Neuronal Preconditioning Requires the Mitophagic Activity of C-Terminus of HSC70-Interacting Protein

This Accepted Manuscript has not been copyedited and formatted. The final version may differ from this version.

Research Articles: Cellular/Molecular

Neuronal Preconditioning Requires the Mitophagic Activity of C-terminus of HSC70-Interacting Protein

Britney N. Lizamaa,b, Amy M. Palubinskya,b, Vineeth A. Raveendranc, Annah M. Moored, Joel D. Federspiele, Simona G. Codreanue, Daniel C. Lieblere and BethAnn McLaughlinb,f,g aNeuroscience Graduate Group bVanderbilt Brain Institute cVanderbilt International Summer Research Academy dVanderbilt Interdisciplinary Graduate Program eDepartments of Biochemistry fNeurology gPharmacology, Vanderbilt University Medical Center, 465 21st Ave S MRB III, Nashville, TN, 37240

DOI: 10.1523/JNEUROSCI.0699-18.2018

Received: 13 March 2018

Revised: 8 June 2018

Accepted: 13 June 2018

Published: 22 June 2018

Author contributions: B.N.L., A.M.P., and B.M. designed research; B.N.L., A.M.P., V.R., A.M., J.F., S.G.C., and D.C.L. performed research; B.N.L., A.M.P., J.F., S.G.C., and B.M. analyzed data; B.N.L. wrote the first draft of the paper; B.N.L., A.M.P., V.R., A.M., J.F., S.G.C., D.C.L., and B.M. edited the paper; B.N.L. and B.M. wrote the paper.

Conflict of Interest: The authors declare no competing financial interests.

The authors thank Dr. Cam Patterson for providing the original CHIP knockout animals; Ms. Sharon Klein, Ms. Arulita Gupta, and Ms. Dominique Szymkiewicz for technical assistance and ImageJ expertise; Dr. Joshua Fessel and Dr. Alice Soragni for electron microscopy expertise; Dr. Christopher E. Wright and Mr. Jeff Duryea for imaging assistance; and Ms. Ama J. Winland for cell culture maintenance. We also thank the Vanderbilt Cell Imaging Shared Resource for transmission electron microscopy preparation and imaging. This work was supported by the Walter and Suzanne Scott Foundation funding of the J.B. Marshall Laboratory (B.M, A.M.P., B.N.L.), the Dan Marino Foundation (B.M., A.M.P.), NIH grants NS050396 (B.M.) and RO1ES022936 (B.M., D.C.L.), a Vanderbilt Brain Institute Scholarship (B.N.L., A.M.P.), a VISRA Program Scholarship (V.R.), and predoctoral fellowships from the AHA 15PRE25100000 (A.M.P.), 14PRE2003500007 (B.N.L.). The authors declare no conflict of interest.

Corresponding author: [email protected]

Cite as: J. Neurosci ; 10.1523/JNEUROSCI.0699-18.2018

Alerts: Sign up at www.jneurosci.org/cgi/alerts to receive customized email alerts when the fully formatted version of this article is published.

Accepted manuscripts are peer-reviewed but have not been through the copyediting, formatting, or proofreading process.

ͳ Neuronal Preconditioning Requires the Mitophagic Activity of C-terminus of HSC70- ʹ Interacting Protein ͵ Ͷ Britney N. Lizamaa,b, Amy M. Palubinskya,b, Vineeth A. Raveendranc, Annah M. Moored, Joel D. Federspiele, ͷ Simona G. Codreanue, Daniel C. Lieblere, & BethAnn McLaughlin*b,f,g ͸ ͹ aNeuroscience Graduate Group, bVanderbilt Brain Institute, cVanderbilt International Summer Research Academy, ͺ dVanderbilt Interdisciplinary Graduate Program, Departments of eBiochemistry, fNeurology and gPharmacology, ͻ Vanderbilt University Medical Center, 465 21st Ave S MRB III, Nashville, TN, 37240 *Corresponding author: ͳͲ [email protected] ͳͳ ͳʹ Abstract: ͳ͵ C-terminus of HSC70-interacting protein (CHIP, STUB1) is a ubiquitously expressed cytosolic E3-ubiquitin ͳͶ ligase. CHIP-deficient mice exhibit cardiovascular stress and motor dysfunction prior to premature death. This ͳͷ phenotype is more consistent with animal models in which master regulators of autophagy are affected rather ͳ͸ than with the mild phenotype of classic E3-ubiquitin ligase mutants. The cellular and biochemical events that ͳ͹ contribute to neurodegeneration and premature aging in CHIP KO models remain poorly understood. ͳͺ Electron and fluorescent microscopy demonstrates that CHIP deficiency is associated with greater numbers of ͳͻ mitochondria, but these organelles are swollen and misshapen. Acute bioenergetic stress triggers CHIP ʹͲ induction and re-localization to mitochondria where it plays a role in the removal of damaged organelles. This ʹͳ mitochondrial clearance is required for protection following low-level bioenergetic stress in neurons. CHIP ʹʹ expression overlaps with stabilization of the redox stress sensor PTEN-inducible kinase 1 (PINK1) and is ʹ͵ associated with increased LC3-mediated mitophagy. Introducing human promoter-driven vectors with ʹͶ mutations in either the E3 ligase or TPR domains of CHIP in primary neurons derived from CHIP-null animals ʹͷ enhances CHIP accumulation at mitochondria. Exposure to autophagy inhibitors suggests the increase in ʹ͸ mitochondrial CHIP is likely due to diminished clearance of these CHIP-tagged organelles. Proteomic analysis ʹ͹ of WT and CHIP KO mouse brains (4 male, 4 female per genotype) reveals proteins essential for maintaining ʹͺ energetic, redox and mitochondrial homeostasis undergo significant genotype-dependent expression changes. ʹͻ Together these data support the use of CHIP deficient animals as a predictive model of age-related ͵Ͳ degeneration with selective neuronal proteotoxicity and mitochondrial failure. ͵ͳ ͵ʹ Significance Statement: Mitochondria are recognized as central determinants of neuronal function and ͵͵ survival. We demonstrate that C-terminus of HSC70-Interacting Protein (CHIP) is critical for neuronal ͵Ͷ responses to stress. CHIP upregulation and localization to mitochondria is required for mitochondrial ͵ͷ autophagy (mitophagy). Unlike other disease-associated E3 ligases such as Parkin and Mahogunin, CHIP ͵͸ controls homeostatic and stress-induced removal of mitochondria. While CHIP deletion results in greater ͵͹ numbers of mitochondria, these organelles have distorted inner membranes without clear cristae. Neuronal ͵ͺ cultures derived from animals lacking CHIP are more vulnerable to acute injuries, and transient loss of CHIP ͵ͻ renders neurons incapable of mounting a protective response following low-level stress. Together these data ͶͲ suggest that CHIP is an essential regulator of mitochondrial number, cell signaling and survival. Ͷͳ Ͷʹ Ͷ͵ Introduction Ͷͺ Garza, 2012). HSP70 binding partners alter the ͶͶ The HSP70 complex is capable of blocking Ͷͻ localization, activity and expression of the Ͷͷ neurodegeneration triggered by a host of genetic ͷͲ chaperone complex itself and its client proteins. E3 Ͷ͸ mutations, physiological events and environmental ͷͳ ligases, such as Parkin and C-terminus of HSC70- Ͷ͹ stressors (Franklin et al., 2005; Gestwicki and ͷʹ Interacting Protein (CHIP), regulate HSP70- ͳ

ͷ͵ mediated protein degradation (Ballinger et al., ͳͲͲ In this work, we sought to understand the ͷͶ 1999; Jiang et al., 2001). CHIP is a 303 amino acid ͳͲͳ underpinnings of the devastating phenotype ͷͷ cytosolic protein that contains an N-terminal ͳͲʹ caused by loss of CHIP by evaluating its role in cell ͷ͸ tetratricopeptide repeat (TPR) domain regulating ͳͲ͵ signaling using neuronal models of acute ͷ͹ HSP/HSC70-docking, a helix-loop-helix domain, ͳͲͶ pathophysiological stress. Using a combination of ͷͺ and a C-terminal Ubox domain that is necessary for ͳͲͷ cellular imaging, proteomics, biochemistry, ͷͻ binding to E2-conjugating enzymes (Ballinger et al., ͳͲ͸ molecular targeting and ultrastructural assays, we ͸Ͳ 1999; Jiang et al., 2001; Xu et al., 2008). CHIP is ͳͲ͹ determined that CHIP deficiency results in radical ͸ͳ unique among E3 ligases in that its expression is ͳͲͺ mitochondrial reorganization not seen in other E3 ͸ʹ induced by acute bioenergetic stress and in post- ͳͲͻ ligase deficient animals. These data reinforce a ͸͵ mortem samples from patients with stroke ͳͳͲ model in which CHIP functions are more in ͸Ͷ (Stankowski et al., 2011; Lizama et al., 2017). CHIP ͳͳͳ keeping with a role as an essential autophagy ͸ͷ is also unique in that it forms asymmetric ͳͳʹ regulator than a standard E3 ligase. Taken together, ͸͸ homodimers, which have not been observed in ͳͳ͵ our data supports the use of these animals as a ͸͹ other E3 ligase proteins and are only formed in ͳͳͶ powerful and robust model of neurological ͸ͺ higher vertebrates, suggesting it may perform ͳͳͷ dysfunction induced by mitochondrial failure and ͸ͻ distinctive roles in cell biology (Zhang et al., 2005; ͳͳ͸ secondary proteotoxicity. ͹Ͳ Ye et al., 2017). ͳͳ͹ ͹ͳ Further evidence of the uniqueness of CHIP can ͳͳͺ Experimental Methods ͹ʹ be found in studies demonstrating that mice ͳͳͻ Materials and Reagents ͹͵ deficient in CHIP have pervasive dysfunction far ͳʹͲ Tissue culture: Fetal bovine serum (SH30070.03) ͹Ͷ more robust than would be predicted based on a ͳʹͳ was obtained from Hyclone. Dulbecco’s modified ͹ͷ primary role as a conventional E3 ligase. CHIP ͳʹʹ Eagle’s medium (DMEM, 11995) with high glucose, ͹͸ insufficiency results in decreased life expectancy ͳʹ͵ minimum essential medium (MEM, 51200), ͹͹ and profound lipid oxidation, as well as poor ͳʹͶ Neurobasal medium, B27 supplement (17504044), ͹ͺ performance on behavioral assessments of motor ͳʹͷ N2 supplement (17502048), 0.25% Trypsin–EDTA, ͹ͻ function and anxiety (McLaughlin et al., 2012; ͳʹ͸ trypan blue stain 0.4%, and penicillin-streptomycin ͺͲ Palubinsky et al., 2015). Notably, CHIP knockout ͳʹ͹ (15140-122) were purchased from Invitrogen. ͺͳ (KO) animals are unable to complete even simple ͳʹͺ LipoJet In Vitro Transfection Kit (Ver. II) was ͺʹ behavioral assessments given their moribund ͳʹͻ purchased from SignaGen. ͺ͵ nature. ͳ͵Ͳ Western blotting: XT-MOPS running buffer, ͺͶ Mutations in CHIP have been identified in ͳ͵ͳ Tris-Glycine transfer buffer, Criterion Bis-Tris gels, ͺͷ patients with an early-onset, recessive form of ͳ͵ʹ Laemmli buffer, and precision plus protein all blue ͺ͸ spinocerebellar ataxia, in which mutations in major ͳ͵͵ standards were purchased from Bio-Rad ͺ͹ structural domains confer loss of CHIP function ͳ͵Ͷ Laboratories. Membrane-blocking solution was ͺͺ (Heimdal et al., 2014; Shi et al., 2014; Synofzik et al., ͳ͵ͷ from Zymed. Hybond P polyvinylidene difluoride ͺͻ 2014; Bettencourt et al., 2015). Patients exhibit some ͳ͵͸ membranes were acquired from GE Healthcare. Gel ͻͲ unifying symptoms including progressive ͳ͵͹ Code Blue Stain Reagent and Western Lightning ͻͳ deterioration in muscle coordination and tone, ͳ͵ͺ Chemiluminescence Reagent Plus were obtained ͻʹ speech difficulty, and cerebellar atrophy. Yet, ͳ͵ͻ from Thermo Scientific. ͻ͵ symptom heterogeneity exists among patients with ͳͶͲ For Western blotting, CHIP (PC711, RRID: ͻͶ CHIP mutations, which include dementia, ͳͶͳ AB_2198058) antibody was purchased from ͻͷ hypogonadism, and epilepsy. Such variable ͳͶʹ Calbiochem. Antibody for p62 was obtained from ͻ͸ observations support a role for CHIP beyond the ͳͶ͵ Cell Signaling Technology (5114). LC3 for Western ͻ͹ cerebellum, where CHIP mutations can cause a ͳͶͶ blotting was obtained from MBL International ͻͺ multidimensional, multisystemic degenerative ͳͶͷ (PD014, RRID: AB_843283), and for ͻͻ disease (Hayer et al., 2017). ͳͶ͸ immunofluorescence LC3 was obtained from ʹ

ͳͶ͹ Abcam (ab62116, RRID:AB_2281379). ͳͻͶ reabsorption and cannibalism of KO animals that ͳͶͺ HSP70/HSP72 (SPA-811, RRID:AB_2120896), and ͳͻͷ are often runted. Genotyping was performed by ͳͶͻ HSC70 antibodies (SPA-816, RRID:AB_312224) ͳͻ͸ PCR with DNA from tail clippings using primers ͳͷͲ were purchased from Enzo Life Science, Inc. PINK1 ͳͻ͹ for the CHIP allele. Primers were purchased from ͳͷͳ (BC100-494, RRID:AB_10127658) and TOM20 ͳͻͺ XXIDT and the sequences of the reverse and ͳͷʹ (H00009804-M01, RRID:AB_2303717) antibodies ͳͻͻ forward primers used are 5’ TGA CAC TCC TCC ͳͷ͵ were purchased from Novus Biologicals. CHIP (sc- ʹͲͲ AGT TCC CTG AG 3’ and 5’ AAT CCA CGA GGC ͳͷͶ 133066, RRID:AB_2286870) antibody for ʹͲͳ TCC GCC TTT 3’, respectively. ͳͷͷ immunofluorescence was purchased from Santa ʹͲʹ ͳͷ͸ Cruz Biotechnology. ʹͲ͵ Rat Primary Neuronal Culture ͳͷ͹ Unless otherwise stated, all other chemicals ʹͲͶ Forebrain cultures were prepared from ͳͷͺ were purchased from Sigma-Aldrich. ʹͲͷ embryonic day 18.5 Sprague-Dawley rats as ͳͷͻ ʹͲ͸ previously described (Stankowski et al., 2011). ͳ͸Ͳ Experimental Design and Statistical Analysis ʹͲ͹ Briefly, forebrain was dissociated and the resultant ͳ͸ͳ Statistical analyses were performed using ʹͲͺ cell suspension was adjusted to 750,000 cells/well in ͳ͸ʹ GraphPad Prism version 5.03 (RRID: ʹͲͻ 6-well tissue culture plates containing five, 12-mm ͳ͸͵ RRID:SCR_002798). Error bars indicate standard ʹͳͲ poly-L-ornithine-treated coverslips/well. Plating ͳ͸Ͷ error of the mean (s.e.m) and statistical significance ʹͳͳ media was composed of 84% (v/v) Dulbecco’s ͳ͸ͷ was assessed by one-way ANOVA followed by ʹͳʹ modified MEM, 8% (v/v) Ham’s F12-nutrients, 8% ͳ͸͸ Bonferroni post-hoc analysis (Fig. 1C, Fig. 3A, Fig. ʹͳ͵ (v/v) fetal bovine serum, 24 U/ml penicillin, 24 ͳ͸͹ 7A, Table 1, Table 2, Table 3), paired two-tailed ʹͳͶ μg/ml streptomycin, and 80 μM L-glutamine. ͳ͸ͺ Students t-test (Table 4), unpaired two-tailed ʹͳͷ Cultures were maintained at 37°C, 5% CO2, and ͳ͸ͻ Student’s t-test (Fig. 1E, Fig. 2D, Fig. 4E, Fig. 5A, ʹͳ͸ media was partially replaced every 2 – 3 days. Glial ͳ͹Ͳ Fig. 6B, Fig. 7B, Fig. 7C), Mann-Whitney analysis ʹͳ͹ cell proliferation was inhibited after two days in ͳ͹ͳ (Fig. 4F), or two-way ANOVA (Fig. 5B). ʹͳͺ culture with 2 μM cytosine arabinoside, and cells ͳ͹ʹ Experimental replicates labeled as “n” were ʹͳͻ were maintained thereafter in Neurobasal media ͳ͹͵ derived from at least 3 independent cell cultures or ʹʹͲ supplemented with 2% B27 and 2% NS21 until 17 ͳ͹Ͷ animals as indicated and are described further in ʹʹͳ days in vitro (Chen et al., 2008). B27 was then ͳ͹ͷ each Methods section. Significance is reported in ʹʹʹ removed from the feeding media and cultures were ͳ͹͸ the Results section, and full statistical results are ʹʹ͵ maintained in Neurobasal supplemented with ͳ͹͹ detailed in figure legends. ʹʹͶ NS21 only. This preparation yields a 98% neuronal ͳ͹ͺ ʹʹͷ culture with a mature complement of NMDA ͳ͹ͻ Maintenance of CHIP WT and KO Mouse Colony ʹʹ͸ receptors (Zeiger et al., 2010). All experiments were ͳͺͲ The Institutional Animal Care and Use ʹʹ͹ conducted between DIV20-27. ͳͺͳ Committee at Vanderbilt University Medical ʹʹͺ ͳͺʹ Center approved all animal husbandry and ʹʹͻ Neuronal Bioenergetic Stress ͳͺ͵ experiments. Parental mouse lines were previously ʹ͵Ͳ Neurons were exposed to brief periods of ͳͺͶ described (Dai et al., 2003). All mice were ʹ͵ͳ bioenergetic stress by removing oxygen and ͳͺͷ maintained on a mixed background of C57BL/6 and ʹ͵ʹ glucose in order to trigger mitochondrial stress ͳͺ͸ 129SvEv for 8 or 9 generations. Further ʹ͵͵ signaling. Oxygen-and-glucose deprivation (OGD) ͳͺ͹ backcrossing exacerbates the early lethality ʹ͵Ͷ was performed as described with minor ͳͺͺ observed in CHIP KO animals resulting in a less ʹ͵ͷ modifications (Palubinsky et al., 2015). Briefly, a ͳͺͻ than 5% survival at birth. As CHIP KO male mice ʹ͵͸ complete media exchange was performed by ͳͻͲ are sterile, heterozygous matings were used to ʹ͵͹ transferring coverslips into wells containing ͳͻͳ maintain the colony. Generating CHIP KO animals ʹ͵ͺ deoxygenated, glucose-free Earle’s balanced salt ͳͻʹ from heterozygous crossings occurs at non- ʹ͵ͻ solution (150 mM NaCl, 2.8 mM KCl, 1 mM CaCl2, ͳͻ͵ Mendelian rates because of preferential ʹͶͲ and 10 mM HEPES; pH 7.3). Cultures were placed ͵

ʹͶͳ in an anaerobic chamber (Billups-Rothberg) for ʹͺͺ in TNEB lysis buffer (50 mM Tris-Cl, pH 7.8, 2 mM ʹͶʹ various durations (15–90 minutes) at 37°C. OGD ʹͺͻ EDTA, 150 mM NaCl, 8 mM β-glycerophosphate, ʹͶ͵ was terminated by moving coverslips into 6- or 24- ʹͻͲ 100 μM sodium orthovanadate, 1% [v/v] Triton X- ʹͶͶ well plates containing MEM, 0.01% (w/v) BSA, ʹͻͳ 100, and protease inhibitor (P8340) diluted 1:1000) ʹͶͷ 2.5% (v/v) HEPES, and 2x N2 ʹͻʹ and extracts were prepared for Western blotting. ʹͶ͸ (MEM/BSA/HEPES/N2). After 18–24 hours, ʹͻ͵ Pilot experiments revealed optimal knockdown of ʹͶ͹ neuronal viability was assessed visually as well as ʹͻͶ CHIP expression at 24 hours (Fig. 5a), and this time ʹͶͺ via LDH assays. ʹͻͷ point was used for all subsequent studies of acute ʹͶͻ ʹͻ͸ CHIP knockdown. ʹͷͲ Viability Assay of Rat Primary Neurons ʹͻ͹ ʹͷͳ Cell viability was assessed using a lactate ʹͻͺ Western Blotting of Primary Neuronal Lysate ʹͷʹ dehydrogenase (LDH)-based in vitro toxicity kit ʹͻͻ For in vitro lysates, all cell lysis and harvesting ʹͷ͵ (Sigma), which colorimetrically determines the ͵ͲͲ steps took place on ice. Cells were washed twice ʹͷͶ amount of LDH released into the culture media ͵Ͳͳ with ice- cold 1x phosphate-buffered saline (PBS), ʹͷͷ from dead and dying neurons. ͵Ͳʹ and following the second wash, 150–400 μL of ʹͷ͸ In order to account for variation in total LDH ͵Ͳ͵ TNEB lysis buffer was added. ʹͷ͹ content, raw LDH values were normalized to the ͵ͲͶ Approximately 100-200 μL of lysate was re- ʹͷͺ toxicity caused by 90-minute OGD, which, at 24 ͵Ͳͷ suspended in an equal volume of Laemmli buffer ʹͷͻ hours, results in 100% neuronal cell death in this ͵Ͳ͸ with β-mercaptoethanol (1:20). Protein samples ʹ͸Ͳ system. Viability is also confirmed visually with ͵Ͳ͹ were heated to 95°C for 10 minutes and stored at - ʹ͸ͳ corresponding brightfield images. ͵Ͳͺ 20°C. Equal protein concentrations were ʹ͸ʹ ͵Ͳͻ determined by DC Protein Assay Kit II and ʹ͸͵ Preconditioning and CHIP siRNA Knockdown in Rat ͵ͳͲ samples were separated on 4-12% Criterion Bis-Tris ʹ͸Ͷ Primary Neurons ͵ͳͳ gels. Proteins were transferred to PVDF ʹ͸ͷ Primary neuronal cultures were grown for 20-25 ͵ͳʹ membranes and blocked in methanol for 5 minutes. ʹ͸͸ days in vitro after which they were exposed to low- ͵ͳ͵ Following 10 minutes of drying, the membranes ʹ͸͹ level bioenergetic stress (15-minute OGD) and ͵ͳͶ were incubated overnight with antibodies detecting ʹ͸ͺ immediately returned to normoxia in ͵ͳͷ CHIP, HSP70, HSC70, TOM20, p62, PINK1, or LC3 ʹ͸ͻ MEM/HEPES/BSA/N2 medium. For studies of ͵ͳ͸ in 5% non-fat dry milk in TBS-Tween (0.1%) or in ʹ͹Ͳ mitophagy, neurons were incubated in bafilomycin ͵ͳ͹ Membrane Blocking Solution at 4°C. Membranes ʹ͹ͳ A1 (BafA; 1 nM) 30 minutes prior to, during, and ͵ͳͺ were washed three times with TBS-Tween, and ʹ͹ʹ after OGD. Twenty-four hours following ͵ͳͻ incubated with horseradish peroxidase conjugated ʹ͹͵ preconditioning, neurons were exposed to 90- ͵ʹͲ secondary antibodies for 1 hour at room ʹ͹Ͷ minute OGD and viability was determined 24 ͵ʹͳ temperature (RT). Following three washes in TBS- ʹ͹ͷ hours later using LDH assays and visual ͵ʹʹ Tween, protein bands were visualized using ʹ͹͸ confirmation. ͵ʹ͵ Western Lightning© chemiluminescence plus ʹ͹͹ CHIP small interfering RNA (siRNA) was ͵ʹͶ enhanced luminol reagent. Western blots were ʹ͹ͺ obtained from QIAGEN Science (50- ͵ʹͷ analyzed using NIH ImageJ (RRID:SCR_002285) to ʹ͹ͻ CCAGCTGGAGATGGA GAGTTA-30) ͵ʹ͸ determine the mean relative densities of each ʹͺͲ (Stankowski et al., 2011). Using the LipoJet ͵ʹ͹ protein band, and fold changes were calculated ʹͺͳ Transfection Kit, CHIP siRNA (25 nM) and green- ͵ʹͺ using untreated cultures as controls. Data represent ʹͺʹ fluorescent protein (GFP; 2 μg) were added ͵ʹͻ results from at least 3 independent experiments as ʹͺ͵ dropwise into each well of a 6-well dish containing ͵͵Ͳ indicated in figure legends. ʹͺͶ neurons in 2 ml of growth medium. Six hours after ͵͵ͳ ʹͺͷ addition of transfection reagents, cells underwent a ͵͵ʹ ʹͺ͸ complete medium change. CHIP siRNA transfected ͵͵͵ CHIP WT and KO Mouse Primary Neuronal Culture ʹͺ͹ cells were harvested 24–72 hours post-transfection ͵͵Ͷ Forebrain cultures were prepared from Ͷ

͵͵ͷ embryonic day 18 mice generated by heterozygous ͵ͺʹ Immunofluorescence of Primary Neurons ͵͵͸ matings as described (Palubinsky et al., 2015). ͵ͺ͵ Cells were fixed in 4% paraformaldehyde (PFA) ͵͵͹ Briefly, mice were decapitated and the entire brain ͵ͺͶ and then permeabilized with 0.1% Triton X-100 as ͵͵ͺ was stored individually in Hibernate E medium ͵ͺͷ previously described (Palubinsky et al., 2015). ͵͵ͻ (HE-Pr; Brain Bits) at 4°C, while tails were ͵ͺ͸ Primary antibody was diluted in 1% BSA overnight ͵ͶͲ processed for genotyping. Once PCR was ͵ͺ͹ at 4°C. Cells were washed with 1x PBS for a total of ͵Ͷͳ concluded, WT and KO brains were pooled by ͵ͺͺ 30 minutes and incubated in Cy-2, Cy-3, or Alexa- ͵Ͷʹ genotype and dissection continued with cortical ͵ͺͻ Fluor 350 secondary antibodies (1:500) for 60 ͵Ͷ͵ tissue digestion in 0.025% trypsin for 20 minutes, ͵ͻͲ minutes at RT. Cells were counterstained with 4,6- ͵ͶͶ followed by mechanical dissociation. Resultant cell ͵ͻͳ diamidino-2-phenylindole (DAPI) to observe ͵Ͷͷ suspensions were plated at 600,000 cells/ml. ͵ͻʹ nuclei. Coverslips were mounted on glass slides ͵Ͷ͸ Cultures were maintained at 37°C, 5% CO2 in ͵ͻ͵ using ProLong Gold (Fisher; P3693), and ͵Ͷ͹ growth media comprising a volume-to-volume ͵ͻͶ fluorescence was observed with a Zeiss Axioplan ͵Ͷͺ mixture of 84% (v/v) DMEM, 8% (v/v) Ham’s F12- ͵ͻͷ microscope at 63x or 40x, where indicated. ͵Ͷͻ nutrients, 8% (v/v) fetal bovine serum, 24 U/ml ͵ͻ͸ ͵ͷͲ penicillin, 24 μg/ml streptomycin, and 80 μM L- ͵ͻ͹ Cell Counting and Colocalization of Immunofluorescent ͵ͷͳ glutamine. Before plating, 2x N2 and 2% NS21 (v/v) ͵ͻͺ Primary Neurons ͵ͷʹ supplements were added to growth media. Glial ͵ͻͻ Cell counts of primary neuronal cultures ͵ͷ͵ proliferation was inhibited after two days in ͶͲͲ prepared for immunofluorescence staining of ͵ͷͶ culture via the addition of 2 μM cytosine ͶͲͳ PINK1 and MAP2, and of CHIP and MAP2, were ͵ͷͷ arabinoside, after which cultures were maintained ͶͲʹ carried out using ImageJ FIJI version 2.0.0-rc- ͵ͷ͸ in Neurobasal medium containing 2% B27 (v/v), 2x ͶͲ͵ 43/1.50e (RRID:SCR_002285). Cell counts were ͵ͷ͹ N2, and 2% NS21 (v/v) supplements with ͶͲͶ performed by two blinded investigators. Statistical ͵ͷͺ antibiotics for two weeks. One week before ͶͲͷ significance was determined by two-tailed ͵ͷͻ experiments, neurons were maintained in ͶͲ͸ unpaired t-test with p<0.05 using GraphPad Prism ͵͸Ͳ Neurobasal medium containing 2% (v/v) NS21 and ͶͲ͹ software. ͵͸ͳ antibiotics only. All experiments were conducted ͶͲͺ Colocalization analysis of mouse primary ͵͸ʹ 20–25 days following dissociation. ͶͲͻ neurons stained with CHIP and TOM20 were ͵͸͵ ͶͳͲ performed using Coloc2 in FIJI. Twenty to 30 fields ͵͸Ͷ CHIP Plasmid Design and Transfection Ͷͳͳ per condition were chosen at random and ͵͸ͷ Plasmids were generated by and purchased Ͷͳʹ measured by blinded investigators. Background ͵͸͸ from ProNovus Bioscience, LLC. Mammalian Ͷͳ͵ was accounted for by using a Rolling-Ball ͵͸͹ expression vectors for hCHIP, hCHIP K30A and ͶͳͶ Background Subtraction of 50. Neurons were ͵͸ͺ hCHIP H260Q are based on pcDNA3.1 (Qian et al., Ͷͳͷ outlined using the polygon drawing tool. Point ͵͸ͻ 2006). The human CHIP promoter sequence was Ͷͳ͸ Spread Function (PSF) was calculated and set to ͵͹Ͳ used in order to induce CHIP expression by Ͷͳ͹ 1.0, and Costes Randomizations was set to 10. ͵͹ͳ endogenous transcription factors. For detection by Ͷͳͺ ͵͹ʹ immunofluorescence, a mCherry expression Ͷͳͻ Generation of WT and CHIP KO Mouse Embryonic ͵͹͵ sequence was added to the 3’ end of the CHIP ͶʹͲ Fibroblast (MEF) Cultures ͵͹Ͷ sequence using a 3xGGGGS linker. Using the Ͷʹͳ Fibroblast cell lines were developed based on ͵͹ͷ LipoJet Transfection Kit, plasmid (5 μg) was added Ͷʹʹ previous protocols (Xu, 2005). Briefly, embryonic ͵͹͸ dropwise into each well of a 6-well dish containing Ͷʹ͵ day 13 mouse pups generated by heterozygous ͵͹͹ neurons in 2 ml of growth medium. Six hours after ͶʹͶ matings were decapitated and internal organs were ͵͹ͺ transfection cells underwent a complete medium Ͷʹͷ removed. Tail samples were taken and all ͵͹ͻ change, and experiments were carried out the Ͷʹ͸ remaining epidermal tissue was minced and placed ͵ͺͲ following day. Ͷʹ͹ in individual conical tubes containing 2 mL of ͵ͺͳ Ͷʹͺ 0.25% trypsin-EDTA and digested at 4°C for 18 ͷ

Ͷʹͻ hours followed by 30 minutes at 37°C. Meanwhile Ͷ͹͸ through a [3:1] (resin: PO) exchange for 3-4 hours, Ͷ͵Ͳ tails were processed for genotyping (Palubinsky et Ͷ͹͹ then were incubated with pure epoxy resin Ͷ͵ͳ al., 2015). Once PCR and digestion was concluded, Ͷ͹ͺ overnight. Samples were then incubated in two Ͷ͵ʹ WT and KO tissue was pooled, an equal volume of Ͷ͹ͻ more changes of pure epoxy resin and allowed to Ͷ͵͵ MEF media (84% (v/v) DMEM, 10% (v/v) fetal ͶͺͲ polymerize at 60qC for 48 hours. Thin sections (70- Ͷ͵Ͷ bovine serum, 24 U/ml penicillin, and 24 mg/ml Ͷͺͳ 80 nm) were cut using an ultramicrotome and Ͷ͵ͷ streptomycin) was added and the tissue was Ͷͺʹ mounted onto copper grids. The sections were Ͷ͵͸ mechanically dissociated using a 5 mL pipette, Ͷͺ͵ stained with 2% uranyl acetate and Reynold’s lead Ͷ͵͹ triturating ~20 times. Any remaining tissue was ͶͺͶ citrate before imaging at 26,000x and 67,000x on an Ͷ͵ͺ allowed to settle while the supernatant was Ͷͺͷ electron microscope (Philips T12 equipped with an Ͷ͵ͻ transferred to a new tube. Five mL of MEF media Ͷͺ͸ AMT CCD camera system, FEI). ͶͶͲ was added to the tissue pellet, trituration was Ͷͺ͹ ͶͶͳ repeated and supernatant was collected and added Ͷͺͺ MitoTracker Labeling and Immunofluorescence of MEFs ͶͶʹ to the initial cell suspension. Trituration and Ͷͺͻ WT and CHIP KO MEFs were grown to 60-80% ͶͶ͵ transfer were repeated a final time and the entire ͶͻͲ confluency on glass coverslips. MitoTracker ͶͶͶ resultant cell suspension was plated in a T75 flask Ͷͻͳ Orange staining was performed as previously ͶͶͷ and maintained at 37°C, and 5% CO2 until cells Ͷͻʹ described (Palubinsky et al., 2015). Briefly, ͶͶ͸ reached 85% confluency at which point they were Ͷͻ͵ MitoTracker (Fisher; M7511) was added to live cell ͶͶ͹ passaged. Immortalization was achieved via ͶͻͶ cultures at a final concentration of 790 nM and ͶͶͺ repeated passaging (roughly 25 passages). Ͷͻͷ incubated at 37°C for 45 minutes. Coverslips were ͶͶͻ Following immortalization, all experiments were Ͷͻ͸ washed with 1x PBS, fixed with 4% (v/v) PFA, ͶͷͲ carried out using MEFs between passages 10-20. Ͷͻ͹ permeabilized with 0.1% Triton X-100, washed Ͷͷͳ Ͷͻͺ again with 1x PBS, and then blocked with 8% (w/v) Ͷͷʹ Transmission Electron Microscopy of MEFs Ͷͻͻ BSA diluted in 1x PBS. After 25 minutes in BSA, Ͷͷ͵ WT and CHIP KO MEFs were grown to 60-80% ͷͲͲ coverslips were incubated overnight at 4°C in ͶͷͶ confluency and fixed in 2.5% glutaraldehyde with ͷͲͳ primary antibody against β-tubulin diluted in 1% Ͷͷͷ 0.1 M cacodylate buffer, pH 7.4 at RT then ͷͲʹ (w/v) BSA. Following primary antibody incubation, Ͷͷ͸ transferred to 4°C overnight. Samples were further ͷͲ͵ cells were washed with 1x PBS for a total of 30 Ͷͷ͹ processed in the Vanderbilt University Medical ͷͲͶ minutes and incubated in Alexa 350 secondary Ͷͷͺ Center Cell Imaging Shared Resource. Cells were ͷͲͷ antibody in 1% BSA for one hour. Cells were next Ͷͷͻ washed, scraped, pelleted, and underwent several ͷͲ͸ washed for a total of 30 minutes in 1x PBS and Ͷ͸Ͳ exchanges of fixative and dehydrating solvents. ͷͲ͹ coverslips were mounted using Prolong Gold. Ͷ͸ͳ Samples were fixed in 2.5% gluteraldehyde in 0.1 ͷͲͺ Using ImageJ FIJI, a total of 120 WT MEF cells Ͷ͸ʹ M cacodylate buffer, pH 7.4 at RT for 1 hour then ͷͲͻ and 169 CHIP KO MEF cells from two individual Ͷ͸͵ transferred to 4ºC overnight. The samples were ͷͳͲ cultures were analyzed. Images were thresholded Ͷ͸Ͷ washed in 0.1 M cacodylate buffer, then incubated ͷͳͳ using an adaptive algorithm and particles were Ͷ͸ͷ 1 hour in 1% osmium tetraoxide at RT then washed ͷͳʹ analyzed for total count, area, and perimeter. Area Ͷ͸͸ with 0.1 M cacodylate buffer. Subsequently, the ͷͳ͵ and perimeter values were then used to calculate Ͷ͸͹ samples were dehydrated through a graded ͷͳͶ mitochondrial circularity (circularity = Ͷ͸ͺ ethanol series and then three exchanges of 100% ͷͳͷ 4π(area/perimeter2), where values closer to 0 are less Ͷ͸ͻ ethanol. Next, the samples were incubated for 5- ͷͳ͸ circular (i.e. ellipses, crescents, and complex Ͷ͹Ͳ minutes in 100% ethanol and propylene oxide (PO) ͷͳ͹ tubular structures) and values close to 1 are Ͷ͹ͳ followed by 2 exchanges of pure PO. Samples were ͷͳͺ perfectly circular. Ͷ͹ʹ then infiltrated with 25% Epon 812 resin and 75% ͷͳͻ Ͷ͹͵ PO for 30 minutes at RT. Next, they were infiltrated ͷʹͲ Proteomic Analysis of WT and CHIP KO Brains Ͷ͹Ͷ with Epon 812 resin and PO [1:1] for 1 hour at RT ͷʹͳ Tissue Preparation and Precipitation Ͷ͹ͷ then overnight at RT. The next day, samples went ͸

ͷʹʹ Whole brains from postnatal day 35 WT mice (4 ͷ͸ͻ pieces were dehydrated in 100% acetonitrile. Next, ͷʹ͵ male, 4 female) and CHIP KO mice (4 male, 4 ͷ͹Ͳ samples were rehydrated in 200 μl of 25 ͷʹͶ female) were removed and immediately placed into ͷ͹ͳ mM AmBic containing 300 ng of trypsin gold ͷʹͷ a glass dounce containing 1 ml of ice-cold TNEB ͷ͹ʹ (Promega) and incubated at 37oC overnight. ͷʹ͸ lysis buffer. Brains were homogenized on ice (35 ͷ͹͵ Following trypsinization, peptides were extracted ͷʹ͹ strokes), sonicated at 6 watts for 10 seconds and ͷ͹Ͷ from the gel via 3, 20 minute washed in 200 μl of ͷʹͺ passed through a 40 micron cell strainer then ͷ͹ͷ 60% aqueous acetonitrile containing 1% formic acid ͷʹͻ through a 0.2 micron filter affixed to a 10 ml ͷ͹͸ and evaporated to dryness in vacuo. Lastly, ͷ͵Ͳ syringe. Protein assays were completed and all ͷ͹͹ peptides were re-suspended in 30% aqueous ͷ͵ͳ samples were adjusted to 2 mg in 1mL of TNEB. ͷ͹ͺ acetonitrile containing 0.1% formic acid and stored ͷ͵ʹ Samples were precipitated by adding 3 mL of ice- ͷ͹ͻ at -80oC (Ham, 2005). ͷ͵͵ cold ethanol, vortexing and incubating on ice for 3 ͷͺͲ ͷ͵Ͷ minutes. Following incubation, samples were spun ͷͺͳ Mass Spectrometry ͷ͵ͷ at 3000 x rpm (1819 g) for 10 minutes at 4°C. Upon ͷͺʹ Protein digests were lyophilized and re- ͷ͵͸ removal of the supernatant, pellets were incubated ͷͺ͵ suspended in water prior to solid phase extraction ͷ͵͹ in a 2:1 mixture of chloroform and ice-cold ͷͺͶ with a Waters Oasis HLB cartridge. Prior to use, ͷ͵ͺ methanol for 3 minutes and centrifuged at 3000 x ͷͺͷ cartridges were activated with 1 ml of acetonitrile ͷ͵ͻ rpm (1819 g) for 10 minutes. Following the spin, ͷͺ͸ and equilibrated with 2 ml of water. Peptides were ͷͶͲ pellets were washed 3 times in 1 ml of ice-cold ͷͺ͹ loaded, washed once with 1 ml water, and eluted ͷͶͳ methanol. After the third wash, pellets were re- ͷͺͺ with 70% acetonitrile containing 0.1% formic acid. ͷͶʹ suspended in 1 ml of 0.5% SDS and sonicated at 6 ͷͺͻ Peptides were evaporated to dryness in vacuo, re- ͷͶ͵ watts for 10 pulses. Following sonication, 10 μl of ͷͻͲ suspended in 10 mM triethylammonium ͷͶͶ sample was added to 8 μl of 4X sample buffer and ͷͻͳ bicarbonate (pH 8.0) and fractionated by bRPLC on ͷͶͷ 2 μl of 1 M DTT. This portion of the sample was ͷͻʹ an Agilent 1260 Infinity LC system equipped with ͷͶ͸ heat denatured at 95°C for 10 minutes while the ͷͻ͵ an XBridge C18 5 μm 4.6 x 250 mm column. Solvent ͷͶ͹ rest of the sample was used for protein assay. ͷͻͶ A was aqueous 10 mM triethylammonium ͷͶͺ ͷͻͷ bicarbonate at pH 7.4 and solvent B was 10 mM ͷͶͻ Peptide Preparation ͷͻ͸ triethylammonium bicarbonate in acetonitrile at a ͷͷͲ Peptides were prepared as previously described ͷͻ͹ flow rate of 0.5 mL/minute. ͷͷͳ (Ham, 2005). Briefly, equal protein amounts of each ͷͻͺ Peptides were further fractionated by a gradient ͷͷʹ sample were loaded onto a 10% Bis-Tris gel with ͷͻͻ in which solvent B was increased from 0% to 5% ͷͷ͵ empty lanes in between samples and separated at ͸ͲͲ from 0 to 10 min, 5% to 35% from 10 to 70 min, 35% ͷͷͶ 180V for 30 minutes. Simply Blue Safe Stain was ͸Ͳͳ to 70% from 70 to 85 min, held at 70% from 85 to 95 ͷͷͷ used to visualize all protein bands and allow for ͸Ͳʹ min, and reduced to 0% from 95 to 105 min. The ͷͷ͸ each sample to be cut horizontally into ~13 ͸Ͳ͵ eluted peptides were collected in 64 fractions, ͷͷ͹ fractions. Each horizontal fraction was then cut ͸ͲͶ which were concatenated to fifteen fractions as ͷͷͺ vertically into ~1 mm cubes and placed in a ͸Ͳͷ described (Wang et al., 2011b). Concatenated ͷͷͻ microcentrifuge tube. One hundred μl of 100 mM ͸Ͳ͸ fractions were evaporated to dryness in vacuo and ͷ͸Ͳ ammonium bicarbonate (AmBic) was added to ͸Ͳ͹ dried samples were re-suspended in 100 uL of 3% ͷ͸ͳ each fraction tube. ͸Ͳͺ acetonitrile with 0.1% formic acid for LC-MS/MS ͷ͸ʹ Samples were reduced with 5 mM DTT in ͸Ͳͻ analysis. ͷ͸͵ AmBic for 30 minutes at 60°C with shaking. ͸ͳͲ LC-MS/MS analyses were performed on a Q ͷ͸Ͷ Samples were then alkylated with 10 mM ͸ͳͳ Exactive Plus mass spectrometer (Thermo Fisher ͷ͸ͷ iodoacetamide in AmBic for 20 minutes in the dark ͸ͳʹ Scientific, Bremen, Germany) equipped with a ͷ͸͸ at RT. Any remaining Safe Stain dye was removed ͸ͳ͵ Proxeon nLC1000 LC (ThermoFisher Scientific) and ͷ͸͹ with additional 100 μl washes in 50 mM AmBic ͸ͳͶ a Nanoflex source (ThermoFisher Scientific). ͷ͸ͺ /acetonitrile (1:1, v/v). Once clear of all dye, gel ͸ͳͷ Peptides were resolved on an 11 cm long column ͹

ͳ͸ with a 75 um internal diameter (New Objective, ͸͸͵ ͳ͹ Woburn, MA, PF360–75-10-N-5) packed with 3 μm ͸͸Ͷ Results ͳͺ particle size and 120 Å pore size ReproSil-Pur C18- ͸͸ͷ ͳͻ AQ resin (Dr. Maisch GmbH, Ammerbuch- ͸͸͸ 1. Low-level Bioenergetic Stress Induces CHIP ʹͲ Entringen, Germany) over a 70 minute gradient at a ͸͸͹ Expression ʹͳ 300nL/min flow rate. The gradient was formed ͸͸ͺ In neuron enriched cultures, bioenergetic stress ʹʹ utilizing 0.1% formic acid in water (solvent A) and ͸͸ͻ induced by 15 minutes of OGD initiates a robust ʹ͵ 0.1% formic acid in acetonitrile (solvent B) whereby ͸͹Ͳ defensive program with sufficient magnitude to ʹͶ the percentage of B was varied as follows: 2% to 5% ͸͹ͳ significantly decrease the cell death associated with ʹͷ in 5 min, 5% to 35% over 55 min, 35% to 90% in 3 ͸͹ʹ a subsequent exposure to a normally lethal 90- ʹ͸ min, followed by 90% for 7 min. ͸͹͵ minute period of OGD (Figure 1A). Consistent with ʹ͹ A single MS1 scan from m/z 300–1800 at 70,000 ͸͹Ͷ prior reports, both bright-field photomicrographs ʹͺ resolution with an automatic gain control (AGC) ͸͹ͷ of live cells (Figure 1B) and LDH release assays ʹͻ value of 3e6 and max injection time of 64 msec was ͸͹͸ (Figure 1C) reveal that low-level bioenergetic stress ͵Ͳ recorded as profile data. A top-12 method was ͸͹͹ protects approximately 45% of neurons from ͵ͳ used, whereby the 12 most intense precursors were ͸͹ͺ subsequent injury (McLaughlin et al., 2003; Brown ͵ʹ automatically chosen for MS2 analysis and a ͸͹ͻ et al., 2010; Zeiger et al., 2010). Importantly, this ͵͵ dynamic exclusion window of 20 s was employed. ͸ͺͲ stressor does not result in neuronal cell death, as ͵Ͷ For each MS2 scan, a resolution of 17,500, an AGC ͸ͺͳ evident by phase-bright somas and intricate ͵ͷ value of 2e5, a max injection time of 100 msec, a 2.0 ͸ͺʹ processes that are indistinguishable from control ͵͸ m/z isolation window, and a normalized collision ͸ͺ͵ cells (Figure 1B). To examine the temporal and ͵͹ energy of 27 was used and centroid data were ͸ͺͶ spatial time-course of CHIP expression in cells ͵ͺ recorded. ͸ͺͷ primed with by low-level bioenergetic event, we ͵ͻ Thermo .raw datafiles from LC-MS/MS runs ͸ͺ͸ harvested neurons at various time points following ͶͲ were converted to mzml format using ͸ͺ͹ 15 minutes of OGD. CHIP expression begins to Ͷͳ Proteowizard version 3.0.5211 (Kessner et al., 2008). ͸ͺͺ increase within 1 hour (1.8-fold ±0.4 above control, Ͷʹ The mzml files were searched using MyriMatch ͸ͺͻ Table 1), peaks 6 hours after stress (2.8-fold ±0.8 Ͷ͵ version 2.1.132 (Tabb et al., 2007) and MS-GF+ ͸ͻͲ above control), and remains elevated at 24 hours ͶͶ version 9517 (Kim and Pevzner, 2014) against the ͸ͻͳ (1.7-fold ±0.3 above control; Figure 1D). Notably, Ͷͷ mouse Refseq database (October 16, 2014). A semi- ͸ͻʹ increases in HSP70 expression lag behind CHIP, Ͷ͸ tryptic search was employed with a maximum of ͸ͻ͵ peaking between 6 and 18 hours (2.5-fold ±0.2 Ͷ͹ four missed cleavages allowed. A fixed ͸ͻͶ above control; Figure 1D) and returning to baseline Ͷͺ carbimidomethyl modification on Cys, a variable ͸ͻͷ levels at 24 hours (0.9-fold of control ±0.4; Figure Ͷͻ oxidation on Met, and a pyro-glu on Gln were ͸ͻ͸ 1D). ͷͲ allowed with a maximum of 2 dynamic ͸ͻ͹ To determine the percentage of neurons that ͷͳ modifications per peptide. Precursor ions were ͸ͻͺ express CHIP in response to bioenergetic stress, we ͷʹ required to be within 15 ppm of expected values ͸ͻͻ stained neurons 24 hours following 15-minute ͷ͵ and fragment ions within 20 ppm. A target-decoy ͹ͲͲ OGD. While we observed an increase in total CHIP ͷͶ search was employed using a reverse sequence ͹Ͳͳ expression via Western blot, by ICC we observed ͷͷ database to allow calculation of FDR for peptide- ͹Ͳʹ that CHIP is expressed in 49% of mildly stressed ͷ͸ spectral matches. The final protein list was ͹Ͳ͵ neurons, which was not significantly different from ͷ͹ assembled following the rule of parsimony in ͹ͲͶ control cultures (Figure 1E). Given that CHIP ͷͺ IDPicker 3 version 3.1.643.0 (Ma et al., 2009) at a ͹Ͳͷ expression undergoes temporal and spatial changes ͷͻ protein FDR of <1%. This protein list was then ͹Ͳ͸ in response to stress (Figure 1C; (Stankowski et al., ͸Ͳ analyzed for spectra which were found to be 2 fold ͹Ͳ͹ 2011; Palubinsky et al., 2015), we suspect that using ͸ͳ or greater different between WT and KO CHIP ͹Ͳͺ this method failed to capture the dynamic turnover ͸ʹ mice. ͺ

Ͳͻ and subcellular localization of CHIP at a static ͹ͷ͸ membrane (Greene et al., 2012; Jin and Youle, ͳͲ timepoint of 24 hours post-OGD. ͹ͷ͹ 2013). The kinase activity of PINK1 recruits E3 ͳͳ In previous work, mitochondria isolated from ͹ͷͺ ligases, such as Parkin, as well as the autophagy ͳʹ CHIP KO animals exhibit dysfunction in response ͹ͷͻ protein LC3, which seeds formation of an ͳ͵ to calcium challenge, suggesting CHIP is important ͹͸Ͳ autophagosome around the mitochondrion ͳͶ for mitochondrial homeostasis during stress ͹͸ͳ (Kawajiri et al., 2010; Klionsky et al., 2016). ͳͷ (Palubinsky et al., 2015). To test this hypothesis, we ͹͸ʹ We hypothesized that stabilization of PINK1 and ͳ͸ used the mitochondrial import receptor, TOM20 to ͹͸͵ mitochondrial clearance may participate in ͳ͹ label mitochondria and co-stained for CHIP in cells ͹͸Ͷ neuroprotection by priming neurons for a ͳͺ exposed to bioenergetic stress. We observed ͹͸ͷ secondary stress. To test this hypothesis, whole cell ͳͻ increased overlap of CHIP with TOM20-positive ͹͸͸ lysate was collected 3, 6, and 24 hours following ʹͲ mitochondria as early as 3 hours after acute stress ͹͸͹ bioenergetic stress and full-length, stabilized ʹͳ (data not shown) that was maximal at 6 hours ͹͸ͺ PINK1 was found to rapidly increase and remain ʹʹ (Figure 1F, overlap of CHIP and TOM20 in white). ͹͸ͻ elevated for 24 hours (3.3-fold ±2.1 above control; ʹ͵ Taken together, these data demonstrate that CHIP ͹͹Ͳ Figure 2B, Table 2). ʹͶ localizes to mitochondria during the period in ͹͹ͳ To further elucidate the extent of ʹͷ which neurons are upregulating endogenous ͹͹ʹ bioenergetically-induced PINK1 stabilization in ʹ͸ protective pathways and HSP70. We next sought to ͹͹͵ neurons, we stained for PINK1 24 hours following ʹ͹ determine how mitochondrial quality control ͹͹Ͷ 15 minutes of OGD and found a significant increase ʹͺ signaling is impacted by low-level bioenergetic ͹͹ͷ in the total number of PINK1 positive cells ʹͻ stressors. ͹͹͸ compared to untreated cultures (28% increase; ͵Ͳ ͹͹͹ Figure 2C & D). The number of MAP2-positive ͵ͳ 2. Low-level Bioenergetic Stress Changes Mitochondrial ͹͹ͺ neurons does not change after PC, consistent with ͵ʹ Morphology and Increases Expression of the Mitophagy ͹͹ͻ previous data (Figure 1); however, we observe that ͵͵ Related Protein PINK1 ͹ͺͲ the underlying neuropil undergoes remodeling that ͵Ͷ Mitochondria are dynamic organelles that ͹ͺͳ is made apparent using immunofluorescence ͵ͷ undergo continuous fission and fusion events, as ͹ͺʹ techniques (Figure 2C). Given that PINK1 is a rapid ͵͸ well as mitophagy. All of these processes are ͹ͺ͵ sensor of mitochondrial dysfunction, this data is ͵͹ induced by various cell stressors, including mild ͹ͺͶ consistent with our prior observation that 15 ͵ͺ hypoxic stress (Archer, 2013). To determine the ͹ͺͷ minutes of OGD, while a mild bioenergetic stress ͵ͻ morphology of mitochondria during low-level ͹ͺ͸ resulting in no neuronal cell death (Figure 1B, 1C), ͶͲ bioenergetic stress, rat primary neurons were ͹ͺ͹ does cause substantive energetic and oxidative Ͷͳ exposed to 15 minutes of OGD then visualized with ͹ͺͺ stress (Stankowski et al., 2011). Ͷʹ TOM20 immunostaining 24 hours later. ͹ͺͻ Ͷ͵ We found that although control and mildly ͹ͻͲ 3. Autophagic Signaling Is Required for Achieving a ͶͶ stressed neurons exhibit healthy, structurally ͹ͻͳ Protective Response Ͷͷ similar processes and nuclei based on MAP2 and ͹ͻʹ Some E3 ligases, such as Parkin, play critical Ͷ͸ DAPi fluorescence, mitochondrial localization is ͹ͻ͵ roles in promoting mitophagy in response to stress, Ͷ͹ profoundly impacted by bioenergetics (Figure 2A). ͹ͻͶ but likely not in the normal clearance of these Ͷͺ Mitochondria are typically well dispersed within ͹ͻͷ organelles (Kubli et al., 2013). We therefore sought Ͷͻ the soma and along processes. However, low-level ͹ͻ͸ to determine the role of CHIP in both ͷͲ stress increased the number of fragmented ͹ͻ͹ pathophysiological and physiological mitophagy in ͷͳ mitochondria along neuronal processes as well as ͹ͻͺ neurons. ͷʹ clustering of mitochondria within the soma. ͹ͻͻ To determine the contribution of mitophagy, we ͷ͵ During cell stress, mitochondrial membrane ͺͲͲ used the autophagy inhibitor, bafilomycin A1 ͷͶ potential is compromised, allowing retention and ͺͲͳ (BafA) to prevent lysosome acidification during our ͷͷ accumulation of PINK1 on the outer mitochondrial ͺͲʹ mild bioenergetic stress (Klionsky et al., 2016). The

ͺͲ͵ next day, we treated rat primary neurons with 90 ͺͷͲ passages under normal growth conditions where ͺͲͶ minutes of OGD, which is typically lethal, and ͺͷͳ both cell lines underwent mitosis after ~30 hours. ͺͲͷ assessed cell death 24 hours later. Neurons treated ͺͷʹ tEM imaging of WT MEFs reveal large, round ͺͲ͸ with 1 nM BafA during the priming event exhibited ͺͷ͵ mitochondria with intact membranes and many ͺͲ͹ increased cell death following secondary stress ͺͷͶ thin cristae (Figure 4A-C). These cells contain large ͺͲͺ compared to those that only received the initial, ͺͷͷ lysosomes and the endoplasmic reticuli appear ͺͲͻ mild bioenergetic priming stress (p<0.05, n=4; ͺͷ͸ tubular throughout the slice. CHIP KO MEFs were ͺͳͲ Figure 3A). ͺͷ͹ appreciably different from WT cells in that the ͺͳͳ By measuring CHIP and HSP70 expression in ͺͷͺ majority of mitochondria had thick, swollen cristae ͺͳʹ this paradigm, we determined both proteins were ͺͷͻ and smaller matrices (Figure 4D-F) consistent with ͺͳ͵ highly expressed in primed neurons in the presence ͺ͸Ͳ mitochondrial stress and swelling even though the ͺͳͶ of BafA, with CHIP expression significantly ͺ͸ͳ cells proliferated normally. KO MEFs also ͺͳͷ increased by 24h (1.4-fold ±0.1 above control; ͺ͸ʹ contained fewer lysosomes and irregularly-shaped ͺͳ͸ Figure 3B, Table 3). PINK1, while increased after ͺ͸͵ endoplasmic reticuli. ͺͳ͹ OGD in the presence or absence of BafA, was not ͺ͸Ͷ To quantify the extent of mitochondrial ͺͳͺ statistically different. To determine autophagic ͺ͸ͷ morphological changes, we performed MitoTracker ͺͳͻ flux, we quantified Western blot band intensities of ͺ͸͸ Orange staining in WT and CHIP KO MEFs and ͺʹͲ LC3-I and LC3-II (Klionsky et al., 2016) and ͺ͸͹ prepared them for ICC with β-tubulin co-stain. ͺʹͳ observed that LC3-II accumulated in cultures ͺ͸ͺ Using ImageJ FIJI, we quantified mitochondrial ͺʹʹ preconditioned with BafA co-treatment, revealing ͺ͸ͻ number, area, and perimeter. Area and perimeter ͺʹ͵ impeded autophagic flux (LC3-II:LC3-I is 2.1-fold ͺ͹Ͳ values were then used to calculate mitochondrial ͺʹͶ ±0.3 above control; Figure 3B, Table 3). We also ͺ͹ͳ circularity, where values closer to 0 are less circular ͺʹͷ observed dramatically increased accumulation of ͺ͹ʹ (i.e. more tubular) and values close to 1 are ͺʹ͸ LC3- and CHIP-positive autophagosomes in ͺ͹͵ perfectly circular. ͺʹ͹ neurons following low-level OGD (Figure 3C: ͺ͹Ͷ CHIP KO MEFs contained significantly greater ͺʹͺ Middle). This effect was further augmented with ͺ͹ͷ numbers of mitochondria than WT (Figure 4G). ͺʹͻ BafA treatment (Figure 3C: Right). Taken together ͺ͹͸ Given the large amount of variability in CHIP KO ͺ͵Ͳ these data suggest that CHIP localization to ͺ͹͹ mitochondrial number (WT: mean= 640, s.e.m.= ͺ͵ͳ mitochondria coincides with mitophagy and that ͺ͹ͺ 33.43; KO: mean= 1307, s.e.m.= 63.96), we ͺ͵ʹ mitophagy is a critical component of the ͺ͹ͻ hypothesize that during the division phases of ͺ͵͵ neuroprotection afforded by preconditioning. ͺͺͲ mitosis, CHIP KO fibroblasts undergo bioenergetic, ͺ͵Ͷ ͺͺͳ mitophaghic and proteostatic pressures which ͺ͵ͷ 4. CHIP Deficiency Increases Mitochondrial Number ͺͺʹ result in highly variable mitochondria not observed ͺ͵͸ and Changes Mitochondrial Morphology ͺͺ͵ in normal fibroblasts. These data are consistent ͺ͵͹ Given the increased CHIP localization to ͺͺͶ with previous reports that mitochondrial mass is ͺ͵ͺ mitochondria during mild bioenergetic stress, we ͺͺͷ largely impacted by proliferation and increased ͺ͵ͻ sought to determine if CHIP also played a role in ͺͺ͸ oxidative stress (Lee et al., 2002). ͺͶͲ mitochondrial homeostasis during physiological ͺͺ͹ Mitochondrial shape is a function of cellular ͺͶͳ conditions. Mouse embryonic fibroblasts (MEFs) ͺͺͺ metabolic demand and mitosis (Mishra et al., 2015; ͺͶʹ provide higher resolution imaging of ͺͺͻ Chen and Chan, 2017). Highly proliferative cells ͺͶ͵ mitochondrial morphology than neurons, as MEFs ͺͻͲ contain more small, punctate mitochondria to ͺͶͶ are large, flat, and non-polarized. We hypothesize ͺͻͳ efficiently segregate them into daughter cells (Chen ͺͶͷ that changes in mitochondrial morphology caused ͺͻʹ and Chan, 2017). WT MEFs conform to this ͺͶ͸ by CHIP loss is pervasive and will manifest even in ͺͻ͵ phenotype and have higher frequencies of circular ͺͶ͹ mitotic, peripheral cell types. ͺͻͶ mitochondria than CHIP KO (Figure 4H). We ͺͶͺ WT and CHIP KO MEFs were prepared for ͺͻͷ observed that relative frequency distributions for ͺͶͻ transmission electron microscopy (tEM) after 15-20 ͺͻ͸ mitochondrial circularity are skewed in CHIP KO ͳͲ

ͺͻ͹ MEFs, with higher frequencies for more elongated ͻͶͶ comparable to lethal, 90 minute OGD, alone (Figure ͺͻͺ mitochondria (circularity <0.3) than in WT (Figure ͻͶͷ 5B). ͺͻͻ 4H). This finding of increased non-circular ͻͶ͸ ͻͲͲ mitochondria in KO cells, taken together with our ͻͶ͹ 6. Mitochondrial Localization of CHIP Is Independent of ͻͲͳ tEM results, suggests a mitochondrial ͻͶͺ its E3 Ligase Activity and HSP70 Binding ͻͲʹ ultrastructure that may vary in proliferative and ͻͶͻ To eliminate changes attributed to cell passage ͻͲ͵ non-proliferative CHIP-deficient cells, but ͻͷͲ and mitosis as observed in MEFs, we next used a ͻͲͶ invariably is associated with a profoundly complex ͻͷͳ superior physiological model of acute neuronal ͻͲͷ structure. As EM analysis depends on 2D ͻͷʹ stress, subjecting post-mitotic CHIP KO primary ͻͲ͸ transverse sectioning of cells, branching and length ͻͷ͵ neuronal cultures to a mild bioenergetic stress. We ͻͲ͹ are not captured using this modality. ͻͷͶ predicted that CHIP KO neurons would exhibit ͻͲͺ ͻͷͷ severe deficits in mitochondrial quality and that, ͻͲͻ 5. CHIP Plays a Critical Role in Acute Neuroprotection ͻͷ͸ even when CHIP was reintroduced into KO cells, it ͻͳͲ Given that CHIP plays a critical role in ͻͷ͹ would colocalize to mitochondria in control cells, ͻͳͳ maintaining normal mitochondrial morphology ͻͷͺ given that they had lacked CHIP prior to ͻͳʹ and its expression increases in response to acute ͻͷͻ transfection. ͻͳ͵ stress (Palubinsky et al., 2015), we hypothesized ͻ͸Ͳ To test these hypotheses, we exposed WT and ͻͳͶ that acute CHIP induction is necessary for ͻ͸ͳ CHIP KO neurons to 15 minutes of OGD and ͻͳͷ neuroprotection. Since chronic loss of CHIP causes ͻ͸ʹ assessed mitophagic processes using ͻͳ͸ profound changes in the proteome and intracellular ͻ͸͵ immunofluorescence staining for TOM20 as a ͻͳ͹ milieu (Min et al., 2008; Palubinsky et al., 2015), we ͻ͸Ͷ mitochondrial marker and LC3 as a signal for ͻͳͺ chose to acutely deplete rat primary neurons of ͻ͸ͷ autophagosome formation 6 hours after the ͻͳͻ endogenous CHIP using siRNA to test this ͻ͸͸ bioenergetic stress. WT neurons exposed to acute ͻʹͲ hypothesis. We achieved approximately 67.5% ͻ͸͹ stress exhibited LC3-positive autophagosomes ͻʹͳ knockdown of CHIP 24 hours post-transfection ͻ͸ͺ throughout the cells that contain both mitochondria ͻʹʹ compared to control (p=0.0258, n=4; Figure 5A). By ͻ͸ͻ and CHIP (Figure 6A). In contrast, KO neurons ͻʹ͵ 48 hours post transfection, CHIP expression was ͻ͹Ͳ exhibit much more pervasive somatic ͻʹͶ not significantly different from control (p=0.661, ͻ͹ͳ redistribution of mitochondria following the ͻʹͷ n=4; Figure 5A). Importantly, CHIP siRNA caused ͻ͹ʹ stressor compared to WT controls (Figure 6A). In ͻʹ͸ no changes in cellular morphology or cell viability ͻ͹͵ CHIP-deficient cells, we also observed elongated ͻʹ͹ (Figure 5B) compared to neurons that were treated ͻ͹Ͷ mitochondria that colocalize with LC3 after OGD, ͻʹͺ with vehicle. ͻ͹ͷ with fewer rounded autophagosome structures. ͻʹͻ To test the effect of CHIP deficiency during ͻ͹͸ Taken together with Figure 4, these data suggest ͻ͵Ͳ neuroprotective priming, we transfected neurons ͻ͹͹ that CHIP is necessary for maintaining ͻ͵ͳ with CHIP siRNA followed 24 hours later (during ͻ͹ͺ mitochondrial morphology and suggest a role for ͻ͵ʹ optimal CHIP knockdown) by a mild bioenergetic ͻ͹ͻ CHIP in the mitophagic removal of dysfunctional ͻ͵͵ stress. CHIP depletion did not impact viability after ͻͺͲ organelles. ͻ͵Ͷ low-level (15 minute) OGD compared to non- ͻͺͳ Quantification of colocalization was performed ͻ͵ͷ transfected neurons (Figure 5B). We next sought to ͻͺʹ using FIJI Coloc2 on immunofluorescence images ͻ͵͸ determine the effect of CHIP depletion on ͻͺ͵ of WT neurons expressing TOM20 and endogenous ͻ͵͹ neuroprotection against a lethal stressor. We ͻͺͶ CHIP. We found that some CHIP:TOM20 ͻ͵ͺ subjected CHIP-depleted cultures to low-level ͻͺͷ colocalization is maintained at basal levels (Figure ͻ͵ͻ OGD followed by 90-minute OGD 24 hours later. ͻͺ͸ 6B). After mild bioenergetic stress, we find that ͻͶͲ Although CHIP expression was uninhibited during ͻͺ͹ CHIP:TOM20 colocalization does not change ͻͶͳ 90 minutes of OGD, cells that were depleted of ͻͺͺ relative to control levels. We suspect that ͻͶʹ CHIP during the priming event were not afforded ͻͺͻ mitophagy occurs rapidly after the initiation of ͻͶ͵ neuroprotection and exhibited levels of cell death ͻͻͲ OGD.

ͳͳ

ͻͻͳ As CHIP does not contain a canonical ͳͲ͵ͺ colocalized with mitochondria that were in close ͻͻʹ mitochondrial targeting sequence, we sought to ͳͲ͵ͻ proximity to LC3 labeling (Figure 7B). These data ͻͻ͵ determine the structural domains necessary for ͳͲͶͲ suggest that neither E3 ligase activity nor HSP ͻͻͶ CHIP localization to mitochondria. CHIP contains ͳͲͶͳ family binding is necessary for CHIP to localize to ͻͻͷ two primary functional domains: the N-terminal ͳͲͶʹ mitochondria. ͻͻ͸ Tetracopeptide Repeat (TPR) domain that binds to ͳͲͶ͵ Quantification of colocalization was performed ͻͻ͹ HSPs and other TPR domain-containing enzymes ͳͲͶͶ using FIJI Coloc2 on immunofluorescence images ͻͻͺ and the C-terminal Ubox domain required for E3 ͳͲͶͷ of KO neurons expressing TOM20 and transfected ͻͻͻ ligase activity (Ballinger et al., 1999; Jiang et al., ͳͲͶ͸ hCHIP in KO neurons. When neurons from CHIP ͳͲͲͲ 2001). Mutations identified in patients with ͳͲͶ͹ KO animals were transfected with either K30A ͳͲͲͳ spinocerebellar ataxia are located in multiple ͳͲͶͺ mutant (Figure 7B) or H260Q mutant CHIP (Figure ͳͲͲʹ domains of CHIP, leading to loss of either ͳͲͶͻ 7C), we observed a significant increase in ͳͲͲ͵ chaperone function or ubiquitin ligase activity ͳͲͷͲ CHIP:TOM20 colocalization after OGD. Similar to ͳͲͲͶ (Heimdal et al., 2014; Shi et al., 2014; Synofzik et al., ͳͲͷͳ WT neurons (Figure 6), we did not observe a ͳͲͲͷ 2014; Bettencourt et al., 2015). If CHIP acts in ways ͳͲͷʹ difference in colocalization between control and ͳͲͲ͸ similar to Parkin, the C-terminal E3 ligase domain ͳͲͷ͵ after bioenergetic stress in KO neurons transfected ͳͲͲ͹ would be necessary for mitophagy and H260Q ͳͲͷͶ with the non-mutated human CHIP plasmid ͳͲͲͺ mutants would accumulate LC3-positive organelles ͳͲͷͷ (Figure 7A). hCHIP does, however, localize to ͳͲͲͻ but fail to remove them. We also introduced a TPR ͳͲͷ͸ mitochondria even under control conditions, which ͳͲͳͲ K30A mutation and hypothesized that if ͳͲͷ͹ we hypothesize is due to the severe disruption of ͳͲͳͳ HSC/HSP70 binding is required for mitophagy, ͳͲͷͺ mitochondrial morphology, as seen in CHIP KO ͳͲͳʹ K30A transfected cells would have no ͳͲͷͻ MEFs and tEM (Figure 4), as well as the limited ͳͲͳ͵ colocalization of constructs with mitochondria ͳͲ͸Ͳ calcium buffering ability noted in isolated CHIP ͳͲͳͶ under baseline conditions or during bioenergetic ͳͲ͸ͳ KO mitochondria (Palubinsky et al., 2015). ͳͲͳͷ stress. ͳͲ͸ʹ Because we see no significant difference in ͳͲͳ͸ Primary forebrain neurons from CHIP KO mice ͳͲ͸͵ mitochondrial localization of endogenous (Figure ͳͲͳ͹ were cultured and transfected with plasmids ͳͲ͸Ͷ 6B) or non-mutated CHIP after OGD (Figure 7A), ͳͲͳͺ expressing non-mutated CHIP driven by a human ͳͲ͸ͷ we suspect that CHIP-mediated mitophagy may be ͳͲͳͻ promoter (hCHIP), Ubox mutant hCHIP (H260Q), ͳͲ͸͸ masked by the rapid turnover of organelles. To test ͳͲʹͲ or TPR mutant hCHIP (K30A) - each fused to a ͳͲ͸͹ this, KO neurons transfected with non-mutated ͳͲʹͳ mCherry expression sequence. Transfected neurons ͳͲ͸ͺ hCHIP were treated with BafA to block ͳͲʹʹ were then exposed to mild bioenergetic stress (15 ͳͲ͸ͻ autophagosome fusion with lysosomes following ͳͲʹ͵ minute OGD) and mCherry localization was ͳͲ͹Ͳ bioenergetic stress (Figure 7A). Indeed, hCHIP ͳͲʹͶ assessed after 6 hours of recovery via ICC for ͳͲ͹ͳ increases colocalization with mitochondria after ͳͲʹͷ TOM20 and LC3. We observed transfection of ~20% ͳͲ͹ʹ low-level stress when autophagy is blocked ͳͲʹ͸ of neurons. ͳͲ͹͵ compared to control and low-level OGD alone, ͳͲʹ͹ Under baseline conditions TOM20-positive ͳͲ͹Ͷ suggesting that CHIP-tagged mitochondria are ͳͲʹͺ mitochondria were well-dispersed, elongated, and ͳͲ͹ͷ directed for autophagy (Figure 7A). Taken ͳͲʹͻ within a network along processes. Six hours after a ͳͲ͹͸ together, these data suggest that functional Ubox ͳͲ͵Ͳ mild bioenergetic stress, staining for all inducible ͳͲ͹͹ and TPR domains are essential for clearance of ͳͲ͵ͳ constructs of CHIP under the human promoter was ͳͲ͹ͺ mitochondria after bioenergetic stress but not for ͳͲ͵ʹ evident (Figure 7A-C). Both human promoter-ͳͲ͹ͻ mitochondrial localization. ͳͲ͵͵ driven, non-mutated CHIP and the same construct ͳͲͺͲ ͳͲ͵Ͷ with a H260Q Ubox inactivation mutation ͳͲͺͳ 7. Loss of CHIP Changes the Expression of Critical ͳͲ͵ͷ colocalized with mitochondria following stress ͳͲͺʹ Bioenergetic Enzymes and Mitochondrial Quality ͳͲ͵͸ (Figure 7A, 7C). Interestingly, transfected CHIP ͳͲͺ͵ Control Proteins ͳͲ͵͹ with a K30A HSP70-interaction site mutation also ͳͲͺͶ Mitochondrial transport failure, altered ͳʹ

ͳͲͺͷ organelle dynamics and changes in structure and ͳͳ͵ʹ given our analyses these proteins are also unable to ͳͲͺ͸ function have been implicated in the primary ͳͳ͵͵ drive this process in the absence of CHIP. ͳͲͺ͹ pathology of an increasing number of ͳͳ͵Ͷ CHIP deficiency also resulted in a 2.3-fold ͳͲͺͺ neurodegenerative diseases (Morotz et al., 2012; ͳͳ͵ͷ increase in the expression of the mitochondrial ͳͲͺͻ Yan et al., 2013; Serrat et al., 2014; Zhang et al., ͳͳ͵͸ trafficking protein Miro (Rhot1). This Rho GTPase ͳͲͻͲ 2015). Proteomic analyses of whole-brains ͳͳ͵͹ is an outer mitochondrial membrane protein ͳͲͻͳ harvested from PND35 WT and CHIP KO mice ͳͳ͵ͺ involved in regulating mitochondrial morphology ͳͲͻʹ revealed a number of proteins essential for ͳͳ͵ͻ and intra- and intercellular trafficking. Miro is ͳͲͻ͵ mitochondrial trafficking and energetic ͳͳͶͲ essential for ensuring ATP availability for ͳͲͻͶ homeostasis that are significantly altered in CHIP ͳͳͶͳ energetically demanding neuronal signaling and ͳͲͻͷ KO animals compared to WT (Table 4). ͳͳͶʹ communication (Tang, 2015). Overexpression of ͳͲͻ͸ The greatest change observed was a 10-fold ͳͳͶ͵ Miro1 or Miro2 increases mitochondrial length ͳͲͻ͹ increase in the expression of ganglioside-induced ͳͳͶͶ (Saotome et al., 2008), a finding that is consistent ͳͲͻͺ differentiation associated protein (Gdap-1). Like ͳͳͶͷ with the elongated phenotype of these organelles in ͳͲͻͻ CHIP, Gdap-1 lacks a canonical mitochondrial ͳͳͶ͸ CHIP deficient cells. In addition to changing ͳͳͲͲ targeting sequence, but functions as an essential ͳͳͶ͹ mitochondrial shape, increased Miro expression ͳͳͲͳ mitochondrial outer membrane protein regulating ͳͳͶͺ may be causally linked to changes in CHIP ͳͳͲʹ mitochondrial structure and function (Niemann et ͳͳͶͻ expression. Phosphorylation of Miro by PINK1 ͳͳͲ͵ al., 2005; Otera and Mihara, 2011). Artificially ͳͳͷͲ (Wang et al., 2011a) targets the protein for ͳͳͲͶ increasing Gdap-1 expression in cell lines results in ͳͳͷͳ degradation by the E3 ubiquitin ligase Parkin (Liu ͳͳͲͷ mitochondrial hyper-fission without overt toxicity, ͳͳͷʹ et al., 2012; Birsa et al., 2014). Given the number of ͳͳͲ͸ while driving the expression of fusion proteins ͳͳͷ͵ overlapping substrates between CHIP and Parkin ͳͳͲ͹ such as mitofusin 1 can reverse this phenotype ͳͳͷͶ (Lizama et al., 2017), it is possible that CHIP ͳͳͲͺ (Niemann et al., 2005). The 10-fold increase in ͳͳͷͷ promotes Miro degradation as well. Taken together ͳͳͲͻ Gdap-1 observed in CHIP KO mice should drive ͳͳͷ͸ with increased Gdap-1, Dnm3, and Drp1, we ͳͳͳͲ mitochondrial fission and therefore manifest as ͳͳͷ͹ suspect that CHIP is a critical link between ͳͳͳͳ small circular mitochondria. Yet our imaging ͳͳͷͺ mitochondrial fission and subsequent mitophagy, ͳͳͳʹ analysis and quantification of mitochondria from ͳͳͷͻ although more studies are needed to confirm direct ͳͳͳ͵ CHIP KO MEFs demonstrate elongated and ͳͳ͸Ͳ protein-protein interactions between CHIP and ͳͳͳͶ misshapen organelles. This suggests that Gdap-1 ͳͳ͸ͳ these mitochondrial targets. ͳͳͳͷ upregulation, while significant, is unsuccessful at ͳͳ͸ʹ In addition to molecules related to fission-fusion ͳͳͳ͸ driving fission in a CHIP-deficient model. ͳͳ͸͵ dynamics, we also identified several mitochondrial ͳͳͳ͹ While less well understood in the context of ͳͳ͸Ͷ enzymes increased in brains from CHIP KO ͳͳͳͺ mitochondrial dynamics, the dynamin protein ͳͳ͸ͷ animals. Alcohol dehydrogenase (ADH), also ͳͳͳͻ family is essential for endocytosis, organelle fission ͳͳ͸͸ termed s-nitroso glutathione terminase, is elevated ͳͳʹͲ and fusion and vesicle formation (Antonny et al., ͳͳ͸͹ 3-fold in CHIP deficient animals. ADH eliminates ͳͳʹͳ 2016). Dynamin 1 and 3 are CNS-specific and are ͳͳ͸ͺ neurotoxic aldehydes generated by lipid ͳͳʹʹ highly expressed in neurons (Romeu and Arola, ͳͳ͸ͻ peroxidation (O'Brien et al., 2005), consistent with ͳͳʹ͵ 2014). We observed a 7.5-fold increase in ͳͳ͹Ͳ our prior report that CHIP deficiency significantly ͳͳʹͶ expression levels of Dynamin 3 (Dnm3) in CHIP ͳͳ͹ͳ increases CNS F2t-Isoprostanes and ͳͳʹͷ KO brains. These data are consistent with our ͳͳ͹ʹ Neuroprostanes (Palubinsky et al., 2015). These ͳͳʹ͸ previous report of increased levels of Dynamin-ͳͳ͹͵ species are derived from the non-enzymatic ͳͳʹ͹ related protein 1 (Drp1) in CHIP KO mice (Lizama ͳͳ͹Ͷ reduction of the membrane lipids arachidonic acid ͳͳʹͺ et al., 2017) as well as a gene dose-dependent ͳͳ͹ͷ and docosahexaenoic acid. In addition to ͳͳʹͻ change in Drp1 oxidation (Palubinsky et al., 2015). ͳͳ͹͸ detoxifying aldehydes, ADH can bind A-beta ͳͳ͵Ͳ Like Gdap-1, increased Dnm3 and Drp1 should ͳͳ͹͹ peptide, forming a mitotoxic species that decreases ͳͳ͵ͳ manifest as significant mitochondrial fission, but ͳͳ͹ͺ Complex IV activity, oxygen and neural glucose ͳ͵

ͳͳ͹ͻ utilization and ATP formation, as well as ͳʹʹ͸ to mitochondria in neural cells after bioenergetic ͳͳͺͲ production of ROS (Yao et al., 2011). ͳʹʹ͹ stress and is secreted by injured neurons (Kaneko ͳͳͺͳ The Krebs cycle enzyme malate dehydrogenase ͳʹʹͺ et al., 2014). Loss of DJ-1 increases susceptibility to ͳͳͺʹ (MDH) is increased 2.2-fold in brains from CHIP ͳʹʹͻ cell death following acute ischemia-reperfusion in ͳͳͺ͵ KO animals. MDH is the final acceptor in the Krebs ͳʹ͵Ͳ mice (Dongworth et al., 2014). As such, novel DJ-1- ͳͳͺͶ cycle and is increased in neural cell culture models ͳʹ͵ͳ binding compounds are currently being studied in ͳͳͺͷ exposed to oxidative stress (Bubber et al., 2005; Shi ͳʹ͵ʹ the context of Parkinson’s disease therapy (Inden et ͳͳͺ͸ and Gibson, 2011). The mitochondrial-specific ͳʹ͵͵ al., 2017) and have the potential to be cross- ͳͳͺ͹ isoform, MDH2, interacts with the malate-aspartate ͳʹ͵Ͷ purposed as neuroprotective strategies against ͳͳͺͺ shuttle, which facilitates the transfer of reducing ͳʹ͵ͷ acute oxidative stress. These data support a model ͳͳͺͻ equivalents from the cytosol to the mitochondria ͳʹ͵͸ in which CHIP is a critical regulator of redox ͳͳͻͲ for oxidation (Musrati et al., 1998). Pharmacological ͳʹ͵͹ homeostasis as DJ-1 decreases significantly in its ͳͳͻͳ activators of MDH and the malate-aspartate shuttle ͳʹ͵ͺ absence, refortifying an environment of protein and ͳͳͻʹ are protective in cardiac ischemia-reperfusion ͳʹ͵ͻ lipid oxidation. ͳͳͻ͵ injury models (Stottrup et al., 2010), suggesting that ͳʹͶͲ Taken together, these data reveal that CHIP ͳͳͻͶ this pathway is rate-limiting and increased ͳʹͶͳ expression is critical to mitochondrial quality ͳͳͻͷ expression and activity may be part of an effort to ͳʹͶʹ control proteins, bioenergetic enzymes, and redox ͳͳͻ͸ adapt to the bioenergetic dysfunction caused by ͳʹͶ͵ sensors. These results contribute to our ͳͳͻ͹ chronic loss of CHIP. ͳʹͶͶ understanding of the severe abnormalities in ͳͳͻͺ The observed decreases in Glutathione S-ͳʹͶͷ mitochondrial shape, cristae structure, and calcium ͳͳͻͻ Transferase (GST) speak to the intense oxidative ͳʹͶ͸ buffering capacity, as well as increased neuronal ͳʹͲͲ stress observed in CHIP deficient brain. ͳʹͶ͹ vulnerability to mild bioenergetic stress when ͳʹͲͳ Glutathione (GSH) is the most abundant cellular ͳʹͶͺ CHIP is absent that we observed. ͳʹͲʹ antioxidant, and maintaining a high ratio of ͳʹͶͻ ͳʹͲ͵ reduced to oxidized glutathione (GSH:GSSG) is ͳʹͷͲ ͳʹͲͶ essential for cellular health (Halliwell and ͳʹͷͳ Discussion ͳʹͲͷ Gutteridge, 1989). GST catalyzes the conjugation of ͳʹͷʹ The interactions between proteins involved in ͳʹͲ͸ GSH to endogenous and exogenous oxidized lipids ͳʹͷ͵ mitochondrial bioenergetics, redox tone, structure, ͳʹͲ͹ and proteins for detoxification (Strange et al., 2001). ͳʹͷͶ signaling, and autophagy are becoming ͳʹͲͺ CHIP KO animals have a 4.1-fold decrease in ͳʹͷͷ increasingly well understood, particularly in ͳʹͲͻ GSTA4 expression. This antioxidant is particularly ͳʹͷ͸ regards to mediating responses to neurological ͳʹͳͲ interesting in that it is dually located in the cytosol ͳʹͷ͹ injury and disease. Redox stress sensors like ͳʹͳͳ and mitochondria (Raza, 2011; Al Nimer et al., ͳʹͷͺ PINK1, DJ-1, and apoptosis signaling kinases play ͳʹͳʹ 2013). Within the mitochondria, GSTA4 detoxifies ͳʹͷͻ critical roles in sensing oxidative stress, triggering ͳʹͳ͵ cells from the lipid peroxidation product 4-ͳʹ͸Ͳ mitophagy and initiating fission and fusion of ͳʹͳͶ hydroxynoneal generated in response to oxidative ͳʹ͸ͳ organelles. The expression and distribution of these ͳʹͳͷ stress (Singhal et al., 2015). The decrease in GSTA4 ͳʹ͸ʹ proteins are regulated by chaperones, protein- ͳʹͳ͸ expression we identified via proteomics ͳʹ͸͵ protein interactions, redox sensitivity, ͳʹͳ͹ corroborates our previous findings of increased ͳʹ͸Ͷ ubiquitination, and degradation. In this work, we ͳʹͳͺ protein oxidation and lipid peroxidation in CHIP ͳʹ͸ͷ show that the chaperone-binding protein, CHIP, is ͳʹͳͻ deficient animals (Palubinsky et al., 2015). ͳʹ͸͸ unique among disease-associated proteins in that it ͳʹʹͲ Other aspects of cellular protein and lipid ͳʹ͸͹ is both an E3 ligase as well as an essential regulator ͳʹʹͳ oxidation profiles are also clearly impacted by ͳʹ͸ͺ of mitochondrial number and morphology in both ͳʹʹʹ CHIP deficiency, including a 2.8-fold decrease in ͳʹ͸ͻ physiological and acute pathophysiological stress ͳʹʹ͵ DJ-1 expression. DJ-1 is a redox sensor that ͳʹ͹Ͳ signaling. ͳʹʹͶ functions in repairing proteins damaged by ͳʹ͹ͳ CHIP was identified as a potential regulator of ͳʹʹͷ glycation during oxidative stress. DJ-1 translocates ͳʹ͹ʹ acute neuronal stress in studies from our lab in ͳͶ

ͳʹ͹͵ which we observed increases in CHIP and HSP70 ͳ͵ʹͲ absent, and CHIP KO cells contain more ͳʹ͹Ͷ expression that correlate with the time windows of ͳ͵ʹͳ mitochondria that are hyperfused. Increased ͳʹ͹ͷ neuroprotection observed in ischemic ͳ͵ʹʹ mitochondrial fusion occurs in cells with high ͳʹ͹͸ preconditioning (Stankowski et al., 2011). Cells ͳ͵ʹ͵ energetic demand and dependence on oxidative ͳʹ͹͹ given a low (non-lethal) level of stress are able to ͳ͵ʹͶ phosphorylation (Mishra et al., 2015). Given that ͳʹ͹ͺ upregulate endogenous protective pathways and ͳ͵ʹͷ CHIP KO cells have poor bioenergetics stores, ͳʹ͹ͻ withstand any number of subsequent injuries ͳ͵ʹ͸ increased oxidative stress, and elevated pro-fission ͳʹͺͲ including OGD, hypoxia, and oxidative stress ͳ͵ʹ͹ proteins as well as hyperfused mitochondria, we ͳʹͺͳ (Stetler et al., 2014). We sought to evaluate the ͳ͵ʹͺ conclude that CHIP is necessary for further ͳʹͺʹ importance of CHIP using a preconditioning ͳ͵ʹͻ processing of basal mitochondrial fission and ͳʹͺ͵ model, as understanding the proteins and ͳ͵͵Ͳ subsequent mitophagy. ͳʹͺͶ pathways that control this endogenous form of ͳ͵͵ͳ Taken together with biophysical data, it is ͳʹͺͷ protection could be leveraged therapeutically. ͳ͵͵ʹ perhaps not surprising that CHIP is unique among ͳʹͺ͸ Primary neuronal cultures exposed to a short ͳ͵͵͵ E3 ligases, as it forms asymmetric homodimers ͳʹͺ͹ period of OGD proved to be a highly effective ͳ͵͵Ͷ with protruding TPR domains. These TPR domains ͳʹͺͺ means to study CHIP redistribution. Here we show ͳ͵͵ͷ transiently bind to the C-terminal domain of ͳʹͺͻ that CHIP expression increases rapidly in neurons ͳ͵͵͸ HSC/P70. The interfacing Ubox domains then add ͳʹͻͲ in response to bioenergetic stress, functioning as a ͳ͵͵͹ poly-ubiquitin chains to exposed K48 residues on ͳʹͻͳ critical regulator of neuroprotection and ͳ͵͵ͺ client proteins (Scheufler et al., 2000; Schulman and ͳʹͻʹ mitochondrial autophagic clearance. Within six ͳ͵͵ͻ Chen, 2005; Stankiewicz et al., 2010). Not only does ͳʹͻ͵ hours of acute stress, CHIP relocalized to ͳ͵ͶͲ CHIP lack a canonical mitochondrial localization ͳʹͻͶ mitochondrial membranes. Silencing CHIP ͳ͵Ͷͳ signal, but protein relocalization continued to occur ͳʹͻͷ expression with siRNA significantly increased ͳ͵Ͷʹ even when we introduced mutations that arrest E3 ͳʹͻ͸ neuronal death in response to subsequent stress. ͳ͵Ͷ͵ ligase function (H260Q) or block HSP70-binding ͳʹͻ͹ These data suggest that CHIP induction, and ͳ͵ͶͶ (K30A). ͳʹͻͺ redistribution, are critical events in neurons in ͳ͵Ͷͷ Surprisingly, mutating sites controlling HSP70 ͳʹͻͻ order to fend off potentially toxic events. ͳ͵Ͷ͸ interactions or ubiquitin chain expansion enhanced ͳ͵ͲͲ We initially hypothesized that when ͳ͵Ͷ͹ CHIP association with mitochondria in response to ͳ͵Ͳͳ redistributed to the mitochondria, CHIP promotes ͳ͵Ͷͺ acute stress. One interpretation of this data is that ͳ͵Ͳʹ mitophagy in a manner akin to Parkin. Parkin is a ͳ͵Ͷͻ the HSP-bound/ubiquitin ligase functions of CHIP ͳ͵Ͳ͵ well-characterized E3-ubiquitin ligase with similar ͳ͵ͷͲ are predominantly cytosolic and that the mutations ͳ͵ͲͶ functional domains to CHIP and, based on cell free ͳ͵ͷͳ in H260Q or K30A increased the free pool of CHIP ͳ͵Ͳͷ assays, some redundant client proteins that are ͳ͵ͷʹ capable of moving to mitochondria. An alternative ͳ͵Ͳ͸ targeted for degradation (Imai et al., 2002; Kumar ͳ͵ͷ͵ interpretation is that HSP70 binding and ubiquitin ͳ͵Ͳ͹ et al., 2012). In vitro studies of Parkin revealed its ͳ͵ͷͶ ligase sites are dispensable for relocalization to ͳ͵Ͳͺ roles in mitochondrial quality control and ͳ͵ͷͷ mitochondria, but are required to drive ͳ͵Ͳͻ mitophagy (Grunewald et al., 2010; Matsuda et al., ͳ͵ͷ͸ downstream mitophagy signaling to degrade ͳ͵ͳͲ 2010; Narendra et al., 2010; Rana et al., 2013; Seirafi ͳ͵ͷ͹ CHIP/mitochondrial complexes, which would ͳ͵ͳͳ et al., 2015). ͳ͵ͷͺ appear as a decrease in CHIP/mitochondrial ͳ͵ͳʹ The ability of CHIP to compensate for Parkin ͳ͵ͷͻ association. ͳ͵ͳ͵ dysfunction was first reported over a decade ago ͳ͵͸Ͳ HSC/HSP70 C-terminal CHIP-binding mutants ͳ͵ͳͶ (Imai et al., 2002), and new studies in Drosophila ͳ͵͸ͳ increase cell vulnerability to stress-induced ͳ͵ͳͷ show that CHIP functions downstream of PINK1 ͳ͵͸ʹ apoptosis (Mosser et al., 2000), and CHIP rescues ͳ͵ͳ͸ signaling, with CHIP overexpression rescuing ͳ͵͸͵ mitochondrial dysfunction in a Ubox-dependent ͳ͵ͳ͹ mitophagy when Parkin is deficient (Chen et al., ͳ͵͸Ͷ manner in PINK1 KO drosophila models (Chen et ͳ͵ͳͺ 2017). In our current work, we reveal that pro-ͳ͵͸ͷ al., 2017). ͳ͵ͳͻ fission proteins are upregulated when CHIP is ͳ͵͸͸ In the absence of a canonical mitochondrial ͳͷ

ͳ͵͸͹ targeting sequence, our data suggest that the ͳͶͳͶ Goldberg et al., 2003; Kaneko et al., 2010). ͳ͵͸ͺ charged helix domains of CHIP may mediate ͳͶͳͷ However, recapitulating neuronal death, ͳ͵͸ͻ mitochondrial localization, or that the TPR and ͳͶͳ͸ particularly in Parkin knockout mouse models, has ͳ͵͹Ͳ Ubox domains may contain unidentified regions ͳͶͳ͹ largely proven unsuccessful (Perez and Palmiter, ͳ͵͹ͳ that bind with proteins or lipids that facilitate ͳͶͳͺ 2005). ͳ͵͹ʹ mitochondrial localization. Indeed, CHIP was ͳͶͳͻ Our data suggest that CHIP performs a more ͳ͵͹͵ recently shown to bind to lipid membranes ͳͶʹͲ vital role in neuronal homeostasis than Parkin. ͳ͵͹Ͷ enriched in phosphotidylinositol and phosphotidic ͳͶʹͳ Indeed, CHIP KO animals, neurons and MEFs have ͳ͵͹ͷ acid, and these lipid substrates compete with ͳͶʹʹ high baseline levels of oxidative stress, bioenergetic ͳ͵͹͸ chaperones for binding at the TPR domain (Kopp et ͳͶʹ͵ dysfunction and a hair-trigger response to injuries ͳ͵͹͹ al., 2017). Interestingly, however, Kopp et al. found ͳͶʹͶ that should be non-toxic. Mitochondria are severely ͳ͵͹ͺ that purified CHIP protein did not bind strongly to ͳͶʹͷ deformed with swollen cristae and are present in ͳ͵͹ͻ cardiolipin using cell-free lipid-binding assays. ͳͶʹ͸ greater numbers when CHIP has been genetically ͳ͵ͺͲ Cardiolipin, an inner mitochondrial membrane ͳͶʹ͹ deleted. This phenotype is unique compared to ͳ͵ͺͳ lipid, externalizes to the outer mitochondrial ͳͶʹͺ observations from other E3 ligase mutants (Stevens ͳ͵ͺʹ membrane in response to mitochondrial injury, ͳͶʹͻ et al., 2015; Mukherjee and Chakrabarti, 2016; Roy ͳ͵ͺ͵ interacts with LC3, and facilitates mitophagy (Chu ͳͶ͵Ͳ et al., 2016). CHIP deficiency is debilitating, as even ͳ͵ͺͶ et al., 2013). Based on our findings, further study of ͳͶ͵ͳ heterozygous mice have significant motor ͳ͵ͺͷ CHIP binding with mitochondrial lipids, such as ͳͶ͵ʹ dysfunction (McLaughlin et al., 2012). ͳ͵ͺ͸ cardiolipin, could prove useful to detect lipid-ͳͶ͵͵ Homozygous-null mice are too frail to participate ͳ͵ͺ͹ protein interactions that may underlie the ͳͶ͵Ͷ in behavioral testing and have significant oxidative ͳ͵ͺͺ mitochondrial morphology and dysfunction ͳͶ͵ͷ dysfunction in the CNS and periphery before dying ͳ͵ͺͻ associated with CHIP depletion. ͳͶ͵͸ prematurely (Min et al., 2008; McLaughlin et al., ͳ͵ͻͲ While H260Q and K30A mutations were ͳͶ͵͹ 2012; Palubinsky et al., 2015). These data stand in ͳ͵ͻͳ selected based on known functions of CHIP ͳͶ͵ͺ stark contrast to Parkin deficient models, which are ͳ͵ͻʹ primary domains, next-generation sequencing of ͳͶ͵ͻ neurochemically, behaviorally and structurally ͳ͵ͻ͵ patients with spinocerebellar ataxias has identified ͳͶͶͲ indistinguishable from littermates (Goldberg et al., ͳ͵ͻͶ numerous mutations in the STUB1 gene encoding ͳͶͶͳ 2003; Palacino et al., 2004; Perez and Palmiter, 2005; ͳ͵ͻͷ CHIP protein. In patients with SCAR16, ͳͶͶʹ Lizama et al., 2017). ͳ͵ͻ͸ homozygous or compound heterozygous point ͳͶͶ͵ The morphological and biochemical changes ͳ͵ͻ͹ mutations are spread throughout the major ͳͶͶͶ observed in CHIP deficient mitochondria do, ͳ͵ͻͺ structural domains of CHIP, resulting in a loss of ͳͶͶͷ however, strikingly parallel mutations in ͳ͵ͻͻ function (Heimdal et al., 2014; Shi et al., 2014; ͳͶͶ͸ autophagy protein 5 (Atg5) that disrupt formation ͳͶͲͲ Synofzik et al., 2014; Bettencourt et al., 2015; ͳͶͶ͹ of the Atg12-Atg5 complex. The Atg12-Atg5 ͳͶͲͳ Kawarai et al., 2016). STUB1 mutations are rare and ͳͶͶͺ conjugate exhibits E3 ligase–like activity, ͳͶͲʹ the clinical phenotype and neuropathological ͳͶͶͻ facilitating lipidation of LC3 family members ͳͶͲ͵ features of these mutations remain poorly ͳͶͷͲ (Otomo et al., 2013). While mice with complete ͳͶͲͶ described. CHIP likely controls critical functions ͳͶͷͳ deletion of Atg5 die shortly after birth (Yoshii et al., ͳͶͲͷ beyond the cerebellum, as patients present not only ͳͶͷʹ 2016), Atg5 mutations have been identified in ͳͶͲ͸ with progressive limb ataxia, gait instability, and ͳͶͷ͵ patients with ataxia and are associated with loss of ͳͶͲ͹ cerebellar atrophy but also with dysarthria and ͳͶͷͶ autophagic signaling in conditional KO models ͳͶͲͺ mild cognitive impairment (Heimdal et al., 2014; ͳͶͷͷ (Kim et al., 2016). Our data support an ‘Atg-5-like’ ͳͶͲͻ Kawarai et al., 2016). ͳͶͷ͸ role for CHIP in normal as well as pathological ͳͶͳͲ Genetic mutations in E3 ligases such as X-linked ͳͶͷ͹ neuronal mitophagy. As CHIP-null animals are not ͳͶͳͳ inhibition of apoptosis protein (XIAP), Hrd1, and ͳͶͷͺ embryonic lethal, CHIP deficient mice may also ͳͶͳʹ Parkin are associated with Parkinson’s disease and ͳͶͷͻ provide a rigorous animal model for studying ͳͶͳ͵ other neurological disorders (Harlin et al., 2001; ͳͶ͸Ͳ premature aging and susceptibility to neurological ͳ͸

ͳͶ͸ͳ injury. ͳͶ͹Ͷ transmission electron microscopy preparation and ͳͶ͸ʹ ͳͶ͹ͷ imaging. This work was supported by the Walter ͳͶ͸͵ Acknowledgements ͳͶ͹͸ and Suzanne Scott Foundation funding of the J.B. ͳͶ͸Ͷ The authors thank Dr. Cam Patterson for ͳͶ͹͹ Marshall Laboratory (B.M, A.M.P., B.N.L.), the Dan ͳͶ͸ͷ providing the original CHIP knockout animals; Ms. ͳͶ͹ͺ Marino Foundation (B.M., A.M.P.), NIH grants ͳͶ͸͸ Sharon Klein, Ms. Arulita Gupta, and Ms. ͳͶ͹ͻ NS050396 (B.M.) and RO1ES022936 (B.M., D.C.L.), ͳͶ͸͹ Dominique Szymkiewicz for technical assistance ͳͶͺͲ a Vanderbilt Brain Institute Scholarship (B.N.L., ͳͶ͸ͺ and ImageJ expertise; Dr. Joshua Fessel and Dr. ͳͶͺͳ A.M.P.), a VISRA Program Scholarship (V.R.), and ͳͶ͸ͻ Alice Soragni for electron microscopy expertise; Dr. ͳͶͺʹ predoctoral fellowships from the AHA ͳͶ͹Ͳ Christopher E. Wright and Mr. Jeff Duryea for ͳͶͺ͵ 15PRE25100000 (A.M.P.), 14PRE2003500007 ͳͶ͹ͳ imaging assistance; and Ms. Ama J. Winland for ͳͶͺͶ (B.N.L.). The authors declare no conflict of interest. ͳͶ͹ʹ cell culture maintenance. We also thank the ͳͶ͹͵ Vanderbilt Cell Imaging Shared Resource for Ͷͺͷ Ͷͺ͸ Ͷͺ͹ Ͷͺͺ Ͷͺͻ Figure Legends: ͶͻͲ Ͷͻͳ Ͷͻʹ Figure 1: Low-Level Bioenergetic Stress Induces CHIP Expression. (A) Primary rat forebrain cultures were Ͷͻ͵ exposed to 15’ OGD. Twenty-four hours later, neurons were exposed to 90’ OGD, following which cells were ͶͻͶ returned to growth media. Cell survival was assessed via LDH 24h later. (B) Representative photomicrographs Ͷͻͷ taken 24h after 15’ OGD (PC) reveal many phase-bright neuronal somas and intact processes that are Ͷͻ͸ indistinguishable from Control. Twenty-four hours after 90’ OGD, there were no phase-bright neurons, Ͷͻ͹ indicative of total neuronal death. Cultures that were preconditioned (PCÆ90’OGD) contained more phase- Ͷͻͺ bright neurons and intact processes compared to 90’ OGD. (C) LDH release was measured, and values were Ͷͻͻ normalized to 90’ OGD (100% cell death) in naive cells. Data were analyzed by one-way ANOVA with ͷͲͲ Bonferroni post-hoc testing (R2=0.7123, F=46.21, p<0.0001, n=4). *Denotes statistical significance (p<0.05), ͷͲͳ comparing 90' OGD to control cultures. Δ denotes statistical significance comparing PC➝90' and all other ͷͲʹ groups (p<0.05). (D) Neurons exposed to 15' OGD were harvested 1, 3, 6, 18, or 24h later. Whole cell lysates ͷͲ͵ were probed with antibodies against HSP70 and CHIP, as well as HSC70 and β-Tubulin as loading controls. ͷͲͶ Quantification is summarized in Table 1. (E) Twenty-four hours after PC, cultures were probed with antibodies ͷͲͷ to CHIP, MAP2, and DAPI. Cell counts were performed by blinded investigators. Basal CHIP expression is ͷͲ͸ observed in 49% of MAP2-positive neurons and is not significant from control post-PC (t=0.367, df=4, p=0.732, ͷͲ͹ n=3, Student’s t-test). (F) Preconditioned neurons were PFA-fixed after 6h of recovery and probed for CHIP ͷͲͺ (magenta), TOM20 (green), and DAPI (blue). ͷͲͻ ͷͳͲ ͷͳͳ ͷͳʹ Figure 2: Low-Level Bioenergetic Stress Results in Changes in Mitochondrial Morphology and Increased ͷͳ͵ Expression of the Mitophagy Related Protein PINK1. (A) Rat primary forebrain neurons were exposed to 15’ ͷͳͶ OGD (PC), probed with antibodies recognizing TOM20 (magenta) and MAP2 (green), and were counterstained ͷͳͷ with nuclei marker DAPI (blue). PC was associated with fragmented mitochondria (magenta) while neuronal ͷͳ͸ processes (green) remained intact, consistent with the preservation of cell viability. (B) Mitochondrial stress ͷͳ͹ was assessed by evaluating stabilized PINK1 3, 6, or 24 hours after PC. Whole cell lysates were probed with ͷͳͺ antibodies against PINK1 and the mitochondrial outer membrane protein TOM20, as well as HSC70 and β-

ͳ͹

ͷͳͻ Tubulin as loading controls. Quantification is summarized in Table 2. (C) Using immunofluorescence, PINK1 ͷʹͲ (magenta) and MAP2 (green) staining revealed high levels of somatic PINK1 staining around nuclei (blue) of ͷʹͳ preconditioned cells, consistent with the overall elevation in PINK1 signal in panel (B). (D) Cell counts of ͷʹʹ PINK1 positive neurons were performed by blinded investigators. PINK1 can be observed in 38% of MAP2- ͷʹ͵ positive neurons at baseline. After PC, the number of neurons expressing stabilized PINK1 significantly ͷʹͶ increases to 68% of neurons (t=4.6, df=4, p=0.01, n=3, Student’s t-test). ͷʹͷ ͷʹ͸ ͷʹ͹ ͷʹͺ Figure 3: Autophagic Signaling Is Required for Achieving a Protective Response. Neurons were exposed to 15' ͷʹͻ OGD in the presence or absence of the Bafilomycin A1 (BafA, 1 nM), a drug that blocks vacuolar ATPases, ͷ͵Ͳ inhibiting fusion between autophagosomes and lysosomes. (A) Survival was assessed by LDH release 24h after ͷ͵ͳ subsequent 90' OGD as described in Figure 1A. Blocking autophagy resulted in 85% cell death, a significant ͷ͵ʹ increase compared to PC neurons without BafA. Values were normalized to 90’ OGD. Data were analyzed by ͷ͵͵ one-way ANOVA with Bonferroni post-hoc testing. *denotes significance compared to preconditioned cells ͷ͵Ͷ (R2=0.9041, F=42.43, p<0.05, n=4). (B) Control neurons, neurons treated with BafA, preconditioned neurons (PC) ͷ͵ͷ and neurons preconditioned plus BafA (PC+BafA) were harvested for Western blotting 24 hours after ͷ͵͸ treatment. Whole cell lysate was probed for CHIP, HSP70, LC3 (LC3-I, 18kDa; LC3-II, 16 kDa), PINK1, p62, ͷ͵͹ TOM70, HSC70, and β-Tubulin. Accumulation of LC3-II and CHIP was evident in PC+BafA. Quantification is ͷ͵ͺ summarized in Table 1. (C) Preconditioned neurons were fixed 6h after PC in the presence or absence of BafA, ͷ͵ͻ and CHIP localization was evaluated by immunofluorescence (magenta). Autophagosomes were labeled with ͷͶͲ LC3 (green), and mitochondria labeled with TOM20 (blue). Images were taken at 63x; teal boxes were ͷͶͳ magnified to show regions of LC3 staining around CHIP-positive mitochondria. Overlap of the three labels ͷͶʹ results in white. ͷͶ͵ ͷͶͶ ͷͶͷ ͷͶ͸ Figure 4: CHIP Deficiency Results in Significantly Increased Mitochondrial Number with Abnormal ͷͶ͹ Morphological Features. Immortalized MEFs were cultured from WT and CHIP KO mice. (A,B,D,E) MEFs ͷͶͺ were grown to 60-70% confluency and fixed for tEM. Sections (70-80 nm) were imaged at 26000x (A, D; Scale ͷͶͻ bar, 500 nm) and 67000x (B, E; Scale bar, 200 nm). (A, B) WT MEFs have well-defined mitochondrial ͷͷͲ membranes and thin inner membranes forming cristae. (D, E) KO MEFs have smaller mitochondria containing ͷͷͳ unstructured cristae and thick inner membranes. (Endoplasmic Reticulum, E; Lysosome, L; Mitochondria, M). ͷͷʹ (C, F) MEFs were stained with MitoTracker Orange and PFA-fixed. Images used for measuring mitochondrial ͷͷ͵ Count, Area, and Perimeter were taken using a Zeiss Apotome (63x) and analyzed via ImageJ. Scale bar, 20 ͷͷͶ μm. (G) KO MEFs exhibit significant increases in MitoTracker Orange-positive mitochondria per cell (mean = ͷͷͷ 1307 mitochondria) compared to WT (mean = 641 mitochondria) (Mann-Whitney U= 4236, p<0.0001, WT n=120, ͷͷ͸ KO n=169; *denotes significance). (H) Circularity was calculated using Area and Perimeter measurements ͷͷ͹ where values closer to 0 indicate elongated or tubular mitochondria, and a value of 1 indicates a perfect circle. ͷͷͺ Values were multiplied by 100 in order to generate histograms. KO MEFs exhibit significantly increased ͷͷͻ numbers of elongated mitochondria (Mann-Whitney U= 5.66E+08, p<0.0001, WT n=36042, KO n=38244; ^ ͷ͸Ͳ denotes a shift in relative frequency compared to WT). Graphs represent data from duplicate coverslips per ͷ͸ͳ genotype from two independent cultures, with significance determined via Mann-Whitney analysis. ͷ͸ʹ ͷ͸͵ ͷ͸Ͷ ͷ͸ͷ

ͳͺ

ͷ͸͸ ͷ͸͹ Figure 5: CHIP Expression Is Required for Achieving a Protective Response. Neurons were treated with vehicle ͷ͸ͺ (C), transfected with GFP (G), or transfected with GFP and 20 nM of CHIP siRNA (si). Neurons were harvested ͷ͸ͻ either 24 or 48 hours later, and Western blotting was performed for CHIP using HSC70 and β-Tubulin as ͷ͹Ͳ loading controls. (A) Maximal silencing of CHIP expression (67.5% of Control; t=4.125, df= 3, p=0.0258, n=4, ͷ͹ͳ Student’s t-test) was achieved 24 hours post-transfection, and this time point was chosen to determine whether ͷ͹ʹ blocking CHIP attenuated PC-induced neuroprotection. CHIP expression by 48h was not significantly ͷ͹͵ different from control (t=0.484, df= 3, p=0.661, n=4, Student’s t-test). No significant differences in HSC70 ͷ͹Ͷ (R2=0.044, F=0.166, p=0.972, n=4, ANOVA) or β-Tubulin expression were observed (R2=0.189, F=0.791, p=0.571, ͷ͹ͷ n=4, ANOVA). (B) LDH released from dead or dying cells was assessed 24 hours after 90' OGD as described in ͷ͹͸ Figure 1. Raw absorbance values were normalized to values obtained from 90’ OGD, which leads to 100% cell ͷ͹͹ death. LDH release data revealed that CHIP-deficient neurons were far less likely to survive 90' OGD, even ͷ͹ͺ though they had been preconditioned (Mean difference compared to non-transfected = 48.08, SE of difference = ͷ͹ͻ 5.706, t = 8.425, df = 123). Data represent the mean from 3 independent experiments, with significance ͷͺͲ determined by two-way ANOVA and Bonferroni post-hoc analysis. One significant outlier was excluded by ͷͺͳ the Grubbs Outlier Test. Significant siRNA transfection-treatment interaction was observed: F(3,123) = 14.62, ͷͺʹ p<0.0001. No significant differences in viability were observed between non-transfected and CHIP siRNA- ͷͺ͵ transfected Control cultures (Mean difference = 8.595, SE of difference = 5.622, t = 1.529, df = 123) and after PC ͷͺͶ (Mean difference = 7.213, SE of difference = 5.622, t = 1.283, df = 123) or 90’ OGD (Mean difference = -0.006, SE ͷͺͷ of difference = 5.622, t = 0.001, df = 123). *denotes significance as compared to non-transfected cells with p<0.05. ͷͺ͸ ͷͺ͹ ͷͺͺ ͷͺͻ Figure 6: CHIP Localizes to Mitochondria At Baseline and After Bioenergetic Stress. (A) WT and CHIP KO ͷͻͲ primary neurons were treated with 15' OGD and fixed 6h later. Cultures were probed with antibodies ͷͻͳ targeting TOM20 (blue), CHIP (red), and LC3 (green). Scale bar, 20 μm. CHIP and TOM20 colocalization was ͷͻʹ analyzed using Coloc2 (FIJI). KO neurons exhibit elongated mitochondria that colocalize with LC3 after OGD, ͷͻ͵ with fewer rounded autophagosome structures compared to WT. (B) WT neurons exhibit a baseline ͷͻͶ colocalization of CHIP and mitochondria with an average Pearson Correlation Coefficient (PCC) of 0.15, which ͷͻͷ was not significantly different post-OGD (PCC=0.12; t=0.947, df= 65, p=0.347, Student’s t-test). ͷͻ͸ ͷͻ͹ ͷͻͺ ͷͻͻ Figure 7: CHIP is Necessary for Maintaining Mitochondrial Quality Control After Stress, Yet Mitochondrial ͸ͲͲ Localization of CHIP is Independent of the TPR and Ubox Domains. (A) CHIP KO neurons were transfected ͸Ͳͳ with CHIP (hCHIP) (A), K30A (B), or H260Q (C). Twenty-four hours after transfection, cultures were exposed ͸Ͳʹ to 15' OGD or 15’ OGD +BafA (1 nM) and PFA-fixed at 6h. Cultures were probed with antibodies targeting ͸Ͳ͵ TOM20 (blue) and LC3 (green). Red fluorescence is emitted from the mCherry tag linked to transfected CHIP. ͸ͲͶ Scale bar, 20 μm. (A) KO neurons transfected with non-mutated hCHIP (KO+hCHIP) exhibit baseline ͸Ͳͷ CHIP:TOM20 colocalization (PCC=0.26) that does not significantly differ post-OGD (PCC=0.21). Cotreatment ͸Ͳ͸ with BafA significantly increases CHIP:TOM20 colocalization post-OGD (PCC=0.38) compared to control and ͸Ͳ͹ OGD alone (R2=0.09, F=3.195, p=0.023, ANOVA). (B) KO neurons transfected with K30A (KO+K30A) ͸Ͳͺ demonstrate increased colocalization post-OGD (PCC=0.43) compared to control (PCC=0.20; t=5.0, df=41, ͸Ͳͻ p<0.0001, Student’s t-test). (C) KO neurons transfected with H260Q (KO+H260Q) show increased ͸ͳͲ colocalization post-OGD (PCC=0.42) compared to control (PCC=0.28; t=2.813, df=45, p=0.007, Student’s t-test). ͸ͳͳ *Denotes significance of p<0.01, and ** p<0.0001. ͸ͳʹ

ͳͻ

ͳ͵ ͳͶ Table 1: Quantification of CHIP and HSP70 After Mild Bioenergetic Stress. Neurons exposed to 15' OGD were ͳͷ harvested 1, 3, 6, 18, or 24h later. Whole cell lysates were probed with antibodies against HSP70 and CHIP, ͳ͸ with HSC70 and β-Tubulin as loading controls (Figure 1D). Bands were quantified by ImageJ FIJI densitometry ͳ͹ analysis. Values summarized in the table are the average band intensity relative to Control ± s.e.m. ͳͺ Significance was determined using one-way ANOVA with Bonferonni post-hoc analysis. Increased HSP70 ͳͻ expression following mild bioenergetic stress was statistically significant at 6h post-OGD (R2=0.520, F=3.897, ʹͲ p=0.014). CHIP increases early following OGD and remains increased above control levels, although the ʹͳ differences were not statistically significant (R2=0.305, F=1.577, p=0.217). No significant differences were ʹʹ observed in loading controls HSC70 (R2=0.052, F=0.296, p=0.960) and β-Tubulin (R2=0.1255, F=0.517, p=0.760). ʹ͵ ʹͶ ʹͷ Table 2: Quantification of PINK1 and TOM20 after Mild Bioenergetic Stress. Neurons exposed to 15' OGD were ʹ͸ harvested 3, 6, or 24h later. Whole cell lysates were probed with antibodies against PINK1, TOM20, with ʹ͹ HSC70 and β-Tubulin as loading controls (Figure 2B). Bands were quantified by ImageJ FIJI densitometry ʹͺ analysis. Values summarized in the table are the average band intensity relative to Control ± s.e.m. ʹͻ Significance was determined using one-way ANOVA with Bonferonni post-hoc analysis. PINK1 expression ͵Ͳ increases by 3h following OGD and remains increased above control levels, although the differences were not ͵ͳ statistically significant (R2=0.18, F=0.588, p=0.640). TOM20 expression did not change significantly following ͵ʹ OGD (R2=0.223, F=1.146, p=0.370). No significant differences were observed in loading controls HSC70 ͵͵ (R2=0.052, F=0.296, p=0.960) and β-Tubulin (R2=0.1255, F=0.517, p=0.760). ͵Ͷ ͵ͷ ͵͸ Table 3: Quantification of Proteins Following Autophagy Inhibition During Mild Bioenergetic Stress. Control ͵͹ neurons and neurons exposed to bafilomycin A (BafA, 1 nM), 15' OGD, or 15’ OGD +BafA were harvested 24h ͵ͺ post-treatment. Whole cell lysates were probed with antibodies against HSP70, CHIP, PINK1, TOM20, p62, ͵ͻ and LC3, with HSC70 and β-Tubulin as loading controls (Figure 3B). Bands were quantified using ImageJ FIJI ͶͲ using densitometry analysis. Values summarized in the table are the average band intensity relative to Control Ͷͳ ± s.e.m. Significance was determined using one-way ANOVA with Bonferonni post-hoc analysis. While HSP70, Ͷʹ PINK1, TOM20, and p62 expression increase following OGD and BafA treatment, they are not found to be Ͷ͵ statistically significant (HSP70: R2=0.354, F=2.193, p=0.142; PINK1: R2=0.308, F=1.778, p=0.205; TOM20: R2=0.18, ͶͶ F=0.587, p=0.641; p62: R2=0.217, F=0.738, p=0.559). The ratio of LC3-II to LC3-I increases following mild Ͷͷ bioenergetic stress and was statistically different when co-treated with BafA (R2=0.501, F=4.021, p=0.034). CHIP Ͷ͸ expression also increases after OGD with BafA co-treatment (R2=0.722, F=6.92, p=0.013). No significant Ͷ͹ differences were observed in loading controls HSC70 (R2=0.38, F=2.453, p=0.114) and β-Tubulin (R2=0.208, Ͷͺ F=1.048, p=0.407). Ͷͻ ͷͲ ͷͳ Table 4: CHIP Loss Changes Expression of Proteins Involved In Mitochondrial Dynamics and Energetics. ͷʹ Whole brain lysates from WT and CHIP KO mice (4 male, 4 female for each genotype) were analyzed by LC- ͷ͵ MS/MS. Proteins listed are >2-fold increased or decreased in KO compared to WT based on spectral count data. ͷͶ Statistical significance was determined by Student’s t-test: Gdap-1 (t= -16.053, df= 7, p= 8.846E-07); Dynamin 3 ͷͷ (t= -29.646, df= 7, p= 1.280E-08); Miro (t= -7.751, df= 7, p= 0.0001); Alcohol dehydrogenase (t= -10.758, df= 7, p= ͷ͸ 1.319E-05); Malate dehydrogenase (t= -8.900, df= 7, p= 4.587E-05); Glutathione S-transferase α4 (t= 14.934, df= 7, ͷ͹ p= 1.448E-06); DJ-1 (t= 8.220, df= 7, p= 7.664E-05). Statistical analysis by two-way ANOVA revealed no ͷͺ significant sex-genotype interaction for these proteins. ͷͻ ʹͲ

͸Ͳ ͸ͳ References: ͸ʹ ͸͵ ͸Ͷ ǡǡǡǡǡ ǡǡ ȋʹͲͳ͵Ȍ ͸ͷ ǦǦ Ͷ ͸͸ ǤƬͳͺǣ͹ͺͶǦ͹ͻͶǤ ͸͹ ǡǡǡǡǡ ǡ ǡ ǡ ǡ ǡ ͸ͺ ǡǡǡ ǡ ǡǡǡ ǡ ͸ͻ ǡ ȋʹͲͳ͸Ȍ ǣ ͹Ͳ Ǥ͵ͷǣʹʹ͹ͲǦʹʹͺͶǤ ͹ͳ ȋʹͲͳ͵Ȍ ǦǦ Ǥ ͹ʹ ͵͸ͻǣʹʹ͵͸ǦʹʹͷͳǤ ͹͵ ǡǡǡ ǡ ǡǡȋͳͻͻͻȌ ǡ ͹Ͷ Ǧ ͹ͷ Ǥ ͳͻǣͶͷ͵ͷǦͶͷͶͷǤ ͹͸ ǡ ǡǦ ǡǡǡǡ ǡ ǡ ͹͹ ǡǡ ǡǡ ǡǡ ǡȋʹͲͳͷȌ ͹ͺ ͹ͻ ͳǤͳͶǣ͵͹ͺǦ͵ͺͳǤ ͺͲ ǡ ǡ ǡ ǡ ǡ ǡ ǡ ǡ ǡǦ ͺͳ ǡ ȋʹͲͳͶȌʹ͹ ͺʹ ͸ͷ Ǥ ͺ͵ ʹͺͻǣͳͶͷ͸ͻǦͳͶͷͺʹǤ ͺͶ ǡǡ ǡ ǡ ǡ ǡǡ ǡ ȋʹͲͳͲȌ ͺͷ Ǧ͸͸ ͺ͸ Ǥ ǣ ͺ͹ ͵ͲǣͷʹͶʹǦͷʹͷʹǤ ͺͺ ǡ ǡ ǡ ǡ ȋʹͲͲͷȌ ͺͻ ǣ Ǥͷ͹ǣ͸ͻͷǦ͹Ͳ͵Ǥ ͻͲ ǡ ȋʹͲͳ͹Ȍ ͻͳ Ǥʹ͸ǣ͵ͻǦͶͺǤ ͻʹ ǡ ǡ ǡ ǡ ǡ ȋʹͲͳ͹Ȍ ͻ͵ ͳ Ǥ ǣ ͻͶ ͵ͳǣͷʹ͵ͶǦͷʹͶͷǤ ͻͷ ǡ ǡ ǡ ǡ ǡ ȋʹͲͲͺȌ ʹͳǣ Ǧ ͻ͸ ʹ͹ Ǥ ͳ͹ͳǣʹ͵ͻǦʹͶ͹Ǥ ͻ͹ Ǥ ȋʹͲͳ͵Ȍ ͻͺ Ǥͳͷǣͳͳͻ͹ǦͳʹͲͷǤ ͻͻ ǡǡǡ ǡǡ ǡ ǡ ǡǡ ǡ ͲͲ ǡ ȋʹͲͲ͵Ȍ ͳ Ͳͳ Ǥ ʹʹǣͷͶͶ͸ǦͷͶͷͺǤ Ͳʹ ǡǡ ǡǡǡǡǡ ǡǡ Ͳ͵ ǡ ȋʹͲͳͶȌ Ǧͳ Ǧ ͲͶ ǤͷǣͳͲͺʹǤ Ͳͷ ǡǦǡǡǡȋʹͲͲͷȌ Ͳ͸ ͹Ͳ ʹ͹ Ǥ ʹͳǣ͵͹ͻǦ Ͳ͹ ͵ͻʹǤ Ͳͺ ǡ ȋʹͲͳʹȌ Ǥ Ͳͻ ͳͲ͹ǣ͵ʹ͹Ǧ͵ͷ͵Ǥ ʹͳ

ͳͲ ǡ ǡ ǡǡ ǡǡ ǡǡ ǡ ͳͳ ǡ ǡǡ ǡǡǡ ȋʹͲͲ͵ȌǦ ͳʹ Ǥ ͳ͵ ʹ͹ͺǣͶ͵͸ʹͺǦͶ͵͸͵ͷǤ ͳͶ ǡ ǡ ǡ ǡ ǡ ǡ ǡ ǡ ͳͷ ȋʹͲͳʹȌ ͳ ǡ ͳ͸ Ǥͳ͵ǣ͵͹ͺǦ͵ͺͷǤ ͳ͹ ǡ ǡ ǡ ǡ ǡ ǡ ǡ ǡ ͳͺ ǡ ǡ ǡ ǡ ȋʹͲͳͲȌ ͳͻ Ǥͷǣͳʹͻ͸ʹǤ ʹͲ ǡ ȋͳͻͺͻȌ ǡʹǤ ʹͳ ǣǢ ʹʹ Ǥ ʹ͵ ȋʹͲͲͷȌ Ǥ ǣ ȋǡ ʹͶ ǡȌǡͳͲǦͳ͹ǤǡǡǣǤ ʹͷ ǡǡ ǡǡȋʹͲͲͳȌ Ǧ Ǥ ʹ͸ ʹͳǣ͵͸ͲͶǦ͵͸ͲͺǤ ʹ͹ ǡ ǡǡǡ ǡ ǡ ǡ ǡ ǡ ǡ ʹͺ ȋʹͲͳ͹Ȍ ͳȀ ʹͻ ǣ Ǥ ͵Ͳ ͳʹǣ͵ͳǤ ͵ͳ ǡ Ǧ ǡ ǡ ǡǡ ǡ ǡ ǡ ͵ʹ ǡ ǡ ǡ ǡ ǡ ǡ ǡ ǡ ͵͵ ȋʹͲͳͶȌ ͳ Ǧ Ǧ ͵Ͷ Ǥ ͻǣͳͶ͸Ǥ ͵ͷ ǡ ǡ ǡ ǡ ǡ ǡ ȋʹͲͲʹȌ ͵͸ ǡ ̵ ǡ ͵͹ Ǥ ͳͲǣͷͷǦ͸͹Ǥ ͵ͺ ǡ ǡ ǡ ǡ ȋʹͲͳ͹Ȍ Ǧͳ ͵ͻ Ǥ ͶͲ ͳͲ͵͹ǣͳͺ͹ǦʹͲʹǤ Ͷͳ ǡǡǡǡǡ ǡȋʹͲͲͳȌ ǦǦ͵ Ͷʹ ǣ ͹Ͳ Ǥ Ͷ͵ ʹ͹͸ǣͶʹͻ͵ͺǦͶʹͻͶͶǤ ͶͶ ǡ ȋʹͲͳ͵Ȍ Ͷͷ ͳ ʹȀǦ Ǥ Ͷ͸ ͻǣͳ͹ͷͲǦͳ͹ͷ͹Ǥ Ͷ͹ ǡ ǡ ǡ ǡ ǡ ȋʹͲͳͲȌ ͳǦ Ͷͺ Ǧ Ǥ Ͷͻ ǣ ͵Ͳǣ͵ͻʹͶǦ͵ͻ͵ʹǤ ͷͲ ǡ ǡ ǡ ȋʹͲͳͶȌ Ǧ Ǧ ͷͳ Ǧͳ ǣ Ǥ ͷʹ ʹͲǣʹ͹ͷǦʹͺͳǤ ͷ͵ ǡǡǡ ǡ ǡ ǡ ȋʹͲͳͲȌ ͳ ͷͶ ͵Ǥ ͷͺͶǣͳͲ͹͵ǦͳͲ͹ͻǤ ͷͷ ǡ ǡ ǡ ǡ ȋʹͲͳ͸Ȍ ǡ ǡ ͷ͸ ͵ ͳ͸Ǥ ͷ͹ ͹͵ǣͺͺͺǦͺͻͲǤ ͷͺ ǡǡǡǡ ȋʹͲͲͺȌǣ ͷͻ Ǥ ʹͶǣʹͷ͵ͶǦʹͷ͵͸Ǥ ʹʹ

͹͸Ͳ ǤȋʹͲͳ͸Ȍ ͷ Ǥ ͹͸ͳ ͷǤ ͹͸ʹ ǡ ȋʹͲͳͶȌ Ǧ Ϊ ͹͸͵ Ǥͷǣͷʹ͹͹Ǥ ͹͸Ͷ ǤȋʹͲͳ͸Ȍ ͹͸ͷ ȋ͵ȌǤͳʹǣͳǦʹʹʹǤ ͹͸͸ ǡ ǡ ǡǦǡ ǡǡ ǡȋʹͲͳ͹Ȍ ͹͸͹ ǦǤ͸Ǥ ͹͸ͺ ǡ ǡ ǡ ǡ ǡ ǡ ǡ ǡ ǡ ͹͸ͻ ȋʹͲͳ͵Ȍ ͹͹Ͳ Ǥ ʹͺͺǣͻͳͷǦͻʹ͸Ǥ ͹͹ͳ ǡǡǡǡ ȋʹͲͳʹȌǦ ͵ ͹͹ʹ Ǧ ͹ͲǦ Ǥ ͹͹͵ ͳʹͲǣ͵ͷͲǦ͵͹ͲǤ ͹͹Ͷ ǡ ǡ ǡ ȋʹͲͲʹȌ ͹͹ͷ Ǥ ͻǣͷͳ͹Ǧͷʹ͸Ǥ ͹͹͸ ǡ ǡ ǡ ǡ ǡ ǡ ǡ ǡ ǡ ǡ ǡ ͹͹͹ ǡ ǡȋʹͲͳʹȌ̵Ǧ ͳ ͹͹ͺ Ǥ ͺǣͳͲͲʹͷ͵͹Ǥ ͹͹ͻ ǡǡ ȋʹͲͳ͹Ȍ͵ ͹ͺͲ Ǥ Ǥ ͹ͺͳ ǡǡǡǡ ǡ ǡ ǡ ǡǡ ͹ͺʹ ǡ ȋʹͲͲͻȌ ʹǤͲǣ ͹ͺ͵ Ǥ ͺǣ͵ͺ͹ʹǦ͵ͺͺͳǤ ͹ͺͶ ǡǡǡǡǡ ǡǡǡǡ ǡǡ ͹ͺͷ ǡ ǡȋʹͲͳͲȌ ͳ ͹ͺ͸ Ǥ ͹ͺ͹ ͳͺͻǣʹͳͳǦʹʹͳǤ ͹ͺͺ ǡ ǡ ǡ ǡ ǡ ȋʹͲͳʹȌ ͹ͺͻ ͵ Ǧ ͹Ͳ ͹ͻͲ ȋ Ȍ Ǥ͹ǣ͵͸͵ͶͲǤ ͹ͻͳ ǡ ǡ ǡ ǡ ǡ ǡ ȋʹͲͲ͵Ȍ ͵ ͹ͻʹ Ǥ ͹ͻ͵ ͳͲͲǣ͹ͳͷǦ͹ʹͲǤ ͹ͻͶ Ǧǡǡǡ ǡǡȋʹͲͲͺȌ ͹ͻͷ ǡ Ǥ ͹ͻ͸ ʹͺǣͶͲͳͺǦͶͲʹͷǤ ͹ͻ͹ ǡ ǡ ǡȋʹͲͳͷȌ ͹ͻͺ Ǥ ͹ͻͻ ʹʹǣͳͲ͵͵ǦͳͲͶͶǤ ͺͲͲ ǡ ǡǡ ǡǡȋʹͲͳʹȌ Ǧ ͺͲͳ ͷ͸ ͺͲʹ Ǥ ʹͳǣͳͻ͹ͻǦͳͻͺͺǤ ͺͲ͵ ǡǡǡǡǡ ǡȋʹͲͲͲȌ ͺͲͶ ͹Ͳ Ǧ Ǥ ͺͲͷ ʹͲǣ͹ͳͶ͸Ǧ͹ͳͷͻǤ ͺͲ͸ ǡ ȋʹͲͳ͸Ȍ ͳ Ǧͳ ͺͲ͹ Ǥ ͳͺ͸͵ǣ͵Ͳ͸ͷǦ͵Ͳͺ͵Ǥ ͺͲͺ ǡǡǡȋͳͻͻͺȌǣǡ ͺͲͻ Ǥ ͳ͹ǣͳͻ͵ǦʹͳͲǤ ʹ͵

ͺͳͲ ǡ ǡ ǡ ǡ ǡ ǡ ǡ ȋʹͲͳͲȌ ͳ ͺͳͳ ǤͺǣͳͲͲͲʹͻͺǤ ͺͳʹ ǡ ǡ ǡ ǡ ȋʹͲͲͷȌ Ǧ ͺͳ͵ ͳ ǣ Ǧ ͺͳͶ ǦǤ ͳ͹ͲǣͳͲ͸͹ǦͳͲ͹ͺǤ ͺͳͷ ̵ ǡ ǡȋʹͲͲͷȌ ǡǡ ǡ ͺͳ͸ Ǥ ͵ͷǣ͸ͲͻǦ͸͸ʹǤ ͺͳ͹ ǡȋʹͲͳͳȌ Ǥ ͺͳͺ ͳͶͻǣʹͶͳǦʹͷͳǤ ͺͳͻ ǡ ǡ ǡ ȋʹͲͳ͵Ȍ ͳʹ̱ ͷ ͺʹͲ ͵Ǥ ʹͲǣͷͻǦ͸͸Ǥ ͺʹͳ ǡ ǡ ǡ ǡ ǡ ǡ ǡ ȋʹͲͲͶȌ ͺʹʹ Ǧ Ǥ ͺʹ͵ ʹ͹ͻǣͳͺ͸ͳͶǦͳͺ͸ʹʹǤ ͺʹͶ ǡ ǡǡ ǡ ǡǡ ǡ ͺʹͷ ȋʹͲͳͷȌ Ǥ ͺʹ͸ Ƭʹ͵ǣͷ͵ͷǦͷͶͻǤ ͺʹ͹ ǡȋʹͲͲͷȌǦ Ǥ ͺʹͺ ͳͲʹǣʹͳ͹ͶǦʹͳ͹ͻǤ ͺʹͻ ǡ ǡ ǡ ǡ ȋʹͲͲ͸Ȍ Ǧ ͺ͵Ͳ ͹ͲǤͶͶͲǣͷͷͳǦͷͷͷǤ ͺ͵ͳ ǡ ǡ ȋʹͲͳ͵Ȍ ǡ ͺ͵ʹ ǡǤ ͺ͵͵ ͳͳͲǣͺ͸͵ͺǦͺ͸Ͷ͵Ǥ ͺ͵Ͷ ȋʹͲͳͳȌ Ǧ ǣ ͺ͵ͷ ǡ Ǥ ʹ͹ͺǣͶʹͶ͵ǦͶʹͷͳǤ ͺ͵͸ ǡȋʹͲͳͶȌ ͳ͵ ͺ͵͹ Ǥ͹ǣͳͺͺǤ ͺ͵ͺ ǡ ǡ ǡ ȋʹͲͳ͸Ȍ ͳǦ Ǥ ͺ͵ͻ Ͷ͹ͷǣʹͺ͵ǦʹͺͺǤ ͺͶͲ ǡǡ ǡǡ ǡǡǡ ȋʹͲͲͺȌ ͺͶͳ ʹΪǦ Ǥ ͺͶʹ ͳͲͷǣʹͲ͹ʹͺǦʹͲ͹͵͵Ǥ ͺͶ͵ ǡ ǡ ǡ ǡ ǡ ǡ ǡ ȋʹͲͲͲȌ ͺͶͶ Ǧ ǣ ͹ͲǦ ͺͶͷ ͻͲ ǤͳͲͳǣͳͻͻǦʹͳͲǤ ͺͶ͸ ǡ ȋʹͲͲͷȌ ǣ Ǥ ͺͶ͹ ʹͲǣ͸ͷ͵Ǧ͸ͷͷǤ ͺͶͺ ǡ ǡ ȋʹͲͳͷȌ Ǥ ʹͺʹǣʹͲ͹͸ǦʹͲͺͺǤ ͺͶͻ ǡǡ Ǧ ǡǦǡǡǦ ǡǡ ǡ ͺͷͲ Ǧ ǡ ǡ ȋʹͲͳͶȌ ͳͲȀ ǣ ǡ ͺͷͳ Ǧ ͺͷʹ Ǥͷǣͳͳ͸͵Ǥ ͺͷ͵ ǡ ǡǡǡǡ ǡǡǡǡǡ ǡ ͺͷͶ ǡǡǡǡȋʹͲͳͶȌ ͺͷͷ Ǥ ʹ͵ǣͳͲͳ͵Ǧ ͺͷ͸ ͳͲʹͶǤ ͺͷ͹ ǡ ȋʹͲͳͳȌǦ ͺͷͺ Ǧ͹Ͷ͵Ǥ ͳͳͺǣͶͶͲǦͶͶͺǤ ͺͷͻ ǡǡǡ ǡ ǡȋʹͲͳͷȌǦ ʹͶ

ͺ͸Ͳ ǣ ͶǦ ǡ Ǧ Ǥ ͺ͸ͳ ʹͺͻǣ͵͸ͳǦ͵͹ͲǤ ͺ͸ʹ ǡ ǡ ǡ ȋʹͲͳͲȌ ͺ͸͵ ͹ͲǦǤ ʹ͹͹ǣ͵͵ͷ͵Ǧ͵͵͸͹Ǥ ͺ͸Ͷ ǡǡǡ ǡ ǡ ȋʹͲͳͳȌǦ ͺ͸ͷ ͹Ͳ ͺ͸͸ ǤƬͳͶǣͳ͹ͺ͹ǦͳͺͲͳǤ ͺ͸͹ ǡǡ ǡǡ ǡ ǡ ǡ ǡ ǡ ȋʹͲͳͶȌ ͺ͸ͺ ǣ Ǥ ͺ͸ͻ ͳͳͶǣͷͺǦͺ͵Ǥ ͺ͹Ͳ ǡǡ ǡǡ ǡǡ ǡǡǡǡ ǡ ͺ͹ͳ ȋʹͲͳͷȌ Ǧ ͺ͹ʹ Ǥ ͺ͹͵ ͳͳʹǣͳͳ͸ͻ͸Ǧͳͳ͹ͲͳǤ ͺ͹Ͷ ǡ ǡ ǡ ǡ ǡ ǡ ǡ ǡ ͺ͹ͷ ǡ ǡ ȋʹͲͳͲȌ Ǧ Ǧ ͺ͹͸ Ǥ ͺͺǣʹͷ͹Ǧ ͺ͹͹ ʹ͸͸Ǥ ͺ͹ͺ ǡǡ ǡ ȋʹͲͲͳȌ ǦǦǤ ͺ͹ͻ ͶͺʹǣʹͳǦʹ͸Ǥ ͺͺͲ ǡ ǡ ǡ Ǧ ǡ ǡ ǡ Ǧ ǡ ͺͺͳ ǡ ǡ ǡ ǡ ȋʹͲͳͶȌ ͳ ͺͺʹ ǣ Ǧ Ǥ ͺͺ͵ ͻǣͷ͹Ǥ ͺͺͶ ǡ ǡȋʹͲͲ͹Ȍ ǣ ͺͺͷ Ǥ ͸ǣ͸ͷͶǦ͸͸ͳǤ ͺͺ͸ ȋʹͲͳͷȌ ǡ ǤͷǤ ͺͺ͹ ǡǡ ǡ ǡǡǡ ǡ ǡ ǡ ȋʹͲͳͳȌ ͺͺͺ ͳ ͺͺͻ ǤͳͶ͹ǣͺͻ͵ǦͻͲ͸Ǥ ͺͻͲ ǡ ǡ ǡ ǡ ǡ ǡ ǡ ǡ Ǧ ǡ ǡ ͺͻͳ ǡ ǡ ǡ ʹǡ ȋʹͲͳͳȌ Ǧ ͺͻʹ ͳͲ Ǥ ͺͻ͵ ͳͳǣʹͲͳͻǦʹͲʹ͸Ǥ ͺͻͶ ȋʹͲͲͷȌǡ ǡ Ǥ ͺͻͷ ʹͺǣʹͺʹͳǤ ͺͻ͸ ǡǡ ǡǡ ǡȋʹͲͲͺȌ ͺͻ͹ Ǥ ͺǣʹ͸Ǥ ͺͻͺ ǡ ǡ ȋʹͲͳ͵Ȍ ͺͻͻ Ǥ Ƭ ͸ʹǣͻͲǦͳͲͳǤ ͻͲͲ ǡ ǡǡ ǡǡ ǡ ǡǡǡ ǡ ǡ ǡ ͻͲͳ ȋʹͲͳͳȌ ǦȋȌǦ Ǧ ͻͲʹ ͻͲ͵ ̵ Ǥ ǣ ͻͲͶ ͵ͳǣʹ͵ͳ͵Ǧʹ͵ʹͲǤ ͻͲͷ ǡ ǡ ǡ ǡ ǡ ǡ ǡ ȋʹͲͳ͹Ȍ ͻͲ͸ ͵ Ǥ ͻͲ͹ ͹ǣͳ͹ͺͻǤ ͻͲͺ ǡǡǡ ǡǡǡ ǡǡ ǡǡ ͻͲͻ ȋʹͲͳ͸Ȍ ͷǦ ʹͷ

ͻͳͲ ͷǤ͵ͻǣͳͳ͸Ǧͳ͵ͲǤ ͻͳͳ ǡ ǡ ǡ ǡ ǡ ȋʹͲͳͲȌ ͻͳʹ Ǧ Ǥ ͻͳ͵ ͳͺͲʹǣͳͲͻͷǦͳͳͲͶǤ ͻͳͶ ǡ ǡ ǡǡ ǡ ǡ ǡǡ ȋʹͲͳͷȌͳ ͻͳͷ Ǥ ͹ǣͳͲͲǤ ͻͳ͸ ǡ ǡ ǡ ǡ ǡ ǡ ȋʹͲͲͷȌ ͻͳ͹ ǦǦ ͵ Ǧ ͳ͵Ǧͳ ͻͳͺ Ǥ ʹͲǣͷʹͷǦͷ͵ͺǤ ͻͳͻ

ʹ͸

Fold Change Normalized to Control N 1h 3h 6h 18h 24h HSC70 4 0.9 ± 0.1 0.9 ± 0.2 1.3 ± 0.5 1.0 ± 0.3 1.1 ± 0.4 HSP70 4 1.4 ± 0.4 2.1 ± 0.3 2.5 ± 0.2 1.0 ± 0.7 1.3 ± 0.3 CHIP 4 1.8 ± 0.4 2.2 ± 0.4 2.8 ± 0.8 2.5 ± 0.8 1.7 ± 0.3 β-Tubulin 4 1.2 ± 0.1 1.3 ± 0.1 1.2 ± 0.2 1.2 ± 0.2 1.2 ± 0.1

Fold Change Normalized to Control N 3h 6h 24h PINK1 3 2.5 ± 1.0 2.6 ± 0.9 3.3 ± 2.1 TOM20 4 1.3 ± 0.1 1.3 ± 0.2 1.2 ± 0.2 HSC70 4 0.9 ± 0.2 1.3 ± 0.5 1.1 ± 0.4 β-Tubulin 4 1.3 ± 0.1 1.2 ± 0.2 1.2 ± 0.1

Fold Change Normalized to Control N Con +BafA OGD OGD +BafA HSC70 4 1.0 ± 0.1 1.1 ± 0.1 1.2 ± 0.1 HSP70 4 1.1 ± 0.3 1.7 ± 0.3 1.4 ± 0.1 CHIP 3 1.0 ± 0.1 1.2 ± 0.0 1.4 ± 0.1 PINK1 4 0.8 ± 0.2 1.8 ± 0.6 1.9 ± 0.6 TOM20 3 2.1 ± 0.8 1.8 ± 0.4 2.8 ± 1.7 p62 3 1.5 ± 0.1 1.2 ± 0.4 2.0 ± 0.9 LC3-I 4 0.8 ± 0.2 1.0 ± 0.3 1.3 ± 0.6 LC3-II 4 0.9 ± 0.2 1.4 ± 0.4 2.3 ± 0.9 LC3-II:LC3-I 4 1.0 ± 0.1 1.5 ± 0.4 2.1 ± 0.3 β-Tubulin 4 1.1 ± 0.1 1.4 ± 0.1 1.3 ± 0.3

Significantly Altered Proteins (Fold Change) Increased in CHIP KO Mice Protein Fold Change Gdap-1 10.2 Dynamin 3 7.5 Miro 2.3 Alcohol dehydrogenase 3.0 Malate dehydrogenase 2.2 Decreased in CHIP KO Mice Protein Fold Change Glutathione S Transferase α4 4.1 DJ-1 2.8