Characterizing the Bcl-2 Associated Athanogene 5 Interactome in the Context of Parkinson’S Disease

Home , Ankycorbin, ARHGAP31, BAG5, PHLDB1, Prefoldin subunit 6, RAVER1

Characterizing the Bcl-2 Associated Athanogene 5 Interactome in the Context of Parkinson’s Disease

Erik Loewen Friesen

A thesis submitted in conformity with the requirements for the degree of Master of Science Department of Laboratory Medicine & Pathobiology University of Toronto

Characterizing the Bcl-2 Associated Athanogene 5 Interactome in the Context of Parkinson’s Disease

Erik Loewen Friesen

Master of Science

Department of Laboratory Medicine & Pathobiology University of Toronto

2018 Abstract

Aberrant alpha-synuclein aggregation is associated with the onset and progression of Parkinson’s disease (PD). This has made molecular chaperones, a class of proteins responsible for maintaining proteostasis, an enticing therapeutic target. BAG5 is a co-chaperone protein that inhibits the chaperone, Hsp70, and promotes PD-like alpha-synuclein aggregation and neurodegeneration. The mechanisms of how BAG5 impairs proteostasis and promotes apoptosis are unclear. The purpose of this project was to characterize the BAG5 interactome to guide further studies of its role in physiological and disease states. A novel interaction between BAG5 and the autophagy adaptor protein, p62, was discovered and investigated, as it pointed to a potential mechanism by which BAG5 could modify alpha-synuclein aggregation. p62 reduced and BAG5 enhanced alpha-synuclein self-association in vitro. BAG5 also promoted p62 stability, suggesting a function of the interaction on other p62-dependent proteostasis pathways.

Acknowledgments

I would like to thank all of the members of the Kalia and Lozano labs for their continuous support throughout this process. This has been a tremendous learning experience that was made very positive by the team members that have helped me along the way. A special thanks goes to Hien Chau for teaching me virtually every basic science technique needed to complete this thesis, and being a great mentor to me for over three years. Thanks also goes to my fellow graduate students, Mitch, Greg, Krystal, Shirley, Stanley and Kevin, for all the support and good times we have had throughout my time in the lab. I would also like to acknowledge the entire group of graduate students and postdocs on the 8th floor of Krembil, who continuously made this process fun and worthwhile. A special thanks also goes to Mitch De Snoo for his extensive technical assistance and helping me get through the thick and thin of laboratory research.

The continuous support from both Suneil and Lorraine Kalia has been instrumental to my success in this program. Thank you for having the trust and patience to allow me to independently explore new ideas and providing guidance when needed. Thanks also to the members of the Schmitt-Ulms lab, namely Gerold, Louisa, and Declan, for the time and effort you put into taking my proteomics data to the next level.

Lastly, I would like to thank the people who supported me on the home front throughout this process. Thank you Katrina for putting up with my ramblings about obscure topics in neuroscience research, and understanding when I had to put in long hours at the lab. Thanks to my parents, Brad, Shelagh, Louise and Peter for your continuous support on multiple fronts.

iii

Table of Contents

Acknowledgments ...... iii

Table of Contents ...... iv

List of Abbreviations ...... vii

List of Tables ...... viii

List of Figures ...... ix

List of Appendices ...... xi

Chapter 1 Introduction ...... 1

1.1 Parkinson’s Disease ...... 1

1.1.1 Overview ...... 1

1.1.2 Familial PD ...... 2

1.1.3 Alpha-synuclein and Proteostasis in PD ...... 5

1.1.4 Mitochondrial Dysfunction in PD ...... 8

1.2 Molecular Chaperones in PD ...... 10

1.2.1 The Nature and Function of Molecular Chaperones ...... 10

1.2.2 Molecular Chaperones and Alpha-Synuclein Pathology ...... 15

1.2.3 Molecular Chaperones and Mitochondrial Dysfunction ...... 17

1.2.4 BAG Family Co-chaperones ...... 18

1.2.5 BAG5 ...... 21

1.3 Summary and Research Objectives ...... 23

Chapter 2 Characterizing the BAG5 Interactome ...... 25

2.1 Introduction ...... 25

2.2 Materials & Methods ...... 31

2.2.1 Antibodies & Reagents ...... 31

2.2.2 Cell Culture ...... 32

2.2.3 Western Blotting ...... 32

2.2.4 Generation of the H4 Stable Cell Lines ...... 32

2.2.5 Immunoprecipitation and Mass Spectrometry: H4 Cells ...... 33

2.2.6 Generation of the SH-SY5Y Stable Cell Lines ...... 33

2.2.7 Immunoprecipitation and Mass Spectrometry: SH-SY5Y Cells ...... 34

2.2.8 Bioinformatic Analysis ...... 35

2.3 Results ...... 36

2.3.1 Characterization of the BAG5 Interactome: H4 Cells ...... 36

2.3.2 Characterization of the BAG5 Interactome: SH-SY5Y Cells ...... 41

2.4 Discussion ...... 44

Chapter 3 Validating the Interaction Between BAG5 and p62 ...... 48

3.1 Introduction ...... 48

3.2 Materials & Methods ...... 52

3.2.1 Antibodies & Reagents ...... 52

3.2.2 Cell Culture ...... 53

3.2.3 Western Blotting ...... 53

3.2.4 GST Pull-down Assay ...... 53

3.2.5 Immunoprecipitation ...... 54

3.2.6 Immunohistochemistry ...... 55

3.3 Results ...... 55

3.3.1 Validation and Visualization of the BAG5-p62 Interaction ...... 55

3.3.2 p62 Interacts with BAG5 via its C-terminal Domains ...... 56

3.4 Discussion ...... 61

Chapter 4 Investigating the BAG5-p62 Interaction in the Context of Alpha-synuclein aggregation ...... 63

4.1 Introduction ...... 63

4.2 Materials & Methods ...... 66

4.2.1 Antibodies & Reagents ...... 66

4.2.2 Cell Culture ...... 66

4.2.3 Western Blotting ...... 66

4.2.4 Alpha-synuclein Protein Complementation Assay ...... 66

4.3 Results ...... 67

4.3.1 p62 reduces the presence of soluble alpha-synuclein and oligomers ...... 67

4.3.2 BAG5 KD reduces alpha-synuclein oligomerization but does not impact p62 .....69

4.3.3 BAG5 Stabilizes Endogenous p62 ...... 73

4.3.4 Discussion ...... 75

Chapter 5 General Discussion & Future Directions ...... 78

5.1 Summary ...... 78

5.2 Study Limitations ...... 79

5.3 Future Directions: BAG5, p62 and Proteostasis ...... 83

5.4 Future Directions: BAG5 and Cell Death ...... 85

5.5 Conclusion ...... 88

References ...... 89

Appendices ...... 103

List of Abbreviations

AAV: adeno-associated virus LN: Lewy neurite AD: Alzheimer’s disease LRRK2: leucine rich repeat kinase 2 ADP: adenosine diphosphate Mcl-1: induced myeloid leukemia cell ALP: autophagy lysosome system differentiation protein Mcl-1 ALS: amyotrophic lateral sclerosis miRNA: microRNA ATP: adenosine triphosphate MPTP: 1-methyl-4-phenyl-1,2,3,6- BAG: bcl-2 associated athanogene tetrahydropyridine BAG5: bcl-2 associated athanogene 5 NBR1: neighbor of BRCA1 gene 1 Bcl-2: B-cell lymphoma 2 NLS: nuclear localization signal CHIP: C-terminal Hsp70 interacting p62/SQSTM1: sequestosome-1 protein PB1: phox and bem1 CMA: chaperone mediated autophagy PCA: protein complementation assay CMV: cytomegalovirus PCR: polymerase chain reaction DAPI: 4',6-diamidino-2-phenylindole, PD: Parkinson’s disease dihydrochloride PINK1: PTEN-induced kinase 1 DJ-1: protein deglycase DJ-1 ROS: reactive oxygen species DSB: double stand break rtTA: reverse transactivator EDTA: ethylenediaminetetraacetic acid SBD: substrate binding domain EIF4G1: eukaryotic translation initiation SDS-PAGE: sodium dodecyl sulfate factor 4 gamma 1 polyacrylamide gel electrophoresis GFP: green fluorescent protein siRNA: small interfering RNA GO: gene ontology SN: substantia nigra gRNA: guide RNA SNCA: alpha-synuclein GST: glutathione S-transferase SNpc: substantia nigra pars compacta HD: Huntington’s disease TALEN: transcription activator-like Hsc70: heat shock cognate 70 kDa effector nucleases HSF-1: heat shock factor 1 TDP-43: transactive response DNA HSP: heat shock protein binding protein 43 kDa iTRAQ: isobaric tags for relative and TOM20: translocase of outer membrane absolute quantification 20 KD: knockdown TOM22: translocase of outer membrane Keap1: kelch-like ECH-associated protein 22 1 TPR: tetratricopeptide repeat KIR: Keap1 interacting region TRAP1: TNF receptor associated protein 1 KO: knockout TRE: tetracycline response element LAMP2A: lysosome-associated membrane UBA: ubiquitin associated domain protein 2A UPS: ubiquitin proteasome system LB: Lewy body ZFN: zinc finger nuclease LC3: light chain 3 LIR: LC3 interacting region

vii

List of Tables

Table 1 Genes Associated with Familial PD ...... 3

Table 2 Top 10 BAG5 and BAG5+BAG5DARA Interacting Proteins ...... 38

viii

List of Figures

Figure 1 Proposed role of molecular and small molecule chaperones in proteostasis...... 14

Figure 2 Schematic diagram of the six BAG family members...... 20

Figure 3 Generation of SH-SY5Y stable cell lines...... 28

Figure 4 Expression of GFP-tagged constructs in H4 cells...... 39

Figure 5 Bioinformatic analysis of the BAG5 interactome in H4 Cells...... 40

Figure 6 Inducible expression of the GFP transgenes in SH-SY5Y cells...... 42

Figure 7 iTRAQ mass spectrometry strategy allows for the visualization of BAG5 vs. BAG5DARA binding preference...... 43

Figure 8 Bioinformatic analysis of the BAG5 interactome in SH-SY5Y Cells...... 44

Figure 9 p62 facilitates the aggregation and degradation of protein aggregates...... 51

Figure 10 Domain structure of p62 and the deletion constructs generated to map its interaction with BAG5...... 52

Figure 11 Confirmation of the interactions between BAG5 and DNAJC13/p62 by GST pulldown...... 57

Figure 12 Confirmation of the BAG5-p62 interaction via co-immunoprecipitation...... 58

Figure 13 BAG5 and p62 co-localize within perinuclear puncta...... 59

Figure 14 p62 associates with BAG5 independently of its N-terminal PB1 domain...... 60

Figure 15 p62 and BAG5 have complex effects on protein aggregation and degradation pathways...... 65

Figure 16 Alpha-synuclein protein complementation assay (PCA) proof of concept...... 68

Figure 17 p62 reduces the presence of both soluble and oligomeric alpha-synuclein...... 70 ix

Figure 18 p62 requires both its C-terminal PB1 domain and N-terminal LIR+UBA domains to influence alpha-synuclein oligomerization...... 71

Figure 19 p62 and BAG5 have independent effects on synuclein oligomerization...... 72

Figure 20 BAG5 stabilizes endogenous levels of p62...... 74

List of Appendices

Appendix 1 BAG5 Interactome: H4 ...... 103

Appendix 2 BAG5 Interactome: SH-SY5Y ...... 113

Appendix 3 Contributions ...... 119

Chapter 1 Introduction 1.1 Parkinson’s Disease

1.1.1 Overview

Parkinson’s disease (PD) is a neurodegenerative disorder characterized by the progressive loss of dopamine neurons in the substantia nigra pars compacta (SNpc). The loss of these neurons results in the dysfunction of a motor circuit in the basal ganglia involved in fine-tuning controlled movements (Kalia and Lang 2015, Friesen et al. 2017). In turn, this overtly manifests as the classic parkinsonian motor deficits including resting tremor, rigidity, akinesia and postural/gait impairments, as well as non-motor deficits such as cognitive impairments and psychiatric symptoms. Treatment options for patients with PD include dopamine replacement therapies, such as levodopa administration, and surgical interventions that modulate activity within the basal ganglia circuitry using deep brain stimulation (DBS) (Kalia and Lang 2015,

Lozano, Hutchison, and Kalia 2017). Unfortunately, both interventions only address the symptoms of PD but do not alter the disease progression. Therefore, there is a need for treatments that address the underlying causes of the disease (Friesen et al. 2017).

A neuropathological hallmark of PD is the presence of intracellular neuronal protein aggregates called Lewy pathology. Lewy pathology can form in either the soma (Lewy bodies

[LBs]) or dendrites (Lewy neurites [LNs]) of neurons and presents in various brain regions, including the SNpc (Kalia et al. 2013, Spillantini et al. 1997). The most abundant constituent of

LBs/LNs is misfolded alpha-synuclein, which has lead to a significant amount of research into the relationship between alpha-synuclein aggregation and PD (Kalia et al. 2013). However, while

1 2 a correlation between alpha-synuclein aggregation and PD exists, the precise nature of the relationship between the two remains elusive.

In terms of etiology, most cases of PD do not follow an observable inheritance pattern, and are classified as “sporadic”. This, combined with the association of PD with exposure to environmental toxins such as rotenone and paraquat, led many to consider PD a disease brought about by environmental factors (Tanner et al. 2011). However, in 1997, it was found that a mutation to the SNCA gene encoding alpha-synuclein could cause PD (Polymeropoulos et al.

1997). Subsequent research lead to the discovery of other PD-causing alpha-synuclein mutations

(currently A30P, E46K, H50Q, G51D, A53T & A53E (Rosborough, Patel, and Kalia 2017)), as well as a host of other genes that, when mutated, could cause PD in either an autosomal dominant or recessive manner (Deng, Wang, and Jankovic 2017). In light of this research, the focus of PD research shifted to include both genetic and environmental etiologies. Now, 5-10% of PD patients are considered to have hereditary forms of PD also known as “familial PD”, and harbor PD-associated genetic mutations (Rosborough, Patel, and Kalia 2017, Lesage and Brice

2009).

1.1.2 Familial PD

Investigations into PD-causing genes and genetic loci are ongoing, and, relatively speaking, the notion of PD as a heritable disease is in its infancy. Therefore, there is still much debate surrounding which genes are actually disease causing. However, a recent review citing the HUGO Gene Nomenclature Committee (HGNC) indicates that there are currently 23 known

PD-causing genetic loci (Table 1), of which 19 have designated gene names (Deng, Wang, and

Jankovic 2017). Five of these genes have been well replicated in clinical populations (Koprich,

Kalia, and Brotchie 2017). These include the genes encoding alpha-synuclein (PARK1/PARK4)

Table 1 Genes Associated with Familial PD (Adapted from Deng et al. (2017))

Inheritance Locus Gene Name Protein Name Pattern

PARK1 SNCA alpha-synuclein AD PARK2 PRKN parkin AR PARK3 PARK3 AD PARK4 SNCA alpha -synuclein AD PARK5 UCHL1 ubiquitin C-terminal hydrolase L1 AD PARK6 PINK1 PTEN induced putative kinase 1 AR PARK7 PARK7 parkinsonism associated deglycase (DJ-1) AR PARK8 LRRK2 leucine rich repeat kinase 2 AD PARK9 ATP13A2 ATPase 13A2 AR PARK10 PARK10 unclear PARK11 GAGYF2 GRB10 interacting GYF protein 2 AD PARK12 PARK12 X-linked PARK13 HTRA2 HtrA serine peptidase 2 AD PARK14 PLA2G6 Phospholipase A2 group VI AR PARK15 FBXO7 F-box protein 7 AR PARK16 PARK16 unclear PARK17 VPS35 vacuolar protein sorting 35 AD PARK18 EIF4G1 eukaryotic translation initiation factor 4 gamma 1 AD PARK19 DNAJC6 DNAJC6 (Hsp40 family) AR PARK20 SYNJ1 synaptoganin 1 AR PARK21 DNAJC13 DNAJC13 (Hsp40 family member) AD PARK22 CHCHD2 coiled-coil-helix-coiled-coil-helix domain containing 2 AD PARK23 VPS13C vacuolar protein sorting 13C AR

and leucine rich repeat kinase 2 (LRRK2, PARK8), which cause autosomal dominant PD, as well as parkin (PARK2), PTEN-induced putative kinase (PINK1, PARK6), and DJ-1 (PARK7), which cause autosomal recessive PD (Lesage and Brice 2009, Deng, Wang, and Jankovic 2017). While there are similarities in the clinical phenotype brought about by mutations in these genes, there are also differences. For example, most patients with PD-associated parkin mutations and some with LRRK2 mutations do not demonstrate Lewy pathology (Schneider and Alcalay 2017, Kalia

4 et al. 2015). Age of onset also varies significantly between the different PD-causing mutations

(Deng, Wang, and Jankovic 2017).

Since their discovery, the properties of familial PD-associated proteins have been extensively characterized. This has revealed that they tend to converge within several common cellular processes (Kumaran and Cookson 2015, Kalia and Lang 2015). Indeed, almost all of these proteins interact with the proteins and pathways responsible for managing cellular protein homeostasis (proteostasis), including molecular chaperones, the ubiquitin proteasome system

(UPS) and the autophagy lysosome pathway (ALP) (Friesen et al. 2017). Many of them also play a role in the maintenance of health and homeostasis within the mitochondrial network. For example, PINK1 and parkin function in a common pathway, termed ‘mitophagy’, that facilitates the autophagic degradation of damaged mitochondria (Narendra et al. 2008). These familial PD- associated proteins also converge onto other processes, such as vesicle trafficking within the golgi apparatus and endolysosomal system (Kett and Dauer 2016), however, these functions are not immediately relevant to this thesis and will not be discussed in further detail.

Because of their function in protein and mitochondrial homeostasis pathways, it is not surprising that disease-causing mutations in PD-associated proteins often results in the loss of proteostasis and/or mitochondrial health. Interestingly, this parallels the deficits observed in sporadic PD. For example, the pesticide rotenone, the herbicide paraquat, and the synthetic compound 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine (MPTP), which cause PD in the human population and are used as a model of dopaminergic neurodegeneration in laboratory settings, exert their toxic effect by disrupting the mitochondrial electron transport chain (Langston et al.

1983, Tanner et al. 2011). Moreover, a PD-associated loss of proteostasis had long been evident due to the presence of Lewy pathology. Therefore, it has become clear that there is a significant

5 overlap in the pathobiology of both sporadic and familial PD. Today, many of the proteins associated with familial PD, such as LRRK2, parkin, PINK1 and alpha-synuclein, are also considered to be important to the onset and progression of sporadic PD (Beilina et al. 2014). This is encouraging from a clinical perspective, as therapies targeting these proteins and pathways could benefit a larger proportion of PD patients.

1.1.3 Alpha-synuclein and Proteostasis in PD

The protein aggregation phenotype observed in PD is not unique. Many other neurodegenerative diseases, such as Alzheimer’s disease (AD), Huntington’s disease (HD) and

Amyotrophic Lateral Sclerosis (ALS), are also characterized by the presence of abnormal protein aggregates. In turn, these neurological disorders have become referred to as ‘proteinopathies’

(Kalia et al. 2013). The protein aggregates in each of these neurodegenerative proteinopathies are composed of a heterogeneous set of constituents, however, each disease typically has a particular protein that is specific to its aberrant inclusions. Examples include amyloid-beta and tau in AD, huntingtin in HD, TDP-43 in ALS and alpha-synuclein in PD. Therefore, neurodegenerative proteinopathies are further subdivided by the type of protein aggregate they are associated with.

Due to the presence of synuclein-rich aggregates in PD, dementia with Lewy bodies (DLB) and multiple system atrophy (MSA), these diseases are referred to as synucleinopathies (Kalia et al.

2013, Wong and Krainc 2017).

Alpha-synuclein is a 14kDa protein that lacks a stable three-dimensional structure, and is therefore considered ‘disordered’ (Wang et al. 2016). Alpha-synuclein composes 1% of total cytosolic protein in the nervous system (Stefanis 2012), predominantly localizes to pre-synaptic terminals, and plays a role in vesicle dynamics and neurotransmitter release (Bendor, Logan, and

Edwards 2013, Iwai et al. 1995). However, despite two decades of research, there is still much uncertainty about its physiological function.

Alpha-synuclein can exist as a monomer, or associate into larger assemblies, such as tetramers or oligomeric species. Oligomers can further associate into fibrillar structures with a β- pleated sheet conformation (‘fibrils’), and then into the large aggregates observed by neuropathologists (Kalia et al. 2013). In PD, alpha-synuclein has an increased propensity to form larger assemblies, including oligomers, fibrils and insoluble aggregates, which has solidified the notion that alpha-synuclein aggregation contributes to the pathogenesis of PD (Kalia et al. 2013).

Indeed, PD-causing mutations and multiplications of the SNCA gene typically promote alpha- synuclein oligomerization and aggregation (Lazaro et al. 2014). Furthermore, mutations in other genes associated with PD, such as LRRK2, also promote alpha-synuclein aggregation (Xiong et al. 2017). Due to the stereotypical presence of Lewy pathology, it was originally hypothesized that large aggregates underpinned the dopaminergic neurodegeneration observed in the SNpc.

However, as we have gained a better understanding of alpha-synuclein, this hypothesis has become increasingly challenged (Friesen et al. 2017, Kalia et al. 2013).

Many investigators have now demonstrated that the smaller oligomeric forms of alpha- synuclein confer more toxicity than the larger aggregates. Indeed, a synthetic E57K alpha- synuclein mutation, which increases oligomer formation, enhanced dopaminergic neurodegeneration in the rat SN relative to synuclein variants that more quickly associate into fibrils (Winner et al. 2011). In addition, the introduction of A53T mutant alpha-synuclein, which more rapidly forms oligomers and fibrils than wild-type synuclein, enhanced mitochondrial dysfunction and cell death in rat primary cortical neurons, neurite defects in C elegans, and dopaminergic neurodegeneration in Drosophila (Karpinar et al. 2009). These effects were even

7 more pronounced with another alpha-synuclein mutant (A30P/A56P/A76P), which readily forms oligomers but is impaired in transitioning into larger fibrils (Karpinar et al. 2009).

The toxicity of alpha-synuclein oligomers is further illustrated by their capacity to propagate between neurons in a prion-like fashion (Chu and Kordower 2015). For example, following the introduction of synthetic alpha-synuclein fibrils or Lewy body fractions from PD patients into rodent brains, Lewy pathology develops within the endogenous pool of alpha- synuclein (Jones et al. 2015, Karampetsou et al. 2017). Moreover, in human clinical trials that have aimed to enhance dopamine tone by introducing fetal dopaminergic neurons into the striatum of PD patients, Lewy pathology is seen to spread from the host to the grafted cells (Li et al. 2008). However, there is some controversy surrounding this finding as many patients did not demonstrate Lewy pathology in grafted cells, and for those that did, LB like inclusions were often found in a small subset of the grafted cells (Cooper et al. 2009). Extracellular alpha- synuclein oligomers can also form pores in the plasma membrane of neurons (Volles et al. 2001), which may indicate a mechanism by which they propagate between neurons or confer neuronal toxicity. Interestingly, the notion that alpha-synuclein pathology can self-propagate within the nervous system supports the plausibility of Braak’s hypothesis, which postulates that PD starts in the gut and moves up to the brain via the vagus nerve (Visanji et al. 2013).

Despite these advances in understanding, the movement of alpha-synuclein between different aggregation states is a complicated and nuanced process that is not yet fully understood.

Moreover, while it is clear that different alpha-synuclein species have unique physiological functions and toxicities, there is still much to be learned about what these are. Alpha-synuclein may have variable importance to the pathobiology of different types of PD, which is evidenced by the fact that some patients do not demonstrate Lewy pathology post mortem (Schneider and

Alcalay 2017, Kalia et al. 2015). It is also possible that alpha-synuclein oligomerization and aggregation is not itself disease causing, but rather the byproduct of an already dysfunctional cellular environment. It will be important to better understand these dynamics as we move towards disease-modifying therapies that target alpha-synuclein.

1.1.4 Mitochondrial Dysfunction in PD

As mentioned above, mitochondrial dysfunction is another common theme in the pathobiology of both sporadic and familial PD. The hypothesis that mitochondrial dysfunction contributes to disease onset was originally brought about by the observation that PD-associated environmental toxins often impair complex I of the electron transport chain (Tanner et al. 2011,

Schapira et al. 1990). This hypothesis was further supported by recent investigations into the familial PD-associated proteins, PINK1 and parkin, which have important roles in maintaining health within the mitochondrial network.

As mentioned above, PINK1 and parkin promote the clearance of damaged mitochondria via the lysosomes through a process termed mitophagy (Pickrell and Youle 2015). Briefly, at baseline conditions, PINK1, a serine/threonine kinase, is constitutively imported into the mitochondrial inner membrane, where it is cleaved and subsequently degraded. When the mitochondrial membrane potential becomes compromised, PINK1 is stabilized on the surface of the mitochondria where it phosphorylates a number of targets, notably including ubiquitin and parkin. These phosphorylation events trigger the recruitment of parkin to the outer mitochondrial membrane (OMM), where it ubiquitinates numerous OMM proteins. Autophagy ‘adaptor’ proteins, such as p62/SQSTM1 (hereafter p62), NBR1, and optineurin are then recruited to the

OMM, along with additional autophagy machinery. This results in the formation of a lipid

9 bilayer (autophagosome) around the mitochondria, and it is subsequently moved to the lysosomes for degradation. This process was recently reviewed by (Pickrell and Youle 2015).

PD-causing loss-of-function mutations in both PINK1 and parkin impair their function in this pathway, suggesting that the process of clearing damaged mitochondria is important in preventing against dopaminergic neurodegeneration (Pickrell and Youle 2015). This is not altogether surprising, as damaged mitochondria can be highly toxic, releasing reactive oxygen species (ROS) and apoptotic stimuli (ex. cytochrome C) into the intracellular environment. This, in turn, promotes DNA damage, a loss of proteostasis and ultimately, cell death. The oxidative stress caused by damaged mitochondria has been suggested to provide a mechanistic explanation of how mitochondrial dysfunction contributes to PD (Al Shahrani et al. 2017, Jenner et al. 1992).

Dopaminergic neurons are at an increased risk of oxidative stress, as dopamine metabolism itself can generate ROS and impair the electron transport chain (Chen et al. 2008). The combination of mitochondrial dysfunction with the increased oxidative load of dopaminergic neurons may exceed the cell’s antioxidant capacity, and explain why we observe a selective degeneration of dopamine neurons in PD.

The notion that mitochondrial dysfunction contributes to the pathogenesis of PD is not altogether separate from alpha-synuclein aggregation and proteostasis dysfunction. For example,

Devi and colleagues demonstrated that alpha-synuclein is targeted to and imported into the mitochondrial inner membrane, where it impairs complex I activity and enhances the generation of ROS (Devi et al. 2008). This cytotoxic activity of alpha-synuclein was enhanced by the A53T mutation. A more recent study demonstrated that alpha-synuclein impairs the import of mitochondrial proteins by disrupting the association of translocase of the outer membrane 20

(TOM20) with its co-receptor TOM22 (Di Maio et al. 2016). This also resulted in the increased production of ROS and a loss of the mitochondrial membrane potential.

The crosstalk between synuclein pathology and mitochondrial dysfunction illustrates that

PD is likely not caused by a single toxic cellular event or pathway, but rather by a combination of several interrelated dysfunctional processes. This complex nature of PD pathobiology makes it challenging to design disease-modifying therapies, as targeted therapies may fail to address all of the factors contributing to disease progression. It appears, then, that disease-modifying therapeutic strategies may require the simultaneous use of multiple interventions. An alternative strategy would be to identify proteins or pathways that sit at ‘connection points’ between the dysfunctional processes observed in PD. Such proteins likely represent the most efficient and effective therapeutic targets for PD, as multiple disease-contributing pathways could be modified using a single targeted therapy. One class of proteins that stands at the intersection between proteostasis, mitochondrial health and cell death are molecular chaperones. In turn, they have become regarded as a potentially powerful therapeutic target for PD (Friesen et al. 2017, K

Kalia, V Kalia, and J McLean 2010).

1.2 Molecular Chaperones in PD

1.2.1 The Nature and Function of Molecular Chaperones

Note: the following section is an excerpt from Friesen et al. (2017)

Molecular chaperones are highly conserved proteins that function to maintain proteostasis by directing the folding of nascent polypeptide chains, refolding misfolded proteins, and targeting misfolded proteins for degradation. Molecular chaperones are also termed ‘heat shock proteins’ (HSPs), as initial studies found them to be upregulated in response to high temperatures. In eukaryotes, HSPs are a large and heterogeneous group of proteins that have

11 been classified into families based on their molecular weight: Hsp40, Hsp60, Hsp70, Hsp90,

Hsp100, and the small HSPs (Kampinga and Bergink 2016). The activity of HSP family members is modulated by another class of proteins termed ‘co-chaperones’ which can be subdivided based on the presence of a Bcl-2 Associated Athanogene (BAG) domain, a tetratricopeptide (TPR) domain, or a J domain. Each of the families of chaperones and co- chaperones are composed of multiple proteins, which, despite having similar functions and domain compositions, often vary significantly in terms of their expression pattern and subcellular localization (Kampinga and Bergink 2016).

Due to the number and heterogeneity of chaperone and co-chaperone proteins, the nomenclature has become complex, with some chaperones receiving multiple names. As such, a new nomenclature was developed where DNAJ, HSPD, HSPA, HSPC, HSPH, and HSPB are the preferred prefix terms for the Hsp40, Hsp60, Hsp70, Hsp90, Hsp100, and small Hsp family members, respectively (Kampinga et al. 2009). For the purposes of this thesis, ‘Hsp’ will be used when referring to an entire family of Hsp chaperones and the new nomenclature will be used when referring to specific members within a family.

The two main chaperone machines in eukaryotes are Hsp70 and Hsp90, which together account for at least half of the molecular chaperones present in eukaryotic cells (Ciechanover and

Kwon 2017). The Hsp70 family members are the most studied molecular chaperones and have received significant attention in PD due to their abundance in Lewy bodies and their neuroprotective effect in pre-clinical models of the disease (K Kalia, V Kalia, and J McLean

2010). A subset of Hsp70 chaperones, namely HSPA1A, HSPA1B, and HSPA6, show stress- induced expression patterns, whereas the other Hsp70 family members, such as HSPA8 (often

12 referred to as Hsc70), are expressed constitutively at baseline conditions (Kampinga and Bergink

2016).

A signaling pathway involving the transcriptional activator, heat shock factor 1 (HSF-1), regulates the expression of inducible Hsp70 family members following stressful stimuli (Figure

1). At baseline conditions, HSF-1 is bound by Hsp90, maintaining HSF-1 in an inactive monomeric form (Zou et al. 1998). Following proteotoxic stress, HSF-1 dissociates from Hsp90 and translocates to the nucleus where it upregulates transcription of its target genes (Morimoto

1998). Once proteostasis is re-established, Hsp90 again sequesters HSF-1 into its inactive monomeric form, suppressing inducible Hsp70 expression. This crosstalk between chaperones and the presence of both constitutively active and stress-inducible chaperones on a negative feedback loop allows for the cell to execute continuous ‘house-keeping’ tasks in proteostasis, as well as respond to potentially devastating proteotoxic stress.

The primary role of Hsp70 is to ensure proper protein folding. Hsp70 accomplishes this by binding exposed hydrophobic domains on misfolded proteins (‘clients’) via its C-terminal substrate binding domain (SBD) and then undergoing multiple ATP hydrolysis cycles at the N- terminal ATPase domain (Rüdiger et al. 1997, Bukau and Horwich 1998). Hydrolysis of ATP to

ADP stabilizes the Hsp70-client complex, which allows for Hsp70 to hold the client protein and increases the likelihood of spontaneous refolding (Ciechanover and Kwon 2017). Subsequent

ADP-ATP exchange reduces the stability of the Hsp70-client complex, allowing for client dissociation or subsequent ATP hydrolysis cycles. While there are multiple models of the mechanism by which Hsp70 mediates protein refolding, the cycling between ATP and ADP bound states is necessary for this function (Goloubinoff and De Los Rios 2007).

The ATP hydrolysis cycle on Hsp70 is modulated by Hsp40, HSPH2 (Hsp110), the TPR domain-containing Hsp70 interacting protein (Hip), and BAG family co-chaperone proteins.

Hsp40s are important for both client selection and facilitating ATP hydrolysis (Kelley 1999), and

Hip stabilizes the ADP bound state of Hsp70 (Höhfeld and Jentsch 1997). Both BAG family members and HSPH2 act as nucleotide exchange factors (NEFs), promoting the release of ADP from the ATPase domain (Höhfeld and Jentsch 1997, Arakawa et al. 2010, Rampelt et al. 2012).

As such, Hsp40 and Hip promote Hsp70-client stability, whereas BAG family proteins and

HSPH2 destabilize the interaction. Therefore, the relative abundance of co-chaperone proteins play an important role in the dynamics of Hsp70 refolding activity. A complex interplay between the nature of the client protein, the Hsp70 family member, and the co-chaperone proteins present likely determines the efficacy and the mechanism by which a protein becomes refolded.

Outside of their primary function of protein refolding, molecular chaperones also play important roles in cellular processes such as guiding misfolded proteins for degradation through the UPS or ALP, disaggregating protein aggregates, suppressing cell death pathways, and promoting mitochondrial health (Figure 1). Hsp70-mediated protein degradation via the UPS is largely regulated by co-chaperone proteins, namely the C-terminal Hsp70 interacting protein

(CHIP), which is both an Hsp70 co-chaperone and an E3 ubiquitin ligase, thus providing a mechanistic link between the chaperone system and the UPS (Meacham et al. 2001, Murata et al.

2001). HSPA8 (Hsc70), in conjunction with lysosomal-associated membrane protein 2A

(LAMP2A) and multiple co-chaperones, can also facilitate protein degradation via the ALP through a process termed chaperone-mediated autophagy (CMA) (Figure 1) (Xilouri, Brekk, and

Stefanis 2016, Cuervo et al. 2004). Moreover, a chaperone machine composed of Hsp70, HSPH2

(Hsp110), and Hsp40 has a demonstrated ‘disaggregase’ activity by which it can remove

14 misfolded proteins from already formed aggregates (Gao et al. 2015, Nillegoda and Bukau

2015).

Figure 1

Lysosome Geldanamycin Ambroxol 17-AAG Isofagomine SNX Compounds Lysosomal Endoplasmic Re/culum Degrada/on GCase (Chaperone- HSPA5 HSF-1 Hsp90 Mediated (GRP78, BIP) Autophagy)

Proteotoxic Stress HSPA8 LAMP2A Co-chaperones Celastrol Protein Aggrega>on Arimoclomol Misfolded Protein Disaggrega>on Carbenoxolone

HSPA8 HSPH2 HSF-1 DNAJB1

Target Genes Inducible Hsp70 Mitochondria CHIP BAG Family Parkin Nucleus Members Other E3 Ligases HSPA9 (Mortalin) Proteasomal Degrada/on

*Figure from Friesen et al. (2017)

Figure 1 Proposed role of molecular and small molecule chaperones in proteostasis. At baseline, Hsp90 is bound to HSF-1, maintaining its inactive state. In the presence of proteotoxic stress, or the addition of Hsp90 inhibitors (i.e., geldanamycin, 17-AAG, SNX compounds), active HSF-1 dissociates from Hsp90 and translocates into the nucleus where it induces Hsp70 expression. Inducible Hsp70 family members direct proteasomal degradation through a pathway mediated by CHIP, Parkin, and other E3 Ligases. This process is inhibited by BAG family members and promoted by small molecule HSF-1 activators including celastrol and carbenoxolone. In response to proteotoxic stress, chaperones also direct misfolded proteins for degradation via the autophagy lysosome system, through interactions with various co-chaperones (chaperone-mediated autophagy). Chaperone/co-chaperone complexes can also function to disaggregate already formed protein aggregates. The pharmacological chaperones, ambroxol and isofagomine, increase glucocerebrosidase (GCase) activity in the lysosome to further promote the process of chaperone-mediated autophagy. Chaperone functions within the endoplasmic reticulum and mitochondria are regulated by the specific members of the Hsp70 family, HSPA5 and HSPA9, respectively.

1.2.2 Molecular Chaperones and Alpha-Synuclein Pathology

Note: the following section is an excerpt from Friesen et al. (2017)

Early evidence implicating molecular chaperones in the pathobiology of PD stemmed from the observation that Hsp70 overexpression attenuated alpha-synuclein mediated dopaminergic neurodegeneration in a Drosophila model (Auluck et al. 2002). This suggested that

Hsp70 may play a neuroprotective role in PD. Subsequently, McLean and colleagues illustrated that multiple chaperone proteins co-localize with Lewy bodies and that the overexpression of several Hsp40 and Hsp70 family members antagonize the formation of alpha-synuclein aggregates in vitro (McLean et al. 2002).

Molecular chaperones were further implicated in the pathobiology of PD through the observation that polymorphisms within the promoter region upstream of both constitutively expressed and inducible Hsp70 family members increases the risk of PD in a patient population

(Wu et al. 2004). Furthermore, mutations in the mitochondrial Hsp70, HSPA9 (mortalin), were recently suggested to promote the development of PD; however, other groups suggest mutations in HSPA9 are not a frequent cause of early-onset PD as they are also found in patient controls

(De Mena et al. 2009, Wadhwa et al. 2015, Burbulla et al. 2010, Freimann et al. 2013).

Since these initial studies, the capacity of Hsp70 overexpression to ameliorate alpha- synuclein aggregation and toxicity has been well characterized. Independent groups have shown that Hsp70 overexpression can attenuate alpha-synuclein-mediated cell death in yeast (Flower et al. 2005) and reduce high molecular weight aggregates and toxicity in rodent models of PD

(Moloney et al. 2014, Klucken, Shin, Masliah, et al. 2004). Hsp70 overexpression was shown to be protective against cell death mediated by the mitochondrial complex I inhibitor, MPTP, both in vitro (Quigney, Gorman, and Samali 2003) and in vivo (Dong et al. 2005). Although alpha-

16 synuclein aggregation is not a feature of this toxin model, alpha-synuclein is required for MPTP- induced cell death as demonstrated by the resistance of alpha-synuclein null mice to MPTP- induced neurotoxicity (Dauer et al. 2002).

In parallel with the Hsp70 overexpression results, recent studies have demonstrated that microRNA (miRNA) mediated translational repression of Hsp70 exacerbates alpha-synuclein aggregation and toxicity in vitro (Zhang and Cheng 2014) and that miRNAs targeting Hsp70 are upregulated in brain regions with Lewy pathology (Alvarez-Erviti et al. 2013). Furthermore, the

Hsp70 family members HSPA8 (Hsc70) and HSPA9 (mortalin) have lower expression in the SN

(HSPA8/9) (Alvarez-Erviti et al. 2010), and leukocytes (HSPA8) (Papagiannakis et al. 2015,

Sala et al. 2014) of PD patients relative to healthy controls, suggesting that chaperone levels and function may have a role in the pathogenesis of PD.

The mechanism by which Hsp70 attenuates alpha-synuclein aggregation and toxicity seems to be dependent on both its refolding activity and its function in protein degradation via the UPS and ALP. Mutations that alter the ATPase function of Hsp70 (K71S) abolish its protective effect on alpha-synuclein toxicity, indicating that Hsp70 folding activity is necessary for its protective function (Klucken, Shin, Hyman, et al. 2004). Hsp70/co-chaperone complexes also mitigate alpha-synuclein mediated toxicity by promoting the degradation of misfolded alpha-synuclein via either the UPS or ALP. Several studies have suggested that CMA may be playing an important role in mitigating alpha-synuclein toxicity and aggregation (Cuervo et al.

2004, Mak et al. 2010, Xilouri and Stefanis 2015). Enhanced alpha-synuclein expression in both transgenic and paraquat models of PD results in a concurrent enhancement of LAMP2A and

HSPA8 expression, and a greater movement of alpha-synuclein into the lysosomes (Mak et al.

2010). Moreover, both LAMP2A and HSPA8 have lower expression in the SN of PD patients

(Alvarez-Erviti et al. 2010), and a recent study demonstrated a correlation between the loss of

LAMP2A and alpha-synuclein aggregation in post-mortem PD brains (Murphy et al. 2015).

Interestingly, the observed decrease in LAMP2A and HSPA8 expression anatomically overlaps with an increase in miRNAs capable of translationally repressing both LAMP2A and HSPA8

(Alvarez-Erviti et al. 2013), further implicating miRNAs in PD-associated chaperone dysregulation.

Outside of CMA, the Hsp70 co-chaperone, CHIP, plays an important dual function in alpha-synuclein degradation, as it can target alpha-synuclein for degradation by either the proteasome or lysosome via its TPR domain or U-box domain, respectively (Shin et al. 2005).

CHIP may mediate this through ubiquitination of alpha-synuclein and suppression of oligomer formation (Kalia et al. 2011). However, not all Hsp70 co-chaperones promote alpha-synuclein degradation. In contrast, overexpression of the BAG family member, BAG5, antagonizes CHIP- mediated alpha-synuclein ubiquitination, which prevents the ability of CHIP to suppress oligomer formation and also enhances alpha-synuclein-mediated toxicity (Kalia et al. 2011).

Therefore, the balance between multiple co-chaperones may assist Hsp70 in triaging whether to refold or degrade a client substrate, and a disruption in the relative abundance or activity of co- chaperones may compromise the chaperone system and subsequently proteostasis.

1.2.3 Molecular Chaperones and Mitochondrial Dysfunction

The potential role of chaperones in the pathobiology of PD is broadened by their capacity to regulate the stability and function of PD-relevant proteins other than alpha-synuclein, particularly those that relate to mitochondrial dysfunction. For example, Hsp70 and Hsp90 family members regulate the stability of PINK1 and Parkin. Hsp90 regulates the processing and stability of PINK1, and the Hsp90 family member HSPC5, commonly known as TNF Receptor

Associated Protein 1 (TRAP1), promotes mitochondrial health and compensates for the mitochondrial dysfunction caused by PD-associated PINK1 mutations (Zhang et al. 2013).

Moreover, the Hsp70 family member HSPA1L and the co-chaperones BAG2 and BAG4 have all been shown to modulate PINK1-Parkin mediated mitophagy by effecting the translocation of

Parkin to damaged mitochondria (Hasson et al. 2013, Qu et al. 2015). Hsp70 supports Parkin by preventing it from being sequestered and by acting in concert with CHIP to promote its E3 ubiquitin ligase activity following proteotoxic stress (Imai et al. 2002). In contrast, the co- chaperone BAG5 inhibits Parkin E3 activity, which may provide a mechanistic explanation as to how BAG5 enhances dopaminergic neurodegeneration (Kalia et al. 2004b) (discussed below).

Taken together, the capacity of Hsp70 and its co-chaperones to manage toxic alpha- synuclein species and mitochondrial homeostasis pathways, indicates that molecular chaperones have a central and multi-faceted role in the pathobiology of PD. Moreover, since multiple chaperones are downregulated, sequestered into protein aggregates, or face age-related loss-of- function in the brains of PD patients, it seems likely that the depletion and dysfunction of molecular chaperones contributes to the progression of PD. As such, this class of proteins is regarded as being a potentially powerful therapeutic target for disease-modifying therapies in PD and other proteinopathies. Unfortunately, the chaperone network is a large and complex group of proteins, demonstrating context dependent changes in expression and function that are far from being completely understood. Therefore, much work remains to be done in understand how this network can be therapeutically targeted to restore proteostasis in disease.

1.2.4 BAG Family Co-chaperones

The BAG family of Hsp70 co-chaperone proteins are a particularly enticing therapeutic target within the chaperone system as they have not only been shown to play important roles in

19 the management of proteostasis and mitochondrial dynamics, but also in cell survival and death pathways. There are six members of the bcl-2 associated athanogene (BAG) family, named

BAG1-6 based on the order in which they were discovered (Kabbage and Dickman 2008a). The first member of this family, BAG1, was identified in 1995 during a screen for bcl-2 interactors, and was shown to promote cell survival due to its synergistic relationship with the oncogene bcl-

2 (Takayama et al. 1995). Shortly thereafter, BAG1 was deemed a co-chaperone protein, as it was found to interact with and modulate the function of Hsp70 (Takayama et al. 1997).

BAG family members are characterized by the presence of a 110-124 amino acid BAG domain at their C-terminus (Figure 2). The BAG domain consists of three anti-parallel amphipathic alpha helices and mediates an interaction with a multitude of proteins including

ATPase domain of Hsp70 (Kabbage and Dickman 2008b). This interaction allows BAG proteins to modulate the activity of Hsp70 by promoting ADP/ATP exchange, as discussed in section

1.2.1. In terms of cell death, most studies have demonstrated that BAG family members promote cell survival. Indeed, many studies have implicated BAG family members in tumorigenesis, due to their potent anti-apoptotic effect, and, in some cases, their support of gain-of-function p53 mutants (Behl 2016, Yue et al. 2016). However, there are some context dependent exceptions to this phenotype. An increasing body of evidence demonstrates that BAG family members, namely

BAG2 (Qu et al. 2015) and BAG5 (Kalia et al. 2004a), can promote cell death in certain contexts.

The notion that BAG family members can promote cell survival in some contexts and death in others has introduced some confusion surrounding their function in cell death pathways and other homeostatic cellular processes. Rather than definitively promoting cell death or survival, it is emerging that BAG family members may exert context-dependent or modulatory

FIGURE 2

Figure 2 Schematic diagram of the six BAG family members. Adapted from Kabbage et al. (2008). Each BAG family member contains the prototypical C-terminal BAG domain that facilitates their interactions with Hsp70. BAG5 is unique in that it contains five BAG domains rather than one. BAG1 has three isoforms that differ in their length and are therefore termed BAG1S, BAG1M and BAG1L. BAG1 and BAG6 contain an N-terminal ubiquitin- like domain (UBL).

effects on apoptosis (De Snoo et al. in preparation). Therefore, as is the case with the chaperone network as a whole, there is need to better understand the mechanistic nuances of BAG family members in order to understand if and how they can be manipulated to restore cellular health and proteostasis in disease. There is an interesting dynamic at play in the BAG protein literature, as

21 cancer researchers are trying to understand how these proteins can be targeted to promote cell death, while neurodegenerative disease researchers are aiming for the opposite effect.

1.2.5 BAG5

BAG5 is a unique member of the BAG family in that it contains five BAG domains rather than one, however, it is still the C-terminal BAG domain that mediates its association with

Hsp70 (Arakawa et al. 2010). BAG5 is of particular interest because it has been implicated in the pathobiology of PD. In the first characterization of BAG5, Kalia and colleagues found that it inhibits the protein folding capacity of Hsp70, impairs the E3 ubiquitin ligase function of parkin, and enhances dopaminergic neurodegeneration in a rodent model of PD (Kalia et al. 2004a).

Additional unpublished work by our lab suggests that BAG5 may enhance cell death by promoting pro-death kinase signaling pathways. These results have made BAG5 an enticing therapeutic target for PD, as reducing its levels or activity would foreseeably stimulate the anti- apoptotic functions of Hsp70 and parkin, suppress cell-death pathways and, ultimately, promote neuronal survival.

Several other lines of evidence have implicated BAG5 in the pathobiology of PD. First,

BAG5 enhances alpha-synuclein oligomerization by inhibiting the E3 ligase activity CHIP, which normally ubiquitinates alpha-synuclein to promote its proteasomal degradation (Kalia et al. 2011). Second, BAG5 interacts with PINK1, which, in conjunction with its inhibitory effect on parkin, suggests that BAG5 may play a role in the maintenance of mitochondrial homeostasis via mitophagy (Wang et al. 2014). Indeed, preliminary evidence from our lab suggests that

BAG5, like BAG2 and BAG4, does modulate the ability of parkin to translocate to damaged mitochondria (De Snoo et al. in preparation). Third, BAG5 promotes autophagy dysfunction on a

22 larger scale, as it forms a complex with LRRK2 that promotes the clearance of the trans-golgi network: an activity that reduces the functional capacity of the ALP (Beilina et al. 2014).

The fact that BAG5 plays a modulatory role in the ALP, UPS, chaperone system and cell death pathways has solidified the notion that it is likely a valid therapeutic target for PD.

However, despite this potential, the functional characteristics of BAG5 are far from being fully characterized. Like the other BAG family members, while initial evidence demonstrated that

BAG5 promotes cell death, recent studies have shown that BAG5 can promote cell survival in a number of in vitro contexts (Bi et al. 2016b, Bruchmann, Roller, Walther, Schäfer, et al. 2013,

Guo et al. 2015b, Gupta et al. 2016, Ma et al. 2012, Wang et al. 2014). Therefore, some work remains to be done in determining the precise function of BAG5 in both wild type and pathological settings before moving forward with the development of a BAG5-targeted therapy.

One option to better understand the function of BAG5 is to characterize its interactome.

This sort of proteomic screen provides insights into the pathways and processes that a protein interacts with, which can stimulate new hypotheses about its function and relevance to certain disease states. Interactome studies can also serve to uncover new protein interactions that, through additional investigation, lead to the generation of more fine-tuned mechanistic insights of a protein’s function. As such, interactome studies are recognized as a powerful tool in understanding the nuances and relationships within the chaperone network, including those of

BAG5. Several studies have already used such proteomic screens to investigate the ‘chaperome’

(Hadizadeh Esfahani et al. 2018, Taldone et al. 2014). For example, Chen and colleagues characterized the BAG3 interactome, which allowed them to uncover a novel function of BAG3 in modulating proteasome activity (Chen et al. 2013). Therefore, characterizing the BAG5

23 interactome appears to be an appropriate next step in understanding its function both within and outside of the context of PD.

1.3 Summary and Research Objectives

Due to their engagement with multiple cellular processes relevant to the pathobiology of neurodegenerative proteinopathies, molecular chaperones have arisen as a potentially powerful therapeutic target. The feasibility of chaperone-based therapies is complicated by the fact that molecular chaperones exist in a complex and changing proteomic network, making them a difficult target to manipulate pharmaceutically. As such, there is an emerging need to characterize the nuances of the chaperone network in order to best understand how therapies can be tailored to restore proteostasis. While there has been significant progress made in understanding chaperone proteomics, there still remains much work to be done.

BAG5 exemplifies the characteristics of the larger chaperone network in that its effect on proteostasis and cell death changes with different contexts, likely due to its transient and changing interactions with proteins both within and outside of the chaperone network. While BAG5 has been hypothesized to be important to the pathobiology of PD, conflicting results surrounding its role in cell death have confused its suitability as a therapeutic target. As such, as is the case with the molecular chaperone network as a whole, it is imperative to understand the regulatory proteomic network that surrounds BAG5 in order to understand if and how it can be targeted by disease-modifying therapies.

In order to address this need, the primary aim of this thesis is to characterize the BAG5 interactome to get a better sense of the proteins and pathways that BAG5 associates with. We hypothesized that by characterizing the BAG5 interactome we would uncover novel interactions relevant to the regulation of proteostasis in PD. Such interactions would have the potential to

24 extend the role of BAG5 in proteostasis beyond what is currently known, and, in turn, provide insights into the molecular function of BAG5 in both physiological and pathological contexts.

Therefore, the secondary and tertiary aims of this project were to identify novel and interesting

BAG5 interactions and explore the function(s) of these interactions in the context of PD.

This manifested into the following three specific research aims that will be discussed over the next three chapters:

1. Characterize the BAG5 interactome in H4 and SH-SY5Y cells

2. Validate a novel interaction identified between BAG5 and p62

3. Understand the function of the BAG5-p62 interaction in the context of alpha-

synuclein aggregation

Chapter 2 Characterizing the BAG5 Interactome 2.1 Introduction

As discussed in the introductory chapter, interactome analyses can serve as an important starting point for understanding a protein’s function. Because there is a lack of clarity surrounding the function of BAG5 in both wild type and pathological settings, the primary goal of this thesis was to characterize the BAG5 interactome. We opted to first investigate the BAG5 interactome in the H4 neuroglioma cell line, as we, and others, have previously used this cell line to interrogate molecular pathways relevant to PD (Kalia et al. 2011, McLean et al. 2002, McLean et al. 2004). H4 cells lines were stably transfected with GFP or GFP-BAG5, and BAG5 interacting proteins were analyzed by immunoprecipitating the GFP transgenes and identifying co-immunoprecipitated proteins via mass spectrometry.

Because BAG5 strongly associates with Hsp70, we foresaw that assessing the BAG5 interactome would be challenging, as it would be difficult to discern whether identified proteins were true interactors, or transient, non-specific cargo of the Hsp70 machinery. As such, we generated an additional cell line stably expressing GFP-BAG5DARA, a previously described

BAG5 mutant incapable of binding to Hsp70 (Kalia et al. 2004a), to gauge whether or not interactions were dependent on the presence of Hsp70. The DARA mutant has key aspartate and arginine residues in four of the BAG domains mutated to alanine to effectively disrupt its interaction with Hsp70.

Following the characterization of the BAG5 interactome in H4 cells, we aimed to solidify the data and understand whether the results could be generalized to other cell types. In collaboration with the Schmitt-Ulms Lab, we carried out a second analysis of the BAG5

25 26 interactome in SH-SY5Y cells. This dopaminergic neuroblastoma cell line was chosen because it is considered to be a better in vitro model of dopaminergic neurons than other popular immortalized cell lines (HEK293, HeLa, etc.) and more conducive to genetic manipulation than primary neuronal cultures.

SH-SY5Y cell lines stably expressing inducible GFP, GFP-BAG5 and GFP-BAG5DARA were generated by inserting the GFP transgenes into the AAVS1 safe harbor. This insertion strategy was used to account for the effects of random genomic integration of the transgene, which may have confounded the results observed in the H4 interactome. Like the H4 interactome, the GFP-tagged proteins were immunoprecipitated and co-immunopreicpitated proteins were gauged using mass spectrometry. However, unlike the H4 interactome, the SH-

SY5Y mass spectrometry analysis made use of isobaric tags for absolute and relative quantification (iTRAQ) in order to gauge differences in a protein’s binding preference for either

BAG5 or BAG5DARA. The iTRAQ 8-plex reagents allow for peptides from 8 different samples to be covalently linked to 8 unique tags that can be detected by the mass spectrometer. In turn, for every identified protein, the relative abundance of the iTRAQ tags serves as a surrogate measure of the relative abundance of that protein across the eight samples.

The generation of the SH-SY5Y cell lines used for this screen required the use of elegant genetic editing techniques, represented a significant portion of the work for this thesis, and resulted in the production of a sophisticated tool for future in vitro assays used in our lab. As such, a more detailed description of the transgene insertion strategy and rationale is discussed here before moving on to the results of the proteomic screens. The theoretical approach used to insert the transgenes was entirely developed by the lab of Dr. Schmitt-Ulms, with a significant amount of the work done by Xinzhu (Louisa) Wang.

Outline of the GFP transgene Insertion Strategy (see Figure 3)

1. Stably introduce lox71 and lox2272 into intron 1 of the AAVS1 safe harbor of SH-SY5Y

cells using CRISPR-Cas9

2. Generate a replacement plasmid containing inducible GFP, GFP-BAG5 or GFP-BAG5DARA

flanked by lox66 and lox2272

3. Stably insert GFP, GFP-BAG5 or GFP-BAG5DARA into intron 1 of the AAVS1 safe harbor of

SH-SY5Y cells using LE/RE cre-recombination

1. Stably introduce lox71 and lox2272 into intron 1 of the AAVS1 safe harbor of SH-SY5Y cells using CRISPR-Cas9

Selection of the AAVS1 Safe Harbour

In order to minimize the unknown effects of random genomic integration, researchers have made use of genomic safe-harbours. Here, a transgene is inserted into a known genomic location where the effect on endogenous cellular function is minimized. While there is some controversy as to what constitutes a genomic safe harbor (ie. within housekeeping genes, intragenic, extragenic, etc.), there are three safe harbours commonly used in human cell lines:

CCR5 (chromosome 3p21.31), ROSA26 (chromosome 3p25.3) and AAVS1 (chromosome

19q13.42) (Sadelain, Papapetrou, and Bushman 2011).

A transgene can be inserted into these regions by capitalizing on DNA repair machinery that introduces new genetic material (from a replacement plasmid) following a double-strand break (DSB) mediated by endonucleases such as zinc-finger nucleases (ZFNs), transcription activator-like effector nucleases (TALENs), or more recently, Cas9 (Sadelain, Papapetrou, and

Bushman 2011). Considering that several studies have successfully used CRISPR-Cas9 to insert

FIGURE 3

Figure 3 Generation of SH-SY5Y stable cell lines. Schematic of the strategy used to insert the GFP, GFP-BAG5, and GFP-BAG5DARA transgenes into the AAVS1 safe harbour of SH-SY5Y cells. Briefly, lox71, lox2272 and a Kanamycin selection marker (KanR) were inserted into the intron 1 of the AAVS1 safe harbour using CRISPR-Cas9 homology directed repair with two 800 base pair homology regions. The GFP transgenes (with tetracycline- responsive element (TRE) promoter), along with the genetic machinery necessary to allow for tetracycline inducible expression (rtTA3 with cytomegalovirus (CMV) promoter) and a Puromycin selection marker (PuroR), were flanked with lox66 and lox2272 to allow for them to be swapped into the safe harbour in the presence of cre recombinase. The recombination of lox71 & lox66 creates a non-functional lox site, which is referred to as a left element/right element (LE/RE) strategy, and prevents the transgenes from being swapped out of the safe harbour by residual cre recombinase.

29 transgenes into intron 1 of the AAVS1 locus without toxic or gene silencing effects (Oceguera-

Yanez et al. 2016), this site was selected by the Schmitt-Ulms lab to insert the transgene. The

AAVS1 locus encodes the ‘protein phosphatase 1 regulatory subunit 12C’ (PPP1R12C), which currently has no known function.

Use of CRISPR-Cas9 to edit the AAVS1 locus

The parent cell line that we used to generate the inducible GFP, GFP-BAG5 and GFP-

BAG5DARA cell lines were edited with CRISPR-Cas9 by X. Wang of the Scmitt-Ulms lab. A paired nickase strategy to insert two lox sites (lox71 and lox2272, discussed in more detail below) into the AAVS1 locus (Ran et al. 2013). In order to achieve this, two pairs of gRNAs that are known to have little off-target effect were used to generate a double strand break in intron 1 of the AAVS1 locus, and a repair plasmid was inserted into the locus via homology-directed repair.

2. Generate a replacement plasmid containing inducible GFP, GFP-BAG5 or GFP-

BAG5DARA flanked by lox66 and lox2272

Creating an Inducible Transgene using the TetON System

The TetON system is a genetic strategy used to control gene expression through the use of tetracycline or tetracycline analogs, such as doxycycline. There are two genetic features required to allow for inducible transgene expression: (1) a gene driven by a Tet-responsive element (TRE) promoter, and (2) a transcription regulator protein that interacts with the TRE promoter in the presence of tetracycline/doxycycline (Das, Tenenbaum, and Berkhout 2016). In the presence of tetracycline, the transcription regulator protein, called ‘reverse transactivator

(rtTA)’, undergoes a conformational change that allows it to associate with the TRE promoter.

This interaction subsequently drives the expression of the gene downstream of the TRE.

Since the initial development of the TetON system, there has been a substantial improvement in the functional capacity of the system through the introduction of new TREs and rtTAs (Das, Tenenbaum, and Berkhout 2016). As such, these cell lines use the upgraded

‘TREtight’, which reduces the leakiness of the TRE promoter by 4.4X, and rtTA3, which is 25X more sensitive to tetracycline/doxycycline than the original rtTA.

Features of the EGFP-BAG5 replacement plasmid

In order to pursue the LE/RE cre-recombination strategy (discussed below), the entire transgene to be inserted into the AAVS1 locus had to be flanked by lox66 and lox2272. The transgene contained a puromycin resistance gene PuroR for the selection of stable lines.

Downstream and in antisense of the PuroR gene were the GFP transgenes, i.e., GFP, GFP-BAG5, or GFP-BAG5DARA, as well as the tetON components TREtight and rtTA3 (CMV driven). GFP-

BAG5 and GFP-BAG5DARA both had GFP fused to their N-terminus, in order to minimize any effect on the C-terminal BAG domain, which is known to mediate the association between

BAG5 and Hsp70.

3. Stably insert GFP, GFP-BAG5 or GFP-BAG5DARA into intron 1 of the AAVS1 safe harbor of SH-SY5Y cells using cre recombination

Cre-recombinase can be used for a number of different functions depending on the nature of the lox sites present (Sauer 2002). Lox sites are 34 base pair (bp) genetic elements that have two Cre-binding sites separated by an 8 bp spacer. If the appropriate lox sites flank two genes, for example one gene in the genome and one on a plasmid, cre-recombinase has the capacity to

31 exchange the two genes so that a transgene on the plasmid can be effectively swapped into the genome (Sauer 2002).

One problem with cre-recombination events between the genome and a plasmid is the possibility of gene reversion, where the gene that was swapped into the genome is moved back into the plasmid by cre-recombinase. This is especially common when the same pairs of lox sites are used on both the plasmid and genome. The left-element/right-element (LE/RE) strategy can be used to minimize gene reversion (Araki, Araki, and Yamamura 2002). Here, lox66 and lox71 are used as the first (downstream) lox sites on the replacement plasmid and genome, respectively, and a common lox site, lox2272, is used as the second (upstream) lox site. Lox66 and lox71 contain mutations to their right and left Cre-binding sites, respectively. Therefore, prior to gene swapping, both lox66/lox71 have one functional and one non-functional cre- binding site, which allows for them to still act as functional lox sites. However, after the swap, one lox66/71 site will have doubly mutated cre-binding sites, lowering the possibility of gene reversion.

2.2 Materials & Methods

2.2.1 Antibodies & Reagents

Antibodies: Anti-Actin (A2066) was obtained from Sigma-Aldrich. Anti-GFP (A11122) was obtained from Invitrogen. Horseradish Peroxidase Linked ECL Anti-Rabbit and Anti-Mouse secondary antibodies were obtained from GE Healthcare. Reagents: Geneticin (G418, 1013027) was obtained from Gibco. Puromycin hydrochloride (PUR555) was obtained from Bioshop.

Doxycycline hydrochloride (DB0889) was obtained from Biobasic.

2.2.2 Cell Culture

SH-SY5Y and H4 cells were cultured in Dulbecco’s Modified Eagle Medium (DMEM,

Gibco) supplemented with 10% fetal bovine serum (FBS, Gibco), 1% antibiotic/antimycotic

(Gibco), and incubated at 37°C with 5% CO2. SH-SY5Y and H4 cells were exclusively grown on cell+ plates (Sarstedt). SH-SY5Y cells were transfected using Lipofectamine 2000 (Thermo

Fisher), and H4 cells were transfected using the SuperFect Transfection Reagent (Qiagen), as per the manufacturer’s protocol.

2.2.3 Western Blotting

H4 and SH-SY5Y cells were lysed with Triton X-100 based radioimmunoprecipitation assay (RIPA) buffer containing 50 mM Tris, 150 mM NaCl, 0.5% sodium deoxycholate, 1%

Triton X-100 and 1X protease inhibitor cocktail (cOmplete, Roche). Protein concentration was quantified using the DC (Bradford) protein assay (BioRad). 20µg of protein lysate was loaded into each lane of a 4-15% acrylamide gels (BioRad) and transferred onto a polyvinylidine fluoride (PVDF) membrane. Blots were blocked with 5% skim milk diluted in tris-buffered saline + 0.01% Tween-20 (TBS-T) for 30 minutes prior to incubation with primary antibody for either 1 hour at 21°C or overnight at 4°C. Blots were subsequently washed three times in TBS-T (10 minutes per wash), incubated in species specific secondary antibody for 1 hour at 21°C, washed again, and then developed using ECL plus western blotting substrate

(Pierce) and visualized on HyBlot CL autoradiographic film (Denville Scientific).

2.2.4 Generation of the H4 Stable Cell Lines

Wild-type H4 neuroglioma cells were stably transfected with GFP, GFP-BAG5 or GFP-

BAG5DARA plasmids, all of which were originally derived from the pEGFP-C1 plasmid

(Clontech, Accession #: U55763). Transfected cells were transferred to selection media

33 containing 700µg/mL G418 (Geneticin) 24 hours after transfection, and were retained in selection media for 14 days. Colonies that reached the size of 100-200 cells were assessed for

GFP-transgene incorporation using fluorescence microscopy. Colonies stably expressing the transgene were transferred to a 96-well plate and grown up to 10cm plates for characterization.

2.2.5 Immunoprecipitation and Mass Spectrometry: H4 Cells

H4 stable lines containing GFP, GFP-BAG5 and GFP-BAG5DARA were assessed for transgene expression with western blot and similarly expressing clones were chosen to move forward with mass spectrometry. The selected cell lines were lysed with RIPA buffer and 1mg of cell lysate was combined with 25uL GFP-trap bead slurry (Chromotek) and rotated at 4°C for 90 minutes. Beads were washed three times with 1mL of a buffer composed of 10mM Tris Hcl (pH

7.5), 150mM NaCl and 0.5mM EDTA. Beads were immediately frozen at -20°C and transported to the SPARC Biocenter at the Hospital for Sick Children. Protein was trypsin-digested directly off of the beads and analyzed by liquid chromatography and mass spectrometry (LC-MS/MS) on an OrbiTrap Elite.

2.2.6 Generation of the SH-SY5Y Stable Cell Lines

CRISPR-Cas9 edited parent SH-SY5Y cell lines were generously provided to us by the laboratory of Dr. Schmitt-Ulms. Genomic PCR and sequencing analysis done by the Schmitt-

Ulms lab revealed that the parent line was heterozygous for the KanR gene flanked by lox71/2722 sites. The laboratory of Dr. Schmitt-Ulms also provided us with the SH-SY5Y negative control cell line containing doxycycline inducible GFP.

We derived our GFP-BAG5 vectors from a GFP-Tau plasmid that was also provided to us by the laboratory of Dr. Schmitt-Ulms. The plasmid initially contained (1) C-terminal GFP-

Tau with a tetracycline responsive element (TREtight) promoter, (2) rtTA3 with a CMVmini

34 promoter, and (3) a eukaryotic puromycin resistance gene (PuroR), all flanked by lox66 and lox2722. GFP-Tau was excised from the plasmid using the endonucleases BbvCI and AflII

(NEB), as per the manufacturer’s protocol. N-terminal GFP-tagged BAG5 and BAG5DARA inserts were generated via PCR from the N-terminal GFP-BAG5 and GFP-BAG5DARA vectors used to generate the H4 cell lines (originally derived from the pEGFP-C1 vector, Clontech, Accession #:

U55763). The inserts were ligated into the replacement plasmid using the In-Fusion HD Cloning

Kit (Clontech), as per the manufacturer’s protocol. Proper insertion of the transgenes was verified by sequencing (ACGT Corp., Toronto, ON).

The inducible GFP-BAG5 and GFP-BAG5DARA replacement plasmids were transfected into the SH-SY5Y parental cell line alongside a plasmid containing the improved Cre- recombinase (iCre) gene. The ratio of transgene plasmid to iCre plasmid was 2:1. Transfected cells were transferred into selection media containing 1.2µg/mL puromycin 24 hours post- transfection. Stable lines were incubated in selection media for 7 days, and then transferred back into regular media for an additional 7 days to allow for colonies to grow to 100-200 cells.

Successful incorporation of the inducible transgene was assessed by adding 2ug/mL doxycycline to the media for 6 hours and assessing green colonies with fluorescence microscopy. Colonies stably expressing the transgene were transferred into a 96-well plate and grown up to 10cm plates for characterization.

2.2.7 Immunoprecipitation and Mass Spectrometry: SH-SY5Y Cells

SH-SY5Y inducible cell lines (GFP, GFP-BAG5 and GFP-BAG5DARA) were treated with

2μg/mL dox for 18h prior to lysis in a digitonin lysis buffer containing 50μg/ml digitonin, 150 mM Tris-HCl (pH 7.5), 4 μg/ml aprotinin, 5 mM EDTA, 1 mM EGTA, 10 mM NaF, 1 mM Na orthovanadate, 1 mM PMSF, protease inhibitor cocktail (cOmplete, Roche), and phosphatase

35 inhibitor cocktail (PhosSTOP, Roche). 3mg cell lysate was combined with 50uL GFP-trap bead slurry (Chromotek) and rotated at 4°C for 90 minutes. Beads were washed four times with 2mL of a buffer consisting of 150 mM Tris-HCl and 10% glycerol, once with 2mL of 25mM HEPES and once with 2mL of 10mM HEPES. Protein was eluted from the beads using a buffer composed of 20% acetonitrile and 1% trifluoroacetic acid (pH 1.9).

All MS sample preparation and experimentation was done by either X. Wang or D.

Williams of the Scmitt-Ulms lab. Briefly, eluted proteins from 3 GFP-BAG5 IP replicates, 3

GFP-BAG5DARA IP replicates and 3 GFP IP replicates were prepared for mass spectrometry through reduction and alkylation of di-sulfide bonds, trypsin digestion, and labeling with 8-plex iTRAQ reagents (Sigma-Aldrich). Peptides and relative iTRAQ label abundance was subsequently identified by OrbiTrap LC-MS/MS/MS (third MS run required for the identification of the iTRAQ label). Protein content in each sample was normalized to overall abundance of GFP in order to perform relative quantification using the iTRAQ labels. Relative iTRAQ peptide quantification was performed using label 118 as the reference sample.

2.2.8 Bioinformatic Analysis

For the H4 interactome, UniProt IDs from all proteins binding to GFP-BAG5, GFP-

BAG5DARA, or both (total: 402) were converted into UniProt gene names using the UniProt

Retrieve/ID mapping tool (https://www.uniprot.org/uploadlists/). Gene names were subsequently run through the GeneAnalytics software hosted by the GeneCards website

(http://geneanalytics.genecards.org), which calculated Gene Ontology and SuperPath enrichments in addition to several other bioinformatic outputs. The GeneAnalytics software also produced FDR-adjusted p-values demonstrating the significance of each enrichment term.

For the SH-SY5Y interactome, the raw interactome list was filtered to exclude proteins that did not have iTRAQ quantification in all eight channels, those that had an abundance ratio of

>0.7 in channels 119 and 121 (GFP negative controls), and those that did not have a “High” FDR confidence (Exp. q value > 0.01,). The remaining 217 proteins (219 total, BAG5 and BAG5DARA were excluded) were processed using the GeneAnalytics software in the same way as the H4 interactome.

2.3 Results

2.3.1 Characterization of the BAG5 Interactome: H4 Cells

The stable incorporation of the GFP, GFP-BAG5 and GFP-BAG5DARA transgenes into the

H4 genome was confirmed by western blot (Figure 4). Due to random integration of the transgenes into the host genome, there was variability in transgene expression between clones.

Clones with similar expression of GFP, GFP-BAG5 and GFP-BAG5DARA, relative to the loading control, were selected to move forward with the mass spectrometry analysis of the BAG5 interactome (Figure 4).

The interactome identified 402 BAG5 interacting proteins, of which 173 specifically bound to GFP-BAG5, 92 to GFP-BAG5DARA , and 137 to both (Figure 5, top). Of the 137 binding to both BAG5 and BAG5DARA, 13 had a 50% lower affinity for BAG5DARA, by total spectrum counts, and were re-classified as BAG5 interacting proteins (Appendix 1). The top ten proteins binding to BAG5, as well as those binding to both BAG5 and BAG5DARA are listed in

Table 2. The former are considered to be dependent on the association between BAG5 and

Hsp70, and the latter are not. The complete list of BAG5 and BAG5DARA interacting proteins are listed in Appendix 1.

Consistent with the notion that the DARA mutation effectively disrupts the association between BAG5 and Hsp70, BAG5, but not BAG5DARA, was shown to interact with both Hsp70

(Table 2) and the C-terminal Hsp70 interacting protein (CHIP; Appendix 1). BAG5 interacted with 5 proteins associated with familial PD, namely UCH-L1 (PARK5), DJ-1 (PARK7), EIF4G1

(PARK18), DNAJC13 (PARK21) and VPS13C (PARK23). Several of these interactions had already been suggested on protein interaction databases. Indeed, our interactome had a 15% overlap with the BAG5 interacting proteins listed on the GeneCards database, and a 20% overlap with those listed on the STRING database.

Consistent with the previous knowledge of BAG5 function, Gene Ontology and Pathway analysis of the interactome showed an enrichment of terms relating to protein folding, autophagy, modulation of UPS function, and regulation of apoptosis (Figure 5). A novel theme that surfaced in this analysis pointed to a role of BAG5 in nuclear functions such as mRNA splicing.

Table 2 Top 10 BAG5 and BAG5+BAG5DARA Interacting Proteins

Spectrum Counts

Gene Name Identified Proteins GFP GFP-BAG5 GFP-DARA

BAG5 DNAJC13 DnaJ homolog subfamily C member 13 0 83 0 Interacting Proteins BAG5 Isoform 2 of BAG family molecular chaperone 5 regulator 5 0 66 0

TUBB2A Tubulin beta-2A chain 0 49 0

CRYBG3 Very large A-kinase anchor protein 0 31 0

MAP1A Isoform 2 of Microtubule-associated protein 1A 0 30 0

HSPH1 Isoform Beta of Heat shock protein 105 kDa 0 29 0

HSPA1L Heat shock 70 kDa protein 1-like 0 29 0

BAG3 BAG family molecular chaperone regulator 3 0 27 0

SEC16A Isoform 5 of Protein transport protein Sec16A 0 25 0

AHR Aryl hydrocarbon receptor 0 22 0

HSPA6 Heat shock 70 kDa protein 6 0 22 0

BAG5 & TUBB3 Tubulin beta-3 chain 0 38 30 BAG5DARA Interacting TUBB6 Tubulin beta-6 chain 0 34 20 Proteins

SQSTM1 p62/Sequestosome-1 0 26 14

SLC25A5 ADP/ATP translocase 2 0 19 13

TUFM Elongation factor Tu, mitochondrial 0 19 10

LUZP1 Leucine zipper protein 1 0 16 8

TUBA1C Tubulin alpha-1C chain 0 15 26

HLA-A HLA class I histocompatibility antigen, A-3 alpha chain 0 15 11

PABPC1 Isoform 2 of Polyadenylate-binding protein 1 0 9 13

Dolichyl-diphosphooligosaccharide--protein RPN1 0 9 13 glycosyltransferase subunit 1

FIGURE 4

Figure 4 Expression of GFP-tagged constructs in H4 cells. Western blots of H4 stable cell line clones expressing

GFP (A, right), GFP-BAG5 (A, left) and GFP-BAG5DARA (B). Protein expression measured by probing the blots with anti-GFP antibody. Blots were also probed with anti-GAPDH (A) or anti-Actin (B) antibodies as a loading control. Red boxes outline the clones that were chosen for the H4 interactome analysis.

FIGURE 5

Figure 5 Bioinformatic analysis of the BAG5 interactome in H4 Cells. Top: venn diagram illustrating the overlap between BAG5 and BAG5DARA interacting proteins in H4 cells. Middle/bottom: Selection of Gene Ontology (GO): Biological Process and SuperPath terms enriched in the H4 BAG5 interactome. Dark grey bars refer to the % of proteins associated with each GO term in the reference dataset. Light grey bars refer to the % of proteins associated with each GO term in the BAG5 interactome dataset. GO/SuperPath term enrichment and FDR adjusted P-values calculated by the GeneAnalytics software.

2.3.2 Characterization of the BAG5 Interactome: SH-SY5Y Cells

The results from the H4 interactome pointed to an association of BAG5 with a number of novel proteins and pathways. Therefore, we proceeded with our second analysis of the BAG5 interactome in SH-SY5Y neuroblastoma cells using the iTRAQ mass spectrometry strategy. The

GFP, GFP-BAG5 and GFP-BAG5DARA transgenes were successfully inserted into the AAVS1 safe harbor of SH-SY5Y cells using CRISPR-Cas9 and cre recombination, and the expression of each of the three transgenes was comparably induced by doxycycline treatment (Figure 6).

Biological triplicates of GFP-BAG5 and GFP-BAG5DARA immunoprecipitations, along with a duplicated GFP negative control, were labeled with the iTRAQ 8-plex reagents and combined into a single mass spectrometry run (Figure 7A). As discussed in this chapter’s introduction, the use of this iTRAQ strategy allowed for us to gauge whether an identified interacting protein preferentially bound to either GFP-BAG5 or GFP-BAG5DARA. For example,

Figure 7B-D demonstrates that the iTRAQ output can illustrate whether an identified protein has a stronger (B, Hsc70), equal (C, alpha-synuclein), or weaker (D, p62) association with BAG5 relative to BAG5DARA. In agreement with the H4 interactome, HSPA8/Hsc70 had a significantly higher binding affinity for GFP-BAG5 relative to GFP-BAG5DARA (Figure 7B).

For the bioinformatic analysis of the SH-SY5Y BAG5 interactome, proteins that bound to the GFP negative control and those not labeled by all 8 iTRAQ reagents were removed, leaving a final set of 217 high false discovery rate (FDR) confidence BAG5 interacting proteins

(Appendix 2). The same bioinformatic analysis tool used for the H4 interactome was employed to analyze the SH-SY5Y interactome. There were many common bioinformatic themes between the two interactomes. Like the H4 interactome, GO-term and pathway analysis revealed an enrichment of terms relating to protein folding, UPS function, cell death and mRNA splicing

(Figure 8). The SH-SY5Y interactome also validated our previous finding that BAG5 interacts with alpha-synuclein (Figure 7C), and illustrated an enrichment of the GO-term “Negative

Regulation of Inclusion Body Assembly” (FDR Adjusted P-value: 4.16x10-5, Figure 8), which aligns with the known role of BAG5 in PD-associated alpha-synuclein aggregation.

FIGURE 6

Figure 6 Inducible expression of the GFP transgenes in SH-SY5Y cells. Western blot illustrating the induction of

GFP, GFP-BAG5 and GFP-BAG5DARA in the SH-SY5Y stable cell lines following treatment with doxycycline for the specified amount of time. Blot sequentially probed with anti-GFP then anti-Actin antibodies.

FIGURE 7

Figure 7 iTRAQ mass spectrometry strategy allows for the visualization of BAG5 vs. BAG5DARA binding preference. (A) The interactome analysis makes use of an 8plex iTRAQ mass spectrometry workflow, where 8 independent immunoprecipitation (IP) samples are included in a single mass spectrometry run. The numbers 113-

121 indicate the iTRAQ label assigned to each sample. 'DARA' refers to triplicate GFP-BAG5DARA IP samples, and ‘BAG5’ triplicate GFP-BAG5 IP samples. Two GFP IPs are also included as a negative control for the downstream iTRAQ analysis. The 8 samples are combined into a single mixture that is analyzed with liquid chromatography (LC) and mass spectrometry (MS). Three mass spectrometry runs are required to identify the peptide and iTRAQ abundance (MS3). (B-D) Examples of the quantitative output generated by the iTRAQ strategy. Histogram illustrates the relative abundance of HSPA8/Hsc70 (B), alpha-synuclein (C) and p62 (D) in the 8 iTRAQ samples grouped into GFP-BAG5 (3 samples), GFP-DARA (3 samples) and GFP (2 samples). Individual points representing the value of each of the 8 iTRAQ samples relative to one of the BAG5 IP replicates. Bar represents mean +/- standard error (SE). Significant differences in iTRAQ abundance measured by 1-way ANOVA with Bonferroni post-hoc testing. ** = p<0.01.

FIGURE 8

Figure 8 Bioinformatic analysis of the BAG5 interactome in SH-SY5Y Cells. Selection of Gene Ontology (GO): Biological Process and SuperPath terms enriched in the BAG5 interactome. Dark grey bars refer to the % of proteins associated with each GO term in the reference dataset. Light grey bars refer to the % of proteins associated with each GO term in the BAG5 interactome dataset. GO/SuperPath term enrichment and FDR adjusted P-values calculated by the GeneAnalytics software.

2.4 Discussion

To our knowledge, this is the first characterization of the BAG5 interactome. In line with previous knowledge of its co-chaperone activity (Kalia et al. 2004a), BAG5 was shown to be a part of a rich network of chaperone and co-chaperone proteins in both the H4 and SH-SY5Y interactomes. This included many Hsp40 and Hsp70 family members, including the PD-relevant

Hsp40 family member DNAJC13 (PARK21, Table 2). In addition to DNAJC13, BAG5 was found to interact with eleven proteins relevant to familial PD across the two interactomes. This notably included the interaction between BAG5 and alpha-synuclein, which had been previously identified by our lab (Kalia et al. 2011).

The experiments presented in this chapter are associated with several limitations, such as the use of immortalized cell lines that likely deviate from the proteomic reality of in vivo dopaminergic neurons and the issue of false negatives and false positives in proteomic screens.

An in depth analysis of these, and other, limitations can be found in Chapter 5.

Our bioinformatic analysis of the interactome revealed that BAG5 associates with a number of previously unexplored pathways and compartments, such as nuclear mRNA splicing and nucleosome assembly. The association of BAG5 with these nuclear proteins and pathways could be relevant to its role in both cancer and neurodegenerative disease: aberrant splicing activity is a well-known contributor to both tumorigenesis and neurodegeneration. For example, the BCL2L1 gene is alternatively spliced into either BCLXL or BCLXS. The former promotes cell survival while the latter promotes cell death. Several cancers demonstrate a preferential formation of BCLXL, which contributes to disease progression (Chen and Weiss 2015). In PD, shorter splice variants of alpha-synuclein decrease its aggregation propensity (La Cognata et al.

2015), and a splice variant of the alpha-synuclein interacting protein ‘synphilin-1’, synphilin-1A, promotes the formation of alpha-synuclein aggregates and neuron death in vitro (Eyal et al.

2006). Moreover, ALS-associated TDP-43 mutations result in aberrant mRNA splicing patterns

(Arnold et al. 2013), and a splice variant of TDP-43 itself was shown to enhance the formation of cytoplasmic inclusions and neuron death in vitro (Xiao et al. 2015). Therefore, understanding

46 how BAG5 interacts with splicing machinery could shed light on how it contributes to the pathogenesis of multiple diseases.

Another dominant theme in the bioinformatic analysis was the association of BAG5 with proteins relevant to the ubiquitin proteasome system (UPS) and autophagy lysosome pathway

(ALP). As discussed in chapter one, these two processes maintain cellular proteostasis by managing the degradation of ubiquitinated proteins and protein aggregates, respectively. BAG5 is already known to play a modulatory function within the UPS, as it inhibits the activity to the two PD-relevant E3 ubiquitin ligases, parkin and CHIP (Kalia et al. 2004a, Kalia et al. 2011).

However, little is currently known about the role of BAG5 in the ALP, despite other BAG family proteins being implicated in this pathway (Gamerdinger et al. 2009, Qu et al. 2015, Behl 2016).

As such, understanding the association of BAG5 with proteins functioning within the ALP is an important avenue of future study, as it may extend the role of BAG5 in cellular proteostasis mechanisms beyond the UPS and provide further insights into its relevance to the pathobiology of PD.

For this reason, the novel interaction discovered between BAG5 and the autophagy adaptor protein, p62, was of particular interest. p62 was the 3rd most abundant protein binding to both BAG5 and BAG5DARA in the H4 interactome, and was found to bind to BAG5DARA in the

SH-SY5Y interactome. The fact that p62 bound to BAG5DARA in both interactomes suggests that this interaction is not a false positive brought about by a non-specific interaction between p62 and Hsp70. p62 has important and extensively characterized roles in autophagy. Indeed, p62 plays a critical role in the formation and subsequent lysosomal degradation of protein aggregates, including the alpha-synuclein aggregates found in PD (Bitto et al. 2014). Interestingly, another

BAG family member, BAG3, has already been shown to impact ALP-mediated proteostasis by

47 interacting with and stabilizing p62 (Gamerdinger et al. 2009). Therefore, the interaction between p62 and BAG5 could be relevant to the maintenance of proteostasis in physiological and/or pathological contexts and merits further investigation.

Chapter 3 Validating the Interaction Between BAG5 and p62 3.1 Introduction

Our analysis of the BAG5 interactome illustrated that BAG5 interacts with a wide variety of proteins and pathways both within and outside of the chaperone network. A dominant theme was the association of BAG5 with proteins that mediate proteostasis via the UPS and ALP. This included many chaperones and co-chaperones known to function within these pathways, but also included proteins outside of the chaperone network, notably p62. The novelty of the interaction between BAG5 and p62, as well as the importance of p62 in the management of proteostasis, led us to investigate this interaction in more detail.

p62 is a multifaceted protein that has been studied in multiple contexts, but has been most extensively characterized as an ‘adaptor protein’ in the ALP (Bitto et al. 2014). The term

‘adaptor protein’ is given to a set of proteins functioning within the ALP that are able to bind to both ubiquitinated proteins and the autophagosome coating protein, ‘light chain 3’ (LC3, Figure

9). As such, these proteins provide a molecular ‘bridge’ between ubiquitinated cargo and the autophagy machinery responsible for directing the cargo to the lysosomes. Due to the presence of several redundant adaptor proteins, such as NBR1 and optineurin, p62 is not essential for all autophagy processes (Bitto et al. 2014). However, age-related losses of p62 function are associated with numerous cellular disturbances, such as defective protein degradation via the

ALP (Liu et al. 2017).

Structurally, p62 is composed of many unique domains, which has lead to it being referred to as a ‘swiss army knife’ within the proteome (Liu et al. 2017). The domain composition of p62 is outlined in Figure 10 and notably includes an N-terminal ‘Phox and Bem1

(PB1)’ domain and a C-terminal ‘ubiquitin associated (UBA)’ domain. Also found at the C- terminal is the LC3 interacting region (LIR), which, together with the UBA domain, forms the autophagy adaptor portion of p62. Between the LIR and UBA domains is the ‘Kelch-like ECH- associated protein 1 (Keap1)’ interacting region (KIR). This domain allows p62 to have a regulatory role in the Keap1-Nrf2 pathway, an important pathway involved in the cellular response to oxidative stress (Figure 9, discussed in more detail in chapter five) (Jiang et al.

2015). Interspersed between the C-terminal PB1 and N-terminal LIR/UBA domains are a number of other domains, including a ZZ-type zinc finger domain, that play important roles in a diverse set of cellular processes that are beyond the scope of this thesis.

As an adaptor protein within the ALP, p62 plays an integral role in the clearance of protein aggregates and damaged organelles. In order to clear proteins via the ALP, p62 can either form de novo aggregates of proteins containing K27 or K63-linked ubiquitin chains or associate with pre-formed aggregates via its C-terminal UBA domain (Bitto et al. 2014). Phosphorylation of p62 at serine 403 within its UBA domain greatly enhances its capacity to form de novo p62- rich aggregates, which are commonly referred to as ‘sequestosomes’ (Figure 9) (Matsumoto et al. 2011). Interestingly, p62-mediated protein aggregate formation requires p62 itself to oligomerize via self-associations at its PB1 domain (Katsuragi, Ichimura, and Komatsu 2015).

Following the formation of the sequestosome, the LIR domain on p62 recruits LC3, which allows for autophagosome formation and subsequent lysosomal degradation of the aggregate

(Figure 9). Therefore, the function of p62 in autophagy requires both the C-terminal PB1 and N- terminal LIR/UBA domains, with the former being more important for the formation of aggregates and the latter being more important for their degradation.

We hypothesized that BAG5 impairs the function of p62 in forming and degrading alpha- synuclein aggregates via autophagy. BAG5 and p62 have both been shown to modulate alpha- synuclein aggregation (discussed in greater detail in Chapter 4), indicating that they may function in a common pathway to exert these effects. Both BAG5 and p62 are also known constituents of alpha-synuclein rich LBs (Kalia et al. 2004a, Zatloukal et al. 2002). Moreover, the notion that BAG5 may influence proteostasis by modulating the function of p62 is not an unprecedented hypothesis. Gamerdinger and colleagues have already illustrated that another

BAG family member, BAG3, stabilizes p62, which promotes an increase in autophagy that protects against the loss of proteostasis that occurs as cell age (Gamerdinger et al. 2009).

Therefore, in order to characterize the BAG5-p62 interaction, the purpose of this chapter is to confirm and investigate the interaction itself, and the following chapter will move on to investigate potential functional consequences.

FIGURE 9

Figure 9 p62 facilitates the aggregation and degradation of protein aggregates. Schematic illustrating some of the known functions of p62 in proteostasis. Phosphorylation of p62 stimulates p62-mediated aggregation of ubiquitinated proteins into “sequestosomes”. p62 subsequently recruits light chain 3 (LC3), which triggers the formation of an autophagosome around the sequestosome. The autophagosome, containing p62 and the protein aggregate, is then shuttled to the lysosomes for degradation. Top left: p62 also plays a role in the Kelch-like ECH- associated protein 1 (Keap1)/nuclear factor (erythroid-derived-2)-like 2 (Nrf2) pathway, by promoting the autophgic degradation of Keap1, which allows Nrf2 to translocate to the nucleus where it acts as a transcription factor.

FIGURE 10

Figure 10 Domain structure of p62 and the deletion constructs generated to map its interaction with BAG5.

PB1: Phox and Bem1p, ZZ: ZZ-type zinc finger, TB: TRAF6 binding, LIR: LC3-interacting region, KIR: keap1- interacting region & UBA: ubiquitin-associated. p62-N-HA & p62-C-HA generated using site directed mutagenesis to separate the N-terminal PB1 domain from the C-terminal domains. All constructs have a CMV promoter and a C- terminal hemagglutinin (HA) tag.

3.2 Materials & Methods

3.2.1 Antibodies & Reagents

Antibodies: anti-p62/SQSTM1 (610833) was obtained from BD biosciences. Anti-Actin

(A2066) and anti-flag (M2, F1804) were obtained from Sigma-Aldrich. Anti-BAG5 (CSB-

PA890743ESR1HU) was obtained from Cusabio. Anti-HA (11867423001) was obtained from

Roche. Anti-Hsp70/72 (ADI-SPA-810) was obtained from Enzo Life Sciences. Anti-DNAJC13

(ABN1657) was obtained from EMD Millipore. Horseradish Peroxidase Linked ECL Anti-

Rabbit and Anti-Mouse secondary antibodies were obtained from GE Healthcare. Alexa Fluor

488 (anti-Rabbit) and 555 (anti-mouse) were obtained from Thermo Fisher. Reagents: 0.1%

Ponceau S in 5% acetic acid was purchased from BioShop.

3.2.2 Cell Culture

H4 cells were cultured in DMEM (Gibco) supplemented with 10% FBS (Gibco), 1% antibiotic/antimycotic (Gibco), and incubated at 37°C with 5% CO2. H4 cells were exclusively grown on cell+ plates (Sarstedt). H4 cells were transfected using the SuperFect Transfection

Reagent (Qiagen), as per the manufacturer’s protocol.

3.2.3 Western Blotting

Western blotting was performed in the same way as described in Chapter 2.

3.2.4 GST Pull-down Assay

GST and GST-BAG5/BAG5DARA recombinant proteins were generated in Escherichia coli using pGEX and pDEST-15-BAG5/BAG5DARA (Gateway cloning system, Thermo Fisher) plasmids, respectively. Recombinant proteins were conjugated to Gluthionine Sepharose 4B (GE

Healthcare) beads by rotating recombinant protein with bead slurry overnight at 4°C in

Dulbecco’s phosphate-buffered saline (PBS) without calcium or magnesium.

In order to characterize proteins binding to the GST fusion proteins, 500µg of cell lysate, topped up to a final volume 500µL with PBS, was incubated with 10µg of GST-fusion protein beads overnight at 4°C with rotation. Beads were subsequently washed three times with 1mL of

RIPA buffer, and the proteins were recovered from the bead slurry by adding 50uL SDS-PAGE sample buffer (with beta-mercaptoethanol) and heat denaturing the sample at 95°C for 10min.

For GST pulldown assays that included exogenous p62-HA, p62-N-HA or p62-C-HA, the plasmids were transfected into H4 cells 24 hours prior to cell lysis.

p62-HA deletion constructs (p62-N-HA & p62-C-HA) were generated using the Q5 Site-

Directed Mutagenesis Kit (NEB) as per the manufacturer’s protocol. p62-HA was a gift from

Qing Zhong (Addgene plasmid #28027). p62-HA was split into two complimentary portions: (1)

‘p62-N-HA’ (aa1-102 including PB1 domain) and (2) ‘p62-C-HA’ (aa103-440 including LIR and UBA domains). All the HA-p62 constructs contain a CMV promoter and a C-terminal HA- tag. The following primers were used to generate the deletion constructs: p62-N-HA

Forward: TTTCTCTTTAATGTAGATTCGGAAGATGTCATCC

Reverse: TACCCATACGATGTTCCAGATTACGC p62-C-HA

Forward: CATAGAATTCCACCACACTGGACTAG

Reverse: AAAGAGTGCCGGCGGG

3.2.5 Immunoprecipitation

H4 cells were transfected with flag-BAG5, p62-HA or both. 24 hours post transfection cells were lysed with RIPA buffer and 500µg of whole cell lysate, topped up to 500µL with PBS, was incubated with 1µg of anti-flag antibody and rotated at 4°C overnight. 50µL of pre-washed sepharose A (GE Healthcare) beads were then added to the protein+antibody mixture and rotated for an additional 4 hours at 4°C. Beads were washed three times with 1mL RIPA buffer containing 0.01% weight/volume sodium dodecyl sulfate (SDS). Immunopreicpitated flag-

BAG5, including co-immunoprecipitated proteins, were then competed off of the beads by adding 100µL of 0.25mg/mL flag peptide diluted in TBS. 50µL SDS-PAGE sample buffer (with

55 beta-mercaptoethanol) was then added, and the samples were boiled at 95°C for 10 minutes prior to analyzing them via western blot.

3.2.6 Immunohistochemistry

Wild-type H4 cells were plated at 70-80% confluency in 24 well plates containing poly-D lysine treated glass cover slips. 24 hours after plating, cells were washed once with PBS and treated with 4% paraformaldehyde (PFA) for 15 minutes at room temperature. The PFA was washed off with three sequential 5 minute PBS washes and the cells were subsequently treated with 0.2% triton X-100 diluted in PBS for 15 minutes at room temperature. Cells were then washed another three times in PBS and blocked with 5% w/v bovine serum albumin (BSA) diluted in PBS for 45 minutes. The BAG5 and p62 antibodies were then diluted 1:1000 in 5%

BSA and incubated with the cells overnight at 4°C with gentle rocking. The next day, cells were washed in PBS, incubated in species specific Alexa Fluor 488 and 555, and washed again in PBS containing 4',6-Diamidino-2-Phenylindole, Dihydrochloride (DAPI). Finally, the coverslips were mounted onto slides using Dako mounting media.

3.3 Results

3.3.1 Validation and Visualization of the BAG5-p62 Interaction

We first validated the interaction between BAG5 and p62 using a GST pull-down assay.

Consistent with the results obtained from the interactome analysis, p62-HA transiently transfected into wild type H4 cells was pulled down by both GST-BAG5 and GST-BAG5DARA

(Figure 11B). p62-HA did not associate with the GST negative control. Endogenous Hsp70 was only pulled down by GST-BAG5, confirming the efficacy of the DARA mutation in this assay

(Figure 11A+B). In order to confirm the validity of the GST pull-down assay, we chose to validate another novel top hit BAG5 interacting protein from the interactome, DNAJC13 (Table

2), that was shown to specifically bind to BAG5 and not BAG5DARA (Table 1). In line with the results obtained in the interactome analysis, endogenous DNAJC13 was pulled down by GST-

BAG5 and not GST or GST-BAG5DARA (Figure 11A). The interaction between BAG5 and p62 was further validated by immunoprecipitation, where we observed that p62-HA co- immunopreicpitates with flag-BAG5, when the two proteins are simultaneously expressed in H4 cells using transient transfection (Figure 12).

In order to assess the association between BAG5 and p62 in a more physiological context, we visualized endogenous p62 and BAG5 in H4 cells (Figure 13). p62 staining demonstrated large intracellular puncta that were located in the perinuclear region, and as such, fit the known characteristics of sequestosomes (Bjorkoy, Lamark, and Johansen 2006). p62 was largely, but not completely, excluded from the nucleus in H4 cells. Endogenous BAG5 had a more diffuse staining pattern and was predominantly localized in the nucleus. Importantly,

BAG5 immunoreactivity demonstrated a co-localization and enrichment within the perinuclear p62 puncta (Figure 13).

3.3.2 p62 Interacts with BAG5 via its C-terminal Domains

As discussed in this chapter’s introduction, p62 is composed of multiple domains that have a variety of roles in cell signaling pathways. Importantly, p62 assists in the formation and subsequent degradation of protein aggregates, with the former function being more dependent on its N-terminal PB1 domain, and the latter function being more dependent on its C-terminal LIR and UBA domains. Therefore, in order to perform further mapping of the BAG5-p62 interaction, p62 deletion constructs were generated to dissociate the N-terminal PB1 domain (p62-N-HA) from the remainder of the C-terminus (p62-C-HA), containing both the LIR and UBA domains

(Figure 10). Using the GST pull down assay, we found that GST-BAG5 was able to pull down

57 p62-HA and p62-C-HA, but not p62-N-HA (Figure 14), indicating the p62 associates with

BAG5 independently of its N-terminal PB1 domain.

FIGURE 11

Figure 11 Confirmation of the interactions between BAG5 and DNAJC13/p62 by GST pull-down. (A+B) Western blot of GST pulldown assays performed to validate the interaction between BAG5 and either endogenous DNAJC13 (A) or exogenous p62-HA transfected into wildtype H4 cells (B). Far left lane illustrates 10% of the total protein input for the GST pulldown assay. Ponceau S staining was used to visualize the presence of the GST constructs in the pulldown conditions. Blots were probed with anti-DNAJC13 (A) or anti-p62 (B), as well as anti-

Hsp70/72 to ensure that BAG5DARA does not bind to Hsp70. Blots are representative of three independent studies.

FIGURE 12

Figure 12 Confirmation of the BAG5-p62 interaction via co-immunoprecipitation. Western blot of a flag immunoprecipitation (IP) of flag-BAG5. Flag-BAG5 and/or p62-HA were transiently transfected into wildtype H4 cells prior to cell lysis and IP. Top panel illustrates the IP input and bottom two panels illustrate the IP conditions. * indicates the heavy chain of the IP antibody. Blots probed with anti-flag and/or anti-p62 antibody.

FIGURE 13

Figure 13 BAG5 and p62 co-localize within perinuclear puncta. Wildtype H4 cells stained for endogenous BAG5 and p62 using fluorescent secondary antibodies to allow for visualization. 4',6-Diamidino-2-Phenylindole (DAPI) used to visualize nuclei. BAG5 demonstrated a highly nuclear staining, while p62 was largely excluded from the nucleus and clustered into small puncta in the perinuclear space. BAG5 co-localized with most, but not all, of these puncta. Scale bar represents 20µm. Representative of three independent experiments.

FIGURE 14

Figure 14 p62 associates with BAG5 independently of its N-terminal PB1 domain. GST pulldown assay mapping the BAG5-p62 interaction to the C-terminal domains of p62. Left panel illustrates the GST pulldown assay input, which includes p62-HA, p62-N-HA & p62-C-HA (see Figure 10) transiently transfected into H4 cells. Right panels illustrate the GST pulldown conditions, Ponceau S staining was used to visualize the presence of the GST constructs in the pulldown conditions. Blot probed with anti-HA antibody to detect p62 constructs. Representative of three independent studies.

3.4 Discussion

By using a variety of different techniques, the results from this chapter confirm the newly identified interaction between BAG5 and p62. The interaction was demonstrated to likely be independent of the interaction between BAG5 and Hsp70, as p62 was able to interact with

BAG5DARA. The physiological plausibility of the interaction was supported by the co-localization of BAG5 and p62 in H4 cells. Importantly, BAG5 was co-enriched in perinuclear p62 sequestosomes, suggesting that BAG5 may impact p62 in either the formation or degradation of these structures. The observation that BAG5 binds to the C-terminal region of p62, which contains the LIR and UBA domains important for protein aggregate degradation, suggests that it may be the latter of the two.

By interacting with the LIR and UBA domains, BAG5 could impact several functions of p62, such as the recruitment of LC3 during autophagosome formation, or the movement of protein aggregates to the lysosomes. However, the limited nature of our interaction mapping does not indicate which specific domain on the C-terminus of p62 mediates its interaction with

BAG5. It will be important to clarify this through further mapping, which could provide additional insight into the function of the interaction. It remains possible that the interaction is not relevant to protein aggregation or degradation, but rather to one of the many other known functions of p62, such as its role in the Keap1-Nrf2 pathway (Bitto et al. 2014, Jiang et al. 2015,

Katsuragi, Ichimura, and Komatsu 2015). However, the strong co-localization of BAG5 with p62 sequestosomes suggests that the interaction likely does have a function in cellular proteostasis.

Due to technical challenges, we were unable to demonstrate an interaction of the two endogenous proteins in either H4 cells as well as rat brain lysate using co-immunoprecipitation.

This could be explained by the low affinity of our antibodies for endogenous p62 and BAG5 or

62 that the interaction is more transient and lost during the immunoprecipitation wash steps. Future experiments will therefore merit an optimization of the antibodies and immunoprecipitation conditions used. Another explanation of our difficulty visualizing the endogenous interaction is that the interaction may exclusively exist in insoluble protein aggregates that are lost during our cell lysis procedure. This is supported by the co-enrichment of BAG5 and p62 into perinuclear sequestosomes. As such, future experiments will also benefit from a more stringent isolation of intracellular protein that preserves these larger aggregates.

Finally, the nuclear localization of endogenous BAG5 observed in H4 cells is particularly interesting because our bioinformatic analysis of the interactome suggested that BAG5 associates with many nuclear proteins and processes such as mRNA splicing. An analysis of the BAG5 peptide sequence using a nuclear localization sequence (NLS) prediction software (http://nls- mapper.iab.keio.ac.jp/cgi-bin/NLS_Mapper_form.cgi) identified two bi-partite NLSs on the N- terminal of BAG5 (data not shown), validating that endogenous BAG5 can translocate into the nucleus. Combined, these results suggest that investigating the function of BAG5 in the nucleus will likely be an important consideration for future study. Notably, this notion is not irrelevant to the BAG5-p62 interaction, as p62 is able to translocate into the nucleus where it is involved a number of processes, including the maintenance of nuclear proteostasis (Pankiv et al. 2010).

Chapter 4 Investigating the BAG5-p62 Interaction in the Context of Alpha- synuclein aggregation

4.1 Introduction

Both BAG5 and p62 have been implicated in PD-associated alpha-synuclein aggregation and degradation, which is outlined schematically in Figure 15. The effect of BAG5 in this regard has been discussed in previous chapters. Briefly, BAG5 is known to suppress alpha-synuclein degradation by inhibiting CHIP (Kalia et al. 2011), and this effect of BAG5 could be exacerbated by its inhibition of parkin, which is also known to effect alpha-synuclein stability (Lonskaya et al. 2013) (Figure 15). BAG5 may also indirectly promote alpha-synuclein aggregation by inhibiting Hsp70, which antagonizes this process (see Chapter 1) (Figure 15).

The association between p62 and alpha-synuclein aggregation is largely brought about by the fact that p62 is a dominant constituent of alpha-synuclein rich LBs (Zatloukal et al. 2002).

Indeed, many neuropathologists now stain for p62, rather than alpha-synuclein, when visualizing

LBs in human tissue. The association of p62 with disease-related protein aggregates is not restricted to PD but extends to other neurodegenerative diseases such as AD, HD and ALS

(Zatloukal et al. 2002). The presence of p62 within LBs and other disease-related protein aggregates is not surprising considering that p62 is known to promote the aggregation of ubiquitinated proteins (Komatsu et al. 2007). Therefore, it may be plausible that p62 promotes alpha-synuclein aggregation (Figure 15), however, this has not yet been proven.

Despite the physical association of p62 with alpha-synuclein aggregates, only two studies have directly assessed the effect of p62 on alpha-synuclein aggregation and degradation.

Watanabe and colleagues first demonstrated that p62 facilitates the autophagic degradation of alpha-synuclein aggregates in vitro (Watanabe et al. 2012), and Tanji and colleagues later found that p62 KO enhanced the number of alpha-synuclein inclusions in an in vivo model of Lewy body disease (Tanji et al. 2015). Therefore, it seems as though p62 facilitates the autophagic degradation of alpha-synuclein aggregates (Figure 15). However, much work remains to be done in understanding how p62 impacts alpha-synuclein pathology, and the factors that regulate p62 in this regard.

Considering that p62 is known to form large, de novo aggregates from soluble ubiquitinated proteins, we hypothesized that p62 promotes the movement of alpha-synuclein from soluble oligomers into insoluble aggregates that are subsequently degraded by autophagy. In

Chapter 3, we suggested that BAG5 impairs the function of p62 in forming and degrading alpha- synuclein aggregates via autophagy. Combining these two ideas, we hypothesized that BAG5 promotes the presence of alpha-synuclein oligomers by suppressing the capacity of p62 to sequester alpha-synuclein oligomers into large aggregates that are subsequently degraded by autophagy. These hypotheses align with the previously known function of p62 to promote the formation and degradation of protein aggregates, and the previous finding that BAG5 indirectly enhances alpha-synuclein aggregation.

FIGURE 15

Figure 15 p62 and BAG5 have complex effects on protein aggregation and degradation pathways. Schematic illustrating the known and predicted effects of BAG5 and p62 on protein degradation pathways. BAG5 inhibits the activity of ‘Heat shock protein 70’ (Hsp70) which is known to antagonize the formation of pathogenic alpha- synuclein aggregates. BAG5 also inhibits the activity of the ubiquitin E3 ligases ‘C-terminal of Hsp70 Interacting protein’ (CHIP) and parkin, which promote the degradation of alpha-synuclein. P62 is known to facilitate the formation of neurodegenerative disease-related protein aggregates as well as the degradation of alpha-synuclein via the autophagy lysosome pathway. The effect of BAG5 on p62 in these pathways has yet to be elucidated, as illustrated by a ‘?’.

4.2 Materials & Methods

4.2.1 Antibodies & Reagents

Antibodies: anti-alpha-synuclein (610786) and anti-p62/SQSTM1 (610833) were obtained from BD biosciences. Anti-Actin (A2066) was obtained from Sigma-Aldrich. Anti-

BAG5 (CSB-PA890743ESR1HU) was obtained from Cusabio. Reagents: Coelenterazine (303-

5) was obtained from NanoLight Technology.

4.2.2 Cell Culture

H4 and HEK293 cells were cultured in DMEM (Gibco) supplemented with 10% FBS

(Gibco), 1% antibiotic/antimycotic (Gibco), and incubated at 37°C with 5% CO2. H4 cells were exclusively grown on cell+ plates (Sarstedt). H4 cells were transfected using the SuperFect

Transfection Reagent (Qiagen) and HEK293 cells were transfected using lipofectamine 200

(Thermo Fisher), as per the manufacturer’s protocol. Small interfering RNA (siRNA) mediated

BAG5 knockdown (KD), was achieved by transfecting either an siRNA targeting BAG5

(siBAG5, Ambion, #s18285), or non-targeting control (siNTC, Ambion) into either H4 or

HEK293 cells using Lipofectamine RNAiMAX (Thermo Fisher) according to manufacturer’s protocol.

4.2.3 Western Blotting

Western blotting was carried out in the way as Chapter 2.

4.2.4 Alpha-synuclein Protein Complementation Assay

Alpha-synuclein luciferase constructs were generated as previously described (Kalia et al.

2011). syn-N, syn-C and p62-HA/p62-C-HA/p62-N-HA constructs were transfected into

HEK293 cells at a 1:1:0.5 ratio. Empty pcDNA3.1 vector was added as necessary so that all transfection conditions contained the same total DNA concentration. In the case of siRNA

67 mediated BAG5 KD, the siRNA transfection was performed 24 hours prior to transfecting in the overexpression plasmids. 24 hours post-transfection of the alpha-synulclein constructs, cells were scraped in 600μL cold PBS and 100uL of cells were transferred in triplicate to an opaque flat-bottomed 96-well plate (Grenier). The other 300μL of cells were saved for western blot analyses. The plate was then analyzed on a CLARIOstar plate-reader (BMG Labtech), which injected 100μL of 40μM coelenterazine (Nanolight Technology) into each well and shook the plate for 2 seconds prior to reading the bioluminescent signal generated by the Gaussia princeps luciferase.

4.3 Results

4.3.1 p62 reduces the presence of soluble alpha-synuclein and oligomers

We first tested the effect of p62 on alpha-synuclein aggregation. To do so we used a previously described luciferase reporter protein complementation assay (PCA) that allows for the analysis of alpha-synuclein oligomerization in vitro (Kalia et al. 2011, Outeiro et al. 2008,

Putcha et al. 2010, Remy and Michnick 2006). This model makes use of two alpha-synuclein constructs that each contain full-length alpha-synuclein fused to either the C-terminal or N- terminal half of Gaussia princeps luciferase (termed syn-N and syn-C, respectively, Figure 16A).

When the two constructs associate, luciferase is reconstituted and can generate a bioluminescent signal in the presence of the appropriate substrate (Figure 16A-C). As such, measurable bioluminescence can be used as a surrogate measure for alpha-synuclein oligomers.

FIGURE 16

Figure 16 Alpha-synuclein protein complementation assay (PCA) proof of concept. (A) Schematic illustration of the two synuclein constructs used in the PCA. Each construct contains full-length synuclein fused to either the N- terminal (syn-N) or C-terminal half (syn-C) of Gaussia princeps luciferase. When the two constructs associate with each other, luciferase is reconstituted and can generate a bioluminescent signal when exposed to the appropriate substrate (in this case coelenterazine) (B) Western blot demonstrating the presence of syn-N and syn-C for the PCA results presented in (C), probed with anti-alpha-synuclein and anti-actin antibodies. Representative of three independent studies. (C) PCA illustrating that a luminescent signal is only generated when both syn-N and syn-C are present. Data is normalized to the “syn-N + syn-C” condition. Data generated from three independent studies measured in triplicate. Bars illustrate mean +/- SE. Statistical significance calculated using independent samples t- test, ** p<0.01.

Using this model, we found that exogenous p62 reduced luciferase activity by 68.8% (SE

= 1.2%, p<0.0001, Figure 17A+B), relative to pcDNA control, and also served to lower the presence of soluble syn-N & syn-C constructs by 34.5% (SE = 6.9%, p<0.001, Figure 17A+C).

This effect of p62 was dependent on both its PB1 domain and N terminal LIR and UBA domains, as the two p62 deletion constructs (p62-N-HA & p62-C-HA) failed to have any effect on luciferase activity relative to control (Figure 18).

4.3.2 BAG5 KD reduces alpha-synuclein oligomerization but does not impact p62

In order to interrogate the effect of BAG5 on this p62 phenotype, we used small interfering RNA (siRNA) mediated BAG5 knockdown (KD). In the presence of p62-HA, BAG5

KD trended towards further reducing luciferase activity, however; this result failed to reach significance (p62-HA + control siRNA = 37.2% +/- 4.6%; p62-HA + BAG5 siRNA = 23.8% +/-

3.4%; p = 0.12, Figure 19A+B). Similarly, flag-BAG5 overexpression also failed to exert any effect on the p62-mediated reduction in luciferase activity (data not shown). Nonetheless, in the absence of exogenous p62, BAG5 KD reduced the luciferase signal by 31.1% +/- 4.7% relative to control (Figure 19A+B), without lowering the levels of soluble alpha-synuclein (Figure 19A).

FIGURE 17

Figure 17 p62 reduces the presence of both soluble and oligomeric alpha-synuclein. (A) Western blot illustrating the levels of p62-HA and syn-N/syn-C for the PCA presented in (B). Probed with anti-p62, anti-alpha- synulcein and anti-actin antobodies. Representative of three independent studies. (B) PCA from HEK293 cells transfected with syn-N+syn-C and either p62-HA or pcDNA negative control. Data generated from three independent studies measured in triplicate, and normalized to the pcDNA condition. Statistical significance calculated using independent samples t-test, **** = p<0.0001. (C) Quantification of the combined syn-N+syn-C band intensity depicted in (A). Data obtained from four independent studies. Synuclein construct intensity in both the pcDNA and p62-HA conditions were normalized to actin intensity and the p62-HA condition was subsequently normalized to the pcDNA condition. Statistical significance calculated using independent samples t-test, *** p<0.001.

FIGURE 18

Figure 18 p62 requires both its C-terminal PB1 domain and N-terminal LIR+UBA domains to influence alpha-synuclein oligomerization. Top: schematic representation of WT p62-HA and the two deletion constructs (outlined in Figure 10). Bottom: PCA results from HEK293 cells transfected with the two synuclein luciferase constructs (syn-N + syn-C) and either p62-HA, p62-N-HA or p62-C-HA. Data representative of three independent studies performed in triplicate, normalized to the pcDNA condition (lane 1). Bars represent mean +/- SE. Statistical significance calculated using 1-way ANOVA with Bonferroni post hoc test, * p<0.01, ** p<0.01.

FIGURE 19

Figure 19 p62 and BAG5 have independent effects on synuclein oligomerization. (A) Western blot of the PCA presented in (B), probed with anti-p62, anti-BAG5, anti-alpha-synuclein and anti-actin antibodies. Top panel is a longer exposure of the p62 probe presented below it. Representative of three independent studies. * indicates a non- specific band. (B) PCA from HEK293 illustrating the effect of siRNA-mediated BAG5 KD on luciferase activity in the presence or absence of exogenous p62-HA. Data generated from three independent studies measured in triplicate, and normalized to the siRNA alone condition (lane 1). Bars represent mean +/- SE. Statistical significance calculated using 1-way ANOVA with Bonferroni post hoc test, ** p<0.01, **** p<0.0001.

4.3.3 BAG5 Stabilizes Endogenous p62

BAG3 has been previously shown to associate with p62 and increase its expression in ageing cells (Gamerdinger et al. 2009). Stable levels of p62 are important for its function in protein aggregation and degradation, as the loss of endogenous p62 results in a suppression of both inclusion body formation (Komatsu et al. 2007) and degradation (Tanji et al. 2015). Our working hypothesis was that BAG5 impairs these activities of p62, which lead us to further hypothesize that BAG5 reduces the expression of p62. Therefore, while conducting the PCA, we additionally analyzed the effect of BAG5 KD on p62 expression.

BAG5 KD did not have an effect on exogenous p62 transfected into HEK293 cells

(Figure 19A [low exp.]). However, BAG5 KD markedly reduced levels of endogenous p62

(Figure 19A [high exp.]). This result contradicted our hypothesis, and mirrored the known function of BAG3 in supporting p62 expression. Gamerdinger and colleagues reported that this effect of BAG3 on p62 was specific to particular cellular environments, namely ageing cells that were facing a loss of proteostasis (Gamerdinger et al. 2009). Therefore, we wanted to assess whether the observed effect of BAG5 on p62 expression was specific to HEK293 cells transfected with exogenous alpha-synuclein, or could be generalized to other cellular environments. Using wild-type H4 cells, BAG5 KD reduced endogenous p62 levels by 71.8%

(SE = 7.4%) relative to the siRNA control condition (p<0.001) (Figure 20A+B), indicating that

BAG5 supports p62 stability across multiple cell types.

FIGURE 20

Figure 20 BAG5 stabilizes endogenous levels of p62. (A) Western blot illustrating the changes in endogenous p62 stimulated by siRNA-mediated BAG5 KD, probed with anti-p62, anti-BAG5 and anti-actin antibodies. Each lane demonstrates an independent study. * indicates a non-specific band. (B) Quantification of the intensity of p62 in the western blot presented in (A). p62 band intensity was normalized to Actin intensity in both the si-NTC and si- BAG5 conditions and values were then subsequently normalized to the si-NTC condition. Statistical significance measure with independent samples t-test: *** p<0.001.

4.3.4 Discussion

The results from this chapter illustrate that BAG5 enhances and p62 reduces alpha- synuclein oligomer levels, as both BAG5 KD and p62 overexpression decreased luciferase activity as measured by the PCA. p62 overexpression also resulted in a significant reduction of soluble alpha-synuclein. These effects of p62 on alpha-synuclein are significant because they support the findings of the two previous studies that demonstrate that p62 facilitates the clearance of alpha-synuclein aggregates (Tanji et al. 2015, Watanabe et al. 2012). Moreover, these results align with our hypothesis that p62 promotes the movement of alpha-synuclein from soluble oligomers into insoluble aggregates that are subsequently degraded by autophagy. The use of autophagy-lysosome inhibitors, such as Bafilomycin-A1, in future PCA experiments will help to confirm this conclusion.

While this makes p62 appear an enticing therapeutic target for PD, there are many outstanding questions that remain. For example, the finding the p62 KO enhances Lewy body presence by (Tanji et al. 2015) seems to contrast the seminal finding by (Komatsu et al. 2007) that p62 KO attenuates the formation of insoluble protein aggregates in autophagy deficient mice. In addition, due to the limitations of the PCA, our observation that p62 reduces soluble alpha-synuclein oligomer presence may represent a p62-stimulated movement of alpha-synuclein into larger insoluble aggregates, rather than a clearance of the protein. As such, it is necessary to gain a more sophisticated mechanistic understanding of how p62 impacts alpha-synuclein oligomerization, aggregation and degradation, before considering how it can be therapeutically manipulated to treat synucleinopathies.

We also hypothesized that BAG5 promotes the presence of alpha-synuclein oligomers by suppressing the capacity of p62 to transition alpha-synuclein oligomers into large aggregates that

76 are subsequently degraded by autophagy. If this were correct, BAG5 KD would be expected to further reduce luciferase activity in the presence of exogenous p62. BAG5 KD did do this, however, the change failed to meet statistical significance. Therefore, it is possible that our hypothesis is incorrect and BAG5 and p62 operate in parallel pathways to impact alpha- synuclein aggregation. Another possibility is that the robust effect of exogenous p62 in the PCA masked additional, subtler effects of BAG5. As such, future interrogations of this interaction will benefit from more nuanced oligomerization assays, as well as assays that interrogate other aspects of proteostasis to which this interaction may be more relevant. This concept is discussed in more detail in Chapter 5.

While we did not establish a clear functional consequence of the BAG5-p62 interaction in this assay, BAG5 KD did reduce alpha-synuclein oligomer formation as measured by the

PCA. This parallels the previous finding that BAG5 indirectly enhances alpha-synuclein oligomerization via its inhibition of CHIP (Kalia et al. 2011) and solidifies a role of BAG5 in promoting alpha-synuclein aggregation. BAG5 also stabilized endogenous levels of p62 in two cell lines. Because p62 is an important regulator of protein aggregation and degradation via the

ALP and UPS, we hypothesize that BAG5 could modulate proteostasis by impacting homeostatic levels of p62. Indeed, BAG3-mediated p62 stabilization was shown to be important for the maintenance of proteostsis by enhancing autophagy in ageing cells (Gamerdinger et al. 2009).

Therefore, while BAG5 may not have a clear effect on the p62-mediated clearance of alpha- synuclein oligomers, the observed effect of BAG5 on p62 levels suggests that this interaction could be relevant to other functions of p62 in proteostasis.

From a mechanistic standpoint, parkin was recently shown to promote the proteasomal degradation of p62 (Song et al. 2016). BAG5 is known to modulate parkin function (Kalia et al.

2004a), and the pathway “Proteolysis Role of Parkin in the UPS” was enriched in both the H4 and SH-SY5Y interactomes (FDR Adjusted P-value: 4.18x10-13 and 3.33x10-16, respectively,

Figure 9). Therefore, it could be hypothesized that the effect of BAG5 on p62 stability may be mediated by the interaction between BAG5 and parkin.

Chapter 5 General Discussion & Future Directions 5.1 Summary

To summarize, this thesis represents the first characterization of the BAG5 interactome.

Many known BAG5 interactions were confirmed, and many novel interactions were uncovered.

In line with the known function of BAG5 as an Hsp70 co-chaperone, BAG5 was found to interact with many Hsp70 family members and a host of other chaperone and co-chaperone proteins. BAG5 also interacted with many proteins outside of the chaperone network, as illustrated through the use of BAG5DARA in the interactome analyses. Furthermore, BAG5 associated with numerous proteins relevant to PD, both within and outside of the chaperone network, notably including alpha-synuclein.

The bioinformatic analysis pointed to an involvement of BAG5 in the ubiquitin proteasome system and autophagy, which aligns with previous investigations of BAG5 in these proteostasis pathways. A novel interaction discovered between BAG5 and p62 exemplified the observed association between BAG5, the UPS, and autophagy, as p62 has important functions in these pathways. Further in vitro experimentation confirmed the BAG5-p62 interaction and elucidated that p62 binds to BAG5 via its C-terminus, which includes the LIR and UBA domains that allow it to facilitate the degradation of protein aggregates. Endogenous BAG5 and p62 also co-localized in sequestosomes located in the perinuclear region of H4 cells, further implicating

BAG5 in p62-mediated protein aggregate formation and/or degradation.

Both p62 and BAG5 have been shown to modulate alpha-synuclein aggregation and degradation. We demonstrate that p62 reduced both soluble levels and oligomers of alpha- synuclein, whereas BAG5 enhanced alpha-synuclein oligomer formation. Knocking down BAG5

79 enhanced the capacity of p62 to reduce alpha-synuclein oligomers, but this result did not achieve statistical significance. This suggested that the two proteins may act in independent pathways to modulate alpha-synuclein oligomerization. Nonetheless, BAG5 KD did result in the reduction of endogenous p62 in two cell lines, indicating that BAG5 may impact p62 function by modifying its cellular abundance.

5.2 Study Limitations

In terms of the interactome analysis, proteomic screens, such as the one conducted here, can generate false positives that cannot be recapitulated in future screens or experiments. This effect may be compounded in our study by the fact that BAG5 predominantly associates with

Hsp70, and Hsp70 can non-specifically interact with a myriad of misfolded proteins. We attempted to address this issue by carrying out the interactome analysis in two independent cell lines and using the BAG5DARA mutant, which should theoretically indicate which proteins can bind to BAG5 in the absence of Hsp70 (Kalia et al. 2004a). However, the fact that we used two significantly different methodologies between the H4 and SH-SY5Y interactomes makes it difficult to compare the two interactome lists. Indeed, while it would be enticing to suggest that proteins not identified in both lists are not true interactors (i.e., false positives), this is not necessarily the case. For example, we have previously validated and published the interaction between BAG5 and alpha-synuclein (Kalia et al. 2011), but this interaction was exclusively identified in the SH-SY5Y screen. Therefore, as has been the case with other proteomic screens, rigorous replication and the use of additional methodologies to verify individual interactions are necessary before assuming any interaction is real.

It should also be noted that the non-physiological nature of the DARA mutation may disrupt or modify BAG5’s association with proteins other than Hsp70. BAG5DARA does retain

some of the functions of wild-type BAG5. For example, like wild-type BAG5, BAG5DARA retains the capacity to dimerize with either itself or wild-type BAG5, suggesting that the DARA mutation does not significantly alter protein conformation (Kalia et al. 2004a). Moreover,

BAG5DARA also retains the capacity to interact with and inhibit the E3 ubiquitin ligase activity of parkin (Kalia et al. 2004a). Nevertheless, caution should be taken in interpreting the physiological meaning of proteins that either were or were not found to bind to BAG5DARA. An additional method that could be used to determine if an interaction is dependent on the presence of Hsp70 would be to conduct a pull-down or immunoprecipitation assay using recombinant proteins rather than whole cell lysate. Recombinant Hsp70 could then be added or removed to assess whether the interaction is dependent on its presence.

Another limitation of our interactome analysis was the use of immortal cell lines that do not represent the proteomic reality of dopaminergic neurons in vivo. As such, it is difficult to conclude that the observed interactions in these screens are relevant to the in vivo setting. We accept this limitation, as the purpose of this project was to get a preliminary BAG5 interactome to guide further investigations of its function. Nonetheless, it will be important to confirm interactions of interest in more relevant cellular environments. This can be accomplished for individual interactions by performing co-immunoprecipitation studies of endogenous proteins from in vivo samples. Or, the same approach can be used on the scale of the entire BAG5 interactome by immunoprecipitating endogenous BAG5 from a relevant biological sample, such as rodent or human SNpc homogenate, and identifying co-immunoprecipitated proteins by mass spectrometry.

In this study, we have demonstrated that the interaction between p62 and BAG5 is independent of the N-terminal PB1 domain on p62. However, there are still many possible

81 domains on the C-terminus of p62 that may mediate its interaction with BAG5 (see Figure 10).

Our mapping studies also did not illustrate which BAG domain on BAG5 interacts with p62.

Therefore, future analyses of the BAG5-p62 interaction will also benefit from a more fine-tuned mapping of the interaction. Understanding which domain/region of p62 specifically facilitates its interaction with BAG5 may provide some insight into the function of the interaction. For example, if the Keap1 interacting region (KIR) on p62 mediates the interaction, it may be more likely that BAG5 impacts the function of p62 in the Keap1-Nrf2 response to oxidative stress than in proteostasis or autophagy.

The PCA analysis of the effect of BAG5 and p62 on alpha-synuclein oligomer formation is also associated with several limitations. One limitation is that alpha-synuclein ‘aggregates’ can exist in several forms, as they build from misfolded protein, to small oligomeric species, to larger fibrils and finally insoluble aggregates (see Chapter 1). While it has been hypothesized that the luciferase readout of this assay corresponds to the presence of soluble alpha-synulcein oligomers

(Kalia et al. 2011), without visualizing the intracellular aggregates, it is not entirely clear what type of alpha-synuclein species are being observed. Therefore, while BAG5 and p62 both modulate the luciferase output, it is difficult to know how this should be interpreted. For example, the p62-mediated reduction of soluble alpha-synuclein oligomers could be interpreted as (1) p62 facilitating the aggregation of alpha-synuclein into large insoluble aggregates, (2) p62 facilitating the degradation of these aggregates, or, (3) both, as would most likely be the case based on the p62 literature. Therefore, other approaches are necessary in order to get a more precise understanding of how BAG5 and p62 are impacting alpha-synuclein pathology.

The most logical next step would be to visualize alpha-synuclein aggregates using immunohistochemistry and assess the effect of modulating BAG5 and/or p62 levels. One option

82 would be to generate PCA constructs where alpha-synuclein is fused the N/C-terminal halves of

GFP rather than luciferase. Such constructs could then be used to visualize alpha-synuclein oligomers in cells and assess how p62 and BAG5 modulate their presence, morphology or localization. The effect of BAG5 and p62 on the formation of larger alpha-synuclein aggregates could also be assessed using techniques such as size-exclusion chromatography or analyzing the levels of alpha-synuclein in detergent insoluble cell lysate fractions.

Another important consideration is that the unique alpha-synuclein species (oligomers, fibrils and aggregates) are associated with different physiological effects and toxicities (see

Chapter 1). As discussed in the introduction, the current hypothesis is that small oligomers confer more toxicity than the large aggregates, but there is still debate surrounding this issue

(Kalia et al. 2013, Kalia and Lang 2015, Rosborough, Patel, and Kalia 2017). Therefore, just because BAG5 and p62 modulate alpha-synuclein aggregation patterns in our PCA model does not mean they have any relevant or predictable effect on synuclein-mediated toxicity. In future experiments, it will be important to not only assess how BAG5 and p62 modulate alpha- synuclein aggregation, but also understand how these effects relate to cell death. Outside of rudimentary in vitro cell death assays (Annexin V, propidium iodide, MTT, etc.), one way in which this could be accomplished would be to stereotactically introduce alpha-synuclein fibrils or a plasmid packaged into an adeno-associated virus (AAV) into the substantia nigra of rats or mice (Koprich et al. 2010). Levels of p62 and BAG5 could then be manipulated through the use of transgenic knockout mice or viral plasmid delivery to test their effect on synuclein pathology and dopaminergic neurodegeneration relative to the appropriate controls.

Lastly, we observed BAG5 to promote p62 stability. While this result could be recapitulated in multiple cell lines, our results do not clarify how BAG5 exerts this effect or why

83 it is physiologically relevant. Because p62 is an aggregation prone protein, as demonstrated by the formation of perinuclear puncta in H4 cells, it is possible that BAG5 KD is not causing a loss of endogenous p62, but rather promoting its movement into insoluble aggregates. This could be clarified via western blot by analyzing the effect of BAG5 KD on the presence of p62 in detergent insoluble cell lysate fractions. It will also be important to confirm that the observed reduction of p62 levels is not an off-target effect of the siRNA-mediated BAG5 KD procedure.

This can be done by using multiple BAG5 siRNAs, or by rescuing p62 levels through the use of autophagy or proteasome inhibitors (ex. Bafilomycin-A1 or MG132).

5.3 Future Directions: BAG5, p62 and Proteostasis

As has been discussed in the previous chapters, one of the most dominant themes in the interactome analysis was the association of BAG5 with UPS- and ALP-mediated proteostasis pathways. BAG5 is already known to modulate proteostasis via its association with Hsp70

(Arakawa et al. 2010), parkin (Kalia et al. 2004a), and CHIP (Kalia et al. 2011), which facilitate protein degradation via the UPS, and to a certain extent, the ALP. p62 also has important functions in maintaining proteostasis through similar pathways as BAG5 (see Figure 15).

Therefore, we chose to investigate the BAG5-p62 interaction that was discovered in the interactome analysis and hypothesized that BAG5 inhibits the capacity of p62 to form and degrade alpha-synuclein aggregates via autophagy.

Unfortunately, we were unable to establish a clear functional interaction in the context of alpha-synuclein aggregation. However, it led to the observation that BAG5 stabilizes p62. This may indicate that interaction is relevant to some aspect of proteostasis, even if not alpha- synuclein aggregation, as homeostatic levels of p62 are known to be important for the management of protein aggregation/degradation dynamics in a number of different contexts

(Bitto et al. 2014, Komatsu et al. 2007). Therefore, assessing the effect of BAG5 on other proteostatic functions of p62, such as the regulation of autophagic flux (Bjorkoy, Lamark, and

Johansen 2006), the degradation of the 26S proteasome (Cohen-Kaplan et al. 2016), or its effect on the aggregation and degradation of proteins other than alpha-synuclein, such as Tau (Babu,

Geetha, and Wooten 2005) or TDP-43 (Brady et al. 2011), will be important avenues of future study.

Outside of p62, the interactome uncovered many other BAG5 interactions relevant to proteostasis. BAG5 was shown to interact with a wide variety of Hsp70 chaperones and co- chaperone proteins with well-known roles in protein folding and degradation pathways. The interaction between BAG5 and DNAJC13, an Hsp40 family co-chaperone, stands out, as mutations to the DNAJC13 gene (PARK21) were recently identified to be a cause of PD in a

Canadian family (Vilariño-Güell et al. 2014). Therefore, it was exciting to recapitulate this interaction using the GST pull-down assay in H4 cells. Little is currently known about

DNAJC13 function; however, several studies have demonstrated that it plays an important role in trafficking clathrin-coated endocytic vesicles (Chang, Hull, and Mellman 2004, Freeman,

Hesketh, and Seaman 2014, Girard et al. 2005, Girard and McPherson 2008, Shi et al. 2009). As such, it was interesting to observe that BAG5 bound to clathrin heavy chain 1 in both the SH-

SY5Y and H4 interactomes (Appendix 1+2). This places BAG5 in the vicinity of DNAJC13 function, and suggests that this interaction may be relevant to endosomal trafficking.

Considering that the endosome system is closely related to the lysosomes, and by extension, autophagy, it is enticing to hypothesize that this interaction may represent another way in which BAG5 modulates proteostasis by effecting the function of the ALP. A recent study by (Yoshida et al. 2018) demonstrated that the PD-causing N855S mutation to DNAJC13

85 resulted in dysfunctional endosome management, which in turn increased alpha-synuclein aggregation and neurodegeneration in a Drosophila model. Therefore, the interaction between

BAG5 and DNAJC13 presents another potential mechanism by which BAG5 impacts alpha- synuclein aggregation and toxicity, and merits future investigation.

Outside of the specific interaction with DNAJC13, the numerous interactions between

BAG5 and chaperone network have broader scale implications for our understanding of the

‘chaperome’. As mentioned in the introduction, the feasibility of chaperone-based therapies is limited by the fact that molecular chaperones exist in a complex and changing proteomic network, making them a difficult target to manipulate pharmaceutically. In turn, there is an emerging need to characterize the nuances of the chaperone network in order to best understand how therapies can be tailored to restore proteostasis. By characterizing the BAG5 interactome, we have contributed to an understanding of the complex interaction networks that exist between chaperone and co-chaperone proteins. This work may facilitate the capacity to therapeutically harness molecular chaperones to restore disease-related perturbations of proteostasis.

5.4 Future Directions: BAG5 and Cell Death

Another dominant theme revealed by our bioinformatic analyses was the association between BAG5 and proteins functioning within apoptosis pathways. This is not surprising, as

BAG5 has been shown to modulate cell death in several different contexts. Indeed, in the primary characterization of BAG5, BAG5 was found to enhance dopaminergic neurodegeneration in the SN of rats exposed to MPTP (Kalia et al. 2004a). Moreover, the BAG family was initially discovered due to its association with the Bcl-2, a prototypical anti-apoptotic protein (Takayama et al. 1995). However, recent studies have challenged the cytotoxic role of

BAG5, illustrating that it can promote cell survival in some in vitro contexts (Bi et al. 2016a,

Bruchmann, Roller, Walther, Schafer, et al. 2013, Guo et al. 2015a, Gupta et al. 2016, Ma et al.

2012, Wang et al. 2014). In order to follow up on our interactome analysis and clarify the role of

BAG5 in cell death, we carried out a comprehensive investigation of the effect of BAG5 on cell death elicited by a number of different toxins. Using the SH-SY5Y stable cell lines, we found that GFP-BAG5 enhanced cell death mediated by mitochondrial damaging agents and oxidative stress, relative to the GFP control (De Snoo et al, in preparation). We hypothesized that this effect of BAG5 could be mediated by its interaction with the E3 ubiquitin ligase parkin.

As discussed in the introduction, parkin is an E3 ubiquitin ligase that is mutated to cause familial PD. Parkin has generally been considered to be a neuroprotective factor, given that loss- of-function mutations to parkin result in neurodegeneration and ineffective clearance of damaged mitochondria (Narendra et al. 2008). However, a growing body of literature indicates that, in the case of severe mitochondrial damage, parkin switches to promote apoptosis by enhancing the proteasomal degradation of the anti-apoptotic Bcl-2 family member, Mcl-1, rather than promoting mitophagy (Carroll, Hollville, and Martin 2014, Zhang et al. 2014). This has generated an emerging hypothesis that parkin can switch between cytoprotective (mitophagy) and cytotoxic (Mcl-1 degradation) states depending on the degree of mitochondrial damage it is faced with. The factors that direct parkin to switch between these states are not yet elucidated.

Parkin is a known BAG5 interacting protein (Kalia et al. 2004a). Moreover, the bioinformatic analysis of the BAG5 interactome demonstrated that the Gene Ontology:

Biological Process term “Proteolysis Role of Parkin in the Ubiquitin-Proteasomal Pathway” was significantly enriched in both the H4 and SH-SY5Y interactomes, solidifying the notion that

BAG5 modulates parkin activity, as has been previously described (Kalia et al. 2004a).

Therefore, we hypothesized that BAG5 enhanced apoptosis in the context of mitochondrial

87 damage by promoting parkin-mediated Mcl-1 degradation. To confirm this hypothesis, we used the GFP and GFP-BAG5 SH-SY5Y stable cell lines to demonstrate that the pro-apoptotic effect of BAG5 in the presence of a mitochondrial damaging agent is, in fact, accompanied by a significant reduction in Mcl-1 (De Snoo et al. in preparation). This appears to be a promising avenue of investigation to gain a mechanistic understanding of how BAG5 modulates cell survival and death pathways, and was largely guided by the knowledge that we gained from the interactome.

Interestingly, the interaction between BAG5 and p62 may also be relevant in this context.

Like parkin, p62 has been also been implicated in cell death triggered by mitochondrial damage, as it promotes cell survival by up-regulating the Keap1-Nrf2 pathway (Park, Kang, and Bae

2015). This pathway has been well characterized to protect against the toxic effects of oxidative stress. At baseline conditions Keap1 binds to Nrf2, a transcription factor, and prevents its translocation into the nucleus. Following an increase in oxidative load, Keap1 is selectively degraded, allowing Nrf2 to translocate into the nucleus and up-regulate the transcription of numerous genes with antioxidant and cytoprotective function (Jiang et al. 2015). Park and colleagues demonstrated that p62 is necessary for oxidative stress-induced Keap1 degradation and that p62 KO enhances mitochondrial-damage induced neuronal death by suppressing the

Keap1-Nrf2 pathway (Park, Kang, and Bae 2015). As such, p62 is critical in combatting the cytotoxic effects of oxidative stress that accompany mitochondrial damage.

Considering that we observed BAG5 to enhance cell death in the context of both mitochondrial damage and increased oxidative load, we hypothesized that BAG5 inhibits the cellular response to oxidative stress by suppressing function of p62 in the Keap1-Nrf2 pathway.

This is represents a novel mechanism by which BAG5-p62 interaction could be relevant outside of proteostatic pathways discussed throughout this thesis.

5.5 Conclusion

After mapping the interactome of BAG5, we chose to focus on the interaction between

BAG5 and p62 because of its novelty, and the potential to extend our knowledge of BAG5 in

PD-relevant proteostasis pathways. This lead to the findings that BAG5 and p62 co-localize at perinuclear sequestosomes and that BAG5 supports p62 stability, suggesting that the interaction is relevant to cellular proteostasis. We also confirmed our hypothesis that p62 reduces the presence of alpha-synuclein oligomers and illustrated that BAG5 has the opposite effect. Many other novel BAG5 interactions were identified that could be relevant to its role in proteostasis, such as the interaction with DNAJC13. It will be important to investigate these interactions in more detail moving forward, as they will likely provide important insights into the role of BAG5 within the chaperone network, and may inform the design of novel therapeutics for a variety of proteinopathies, including Parkinson’s disease.

References

Al Shahrani, M., S. Heales, I. Hargreaves, and M. Orford. 2017. "Oxidative Stress: Mechanistic Insights into Inherited Mitochondrial Disorders and Parkinson's Disease." J Clin Med 6 (11). doi: 10.3390/jcm6110100.

Alvarez-Erviti, Lydia, Maria C Rodriguez-Oroz, J Mark Cooper, Cristina Caballero, Isidro Ferrer, Jose A Obeso, and Anthony HV Schapira. 2010. "Chaperone-mediated autophagy markers in Parkinson disease brains." Archives of neurology 67 (12):1464-1472.

Alvarez-Erviti, Lydia, Y Seow, A HV Schapira, María Cruz Rodriguez-Oroz, José A Obeso, and JM Cooper. 2013. "Influence of microRNA deregulation on chaperone-mediated autophagy and α-synuclein pathology in Parkinson's disease." Cell death & disease 4 (3):e545.

Arakawa, Akihiko, Noriko Handa, Noboru Ohsawa, Meiri Shida, Takanori Kigawa, Fumiaki Hayashi, Mikako Shirouzu, and Shigeyuki Yokoyama. 2010. "The C-terminal BAG domain of BAG5 induces conformational changes of the Hsp70 nucleotide-binding domain for ADP-ATP exchange." Structure 18 (3):309-319.

Araki, K., M. Araki, and K. Yamamura. 2002. "Site-directed integration of the cre gene mediated by Cre recombinase using a combination of mutant lox sites." Nucleic Acids Res 30 (19):e103.

Arnold, E. S., S. C. Ling, S. C. Huelga, C. Lagier-Tourenne, M. Polymenidou, D. Ditsworth, H. B. Kordasiewicz, M. McAlonis-Downes, O. Platoshyn, P. A. Parone, S. Da Cruz, K. M. Clutario, D. Swing, L. Tessarollo, M. Marsala, C. E. Shaw, G. W. Yeo, and D. W. Cleveland. 2013. "ALS-linked TDP-43 mutations produce aberrant RNA splicing and adult-onset motor neuron disease without aggregation or loss of nuclear TDP-43." Proc Natl Acad Sci U S A 110 (8):E736-45. doi: 10.1073/pnas.1222809110.

Auluck, P. K., H. Y. Chan, J. Q. Trojanowski, V. M. Lee, and N. M. Bonini. 2002. "Chaperone suppression of alpha-synuclein toxicity in a Drosophila model for Parkinson's disease." Science 295 (5556):865-8. doi: 10.1126/science.1067389.

Babu, J. R., T. Geetha, and M. W. Wooten. 2005. "Sequestosome 1/p62 shuttles polyubiquitinated tau for proteasomal degradation." J Neurochem 94 (1):192-203. doi: 10.1111/j.1471-4159.2005.03181.x.

Behl, C. 2016. "Breaking BAG: The Co-Chaperone BAG3 in Health and Disease." Trends Pharmacol Sci 37 (8):672-88. doi: 10.1016/j.tips.2016.04.007.

Beilina, A., I. N. Rudenko, A. Kaganovich, L. Civiero, H. Chau, S. K. Kalia, L. V. Kalia, E. Lobbestael, R. Chia, K. Ndukwe, J. Ding, M. A. Nalls, M. Olszewski, D. N. Hauser, R. Kumaran, A. M. Lozano, V. Baekelandt, L. E. Greene, J. M. Taymans, E. Greggio, and M. R. Cookson. 2014. "Unbiased screen for interactors of leucine-rich repeat kinase 2 supports a common pathway for sporadic and familial Parkinson disease." Proc Natl Acad Sci U S A 111 (7):2626-31. doi: 10.1073/pnas.1318306111.

89 90

Bendor, J. T., T. P. Logan, and R. H. Edwards. 2013. "The function of alpha-synuclein." Neuron 79 (6):1044-66. doi: 10.1016/j.neuron.2013.09.004.

Bi, L., Q. Yang, J. Yuan, Q. Miao, L. Duan, F. Li, and S. Wang. 2016a. "MicroRNA-127-3p acts as a tumor suppressor in epithelial ovarian cancer by regulating the BAG5 gene." Oncol Rep 36 (5):2563-2570. doi: 10.3892/or.2016.5055.

Bi, Linlin, Qian Yang, Jiani Yuan, Qing Miao, Linrui Duan, Feng Li, and Siwang Wang. 2016b. "MicroRNA-127-3p acts as a tumor suppressor in epithelial ovarian cancer by regulating the BAG5 gene." Oncology reports 36 (5):2563-2570.

Bitto, A., C. A. Lerner, T. Nacarelli, E. Crowe, C. Torres, and C. Sell. 2014. "P62/SQSTM1 at the interface of aging, autophagy, and disease." Age (Dordr) 36 (3):9626. doi: 10.1007/s11357-014-9626-3.

Bjorkoy, G., T. Lamark, and T. Johansen. 2006. "p62/SQSTM1: a missing link between protein aggregates and the autophagy machinery." Autophagy 2 (2):138-9.

Brady, O. A., P. Meng, Y. Zheng, Y. Mao, and F. Hu. 2011. "Regulation of TDP-43 aggregation by phosphorylation and p62/SQSTM1." J Neurochem 116 (2):248-59. doi: 10.1111/j.1471-4159.2010.07098.x.

Bruchmann, A., C. Roller, T. V. Walther, G. Schafer, S. Lehmusvaara, T. Visakorpi, H. Klocker, A. C. Cato, and D. Maddalo. 2013. "Bcl-2 associated athanogene 5 (Bag5) is overexpressed in prostate cancer and inhibits ER-stress induced apoptosis." BMC Cancer 13:96. doi: 10.1186/1471-2407-13-96.

Bruchmann, Anja, Corinna Roller, Tamara Vanessa Walther, Georg Schäfer, Sara Lehmusvaara, Tapio Visakorpi, Helmut Klocker, Andrew CB Cato, and Danilo Maddalo. 2013. "Bcl-2 associated athanogene 5 (Bag5) is overexpressed in prostate cancer and inhibits ER-stress induced apoptosis." BMC cancer 13 (1):96.

Bukau, Bernd, and Arthur L Horwich. 1998. "The Hsp70 and Hsp60 chaperone machines." Cell 92 (3):351-366.

Burbulla, Lena F, Carina Schelling, Hiroki Kato, Doron Rapaport, Dirk Woitalla, Carola Schiesling, Claudia Schulte, Manu Sharma, Thomas Illig, and Peter Bauer. 2010. "Dissecting the role of the mitochondrial chaperone mortalin in Parkinson's disease: functional impact of disease-related variants on mitochondrial homeostasis." Human molecular genetics 19 (22):4437-4452.

Carroll, Richard G, Emilie Hollville, and Seamus J Martin. 2014. "Parkin sensitizes toward apoptosis induced by mitochondrial depolarization through promoting degradation of Mcl-1." Cell reports 9 (4):1538-1553.

Chang, H. C., M. Hull, and I. Mellman. 2004. "The J-domain protein Rme-8 interacts with Hsc70 to control clathrin-dependent endocytosis in Drosophila." J Cell Biol 164 (7):1055-64. doi: 10.1083/jcb.200311084.

Chen, J., and W. A. Weiss. 2015. "Alternative splicing in cancer: implications for biology and therapy." Oncogene 34 (1):1-14. doi: 10.1038/onc.2013.570.

Chen, L., Y. Ding, B. Cagniard, A. D. Van Laar, A. Mortimer, W. Chi, T. G. Hastings, U. J. Kang, and X. Zhuang. 2008. "Unregulated cytosolic dopamine causes neurodegeneration associated with oxidative stress in mice." J Neurosci 28 (2):425-33. doi: 10.1523/jneurosci.3602-07.2008.

Chen, Y., L. N. Yang, L. Cheng, S. Tu, S. J. Guo, H. Y. Le, Q. Xiong, R. Mo, C. Y. Li, J. S. Jeong, L. Jiang, S. Blackshaw, L. J. Bi, H. Zhu, S. C. Tao, and F. Ge. 2013. "Bcl2- associated athanogene 3 interactome analysis reveals a new role in modulating proteasome activity." Mol Cell Proteomics 12 (10):2804-19. doi: 10.1074/mcp.M112.025882.

Chu, Y., and J. H. Kordower. 2015. "The prion hypothesis of Parkinson's disease." Curr Neurol Neurosci Rep 15 (5):28. doi: 10.1007/s11910-015-0549-x.

Ciechanover, A., and Y. T. Kwon. 2017. "Protein Quality Control by Molecular Chaperones in Neurodegeneration." Front Neurosci 11:185. doi: 10.3389/fnins.2017.00185.

Cohen-Kaplan, V., I. Livneh, N. Avni, B. Fabre, T. Ziv, Y. T. Kwon, and A. Ciechanover. 2016. "p62- and ubiquitin-dependent stress-induced autophagy of the mammalian 26S proteasome." Proc Natl Acad Sci U S A 113 (47):E7490-e7499. doi: 10.1073/pnas.1615455113.

Cooper, O., A. Astradsson, P. Hallett, H. Robertson, I. Mendez, and O. Isacson. 2009. "Lack of functional relevance of isolated cell damage in transplants of Parkinson's disease patients." J Neurol 256 Suppl 3:310-6. doi: 10.1007/s00415-009-5242-z.

Cuervo, Ana Maria, Leonidas Stefanis, Ross Fredenburg, Peter T Lansbury, and David Sulzer. 2004. "Impaired degradation of mutant α-synuclein by chaperone-mediated autophagy." Science 305 (5688):1292-1295.

Das, A. T., L. Tenenbaum, and B. Berkhout. 2016. "Tet-On Systems For Doxycycline-inducible Gene Expression." Curr Gene Ther 16 (3):156-67.

Dauer, William, Nikolai Kholodilov, Miquel Vila, Anne-Cecile Trillat, Rose Goodchild, Kristin E Larsen, Roland Staal, Kim Tieu, Yvonne Schmitz, and Chao Annie Yuan. 2002. "Resistance of α-synuclein null mice to the parkinsonian neurotoxin MPTP." Proceedings of the National Academy of Sciences 99 (22):14524-14529.

De Mena, Lorena, Eliecer Coto, Elena Sánchez-Ferrero, René Ribacoba, Luis M Guisasola, Carlos Salvador, Marta Blázquez, and Victoria Alvarez. 2009. "Mutational screening of the mortalin gene (HSPA9) in Parkinson’s disease." Journal of neural transmission 116 (10):1289-1293.

Deng, H., P. Wang, and J. Jankovic. 2017. "The genetics of Parkinson disease." Ageing Res Rev 42:72-85. doi: 10.1016/j.arr.2017.12.007.

Devi, L., V. Raghavendran, B. M. Prabhu, N. G. Avadhani, and H. K. Anandatheerthavarada. 2008. "Mitochondrial import and accumulation of alpha-synuclein impair complex I in human dopaminergic neuronal cultures and Parkinson disease brain." J Biol Chem 283 (14):9089-100. doi: 10.1074/jbc.M710012200.

Di Maio, R., P. J. Barrett, E. K. Hoffman, C. W. Barrett, A. Zharikov, A. Borah, X. Hu, J. McCoy, C. T. Chu, E. A. Burton, T. G. Hastings, and J. T. Greenamyre. 2016. "alpha- Synuclein binds to TOM20 and inhibits mitochondrial protein import in Parkinson's disease." Sci Transl Med 8 (342):342ra78. doi: 10.1126/scitranslmed.aaf3634.

Dong, Zhizhong, David P Wolfer, Hans-Peter Lipp, and Hansruedi Büeler. 2005. "Hsp70 gene transfer by adeno-associated virus inhibits MPTP-induced nigrostriatal degeneration in the mouse model of Parkinson disease." Molecular therapy 11 (1):80-88.

Eyal, A., R. Szargel, E. Avraham, E. Liani, J. Haskin, R. Rott, and S. Engelender. 2006. "Synphilin-1A: an aggregation-prone isoform of synphilin-1 that causes neuronal death and is present in aggregates from alpha-synucleinopathy patients." Proc Natl Acad Sci U S A 103 (15):5917-22. doi: 10.1073/pnas.0509707103.

Flower, Todd R, Liudmila S Chesnokova, Clifford A Froelich, Cheryl Dixon, and Stephan N Witt. 2005. "Heat shock prevents alpha-synuclein-induced apoptosis in a yeast model of Parkinson's disease." Journal of molecular biology 351 (5):1081-1100.

Freeman, C. L., G. Hesketh, and M. N. Seaman. 2014. "RME-8 coordinates the activity of the WASH complex with the function of the retromer SNX dimer to control endosomal tubulation." J Cell Sci 127 (Pt 9):2053-70. doi: 10.1242/jcs.144659.

Freimann, Karen, Katja Zschiedrich, Norbert Brüggemann, Anne Grünewald, Heike Pawlack, Johann Hagenah, Katja Lohmann, Christine Klein, and Ana Westenberger. 2013. "Mortalin mutations are not a frequent cause of early-onset Parkinson disease." Neurobiology of aging 34 (11):2694. e19-2694. e20.

Friesen, E. L., M. L. De Snoo, L. Rajendran, L. V. Kalia, and S. K. Kalia. 2017. "Chaperone- Based Therapies for Disease Modification in Parkinson's Disease." Parkinsons Dis 2017:5015307. doi: 10.1155/2017/5015307.

Gamerdinger, M., P. Hajieva, A. M. Kaya, U. Wolfrum, F. U. Hartl, and C. Behl. 2009. "Protein quality control during aging involves recruitment of the macroautophagy pathway by BAG3." Embo j 28 (7):889-901. doi: 10.1038/emboj.2009.29.

Gao, Xuechao, Marta Carroni, Carmen Nussbaum-Krammer, Axel Mogk, Nadinath B Nillegoda, Anna Szlachcic, D Lys Guilbride, Helen R Saibil, Matthias P Mayer, and Bernd Bukau. 2015. "Human Hsp70 disaggregase reverses Parkinson’s-linked α-synuclein amyloid fibrils." Molecular cell 59 (5):781-793.

Girard, M., and P. S. McPherson. 2008. "RME-8 regulates trafficking of the epidermal growth factor receptor." FEBS Lett 582 (6):961-6. doi: 10.1016/j.febslet.2008.02.042.

Girard, M., V. Poupon, F. Blondeau, and P. S. McPherson. 2005. "The DnaJ-domain protein RME-8 functions in endosomal trafficking." J Biol Chem 280 (48):40135-43. doi: 10.1074/jbc.M505036200.

Goloubinoff, Pierre, and Paolo De Los Rios. 2007. "The mechanism of Hsp70 chaperones:(entropic) pulling the models together." Trends in biochemical sciences 32 (8):372-380.

Guo, K., L. Li, G. Yin, X. Zi, and L. Liu. 2015a. "Bag5 protects neuronal cells from amyloid beta-induced cell death." J Mol Neurosci 55 (4):815-20. doi: 10.1007/s12031-014-0433- 1.

Guo, Ke, Liuhong Li, Gang Yin, Xiaohong Zi, and Lei Liu. 2015b. "Bag5 Protects Neuronal Cells from Amyloid β-induced Cell Death." Journal of Molecular Neuroscience 55 (4):815-820.

Gupta, M. K., F. G. Tahrir, T. Knezevic, M. K. White, J. Gordon, J. Y. Cheung, K. Khalili, and A. M. Feldman. 2016. "GRP78 Interacting Partner Bag5 Responds to ER Stress and Protects Cardiomyocytes From ER Stress-Induced Apoptosis." J Cell Biochem 117 (8):1813-21. doi: 10.1002/jcb.25481.

Hadizadeh Esfahani, A., A. Sverchkova, J. Saez-Rodriguez, A. A. Schuppert, and M. Brehme. 2018. "A systematic atlas of chaperome deregulation topologies across the human cancer landscape." PLoS Comput Biol 14 (1):e1005890. doi: 10.1371/journal.pcbi.1005890.

Hasson, Samuel A., Lesley A. Kane, Koji Yamano, Chiu-Hui Huang, Danielle A. Sliter, Eugen Buehler, Chunxin Wang, Sabrina M. Heman-Ackah, Tara Hessa, and Rajarshi Guha. 2013. "High-content genome-wide RNAi screens identify regulators of parkin upstream of mitophagy." Nature 504 (7479):291-295.

Höhfeld, Jörg, and Stefan Jentsch. 1997. "GrpE‐like regulation of the Hsc70 chaperone by the anti‐apoptotic protein BAG‐1." The EMBO journal 16 (20):6209-6216.

Imai, Yuzuru, Mariko Soda, Shigetsugu Hatakeyama, Takumi Akagi, Tsutomu Hashikawa, Kei- Ichi Nakayama, and Ryosuke Takahashi. 2002. "CHIP is associated with Parkin, a gene responsible for familial Parkinson's disease, and enhances its ubiquitin ligase activity." Molecular cell 10 (1):55-67.

Iwai, A., E. Masliah, M. Yoshimoto, N. Ge, L. Flanagan, H. A. de Silva, A. Kittel, and T. Saitoh. 1995. "The precursor protein of non-A beta component of Alzheimer's disease amyloid is a presynaptic protein of the central nervous system." Neuron 14 (2):467-75.

Jenner, P., D. T. Dexter, J. Sian, A. H. Schapira, and C. D. Marsden. 1992. "Oxidative stress as a cause of nigral cell death in Parkinson's disease and incidental Lewy body disease. The Royal Kings and Queens Parkinson's Disease Research Group." Ann Neurol 32 Suppl:S82-7.

Jiang, T., B. Harder, M. Rojo de la Vega, P. K. Wong, E. Chapman, and D. D. Zhang. 2015. "p62 links autophagy and Nrf2 signaling." Free Radic Biol Med 88 (Pt B):199-204. doi: 10.1016/j.freeradbiomed.2015.06.014.

Jones, D. R., M. Delenclos, A. T. Baine, M. DeTure, M. E. Murray, D. W. Dickson, and P. J. McLean. 2015. "Transmission of Soluble and Insoluble alpha-Synuclein to Mice." J Neuropathol Exp Neurol 74 (12):1158-69. doi: 10.1097/nen.0000000000000262.

K Kalia, S, L V Kalia, and P J McLean. 2010. "Molecular chaperones as rational drug targets for Parkinson's disease therapeutics." CNS & Neurological Disorders-Drug Targets (Formerly Current Drug Targets-CNS & Neurological Disorders) 9 (6):741-753.

Kabbage, M., and M. B. Dickman. 2008a. "The BAG proteins: a ubiquitous family of chaperone regulators." Cellular and Molecular Life Sciences 65 (9):1390-1402.

Kabbage, M., and M. B. Dickman. 2008b. "The BAG proteins: a ubiquitous family of chaperone regulators." Cell Mol Life Sci 65 (9):1390-402. doi: 10.1007/s00018-008-7535-2.

Kalia, L. V., S. K. Kalia, H. Chau, A. M. Lozano, B. T. Hyman, and P. J. McLean. 2011. "Ubiquitinylation of alpha-synuclein by carboxyl terminus Hsp70-interacting protein (CHIP) is regulated by Bcl-2-associated athanogene 5 (BAG5)." PLoS One 6 (2):e14695. doi: 10.1371/journal.pone.0014695.

Kalia, L. V., A. E. Lang, L. N. Hazrati, S. Fujioka, Z. K. Wszolek, D. W. Dickson, O. A. Ross, V. M. Van Deerlin, J. Q. Trojanowski, H. I. Hurtig, R. N. Alcalay, K. S. Marder, L. N. Clark, C. Gaig, E. Tolosa, J. Ruiz-Martinez, J. F. Marti-Masso, I. Ferrer, A. Lopez de Munain, S. M. Goldman, B. Schule, J. W. Langston, J. O. Aasly, M. T. Giordana, V. Bonifati, A. Puschmann, M. Canesi, G. Pezzoli, A. Maues De Paula, K. Hasegawa, C. Duyckaerts, A. Brice, A. J. Stoessl, and C. Marras. 2015. "Clinical correlations with Lewy body pathology in LRRK2-related Parkinson disease." JAMA Neurol 72 (1):100-5. doi: 10.1001/jamaneurol.2014.2704.

Kalia, Lorraine V, Suneil K Kalia, Pamela J McLean, Andres M Lozano, and Anthony E Lang. 2013. "α‐Synuclein oligomers and clinical implications for Parkinson disease." Annals of neurology 73 (2):155-169.

Kalia, Lorraine V., and Anthony E. Lang. 2015. "Parkinson's disease." The Lancet 386 (9996):896-912. doi: 10.1016/s0140-6736(14)61393-3.

Kalia, S. K., S. Lee, P. D. Smith, L. Liu, S. J. Crocker, T. E. Thorarinsdottir, J. R. Glover, E. A. Fon, D. S. Park, and A. M. Lozano. 2004a. "BAG5 inhibits parkin and enhances dopaminergic neuron degeneration." Neuron 44 (6):931-45. doi: 10.1016/j.neuron.2004.11.026.

Kalia, Suneil K., Sang Lee, Patrice D. Smith, Li Liu, Stephen J. Crocker, Thorhildur E. Thorarinsdottir, John R. Glover, Edward A. Fon, David S. Park, and Andres M. Lozano. 2004b. "BAG5 inhibits parkin and enhances dopaminergic neuron degeneration." Neuron 44 (6):931-945.

Kampinga, Harm H, and Steven Bergink. 2016. "Heat shock proteins as potential targets for protective strategies in neurodegeneration." The Lancet Neurology 15 (7):748-759.

Kampinga, Harm H, Jurre Hageman, Michel J Vos, Hiroshi Kubota, Robert M Tanguay, Elspeth A Bruford, Michael E Cheetham, Bin Chen, and Lawrence E Hightower. 2009. "Guidelines for the nomenclature of the human heat shock proteins." Cell Stress and Chaperones 14 (1):105-111.

Karampetsou, M., M. T. Ardah, M. Semitekolou, A. Polissidis, M. Samiotaki, M. Kalomoiri, N. Majbour, G. Xanthou, O. M. A. El-Agnaf, and K. Vekrellis. 2017. "Phosphorylated exogenous alpha-synuclein fibrils exacerbate pathology and induce neuronal dysfunction in mice." Sci Rep 7 (1):16533. doi: 10.1038/s41598-017-15813-8.

Karpinar, D. P., M. B. Balija, S. Kugler, F. Opazo, N. Rezaei-Ghaleh, N. Wender, H. Y. Kim, G. Taschenberger, B. H. Falkenburger, H. Heise, A. Kumar, D. Riedel, L. Fichtner, A. Voigt, G. H. Braus, K. Giller, S. Becker, A. Herzig, M. Baldus, H. Jackle, S. Eimer, J. B. Schulz, C. Griesinger, and M. Zweckstetter. 2009. "Pre-fibrillar alpha-synuclein variants with impaired beta-structure increase neurotoxicity in Parkinson's disease models." Embo j 28 (20):3256-68. doi: 10.1038/emboj.2009.257.

Katsuragi, Y., Y. Ichimura, and M. Komatsu. 2015. "p62/SQSTM1 functions as a signaling hub and an autophagy adaptor." Febs j 282 (24):4672-8. doi: 10.1111/febs.13540.

Kelley, William L. 1999. "Molecular chaperones: how J domains turn on Hsp70s." Current Biology 9 (8):R305-R308.

Kett, L. R., and W. T. Dauer. 2016. "Endolysosomal dysfunction in Parkinson's disease: Recent developments and future challenges." Mov Disord 31 (10):1433-1443. doi: 10.1002/mds.26797.

Klucken, J., Y. Shin, B. T. Hyman, and P. J. McLean. 2004. "A single amino acid substitution differentiates Hsp70-dependent effects on alpha-synuclein degradation and toxicity." Biochem Biophys Res Commun 325 (1):367-73. doi: 10.1016/j.bbrc.2004.10.037.

Klucken, Jochen, Youngah Shin, Eliezer Masliah, Bradley T Hyman, and Pamela J McLean. 2004. "Hsp70 reduces α-synuclein aggregation and toxicity." Journal of Biological Chemistry 279 (24):25497-25502.

Komatsu, M., S. Waguri, M. Koike, Y. S. Sou, T. Ueno, T. Hara, N. Mizushima, J. Iwata, J. Ezaki, S. Murata, J. Hamazaki, Y. Nishito, S. Iemura, T. Natsume, T. Yanagawa, J. Uwayama, E. Warabi, H. Yoshida, T. Ishii, A. Kobayashi, M. Yamamoto, Z. Yue, Y. Uchiyama, E. Kominami, and K. Tanaka. 2007. "Homeostatic levels of p62 control cytoplasmic inclusion body formation in autophagy-deficient mice." Cell 131 (6):1149- 63. doi: 10.1016/j.cell.2007.10.035.

Koprich, J. B., T. H. Johnston, M. G. Reyes, X. Sun, and J. M. Brotchie. 2010. "Expression of human A53T alpha-synuclein in the rat substantia nigra using a novel AAV1/2 vector produces a rapidly evolving pathology with protein aggregation, dystrophic neurite

architecture and nigrostriatal degeneration with potential to model the pathology of Parkinson's disease." Mol Neurodegener 5:43. doi: 10.1186/1750-1326-5-43.

Koprich, J. B., L. V. Kalia, and J. M. Brotchie. 2017. "Animal models of alpha-synucleinopathy for Parkinson disease drug development." Nat Rev Neurosci 18 (9):515-529. doi: 10.1038/nrn.2017.75.

Kumaran, R., and M. R. Cookson. 2015. "Pathways to Parkinsonism Redux: convergent pathobiological mechanisms in genetics of Parkinson's disease." Hum Mol Genet 24 (R1):R32-44. doi: 10.1093/hmg/ddv236.

La Cognata, V., V. D'Agata, F. Cavalcanti, and S. Cavallaro. 2015. "Splicing: is there an alternative contribution to Parkinson's disease?" Neurogenetics 16 (4):245-63. doi: 10.1007/s10048-015-0449-x.

Langston, JW, P Ballard, JW Tetrud, and I Irwin. 1983. "Chronic Parkinsonism in humans due to a product of meperidine-analog synthesis." Science 219 (4587):979-980. doi: 10.1126/science.6823561.

Lazaro, D. F., E. F. Rodrigues, R. Langohr, H. Shahpasandzadeh, T. Ribeiro, P. Guerreiro, E. Gerhardt, K. Krohnert, J. Klucken, M. D. Pereira, B. Popova, N. Kruse, B. Mollenhauer, S. O. Rizzoli, G. H. Braus, K. M. Danzer, and T. F. Outeiro. 2014. "Systematic comparison of the effects of alpha-synuclein mutations on its oligomerization and aggregation." PLoS Genet 10 (11):e1004741. doi: 10.1371/journal.pgen.1004741.

Lesage, Suzanne, and Alexis Brice. 2009. "Parkinson's disease: from monogenic forms to genetic susceptibility factors." Human Molecular Genetics 18 (R1):R48-R59. doi: 10.1093/hmg/ddp012.

Li, J. Y., E. Englund, J. L. Holton, D. Soulet, P. Hagell, A. J. Lees, T. Lashley, N. P. Quinn, S. Rehncrona, A. Bjorklund, H. Widner, T. Revesz, O. Lindvall, and P. Brundin. 2008. "Lewy bodies in grafted neurons in subjects with Parkinson's disease suggest host-to- graft disease propagation." Nat Med 14 (5):501-3. doi: 10.1038/nm1746.

Liu, H., C. Dai, Y. Fan, B. Guo, K. Ren, T. Sun, and W. Wang. 2017. "From autophagy to mitophagy: the roles of P62 in neurodegenerative diseases." J Bioenerg Biomembr 49 (5):413-422. doi: 10.1007/s10863-017-9727-7.

Lonskaya, I., N. M. Desforges, M. L. Hebron, and C. E. Moussa. 2013. "Ubiquitination increases parkin activity to promote autophagic alpha-synuclein clearance." PLoS One 8 (12):e83914. doi: 10.1371/journal.pone.0083914.

Lozano, A. M., W. D. Hutchison, and S. K. Kalia. 2017. "What Have We Learned About Movement Disorders from Functional Neurosurgery?" Annu Rev Neurosci 40:453-477. doi: 10.1146/annurev-neuro-070815-013906.

Ma, Mingming, Xuejing Wang, Xuebing Ding, Jing Jing, Yaohua Ma, and Junfang Teng. 2012. "Protective effect of BAG5 on MPP+-induced apoptosis in PC12 cells." Neurological research 34 (10):977-983.

Mak, Sally K, Alison L McCormack, Amy B Manning-Boğ, Ana Maria Cuervo, and Donato A Di Monte. 2010. "Lysosomal degradation of α-synuclein in vivo." Journal of Biological Chemistry 285 (18):13621-13629.

Matsumoto, G., K. Wada, M. Okuno, M. Kurosawa, and N. Nukina. 2011. "Serine 403 phosphorylation of p62/SQSTM1 regulates selective autophagic clearance of ubiquitinated proteins." Mol Cell 44 (2):279-89. doi: 10.1016/j.molcel.2011.07.039.

McLean, P. J., J. Klucken, Y. Shin, and B. T. Hyman. 2004. "Geldanamycin induces Hsp70 and prevents alpha-synuclein aggregation and toxicity in vitro." Biochem Biophys Res Commun 321 (3):665-9. doi: 10.1016/j.bbrc.2004.07.021.

McLean, Pamela J, Hibiki Kawamata, Saadat Shariff, Jeffrey Hewett, Nutan Sharma, Kenji Ueda, Xandra O Breakefield, and Bradley T Hyman. 2002. "TorsinA and heat shock proteins act as molecular chaperones: suppression of α‐synuclein aggregation." Journal of neurochemistry 83 (4):846-854.

Meacham, Geoffrey C, Cam Patterson, Wenyue Zhang, J Michael Younger, and Douglas M Cyr. 2001. "The Hsc70 co-chaperone CHIP targets immature CFTR for proteasomal degradation." Nature cell biology 3 (1):100-105.

Moloney, Teresa C, Rhona Hyland, Daniel O'Toole, Alexia Paucard, Deniz Kirik, Aideen O'Doherty, Adrienne M Gorman, and Eilís Dowd. 2014. "Heat Shock Protein 70 Reduces α‐Synuclein‐Induced Predegenerative Neuronal Dystrophy in the α‐Synuclein Viral Gene Transfer Rat Model of Parkinson's Disease." CNS neuroscience & therapeutics 20 (1):50-58.

Morimoto, Richard I. 1998. "Regulation of the heat shock transcriptional response: cross talk between a family of heat shock factors, molecular chaperones, and negative regulators." Genes & development 12 (24):3788-3796.

Murata, Shigeo, Yasufumi Minami, Michiko Minami, Tomoki Chiba, and Keiji Tanaka. 2001. "CHIP is a chaperone‐dependent E3 ligase that ubiquitylates unfolded protein." EMBO reports 2 (12):1133-1138.

Murphy, Karen E, Amanda M Gysbers, Sarah K Abbott, Adena S Spiro, Akiko Furuta, Antony Cooper, Brett Garner, Tomohiro Kabuta, and Glenda M Halliday. 2015. "Lysosomal‐ associated membrane protein 2 isoforms are differentially affected in early Parkinson's disease." Movement Disorders 30 (12):1639-1647.

Narendra, Derek, Atsushi Tanaka, Der-Fen Suen, and Richard J. Youle. 2008. "Parkin is recruited selectively to impaired mitochondria and promotes their autophagy." The Journal of cell biology 183 (5):795-803.

Nillegoda, N. B., and B. Bukau. 2015. "Metazoan Hsp70-based protein disaggregases: emergence and mechanisms." Front Mol Biosci 2:57. doi: 10.3389/fmolb.2015.00057.

Oceguera-Yanez, F., S. I. Kim, T. Matsumoto, G. W. Tan, L. Xiang, T. Hatani, T. Kondo, M. Ikeya, Y. Yoshida, H. Inoue, and K. Woltjen. 2016. "Engineering the AAVS1 locus for

consistent and scalable transgene expression in human iPSCs and their differentiated derivatives." Methods 101:43-55. doi: 10.1016/j.ymeth.2015.12.012.

Outeiro, T. F., P. Putcha, J. E. Tetzlaff, R. Spoelgen, M. Koker, F. Carvalho, B. T. Hyman, and P. J. McLean. 2008. "Formation of toxic oligomeric alpha-synuclein species in living cells." PLoS One 3 (4):e1867. doi: 10.1371/journal.pone.0001867.

Pankiv, S., T. Lamark, J. A. Bruun, A. Overvatn, G. Bjorkoy, and T. Johansen. 2010. "Nucleocytoplasmic shuttling of p62/SQSTM1 and its role in recruitment of nuclear polyubiquitinated proteins to promyelocytic leukemia bodies." J Biol Chem 285 (8):5941-53. doi: 10.1074/jbc.M109.039925.

Papagiannakis, Nikolaos, Maria Xilouri, Christos Koros, Maria Stamelou, Roubina Antonelou, Matina Maniati, Dimitra Papadimitriou, Marina Moraitou, Helen Michelakakis, and Leonidas Stefanis. 2015. "Lysosomal alterations in peripheral blood mononuclear cells of Parkinson's disease patients." Movement Disorders 30 (13):1830-1834.

Park, J. S., D. H. Kang, and S. H. Bae. 2015. "p62 prevents carbonyl cyanide m-chlorophenyl hydrazine (CCCP)-induced apoptotic cell death by activating Nrf2." Biochem Biophys Res Commun 464 (4):1139-44. doi: 10.1016/j.bbrc.2015.07.093.

Pickrell, Alicia M, and Richard J Youle. 2015. "The roles of PINK1, parkin, and mitochondrial fidelity in Parkinson’s disease." Neuron 85 (2):257-273.

Polymeropoulos, Mihael H, Christian Lavedan, Elisabeth Leroy, Susan E Ide, Anindya Dehejia, Amalia Dutra, Brian Pike, Holly Root, Jeffrey Rubenstein, and Rebecca Boyer. 1997. "Mutation in the α-synuclein gene identified in families with Parkinson's disease." science 276 (5321):2045-2047.

Putcha, P., K. M. Danzer, L. R. Kranich, A. Scott, M. Silinski, S. Mabbett, C. D. Hicks, J. M. Veal, P. M. Steed, B. T. Hyman, and P. J. McLean. 2010. "Brain-permeable small- molecule inhibitors of Hsp90 prevent alpha-synuclein oligomer formation and rescue alpha-synuclein-induced toxicity." J Pharmacol Exp Ther 332 (3):849-57. doi: 10.1124/jpet.109.158436.

Qu, Dianbo, Ali Hage, Katie Don-Carolis, En Huang, Alvin Joselin, Farzaneh Safarpour, Paul C. Marcogliese, Maxime W. C. Rousseaux, Sarah J. Hewitt, and Tianwen Huang. 2015. "BAG2 Gene-mediated Regulation of PINK1 Protein Is Critical for Mitochondrial Translocation of PARKIN and Neuronal Survival." Journal of Biological Chemistry 290 (51):30441-30452.

Quigney, Declan J, Adrienne M Gorman, and Afshin Samali. 2003. "Heat shock protects PC12 cells against MPP+ toxicity." Brain research 993 (1):133-139.

Rampelt, Heike, Janine Kirstein‐Miles, Nadinath B Nillegoda, Kang Chi, Sebastian R Scholz, Richard I Morimoto, and Bernd Bukau. 2012. "Metazoan Hsp70 machines use Hsp110 to power protein disaggregation." The EMBO journal 31 (21):4221-4235.

Ran, F. A., P. D. Hsu, C. Y. Lin, J. S. Gootenberg, S. Konermann, A. E. Trevino, D. A. Scott, A. Inoue, S. Matoba, Y. Zhang, and F. Zhang. 2013. "Double nicking by RNA-guided CRISPR Cas9 for enhanced genome editing specificity." Cell 154 (6):1380-9. doi: 10.1016/j.cell.2013.08.021.

Remy, I., and S. W. Michnick. 2006. "A highly sensitive protein-protein interaction assay based on Gaussia luciferase." Nat Methods 3 (12):977-9. doi: 10.1038/nmeth979.

Rosborough, K., N. Patel, and L. V. Kalia. 2017. "alpha-Synuclein and Parkinsonism: Updates and Future Perspectives." Curr Neurol Neurosci Rep 17 (4):31. doi: 10.1007/s11910- 017-0737-y.

Rüdiger, Stefan, Lothar Germeroth, Jens Schneider‐Mergener, and Bernd Bukau. 1997. "Substrate specificity of the DnaK chaperone determined by screening cellulose‐bound peptide libraries." The EMBO journal 16 (7):1501-1507.

Sadelain, M., E. P. Papapetrou, and F. D. Bushman. 2011. "Safe harbours for the integration of new DNA in the human genome." In Nat Rev Cancer, 51-8. England.

Sala, Gessica, Giovanni Stefanoni, Alessandro Arosio, Chiara Riva, Laura Melchionda, Enrico Saracchi, Silvia Fermi, Laura Brighina, and Carlo Ferrarese. 2014. "Reduced expression of the chaperone-mediated autophagy carrier hsc70 protein in lymphomonocytes of patients with Parkinson's disease." Brain research 1546:46-52.

Sauer, B. 2002. "Cre/lox: one more step in the taming of the genome." Endocrine 19 (3):221-8.

Schapira, A. H., J. M. Cooper, D. Dexter, J. B. Clark, P. Jenner, and C. D. Marsden. 1990. "Mitochondrial complex I deficiency in Parkinson's disease." J Neurochem 54 (3):823-7.

Schneider, S. A., and R. N. Alcalay. 2017. "Neuropathology of genetic synucleinopathies with parkinsonism: Review of the literature." Mov Disord 32 (11):1504-1523. doi: 10.1002/mds.27193.

Shi, A., L. Sun, R. Banerjee, M. Tobin, Y. Zhang, and B. D. Grant. 2009. "Regulation of endosomal clathrin and retromer-mediated endosome to Golgi retrograde transport by the J-domain protein RME-8." Embo j 28 (21):3290-302. doi: 10.1038/emboj.2009.272.

Shin, Youngah, Jochen Klucken, Cam Patterson, Bradley T Hyman, and Pamela J McLean. 2005. "The co-chaperone carboxyl terminus of Hsp70-interacting protein (CHIP) mediates α-synuclein degradation decisions between proteasomal and lysosomal pathways." Journal of Biological Chemistry 280 (25):23727-23734.

Song, P., S. Li, H. Wu, R. Gao, G. Rao, D. Wang, Z. Chen, B. Ma, H. Wang, N. Sui, H. Deng, Z. Zhang, T. Tang, Z. Tan, Z. Han, T. Lu, Y. Zhu, and Q. Chen. 2016. "Parkin promotes proteasomal degradation of p62: implication of selective vulnerability of neuronal cells in the pathogenesis of Parkinson's disease." Protein Cell 7 (2):114-29. doi: 10.1007/s13238-015-0230-9.

100

Spillantini, Maria Grazia, Marie Luise Schmidt, Virginia M-Y Lee, John Q Trojanowski, Ross Jakes, and Michel Goedert. 1997. "α-Synuclein in Lewy bodies." Nature 388 (6645):839-840.

Stefanis, L. 2012. "alpha-Synuclein in Parkinson's disease." Cold Spring Harb Perspect Med 2 (2):a009399. doi: 10.1101/cshperspect.a009399.

Takayama, S., D. N. Bimston, S. Matsuzawa, B. C. Freeman, C. Aime-Sempe, Z. Xie, R. I. Morimoto, and J. C. Reed. 1997. "BAG-1 modulates the chaperone activity of Hsp70/Hsc70." Embo j 16 (16):4887-96. doi: 10.1093/emboj/16.16.4887.

Takayama, S., T. Sato, S. Krajewski, K. Kochel, S. Irie, J. A. Millan, and J. C. Reed. 1995. "Cloning and functional analysis of BAG-1: a novel Bcl-2-binding protein with anti-cell death activity." Cell 80 (2):279-84.

Taldone, T., S. O. Ochiana, P. D. Patel, and G. Chiosis. 2014. "Selective targeting of the stress chaperome as a therapeutic strategy." Trends Pharmacol Sci 35 (11):592-603. doi: 10.1016/j.tips.2014.09.001.

Tanji, K., S. Odagiri, Y. Miki, A. Maruyama, Y. Nikaido, J. Mimura, F. Mori, E. Warabi, T. Yanagawa, S. Ueno, K. Itoh, and K. Wakabayashi. 2015. "p62 Deficiency Enhances alpha-Synuclein Pathology in Mice." Brain Pathol 25 (5):552-64. doi: 10.1111/bpa.12214.

Tanner, Caroline M., Freya Kamel, G. Webster Ross, Jane A. Hoppin, Samuel M. Goldman, Monica Korell, Connie Marras, Grace S. Bhudhikanok, Meike Kasten, Anabel R. Chade, Kathleen Comyns, Marie Barber Richards, Cheryl Meng, Benjamin Priestley, Hubert H. Fernandez, Franca Cambi, David M. Umbach, Aaron Blair, Dale P. Sandler, and J. William Langston. 2011. "Rotenone, Paraquat, and Parkinson’s Disease." Environmental Health Perspectives 119 (6):866-872. doi: 10.1289/ehp.1002839.

Vilariño-Güell, Carles, Alex Rajput, Austen J Milnerwood, Brinda Shah, Chelsea Szu-Tu, Joanne Trinh, Irene Yu, Mary Encarnacion, Lise N Munsie, and Lucia Tapia. 2014. "DNAJC13 mutations in Parkinson disease." Human molecular genetics 23 (7):1794- 1801.

Visanji, N. P., P. L. Brooks, L. N. Hazrati, and A. E. Lang. 2013. "The prion hypothesis in Parkinson's disease: Braak to the future." Acta Neuropathol Commun 1:2. doi: 10.1186/2051-5960-1-2.

Volles, M. J., S. J. Lee, J. C. Rochet, M. D. Shtilerman, T. T. Ding, J. C. Kessler, and P. T. Lansbury, Jr. 2001. "Vesicle permeabilization by protofibrillar alpha-synuclein: implications for the pathogenesis and treatment of Parkinson's disease." Biochemistry 40 (26):7812-9.

Wadhwa, Renu, Jihoon Ryu, Hyo Min Ahn, Nishant Saxena, Anupama Chaudhary, Chae-Ok Yun, and Sunil C Kaul. 2015. "Functional significance of point mutations in stress chaperone mortalin and their relevance to Parkinson disease." Journal of Biological Chemistry 290 (13):8447-8456.

101

Wang, C., C. Zhao, D. Li, Z. Tian, Y. Lai, J. Diao, and C. Liu. 2016. "Versatile Structures of alpha-Synuclein." Front Mol Neurosci 9:48. doi: 10.3389/fnmol.2016.00048.

Wang, X., J. Guo, E. Fei, Y. Mu, S. He, X. Che, J. Tan, K. Xia, Z. Zhang, G. Wang, and B. Tang. 2014. "BAG5 protects against mitochondrial oxidative damage through regulating PINK1 degradation." PLoS One 9 (1):e86276. doi: 10.1371/journal.pone.0086276.

Watanabe, Y., H. Tatebe, K. Taguchi, Y. Endo, T. Tokuda, T. Mizuno, M. Nakagawa, and M. Tanaka. 2012. "p62/SQSTM1-dependent autophagy of Lewy body-like alpha-synuclein inclusions." PLoS One 7 (12):e52868. doi: 10.1371/journal.pone.0052868.

Winner, Beate, Roberto Jappelli, Samir K Maji, Paula A Desplats, Leah Boyer, Stefan Aigner, Claudia Hetzer, Thomas Loher, Marçal Vilar, and Silvia Campioni. 2011. "In vivo demonstration that α-synuclein oligomers are toxic." Proceedings of the National Academy of Sciences 108 (10):4194-4199.

Wong, Y. C., and D. Krainc. 2017. "alpha-synuclein toxicity in neurodegeneration: mechanism and therapeutic strategies." Nat Med 23 (2):1-13. doi: 10.1038/nm.4269.

Wu, Yih-Ru, Cheng-Kuang Wang, Chiung-Mei Chen, Yuying Hsu, Sih-Jing Lin, Yi-Ying Lin, Hon-Chung Fung, Kuo-Hsuan Chang, and Guey-Jen Lee-Chen. 2004. "Analysis of heat- shock protein 70 gene polymorphisms and the risk of Parkinson’s disease." Human genetics 114 (3):236-241.

Xiao, S., T. Sanelli, H. Chiang, Y. Sun, A. Chakrabartty, J. Keith, E. Rogaeva, L. Zinman, and J. Robertson. 2015. "Low molecular weight species of TDP-43 generated by abnormal splicing form inclusions in amyotrophic lateral sclerosis and result in motor neuron death." Acta Neuropathol 130 (1):49-61. doi: 10.1007/s00401-015-1412-5.

Xilouri, M., and L. Stefanis. 2015. "Chaperone mediated autophagy to the rescue: A new-fangled target for the treatment of neurodegenerative diseases." Mol Cell Neurosci 66 (Pt A):29- 36. doi: 10.1016/j.mcn.2015.01.003.

Xilouri, Maria, Oeystein Roed Brekk, and Leonidas Stefanis. 2016. "Autophagy and Alpha‐ Synuclein: Relevance to Parkinson's Disease and Related Synucleopathies." Movement Disorders 31 (2):178-192.

Xiong, Y., S. Neifert, S. S. Karuppagounder, J. N. Stankowski, B. D. Lee, J. C. Grima, G. Chen, H. S. Ko, Y. Lee, D. Swing, L. Tessarollo, T. M. Dawson, and V. L. Dawson. 2017. "Overexpression of Parkinson's Disease-Associated Mutation LRRK2 G2019S in Mouse Forebrain Induces Behavioral Deficits and alpha-Synuclein Pathology." eNeuro 4 (2). doi: 10.1523/eneuro.0004-17.2017.

Yoshida, S., T. Hasegawa, M. Suzuki, N. Sugeno, J. Kobayashi, M. Ueyama, M. Fukuda, A. Ido- Fujibayashi, K. Sekiguchi, M. Ezura, A. Kikuchi, T. Baba, A. Takeda, H. Mochizuki, Y. Nagai, and M. Aoki. 2018. "Parkinson's disease-linked DNAJC13 mutation aggravates alpha-synuclein-induced neurotoxicity through perturbation of endosomal trafficking." Hum Mol Genet. doi: 10.1093/hmg/ddy003.

102

Yue, X., Y. Zhao, G. Huang, J. Li, J. Zhu, Z. Feng, and W. Hu. 2016. "A novel mutant p53 binding partner BAG5 stabilizes mutant p53 and promotes mutant p53 GOFs in tumorigenesis." Cell Discov 2:16039. doi: 10.1038/celldisc.2016.39.

Zatloukal, K., C. Stumptner, A. Fuchsbichler, H. Heid, M. Schnoelzer, L. Kenner, R. Kleinert, M. Prinz, A. Aguzzi, and H. Denk. 2002. "p62 Is a common component of cytoplasmic inclusions in protein aggregation diseases." Am J Pathol 160 (1):255-63. doi: 10.1016/s0002-9440(10)64369-6.

Zhang, Conggang, Schuyler Lee, Yinghua Peng, Eric Bunker, Emilie Giaime, Jie Shen, Zongyao Zhou, and Xuedong Liu. 2014. "PINK1 triggers autocatalytic activation of Parkin to specify cell fate decisions." Current Biology 24 (16):1854-1865.

Zhang, Li, Peter Karsten, Sabine Hamm, Joe H Pogson, A Kathrin Müller-Rischart, Nicole Exner, Christian Haass, Alexander J Whitworth, Konstanze F Winklhofer, and Jörg B Schulz. 2013. "TRAP1 rescues PINK1 loss-of-function phenotypes." Human molecular genetics:ddt132.

Zhang, Zhelin, and Yan Cheng. 2014. "miR-16-1 promotes the aberrant α-synuclein accumulation in parkinson disease via targeting heat shock protein 70." The Scientific World Journal 2014.

Zou, Jiangying, Yongle Guo, Toumy Guettouche, David F Smith, and Richard Voellmy. 1998. "Repression of heat shock transcription factor HSF1 activation by HSP90 (HSP90 complex) that forms a stress-sensitive complex with HSF1." Cell 94 (4):471-480.

Appendices

Appendix 1 BAG5 Interactome: H4

Reclassified as GFP IP GFP-BAG5 IP GFP-DARA IP BAG5 Gene (Spectrum (Spectrum (Spectrum Interacting Identified Proteins Name Counts) Counts) Counts) Protein DnaJ homolog subfamily C member 13 DNAJC13 0 83 0 Isoform 2 of BAG family molecular chaperone BAG5 regulator 5 0 66 0

Tubulin beta-2A chain TUBB2A 0 49 0

Very large A-kinase anchor protein CRYBG3 0 31 0

Isoform 2 of Microtubule-associated protein 1A MAP1A 0 30 0

Isoform Beta of Heat shock protein 105 kDa HSPH1 0 29 0

Heat shock 70 kDa protein 1-like HSPA1L 0 29 0

BAG family molecular chaperone regulator 3 BAG3 0 27 0

Isoform 5 of Protein transport protein Sec16A SEC16A 0 25 0

Aryl hydrocarbon receptor AHR 0 22 0

Heat shock 70 kDa protein 6 HSPA6 0 22 0 Isoform 2 of DnaJ homolog subfamily C member DNAJC7 7 0 19 0

Isoform 3 of Nuclear receptor corepressor 2 NCOR2 0 19 0

Pre-mRNA-processing-splicing factor 8 PRPF8 0 18 0

Heat shock-related 70 kDa protein 2 HSPA2 0 18 0 U5 small nuclear ribonucleoprotein 200 kDa SNRNP200 helicase 0 17 0

Melanoma-associated antigen C1 MAGEC1 0 17 0

Myeloid leukemia factor 2 MLF2 0 16 0 F-box-like/WD repeat-containing protein TBL1XR1 TBL1XR1 0 16 0

182 kDa tankyrase-1-binding protein TNKS1BP1 0 14 0

Microtubule-associated protein 1B MAP1B 0 14 0

DNA-directed RNA polymerase II subunit RPB2 POLR2B 0 12 0

E3 ubiquitin-protein ligase CHIP STUB1 0 12 0

Isoform 2 of Coatomer subunit alpha COPA 0 11 0

RuvB-like 2 RUVBL2 0 11 0

Isoform 4 of WD repeat-containing protein 62 WDR62 0 11 0 Isoform 2 of Rho guanine nucleotide exchange ARHGEF2 factor 2 0 11 0

Forkhead box protein K1 FOXK1 0 10 0 Isoform 4 of Ankyrin repeat and KH domain- ANKHD1 containing protein 1 0 10 0 Isoform 2 of 116 kDa U5 small nuclear EFTUD2 ribonucleoprotein component 0 10 0

Bifunctional glutamate/proline--tRNA ligase EPRS 0 9 0

X-ray repair cross-complementing protein 5 XRCC5 0 9 0

Holliday junction recognition protein HJURP 0 9 0

103 104

Aspartate--tRNA ligase, cytoplasmic DARS 0 8 0

Zinc finger protein 318 ZNF318 0 8 0 Isoform 2 of Enhancer of mRNA-decapping EDC4 protein 4 0 8 0

Isoform 2 of Protein TANC1 TANC1 0 8 0

DnaJ homolog subfamily B member 6 DNAJB6 0 8 0

SNW domain-containing protein 1 SNW1 0 8 0

Pre-mRNA-processing factor 19 PRPF19 0 8 0

DnaJ homolog subfamily B member 4 DNAJB4 0 7 0

Isoform 2 of Histone deacetylase 2 HDAC2 0 7 0

Isoform 3 of Hsp70-binding protein 1 HSPBP1 0 7 0

DnaJ homolog subfamily B member 1 DNAJB1 0 7 0

Histone deacetylase 1 HDAC1 0 7 0 Isoform 2 of Ankyrin repeat domain-containing ANKRD17 protein 17 0 7 0 Isoform Delta 10 of Calcium/calmodulin- CAMK2D dependent protein kinase type II subunit delta 0 6 0

DnaJ homolog subfamily A member 2 DNAJA2 0 6 0

Isoform 2 of E3 ubiquitin-protein ligase RNF213 RNF213 0 6 0

Isoform 2 of CCR4-N CNOT1 0 6 0

Collagen alpha-1(VIII) chain COL8A1 0 6 0

Heat shock 70 kDa protein 4 HSPA4 0 6 0

Coatomer subunit beta COPB1 0 6 0 Isoform 2 of Calcium-binding mitochondrial SLC25A13 carrier protein Aralar2 0 6 0

Isoform 2 of Protein PRRC2C PRRC2C 0 6 0 Isoform 10 of Calcium/calmodulin-dependent CAMK2G protein kinase type II subunit gamma 0 6 0

Glutathione S-transferase Mu 3 GSTM3 0 5 0

Cell division cycle 5-like protein CDC5L 0 5 0

Isoform 2 of Vigilin HDLBP 0 5 0

Isoform 2 of Stromal interaction molecule 2 STIM2 0 5 0 Isoform 3 of Serine/threonine-protein phosphatase ANKRD28 6 regulatory ankyrin repeat subunit A 0 5 0

Isoform 2 of Pleiotropic regulator 1 PLRG1 0 5 0

Isoform 2 of Protein SEC13 homolog SEC13 0 5 0

Kynureninase KYNU 0 5 0

Isoform 2 of MIC IMMT 0 5 0 Transforming acidic coiled-coil-containing protein TACC3 3 0 5 0

Isoform SV of 14-3-3 protein epsilon YWHAE 0 5 0 Isoform Heart of ATP synthase subunit gamma, ATP5C1 mitochondrial 0 4 0 Isoform 2 of KH domain-containing, RNA- KHDRBS1 binding, signal transduction-associated protein 1 0 4 0

Isoform 2 of Stress-induced-phosphoprotein 1 STIP1 0 4 0

Emerin EMD 0 4 0

Endoplasmin HSP90B1 0 4 0

Ras-related protein Rab-13 RAB13 0 4 0

Acyl-protein thioesterase 2 LYPLA2 0 4 0

Isoform 10 of Glucocorticoid receptor NR3C1 0 4 0

105

Isoform 2 of Thrombospondin-1 THBS1 0 4 0

CAD protein CAD 0 4 0

Protein transport protein Sec23A SEC23A 0 4 0 Isoform 2 of Probable E3 ubiquitin-protein ligase HERC4 HERC4 0 4 0

Isoform 3 of Sickle tail protein homolog KIAA1217 0 4 0 Isoform 3 of ADP-ribosylation factor GTPase- ARFGAP1 activating protein 1 0 4 0

Isoform 2 of Polyhomeotic-like protein 3 PHC3 0 4 0

Isoform 2 of Drebrin-like protein DBNL 0 4 0 Serine/threonine-protein phosphatase 6 regulatory PPP6R1 subunit 1 0 4 0

Glutaredoxin-1 GLRX 0 4 0

Melanoma-associated antigen 1 MAGEA1 0 4 0

Protein transport protein Sec23B SEC23B 0 4 0

14-3-3 protein gamma YWHAG 0 4 0

Isoform 2 of A-kinase anchor protein 13 AKAP13 0 3 0

Isoform 2 of Myoferlin MYOF 0 3 0

Cell division cycle protein 20 homolog CDC20 0 3 0

60S ribosomal protein L27 RPL27 0 3 0

Rho-related GTP-binding protein RhoC RHOC 0 3 0 Isoform 2 of Pyruvate dehydrogenase E1 component subunit alpha, somatic form, PDHA1 mitochondrial 0 3 0

Isoform 2 of Elongation factor 1-delta EEF1D 0 3 0

Protein disulfide-isomerase P4HB 0 3 0 Isoform 2 of Aminoacyl tRNA synthase complex- AIMP1 interacting multifunctional protein 1 0 3 0

OSMR 0 3 0

Isoform 2 of Nuclear receptor corepressor 1 NCOR1 0 3 0

HAUS augmin-like complex subunit 5 HAUS5 0 3 0

Protein disulfide-isomerase A3 PDIA3 0 3 0

Trifunctional enzyme subunit alpha, mitochondrial HADHA 0 3 0 Isoform 2 of Serine/threonine-protein phosphatase PPP2R2A 2A 55 kDa regulatory subunit B alpha isoform 0 3 0 Isoform 10 of Aspartyl/asparaginyl beta- ASPH hydroxylase 0 3 0 Isoform 2 of KN motif and ankyrin repeat domain- KANK2 containing protein 2 0 3 0

Isoform 4 of F-box/LRR-repeat protein 18 FBXL18 0 3 0

Isoform 2 of BRCA1-A complex subunit RAP80 UIMC1 0 3 0

Heat shock protein beta-8 HSPB8 0 3 0

2'-5'-oligoadenylate synthase 3 OAS3 0 3 0

Isoform 2 of WD repeat-containing protein 1 WDR1 0 3 0

40S ribosomal protein S23 RPS23 0 3 0

Aldo-keto reductase family 1 member C1 AKR1C1 0 3 0

Mitochondrial glutamate carrier 1 SLC25A22 0 3 0

Cytoplasmic dynein 1 heavy chain 1 DYNC1H1 0 3 0

Glutathione S-transferase P GSTP1 0 3 0

Protein transport protein Sec24C SEC24C 0 3 0

106

Isoform 2 of Rap guanine nucleotide exchange RAPGEF6 factor 6 0 3 0

Cathepsin B CTSB 0 2 0

Histone H1.5 HIST1H1B 0 2 0

Destrin DSTN 0 2 0

Protein PRR14L PRR14L 0 2 0 Isoform 2 of Stomatin-like protein 2, STOML2 mitochondrial 0 2 0 Insulin-like growth factor 2 mRNA-binding IGF2BP3 protein 3 0 2 0 Isoform 2 of Protein arginine N-methyltransferase PRMT5 5 0 2 0

Eukaryotic translation initiation factor 2 subunit 3 EIF2S3 0 2 0

Isoform 2 of AP-2 complex subunit beta AP2B1 0 2 0

Isoform 2 of E3 ubiquitin-protein ligase TRIM22 TRIM22 0 2 0 Isoform 2 of SRA stem-loop-interacting RNA- SLIRP binding protein, mitochondrial 0 2 0

Very-long-chain enoyl-CoA reductase TECR 0 2 0

Isoform 2 of Splicing factor U2AF 65 kDa subunit U2AF2 0 2 0

Cyclin-dependent kinase 6 CDK6 0 2 0

Peptidyl-prolyl cis-trans isomerase-like 1 PPIL1 0 2 0 Isoform 2 of Thioredoxin domain-containing TXNDC5 protein 5 0 2 0 Isoform 2 of Vacuolar protein sorting-associated VPS13C protein 13C 0 2 0 Isoform 2 of Procollagen-lysine,2-oxoglutarate 5- PLOD2 dioxygenase 2 0 2 0

Isoform 2 of Histone deacetylase 3 HDAC3 0 2 0 Isoform 2 of Lysine-specific histone demethylase KDM1A 1A 0 2 0 Isoform 2 of Serine/threonine-protein phosphatase PPP6R2 6 regulatory subunit 2 0 2 0

Actin-like protein 6A ACTL6A 0 2 0

Ferritin light chain FTL 0 2 0

Eukaryotic translation initiation factor 2 subunit 2 EIF2S2 0 2 0

Isoform 10 of Calpastatin CAST 0 2 0

Isoform PML-11 of Protein PML PML 0 2 0

Isoleucine--tRNA ligase, cytoplasmic IARS 0 2 0

Isoform 2 of Large proline-rich protein BAG6 BAG6 0 2 0

Ubiquitin-conjugating enzyme E2 A UBE2A 0 2 0

Isoform 2 of Host cell factor 1 HCFC1 0 2 0

Isoform B2 of Smoothelin SMTN 0 2 0 Isoform 2 of ATP synthase subunit f, ATP5J2 mitochondrial 0 2 0 Serine/threonine-protein phosphatase 2A catalytic PPP2CB subunit beta isoform 0 2 0

40S ribosomal protein S15 RPS15 0 2 0

Isoform 2 of 60S ribosomal protein L31 RPL31 0 2 0

S-phase kinase-associated protein 1 SKP1 0 2 0

Forkhead box protein F1 FOXF1 0 2 0 Aminoacyl tRNA synthase complex-interacting AIMP2 multifunctional protein 2 0 2 0

Isoform 2 of Reticulocalbin-2 RCN2 0 2 0

Isoform 2 of Protein disulfide-isomerase A6 PDIA6 0 2 0

107

Helicase SKI2W SKIV2L 0 2 0

Isoform 2 of Glutamine and serine-rich protein 1 QSER1 0 2 0

Rho GTPase-activating protein 31 ARHGAP31 0 2 0 Isoform 2 of Mitochondrial import inner TIMM50 membrane translocase subunit TIM50 0 2 0

Isoform 2 of Centrosomal protein of 170 kDa CEP170 0 2 0

Zinc finger FYVE domain-containing protein 16 ZFYVE16 0 2 0 Isoform 2 of Pleckstrin homology-like domain PHLDB1 family B member 1 0 2 0

Hermansky-Pudlak syndrome 6 protein HPS6 0 2 0

Spindle and kinetochore-associated protein 3 SKA3 0 2 0 Isoform 3 of Ubiquitin carboxyl-terminal USP7 hydrolase 7 0 2 0

U5 small nuclear ribonucleoprotein 40 kDa protein SNRNP40 0 2 0

Isoform 2 of Lysophospholipid acyltransferase 7 MBOAT7 0 2 0

Isoform 3 of Endophilin-A2 SH3GL1 0 2 0

Isoform 2 of A-kinase anchor protein 2 AKAP2 0 2 0

Isoform 3 of Centrosomal protein P POC5 0 2 0

AP-5 complex subunit zeta-1 AP5Z1 0 2 0

DARA 0 0 471

Actin, alpha cardiac muscle 1 ACTC1 0 0 41

Isoform 6 of Myosin-14 MYH14 0 0 30

Tubulin alpha-1A chain TUBA1A 0 0 27

Tropomyosin alpha-3 chain TPM3 0 0 22

Isoform 5 of Tropomyosin alpha-3 chain TPM3 0 0 21

Myosin phosphatase Rho-interacting protein MPRIP 0 0 19

Nuclear pore complex protein Nup155 NUP155 0 0 15

Isoform 2 of Ankycorbin RAI14 0 0 14

DNA-dependent protein kinase catalytic subunit PRKDC 0 0 12

Myosin light chain 6B MYL6B 0 0 12 Isoform MLC3 of Myosin light chain 1/3, skeletal MYL1 muscle isoform 0 0 11

Keratin, type I cytoskeletal 17 KRT17 0 0 11

Isoform 2 of Collagen alpha-3(VI) chain COL6A3 0 0 7

Isoform 3 of Treacle protein TCOF1 0 0 7 Isoform 10 of LIM and calponin homology LIMCH1 domains-containing protein 1 0 0 6

Proteasome activator complex subunit 1 PSME1 0 0 6 Isoform B of Ras GTPase-activating protein- G3BP2 binding protein 2 0 0 6

Isoform 2 of Caprin-1 CAPRIN1 0 0 6 Isoform 2 of Proteasome activator complex PSME3 subunit 3 0 0 6 Nuclear fragile X mental retardation-interacting NUFIP2 protein 2 0 0 6

T-complex protein 1 subunit zeta CCT6A 0 0 5

Ras-related protein Rab-7a RAB7A 0 0 5

Actin-binding protein anillin ANLN 0 0 5 Isoform 2 of Ribose-phosphate pyrophosphokinase PRPS2 2 0 0 5

Beta-1,3-galactosyltransferase 6 B3GALT6 0 0 5

108

Ribose-phosphate pyrophosphokinase 1 PRPS1 0 0 5 Isoform 2 of Guanine nucleotide-binding protein GNB1 G(I)/G(S)/G(T) subunit beta-1 0 0 4

Proteasome activator complex subunit 2 PSME2 0 0 4

Isoform 2 of Ubiquitin-conjugating enzyme E2 D3 UBE2D3 0 0 4

Tropomodulin-3 TMOD3 0 0 4

Ras-related protein Rab-8A RAB8A 0 0 4

Barrier-to-autointegration factor BANF1 0 0 4 Leucine-rich PPR motif-containing protein, LRPPRC mitochondrial 0 0 4

Crk-like protein CRKL 0 0 4

Brain acid soluble protein 1 BASP1 0 0 4

Uncharacterized protein C19orf43 TRIR 0 0 4

Isoform 2 of Cysteine--tRNA ligase, cytoplasmic CARS 0 0 3 Isoform 2 of Cullin-associated NEDD8-dissociated CAND1 protein 1 0 0 3

Inosine-5'-monophosphate dehydrogenase 2 IMPDH2 0 0 3

Nicotinamide N-methyltransferase NNMT 0 0 3 Isoform B of Histone acetyltransferase type B HAT1 catalytic subunit 0 0 3

Long-chain-fatty-acid--CoA ligase 3 ACSL3 0 0 3

Isoform 3 of 60S ribosomal protein L17 RPL17 0 0 3

S-adenosylmethionine synthase isoform type-2 MAT2A 0 0 3

DnaJ homolog subfamily A member 1 DNAJA1 0 0 3

Isoform 2 of PDZ and LIM domain protein 4 PDLIM4 0 0 3

Vesicular integral-membrane protein VIP36 LMAN2 0 0 3

TAR DNA-binding protein 43 TARDBP 0 0 3 Isoform 2 of Leucine-rich repeat flightless- LRRFIP1 interacting protein 1 0 0 3

Isoform 2 of Kinesin-like protein KIF2C KIF2C 0 0 3 Isoform 2 of BUB3-interacting and GLEBS motif- ZNF207 containing protein ZNF207 0 0 3

Heat shock protein beta-1 HSPB1 0 0 3

Isoform 2 of Integrin beta-1 ITGB1 0 0 3

C-1-tetrahydrofolate synthase, cytoplasmic MTHFD1 0 0 3

Isoform 2 of Multifunctional protein ADE2 PAICS 0 0 3

Valine--tRNA ligase VARS 0 0 3

Importin subunit alpha-1 KPNA2 0 0 3

PRKC apoptosis WT1 regulator protein PAWR 0 0 3 Isoform 2 of Chromosome-associated kinesin KIF4A KIF4A 0 0 3

Beta-2-microglobulin B2M 0 0 2

40S ribosomal protein S6 RPS6 0 0 2

Ras-related protein Rap-1b-like protein 0 0 2 Membrane-associated progesterone receptor PGRMC1 component 1 0 0 2

Actin-related protein 2/3 complex subunit 3 ARPC3 0 0 2

Isoform 2 of Surfeit locus protein 4 SURF4 0 0 2 Isoform sGi2 of Guanine nucleotide-binding GNAI2 protein G(i) subunit alpha-2 0 0 2

X-ray repair cross-complementing protein 6 XRCC6 0 0 2

109

Adenosylhomocysteinase AHCY 0 0 2

Isoform 2 of Ras-related protein Rab-5C RAB5C 0 0 2 Isoform 2 of Nucleosome assembly protein 1-like NAP1L1 1 0 0 2

Isoform 2 of 26S protease regulatory subunit 8 PSMC5 0 0 2

60S ribosomal protein L19 RPL19 0 0 2

Transcription intermediary factor 1-beta TRIM28 0 0 2 Isoform 2 of Spectrin alpha chain, non- SPTAN1 erythrocytic 1 0 0 2

Neutral amino acid transporter B(0) SLC1A5 0 0 2

Glutaredoxin-related protein 5, mitochondrial GLRX5 0 0 2

Isoform 1 of Unconventional myosin-VI MYO6 0 0 2

Histone H3.1 HIST1H3A 0 0 2 Basic leucine zipper and W2 domain-containing BZW2 protein 2 0 0 2

Exportin-1 XPO1 0 0 2

Isoform 3BC of Catenin delta-1 CTNND1 0 0 2 Isoform 3 of Sodium/potassium-transporting ATP1A1 ATPase subunit alpha-1 0 0 2

Splicing factor, proline- and glutamine-rich SFPQ 0 0 2

Isoform 3 of Exportin-2 CSE1L 0 0 2 Isoform 2 of Ubiquitin carboxyl-terminal USP10 hydrolase 10 0 0 2

Isoform 2 of LIM domain only protein 7 LMO7 0 0 2

Isoform 3 of Palladin PALLD 0 0 2 Isoform 5 of Myeloid differentiation primary MYD88 response protein MyD88 0 0 2 Isoform 2 of Protein arginine N-methyltransferase PRMT1 1 0 0 2

Protein syndesmos NUDT16L1 0 0 2

Ras-related protein Ral-A RALA 0 0 2

60S ribosomal protein L26 RPL26 0 0 2

40S ribosomal protein S26 RPS26 0 0 2

Heat shock cognate 71 kDa protein HSPA8 0 54 3 Yes Tubulin beta-3 chain TUBB3 0 38 30

Tensin-3 TNS3 0 35 3 Yes Tubulin beta-6 chain TUBB6 0 34 20

BAG family molecular chaperone regulator 2 BAG2 0 30 2 Yes Sequestosome-1 SQSTM1 0 26 14

Isoform 2 of Dedicator of cytokinesis protein 7 DOCK7 0 23 3 Yes Isoform 6 of Microtubule-associated protein 4 MAP4 0 20 4 Yes ADP/ATP translocase 2 SLC25A5 0 19 13

Elongation factor Tu, mitochondrial TUFM 0 19 10

Leucine zipper protein 1 LUZP1 0 16 8

Tubulin alpha-1C chain TUBA1C 0 15 26 HLA class I histocompatibility antigen, A-3 alpha HLA-A chain 0 15 11

Isoform 2 of Polyadenylate-binding protein 1 PABPC1 0 9 13 Dolichyl-diphosphooligosaccharide--protein RPN1 glycosyltransferase subunit 1 0 9 13

Voltage-dependent anion-selective channel protein VDAC1 0 9 10

110

60S ribosomal protein L4 RPL4 0 9 8 Isoform B of Phosphate carrier protein, SLC25A3 mitochondrial 0 9 5 Isoform LCRMP-4 of Dihydropyrimidinase- DPYSL3 related protein 3 0 8 15

Glycine--tRNA ligase GARS 0 8 9

Ubiquitin carboxyl-terminal hydrolase isozyme L1 UCHL1 0 8 9

Nucleolin NCL 0 8 9

Ubiquitin/ISG15-conjugating enzyme E2 L6 UBE2L6 0 8 7

Lamina-associated polypeptide 2, isoform alpha TMPO 0 8 7 Serine/threonine-protein phosphatase 2A 65 kDa PPP2R1A regulatory subunit A alpha isoform 0 8 6 Pre-mRNA-splicing factor ATP-dependent RNA DHX15 helicase DHX15 0 8 2 Yes Arginine--tRNA ligase, cytoplasmic RARS 0 8 2 Yes Isoform 10 of Tropomyosin alpha-1 chain TPM1 0 7 28

Isoform 2 of T-complex protein 1 subunit gamma CCT3 0 7 8

40S ribosomal protein S3a RPS3A 0 7 7

60S ribosomal protein L7a RPL7A 0 7 7

40S ribosomal protein S17 RPS17 0 7 6 Isoform 2 of DNA replication licensing factor MCM3 MCM3 0 7 5

Nuclease-sensitive element-binding protein 1 YBX1 0 7 5 Isoform Short of Heterogeneous nuclear HNRNPU ribonucleoprotein U 0 7 5 Isoform 2 of Glutamine--fructose-6-phosphate GFPT1 aminotransferase [isomerizing] 1 0 7 4

14-3-3 protein zeta/delta YWHAZ 0 7 4

Isoform 3 of ATP-dependent RNA helicase DDX1 DDX1 0 7 3 Yes Isoform 3 of 26S proteasome non-ATPase PSMD2 regulatory subunit 2 0 7 3 Yes Transferrin receptor protein 1 TFRC 0 6 14

60S ribosomal protein L3 RPL3 0 6 8 Lamina-associated polypeptide 2, isoforms TMPO beta/gamma 0 6 8

Dihydropyrimidinase-related protein 2 DPYSL2 0 6 7

T-complex protein 1 subunit alpha TCP1 0 6 6

60S ribosomal protein L23a RPL23A 0 6 3

Four and a half LIM domains protein 2 FHL2 0 6 3

Isoform 2 of Thioredoxin reductase 1, cytoplasmic TXNRD1 0 6 3

Ubiquitin-like protein ISG15 ISG15 0 6 2 Yes Isoform 2 of Splicing factor 1 SF1 0 6 2 Yes Vesicle-associated membrane protein-associated VAPA protein A 0 6 2 Yes Isoform 2 of Unconventional myosin-Ic MYO1C 0 5 14

Isoform 2 of Cellular nucleic acid-binding protein CNBP 0 5 7

F-actin-capping protein subunit alpha-2 CAPZA2 0 5 7 Isoform A1-A of Heterogeneous nuclear HNRNPA1 ribonucleoprotein A1 0 5 6

Isoform 2 of Y-box-binding protein 3 YBX3 0 5 5

T-complex protein 1 subunit theta CCT8 0 5 4

111

60S ribosomal protein L23 RPL23 0 5 4

Isoform 2 of Mitotic checkpoint protein BUB3 BUB3 0 5 3

Isoform 2 of Histone-binding protein RBBP7 RBBP7 0 5 3

PDZ and LIM domain protein 7 PDLIM7 0 5 3

60S acidic ribosomal protein P2 RPLP2 0 5 2 Yes Unconventional myosin-Ib MYO1B 0 4 26

Aspartyl aminopeptidase DNPEP 0 4 17

Isoform 2 of Reticulon-4 RTN4 0 4 7

60S ribosomal protein L15 RPL15 0 4 5

Isoform ADelta10 of Prelamin-A/C LMNA 0 4 5

40S ribosomal protein S12 RPS12 0 4 5

60S ribosomal protein L10 RPL10 0 4 5

40S ribosomal protein S9 RPS9 0 4 4 Isoform 2 of Dolichyl-diphosphooligosaccharide-- RPN2 protein glycosyltransferase subunit 2 0 4 4 Isoform 1 of Voltage-dependent anion-selective VDAC2 channel protein 2 0 4 4

60S ribosomal protein L14 RPL14 0 4 4

40S ribosomal protein S7 RPS7 0 4 3

Protein deglycase DJ-1 PARK7 0 4 3

DNA replication licensing factor MCM5 MCM5 0 4 3

Isoform 2 of Caldesmon CALD1 0 4 2

Histone H1.4 HIST1H1E 0 4 2 Serine/threonine-protein phosphatase PP1-beta PPP1CB catalytic subunit 0 4 2 Complement component 1 Q subcomponent- C1QBP binding protein, mitochondrial 0 4 2

Dynein light chain 1, cytoplasmic DYNLL1 0 4 2

Isoform 2 of Ataxin-2-like protein ATXN2L 0 3 7

Ras GTPase-activating protein-binding protein 1 G3BP1 0 3 6

40S ribosomal protein S8 RPS8 0 3 5

60S ribosomal protein L13 RPL13 0 3 5

Peroxiredoxin-6 PRDX6 0 3 5

Isoform 2 of Bifunctional coenzyme A synthase COASY 0 3 5

Prohibitin-2 PHB2 0 3 5

Phenylalanine--tRNA ligase beta subunit FARSB 0 3 5

40S ribosomal protein SA RPSA 0 3 4

Isoform 2 of 40S ribosomal protein S24 RPS24 0 3 4 tRNA-splicing ligase RtcB homolog RTCB 0 3 4

ATP-dependent 6-phosphofructokinase, liver type PFKL 0 3 4

60S acidic ribosomal protein P0 RPLP0 0 3 3

Translocon-associated protein subunit delta SSR4 0 3 3

Isoform 2 of ATP-citrate synthase ACLY 0 3 3

60S ribosomal protein L30 RPL30 0 3 3 Isoform 2 of Phenylalanine--tRNA ligase alpha FARSA subunit 0 3 2

60S ribosomal protein L27a RPL27A 0 3 2

60S ribosomal protein L10a RPL10A 0 3 2

112

Solute carrier family 2, facilitated glucose SLC2A1 transporter member 1 0 3 2

Malectin MLEC 0 3 2 Isoform 2 of Asparagine synthetase [glutamine- ASNS hydrolyzing] 0 3 2 Protein transport protein Sec61 subunit alpha SEC61A1 isoform 1 0 3 2

60S ribosomal protein L24 RPL24 0 3 2 Isoform B of Eukaryotic translation initiation EIF4G1 factor 4 gamma 1 0 3 2

Leucine-rich repeat-containing protein 47 LRRC47 0 3 2

FAS-associated factor 2 FAF2 0 3 2

Sideroflexin-1 SFXN1 0 3 2

Leukocyte elastase inhibitor SERPINB1 0 3 2

60S ribosomal protein L8 RPL8 0 3 2

CPOX 0 2 6

Tumor susceptibility gene 101 protein TSG101 0 2 5

Probable ATP-dependent RNA helicase DDX6 DDX6 0 2 4

Vesicle-trafficking protein SEC22b SEC22B 0 2 4

DNA replication licensing factor MCM7 MCM7 0 2 4

Isoform 2 of Protein flightless-1 homolog FLII 0 2 4

E3 ubiquitin/ISG15 ligase TRIM25 TRIM25 0 2 3

26S protease regulatory subunit 6A PSMC3 0 2 3

60S ribosomal protein L35 RPL35 0 2 3

Serpin B6 SERPINB6 0 2 3 Isoform Del-701 of Signal transducer and activator STAT3 of transcription 3 0 2 3

Hypoxanthine-guanine phosphoribosyltransferase HPRT1 0 2 3

Ubiquitin-conjugating enzyme E2 N UBE2N 0 2 3

Isoform 3 of Cytoskeleton-associated protein 5 CKAP5 0 2 3 Isoform 2 of ATPase family AAA domain- ATAD3A containing protein 3A 0 2 3 Isoform Beta of Signal transducer and activator of STAT1 transcription 1-alpha/beta 0 2 3

Isoform 2 of 26S protease regulatory subunit 4 PSMC1 0 2 3

Isoform 2 of 26S protease regulatory subunit 7 PSMC2 0 2 2

Alpha-actinin-4 ACTN4 0 2 2

60S acidic ribosomal protein P1 RPLP1 0 2 2

Elongation factor 1-beta EEF1B2 0 2 2

Protein transport protein Sec61 subunit beta SEC61B 0 2 2 Isoform 2 of Mannosyl-oligosaccharide MOGS glucosidase 0 2 2

RuvB-like 1 RUVBL1 0 2 2

Coatomer subunit gamma-1 COPG1 0 2 2

60S ribosomal protein L13a RPL13A 0 2 2

LanC-like protein 1 LANCL1 0 2 2

113

Appendix 2 BAG5 Interactome: SH-SY5Y

iTRAQ Abundance Relative to BAG5(3)

BAG5 BAG5 BAG5 DARA DARA DARA Description Gene (1) (2) (3) (1) (2) (3) GFP(1) GFP(2)

Isoform 2 of BAG family molecular chaperone regulator 5 BAG5 1.349 0.916 1 0.659 0.741 0.806 0.019 0.01 BAG5DARA mutant DARA 4.14 2.469 1 76.56 85.407 100 0.225 0.01 Heat shock 70 kDa protein 1B HSPA1B 1.227 0.911 1 0.125 0.081 0.131 0.021 0.01 Heat shock cognate 71 kDa protein HSPA8 1.521 0.857 1 0.172 0.077 0.117 0.022 0.01 Tubulin beta chain TUBB 1.103 0.987 1 0.213 0.245 0.275 0.026 0.01 Tubulin beta-4B chain TUBB4B 1.141 0.965 1 0.555 0.221 0.297 0.022 0.01 Heat shock-related 70 kDa protein 2 HSPA2 1.25 1.015 1 0.414 0.151 0.412 0.1 0.133 Tubulin beta-2B chain TUBB2B 1.216 1.147 1 0.767 0.01 0.01 0.01 0.01 Isoform 2 of Tubulin alpha-1B chain TUBA1B 1.166 0.982 1 0.165 0.148 0.181 0.027 0.01 Tubulin alpha-3C/D chain TUBA3D 1.037 0.973 1 0.153 0.211 0.214 0.025 0.01 Tubulin alpha-4A chain TUBA4A 1.084 0.897 1 0.197 0.127 0.17 0.019 0.01 E3 ubiquitin-protein ligase CHIP STUB1 1.106 0.816 1 0.328 0.059 0.137 0.04 0.024 Heat shock 70 kDa protein 6 HSPA6 1.436 0.9 1 0.55 0.072 0.193 0.01 0.01 Non-P NONO 1.27 0.892 1 1.306 0.843 0.732 0.034 0.018 X-ray repair cross-complementing protein 6 XRCC6 1.147 0.981 1 0.885 0.712 0.644 0.022 0.011 40S ribosomal protein S17 RPS17 1.183 1.013 1 1.832 1.71 1.058 0.084 0.068 Poly [ADP-ribose] polymerase 1 PARP1 1.288 0.885 1 1.513 0.38 0.311 0.023 0.01 Nucleolin NCL 1.371 1.115 1 1.345 0.615 0.595 0.044 0.022 Myeloid leukemia factor 2 MLF2 0.993 0.862 1 0.74 0.035 0.349 0.01 0.01 Uncharacterized protein TUBB3 1.97 1.092 1 6.648 1.556 2.743 0.01 0.01 40S ribosomal protein S15 RPS15 1.021 0.956 1 1.472 1.32 0.893 0.082 0.044 78 kDa glucose-regulated protein HSPA5 1.254 1.099 1 0.956 0.218 0.316 0.037 0.045 DnaJ homolog subfamily A member 1 DNAJA1 1.145 0.911 1 0.726 0.583 0.443 0.031 0.01 T-complex protein 1 subunit beta CCT2 1.121 0.954 1 3.123 0.625 0.791 0.01 0.01 U2 small nuclear ribonucleoprotein A' SNRPA1 1.11 0.863 1 2.405 1.104 1.006 0.069 0.033 Elongation factor 1-alpha 1 EEF1A1 1.173 0.926 1 0.703 0.73 0.62 0.02 0.011 X-ray repair cross-complementing protein 5 XRCC5 1.028 0.89 1 3.171 0.589 0.549 0.073 0.053 60S ribosomal protein L3 RPL3 1.267 1.026 1 0.756 0.479 0.338 0.027 0.01 Putative heat shock 70 kDa protein 7 HSPA7 3.058 1.136 1 7.978 0.01 0.01 0.01 0.01 TH ALYREF 1.33 1.017 1 2.551 0.826 0.886 0.064 0.061 60S ribosomal protein L23 RPL23 1.007 0.913 1 1.422 1.276 1.022 0.064 0.073 40S ribosomal protein S10 RPS10 1.103 0.903 1 1.504 1.199 0.793 0.089 0.046 BAG family molecular chaperone regulator 2 BAG2 1.213 0.995 1 0.769 0.076 0.134 0.041 0.059 Serum albumin [Bos taurus] ALB 0.159 0.11 1 0.508 0.122 0.319 0.034 0.024

114

Isoform 2 of Stress-induced- phosphoprotein 1 STIP1 1.108 0.814 1 0.938 0.086 0.178 0.038 0.01 Nucleophosmin NPM1 1.17 1.101 1 1.369 0.791 0.797 0.047 0.01 60S ribosomal protein L4 RPL4 1.157 0.921 1 0.897 0.703 0.459 0.026 0.013 Tubulin beta-8 chain TUBB8 1.566 0.276 1 3.426 2.048 0.316 0.01 0.01 60S ribosomal protein L7a (Fragment) RPL7A 1.15 0.996 1 1.225 1.299 0.927 0.052 0.041 Isoform 2 of Polyadenylate-binding protein 4 PABPC4 1.529 1.032 1 3.49 0.519 0.745 0.2 0.092 40S ribosomal protein S13 RPS13 1 1.026 1 1.818 1.336 0.652 0.099 0.032 HSP90AB Heat shock protein HSP 90-beta 1 1.272 0.881 1 1.329 0.158 0.25 0.18 0.139 60S ribosomal protein L23a RPL23A 1.158 0.969 1 1.879 1.448 1.02 0.06 0.049 40S ribosomal protein S3 RPS3 1.228 0.934 1 1.018 0.563 0.321 0.073 0.027 Insulin receptor substrate 4 IRS4 1.599 1.108 1 1.675 0.01 0.696 0.28 0.221 Polypyrimidine tract-binding protein 1 PTBP1 1.028 0.857 1 0.622 0.538 0.355 0.019 0.01 40S ribosomal protein S15a RPS15A 1.239 1.034 1 2.133 0.918 0.607 0.068 0.037 Clathrin heavy chain 1 CLTC 1.612 1.318 1 2.411 0.874 0.764 0.032 0.034 Insulin-like growth factor 2 mRNA- binding protein 1 IGF2BP1 1.242 1.268 1 1.862 1.211 1.856 0.226 0.056 60S ribosomal protein L6 RPL6 1.208 0.916 1 1.128 0.973 0.802 0.056 0.043 40S ribosomal protein S18 RPS18 1.113 0.985 1 2.703 2.329 1.504 0.069 0.052 40S ribosomal protein S8 RPS8 1.266 1.016 1 1.027 0.798 0.648 0.227 0.02 60S ribosomal protein L17 RPL17 1.255 0.884 1 3.022 0.819 0.576 0.01 0.01 40S ribosomal protein S19 RPS19 1.138 0.955 1 2.16 1.599 1.272 0.103 0.067 HIST1H1 Histone H1.2 C 0.989 1.048 1 1.87 1.175 0.944 0.063 0.026 HIST1H1 Histone H1.4 E 1.572 6.325 1 10.567 0.01 0.01 0.01 0.01 Non-P NONO 2.008 0.935 1 1.18 0.01 1.581 0.01 0.01 Tubulin alpha-4A chain (Fragment) TUBA4A 1.668 1.086 1 5.575 0.795 0.956 0.01 0.01 Uncharacterized protein 0.885 1.714 1 2.295 1.801 1.34 0.033 0.035 40S ribosomal protein S7 RPS7 1.108 1.047 1 1.119 1.083 0.741 0.065 0.049 Nuclease-sensitive element-binding protein 1 (Fragment) YBX1 1.259 1.089 1 1.776 1.051 1.263 0.043 0.053 60S ribosomal protein L9 RPL9 1.23 1.028 1 0.823 0.586 0.554 0.035 0.046 40S ribosomal protein S24 RPS24 1.126 1.04 1 1.69 1.026 0.515 0.043 0.01 Isoform 7 of Interleukin enhancer- binding factor 3 ILF3 1.258 0.872 1 1.736 0.461 0.51 0.039 0.044 Ubiquitin-40S ribosomal protein S27a RPS27A 1.075 0.915 1 0.719 0.836 0.885 0.104 0.05 Polyubiquitin-C (Fragment) UBC 2.153 0.705 1 2.38 0.01 1.401 0.477 0.487 Pre-mRNA-processing factor 19 PRPF19 1.249 0.982 1 0.165 0.213 0.152 0.01 0.037 Plasminogen activator inhibitor 1 RNA-binding protein SERBP1 0.927 0.707 1 0.987 0.564 0.72 0.052 0.052 Beta-casein CSN2 1.42 0.794 1 4.549 0.793 1.51 0.166 0.017 40S ribosomal protein S14 RPS14 1.168 1.058 1 1.69 0.631 0.83 0.01 0.01 Heterogeneous nuclear ribonucleoprotein D0 (Fragment) HNRNPD 1.076 0.857 1 0.41 0.322 0.396 0.038 0.026 Y-box-binding protein 2 YBX2 0.677 0.442 1 2.132 0.776 0.737 0.01 0.01 Alpha-S1-casein CSN1S1 1.544 0.769 1 12.379 1.322 2.056 0.355 0.01 60S ribosomal protein L24 RPL24 1.061 0.953 1 1.615 1.276 0.819 0.057 0.03 40S ribosomal protein S3a RPS3A 0.966 0.858 1 1.213 0.693 0.706 0.041 0.013

115

60S ribosomal protein L31 (Fragment) RPL31 0.995 0.948 1 1.245 1.18 0.855 0.03 0.013 60S ribosomal protein L13 RPL13 1.122 1.055 1 1.649 1.374 1.23 0.062 0.031 Cell division cycle 5-like protein CDC5L 1.15 1.216 1 1.653 0.01 1.715 0.01 0.116 60S ribosomal protein L10a RPL10A 1.076 2.011 1 1.454 1.015 0.89 0.047 0.039 Heterogeneous nuclear ribonucleoprotein F HNRNPF 1.33 1.01 1 1.675 0.809 1.058 0.01 0.01 T-complex protein 1 subunit delta CCT4 1.299 1.093 1 1.819 0.695 0.493 0.01 0.01 40S ribosomal protein S19 (Fragment) RPS19 1.269 0.294 1 8.54 1.673 0.621 0.516 0.154 T-complex protein 1 subunit alpha TCP1 1.114 0.95 1 1.559 0.629 0.403 0.01 0.01 Isoform 2 of Histone deacetylase 6 HDAC6 0.87 0.668 1 3 0.826 1.836 0.01 0.01 40S ribosomal protein S2 RPS2 1.172 0.978 1 1.45 0.644 0.451 0.041 0.014 60S ribosomal protein L27 (Fragment) RPL27 1.253 1.098 1 2.106 1.791 1.261 0.082 0.058 40S ribosomal protein S23 RPS23 1.079 0.961 1 1.052 1.056 0.864 0.094 0.056 60S ribosomal protein L8 RPL8 1.308 1.086 1 1.846 1.038 0.898 0.121 0.017 Isoform B of DnaJ homolog subfamily B member 6 DNAJB6 1.401 1.038 1 0.146 0.071 0.125 0.014 0.01 Ras GTPase-activating protein- binding protein 1 G3BP1 1.555 1.243 1 4.406 0.785 0.725 0.098 0.056 Interleukin enhancer-binding factor 2 ILF2 1.525 1.229 1 1.382 0.696 0.434 0.01 0.083 Ribosomal protein L15 (Fragment) RPL15 1.172 0.85 1 1.332 0.892 0.569 0.034 0.037 T-complex protein 1 subunit zeta CCT6A 1.247 0.927 1 1.514 0.355 0.404 0.01 0.09 60S ribosomal protein L7a RPL7A 0.865 0.818 1 1.672 1.08 1.011 0.067 0.05 40S ribosomal protein S25 RPS25 0.896 0.875 1 2.436 1.542 0.816 0.046 0.035 T-complex protein 1 subunit theta CCT8 1.422 1.018 1 1.136 0.76 0.855 0.022 0.01 Eukaryotic translation initiation factor 2 subunit 2 EIF2S2 1.281 0.762 1 6.383 1.324 1.764 0.01 0.01 60S ribosomal protein L7 RPL7 1.01 1.107 1 3.895 1.301 0.907 0.376 0.083 HIST1H1 Histone H1.1 A 2.408 0.376 1 1.948 1.492 5.125 0.523 0.01 HIST1H1 Histone H1t T 0.644 0.741 1 1.204 0.517 0.552 0.01 0.01 Transmembrane protein 263 C12orf23 0.946 1.101 1 1.903 1.633 1.012 0.125 0.079 ANKHD1 - Isoform 6 of Ankyrin repeat and KH EIF4EBP domain-containing protein 1 3 2.086 2.166 1 3.391 0.75 0.963 0.01 0.01 40S ribosomal protein S11 RPS11 1.051 0.944 1 1.261 1.138 0.934 0.054 0.05 60S ribosomal protein L18 (Fragment) RPL18 1.223 1.003 1 1.225 0.953 0.668 0.046 0.082 60S acidic ribosomal protein P2 (Fragment) RPLP2 1.199 0.849 1 1.993 1.278 1.13 0.131 0.065 Heterogeneous nuclear HNRNPH ribonucleoprotein H 1 1.413 2.078 1 0.01 0.01 0.01 0.01 0.01 Isoform 2 of Heterogeneous nuclear ribonucleoprotein K HNRNPK 1.682 1.329 1 0.814 0.918 0.884 0.11 0.01 60S ribosomal protein L22 RPL22 1.256 1.088 1 2.004 1.496 0.739 0.09 0.051 Heterogeneous nuclear ribonucleoprotein U HNRNPU 2.148 1.136 1 7.176 1.061 1.503 0.055 0.183 DNA-dependent protein kinase catalytic subunit PRKDC 1.541 0.967 1 5.694 0.672 0.686 0.239 0.267 Ribonucleoprotein PTB-binding 1 RAVER1 1.108 0.968 1 1.639 0.514 0.477 0.204 0.06 60S ribosomal protein L38 RPL38 1.032 0.844 1 2.037 2.507 2.111 0.124 0.201 Insulin-like growth factor 2 mRNA- binding protein 3 IGF2BP3 0.738 0.668 1 2.205 1.352 0.24 0.01 0.01 60S ribosomal protein L14 RPL14 1.467 0.919 1 1.753 1.036 0.872 0.097 0.01

116

40S ribosomal protein S5 RPS5 0.874 1.283 1 0.748 0.294 0.474 0.25 0.099 Coiled-coil domain-containing protein 124 CCDC124 0.375 0.327 1 0.402 0.359 0.406 0.056 0.077 Elongation factor 2 EEF2 2.087 1.213 1 1.104 0.463 0.502 0.01 0.01 Prefoldin subunit 2 PFDN2 0.883 0.734 1 0.873 0.01 0.388 0.01 0.01 Heterogeneous nuclear HNRNPA ribonucleoprotein A1-like 2 1L2 0.993 1.032 1 0.745 0.705 0.589 0.153 0.111 60S ribosomal protein L36 RPL36 1.26 1.023 1 2.238 0.624 0.719 0.066 0.037 Peroxiredoxin-1 (Fragment) PRDX1 1.736 1.181 1 1.575 1.332 1.305 0.196 0.158 Programmed cell death protein 5 PDCD5 1.182 1.033 1 8.588 2.217 2.424 0.116 0.166 Protein AF1q MLLT11 3.838 0.668 1 4.017 2.367 2.104 0.163 0.193 Isoform 5 of Ubiquitin-associated protein 2-like UBAP2L 1.223 0.868 1 2.504 0.758 0.691 0.655 0.058 60S ribosomal protein L21 RPL21 1.022 0.89 1 1.727 0.879 0.76 0.01 0.025 Heterogeneous nuclear HNRNPA ribonucleoprotein A1 (Fragment) 1 1.681 1.708 1 4.744 7.169 1.493 0.01 0.01 40S ribosomal protein S30 FAU 1.028 1.034 1 1.611 1.382 1.062 0.103 0.096 60S ribosomal protein L10 RPL10 1.222 0.851 1 1.791 0.587 0.585 0.348 0.354 Isoform 2 of 60S ribosomal protein L15 RPL15 3.851 1.82 1 16.748 0.01 0.01 0.01 0.01 Ankyrin repeat domain-containing ANKRD1 protein 17 7 1.273 1.16 1 1.925 2.343 0.361 0.108 0.123 40S ribosomal protein S6 RPS6 1.262 0.99 1 1.892 0.846 0.849 0.039 0.015 Fatty acid synthase FASN 1.462 0.789 1 2.147 0.619 0.79 0.601 0.233 40S ribosomal protein S27 RPS27L 1.271 0.86 1 0.704 0.01 0.164 0.01 0.01 DnaJ homolog subfamily B member 1 DNAJB1 1.377 0.802 1 1.696 0.814 1.307 0.252 0.068 60S ribosomal protein L27a RPL27A 1.205 1.004 1 1.718 1.199 0.842 0.078 0.035 60S ribosomal protein L35 RPL35 1.397 1.243 1 3.113 1.585 1.425 0.03 0.033 Transforming protein RhoA RHOA 1.256 0.748 1 5.744 0.01 2.167 0.01 0.01 Cytoskeleton-associated protein 5 CKAP5 1.332 0.857 1 2.781 0.685 0.675 0.169 0.072 Putative heat shock protein HSP 90- HSP90A alpha A4 A4P 0.149 0.01 1 0.938 0.01 0.913 0.01 0.01 Microtubule-associated protein MAP4 0.968 0.705 1 0.324 0.091 0.137 0.066 0.01 Clathrin heavy chain 2 CLTCL1 0.387 0.623 1 2.961 0.675 0.576 0.01 0.01 Ubiquitin-60S ribosomal protein L40 (Fragment) UBA52 2.372 0.829 1 0.536 0.658 0.01 0.01 0.01 Alpha-2-HS-glycoprotein AHSG 1.827 1.091 1 2.119 1.543 2.053 0.264 0.052 Heterogeneous nuclear HNRNPA ribonucleoprotein A0 0 1.494 1.996 1 6.676 0.01 2.127 0.01 0.01 ATP-dependent RNA helicase DDX3X DDX3X 1.115 0.797 1 2.919 0.873 0.919 0.01 0.01 Isoform 6 of Protein PRRC2C PRR2C 2.439 1.299 1 4.894 0.481 0.355 0.238 0.215 Heterogeneous nuclear HNRNPA ribonucleoproteins A2/B1 2B1 0.368 0.398 1 3.554 0.01 0.01 0.01 0.01 Heterogeneous nuclear ribonucleoprotein Q SYNCRIP 1.141 0.78 1 4.933 0.679 0.975 0.352 0.205 Insulin-like growth factor 2 mRNA- binding protein 2 IGF2BP2 2.059 0.765 1 0.01 0.01 0.01 0.01 0.01 40S ribosomal protein S26 RPS26 1.221 1.003 1 2.401 1.971 1.596 0.393 0.113 40S ribosomal protein S27 RPS27 1.576 1.299 1 0.607 0.273 0.384 0.01 0.01 Nucleosome assembly protein 1-like 1 (Fragment) NAP1L1 1.124 0.264 1 1.88 0.01 0.01 0.01 0.01 Matrin-3 MATR3 1.249 1.148 1 7.784 0.01 1.473 0.118 0.116 Protein dpy-30 homolog DPY30 0.634 0.563 1 0.412 0.342 0.29 0.01 0.01

117

Prefoldin subunit 6 PFDN6 0.718 0.367 1 1.836 0.59 0.241 0.018 0.023 116 kDa U5 small nuclear ribonucleoprotein component EFTUD2 1.161 1.062 1 0.625 0.137 0.186 0.01 0.01 Programmed cell death protein 5 PDCD5 1.636 0.639 1 14.293 0.01 0.01 0.01 0.01 Interleukin enhancer-binding factor 3 (Fragment) ILF3 0.01 0.01 1 4.302 0.01 0.01 0.01 0.01 Isoform 6 of Splicing factor 1 SF1 1.761 0.513 1 2.668 0.01 0.449 0.01 0.01 Myosin-9 MYH9 0.561 0.496 1 4.641 1.622 1.012 0.01 0.01 RuvB-like 1 RUVBL1 1.226 1.019 1 1.275 0.01 0.285 0.065 0.01 Transcription factor BTF3 BTF3 4.971 1.24 1 5.528 3.611 2.818 0.01 0.01 HIST1H2 Histone H2B BN 2.169 7.418 1 2.425 1.293 3.248 0.01 0.01 Hemoglobin subunit alpha HBA2 3.836 2.897 1 3.759 2.839 1.713 0.132 0.199 Histone H2AX H2AFX 0.01 1.064 1 14.555 0.01 0.01 0.01 0.01 60S ribosomal protein L18a RPL18A 1.078 0.96 1 1.667 1.379 1.267 0.043 0.037 Probable ATP-dependent RNA helicase DDX5 DDX5 1.39 0.662 1 6.73 3.312 0.01 0.01 0.01 Serum albumin ALB 1.94 0.954 1 0.01 0.01 2.57 0.01 0.01 Isoform 2 of TAR DNA-binding protein 43 TARDBP 0.01 0.01 1 8.349 0.01 0.01 0.01 0.01 Heterogeneous nuclear HNRNPA ribonucleoprotein A/B B 1.121 0.909 1 0.82 0.322 0.769 0.01 0.01 Heterogeneous nuclear HNRNP ribonucleoprotein M M 1.725 1.1 1 6.787 0.85 1.355 0.337 0.346 Isoform 2 of Reticulocalbin-2 RCN2 1.656 1.901 1 0.01 0.01 0.01 0.01 0.01 60S ribosomal protein L13a RPL13A 1.431 0.983 1 1.312 0.767 0.922 0.078 0.032 60S ribosomal protein L26 RPL26 1.093 0.922 1 1.93 0.983 0.78 0.03 0.074 40S ribosomal protein S16 RPS16 1.008 1.001 1 0.995 1.1 0.568 0.041 0.04 Nascent polypeptide-associated complex subunit alpha-2 NACA2 1.996 0.809 1 4.87 4.802 3.916 0.01 0.622 HSP90A Heat shock protein HSP 90-alpha A2 A2 0.929 1.012 1 0.781 0.01 0.398 0.01 0.01 60S ribosomal protein L29 RPL29 1.021 0.969 1 1.877 1.274 1.127 0.046 0.068 Isoform 2 of Putative eukaryotic translation initiation factor 2 subunit 3-like protein EIF2S3L 2.77 1.887 1 64.717 2.874 3.095 0.01 0.01 Isoform 2 of T-complex protein 1 subunit epsilon CCT5 1.945 1.987 1 2.727 2.273 0.01 0.01 0.01 Isoform 2 of Eukaryotic translation initiation factor 3 subunit B EIF3B 2.22 2.912 1 20.221 0.01 0.01 0.01 0.01 Tetratricopeptide repeat protein 1 TTC1 4.53 2.298 1 9.933 0.01 0.01 0.01 0.01 LOC1022 Uncharacterized protein C11orf98 88414 1.46 1.132 1 4.924 0.01 0.01 0.01 0.01 Paraspeckle component 1 PSPC1 0.696 0.879 1 1.834 0.235 0.827 0.01 0.01 Isoform 2 of 40S ribosomal protein S20 RPS20 0.983 0.848 1 0.944 0.568 0.584 0.057 0.059 Isoform 4 of Myosin-10 MYH10 1.596 1.106 1 7.394 0.243 0.607 0.088 0.01 Dermcidin DCD 1.109 1.134 1 1.391 0.684 3.94 0.332 0.14 60S ribosomal protein L34 RPL34 1.136 0.983 1 1.615 1.117 0.715 0.053 0.042 DnaJ homolog subfamily C member 9 DNAJC9 0.979 1.05 1 0.139 0.01 0.293 0.01 0.01 Probable ATP-dependent RNA helicase DDX17 DDX17 0.636 1.037 1 4.605 0.01 0.252 0.01 0.01 Myosin light polypeptide 6 MYL6 1.584 1.006 1 4.213 1.433 0.751 0.01 0.01 Heterogeneous nuclear HNRNPD ribonucleoprotein D-like L 1.001 0.872 1 1.777 0.678 0.845 0.054 0.01 P POTEF 1.098 1.058 1 2.681 1.356 1.045 0.01 0.01

118

40S ribosomal protein S9 RPS9 1.169 1.166 1 2.863 0.708 1.078 0.01 0.01 Eukaryotic translation initiation factor 3 subunit F EIF3F 1.366 0.97 1 1.4 0.01 0.01 0.01 0.01 SNW domain-containing protein 1 SNW1 0.987 0.787 1 1.891 0.711 0.42 0.684 0.01 60S ribosomal protein L11 RPL11 1.355 1.262 1 1.669 0.628 0.634 0.01 0.01 60S ribosomal protein L32 RPL32 1.016 0.986 1 2.872 1.358 0.734 0.068 0.01 60S ribosomal protein L37a RPL37A 0.972 1 1 3.171 1.797 1.174 0.058 0.038 ATP-dependent RNA helicase A DHX9 1.427 1.723 1 8.046 1.623 2.535 0.01 0.01 Bifunctional glutamate/proline-- tRNA ligase EPRS 1.287 0.977 1 4.894 0.51 1.81 0.01 0.01 14-3-3 protein zeta/delta (Fragment) YWHAZ 0.383 0.552 1 3.151 1.613 0.386 0.125 0.01 Signal recognition particle 14 kDa protein SRP14 1.188 1.059 1 2.628 0.997 0.626 0.01 0.01 Nucleolar RNA helicase 2 DDX21 0.763 0.605 1 3.985 0.355 0.605 0.22 0.375 Histone deacetylase 6 (Fragment) HDAC6 1.227 0.816 1 2.527 2.01 1.588 0.125 0.15 HIST1H4 Histone H4 A 1.705 0.996 1 4.911 1.07 0.946 0.269 0.339 HIST1H2 Histone H2B type 1-A BA 0.991 1.639 1 2.078 2.232 1.48 0.094 0.055 40S ribosomal protein S12 RPS12 1.25 1.127 1 0.791 0.852 0.741 0.01 0.01 Ubiquitin-associated protein 2 UBAP2 0.667 0.455 1 0.748 0.523 0.483 0.213 0.426 Histone deacetylase complex subunit SAP18 SAP18 0.861 0.01 1 3.34 0.01 0.01 0.01 0.01 Lupus La protein SSB 0.96 0.949 1 2.009 0.892 0.348 0.437 0.393 Spermatid perinuclear RNA-binding protein STRBP 0.622 0.413 1 1.906 0.214 0.71 0.01 0.261 Histone-binding protein RBBP4 RBBP4 0.741 0.927 1 0.01 0.01 0.01 0.01 0.01 Alpha-synuclein SNCA 0.738 1.431 1 2.961 0.595 0.836 0.01 0.01 Double-stranded RNA-binding protein Staufen homolog 1 STAU1 0.665 0.656 1 2.134 0.526 0.638 0.107 0.054 60S ribosomal protein L35a RPL35A 1.196 1.03 1 1.59 1.151 0.599 0.01 0.044 Ribonucleoprotein PTB-binding 1 (Fragment) RAVER1 2.074 0.748 1 16.29 2.728 1.702 0.01 0.01 Eukaryotic translation initiation factor 4 gamma 1 EIF4G1 1.343 1.144 1 3.584 0.803 0.454 0.01 0.01 Small nuclear ribonucleoprotein Sm D1 SNRPD1 0.912 1.095 1 2.131 0.01 0.01 0.01 0.01

119

Appendix 3 Contributions

Xinzhu (Louisa) Wang generated the CRISPR-Cas9 edited SH-SY5Y cells and created the backbone of the replacement plasmid used by me to subsequently insert the GFP-BAG5 transgenes into the AAVS1 safe harbor (outlined in Chapter 2).

Declan Williams conducted the mass spectrometry sample preparation, the iTRAQ mass spectrometry run and the subsequent peptide identification (Chapter 2).

Shirley Zhang helped to generate the p62-HA deletion constructs outlined in Figure 10. She also provided significant technical assistance with the immunohistochemistry experiments, helped to create Figure 13, and contributed to the PCA assay results presented in Chapter 4.

Victoria Agapova helped to generate the p62-HA deletion constructs outlined in Figure 10.

Mitch De Snoo provided technical assistance with cell culture and western blotting.

Figure 1 and Sections 1.2.1, 1.2.2 & 1.2.3 were all derived from the following source:

Erik L. Friesen, Mitch L. De Snoo, Luckshi Rajendran, Lorraine V. Kalia, and Suneil K. Kalia, “Chaperone-Based Therapies for Disease Modification in Parkinson’s Disease,” Parkinson’s Disease, vol. 2017, Article ID 5015307, 11 pages, 2017. doi:10.1155/2017/5015307.

120