Tuning Hsp104 specificity to selectively detoxify α-synuclein

Summary

Hsp104 is an AAA+ protein disaggregase that solubilizes and reactivates proteins trapped in aggregated states. We have engineered potentiated Hsp104 variants to mitigate toxic misfolding of α-synuclein, TDP-43, and FUS implicated in fatal neurodegenerative disorders. Though potent disaggregases, these enhanced Hsp104 variants lack substrate specificity, and can have unfavorable off-target effects. Here, to lessen off-target effects, we engineer substrate-specific Hsp104 variants. By altering Hsp104 pore loops that engage substrate, we disambiguate Hsp104 variants that selectively suppress α-synuclein toxicity but not TDP-43 or FUS toxicity. Remarkably, α-synuclein-specific Hsp104 variants emerge that mitigate α-synuclein toxicity via distinct ATPase-dependent mechanisms, involving α-synuclein disaggregation or detoxification of soluble α-synuclein conformers. Importantly, both types of α-synuclein-specific Hsp104 variant reduce dopaminergic neurodegeneration in a C. elegans model of Parkinson’s disease more effectively than non-specific variants. We suggest that increasing the substrate specificity of enhanced disaggregases could be applied broadly to tailor therapeutics for neurodegenerative disease.

Graphical Abstract

eTOC blurb

In this paper, Mack et al. engineer Hsp104 variants with enhanced selectivity for α-synuclein that reduce dopaminergic neurodegeneration more effectively than non-specific Hsp104 variants. These findings suggest that increasing the substrate specificity of enhanced protein disaggregases could have broad applications in tailoring therapeutics for neurodegenerative disease.

Introduction

There are no effective therapies for several fatal neurodegenerative diseases including Parkinson’s disease (PD), amyotrophic lateral sclerosis (ALS), and frontotemporal dementia (FTD). PD is marked by misfolding of α-synuclein (α-syn), a protein normally found in presynaptic terminals, which may function in synaptic vesicle recycling.^1,2 α-Syn misfolds into toxic oligomers and amyloid fibrils, accumulating in characteristic Lewy bodies in the cytoplasm of dopaminergic neurons that degenerate in PD.^3–5 Indeed, protein misfolding and aggregation unite a spectrum of fatal neurodegenerative diseases.⁶ Thus, it is important to develop therapeutics that directly antagonize the underlying toxic protein-misfolding events in neurodegenerative disease. In this way, proteins can be restored to their native conformation and function, which may halt the debilitating trajectory of neurodegeneration.^7–11

Hsp104, an asymmetric, hexameric AAA+ protein disaggregase from yeast, is an intriguing therapeutic agent to directly target toxic protein-misfolding events in neurodegenerative disease.^12,13 Hsp104 uses energy from ATP binding and hydrolysis, and collaboration with Hsp70 and Hsp40, to reactivate proteins trapped in insoluble states.^13–16 Hsp104 hexamers are dynamic and adopt open "lock-washer" spiral states and closed ring structures that translocate polypeptides across the central channel.^14,17–20 During protein disaggregation, pore-loop tyrosines grip substrate and ATP hydrolysis-driven conformational changes at the spiral seam ratchet substrate either partially or completely through the channel.^{16,17,19,21–26} Thus, Hsp104 liberates individual polypeptides from soluble toxic oligomers, amorphous aggregates, stress-induced condensates, and amyloid fibrils, which can then regain their functional form.^14–16,27 Curiously, Hsp104 is not found in metazoa, but is conserved in eubacteria, algae, fungi, protozoa, and plants.²⁸ As humans lack an Hsp104 homolog, they have limited capacity to effectively counter overwhelming protein-misfolding events that underlie neurodegenerative disease.^7,29 Introduction of Hsp104 into animal models (e.g. worm, fly, mouse, and rat) protects against deleterious protein misfolding and neurodegeneration connected to PD and polyglutamine-expansion disorders.^27,30–33 Nonetheless, the ability of Hsp104 to counter the misfolding and toxicity of human disease-linked proteins can be limited, requiring high Hsp104 concentrations.^14,27

To address this issue, we have engineered enhanced versions of Hsp104 that more effectively disaggregate various disease-linked proteins, including α-syn (linked to PD), and TDP-43 and FUS (linked to ALS/FTD) under conditions where wild-type (WT) Hsp104 is ineffective.^34–42 Select potentiated variants reduce dopaminergic neurodegeneration in a C. elegans model of PD³⁴ and reverse FUS aggregation and toxicity in mammalian cells.⁴³ Typically, potentiated Hsp104 variants exhibit elevated ATPase activity, altered protomer cooperativity, altered substrate recognition, and prolonged substrate interactions, which enables more productive disaggregase activity.^26,34,35,44 Though powerful disaggregases, these potentiated Hsp104 variants can exhibit off-target toxicity, which may limit their progression along the therapeutic pipeline through more complex model systems.^34,45 This toxicity is likely caused by aberrant unfolding of essential substrates, resulting in unwanted off-target effects.³⁵ Thus, enhanced substrate specificity is a desirable attribute for Hsp104 variants to effectively translate into higher organisms as therapeutics.⁴⁶

Here, we hypothesized that specifically mutating Hsp104 residues known to contact substrate in a potentiated variant background would couple increased substrate-specificity to enhanced disaggregase activity. We found that specific alterations to pore-loop tyrosines that engage substrate directly endowed Hsp104 with the ability to selectively mitigate α-syn toxicity. Surprisingly, two classes of α-syn-specific Hsp104 variant emerged. The first class mitigated α-syn toxicity via ATPase-dependent disaggregation of α-syn inclusions. By contrast, an unanticipated second class mitigated α-syn toxicity via ATPase-dependent detoxification of soluble α-syn conformers without disaggregation of α-syn inclusions. Importantly, both types of α-syn-specific Hsp104 variant reduced dopaminergic neurodegeneration in a C. elegans model of PD more effectively than non-specific Hsp104 variants. Thus, we establish an important concept: specializing protein disaggregases against individual disease-associated substrates can improve their therapeutic utility. We anticipate that this concept can be applied broadly to diverse protein disaggregases and specific targets implicated in protein-misfolding disorders. In this way, specific toxic misfolding events could be remediated with tailor-made therapeutic disaggregases.

Results

Targeting Hsp104 pore-loop tyrosines via a rational engineering approach

Hsp104 consists of an N-terminal domain (NTD), nucleotide-binding domain 1 (NBD1), a middle domain (MD), nucleotide-binding domain 2 (NBD2), and a C-terminal domain (CTD) (Figure 1A).⁴⁷ Hsp104 forms an asymmetric, ring-like hexamer, and threads substrate through its central channel, which is lined with substrate-binding pore loops (Figure 1B–D).^17,18,21,22 Each NBD contains a highly-conserved tyrosine embedded within a pore-loop motif: ‘KYKG’ in NBD1 (residues 256-259) comprises pore-loop 1, and ‘GYVG’ in NBD2 (residues 661-664) comprises pore-loop 2^17,18,21,22 (Figure 1A–D).^17,18,21,22 The pore-loop tyrosines, Y257 in pore-loop 1 and Y662 in pore-loop 2, are essential for disaggregation and substrate threading.^{14,16–18,21,22,24} The pore loops are arranged in the Hsp104 axial channel as a ‘spiral staircase’, allowing for substrate to be translocated through the channel (Figure 1B–D).^17,18 Y257 and Y662 contact substrate directly.¹⁷ We reasoned that subtly changing the properties of these highly-conserved tyrosines would enable us to tune the substrate repertoire of Hsp104. We altered each pore-loop tyrosine to a series of hydrophobic, aromatic, or uncharged polar residues to preserve different features of the original tyrosine residue, which was found to be optimal for protein disaggregation.^24,34 Using this rational engineering approach, we isolated α-syn-specific Hsp104 variants.

Figure 1. Pore-loop substitutions in Hsp104^A503S do not generate substrate-specific Hsp104 variants.

(A) Domain map of Hsp104 showing N-terminal domain (N; green), nucleotide-binding domain 1 (NBD1; purple), middle domain (MD; maroon), nucleotide-binding domain 2 (NBD2; light blue), and C-terminal domain (C, orange). Location of residues assessed in this work are shown. (B) Structure of Hsp104 hexamer (side-view; left, top-down view; right) with residues from this study indicated. Colors of domains in protomer correspond to (A). PDB: 5VY9. (C) Zooms into Hsp104 structure highlighting K358:D484 interaction, and location of these residues in relation to A503. Substrate-binding pore-loop tyrosine residues are shown (Y257 and Y662) relative to polypeptide (casein) substrate (tan) in the Hsp104 channel. PDB: 5VY9. (D) Structure of Hsp104 hexamer (gray) with two subunits omitted to reveal polypeptide (casein) substrate (yellow) in the Hsp104 channel. Pore loops (pink; Y257 in NBD1 and purple; Y662 in NBD2) line the channel of Hsp104 in a “spiral staircase” manner. Pore-loop residues are essential for substrate translocation, as they establish the main contacts between Hsp104 and substrate. PDB: 5VY9. (E) Δhsp104 yeast integrated with α-syn-YFP (top), FUS (middle), or TDP-43 (bottom) on a galactose-inducible promoter were transformed with Hsp104 variants or an empty vector control. Yeast were spotted onto glucose (uninducing, off) and galactose (inducing, on) media in a five-fold serial dilution. Stars indicate substitution to pore loop in NBD1 (pink), NBD2 (purple). (F) Integrated strains from (E) were induced in the presence of Hsp104 variants or empty vector control for 5 hours (FUS, TDP-43) or 8 hours (α-syn). Yeast were lysed and processed for Western blot. 3-Phosphoglycerate kinase (PGK) is a loading control.

Mutating pore-loop tyrosines in Hsp104^A503S does not yield α-syn-specific variants

Mutating pore-loop tyrosines in WT Hsp104 does not enable suppression of α-syn, TDP-43, or FUS toxicity in yeast.³⁴ Thus, we first set out to engineer substrate-specific Hsp104 variants using a generally potentiated Hsp104 variant, Hsp104^A503S, as a starting scaffold.³⁴ Our goal was to leverage the elevated activity of Hsp104^A503S as a starting point to introduce substitutions that alter Hsp104 substrate selectivity. We assessed Hsp104 variant activity in powerful yeast models of α-syn, TDP-43, and FUS proteinopathy, which faithfully recapitulate several aspects of neurodegenerative disease, including protein aggregation and toxicity.^48–50 Importantly, these valuable yeast models have enabled identification of genetic suppressors and drug candidates that mitigate neurodegeneration in C. elegans, fly, mouse, rat, and human patient-derived neuronal models of disease.^34,51–57

Hsp104^A503S potently suppressed α-syn, TDP-43, and FUS toxicity in yeast (Figure 1E). ³⁴ In an effort to confer substrate specificity, we first introduced hydrophobic mutations to pore-loop 1 Y257 or pore-loop 2 Y662 to maintain the hydrophobic character of the tyrosine. Thus, we substituted a leucine at pore-loop 1 Y257. Although Y257 is highly conserved, leucine is also found rarely at this position in ~0.5% of Hsp104 homologues according to our Generative Regularized ModeLs of proteINs (GREMLIN) analysis of 5,812 Hsp104 species variants.⁵⁸ Relative to Hsp104^A503S, Hsp104^A503S:Y257L displayed reduced activity, as it suppressed α-syn and FUS toxicity, but not TDP-43 toxicity in yeast (Figure 1E). Though substitution of a hydrophobic residue at pore-loop 1 Y257 reduced toxicity suppression, introducing a hydrophobic methionine residue at pore-loop 2 Y662 had a different effect. Hsp104^A503S:Y662M suppressed toxicity of α-syn, TDP-43, and FUS, but not as strongly as Hsp104^A503S (Figure 1E). Combining these pore-loop mutations in Hsp104^{A503S:Y257L:Y662M} diminished any toxicity-suppression activity (Figure 1E).

Since hydrophobic substitutions at each pore loop did not lead to enhanced substrate specificity, we evaluated the effect of an uncharged, polar variant at Y257, and so introduced a threonine substitution. Hsp104^A503S:Y257T was unable to suppress toxicity of α-syn, TDP-43, or FUS (Figure 1E). Expressing each pore-loop variant did not noticeably affect expression levels of Hsp104 or disease-associated substrates (Figure 1F). Altogether, slightly altering Hsp104 pore-loop tyrosines in the potentiated Hsp104^A503S background did not generate any variants with the desired substrate specificity. Indeed, several pore-loop mutations diminished the ability of Hsp104^A503S to mitigate proteotoxicity.

Hsp104^K358D can be tailored via tuning pore loops for substrate specificity

Hsp104^A503S is potentiated via disruption of interprotomer contacts between helix L1 and helix L3 of the MD^17,26,40,59, but was not amenable for tuning substrate specificity (Figure 1E). Thus, we wondered whether substrate selectivity could be more effectively engineered in a Hsp104 variant potentiated by a different mechanism than Hsp104^A503S. We first considered Hsp104 with a K358D mutation in NBD1 as a starting scaffold. Hsp104^K358D has a mutation that disrupts interactions between D484 in the MD and K358 in NBD1 (Figure 1C) similar to the previously-studied Hsp104^K358E.^17,60 Hsp104^K358E has elevated ATPase activity relative to Hsp104, but has been reported to be extremely toxic to yeast.⁶⁰ By contrast, Hsp104^K358D was not toxic to ∆hsp104 yeast at 30°C (Figure S1). Thus, we explored whether this Hsp104 variant could suppress toxicity of neurodegenerative disease-linked substrates, and whether we could now tune pore-loop residues to engender substrate specificity.

Hsp104^K358D suppressed α-syn, FUS, and TDP-43 toxicity in yeast (Figure 2A). Thus, Hsp104^K358D can antagonize proteotoxicity connected to PD and ALS/FTD. Using Hsp104^K358D as a starting scaffold, we systematically altered the substrate-binding, pore-loop tyrosines, Y257 and Y662, as with Hsp104^A503S. Interestingly, Hsp104^K358D:Y257L suppressed toxicity of α-syn, FUS, and TDP-43 more effectively than Hsp104^K358D (Figure 2A). Thus, a leucine at position 257 in pore-loop 1 is more productive than a tyrosine in the Hsp104^K358D background in striking contrast to Hsp104^A503S (Figure 1E, 2A). Interestingly, Hsp104^K358D:Y662L, with a leucine at pore-loop 2 Y662, did not suppress FUS or TDP-43 toxicity, but mildly suppressed α-syn toxicity (Figure 2A). Thus, increased α-syn selectivity can originate by altering Y662 in the K358D background.

Figure 2. Hsp104^K358D:Y662M selectively suppresses α-syn toxicity.

(A) Δhsp104 yeast integrated with α-syn-YFP (top), FUS (middle), or TDP-43 (bottom) on a galactose-inducible promoter were transformed with Hsp104 variants or an empty vector control. Yeast were spotted onto glucose (uninducing, off) and galactose (inducing, on) media in a five-fold serial dilution. Stars indicate substitution to pore loop in NBD1 (pink), NBD2 (purple). (B) Integrated strains from (A) were induced in the presence of Hsp104 variants or empty vector control for 5 hours (FUS, TDP-43) or 8 hours (α-syn). Yeast were lysed and lysates visualized via Western blot. 3-Phosphoglycerate kinase (PGK) is a loading control. See also Figure S1–S4.

We next explored hydrophobic, aromatic substitutions at pore loops 1 and 2. Hsp104^{K358D:Y257F/W} and Hsp104^{K358D:Y662F/W} suppressed toxicity of α-syn, FUS, and TDP-43 (Figure 2A). We also substituted hydrophobic residues isoleucine and valine at the pore-loop tyrosines. Hsp104^{K358D:Y257I/V} slightly suppressed α-syn toxicity and did not suppress FUS or TDP-43 toxicity (Figure 2A). Thus, α-syn selectivity can also emerge by mutating Y257 in the K358D background. By contrast, Hsp104^{K358D:Y662I/V} did not suppress α-syn, FUS, or TDP-43 toxicity (Figure 2A). Interestingly, even a minor change from leucine to isoleucine at position 662 in the K358D background resulted in diminished ability to suppress α-syn toxicity.

Generally, our pore-loop Hsp104 variants did not grossly affect α-syn or TDP-43 expression levels (Figure 2B). A subset of these variants mildly reduced FUS expression (Figure 2B), but we have shown before that this mild reduction in FUS levels is not required to suppress FUS toxicity by enhanced Hsp104 variants.^35,40 Collectively, our data suggest that engineered pore-loop variants do not suppress disease-associated substrate toxicity by severely lowering substrate levels.

Given the changes in toxicity suppression when a single pore-loop tyrosine was mutated in the Hsp104^K358D background, we next tested the effect of double pore-loop tyrosine mutations. As Hsp104^K358D:Y257L showed robust suppression of α-syn, FUS, and TDP-43 toxicity, we used this variant as a starting scaffold to tune the second pore-loop tyrosine. Hsp104^{K358D:Y257L:Y662L} suppressed toxicity of α-syn, but only slightly reduced FUS and TDP-43 toxicity (Figure 3A). Hsp104^{K358D:Y257L:Y662F/W} suppressed toxicity of all three substrates, consistent with the finding that phenylalanine and tryptophan substitutions at either pore loop alone do not affect the activity of enhanced Hsp104 variants (Figure 3A).³⁴ Interestingly, introducing polar, uncharged residues (S, T, Q, N) at pore-loop 2 Y662 ablated the toxicity suppression activity of Hsp104^K358D:Y257L, again suggesting that pore-loop 2 Y662 is not as amenable to mutation as pore-loop 1 Y257 (Figure 3A). Overall, the double pore-loop Hsp104 variants were similar to those with single pore-loop mutations in that aside from a set of variants that mildly reduced FUS expression, the variants did not grossly affect disease-associated substrate levels (Figure 3B).

Figure 3. Hsp104^{K358D:Y257L:Y662M} selectively suppresses α-syn toxicity.

(A) Δhsp104 yeast integrated with α-syn-YFP (top), FUS (middle), or TDP-43 (bottom) on a galactose-inducible promoter were transformed with Hsp104 variants or an empty vector control. Yeast were spotted onto glucose (uninducing, off) and galactose (inducing, on) media in a five-fold serial dilution. Stars indicate substitution to pore loop in NBD1 (pink), NBD2 (purple). (B) Integrated strains from (A) were induced in the presence of Hsp104 variants or empty vector control for 5 hours (FUS, TDP-43) or 8 hours (α-syn). Yeast were lysed and lysates visualized via Western blot. 3-Phosphoglycerate kinase (PGK) is a loading control. See also Figure S3 and S4.

Our rational approach to altering the substrate repertoire of Hsp104 revealed several trends for pore-loop substitutions that are favorable for general potentiation. In the enhanced Hsp104^K358D background, aromatic residues (W, F) at Y257 or Y662 maintained enhanced toxicity suppression activity (Figure 2A). Substituting aromatic residues for Y662 in the potentiated Hsp104^K358D:Y257L background also maintained potentiated activity (Figure 3A). Hydrophobic residues (I, V) at Y257 in the Hsp104^K358D background were not favorable for potentiated activity against TDP-43 or FUS but did yield more selective variants that mildly suppressed α-syn toxicity (Figure 2A). Additionally, uncharged, polar residues at Y662 in Hsp104^K358D:Y257L eliminated potentiated activity (Figure 3A).

Y662M in pore-loop 2 of Hsp104^K358D or Hsp104^K358D:Y257L confers α-syn selectivity

We unearthed further α-syn-specific variants by substituting a Met residue at the NBD2 pore-loop tyrosine. Thus, a Y662M substitution at pore-loop 2 in the Hsp104^K358D or Hsp104^K358D:Y257L created variants that selectively suppressed α-syn toxicity (Figure 2A, 3A). Indeed, Hsp104^K358D:Y662M and Hsp104^{K358D:Y257L:Y662M} suppressed α-syn toxicity but were ineffective against FUS or TDP-43 toxicity (Figure 2A, 3A). These variants did not grossly affect α-syn expression levels (Figure 2B, 3B). Interestingly, introducing a Met residue at pore-loop 1 (Y257) in the Hsp104^K358D background did not confer substrate specificity, but rather strengthened suppression of α-syn, FUS, and TDP-43 toxicity (Figure 2A). Thus, replacing Tyr with Met in pore-loop 1 versus pore-loop 2 has distinct effects on the substrate selectivity of Hsp104.

Hsp104^D484K can also be tailored via tuning pore loops for substrate specificity

In addition to Hsp104^K358D, Hsp104^D484K is another Hsp104 variant that disrupts MD-NBD1 interactions normally mediated by K358 in NBD1 and D484 in helix L2 of the MD (Figure 1C).⁶⁰ To our surprise, Hsp104^D484K was not toxic to yeast at 30°C (Figure S1). Thus, neither Hsp104^K358D nor Hsp104^D484K are overtly toxic to yeast at 30°C. We generated a set of pore-loop Hsp104^D484K variants, which displayed very similar effects on α-syn, FUS, and TDP-43 toxicity and expression to those studied in the Hsp104^K358D background (Figure S2A, B). For example, like Hsp104^{K358D:Y257L/M}, Hsp104^{D484K:Y257L/M} enhanced activity against α-syn, FUS, and TDP-43 (Figure S2A). Furthermore, mutation of Y257 to I or V, or Y662 to L or M yielded Hsp104^D484K variants that selectively suppressed α-syn toxicity (Figure S2A). Thus, in potentiated backgrounds created by disrupting NBD1 and MD contacts (e.g., ^K358D and D484K) we establish general rules for: (a) mutating pore loops to enhance suppression of α-syn, FUS, and TDP-43 toxicity; and (b) mutating pore loops to selectively suppress α-syn toxicity. Specifically, enhanced activity is conferred by Y257L/M mutations and α-syn-selectivity is conferred by Y257I/V mutation or Y662L/M mutations.

Y257T/Q in pore-loop 1 of Hsp104^K358D confer α-syn selectivity

We next introduced alternative uncharged, polar residues Thr, Gln, Ser, or Asn at pore-loop 1, Y257. As these substitutions still maintain the polar, uncharged character of tyrosine we reasoned they might tune the substrate repertoire of Hsp104. Remarkably, Hsp104^{K358D:Y257T/Q} specifically suppressed α-syn toxicity and did not suppress FUS or TDP-43 toxicity (Figure 4A). Furthermore, Hsp104^{K358D:Y257T/Q} did not reduce α-syn expression levels (Figure 4B). Interestingly, Hsp104^{K358D:Y257S/N} did not effectively suppress α-syn, TDP-43, or FUS toxicity (Figure 4A). We focused on Hsp104^K358D:Y257T as a representative among the uncharged, polar (Hsp104^{K358D:Y257T/Q}) and hydrophobic (Hsp104^{K358D:Y257I/V}) α-syn-specific variants. This variant suppressed α-syn toxicity to a lesser extent than Hsp104^K358D:Y662M (Figure 2A, 4A), revealing that even among α-syn-specific variants, we can tune the degree of toxicity suppression.

Figure 4. Hsp104^{K358D:Y257T/Q} selectively suppress α-syn toxicity.

(A) Δhsp104 yeast integrated with α-syn-YFP (top), FUS (middle), or TDP-43 (bottom) on a galactose-inducible promoter were transformed with Hsp104 variants or an empty vector control. Yeast were spotted onto glucose (uninducing, off) and galactose (inducing, on) media in a five-fold serial dilution. Stars indicate substitution to pore loop in NBD1 (pink), NBD2 (purple). (B) Integrated strains from (A) were induced in the presence of Hsp104 variants or empty vector control for 5 hours (FUS, TDP-43) or 8 hours (α-syn). Yeast were lysed and lysates visualized via Western blot. 3-Phosphoglycerate kinase (PGK) is a loading control. See also Figure S3 and S4.

α-Syn-specific Hsp104 variants suppress α-syn toxicity through an ATPase-dependent mechanism

We next determined whether α-syn-specific Hsp104 variants suppressed α-syn toxicity via an ATP hydrolysis-dependent process as for other potentiated Hsp104 variants.^34,39 Thus, we made double Walker A (DWA) mutations (K218T and K620T) that render Hsp104 unable to bind ATP at NBD1 and NBD2.^{14,39,61–64} The DWA mutations inhibited potentiated variants Hsp104^A503S, Hsp104^K358D, and Hsp104^K358D:Y257L from suppressing toxicity of α-syn, FUS, and TDP-43 (Figure S3A). DWA mutations in Hsp104^K358D:Y257T diminished suppression of α-syn toxicity (Figure S3A), suggesting Hsp104^K358D:Y257T also employs an ATP hydrolysis-dependent mechanism. Hsp104^K358D:Y662M and Hsp104^{K358D:Y257L:Y662M} were also inactivated by DWA mutations (Figure S3A). Thus, these α-syn-specific Hsp104 variants also work through a mechanism reliant on ATP hydrolysis. The reduction in therapeutic Hsp104 activity conferred by DWA mutations was not due to reduced Hsp104 levels (Figure S3B). Overall, these findings suggest that α-syn-specific Hsp104 variants suppress α-syn toxicity through an ATPase-dependent mechanism and not via a passive mechanism as with select Hsp104 species variants.⁶⁵

α-Syn-specific Hsp104 variants do not suppress toxicity of PD-linked α-syn^E46K or α-syn^A53T

We next tested whether α-syn-specific Hsp104 variants could suppress the toxicity of α-syn variants, α-syn^E46K and α-syn^A53T, which cause rare familial forms of PD (Figure S4A).^66,67 None of the Hsp104 variants grossly affected Hsp104 or α-syn expression (Figure S4B). As anticipated, Hsp104 was unable to mitigate toxicity of α-syn^E46K and α-syn^A53T, whereas Hsp104^A503S reduced α-syn^E46K and α-syn^A53T toxicity (Figure S4B).³⁵ By contrast, Hsp104^K358D weakly reduced α-syn^E46K toxicity, but not α-syn^A53T toxicity, indicating that α-syn^A53T toxicity is more challenging to suppress (Figure S4B). Addition of the Y257L mutation to the Hsp104^K358D background enabled suppression of α-syn^E46K and α-syn^A53T toxicity (Figure S4B). This finding provides further evidence that a leucine at position 257 in pore-loop 1 is more productive than a tyrosine for mitigating α-syn toxicity in the Hsp104^K358D background. Interestingly, the α-syn-specific Hsp104 variants, Hsp104^K358D:Y257T, Hsp104^K358D:Y662M, and Hsp104^{K358D:Y257L:Y662M} could not mitigate α-syn^E46K or α-syn^A53T toxicity (Figure S4B). Thus, the PD-linked E46K and A53T mutations in the amphipathic region of α-syn reduce the efficacy of Hsp104^K358D:Y257T, Hsp104^K358D:Y662M, and Hsp104^{K358D:Y257L:Y662M}.

Determinants in the α-syn C-terminal acidic domain enable toxicity suppression by α-syn-selective Hsp104 variants

Hsp104-binding peptides can be enriched in aromatic residues, uncharged polar residues, as well as basic and acidic residues²¹. Hence, we wondered whether the acidic C-terminal domain of α-syn might enable toxicity mitigation by Hsp104 variants. Thus, we expressed α-syn constructs encompassing residues 1-95, 1-115, or 1-125 in yeast (Figure S4A). α-Syn^1-95 lacks the entire C-terminal acidic region (amino acids 95-140), whereas α-syn^1-115 and α-syn^1-125 lack C-terminal portions of the acidic region and are naturally occurring fragments of α-syn that are more aggregation prone (Figure S4A).⁶⁸ α-Syn^1-95, α-syn^1-115, and α-syn^1-125 were robustly expressed and were toxic in yeast (Figure S4C). This toxicity could be mitigated by the generally potentiated Hsp104 variants, Hsp104^A503S and Hsp104^K358D, but not by Hsp104 (Figure S4D). By contrast, Hsp104^K358D:Y257L reduced toxicity of α-syn^1-115 and α-syn^1-125 but was not effective against α-syn^1-95 (Figure S4D). Thus, mutation of Y257 in the ^K358D background can reduce the ability to mitigate α-syn^1-95 toxicity. Indeed, the α-syn-selective variants, Hsp104^K358D:Y257T, Hsp104^K358D:Y662M, and Hsp104^{K358D:Y257L:Y662M} were unable to mitigate the toxicity of α-syn^1-95 but could partially mitigate toxicity of α-syn^1-115 and α-syn^1-125 (Figure S4D). These findings indicate that determinants between residues ^96-115 in the C-terminal acidic region of α-syn are critical for Hsp104^K358D:Y257L, and the α-syn-selective Hsp104 variants to suppress α-syn toxicity. Thus, this region of α-syn acquires increased importance for toxicity mitigation by Hsp104^K358D:Y257L, as well as the α-syn-selective Hsp104 variants Hsp104^K358D:Y257T, Hsp104^K358D:Y662M, and Hsp104^{K358D:Y257L:Y662M}, but is less critical for Hsp104^A503S and Hsp104^K358D.

α-Syn-specific Hsp104 variants do not suppress FUS aggregation

We next evaluated the ability of the α-syn-specific Hsp104 variants to antagonize FUS aggregation in yeast. α-Syn-specific Hsp104 variants should not affect FUS aggregates in yeast, as FUS is no longer recognized as a substrate. In ALS/FTD and in yeast proteinopathy models, FUS mislocalizes to cytoplasmic aggregates^{50,69,70.50,69,70} Enhanced Hsp104 variants, Hsp104^K358D and Hsp104^K358D:Y257L, effectively suppressed FUS aggregation compared to Hsp104, which had no effect (Figure S5A, B). In agreement with our toxicity-suppression findings, α-syn-specific variants Hsp104^K358D:Y257T, Hsp104^K358D:Y662M, and Hsp104^{K358D:Y257L:Y662M} did not antagonize FUS aggregation (Figure S5A, B). Thus, our α-syn-specific variants do not suppress FUS toxicity or aggregation in yeast, strongly advocating for tailored substrate specificity.

α-Syn-specific Hsp104 variants do not restore TDP-43 to the nucleus

As with FUS, we expect α-syn-specific Hsp104 variants should not suppress mislocalization of TDP-43, as these variants have an altered substrate repertoire. TDP-43 is normally localized to the nucleus in human cells, but mislocalizes to cytoplasmic aggregates in ALS/FTD pathology, which is recapitulated in yeast.^48,51,71,72 Hsp104 was unable to return TDP-43 to the nucleus, whereas enhanced variants, Hsp104^K358D and Hsp104^K358D:Y257L, significantly restored nuclear localization to TDP-43, as ~39% and ~50% of cells respectively harbored nuclear TDP-43 (Figure S5C, D). Importantly, α-syn-specific variants Hsp104^K358D:Y662M and Hsp104^{K358D:Y257L:Y662M} did not restore TDP-43 to the nucleus, and α-syn-specific variant Hsp104^K358D:Y257T had only a slight effect on TDP-43 localization (Figure S5C, D). Thus, α-syn-specific Hsp104 variants do not suppress TDP-43 toxicity or restore TDP-43 to the nucleus, consistent with increased substrate selectivity.

α-Syn-specific Hsp104 variants differentially affect cytoplasmic α-syn foci

α-Syn initially localizes to the plasma membrane in yeast, but eventually forms toxic cytoplasmic inclusions that are detergent-insoluble, contain high molecular weight α-syn species, can react with Thioflavin-S (an amyloid-diagnostic dye), and cluster cytoplasmic vesicles reminiscent of aspects of Lewy pathology in PD.^{4,5,34,49,73–77} Hsp104 is unable to antagonize formation of cytoplasmic α-syn foci and return α-syn to the plasma membrane, whereas enhanced variants Hsp104^K358D and Hsp104^K358D:Y257L robustly clear α-syn foci (Figure 5A, B). Only ~18% cells had α-syn to the plasma membrane with Hsp104, whereas Hsp104^K358D and Hsp104^K358D:Y257L promoted α-syn plasma membrane localization in ~62% and ~48% of cells respectively (Figure 5B). α-Syn-specific variants Hsp104^K358D:Y662M and Hsp104^{K358D:Y257L:Y662M} also eradicated α-syn foci (Figure 5A, B). Unexpectedly, α-syn-specific variant Hsp104^K358D:Y257T did not significantly eliminate α-syn foci relative to Hsp104 (Figure 5A, B). Thus, Hsp104^K358D:Y257T is likely not suppressing α-syn toxicity by resolving α-syn foci. The uncoupling of toxicity suppression and clearance of α-syn foci is a novel feature of Hsp104^K358D:Y257T, as we have not previously engineered Hsp104 in this way. Thus, Hsp104^K358D:Y257T emerges as a substrate-specific Hsp104 variant that is distinct from other α-syn-specific variants Hsp104^K358D:Y662M and Hsp104^{K358D:Y257L:Y662M}.

Figure 5. α-Syn-specific Hsp104 variants differentially eradicate α-syn inclusions.

(A) Representative fluorescence microscopy images of Δhsp104 yeast integrated with α-syn-YFP and transformed with Hsp104 variants or an empty vector control. Bar, 4µm. (B) Localization of α-syn-YFP in yeast was quantified by counting the number of cells with plasma membrane-localized α-syn-YFP or α-syn-YFP foci in the cytoplasm. Values represent means ± SEM (n = 3–6). **P<0.01, ***P<0.001, ****P<0.0001; One-way ANOVA with Dunnett’s post-hoc test. (C) Representative fluorescence microscopy images of Δhsp104 yeast integrated with α-syn-YFP and transformed with mCherry – Hsp104 variants or mCherry alone as a control. Bar, 4µm. (D) Colocalization of mCherry–Hsp104 variants or mCherry alone with α-syn-YFP foci was quantified. Values represent means ± SEM (n = 2–6). *P<0.05; One-way ANOVA with Dunnett’s post-hoc test. (E) Δhsp104 yeast integrated with α-syn-YFP on a galactose-inducible promoter were transformed with mCherry – Hsp104 variants or mCherry alone. Yeast were spotted onto glucose (uninducing, off) and galactose (inducing, on) media in a five-fold serial dilution.

See also Figure S5.

Hsp104^K358D:Y257T colocalizes with α-syn foci more frequently than Hsp104

As Hsp104^K358D:Y257T suppressed α-syn toxicity in yeast but did not eliminate α-syn foci, we were curious if this variant colocalized with α-syn foci. One explanation for the uncoupling of toxicity suppression and clearance of α-syn foci is that Hsp104^K358D:Y257T is still able to recognize and colocalize with α-syn foci to mitigate their toxicity. We tagged Hsp104 and Hsp104^K358D:Y257T with mCherry to visualize their cellular localization in yeast expressing α-syn-YFP. C-terminally mCherry-tagged Hsp104^K358D:Y257T more frequently colocalized with α-syn foci than mCherry-tagged Hsp104 or mCherry alone (Figure 5C, D). mCherry-Hsp104^K358D:Y257T colocalized with α-syn foci in ~25% of cells (Figure 5D) and suppressed α-syn toxicity (Figure 5E). We also assessed whether the generally potentiated variants, mCherry-Hsp104^K358D or mCherry-Hsp104^K358D:Y257L, colocalized with α-syn foci. We find that cells with α-syn foci typically expressed very low levels of mCherry-Hsp104^K358D, mCherry-Hsp104^A503S, or mCherry-Hsp104^K358D:Y257L, and we do not observe any colocalization with α-syn foci (Figure 5D). By contrast, cells that robustly express α-syn-YFP and mCherry-Hsp104^K358D, mCherry-Hsp104^A503S, or mCherry-Hsp104^K358D:Y257L, do not contain α-syn-YFP foci and suppress α-syn toxicity (Figure 5E). Thus, the colocalization with α-syn foci is a specific feature of Hsp104^K358D:Y257T, which we suggest contributes to the ability of Hsp104^K358D:Y257T to mitigate α-syn toxicity. Hsp104^K358D:Y257T may remodel α-syn conformers to a less toxic conformation or otherwise quench their toxicity (e.g., by shielding surfaces from interacting with other essential cell components, such as intracellular vesicles). Furthermore, Hsp104^K358D:Y257T likely engages α-syn foci more efficiently than Hsp104, as it has enhanced substrate specificity.

α-Syn-specific Hsp104 variants have altered ATPase activity

To elucidate the underlying mechanism for substrate-specific toxicity suppression, we surveyed various biochemical properties of a representative set of Hsp104 variants. We first tested the ATPase activity to determine if α-syn-specific variants differ in ability to hydrolyze ATP relative to Hsp104 and potentiated variants. We purified the generally potentiated MD variant, Hsp104^A503S, and the generally potentiated NBD1 variant, Hsp104^K358D, both of which exhibit elevated ATPase activity compared to Hsp104 (Figure 6A).^26,34,60 We also purified a generally potentiated pore-loop variant, Hsp104^K358D:Y257L, and three α-syn specific variants, Hsp104^K358D:Y257T, Hsp104^K358D:Y662M, and Hsp104^{K358D:Y257L:Y662M}. Interestingly, Hsp104^K358D:Y257L exhibited reduced ATPase activity compared to Hsp104^K358D, but slightly higher ATPase activity than Hsp104 (Figure 6A). Thus, Hsp104^K358D:Y257L provides another example of a potentiated Hsp104 that does not have greatly elevated basal ATPase activity as with Hsp104^D504C.³⁴ Nonetheless, this effect was unexpected as mutating Y257 in WT Hsp104 does not typically reduce basal ATPase activity.^14,21,39 This reduction in ATPase activity was even more pronounced for the α-syn-specific variant Hsp104^K358D:Y257T, which had ATPase activity similar to Hsp104, but reduced ATPase activity compared to Hsp104^K358D (Figure 6A). Thus, the K358D mutation sensitizes Hsp104 ATPase activity to Y257 mutations, perhaps due to destabilization of the small domain of NBD1 similar to another enhanced NBD1 variant, Hsp104^E360R, which also bears a charge-reversing mutation in the vicinity of K358^26,40.26,40

Figure 6. α-Syn-specific Hsp104 variants have distinct biochemical properties.

(A) ATPase activity of Hsp104 variants. Values represent means±SEM (n=4). (B) HAP variants (0.167μM hexamer) plus ClpP (21μM monomer) were incubated with FITC-casein, and FITC-casein degradation was measured by fluorescence. Negative controls FITC-casein alone (black) and ClpP alone control (dark brown) were included. Values were normalized to HAP plus ClpP at 60 minutes and represent means±SEM (n=3–6). (C) RepA_1-25-GFP (0.7μM) was incubated with Hsp104 variants (2.1μM hexamer) plus GroEL^trap (2.5μM tetradecamer), and RepA_1-25-GFP unfolding was measured by fluorescence. Negative control of RepA_1-25-GFP alone included. A representative trial is shown (n=3). (D) Hsp104 variants alone (1μM hexamer, solid bars) or with chaperone pairs Hsc70 and Hdj2 (0.167μM each, checkered bars), or Hsp72 and Hdj1 (0.167μM each, dotted bars), were incubated with urea-denatured luciferase aggregates. Reactivation activity was measured by luminescence. Negative control of luciferase aggregates alone included. Values represent means±SEM (n=2–8). (E) α-Syn fibrils (3µM monomer) were incubated without or with Hsp104, Hsp104^A503S, Hsp104^K358D, Hsp104^K358D:Y257T or Hsp104^K358D:Y662M (3µM) plus Hsc70 (3µM), Hdj1 (3µM), ATP (20 mM) and an ATP regeneration system (20mM creatine phosphate and 0.5µM creatine kinase) for 2h at 30°C. Disaggregation was assessed by Thioflavin-T (ThT) fluorescence. Values represent means±SEM (n=3). (F) TDP-43 fibrils (3µM monomer) were incubated without or with Hsp104, Hsp104^A503S, Hsp104^K358D, Hsp104^K358D:Y257T or Hsp104^K358D:Y662M (3µM) plus Hsc70 (3µM), Hdj1 (3µM), ATP (20mM) and an ATP regeneration system (20mM creatine phosphate and 0.5µM creatine kinase) for 2h at 30°C. Disaggregation was assessed by turbidity (absorbance at 350nm). Values represent means±SEM (n=3). (G) Soluble α-syn oligomers (0.5µM monomer) were incubated without or with Hsp104, Hsp104^A503S, Hsp104^K358D, Hsp104^K358D:Y257L, Hsp104^K358D:Y257T or Hsp104^K358D:Y662M (1µM) plus ATP (20mM) and an ATP regeneration system (20mM creatine phosphate and 0.5µM creatine kinase) for 1h at 37°C. Reactions were then fractionated through a Microcon YM-100 (100-kDa molecular weight cut off) filter. The amount of α-syn in the filtrate fraction was then determined. Values represent means±SEM (n=3).

See also Figure S6.

The other α-syn-specific variants, Hsp104^K358D:Y662M and Hsp104^{K358D:Y257L:Y662M}, exhibited ATPase activity equal to or greater than the activity of Hsp104^A503S and Hsp104^K358D (Figure 6A). Indeed, the Y662M mutation stimulated ATPase activity in the ^K358D background, and even counteracted the effect of the Y257L mutation such that Hsp104^{K358D:Y257L:Y662M} had higher ATPase activity than Hsp104^K358D:Y257L and similar ATPase activity to Hsp104^K358D (Figure 6A). This result was also unanticipated as Y662 mutations do not typically affect WT Hsp104 ATPase activity,^14,24,39 indicating that the K358D mutation sensitizes Hsp104 ATPase activity to Y662 mutations.

The difference in ATPase activity between α-syn-specific variants is a key biochemical distinction between them. These findings further suggest that α-syn-specific variants likely employ different toxicity-suppression mechanisms. Thus, the lower ATPase activity of Hsp104^K358D:Y257T may limit clearance of α-syn foci in yeast, whereas elevated ATPase activity of Hsp104^K358D:Y662M and Hsp104^{K358D:Y257L:Y662M} likely enables clearance of α-syn foci.

α-Syn-specific Hsp104 variants have altered translocase activity

Next, we assessed translocase activity of each Hsp104 variant. Hsp104 translocates substrates across its central channel during disaggregation.^13,14 To assess substrate translocation, we evaluated each Hsp104 variant in the HAP background in vitro. HAP is an Hsp104 variant with three missense mutations (G739I:S740G:K741F) that enable interaction with the chambered protease, ClpP.²² Thus, HAP but not Hsp104 can translocate substrate into the proteolytic chamber of ClpP where it is degraded. We monitored translocation of a model unfolded substrate fluorescein isothiocyanate (FITC)-labeled casein through HAP via degradation by ClpP, which liberates FITC and increases fluorescence.¹⁴ Mutating each pore-loop Tyr to Ala (Y257A:Y662A), which ablates substrate translocation and disaggregation,^14,21,22,39 resulted in the DPLA (‘double pore-loop alanine’) HAP variant that lacks translocation activity (Figure 6B). Thus, minimal passive translocation occurs in this system. We previously found that HAP^A503V translocated substrate more rapidly than HAP³⁴. Likewise, HAP^A503S and HAP^K358D translocated substrate more rapidly than HAP (Figure 6B). This accelerated substrate translocation may promote enhanced disaggregase activity.

Interestingly, α-syn-specific variants exhibited distinct activities. HAP^K358D:Y662M translocated FITC-casein more effectively than HAP, whereas HAP^{K358D:Y257L:Y662M} was less effective (Figure 6B). In the context of Hsp104, these variants have similar ATPase activity (Figure 6A). Thus, HAP^{K358D:Y257L:Y662M} likely has reduced ability to grip FITC-casein resulting in decelerated translocation. HAP^K358D:Y257L and HAP^K358D:Y257T translocated FITC-casein at a similar rate but were less effective than HAP (Figure 6B). These findings suggest that interfering with Y257 of pore-loop 1 likely weakens grip of casein and thereby reduces translocase activity. Changes in substrate grip and translocase activity likely contribute to altered patterns of substrate specificity, as they could enable fine-tuning of the force used by Hsp104 to disaggregate different substrates. Indeed, Hsp104 variants with low translocase activity may only partially translocate substrate, thus changing the way a substrate is disaggregated.

α-Syn-specific Hsp104 variants have altered unfoldase activity

We next evaluated the unfoldase activity of each Hsp104 variant using RepA_1-25-GFP as a model substrate. In these reactions, we included GroEL^trap, which captures newly unfolded RepA_1-25-GFP and prevents its refolding.⁷⁸ Hsp104 is unable to unfold RepA_1-25-GFP under these conditions, whereas potentiated Hsp104 variants, Hsp104^A503S and Hsp104^K358D, rapidly unfold RepA_1-25-GFP³⁴ (Figure 6C). Remarkably, α-syn-specific Hsp104 variants showed distinct unfoldase activity from one another. α-Syn-specific Hsp104^K358D:Y662M unfolded RepA_1-25-GFP almost as rapidly as Hsp104^A503S and Hsp104^K358D (Figure 6C). By contrast, α-syn-specific Hsp104^K358D:Y257T and Hsp104^{K358D:Y257L:Y662M}, as well as generally potentiated pore-loop variant Hsp104^K358D:Y257L, only modestly unfolded RepA_1-25-GFP (Figure 6C). Here too, it is likely that pore-loop 1 mutations weaken substrate gripping by Hsp104 and limit unfolding of the RepA_1-25-GFP substrate. As Hsp104^K358D:Y662M possessed strong unfoldase activity similar to generally potentiated Hsp104 variants, enhanced substrate recognition or unfolding power may contribute to its mechanism of toxicity suppression. The stark difference in unfoldase activity between α-syn-specific Hsp104 variants is another distinction in their biochemical properties that could contribute to different modes of α-syn toxicity suppression in yeast.

α-Syn-specific Hsp104 variants have altered luciferase disaggregase activity

Next, we assessed disaggregase activity of each Hsp104 variant against aggregated luciferase. Differences in the ability to reactivate this model substrate could help to illuminate general disaggregase properties of substrate-selective Hsp104 variants. Hsp104 was unable to reactivate disordered luciferase aggregates on its own, and required Hsp70 and Hsp40 (Figure 6D).^15,34 Generally potentiated variants, Hsp104^A503S and Hsp104^K358D, exhibited similar activities, and had high luciferase reactivation activity even in the absence of Hsp70 and Hsp40 (Figure 6D). Interestingly, the generally potentiated pore-loop variant Hsp104^K358D:Y257L was less active than Hsp104, indicating that Y257L mutation is inhibitory with respect to specifically disaggregating luciferase (Figure 6D). Indeed, it appears that Y257L alters activity against model substrates in vitro, but not against disease substrates in yeast. Of the three α-syn specific Hsp104 variants, only Hsp104^K358D:Y662M had luciferase reactivation activity (Figure 6D). Hsp104^K358D:Y662M was more active than Hsp104 and displayed comparable activity in the absence and presence of Hsp70 and Hsp40 (Figure 6D). Notably, each Hsp104 variant showed similar activity with two different human Hsp70 and Hsp40 pairs, Hsc70/Hdj2, and Hsp72/Hdj1, suggesting this activity is not dependent on the presence of specific Hsp70 or Hsp40 variants (Figure 6D). The distinct luciferase reactivation activity of Hsp104^K358D:Y662M and lack of activity of Hsp104^K358D:Y257L, Hsp104^K358D:Y257T, and Hsp104^{K358D:Y257L:Y662M}, likely result from a specific effect on luciferase from perturbing pore-loop 1. Altering pore-loop 1 in the context of the enhanced Hsp104^K358D background must interfere with disaggregation of disordered luciferase aggregates. Indeed, these variants may no longer recognize luciferase effectively as a result of altered substrate-specificity.

Hsp104^K358D:Y662M disaggregates α-syn fibrils but not TDP-43 fibrils

Next, we assessed the ability of Hsp104 variants to disaggregate preformed α-syn and TDP-43 fibrils.³⁴ Hsp104 displayed limited ability to dissociate α-syn or TDP-43 fibrils under these conditions, whereas Hsp104^A503S and Hsp104^K358D effectively dismantled α-syn and TDP-43 fibrils (Figure 6E, F). Hsp104^K358D:Y257L also effectively disaggregated α-syn and TDP-43 fibrils (Figure 6E, F). Thus, although ineffective against luciferase (Figure 6D), Hsp104^K358D:Y257L was active against α-syn and TDP-43 fibrils (Figure 6E, F). By contrast, Hsp104^K358D:Y257T was unable to dissolve α-syn or TDP-43 fibrils (Figure 6E, F). Thus, the ability of Hsp104^K358D:Y257T to reduce α-syn toxicity is separated from α-syn disaggregation. Finally, Hsp104^K358D:Y662M could disassemble α-syn fibrils, but not TDP-43 fibrils (Figure 6E, F), indicating a change in substrate specificity that explains the ability to reduce α-syn toxicity, but not TDP-43 toxicity.

Hsp104^K358D:Y257T and Hsp104^K358D:Y662M disassemble soluble α-syn oligomers

We next assessed the ability of Hsp104 variants to disassemble preformed, soluble α-syn oligomers, which are a toxic species that accumulate in yeast.^76,79 Under our conditions, Hsp104 displayed limited ability to dissociate soluble α-syn oligomers, whereas Hsp104^A503S, Hsp104^K358D, and Hsp104^K358D:Y257L effectively disassembled soluble α-syn oligomers (Figure 6G). Importantly, the α-syn-selective variants, Hsp104^K358D:Y257T and Hsp104^K358D:Y662M, also effectively disassembled soluble α-syn oligomers (Figure 6G). Thus, these α-syn-selective variants directly remodel toxic α-syn conformers, which may enable mitigation of α-syn toxicity in yeast.

α-Syn-specific Hsp104 variants show distinct toxicity phenotypes at 37°C

As the α-syn-specific Hsp104 variants had differing biochemical profiles, we wondered if these variants showed differences in their off-target toxicity to yeast. Thus, we evaluated their toxicity at 37°C in the absence of neurodegenerative disease-associated substrate. Yeast are mildly stressed at 37°C as essential proteins are slightly unfolded. Potentiated Hsp104 variants can nonspecifically target key proteins that are slightly unfolded, resulting in a strong toxic phenotype when expressed in yeast.³⁵ Thus, this assay can be used to uncover Hsp104 variants that exhibit less off-target toxicity that may be advanced to more complex model systems. Consistent with previous results,³⁴ the potentiated variants Hsp104^A503V and Hsp104^K358D were toxic to yeast relative to Hsp104 at 37°C but not 30°C (Figure S6A). Interestingly, generally potentiated Hsp104^K358D:Y257L and α-syn-specific Hsp104^K358D:Y257T virtually eliminated this toxic phenotype (Figure S6A). Low toxicity to yeast could result from reduced targeting of essential, unfolded substrates in yeast and enhanced recognition of toxic substrates. It appears that the Y257L mutation reduces Hsp104 activity against off-target yeast substrates, but not against disease substrates. The ability to suppress toxicity of neurodegenerative disease-associated substrates, coupled with very low inherent toxicity to yeast, are desirable attributes for translating these disaggregases into effective therapeutic agents. Surprisingly, however, both Hsp104^K358D:Y662M and Hsp104^{K358D:Y257L:Y662M} were noticeably toxic to yeast at 37°C (Figure S6A). We suggest these variants could be targeting a substrate under mild stress conditions that is essential for yeast. It seems the Y662M mutation reduces Hsp104 activity specifically against TDP-43 and FUS, but not against α-syn, or off-target yeast substrates.

α-Syn-specific Hsp104 variants confer yeast thermotolerance to different extents

Next, we assessed the ability of α-syn-specific Hsp104 variants to confer thermotolerance in yeast. One of the central functions of Hsp104 is to help yeast survive heat shock by resolubilizing proteins that aggregate during stress.^62,80–84 We hypothesized that substrate-specific Hsp104 variants should be defective in conferring thermotolerance, in that they are no longer able to recognize a broad array of substrates for solubilization. We evaluated each Hsp104 variant under the native heat shock element (HSE) Hsp104 promoter relative to Hsp104 in their ability to confer thermotolerance to yeast. We first induced Hsp104 expression at 37°C, then heat-shocked yeast at 50°C (Figure S6B). All Hsp104 variants were expressed at similar levels (Figure S6C). Empty vector conferred no thermotolerance (very few yeast survive a 30-minute heat shock at 50°C), whereas Hsp104 transformed into a Δhsp104 strain effectively complemented the yeast survival observed in yeast expressing Hsp104 endogenously (Figure S6B). Potentiated variants Hsp104^A503V and Hsp104^K358D did not retain full Hsp104 activity in conferring thermotolerance (Figure S6B). Likewise, Hsp104^K358D:Y662M and Hsp104^{K358D:Y257L:Y662M} also conferred reduced thermotolerance (Figure S6B). This reduced activity is perhaps due to off-target toxicity at the 37°C pretreatment step (Figure S6A). Interestingly, Hsp104^K358D:Y257L and Hsp104^K358D:Y257T conferred similar levels of reduced thermotolerance (Figure S6B). For these variants, which are not toxic at 37°C (Figure S6A), the reduced thermotolerance indicates that they do not disaggregate the complete repertoire of Hsp104 substrates (Figure S6B).

Several α-syn-specific Hsp104 variants prevent α-syn-induced neurodegeneration in C. elegans

We next tested whether the Hsp104 variants that selectively suppressed α-syn toxicity in yeast can prevent neurodegeneration in a C. elegans model of PD, which has successfully validated PD-relevant modifiers of α-syn toxicity.^{34,55,56,85,86} Preventing α-syn-induced neurodegeneration in this model is a key step in translating our findings in yeast to a full metazoan nervous system. We tested α-syn specific Hsp104^K358D:Y257T, Hsp104^K358D:Y662M, and Hsp104^{K358D:Y257L:Y662M}, as well as generally potentiated Hsp104^K358D and Hsp104^K358D:Y257L. In this transgenic model, expression of human α-syn and Hsp104 variants is driven selectively in the dopaminergic neurons of worms using the endogenous promoter of the dopamine transporter gene, P_dat-1. When α-syn is expressed alone, ~23% of worms retained the full complement of dopaminergic neurons at day 7, and ~10% at day 10 post hatching (Figure 7A), which recapitulates the progressive α-syn neurotoxicity in human α-synucleinopathies. Hsp104 was unable to protect against α-syn-induced neurodegeneration (Figure 7A).³⁴ Surprisingly, generally potentiated Hsp104^K358D and Hsp104^K358D:Y257L also did not significantly protect against neurodegeneration (Figure 7A). Moreover, the α-syn-specific Hsp104^{K358D:Y257L:Y662M} did not protect against neurodegeneration, suggesting that this variant also did not translate from yeast to worm.

Figure 7. Several α-syn-specific Hsp104 variants prevent dopaminergic neuron degeneration in a C. elegans model of PD.

(A) Hsp104 variants and α-syn were co-expressed in the dopaminergic (DA) neurons of C. elegans. Animals expressing Hsp104^K358D:Y257T are significantly protected against α-syn-induced DA neuron degeneration at day 7 post-hatching (left). DA neurodegeneration was exacerbated at day 10 post-hatching (right). Though not protective at day 7 post-hatching, overexpression of Hsp104 variants Hsp104^DPLF:A503V, Hsp104^A503S, and Hsp104^K358D:Y662M protected against α-syn-induced DA neurodegeneration at day 10 post-hatching (right). Notably, the Hsp104^K358D:Y257T variant consistently exhibits neuroprotection at both the earlier and later timepoints. Values represent means±SEM, n=30 per replicate, three independent experiments were performed/variant, and three distinct worm stable lines were generated for each Hsp104 variant. *P<0.05, **P<0.01, ***P<0.001; One-way ANOVA with Dunnett’s post-hoc test (compared to α-syn alone control). (B) A representative image of C. elegans DA neurons in worms expressing α-syn alone (left) or α-syn + Hsp104^K358D-Y257T (right). Nematodes have six anterior DA neurons (4 CEP and 2 ADE), which were scored at day 7 and 10 post-hatching. Left, the worm has only two normal neurons (2 CEP), where the other four neurons have degenerated. Right, the full complement of six anterior DA neurons expressing Hsp104^K358D-Y257T + α-syn, indicating a protective activity against α-syn toxicity. Triangles show normal neurons while arrows depict regions where there are degenerating or missing neurons. Bar, 10µm.

See also Figure S7.

Remarkably, α-syn-specific variant, Hsp104^K358D:Y257T, strongly protected against α-syn-induced neurodegeneration at days 7 and 10 (Figure 7A, B). Despite only moderate α-syn toxicity suppression in yeast, Hsp104^K358D:Y257T displayed the strongest protection in the context of a metazoan nervous system, as ~43% of worms retained the full complement of DA neurons at day 7, and ~27% at day 10 (Figure 7A, B). Hsp104^K358D:Y257T is more effective in the C. elegans PD model than previously engineered potentiated variants Hsp104^A503S and Hsp104^{DPLF:A503V 34} (Figure 7A, B). Hsp104^K358D:Y257T expressed lower mRNA levels than generally potentiated variants in transgenic C. elegans and did not affect α-syn expression levels (Figure S7). This finding suggested that the reduced neurodegeneration is due to the enhanced activity of this α-syn-specific Hsp104 variant.

Interestingly, another α-syn-specific variant, Hsp104^K358D:Y662M, conferred neuroprotection but was only significantly protective against neurodegeneration at a later timepoint (day 10), where ~30% of worms retained the full complement of WT neurons (Figure 7A). As Hsp104^K358D:Y662M did not prevent neurodegeneration at an earlier time point (day 7), this variant did not appear able to delay disease onset, but instead limited the progression of α-syn neurotoxicity at a threshold of greater severity. Delayed prevention of α-syn-induced neurodegeneration could stem from a unique remodeling activity of Hsp104^K358D:Y662M. Using an Hsp104 variant that stops disease progression earlier (Hsp104^K358D:Y257T), followed by Hsp104^K358D:Y662M which works more effectively later in disease, may be an advantageous combination strategy for mitigating α-syn toxicity in dopaminergic neurons.

Discussion

Here, we engineered Hsp104 variants that selectively suppress toxicity of α-syn. Unexpectedly, the Hsp104 background used to engineer enhanced substrate specificity was critical. For example, we found that Hsp104^A503S, a potentiated Hsp104 variant that breaks MD-MD contacts,^26,40 was not amenable for introducing substrate specificity via mutation of substrate-binding, pore-loop tyrosines. By contrast, α-syn-selective Hsp104 variants could be isolated by mutating pore-loop tyrosines in enhanced backgrounds that instead break NBD1-MD contacts, e.g. Hsp104^K358D or Hsp104^D484K.⁶⁰ Thus, the precise nature of the underlying potentiation mechanism is an important aspect to consider when tuning Hsp104 variants²⁶.

By making subtle changes to the substrate-binding, pore-loop tyrosines that line the axial channel of Hsp104, we re-wired enhanced Hsp104 variants (Hsp104^K358D or Hsp104^D484K) into substrate-specific variants. Thus, we shifted the substrate repertoire of Hsp104 such that α-syn toxicity could be mitigated, whereas TDP-43 or FUS toxicity could not. α-Syn-selectivity could be conferred by Y257I/V/T/Q mutations in pore-loop 1 or by Y662L/M mutations in pore-loop 2 in the K358D background. Surprisingly, two classes of α-syn-specific Hsp104 variant emerged that reduced α-syn toxicity via distinct mechanisms.

The first class, which includes Hsp104^K358D:Y662M and Hsp104^{K358D:Y257L:Y662M}, mitigated α-syn toxicity via ATPase-dependent disaggregation of α-syn inclusions. Similar to our previously engineered, potentiated Hsp104 variants, Hsp104^K358D:Y662M and Hsp104^{K358D:Y257L:Y662M} exhibited elevated ATPase activity compared to Hsp104. Hsp104^K358D:Y662M and Hsp104^{K358D:Y257L:Y662M} cleared α-syn foci in yeast but did not affect TDP-43 or FUS aggregation. Importantly, Hsp104^K358D:Y662M dissolved α-syn fibrils but not TDP-43 fibrils in vitro. Thus, these Hsp104 variants have altered substrate specificity, which reduces disaggregase activity against some substrates (e.g., TDP-43) but permits effective α-syn disaggregation.

The second class, which includes Hsp104^K358D:Y257T, mitigated α-syn toxicity via ATPase-dependent detoxification of α-syn conformers without disaggregation of α-syn inclusions. Intriguingly, Hsp104^K358D:Y257T exhibited similar ATPase activity to WT Hsp104. However, Hsp104^K358D:Y257T was unable to disaggregate luciferase in vitro, but could confer some level of thermotolerance in vivo indicating that it retains some disaggregase activity. Unexpectedly, Hsp104^K358D:Y257T neither eliminated α-syn foci in yeast nor disaggregated α-syn fibrils in vitro. Importantly, Hsp104^K358D:Y257T was more effective than WT Hsp104 in disassembling soluble α-syn oligomers in vitro, which are a neurotoxic species^76,79,87,88. Moreover, unlike WT Hsp104, Hsp104^K358D:Y257T colocalized with α-syn foci. Hsp104^K358D:Y257T may partially remodel α-syn inclusions into less toxic structures. For example, Hsp104^K358D:Y257T could selectively disassemble soluble α-syn oligomers throughout the cell, including any associated with α-syn inclusions. In addition, Hsp104^K358D:Y257T might extract essential proteins trapped in α-syn foci, or mitigate toxic interactions between α-syn and intracellular vesicles or organelles.^5,73,74,89

Hsp104 variants from each class (Hsp104^K358D:Y662M and Hsp104^K358D:Y257T) prevented dopaminergic neuron degeneration in a C. elegans model of PD more effectively than previous non-α-syn-specific, enhanced Hsp104 variants. Hsp104^K358D:Y257T protected against neurodegeneration throughout the course of disease in this model, whereas Hsp104^K358D:Y662M protection was restricted to the later time point. These data highlight the utility of the transgenic worm model in enabling temporal distinctions in neuromodulation to be parsed in vivo. Thus, substrate-specific Hsp104 variants effectively suppress neurodegeneration in the context of an intact metazoan nervous system, thereby translating the therapeutic benefits of these variants from yeast to metazoa. These findings bode well for subsequent validation, given the proven translational efficacy of the C. elegans α-syn model used in these studies, which has reproducibly demonstrated a predictive capacity to yield outcomes representative of mammalian models of PD.⁹⁰

We have also established that the level of off-target toxicity in yeast inherent to potentiated Hsp104 variants does not necessarily predict whether these variants will be neuroprotective in C. elegans. For example, Hsp104^K358D:Y662M exhibited off-target toxicity in yeast, but mitigated α-syn-induced dopaminergic neurodegeneration at later time points in C. elegans. Hsp104^K358D:Y662M could target essential yeast proteins for unfolding, but these proteins may not be as crucial or may be absent from C. elegans. By contrast, Hsp104^{K358D:Y257L:Y662M} exhibited off-target toxicity in yeast but did not prevent neurodegeneration in C. elegans. The challenge of translating our findings from yeast to worm likely reflects key differences between the two model systems. One important distinction is that the proteome of dopaminergic neurons in C. elegans likely has crucial differences from the yeast proteome. For example, dopamine modulates the propensity of α-syn to oligomerize and impacts neurodegeneration in mice, dopaminergic neuron culture, and C. elegans.^91,92 Thus, the absence of dopamine in yeast, as well different sets of Hsp104-interacting proteins in yeast versus worm renders the direct translation of Hsp104 variants to metazoa challenging. Nevertheless, our efforts to engineer α-syn-specific Hsp104 variants have yielded Hsp104^K358D:Y662M and Hsp104^K358D:Y257T, which outperform prior potentiated variants. It will be important to advance these Hsp104 variants to mammalian models of α-synucleinopathies.

It will also be of great interest to engineer substrate-specific Hsp104 variants to specifically target a range of other misfolded proteins in neurodegenerative disease, including TDP-43 and FUS. At a minimum, our findings suggest that engineering substrate-specific disaggregases can improve their ability to confer neuroprotection. We envision that increasing the substrate specificity of enhanced disaggregases could be applied broadly to tailor therapeutics for neurodegenerative disease. Beyond Hsp104, we anticipate that fine-tuning disaggregases found in humans, such as Hsp110, Hsp70, and Hsp40, nuclear-import receptors, Skd3, TRIMs, and polyD/E proteins like DAXX will also be immensely valuable.^{7,29,46,93–97} We suggest that highly tuned, specific disaggregases that reverse targeted toxic misfolding events represent an exciting avenue for therapeutic agents in neurodegenerative disease.^7,11,46

Limitations of the Study

We have not assessed the substrate selectivity of α-synuclein-specific Hsp104 variants in a proteome-wide manner, which will be important to address in future studies. Moreover, we have not assessed α-synuclein-specific Hsp104 variants in mammalian models. In future studies, it will be important to advance the α-synuclein-specific Hsp104 variants to mammalian models of α-synucleinopathies, including human neurons in culture and mouse models.

STAR Methods

RESOURCE AVAILABILITY

Lead contact

Further information and requests for resources and reagents should be directed to and will be fulfilled by the lead contact, James Shorter (jshorter@pennmedicine.upenn.edu).

Materials availability

Plasmids generated in this study will be made readily available to the scientific community. We will honor requests in a timely fashion. Material transfers will be made with no more restrictive terms than in the Simple Letter Agreement or the Uniform Biological Materials Transfer Agreement and without reach through requirements.

Data and code availability

• All raw images of yeast spotting, western blots, and C. elegans experiments are available in Mendeley Data. All data are publicly available as of the date of publication. Accession numbers and DOI are listed in the key resources table.

• All custom code has been deposited to GitHub and Zenodo. DOI are listed in the key resources table.

• Any additional information required to reanalyze the data reported in this paper is available from the lead contact upon request.

Key resources table

REAGENT or RESOURCE	SOURCE	IDENTIFIER
Antibodies
Rabbit anti-Hsp104 polyclonal	Enzo Life Sciences	Cat#ADI-SPA-1040-F; RRID:AB_2039208
Rabbit anti-FUS polyclonal	Bethyl Laboratories	Cat# A300-302A; RRID:AB_309445
Rabbit anti-TDP-43 polyclonal	Proteintech	Cat#10782; RRID: AB_615042
Rabbit anti-GFP polyclonal	Sigma-Aldrich	Cat# G1544; RRID: AB_439690
Mouse anti-α-Synuclein monoclonal (Syn303)	Kelvin Luk (University of Pennsylvania)	RRID:AB_2315395
Mouse anti-PGK1 monoclonal	Thermo Fisher	Cat#459250; RRID:AB_2532235
IRDye 800CW Goat anti-Mouse IgG secondary antibody	LI-COR	Cat# 926–32210; RRID:AB_621842
IRDye 680RD Goat anti-Rabbit IgG secondary antibody	LI-COR	Cat# 926–68071; RRID:AB_10956166
Bacterial and virus strains
Escherichia coli DH5α competent cells	Thermo Fisher	Cat#18265017
Escherichia coli BL21-CodonPlus (DE3)-RIL competent cells	Agilent	Cat#230245
Chemicals, peptides, and recombinant proteins
Creatine phosphate	Roche	Cat#10621722001
Adenosine 5’-triphosphate disodium salt hydrate (ATP)	Sigma-Aldrich	Cat#A3377
cOmplete, Mini, EDTA-free Protease Inhibitor Cocktail	Sigma-Aldrich	Cat#4693159001
Casein fluorescein isothiocyanate from bovine milk (FITC-Casein)	Sigma-Aldrich	Cat#C0528
Creatine kinase	Roche	Cat#10127566001
Firefly luciferase	Sigma-Aldrich	Cat#L9506
Lysozyme	Sigma-Aldrich	Cat#L6876
His-TEV protease	Cupo and Shorter98	N/A
Hsp104	Jackrel et al.34	N/A
Hsp104A503S	Jackrel et al.34	N/A
Hsp104K358D	This paper	N/A
Hsp104K358D:Y257L	This paper	N/A
Hsp104K358D:Y257T	This paper	N/A
Hsp104K358D:Y662M	This paper	N/A
Hsp104K358D:Y257L:Y662M	This paper	N/A
HAP	Jackrel et al.34	N/A
HAPA503S	This paper	N/A
HAPK358D	This paper	N/A
HAPK358D:Y257L	This paper	N/A
HAPK358D:Y257T	This paper	N/A
HAPK358D:Y662M	This paper	N/A
HAPK358D:Y257L:Y662M	This paper	N/A
GroELtrap	Jackrel et al.34	N/A
RepA1-25-GFP	Lopez et al.99	N/A
ClpP	Jackrel et al.34	N/A
Hsc70	Enzo Life Sciences	Cat#ADI-SPP-751-F
Hdj2	Enzo Life Sciences	Cat#ADI-SPP-405-F
Hsp72	Michalska et al.20	N/A
Hdj1	Michalska et al.20	N/A
α-Syn	Jackrel et al.34	N/A
TDP-43	Jackrel et al.34	N/A
Critical commercial assays
ATPase Activity Kit (Colorimetric)	Innova Biosciences	Cat#601–0120
Luciferase Assay Reagent	Promega	Cat#E1483
Human α-Synuclein ELISA Kit	Thermo Fisher	Cat#KHB0061
QuikChange Site-Directed Mutagenesis Kit	Agilent	Cat# 200518
Deposited data
Raw images of yeast spotting, western blots, and C. elegans experiments	Mendeley data	doi: 10.17632/t7k9w8txzk.1
Experimental models: Organisms/strains
S. cerevisiae W303a (MATa, can1-100, his3-11, 15, leu2-3, 112, trp1-1, ura3-1, ade2-1)	Jackrel et al.34	N/A
S. cerevisiae W303a∆hsp104 (MATa, can1-100, his3-11, 15, leu2-3, 112, trp1-1, ura3-1, ade2-1, hsp104::KanMX)	Jackrel et al.34	N/A
S. cerevisiae W303aΔhsp104-pAG303GAL-α-syn-YFP-pAG304GAL-α-syn-YFP	Jackrel et al.34	N/A
S. cerevisiae W303aΔhsp104-pAG303GAL-FUS	Jackrel et al.34	N/A
S. cerevisiae W303aΔhsp104-pAG303GAL-TDP-43	Jackrel et al.34	N/A
S. cerevisiae W303aΔhsp104-pAG303GAL-TDP-43-GFPS11-pAG305GAL-GFPS1-10	Jackrel et al.34	N/A
S. cerevisiae W303aΔhsp104-pAG303GAL-FUS-GFP	Jackrel et al.34	N/A
C. elegans UA44 (baln11 [Pdat-1::α-syn, Pdat-1: :GFP])	Cao et al.85	N/A
C. elegans UA367 (baEx198 a,b,c [Pdat-1::HSP104-K358D, rol-6];baln11 [Pdat-1:: α-syn, Pdat-1 ::GFP])	This study	N/A
C. elegans UA368 (baEx199 a,b,c [Pdat-1::HSP104-K358D:Y257L, rol-6];baln11 [Pdat-1:: α-syn, Pdat-1 ::GFP])	This study	N/A
C. elegans UA369 (baEx200 a,b,c [Pdat-1::HSP104-K358D:Y257T, rol-6];baln11 [Pdat-1::α-syn, Pdat-1 ::GFP])	This study	N/A
C. elegans UA370 (baEx201 a,b,c [Pdat-1::HSP104-K358D:Y257L:Y662M, rol-6];baln11 [Pdat-1:: α-syn, Pdat-1::GFP])	This study	N/A
C. elegans UA371 (baEx202 a,b,c [Pdat-1::HSP104-K358D: Y662M, rol-6];baln11 [Pdat-1:: α-syn, Pdat-1 ::GFP])	This study	N/A
C. elegans UA256 (baEx147 a,b,c [Pdat-1::Hsp104 WT, Pmyo-2::mCherry]; baIn11 [Pdat-1:: α-syn, Pdat-1::GFP])	Jackrel et al.34	N/A
C. elegans UA259 (baEx150 a,b,c [Pdat-1::Hsp104-A503S, Pmyo-2::mCherry]; baIn11 [Pdat-1::α-syn, Pdat-1::GFP])	Jackrel et al.34	N/A
C. elegans UA262 (baEx153 a,b,c [Pdat-1::Hsp104-DPLF:A503V, Pmyo-2::mCherry]; baIn11[Pdat-1::α-syn, Pdat-1::GFP])	Jackrel et al.34	N/A
Oligonucleotides
See Table S1
Recombinant DNA
pRS416GAL	Jackrel et al.34	N/A
pRS416GAL-Hsp104	Jackrel et al.34	N/A
pRS416GAL-Hsp104A503S	Jackrel et al.34	N/A
pRS416GAL-Hsp104A503S:Y257L	This study	N/A
pRS416GAL-Hsp104A503S:Y257T	This study	N/A
pRS416GAL-Hsp104A503S:Y662M	This study	N/A
pRS416GAL-Hsp104A503S:Y257L:Y662M	This study	N/A
pRS416GAL-Hsp104K358D	This study	N/A
pRS416GAL-Hsp104K358D:Y257L	This study	N/A
pRS416GAL-Hsp104K358D:Y257F	This study	N/A
pRS416GAL-Hsp104K358D:Y257W	This study	N/A
pRS416GAL-Hsp104K358D:Y257I	This study	N/A
pRS416GAL-Hsp104K358D:Y257V	This study	N/A
pRS416GAL-Hsp104K358D:Y257M	This study	N/A
pRS416GAL-Hsp104K358D:Y662L	This study	N/A
pRS416GAL-Hsp104K358D:Y662F	This study	N/A
pRS416GAL-Hsp104K358D:Y662W	This study	N/A
pRS416GAL-Hsp104K358D:Y662I	This study	N/A
pRS416GAL-Hsp104K358D:Y662V	This study	N/A
pRS416GAL-Hsp104K358D:Y662M	This study	N/A
pRS416GAL-Hsp104D484K	This study	N/A
pRS416GAL-Hsp104D484K:Y257L	This study	N/A
pRS416GAL-Hsp104D484K:Y257F	This study	N/A
pRS416GAL-Hsp104D484K:Y257W	This study	N/A
pRS416GAL-Hsp104D484K:Y257I	This study	N/A
pRS416GAL-Hsp104D484K:Y257V	This study	N/A
pRS416GAL-Hsp104D484K:Y257M	This study	N/A
pRS416GAL-Hsp104D484K:Y662L	This study	N/A
pRS416GAL-Hsp104D484K:Y662F	This study	N/A
pRS416GAL-Hsp104D484K:Y662W	This study	N/A
pRS416GAL-Hsp104D484K:Y662I	This study	N/A
pRS416GAL-Hsp104D484K:Y662V	This study	N/A
pRS416GAL-Hsp104D484K:Y662M	This study	N/A
pRS416GAL-Hsp104Y257L	This study	N/A
pRS416GAL-Hsp104Y662L	This study	N/A
pRS416GAL-Hsp104K358D:Y257L:Y662L	This study	N/A
pRS416GAL-Hsp104K358D:Y257L:Y662F	This study	N/A
pRS416GAL-Hsp104K358D:Y257L:Y662W	This study	N/A
pRS416GAL-Hsp104K358D:Y257L:Y662M	This study	N/A
pRS416GAL-Hsp104K358D:Y257L:Y662S	This study	N/A
pRS416GAL-Hsp104K358D:Y257L:Y662T	This study	N/A
pRS416GAL-Hsp104K358D:Y257L:Y662Q	This study	N/A
pRS416GAL-Hsp104K358D:Y257L:Y662N	This study	N/A
pRS416GAL-Hsp104K218T:K620T	Torrente et al.39	N/A
pRS416GAL-Hsp104K218T:A503S:K620T	This study	N/A
pRS416GAL-Hsp104K218T:K358D:K620T	This study	N/A
pRS416GAL-Hsp104K218T:Y257L:K358D:K620T	This study	N/A
pRS416GAL-Hsp104K218T:Y257T:K358D:K620T	This study	N/A
pRS416GAL-Hsp104K218T:K358D:K620T:Y662M	This study	N/A
pRS416GAL-Hsp104K218T:Y257L:K358D:K620T:Y662M	This study	N/A
pRS416GAL-Hsp104K358D:Y257S	This study	N/A
pRS416GAL-Hsp104K358D:Y257T	This study	N/A
pRS416GAL-Hsp104K358D:Y257Q	This study	N/A
pRS416GAL-Hsp104K358D:Y257N	This study	N/A
pRS416GAL-mCherry	This study	N/A
pRS416GAL-Hsp104-mCherry	This study	N/A
pRS416GAL-Hsp104K358D-mCherry	This study	N/A
pRS416GAL-Hsp104A503S-mCherry	This study	N/A
pRS416GAL-Hsp104K358D:Y257T-mCherry	This study	N/A
pRS416GAL-Hsp104K358D:Y257L-mCherry	This study	N/A
pRS416HSE-Hsp104	Jackrel et al.34	N/A
pRS416HSE-Hsp104A503V	Jackrel et al.34	N/A
pRS416HSE-Hsp104K358D	This study	N/A
pRS416HSE-Hsp104K358D:Y257L	This study	N/A
pRS416HSE-Hsp104K358D:Y257T	This study	N/A
pRS416HSE-Hsp104K358D:Y662M	This study	N/A
pRS416HSE-Hsp104K358D:Y257L:Y622M	This study	N/A
pAG413GAL	Alberti et al.100	N/A
pAG413GAL-α-syn	This study	N/A
pAG413GAL-α-synE46K	This study	N/A
pAG413GAL-α-synA53T	This study	N/A
pAG413GAL-α-syn1-95	This study	N/A
pAG413GAL-α-syn1-115	This study	N/A
pAG413GAL-α-syn1-125	This study	N/A
pAG415GAL	Alberti et al.100	N/A
pAG415GAL-α-syn	This study	N/A
pAG415GAL-α-synE46K	This study	N/A
pAG415GAL-α-synA53T	This study	N/A
pAG415GAL-α-syn1-95	This study	N/A
pAG415GAL-α-syn1-115	This study	N/A
pAG415GAL-α-syn1-125	This study	N/A
pNOTAG-Hsp104	Jackrel et al.34	N/A
pNOTAG-Hsp104A503S	Jackrel et al.34	N/A
pNOTAG-Hsp104K358D	This study	N/A
pNOTAG-Hsp104K358D:Y257L	This study	N/A
pNOTAG-Hsp104K358D:Y257T	This study	N/A
pNOTAG-Hsp104K358D:Y662M	This study	N/A
pNOTAG-Hsp104K358D:Y257L:Y662M	This study	N/A
pNOTAG-HAP	Jackrel et al.34	N/A
pNOTAG-HAPA503S	This study	N/A
pNOTAG-HAPK358D	This study	N/A
pNOTAG-HAPK358D:Y257L	This study	N/A
pNOTAG-HAPK358D:Y257T	This study	N/A
pNOTAG-HAPK358D:Y662M	This study	N/A
pNOTAG-HAPK358D:Y257L:Y662M	This study	N/A
pTrc99A-GroELtrap	Jackrel et al.34	N/A
pBAD-RepA1-25-GFP	Lopez et al.99	N/A
pET28b-ClpP-his	Jackrel et al.34	N/A
pE-SUMO-Hsp72	This study	N/A
pE-SUMO-Hdj1	Michalska et al.20	N/A
pHis-TEV	Cupo and Shorter98	N/A
pT7-7 α-syn	Lo Bianco et al.27	N/A
pDUET-GST-TEV-TDP-43	Johnson et al.71	N/A
Software and algorithms
Prism 9	GraphPad	N/A
ImageJ	Rueden et al.101	N/A
Other
Custom code for a semi-automated procedure was used to quantify colocalization between the red and green images.	Zenodo	doi: 10.5281/zenodo.8185164

EXPERIMENTAL MODEL AND STUDY PARTICIPANT DETAILS

Yeast strains

Yeast strains used were wild-type W303a (MATa, can1-100, his3-11, 15, leu2-3, 112, trp1-1, ura3-1, ade2-1) or the isogenic strain W303aΔhsp104.³⁴ The yeast strains W303aΔhsp104-pAG303GAL-α-syn-YFP-pAG304GAL-α-syn-YFP, W303aΔhsp104-pAG303GAL-FUS, W303aΔhsp104-pAG303GAL-TDP-43, W303aΔhsp104-pAG303GAL-TDP-43-GFPS11-pAG305GAL-GFPS1-10, and W303aΔhsp104-pAG303GAL-FUS-GFP have been described previously.^34,35,102 Yeast were grown in rich medium (YPD) or in synthetic media without amino acids used for selection. 2% sugar (dextrose, raffinose, or galactose) was added to synthetic media.

C. elegans

Nematodes were grown and maintained using well-established procedures.¹⁰³ Plasmid constructs were injected into worms to generate transgenic animals using previously described methods.¹⁰⁴ Strains UA367 (baEx198 a,b,c [P_dat-1::HSP104-K358D, rol-6];baln11 [P_dat-1::α-syn, P_dat-1::GFP]), UA368 (baEx199 a,b,c [P_dat-1::HSP104-K358D:Y257L, rol-6];baln11 [P_dat-1::α-syn, P_dat-1::GFP]), UA369 (baEx200 a,b,c [P_dat-1::HSP104-K358D:Y257T, rol-6];baln11 [P_dat-1::α-syn, P_dat-1::GFP]), UA370 (baEx201 a,b,c [P_dat-1::HSP104-K358D:Y257L:Y662M, rol-6];baln11 [P_dat-1::α-syn, P_dat-1::GFP]), and UA371 (baEx202 a,b,c [P_dat-1::HSP104-K358D:Y662M, rol-6];baln11 [P_dat-1::α-syn, P_dat-1::GFP]) were generated by injecting 50 ng/μl of corresponding plasmid construct into UA44 (baln11 [P_dat-1::α-syn, P_dat-1::GFP]) with phenotypic marker (rol-6, 50 ng/μl, for roller expression). Three independent stable lines were created for each group (a, b, c). Strain UA256 (baEx147 a,b,c [P_dat-1::Hsp104 WT, P_myo-2::mCherry]; baIn11 [P_dat-1::α-syn, P_dat-1::GFP]), UA259 (baEx150 a,b,c [P_dat-1::Hsp104-A503S, P_myo-2::mCherry]; baIn11 [Pdat-1::α-syn, P_dat-1::GFP]), and UA262 (baEx153 a,b,c [P_dat-1::Hsp104-DPLF:A503V, P_myo-2::mCherry]; baIn11[P_dat-1:: α-syn, P_dat-1::GFP]) were previously generated.³⁴

METHOD DETAILS

Yeast plasmids

Hsp104 variants were under control of a galactose-inducible promoter on pRS416GAL plasmids, except for in thermotolerance assays where they were under control of the HSE promoter on pRS416HSE plasmids. For mCherry-tagged Hsp104 variants, mCherry is located at the C-terminal end of Hsp104, separated by a glycine-serine linker (also on pRS416GAL plasmids). For experiments in Figure S4, α-syn variants (α-syn, α-syn^E46K, α-syn^A53T, α, α-syn^1-115, or α-syn^1-125) were under control of a galactose-inducible promoter on pAG413GAL and pAG415GAL plasmids.

Site-directed mutagenesis

Mutations were introduced into Hsp104 through QuikChange site-directed mutagenesis (Agilent) and confirmed by DNA sequencing.

Yeast transformation and spotting assays

Plasmids containing Hsp104 variants were transformed into yeast using a standard lithium acetate and polyethylene glycol procedure.¹⁰⁵ For spotting assays, yeast cultures were grown to saturation overnight at 30°C in dropout media containing raffinose. Raffinose cultures were then normalized according to OD₆₀₀, five-fold serial diluted, and spotted onto glucose and galactose plates using a 96-bolt replicator tool. Plates were grown at 30°C for 2–3 days and imaged.

Toxicity spotting assay

pRS416GAL plasmids containing Hsp104 variants were transformed into W303aΔhsp104 yeast. Yeast cultures were grown to saturation overnight at 30°C in dropout media containing raffinose. Raffinose cultures were then normalized according to OD₆₀₀ and five-fold serial diluted. The cultures were spotted onto two sets of glucose and galactose plates using a 96-bolt replicator tool. One set of plates was grown at 30°C, and the other at 37°C, for 2–3 days and subsequently imaged.

Western blotting

Hsp104 variants transformed into appropriate yeast strains were grown to saturation overnight at 30°C in dropout media containing raffinose. Cultures were normalized to OD₆₀₀ = 0.3 and grown in galactose dropout media at 30°C to induce Hsp104 and disease substrate expression (α-syn cultures induced for 8h, TDP-43 and FUS cultures induced for 5h). Galactose cultures were then normalized according to OD₆₀₀ and the equivalent of 6ml culture with an OD₆₀₀ = 0.6 were harvested by centrifugation. Media was aspirated, and the cell pellets were resuspended in 0.1M NaOH and incubated at room temperature for 5min. Cells were pelleted again by centrifugation, supernatant removed, and pellet was resuspended in 100µL 1X SDS sample buffer and boiled for 4–5min. Samples were separated via SDS-PAGE (4–20% gradient, Bio-Rad) and transferred to a PVDF membrane (Millipore) using a Trans-Blot SD Semi-Dry Transfer Cell (Bio-Rad). Membranes were blocked for at least 1h at room temperature and then incubated with primary antibodies (rabbit anti-Hsp104 polyclonal (Enzo Life Sciences); rabbit anti-FUS polyclonal (Bethyl Laboratories); rabbit anti-TDP-43 polyclonal (Proteintech); rabbit anti-GFP polyclonal (Sigma-Aldrich); mouse anti-PGK1 monoclonal (Thermo Fisher)) at 4°C overnight. Membranes were washed multiple times with PBS-T, incubated with secondary antibodies (goat anti-mouse and goat anti-rabbit, LI-COR) for 1h at room temperature, and washed again multiple times with PBS-T (final wash with PBS). Membranes were imaged using a LI-COR Odyssey FC Imaging system.

Thermotolerance assay

Hsp104 variants under the HSE promoter were transformed into W303aΔhsp104 yeast. Yeast cultures were grown to saturation overnight at 30°C in glucose dropout media. Cultures were normalized to OD₆₀₀ = 0.3 and grown in glucose dropout media at 30°C for at least 4h, after which the equivalent of 6 ml culture with an OD₆₀₀ = 0.6 was grown at 37°C for 30 min (if assessing Hsp104 expression, samples would be harvested at this stage for western blot as described above). Cultures were then heat-shocked at 50°C in 1.5ml Eppendorf tubes in an Eppendorf Thermomixer for 30min and incubated on ice for 2min. Cultures were diluted appropriately, plated on glucose dropout media, and incubated at 30°C. After 2–3 days, colonies were counted using an aCOLyte colony counter and software (Synbiosis). W303a yeast were transformed with an empty pRS416Gal vector and treated as above for endogenous Hsp104 control.

Fluorescence microscopy

Yeast strains for fluorescence microscopy were as in Jackrel et al.³⁴ Hsp104 variants on pRS416GAL plasmids were transformed into each yeast strain (α-syn-YFP, FUS-GFP, or TDP43-GFPS11 from Jackrel et al.³⁴) and grown to saturation overnight at 30°C in dropout media containing raffinose. Cultures were normalized to OD₆₀₀ = 0.3 and grown at 30°C in galactose dropout media to induce Hsp104 and disease substrate expression (α-syn cultures induced for 8h, TDP-43 and FUS cultures induced for 5h). For α-syn and FUS cultures, cells were harvested and processed for microscopy, and imaged as live cells. For TDP-43 cultures, cells were harvested and fixed by spinning cells down, resuspending in 1ml cold 70% ethanol, and immediately pelleting cells again. Cells were washed 3 times with cold PBS and resuspended in Vectashield mounting medium with 4’,6-diamidino-2-phenylindole (DAPI) (Vector Laboratories). Cells were imaged at 100X magnification using a Leica DM IRBE microscope. Cells were analyzed using ImageJ software, and a minimum of 200 cells were quantified per sample in at least three independent trials. One-way ANOVA with Dunnett’s post hoc test was performed using GraphPad Prism Software.

For fluorescence microscopy to assess mCherry-GFP localization, Hsp104 variants with a C-terminal mCherry tag, or mCherry alone (on pRS416GAL plasmids) were transformed into the α-syn-YFP strain described in Jackrel et al.³⁴ Cultures were treated as α-syn cultures described above. After induction, cells were harvested, stained with Hoechst dye (channel not shown in Figure 5C) and imaged at 100X magnification using a Leica DM IRBE microscope. For analysis of mCherry colocalization with GFP, cells were imaged in red (TX2) and green (L5) channels. A semi-automated procedure was used to quantify colocalization between the red and green images (code available at https://github.com/bensondaled/mack-cell-annotation). Cells containing α-syn-YFP foci were manually identified and selected using a custom interface. A colocalization metric for each cell was then defined as the Pearson correlation between the pixels in the red and green channels. A cell was defined to exhibit colocalization if this measure exceeded 0.85, a uniform threshold applied to all images and chosen for its agreement with manual colocalization assessments. Alternatively, colocalization was determined manually. A minimum of 200 cells were quantified per sample in at least three independent trials. One-way ANOVA with Dunnett’s post hoc test was performed using GraphPad Prism Software.

C. elegans dopaminergic neurodegeneration

For dopaminergic neurodegeneration analyses, the transgenic animals were scored as described previously.¹⁰⁶ Briefly, on the day of analysis, the six anterior dopaminergic neurons [four CEP (cephalic) and two ADE (anterior deirid)] were examined in 30 animals randomly. Each nematode was considered normal when all six anterior DA neurons were present. However, if a worm exhibited any degenerative phenotype, such as a missing dendritic process, cell body loss, or a blebbing neuronal process, it was scored as degenerating. Three independent transgenic worm lines were analyzed per genetic background for a total of ninety animals per transgenic line and an average of total percentage of worms with normal neurons was reported in the study. A one-way ANOVA, followed by a Dunnett’s post hoc test, was performed for statistical analysis using GraphPad Prism Software.

Quantitative Real-Time PCR in nematodes

RNA isolation and qRT-PCR were performed on worms using previously published methods.^106,107 Briefly, total RNA was isolated from 100 young adult (day 4 post hatching) nematodes from corresponding transgenic line using TRI reagent (Molecular Research Center). 1μg of RNA was used for cDNA synthesis using the iScript Reverse Transcription Supermix for qRT-PCR (Bio-Rad, Hercules, CA, USA). qRT-PCR was performed using IQ-SYBR Green Supermix (Bio-Rad) with the Bio-Rad CFX96 Real-Time System. Hsp104 expression levels were normalized to two reference genes (snb-1 and y45). α-syn expression levels were normalized to three reference genes (ama-1, cdc-42, and tba-1). The reference target stability was analyzed by GeNorm and passed for all reference genes listed above. Three technical replicates with three biological replicates were performed in this study. Each primer pair was confirmed for at least 90–100% efficiency by standard curve analyses. The following primers were used for the assays:

hsp104 Forward: GGCCATGAAGAATTGACACAAG

hsp104 Reverse: CCTTAAATCAGCGGCGGTAG

α-syn Forward: ATGTAGGCTCCAAAACCAAGG

α-syn Reverse: ACTGCTCCTCCAACATTTGTC

snb-1 Forward: CCGGATAAGACCATCTTGACG

snb-1 Reverse: GACGACTTCATCAACCTGAGC

y45 Forward: GTCGCTTCAAATCAGTTCAGC

y45 Reverse: GTTCTTGTCAAGTGATCCGACA

ama-1 Forward: TCCTACGATGTATCGAGGCAA

ama-1 Reverse: CTCCCTCCGGTGTAATAATGA

cdc-42 Forward: CCGAGAAAAATGGGTGCCTG

cdc-42 Reverse: TTCTCGAGCATTCCTGGATCAT

tba-1 Forward: ATCTCTGCTGACAAGGCTTAC

tba-1 Reverse: GTACAAGAGGCAAACAGCCAT

Protein purification

Hsp104 and HAP

Hsp104 and HAP proteins were purified as described³⁴ with the following modifications. Eluate from Affi-Gel Blue Gel was equilibrated to a low-salt buffer Q (~100mM NaCl, 20mM TRIS pH 8.0, 5mM MgCl₂, 0.5mM EDTA) and purified via ResourceQ anion exchange chromatography. Buffer Q (20mM TRIS pH 8.0, 50mM NaCl, 5mM MgCl₂, 0.5mM EDTA) was used as running buffer, and the protein was eluted with a linear gradient of buffer Q+ (20mM TRIS pH 8.0, 1M NaCl, 5mM MgCl₂, 0.5mM EDTA). The eluted protein was buffer-exchanged into high-salt storage buffer (40mM HEPES-KOH pH 7.4, 500mM KCl, 20mM MgCl₂) plus 10% glycerol and 1mM DTT and snap-frozen.

GroEL_trap

pTrc99A-GroEL_trap was transformed into DH5α competent E. coli cells (Thermo Fisher). Cells were grown in 2xYT medium with appropriate antibiotics at 37°C with shaking until OD₆₀₀ reached ~ 0.4 – 0.6. Protein overexpression was induced with 1mM IPTG, and cells were grown at 37°C until OD₆₀₀ ~ 2.0. Cells were harvested by spinning (5000g, 4°C, 15 min) and pellet was resuspended in 50mM sodium phosphate buffer and centrifuged (5000g, 4°C, 15min). Pellet was resuspended in low-salt buffer (50mM Tris-HCl pH 7.5, 1mM EDTA, 1mM DTT, 50mM NaCl) and 10mg lysozyme per g cell pellet. Sample stirred gently for 5min, lysed through sonication, and centrifuged (17,000 g, 4°C, 30min). Clarified lysate loaded onto HiTrap Q HP column (GE Healthcare) and eluted through salt gradient using low-salt buffer (as described above) and high-salt buffer (50mM Tris-HCl pH 7.5, 1mM EDTA, 1mM DTT, 500mM NaCl¹⁰⁸). Collected fractions were exchanged into the following TKME-100 buffer: 20mM TRIS-HCl pH 7.5, 100mM KCl, 10mM MgCl₂, 0.1mM EDTA, 5mM DTT, 10% glycerol, 0.005% Triton X-100 and snap-frozen.

RepA_1-25-GFP

pBAD-RepA_1-25-GFP was transformed into BL21 (DE3)-RIL cells. Cells were inoculated in 2xYT medium with appropriate antibiotics at 37°C with shaking until OD₆₀₀ reached ~ 0.6 – 0.8. Protein overexpression was induced with 1mM IPTG, and cells were grown at 30°C for 4h. Cells were harvested by spinning (4000rpm, 25min) and pellet was resuspended in 40mM HEPES-KOH pH 7.4 plus 2mM 2-Mercaptoethanol (BME) and EDTA-free protease inhibitors. Cells were lysed using a sonicator and centrifuged (16,000rpm, 20min). The resulting pellet was washed twice with HM buffer (40mM HEPES-KOH pH 7.4, 20mM MgCl₂) plus 2mM BME. After each wash, cells were centrifuged (16,000 rpm, 20min). Pellet was then resuspended in buffer containing 8M urea, 40mM Tris-HCl pH 6.8, 500mM NaCl, 10% glycerol (v/v) and agitated slowly overnight at 25°C. Sample was centrifuged (16,000rpm, 20min) and supernatant was kept. Supernatant was incubated with Ni-NTA beads (HisPur^™ Ni-NTA Resin, Thermo Scientific) pre-equilibrated in buffer containing 8M urea, 40mM Tris pH 6.8, 500mM NaCl, 10% glycerol (v/v) for 2h on a spinning wheel at 25°C at the lowest speed. Sample washed 5 times with buffer containing 8M urea, 40mM Tris-HCl pH 6.8, 500mM NaCl, 20mM imidazole, 10% glycerol (v/v). Sample then washed 5 times with buffer containing 8M urea, 40mM Tris-HCl pH 6.8, 500mM NaCl, 40mM imidazole, 10% glycerol (v/v). Sample eluted with buffer containing 8 M urea, 40mM Tris-HCl pH 6.8, 500mM NaCl, 500mM imidazole, 10% glycerol (v/v). Sample was dialyzed overnight into buffer containing 40mM HEPES-KOH pH 7.4, 20mM imidazole, 150mM KCl, 2mM BME, 10% glycerol (v/v) at 4°C and re-loaded onto a Ni-NTA column (HisTrap^™ HP, GE Healthcare). An imidazole gradient was applied (from 20mM to 500mM) over 20CV in buffer containing 40mM HEPES-KOH pH 7.4, 20mM imidazole, 150mM KCl, 2mM BME, 10% glycerol (v/v). The purity of eluted fractions was assessed using SDS-PAGE. Collected fractions were buffer-exchanged into HKM-150 buffer (40mM HEPES-KOH pH 7.4, 150mM KCl, 20mM MgCl₂) plus 2 mM BME and 10% glycerol (v/v) and snap-frozen.

ClpP

C-terminally His-tagged ClpP was purified as described³⁴.

Hsp70 and Hsp40

Hsc70 and Hdj2 were from Enzo Life Sciences. Hdj1 was purified as in Michalska et al.²⁰ Hsp72 was purified as for Hsc70 in Michalska et al.²⁰

α-Syn and TDP-43

α-Syn and TDP-43 were purified as described.³⁴

ATPase assay

Hsp104 (0.25µM monomer in Hsp104 buffer, see Hsp104 purification) was incubated with ATP (1 mM) for 5min at 25°C in luciferase-refolding buffer (LRB: 25mM HEPES-KOH pH 7.4, 150mM KAOc, 10mM MgAOc, 10mM DTT). The final reaction buffer contained 6.3% of HKM-500 buffer. ATPase activity was evaluated by the release of inorganic phosphate, which was measured using a malachite green phosphate detection kit (Innova Biosciences). Background hydrolysis at time zero was subtracted.

FITC-casein degradation assay

FITC-casein (0.1µM) was incubated with HAP or HAP variants (1µM monomer), ClpP (21µM monomer), ATP (5mM), and an ATP regeneration system (1mM creatine phosphate and 0.25µM creatine kinase). ClpP was first buffer-exchanged into luciferase-refolding buffer (LRB: 25mM HEPES-KOH pH 7.4, 150mM KAOc, 10mM MgAOc, 10mM DTT) at 25°C. Reaction mixtures were assembled 25°C in LRB and degradation of FITC-casein was monitored by measuring fluorescence (excitation 490nm, emission 520nm) using a Tecan Infinite M1000 plate reader.

RepA_1-25-GFP unfoldase assay

RepA_1-25-GFP (0.7µM) was incubated with Hsp104 or Hsp104 variants (2.1µM hexamer), ATP (4mM), ARS (20mM creatine phosphate, 0.06µg/µl creatine kinase). GroEL_trap (2.5µM tetradecamer) was included to prevent refolding of unfolded RepA_1-25-GFP. Hsp104 variants were buffer-exchanged into TKME-100 buffer at 25°C. Reactions were assembled on ice in TKME-100 buffer plus 20µg/ml BSA. RepA_1-25-GFP unfolding was measured by fluorescence (excitation 395nm, emission 510nm) using a Tecan Safire², which was heated to 30°C prior to reading.

Luciferase reactivation assay

Aggregated luciferase (100nM) was incubated with Hsp104 or Hsp104 variants (1µM monomer), ATP (5mM), and an ATP regeneration system (10mM creatine phosphate, 0.25µM creatine kinase) in the presence or absence of co-chaperones Hsp70 (Hsc70 or Hsp72, 0.167µM) and Hsp40 (Hdj2 or Hdj1, 0.167µM) for 90min at 25°C in LRB. The final reaction buffer contained 8.6% of HKM-500 buffer. After 90min, luciferase activity was measured with a luciferase assay reagent (Promega). Recovered luminescence was measured using a Tecan Infinite M1000 plate reader.

α-Syn and TDP-43 fibril disaggregation

α-Syn (80µM) was assembled into fibrils via incubation in 40mM HEPES-KOH (pH 7.4), 150 mM KCl, 20 mM MgCl₂, and 1 mM dithiothreitol for 48h at 37°C with agitation. α-Syn fibrils (3µM monomer) were incubated without or with Hsp104, Hsp104^A503S, Hsp104^K358D, Hsp104^K358D:Y257L, Hsp104^K358D:Y257T or Hsp104^K358D:Y662M (3µM) plus Hsc70 (3µM), Hdj1 (3µM), ATP (20mM) and an ATP regeneration system (20mM creatine phosphate and 0.5µM creatine kinase) for 2h at 30°C. Disaggregation was assessed by Thioflavin-T fluorescence. ²⁷ To generate TDP-43 fibrils, GST-TEV-TDP-43 (6µM) was incubated with TEV protease in 40mM HEPES-KOH (pH 7.4), 150mM KCl, 20mM MgCl₂, and 1mM dithiothreitol for 16h at 25°C with agitation, by which time all the TDP-43 had aggregated. TDP-43 fibrils (3µM monomer) were incubated without or with Hsp104, Hsp104^A503S, Hsp104^K358D, Hsp104^K358D:Y257L, Hsp104^K358D:Y257T or Hsp104^K358D:Y662M (3µM) plus Hsc70 (3µM), Hdj1 (3µM), ATP (20 mM) and an ATP regeneration system (20mM creatine phosphate and 0.5µM creatine kinase) for 2h at 30°C. Disaggregation was assessed by turbidity (absorbance at 350nm).

Soluble α-syn oligomer disassembly

Soluble α-syn oligomers were purified by gel filtration as described.¹⁰⁹ For disassembly experiments, soluble α-syn oligomers (0.5 μM monomer) were incubated without or with Hsp104, Hsp104^A503S, Hsp104^K358D, Hsp104^K358D:Y257L, Hsp104^K358D:Y257T or Hsp104^K358D:Y662M (1µM) in KHM buffer (40 mM HEPES-KOH, pH 7.4, 150 mM KCl, 20 mM MgCl2, 1 mM DTT) plus ATP (20 mM) and an ATP regeneration system (20mM creatine phosphate and 0.5µM creatine kinase) for 1 hour at 37°C. Reactions were then fractionated through a Microcon YM-100 (100-kDa molecular weight cut off) filter (Milllipore). Soluble α-syn oligomers are trapped by the filter, and only α-syn monomers enter the filtrate fraction.²⁷ The amount of α-syn in the filtrate fraction was then determined using α-syn ELISA Kit (Thermofisher).

QUANTIFICATION AND STATISTICAL ANALYSIS

Quantification is as described in the figure legends. Statistical analyses were performed using GraphPad Prism (GraphPad Software, Inc.; La Jolla, CA, USA) as described in figure legends.

Highlights.

• Tuning Hsp104 specificity to selectively detoxify α-synuclein (α-syn)

• Hsp104 variants with enhanced selectivity for α-syn oligomers and fibrils

• Hsp104 variants with enhanced selectivity for soluble α-syn oligomers

• α-syn-specific Hsp104 variants confer increased neuroprotection in C. elegans