{"paper_id":"275ab50f-fd89-42d2-b9bf-4474e6d88006","body_text":"1 \n \nCHIP protects lysosomes from CLN4 mutant-induced membrane damages \n \nJuhyung Lee1, Jizhong Zou3, Wan Nur Atiqah Binti Mazli2, Natalie Chin1, Michal Jarnik4, Layla Saidi1, \nYue Xu1, John Replogle5, Michael Ward5, Juan Bonifacino4, Wei Zheng6, Ling Hao2, Yihong Ye1 \n \n1 Laboratory of Molecular Biology, National Institute of Diabetes, Digestive, and Kidney Diseases, \nNational Institutes of Health, Bethesda, MD 20892, USA \n2 Department of Chemistry and Biochemistry, the University of Maryland, College Park, MD 20742, USA \n3 iPSC Core, National Heart, Lung, and Blood Institute, National Institutes of Health, Bethesda, MD \n20892, USA \n4 Neurosciences and Cellular and Structural Biology Division, National Institute of Child Health and \nHuman Development, National Institutes of Health, Bethesda, MD 20892, USA \n5 Neurogenetics Branch, National Institute of Neurological Disorders and Stroke, National Institutes of \nHealth, Bethesda, MD 20892, USA \n6 Therapeutic Development Branch, National Center for Advancing Translational Sciences, National \nInstitutes of Health, Bethesda, MD 20850, USA \n \n \nCorresponding authors \nyihongy@mail.nih.gov \n \nShort title:  CHIP-mediated microautophagy preserves lysosome membrane integrity \n \nKey words: CHIP/STUB1, DNAJC5/CSPα, Ceroid Lipofuscinosis Neuronal/CLN4, lysosome membrane \ndamage, autophagy/microautophagy, ubiquitin, lysosome storage disease/LSD, neurodegenerative disease, \nDrosophila disease model \n \n  \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n2 \n \nUnderstanding how cells mitigate lysosomal damage is critical for unraveling pathogenic mechanisms \nof lysosome -related diseases.  Here we use organelle-specific proteomics in iPSC-derived neurons \n(i3Neuron) and an in vitro lysosome-damaging assay to demonstrate that lysosome damage, caused \nby the aggregation of Ceroid Lipofuscinosis Neuronal 4 ( CLN4)-linked DNAJC5 mutants on \nlysosomal membranes, serves as a critical pathogenic linchpin in CLN4 -associated \nneurodegeneration. Intriguingly, in non-neuronal cells, a ubiquitin-dependent microautophagy \nmechanism downregulates CLN4 aggregates to counteract  CLN4-associated lysotoxicity. Genome-\nwide CRISPR screens identify the ubiquitin ligase CHIP as a central microautophagy regulator that \nconfers ubiquitin-dependent lysosome protection. Importantly, CHIP’s lysosome protection function \nis transferrable, as ectopic CHIP  improves lysosomal function in CLN4 i3Neurons, and effectively \nalleviates lipofuscin accumulation and neurodegeneration in a Drosophila CLN4 disease model. Our \nstudy establishes CHIP-mediated microautophagy as a key organelle damage guardian that preserves \nlysosome integrity, offering new insights into therapeutic development for CLN4 and other lysosome-\nrelated neurodegenerative diseases.   \n \n \n  \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n3 \n \nCeroid lipofuscinosis neuronal (CLN) refers to a group of genetically inherited lysosomal storage diseases \n(LSD) featured with massive neurodegeneration 1, 2. These diseases are caused by mutations in a \ncollection of genes (e.g., CLN1, CLN2, CLN3), which lead to lysosomal dysfunction and the buildup of \ntoxic substances in the brain and other tissues 3, 4. A common pathologic hallmark of CLN diseases is the \naccumulation of autofluorescent lipopigments in the form of ceroid or lipofuscin in both neuronal and \nnon-neuronal cells. However, the disease symptoms are mostly manifested in the central nervous system \n5. CLN, as an uncurable disease, primarily affects children with common progressive symptoms including \nvision loss, seizures, motor deterioration, cognitive decline, and eventually, profound neurological \nimpairment 1, 3, 5. While most CLN cases are autosomal recessive, CLN4 (also known as Kufs disease) is \ncaused by dominant mutations in DNAJC5 6-10, a gene regulating protein folding and cellular homeostasis.  \nDNAJC5, also named Cysteine String Protein α (CSPα), encodes a member of the DnaJ/Hsp40 \nfamily of molecular chaperones known as DNAJC5. The Hsp40 family members are featured by the \npresence of a J-domain, which interacts with HSC70 to promote its ATPase cycle 11, 12. DNAJC5 also \ncontains a cysteine string domain, which undergoes palmitoylation to facilitate its membrane interactions \n13. In neurons, DNAJC5 is primarily localized to synaptic vesicles, but it is also found in other membrane \ncompartments such as Golgi-associated vesicles, plasma membrane (PM), and lysosomes 4. Owing to its \nco-chaperone activity 14, DNAJC5 plays critical roles in many cellular processes, including calcium \nregulation 15, membrane fusion and exocytosis 14, 16, 17, and unconventional secretion of misfolded \nproteins18-20 where it aids in the transport of misfolded cytosolic proteins out of the cell 21. DNAJC5 also \nhas a role in microautophagy, a form of autophagy that engulfs smaller portions of the cytoplasm \nincluding damaged proteins and organelles by late endosomes/lysosomes 22. The diverse roles of DNAJC5 \nunderscore a critical neuroprotective function linked to cellular homeostasis regulation 23. \nNumerous studies have characterized the biochemical properties of the CLN4-associated \nDNAJC5 mutants in vitro and in cells. It is known that these mutations abolish DNAJC5 palmitoylation \n24, causing the mutant proteins to be more prone to aggregation 25-27. Moreover, these mutations disrupt \nthe association of DNAJC5 with a Golgi-associated compartment 22, abolishing its activity in \nunconventional protein secretion 20, 22 while enhancing its association with the lysosomes 22, 27. However, \nit is unclear how these changes disrupt cell homeostasis to cause lipofuscinosis and neuronal cell death. \nIn this study, we identify lysosomal membrane damage as a key pathogenic mechanism in CLN4 \ndisease. We uncover a ubiquitin- and microautophagy-dependent lysosome protective mechanism that \nmasks CLN4 mutants’ lysotoxicity in non-neuronal cells, which is mediated by the ubiquitin ligase CHIP. \nAccordingly, CHIP overexpression restores lysosome function in CLN4 mutation-bearing i\n3Neurons and \nreduces lipofuscin accumulation and neurodegeneration in a Drosophila CLN4 disease model. These \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n4 \n \nfindings establish CHIP-mediated microautophagy as a promising target of therapeutic intervention in \ntreating lysosome damage-related neurodegenerative diseases. \n \nResults \nGeneration and characterization of human iPSC-derived CLN4 disease models \nTo model CLN4 disease, we used CRISPR-mediated gene editing to introduce L115R or L116Δ mutation \ninto the endogenous DNAJC5 locus in a well-characterized iPSC line with verified karyotype and \npluripotency (Extended Data Fig. 1a) 28. Genomic sequencing confirmed several clones heterozygous \n(HT) for the L115R or L116Δ allele and a homozygous (HM) L116Δ clone (Extended Data Fig. 1b-d). We \nselected clones with normal karyotypes and confirmed pluripotency (Extended Data Fig. 1e, f) and \ntransduced them with a lentiviral vector that enabled their differentiation into excitatory neurons \n(hereafter referred to as i3Neurons) via tetracycline-induced expression of neurogenin-2 (NGN2) \n(Extended Data Fig. 1g) 29. \nNext, we characterized these cells focusing on phenotypes relevant to the CLN4 disease. CLN4 \nDNAJC5 mutants are known to aggregate in an iron-sulfur cluster-, ATP-, and HSC70-dependent manner \n24. Immunoblotting confirmed that ~50% of DNAJC5 existed in an SDS-resistant, high molecular weight \n(HMW) form in cells heterozygous for L115R or L116Δ. In L116Δ HM i3Neurons, DNAJC5 was almost \nentirely present in the HMW form (Fig. 1a, Extended Data Fig. 2a). As expected, substantial amount of \nthe HMW DNAJC5 mutants were present in the NP40-insoluble fractions, consistent with their \naggregation properties. While we saw no difference in cell morphology or growth rate between wild-type \n(WT) and mutant iPSCs, significantly more L115R and L116Δ cells underwent apoptosis compared to \nWT cells after 16 days (d16) in differentiation, and L116Δ HM cells exhibited the most severe phenotype \n(Fig. 1b, c). \nSince CLN4 is an LSD, we characterized the lysosomes by LysoTracker staining, which \npreferentially labels acidic lysosomes. Fluorescence intensity measurement indicated a significant \nreduction of LysoTracker signal in mutant cells compared to the WT control, with L116Δ HM i\n3Neurons \nshowing the most pronounced decrease (Fig. 1d). Interestingly, while imaging live cells stained with \nLysoTracker confirmed the pH increase phenotype in d16 L116Δ HM i3Neurons, no significant change in \nlysosomal pH was observed in d12 L116Δ HM i3Neurons (Extended Data Fig. 2b). Staining with Magic \nRed, a fluorogenic dye indicative of the lysosomal protease Cathepsin B (CTSB) activity, showed reduced \nCTSB activity in mutant cells, mirroring the LysoTracker result (Extended Data Fig. 2c). These findings \nhighlight lysosomal dysfunction as a primary defect dose-dependently induced by CLN4 mutant alleles, \nlike via a yet-to-be defined gain-of-toxic activity suggested previously 22, 27, 30. \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n5 \n \nTo further investigate the molecular basis of CLN4 mutant-induced neurotoxicity, we focused on \nL116Δ HM i3Neurons due to their severe phenotype. Immunostaining for LAMP1, a lysosomal membrane \nprotein, in d16 i3Neurons showed increased LAMP1 signal in L116Δ HM i3Neurons compared to WT \ncontrol (Fig. 1e, right panels); L116Δ HM i3Neurons also contained many enlarged, irregularly shaped \nlysosomes. Interestingly, these abnormalities were not obvious at d12 (Fig. 1e, f). Transmission electron \nmicroscopy (TEM) revealed spherical lysosomes with dense core of 0.1–1 µm in diameter in WT \ni3Neurons, while in L116Δ HM cells, we frequently detected giant lysosomes (2–3 µm in diameter) \ncontaining undigested materials. These enlarged lysosomes often had rough or ruptured membranes (Fig. \n1g, Extended Data Fig. 2d). These observations, together with the reduced LysoTracker staining and \nCTSB activity, suggest that lysosomes might be damaged in CLN4 mutant i3Neurons. \nA hallmark of CLN diseases is the accumulation of autofluorescent storage material (AFSM). \nIndeed, confocal microscopy revealed many autofluorescent puncta in L116Δ HM i3Neurons but not WT \ncells (Fig. 1h). Most AFSM puncta were colocalized with LysoTracker signals, indicating lysosome as \ntheir source of origin. \nTo further elucidate the organelle homeostasis defects in L116Δ i3Neurons, we performed an \norganelle-based proteomics study. We isolated microsomes from WT and L116Δ HM i3Neurons using \ngradient centrifugation and analyzed DNAJC5-enriched membrane fractions by mass spectrometry (Fig. \n1i, Extended Data Fig. 2e, f). Based on the known subcellular localizations of DNAJC5 4, these DNAJC5-\npositive membranes should include lysosomes, synaptic vesicles, Golgi-associated vesicles, and the PM. \nMass spectrometry analysis identified 265 proteins up-regulated by at least 1.5-fold and 86 proteins \ndown-regulated similarly on membranes of L116Δ HM i3Neurons (Fig. 1j, Supplementary Table 1). Gene \nOntology (GO) pathway analysis showed that downregulated proteins were predominantly synaptic and \nPM proteins linked to synaptic transmission (Fig. 1k). By contrast, up-regulated proteins were mostly \ninvolved in pathways associated with lysosomal homeostasis including ubiquitin-dependent \nmicroautophagy (also named multivesicular body or MVB) and mTOR signaling (Fig. 1l, Extended Data \nFig. 2g). These results suggest that CLN4 mutants disrupt synaptic membrane proteome and lysosome \nhomeostasis. \n     \nCLN4-associated DNAJC5 aggregates damage lysosomal membranes \nSeveral lines of evidence suggested that the observed lysosomal defects in CLN4 i\n3Neurons likely \nresulted from membrane destabilization rather than the inhibition of the vacuolar ATPase (v-ATPase) \ncomplex. First, while co-immunoprecipitation readily detected an interaction between WT DNAJC5 and \nendogenous ATP6V1G2, a cytoplasmic subunit of the v-ATPase complex, as shown previously 31, the \ninteraction of ATP6V1G2 with CLN4 mutants was barely detectable (Extended data Fig. 3a), even though \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n6 \n \nthe CLN4 mutants are known to associate with the lysosomes better than WT DNAJC5 22. Thus, it \nseemed unlikely that CLN4 mutants could co-aggregate with the v-ATPase complex. Importantly, electron \nmicroscopy revealed many abnormal lysosomes with disrupted membranes in L116∆ HM i3Neurons \n(Extended data Fig. 2d), consistent with the lysosomal recruitment of ESCRT components, a phenotype \nknown as a cellular stress response to lysosome damages 32-36. \nWe postulated that lysosome-associated CLN4 aggregates might have a lysosome damaging \nactivity. To test this idea, we developed an in vitro assay using semi-permeabilized U2OS cells (Fig. 2a). \nWe labeled functional lysosomes with a LysoTracker dye and then permeabilized the PM in a fraction of \nthe cells with the pore forming toxin streptolysin O (SLO). After PM permeabilization and removal of the \ncytosol, cells were incubated with recombinant DNAJC5 together with a membrane impermeable dye (to \nlabel permeabilized cells), ATP, and concentrated cow live cytosol. We included cytosol in this \nexperiment because immunoblotting demonstrated that incubation with ATP and cytosol caused CLN4 \nmutants to form HMW aggregates (Fig. 2b), analogous to those observed in cells. Confocal microscopy \nshowed that addition of L115R or L116∆ to SLO-treated cells reduced LysoTracker signal specifically in \npermeabilized cells, whereas in buffer or WT DNAJC5-treated cells, permeabilized and unpermeabilized \ncells had similar LysoTracker signal (Fig 2c, d). These results further suggested that CLN4 DNAJC5 \nmutants can directly impair lysosomes. \n  To conclusively demonstrate the membrane destabilizing activity of the CLN4 mutants, we \nrepeated the lysosome-damaging experiment after loading LysoTracker green-stained cells with Alexa568-\nlabeled low molecular weight Dextran (Fig. 2e). When these cells were treated with the lysosome-\ndamaging compound LLOMe, both LysoTracker and Dextran signals were lost due to membrane damage \n(Extended Data Fig. 3b). On the other hand, when SLO-permeabilized cells were incubated in a buffer \nwithout ATP, v-ATPase was immediately inhibited but lysosomes remained intact. Consequently, we \nobserved a rapid reduction of the LysoTracker signal, but the Dextran signal remained unaffected \n(Extended Data Fig. 3c, middle panels). As expected, the addition of ATP maintained the v-ATPase \nactivity, allowing cells to retain both LysoTracker and Dextran signals (bottom panels). Thus, this \nsensitive assay is capable of distinguishing membrane damage from v-ATPase inhibition. We then \nincubated SLO-treated cells preloaded with Dextran and LysoTracker with cytosol, ATP, and CLN4 \nmutants, and observed reduced LysoTracker and Dextran signals in CLN4 mutant-treated cells compared \nto mock-treated cells (buffer) (Fig. 2f, g). These results confirmed that both L115R and L116∆ mutants \ncan destabilize lysosomal membrane to cause membrane leakage.  \n To see whether it is the aggregated CLN4 species that induces lysotoxicity, we first incubated \npurified L116∆ with cow liver cytosol and ATP, and then isolated the HMW species by centrifugation \nthrough a sucrose cushion. We incubated SLO-permeabilized cells with either untreated L116∆ (mostly \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n7 \n \nmonomer and dimer) or purified L116∆ HMW species in the absence of cytosol. Under this condition, \nonly HMW L116∆ could reduce LysoTracker signal (Fig. 2h, i), suggesting that L116∆ aggregates are \nresponsible for the lysotoxic activity, and the role of the cytosol is to promote CLN4 aggregation. Since \nlysosomal defects were only seen in mature CLN4 neurons but not in immature neurons (Fig. 1e, f) or \nU2OS cells overexpressing L116∆ (Extended Data Fig. 4a), it seems that non-neuronal cells have evolved \na mechanism to counteract CLN4 mutants’ lysotoxicity.  \n   \nUbiquitin-dependent microautophagy safeguards lysosomes from CLN4-induced membrane \ndamages in non-neuronal cells \nThe recruitment of the ESCRT components to lysosomes in L116∆ HM i3Neurons prompted us to \ninvestigate whether CLN4 mutants could activate ubiquitin-dependent microautophagy in non-neuronal \ncells. Indeed, immunostaining detected many ubiquitin-positive cytoplasmic puncta in U2OS cells \nexpressing mCitrine (Ci)-tagged L115R or L116∆; most puncta also contained DNAJC5, revealing their \nidentity as lysosomes (Fig. 3a, b). By contrast, few ubiquitin puncta were observed in cells expressing \nWT Ci-DNAJC5 or mCitrine. Moreover, the ubiquitin-binding component of the ESCRT0 complex \nhepatocyte growth factor-regulated tyrosine kinase substrate (HGS) was also recruited to CLN4-positive \npuncta (Fig. 3c, d). Denatured immunoprecipitation demonstrated that membrane-associated CLN4 \nmutants were more ubiquitinated than WT DNAJC5 in non-neuronal cells overexpressing these proteins \n(Extended Data Fig. 4b).  Together, these results suggest that CLN4 mutants induce ubiquitin build-up on \nlysosomes with some conjugates attached to the mutant proteins. Lysosome-associated ubiquitination then \nrecruits HGS and activates microautophagy. \n To see whether CLN4 mutants are microautophagy substrates, we expressed Keima-tagged \nDNAJC5 variants in HEK293T cells. Keima is a pH sensitive green fluorescence protein that displays \ndistinct fluorescent spectrum in different pH environments 37. Flow cytometry showed that CLN4 mutants \nhad increased lysosomal translocation activity compared to WT DNAJC5, but like WT DNAJC5, their \nlysosomal translocation was inhibited by an ATPase inactive VPS4 mutant (VPS4 E228Q) that blocks \nESCRT-dependent microautophagy, and by the ubiquitin E1 inhibitor TAK-243 (Fig.  3e, f). \nImmunoblotting confirmed that in cells treated with TAK-243, CLN4 mutants were accumulated in HMW \nform (Fig. 3g). Together, these results suggest that CLN4 mutants are targeted for degradation by \nubiquitin- and ESCRT-dependent microautophagy in non-neuronal cells.   \nIn addition to ESCRT-dependent microautophagy, mammalian cells also use ESCRT-independent \nmicroautophagy to degrade abnormal cytosolic proteins 38. To test whether ESCRT-independent \nmicroautophagy also targets CLN4 mutants, we treated cells expressing CLN4 mutants with GW4869. \nGW4869 is a neutral sphingomyelinase inhibitor that blocks ceramide synthesis, which is required for \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n8 \n \nESCRT-independent microautophagy 38. Like TAK-243, GW4869 treatment also reduced the lysosomal \ntranslocation of CLN4 mutants, causing these proteins to accumulate in cytoplasmic aggregates \n(Extended Data Fig. 4c, d). \nTo test whether microautophagy counteracts CLN4 mutants’ lysotoxicity in non-neuronal cells, \nwe treated U2OS cells transfected with the CLN4 mutants with TAK-243 to inhibit microautophagy. \nUnlike in untreated cells, we now detected a reduction in LysoTracker signal in L115R- or L116∆-\nexpressing cells, and to a lesser extent, also in WT DNAJC5-expressing cells compared to untransfected \ncells (Fig. 3h, i). As expected, cells co-expressing VPS4 E228Q with CLN4 mutants also had reduced \nLysoTracker signal (Fig. 3j, k), so were the DNAJC5 L116∆-expressing cells exposed to GW4869 \ntreatment (Extended Data Fig. 4e, f). Thus, both ESCRT-dependent and ESCRT-independent \nmicroautophagy can target CLN4 mutants, ameliorating lysosome dysfunction in non-neuronal cells \nexpressing these mutants.  \n \nCRISPR screens identify CHIP-mediated microautophagy as a lysosome guardian in non-neuronal \ncells \nTo identify the factors conferring ubiquitin-dependent lysosome protection, we conducted two CRISPR \nscreens using HEK293T cells stably expressing Keima-WT DNAJC5 and a Keima-tagged DNAJC5 \nmutant lacking the HSC70 binding J domain (∆J). These proteins undergo microautophagy similarly as \nCLN4 mutants 22, but are less toxic, and therefore, more suitable for long-term overexpression. We \ninfected Keima-DNAJC5 and Keima-DNAJC5 ∆J mutant cells with a lentiviral library targeting the \n~20,000 human genes each with six sgRNAs (Extended Data, Fig. 5a) 39. We then used FACS to isolate \ncells with increased neutral-to-acidic Keima ratio, indicative of impaired microautophagy. High-\nthroughput sequencing in two biological repeats for each screen identified sgRNAs and the corresponding \ntarget genes enriched in cells with defective microautophagy (Fig. 4a). Among them, 85 were deemed as \nhigh-confident hits because they were statistically significant in both WT and ∆J screens (Fig. 4b).  \nMoreover, many hits in this list are known microautophagy regulators such as components of the ESCRT \ncomplexes (Fig. 4c). We also identified ATP6V0C, an integral membrane component of the v-ATPase \ncomplex, whose inactivation is expected to deacidify the lysosomes. Intriguingly, a ubiquitin ligase \nnamed CHIP/STUB1 and its cognate conjugating enzyme UBE2N were also among the high confident \nlist, and small hairpin RNA (shRNA)-mediated knockdown validated them as positive regulators of \nDNAJC5 microautophagy (Fig. 4d). We decided to focus our studies on CHIP because it was annotated \nby BioGRID as a component of the DNAJC5 interactome (Fig. 4b) \n31. Indeed, co-immunoprecipitation \nconfirmed that CHIP could bind both WT DNAJC5 and the CLN4 mutants independent of the J domain \n(Fig. 4e). \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n9 \n \nCHIP (C-terminus of HSC70-Interacting Protein) is a U-box containing ubiquitin ligase that \ninteracts with HSP90 and HSC70 via several TRP motifs. Mutations in CHIP have been widely linked to \nneurodegenerative diseases, but the underlying mechanisms are unknown 40, 41. To better characterize the \nrole of CHIP in microautophagy, we used CRISPR to generate CHIP knockout (KO) cells.  As a positive \ncontrol, we also generated CHMP6 (a component of ESCRT-III) KO cells. Depletion of either CHMP6 or \nCHIP significantly reduced lysosomal translocation of Keima-DNAJC5 (Extended Data Fig. 5b). \nNoticeably, the lysosomal translocation of Keima-tagged CLN4 mutants was also diminished in CHIP KO \ncells (Extended Data Fig. 5c), further confirming its role in microautophagy. \nTo further characterize the role of CHIP in microautophagy, we used confocal microscopy to test \nwhether CHIP is required for lysosomal accumulation of ubiquitin and HGS in U2OS cells expressing \nCLN4 mutants using L116∆ as a representative. Immunostaining showed that unlike WT cells, the \nlysosomal accumulation of ubiquitin and HGS in L116∆-expressing CHIP KO cells were not obvious, but \nre-expressing CHIP rescued this phenotype (Fig. 4f-i). Furthermore, denatured immunoprecipitations \nshowed that CHIP depletion reduced the ubiquitination of both WT and the L116∆ DNAJC5 mutant \n(Extended data Figure 5d, e). Thus, CHIP promotes DNAJC5 ubiquitination, particularly those associated \nwith CLN4 mutants, resulting in HGS recruitment to lysosomes. \nTo see whether CHIP-mediated ubiquitination counteracts CLN4-mutants’ lysotoxicity, we \ntransfected WT or CHIP KO U2OS cells with Ci-L116∆ and then stained these cells with LysoTracker. \nLike in E1 inhibitor-treated cells, Ci-L116∆ expression in CHIP KO cells also destabilized lysosomes to \nreduce LysoTracker staining. This phenotype was rescued when WT CHIP was co-transfected (Fig. 4j, k). \nTogether, these results suggest that in non-neuronal cells, CHIP stimulates ubiquitination on CLN4-\ncontaining lysosomes, activating microautophagy to counteract the membrane damaging activity of the \nCLN4 mutants.  \n \nCHIP rescues lysosomal defects and cell death in L116∆ i3Neurons \nWhat accounts for the differential sensitivity to CLN4-mediated lysotoxicity between immature and \nmature i3Neurons? To address this question, we used immunoblotting to compare the expression of CHIP \nand DNAJC5 during the differentiation of WT and L116∆ HM i3Neurons. Our data suggested that \nDNAJC5 expression was increased after d8 in both WT and L116∆ HM cells, and as expected, L116∆ \nwas mostly in the HMW form (Fig. 5a, Extended Data Fig. 6a). By contrast, CHIP protein solubilized by \nthe detergent NP40 was largely unchanged throughout the differentiation and between WT and L116∆ \ncells. However, we observed an increase of CHIP in NP40-insoluble fractions, starting at d8 and more \npronounced in L116∆ cells. The accumulation of CHIP in NP40-insoluble fractions might reflect a change \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n10 \n \nin its activity because this phenotype correlated with the accumulation of HMW L116∆ aggregates (Fig. \n5a, Extended Data Fig. 6a). Moreover, transducing i3Neurons with Keima-DNAJC5-expressing \nlentiviruses revealed a modest reduction of DNAJC5 lysosomal translocation after d12, suggesting \nreduced microautophagy (Extended Data Fig. 6b, c). Collectively, these results suggested a reduction in \nCHIP activity during late stage of L116∆ i3Neuron differentiation, which diminishes microautophagy and \ncauses L116∆ aggregate to accumulate. These changes explain why significant lysosome defects and cell \ndeath were only noticed in L116∆ HM cells after d12 in differentiation.  \nTo further test the role of microautophagy in neuroprotection, we asked whether ectopically \nexpressing CHIP in i3Neuron could alleviate lysosomal damage caused by endogenous L116∆. To avoid \noverexpression artifact, we used a neuron-specific synapsin promoter activated after d8 in differentiation \nto drive CHIP expression in i3Neurons. Immunoblotting revealed a modest, but reproducible reduction of \ntotal L116∆ aggregate in L116∆ HM cells expressing CHIP compared to cells with no ectopic CHIP \n(Extended Data Fig. 6d, e).  This phenotype was more pronounced when affinity-purified lysosomes from \ncontrol and CHIP-expressing L116∆ HM i3Neurons were analyzed (Extended Data Fig. 6f).  \nAs anticipated, CHIP-expressing L116∆ i3Neurons had increased LysoTracker signal in both \nsoma and neurites (Fig. 5b, c). Consistent with improved acidity, DQ-BSA-based lysosomal activity assay \nshowed partially restored hydrolase activity in L116∆ HM i3Neurons by WT but not the ∆Ubox CHIP \nmutant lacking the U-box (Fig. 5d, e). Importantly, CHIP knockdown in L116∆-overexpressing i3Neurons \nenhanced L116∆-induced cell death (Fig. 5f, g), whereas ectopic expression of CHIP in L116∆ HM \ni3Neurons reduced differentiation-associated cell death (Fig. 5h). Collectively, these results demonstrated \nthat ectopic CHIP improves lysosome homeostasis and mitigates cell death in L116∆ HM i3Neurons. \n \nCHIP rescues lipofuscin accumulation and neurodegeneration in a Drosophila CLN4 model \nTo see whether CHIP could rescue CLN4 mutant-associated disease phenotypes in vivo, we tested its \nfunction in a recently established Drosophila CLN4 disease model 27. We first generated a transgenic line \ncarrying Keima-Csp1, the Drosophila homolog of DNAJC5 (dDNAJC5). We also generated transgenic \nflies bearing Drosophila WT CHIP (dCHIP) or dCHIP lacking the TRP or U-box-coding sequences. \nThese transgenes were all placed downstream of the upstream activating sequence (UAS) to achieve \ntissue specific gene expression. We first compared the relative microautophagy activities in various fly \ntissues using a Gal4 line driven by the heat shock promoter. Measuring the ratio between acidic (red) and \nneutral (green) Keima-CSPα signal showed high lysosomal translocation in intestine epithelial cells \nfollowed by ventral nerve chord, and the brain, while fat body had the lowest microautophagy activity \n(Fig. 6a, Extended Data Fig. 7a). The microautophagy activity of CSPα was modest in larval \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n11 \n \nphotoreceptor cells, but it could be enhanced when dCHIP was co-expressed (Fig. 6b). Thus, the CHIP’s \nfunction in microautophagy is conserved in flies. \nWe next crossed the UAS-CHIP lines to flies expressing human CLN4 L116∆ in larval \nphotoreceptor cells via the GMR promoter. High level expression of human CLN4 L116∆ caused \nlipofuscin accumulation, enlarged lysosomes decorated with ubiquitin, and neuronal cell death 27, \nrecapitulating phenotypes seen in CLN4 patient 5. Indeed, when imaginal eye discs from GMR>L116∆ \nthird instar larva were stained with ubiquitin and DNAJC5 antibodies, photoreceptor cells expressing \nCLN4-L116∆ had more ubiquitin-positive puncta than those in WT flies (W1118) (Fig. 6c, d), and like in \nmammalian cells, many ubiquitin positive puncta were also positive for L116∆ (Extended data Fig. 7b). \nConfocal microscopy also detected many autofluorescent puncta resembling ceroid or lipofuscin. \nInterestingly, these structures could be stained by an amyloid specific dye, suggesting that they contain \nprotein aggregates (Fig. 6e, Extended data Fig. 7c). As expected, acridine orange staining showed high \nlevel of cell death in L116∆-positive but not in WT eye discs (Fig. 6g).  \nCo-expression of dCHIP in photoreceptor cells significantly reduced L116∆ protein level, while \nincreasing lysosome-associated ubiquitin puncta (Fig. 6c, d), suggesting that dCHIP promotes lysosome-\nassociated ubiquitination to down-regulate L116∆. Neither dCHIP ∆U-box or dCHIP ∆TPR mutant could \ndownregulate L116∆, although only dCHIP ∆U-box failed to stimulate L116∆-associated ubiquitination \n(Fig. 6c, d). Thus, additional factor(s) acting through the CHIP TPR domain are also involved in L116∆ \ndownregulation. As expected, dCHIP but not the ∆TRP or ∆U-box CHIP mutant rescued lipofuscin \naccumulation (Fig. 6e, f) and mitigated L116∆-associated neurodegeneration (Fig. 6g).  \n As expected from CHIP’s ability to downregulate L116∆ and reduce lipofuscin accumulation and \ncell death, WT dCHIP but not the U-box or TRP-deleted CHIP mutant suppressed the L116∆-induced \nrough eye phenotype (Fig. 6h, Extended Data Fig. 7d). The rescue effect was comparable to that caused \nby HSC70 knockdown 27, which presumably prevents iron-sulfur cluster-mediated L116∆ aggregation \n(Extended Data Fig. 7d) 24. Conversely, knockdown of CHIP or the microautophagy regulator Tsg101 \nexacerbated the rough eye phenotype in L116∆-expressing flies (Extended Data Fig. 7e). Notably, CHIP \noverexpression did not affect the rough eye phenotype induced by polyQ aggregates (Extended Data Fig. \n7f), suggesting that its activity is specific towards lysotoxic proteins. Collectively, these data demonstrates \nthat dCHIP enhances L116∆-associated ubiquitination to downregulate L116∆, and therefore, rescues \nlipofuscin accumulation and neurodegeneration in flies.  \n Lastly, we tested whether Tsg101-dependent microautophagy is required for CHIP’s \nlysoprotective activity. To this end, we generated GMR> hDNAJC5 L116∆; dCHIP flies with or without \nTsg101 shRNA. Knockdown of Tsg101 significantly reversed the eye improvement by dCHIP, although \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n12 \n \non its own, Tsg101 knockdown did not change the eye morphology (Fig. 7a). Immunostaining showed \nthat dCHIP-mediated down-regulation of L116∆ was partially abolished when Tsg101 was knocked down \n(Fig. 7b, c). Together, these results suggested that dCHIP downregulates L116∆ at least in part via \nTsg101-mediated microautophagy, which mitigates L116∆-associated neurodegeneration. \n    \nDiscussion \nLSDs are increasingly recognized as neurodegenerative disorders with lysosome deficiency as a common \npathological hallmark. Accordingly, most CLN diseases are associated with recessive mutations in genes \nencoding lysosomal enzymes or factors essential for transporting these enzymes to lysosomes 4. In this \nregard, CLN4 is unique as it is linked to autosomal dominant mutations in DNAJC5 presumed to confer a \ngain-of-toxic function, although the underlying mechanism has been unclear.   \nIn i3Neurons bearing the L115R or L116Δ mutation in DNAJC5, we observed a dose-dependent \nlysotoxicity from the encoded mutant proteins, causing characteristic lysosomal abnormalities including \nenlarged lysosomes with damaged or ruptured membranes, diminished hydrolase activity, and AFSM \naccumulation. Our in vitro studies show that CLN4 mutants can oligomerize in an ATP- and cytosol-\ndependent manner, consistent with the reported role of cytosolic iron-sulfur clusters in CLN4 aggregation \n24. Intriguingly, CLN4 aggregates generated in vitro can directly damage lysosomes from the cytosolic \nside in permeabilized cells. Since several studies reported that endocytosed protein aggregates can also \ndamage lysosomal membranes but from the luminal side 42, 43, lysosome membranes may be intrinsically \nvulnerable to protein aggregate-induced membrane instability. \nNotably, it was reported that polyQ-containing protein aggregates can disrupt and deform the \nendoplasmic reticulum membranes 44. How cytosolic protein aggregates compromise distinct membrane \ncompartments remains to be elucidated, but these processes likely involve specific membrane adaptors \nthat bring protein aggregates to the target membrane. Once in proximity to the membranes, protein \naggregates may destabilize the lipid bilayer directly, form a membrane-embedded pore through \noligomerization, or damage membranes via reactive oxygen species (ROS) generation. Lysosomal \nmembrane rupture is known to trigger cascading cellular damage, activating inflammation, and releasing \nlysosomal enzymes, which ultimately leads to cell death \n45, 46. \nCells deploy several protective mechanisms to maintain lysosomal integrity, which include \nESCRT-mediated membrane repair and lysophagy 36, 47. Our findings identify ubiquitin-mediated \nmicroautophagy as an additional organelle damage control mechanism that safeguards lysosome \nhomeostasis (Fig. 7d). Microautophagy can target toxic cytosolic protein aggregates for degradation, \ntherefore, protects lysosomes from aggregate-induced membrane damage. Additionally, microautophagy \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n13 \n \nmay facilitate membrane repair, since ESCRT-mediated MVB formation is known to remodel lyososomal \nmembranes, which can aid in the repair of damaged lysosomal membranes 32-35. \nOur findings indicate that CHIP-dependent ubiquitination at CLN4-positive lysosomes serves as a \nmolecular linchpin in ubiquitin-dependent lysosomal protection. In non-neuronal cells, CHIP-mediated \nmicroautophagy directs aggregates and possibly damaged membranes toward the degradation pathway, \nwhich explains the lack of significant lysosomal defect in fibroblast cells bearing CLN4 disease mutations \n30. However, during neuronal differentiation, this protective function appears impaired when CLN4 \nmutants are present, as CHIP becomes sequestered within insoluble aggregates, which correlates with \nincreased CLN4 aggregation and lysosome destabilization during i3Neurons differentiation. Whether \nCHIP activity is similarly downregulated during neuronal differentiation in animals is unclear. However, \nsince CHIP expression is relatively low in both excitatory and inhibitory neurons according to the Human \nProtein Atlas (proteinatlas.org) 48, we presumed that diminished CHIP expression/function in neurons may \ncause the vulnerability to CLN4-induced lysotoxicity. Consistent with this notion, ectopic CHIP \nexpression can partially restore lysosomal integrity and cell viability in CLN4 i3Neurons and mitigate \nCLN4 pathology in a Drosophila model of CLN4 disease. The rescue effect of CHIP in Drosophila is \nbetter than in i3Neurons, probably because CHIP was co-expressed with L116Δ in flies using the same \npromoter, while in i3Neurons, CHIP expression driven by the Synapsin promoter lags behind L116Δ \naggregation.  \nCHIP has long been recognized as a critical component of the cellular protein quality control \nsystem, since it can interact with cytosolic chaperones HSC70 and HSP90. Consequently, CHIP can \npreferentially ubiquitinate misfolded or aggregation-prone proteins and target them for degradation 49. \nCHIP’s vital role in quality control and homeostasis regulation is underscored by numerous genetic links \nto neurodegenerative diseases such as spinocerebellar ataxia 49. Intriguingly, recent studies suggested that \nCHIP’s activity is regulated by a monomer-to-dimer switch: dimeric CHIP mediates chaperone-assisted \nturnover via the proteasome, while monomeric CHIP promotes the turnover of membrane-bound proteins, \nas observed in C. elegans 50. Whether CHIP’s role in lysosome homeostasis regulation requires the \nmonomeric or dimeric form and how CHIP-associated chaperones aid in this process remain to be tested. \nThe relevance of our findings to broader neurodegenerative disease mechanisms is compelling. \nLysosomal membrane damage is commonly reported in neurodegenerative diseases such as Parkinson’s \nand Alzheimer’s diseases, where internalized protein aggregates are major contributors to membrane \ndestabilization and lysosome leakage 42, 43, 51. These observations suggest that CHIP’s role in \nmicroautophagy may also exploited to develop therapeutics for lysosome damage-associated disorders. \nSupporting this idea, CHIP overexpression was shown to promote the degradation of an aggregated form \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n14 \n \nof α-synuclein in human H4 neuroglioma cells 52, though whether this process involves microautophagy is \nunclear.  \nIn summary, our study provides novel insights into the pathological role of CLN4-associated \nmutant DNAJC5 aggregates, establishing lysosomal damage as a central feature of the disease. Our work \nreveals a critical role for CHIP-dependent ubiquitination and microautophagy in lysosomal damage \ncontrol and suggests that modulation of this lysosome protection mechanism may hold therapeutic \npromise for ameliorating lysosomal dysfunction in CLN4 and other aggregate-associated \nneurodegenerative diseases. \n \nMaterials and Methods \nCell lines, plasmids, antibodies, and other reagents \nU-2 OS (U2OS) and 293T cells were purchased from ATCC (catalog no. HTB-96 and CRL-3216). Cells \nwere grown in DMEM (Corning, catalog no. 10-013-CV) supplemented with 10% FBS (Corning, catalog \nno. 35-011-CV) and 100 µg/mL penicillin-streptomycin (Gibco, catalog no. 15140-122), and maintained \nat 37 °C, 5% CO\n2, and 95% humidity. Lipofectamine 2000 (Invitrogen, catalog no. 11668027) and \nTransIT-293 Transfection Reagent (Mirus Bio, catalog no. MIR 2700) were used for transfection in U2OS \nand 293T, respectively, according to the manufacturer’s protocol. For siRNA transfection, Lipofectamine \nRNAiMAX (Invitrogen, catalog no. 13778075) was used according to the manufacturer’s protocol. All \nplasmids used in this study were constructed by standard molecular biology methods and sequenced \nconfirmed. All plasmids will be deposited to Addgene. Plasmids, antibodies and reagents used in this \nstudy are listed in Supplementary Table 2. \n \nLentivirus production \nTo prepare lentiviruses, HEK293FT cells (Thermo Fisher Scientific, catalog no. R70007) were transfected \nwith pVSV-G (Addgene #8454), psPAX2 (Addgene #12260), and lentiviral expression vectors (e.g., \npLenti CMV Hygro or FSW) in a 1:1.5:2 mass ratio. After 24 h, the media were replaced with fresh media \nand incubated for 48 h to collect viruses. The viral supernatants were harvested, centrifuged at 300 g for \n10 minutes, filtered through a 0.45 µm PVDF syringe filter unit (Sigma-Aldrich, catalog no. \nSLHVM33RS), and concentrated ~100x using Lentivirus Precipitation Solution (ALSTEM, catalog no. \nVC100) according to the manufacturer’s protocol. The resulting virus pellets were resuspended in DPBS \n(Gibco, catalog no. 14190-144) and Virus Protection Medium (ALSTEM, catalog no. VF050), aliquoted, \nflash-frozen in liquid nitrogen, and stored at -80 °C. \n \nCRISPR-CAS9 gene editing \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n15 \n \nCHIP and CHMP6 knock-out in 293T and CHIP knock-out in U2OS cell lines were carried out by \ninfection with lentiviruses carrying Cas9 and the corresponding sgRNA in pLenti CRISPRv2 (Addgene \n#52961), as described 39. The following gRNAs, designed using the ATUM website \n(https://www.atum.bio/eCommerce/cas9/input), were used: \nsg-hCHIP forward: 5’-caccgTGTATTACACCAACCGGGCC-3’, sg-hCHIP reverse: 5’-\naaacGGCCCGGTTGGTGTAATACAc-3’; sg-hCHMP6 forward: 5’- \ncaccgGCTCAAGAAGAAGCGATACC-3’, sg-hCHMP6 reverse: 5’- \naaacGGTATCGCTTCTTCTTGAGCc-3’. Infected cells were cultured for three days followed by \npuromycin (gibco, catalog no. A1113803) selection (0.4 µg/mL for 293T and 1 µg/mL for U2OS cells) \nuntil the parallel non-infected control cells were all killed. After one passage in medium without \npuromycin, cells were infinitely diluted and seeded at 1 cell per well into 96 well plates either manually \nor using a BD Fusion cell sorter. Knock-out clones were verified by immunoblotting.      \n To generate L115R and L116∆ knockin iPSC cells, 0.8 million HT727D iPSCs 28 were transfected \nwith premixed 60 pmol HiFiCas9 V3 (IDT, catalog no. 1081061), 60 pmol dCas9 V3 (IDT, catalog no. \n1081067), 200 pmol sgRNA (Synthego), and 200 pmol single strand oligodeoxynucleotides \n(ssODN) (IDT) using Nucleofector 4D with buffer P3 and program CA-137 (Lonza, catalog no. V4XP-\n3024) and then plated onto one well of rhLaminin-521 (gibco, catalog no. A29249) coated 6-well plate \nwith StemFlex (gibco, catalog no. A33493-01), RevitaCell (gibco, catalog no. A2644501) and HDR \nenhancer V2 (IDT, catalog no. 10007921). Transfected cells were cultured in 32 °C incubator for 3 days \nbefore moving to 37 °C incubator, and fresh medium was changed daily after transfection. Single-cell \nsubcloning was done using manual serial dilution to 1 cell/96-well Matrigel coated plate with StemFlex \nand CloneR2 (Stem Cell Technologies, catalog no. 100-0691). 10 days later, single-cell clones were \npicked and confirmed by genomic PCR, Sanger sequencing, and ICE analysis (Synthego). Besides L115R \nand L116∆ mutation knock-in clones, biallelic DNAJC5 knockout clones and unedited isogenic control \nclones were also saved.  The sgRNA target, single-stranded oligodeoxynucleotides (ssODNs), and PCR \nprimers used for p.L115R (c.344T>G) and p.L116∆ (c.346-348∆) mutation knock-in in DNAJC5 are \nlisted below: \np.L115R (c.344T>G), sgRNA target: AGCAGTAGCAGCACGTGAGG, ssODN: \naacaccaccttcttctccccccagGCCCTGTTTGTCTTCTGCGGCCGCCTGACGTGCTGCTACTGCTGCTGCT\nGTCTGTGCTGCTGCTTCAACTGCTGCT (Note: this sequence contains a L116 CTC to CTG silent \nmutation to avoid recutting after L115R KI). \np.L116del (c.346-348del), sgRNA target: AGCAGTAGCAGCACGTGAGG, ssODN: \naacaccaccttcttctccccccagGCCCTGTTTGTCTTCTGCGGCCTGACGTGCTGCTACTGCTGCTGCTGTC\nTGTGCTGCTGCTTCAACTGCTGCT (contains L115 CTC>CTG silent mutation) \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n16 \n \nPCR forward primer: TGCTTTTCTTTAAGCTGCGGG \nPCR reverse primer: TACCTCAGGGTCCACGTTCA \n \nCell lysis, membrane fractionation, and immunoblotting \nTo lyse cells, ~2-3 million 293T cells were washed with phosphate-buffered saline (PBS) two times and \nthen treated with 300 µL the NP40 lysis buffer containing 0.5% Nonidet P-40 (Sigma-Aldrich, catalog no. \n56741-50ML-F], 50 mM Tris-HCl, pH 7.4, 150 mM NaCl, 2 mM MgCl2, 1 mM EDTA, 1 mM TCEP \n(Thermo Fisher Scientific, catalog no. 20490), and a protease inhibitor cocktail. Cells were incubated on \nice for 15 min with occasional mixing. The lysates were cleared by centrifugation (17,000 g, 10 min at 4 \n°C). The supernatant (the NP40-soluble fraction) was mixed with 4x Laemmli buffer (BioRad, catalog no. \n1610747) and then heated at 95 °C for 5 min. The remaining pellets were washed once with the NP40 \nlysis buffer and then centrifuged at 17,000 g for 10min at 4 °C. The pellets were resuspended in 100 µL \nPBS by gentle pipetting and then mixed with equal volume of pre-heated 2x Laemmli buffer (BioRad). \nThe samples were immediately heated at 95 °C for 20 min to obtain the NP40-insoluble fraction.  \nCytosol-membrane fractionation was performed as described previously 22. Cells harvested in \nPBS were pelleted by centrifugation at 500 g for 5 min at 4°C. Cell pellets were treated with a \npermeabilization buffer (PB) containing 0.025% digitonin (Sigma-Aldrich, catalog no. D141), 230 mM \npotassium acetate (Sigma-Aldrich, catalog no.P1190), 10 mM sodium acetate (Sigma-Aldrich, catalog no. \nS2889), 50 mM HEPES, pH7.3, 5 mM MgCl2, 1 mM EGTA (Sigma-Aldrich, catalog no. 324626), 1 mM \nTCEP (Thermo Fisher Scientific), and a protease inhibitor cocktail on ice for 5 min. Plasma membrane \npermeabilization was confirmed by trypan blue (Thermo Fisher Scientific, catalog no. 15250061) \nstaining. Cells were spun at 17,000 g for 5 min. The supernatant was saved as the cytosol fraction. The \nresulting membrane pellets were washed with 1x PB buffer followed by centrifugation and then further \nlysed by a CHAPS lysis buffer (1% CHAPS [Sigma-Aldrich, catalog no. 10810118001], 50 mM HEPES, \npH 7.4, 100 mM NaCl, 1 mM TCEP, and protease inhibitors). The lysates were cleared by centrifugation \nat 17,000 g for 5 min and the supernatant fractions were collected as membrane fraction.  \nFor immunoblotting, samples were loaded onto 4-12% Bis-Tris gels (Invitrogen) followed by \nSDS-PAGE with MES SDS Running Buffer (Invitrogen, catalog no. B0002). The proteins were \ntransferred to Nitrocellular membranes (0.45 µm, BioRad, catalog no. 1620115), which were stained \nusing the Ponceau S. solution (Sigma-Aldrich, catalog no. P7170). Membranes were then blocked with \nPBS containing 5% non-fat milk, washed with PBS three times, and then incubated with primary \nantibodies in 5 % BSA (Sigma-Aldrich, catalog no. A9418) in PBS supplemented with sodium azide \n(0.03%) for overnight in 4 °C. The membranes were washed three times with PBS and further incubated \neither with goat anti-Mouse DyLight™ 680, goat anti-Rabbit DyLight™ 800 (Thermo Fisher Scientific), \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n17 \n \nor HRP-labeled goat anti-Mouse or Rabbit IgG (Sigma-Aldrich) at room temperature for 1 h. After \nthorough wash with PBS for three times, fluorescence or chemiluminescence signal was detected using a \nChemiDoc MP scanner (Biorad) and quantified using the ImageLab software (BioRad). \n \nImmunoprecipitation \nFor co-immunoprecipitation assay, 293T cells transfected with FLAG-DNAJC5 were fractionated into \ncytosol and membrane fractions as described above. ANTI-FLAG® M2 Affinity Gel (Sigma Aldrich, \ncatalog no. A2220) was added to the fractions and incubated for 1 h at 4 °C. A portion of the extracts were \nsaved as input before adding the beads. The beads were washed by PBS three times and the proteins \nbound were eluted by 1x Laemmli buffer and heating at 95 °C for 5 min before SDS-PAGE and \nimmunoblotting analysis.  \nDenatured immunoprecipitation was performed to detect ubiquitination on DNAJC5. To this end, \nthe cells were transfected or infected with lentiviruses to achieve the expression of tagged-DNAJC5 \nvariants.  For lentivirus-infected cells, drug selection was conducted to make cells stably expressing these \nDNAJC5 variants. Cells were lysed in a buffer (150 µL) containing 0.5 % NP40, Tris-HCl 7.4 50 mM, \n150 mM NaCl, 2 mM MgCl\n2, 1 mM DTT, 2 mM NEM and protease inhibitor.  After centrifugation, \ncleared cell extracts were adjusted with SDS and DTT to contain 1% SDS and 5 mM DTT. Cell lysates \nwere heated at 95 °C for 5 min. Lysate were diluted with the NP40 lysis buffer by 10-fold and then \ncentrifuged.  The cleared lysates were incubated with FLAG beads or DNAJC5 antibody-containing \nbeads to purify DNAJC5. Bound proteins were washed and then eluted with 1x Laemmli buffer by \nheating at 95 °C for 5 min. Eluted proteins were analyzed by SDS-PAGE and immunoblotting with anti-\nubiquitin.   \n \nHuman iPSC culture and neuronal differentiation \nHuman WT and CLN4 mutation-bearing iPSC cells were transduced with a doxycyclin-inducible NGN2 \nas described previously 53. These cells were cultured on Matrigel-coated dishes in StemFlex medium \n(Thermo Fisher Scientific) according to the manufacturer’s instruction. Briefly, cells were passaged when \nthey reach 70% confluent using StemPro Accutase (gibco, catalog no. A11105-01) Cells were seeded into \nMatrigel (Corning, catalog no. CLS354277)-coated dishes with a density of 10,000 cells per cm2 in the \npresence of 1 µM Chroman 1 (MedChemExpress, catalog no. HY-15392). The chroman 1 was removed \non the following day and medium change was performed every day. If necessary, EZ-Lift Stem Cell \nPassaging Reagent (Sigma-Aldrich, catalog no. SCM139) was used to eliminate spontaneously \ndifferentiated cells from the culture according to the manufacturer’s protocol. Differentiation of iPSCs to \ni\n3Neurons was performed as previously described 53 with minor modifications. On day 0, iPSCs were \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n18 \n \nseeded to Matrigel-coated dishes at 80,000 cells per cm2 using the KnockOut DMEM/F-12 (Thermo \nFisher Scientific, catalog no. 12660012)-based induction medium containing 1x N-2 supplement (Thermo \nFisher Scientific, catalog no. 17502048), 1x GlutaMax (gibco, catalog no. 35050061), 1x non-essential \namino acids (NEAAs; gibco, catalog no. 11140050), 1 µM Chroman 1 and 2 µg/mL doxycycline (Sigma-\nAldrich, catalog no. D5207). Medium change was performed on Day 1, 2 and 3 using induction medium \nwithout Chroman1. On day 3, 1 µM 5-Fluore-2'-deoxyuridine (Sigma-Aldrich, catalog no. F0503) and 1 \nµM Uridine (Sigma-Aldrich, catalog no. U3003) were added for overnight treatment to eliminate \nundifferentiated cells. On day 3, ibidi 8-well glass chamber (for imaging) or corning dishes (for \nbiochemical assays) were coated with 100 µg/mL Poly-L-Ornithine (Sigma-Aldrich, catalog no. P3655) \nfor overnight at 37 °C. Next day (day 4), the wells were washed with DPBS twice and further coated with \n1 µg/mL of rhLaminin-521 (gibco) for 1 h at 37 °C. At the day 4, the differentiated neurons were \ndissociated using Accutase, counted and plated with a density of 100,000 cells per cm2 (for imaging) or \n200,000 cells per cm2 (for biochemical assays) using neuronal culture medium containing 1:1 mix of \nBrainPhys medium (STEMCELL, catalog no. 5790) and KnockOut DMEM/F-12 supplemented with 1x \nN21-Max (R&D Systems, catalog no. AR008), 10 ng/mL BDNF (Peprotech, catalog no. 450-02), 10 \nng/mL GDNF (Peprotech, catalog no. 450-10), 10 ng/mL NT-3 (Peprotech, catalog no. 450-03), 0.5 \nµg/mL rhLaminin (gibco) and 2 µg/mL doxycycline (Sigma-Aldrich). At day 6, the medium was changed \nwith fresh neuronal culture medium without KnockOut DMEM/F-12 and Doxycyclin. Since then, half of \nthe medium was removed every 3 days, and an equal volume of fresh medium was added for \nmaintenance. If necessary, iPSC or day 4 i3Neurons were frozen using CryoStor® CS10 medium \n(STEMCELL, catalog no. 100-1061) according to the manufacturer’s protocol and kept in liquid nitrogen \ntank for storage.  \n \nTransmission Electron Microscopy \nCells were fixed in 2.5 % Glutaraldehyde, 2 % Formaldehyde, in 0.1 M Cacodylate/2 mM Calcium \nchloride pH 7.4 (CacCl) for 15 min at room temperature followed by 45 min incubation on ice. Coverslips \nwere washed 3 times for 5 min each with CacCl, post-fixed for 30 min in 0.5 % Osmium tetroxide/0.5% \npotassium ferrocyanide in the same buffer, washed and treated with 1 % tannic acid for 30 min.  Cells \nwere washed 2 times with CacCl, 2 times with 50 mM sodium acetate (pH 5.2) and stained overnight with \n2% Uranyl acetate. After washing 2 times with the acetate buffer and 2 times with water, the samples \nwere dehydrated through a series of increasing concentration of ethanol (50%, 75%, 90%, 3 times in \n100% anhydrous) and embedded in EMBed\n812 epoxy resin (EMS, catalog no. 14120)). After resin \npolymerization, the coverslip was removed by hydrofluoric acid. Sample blocks were cut out and \nmounted on a holder. Ultrathin sections (80 nm thick) were cut parallel to the plane of the coverslip and \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n19 \n \nmounted on formvar/carbon coated EM grids. Sections were stained with lead citrate and imaged in FEI \nTecnai 20 transmission electron microscope operated at 120 kV . Images were recorded on AMT \nNanosprint 12 wide field CCD camera.  \n \nCell staining, confocal imaging, and data analysis \nmKeima fluorescence and autofluorescence (AFSM) were detected in live cells following previously \nreported protocols \n22. For immunostaining, i3Neurons or U2OS cells cultured on µ-Slide 8 Well high \nGlass Bottom (ibidi, catalog no. 80807) were fixed in PBS containing 4 % paraformaldehyde (Thermo \nFisher Scientific, catalog no. 28908) for 10 min at room temperature. Cells were then washed with PBS \ntwice and permeabilized with a PBS-based staining solution containing 0.2 % saponin (Sigma-Aldrich, \ncatalog no. 47036) and 10 % FBS for 10 min at room temperature. For methanol fixation, cold methanol \nwas added to PBS-washed cells. Cells were incubated at -20 °C for 10 min and washed with PBS three \ntimes and blocked with 5% FBS in PBS for 10 min at room temperature. Cells were stained by primary \nantibodies diluted in the staining solution overnight at 4 °C and washed with PBS three times. Alexa \nFluor® labeled secondary antibodies (Thermo Fisher Scientific) were diluted in the staining solution and \nadded to cells for 1 h at room temperature. Where indicated, 1 μg/mL DAPI (Sigma-Aldrich, catalog no. \nD9542) was included in the staining solution to label nuclei. The slides were washed three times with \nPBS. The stained slides were imaged by either LSM780 laser scanning confocal microscopy (Zeiss) or \nCSU-W1 SoRa spinning disk super-resolution microscopy (Nikon).  \nFor LysoTracker and Magic Red staining in live i\n3Neuron, LysoTracker™ Red DND-99 (1: \n5,000; Invitrogen, catalog no. L7528) or Magic Red (1: 250; Antibodies Inc., catalog no. 938) was added \nto cells and stained for 30min. For the DQ-BSA assay, i3Neurons were incubated with DQ red BSA \n(Invitrogen, catalog no. D12051) at 40 µg/mL for 6 h. Where indicated, Hoechst33442 (1:10,000; \nInvitrogen, H3570) was added to the media for labeling nuclei. Random-fielded images were collected as \nz-section to cover entire cell volume using the Nikon CSU-W1 SoRa confocal microscope equipped with \n60x TIRF objective (NA=1.6) and a heating and CO2-perfused chamber.  \nFor data quantifications, ImageJ software was used to merge all the z-sections, generating \nmaximum projected images. Fluorescence intensity was measured by using ROIs covering the soma of \ni3Neurons. A portion of soma with no signal in each image was selected for background subtract. The total \nadjusted intensity was calculated by multiply individual mean intensity with the cell area. To measure the \nlysosome volume or the area of immunostained signals, images were subject to programed image \nthresholding. Automate single particle analysis was used to identify lysosomes and measure their size.  \nFor cell viability assay in i\n3Neuron, the LIVE/DEAD™ Viability/Cytotoxicity Kit (Invitrogen, \ncatalog no. R37601) was used according to the manufacturer’s protocol. The number of live neurons \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n20 \n \n(green) and dead neurons in randomly selected fields were determined by using a threshold and watershed \nfunction in ImageJ. The cell death index (normalized red) was calculated as the number of red cells \ndivided by the number of green plus red cells (Total cells). For 3-D visualization, we used Imaris (Oxford \ninstruments). Some images were processed using Nikon NIS-element plugin Denoise.ai for background \nreduction. \nFor plate reader-based fluorescence measurement, i\n3Neurons cultured in 96 well plates were \nmeasured by GloMax® Explorer Multimode Microplate Reader (Promega). Wells containing no staining \ndyes were used for background subtraction. For area or intensity analyses, untransfected cells next to \ntransfected cells were used as internal controls to measure the relative fold-change.  \n \nMembrane fractionation and lysosome isolation \nLysosome Enrichment Kit (Thermo Fisher Scientific, catalog no. 89839) was used for fractionating \norganelles from i\n3Neurons according to the manufacturer’s protocol. Briefly, two p100 dishes containing \ni3Neurons at day16 per condition were washed with DPBS twice, and neurons were harvested by \ncentrifugation at 1,000 g for 10 min. After removal of DPBS, cells were resuspended in 500 µL Buffer A \nwith protease inhibitors, vortexed for 5 seconds, and incubated on ice for 5 min. Plasma membranes were \nbroken by 60 strokes using a 2 mL glass Dounce homogenizer with a tight pestle. 500 µL buffer B with \nprotease inhibitors was added. The extracts were cleared by centrifugation at 500 g for 10 min at 4 °C. \nThe resulting supernatants with 15% of OptiPrep™ cell separation media were divided into three equal \nportions and each portion was overlayed on a discontinued OptiPrep™ media gradient (17, 20, 23, 27, \nand 30 %) in ultra-clear centrifuge tubes (Beckman Coulter, catalog no. 344062) following the provided \nprotocol.  After ultracentrifugation at 145,000 g (37,600 rpm) for 2 h at 4 °C, 0.5 mL fractions were \ncarefully collected from the top of gradients and gently mixed with 1 mL PBS. The fractions were \ncentrifuged at 18,000 g for 30 min at 4 °C. The supernatant fractions were removed. PBS (0.2 mL) was \nadded to each fraction to resuspend the membrane pellet. The same fractions were combined and then \ncentrifuged at 18,000 g for 30 min at 4 °C again. The supernatant fractions are removed and stored at -80 \n°C until immunoblotting and mass spectrometry analyses. \nTo immune purify lysosomes, lyso-IP method was used as previously described with \nmodifications \n54. Briefly, i3Neurons at day6 were infected by lentivirus to express TMEM192-HA with \nCHIP-Myc or empty vector. At day16, neurons were washed once with Tris-buffered saline (TBS) and \nharvested by centrifugation at 450 g for 5 min. TBS is removed and cell pellets were resuspended with \ncold fresh TBS with protease inhibitors and lysed by 20 strokes using a 2 mL glass Dounce homogenizer \nwith a tight pestle. The lysates were transferred and mixed with 8 mM CaCl2 and vortexed. After \ncentrifugation at 1,150 g for 3 min in 4 °C, supernatants were carefully transferred into new tubes and \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n21 \n \nused for immunoprecipitation using Anti-HA Magnetic Beads (Thermo Fisher Scientific, catalog no. \n88836) pre-equilibrated with the same buffer. A small fraction was saved before adding the beads for \ninput. The beads were incubated on a rocker for 4 h in 4 °C and washed with TBS with 8 mM CaCl2 and \nprotease inhibitors six times using DynaMag™-2 Magnet (Invitrogen, catalog no. 12321D ). Lysosomes \nwere eluted and lysed by heating at 65 °C for 10 min in the presence of 1x Laemmli buffer. \n \nOrganelle proteomics sample preparation \nMembrane pellets prepared as described above were reconstituted in ice-cold lysis buffer (500 mM NaCl, \n0.1% SDS, 1% Triton, 5 mM TCEP) and sonicated using QSonica sonicator (QSonica, catalog no. \nQ800R) for 10 min in an ice-cold water bath with alternating 40 s-on and 20 s-off cycles. Protein \nconcentration was determined using a Bio-Rad Detergent Compatible (DC) protein assay. Protein \nreduction, alkylation, and digestion were conducted with an automated SP3 method in an automated \nKingFisher sample preparation system as described previously \n55. Briefly, samples were reduced with 5 \nmM TCEP at room temperature for 40 min, alkylated with 15 mM IAA for 40 min at 37 °C, and quenched \nwith 15 mM DTT for 15 min at 37 °C. Acetonitrile (ACN) was added to 80% percentage in volume in \neach sample, and 10 µL of SP3 beads (Cytiva) was added in the sample followed by 10 min of incubation \nto induce protein binding. Beads were washed three times with 95% ACN, two times with 70% ethanol, \nand then released in 100 µL of 50 mM ammonium bicarbonate buffer. Proteins were digested with \nTrypsin/Lys-C mix (1:25, enzyme: protein) for 1 h at 47 °C. Beads were washed again in 100 µL of LC-\nMS grade water. Both supernatants of peptides elution were combined. Residual detergents from digested \npeptides were tested and removed using the ContamSPOT assay prior to LC-MS analysis, as described \npreviously \n56. Peptide samples were dried down under SpeedVac and kept in -30 °C.  \n \nLC-MS/MS analysis \nDried peptide samples were reconstituted in 0.1% formic acid (FA), 2% ACN in LC-MS grade water and \nanalyzed on a Dionex Ultimate 3000 RSLCnano system coupled with a Thermo Fisher Scientific \nScientific Q-Exactive HF-X Orbitrap mass spectrometer. Peptides were separated on an Easy-Spray \nPepMap RSLC C18 column (2 µm, 100 Å, 75 µm × 50 cm) with a 180 min LC gradient and a flow rate of \n0.25 µL/min. Mobile phase A was 0.1% FA in water, and mobile phase B was 0.1% FA in ACN. Samples \nwere analyzed in a staggered data-independent acquisition mode with 75 sequential scans and an m/z 8.0 \nisolation window. MS1 scanned from m/z 400 to 1,000 with a resolving power of 60K, an automatic gain \ncontrol (AGC) target of 1E6, and a maximum injection time (maxIT) of 60 ms. MS2 resolving power was \n15K, AGC target was 2E5, and maxIT was 40 ms. Normalized collision energy was 30%.  \n \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n22 \n \nProteomics data analysis  \nProteomics raw data was analyzed in the Spectronaut software (v18.1, Biognosys). Proteins and peptides \nwere identified with a 1% false discovery rate cut off using the Swiss-Prot Homo sapiens database and \nneuron-specific contaminant library \n57. Missed cleavages were set up to 2 and variable modifications up to \n3.  Cysteine carbamidomethylation was set as fixed modifications, and methionine oxidation, and protein \nN-terminal acetylation was set as variable modifications. Precursor intensities below 1000 were removed \nfrom the Spectronaut protein report file. Statistical analysis was conducted using a two-tailed Student’s t-\ntest. \n \nIn vitro lysosome-damaging assay \nTo test the lysosome damaging activity of CLN4 mutants, 25,000 cells of U2OS cells were seeded in an \n8-well Ibidi cell chamber two days before the experiment. We then labeled the cells with Lysotracker\nTM \nRed DND-99 (Thermo Fisher Scientific) at 1 µM for 30 min at 37 °C. Cells were washed four times with \nice-cold phosphate-buffered buffer saline containing 2 mM MgCl2 (PBS-Mg) and then treated with the \nsame buffer (300 µL) containing 100 units of Streptolysin O (Sigma-Aldrich, catalog no. SAE0089) on \nice for 10 min.  Cells were then washed three times with the PBS-Mg buffer and the incubated with 200 \nµL reagent mixture containing 150 µL PBS-Mg, 50 µL cow liver cytosol or buffer, 2 mM ATP, 1 mM \nDTT, 0.5 µM LysoTracker Red, a cell impermeable dye and a protease inhibitor cocktail. WT or CLN4 \nmutant proteins were added at 2 µM. Cells were incubated at 37 °C for 30 min and then imaged by a \nNikon CSU-W1 SoRa super-resolution confocal microscope. \n To measure lysosome by Dextran leakage, U2OS cells were first loaded with Dextran Alexa \nFluor™ 568 10,000 MW (100 µg/mL; Invitrogen, catalog no. D22912) at 37 °C for 4 h, and then \nincubated in a Dextran-free medium for 3 h. Cells were then labeled with a LysoTrackerTM Green dye \n(Invitrogen, catalog no. L7526) at 1 µM for 30 min before treated with Steptolysin O.  \n \nRecombinant protein purification \nTo purify recombinant human DNAJC5 (hDNAJC5) proteins, hDNAJC5 WT, L115R and L116Δ proteins \ncontaining a TEV cleavage site between a GST tag and the protein were expressed in BL21(DE3) \ncompetent E.coli (NEB, catalog no. C2527) by adding 0.5 mM IPTG to 2 liters LB cultures at 16 °C for \novernight. Cells were harvested by centrifugation at 6,000 g for 20 min at 4 °C, and the pellets were \nresuspended in 35 mL PBS containing 2 mM TCEP and protease inhibitors. Cells were lysed by \nsonication (30 % amplitude) on ice with the setting of 10 sec ON / 20 sec OFF for 10 min of total ON \ntime. The lysates were cleared by centrifugation at 40,000 g (~18,000 rpm) for 30 min in 4 °C and the \nsupernatants were incubated with 2 mL (1 mL beads bed volume) of Glutathione Sepharose™ 4 Fast \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n23 \n \nFlow (Cytiva, catalog no. 17513202) pre-equilibrated with PBS on a shaker for overnight. The beads were \ntransferred to glass chromatography columns (Bio-Rad) and washed three times by 15 mL of PBS with 2 \nmM TCEP. To elute proteins from the beads, 1 mL PBS containing 400 unit of biotin-tagged TEV \nprotease (Sigma-Aldrich, catalog no. SAE0118) was added to the beads and incubated at 4 °C overnight. \nElutes containing proteins were collected by gravity flow and further incubated with 100 µL of 1:1 \nmixture of Glutathione Sepharose (Cytiva) and High-Capacity Streptavidin Agarose (Thermo Fisher \nScientific, catalog no. 20357) for 1 h at 4 °C to remove the TEV protease. The protein purity was \nconfirmed by SDS-PAGE and Coomassie staining. The proteins were aliquoted, flash-frozen in liquid \nnitrogen and stored at -80 °C. \n \nGenome-wide CRISPR/CAS9 knockout screen \nThe GeckoV2 library was purchased from Addgene (1-000-000-048) and amplified according to the \nonline protocol from Feng Zhang’s lab 39. The complexity of the sgRNA library was verified by high-\nthroughput sequencing by the NIDDK Genomic Core. Lentiviruses containing the sgRNA library and \nCas9 were generated and used for transduction via spinfection. Briefly, 60 million of 293T stably \nexpressing mKeima-tagged human DNAJC5 WT or ΔJ mutant supplemented with 8 μg/mL polybrene \n(Sigma-Aldrich, catalog no. TR-0003) were seeded in two 12-well plates at a density of 3 million cells \nper well. Concentrated GeCKO v2 lentivirus (library A) was added to each well at a multiplicity of \ninfection (MOI) at 0.3. Cells were then spun at 1,000 g at room temperature for 2 h followed by \nincubation at 37 °C in a humidified incubator for 1 h. After the medium was removed, fresh growth \nmedium was added, and cells were incubated for 48 h before the start of the selection for lentiviral \nintegration using puromycin (0.3 μg/mL). The transduced cells were subcultured in medium \nsupplemented with puromycin every 2 days for a total of 8 days and 80 million cells were maintained for \neach passage. After puromycin treatment, cells were recovered in a medium lacking puromycin for 24 h \nbefore cell sorting. In total, 100 million cells were sorted into mKeima neutral (3% of total cells) or \nmKeima acidic (97% of total cells) cell populations by a FACSAria™ Fusion Flow Cytometer (BD \nBioscience). Genomic DNA was extracted from each cell population using a QIAGEN Blood Maxi kit \n(for mKeima acidic cells; Qiagen, catalog no. 51192) or QIAGEN Blood Midi kit (for mKeima neutral \ncells; Qiagen, catalog no. 51183) according to the manufacturer’s protocols. sgRNAs sequences were \namplified from genomic DNA samples using the Herculase II Fusion DNA Polymerase (Agilent, catalog \nno. 600675) in two PCR steps as follows. In the first PCR, the genomic region containing sgRNAs were \namplified from ~250 μg and ~20 μg of total DNA from acidic and neutral samples respectively, using the \nfollowing primers: Forward: 5′-AATGGACTATCATATGCTTACCGTAACTTGAAAGTATTTCG-3′; \nReverse: 5′- TTCAAAAAAGCACCGACTCGGTGCCACTTTTTCAAGTTGATAAC-3′. In total, 32 and \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n24 \n \nthree PCR reactions for acidic and neutral samples respectively were performed in parallel, with 8 μg \ngenomic DNA in each reaction using Herculase II Fusion DNA Polymerase (Agilent) for 18 cycles and \nthe resulting PCR products were combined. In the second step PCR, 5 μL of first PCR product was used \nin a 100 μL reaction volume and 12 PCR cycles were used. The primers used for the second PCR include \nstagger sequences of variable lengths and a 6 bp barcode for multiplexing of different biological samples. \nThe following PCR primers were used in the second step: Forward: 5′-\nAATGATACGGCGACCACCGAGATCTACACTCTTTCCCTACACGACGCTCTTCCGATCT-(1-7bp \nvariable length sequences)-(6bp barcode)-TTGTGGAAAGGACGAAACACCG -3’; Reverse: 5′-\nCAAGCAGAAGACGGCATACGAGAT-(6bp barcode)-\nGTGACTGGAGTTCAGACGTGTGCTCTTCCGATCTTCAAGTTGATAACGGACTAGCC-3′. The \nfinal PCR products were gel extracted, quantified, and sequenced using a NovaSeq sequencer (Illumina) \nby the NHLBI DNA Sequencing and Genomics Core. sgRNA sequences were obtained per sample by \nextracting 20 bps followed by the index sequence of “TTGTGGAAAGGACGAAACACCG” on the de-\nmultiplexed FASTQ files from Illumina’s NGS sequencer using the Cutadapt software, version \n2.8 (https://doi.org/10.14806/ej.17.1.200). FASTQC, version 0.11.9 was used to assess the sequencing \nquality (http://www.bioinformatics.babraham.ac.uk/projects/fastqc/). MAGeCK, version 0.5.9, was used \nto quantify and to identify differentially expressed sgRNAs. MAGeCK count command was run on the \nGECKO library A to quantify. Differentially expressed sgRNAs with statistical significance were \ndetermined by running the MAGeCK test command in unpaired mode. Genes were ranked based on the \nnumber of unique sgRNA enriched in the mKeima neutral population versus the mKeima acidic \npopulation. Data was derived from two biological repeats for each screen. \n \nF\nlow cytometry \nmKeima-expressing cells were dissociated with fresh DMEM medium by gentle pipetting and passed \nthrough a cell strainer cap filter (Thermo Fisher Scientific, catalog no. 08–771-23). Flow cytometry was \nperformed on an LSRII Fortessa analyzer (Becton Dickinson). The gate for acidic (Ex586/Em620)/neutral \n(Ex440/Em620) intensities of individual cells (>10,000 cells) were determined manually using Bafilomycin \nA1 (100 nM for 2–4 h)-treated samples as a reference. Bafilomycin A1 treatment converts ~99 % of cell \npopulation to the neutral gate. Flow data were analyzed using FlowJo 10.9 software (FlowJo LLC). \n \nGeneration of DNAJC5 antibody \nGST-tagged hDNAJC5 purified from E. coli was used for injection to rabbits by a commercial service \n(LAMPIRE). The GST-DNAJC5 proteins were crosslinked to CNBr-activated sepharose 4 fast flows for \nantibody purification. In detail, ~0.85 g of CNBr Activated Sepharose™ 4 Fast Flow (Cytiva, catalog no. \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n25 \n \n17-0981-01) in 30 mL of ice cold 1 mM HCl was loaded onto a chromatography column (BioRad). The \nHCl solution is removed by gravity flow and beads were thoroughly washed with 20 mL of ice cold 1 mM \nHCl three times and 20 mL of coupling buffer (0.1 M NaHCO\n3, 0.5 M NaCl, pH 8.3). GST-hDNAJC5 \nprotein (6 mg) in 15 ml coupling buffer was incubated with the beads for 2 h at room temperature with \ngentle shaking. The unbound protein was removed by allowing the solution to drain. After washing the \nbeads with 10 mL coupling solution, 20 mL of 0.1 M Tris 8.0 was incubated with the beads for 3 h with \nshaking at room temperature to block the uncoupled active sites. The buffer was removed and beads were \nwashed 0.1 M glycine pH 3.0, and then with PBS.  \nTo purify antibodies, rabbit anti-DNAJC serum (15 mL) was applied to the column and incubated \nfor overnight with shaking in 4 °C. The antibody was eluted by applying 14 mL of 0.1 M Glycine pH 3.0. \nThe received fractions were combined and applied to Vivaspin 20 centrifugal concentrator (3,000 \nMWCO; Vivaproducts, catalog no. VS2091) and centrifuged at 4,200 g for 30 min in 4 °C. The resulting \nsolution was dialyzed by PBS with 200 mM NaCl using Slide-A-Lyzer™ Dialysis Cassettes (10K \nMWCO; Thermo Fisher Scientific, catalog no. 66382) for overnight at 4 °C. The antibody was aliquoted, \nflash-frozen in liquid nitrogen and stored at -80 °C. \n \nDrosophila experiments \nFly strains bearing shRNA-expressing cassettes downstream of UAS sequences targeting CHIP/STUB1 \n(33938#), Tsg101 (35710#), Hsp70 (34836#) are from the Bloomington Drosophila Stock Center \n(BDSC). The flies expressing human DNAJC5 L116Δ were described previously \n27. The strain expressing \nGAL4 under the heat shock promoter was also purchased from BDSC (2077#). All cultures were \nmaintained on BDSC cornmeal food (Lab-express) in 25 °C incubators equipped with a programmable \nLED light. \n Transgenic flies carrying UAS-dCHIP WT, UAS-dCHIP \n∆-Ubox, or UAS-dCHIP-∆TPR \ntransgenes were made by injecting the corresponding plasmids (a generous gift from Dr. Tang, B. and \nDuan, R. of Central South University, China) into the R9752 line by Rainbowgene. Stable transgenic lines \nwere crossed to the Sp/Cyo; Tm2/Tm6 double balancer strain to determine which chromosome harbors the \ntransgene. For most experiments, balanced lines were back-crossed to W\n1118 to obtain homozygous lines \nwithout balancer chromosome. UAS-Keima-dDNAJC5 (Csp1) flies were made using a similar strategy. \nFor imaging fly eyes, 10-20 adult flies were fixed in PBS containing 4% formaldehyde for 1 h, \nrinsed with PBS. The flies were then dehydrated by soaking sequentially in 30%, 50%, 70%, 90%, and \n100% ethanol. The flies were air-dried and dried flies were mounted in an Ibidi imaging chamber by \nVaseline. Fly eyes were scanned by a Nikon CSU-W1 confocal microscope using Ex\n488/Em520 nm. Shown \nis the maximum projection view of the scanned Z-section images. Alternatively, flies were fixed on a petri \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n26 \n \ndish with Vaseline and then imaged directly with a dissecting microscope equipped with a AmScope \nMU1803 digital camera.  \nDissecting and immunostaining were performed as previously described 58. Briefly, imaginal eye \ndiscs were dissected from third instar lava in PBS, fixed with 4% paraformaldehyde in PBS for 20 min at \nroom temperature. Fixed discs were washed with PBS four times, permeabilized with a PBS-based \nstaining buffer containing 0.2% Saponin and 5% FBS in PBS. Discs were then incubated with primary \nantibodies in the staining buffer at 4 °C overnight. The primary antibodies used are FK2 (1:250) and \nDNAJC5 (1:500). Discs were washed three times with PBS and then stained with corresponding \nsecondary antibody labeled with either Alexa Fluor 488 or 568.   \nTo detect AFSM and protein aggregates in photoreceptor cells, imaginal eye discs were dissected \nfrom third instar larva and incubated in PBS with AmyTracker 680 (Ebba biotech) at 2 µg/mL at room \ntemperature for 30 min and then mounted in Sang M3 medium (Sigma-Aldrich, catalog no. S3652) with \n5% FBS and 20 % glycerol.  AFSM was detected using Ex405 nm/Em480 nm by a Zeiss LSM780 \nconfocal microscope. To detect apoptotic cells, dissected eye discs were stained with acridine orange \n(Invitrogen, catalog no. A1301) in PBS at a concentration of 1 µg/mL for 5 min. Discs were washed once \nwith PBS and then mounted for imaging immediately.  \n \nStatistics and reproducibility \nAll statistical analyses were conducted with GraphPad Prism v10. Statistical methods and the number of \ncells or Drosophila eye discs (N) are indicated in figure legends or shown in figures as individual data \npoint. Biological repeats (n) are specified in figure legends. For statistically significant comparisons with \nP-value larger than 0.0001, we provided the exact P-values in figures. We did not predetermine sample \nsize. The sample sizes are consistent with similar studies reported in the literature. No biological repeat \nwas excluded from the analyses. For individual cell analyses, a few data points outside of 1.5 times the \ninterquartile range were considered as outliers and were excluded. All experiments were repeated at least \ntwice with individual data point labeled in figures. The immunoblotting and flow cytometry data, \nwhenever shown, are representative of similar results from at least two independent biological replicates \nunless specified in figure legends. For imaging analyses, cells in randomly selected field were analyzed. \nThe researchers were not blinded.  For EM study, lysosomes were identified based on their typical \nmorphology and the presence of intraluminal contents. All identified lysosomes are included in the \nanalysis. Data shown are representative of two biological repeats. For CRISPR screens, two biological \nrepeats were performed with each cell line and the data were pooled to identify statistically significant \nhits. For organelle-based proteomic study, membranes collected from 3 biological repeats were processed \nsimultaneously. For Drosophila experiments, sex was not considered as a variable, and larva were \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n27 \n \ndissected without pre-determination of the sex. All iPSC cell lines were derived from a single male \nindividual. Figures were prepared using ImageJ 1.54f, Adobe Photoshop v25.12.1, and Adobe Illustrator \n28.7.4. \n \nData Availability  \nCRISPR screen raw data will be deposited to GEO. All proteomics MS raw files have been deposited to \nthe ProteomeXchange Consortium and are available through the MassIVE repository (Identifier: \nPXD058788). Other data in support of the conclusions is available in either main figures, extended data \nfigures, or supplementary tables.   \n \nReferences \n1. Mole, S.E. & Cotman, S.L. Genetics of the neuronal ceroid lipofuscinoses (Batten disease). \nBiochim Biophys Acta 1852, 2237-2241 (2015). \n2. Chabrol, B., Caillaud, C. & Minassian, B. Neuronal ceroid lipofuscinoses. Handb Clin Neurol 113, \n1701-1706 (2013). \n3. Johnson, T.B. et al. Therapeutic landscape for Batten disease: current treatments and future \nprospects. Nat Rev Neurol 15, 161-178 (2019). \n4. Lee, J., Xu, Y . & Ye, Y . Safeguarding lysosomal homeostasis by DNAJC5/CSPalpha-mediated \nunconventional protein secretion and endosomal microautophagy. Frontiers in Cell and \nDevelopmental Biology (2022). \n5. Anderson, G.W., Goebel, H.H. & Simonati, A. Human pathology in NCL. Biochim Biophys Acta \n1832, 1807-1826 (2013). \n6. Noskova, L. et al. Mutations in DNAJC5, encoding cysteine-string protein alpha, cause autosomal-\ndominant adult-onset neuronal ceroid lipofuscinosis. Am J Hum Genet 89, 241-252 (2011). \n7. Burneo, J.G. et al. Adult-onset neuronal ceroid lipofuscinosis (Kufs disease) with autosomal \ndominant inheritance in Alabama. Epilepsia 44, 841-846 (2003). \n8. Naseri, N., Sharma, M. & Velinov, M. Autosomal dominant neuronal ceroid lipofuscinosis: Clinical \nfeatures and molecular basis. Clin Genet 99, 111-118 (2021). \n9. Benitez, B.A. et al. Exome-sequencing confirms DNAJC5 mutations as cause of adult neuronal \nceroid-lipofuscinosis. PLoS One 6, e26741 (2011). \n10. Velinov, M. et al. Mutations in the gene DNAJC5 cause autosomal dominant Kufs disease in a \nproportion of cases: study of the Parry family and 8 other families. PLoS One 7, e29729 (2012). \n11. Braun, J.E., Wilbanks, S.M. & Scheller, R.H. The cysteine string secretory vesicle protein activates \nHsc70 ATPase. J Biol Chem 271, 25989-25993 (1996). \n12. Chamberlain, L.H. & Burgoyne, R.D. The molecular chaperone function of the secretory vesicle \ncysteine string proteins. J Biol Chem 272, 31420-31426 (1997). \n13. Burgoyne, R.D. & Morgan, A. Cysteine string protein (CSP) and its role in preventing \nneurodegeneration. Semin Cell Dev Biol 40, 153-159 (2015). \n14. Bronk, P . et al. Drosophila Hsc70-4 is critical for neurotransmitter exocytosis in vivo. Neuron 30, \n475-488 (2001). \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n28 \n \n15. Dawson-Scully, K., Bronk, P ., Atwood, H.L. & Zinsmaier, K.E. Cysteine-string protein increases the \ncalcium sensitivity of neurotransmitter exocytosis in Drosophila. J Neurosci 20, 6039-6047 \n(2000). \n16. Zhang, H., Kelley, W.L., Chamberlain, L.H., Burgoyne, R.D. & Lang, J. Mutational analysis of \ncysteine-string protein function in insulin exocytosis. J Cell Sci 112 ( Pt 9), 1345-1351 (1999). \n17. Sharma, M., Burre, J. & Sudhof, T.C. CSPalpha promotes SNARE-complex assembly by \nchaperoning SNAP-25 during synaptic activity. Nat Cell Biol 13, 30-39 (2011). \n18. Fontaine, S.N. et al. DnaJ/Hsc70 chaperone complexes control the extracellular release of \nneurodegenerative-associated proteins. EMBO J 35, 1537-1549 (2016). \n19. Xu ,  Y. et al. DNAJC5 facilitates USP19-dependent unconventional secretion of misfolded cytosolic \nproteins. Cell Discov 4, 11 (2018). \n20. Wu, S. et al. Unconventional secretion of alpha-synuclein mediated by palmitoylated DNAJC5 \noligomers. Elife 12, e85837 (2023). \n21. Lee, J.G., Takahama, S., Zhang, G., Tomarev, S.I. & Ye, Y . Unconventional secretion of misfolded \nproteins promotes adaptation to proteasome dysfunction in mammalian cells. Nat Cell Biol 18, \n765-776 (2016). \n22. Lee, J. et al. Abnormal triaging of misfolded proteins by adult neuronal ceroid lipofuscinosis-\nassociated DNAJC5/CSPalpha mutants causes lipofuscin accumulation. Autophagy 19, 204-223 \n(2023). \n23. Schmitz, F. et al. CSPalpha-deficiency causes massive and rapid photoreceptor degeneration. \nProc Natl Acad Sci U S A 103, 2926-2931 (2006). \n24. Naseri, N.N. et al. Aggregation of mutant cysteine string protein-alpha via Fe-S cluster binding is \nmitigated by iron chelators. Nat Struct Mol Biol 27, 192-201 (2020). \n25. Diez-Ardanuy, C., Greaves, J., Munro, K.R., Tomkinson, N.C. & Chamberlain, L.H. A cluster of \npalmitoylated cysteines are essential for aggregation of cysteine-string protein mutants that \ncause neuronal ceroid lipofuscinosis. Sci Rep 7, 10 (2017). \n26. Henderson, M.X. et al. Neuronal ceroid lipofuscinosis with DNAJC5/CSPalpha mutation has PPT1 \npathology and exhibit aberrant protein palmitoylation. Acta Neuropathol 131, 621-637 (2016). \n27. Imler, E. et al. A Drosophila model of neuronal ceroid lipofuscinosis CLN4 reveals a hypermorphic \ngain of function mechanism. Elife 8 (2019). \n28. Xu, X. et al. Four induced pluripotent stem cell lines (TRNDi021-C, TRNDi023-D, TRNDi024-D and \nTRNDi025-A) generated from fibroblasts of four healthy individuals. Stem Cell Res 49, 102011 \n(2020). \n29. Zhang, Y . et al. Rapid single-step induction of functional neurons from human pluripotent stem \ncells. Neuron 78, 785-798 (2013). \n30. Benitez, B.A. & Sands, M.S. Primary fibroblasts from CSPalpha mutation carriers recapitulate \nhallmarks of the adult onset neuronal ceroid lipofuscinosis. Sci Rep 7, 6332 (2017). \n31. Oughtred, R. et al. The BioGRID database: A comprehensive biomedical resource of curated \nprotein, genetic, and chemical interactions. Protein Sci 30, 187-200 (2021). \n32. Radulovic, M. et al. ESCRT-mediated lysosome repair precedes lysophagy and promotes cell \nsurvival. EMBO J 37 (2018). \n33. Skowyra, M.L., Schlesinger, P .H., Naismith, T.V. & Hanson, P .I. Triggered recruitment of ESCRT \nmachinery promotes endolysosomal repair. Science 360 (2018). \n34. Ogura, M. et al. Microautophagy regulated by STK38 and GABARAPs is essential to repair \nlysosomes and prevent aging. EMBO Rep 24, e57300 (2023). \n35. Kuchitsu, Y . & Taguchi, T. Lysosomal microautophagy: an emerging dimension in mammalian \nautophagy. Trends Cell Biol 34, 606-616 (2024). \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n29 \n \n36. Meyer, H. & Kravic, B. The Endo-Lysosomal Damage Response. Annu Rev Biochem 93, 367-387 \n(2024). \n37. Violot, S., Carpentier, P ., Blanchoin, L. & Bourgeois, D. Reverse pH-dependence of chromophore \nprotonation explains the large Stokes shift of the red fluorescent protein mKeima. J Am Chem \nSoc 131, 10356-10357 (2009). \n38. Trajkovic, K. et al. Ceramide triggers budding of exosome vesicles into multivesicular endosomes. \nScience 319, 1244-1247 (2008). \n39. Sanjana, N.E., Shalem, O. & Zhang, F. Improved vectors and genome-wide libraries for CRISPR \nscreening. Nat Methods 11, 783-784 (2014). \n40. Shi, Y . et al. Identification of CHIP as a novel causative gene for autosomal recessive cerebellar \nataxia. PLoS One 8, e81884 (2013). \n41. Shi, C.H. et al. Ataxia and hypogonadism caused by the loss of ubiquitin ligase activity of the U \nbox protein CHIP . Hum Mol Genet 23, 1013-1024 (2014). \n42. Rose, K. et al. Tau fibrils induce nanoscale membrane damage and nucleate cytosolic tau at \nlysosomes. Proc Natl Acad Sci U S A 121, e2315690121 (2024). \n43. Riera-Tur, I. et al. Amyloid-like aggregating proteins cause lysosomal defects in neurons via gain-\nof-function toxicity. Life Sci Alliance 5 (2022). \n44. Bauerlein, F.J.B. et al. In Situ Architecture and Cellular Interactions of PolyQ Inclusions. Cell 171, \n179-187 e110 (2017). \n45. Seoane, P .I. et al. The NLRP3-inflammasome as a sensor of organelle dysfunction. J Cell Biol 219 \n(2020). \n46. Aits, S. & Jaattela, M. Lysosomal cell death at a glance. J Cell Sci 126, 1905-1912 (2013). \n47. Yang, H. & Tan, J.X. Lysosomal quality control: molecular mechanisms and therapeutic \nimplications. Trends Cell Biol 33, 749-764 (2023). \n48. Uhlen, M. et al. Proteomics. Tissue-based map of the human proteome. Science 347, 1260419 \n(2015). \n49. Wang, T. et al. The E3 ubiquitin ligase CHIP in normal cell function and in disease conditions. Ann \nN Y Acad Sci 1460, 3-10 (2020). \n50. Balaji, V. et al. A dimer-monomer switch controls CHIP-dependent substrate ubiquitylation and \nprocessing. Mol Cell 82, 3239-3254 e3211 (2022). \n51. Udayar, V., Chen, Y ., Sidransky, E. & Jagasia, R. Lysosomal dysfunction in neurodegeneration: \nemerging concepts and methods. Trends Neurosci 45, 184-199 (2022). \n52. Tetzlaff, J.E. et al. CHIP targets toxic alpha-Synuclein oligomers for degradation. J Biol Chem 283, \n17962-17968 (2008). \n53. Pantazis, C.B. et al. A reference human induced pluripotent stem cell line for large-scale \ncollaborative studies. Cell Stem Cell 29, 1685-1702 e1622 (2022). \n54. Abu-Remaileh, M. et al. Lysosomal metabolomics reveals V-ATPase- and mTOR-dependent \nregulation of amino acid efflux from lysosomes. Science 358, 807-813 (2017). \n55. Li, H. et al. Benchmarking and Automating the Biotinylation Proteomics Workflow. Res Sq (2024). \n56. Smeriglio, N., Li, H., Mazli, W., Bendel, K. & Hao, L. Contaminant Spot Check and Removal Assay \n(ContamSPOT) for Mass Spectrometry Analysis. Anal Chem 96, 2574-2581 (2024). \n57. Frankenfield, A.M., Ni, J., Ahmed, M. & Hao, L. Protein Contaminants Matter: Building Universal \nProtein Contaminant Libraries for DDA and DIA Proteomics. J Proteome Res 21, 2104-2113 \n(2022). \n58. Ye, Y . & Fortini, M.E. Characterization of Drosophila Presenilin and its colocalization with Notch \nduring development. Mech Dev 79, 199-211 (1998). \n \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n30 \n \nAcknowledgements \nWe thank the Advanced Light Microscope Core at NIDDK for assistance with imaging, NHLBI flow \ncytometry core for cell sorting, and NHLBI Genomic Core for high-throughput sequence, S. Yun at NIDDK \ngenomic core for analyzing the CRISPR screen data, K. Zinsmaier (U. Arizona) for GMR-L116∆ flies, R.  \nPuertollano (NHLBI) for critical reading of the manuscript . The research is supported by the  intramural \nresearch program of NIDDK (Y . Ye), of NICHD (J. Bonifacino), of NINDS (M. Ward), of NCATS (W. \nZheng), and by an NIH grant R01NS121608 (L. Hao).  \n \nAuthor contributions \nJ. Lee, W. Binti Maxli, N. Chin, M. Jarnik, L. Saidi, Y . Xu, and Y . Ye performed the experiments and \nanalyzed the data. J. Zou and W. Zheng provided the knock-in iPSCs, J. Replogle and M. Ward assisted in \ni\n3Neuron platform set-up, W. Binti Maxli and L. Hao conducted the mass spectrometry analysis, M. Jarnik \nand B. Juan conducted the EM analys is. J. Lee, M. Jarnik, L. Hao,  and Y . Ye wrote the paper. All authors \nhelped edit the manuscript. \n \nEthics declarations \nCompeting interests \nThe authors declare no competing financial interest. \n  \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n31 \n \nFigure and figure legends \n \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n32 \n \nFig. 1: Disrupted lysosome homeostasis in iPSC-derived i3Neuron bearing CLN4 mutations. \na, Representative immunoblots of DNAJC5 in i 3Neurons of the indicated genotypes at day 16 in \ndifferentiation (d16). The blots represent 4 independent experiments. HMW, high molecular weight species, \nWT, wild-type; HT, heterozygous; HM, homozygous. b, i3Neurons of the indicated genotypes at d16 were \nstained by green -fluorescent calcein -AM to indicate intracellular esterase activity (live cells) and red -\nfluorescent ethidium homodimer-1 to indicate loss of plasma membrane integrity (dead cells).  Scale bars, \n50 µm. c, Quantification of the experiments in b. Each dot represents a randomly selected field. P-values \nwere determined by one -way ANOV A. n=2 biological repeats. d, i3Neurons of the indicated genotypes at \nd16 were stained with LysoTracker Red (0.2 µM) and Hoechst33442 (1 µ g/mL). The  fluorescence \nintensities were measured by a plate reader. Cells treated with Bafilomycin A1 (Baf.A1) serve as a positive \ncontrol. AU, arbitrary unit. P-values were determined by one-way ANOV A; n=4 biological repeats. e, WT \nand L116∆ homozygous (HM) i3Neurons at the indicated differentiation stage were stained by anti-LAMP1 \nantibodies (green) and Hoechst  (blue). Scale bars, 5 µm. f, Quantification of LAMP1-positive area in \nrandomly selected cells (n>20) in three independent experiments represented by e. P-value by unpaired \nStudent’s t-test. g, Representative EM images of lysosomes in WT and L116∆ HM i3Neurons at d16. h, WT \nand L116∆ HM i3Neurons at d16 were stained by LysoTracker Red and imaged . i, A fraction of DNAJC5 \npositive membrane fractions from the indicated i3Neurons were analyzed by immunoblotting. j, A heat map \nshows the proteins up- or down-regulated on the DNAJC5-positive organelles in L116∆ HM i3Neurons, as \ndetermined by three mass spectrometry analyses. k, A Venn diagram shows the three major categories of \nthe proteins down-regulated on the DNAJC5-positive organelles by L116∆ HM mutation. l, A volcano plot \nhighlighting the proteins up- and down -regulated on the DNAJC5 -positive organelles by L116∆  HM \nmutation. Error bars in c, d, f represent mean ± standard error of the mean (s.e.m.).  \n \n \n  \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n33 \n \n \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n34 \n \nFig. 2: CLN4 mutant aggregates can damage lysosomal membranes. \na, A schematic diagram showing the in vitro lysosome-damaging assay. SLO, Streptolysin O, CLC; calf \nliver cytosol. b, Purified WT DNAJC5 or L116∆ mutant incubated with buffer or CLC with or without ATP \nat 37 ° C for 30 min were analyzed by immunoblotting. HMW, high molecular weight species. c,  U2OS \ncells stained with LysoTracker Red (LysoT Red) were treated with SLO and then incubated with either \nbuffer or the indicated DNAJC5 proteins with CLC, ATP and a cell impermeable dye NucSpot green at 37 \n°C for 30 min. Cells were imaged immediately. Scale bars, 10 µm. d, Quantification of the relative LysoT-\npositive areas in individual permeabilized (PM) cells treated with buffer or the indicated DNAJC5 proteins. \nIntact cells as indicated by the lack of NucSpot staining serves as a reference. P-values were determined by \none-way ANOV A from three biological repeats. e, The modified lysosome damaging assay using Dextran \n(10 kDa)-loaded cells stained with LysoTracker green. f, U2OS cells loaded with Dextran (10 kDa, magenta) \nwere stained with LysoTracker green (green ) and then treated with SLO. Permeabilized cells were then \ntreated CLC with ATP and the indicated CLN4 mutant proteins or buffer as a negative control. Dashed lines \nindicate intact cells. Scale bars, 10 µm. g, Quantification of the relative LysoT positive areas (top panel) or \nDextran-positive areas (bottom panel) in individual permeabilized (PM) cells treated with buffer or the \nindicated DNAJC5 proteins plus ATP and CLC. P-values were determined by one-way ANOV A from two \nindependent experiments. h, As in c, excepted that permeabilized cells were treated with monomeric L116∆ \n(left) or purified L116∆ HMW species (right) plus ATP but in the absence of cytosol. Scale bars, 10 µm. i, \nQuantification of the relative LysoT-positive areas in experiment h. P-values were determined by unpaired \nstudent’s t-test from two  biological repeats. Error bars in d and g represent mean ± standard error of the \nmean (s.e.m.). Confocal images in c, f, and h were processed using maximum intensity projection to \nrepresent the fluorescence signal across the z-stacks. \n  \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n35 \n \n \n \nFig. 3: A ubiquitin- and ESCRT-dependent mechanism counteracts CLN4 -induced lysotoxicity in \nnon-neuronal cells. \na, U2OS cells transfected mCitrine (Ci) or Ci -tagged DNAJC5 variants were methanol-fixed and stained \nwith ubiquitin antibody (FK2) (magenta). The box-indicated area in the bottom panels is enlarged to show \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n36 \n \nthe partial co -localization of DNAJC5 with ubiquitin (Ub)  (right panels) . Green and cyan -dotted lines \nhighlight cell boundaries and nuclei of transfected cells, respectively. Scale bars, 10 µm. b, Quantification \nof the relative cytoplasmic Ub  puncta areas in cells transfected with the indicated DNAJC5 variants. \nUntransfected (UT) cells serve as a reference. P -values were determined by one-way ANOV A from two \nbiological repeats. c, As in a, except that cells transfected with the indicated DNAJC5 variants were stained \nwith antibodies against HGS. Scale bars, 10 µm. d, Quantification of the relative cytoplasmic HGS puncta \nareas as shown in c. P -values were determined by one-way ANOV A from  two biological repeats. e,  \nHEK293T cells stably expressing Keima-tagged WT DNAJC5 or the indicated CLN4 mutants were \ntransfected with a dominant negative form of VPS4 (E228Q) and analyzed by flow cytometry. f, HEK293T \ncells expressing Keima-tagged DNAJC5 variants were treated with Baf.A1 (100 nM) or TAK-243 (1 µM) \nfor 4 h and then analyzed by flow cytometry. g, HEK293T cells expressing Keima-tagged DNAJC5 variants \nwere treated with DMSO as a control or TAK-243 (T243, 1 µM)  for 16 h . NP40-soluble and -insoluble \nfractions were analyzed by immunoblotting.  h, U2OS cells transfected with Ci -tagged DNAJC5 variants \nwere treated with TAK-243 (1 µM, 4 h) and stained with LysoTrack Red (magenta) and Hoechst. Scale bars, \n10 µm. i, Quantification of the relative LysoT -positive areas in transfected cells after 4 h treatment with \nTAK-243 (1 µM). Untransfected cells serve as internal negative controls. P-values from unpaired student’s \nt-test (left) or one-way ANOV A (right) from two biological repeats. j, As in h, except that cells transfected \nwith Celurean  (Ce) or Ce-tagged DNAJC5 variants together with VPS4 E228Q-HA were stained with \nLysoTracker Red (magenta). Scale bars, 10 µm. k, Quantification of the relative LysoT positive areas in \ncells expressing the indicated proteins. P -values are determined by one -way ANOV A from two biological \nrepeats. Error bars in b, d, i and k represent mean ± standard error of the mean (s.e.m.). In a, c, h, and j, z -\nstack images were processed using maximum intensity projection. \n      \n \n  \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n37 \n \n \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n38 \n \nFig. 4: CHIP-mediated ubiquitination protects lysosomes from DNAJC5 L116∆-induced lysosome \ndamage. \na,  A volcano plot shows the selected sgRNAs enriched in the neutral population from the Keima-DNAJC5 \nWT-based CRISPR screen. b, A Venn diagram shows CHIP and RAB7A as the only overlapped genes from \nthe two CRISPR screens, which were also identified in the DNAJC5-interactome. c, A heat map shows the \nrelative significance of the positive hits from the CRISPR screens. d, HEK293T cells expressing Keima-\nDNAKC5 WT were transfected with shRNA-expressing constructs targeting the indicated genes and \nanalyzed by flow cytometry. Con, Control. e, HEK293T cells were transfected with FLAG-tagged DNAJC5 \nvariants as indicated or an empty vector (EV) together with MYC-tagged CHIP.  FLAG beads were used to \npull down DNAJC5. Bound proteins were analyzed by immunoblotting together with a fraction of the input \nsamples. f, WT or CHIP knockout (KO) cells were transfected with Ci-tagged L116∆, methanol-fixed, and \nstained with ubiquitin antibodies. In the right panels, CHIP KO cells were co -transfected with Ci -L116∆ \nand CHIP-MYC. Green- and blue-dotted lines indicate transfected cells and nuclei respectively. Note that \nCHIP expression rescues the defect in lysosome-associated ubiquitination in CHIP KO cells. Scale bars, 10 \nµm. g, Quantification of the positive cytoplasmic ubiquitin area normalized by that in untransfected (UT) \ncells in f. n= 2 biological repeats. h, As in f, excepted that cells were stained with HGS and MYC antibodies. \nScale bars, 10 µm. i, Quantification of h as in g. P -values in g and I are determined by one -way ANOVA \nfrom two biological repeats. j, WT and CHIP KO U2OS cells were transfected with stained with Ci-L116∆, \nstained with LysoTracker Red (magenta) and imaged. Where indicated, CHIP KO cells were co-transfected \nwith Ci-L116 ∆ and CHIP -MYC. Note that only in CHIP KO background, L116∆ expression reduced \nLysoTracker signal. Scale bars, 10 µm. k, Quantification of the relative LysoTracker signal in j. P-values \nare determined by one -way ANOV A from two biological repeats.  In f, h, and j, z -stack images were \nprocessed using maximum intensity projection. \n \n \n  \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n39 \n \n \n \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n40 \n \nFig. 5: CHIP expression partially rescues the lysosomal defect and cell death phenotypes in L116∆ \ni3Neurons. \na, Quantification of the levels of the indicated proteins in NP40-soluble or insoluble fractions during \ni3Neuron differentiation. Representative blots are shown in Extended data Fig. 6a. Error bars represent s.e.m. \nof three biological repeats. b, WT and L116∆ HM i3Neurons were mock infected (left and middle panels) \nor infected with CHIP-expressing lentiviruses at d6. Cells were stained with LysoTracker Red and Hoechst \nat d16 and imaged. Shown are 3D reconstructed view (top panels) of soma or a z section of a neurite -\nenriched region (bottom panels). Scale bars, 5 µm. c, Quantification of LysoT signal in soma as shown in \nb. P-value is by unpaired student’s t -test. Error bars represent s.e.m. of individual cells (n>20) from two \nbiological repeats. d, As in b except that L116∆ HM i3Neurons were also infect with a lentivirus expressing \nCHIP ∆U-box for comparison and that cells at d16 were treated with DQ-BSA (40 µg/ml) for 6 h and then \nstained with Hoechst. Z-stack images were processed using maximum intensity projection.  Scale bars, 10 \nµm. e, Quantification of the relative DQ-BSA fluorescence intensity as shown in d. Error bars represent \ns.e.m of individual cells from two biological repeats. P-values are determined by one-way ANOV A. f, WT \nor L116∆ HM i3Neurons were infected with lentiviruses expressing either control (sh-con) or CHIP specific \nshRNA. Cells at d13 were stained by green calcein-AM and red ethidium homodimer-1. Scale bars, 50 µm. \ng, Quantification of cell death as shown in f from two  biological repeats. Each dot represents a randomly \nselected field.  Error bars represent means ± s.e.m. P-value are determined by one-way ANOV A. h, WT or \nL116∆ HM i3Neurons were infected with lentiviruses carrying either an empty vector or a CHIP-expressing \ncassette. Cell death at d16 were measured as in f. Shown is the quantification results as in g.  Error bars \nrepresent means ± s.e.m. of two  biological repeats.  Each data point represents a randomly selected field.  \nP-value are determined by one-way ANOV A. \n  \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n41 \n \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n42 \n \nFig. 6: CHIP downregulates L116∆ and rescues lipofusin accumulation and neurodegeneration in a \nDrosophila CLN4 disease model. \na, Representative confocal images of an eye disc and a segment of mid -gut from third insta larva of flies \nexpressing Keima-dDNAJC5 (Csp1) by heat shock (HS)-Gal4. Scale bars, 100 µm. b, Confocal images of \neye discs from third instar larva of GMR>Keima-dDNAJC5 (Control) or GMR>Keima-dDNAJC5; dCHIP \n(dCHIP). Scale bars, 10 µm. The graph shows the quantification of the acidic/neutral Keima signal ratio. \nError bars represent means ± s.e. m., N=12 discs. P-value was determined by unpaired student’s t -test. c, \nRepresentative eye discs from larvae of the indicated genotypes were stained with anti-Ubiquitin (top panels) \nor DNAJC5 antibodies. Shown are maximum projected view of confocal sections of entire tissues. Scale \nbars, 10 µm. d, Quantification of the ubiquitin signal (top) or L116∆ levels (bottom) in individual eye discs. \nError bars represent means ± s.e. m. P-values were determined by one-way ANOV A.  e, Eye discs from \nlarvae of GMR> Keima-DNAJC5 plus the indicated transgenes were stained with AmyTracker (2 µg/mL) \nand imaged. Scale bars, 10 µm. f, Quantification of the AmyTracker -positive punctae in e. Error bars \nrepresent means ± s.e.m. P -values were determined by one-way ANOV A. g, Eye discs from third insta \nlarvae of the indicated genotypes were stained with acridine orange (1 µg/mL ) and imaged. Scale bars, 10 \nµm. The graph shows the quantification of the experiment. Error bars represent means ± s.e. m. P-values \nwere determined by one-way ANOV A. h. Eyes of flies with the indicated genotypes were either directly \nimaged by a CCD camera mounted on a dissecting microscope (top panels) or first fixed and then scanned \nby a confocal microscope (bottom panels).  Confocal images in this figure were processed using maximum \nintensity projection to reconstruct the in-depth view of tissues. \n  \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n43 \n \n \n \nFig. 7: CHIP’s neuroprotective activity depends on ESCRT. \na, Representative images of the eyes of flies with the indicated genotypes. Ctrl, Control; KD, knockdown. \nScale bars, 100 µm. b, Eye discs from third instar larvae of the indicated genotypes were stained with \nDNAJC5 antibodies and DAPI (blue) to label DNA.  Z-stack i mages were processed using maximum \nintensity projection. The white-dotted lines indicate the morphogenic furrow. A, anterior part. Scale bars, \n10 µm. c, Quantification of the relative DNAJC5 fluorescence intensity in individual eye discs as shown in \nb. Error bars represent means ± s.e.m. P-values were determined by one-way ANOV A. d, A model depicts \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint \n\n44 \n \nthe role of CHIP - and ubiquitin-mediated microautophagy in counteracting CLN4 mutant -induced \nlysosome damage.  \n105 and is also made available for use under a CC0 license. \n(which was not certified by peer review) is the author/funder. This article is a US Government work. It is not subject to copyright under 17 USC \nThe copyright holder for this preprintthis version posted February 19, 2025. ; https://doi.org/10.1101/2025.02.18.638932doi: bioRxiv preprint","source_license":"Public-Domain","license_restricted":false}