DNAJC5 facilitates USP19-dependent unconventional secretion of misfolded cytosolic proteins

Abstract

Cell-to-cell transmission of misfolded proteins propagates proteotoxic stress in multicellular organisms when transmitted polypeptides serve as a seeding template to cause protein misfolding in recipient cells, but how misfolded proteins are released from cells to initiate this process is unclear. Misfolding-associated protein secretion (MAPS) is an unconventional protein-disposing mechanism that specifically exports misfolded cytosolic proteins including various neurodegenerative disease-causing proteins. Here we establish the HSC70 co-chaperone DNAJC5 as an essential mediator of MAPS. USP19, a previously uncovered MAPS regulator binds HSC70 and acts upstream of HSC70 and DNAJC5. We further show that as a membrane-associated protein localized preferentially to late endosomes and lysosomes, DNAJC5 can chaperone MAPS client proteins to the cell exterior. Intriguingly, upon secretion, misfolded proteins can be taken up through endocytosis and eventually degraded in the lysosome. Collectively, these findings suggest a transcellular protein quality control regulatory pathway in which a deubiquitinase-chaperone axis forms a "triaging hub", transferring aberrant polypeptides from stressed cells to healthy ones for disposal.

Fig. 1. USP19 promotes the secretion of neurotoxic proteins.

a USP19 overexpression did not alter polyubiquitination profile. Whole cell extracts (WCE) from cells transfected with either an empty vector (E.V.), or wild type (WT) USP19, or USP19 C506S (CS) plasmids for 48 h were analyzed by immunoblotting (IB) Ub-S, ubiquitinated substrates. b USP19 overexpression does not affect cell viability. The viability of cells transfected with the indicated plasmids was monitored 48 h post-transfection (mean ± s.e.m., n = 3). c USP19 stimulates Tau secretion. Conditioned medium and cell lysate from 293T cells transfected as indicated were analyzed by immunoblotting (IB). Note that the secretion of the conventional ER cargo Clusterin is not affected by USP19 overexpression. d USP19 promotes the secretion of Atx3-Q84. As in c, except that Flag-Atx3-Q84 was used and that the medium was subject to immunoprecipitation (IP) by a Flag antibody followed by immunoblotting. Asterisk, IgG. e USP19 promotes the secretion of TDP-43. As in d, except that GFP-TDP-43 was tested. Asterisk, IgG

Fig. 2. Hsp90 interaction is dispensable for USP19-stimulated secretion.

a A schematic illustration of the USP19 constructs used in the study. UI UBL-containing insertion. b Cell lysates prepared from 293T cells expressing the indicated USP19 variants were subject to immunoprecipitation (IP) followed by immunoblotting. As a negative control, cells were transfected with an empty vector (E.V.). c A USP19 mutant lacking the UBL-containing insertion (UI) is a stronger MAPS stimulator. Secretion of α-Syn-Flag from cells expressing the indicated USP19 variants was analyzed by immunoblotting. The graph shows the relative level of α-Syn secretion normalized by α-Syn in cell lysates (mean ± s.e.m., n = 3)

Fig. 3. USP19 functions together with HSC70 and DNAJC5 in MAPS.

a Endogenous USP19 interacts with HSC70. USP19 immunoprecipitated (IP) from lysates of control (Ctrl.) or USP19 null CRISPR cells was analyzed by immunoblotting. The graph shows the quantification of two independent experiments. b USP19 interacts with HSC70 via the USP domain. c DNAJC5 overexpression increases the entry of mCherry-GFP1-10 (mCh-GFP1-10) into vesicles in a perinuclear region. COS7 cells transfected with the indicated plasmids together with mCitrine-Rab9 (mCi-Rab9) for 16 h were subject to 5 rounds of photobleaching in the indicated regions. The graph shows the number of GFP1-10 vesicles in the absence and presence of DNAJC5 (mean ± s.e.m., ***p < 0.005, cell number n = 20 for control and 21 for USP19-expressing cells). d DNAJC5 induces the entry of GFP1-10 into vesicles that contain Rab9. Cells co-transfected with mCh-GFP1-10, mCi-Rab9, and Flag-DNAJC5 were photobleached and imaged. e Colocalization of mCh-GFP1-10 with EGFP-α-Syn. Dual color photobleaching was performed using cells expressing both mCh-GFP1-10 and EGFP-α-Syn together with Flag-DNAJC5. Cells were then imaged by time lapse fluorescence confocal microscopy

Fig. 4. DNAJC5-stimulated MAPS is independent of USP19.

a DNAJC5 induces α-Syn secretion in both WT control (Ctrl.) and USP19 null CRISPR 293T cells. b The graph shows the quantification of three independent experiments in (a) (mean ± s.e.m., n = 3). For DNAJC5-overexpressing cells, the level of secretion is further normalized by DNAJC5 expression in cell lysate using control CRISPR cells as the reference. c DNAJC5 induces Tau secretion in both WT and USP19 null cells. d Flag-DNAJC5 expression in WT and USP19 null cells as determined by quantitative immunoblotting (mean ± s.e.m., n = 6). e As in (c) except that different amount of DNAJC5 plasmid were used. f Knockout of USP19 does not reduce DNAJC5-induced Tau secretion. Shown is the relative amount of Tau in medium (med.) normalized first by Tau and then by DNAJC5 in whole cell extract (WCE) from seven transfections in three independent experiments

Fig. 5. HSC70 and DNAJC5 function downstream of USP19 in MAPS.

a Knockdown of HSC70 reduces USP19-induced secretion of α-Syn. b HSC70 knockdown efficiency in the secretion experiments as determined by quantitative immunoblotting analyses (mean ± s.e.m., n = 3). c Quantification of α-Syn secretion in control (Ctrl.) and HSC70 knockdown cells that expressed USP19. Shown is the level of α-Syn secretion normalized by α-Syn in lysate (mean ± s.e.m., n = 3). d HSC70 knockdown reduces Tau secretion from cells overexpressing USP19. e Knockdown of DNAJC5 reduces USP19-mediated secretion of α-Syn. As in (a), except that DNAJC5 was knocked down. The graph shows the relative levels of α-Syn secretion from control and DNAJC5 knockdown cells that also express USP19. α-Syn in media is normalized by α-Syn and USP19 in lysate (mean ± s.e.m., n = 3). f DNAJC5 knockdown reduces USP19-induced Tau secretion. As in (e), except that F-Tau was expressed instead of α-Syn

Fig. 6. DNAJC5 is a late endosome-associated protein that promotes membrane association of a MAPS substrate.

a Endogenous DNAJC5 is mostly associated with membranes. 293T cells were fractionated into a cytosol (C) and a membrane (M) fraction for analysis by quantitative immunoblotting (IB) (mean ± s.e.m., n = 3). b Endogenous DNAJC5 is localized to compartments positive for Rab9 or LAMP1. 1–3 COS7 cells transfected with mCitrine-Rab9 (mCi-Rab9) were stained with a DNAJC5 antibody in red. The insets show an example of DNAJC5 co-localization with Rab9. 4–6, COS7 cells were stained with DNAJC5 antibody in red and LAMP1 antibody in green. 7–9 is an enlarged view of the boxed area in panel 6. c Membrane association requires the cysteine string domain (CS) of DNAJC5. COS7 cells expressing mCherry-Rab9 (mCh-Rab9) together with photoactivable (pA) GFP-tagged wild type (WT) DNAJC5 or a mutant lacking the CS segment were imaged after activation of DNAJC5 fluorescence in the indicated areas. d The CS domain is required to stimulate MAPS. Secretion of α-Syn from 293T cells expressing the indicated proteins was quantified and normalized to the levels of DNAJC5 in lysates (mean ± s.e.m., n = 3). e Endogenous DNAJC5 is co-localized with a MAPS substrate. Cells expressing mCerulean-USP19 (mCe-USP19) and mCh-GFP1-10 were permeabilized, fixed and stained with DNAJC5 in red, Flag in green and DAPI in blue. Arrowheads indicate lysosome-localized DNAJC5, which is not colocalized with GFP1-10. The two indicated regions that contain many small DNAJC5-positive vesicles were enlarged and shown in the right panels

Fig. 7. DNAJC5 is co-released with MAPS substrates.

a Interaction of DNAJC5 with secreted GFP1-10 in conditioned media. Conditioned media from transfected 293T cells was either directly analyzed by immunoblotting (IB), or first subject to IP by a Flag antibody before IB. Asterisk indicates the IgG bands. b Most GFP1-10 molecules secreted are in complex with DNAJC5. Conditioned media were either directly analyzed by IB (input), or first separated into a DNAJC5-positive and a DNAJC5 free fraction by IP with Flag antibody before IB. c USP19 enhances DNAJC5 secretion. Secretion of FALG-DNAJC5 from cells co-transfected with the indicated plasmids was analyzed. Where indicated, cell lysates were also analyzed by immunoblotting. The graph shows normalized DNAJC5 secretion (mean ± s.e.m., n = 3)

Fig. 8. Uptake and degradation of a misfolded protein by endocytosis.

a GFP1-10 is taken up into EEA1 positive early endosomes. Recombinant GFP1-10 was incubated with COS7 cells at 37 °C for the indicated time periods. Cells were stained with an EEA1 antibody (green) and GFP antibody (red). b Degradation of recombinant GFP1-10 by lysosomes. 293T cells were incubated with GFP1-10 and then washed. Cells were incubated in fresh medium in the presence of the indicated inhibitors. At the indicated time points, a fraction of the cells were lysed and analyzed by IB. The gel shown is a representative of 3 independent experiments. The graph shows the quantification of the experiment. PI, protease inhibitors. c Uptake of secreted α-synuclein by COS7 cells. EGFP-α-Syn-expressing cells were transfected with Flag-DNAJC5. Conditioned medium between 34 and 48 h post-transfection was harvested and concentrated by 5-fold. COS7 cells transfected with mCh-Rab5 were then incubated with conditioned medium for 2 h before live imaging by a confocal microscopy. The dashed line indicates the nucleus (N). The arrow indicates peri-nuclear vesicles where internalized α-Syn is localized. Scale bars, 5 µm

Fig. 9. A coordinated PQC mechanism allows cells to share a proteotoxic stress burden.

A model of protein quality control via intercellular transfer of misfolded proteins. PM, plasma membrane; ER, endoplasmic reticulum; LE, late endosome. Misfolded cytosolic proteins can be secreted by the following steps. In step 1, misfolded proteins (in red) are recruited to the ER surface by binding to a receptor (e.g., USP19) that has a chaperone activity. In step 2, misfolded proteins are transferred to DNAJC5, a HSC70 co-chaperone that is associated with the membrane of LEs and lysosomes, which are in close contact with the ER. In step 3, the complex of DNAJC5 and misfolded proteins are moved into the lumen of late endosomes probably via a protein translocation channel. In step 4, LEs contact and fuse with the PMs to release misfolded proteins. Once secreted, misfolded proteins can be taken up by another cell (Step 5) and degraded in the lysosomes (Step 6)