Solid-to-liquid phase transition in the dissolution of cytosolic misfolded-protein aggregates

Abstract

Accumulation of protein aggregates is a hallmark of cellular aging and degenerative disorders. This could result from either increased protein misfolding and aggregation or impaired dissolution of aggregates formed under stress, the latter of which is poorly understood. In this study, we employed quantitative live-cell imaging to investigate the dynamic process of protein disaggregation in yeast. We show that protein aggregates formed upon heat stress are solid condensates, but after stress attenuation these protein aggregates first transition into a liquid-like state during their dissolution. This solid-to-liquid phase transition (SLPT) accompanies the reduction in aggregate number due to the fusion of the liquid condensates. The chaperone activity of Hsp104, a Clp/HSP100 family chaperone, is required for both SLPT and subsequent dispersal of the liquid condensates. Sse1, a yeast HSP110 chaperone, also facilitates SLPT. These results illuminate an unexpected mechanistic framework of cellular control over protein disaggregation upon stress attenuation.

Keywords: Biochemistry; Biological sciences; Cell biology.

Graphical abstract

Figure 1

Protein aggregates undergo a solid to liquid-like phase transition during the dissolution process (A) Time-lapse confocal imaging of a single yeast cell expressing FlucSM-GFP labeled aggregates after a 30 min heat shock (HS) at 42°C and shifted back to their permissive temperature of 30°C for 90 min. Scale bar, 2 μm. (B) Zoom in of individual rugged aggregates (pink box), aggregates that are beginning to come together and cluster (blue box), and aggregates once they have fused together (green box) displayed with fire LUTs. Scale bar, 0.5 μm. (C) Quantification of aggregate circularity after HS during dissolution. Each dot represents the circularity of a single aggregate. N = 3 movies. Error bars represent the SD. Statistical analysis was performed via a one-way ANOVA with a p value < 0.05 followed by a Tukey post hoc test with the following significance cut-off: not significant (ns): p > 0.05, ∗: p ≤ 0.05, ∗∗: p ≤ 0.01, ∗∗∗: p ≤ 0.001, ∗∗∗∗: p ≤ 0.0001. (D) Lineage trace of FlucSM-GFP labeled aggregates (top) and aggregate number (bottom) over time from the representative time-lapse movie of the cell from (A). Aggregates were detected and tracked in Imaris so that each line represents a track from an individual spot at each time point. Pink arrows represent events where a clustered aggregate moves apart and green arrows represent fusion events. The color represents the mean intensity of the aggregate in the track; pink denotes high intensity and blue denotes low intensity aggregates. (E) Normalized mean aggregate intensity, whole cell intensity, and aggregate number per cell of FlucSM-GFP labeled aggregates following HS. Time represents the number of minutes after shifting back to permissive temperature of 30°C (n = 3 movies and 96 cells). Error bars represent the SD. (F) Representative images from FlucSM-mEOS3.2 photoconversion experiment. Photoconverted aggregate after HS is circled with a pink circle and the non-converted aggregate is circled in green. Over time, these aggregates come together and the fused aggregate is circled in white. Time is represented in minutes. Scale bar, 0.5 μm. (G and H) Representative kymograph through the bleached half and non-bleached aggregate half of a FlucSM-GFP labeled aggregate at 0 min (G) and 30 min (H) after HS. For 0 min after HS: the bleached region (black line) and non-bleached region (gray line). For 30 min after HS: the bleached region (pink line) and non-bleached region (light pink line). Scale bar, 0.5 μm. (I) Quantification of the fluorescence intensity decay of the bleached half and non-bleached half of FlucSM-GFP aggregates at 0 and 30 min after HS depicted in G and H. (J) Normalized mean aggregate intensity and SEM for FLIP experiments acquired 0, 20, and 30 min after HS (n = 20, n = 12, and n = 25 respectively). Also included is the control normalized fluorescence intensity of a non-bleached aggregate in the same field as bleached aggregates at 0 min time point (n = 14 aggregates). Decay rate k (K), half-life (L), and immobile fraction (M) from the non-bleached half of the aggregate whose other half was being photobleached. Individual data points represent a single bleached aggregate. Statistical analysis was performed via an unpaired two-tailed t-test with the following significance cut-off: not significant (NS): p > 0.05, ∗: p ≤ 0.05, ∗∗: p ≤ 0.01, ∗∗∗: p ≤ 0.001, ∗∗∗∗: p ≤ 0.0001.

Figure 2

Molecular determinants of aggregate dissolution (A) Screen schematic to identify proteins involved in aggregate dissolution. Aggregate enriched proteins were knocked out, labeled with FlucSM-GFP, subjected to a 30 min HS and imaged at 0 and 30 min after HS. Aggregate number was quantified (as in Figure S3) and compared between KO mutants and WT. (B) Representative images for top hits with dissolution and fusion defects. FlucSM-GFP aggregates are shown at 0 and 30 min after HS. Note xrn1Δ intensity is scaled differently than WT and other hits since the aggregates are dimmer. (C and D) Normalized mean aggregate intensity (C) and normalized aggregate number per cell (D) of FlucSM-GFP labeled aggregates in KO strains after HS. Time represents the number of minutes after shifting back to permissive temperature. Error bars represent the SD (WT n = 3 movies and 165 cells, hsp104Δ n = 3 movies and 141 cells, pfk1Δ n = 2 movies and 90 cells, sse1Δ n = 3 movies and 91 cells, xrn1Δ n = 3 movies and 87 cells). (E and F) Lineage trace (F) of xrn1Δ FlucSM-GFP labeled aggregates (top) and aggregate number (bottom) over time from time course images (E). Each line represents an aggregate track and the corresponding number of aggregates per cell at each time point. Due to the increased aggregate number per cell, only a subset of tracks are shown. Pink arrows represent events where a clustered aggregate falls apart. The color represents the mean intensity of the aggregate in the track; pink represents high intensity while blue represents low intensity aggregates. (G and H) Lineage trace (H) of pfk1Δ FlucSM-GFP labeled aggregates (top) and aggregate number (bottom) over time from time course images (H). Only a subset of aggregate tracks are depicted. Statistically colored traces and arrow coloring same as in (E & F). Scale bar for all panels, 2 μm.

Figure 3

Hsp104 is required for the solid to liquid-like transition (A) Dissolution of FlucSM-GFP labeled aggregates in hsp104Δ cells during a 90 min time course showing not only failure for aggregates to dissolve, but also lack of fusion. Pink arrows indicate individual aggregates and teal arrows aggregates that cluster together but do not stay together. Scale bar, 2 μm. (B) Lineage trace and aggregate number over time for images in (A). Pink arrows represent events where a clustered aggregate moves apart and green arrows represent when they come together. The color represents the mean intensity of the aggregate in the track; pink represents high intensity while blue represents low intensity aggregates. (C and D) Normalized mean whole cell and aggregate intensity (C) and normalized aggregate number per cell (D) of FlucSM-GFP labeled aggregates following HS. Time represents the number of min after shifting back to permissive temperature of 30°C (WT n = 3 movies and 96 cells and hsp104Δ n = 3 movies and 110 cells). Error bars represent SD. Images were taken alongside and compared to WT from Figure 1D. (E) Representative FLIP kymograph of a FlucSM-GFP labeled aggregate in hsp104Δ cells 0 and 30 min after HS. For 0 min after HS: the bleached region (black line) and non-bleached region (gray line). For 30 min after HS: the bleached region (pink line) and non-bleached region (light pink line) and the fluorescence loss in these two examples is quantified in (Figures S4G–S4J). Scale bar, 1 μm. (F) Mean normalized aggregate intensity and SEM for FLIP experiments (n = 8 for hsp104Δ 0 min, and n = 14 for hsp104Δ 30 min). Control normalized fluorescence intensity of a non-bleached aggregate in the same field as the bleached aggregates at 0 min time point in hsp104Δ cells (n = 8 aggregates). (G) Immobile fraction in hsp104Δ cells compared to WT (Figure 1) and error bars represent SD. Individual data points represent the immobile fraction of a single non-bleached aggregate half. Statistical analysis was performed via a one-way ANOVA with a p value < 0.05 followed by a Tukey post hoc test with the following significance cut-off: not significant (ns): p > 0.05, ∗: p ≤ 0.05, ∗∗: p ≤ 0.01, ∗∗∗: p ≤ 0.001, ∗∗∗∗: p ≤ 0.0001. (H) Schematic of Hsp104 induction experiment in hsp104Δ background. (I) Aggregates were allowed to form in the absence of Hsp104 and Hsp104 expression was induced with estradiol (ethanol as a solvent control) and images from 30 min time intervals after HS. Scale bar, 2 μm. Note the intensity is scaled differentially for t = 1 to prevent overexposure of following images. (J) Mean aggregate number per cell decreases with the induction of Hsp104 expression in hsp104Δ. Each dot represents the average number of aggregates per cell in one field (n = 3 fields each). Error bars represent the SD. (K and L) Representative FLIP kymograph of hsp104Δ cells induced with Hsp104 expression for 2 h with estradiol (K) and quantification of the mobile vs. immobile fraction (L) (n = 13 aggregates). Scale bar, 1 μm.

Figure 4

Hsp104 activity is required for the solid to liquid-like transition after HS (A) Schematic of Hsp104 and point mutations that affect chaperone activity that are used in our study. Color labels are consistent for these mutants in graphs C, D, I, and J. (B) Montages of FlucSM-GFP labeled aggregates in cells with indicated Hsp104 mutations recovering from HS over 150 min time lapse movie. Scale bar, 2 μm. (C and D) Normalized mean aggregate intensity (C) and normalized aggregate number per cell (D) of FlucSM-GFP labeled aggregates following HS. Error bars represent SD. (Hsp104 WT n = 3 movies and 132 cells, hsp104^A503V n = 3 movies and 132 cells, hsp104^Y257A n = 3 movies and 146 cells, and hsp104^DWB n = 3 movies and 143 cells). (E and H) Representative kymographs of FlucSM-GFP labeled aggregates in Hsp104 variant backgrounds at 0 min and 30 min after HS in cells expressing WT Hsp104 (E), Hsp104^A503V (F), Hsp104^Y257A (G), Hsp104^DWB (H). For 0 min after HS: the bleached region (black line) and non-bleached region (gray line). For 30 min after HS: the bleached region (pink line) and non-bleached region (light pink line). (I) Normalized mean aggregate intensity and SEM for FLIP experiments acquired in cells expressing WT Hsp104 (n = 15 and 20), Hsp104^A503V (n = 20 and 19), Hsp104^Y257A (n = 10 and 13), and Hsp104^DWB (n = 21 and 15) 0 and 30 min after HS. Scale bar, 0.5 μm. Control normalized fluorescence intensity of a non-bleached aggregate in the same field as bleached aggregates at 0 min time point in WT Hsp104 cells (n = 8 aggregates). (J) Immobile fraction from the non-bleached half from FLIP experiments. Individual data points represent a single non-bleached aggregate half. Statistical analysis was performed via a one-way ANOVA with a p value < 0.05 followed by a Tukey post hoc test with the following significance cut-off: not significant (ns): p > 0.05, ∗: p ≤ 0.05, ∗∗: p ≤ 0.01, ∗∗∗: p ≤ 0.001, ∗∗∗∗: p ≤ 0.0001.

Figure 5

The role of Hsp70 chaperones in aggregate dissolution (A) Time-lapse images of FlucSM-GFP labeled aggregates in cells in indicated strain backgrounds over 2 h. Scale bar, 2 μm. (B and C) Normalized mean aggregate intensity (B) and normalized aggregate number per cell (C) of FlucSM-GFP labeled aggregates following HS. Error bars represent SD. (WT n = 3 movies and 96 cells, hsp104Δ n = 3 movies and 110 cells, the ssa1^tsssa2Δ ssa3Δ ssa4Δ n = 3 movies and 91 cells, sse1Δ n = 3 movies and 91 cells). ssa1^tsssa2Δ ssa3Δ ssa4Δ plotted alongside WT (Figure 1), sse1 (Figure 2), and hsp104 (Figure 3). (D) Half-time of aggregate dissolution for aggregate number and intensity. Statistical analysis was performed via a one-way ANOVA with a p value < 0.05 followed by a Tukey post hoc test with the following significance cut-off: not significant (ns): p > 0.05, ∗: p ≤ 0.05, ∗∗: p ≤ 0.01, ∗∗∗: p ≤ 0.001, ∗∗∗∗: p ≤ 0.0001. (E and F) Kymographs of bleached aggregates for FLIP experiments in ssa1^tsssa2Δ ssa3Δ ssa4Δ (E) and sse1Δ (F). Scale bar is 1 μm. (H and I) Normalized fluorescence intensity for FLIP experiments (H) and (I) respectively. (ssa1^tsssa2Δ ssa3Δ ssa4Δ n = 21 and n = 26 for 0 and 30 and for sse1Δ n = 11 and n = 10 for 0 and 30). Control normalized fluorescence intensity of a non-bleached aggregate in the same field as bleached aggregates at 0 min time point in ssa1^tsssa2Δ ssa3Δ ssa4Δ cells (H, n = 14 aggregates) and sse1Δ cells (I, n = 11 aggregates). (G and J) Half-life (G) and immobile fraction (J) plotted alongside WT (Figure 1) and hsp104Δ (Figure 3). Statistical analysis was performed via an unpaired two-tailed t-test (G) and a one-way ANOVA (J) with the same significance cut-offs as in (D).

Figure 6

Aggregates must first undergo a Hsp104 dependent phase transition before dissolution (A) Schematic illustrating that aggregates were allowed to form in WT cells, and 0 min after HS, Hsp104 was inhibited with GdnHCl. Representative maximum intensity projection images are shown every 30 min. Scale bar, 2 μm. Pink arrows represent individual aggregates and teal arrows aggregates that cluster together. (B) Same as (A) but Hsp104 was inhibited 30 min after HS. (C and D) Average aggregate intensity does not decrease with Hsp104 inhibition when added at either time point (C) nor does the normalized aggregate number per cell with Hsp104 inhibition (D) (WT 0 min n = 2 and 110 cells, 0 min + GdnHCl n = 3 and 140 cells, WT 30 m in n = 2 and 106 cells, and 30 min + GdnHCl n = 3 and 114 cells). (E and F) Representative FLIP kymograph of a FlucSM-GFP labeled aggregate when Hsp104 was inhibited with GdnHCl for 1 h after 0 min (E) and 30 min after HS (F). For 0 min after HS: the bleached region (black line) and non-bleached region (gray line). For 30 min after HS: the bleached region (pink line) and non-bleached region (light pink line). Scale bar, 0.5 μm. (G) Quantification of the fluorescent loss of the bleached half and non-bleached half of aggregates in (E&F). (H) Normalized mean aggregate intensity and SEM for FLIP experiments acquired in cells with Hsp104 inhibited 0 min and 30 min after HS (n = 14 and n = 25 respectively). Control normalized fluorescence intensity of a non-bleached aggregate in the same field as bleached aggregates at 0 min time point in cells inhibited with GdnHCl at 0 min (n = 10 aggregates). (I–K) Decay rate k (I), half-life (J), and immobile fraction (K) from the non-bleached half from FLIP experiments compared to WT and hsp104Δ. Individual data point represent a single bleached aggregate and statistical analysis was performed via an unpaired two-tailed t-test (I-J) and a one-way ANOVA (K) with the following significance cut-off: not significant (NS): p > 0.05, ∗: p ≤ 0.05, ∗∗: p ≤ 0.01, ∗∗∗: p ≤ 0.001, ∗∗∗∗: p ≤ 0.0001.