF1-TPR Interaction Facilitates Export
For experiments analyzing endogenous HSP70 mRNA levels, the following primers were used: human HSP70, 5Ј-caccttgccgtgttgga-3Ј and 5Ј-ttctcgcggatccagtg-3Ј; and human L32, 5Ј-catctccttctcggcatca-3Ј and 5Ј-aaccctgttgtcaatgcctc-3Ј. The data presented represent two independent experiments in which values obtained utilizing the formula 2 Ϫ⌬⌬Ct were averaged. The results are graphed as relative differences between cytoplasmic and nuclear HSP70 mRNA levels compared with the control samples (GFP alone), which were set to a value of 1.

RESULTS
To further the understanding of the regulation and function of HSF1, we performed a yeast two-hybrid screen using the HSF1 protein as a bait. Two of the interacting clones that were obtained from this screen represented two different regions of the TPR protein. One of the HSF1-interacting regions of TPR identified by the yeast two-hybrid screen comprises a sequence near the N terminus (amino acids 14 -117), whereas the other is located close to the middle of the protein (amino acids 1218 -1320) (Fig. 1, A and B). As an independent test of the interaction between HSF1 and these two regions of TPR and to determine whether the interaction is direct, in vitro binding experiments were performed in which in vitro translated 35 S-labeled TPR-(14 -117) and TPR-(1218 -1320) were incubated with GST-HSF1 or GST bound to glutathione-agarose beads. The results confirmed the ability of both regions of TPR to interact with HSF1 (Fig. 1C).
To determine whether endogenous HSF1 and TPR proteins interact and, if so, whether the interaction between these proteins is regulated in a stress-dependent manner, immunoprecipitation analysis was performed using extracts of cells kept at 37°C or cells subjected to heat stress at 42°C for 1 h. Fig. 2 shows that endogenous HSF1 and TPR did associate and that more HSF1-TPR complex was observed in extracts of stressed cells compared with non-stressed cells.
In multicellular eukaryotes, HSF1 binds to heat shock gene promoters in response to stress conditions. Therefore, the data presented in Fig. 2 indicating that HSF1 interacts with TPR in a stress-induced manner prompted the question of whether TPR might also associate with the promoter of the stress-inducible HSP70 gene when cells are exposed to stress. We tested this hypothesis using the ChIP assay. The results demonstrate that a low level of TPR association was detected within cross-linking distance of the HSP70 promoter in cells kept at 37°C and that a higher level of TPR was associated with the HSP70 promoter in cells that were subjected to stress treatment at 42°C for 1 h (Fig.  3A, upper panels). TPR was not found to associate with the promoter region of the histone H4 gene, indicating the specificity of its HSP70 promoter association (Fig. 3A, lower panels). As a complementary approach, we repeated this experiment and used quantitative real-time PCR for the analysis. The results (Fig. 3B) are consistent with the finding of increased association between TPR and the HSP70 promoter in response to exposure to stress conditions.
On the basis of previous results indicating a role for TPR in mRNA export (13,18,19), including the finding that the yeast TPR ortholog Mlp1p interacts with the mRNA export heterogeneous nuclear ribonucleoprotein Nab2p (19), we hypothesized that the recruitment of TPR to the HSP70 promoter might function as a way to specifically promote association between  TPR and the stress-induced transcripts that arise from this gene. We tested this hypothesis using an RNA immunoprecipitation approach. In this experiment, HeLa cells were transfected with expression constructs in which the luciferase gene is transcribed from either the stress-inducible human HSP70 gene promoter or the RSV promoter. The transfected cells were subjected to 1 h of heat shock treatment (42°C), after which they were incubated with the chemical cross-linking agent paraformaldehyde, and extracts of these cells were then immunoprecipitated using anti-TPR antibody. RNA isolated from the TPR-containing complexes was reverse-transcribed into cDNAs, which were then analyzed by quantitative real-time PCR using a luciferase primer pair. The results of this experiment (Fig. 4) indicate that significantly more luciferase mRNA transcripts generated from the HSP70 promoter were associated with the TPR protein compared with luciferase mRNAs transcribed from the RSV promoter. The results described above indicate that stress conditions result in increased interaction between HSF1 and TPR, increased association of TPR with the HSP70 promoter, and the preferential association of TPR with mRNAs arising from transcription from the HSP70 promoter. In light of the data suggesting a role for TPR in mRNA export (13,18,19), we hypothesized that these events could be part of a mechanism for specifically enhancing the export of mRNAs transcribed from heat shock gene promoters. To test this hypothesis, we sought to deter-  mine whether export of the HSP70 promoter-driven luciferase mRNAs described above is affected by cotransfecting the cells with expression constructs encoding the two regions of the TPR protein (amino acids 14 -117 and 1218 -1320) that were shown by our data to interact with HSF1 (Fig. 1). We reasoned that if HSF1-TPR interaction is important for export of mRNAs expressed from HSP gene promoters, then expressing either of these two HSF1-binding regions of TPR could inhibit export of these mRNAs expressed from HSP gene promoters by decreasing the ability of HSF1 and TPR to associate. Fig. 5A shows that both GFP fusion constructs were expressed at levels similar to that of GFP alone. Fluorescence microscopy analysis of cells transfected with these constructs revealed that a significant proportion of each GFP-TPR fragment fusion construct was found in the nuclei of these cells, where they would need to be to exert their effects in this experiment (supplemental Fig. S1).

α-Tpr
Co-immunoprecipitation analysis confirmed that transfection of the GFP-TPR-(14 -117) and GFP-TPR-(1218 -1320) constructs, but not the GFP expression construct, resulted in decreased levels of the HSF1-TPR complex in stressed cells (Fig. 5B). Next, cells were cotransfected with the GFP-TPR-(14 -117), GFP-TPR-(1218 -1320), or GFP-alone construct along with the construct containing luciferase expressed from either the HSP70 promoter or the RSV promoter (non-heat shock element-containing) and subjected to heat shock treatment at 42°C for 60 min, and mRNA from the cytoplasmic and nuclear fractions of these cells was then analyzed by RNase protection assay using a probe that detects luciferase mRNA or mRNA encoding the L32 ribosomal protein (as a control). The results indicate that cells transfected with GFP-TPR-(14 -117) or GFP-TPR-(1218 -1320) exhibited a decrease in the cytoplasmic levels of luciferase mRNA expressed from the heat shock elementcontaining HSP70 promoter compared with cells transfected with GFP (Fig. 5C, upper panels). Transfection of GFP-TPR- (14 -117) or GFP-TPR-(1218 -1320) did not change the nuclear versus cytoplasmic levels of luciferase mRNA expressed from the non-heat shock element-containing RSV promoter compared with transfection of GFP alone, indicating the HSP70 promoter selectivity of the effect (Fig. 5C, lower panels).
Finally, to test whether expression of these two regions of the TPR protein is able to inhibit export of endogenous HSP70i mRNAs, we performed an experiment similar to the one shown in Fig. 5D, except that the cells were transfected with only the GFP-TPR-(14 -117), GFP-TPR-(1218 -1320), or GFP expression construct (no luciferase reporter constructs). After subjecting the transfected cells to heat shock treatment at 42°C for 60 min, mRNA from the cytoplasmic and nuclear fractions of these cells was reverse-transcribed and then analyzed by quantitative real-time PCR using primers that amplify the HSP70i sequence or the L32 sequence (normalizing control). The results of this experiment (Fig. 5D) are presented graphically as  (1218 -1320), or GFP alone were subjected to heat treatment at 42°C for 60 min, and HSF1 immunoprecipitates from extracts of these cells were then subjected to anti-TPR Western blotting. C, HeLa cells were cotransfected with the GFP-TPR-(14 -117), GFP-TPR-(1218 -1320), or GFP-alone construct along with either a HSP70 promoter-driven (upper panels) or RSV promoter-driven (lower panels) reporter plasmid and subjected to heat shock treatment at 42°C for 60 min, and mRNA from the cytoplasmic (C) and nuclear (N) fractions of these transfected cells was then analyzed by RNase protection assay using a probe that detects luciferase (Luc) mRNA or L32 ribosomal protein mRNA (control). D, HeLa cells were transfected with the GFP-TPR-(14 -117), GFP-TPR-(1218 -1320), or GFP-alone construct and subjected to heat shock treatment at 42°C for 60 min, and mRNA from the cytoplasmic and nuclear fractions of the transfected cells was then reverse-transcribed into cDNA and analyzed by quantitative real-time PCR using primers that amplify a segment of the HSP70i nucleotide sequence or the L32 ribosomal protein nucleotide sequence (normalizing control). The results of this experiment are presented as the relative levels of endogenous HSP70 mRNA in the nucleus or cytoplasm, normalized for L32 mRNA levels, compared with the results for the control cells (transfected with GFP alone), which were set to a value of 1. Data are from duplicate experiments. Data are shown as means Ϯ S.E. the relative levels of endogenous HSP70 mRNA in the nucleus or cytoplasm, normalized for L32 mRNA levels, compared with the results for the control cells (transfected with GFP alone), which were set to a value of 1. The results indicate that the export of endogenous HSP70i mRNA, like that of the heat shock element-driven luciferase mRNAs above, was decreased in cells transfected with GFP-TPR-(14 -117) or GFP-TPR-(1218 -1320) compared with cells transfected with GFP (Fig. 5D).

DISCUSSION
The results presented in this work indicate that, in response to stress, the TPR protein interacts with the stress gene transcriptional regulator HSF1, is recruited to the HSP70 promoter region, and preferentially associates with mRNAs transcribed from this promoter compared with those expressed from a non-stress-induced promoter and that the HSF1-TPR interaction is required for efficient export of HSP mRNAs from the nucleus during stress. The association of TPR with these mRNAs may be assisted by its interaction with mRNA-binding heterogeneous nuclear ribonucleoproteins such as Nab2p (19). The TPR protein is known to be able to form homodimers (25). Thus, once it is complexed with HSP mRNAs, TPR could facilitate their export by docking with TPR found at the nucleoplasmic face of nuclear pore complexes (10 -17).
These results reveal the existence of a direct functional connection between the first and last nuclear steps in the gene expression pathway, transcription and export of mRNAs from the nucleus. The HSF1-TPR interaction and its downstream events could serve as a mechanism for bypassing the inhibition of mRNA export that occurs in response to stress and/or to increase the kinetics of export of HSP mRNAs so that cells can express these crucial cytoprotective proteins as soon as possible. One intriguing question for future studies is whether the TPR protein is recruited to other genes to aid in the export of their mRNAs from the nucleus.