The Hsp90 molecular chaperone governs client proteins by targeting intrinsically disordered regions

Abstract

Molecular chaperones govern proteome health to support cell homeostasis. An essential eukaryotic component of the chaperone system is Hsp90. Using a chemical-biology approach, we characterized the features driving the Hsp90 physical interactome. We found that Hsp90 associated with ∼20% of the yeast proteome using its three domains to preferentially target intrinsically disordered regions (IDRs) of client proteins. Hsp90 selectively utilized an IDR to regulate client activity as well as maintained IDR-protein health by preventing the transition to stress granules or P-bodies at physiological temperatures. We also discovered that Hsp90 controls the fidelity of ribosome initiation that triggers a heat shock response when disrupted. Our study provides insights into how this abundant molecular chaperone supports a dynamic and healthy native protein landscape.

Keywords: Hsp90; fidelity of translation initiation; intrinsically disordered regions; molecular chaperone; proteostasis.

Figure 1.. Establishing an Hsp90 physical interactome.

(A) Topology maps of Hsp90 showing the interaction areas with the indicated cochaperones or Cdk4 client (left), hydrophobicity surfaces (right) or amino acid positions that were engineered to Bpa residues (center). (B) Hsp90 interactors identified by mass spectrometry that were Bpa-dependent (dark spots) or Bpa-independent (light spot) were separated into gene ontology categories, each hit was assigned to an initial cell process and displayed. (C) The Bpa-hits demonstrate a substantial physical connection between Hsp90 and numerous other molecular chaperones and cochaperones, which we will refer to as (co)chaperone interactions. (See also Figure S1, Table S1)

Figure 2.. Combinatorial use of Hsp90’s 3 domains support a broad physical interactome.

(A) A Venn diagram with the total number of hits for each Hsp90 Bpa site is shown as well as the overlap between the hits linked to the various sites. (B) Bpa site-specific hits were compared to the established physical and genetic interactors of yeast Hsp90s. (See also Figure S2, Table S2)

Figure 3.. Hsp90 associates with a specific IDR to regulate client activity.

(A) The propensity of Hsp90-bound peptides (numbers in parentheses) to be within structured (e.g., α-helix or β-sheet) or IDR regions of a target protein was determined for known (co)chaperones and potential clients, as indicated including total numbers for each, using D²P² and AlphaFold software. (B) The D²P² (top; the rows of the colored bars represent the prediction of disorder by the 9 different algorithms used by D²P², each row/color corresponds to a single predictor and the bars within the row are a prediction of disorder) and AlphaFold (bottom) predictions for the Hsp90 client Rsc3 is shown as well as the position of the Hsp90-linked peptide (highlighted in violet). (C) The influence of Hsp90 (1, 4, 16 μM) on the DNA binding activity of Rsc3 and Rsc3 missing the Hsp90-associated IDR (Rsc3ΔIDR) was monitored by EMSA using purified recombinant proteins and a radiolabeled dsDNA RSC3 consensus oligonucleotide. (D) The capacity of Rsc3-AD and Rsc3ΔIDR-AD to drive the expression of a HIS3 reporter through an upstream RSC3 consensus motif was monitored in yeast transformants carrying an empty vector (−) or expressing the yeast Hsp90 homologs Hsc82 or Hsp82, as marked. (See also Figure S3, Table S3, Table S4)

Figure 4.. Hsp90 regulates translation initiation, which controls an HSR, by influencing the physical state of the Ded1 RNA helicase.

(A) Hsp90-hits functioning in various steps of the protein translation process are shown. Each protein is shaded based on its relative percent of IDR with dark green (50% IDR), medium green (33%), light green (25%), and white (<25%). (B) The fidelity of translation initiation was determined using the dual luciferase assay and yeast expressing WT or the G170D ts allele as a sole source of Hsp90. Data is presented as the mean ± SEM and the p-values are the following: 1 h 0.165, 2 h 0.147, and 4 h 0.006. Fold increase in luciferase expression of firefly (UUG noncanonical) relative to Renilla (AUG) in G170D was normalized to WT following incubation at 37.5°C. (C) Formation of Ded1 SGs were visualized in yeast expressing GFP-Ded1 that were treated with DMSO (white bars) or Radicicol (black bars) and shifted to 40° or 46°C, as marked for the indicated times. Data are presented as mean ± SEM and the p-values are the following: 40°C 4 min 0.90, 7 min 0.41, 10 min 0.08, 15 min 0.21, 30 min 0.04 and 46°C 4 min 0.75, 7 min 0.34, 10 min 0.41, 15 min 0.35, 30 min 0.86. (D) The physical status of the SG component Dcg1 and P-Body factor Ngr1 were followed using GFP-fusions in yeast treated with DMSO (white bars) or Radicicol (black bars) exposed to 40°C for 30 min. Data are presented as mean ± SEM and the p-values are the following: Dcp1 0.0001 and Ngr1 0.0037.

Figure 5.. The HSR response triggered by Hsp90 loss is translation dependent.

(A) Growth rates were determined for parental (circles) or G170D (squares) yeast grown at 30°, 35°, or 37.5°C in the absence (open) or presence (closed) of MG132 for the marked times. SEM for each data point derived from biological triplicates are included. (B) The induction of SSA4 expression in yeast exposed to cycloheximide (C), 42°C (HS), Azetidine (A), and/or Radicicol (R) was measured by qRT-PCR relative to cells maintained at 30°C. Data are presented as mean ± SEM. (C) The induction of HSP70 expression in human HeLA cells exposed to 42°C (HS) or Ganetespib (G25 – 25 μM; G250 – 250 μM) with or without cycloheximide (C) was measured by qRT-PCR relative to cells maintained at 37°C. Data are presented as mean ± SEM and the p-values are the following: HS vs C 0.0001, HS+C vs HS 0.2930, G25 vs C 0.0001, G25+C 0.1375, G250 vs C 0.0001, and G250+C 0.0523. (See also Figure S4)

Figure 6.. Client protein recognition and regulation by the Hsp90 chaperone.

The well-established ATP-dependent conformations of Hsp90⁷ further add to the client recognition capacity of Hsp90 by providing distinct binding platforms for different target proteins (A) and mediating a physical manipulation of a bound client polypeptide as Hsp90 progresses through its ATPase cycle (B).