
==== Front
Neoplasia
Neoplasia
Neoplasia (New York, N.Y.)
1522-8002
1476-5586
Neoplasia Press

S1476-5586(24)00096-4
10.1016/j.neo.2024.101055
101055
Original Research
Unveiling the impact of SUMOylation at K298 site of heat shock factor 1 on glioblastoma malignant progression
Li Xiang abc1
Wang Zongqi ab1
Gao Bixi ab
Dai Kun ab
Wu Jiang ab
Shen Kecheng ab
Li Guangzhao ab
Niu Xiaowang ab
Wu Xin ab
Li Longyuan ab
Shen Haitao ab
Li Haiying ab
Yu Zhengquan ab
Wang Zhong wangz8761@163.com
ab⁎
Chen Gang gangchen@suda.edu.cn
ab⁎
a Department of Neurosurgery & Brain and Nerve Research Laboratory, The First Affiliated Hospital of Soochow University, Soochow University, Suzhou 215006, China
b Institute of Stroke Research, Soochow University, Suzhou 215006, China
c Department of Neurosurgery, Xinghua People's Hospital Affiliated to Yangzhou University, Xinghua 225700, China
⁎ Corresponding authors at: Department of Neurosurgery & Brain and Nerve Research Laboratory, The First Affiliated Hospital of Soochow University, 188 Shizi Street, Suzhou 215006, China wangz8761@163.comgangchen@suda.edu.cn
1 These authors contributed equally to this work.

10 9 2024
11 2024
10 9 2024
57 1010554 7 2024
3 9 2024
5 9 2024
© 2024 The Authors. Published by Elsevier Inc. CCBYLICENSE.
2024

https://creativecommons.org/licenses/by-nc/4.0/ This is an open access article under the CC BY-NC license (http://creativecommons.org/licenses/by-nc/4.0/).
Background

Glioblastoma (GBM) poses a significant medical challenge due to its aggressive nature and poor prognosis. Mitochondrial unfolded protein response (UPRmt) and the heat shock factor 1 (HSF1) pathway play crucial roles in GBM pathogenesis. Post-translational modifications, such as SUMOylation, regulate the mechanism of action of HSF1 and may influence the progression of GBM. Understanding the interplay between SUMOylation-modified HSF1 and GBM pathophysiology is essential for developing targeted therapies.

Methods

We conducted a comprehensive investigation using cellular, molecular, and in vivo techniques. Cell culture experiments involved establishing stable cell lines, protein extraction, Western blotting, co-immunoprecipitation, and immunofluorescence analysis. Mass spectrometry was utilized for protein interaction studies. Computational modeling techniques were employed for protein structure analysis. Plasmid construction and lentiviral transfection facilitated the manipulation of HSF1 SUMOylation. In vivo studies employed xenograft models for tumor growth assessment.

Results

Our research findings indicate that HSF1 primarily undergoes SUMOylation at the lysine residue K298, enhancing its nuclear translocation, stability, and downstream heat shock protein expression, while having no effect on its trimer conformation. SUMOylated HSF1 promoted the UPRmt pathway, leading to increased GBM cell proliferation, migration, invasion, and reduced apoptosis. In vivo studies have confirmed that SUMOylation of HSF1 enhances its oncogenic effect in promoting tumor growth in GBM xenograft models.

Conclusion

This study elucidates the significance of SUMOylation modification of HSF1 in driving GBM progression. Targeting SUMOylated HSF1 may offer a novel therapeutic approach for GBM treatment. Further investigation into the specific molecular mechanisms influenced by SUMOylated HSF1 is warranted for the development of effective targeted therapies to improve outcomes for GBM patients.

Graphical abstract

Image, graphical abstract

Keywords

Glioblastoma
SUMOylation
Heat shock factor 1
Mitochondrial unfolded protein response
Heat shock protein 60
Mitochondrial heat shock protein 70
==== Body
pmcIntroduction

Glioblastoma (GBM) is one of the most common and aggressive cancers affecting the central nervous system, classified by the World Health Organization as a grade IV astrocytic tumor with a poor prognosis [1]. While there have been improvements in the median life expectancy and 2-year survival rate for GBM patients [2], the 5-year overall survival rate remains low at only 9.8 %, even with a combination of concomitant adjuvant temozolomide therapy and radiotherapy [3]. Consequently, GBM presents a complex medical challenge that warrants further comprehensive and methodical investigation into mechanisms aimed at enhancing treatment outcomes.

During the uncontrolled growth of cancer cells, they acquire various vulnerabilities that must be supported by essential mechanisms [4]. Mitochondria play a crucial role in sustaining cancer growth and survival at different stages [5]. The mitochondrial unfolded protein response (UPRmt), a stress response found in mitochondria of both C. elegans and mammals [6,7], is vital for maintaining mitochondrial health and facilitating tumor development [8,9]. Healthy cells activate the UPRmt to regulate mitochondrial protein balance, promoting cellular stability [10]. Cancer cells exploit this unique pathway to bolster their long-term survival, facilitating cancer progression and metastasis [8,11]. The UPRmt is hyperactive in cancer, and its components are elevated in various cancer types [[11], [12], [13]]. Two chaperone systems, Heat Shock Protein 60 (HSP60) and Mitochondrial Heat Shock Protein 70 (mtHSP70), aid in protein folding within the mitochondrial matrix [14,15]. Studies indicated that HSP60 and mtHSP70 are abundantly present in GBM and are correlated with tumor grading according to the WHO classification [16,17].

Heat Shock Factor 1 (HSF1) is an essential, evolutionarily conserved transcription factor that triggers the expression of downstream heat shock proteins. Its activity is meticulously regulated through post-translational modifications, including phosphorylation, ubiquitination, and acetylation [[18], [19], [20]]. Among these modifications, SUMOylation, occurring at lysine residues, plays a critical role in mediating protein-protein interactions, intracellular localization, and cellular activities [21,22]. Initial observations of heightened SUMOylation were noted in GBM and other astrocytic malignancies [23]. Yang et al. reported a nearly 30-fold increase in SUMO1 and SUMO2/3-linked proteins in GBM patient samples. The levels of SUMOylation were particularly elevated in grade II and III astrocytomas, peaking in GBM [23]. Importantly, this increase in SUMOylation has been consistently observed in patient samples and various human GBM cell lines [24]. However, the mechanism underlying the SUMOylation modification of HSF1, its impact on downstream heat shock protein expression, especially its interplay with the UPRmt mediated by HSP60 and mtHSP70, and its influence on glioblastoma proliferation, apoptosis, migration, invasion, and other processes, remains elusive.

The SUMO (Small Ubiquitin-like Modifier) polypeptide, also known as Sentrin, is approximately 12 kDa in size and forms reversible covalent bonds with target proteins at lysine residues. Lysine residues targeted for SUMOylation in proteins are typically situated within a consensus recognition motif, ΨKxE, where Ψ represents a large hydrophobic amino acid, X represents any amino acid, and K denotes the site for SUMO conjugation [25]. Mammals have three main SUMO paralogs: SUMO1, SUMO2, and SUMO3. While SUMO1 shares only 47 % homology with SUMO2, SUMO2 and SUMO3 exhibit 97 % homology and are indistinguishable by antibodies [25].

Our study revealed that HSF1 can be modified by SUMO1 and SUMO2/3, with this modification predominantly occurring at lysine residue K298, with SUMO1 modification prevailing. SUMOylation of HSF1 facilitates its nuclear translocation from the cytoplasm, shielding it from degradation by the ubiquitination system and enhancing protein stability, while not affecting its trimer conformation. This modification also amplifies the expression of mtHSP70 and HSP60, thereby promoting the UPRmt, which in turn fuels glioblastoma proliferation, migration, and invasion, while dampening apoptosis.

Materials and methods

Cell culture

Cell lines were acquired from Zhongqiao Xinzhou Biotechnology Co., Ltd. in Shanghai, China. Glioma cell lines U87, T98G, U118MG, SF26, LN229, and U251, along with human astrocytes (HA), were cultured in DMEM with 10 % FBS and 1 % penicillin/streptomycin. For the culture of HA, astrocyte growth factors were included in the complete medium. The cells were then incubated in a humidified chamber set at 37°C with 5 % CO2 levels.

Mass spectrometry analysis

The samples were analyzed on Thermo Fisher LTQ Obitrap ETD mass spectrometry. Briefly, the samples were loaded onto an high-performance liquid chromatography (HPLC) chromatography system (Thermo Fisher Easy-nLC 1000) equipped with a C18 column (1.8 mm, 0.15 × 1,00 mm). Solvent A contained 0.1 % formic acid and solvent B contained 100 % acetonitrile. The elution gradient was from 4 % to 18 % in solvent A for 182 min and 18 % to 90 % in solvent B for 13 min at a flow rate of 300 nL/min. Mass spectrometry analysis was carried out at the AIMS Scientific Co.,Ltd. (Shanghai, China) in the positive-ion mode with an automated data-dependent MS/MS analysis with full scans (350–1,600 m/z) acquired using Fourier transform mass spectrometer at a mass resolution of 30,000 and the 10 most intense precursor ions were selected for MS/MS. The MS/MS was acquired using higher-energy collision dissociation at 35 % collision energy at a mass resolution of 15,000.

Plasmid construction and transfection

Using FUGW-CMV-mCherry-polyA-hUBC-EGFP (NheI&HpaI) as the linear vector, we employed seamless cloning to individually insert the fragments of HSF1-WT, HSF1-K91R, HSF1-K126R, and HSF1-K298R to obtain overexpression lentiviral plasmids. These plasmids can be directly transfected into cells for transient overexpression or packaged into lentivirus to infect target cells and obtain stable overexpression cell lines. DNA sequencing analysis confirmed all structures.

Plasmids were transfected into cells using Lipofectamine 3000 (Invitrogen, NY, USA) according to the manufacturer's guidelines when the cell confluence reached 80–90 %.

Generation of stable cell lines

Using the lentiviral plasmid system, HEK293T cells were conventionally resuscitated, passaged and expanded to 10 cm cell culture dishes. When the cells reached approximately 80 % confluency, 10 ml of fresh culture medium was replaced in each dish. Opti-MEM (750 μl), lentiviral vector plasmid (FUGW, 17 μg), PspAx2 (11 μg), PmD2G (6 μg), and Lipo8000 (29 μl) were mixed gently in an EP tube and added to the culture dish. After gently shaking to mix, the cells were incubated for an additional 48 h, and the liquid above was gathered. 4 × universal virus concentrate was added in the right quantity, and the mixture was then refrigerated at 4°C overnight. After following the directions, the potent virus was retrieved and placed in storage at a temperature of -80°C.

Healthy U87 and U251 cells were passaged at a 1:3 ratio into 24-well plates. After 24 h of incubation, when the cells reached approximately 60 % confluency, an appropriate amount of virus suspension was added. After mixing, the cells were cultured for an additional 48 h. Green fluorescence from EGFP was detected using a fluorescence microscope. After reaching a confluency of 90 % or higher, the cells were then passaged and grown in 6-well plates. When the cells reached over 90 % confluency, they were passaged into T25 culture flasks in preparation for flow cytometry.

BD flow cytometry was used to sort the green fluorescent cells. Each cell strain was sorted for approximately 5 million cells, cultured in 24-well plates with DMEM high glucose medium containing 3 % dual antibodies. When the cell confluency reached 90 % or more, routine passaging and expansion were performed, and cells were promptly cryopreserved. Early passage cells were used for each experiment.

Protein aggregation assays

Aggregated proteins were detected with the PROTEOSTAT Aggresome Detection Kit (ENZ-51035; Enzo Life Sciences, Farmingdale, NY, USA) according to the manufacturer's instructions. In short, U87 and U251 cells that had been stably transfected were placed in 24-well culture plates. Upon reaching approximately 80 % confluence, the cells were treated with 4 % formaldehyde at 25°C for 30 min, followed by permeabilization using 0.5 % Triton X-100 and 3 mm EDTA (pH 8.0) on ice for another 30 min. Afterward, the cells were treated with a dual detection solution that included Hoechst 33,342 (for nuclear staining) and PROTEOSTAT dye reagent. The specimens were shielded from light and left to incubate at ambient temperature for half an hour. Following incubation, the cells were rinsed, and then the coverslip was positioned onto a glass slide for observation under a microscope. Aggregated protein imaging was performed using fluorescence microscopy.

Nuclear‑cytoplasm fractionation

Nuclear and cytoplasmic protein fractionation was conducted using the Nuclear and Cytoplasmic Protein Extraction Kit (P0033, Beyotime, Shanghai, China) following the manufacturer's instructions and detected by Western blot.

Stability assay

Lentiviral stably transfected U87 cells were treated with cycloheximide (100 µg/ml). At different times after treatment, cells were analyzed by Western blot with the indicated antibodies (Table S1). Bands intensity was measured using ImageJ software. Flag-HSF1 bands intensity was normalized to a housekeeping protein.

Molecular docking

Using AlphaFold2 to predict and construct the structures of human HSF1 (WT, UniProt ID: Q00613) and KPNA6 (MUT, 298R) [26]. Next, protein structure optimization and adjustment are performed. The first step involves protonation processing at a neutral pH of 7, which is carried out using the H++3 online tool [27]. Subsequently, UCSF Chimera software [28] is employed to assign Amber14SB charges. Next, the trimeric protein molecular docking is conducted using the professional protein-protein docking tool HDOCK [29]. The empirically derived iterative scoring function ITScorePP is utilized for performing molecular docking and configuration scoring. Negative scores indicate molecular binding, and larger absolute values imply stronger binding capabilities. Docking has a maximum output configuration number of 100, with the top 10 configurations being scored. The top 10 configurations are evaluated based on their docking and confidence scores, with the highest scoring one being chosen for additional analysis. PyMOL 2.41 [30] is utilized for 3D visualization analysis, while the academic version of LigPlot 2.1 [31] is employed for visual representation.

Mitochondrial membrane potential (MMP) assay

The Mitochondrial membrane potential [32] was detected using the Enhanced mitochondrial membrane potential JC-1 assay kit (C2003S, Beyotime). Briefly, cells were exposed to JC-1 staining solution for 20 min in a humid incubator at 37°C with 5 % CO2 and then examined using fluorescence microscope and ELISA reader.

Transwell assay

The migration/invasion assay was performed using the Boyden chambers (8-μm pore size, Corning). The treated cells (2 × 104/well) in 100 μL non-FBS medium were plated into the inserts and the lower compartment was filled with medium augmented with 10 % FBS. After reacting for 12 h, the cells on top of the chambers were cleaned with Q-tips and cells on the lower compartment were fixed in 4 % paraformaldehyde (DF0135, Leagene, China) for 15 min before staining with 0.1 % crystal violet (548-62-9, Qiangshun, China) for 0.5 h. Using an optical microscope (AE 2000, Motic, Germany), we obtained images of the membrane's bottom side. For the migration assay, except that the transwell membranes were precoated with Matrigel (R&D Systems, USA) and the cells were incubated for 18 h at 37°C in 5 % CO2. By using ImageJ software, the number of invaded and migrated cells was quantified .

Wound healing analysis

Briefly, 5 × 105 cells per well were seeded in 6-well plates and incubated overnight at 37°C with 5 % CO2. Afterwards, we used a 0.1 ml sterile yellow pipette tip to create a gap. Following a 24-h incubation at 37 degrees Celsius with 5 % carbon dioxide, the outcomes were examined and images were captured using an optical microscope at two different time intervals (0 and 24 h).

Subcutaneous tumor implantation

Male BALB/c nude mice (aged 4 weeks, weight 20–22 g, from Shanghai laboratory animal Center, Shanghai, China) were subcutaneously inoculated with U87 cell suspension (5 × 106 in 100 μL sterilized PBS) . Tumor size was measured every week by a caliper with the formula 0.52 × L × W2, where L indicates length and W indicates width. For tumorigenesis assay, all mice were sacrificed and tumors were harvested 3 weeks later, followed by photography and weighing.

Statistical analysis

Statistical analyses were performed using the Prism 8.0.2 software (GraphPad Software, USA). Quantitative data were compared using a student's t-test between two samples or one-way analysis of variance (ANOVA) for multiple samples. The data of subcutaneous tumor diameter and protein stability in each group were analyzed using two-way analysis of variance (ANOVA). All results were indicated as the mean ± S.D. and repeated in at least three independent experiments. P-value < 0.05 was considered statistically significant.

Results

HSF1 is modified by SUMOylation at lysine residues K298

To ascertain whether HSF1 can be SUMOylated, we conducted affinity purification and mass spectrometry. Analysis of the mass spectrometry results revealed that HSF1 can interact with SUMO1, SUMO2, and SUMO3, indicating that HSF1 can be SUMOylated in glioblastoma cells (Figs. 1A and S1D).Fig. 1 HSF1 is modified by SUMOylation at lysine residues K298 (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article).

Fig 1(A) U87 cells were lysed, and HSF1 protein was affinity purified using magnetic beads. The interacting proteins were identified by mass spectrometry. The raw mass spectrometry data were searched against the uniprot-Bifidobacterium bifidum ATCC 29521.fasta database using Proteome Discoverer 1.4 software (Thermo). Analysis of the detection results revealed the presence of SUMO1, SUMO2, and SUMO3 peptide segments among the interacting proteins, and mass spectra were generated. (B) U87 cells were lysed, and the same lysate was subjected to immunoprecipitation using IgG, SUMO1, and SUMO2/3 antibodies, followed by Western blot analysis using anti-HSF1 antibody. (C) Flag-HSF1 plasmids were transfected into U87 cells. After 24 h, the cells were lysed, and immunoprecipitation was performed using Flag antibody, followed by Western blot analysis using anti-SUMO1 antibody. (D) The complete amino acid sequence of HSF1 was exported from the Uniprot website. Six potential SUMOylation sites were identified based on the SUMOylation consensus sequence: ψkxE (highlighted in red box). (E) Flag-HSF1 WT, Flag-HSF1 K91R, Flag-HSF1 K126R, and Flag-HSF1 K298R plasmids were transfected into U87 cells. After 24 h, the cells were lysed, and immunoprecipitation was performed using Flag antibody, followed by Western blot analysis using anti-SUMO1 antibody. (F) The 298th lysine residue site of HSF1 protein is highly conserved among different species. Amino acid alignment near the K298 position of HSF1 sequences from humans and various species.

In order to confirm the findings from mass spectrometry and show the in vivo SUMOylation of HSF1, we utilized a technique that enables the identification of highly concentrated SUMO targets in natural settings to assess the SUMOylation of native HSF1 protein [33]. As shown in the figure (Fig. 1B), we detected SUMOylated bands of HSF1 in the SUMO immunoprecipitation extracts, but not in the IgG immunoprecipitation extracts, indicating that endogenous HSF1 is SUMOylated in cells. By comparing the bands of SUMO-1 and SUMO-2/3, it was observed that the conjugation level of SUMO-1 was higher. Therefore, we focused on studying the SUMOylation modification of HSF1 primarily with SUMO1.

Due to the subsequent experiments requiring the use of plasmids and viruses, we further verified whether exogenous HSF1 protein could be modified by endogenous SUMO1. We observed multiple bands representing SUMO1-modified HSF1 once again (Fig. 1C), indicating that exogenous HSF1 protein can also be modified by endogenous SUMOylation enzymes.

We extracted the complete amino acid sequence of HSF1 from Uniprot (https://www.uniprot.org/uniprot/) to pinpoint the lysine residues that function as SUMO acceptor sites. Six possible SUMOylation sites were identified in the 529 amino acid sequence of HSF1, following the ψkxE SUMOylation consensus motif, with positions at 91, 126, 150, 162, 298, and 381 (Fig. 1D). Abgent SUMOplot (http://www.abgent.com/sumoplot/) and JASSA were used for analyzing potential SUMOylation sites in HSF1, with the findings aligning with our screening results (Fig. S1B and C). Combining both results, we ultimately selected positions 91, 126, and 298 to construct single mutant plasmids, mutating the lysine residues at these sites to arginine, and transfected them into U87 cells to test the in vivo SUMOylation capability of the mutants. Compared to the HSF1 WT plasmid, no changes in SUMO1 modification were observed with the K91R and K126R mutants. However, upon mutation of lysine K298 to arginine, SUMO1 modification was virtually abolished, thus identifying the lysine residue at position 298 as the primary SUMO1 acceptor site (Fig. 1E). Further comparison revealed that the corresponding lysine residue K298 in homologous proteins of different species is highly conserved (Fig. 1F). In summary, we proposed that HSF1 is a novel target for SUMO modification, with lysine residue 298 of HSF1 being preferentially modified by SUMO1.

Effect of SUMOylation modification on subcellular localization, stability, and trimer conformation of HSF1

To investigate the impact of SUMOylation on HSF1, we utilized lentiviral transfection to establish stable expression systems of Flag-tagged HSF1 WT and HSF1 K298R in U87 and U251 cells for subsequent studies. Prior to formal experiments, to eliminate the potential influence of protein expression levels on subsequent experimental results, we lysed the cells and performed Western blot analysis using anti-Flag antibodies. The results showed no statistically significant difference in expression levels between HSF1 WT and HSF1 K298R (Fig. S2A and B). Next, we continued to perform Western blot analysis using anti-HSF1 antibodies to further investigate the effects of different HSF1 protein expression levels on the biological behavior of tumor cells between the VECTOR and WT groups. The results indicated a significant difference in HSF1 protein levels between these two groups, and this difference was statistically significant. No statistically significant difference in the total amount of HSF1 protein was found between the WT and K298 groups, consistent with the conclusions obtained using anti-Flag antibodies (Fig. S2C and D).

We first investigated whether SUMOylation of HSF1 alters its subcellular localization. Immunofluorescence staining indicated the presence of HSF1 in both the cytoplasm and nucleus (Fig. 2A). In the nucleus, the levels of Flag-HSF1 WT were higher compared to Flag-HSF1 K298R, while the opposite was observed in the cytoplasm. Using a subcellular fractionation kit for separation and quantitative analysis of cytoplasmic and nuclear proteins, the results indicated statistically significant differences in protein levels between the cytoplasm and nucleus for each group of Flag-HSF1 (Fig. 2B and C). This suggested that SUMOylation of HSF1 promotes its directional translocation from the cytoplasm to the nucleus.Fig. 2 The effect of SUMOylation modification on the subcellular localization, stability, and trimer conformation of HSF1 (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article).

Fig 2(A) Immunofluorescence staining reveals the subcellular localization of HSF1 in the cytoplasm and nucleus. Flag antibody was used to stain exogenous HSF1 in stably transfected U87 cells (red), and the nucleus was stained with DAPI (blue). (B) Cytoplasmic and nuclear proteins from stably transfected U87 cells were separated using a kit, followed by Western blot analysis using anti-Flag antibody. (C) Quantitative analysis of Flag-HSF1 protein levels in the cytoplasm and nucleus. (D) Different transfected U87 cells were treated with the protein synthesis inhibitor cycloheximide (CHX), and at specified time points, protein extracts were subjected to Western blot analysis using Flag antibody. (E) Quantitative analysis of changes in Flag-HSF1 levels over time after cycloheximide (CHX) treatment in different transfected U87 cells. (F) U87 cells with different transfections were lysed, immunoprecipitated using anti-IgG and anti-Flag antibodies, and then subjected to Western blot analysis using anti-ubiquitination antibody. (G) Schematic representation of wild-type HSF1 binding. On the left is a 3D schematic, where each color represents each individual. On the right is a 2D representation, where green lines represent hydrogen bonds, red represents van der Waals forces, and others represent hydrophobic forces. (H) Schematic representation of mutant HSF1 binding. On the left is a 3D schematic, where each color represents each individual. On the right is a 2D representation, where green lines represent hydrogen bonds, red represents van der Waals forces, and others represent hydrophobic forces. Data are presented as mean ± SD. n = 3. '*', P < 0.05; '**', P < 0.01; 'ns', not significant.

To test whether SUMOylation regulates the stability of HSF1, we treated U87 cells stably expressing HSF1 WT and HSF1 K298R with cycloheximide to determine the half-life of these proteins. Interestingly, within 24 h of treatment, the protein levels of HSF1 K298R decreased significantly compared to HSF1 WT, indicating reduced stability. As time progressed, the difference between the two became more pronounced (Fig. 2D and E). To further elucidate whether the accelerated degradation of HSF1 K298R is related to ubiquitination, we performed ubiquitination immunoprecipitation assays. The findings indicated a notable rise in ubiquitination attaching to K298R HSF1, clarifying the reason for its accelerated breakdown rate (Fig. 2F). These findings suggested that SUMOylation at position K298 enhances the stability of HSF1 WT by reducing its ubiquitination and subsequent proteasomal degradation.

Three HSF1 monomers assemble to form a trimer, which then binds to target genes to exert transcription factor function. We used molecular docking techniques to examine whether the lysine residue at position 298 affects protein-protein interactions, namely the stability of the trimeric conformation (Figs. 2G and H, S2E and F). Molecular docking was performed using the HDOCK tool, generating a total of 10 binding conformations. We then calculated the docking scores and confidence levels of the top 10 conformations, with higher negative scores indicating tighter binding (relative values, no standard reference). Analysis of the selected conformations revealed that SUMOylation modification at the K298 position of HSF1 did not the stability of its trimeric conformation(Fig. S2G and H).

As a transcription factor, HSF1 needs to bind to target genes to exert its transcriptional activity. There are also some reports on how SUMOylation regulates the binding and dissociation of HSF1 with target genes. One study found that mutating the lysine at position 298 of HSF1 significantly decreases its stress-induced transcriptional activity in vivo [34], suggesting that SUMOylation modification may regulate HSF1′s transcriptional activity. Kmiecik et al. [35] analyzed the SUMOylation of HSF1 in vitro using purified components and found that HSF1 can be SUMOylated when bound to DNA. SUMOylation of HSF1 neither changes its DNA-binding affinity nor affects its subsequent dissociation from DNA, indicating that SUMOylation acts at the transcriptional level of HSF1. Additionally, they also observed that the trimerization of HSF1 and its phosphorylation at serines 303 and 307 increased the efficiency of SUMOylation, consistent with previous research results [36]. Therefore, the above research results suggest that SUMOylation modification may also regulate HSF1 by enhancing its transcriptional activity.

SUMOylation modification of HSF1 can enhance the UPRmt

Acute mitochondrial protein folding stress leads to a significant increase in mitochondrial chaperone proteins within the mitochondrial matrix [37,38], suggesting that these chaperone proteins play a crucial role in the UPRmt process. HSF1 is essential for the activation of mitochondrial chaperone protein genes, including HSP60 and mtHSP70. During various levels of UPRmt, HSF1 continuously binds to the promoters of mitochondrial chaperone protein genes, and its binding rate is significantly enhanced. Additionally, HSF1 supports the maintenance of mitochondrial function [39]. These findings indicate that HSF1 is involved in UPRmt by promoting the expression of relevant mitochondrial chaperone proteins and maintaining mitochondrial function through UPRmt.

To verify the above hypothesis and explore the specific mechanism by which SUMOylation of HSF1 enhances UPRmt, we lysed U87 cells and performed immunoblot analysis. Since mtHSP70 and HSP60 are downstream proteins of HSF1 and are also the main heat shock proteins in the mitochondrial unfolded protein response, we used anti-mtHSP70 and anti-HSP60 antibodies for the immunoblot analysis. The findings indicated that the SUMOylation alteration at the K298 site of HSF1 boosts its ability to transcription, resulting in elevated levels of downstream heat shock proteins (Fig. 3A-C). This indicated that SUMOylation modification of HSF1 enhances the UPRmt by increasing the expression of mtHSP70 and HSP60 in the mitochondria.Fig. 3 SUMOylation modification of HSF1 can enhance the UPR mt.

Fig 3(A) Lysis of U87 cells transfected with different viruses, followed by Western blot analysis using anti-mtHSP70 antibody and anti-HSP60 antibody to determine the impact of HSF1 SUMOylation on downstream mtHSP70 and HSP60 expression. (B-C) Quantitative analysis of the impact of HSF1 SUMOylation on downstream mtHSP70 and HSP60 expression. (D-E) Treatment of U87 and U251 cells transfected with different viruses using JC-1, followed by detection of aggregate and monomer fluorescence signals at specified wavelengths using ELISA reader, normalized to the VECTOR group. Panel D represents U87 cells, and Panel E represents U251 cells. (F) Treatment of U87 and U251 cells transfected with different viruses using JC-1, followed by observation of mitochondrial polymer fluorescence intensity (red) between groups using fluorescence microscopy. (G-H) Assessment of protein aggregates in U87 and U251 cells transfected with different viruses using a specialized assay kit. Protein aggregates inside the cells are stained in red, while the nucleus is stained in blue. Data are presented as mean ± SD. n = 3. '*', P < 0.05; '**', P < 0.01; 'ns', not significant.

MMP is a critical indicator of mitochondrial function. To further investigate the relationship between HSF1 SUMOylation and mitochondrial function, we used JC-1 staining to assess the impact of HSF1 SUMOylation on MMP. Our results demonstrated that mutations at the HSF1 K298 site induce a conversion of JC-1 aggregates to monomers (Fig. 3D-F). These findings suggested that SUMOylation modification of HSF1 enhances mitochondrial function in U87 and U251 cells.

Due to the vigorous metabolism and proliferation of tumor cells, a large amount of abnormal proteins is produced, leading to the formation of protein aggregates. The primary role of UPRmt is to clear protein aggregates and maintain the function of tumor cells. To investigate this, we used a protein aggregate detection kit to assess the presence of aggregates in the cells, as shown in the figure (Fig. 3G and H). A substantial amount of aggregates was found in normal U87 and U251 cells, which gradually decreased with increasing HSF1 expression. By comparing the two groups of HSF1 WT and HSF1 K298R, We can conclude that SUMOylation modification of HSF further enhances the function of UPRmt.

HSF1 SUMOylation regulated the malignant progress of U87 cells and U251 cells

MMP reduction serves as a potential signal for early cell apoptosis [40]. We have demonstrated that SUMOylation modification of HSF1 can inhibit MMP reduction in U87 and U251 cells. Further investigation into the effect of this modification on cell proliferation was conducted through plate colony formation assays, revealing that SUMOylation modification of HSF1 promotes proliferation in both U87 and U251 cell lines (Fig. 4A-D). We observed the proliferation of U87 and U251 cells in three groups over 72 h using the CCK8 assay, and constructed proliferation curves for quantitative analysis (Fig. S3A and B). The results were consistent with the plate colony formation assay. Comparing the VECTOR and WT groups, it was evident that in both U87 and U251 cells, the proliferation ability significantly increased with the higher expression of HSF1.Fig. 4 HSF1 SUMOylation regulated the malignant progress of U87 cells and U251 cells.

Fig 4(A, B) Plate cloning experiments to assess the impact of HSF1 SUMOylation on the proliferation capacity of U87 and U251 cells. (C, D) Quantitative analysis of the results of plate cloning experiments in U87 and U251 cells. (E, F) Quantitative analysis of scratch assay results in U87 and U251 cells. (G, H) Scratch assay to evaluate the impact of HSF1 SUMOylation on the migration ability of U87 and U251 cells. (I, J) Quantitative analysis of Transwell assay results. (K, L) Transwell assay to assess the impact of HSF1 SUMOylation on the migration and invasion abilities of U87 and U251 cells. Data are presented as mean ± SD. n = 3. '*', P < 0.05; '**', P < 0.01; 'ns', not significant

In the previous text, we demonstrated that HSF1 SUMOylation enhances UPRmt by increasing the expression of mtHSP70 and HSP60 proteins in cells. To verify that the proliferative effect of HSF1 SUMOylation is mediated through UPRmt by mtHSP70 and HSP60, we conducted experimental validation. U87 cells from the K298 group were co-transfected with mtHSP70 and HSP60 plasmids. After 24 h, cells were lysed, and immunoblot analysis was performed using anti-mtHSP70 and anti-HSP60 antibodies, comparing with the VECTOR and WT groups (Fig. S3C-E). The CCK8 assay showed that, compared to the WT group, the proliferation ability difference between the WT and K298R groups disappeared with the loss of differential expression of mtHSP70 and HSP60(Fig. S3F).

Additionally, we evaluated the role of SUMOylation modification of HSF1 in cell migration and invasion through scratch wound healing and Transwell assays. In both U87 and U251 cells, SUMOylation modification of HSF1 enhanced cell migration and invasion (Fig. 4E-L).

SUMOylation modification of HSF1 promoted the growth of glioblastoma in vivo

To evaluate how SUMOylation modification at the K298 site of the HSF1 protein affects cell growth in vivo, we utilized a xenograft model where U87 cells were injected subcutaneously into the right flank of nude mice (Fig. 5A). Tumor volume was measured every 7 days starting from day 7 post-injection. The results showed a significant increase in tumor growth upon SUMOylation modification of HSF1 (Fig. 5E). After 21 days post-injection, tumors were dissected, photographed (Fig. 5B and C), and weighed (Fig. 5D). Tumors generated from HSF1 K298R U87 cells were noticeably smaller and lighter compared to those generated from HSF1 WT U87 cells (Fig. 5D and E). Furthermore, staining with Ki67 and Bax revealed weaker proliferative activity and increased apoptosis in tumors derived from HSF1 K298R U87 cells (Fig. 5F-I). These results suggested that SUMOylation modification of HSF1 enhances cell growth in glioblastoma in vivo.Fig. 5 SUMOylation modification of HSF1 promoted the growth of glioblastoma in vivo.

Fig 5(A) Flowchart of the subcutaneous tumor experiment in mice. (B) Formation of subcutaneous tumors by U87 cells in nude mice at week 3 (n = 6/group). (C) After the third week post-implantation, mice were euthanized, and subcutaneous tumors were dissected and removed. (D) Measurement of tumor weight and quantitative analysis at the endpoint of the study. (E) Weekly recording of tumor growth curves and quantitative analysis after subcutaneous injection in nude mice. (F, G) Slicing of subcutaneous tumors from each group of mice, Ki67 staining to measure the proliferation ability of U87 cells in vivo, followed by quantitative analysis of experimental results. (H, I) Slicing of subcutaneous tumors from each group of mice, Bax staining to measure the apoptosis of U87 cells in vivo, followed by quantitative analysis of experimental results. Data are presented as mean ± SD. n = 6. '*', P < 0.05; '**', P < 0.01; 'ns', not significant.

Discussion

In this study, we show that endogenous HSF1 can undergo SUMOylation modifications by SUMO1, SUMO2, and SUMO3, while SUMO1 being the most common type in glioblastoma. Furthermore, through plasmid transfection, we found that exogenous HSF1 can also be modified by endogenous SUMOylation enzymes. This suggested that under conditions where endogenous SUMOylation enzymes are present, both endogenous and exogenous HSF1 can be SUMO-modified, indicating the physiological relevance of HSF1 SUMOylation. Mutating the lysine residue at position 298 to arginine in the amino acid sequence of HSF1 almost abolishes its SUMOylation, indicating that K298 is the major modification site of HSF1.

A variety of effects can be induced by SUMOylation on target proteins, including the regulation of their subcellular localization, stability, transcriptional activity, and interactions with other proteins [41]. To investigate the mechanism of action of SUMOylation on HSF1, we compared the subcellular localization, stability, and trimer conformation of HSF1 WT and HSF1 K298R in cells. Despite similar total levels of exogenous protein, their distribution in the cytoplasm and nucleus differed: HSF1 WT was more abundant in the nucleus, while HSF1 K298R showed higher levels in the cytoplasm. Compared to HSF1 WT, HSF1 K298R exhibited reduced stability and increased degradation rate, associated with enhanced binding to ubiquitin. Molecular docking analysis revealed that the mutation at the K298 site did not affect the stability of HSF1′s trimeric conformation.

We found that SUMOylation modification of HSF1 reduces the aggregation of misfolded proteins and improves mitochondrial function by enhancing UPRmt, which is achieved through increased expression of downstream mtHSP70 and HSP60. Finally, we confirmed that SUMOylation of HSF1 can enhance the migration and invasion of U87 and U251 cells, both in vitro and in vivo, and promote cell proliferation while inhibiting apoptosis.

The SUMOylation modification of HSF1 promotes its directed translocation into the nucleus, while ubiquitination modification facilitates the degradation of HSF1. Lately, there has been a thorough examination of the relationship between SUMOylation and ubiquitination. The ubiquitination of a variety of substrates is affected by SUMOylation, such as androgen receptor [41], phosphatase and tensin homolog [42], p53 [43], and the aberrations of these modifications in the substrates may lead to the occurrence and progression of cancer. Interestingly, the simultaneous action of SUMOylation and ubiquitination on the lysine residues of the substrate has demonstrated that SUMOylation at the identical site of the substrate can inhibit its ubiquitination. Under certain circumstances, SUMOylation of the substrate can enhance its later ubiquitination. SUMO-targeted ubiquitin ligases, like RNF4, have the ability to enhance the ubiquitination of SUMOylated targets [44]. The findings indicate that SUMOylation significantly impacts the ubiquitin–proteasome pathway. We speculate that subcellular localization may be another factor influencing the degradation of HSF1, as SUMOylation occurs in the nucleus, while ubiquitination mainly occurs in the cytoplasm [45]. Both HSF1 WT and HSF1 K298R have similar abilities to translocate from the cytoplasm to the nucleus. However, HSF1 WT undergoes SUMOylation modification in the nucleus, which restricts its export from the nucleus, while it has no effect on HSF1 K298R. This results in differential distribution of HSF1 WT and HSF1 K298R in the nucleus. Since HSF1 K298R is more abundant in the cytoplasm, it is more susceptible to ubiquitination modification and subsequent degradation.

Mammals have five domains within HSF1. Within the N-terminus lies the DNA-binding domain, abbreviated as DBD. After the oligomerization domain comes the leucine zipper repeats 1–3 (LZ1–3; also known as heptad repeats HR-A and HR-B), the regulatory domain (RD), and a fourth leucine zipper repeat (LZ4 or HR-C) domain. The C-terminus includes the transactivation domain, which interacts with the general transcription machinery to facilitate the release of paused RNA Pol II near the promoter and drive transcription elongation [46,47]. Trimerization of HSF1 is facilitated by hydrophobic repeat regions (HR-A, HR-B, and HR-C), also known as leucine zipper domains (LZ-1, LZ-2, LZ-3, LZ-4), which enable a parallel alignment of HSF1 monomers into active trimers [48]. From a structural analysis standpoint, K298 is situated within the RD, and its mutation does not impact trimer formation, aligning with our experimental findings.

The UPRmt is comprised of two protein quality control mechanisms: the protein folding apparatus, which includes HSP60 and mtHSP70 as the primary chaperone systems, and the proteolysis apparatus, which includes ClpP, LONP1, and numerous other proteases responsible for breaking down misfolded proteins [[49], [50], [51]]. Both HSP60 and mtHS70 play a crucial role in the folding of proteins within the mitochondria [49,50]. Research has consistently demonstrated that the activation of the UPRmt is essential for the growth and advancement of cancer [52]. Mitohormesis-induced UPRmt is crucial for promoting the invasion and metastasis of cancer cells [11]. In breast cancer patients, a strong correlation exists between elevated levels of UPRmt-related genes and decreased overall survival as well as metastasis-free survival [11]. Research indicating the increase of UPRmt elements in breast cancer implies that the activation of UPRmt plays a role in the advancement of breast cancer [53]. Furthermore, polymorphisms in genes encoding UPRmt components are associated with an increased risk of head and neck cancer [54]. Remarkably, under mitochondrial stress, the UPRmt pathway is responsible for releasing the mitokine GDF15, leading to the enhancement of thyroid cancer cell invasion [55]. HSP60 and mtHSP70 are protective proteins found in mitochondria and serve as the primary molecular chaperones activated by UPRmt. Studies have demonstrated that cancer progression relies on HSP60 and mtHSP70, aiding in the restructuring of proteins and breaking down aggregated proteins [56]. Additionally, there is growing evidence suggesting elevated levels of HSP60 and mtHSP70 in different types of tumors [[57], [58], [59]]. Silencing of HSP60 leads to reduced cell growth in pancreatic cancer [60], ovarian cancer [61], and glioblastoma [62,63]. Knocking down HSP60 has been shown to impede the growth of tumors in xenograft models of both pancreatic cancer [60] and glioblastoma [63]. Reducing HSP60 levels results in significant impairments in mitochondrial activities, ultimately impeding the growth and viability of cells. In glioblastoma, a decrease in HSP60 leads to an increase in ROS levels, subsequently triggering the AMPK pathway to suppress protein synthesis and impede cell growth [62]. Silencing of mtHSP70 suppresses the growth, movement, and infiltration of different types of cancer cells [64,65]. Additionally, reducing mtHSP70 results in decreased amounts of phosphorylated ERK1/2 and p‐c‐Raf [65], as well as the receptor tyrosine kinase RET, along with lower levels of the anti-apoptotic proteins Bcl-2, Bcl-xL, and Mcl-1 [64].

Given the promoting role of HSP60 and mtHSP70 in cancer development, we have reason to believe that a significant portion of UPRmt support for cancer is achieved through these two mitochondrial heat shock proteins. Moreover, as integral components of UPRmt, the expression levels of HSP60 and mtHSP70 can serve as indicators of UPRmt severity. The increase in HSP60 and mtHSP70 expression, as evidenced by JC-1 staining showing enhanced mitochondrial function and reduced aggregates, indicates a positive correlation between these heat shock proteins and UPRmt. Cell experiments have demonstrated that SUMOylation of HSF1 enhances the migration and invasion capabilities of U87 and U251 cells, with both in vitro and in vivo experiments confirming its role in promoting cell proliferation and inhibiting apoptosis. This effect is achieved indirectly by enhancing UPRmt through increased expression of HSP60 and mtHSP70. However, further research is needed to investigate the involvement of other factors.

Here, we have elucidated the mechanism of action of SUMOylation on HSF1, and by increasing the expression of HSP60 and mtHSP70 to enhance UPRmt and promote glioblastoma progression, similar to the HSF1 SUMOylation—UPRmt axis. However, as a transcription factor, we have not further studied whether the mutation at the K298 site affects its binding and dissociation with target genes, which is the focus of our future research. In summary, we believe that the novel regulatory mechanism of SUMOylation on HSF1 may become a new therapeutic target for glioblastoma.

Data availability statement

All data accessed from external sources and prior publications have been referenced in the text and corresponding figure legends. Additional data used and/or analyzed during the current study are available from the corresponding author on reasonable request.

Ethics statement

All animal experimental protocols were carried out according to the guidelines of The first affiliated hospital of Suzhou University for animal research and were approved by Ethic Committee (Project No. 82002643).

Funding

This work was supported by the 10.13039/501100001809 National Natural Science Foundation of China under Grant [82002643 ], Science and Technology Special Fund of Jiangsu Province (Key Research and Development Program for Social Development) [BE2023712] and Suzhou Medical Technology Innovation Project-Clinical Frontier [SKY2022002].

CRediT authorship contribution statement

Xiang Li: Writing – original draft, Formal analysis, Data curation, Conceptualization. Zongqi Wang: Funding acquisition, Formal analysis, Data curation, Conceptualization. Bixi Gao: Formal analysis, Data curation, Conceptualization. Kun Dai: Project administration, Methodology, Investigation. Jiang Wu: Project administration, Methodology, Investigation. Kecheng Shen: Project administration, Methodology, Investigation. Guangzhao Li: Project administration, Methodology, Investigation. Xiaowang Niu: Project administration, Methodology, Investigation. Xin Wu: Software, Resources. Longyuan Li: Software, Resources. Haitao Shen: Validation, Supervision. Haiying Li: Validation, Supervision. Zhengquan Yu: Supervision, Funding acquisition. Zhong Wang: Writing – review & editing, Supervision, Conceptualization. Gang Chen: Writing – review & editing, Supervision, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix Supplementary materials

Image, application 1

Image, application 2

Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.neo.2024.101055.
==== Refs
References

1 Louis D.N. Ohgaki H. Wiestler O.D. The 2007 WHO classification of tumours of the central nervous system Acta Neuropathol. 114 2 2007 97 109 17618441
2 Stupp R. Mason W.P. van den Bent M.J. Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma N. Engl. J. Med. 352 10 2005 987 996 15758009
3 Stupp R. Hegi M.E. Mason W.P. Effects of radiotherapy with concomitant and adjuvant temozolomide versus radiotherapy alone on survival in glioblastoma in a randomised phase III study: 5-year analysis of the EORTC-NCIC trial Lancet Oncol. 10 5 2009 459 466 19269895
4 Luo J. Solimini N.L. Elledge S.J. Principles of cancer therapy: oncogene and non-oncogene addiction Cell 136 5 2009 823 837 19269363
5 Vyas S. Zaganjor E. Haigis M.C. Mitochondria and cancer Cell 166 3 2016 555 566 27471965
6 Fiorese C.J. Schulz A.M. Lin Y.F. Rosin N. Pellegrino M.W. Haynes C.M. The transcription factor ATF5 mediates a mammalian mitochondrial UPR Curr. Biol. 26 15 2016 2037 2043 27426517
7 Martinus R.D. Garth G.P. Webster T.L. Selective induction of mitochondrial chaperones in response to loss of the mitochondrial genome Eur. J. Biochem. 240 1 1996 98 103 8797841
8 Kenny T.C. Gomez M.L. Mitohormesis G.D. UPR(mt), and the complexity of mitochondrial DNA landscapes in cancer Cancer Res. 79 24 2019 6057 6066 31484668
9 Naresh N.U. Haynes C.M. Signaling and regulation of the mitochondrial unfolded protein response Cold Spring. Harb. Perspect. Biol. 11 6 2019 a033944
10 Tian Y. Garcia G. Bian Q. Mitochondrial stress induces chromatin reorganization to promote longevity and UPR(mt) Cell 165 5 2016 1197 1208 27133166
11 Kenny T.C. Craig A.J. Villanueva A. Germain D. Mitohormesis primes tumor invasion and metastasis Cell Rep. 27 8 2019 2292 2303 e6 31116976
12 Haynes C.M. Fiorese C.J. Lin Y.F. Evaluating and responding to mitochondrial dysfunction: the mitochondrial unfolded-protein response and beyond Trends Cell Biol. 23 7 2013 311 318 23489877
13 Kenny T.C. Manfredi G. Germain D. The mitochondrial unfolded protein response as a non-oncogene addiction to support adaptation to stress during transformation in cancer and beyond Front. Oncol. 7 2017 159 28798902
14 Inigo J.R. Kumar R. Chandra D. Targeting the mitochondrial unfolded protein response in cancer: opportunities and challenges Trends Cancer 7 12 2021 1050 1053 34580036
15 Shin C.S. Meng S. Garbis S.D. LONP1 and mtHSP70 cooperate to promote mitochondrial protein folding Nat. Commun. 12 1 2021 265 33431889
16 Alexiou G.A. Karamoutsios A. Lallas G. Expression of heat shock proteins in brain tumors Turk. Neurosurg. 24 5 2014 745 749 25269047
17 Beaman G.M. Dennison S.R. Chatfield L.K. Phoenix D.A. Reliability of HSP70 (HSPA) expression as a prognostic marker in glioma Mol. Cell Biochem. 393 1-2 2014 301 307 24833463
18 Gomez-Pastor R. Burchfiel E.T. Thiele D.J. Regulation of heat shock transcription factors and their roles in physiology and disease Nat. Rev. Mol. Cell Biol. 19 1 2018 4 19 28852220
19 Barna J. Csermely P. Vellai T. Roles of heat shock factor 1 beyond the heat shock response Cell Mol. Life Sci. 75 16 2018 2897 2916 29774376
20 Li J. Labbadia J. Morimoto R.I. Rethinking HSF1 in stress, development, and organismal health Trends Cell Biol. 27 12 2017 895 905 28890254
21 Seeler J.S. Dejean A. SUMO and the robustness of cancer Nat. Rev. Cancer 17 3 2017 184 197 28134258
22 Gareau J.R. Lima C.D. The SUMO pathway: emerging mechanisms that shape specificity, conjugation and recognition Nat. Rev. Mol. Cell Biol. 11 12 2010 861 871 21102611
23 Yang W. Wang L. Roehn G. Small ubiquitin-like modifier 1-3 conjugation [corrected] is activated in human astrocytic brain tumors and is required for glioblastoma cell survival Cancer Sci. 104 1 2013 70 77 23078246
24 Bernstock J.D. Ye D. Gessler F.A. Topotecan is a potent inhibitor of SUMOylation in glioblastoma multiforme and alters both cellular replication and metabolic programming Sci. Rep. 7 1 2017 7425 28785061
25 Martin S. Wilkinson K.A. Nishimune A. Henley J.M. Emerging extranuclear roles of protein SUMOylation in neuronal function and dysfunction Nat. Rev. Neurosci. 8 12 2007 948 959 17987030
26 Jumper J. Evans R. Pritzel A. Highly accurate protein structure prediction with AlphaFold Nature 596 7873 2021 583 589 34265844
27 Anandakrishnan R. Aguilar B. Onufriev A.V. H++ 3.0: automating pK prediction and the preparation of biomolecular structures for atomistic molecular modeling and simulations Nucleic. Acids. Res. 40 Web Server issue 2012 W537 W541 22570416
28 Pettersen E.F. Goddard T.D. Huang C.C. UCSF Chimera–a visualization system for exploratory research and analysis J. Comput. Chem. 25 13 2004 1605 1612 15264254
29 Yan Y. Zhang D. Zhou P. Li B. Huang S.Y. HDOCK: a web server for protein-protein and protein-DNA/RNA docking based on a hybrid strategy Nucleic. Acids. Res. 45 W1 2017 W365 W373 28521030
30 Mooers B. Shortcuts for faster image creation in PyMOL Protein Sci. 29 1 2020 268 276 31710740
31 Laskowski R.A. Swindells M.B. LigPlot+: multiple ligand-protein interaction diagrams for drug discovery J. Chem. Inf. Model. 51 10 2011 2778 2786 21919503
32 Teodoro J.S. Palmeira C.M. Rolo A.P. Mitochondrial membrane potential (ΔΨ) fluctuations associated with the metabolic states of mitochondria Methods Mol. Biol. 1782 2018 109 119 29850996
33 Barysch S.V. Dittner C. Flotho A. Becker J. Melchior F. Identification and analysis of endogenous SUMO1 and SUMO2/3 targets in mammalian cells and tissues using monoclonal antibodies Nat. Protoc. 9 4 2014 896 909 24651501
34 Hong Y. Rogers R. Matunis M.J. Regulation of heat shock transcription factor 1 by stress-induced SUMO-1 modification J. Biol. Chem. 276 43 2001 40263 40267 11514557
35 Kmiecik S.W. Drzewicka K. Melchior F. Mayer M.P. Heat shock transcription factor 1 is SUMOylated in the activated trimeric state J. Biol. Chem. 296 2021 100324
36 Hietakangas V. Ahlskog J.K. Jakobsson A.M. Phosphorylation of serine 303 is a prerequisite for the stress-inducible SUMO modification of heat shock factor 1 Mol. Cell Biol. 23 8 2003 2953 2968 12665592
37 Münch C. Harper J.W. Mitochondrial unfolded protein response controls matrix pre-RNA processing and translation Nature 534 7609 2016 710 713 27350246
38 Sutandy F. Gößner I. Tascher G. Münch C. A cytosolic surveillance mechanism activates the mitochondrial UPR Nature 618 7966 2023 849 854 37286597
39 Katiyar A. Fujimoto M. Tan K. HSF1 is required for induction of mitochondrial chaperones during the mitochondrial unfolded protein response FEBS Open. Bio 10 6 2020 1135 1148
40 Zhu C. Zhang C. Cui X. Wu J. Cui Z. Shen X. Trichosanthin inhibits cervical cancer by regulating oxidative stress-induced apoptosis Bioengineered 12 1 2021 2779 2790 34233587
41 Yang N. Liu S. Qin T. SUMO3 modification by PIAS1 modulates androgen receptor cellular distribution and stability Cell Commun. Signal. 17 1 2019 153 31752909
42 Bawa-Khalfe T. Yang F.M. Ritho J. Lin H.K. Cheng J. Yeh E.T. SENP1 regulates PTEN stability to dictate prostate cancer development Oncotarget 8 11 2017 17651 17664 27852060
43 Ashikari D. Takayama K. Tanaka T. Androgen induces G3BP2 and SUMO-mediated p53 nuclear export in prostate cancer Oncogene 36 45 2017 6272 6281 28692047
44 Kumar R. Sabapathy K. RNF4-A paradigm for SUMOylation-mediated ubiquitination Proteomics 19 21-22 2019 e1900185
45 Ergünay T. Ayhan Ö. Celen A.B. Sumoylation of Cas9 at lysine 848 regulates protein stability and DNA binding Life Sci. Alliance 5 4 2022 e202101078
46 Neudegger T. Verghese J. Hayer-Hartl M. Hartl F.U. Bracher A. Structure of human heat-shock transcription factor 1 in complex with DNA Nat. Struct. Mol. Biol. 23 2 2016 140 146 26727489
47 Vihervaara A. Mahat D.B. Guertin M.J. Transcriptional response to stress is pre-wired by promoter and enhancer architecture Nat. Commun. 8 1 2017 255 28811569
48 Jaeger A.M. Makley L.N. Gestwicki J.E. Thiele D.J. Genomic heat shock element sequences drive cooperative human heat shock factor 1 DNA binding and selectivity J. Biol. Chem. 289 44 2014 30459 30469 25204655
49 Bukau B. Horwich A.L. The Hsp70 and Hsp60 chaperone machines Cell 92 3 1998 351 366 9476895
50 Wiedemann N. Pfanner N. Mitochondrial machineries for protein import and assembly Annu. Rev. Biochem. 86 2017 685 714 28301740
51 Pryde K.R. Taanman J.W. Schapira A.H. A LON-ClpP proteolytic axis degrades complex I to extinguish ROS production in depolarized mitochondria Cell Rep. 17 10 2016 2522 2531 27926857
52 Wang S.F. Chen S. Tseng L.M. Lee H.C. Role of the mitochondrial stress response in human cancer progression Exp. Biol. Med. (Maywood) 245 10 2020 861 878 32326760
53 Chen F.M. Huang L.J. Ou-Yang F. Kan J.Y. Kao L.C. Hou M.F. Activation of mitochondrial unfolded protein response is associated with Her2-overexpression breast cancer Breast. Cancer Res. Treat. 183 1 2020 61 70 32601970
54 Ahmed M.W. Mahjabeen I. Gul S. Khursheed A. Mehmood A. Kayani M.A. Relationship of single nucleotide polymorphisms and haplotype interaction of mitochondrial unfolded protein response pathway genes with head and neck cancer Future Oncol. 15 33 2019 3819 3829 31651195
55 Kang Y.E. Kim J.M. Lim M.A. Growth differentiation factor 15 is a cancer cell-induced mitokine that primes thyroid cancer cells for invasiveness Thyroid 31 5 2021 772 786 33256569
56 Czarnecka A.M. Campanella C. Zummo G. Cappello F. Mitochondrial chaperones in cancer: from molecular biology to clinical diagnostics Cancer Biol. Ther. 5 7 2006 714 720 16861898
57 Castilla C. Congregado B. Conde J.M. Immunohistochemical expression of Hsp60 correlates with tumor progression and hormone resistance in prostate cancer Urology 76 4 2010 1017 e1-6
58 Na Y. Kaul S.C. Ryu J. Stress chaperone mortalin contributes to epithelial-mesenchymal transition and cancer metastasis Cancer Res. 76 9 2016 2754 2765 26960973
59 Yi X. Luk J.M. Lee N.P. Association of mortalin (HSPA9) with liver cancer metastasis and prediction for early tumor recurrence Mol. Cell Proteomics 7 2 2008 315 325 17934217
60 Zhou C. Sun H. Zheng C. Oncogenic HSP60 regulates mitochondrial oxidative phosphorylation to support Erk1/2 activation during pancreatic cancer cell growth Cell Death. Dis. 9 2 2018 161 29415987
61 Guo J. Li X. Zhang W. HSP60-regulated mitochondrial proteostasis and protein translation promote tumor growth of ovarian cancer Sci. Rep. 9 1 2019 12628 31477750
62 Tang H. Li J. Liu X. Wang G. Luo M. Deng H. Down-regulation of HSP60 suppresses the proliferation of glioblastoma cells via the ROS/AMPK/mTOR pathway Sci. Rep. 6 2016 28388 27325206
63 Ghosh J.C. Siegelin M.D. Dohi T. Altieri D.C. Heat shock protein 60 regulation of the mitochondrial permeability transition pore in tumor cells Cancer Res. 70 22 2010 8988 8993 20978188
64 Starenki D. Hong S.K. Lloyd R.V. Park J.I. Mortalin (GRP75/HSPA9) upregulation promotes survival and proliferation of medullary thyroid carcinoma cells Oncogene 34 35 2015 4624 4634 25435367
65 Hu Y. Yang L. Yang Y. Oncogenic role of mortalin contributes to ovarian tumorigenesis by activating the MAPK-ERK pathway J. Cell Mol. Med. 20 11 2016 2111 2121 27374312
