The mitochondrial unfolded protein response (UPRmt) is characterized by the transcriptional induction of mitochondrial chaperone and protease genes in response to impaired mitochondrial proteostasis. Here, we show that heat shock transcription factor (HSF1) is required for activation of mitochondrial chaperone genes and supports the maintenance of mitochondrial function in mouse cells during the UPRmt.
activating transcription factor
CAAT enhancer‐binding protein
C/EBP homology protein
heat shock transcription factor 1
heat shock protein
heat shock response
single‐stranded DNA binding protein 1
unfolded protein response
Protein homeostasis or proteostasis within a cell is adjusted mainly at the levels of protein synthesis, folding, and degradation, and its maintenance is essential for cellular functions. Environmental and metabolic stresses constantly induce protein misfolding and challenge proteostasis capacity. To cope with these proteotoxic stresses, cells are equipped with adaptive mechanisms accompanied by changes in gene expression . Among these, the heat shock response (HSR) is an evolutionarily conserved mechanism that is characterized by the induction of a set of heat shock proteins (HSPs) or chaperones, including HSP110, HSP90, HSP70, HSP40, and HSP27, which assist with protein folding, and some non‐HSP proteins involved in protein degradation . The HSR is regulated mainly at the transcriptional level by heat shock transcription factor 1 (HSF1) in mammalian cells, and it maintains proteostasis capacity in both the nucleus and cytoplasm .
In contrast, analogous adaptive responses against protein misfolding in the endoplasmic reticulum (ER) and mitochondria are called as unfolded protein response in the ER (UPRER) and mitochondrial UPR (UPRmt ), respectively [4, 5]. The latter response is characterized by the induction of mitochondrial chaperones and proteases, which localize and act in the mitochondria in response to the accumulation of misfolded proteins or an imbalance in mitochondrial and nuclear‐encoded proteins in the mitochondria [6, 7, 8]. This response is regulated by the basic leucine zipper (bZIP) transcription factor ATFS‐1 in C. elegans . ATFS‐1, which is localized to the mitochondrial matrix in normal conditions, accumulates in the nucleus and activates the UPRmt genes in response to mitochondrial proteotoxic stress. In addition, several factors including a mitochondrial transporter, transcription factors, and histone‐modifying enzymes are also involved in the UPRmt [10, 11]. In particular, histone demethylases JMJD‐3.1 and JMJD‐1.2 are necessary, and their overexpression is sufficient for the UPRmt . In mammals, the bZIP transcription factor ATF5 is regulated similarly to ATFS‐1 and activates the UPRmt genes during accumulation of truncated ornithine transcarbamylase (ΔOTC) in the mitochondria . Another bZIP transcription factor CHOP in complex with C/EBP also activates the UPRmt genes, and its expression is induced via activation of JUN, which is mediated by c‐Jun N‐terminal kinase 2 during accumulation of ΔOTC [7, 14].
At first, synthesis of a mammalian homolog of the bacterial GroEL protein was found to be elevated during heat shock and was referred to as HSP58 (thereafter HSP60), whereas that of a mitochondrial member of HSP70 family was increased in cells deprived of glucose and was referred to as glucose regulated protein GRP75 (also known as mtHSP70) . Mammalian HSP60 and HSP10 genes are linked head‐to‐head and share a bidirectional promoter, which is activated during heat shock [16, 17]. However, HSF1 was not thought to be involved in the upregulation of HSP60 and HSP10 during the UPRmt , because HSP70 was not upregulated simultaneously [6, 7, 16]. Recently, it was suggested that HSF1 in complex with a coactivator, mitochondrial single‐stranded DNA binding protein 1 (SSBP1), regulates the expression of mitochondrial chaperones, including HSP60, HSP10, and mtHSP70, during heat shock . Of note, not only HSF1 but also mitochondrial SSBP1 accumulates in the nucleus and binds to the promoters of these genes on heat shock conditions . Therefore, it should be determined whether HSF1 and SSBP1 play an indispensable role in the UPRmt. In this study, we showed that HSF1 is required for expression of nuclear‐encoded mitochondrial chaperones, HSP60, HSP10, and mtHSP70, but not for that of Lon protease, in response to impaired mitochondrial proteostasis, whereas SSBP1 is partially required for the induction. Furthermore, HSF1 promoted the maintenance of mitochondrial function during the UPRmt.
Immortalized wild‐type (clone #10) and HSF1‐null (clone #4) mouse embryonic fibroblasts (MEF) , HeLa (ATCC CCL‐2) cells, and HEK293 (ATCC CRL‐1573) cells were maintained at 37 °C in 5% CO2 in Dulbecco's modified Eagle's medium containing 10% fetal bovine serum (Sigma‐Aldrich St. Louis, MO, USA). Cells were treated with mitochondria‐specific stress reagents, 10 μm gamitrinib‐triphenylphosphonium (GTPP) (a kind gift from D. C. Altieri), 5 μm synthetic triterpenoid 2‐cyano‐3, 12‐dioxooleana‐1, 9(11)‐dien‐28‐oic acid (CDDO) (Cayman Chemicals, Ann Arbor, MI, USA), and 20 μm rotenone (Sigma‐Aldrich, St. Louis, MO, USA) for 6 h.
Total RNA was isolated from cells using TRIzol (Ambion, Carlsband, CA, USA). First‐strand cDNA was synthesized using PrimeScript II Reverse Transcriptase and oligo‐dT primer in accordance with the manufacturer's instructions (TAKARA, Kusatsu, Japan). Real‐time quantitative PCR (qPCR) was performed using StepOnePlus (Applied Biosystems, Foster City, CA, USA) with the Power SYBR Green PCR Master Mix (Applied Biosystems) using primers for mouse HSP60 (HSPD1), HSP10 (HSPE1), mtHSP70 (HSPA9), Lon, and HSP70 (HSPA1A and HSPA1B) (Table S1). Relative quantities of mRNAs were normalized against GAPDH or RPLPO (large ribosomal protein) mRNA levels. All reactions were performed in triplicate with samples derived from three experiments.
To generate adenovirus vectors expressing short hairpin RNAs against mouse HSF1, SSBP1 and TRAP1, oligonucleotides containing each target sequence (Table S2) were annealed and inserted into pCR2.1‐hU6 at the BamHI/HindIII sites, and then, XhoI/HindIII fragments containing hU6‐shRNA were inserted into a pShuttle‐CMV vector (Stratagene) . To knock down HSF1, SSBP1, or TRAP1, MEF cells were infected with Ad‐sh‐mHSF1‐KD, Ad‐sh‐mSSBP1‐KD, or Ad‐sh‐mTRAP1‐KD (1 × 108 pfu·mL−1) for 2 h and maintained in normal medium for 70 h. As a control, the cells were infected with an adenovirus vector expressing scrambled RNA (Ad‐sh‐SCR).
Cells pellets were lysed with NP‐40 buffer (150 mm NaCl, 1.0% Nonidet P‐40, 50 mm Tris/HCl, pH 8.0) containing protease inhibitors (1 mm phenylmethylsulfonyl fluoride, 1 μg·mL−1 leupeptin, and 1 μg·mL−1 pepstatin) on ice for 10 min. After centrifugation at 16 000 g for 10 min, supernatants were subjected to SDS/PAGE. For HSP10 blot, a Real Gel Plate with 10‐20% polyacrylamide gel (MDG‐296; BIO CRAFT, Tokyo, Japan) was used. After proteins were transferred onto nitrocellulose or PVDF (SSBP1 blot) membranes, the membranes were blocked in PBS/5% milk at a room temperature for 1 h and then were immunoblotted using rabbit antibodies against HSF1 (anti‐mHSF1j, Millipore ABE1044; dilution, 1 : 1000) , TRAP1 (anti‐mTRAP1a; dilution, 1 : 1000) (this study) HSP60 (anti‐HSP60‐1; 1 : 2000) , HSP10 (Santa Cruz, CA, USA sc‐20958; 1 : 1000), mtHSP70 (or GRP75) (Santa Cruz sc‐13967; 1 : 1000), and SSBP1 (anti‐mSSBP1x; dilution, 1 : 1000) (this study), and mouse antibody for HSP70 (Santa Cruz W27; 1 : 1000) and β‐actin (AC‐15; Sigma, St. Louis, MO, USA) diluted in PBS/2% milk at a room temperature for 1 h or at 4 °C overnight. The membranes was washed three times with PBS for 5 min and incubated at room temperature for 1 h with secondary antibodies: peroxidase‐conjugated goat anti‐rabbit or anti‐mouse IgG. After washing with PBS/0.1% Tween‐20 three times, chemiluminescent signals from ECL detection reagents (GE Healthcare, Buckinghamshire, UK) were captured on X‐ray film (Super RX; Fujifilm, Tokyo, Japan). Intensity of the bands was quantified using NIH imagej (NIH, Washington, DC, USA). We generated rabbit antisera against mouse TRAP1 (anti‐mTRAP1a) and SSBP1 (anti‐mSSBP1x) by immunizing rabbits using TiterMax Gold adjuvant (CytRx, Los Angeles, CA, USA) with bacterially expressed recombinant GST‐mTRAP1 (full‐length protein) and GST‐mSSBP1 (full‐length protein), respectively.
Mouse embryonic fibroblasts cells were treated with 10 μm GTPP, 5 μm CDDO, or 20 μm rotenone for 6 h, or heat shock at 42 °C for 30 min. Whole‐cell extracts were prepared in buffer C (0.42 m NaCl, 20 mm HEPES/NaOH, pH 7.9, 25% glycerol, 1.5 mm MgCl2, 0.2 mm EDTA) containing protease inhibitors . Aliquots containing 40 μg protein were mixed with a 0.05 volume of 100 mm disuccinimidyl glutarate (DSG) (final concentration of 5 mm) at room temperature for 30 min and were subjected to western blotting using HSF1 antibody (anti‐mHSF1j).
HeLa cells were grown on coated glass coverslips in 35 mm culture dishes for 16 h at 37 °C in 5% CO2. Cells were fixed with 100% methanol at −20 °C for 15 min and then washed three times with PBS for 5 min each. Subsequently, they were permeabilized and blocked with PBS/0.1% Triton X‐100/5% goat serum at room temperature for 1 h. After washing with PBS once, the coverslips were incubated with rat monoclonal IgG for HSF1 (10H8, ab61382; Abcam, Cambridge, UK) (1 : 200 dilution) at 4 °C overnight and washed three times with PBS. They were then incubated with FITC‐conjugated goat anti‐rabbit IgG (Cappel) (1 : 200 dilution in PBS/2% milk) or Alexa Flour 546‐conjugated goat anti rat IgG (Molecular Probes, Eugene, OR, USA) (1 : 200 dilution) at room temperature for 1 h. Coverslips were washed three times with PBS for 5 min each and then mounted in a VECTASHIELD with 4′,6‐diamino‐2‐phenylindole (DAPI) mounting medium (Vector Laboratories, Burlingame, CA, USA). High‐resolution (×63 objective magnification) confocal images were taken using LSM510 META confocal microscope (Carl Zeiss, Jena, Germany) and were quantified by using Zen lite software (Carl Zeiss). HSF1 fluorescence signals in a total cell and a nucleus were estimated by measuring the average intensities of pixels by manually tracing cellular periphery and the region stained with DAPI, respectively. Percentage of HSF1 fluorescence signal localized in the nucleus was calculated by normalizing the nuclear signal intensity to total fluorescence intensity from the cell.
ChIP experiments were performed using a kit in accordance with the manufacturer's instructions (EMD Millipore, Burlington, MA, USA). The antibody used for ChIP assays was anti‐mHSF1j. Real‐time qPCR of ChIP‐enriched DNAs in HSP60, mtHSP70, HSP70 (HSPA1A ), and its intergenic region was performed using the primers listed in Table S3. Percentage input was determined by comparing the cycle threshold value of each sample to a standard curve generated from a 5‐point serial dilution of genomic input and compensated by values obtained using normal IgG. IgG‐negative control immunoprecipitations for all sites yielded < 0.05% input. All reactions were performed in triplicate with samples derived from three experiments.
Mouse embryonic fibroblast cells, which were infected with Ad‐sh‐mHSF1‐KD or Ad‐sh‐SCR, were seeded into plastic 96 well plates at a density of 5 × 104 cells/well and grown for 16 h. After treatment with each inhibitor for 3 h, the cells were stained with MitoTracker Red CMXRos (Molecular Probes) for 30 min. The wells were washed twice with PBS to remove excess fluorescent dye, and fluorescence signals were measured at 540 nm/615 nm (excitation/emission) using an ARVO X4 multilabel plate reader (PerkinElmer, Inc., Waltham, MA, USA). Alternatively, cells infected with Ad‐sh‐mHSF1‐KD or Ad‐sh‐SCR were grown on glass coverslips in 35 mm culture dishes for 16 h, treated as described above, and were fixed with 100% methanol at −20 °C for 15 min. Coverslips were washed three times with PBS for 5 min each and then mounted in a VECTASHIELD with DAPI mounting medium (Vector Laboratories). High‐resolution (×63 objective magnification) confocal images were taken using LSM510 META confocal microscope (Carl Zeiss).
Oxygen consumption was examined by using MitoXpress Xtra Oxygen Consumption Assay (Agilent, Chicopee, MA, USA) in accordance with the manufacturer's instructions. MEF cells were treated as described above in plastic 96 well plates and maintained at 37 °C on a thermoregulator. The cells were loaded with a reagent containing the oxygen‐sensitive MitoXpress Xtra fluorescent probe and treated with or without 500 nm FCCP (Cayman Chemical) or 5 μm antimycin A (Abcam), and were covered by mineral oil. Each sample well was then measured at 340 nm/642 nm (excitation/emission) repetitively every 5 min over 120 min using an ARVO X4 multilabel plate reader (PerkinElmer, Inc.), by taking TR‐F intensity readings at delay time of 30 and 70 μs and gate time 100 μs. Measured TR‐F intensity signals (counts·s−1) were converted into lifetime signals (μs). Relative oxygen consumption rate (OCR) was estimated as a value of MitoXpress Xtra fluorescence lifetime signal per hour per mg of protein (μs·h−1·mg−1).
Data were analyzed using Student's t‐test for comparisons between two groups. Multiple‐group differences were assessed by one‐way ANOVA test, followed by the Tukey post hoc test (jmp pro 14 software; SAS Institute Inc., Cary, NC, USA). Asterisks in figures indicate that differences were significant (P < 0.01 or 0.05). Error bars represent the standard deviations for more than three independent experiments.
To examine the roles of HSF1 in the UPRmt , we treated immortalized MEF cells with three reagents that target mitochondrial proteins and impair mitochondrial proteostasis. GTPP inhibits the matrix HSP90 chaperone TRAP1 [23, 24], and CDDO inhibits the matrix protease Lon . Rotenone is an inhibitor of the electron transfer complex 1 (ETC1) and increases production of reactive oxygen species (ROS). Protein levels of HSP60 and HSP10 were increased by treatment with GTPP or CDDO at concentrations of 5–20 μm , but were not by treatment with rotenone (Fig. 1A). In contrast, mtHSP70 protein levels were increased by treatment with 10–50 μm rotenone and were slightly increased by treatment with GTPP or CDDO. HSP60 mRNA levels were also increased by treatment with GTPP or CDDO, and mtHSP70 mRNA levels were increased by treatment with rotenone (Fig. 1B). Thus, the treatment of MEF cells with these reagents induced at least some mitochondrial HSPs in a dose‐dependent manner, as reported previously [26, 27]. We then treated wild‐type and HSF1‐null MEF cells with 10 μm GTPP, 5 μm CDDO, or 20 μm rotenone for 6 h and found that HSP60, HSP10, and mtHSP70 mRNA levels were increased by 1.2‐ to 2.0‐fold in wild‐type cells treated with GTPP and CDDO, and only mtHSP70 mRNA levels were significantly increased in cells treated with rotenone (Fig. 1C). Remarkably, mRNA levels of these genes were not induced in HSF1‐null cells at all. mRNA levels of HSP70 were simultaneously increased by 10‐ to 45‐fold in wild‐type cells in a HSF1‐dependent manner, suggesting that cytoplasmic proteostasis was also impaired in these conditions (Fig. 1C) . In marked contrast, expression of mitochondrial protease Lon mRNA was induced in both wild‐type and HSF1‐null cells during the treatment (Fig. 1D). To exclude nonspecific effects of GTPP, we knocked down TRAP1 and confirmed that both HSP60 and HSP70 protein levels were increased in TRAP1‐knockdown cells (Fig. 1E) [27, 29]. HSP60, HSP10, and HSP70 mRNA levels were also increased by about 1.5‐fold (Fig. 1F). However, they were not increased at all in TRAP1‐knockdown cells deficient in HSF1. These results demonstrated that HSF1 is required for activation of mitochondrial chaperone genes, but not for that of Lon protease gene, during the UPRmt in mouse cells, when mitochondrial proteostasis is impaired by targeting a mitochondrial chaperone or protease, or an ETC component.
We then investigated the effects of SSBP1 on the activation of UPRmt genes in response to impaired mitochondrial proteostasis. MEF cells were infected for 72 h with an adenovirus expressing short hairpin RNA for SSBP1 or HSF1, or scrambled RNA (SCR) as a control, and protein level of SSBP1 or HSF1 was transiently reduced (Fig. 2A). We confirmed that the expression of HSP60, HSP10, and mtHSP70 mRNAs as well as HSP70 mRNA was hardly increased in HSF1‐knockdown cells during treatment with GTPP, CDDO, or rotenone (Fig. 2B–D, black bars). In SSBP1‐knockdown cells, the expression of HSP70 mRNA was partially increased during the same treatment (Fig. 2B–D, gray bars), like during treatment with heat shock . In marked contrast, HSP60 mRNA expression was not increased at all in SSBP1‐knockdown cells during treatment with GTPP or CDDO. Similarly, HSP10 mRNA expression was less increased in SSBP1‐knockdown cells during GTPP and CDDO treatment than scrambled RNA‐treated cells (Fig. 2B,C, gray bars). On the other hand, mtHSP70 mRNA expression was fully increased in SSBP1‐knockdown cells during CDDO treatment, whereas it was less increased in the same cells treated with GTPP or rotenone (Fig. 2B–D, gray bars). These results suggested different requirements of SSBP1 on the activation of mitochondrial chaperone genes during the UPRmt.
We investigated whether HSF1 is activated directly or indirectly during treatment with GTPP, CDDO, or rotenone. HSF1 activation involves its nuclear translocation, trimer formation, and phosphorylation of a specific residue [30, 31]. First, we performed immunofluorescence analysis of HeLa cells using confocal microscopy because HSF1 localization was intensively studied in the cells . We found that HSF1 localizes to both the cytoplasm and nucleus in unstressed cells and slightly accumulates in the nucleus during treatment with GTPP, CDDO, or rotenone (Fig. 3A,B). Nuclear foci termed HSF1 granules were detected in cells treated with heat shock but not in cells treated with these reagents. Second, we examined the oligomeric form of HSF1 using DSG cross‐linking experiments. Monomeric HSF1 shifted to a trimeric form during treatment of MEF cells with heat shock and was partly shifted to a trimeric form during treatment with GTPP, CDDO, or rotenone (Fig. 3C). Third, we studied HSF1‐Ser326 phosphorylation, which is an active mark of HSF1 transcriptional activity . Because a specific antibody for human HSF1‐Ser326, but not for mouse HSF1‐Ser326, is available, we replaced endogenous HSF1 with human HSF1 in MEF cells. It was revealed that hHSF1‐Ser326 was phosphorylated at lower levels in cells treated with GTPP, CDDO, or rotenone than in cells treated with heat shock at 42 °C 90 min (Fig. 3D). Hyperphosphorylation of HSF1, which is detected as retarded bands on a gel, is often correlated with the activation of HSF1, but was not evident in the same cells. These results suggested that HSF1 is modestly activated in response to impaired mitochondrial proteostasis.
It was assumed that HSF1 mildly occupies mitochondrial chaperone gene promoters in vivo in impaired mitochondrial proteostasis conditions, because it is activated only modestly. As shown previously, HSF1 heavily bound to HSP60/HSP10 promoter as well as HSP70 (HSPA1A) promoter, and a little to mtHSP70 promoter during heat shock at 42 °C for 30 min (Fig. 4A,B). In contrast, HSF1 moderately bound to HSP70 promoter in cells treated with GTPP and CDDO and bound to it at a lower level in cells treated with rotenone. HSF1 constitutively bound to mtHSP70 promoter to some extent, and its binding was induced moderately in cells treated with rotenone and was little induced in cells treated with GTPP and CDDO. HSF1 also constitutively bound to HSP60/HSP10 promoter to some extent, and the levels of HSF1 binding were little induced in cells treated with rotenone. Contrary to our expectation, the levels of HSF1 binding were heavily induced in cells treated with GTPP and CDDO, like in cells treated with heat shock (Fig. 4A,B). Furthermore, we confirmed that levels of HSF1 binding to HSP60/HSP10 promoter were markedly induced in TRAP1‐knockdown cells (Fig. 4C). These results indicated that HSF1 occupancy on the mitochondrial chaperone gene promoters is induced at different levels. HSF1 occupancy in HSP60/HSP10 promoter was remarkably high during the treatment with GTPP and CDDO, whereas that in mtHSP70 or HSP70 promoter was moderate.
To test whether HSF1‐mediated expression of UPRmt genes is related with mitochondrial function, we first examined mitochondrial membrane potential using a fluorescent probe MitoTracker Red. The intensity of MitoTracker fluorescence was not affected when MEF cells were treated with 10 μm GTPP, 5 μm CDDO, or 20 μm rotenone for 3 h (Fig. 5A). However, it was significantly reduced in HSF1‐knockdown cells treated with GTPP or rotenone, but not in those cells treated with CDDO. We next examined the basal (−FCCP) and maximal (+FCCP) oxygen consumption in the same cells (Fig. 5B). The relative oxygen consumption rate (OCR) was not significantly reduced by HSF1 knockdown, but was reduced in cells treated with GTPP, CDDO, or rotenone for 3 h. Remarkably, the levels of basal and maximal respiration in the presence of GTPP, CDDO, or rotenone were more reduced in HSF1‐knockdown cells than those in scrambled RNA‐treated cells. These results suggested that HSF1 promotes the maintenance of mitochondrial function in response to impaired mitochondrial proteostasis.
Mitochondria are the central hub of metabolic and signaling processes including ATP production and apoptotic cell death [33, 34], and declines in mitochondrial function are associated with aging and disorders, such as neurodegenerative diseases and cancer [35, 36]. Cells must adapt to a large variety of mitochondrial dysfunctions by changing nuclear‐encoded mitochondrial gene expression. Among these homeostatic mechanisms, the UPRmt is an adaptive response to accumulation of misfolded proteins in mitochondria. ATF5 and CHOP have been shown to be required for the activation of UPRmt genes during accumulation of ΔOTC in human HEK293 and monkey COS‐7 cells, respectively [7, 13]. In this study, we used immortalized MEF cells for analysis of the UPRmt, and mechanisms of the UPRmt were analyzed during treatment with GTPP, CDDO, or rotenone [26, 27, 37], which induces the expression of HSP60, HSP10, mtHSP70, or Lon as well as cytoplasmic HSP70. We showed that both disruption of HSF1 gene and transient HSF1 knockdown abolished the upregulation of mitochondrial chaperone genes, but not for that of protease Lon, during the UPRmt (Figs 1 and 2). In contrast, SSBP1 is required for the upregulation of only HSP60. Even in unstressed conditions, HSF1 constitutively occupied HSP60/HSP10 and mtHSP70 promoters (Fig. 4). Furthermore, a very small part of HSF1 accumulated in the nucleus, shifted to a trimeric form, and was phosphorylated at Ser326, suggesting that HSF1 was activated directly or indirectly in response to impaired mitochondrial proteostasis (Fig. 3). Although treatment with the inhibitors may also cause proteostasis impairment in the cytoplasm, our observation indicated that HSF1 is required for activation of mitochondrial chaperone genes during the UPRmt.
HSF1 has been shown to plays roles in the maintenance of mitochondrial function through different pathways. HSF1 deficiency causes reduced constitutive expression of cytoplasmic HSPs including HSP25, which is associated with a decrease in cellular GSH/GSSG ratio and an increase in mitochondrial oxidative stress in the heart, kidney, and oocytes [38, 39, 40, 41]. Induction of HSPs, including HSP60 and HSP10, by HSF1 and SSBP1 promotes the maintenance of mitochondrial membrane potential in proteotoxic stress conditions, which are caused by heat shock or proteasome inhibition . Furthermore, activation of HSF1 is associated with increased mitochondrial function by enhancing the expression of PGC1α, which is a central regulator of mitochondrial biogenesis and function . Consistently, mitochondrial function such as mitochondrial membrane potential is suggested to be more reduced by the expression of an aggregation‐prone polyglutamine protein in HSF1‐knockdown cells . Here, we showed that mitochondrial membrane potential or relative OCR were more reduced in HSF1‐knockdown cells than those in scrambled RNA‐treated cells during treatment with GTPP, CDDO, or rotenone (Fig. 5). Our observations suggested that mitochondrial function in conditions of impaired mitochondrial proteostasis is maintained in part by the HSF1‐dependent upregulation of mitochondrial chaperone genes.
It is worth noting that HSP60 and HSP10 uniquely share a bidirectional promoter containing an HSE, which consisted of at least four inverted repeats of an exceptionally conserved consensus nGAAn unit [16, 17]. ChIP‐seq and ChIP‐qPCR data analysis showed that HSF1 constitutively binds to the bidirectional promoter at a much higher level than to the promoters of other HSP genes including HSP70 in MEF cells, and the level of HSF1 binding to this promoter was dramatically elevated during heat shock [18, 19]. HSF1 was mostly converted to a DNA‐binding trimer during heat shock, whereas a small part of HSF1 shifted to a trimer during the UPRmt (Fig. 3C). Unexpectedly, in vivo HSF1 binding to the bidirectional promoter was induced in cells treated with GTPP and CDDO at the same levels as that in cells treated with heat shock (Fig. 4), although level of HSF1 binding to this promoter was little elevated in cells treated with rotenone. Thus, analysis of in vivo HSF1 binding to the unique bidirectional promoter of HSP60/HSP10 could be a sensitive marker of the UPRmt.
The authors would like to thank Drs D.C. Altieri and Y.C. Chae (The Wistar Institute, Philadelphia, PA) for a gift of GTPP. This work was supported by JSPS KAKENHI grant numbers 15H04706 and 18H02625 (to AN), the Uehara Memorial Foundation (to AN), and the Yamaguchi University ‘Pump‐Priming Program’ (to AN). AK was supported by Otsuka Toshimi Scholarship Foundation, Sojinkai Foundation, and Yamaguchi University Foundation.