An Hsp90 co-chaperone links protein folding and degradation and is part of a conserved protein quality control

In this paper, we show that the essential Hsp90 co-chaperone Sgt1 is a member of a general protein quality control network that links folding and degradation through its participation in the degradation of misfolded proteins both in the cytosol and the endoplasmic reticulum (ER). Sgt1-dependent protein degradation acts in a parallel pathway to the ubiquitin ligase (E3) and ubiquitin chain elongase (E4), Hul5, and overproduction of Hul5 partly suppresses defects in cells with reduced Sgt1 activity. Upon proteostatic stress, Sgt1 accumulates transiently, in an Hsp90-and proteasome-dependent manner, with quality control sites (Q-bodies) of both yeast and human cells that co-localize with Vps13, a protein that creates organelle contact sites. Mis-folding disease proteins, such as synphilin-1 involved in Parkinson’s disease, are also sequestered to these compartments and require Sgt1 for their clearance.

In brief Eisele et al. demonstrate that the Hsp90 co-chaperone Sgt1 has a widespread role in the folding and degradation of misfolded proteins.Furthermore, Sgt1 localizes upon heat stress to early qualitycontrol sites in an Hsp90-and proteasome-dependent manner, a process that is necessary for efficient resolution of protein inclusions.
... A balanced interplay between the systems involved in protein folding and protein degradation is key for cellular fitness, longevity, and prevention of neurodegenerative diseases.These two processes are often collectively referred to as protein homeostasis or proteostasis (Labbadia and Morimoto, 2015).The folding process is dependent on a network of molecular chaperones (Hartl et al., 2011), whereas the degradation process relies on the ubiquitin proteasome system (UPS) (Amm et al., 2014).A second line of defense in proteostasis, called spatial Protein Quality Control (spatial PQC), sequesters oligomers and aggregated proteins into large inclusions at certain protective locations within the cell (Sontag et al., 2017), a process limiting the toxicity of aberrant proteins.Specifically, upon proteostatic stress of yeast cells, the misfolded proteins initially accumulate at multiple sites, called CytoQs (Miller et al., 2015b), stress foci (Spokoini et al., 2012), or Q-bodies (Escusa-Toret et al., 2013), which later coalesce into larger inclusions at, at least, four distinct spatial quality control sites: the juxtanuclear quality control site (JUNQ); the intranuclear quality control site (INQ); the peripheral, vacuole-associated Insoluble Protein Deposit (IPOD) (Kaganovich et al., 2008;Miller et al., 2015a); and a site adjacent to mitochondria (Braun and Westermann, 2017).
The Hsp90 chaperones represent a major class of chaperones that work together with a variety of co-chaperones (Sahasrabudhe et al., 2017), many of which show high specificity for a small set of client proteins.The Hsp90 co-chaperone Sgt1 (Catlett and Kaplan, 2006) is known to possess such a specific role as an adaptor protein linking the Skp, Cullin, F-box (SCF) ubiquitin ligase complex to the Hsp90 system and to aid in the assembly of the kinetochore complex (Bansal et al., 2009;Kadota et al., 2010;Kitagawa et al., 1999).Recently, Sgt1 was shown to affect a relatively broad range of client proteins (Sahasrabudhe et al., 2017) indicating a more widespread role in Hsp90 folding processes.
Following two genome-wide screens, we here report that the Sgt1 co-chaperone is a much more general player in cellular PQC than previously anticipated and acts as an important linker, together with Hsp90, between the proteolytic and the folding branches of the proteostatic network.

Genome-wide identification of genes required for the degradation of misfolded proteins in the cytosol
To identify components of the PQC system required for the degradation of cytosolic misfolded proteins, we performed a systematic genome-wide screen following the SGA (Synthetic Genetic Array) methodology in the complete EUROSCARF deletion library and a collection containing temperature-sensitive (ts) alleles of most essential genes.In both collections, we expressed the misfolding cytosolic mutant of the yeast carboxypeptidase Y (DssCPY*) fused to Leu2, an often-used prototrophic marker in Saccharomyces cerevisiae laboratory strains, tagged with a C-terminal myc tag (DssCL*).For a second screen, we introduced Von Hippel-Lindau (VHL)-LEU2, encoding the misfolding mammalian VHL tumor suppressor protein (McClellan et al., 2005) fused to Leu2.Due to the misfolding of DssCL* and VHL-Leu2, cells degrade these proteins rapidly (Eisele and Wolf, 2008;McClellan et al., 2005).Thus, mutants in which the misfolded proteins are stabilized can easily be scored for, because they grow better than wildtype cells on media lacking leucine.
Spatial Analysis of Functional Enrichment (SAFE) (Baryshnikova, 2016;Usaj et al., 2017) of common hits of the two screens showed that four major functional groups among the essential genes were required for proper degradation of the two misfolded proteins.Among the non-essential genes, those linked to the functional groups' transcription, chromatin, and protein turnover were markedly enriched (Figure 1A; Table S1).In contrast, cells with mutations in single genes, essential or non-essential, functioning as chaperones and/or co-chaperones did not display  S1.
defects in the degradation of misfolded cytosolic proteins, suggesting that the chaperone/co-chaperone network of the cell is genetically well buffered.However, there was one noteworthy exception: the Hsp90 co-chaperone Sgt1.
Cells with reduced activity of Sgt1 displayed enhanced growth on plates lacking leucine and grew as well as cells lacking the ubiquitin ligase Ubr1, which has previously been shown to be a key factor in tagging misfolded DssCPY* for degradation (Eisele and Wolf, 2008;Heck et al., 2010).Combining the sgt1-3 mutation with an UBR1 deletion did not result in additive effects, demonstrating that they act in common pathways (Figure 1B).The sgt1-3-dependent stabilization of DssCL* could be complemented by extra-chromosomal SGT1 expression, and we confirmed that the Hsp90 system and Sgt1 are not involved in the folding or degradation of the Leu2-myc domain alone (Figures S1A and S1B).
SGT1 mutants displayed an enhanced sensitivity to the proline analog, AZC, which causes severe and general protein misfolding (Trotter et al., 2001) (Figure S1C).The data suggest that Sgt1 might be a more general player in cytosolic quality control than previously anticipated with a functional link to degradation of misfolded cytosolic proteins.Further, we observed that the sgt1-3 ubr1D double mutant showed a reduced heat sensitivity compared with the sgt1-3 mutant alone (Figure S1D).Ubr1 is known to have a role in degradation of several ts mutants (Khosrow-Khavar et al., 2012), and we could demonstrate that Sgt1 mutant protein also is a substrate of this E3 ligase (Figure S1E).
The Hsp90 co-chaperone Sgt1 physically interacts with cytosolic misfolded proteins and is required for their degradation Cycloheximide chase experiments demonstrated that the rate of degradation of DssCL* was delayed in cells with reduced Sgt1 activity to almost the same extent as in cells lacking Ubr1 (Figure 1C).However, degradation of DssCL* was not much altered in a mutant of CDC34, the ubiquitin-conjugating enzyme (E2) of the SCF complex, demonstrating a Sgt1 role independent of this complex (Figure S1F).Moreover, Sgt1 and DssCL* interacted physically, indicating a direct role for Sgt1 in modulating the stability of this misfolded substrate (Figure S1G).Degradation of VHL in the sgt1-3 mutant was also retarded (Figure S1H).Previously, VHL degradation was shown to be dependent on Hsp90s (McClellan et al., 2005); however, cycloheximide chase experiments of DssCL* in the presence of Hsp90 inhibitor geldanamycin (GA) in hsc82D or sgt1-3 strains did not support a role for Hsp90 in degradation of this substrate (Figure S1I).Also, the physical interaction between Sgt1 and DssCL* was not altered when Hsp90 activity is inhibited, neither by deletion of Hsc82 nor by addition of GA (Figure S1G).
A GFP-tagged version of misfolded cytosolic CPY* (DssCG*) has previously been shown to be dependent on both the nuclear localized E3 ligase San1 and the cytoplasmic Ubr1 for its proteasomal degradation (Heck et al., 2010).If one of the two ligases was deleted, degradation of DssCG* was only slightly delayed.However, double deletions of UBR1 and SAN1 led to strong stabilization of this substrate (Heck et al., 2010) (Figure 2C).We found that sgt1-3 mutation stabilized DssCG* more than either one of the single ubr1D and san1D mutations (Figure 1D).Combining sgt1-3 with single mutants in UBR1 or SAN1 did not result in additive effects, suggesting that Sgt1 is a key co-factor in the Ubr1/San1dependent degradation of DssCG*.A summary of factors involved in quality control of established fusion model substrates used in this study can be found in Table 1.
Sgt1 is required for endoplasmic reticulum (ER)associated degradation (ERAD) in parallel with the multiubiquitin chain assembly factor Hul5 We noticed that fragments derived from DssCL* degradation accumulated in sgt1-3 cells with sizes around 100 and 75 kDa (bands 2 and 3, respectively; Figure 1C).A similar accumulation of fragments of the ERAD substrate CTL*, a model substrate consisting of the ER lumenal ERAD substrate CPY* fused to a transmembrane domain and Leu2 (Medicherla et al., 2004), has been reported previously in cells lacking Hul5 (Kohlmann et al., 2008), a ubiquitin ligase (E3)/ubiquitin chain elongating enzyme (E4) (Crosas et al., 2006).We found that similar fragments of DssCL* accumulated in hul5D cells (Figures 2A and  S2A).Like for CTL* (Kohlmann et al., 2008), deletion of HUL5 had no effect on degradation of full-length DssCL*.Deleting HUL5 in the sgt1-3 mutant generated an additive effect on the stabilization of fragments, and overproduction of Hul5 in sgt1-3 cells suppressed such accumulation of fragments (Figures 2A and S2A).The accumulation of DssCL* fragments containing Leu2 in sgt1-3 and hul5D strains explains their enhanced growth on media lacking leucine (Figure S2B).We discovered more ubiquitylated DssCL* was accumulating in the sgt1-3 mutant.Overexpression of HUL5 led to a wild-type-like ubiquitylation state and to suppression of the enhanced ubiquitylation phenotype of the sgt1-3 strain.A HUL5 deletion led to comparably lower ubiquitylation levels than the sgt1-3 mutant.As expected, the UBR1 deletion strain displayed a markedly reduced ubiquitylation of this substrate (Figure S2C).
Because Hul5 has been shown to be required for the degradation of some ERAD substrates (Kohlmann et al., 2008), we tested if this was true also for Sgt1.Indeed, CTL* was stabilized in sgt1-3 cells, and two fragments, corresponding to those accumulating in hul5D cells (Kohlmann et al., 2008), were found to accumulate also in sgt1-3 cells (bands 2 and 3; Figure 2B).Again, stabilization of CTL* fragment accumulation and enhanced growth on plates lacking leucine of sgt1-3 cells could be suppressed by elevating Hul5 levels (Figures 2B and S2D).The ER lumenal misfolded ERAD substrate CPY*-HA (Taxis et al., 2002) was also dependent on proper SGT1 functionality for its degradation, and HUL5 overexpression improved the degradation of this substrate in the sgt1-3 mutant (Figure 2C).In contrast, the previously described cytosolic Hsp70-dependent degradation of the ERAD substrate CTG* (Taxis et al., 2003) was not dependent on Sgt1 (Figure S2E), ruling out the possibility that mutation in SGT1 leads to a general inhibition of the 26S proteasome.Taken together, the data suggest that Sgt1 and Hul5 act in parallel pathways and display similar mechanisms involved in protein degradation.

Overproduction of Hul5 increases the fitness and longevity of cells mutated in SGT1
Mutation in SGT1 caused severe fitness defects, but overexpression of HUL5 could partially compensate for reduced Sgt1 activity Cell Reports 35, 109328, June 29, 2021 3 Report up to 33 C (Figure 2D).Similarly, the rate of replicative aging was drastically accelerated in sgt1-3 cells, and overproduction of Hul5 could retard such aging to a certain degree (Figure 2E).In contrast, Hul5 overproduction alone accelerated replicative aging (Figure 2F).Deletion of HUL5 had no effect, while overexpression of SGT1 caused a modest decrease in lifespan (Figure S2F).

Sgt1 is required for aggregate coalescence and clearance
Using a GFP-tagged version of the yeast disaggregase Hsp104, which specifically binds to aggregated proteins (Erjavec et al., 2007;Spokoini et al., 2012), we found that a larger fraction of sgt1-3 cells growing at 30 C contained aggregates compared with wild-type cells (Figure 3A).Upon heat shock at 38 C for 90 min, sgt1-3 cells failed to form the typical one to two protein inclusions seen in wild-type cells and instead exhibited multiple (three or more; type 3 cells) aggregates in the cytoplasm (Figures 3A and S3A).Moreover, when cells were incubated at 42 C for 30 min and then allowed to recover at 30 C for 90 min, the rate of clearance of aggregates was markedly delayed in sgt1-3 cells (Figure S3B).This delay of recovery did not correlate with increased cell death in the ts sgt1-3 mutant (Figure S3C).The disaggregation of Ubc9ts-GFP, a fusion protein of a ts allele of UBC9 that misfolds upon a 38 C heat shock (Kaganovich et al., 2008), was also delayed in cells with reduced Sgt1 activity (Figures S3D and S3E).Sgt1 was found to be crucial also for alleviating accumulation of aggregates during aging (Figure S3F).Moreover, although the ''middle-aged'' (10-15 generations old) wild-type cells displayed mainly one or two inclusions, middle-aged sgt1-3 cells contained multiple aggregates (Figure S3F).

Sgt1 clusters in a quality control compartment dependent on Hsp90 and proteasome activity
Spatial quality control sites in yeast include JUNQ, INQ, IPOD, and CytoQs (Escusa-Toret et al., 2013;Kaganovich et al., 2008;Miller et al., 2015aMiller et al., , 2015b;;Spokoini et al., 2012).We used a GFP-tagged version of SGT1 to elucidate if Sgt1 accumulates at any such sites.By crossing GFP-SGT1 with a sgt1-3 strain, we found that the sgt1-3 allele was fully complemented demonstrating the functionality of the tagged GFP-Sgt1 protein (Figure S3G).We found that GFP-Sgt1 normally displayed a uniform distribution in the cytosol and nucleus.However, when cells were subjected to a heat shock at 42 C, the fusion protein formed multiple small foci (Figure 3B).Because Sgt1 is a known co-chaperone of Hsp90s (Hsp82 and Hsp82), we tested if Hsp82-GFP accumulated in similar foci (Figure 3B) and found overlapping GFP-Sgt1 and Hsp82-Ruby foci (Figure S3H).Deleting the constitutive Hsp90 gene, HSC82, drastically reduced formation of the GFP-Sgt1 foci upon heat shock (Figure 3C), and inhibiting Hsp90 activity with GA prevented both Sgt1 and Hsp82 foci formation (Figure 3D).In addition, Sgt1 and Hsp82 foci formation required the Hsp70s Ssa1 and Ssa2 and the disaggregase Hsp104 (Figure 3C), but not the small heat shock protein, Hsp42, which is an essential factor for sequestration of misfolded proteins to peripheral aggregates and Q-bodies (Escusa-Toret et al., 2013;Specht et al., 2011) (Figure S3I).By co-expressing Hsp42-Ruby with GFP-Sgt1, we found that both foci partially overlap after a heat shock at 42 C, but after recovery at 30 C, Sgt1 foci resolve quickly, while Hsp42 foci remain and coalesce into few inclusions free of Sgt1 (Figure S3J).We also found that proteasome activity was required for Sgt1 and Hsp90 foci formation (Figure 3E).Therefore, we refer to these foci as Hsp70/90/ 104 and proteasome-dependent heat-induced inclusions (HAPIs) that are formed transiently during proteostatic stress and that do Report not appear to coalesce (Figures 3F and S3K).HAPIs seem to be unstable because they cannot be fixed with formaldehyde, which suggests that they are dynamic complexes that transiently colocalize with Hsp42-containing Q-bodies before such bodies coalesce into IPODs (Figure S3L).No foci were formed after heat shock by Skp1-GFP, the SCF complex backbone protein, suggesting no involvement of this complex in HAPI formation (Figure S3M).Immunoelectron microscopy of cryofixed GFP-Sgt1 cells grown at 30 C (Figure 3G) showed a rather random distribution of gold particles after labeling of the sectioned cells with GFPspecific and gold-conjugated secondary antibodies (Figure 3G).However, cells that were heat shocked displayed protein aggregation that appears as electron dense clusters in the electron micrographs.We found that gold particles displayed a higher density in these areas of protein aggregation both in the cytosol and the nucleus (Figures 3G and 3H).
Because no transient foci have been described yet for Hsp90s or its co-chaperones, we tested the behavior of GFP-tagged human Sgt1 homolog SUGT1 in HeLa cells.At 37 C, GFP-SUGT1 was evenly distributed in the cytosol but formed HAPI-likeshaped foci when cells were switched to 40 C. Again, cells treated with Hsp90 or proteasome inhibitor were unable to form foci (Figure 3I), demonstrating conservation of this phenomenon.
We performed pull-down experiments of GFP-Sgt1 from yeast cells growing at 30 C and cells heat shocked for 30 min at 42 C, and we used mass spectrometry to identify proteins that bind to GFP-Sgt1 at both 30 C and 42 C.In this group, components of the SCF complex and the Hsp90 machinery were strongly enriched (Table S2).Proteins enriched in binding to GFP-Sgt1 exclusively at 42 C have roles in endosome formation and trafficking, and many are found to be associated with the plasma membrane and nuclear periphery (Table S2; Figure 4A).One protein of this group is the heat-stress-induced Btn2, which plays an important role as a chaperone and sequestrase for misfolding proteins, a triage between refolding and degradation (Ho et al., 2019;Malinovska et al., 2012;Miller et al., 2015a).Also, Vps13 was co-purified with Sgt1 exclusively at 42 C.In human, nonfunctional Vps13 can cause the progressive movement disorder chorea-acantho-cytosis, the developmental disorder Cohen syndrome, Parkinson's disease, or spastic ataxia (Kolehmainen et al., 2003;Lesage et al., 2016;Rampoldi et al., 2001;Seong et al., 2018).Because Vps13 is concentrated at membrane contact sites, Vps13 foci can already be observed before heat shock in a strain harboring a Vps13-GFP (Lang et al., 2015).These foci increased drastically in number after heat shock at 42 C and co-localized transiently with Hsp82 (Figure 4B).However, different from Hsp82, increased foci formation of Vps13 could also be observed with the Hsp90 inhibitor GA (Figure 4B).During recovery from heat shock, Vps13 foci were rapidly reduced to normal levels.However, under conditions when Hsp82 or Sgt1 was not able to form heat-induced foci, e.g., GA treatment, the clearance of Vps13 foci was markedly retarded (Figure 4C).Similar results were obtained when the clearance of foci following a heat shock was examined using the disaggregase Hsp104 (Figure S4A) and the aggregation-prone protein Guk1-7 (Babazadeh et al., 2019) (Figure S4B).
A protein interacting with Sgt1 at 30 C and 42 C was the Hsp70/Hsp90 co-chaperone Sti1 (Wolfe et al., 2013) (Table S2).With its TPR domain-containing regions, Sti1 and its mammalian ortholog HOP (Hsp70-Hsp90 organizing protein) have been shown to bind to the EEVD motives of Hsp70s and Hsp90s to form a ternary complex (Herna ´ndez et al., 2002).No HAPI formation could be observed in strains deleted for STI1 (Figure S4C), indicating the need for this Hsp70-and Hsp90-linking protein in the recruitment of the Hsp90 machinery to HAPIs.
The aggregated form of the Parkinson's disease protein, synphilin-1, co-localizes with Sgt1 and requires Sgt1 for its clearance When the recombinant synphilin-1 protein of Parkinson's disease (Wakabayashi et al., 2000) is expressed in exponentially growing yeast cells, only a small fraction of synphilin-1 forms inclusions (B€ uttner et al., 2010).Upon shift to 42 C, all synphilin-1 formed foci, which were rapidly cleared when the temperature was lowered to 30 C, and this clearance was again dependent on a functional Hsp90 system (Figure 4D).When co-expressed with GFP-tagged Sgt1, a high number of heat-shocked cells displayed co-localization of synphilin-1 and Sgt1.This  S4D).In sgt1-3 cells, resolution of synphilin-1 foci was markedly retarded (Figure 4E), suggesting that Sgt1 needs to be fully functional to ensure rapid clearance of these inclusions.

DISCUSSION
The Hsp90s make up 1%-2% of total cellular protein and increase to 4%-6% in stressed cells (Prodromou, 2016), but their exact role in handling misfolded proteins and aggregates remains obscure.Although degradation of VHL was shown to be delayed in Hsp90 mutant cells (McClellan et al., 2005), degradation of variants of the misfolding cytosolic model substrate DssCPY* was shown to be independent of Hsp90s (Park et al., 2007).Here we show that the Hsp90 co-chaperone Sgt1 is required for degradation and together with Hsp90s is necessary for aggregate removal of misfolded proteins.Moreover, we found that Hsp90 and Sgt1 are recruited to a transient site early  S2.
during heat shock, which appears to be distinct from Q-bodies and IPODs.The peculiarity of HAPI entails the requirement for both proteasome and Hsp70/90/104 activity for HAPI to form (Figures 3C-3E).In contrast, Q-bodies and IPODs appear in higher number and are more persistent in cells with reduced Hsp90 activity and upon inhibition of the 26S proteasome (Escusa-Toret et al., 2013).The foci formed by Sgt1 and Hsp82 are transient; when cells are switched back to 30 C, almost all HAPI are rapidly cleared, while Hsp104-associated foci persist longer (Figure 3F), undergo coalescence into IPODs, and cannot be inhibited by addition of GA (Figure S4A).The observation that HAPI cannot be visualized when cells are fixed after heat shock substantiates its different physical properties (Figure S3L).Previously, an Hsp82-GFP fusion has been used as a marker for Q-bodies formed after heat shock (Saarikangas and Barral, 2015), which raises the question of whether the foci analyzed were Q-bodies, HAPIs, or both.We believe that they may have been both because we found that Sgt1 foci co-localize, like Q-bodies, with all the reporters of aggregates used, and using EM we found that Sgt1 was always associated with aggregates in the cytosol (and nucleus) early during a heat shock.Thus, we propose that HAPIs are transiently associated with Q-bodies during heat stress but, in contrast with Q-bodies, require full Hsp90 and proteasome activity to form.That HAPIs, but not Q-bodies, are formed in Hsp42 mutants indicates that these two subcellular complexes are formed independently before their association.
The role of Sgt1 in aggregate clearance appears mechanistically linked to its role in the degradation of misfolded proteins, and reduced Sgt1 activity results in a drastically impaired proteasomal degradation of misfolded proteins.This process can be partially rescued by overexpression of the proteasome-bound E3/E4 ubiquitin ligase Hul5, which was shown previously to increase the processivity of the proteasome (Crosas et al., 2006;Kuo and Goldberg, 2017).Partial degradation products of the substrates DssCL* and CTL* accumulate in sgt1-3 cells just like in hul5D cells (Figures 2A and 2B), and even though the misfolding protein DssCL* becomes highly ubiquitylated in the sgt1-3 mutant, its degradation is markedly inhibited.This indicates that proper Sgt1 function is needed also for the transfer of substrates to the 26S proteasome.HUL5 overexpression can rescue this accrual of ubiquitylated substrates, probably by its own ubiquitylation activity (Figure S2C).With reduced activity of Sgt1, Hsp90-dependent protein folding cycles are most likely slowed down, but substrates that are already marked with ubiquitin may become further ubiquitylated by Hul5 activity (Fang et al., 2011), which causes transfer from the folding machinery to degradation by the proteasome.Interestingly, aggregate formation is drastically increased in mutants of SGT1 during replicative aging, and the replicative lifespan of SGT1 mutant cells is reduced.Overexpression of HUL5 partly counteracted accelerated aging in Sgt1-deficient cells, indicating that an interplay between the Hsp90/Sgt1 and Hul5 systems is important for longevity assurance.
The role of the Hsp90 co-chaperone Sgt1 in aggregate management and protein degradation is interesting in view of data linking Hsp90 activity to neurodegenerative diseases.Several studies have proposed that Hsp90s stabilize aberrant disease-associated proteins, and that inhibition of these chaperones could redirect neuronal aggregated proteins for degradation (Luo et al., 2010).Other studies have demonstrated that the Hsp90 co-chaperones, CyP40 and PP5, reduce tau pathology but are repressed in aged and Alzheimer disease patients (Shelton et al., 2017).Interestingly, human Sgt1 concentrations have been shown to be drastically reduced in the brains of Alzheimer disease patients (Spiechowicz et al., 2006).In contrast, a recent study showed that Sgt1 increases in the brain of Parkinson's disease patients (Bohush et al., 2019).The impact of such an increase or decrease is not known, but we show here that aggregates of the Parkinson's disease protein, synphilin-1, co-localize with Sgt1 in HAPIs, and that Sgt1 deficiency and conditions that prevent HAPI formation retard clearance of synphilin-1 aggregates.Based on these observations, we believe the role of Hsp90s, Sgt1, and HAPI formation deserves future attention in the field of neurodegeneration and PQC.

Report
Tris/HCl pH 6.8, 2,5 mM EDTA/NaOH pH 8.0, 2% (w/v) Na-Dodecylsulfat (SDS), 0.05% (w/v) bromphenol blue).Prior to use 1% (v/v) beta-mercaptoethanol and complete protease inhibitor (Roche) was added to urea loading buffer.After incubation for 10 min at 70 C, samples were centrifuged at 13000xg for 1 min.15 ml were loaded on a 4%-12% gradient 26 well Criterion XT Bis-Tris Protein gel (Bio-Rad).Gels were transferred on PVDF membrane (Millipore) with a wet blotting system (Criterion Blotter, Bio-Rad).The blots were incubated in Odyssey blocking buffer (LI-COR) for 1 h at room temperature prior to probing with primary antibodies in PBS-T.Membranes were washed and incubated with the appropriate secondary antibody (LI-COR, IRdye secondary antibodies).Membranes were scanned on a LI-COR Odyssey scanner, and western blots were quantified using ImageJ (NIH).Please see Key Resources Table for antibodies, reagents, and equipment details.

Replicative lifespan assay
Exponentially growing cells were plated on YPD plates and allowed to recover before assayed for replicative lifespan.A micromanipulator (MSM 400, Singer instruments) was used to select mother cells and to remove their daughters for assessing of replicative age.

Isolation of old mother cells
Old cells expressing Hsp104-GFP were isolated using the magnetabind biotin-streptavidin system according to established protocols (Hill et al., 2016;Sinclair and Guarente, 1997;Smeal et al., 1996).Biotin labeled cells were isolated after culturing for one day.The median age of the old cells was determined by counting of bud scars after formaldehyde fixation and staining cells with 10 mg/mL Wheat Germ Agglutinin Alexa Fluorâ 555 conjugate (WGA, Life Technologies).
Assessment of colony forming units (CFU) Cells were grown in CSM to mid-exponential phase at 30 C. Cells were shifted to 42 C for 30 min, and then allowed to recover at 30 C for 90 min.Control cells were kept growing at 30 C during that period and then adjusted to same optical density.Cells were diluted sequentially one to hundred, followed by a two or ten-fold dilution.100 ml cell suspension were plated on CSM agar plates and number of CFU assessed after growth for 5 days at 22 C.
Hsp104, Hsp42 and Guk1-7 protein aggregation induction Fluorescent protein tagged cells were grown to mid-exponential phase at 30 C in CSM and then shifted to 38 C, or to 42 C for 30 min and then back to 30 C. Samples were taken at indicated time points, fixed with formaldehyde (3.7% final concentration) for 30 min, washed three times with PBS and observed by fluorescence microscopy (see below).
Sgt1, Hsp82, Vps13, Hsp42-, and synphilin-1 foci formation assay Fluorescent protein tagged cells were grown in CSM (or SD-Ura in case of dsRed-synphilin-1) to mid-exponential phase at 30 C. If cells were to be treated with 26S proteasome inhibitor MG132 (75 mM final concentration, Enzolifesciences), SDS was added to growing cells to a final concentration of 0.003% three hours prior to addition of the drug and heat shock (Liu et al., 2007).Geldanamycin (GA) was added to a final concentration of 70 mM prior to heat shock (Theodoraki et al., 2012).For induction of HAPI or Vps13 and synphilin-1 foci, cells were shifted to 42 C water bath for 30 min.Then cells were briefly spun at 5000xg and observed with a Zeiss Axio Observer Z1 inverted fluorescence microscope, using Plan Apo 100X oil objective NA:1.4 and the following filter sets: 38 HEeGFP, 45 HQ TexasRed.

Electron microscopy
High pressure freezing and freeze substitution of yeast cells Exponentially growing GFP-Sgt1 expressing yeast cells were grown at 30 C or shifted to 42 C for 30 min prior to harvesting by filtering and high pressure freezing in a Wohlwend Compact 03 (M.Wohlwend GmbH, Sennwald, Switzerland).Freeze substitution was carried out in a Leica EM AFS2 (Leica Microsystems, Vienna, Austria) using 2% uranyl acetate dissolved in 10% methanol and 90% acetone for 1 h at À90 C (Ho ¨o ¨g et al., 2014).The temperature was raised to À50 C, 2.9 C per hour, during two washes in acetone.The cells were infiltrated with Lowicryl HM20 (Polysciences, Warrington, PA) mixed with acetone (1:4, 2:3, 1:1, 4:1) and three times with pure Lowicryl, each step lasting 2 hours.The resin was polymerized with UV light 72 h at À50 C followed by 24 hr at room temperature.70 nm ultra-thin sections were produced using a Reichert-Jung Ultracut E Ultramicrotome (C.Reichert, Vienna, Austria) and an ultra 45 diamond knife (Diatome, Biel, Switzerland).The thin sections were placed on cupper grids coated with 1% Formvar and on-section contrast stained.Micrographs were taken on a Tecnai T12 electron microscope equipped with a Ceta CMOS 16M camera (FEI Co., Eindhoven, the Netherlands) operated at 120 kV.

Immunoelectron microscopy
Grids were fixed in 1% paraformaldehyde in PBS for 10 min at room temperature, washed 3 3 1 min in PBS, and blocked for 1 h in 0.8% BSA + 0.1% fish skin gelatin in PBS at room temperature.For detection of GFP-Sgt1, grids were incubated in a 1 to 30 dilution of rabbit anti-GFP (ab6556, abcam, Cambridge, UK) at 4 C over-night followed by a 1 to 20 dilution of goat anti-rabbit 10 nm gold (#25108, Electron Microscopy Sciences, Hatfield, PA) for 1 h at room temperature.3 3 20 min wash steps were carried out in PBS after incubations with each antibody.Antibodies were fixed in 1% glutaraldehyde in dH 2 O for 1 h and washed 3 3 1 min in dH 2 O. 2% uranyl acetate and Reynold's lead citrate were used for on-section contrast staining (Reynolds, 1963).Imaging as above.

Mass spectrometry
Sample preparation and TMT labeling Cysteines were reduced with dithiothreitol (Biomol) at 56 C for 30 min (10 mM in 50 mM HEPES (Biomol), pH 8.5), and further alkylated with iodoacetamide (Merck) at room temperature, in the dark for 30 min (20 mM in 50 mM HEPES, pH 8.5).For sample clean up and digestion, the SP3 protocol (Hughes et al., 2014) was used and trypsin (sequencing grade, Promega) was added (enzyme to protein ratio 1:20) for overnight digestion at 37 C. Next day, peptides were extracted and labeled with TMT10plex Isobaric Label Reagent (ThermoFisher) according the manufacturer's instructions.For further sample clean up an OASISâ HLB mElution Plate (Waters) was used.Offline high pH reverse phase fractionation was carried out on an Agilent 1200 Infinity high-performance liquid chromatography system, equipped with a Gemini C18 column (3 mm, 110 A ˚, 100 3 1.0 mm, Phenomenex).
Trapping was carried out with a constant flow of solvent A (0.1% formic acid in water) at 30 mL/min onto the trapping column for 6 min.Subsequently, peptides were eluted via the analytical column with a constant flow of 0.3 mL/min with increasing percentage of solvent B (0.1% formic acid in acetonitrile) from 2% to 4% in 4 min, from 4% to 8% in 2 min, then 8% to 28% for a further 96 min, and finally from 28% to 40% in another 10 min.The outlet of the analytical column was coupled directly to a QExactive plus (Thermo Fisher) mass spectrometer using the proxeon nanoflow source in positive ion mode.The peptides were introduced into the QExactive plus via a Pico-Tip Emitter 360 mm OD x 20 mm ID; 10 mm tip (New Objective) and an applied spray voltage of 2.3 kV.The capillary temperature was set at 320 C. Full mass scan was acquired with mass range 375-1200 m/z in profile mode in the FT with resolution of 70000.The filling time was set at maximum of 250 ms with a limitation of 3x10 6 ions.Data dependent acquisition (DDA) was performed with the resolution of the Orbitrap set to 35000, with a fill time of 120 ms and a limitation of 2x10 5 ions.A normalized collision energy of 32 was applied.A loop count of 10 with count 1 was used and a minimum AGC trigger of 2e 2 was set.Dynamic exclusion time of 30 s was used.The peptide match algorithm was set to 'preferred' and charge exclusion 'unassigned', charge states 1, 5 -8 were excluded.MS 2 data was acquired in profile mode.

MS data analysis
IsobarQuant (Franken et al., 2015) and Mascot (v2.2.07) were used to process the acquired data, which was searched against a Uniprot Saccharomyces cerevisiae proteome database (UP000002311) containing common contaminants and reversed sequences.The following modifications were included into the search parameters: Carbamidomethyl (C) and TMT10 (K) (fixed modification), Acetyl (N-term), Oxidation (M) and TMT10 (N-term) (variable modifications).For the full scan (MS1) a mass error tolerance of 10 ppm and for MS/MS (MS2) spectra of 0.02 Da was set.Further parameters were set: Trypsin as protease with an allowance of maximum two missed cleavages: a minimum peptide length of seven amino acids; at least two unique peptides were required for a protein identification.The false discovery rate on peptide and protein level was set to 0.01.The protein output files of IsobarQuant were processed using the R programming language (https://www.r-project.org).As a quality filter, only proteins that were quantified with at least 2 unique peptides were used for the downstream analysis.Raw TMT reporter ion signals (signal_sum columns) were first batch-cleaned using the removeBatchEffect function from the limma package (Ritchie et al., 2015) and further normalized using the vsn package (variance stabilization normalization; Huber et al., 2002).Missing values were imputed using the knn option of the impute function using the Msnbase package (Gatto and Lilley, 2012).Proteins were tested for differential expression using limma again and called hits with a false discovery smaller 1% and a fold-change cut-off of 100%.The mass spectrometry proteomics data have been deposited to the ProteomeXchange Consortium via the PRIDE (Perez-Riverol et al., 2019) partner repository with the dataset identifier PXD016174 (See also Key Resources Table ).

QUANTIFICATION AND STATISTICAL ANALYSIS
To identify network regions significantly enriched for DssCL* and VHL-Leu2 stabilization, as well as for physical interaction with Sgt1 at 30 C and 42 C, hits were analyzed with the help of TheCellMap.org.
Lifespan assays were done in at least two replicates.Statistical analysis was performed using Logrank (Mantel-Cox test) in Graph-Pad Prismâ 8.2.1 (Figures 2E,2F,and S2F).Statistical analysis of Sgt1 localization by immunoelectron microscopy performed using Wilcoxon matched-pairs signed rank test in GraphPad Prismâ 8.2.1 (Figure 3H).
All data in the bar graphs are presented as an average of n R 3 replicates ± SEM.In figures, asterisks denote statistical significance as calculated by Student's t test *p < 0.05, **p < 0.005, ***p < 0.0005, unpaired two tailed t test using MS Excel.

Figure 1 .
Figure 1.Identification of the Hsp90 co-chaperone Sgt1 as an essential protein required for the degradation of misfolded cytosolic proteins (A) SAFE analysis performed of common hits from two SGA screens of strains expressing one of the chimeric cytosolic substrates, DssCL* or VHL-LEU2, of the 26S proteasome.A cut-off value of 5eÀ2 was chosen for visualization at https://thecellmap.org.(B) Spot test of different strains expressing DssCL*myc on media lacking leucine and uracil (-Leu -Ura) or uracil (-Ura) only as a control.(C and D) Analysis of DssCL* and DssCG* degradation during a cycloheximide chase at 30 C. Pgk1 served as a loading control.(C) Band 1 indicates full-length and bands 2 and 3 indicate C-terminal partial degradation products of DssCL* detected with antibodies specific for the C-terminal myc tag.(D) DssCG* was detected with antibodies specific for its C-terminal GFP tag.(C and D) Average of three biological replicates was used for quantification.Error bars indicate standard error of the mean (SEM).See also Figure S1 and TableS1.

Figure 2 .
Figure 2. Reduction in processing of misfolded proteins of the cytoplasm and the ER and of fitness and replicative lifespan caused by SGT1 deficiency can be partially suppressed by overexpressing HUL5 (A-C) Cycloheximide chase analysis of DssCL* (A), CTL* (B), and CPY*-HA (C) in indicated strains grown at 30 C. Full-length band 1 of DssCL* and CTL* detected with anti-myc, CPY*-HA with anti-HA antibodies.Pgk1 served as a loading control.Bars represent relative average protein levels, and error bars indicate SEM of three biological replicates.(D) 5-fold dilution series of indicated strains on YPD incubated at indicated temperatures.(E) Replicative lifespan of sgt1-3 (n = 96) and sgt1-3 HUL5 overexpression strain (n = 101; p < 0.0001, log-rank [Mantel-Cox] test).(F) Replicative lifespan of wild-type (Wt) (n = 102) and HUL5 overexpression strain (n = 124; p < 0.0001, log-rank [Mantel-Cox] test).Numbers in parentheses indicate the median replicative lifespan of each strain.See also Figure S2.

Figure 3 .
Figure 3. Sgt1 is required for aggregate coalescence and clearance during stress, and it clusters upon heat stress together with Hsp82 in an Hsp90 and proteasome activity-dependent process to Q-bodies (A) Exponentially growing cells expressing Hsp104-GFP were shifted from 30 C to 38 C for 90 min.(A and F) Bars display total amount of cells with foci and are divided into three types dependent on the amount of foci per cell (type 1 cells [blue] containing one, type 2 cells [red] containing two, or type 3 cells [green] containing three or more foci).(B) Representative pictures of exponentially growing GFP-Sgt1-or Hsp82-GFP-expressing cells were shifted from 30 C to 42 C for 30 min and analyzed by fluorescence microscopy for foci formation.Scale bar indicates 5 mm.(C) Percentage of cells forming GFP-Sgt1 or Hsp82-GFP foci in indicated strains before and after heat shock.(D and E) GFP-Sgt1 or Hsp82-GFP foci formation upon inhibition of Hsp90 activity by adding 70 mM geldanamycin (GA) (D) and proteasome inhibition using 150 mM MG132 when shifted to 42 C (E). DMSO served as a solvent of both drugs and as a negative control.(F) Clearance of heat shock-induced (42 C for 30 min) Hsp104-GFP and GFP-Sgt1 foci after 30 and 60 min recovery at 30 C. Bars are described under (A).(G) Thin sections of GFP-Sgt1-expressing yeast cells grown at 30 C or shifted to 42 C for 30 min.Sections were immunogold labeled for visualization of Sgt1 localization.Gold particles are marked with arrowheads.A, electron-dense regions representing protein aggregates; C, cytosol; M, mitochondria, N, nucleus; V, vacuole.Scale bars indicate 100 nm.(H) Sgt1 localization (gold labeling density) inside and outside of cytosolic and nuclear protein aggregates (electron-dense clusters) were assessed at 30 C (n = 20 cells) and 42 C (n = 19 cells).Wilcoxon matched-pairs signed rank test, ****p < 0.0001, *p = 0.0105.(I) Foci formation of human Sgt1 (EGFP-SUGT1) in HeLa cells.Cells were shifted to 40 C and mock treated (DMSO), or GA or MG132 treated.Inset shows a zoom for better visualization of EGFP-SUGT1 foci.EGFP-SUGT1-expressing cells grown at 37 C (no HS) or EGFP after 30 min heat shock at 40 C are shown as negative controls.All data presented in the bar graphs are an average of n R 3 biological replicates ± SEM.See also Figure S3.

Figure 4 .
Figure 4. Physical interactors of Sgt1 at 30 C and 42 C and how their properties are influenced by HAPI formation (A) Mass spectrometry analysis of GFP-Sgt1 co-purifying proteins from cells grown at 30 C or heat shocked at 42 C for 30 min (n = 4 biological replicates).SAFE analysis of hits from the screen that bind at 30 C and 42 C (green), exclusively at 30 C (purple) and at 42 C (yellow).A cut-off value of 5eÀ2 was chosen for visualization at https://thecellmap.org.(B) Vps13-GFP foci partially overlap with Hsp82-Ruby foci when allowed to be formed (DMSO).Line intensity plot of Vps13-GFP and Hsp82-Ruby shows the distribution of relative fluorescence after heat shock across the white line.Scale bar indicates 5 mm.(C and D) Percentage of four or more Vps13-GFP foci-containing cells (C) or synphilin-1 foci-forming cells (D) was analyzed at 42 C and recovery at 30 C for indicated time points in conditions that allow HAPI formation (DMSO) compared with conditions that do not allow HAPI formation (GA).(E) Percentage of cells showing synphilin-1 foci in Wt and sgt1-3 mutant cells was analyzed at 42 C and recovery at 30 C. All data in the bar graphs presented are an average of n R 3 biological replicates ± SEM. *p < 0.05, **p < 0.005, unpaired two-tailed t test.See also Figure S4 and TableS2.