Cells increase their size in response to chronic stress.

(A) Flow cytometric quantification of cell viability under chronic HS for 7 days (HS = 39°C for HEK and HCT116 cells; HS = 40 °C for A549 and RPE1 cells) (n = 4 biologically independent samples). (B) Flow cytometric quantification of cell size after 7 days of chronic HS (biologically independent samples: n = 6 for HEK and A549; n = 5 for RPE1; n = 4 for HCT116). (C and D) Flow cytometric analysis of cell cycle (n = 3 biologically independent samples). (E) Flow cytometric quantification of cell size during different time intervals of chronic HS (n = 4 biologically independent samples). (F) Proliferation of HEK and A549 cells at the indicated temperature for the indicated period presented as cell numbers. Cells were seeded at a density of 5 x 106 and 3 x 106 per 15 cm plate for HEK and A549 cells, respectively. The numbers of live cells counted after 7 days are plotted (n = 5 biologically independent experiments). (G) Proliferation of A549 and RPE1 cells measured with a crystal violet assay (n = 3 biologically independent experiments). The adapted cells were maintained at 40 °C for one week before this experimental start point and continued at 40 °C during the experiment. See scheme of the experiment on the right. (H) Flow cytometric quantification of cell size in chronic HS and recovery (n = 3 biologically independent samples). (I) Flow cytometric analysis of cell cycle in chronic HS and post HS recovery (n = 3 biologically independent samples). The data are represented as mean values ± SEM for all bar and line graphs. The statistical significance between the groups was analyzed by two-tailed unpaired Student’s t-tests.

Cells are unable to adapt to chronic stress in the absence of one of the cytosolic Hsp90 isoforms.

(A) Immunoblots of Hsp90α and Hsp90β in WT HEK and A549 cells, and their respective Hsp90α/β KO cells. GAPDH serves as the loading control (α KO; Hsp90α KO and β KO; Hsp90β KO) (representative images of n = 4 biologically independent experiments). (B) Flow cytometric quantification of cell viability of HEK, A549, and their respective Hsp90α/β KO cells in chronic HS at different time points during a period of 7 days (n = 5 biologically independent samples). Note that the X axis does not have a linear scale and that lines connecting the data points are drawn as a visual aid. (C) Flow cytometric quantification of cell size in chronic HS at different time points during a period of 7 days (HS = 39 °C for HEK and 40 °C for A549) (n = 4 biologically independent samples) The data are represented as mean values ± SEM for all bar graphs. The statistical significance between the groups was analyzed by two-tailed unpaired Student’s t-tests. (D) Fluorescence microscopy images of A549 WT and Hsp90α/β KO cells after 4 days of chronic HS. The cytoskeleton is stained with phalloidin-Alexa488 (green), and the nucleus is stained with DAPI (blue). Images were captured with a fluorescence microscope (Zeiss, Germany). The scale bars on the images on the far right are 50 μM (images are representative of n = 2 biologically independent experiments).

Hsf1 regulates cell size in response to stress.

(A) Fold change of Hsf1 activity of HEK WT, A549 WT, and their respective Hsp90α/β KO cells at 37 °C as measured by luciferase reporter assay (n = 3 biologically independent samples and 2 experimental replicates each time). (B) Immunoblots of Hsf1 in the cytosolic and nuclear fractions of HEK WT and Hsp90α/β KO cells (α KO, Hsp90αKO; β KO, Hsp90βKO). GAPDH and lamin B1 serve as loading controls (representative blots of n = 2 biologically independent experiments). (C) Fold change of Hsf1 activity of HEK WT, A549 WT, and their respective Hsp90α/β KO cells in chronic HS as measured by luciferase reporter assay (n = 3 biologically independent samples, and 2 experimental replicates each time for HEK; n = 3 biologically independent samples, and 4 experimental replicates each time for A549). (D) Volcano plots of the normalized fold changes in protein levels of some core Hsf1 target genes (list obtained from https://hsf1base.org/) in chronic HS, determined by quantitative label-free proteomic analysis of Hsp90α/β KO and WT HEK cells. Molecular chaperones, whose expression is regulated by Hsf1, are excluded from this dataset. Each genotype was compared with its respective 37 °C control (n = 3 biologically independent samples). Log2 fold changes of > 0.5 or < -0.5 with a p-value of < 0.05 were considered significant differences for a particular protein. (E) Flow cytometric quantification of cell size of HEK, A549, and RPE1 cells upon overexpression of WT Hsf1 (with plasmid pcDNA-Flag HSF1 wt) or mutant Hsf1 (pcDNA-Flag HSF1 C205 (Kijima et al., 2018), retaining only the first 205 amino acids), and with plasmid pcDNA3.1(+) as empty vector control. Transfected cells to be measured were identified on the basis of their coexpression of EGFP (n = 4 biologically independent experiments). (F) Flow cytometric quantification of cell size in chronic HS after knockdown of Hsf1 in A549 WT and Hsp90αKO cells. Here the chronic HS for A549 cells is at 39 °C instead of 40 °C to reduce HS-induced damage in Hsf1 knockdown conditions (n = 3 biologically independent samples). (G) Flow cytometric quantification of cell cycle in HS after knockdown of Hsf1 in A549 WT and Hsp90αKO cells. The data are represented as mean values ± SEM for all bar graphs. The statistical significance between the groups was analyzed by two-tailed unpaired Student’s t-tests.

Hsp90α/β KO cells maintain chaperones, co-chaperones, and Hsp90 interactors during chronic stress adaptation.

(A) Volcano plots of the normalized fold changes of molecular chaperones and co-chaperones of cells subjected to 1 and 4 days of chronic HS determined by quantitative label-free proteomic analysis of Hsp90α/β KO and WT HEK cells. Each genotype was compared with its respective 37 °C control (n = 3 biologically independent samples). (B) Immunoblots of different molecular chaperones in HEK WT and Hsp90α/β KO cells (α KO, Hsp90αKO; β KO, Hsp90βKO). GAPDH serves as the loading control for all three panels (representative of n = 2 independent experiments). (C) Volcano plots of the normalized fold changes of Hsp90 interactors (list obtained from https://www.picard.ch/Hsp90Int) in 4 days of chronic HS determined by quantitative label-free proteomic analysis of Hsp90α/β KO and WT HEK cells. Each genotype was compared with its respective 37 °C control (n = 3 biologically independent samples). For all volcano plots, Log2 fold changes of > 0.5 or < -0.5 with a p-value of < 0.05 were considered significant differences for a particular protein. (D and E) In vivo refolding of heat-denatured luciferase of control cells (blue line) and cells heat-adapted to 39 °C (orange line). Luciferase activity before the acute HS (at 43 °C) is set to 100% (n = 3 biologically independent samples). See scheme of the experiment below. Note the different scales of the Y axes of the bar graphs in panel E. The data are represented as mean values ± SEM for all bar graphs. The statistical significance between the groups was analyzed by two-tailed unpaired Student’s t-tests. The p-values for Hsp90α and Hsp90β KO cells are in blue and red, respectively. All the p-values are for comparisons to the respective WT.

Hsp90α/β KO cells suffer from cytoplasmic protein dilution during adaptation to chronic stress.

(A) FRAP experiments with control and heat-adapted live cells expressing EGFP. The respective box plots show the t-half values of recovery of EGFP fluorescence and the apparent EGFP diffusion coefficients (n= 10 cells from 2 biologically independent experiments). (B) Fold change of cell size (represented by the FSC-MFI values) and total proteins (determined as MFI-FL1 values) in chronic HS as analyzed by flow cytometry. Cells were fixed, and total proteins were stained using Alexa Fluor 488 NHS ester (n = 3 biologically independent experiments). Lines connecting the data points are drawn as a visual aid. (C) Volcano plots of the normalized fold changes of the ribosomal proteins (list obtained from http://ribosome.med.miyazaki-u.ac.jp/) after 4 days of chronic HS determined by quantitative label-free proteomic analysis in Hsp90α/β KO and WT HEK cells. Each genotype was compared with its respective 37 °C control (n= 3 biologically independent samples). Log2 fold changes of > 0.6 or < -0.6 with a p-value of < 0.05 were considered significant differences for a particular protein. The box below the volcano plot shows the corresponding names of the proteins that were significantly downregulated. (D) Flow cytometric analysis of total translation of HEK WT and Hsp90α/β KO cells at 37 °C and after 4 days of chronic HS (see scheme of the experiment on the top). Nascent polypeptide chains were labeled with OP-puromycin during cell culture, and the incorporation of puromycin at different time points was analyzed (n = 4 experimental samples). (E) Representative polysome profiles of HEK WT cells at 37 °C and after 4 days of chronic HS (representative of n = 2 biologically independent experiments). (F and G) Immunoblots of some of the translation-related proteins. GAPDH and β-actin serve as loading controls (images are representative of n = 2 independent biological samples). (H) Relative fold changes of total translation and cell size in the early phase of adaptation to chronic HS (see schemes of experiments on the right) for A549 WT and Hsp90α/β KO cells. The data are represented as mean values ± SEM for all bar graphs. The statistical significance between the groups was analyzed by two-tailed unpaired Student’s t-tests.

Hsp90 is crucial for cellular proteostasis during adaptation to chronic stress.

(A) Flow cytometric determination of the in vivo UPS activity in chronic HS compared to 37 °C, using the Ub-M-GFP and Ub-R-GFP reporter proteins (n = 4 biologically independent samples). (B) Flow cytometric measurement of autophagic flux in chronic HS compared to 37 °C, using a mCherry-GFP-LC3 reporter. Flux is calculated as the ratio of the mean fluorescence intensities of mCherry and GFP-positive cells (n = 4 biologically independent samples). (C) In vitro steady-state proteasomal activity with lysates of HEK WT, and Hsp90α and Hsp90β KO cells determined by measuring fluorescence of the cleaved substrate suc-LLVY-AMC (n = 2 biologically independent samples). (D) Fluorescence micrographs of cells expressing the fusion protein EGFP-Q74 visible as aggregates with green fluorescence. The scale bar in micrographs indicates 10 μm (images are representative of n = 2 independent biological samples). The data are represented as mean values ± SEM for all bar graphs. The statistical significance between the groups was analyzed by two-tailed unpaired Student’s t-tests.

Enlarged cells are more resistant to additional stress.

(A) Scheme of cell size enlargement or reduction experiments. CHX, cycloheximide; CDKi, CDK4/6 inhibitor. (B) Cell size was first enlarged by treating cells with 100 nM CDKi for 3 days; then, cells were washed and subjected to chronic HS at 40 °C for 3 more days (CDKi > HS). Cell size (% FSC-MFI; grey part of the bars) and cell death (% annexin V and PI-positive; red part of the bars) were measured by flow cytometry. The values for cell size and death in the different experimental conditions are normalized to the respective 37 °C controls (n = 3 biologically independent experiments). (C) Cells were first pretreated with 7.5 nM rapamycin (Rapa) for 3 days to reduce the cell size. After that, the cells were subjected to chronic HS at 40 °C for 3 days (Rapa > HS). HS > Rapa, the two treatments were done the other way around. The cell size (% FSC-MFI) and relative cell death (% annexin V and PI-positive) were quantified by flow cytometry. The values for cell size and death in different experimental conditions are normalized to the respective 37 °C control (n = 3 biologically independent experiments). (D) Scheme of experiments aimed at determining impact of limiting physical space on cell size increase. (E and F) Phase-contrast micrographs of RPE1 cells seeded in different numbers to restrict the space for cell size increase during adaptation to chronic HS (representative images of n = 4 biologically independent experiments). The cell size (% FSC-MFI) and relative cell death (% annexin V-PI positive) are quantified by flow cytometry. For the bar graphs, the values for cell size and death in different conditions are normalized to the low density (dns) cell population at 37 °C day 0 (n = 4 biologically independent experiments). The data are represented as mean values ± SEM for all bar graphs. (G) Immunoblots in the lower panels show the endogenous Hsp90α and Hsp90β, and the exogenously overexpressed larger fusion proteins of Hsp90α (as mCherry-Hsp90α) and Hsp90β (as EGFP-Hsp90β), with red boxes highlighting samples with exogenous Hsp90. Images of the Coomassie-stained gels in the upper panels show the corresponding levels of total proteins with red boxes indicating lanes for samples from cells subjected to chronic HS (representative images of n = 2 biologically independent experiments). (H) Schematic representation of impact of chronic mild stress on cells. Wild-type cells initially adapt by enlarging their size and increasing total protein to maintain a minimum threshold level of functional proteins. The right part of the scheme (surrounded by a stippled box), shows what may happen if stress persists for much longer: cell size enlargement and translation are uncoupled, and because of protein damage, which continues to accumulate, cells become senescent and/or die.

Cells increase their size in response to different types of chronic stress.

(A) Flow cytometric quantification of cell viability after 4 days of 10 μM sodium arsenite (Ars), 1% hypoxia (Hypo), or L-azetidine-2-carboxylic acid (AZC) treatment (5 μM) (n = 3 biologically independent samples for Ars, Hypo, and AZC, respectively, for all cell lines). (B) Flow cytometric quantification of cell size after 4 days of treatment with 10 μM sodium arsenite (Ars), 1% hypoxia (Hypo), or 5 µM L-azetidine-2-carboxylic acid (AZC) (n = 4, 3, and 4 for Ars, Hypo, and AZC, respectively for HEK; for A549, n = 3, 3, and 4 for Ars, Hypo, and AZC, respectively; for RPE1, n = 3, 3, and 3 biologically independent samples for Ars, Hypo, and AZC, respectively). (C) Cell diameters of cells subjected to 7 days of mild HS, measured with an automated cell counter; the bar graph on the left shows absolute values whereas the one on the right shows the same data as % changes relative to cells left at control conditions (samples C); n = 4 samples. (D) Scanned images of plates with crystal violet-stained cells (representative images of n = 3 independent experiments). (E) Flow cytometric analysis of cell cycle after 1 week in chronic HS and post HS recovery (n = 3 biologically independent samples). The data are represented as mean values ± SEM for all bar and line graphs. The statistical significance between the groups was analyzed by two-tailed unpaired Student’s t-tests.

Cells are unable to adapt to chronic stress in the absence of one of the Hsp90 isoforms.

(A) Flow cytometric quantification of cell death of A549 WT and Hsp90α/β KO cells after 4 days in 1% hypoxia (Hypo) and 10 μM sodium arsenite (Ars) treatment (n = 3 biologically independent samples). (B) Flow cytometric quantification of cell size after 4 days in 1% hypoxia (Hypo) and 10 μM sodium arsenite (Ars) treatment (n = 4 biologically independent samples). (C) Fluorescence microscopy images of HEK WT and Hsp90α/β KO cells after 4 days of chronic HS. The cytoskeleton is stained with Phalloidin-Alexa488 (green), and the nucleus is stained with DAPI (blue). The scale bar in the top right micrograph is 50 μM (images are representative of n = 2 biologically independent experiments). The data are represented as mean values ± SEM for all bar graphs. The statistical significance between the groups was analyzed by two-tailed unpaired Student’s t-tests.

Hsf1 induces cell size in response to stress.

(A) Volcano plots of the normalized fold changes in protein levels of some of the core Hsf1 target genes (list obtained from https://hsf1base.org/) in Hsp90α/β KO cells compared to WT HEK as determined by quantitative label-free proteomic analysis. Molecular chaperones, whose expression is regulated by Hsf1, are excluded from this dataset (n = 3 biologically independent samples). Log2 fold changes of > 0.5 or < -0.5 with a p-value of < 0.05 were considered significant differences for a particular protein. (B) Fold change of Hsf1 activity of HEK, A549, and RPE1 cells upon overexpressing WT and mutant Hsf1 in combination with EGFP, as measured with the Hsf1 luciferase reporter. Control is transfected with only Hsf1 reporter plasmid and pEGFP-C1, those are common to all the experimental conditions; (n = 3 biologically independent samples). (C) Fold change of Hsf1 activity in A549 WT, Hsp90α KO, and Hsp90β KO cells after 4 days of capsaicin treatment as measured by luciferase reporter assay (n = 3 biologically independent samples). (D) Flow cytometric quantification of cell size after 4 days of capsaicin treatment of Hsp90α/β KO and WT A549 cells (n = 3 biologically independent samples). (E) Immunoblots of Hsf1 after Hsf1 knockdown in A549 WT and Hsp90α KO cells. β-actin serves as the loading control (representative of n = 2 independent experiments). (F) Fold change of Hsf1 activity in A549 WT and Hsp90α KO cells in chronic HS after Hsf1 knockdown as measured by luciferase reporter assays. Here the chronic HS for A549 cells is 39 °C to instead of 40 °C to reduce HS-induced damage in Hsf1 knockdown conditions (n = 3 biologically independent samples). (G) Flow cytometric quantification of cell size of mouse fibroblast. (90αKO, 90β HET), homozygous hsp90α KO, heterozygous hsp90β KO cells (n = 6 biologically independent samples). (H) Fold change of Hsf1 activity in (90αKO, 90β HET) MAFs compared to WT at 37 °C, as measured by luciferase reporter assays (n = 3 biologically independent samples). (I) Fold change of Hsf1 activity in MAFs subjected to chronic HS (orange bars) compared to 37 °C (blue bars), as measured by luciferase reporter assay (n = 3 biologically independent samples). (J) Flow cytometric quantification of cell size of MAFs subjected to chronic HS (orange bars) by comparison to 37 °C (blue bars) (n = 4 biologically independent samples). The data are represented as mean values ± SEM for all bar graphs. Where indicated, the statistical significance between the groups was analyzed by two-tailed unpaired Student’s t-tests.

Hsp90α/β KO cells maintain total proteins and Hsp90 interactors in chronic stress.

(A) Immunoblots of some molecular chaperones of A549 WT and Hsp90α/β KO cells (α KO, Hsp90αKO; β KO, Hsp90βKO). GAPDH and the Ponceau S-stained nitrocellulose filter serve as loading controls. (B) Volcano plots of the normalized fold changes of total proteins of cells subjected to chronic HS for 1 and 4 days determined by quantitative label-free proteomic analysis of Hsp90α/β KO and WT HEK cells. Each genotype was compared with its respective 37 °C control (n = 3 biologically independent samples) (C) Volcano plots of the normalized fold changes of Hsp90 interactors (list obtained from https://www.picard.ch/Hsp90Int) in Hsp90α/β KO HEK cells compared to WT as determined by quantitative label-free proteomic analysis (n = 3 biologically independent samples). (D) Volcano plots of the normalized fold changes of Hsp90 interactors in cells subjected to 1 day of chronic HS determined by quantitative label-free proteomic analysis. Each genotype was compared with its respective 37 °C control (n = 3 biologically independent samples). For all volcano plots Log2 fold changes of > 0.5 or < -0.5 with a p-value of < 0.05 were considered significant differences for a particular protein.

Cells maintain cytoplasmic density and total protein ratio during stress-induced cell size increase.

(A) Scheme of the FRAP experiments. (B to D) FRAP experiments with live cells expressing EGFP. The respective box plots represent the t-half values of recovery of EGFP fluorescence and the apparent EGFP diffusion coefficients (n= 10 cells from 2 biologically independent experiments). (E) Protein amount in cell lysates (5,000 cells lysed in 200 μl lysis buffer) of control cells and cells adapted to chronic HS as measured by Bradford assay. (F) Flow cytometric analysis of global translation of HEK WT and Hsp90α/β KO cells at 37 °C and after 1 day under chronic HS. See scheme of the experiment on the top. Nascent polypeptide chains were labeled with OP-puromycin during cell culture, and the incorporation of puromycin at different time points was analyzed (n = 4 experimental samples). (G) Total translation of HEK WT cells during a 4 h time span during different phases of chronic HS. (H) Immunoblot analysis of global translation as indicated by incorporation of puromycin into nascent polypeptides during the first 4 h of shifting cells to chronic HS conditions, compared to cells remaining at 37 °C. The 0 h time point of puromycin labelling serves as a negative control, and the Ponceau S-stained nitrocellulose filter as loading control (representative images from n = 2 biologically independent experiments) (I) Immunoblots of some of the translation-related proteins in A549 cells. β-actin serves as the loading control.

Hsp90α/β KO cells maintain WT levels of protein degradation activities in unstressed conditions.

(A) Flow cytometric determination of the in vivo UPS activity using the Ub-M-GFP and Ub-R-GFP reporter plasmids (n = 4 biologically independent samples). (B) Flow cytometric measurement of autophagic flux using a mCherry-GFP-LC3 reporter. Flux is calculated as the ratio of the mean fluorescence intensities of mCherry and GFP-positive cells (n = 4 biologically independent samples). (C) Quantification of the number of EGFP-Q74 aggregates from 6 representative micrographs such as the ones of Fig. 6D. For each bar data obtained from six representative macrographs. The data are represented as mean values ± SEM for all bar graphs, standardized to the respective WT values set to 1. The statistical significance between the groups was analyzed by two-tailed unpaired Student’s t-tests.

Smaller cells are more susceptible to additional stress.

(A) Cell size was enlarged by treating cells with 100 nM CDKi for 3 days. Fold change of cell size (represented by the FSC-MFI values) and total proteins (determined as MFI-FL1 values) were analyzed by flow cytometry. Cells were fixed, and total proteins were stained using Alexa Fluor 488 NHS ester (n = 3 biologically independent experiments). (B) Coomassie-stained protein gel showing protein samples from equal numbers of cells treated or not with CDKi (representative images of n = 2 biologically independent experiments). (C) Cell size was enlarged by treating cells with CDKi for 3 days, before cells were treated with sodium arsenite (40 μM) for 3 days. Cell size (% MFI) and cell death (% annexin V and PI-positive) are measured by flow cytometry. α KO, Hsp90α KO; β KO, Hsp90β KO cells. (D) Order of treatment experiment as in Fig. 7C, but with HEK cells. Cell size (% MFI) and relative cell death (% annexin V and PI-positive) are quantified by flow cytometry. The values for cell size and death in the different experimental conditions are normalized to the respective 37 °C controls (n = 3 biologically independent experiments). (E) Cell size was reduced by serum starvation (starved) for 3 days before subjecting cells to chronic HS for 3 additional days (starved > HS). HS > starved, the two treatments were done the other way around (HS > starved). Cell size (% MFI) and relative cell death (% annexin V-PI positive) were quantified by flow cytometry and normalized to the respective 37 °C controls (n = 3 experimental samples). (F) Order of treatment experiment with CHX and rapamycin to reduce cell size first for 3 days and then subjecting cells to oxidative stress with 10 µM sodium arsenite (Ars) for 1 day. Cell size (% MFI) and relative cell death (% annexin V and PI-positive) were quantified by flow cytometry and normalized to the Ars single treatment controls (n = 3 experimental samples).

Schematic representation of the flow cytometric strategies for cell size and cell cycle analyses.

(A) Gating and analysis strategy for cell size; relevant to Fig. 1B, E and H, Fig. 2C, Fig. 5B and H, Fig. 7B-C and F, figure supplements 1B, 2B, 3D, G and J, and 7A and C-F. For size measurements, cell populations were gated based on the values of the forward scatter (FSC). (B) Gating and analysis strategy for cell cycle; relevant to Fig. 1C-D and I, and figure supplement 1D.

Schematic representation of the flow cytometric strategies to measure translation.

Relevant to Fig. 5D and H, and figure supplement 5F-G.

Schematic representation of the flow cytometric strategies for measuring autophagic flux and in vivo UPS activities.

(A) Gating strategy for autophagic flux measurements, relevant to Fig. 6B and figure supplement 6B. (B) Gating strategy for measuring UPS activity; related to Fig. 6A and figure supplement 6A.

List of oligonucleotides used to generate the expression vectors for the shRNAs shHSF1-1 and shHSF1-2.

List of antibodies used in this study.