SUPPLEMENTAL INFORMATION
Senescence inhibits the chaperone response to thermal stress
Jack Llewellyn1, 2, Venkatesh Mallikarjun1, 2, 3, Ellen Appleton1, 2, Maria Osipova1, 2, Hamish TJ Gilbert1, 2,
Stephen M Richardson2, Simon J Hubbard4, 5 and Joe Swift1, 2, 5
(1) Wellcome Centre for Cell-Matrix Research, Oxford Road, Manchester, M13 9PT, UK.
(2) Division of Cell Matrix Biology and Regenerative Medicine, School of Biological Sciences, Faculty of
Biology, Medicine and Health, Manchester Academic Health Science Centre, University of
Manchester, Manchester, M13 9PL, UK.
(3) Current address: Department of Biomedical Engineering, University of Virginia, Box 800759, Health
System, Charlottesville, VA, 22903, USA.
(4) Division of Evolution and Genomic Sciences, School of Biological Sciences, Faculty of Biology,
Medicine and Health, Manchester Academic Health Science Centre, University of Manchester,
Manchester, M13 9PL, UK.
(5) Correspondence to SJH ([email protected]) or JS ([email protected]).
Page 1 of 11 Supplemental Information: Llewellyn et al. Chaperone stress response in senescence
CONTENTS
Supplemental figures S1 – S5 … … … … … … … … 3
Supplemental table S6 … … … … … … … … 10
Supplemental references … … … … … … … … 11
Page 2 of 11
Supplemental Information: Llewellyn et al. Chaperone stress response in senescence
SUPPLEMENTAL FIGURES
Figure S1.
A EP (passage 3) LP (passage 16)
200 µm 200 µm
1.5 3 B Mass spectrometry proteomics (n = 4) C mRNA (n = 4) D 100k EP 1.0 2 p < 0.0001 p < 0.0001 LP p < 0.0001 p < 0.0001 ) 0.5 1 2 p < 0.0001 p < 0.0001 10k 0.0 0
-0.5 -1 Cell area (µm Cell area fold change vs. EP fold change vs. EP 2 2 log -1.0 log -2 1k
-1.5 -3 GLB1 LMNB1 LMNB1 WH211 WH226 WH230
Figure S1. Characteristics of senescence in primary human mesenchymal stem cells (hMSCs) cultured to late passage (LP) were increased relative to those in donor-matched cells at early passage (EP). LP cells were examined at the passage number where onset of senescence occurred i.e. further cell proliferation was halted. This point was found to vary between donors (see Table 1 in the Materials and Methods section).
Comparisons were made between donor-matched samples of senescent LP and EP cells at passage 3: (A)
LP hMSCs showed positive staining for β-galactosidase. (B) Quantitative mass spectrometry proteomics (Fig.
1A) showed a corresponding significant increase in β-galactosidase protein (GLB1, p < 0.0001) in LP vs. EP
cells, and significant loss of the senescence marker lamin-B1 (LMNB1, p < 0.0001) (Shimi et al, 2011). Log2 fold-changes and significance values corrected for Benjamini-Hochberg false discovery rate (BH-FDR) were calculated as described previously (Mallikarjun et al, 2020); n = 4 primary donors. (C) RT-qPCR also showed a corresponding significant decrease of LMNB1 transcript in LP vs. EP hMSCs (p < 0.0001 established by t- test; normalisation to housekeeping gene PPIA; n = 4 primary donors). (D) An analysis of cell morphology showed that LP hMSCs had significantly larger spread areas than EP cells (p < 0.0001 established by ANOVA testing, with ≥ 39 cells imaged per condition), consistent with established senescence phenotype
(Hernandez-Segura et al, 2018). Solid red lines indicate medians; dashed red lines indicate quartiles.
Page 3 of 11
Supplemental Information: Llewellyn et al. Chaperone stress response in senescence
Figure S2.
TTC8 A CCT8L2 B FKBP14 BBS4 PFDN5 PFDN1 CCT6B CCT8 MKKS PFDN2 UNC45A SGTA CCT4 AIP CCT3 PFDN6 BBS10 CCT5 VBP1 TXNRD1 FKBP5 TCP1 UNC45B FKBP6 CCT7 CDC23 PFDN4 ANAPC5 PPP5C CCT2 CCT6A TXNRD3 PPIL6 FKBPL FKBP9 ERP44 CDC27 TTC4 TXN PRDX4 CDC16 CLPB HSPE1 PPIG NKTR PPID QSOX1 ANAPC7 STUB1 FKBP4 TTC30A LONRF1 DNAJB4 TTC12 DNAJB6 ST13 DNAJB12 HSP90AA1 DNAJB7 DNAJB5 HSPH1 TTC21B TTC26 SUGT1 DNAJB1 TRAP1 STIP1 HSP90AB1 CDC37L1 TOR4A AHSA2 HSPA2 DNAJC5 BAG1 PTGES3 TTC28 DNAJA3 GRPEL2 GRPEL1 0 1 PTGDS CDC37 AHSA1 STRING interaction score HSPA13 HSPA12B DNAJA1 HSPA1L TTC30B IFT88 DNAJA2 HSPA14 DNAJC6 HSCB DNAJB9 BAG2 BAG3 HSPD1 FKBP8 HSPA1A PPIH DNAJC5B HSP90B1 HSPA4 HSPA8 HSPA9 HSPA12A KLC1 DNAJA4 IFIT1 BCS1L DNAJB13 HSPA6 HSPA1B DNAJC3 TOMM70A HSPA5 HSPA4L IFIT2 CWC27 BAG4 KLC3 HSPB1 HYOU1 DNAJC2 DNAJC19 PRPF6 VCP DNAJB11 GAK DNAJB2 KLC4 DNAJB14 XAB2 BAG5 DNAJC10 DNAJC10 DNAJC1 IFIT3 PPIE PPIL1 DNAJC7 KLC2 PPWD1 CANX DNAJB8 CRNKL1 DNAJC8 PDIA6 TXNDC5 SRP68 TOR1B NOXA1 CRYAB PDIA3 PDIA4
CLGN CALR SIL1 ERO1L SRP72 TOR1A NCF2 CRYAA ERLEC1 EDEM1 P4HB ERO1LB ASPH NAA15 PHB LEPREL1 ERP27
PLOD3 P4HA1 PLOD2 P4HA3
P4HA2 FKBP1B NAA16 PHB2 1 64 PLOD1 PPIB SERPINH1 Number of interactions
Figure S2. Application of network analysis to define functional groups within the human chaperome.
(A) The 332 protein human chaperone network (Brehme et al, 2014), with nodes coloured according to chaperone degree as a measure of importance to the function of the network (Rubinov & Sporns, 2010). The most highly connected proteins in the network were: HSPA8 (degree = 64); HSP90AA1 (53); HSPA5 (52);
HSPA9 (48); and, HSPA1A (47). (B) The adjacency matrix of the human chaperone network following modularity analysis. Rows and columns are the 332 nodes in the network, with entries representing the
STRING interaction score (Szklarczyk et al, 2019). Interactions have been filtered to only include those of the highest confidence (score ≥ 0.9).
Page 4 of 11
Supplemental Information: Llewellyn et al. Chaperone stress response in senescence
Figure S3.
EP LP EP LP 37°C 37°C 42°C 42°C A (n = 2 donors) B (n = 2 donors) C HSPA4 D HSPA4 BAG5 DNAJC3 BAG5 DNAJC3 14 14 BAG2 HSPA6 BAG2 HSPA6 sig. up 10 sig. up 30 sig. down 22 sig. down 118 HSPA1A DNAJA2 HSPA1A DNAJA2 12 12 DNAJC5 HSPH1 DNAJC5 HSPH1 10 10 HSPB1 DNAJB4 HSPB1 DNAJB4 HSP70 HSP70 8 8 HSPA8 HSPA4L HSPA8 HSPA4L machinery machinery 6 6 DNAJC10 DNAJA1 DNAJC10 DNAJA1
HSPD1 GRPEL1 HSPD1 GRPEL1 4 4 FDR-corrected p-value FDR-corrected p-value FDR-corrected
10 10 DNAJB1 HSPA9 DNAJB1 HSPA9
2 2 FKBP8 HSPA12A FKBP8 HSPA12A
-log -log TOMM70A BAG3 TOMM70A BAG3 HSPA5 HSPA5 0 0 PES vs. control PES vs. control -3 -2 -1 0 1 2 3 -3 -2 -1 0 1 2 3 LP at 37°C LP at 42°C log2 fold-change log2 fold-change geldanamycin vs. control radicicol vs. control
Figure S3. Proteomic analysis of early passage (EP) and late passage (LP) primary human mesenchymal stem cells (hMSCs) treated with chaperone inhibitors and thermal stress. Volcano plots showing changes to the proteomes of EP hMSCs treated with the HSP90-inhibitors (A) geldanamycin, or (B) radicicol, at 37 °C. Red and blue points satisfy a p-value < 0.05. p-values were calculated using empirical
Bayes-modified t-tests with Benjamini–Hochberg false discovery rate (FDR) correction (Mallikarjun et al,
2020). (C) Network of proteins in the HSP70 chaperome module, showing changes induced by HSP70- inhibitor 2-phenylethynesulfonamide (PES) in LP hMSCs under control conditions. (D) Changes to protein levels in the HSP70 chaperome module induced by PES treatment in LP hMSCs subjected to 2-hour heat treatment at 42 °C. In network diagrams, node size is indicative of chaperone degree, while edge weight indicates interaction score between chaperones. The central node is coloured according to the mean abundance change of chaperones associated with the module. Descriptions of the chaperone modules can be found in Fig. 2A of the main paper.
Page 5 of 11
Supplemental Information: Llewellyn et al. Chaperone stress response in senescence
Figure S4.
1.5 1.5 A EP B 1.0 LP 1.0
0.5 0.5
0.0 0.0 (nucleus) (nucleus) LP vs. EP -0.5 -0.5 p-value p-value (EP ≠ LP) (vs. t = -2 hrs) 0.0001 0.0001 HSPA1A mean IF intensity HSPA1A HSPA1A mean IF intensity HSPA1A 0.001 2 2 -1.0 -1.0 0.001 0.01 0.01 log
log > 0.05 (n = 3 donors) > 0.05 (n = 3 donors) -1.5 -1.5 -2 0 4 8 12 16 20 24 -2 0 4 8 12 16 20 24 Time after heat shock (hours) Time after heat shock (hours)
1.5 1.5 C p-value D (vs. t = -2 hrs) 1.0 0.0001 1.0 0.001 0.01 0.5 > 0.05 0.5
0.0 0.0 LP vs. EP -0.5 -0.5 p-value (EP ≠ LP) nuclear:cytoplasmic nuclear:cytoplasmic 0.0001 HSPA1A mean IF intensity HSPA1A HSPA1A mean IF intensity HSPA1A 0.001 2 -1.0 2 -1.0 EP 0.01 log (n = 3 donors) LP log (n = 3 donors) > 0.05 -1.5 -1.5 -2 0 4 8 12 16 20 24 -2 0 4 8 12 16 20 24 Time after heat shock (hours) Time after heat shock (hours)
1.5 E F EP 1.0 LP
Dapi EP Phalloidin 0.5
0.0 (nucleus) 100 µm HSF1 -0.5 p-value (vs. t = -2 hrs) HSF1 mean IF intensity
2 0.0001 -1.0 0.001
log 0.01 (n = 3 donors) > 0.05 -1.5 Dapi -2 0 4 8 12 16 20 24 Phalloidin LP Time after heat shock (hours)
G 1.5 100 µm HSF1 1.0 -2 -1 0 0.5 1 2 4 8 24 0.5 Time after heat shock (hours)
0.0 LP vs. EP -0.5 p-value (EP ≠ LP) 0.0001 -1.0 0.001 0.01 HSF1 mean IF intensity (nucleus)
2 (n = 3 donors) > 0.05 -1.5 log -2 0 4 8 12 16 20 24 Time after heat shock (hours)
Page 6 of 11
Supplemental Information: Llewellyn et al. Chaperone stress response in senescence
Figure S4. Analysis of immunofluorescence (IF) images of heat shock protein 70 kDa (HSPA1A) and heat shock factor 1 (HSF1) in early passage (EP) and late passage (LP) human mesenchymal stem cells
(hMSCs) subjected to heat treatment. hMSCs were imaged before, during and over a 24-hour period following a 2-hour heat treatment at 42 °C (see Fig. 4A in the main paper). Mean HSPA1A intensities were determined in nuclear and cytoplasmic cellular regions, defined by areas of DAPI and phalloidin staining, respectively. (A) Levels of HSPA1A protein in the nuclei of EP and LP hMSCs before, during and after heat stress. (B) Ratios of mean intensities of HSPA1A in the nuclei of LP vs. EP hMSCs subjected to heat stress.
(C) Nuclear-to-cytoplasmic ratios of HSPA1A in EP and LP hMSCs subjected to heat stress. (D) Nuclear-to- cytoplasmic ratios of HSPA1A in LP vs. EP hMSCs subjected to heat stress. (E) Representative IF images of
HSF1 in EP and LP hMSCs before, during and for 24 hours following a 2-hour heat treatment at 42 °C. (F)
Quantification of levels of HSF1 protein in the nuclei of EP and LP hMSCs before, during and after heat stress.
(G) Ratios of mean intensities of HSF1 in the nuclei of LP vs. EP hMSCs subjected to heat stress. In figure panels (A), (C) and (F), point size indicates the significance of changes vs. pre-heating, with solid points showing p < 0.05 (from ANOVA testing, n = 3 donors). In panels (B), (D) and (G), point size indicates significance where EP ≠ LP, with solid points showing p < 0.05 (from ANOVA testing, n = 3 donors).
Page 7 of 11
Supplemental Information: Llewellyn et al. Chaperone stress response in senescence
Figure S5.
1.5 0.010 15 A 1.00 B C HSPA1A 1.0 10 ) 0.75 0.005 log 2
0.5 5 transcript ( fold-change in silico 0.50 0.0 0.000 0 fold-change 2
-0.5 -5 in vitro HSP70 fold-change log 2 -0.005 0.25 free HSF1 ( HSPA1A (normalised)
log -1.0 in vitro -10
in vitro ) in silico in silico 0.00 -1.5 -0.010 -15 0 4 8 12 16 20 24 -2 0 4 8 12 16 20 24 -2 0 4 8 12 16 20 24 Time (hours) Time after heat shock (hours) Time after heat shock (hours)
D 15 E F 3.0 (n = 4 donors) MFP HSPA1A 4 10 CHIP M) total HSPA1A µ 2 5 2.0 free HSPA1A 0 0 abundance of 2 -5 active HSF1 1.0 -2 (concentration, Median peptide log 10 signal intensity (AU) mBBr-adduct peptides -10 log -4 HSF1-bound HSPA1A -15 0.0 0 hrs 2 hrs 0 120 240 360 480 600 TH269 WH226 WH211 TH215 Simulation time (minutes) Donor identifier
G H I
Page 8 of 11
Supplemental Information: Llewellyn et al. Chaperone stress response in senescence
Figure S5. Results supporting development and characterisation of an ordinary / delayed differential equation (ODE/DDE) model of the cellular response to heat stress. (A) Decay of HSPA1A levels simulated by removing HSPA1A synthesis from the ODE/DDE model at equilibrium. The in silico half-life of HSPA1A was 100 minutes. (B) ODE/DDE model simulation of the stress response with adjustments made to resemble senescent cells. The in silico model was given sufficient time to reach a stable equilibrium, before a proteotoxic stress was simulated by increasing the rate at which misfolded proteins were generated within the model for 120 minutes. At each time interval, the in silico concentration of HSPA1A was recorded, and is shown overlaid with data acquired by immunofluorescence from late passage (LP) hMSCs in vitro (see Fig.
4B of the main paper). (C) As (B), but showing concentrations of modelled active HSF1 overlaid, with a scaling factor, onto experimentally-derived HSPA1A transcript levels in early passage (EP) hMSCs (see Fig. 4D of the main paper). (D) Abundances of monobromobimane (mBBr)-tagged peptides in mass spectrometry data pre- stress and immediately following a 2-hour treatment at 42 °C. Dotted lines indicate sample medians. (E)
Convergence of populations in the ODE/DDE model to a stable equilibrium prior to stress. (F) HSPA1A and
CHIP protein signal intensities quantified by mass spectrometry, in the absence of stress, in EP hMSCs from four human donors. (G-I) Demonstration of the optimisation of free model parameters. The ODE/DDE model was used to simulate the stress response over a range of parameter values, and the sum of squared errors
(SSE) compared to experimentally derived data (measurements of HSPA1A by immunofluorescence) at each parameter value. The parameter values chosen by optimisation are shown overlaid in black over the SSE values.
Page 9 of 11
Supplemental Information: Llewellyn et al. Chaperone stress response in senescence
SUPPLEMENTAL TABLE
Parameter Description Value Unit
� Maximal HSPA1A transcription rate 6.4294 × 10! ���"#
� (Modified value in senescent cells) 4.5865 × 10! ���"#
� Hill coefficient 3 (dimensionless)
�$ HSF1 dissociation constant 3.0534 ��
"% "# "# �! HSF1-HSPA1A binding rate 3.9827 × 10 �� ���
"% "# "# �& HSPA1A-MFP binding rate 3.9827 × 10 �� ���
"& "# "# �% CHIP-mediated HSPA1A turnover rate 1.2400 × 10 �� ���
"! "# �' HSPA1A-mediated MFP refolding rate 5.0199 × 10 ���
"& "# "# �( CHIP-mediated MFP turnover rate 2.1200 × 10 �� ���
"# �) Rate of protein misfolding at 37 °C 69.1068 �� ���
� Rate of protein misfolding at 42 °C 4.2651 × 10! �� ���"#
� Simulation time elapsed ���
� HSF1 activation/HSPA1A synthesis delay ~�[60, 180] ���
Supplemental Table S6. Optimised and approximated parameters for the ordinary/delay differential equation
(ODE/DDE) mathematical model of the heat stress response. MFP = misfolded protein; CHIP = E3 ubiquitin- protein ligase CHIP.
Page 10 of 11
Supplemental Information: Llewellyn et al. Chaperone stress response in senescence
SUPPLEMENTAL REFERENCES
Brehme M, Voisine C, Rolland T, Wachi S, Soper JH, Zhu Y, Orton K, Villella A, Garza D, Vidal M, Ge H, Morimoto RI (2014) A chaperome subnetwork safeguards proteostasis in aging and neurodegenerative disease. Cell Reports 9: 1135-1150
Brocchieri L, Karlin S (2005) Protein length in eukaryotic and prokaryotic proteomes. Nucleic Acids Research 33: 3390-3400
Hernandez-Segura A, Nehme J, Demaria M (2018) Hallmarks of cellular senescence. Trends in Cell Biology 28: 436-453
Mallikarjun V, Richardson SM, Swift J (2020) BayesENproteomics: Bayesian elastic nets for quantification of peptidoforms in complex samples. Journal of Proteome Research 19: 2167-2184
Rubinov M, Sporns O (2010) Complex network measures of brain connectivity: Uses and interpretations. Neuroimage 52: 1059-1069
Shimi T, Butin-Israeli V, Adam SA, Hamanaka RB, Goldman AE, Lucas CA, Shumaker DK, Kosak ST, Chandel NS, Goldman RD (2011) The role of nuclear lamin B1 in cell proliferation and senescence. Genes & Development 25: 2579-2593
Szklarczyk D, Gable AL, Lyon D, Junge A, Wyder S, Huerta-Cepas J, Simonovic M, Doncheva NT, Morris JH, Bork P, Jensen LJ, Mering C (2019) STRING v11: protein-protein association networks with increased coverage, supporting functional discovery in genome-wide experimental datasets. Nucleic Acids Research 47: D607-D613
Page 11 of 11