Engineering a membrane protein chaperone to ameliorate the proteotoxicity of mutant huntingtin

Abstract

Toxic protein aggregates are associated with various neurodegenerative diseases, including Huntington's disease (HD). Since no current treatment delays the progression of HD, we develop a mechanistic approach to prevent mutant huntingtin (mHttex1) aggregation. Here, we engineer the ATP-independent cytosolic chaperone PEX19, which targets peroxisomal membrane proteins to peroxisomes, to remove mHttex1 aggregates. Using yeast toxicity-based screening with a random mutant library, we identify two yeast PEX19 variants and engineer equivalent mutations into human PEX19 (hsPEX19). These variants effectively delay mHttex1 aggregation in vitro and in cellular HD models. The mutated hydrophobic residue in the α4 helix of hsPEX19 variants binds to the N17 domain of mHttex1, thereby inhibiting the initial aggregation process. Overexpression of the hsPEX19-FV variant rescues HD-associated phenotypes in primary striatal neurons and in Drosophila. Overall, our data reveal that engineering ATP-independent membrane protein chaperones is a promising therapeutic approach for rational targeting of mHttex1 aggregation in HD.

Fig. 1. Identification of scPEX19 variants that suppress cellular toxicity of mHttex1.

a Yeast toxicity-based screening to identify mHttex1 suppressors. scPEX19 plasmid library generated by random mutagenesis was transformed into Httex1-97QΔP-GFP-integrated yeast cells. Both scPEX19 variants and Httex1-97QΔP-GFP are under the control of the GAL1 promoter. The sequences of identified scPEX19 mutants, m1, and m2, are shown, and the common mutation sites are highlighted in red. Created in BioRender. Hyunju, C. (2025) https://BioRender.com/o38y171. b, c Growth test of Httex1-25QΔP-GFP- and Httex1-97QΔP-GFP-integrated yeast cells expressing scPEX19-WT and its scPEX19 variants. Five-fold serial dilutions of cells were spotted on galactose plates to coexpress Httex1-25QΔP or Httex1-97QΔP and the indicated scPEX19 proteins (Right) or on glucose plates as loading controls (Left). Representative images from three biological replicates (n = 3). d, e (d) Confocal microscopy images of Httex1-25QΔP-GFP and Httex1-97QΔP-GFP cells upon coexpression of scPEX19-WT, scPEX1.9-FV, and scPEX19-FI. Scale bar: 10 µm. The percentage of cells containing 97QΔP aggregates was quantified in (e). f, g (f) Representative image of Western blot monitoring SDS-insoluble and SDS-soluble Httex1-97QΔP-GFP proteins and (g) quantification of SDS-insoluble protein in (f). Yeast cells were lysed using glass beads, and then the total cell lysates were analyzed using Western blot. N-terminally FLAG-tagged Httex1-97QΔP-GFP was probed using FLAG antibody. PGK1 serves as a loading control. Data in (e), (g) are shown as mean ± SD, with three biological replicates (n = 3). Pairwise comparisons are shown as indicated, where ****p < 0.0001 by the ordinary one-way ANOVA with Tukey post-hoc test. Source data are provided as a Source Data file.

Fig. 2. hsPEX19 variants suppress mHttex1 aggregation in vitro and in mammalian cells.

a Multiple sequence alignment of the α4 helix sequences of PEX19 across various species. The alignment was performed using the Clustal Omega and displayed with ESPript 3^,. Conserved sequences of scPEX19-L288 and scPEX19-E292 are highlighted as green boxes. b NMR structure of hsPEX19-CTD (161–299 aa) (PDB 5LNF). Two conserved residues, M255 and Q259, are located in the α4 helix of hsPEX19 and are shown in magenta. The M255 residue of hsPEX19 is known to bind its C-terminally modified farnesyl group (cyan). c, d In vitro aggregation assay of Httex1-51Q in the absence and presence of hsPEX19 proteins. 3 μM of GST-TEV-Httex1-51Q-Stag and 1.5 μM (0.5×) or 3 μM (1×) of PEX19 proteins were incubated at 30 °C, and after the addition of TEV protease, samples were quenched at the indicated time points. SDS-insoluble Httex1-51Q aggregates in (c) and their replicates were quantified and shown in (d) (n = 3, mean ± SD). e ThioflavinT fluorescence assay to measure fibril formation of Httex1-51Q. The fluorescence intensity was measured every 15 min. Data are shown as mean ± SD with n  =  3 (technical replicates). f Negatively stained transmission electron micrograph (TEM) images of Httex1-51Q in the absence and presence of hsPEX19 proteins. Scale bar: 500 nm g, h (g) Confocal microscopy images of HEK293T cells coexpressing Httex1-19Q-GFP or Httex1-134Q-GFP and hsPEX19. Empty vector control (denoted as vector control) was used as a negative control. Scale bar: 50 µm. h The percentage of cells containing 134Q aggregates was quantified. i (top) Representative image of the filter trap assays monitoring the SDS-insoluble Httex1-134Q-GFP aggregates in HEK293T cells upon coexpression of hsPEX19 proteins. (bottom) Quantification of the images and their replicates. Data in (h, i) are shown as mean ± SD, with three biological replicates (n = 3). Pairwise comparisons are shown as indicated, where **p < 0.01, ***p < 0.001, ****p < 0.0001 by the ordinary one-way ANOVA with Tukey post-hoc test. Source data are provided as a Source Data file.

Fig. 3. The α4 helix of hsPEX19 variants directly interacts with the N17 domain of mHttex1.

a N-terminal GST-tagged Httex1-51Q-WT or Httex1-51Q-ΔN were used to monitor Htt51Q aggregation or interaction between hsPEX19 and Httex1-51Q proteins in vitro. The helical wheel illustrates the distribution of hydrophobic amino acids in the amphipathic helix of the N17 domain of Httex1-51Q. The sequences of hydrophobic amino acids in the N17 domain are also highlighted in red. “*” represents the Bpa incorporation site on Httex1-51Q. b, c In vitro aggregation assay of Httex1-51Q-ΔN in the absence and presence of hsPEX19 variants. SDS-insoluble aggregates of Httex1-51Q in (b) and their replicates were quantified and shown in (c) (n = 3, mean ± SD). d Schematic illustration of Ataxin3-78Q-WT and N17-Ataxin3-78Q. Ataxin3 consists of an N-terminal Josephin domain, several ubiquitin-interacting motifs (UIMs), and a polyQ repeat domain. The N17 domain of Httex1 was fused to the N-terminus of Ataxin3-78Q. e–g In vitro aggregation assay of Ataxin3-78Q-WT and N17-Ataxin3-78Q in the absence and presence of hsPEX19 variants. 30 µM of Ataxin3-78Q was incubated with 30 µM of hsPEX19 proteins at 37 °C. SDS-insoluble aggregates of Ataxin3-78Q-WT and N17-Ataxin3-78Q in (e) and their replicates were quantified and shown in (f) and (g), respectively (n = 3, mean ± SD). h, i Bpa crosslinking assay to monitor the direct association of hsPEX19^Bpa with Httex1-51Q-WT or Httex1-51Q-ΔN. 3 μM of hsPEX19-FV^Bpa or hsPEX19-FI^Bpa was incubated with an equimolar concentration of Httex1-51Q-WT or Httex1-51Q-ΔN for 3 h at 30 °C. Crosslinked samples were analyzed using Western blots probed with PEX19 (h) and S-tag (51Q) (i) antibodies. j, k Bpa crosslinking assay to monitor the intermolecular interaction of Httex1-51Q-F11^Bpa with hsPEX19-WT and its hsPEX19 variants. 3 μM of Httex1-51Q-F11^Bpa was incubated with 1.5 μM of hsPEX19 proteins for 3 h at 30 °C. Crosslinked samples were subjected to Western blot analysis against PEX19 (j) and S-tag (51Q) (k) antibodies. All Bpa crosslinking assays in (h–k) were performed twice or three times independently (n = 3 for (h and i) or 2 for (j and k)). Source data are provided as a Source Data file.

Fig. 4. The α4 helix of hsPEX19 variants is a specific binding site for the N17 domain of Httex1-51Q.

a Sequences of the N17 domain of Httex1-51Q and TMDs of peroxisomal and mitochondrial membrane proteins and their Grand Average of Hydropathy (GRAVY) scores. All listed membrane proteins are known to interact with hsPEX19 during their targeting to peroxisome or mitochondria^,,,. PEX26, PEX11B, and ACBD5 are peroxisomal tail-anchored membrane proteins (TAs), whereas Fis1 is a dually localized TA in mitochondria and peroxisomes^,. PEX13 and PMP34 are multi-spanning peroxisomal membrane proteins (PMPs)^,. ATAD1 is an N-terminal signal-anchored membrane protein localized in both mitochondria and peroxisome. b Schematic representation of Httex1-51Q, N17-MBP, MBP-WT, and PEX26. The isolated N17 sequence was N-terminally fused to MBP (maltose binding protein). The recombinant PEX26 protein contains the N-terminal 2 × Strep-tagged SUMO domain and the PEX26 targeting sequences (237–305 aa) encompassing the TMD and C-terminal charged tail of PEX26. c, d Bpa crosslinking assay of hsPEX19^Bpa with MBP-WT, N17-MBP, N17-MBP variants. e, f In vitro aggregation assay to monitor the chaperone activity of hsPEX19-FV and SGTA toward Httex1-51Q. SDS-insoluble Httex1-51Q aggregates in (e) and their replicates were quantified and shown in (f) (n = 3, mean ± SD). g, h Representative images of His₆-PEX19 pulldown assay with PEX26 in (g) and their quantification in (h). I, FT, and E denote total input, flowthrough, and elution, respectively. The amounts of PEX26 bound to hsPEX19 were analyzed by western blot against Strep (PEX26) and PEX19 antibodies. Data in (h) are shown as mean ± SD, with n = 3 (technical replicates). i–k Bpa crosslinking assay of hsPEX19^Bpa with either Httex1-51Q or PEX26. The Bpa crosslinking assays with Httex1-51Q were carried out under the same conditions as Fig. 3h and i. Prior to the Bpa crosslinking assay, 0.75 µM of PEX26 was incubated with 3 µM of hsPEX19^Bpa at room temperature for 5 min. “*” represents the SDS-resistant PEX26 dimers in (k). All Bpa crosslinking assays were performed three times independently (n = 3). Source data are provided as a Source Data file.

Fig. 5. hsPEX19-FV mitigates mHttex1-induced neurodegenerative phenotypes.

a Confocal microscopy images of mouse striatal neurons coexpressing Httex1-19Q-GFP and Httex1-134Q-GFP with vector control, hsPEX19-WT, or hsPEX19-FV. Tuj1 (neuron marker) was stained using Tuj1 antibody. Scale bar: 20 µm. b The degrees of fragmentation in a primary neuron are classified as four fragmentation scores: score 1 (fragmented areas in a single neuron are less than 5%), score 2 (5% < fragmented areas < 50%), score 3 (50% < fragmented areas < 90%), score 4 (fragmented areas > 90%). The heatmap shows the population of fragmentation scores for each condition (from left to right, n = 23, 19, 25, 27 neurons). c Climbing ability of 12-day-old adult flies (W¹¹¹⁸, Httex1-20Q, and Httex1-93Q) expressing vector control, hsPEX19-WT, and hsPEX19-FV in motor neurons. The data in (c) are shown as violin plots with mean and quartiles (from left to right, n = 102, 120, 119, 105, 103, 110, 101, 112, 105 adult flies). Climbing index (5 cm/5 sec). Statistical significance was evaluated using the two-way ANOVA with Tukey post-hoc test. **p < 0.01, ****p < 0.0001, ns = not significant. d Climbing ability of 10-day-old adult flies (W¹¹¹⁸, Httex1-20Q, and Httex1-93Q) expressing vector control, hsPEX19-WT, and hsPEX19-FV in pan-neurons. Climbing index (5 cm/5 sec). The data in (d) are shown as violin plots with mean and quartiles (from left to right, n = 109, 102, 91, 251, 205, 152, 125, 103, 130 adult flies). Statistical significance was evaluated using two-way ANOVA with Tukey post-hoc test. **p < 0.01, ****p < 0.0001, ns = not significant. e–g Lifespan analysis of W¹¹¹⁸, Httex1-20Q, and Httex1-93Q Drosophila expressing vector control, hsPEX19-WT, and hsPEX19-FV in pan-neurons. Lifespan data were plotted as Kaplan-Meier survival curves, and p-values were determined using the Log-rank (Mantel-Cox) test and Gehan-Breslow-Wilcoxon test. Vector control, hsPEX19-WT, hsPEX19-FV; (e) n = 140, 148, 167, (f) n = 159, 114, 117, (g) n = 164, 151, 121 adult flies. *p < 0.05, **p < 0.01, ****p < 0.0001, ns = not significant. Genotypes of Drosophila used in (c–g) are listed in Supplementary Table 1. Source data are provided as a Source Data file.