Submit to Movement Disorders

The Pennsylvania State University

The Graduate School

College of Medicine

MECHANISTIC STUDIES ON BIF-1 FUNCTION IN

OBESITY AND ATG9 TRAFFICKING

A Dissertation in Pharmacology by Ying Liu

 2016 Ying Liu

Submitted in Partial Fulfillment of the Requirements for the Degree of Doctor of Philosophy

December 2016

The dissertation of Ying Liu was reviewed and approved* by the following:

Hong-Gang Wang Lois High Berstler Professor, Pediatrics and Pharmacology Dissertation Advisor Chair of Committee

Richard Mailman Professor of Pharmacology

Thomas Spratt Associate Professor of Biochemistry and Molecular Biology

Yoshinori Takahashi Assistant Professor of Pharmacology

Yuguang (Roger) Shi Former Professor of Cellular and Molecular physiology

Jong K. Yun Associate Professor and Graduate Program Direct of Pharmacology

*Signatures are on file at the Graduate School

ABSTRACT

Autophagy, the catabolic process whereby intracellular components are degraded by the lysosome, is long recognized for its roles in maintaining cellular energetic balance by breaking down dysfunctional proteins and recycling the amino acids. Adding to this classic view, recent studies have found that autophagy also directly impacts energetics by degrading lipids. This process, referred to as lipophagy, serves to prevent excessive fat accumulation in multiple tissues such as the liver and skeletal muscle, and modulates the hormone production in hypothalamus to affect food intake. The molecular mechanisms of lipophagy, particularly those related to the biogenesis of autophagosome to sequester the substrates, are unclear due to poor understanding of the responsible molecular interactions.

Moreover, in the largest lipid storage organ - adipose tissue – it is unknown whether autophagy plays a role in fat degradation. It has been established that autophagy is essential for early development including embryonic adipogenesis, but it is not known if impairment of this catabolic process in an individual with mature adipose tissue leads to obesity.

Bif-1 is a previously identified positive regulator of autophagy that interacts with

UVRAG, and modulates Atg9 trafficking to promote the formation of autophagosomes.

The current study identified Bif-1 as a novel regulator in lipid catabolism that will prevent the development of obesity and insulin resistance upon aging or dietary challenge. My data show that Bif-1 deficiency promotes the expansion of adipose tissue mass without altering food intake or physical activities. Although Bif-1 is dispensable for adipose tissue development, its deficiency reduces the rate of adipose tissue lipolysis, lowers the iii

degradation of autophagy adaptor and substrate p62, and results in adipocyte hypertrophy upon aging. This function of Bif-1 in lipid turnover is not limited to adipose tissue since lipid droplet clearance is also attenuated by Bif-1 loss in the liver of mice that were starved and re-fed. Interestingly, obesity induced by a high fat-diet or Bif-1 deficiency downregulates the protein levels of Atg9 and a lysosomal protein Lamp1 in the adipose tissue.

Together, my research discovered a new role for Bif-1 in regulating lipid metabolism likely via lipophagy. Many questions remain, for example, does Bif-1-mediated lipid degradation depend on Atg9, and what other potential interactors of the Atg9-Bif-1 complex promote the biogenesis of autophagosomes? In order to address this question, we performed an Atg9 interactome profiling using inducible protein crosslinking coupled with affinity purification and proteomics. This identified VCP/p97 as a novel interactor with Atg9. Moreover, genetic knockdown of VCP in cells appears to disrupt autophagosome maturation.

Although further studies are needed to precisely dissect the function of VCP in the autophagy machinery, I propose a potential role for VCP in the retrieval of autophagy proteins from the autophagosome.

In summary, my research has increased our understanding in the molecular mechanisms of autophagy and lipophagy. I identified Bif-1 as a novel player in lipid homeostasis and obesity, and I hypothesize that targeting lipophagy may provide new avenues for the treatment of obesity and its related metabolic complications. My findings also broaden our knowledge of the Atg9 interaction network through the identification of

VCP as an Atg9 interactor.

Table of Contents

LIST OF FIGURES ...... vii

LIST OF ABBREVIATIONS ...... ix

ACKNOWLEDGEMENTS ...... xiii

Chapter 1. Literature Review ...... 1

1.1. Autophagy ...... 2 1.1.1. Molecular Machinery of Autophagy ...... 2 1.1.2. Regulation of Autophagy by Nutrient and Energy Status ...... 10

1.2. Obesity and Autophagy ...... 13 1.2.1. Links Between Obesity and Autophagy...... 13 1.2.2. Lipophagy Mechanisms ...... 19

1.3. Bif-1 as A Multifunctional Protein ...... 24 1.3.1. Overview of Bif-1 Structure, Expression, and Function ...... 24 1.3.2. Bif-1 and Cancers ...... 28

1.4. Atg9 as An Essential Molecule in Autophagy ...... 29 1.4.1. Overview of Atg9 Biology and Importance for Autophagy ...... 29 1.4.2. Yeast Atg9 Trafficking and Interactions ...... 30 1.4.3. Mamalian Atg9 Trafficking and Interaction Network ...... 31

1.5. VCP as A Novel Regulator of Autophagy ...... 36 1.5.1. VCP Structure and Function Overview ...... 36 1.5.2. Reported Roles of VCP in Autophagy ...... 37 1.5.3. VCP and Human Diseases ...... 38

Chapter 2. Roles of Bif-1 in Obesity ...... 40

2.1. Abstract ...... 41

2.2. Results ...... 42 2.2.1. Ablation of Bif-1 Promotes Aging- and Diet-induced Weight Gain and Fat Accumulation...... 42 2.2.2. Bif-1 KO Mice Develop Insulin Resistance Following Excessive

Adiposity...... 46 2.2.3. Bif-1 Does Not Regulate Food Intake, Physical Activities or Fuel Preference...... 49 2.2.4. Development of Obesity With Bif-1 Deficiency is Independent of Adipogenesis...... 52 2.2.5. Bif-1 Loss Impairs Lipid Turnover in Adipose Tissue and Liver...... 55 2.2.6. Bif-1 Deficiency Reduces Autophagic Flux and Downregulates Autophagic- Lysosomal Proteins...... 59

2.3. Discussion ...... 63

2.4. Material and Methods ...... 66

Chapter 3. Identification of Atg9 Interacting Proteins ...... 74

3.1. Abstract ...... 75

3.2. Results ...... 76 3.2.1. Purification of Atg9-cPTP Complexes Through Tandem Affinity Purification ...... 76 3.2.2. Interaction Candidates Identified by Proteomics ...... 80 3.2.3. Autophagic Function Screen in Atg9 Interacting Top Candidates ...... 82 3.2.4. Validation of VCP and Atg9 Interaction ...... 84

3.3. Discussion ...... 85

3.4. Material and Methods ...... 87

Chapter 4. Summary, Future Directions, and Significance ...... 93

4.1. Role of Bif-1 in Lipid Homeostasis and Obesity (Chapter 2) ...... 94 4.1.1. Summary of Findings ...... 94 4.1.2. Open Questions and Future Directions ...... 95

4.2. Atg9 and the Identification of Its Interaction Partners (Chapter 3) ...... 98 4.2.1. Summary of Findings ...... 98 4.2.2. Open Questions and Future Directions ...... 99

4.3. Overall Significance and Perspectives ...... 101

REFERENCES ...... 105 vi

LIST OF FIGURES

Figure 1.1. Overview of the Autophagy Process and Molecular Machinery. 8

Figure 1.2. Model of Autophagosome – Late Endosome/Lysosome Fusion. 9

Figure 1.3. Model of Lipophagy Mechanism. 23

Figure 1.4. Mammalian Atg9 Trafficking Pathways and Interaction Network. 35

Figure 2.1. Loss of Bif-1 induces obesity. 44

Figure 2.2. Metabolic characteristics of male mice fed with indicated diet. 45

Figure 2.3. Plasma insulin and leptin levels in 10-week-old mice. 47

Figure 2.4. Obesity induced by Bif-1 loss leads to insulin resistance. 48

Figure 2.5. Effects of Bif-1 on food intake, physical activities, and fuel 51

preference.

Figure 2.6. Respiratory exchange ratio (RER) of WT and Bif-1 KO mice. 52

Figure 2.7. Bif-1 deficiency does not affect adipogenesis. 54

Figure 2.8. Bif-1 regulates lipid turn over in adipose tissue and liver. 57

Figure 2.9. Body weight and body fat percentage in 6-week-old mice. 58

Figure 2.10. Bif-1 is important for autophagy in adipose tissue. 61

vii

Figure 2.11. Quantifications for immunoblots in Figure 2.10. 62

Figure 2.12. Model of Bif-1 function in lipid catabolism. 65

Figure 3.1. Atg9-cPTP is Functional for Autophagy Activity. 77

Figure 3.2. Illustration of UV-inducible protein crosslink using photo amino 78

acids.

Figure 3.3. Analysis of fractions collected in tandem affinity purification of 79

Atg9-cPTP complex.

Figure 3.4. Stratification of Atg9-interacting hits. 81

Figure 3.5. VCP depletion disrupts autophagosome maturation and 83

accumulates p62.

Figure 3.6. Co-Immunoprecipitation of VCP-EGFP and Atg9. 84

viii

LIST OF ABBREVIATIONS

AgRP Agouti-related peptide

Ambra1 Autophagy and beclin 1 regulator 1

Atg Autophagy-related genes

AUP1 Ancient ubiquitous protein

ATGL Adipose triglyceride lipase

BAR Bin-Amphiphysin-Rvs

BAT Brown adipose tissue

Bax Bcl-2-associated X

Bcl2 B-cell lymphoma 2

Bif-1 Bax-interacting factor 1

CD Control/standard chow diet

Cdk5 Cyclin-dependent kinase 5

CDT1 Chromatin Licensing And DNA Replication Factor 1

CGI-58 Comparative gene identification-58

CMA Chaperone-mediated autophagy

COPI Coat protein complex I

Cvt Cytoplasm to vacuole targeting

Dapk-1 Death-associated protein kinase 1

DDB2 DNA damage-binding protein 2

DFCP1 Double FYVE-Containing Protein 1

DNM2 Dynamin 2

EEA1 Early endosome antigen 1

EGFR Epidermal growth factor receptor eIF-2α Eukaryotic Initiation Factor 2

ERK Extracellular signal–regulated kinases

ESCRT Endosomal sorting complexes required for transport

FFA Free fatty acid

FIP200 Focal adhesion family kinase-interacting protein of 200 kDa

GABARAP Gamma-aminobutyric acid receptor-associated protein

HDAC6 Histone deacetylase 6

HFD High fat diet

HOPS Homotypic fusion and protein sorting

HSC70 Heat shock cognate 70 kDa protein

HSL Hormone-sensitive lipase

IBMPFD Inclusion body myopathy with Paget's disease of bone and frontotemporal dementia

IkBα Inhibitor of kappa B

ILVs Intraluminal vesicles

JNK c-Jun N-terminal kinase

LC3 Microtubule-associated protein 1A/1B-light chain 3

LD Lipid droplets

Lamp-1 Lysosome-associated membrane protein 1

Lamp-2A Lysosome-associated membrane protein 2A

MEF Mouse embryonic fibroblast

MGL Monoacylglycerol lipase

MRLC Myosin regulatory light chain mTOR Mammalian target of rapamycin

MVB Multivesicular body

NFκB Nuclear factor κB

NGF Neuronal growth factor

ORO Oil Red O

ORP1L Oxysterol binding protein related protein 1L

PAS Pre-autophagosomal structure or phagophore assembly site p38IP p38-interacting protein

PI3KC3 Class III phosphatidylinositol 3-kinase

PKR Protein Kinase R

PLIN Perilipin

POMC Proopiomelanocortin

PPARγ Peroxisome-proliferator activated receptor gamma

PRR Proline-rich region

RER Respiratory exchange ratio

RILP Rab-interacting lysosomal protein

ROS Reactive Oxygen Species

SH3 Src-homology 3

STAT3 Signal transducer and activator of transcription 3

STX17 Syntaxin 17

SNAP-29 Synaptosomal-associated protein 29

TAG Triacylglyceride

TFEB Transcription factor EB

TGN Trans-Golgi network

TrkA Tropomyosin receptor kinase A

Ulk1/2 Unc-51-like kinase 1/2

UVRAG UV Radiation Resistance Associated

VCP Valosin containing protein

VMP1 Vacuole membrane protein 1

VSV Vesicular stomatitis virus

WIPI2 WD Repeat Domain Phosphoinositide Interacting 2

ZIPK Zipper-interacting protein kinase

α-MSH α-Melanocyte-stimulating hormone

xii

ACKNOWLEDGEMENTS

First and foremost, I would like to thank Dr. Hong-Gang Wang, Dr. Yoshinori

Takahashi, Dr. Richard Mailman, Dr. Thomas Spratt, Dr. Shantu Amin, and Dr. Roger Shi for serving as my dissertation committee. This dissertation could not have been completed without your invaluable suggestions and feedback.

I would especially like to thank my mentor Dr. Hong-Gang Wang for guiding and supporting me over the years. You have set an example of excellence as a scientist, an advisor, and a mentor. This learning experience will benefit me profoundly in my future career.

I would also like to extend my thanks and gratitude to Yoshi for all of his advice, help, constant enthusiasm and encouragement.

My dissertation is truly a team effort and I would like to thank all members of the

Wang lab for their important contribution and discussion.

I am also thankful to my previous mentor Dr. Philip Lazarus, who helped me become a better researcher and enlightened me deeply in this learning journey.

Finally, I am very grateful to my amazing family for their endless love and encouragement. In particular, I would like to thank my husband, Kai Cao. You have been taking over every ounce of burden away from my shoulder and walking me through all the difficult times during the past many years. Your unlimited support has contributed significantly to my career, my academic pursuits, and my spiritual growth. I undoubtedly could not have done this without you.

xiii

The work included in Chapter 2 was originally published in Scientific Reports.

Ying Liu, Yoshinori Takahashi, Neelam Desai, Jun Zhang, Jacob M. Serfass, Yu-Guang

Shi, Christopher J. Lynch & Hong-Gang Wang. Bif-1 deficiency impairs lipid homeostasis and causes obesity accompanied by insulin resistance. Scientific Reports 6, Article number:

20453 (2016).

xiv

Chapter 1. LITERATURE REVIEW

1.1. Autophagy

1.1.1. Molecular Machinery of Autophagy

Autophagy is a collection of tightly regulated catabolic processes, where cytoplasmic components are delivered to the lysosome for degradation.1-4 There are broadly three types of autophagy: macroautophagy, microautophagy and chaperone- mediated autophagy (CMA) (Figure 1.1). Macroautophagy involves the formation of the double-membraned vesicles termed autophagosomes that engulf and sequester the cytoplasmic cargo and transport them to the lysosome. Microautophagy implicates the invagination of the lysosomal membrane, which directly engulfs, traps, and eventually degrades the substrates. CMA is distinct from macroautophagy and microautophagy in that cargo is not sequestered within a membrane bound vesicle. In CMA, the autophagy substrates are proteins containing a KFERQ-like motif that is recognized by the chaperone heat shock cognate 70 kDa protein (HSC70); HSC70 promotes the translocation of these substrates across the lysosomal membranes into the lysosomal lumen through lysosome-associated membrane protein 2A (LAMP2A) receptor.1,4,5

Macroautophagy (hereafter referred to as autophagy) is the most extensively studied among the three types of autophagy, and is also the focus of this dissertation.

The molecular machinery of autophagy is highly conserved from yeast to mammals, and is tightly controlled by a cascade of over 30 autophagy-related (Atg) genes.5 The key steps in the autophagy process include initiation, vesicle nucleation (the

formation of a cup-shaped isolation membrane), vesicle elongation, closure, fusion and degradation (Figure 1.1).1,5

Initiation of autophagy:

Autophagy is initiated through the serine-threonine protein kinase Ulk1/2 (Unc-

51-like kinase 1/2) complex comprising Ulk1/2, Atg13, FIP200, and Atg101.1 In basal states, Ulk1/2 and Atg13 are phosphorylated by mTORC1 (mechanistic target of rapamycin complex 1) and this inhibits autophagy.6 Upon nutrient deprivation, the energy sensor AMPK (AMP-activated protein kinase) gets activated, which directly phosphorylates and inhibits mTORC1. The mTORC1-dependent phosphorylation sites in

Ulk1/2 are then rapidly dephosphorylated by phosphatase PP2A, causing the dissociation of Ulk1/2 from the mTORC1 complex.1,7 Ulk1/2 can then be directly phosphorylated by

AMPK and rendered enzymatically active.7 The activated Ulk1/2 autophosphorylates itself and further phosphorylates Atg13, FIP200, Ambra1, and Beclin 1 to start the downstream process of autophagy.8,9

Nucleation:

The Class III phosphoinositide 3-kinase (PIK3C3) complex consisting of PIK3C3

(Vps34), PIK3R4 (p150), Atg14, Beclin 1, and Nrbf2 (Atg38), participates in the nucleation step.1 Under autophagy-inducing conditions, the activated Ulk1/2 complex recruits the PIK3C3 complex to the autophagosome formation site through phosphorylation of Ambra1 (Autophagy and beclin 1 regulator 1).9 Ulk1/2 also enhances

the activity of the PIK3C3 complex through direct phosphorylation of Beclin-1.8 The activated PIK3C3 complex generates phosphatidylinositol 3-monophosphate (PI3P). As the BATS (Barkor/Atg14L autophagosome-targeting sequence) domain of Atg14 favors highly curved membranes containing PI3P, it is proposed that the BATS domain senses membrane curvature and binds to the isolation membrane, thereby targeting the rest of the Atg14 containing PI3KC3 complex to the curved isolation membrane to efficiently produce more PI3P.10 This results in a feed-forward loop of PI3P production and membrane curvature that further recruits other PI3P interacting proteins DFCP1 (Double

FYVE domain–containing protein 1) and WIPI1/2 (WD repeat domain phosphoinositide- interacting protein 1/2).11,12 Nrbf2/Atg38 was identified only recently as a component of the PIK3C3 complex.13,14 However, whether Nrbf2/Atg38 promotes or suppresses

PIK3C3 activity remains unclear and future studies are needed to dissect its precise role for autophagy. Moreover, Bax-interacting factor 1 (Bif-1) interacts with Beclin 1 through

UVRAG (ultraviolet radiation resistance-associated gene) to positively regulate PIK3C3 activity and autophagy.15 Bif-1 has membrane curvature-inducing activity,16 therefore it may also play a role in the induction and/or stabilization of autophagosome membrane curvature.

Elongation:

The elongation and expansion of the autophagosomal membrane are mediated by two ubiquitin-like conjugation systems: ATG12–ATG5–ATG16L and ATG8 (LC3)– phosphatidylethanolamine.1,17 In the first system, ATG12 is conjugated to ATG5 through

an E1-like ubiquitin-activating enzyme ATG7 and an E2-like ubiquitin-conjugating enzyme ATG10. The ATG12–ATG5 conjugate binds to ATG16L, forming the ATG12–

ATG5–ATG16L complex, which then participates in LC3–phosphatidylethanolamine conjugation. To prime proLC3 for lipidation, the protease Atg4B cleaves the C-terminus of this molecule exposing a glycine residue critical for conjugation.18,19 The cleaved form of LC3 is referred to as LC3-I, which is then conjugated to phosphatidylethanolamine by

ATG7, another E2-like enzyme ATG3 and the ATG12–ATG5–ATG16L complex that act as an E3-like enzyme complex.20 The exact function of LC3 in the expansion of autophagosomal membranes remains unclear. Nonetheless, LC3-II is the only protein known to be directly and stably associated with the autophagosomal membrane until fusion with a lysosome and thus is a widely used marker of autophagosomes.21

Closure:

The closure of the autophagosomal membranes is the last part of autophagosome formation, and is the least understood step; it has been indicated that the membrane edges of an open autophagosome need to come close and go through membrane scission in order to allow the vesicle to seal.22 Recent studies have identified GABARAP (gamma- aminobutyric acid receptor-associated protein), Atg2A/B, and VMP1 (vacuole membrane protein 1) as potential regulators of closure. GABARAP is a mammalian isoform of LC3, and its conjugation to PE appears to be essential for autophagosome closure.23 Atg2 appears to regulate autophagosome sealing, as loss of Atg2 in both yeast and mammalian cells accumulates immature autophagic structures.24,25 VMP1 appears to be critical in 5

autophagosome closure, as loss of VMP1 accumulates open structures positive for

Atg16L, LC3 and p62.25

Fusion and degradation:

The final step in autophagy is the fusion of closed autophagosomes with late endosomes and/or lysosomes, where the encapsulated cargo is degraded by acidic hydrolases, proteases, and lipases. It is important to note that the fusion step is a precisely and tightly regulated process, and only occurs with completed autophagosomes.26 The mechanism for such specificity remains unknown, although it has been proposed that the retrieval of the protein complexes recruited during the initiation and elongation phases may serve as a potential trigger to commence fusion, as sealed autophagosomes are stripped of Ulk1/2, Beclin 1, Atg5 and Atg16.27

The mechanism of autophagosome-lysosome fusion is beginning to be understood, and it involves the interactions between a plethora of proteins including some specific for autophagosome/lysosome fusion and others involved in vesicular fusion more generally (Figure 1.2). Microtubules and the dynein-dynactin motor complex facilitate the transport of peripherally generated autophagosomes to perinuclearly located late endosomes and lysosomes.28-30 This enables the action of the SNAREs, which are membrane-anchored proteins that promote membrane fusion through the assembly of a trans-SNARE complex consisting of one vesicle v-SNARE on the donor membrane and t-SNAREs on the acceptor membrane. Tethering autophagosomes to lysosomes is mediated by the Rab7 effector proteins ORP1L (oxysterol binding protein related protein 6

1L) and RILP (Rab-interacting lysosomal protein), which recruit the dynein-dynactin motor and interacts with the HOPS (homotypic fusion and protein sorting) complex. This interaction is tightly controlled by ORP1L that senses cholesterol levels and, under low cholesterol conditions, prevents the recruitment of HOPS to Rab7-RILP and thus inhibits membrane tethering.31 The HOPS complex will then bring lysosomes and autophagosomes together by cross-linking Rab7 on lysosomes to syntaxin 17 (STX17) on autophagosomes in a process promoted by Atg14.32 Moreover, STX17 is also the Qa-

SNARE essential for autophagosome-lysosome fusion.33 After targeted to the autophagosome, STX17 recruits the Qbc-SNARE synaptosomal-associated protein 29

(SNAP-29) and the complex binds to oligomeric Atg14 and then interacts with the R-

SNARE VAMP8 on the lysosomes.

Atg9 Vesicles:

Atg9 is the only transmembrane protein among the core Atg proteins.34 Since it localizes to membrane vesicles and is transported to autophagosome formation sites, it has been proposed that Atg9 vesicles could supply membrane sources for autophagosome biogenesis.35-37 This possibility and the detailed Atg9 trafficking pathways will be further discussed in later sections.

Figure 1.1 Overview of the Autophagy Process and Molecular Machinery. There are broadly three types of autophagy: macroautophagy, microautophagy and chaperone-mediated autophagy (CMA). Key steps in the autophagy process include initiation, vesicle nucleation, vesicle elongation, closure, fusion and degradation. Autophagy is initiated through the serine-threonine protein kinase Ulk1/2 complex comprising

Ulk1/2, Atg13, FIP200, and Atg101. The Class III phosphoinositide 3-kinase (PIK3C3) complex consists of PIK3C3 (Vps34), PIK3R4 (p150), Atg14, and Beclin 1, participates in the nucleation step. The elongation and expansion of the autophagosomal membrane are mediated by two ubiquitin-like conjugation systems: ATG12–ATG5–ATG16L and LC3–phosphatidylethanolamine. The autophagosome eventually close and fuses with late endosomes and/or lysosomes, where the encapsulated cargo is degraded by acidic hydrolases, proteases, and lipases. (Adapted from: Autophagy at the crossroads of catabolism and anabolism, Nat Rev Mol Cell Biol. 2015 , with modifications.)

Figure 1.2 Model of Autophagosome – Late Endosome/Lysosome Fusion. Tethering of autophagosomes with late endodome (LE) is orchestrated by Rab7 and its effectors ORP1L (cholesterol sensor) and RILP, that bind to the HOPS complex. The SNARE protein Syntaxin 17 on autophagosome binds HOPS and mediates fusion together with the SNAREs SNAP29 and VAMP8, and ATG14. Upon fusion, RILP recruits dynactin–dynein to amphisomes to promote minus-end transport. (Adapted from: Cholesterol and ORP1L- mediated ER contact sites control autophagosome transport and fusion with the endocytic pathway, Nat

Commun. 2016.)

1.1.2. Regulation of Autophagy by Nutrient and Energy Status

Autophagy functions as a survival mechanism under conditions of stress, maintaining cellular homeostasis and integrity by regenerating metabolic precursors and clearing dysfunctional proteins or organelles.38 Proteins, lipids, and carbohydrates as three main categories of energy macromolecules can all be degraded by autophagy; and in turn, amino acids, fatty acids, and glucose all serve as metabolic cues to induce or inhibit autophagy activity.

Regulation of autophagy by amino acids: Most extensively studied

Seminal studies showing that autophagy is responsive to fluctuations in amino acids predate the identification and cloning of the ATG genes. In 1977, Schworer and colleagues showed that perfusion of rat liver in the absence of amino acids rapidly increased the number of autophagosomes.39 Conversely, high concentration of amino acids inhibited autophagy.39,40 It was unknown at that time how amino acids repress autophagy, and more recent studies found this process is mediated by mechanistic target of rapamycin (mTOR).41-43 mTOR is a highly conserved serine/threonine kinase that signals in response to multiple stimuli including amino acids, energy levels, oxygen, growth factors, and stress to regulate cell growth and maintain metabolic homeostasis.44

In mammalian cells, mTOR forms two functionally distinct complexes, mTORC1

(mTOR complex 1) and mTORC2 (mTOR complex 2). mTORC1 is enzymatically active in the presence of amino acids. This activation process starts with the Ragulator complex sensing amino acids and activating the Rag GTPase complexe that then tethers mTORC1

to the lysosome. The recruitment of mTORC1 to the lysosome brings it into proximity with another small GTPase Rheb that is critical for mTORC1 activation.45 As Rheb is active only in the presence of growth factors, full activation of mTORC1 can only be achieved when both amino acids and growth factors are sufficient.

Regulation of autophagy by lipids and fatty acids: Significance is beginning to be appreciated

As the primary source of energy in mammals, lipids also exert marked effects on autophagy. Excessive concentrations of lipids are reported to inhibit autophagy by blocking the fusion of the autophagosome with the lysosome or by impairing lysosomal acidification and hydrolase activity46. By contrast, free fatty acids, such as palmitic acid and oleic acid, induce autophagy through inhibition of mTORC1 or via the PKR-JNK

(Protein Kinase R- c-Jun N-terminal kinase) pathway.47 In addition, palmitic acid disrupts interactions between STAT3 and PKR, leading to PKR-dependent phosphorylation of eIF-2α and subsequent induction of autophagy.48 In mouse embryonic fibroblasts

(MEFs), palmitic acid can also induce autophagy by increasing diacylglycerol content and activating protein kinase C independent of mTORC1.49 Moreover, the degree of saturation of free fatty acids can exert diverse effects on autophagy.50

Regulation of autophagy by glucose: Mechanisms remain elusive and controversial

As an important determinant of the cellular energy status, glucose regulates autophagy with complex mechanisms. It was originally found that glucose deprivation

induces autophagy by increasing the AMP/ATP ratio that activates AMPK (5' AMP- activated protein kinase), as a mechanism to restore ATP from cellular components.38

AMPK activates autophagy through TSC (tuberous sclerosis complex) phosphorylation and mTORC1 inhibition. Moreover, AMPK phosphorylates and activates p53. Activated p53 translocates to the nucleus, where it promotes the expression of different genes implicated in autophagy such as Sestrin2 and Dapk-1 (death-associated protein kinase

1).51 Sestrin2 in turn activates AMPK and thus starts a positive feedback loop;52 and

Dapk-1 phosphorylates Beclin1 and prevents the formation of the Beclin1/Bcl-2 complex that would otherwise inhibit autophagy.53

It is important to note that, glucose regulates autophagy not only necessarily through the sensing of ATP levels. Both excessively low and high levels of glucose can alter the oxidative phosphorylation in cells leading to mitochondrial hyperpolarization54 and accumulation of ROS55. The increase in ROS levels can be sensed by a cascade of proteins such as AMPK, p38 MAPK (P38 mitogen-activated protein kinase),56,57 JNK58,59 and ERK (Extracellular signal–regulated kinase)60; and the activation of these kinases eventually induce autophagy.

Regulation of autophagy by hormones: Of ultimate pathological implications

In addition to the nutrients described above, hormones, in particular insulin and glucagon, are key regulators of autophagy. Insulin inhibits39, whereas glucagon stimulates this catabolic process61. Autophagy's tight regulation by both nutrient availability and hormones, along with its roles in whole-body metabolism and cellular 12

adaptation to stress, imply its critical involvement in the pathophysiology of common metabolic disorders, such as obesity and type II diabetes. In recent years, emerging studies have explored the roles of autophagy in obesity and its complications, and the underlying links and mechanisms are beginning to be understood.

1.2. Obesity and Autophagy

1.2.1. Links between Obesity and Autophagy

Obesity is a complex disorder involving an excessive amount of body fat. It has become a worldwide health problem as a consequence of high-calorie intake and the adoption of sedentary lifestyles. In 2012, nearly 17% of children and 35% of adults aged

20 years or older are obese in the US62. Obesity is associated with various health effects such as heart disease, type II diabetes, stroke, osteoarthritis, and several types of cancer63.

However, controlling obesity and its associated health conditions remains a major challenge due to an incomplete understanding of its molecular machinery.

In recent years, it has been reported that autophagy plays complex roles in obesity and insulin resistance by regulating food intake, adipose tissue development, as well as fat metabolism in the liver, pancreatic β cells and skeletal muscles. Here I will review the current understanding on the implication of autophagy in these different aspects of obesity pathophysiology.

Food Intake

The hypothalamus is a central site of regulation for appetite and whole-body energy expenditure. There are two main neuronal populations of the hypothalamic arcuate nucleus that directly involve autophagy: the agouti-related peptide (AgRP)- expressing neurons and the proopiomelanocortin (POMC)-expressing neurons. AgRP neurons produce orexigenic hormones, such as AgRP and neuropeptide Y, whereas

POMC neurons produce anorexigenic hormones such as POMC and α-melanocyte- stimulating hormone (α-MSH). Autophagy appears to play distinct roles in these two populations in relation to obesity.

POMC neuron-specific Atg7 knockout (KO, Atg7 ΔPOMC) in mice caused increased food intake, decreased energy expenditure, and these mice developed obesity after being challenged with a high-fat diet and displayed impaired glucose tolerance and insulin resistance.64 These metabolic phenotypes in Atg7ΔPOMC mice were caused by decreased production and secretion of α-melanocyte-stimulating hormone (α-MSH).

Moreover, loss of Atg7 in POMC neuron also impaired leptin signaling that also influences feeding behavior. In addition, impaired leptin signaling in POMC neurons of the Atg7 KO mice might also reduce peripheral lipolysis.

In contrast, mice with Atg7 deletion in the AgRP neurons (Atg7ΔAgRP) had impaired induction in the expression of AgRP in response to the free fatty acids released during starvation.65 Concurrently, these mice express increased levels of the anorexigenic hormones POMC and α-MSH, and consequently display a lean phenotype. These studies

suggest that metabolic influence of hypothalamic autophagy is dependent on the types of neurons affected and is regulated by complicated interactions between different neuron population.

Adipose Tissue Development

It is well established that autophagy is essential for overall early development – from embryogenesis,66-69 over stem cell differentiation into various specialized cell types,70-72 up to the adaptation and survival of neonates73,74. Initial evidence for a role of autophagy in adipose tissue development specifically, was obtained from the post mortem analysis of the subcutaneous tissues in neonatal Atg5 whole-body KO mice that die soon after birth.75 These Atg5 KO animals had significantly reduced numbers of adipocytes in comparison to their wild-type counterparts. Later studies using either in-vitro adipogenesis or in-vivo tissue-specific KO approaches have emerged, in order to more specifically delineate the role of autophagy in adipose tissue development. In-vitro assays using the 3T3-L1 cell line or mouse embryonic fibroblasts (MEFs) revealed that autophagy appears essential for the completion of adipogenesis75. Cells depleted of Atg5 or Atg7 fail to differentiate into mature adipocyte-like cells, even though they do initially form fewer and smaller lipid-droplets in comparison to control cells upon differentiation induction. Consistently, these autophagy deficient cells have lower degree of induction in the mRNA and protein levels of Pparγ (peroxisome proliferator-activated receptor-γ) and

Cebpα (CCAAT/enhancer binding protein-α) than wild-type cells.76,77 Similar results

were observed when autophagy was inhibited pharmacologically using chloroquine or 3- methyladenine.76

In mice with a targeted deletion of Atg7 in the adipose tissue (Atg7ΔAT), white adipose tissue mass is markedly reduced and presents features of brown adipose tissue

(BAT).77 The Atg7ΔAT mice are protected from diet-induced obesity, and showed improved insulin and glucose sensitivities. In addition to white adipose tissue, a role for autophagy in brown adipose tissue development has also been explored. Mice having brown adipose tissue-specific Atg7 deletion show defective BAT differentiation and impaired thermogenic response to cold exposure.78

Changes in adipose tissue autophagy have also been examined in the context of obesity. Elevated expression levels of autophagy-related genes in adipose tissue appear to correlate with the degree of adiposity in both patients with obesity and obese mice.79,80

Autophagic activity also positively correlates with adipose-tissue inflammation as characterized by macrophage infiltration. However, inhibition of autophagy results in a marked increase in the expression of proinflammatory cytokines in adipose tissue of obese mice or patients with obesity. These results appear counter-intuitive to each other, but their discrepancy indicates that autophagy may play a compensatory role to attenuate obesity-induced inflammation and insulin resistance.

Liver Complications

Liver is a major metabolism organ, and therefore the role of hepatocyte autophagy in body metabolism has been extensively investigated. In the context of lipid metabolism,

recent studies have discovered that autophagy mediates the lipolysis of hepatic triacylglycerides (TAGs) stored in lipid droplets (LDs) in a process referred to as lipophagy.81 During lipophagy, a portion of a large LD or an entire small LD are encapsulated into an autophagosome and delivered to the lysosome where TAGs and lipid droplet coating proteins, such as perilipins, are hydrolyzed by lysosomal lipases and proteases, respectively. Singh R. et al. found that suppression of lipophagy by liver- specific depletion of Atg5, Atg7, has the net effect of promoting hepatic lipid accumulation. In the mice lacking lipophagy in liver, a reduction in hepatocyte mitochondrial fatty acid β-oxidation also occurred presumably secondary to the reduced supply of free fatty acids.81

In addition, liver specific knockdown of Atg14 and knockout of the positive regulator of autophagy transcription factor EB (TFEB) increased hepatic TAGs82, whereas adenoviral overexpression of Atg7 or Atg14 successfully reversed hepatic steatosis in genetic obesity models. Conversely, hepatocyte deletion of focal adhesion family kinase-interacting protein of 200 kDa (FIP200), a core component of the autophagy initiating complex, reduced hepatic TG content after starvation or HFD feeding.83 It is possible that loss of FIP200 may have altered lipid metabolism through off-target effects independent of autophagy as FIP200 deficiency affected many other genes involved in lipid metabolism.

The roles of hepatic autophagy in insulin sensitivity and type II diabetes mellitus have also been explored. Reduced expression of Atg7 in the liver leads to increased ER stress and worsened insulin resistance in lean mice, whereas adenoviral overexpression of 17

Atg7 relieves the metabolic complications in obese mice. Similarly, adenoviral overexpression of TFEB ameliorates insulin resistance in mice fed a high-fat diet and in ob/ob mice.82 Finally, obesity and the associating insulin resistance impair hepatic autophagy through either reducing the expression of autophagy-related genes, or promoting the post-translational cleavage of core autophagy proteins by the protease calpain.

Pancreas Complications

The understanding of β-cell autophagy in body metabolism came from studies using β cell-specific autophagy deficient mice. Atg7Δβ cell mice display hyperglycemia, glucose intolerance and hypoinsulinemia characteristics due to reduced β-cell mass and insulin secretion.84 However, these mice do not develop diabetes, which is in drastic contrast with the Atg7Δβ cell -ob/ob mice that develop severe diabetes mellitus accompanied by markedly increased β cell apoptosis, reduced β cell mass, elevated oxidative stress in pancreatic islets and severely impaired β cell function.85 This phenotype difference indicates that autophagy in β cells is likely dispensable for the development of diabetes per se, but it could prevent the progression from obesity to diabetes mellitus. In the pancreatic islets of obese mice, increased numbers of autophagosomes are observed; however, autophagic activity has not been determined due to lack of appropriate marker for autophagic degradation in tissues collected at a static time. Thus, future studies are necessary to assess the effect of obesity and type II diabetes on the autophagic activity in β cells and pancreatic islets.

Skeletal Muscle Complications

The role of skeletal muscle autophagy in glucose utilization or insulin sensitivity has been investigated using skeletal muscle-specific Atg7-KO (Atg7Δsm) mice. The KO mice have reduced lean and fat mass, enhanced glucose clearance, and consequently, develop muscle atrophy with decreased muscle force as compared to the wild-type mice.86 Moreover, skeletal muscle-specific Atg7 depletion protected mice from obesity and insulin resistance upon a high-fat diet challenge.

Lipid accumulation in skeletal muscle is a prominent feature of obesity and type

II diabetes mellitus. Because autophagy participates in lipid breakdown, modulating autophagy in skeletal muscle of animals and humans with obesity could potentially alleviate ectopic lipid burdens. One study found that the expression levels of Beclin-1 and

LC3-II are upregulated in the skeletal muscle of ob/ob mice; however, the autophagic activity has not been determined in these mice. Whether obesity inhibits autophagy in skeletal muscle and thereby promotes lipid accumulation or autophagy is upregulated to compensate for lipid overload in skeletal muscle, remains to be determined.

1.2.2. Lipophagy Mechanisms

Lipid droplets and the traditional hydrolysis pathway: the old

All cells store lipids in the form of lipid droplets (LDs). LD is now considered a subcellular organelle consisting of a neutral lipid core of triglycerides and cholesteryl esters that is limited by a monolayer of phospholipids and a family of LD surface decorating proteins, classified as perilipins (PLINs)87. These PLINs play dual roles in

lipolysis. Under basal state, PLINs act as gatekeepers to prevent the lipases from degrading LD88, whereas in demand of energy supply, PLINs serve as anchoring sites for lipase to facilitate the LD breakdown89. In most cell types, LD ranges from 0.1 to 10 μm in size, whereas in adipocytes, LDs are 10–100 times larger. The traditional lipolysis is hormone induced and involves the coordinated activities of neutral lipases including adipose triglyceride lipase (ATGL), hormone-sensitive lipase (HSL) and monoacylglycerol lipase (MGL)90. During lipolysis, which is best characterized in the adipocyte, hormones such as epinephrine activate Protein Kinase A, leading to the phosphorylation and proteasomal degradation of perilipin 1 (PLIN1). The phosphorylation of PLIN1 causes in the release of comparative gene identification-58

(CGI-58) that then activates ATGL and initiates triglyceride breakdown; the resulting diacylglyerides and monoacylglycerides are then hydrolyzed by HSL and MGL, respectively.

Lipophagy as a ubiquitous way of lipid hydrolysis: the new

In addition to the traditional lipolysis pathway that breaks down fat, as discussed above, autophagy also plays critical roles in lipid turnover to control fat accumulation in various insulin responsive tissues. This process, termed lipophagy, has now been demonstrated in more diverse cell types such as striatal neurons,91 glial cells,91 macrophage foam cells,92 enterocytes,93 T cells94, fibroblasts,95 adipose-resident macrophages92, prostate carcinoma cells96, as well as in other organisms including

Saccharomyces cerevisiae,97 Caenorhabditis elegans,98 and certain fungal species99. The

ubiquitous nature of lipophagy hints at its role as a likely conserved mechanism to counteract excessive lipid build-up or to rapidly mobilize lipid reserves for diverse cellular and tissue functions.

Lipid droplet targeting and sequestration mechanisms: unclear

The mechanisms of lipophagy, particularly those related to how autophagy machinery selectively targets LD, and how autophagosome biogenesis occurs to sequester the substrates, remain elusive. While polyubiquitination is a well-established mechanism to target proteins and organelles destined for degradation, it remains unclear whether polyubiquitination or similar modification events serve to tag and degrade LDs.

It has been proposed that the LD localized ancient ubiquitous protein (AUP1) and its interaction partner E2 ubiquitin-conjugating enzyme Ube2g2 may form a complex that labels LD for degradation.100 Future studies are needed to test this hypothesis, and to exclude the possibility that AUP1 is not just sequestered at LD similar to histone proteins101 as a mechanism of its function inactivation.

Regarding the sequestration mechanism, it has been suggested that in lipophagy an autophagosome either nibbles or swamps an LD. In the study of Singh et al., the authors observed that both LC3-I and LC3-II were found in lipid fraction of the cell,81 concluding that the lipidation of LC3 and potentially the de novo biogenesis of the autophagosomes to sequester LD may occur at the lipid droplet surface. In line with these findings, Dupont et al. found that a neural lipase PNPLA5 (Patatin-like phospholipase domain-containing protein 5) converts TAG to DAG, which is further metabolized to

phospholipids; the phospholipids supply lipid sources for autophagosome biogenesis; and

presence of LD enhances the autophagic degradation of other diverse substrates.102

Chaperon mediated autophagy participates in lipophagy: paves the way for macroautophagy of LD

Alternative mechanisms by which autophagy may regulate hepatocyte lipid

metabolism have been demonstrated for chaperon mediated autophagy (CMA). Mice

having liver-specific CMA deficiency by deletion of Lamp-2A (Lysosome-associated

membrane protein 2) developed hepatic steatosis accompanied by abnormal carbohydrate

and lipid metabolism. Moreover, perilipin 2 (PLIN2) and perilipin 3 (PLIN3), two LD

coating proteins, were identified as CMA substrates that are degraded prior to the start of

lipolysis103. Blocking CMA in cultured cells and mouse liver reduced the LD association

of adipose triglyceride lipase ATGL and macroautophagy-related proteins, decreased

lipid oxidation, and promoted the accumulation of lipid droplets. These studies suggest

that degradation of LD surface perilipins by CMA may act to prime hormonal lipolysis

and macroautophagy of lipid droplets (Figure 1.3).

Figure 1.3. Model of Lipophagy Mechanisms. (A) Autophagosome engulfs an entire small lipid droplet or nibbles portions of a large lipid droplet, and delivers the lipid cargo to lysosomes for degradation. Fatty acids are released into the cytosol and used by the mitochondria in β-oxidation. (B) Chaperone-mediated autophagy (CMA) degrades the lipid droplet coat proteins PLIN2 and PLIN3 through coordinated action of Hsc70 and the LAMP-2A receptor, which primes the access for lipase ATGL and the autophagic machinery. (Adapted from: Autophagy, lipophagy and lysosomal lipid storage disorders, Biochim Biophys Acta. 2016.) 23

1.3. Bif-1 as a Multifunctional Protein

Bax interacting factor 1 (Bif-1, also known as SH3GLB1 or Endophilin B1) is a member of the membrane curvature-driving endophilin protein family. Previous studies have shown that Bif-1 serves diverse cellular function from autophagy15,37,104, apoptosis105, endocytosis106 107, over mitochondrial quality control108, to cytokinesis109.

Before discussing the novel role of Bif-1 in lipid metabolism and obesity being discovered in the dissertation study, here I will first review the basic information of this molecule and some of its previously established function.

1.3.1. Overview of Bif-1 Structure, Expression, and Function

Structure:

The endophilin proteins can be categorized into two subfamilies based on sequence homology: Endophilin A (Endophilin A1-3) and Endophilin B (Bif-

1/SH3GLB1/Endophilin B1 and SH3GLB2/Endophilin B2). Both endophilin A and B proteins contain a C-terminal SH3 (Src-homology 3) domain that interacts with proteins containing a proline-rich region (PRR), and an N-terminal BAR (Bin-Amphiphysin-Rvs) domain that binds and bends the lipid bilayers.110 The N-BAR domain consists of three major regions: an amphipathic helix 0 (H0), an amphipathic helix inserted in helix 1

(H1I), and a main body. Two neighboring N-BAR domains can interact through a coiled- coil region in the main body to form a crescent-shaped dimer. To drive membrane curvature, the H0 regions bind to membranes, and the H1I regions insert into and then pull out the membranes in concert with the dimerized main bodies.37,111 Moreover, the

H1I region appears to be needed for correct localization of specific endophilin, as replacing H1I of Bif-1 with that of endophilin A1 erroneously targets Bif-1 to plasma membrane instead of cytosol.112,113

Expression:

The Bif-1/endophilin B1 gene has four splice variants including endophilin B1a,

B1b and B1c and endophilin B1t. Endophilin B1a has widespread tissue distribution; B1b and B1c appear to be brain-specific,114 and have a short (s) and long (l) exon 6 (relative to endoB1a), respectively; B1t is highly expressed in testis and lacks the C-terminal 10 amino acids due to a premature stop codon.115 This dissertation study is focused on endophilin B1a, hereafter referred to as Bif-1.

Bif-1 protein is evolutionarily conserved, with the human and mouse homologues sharing 96% amino acid identity.116 It has been reported to express in most tissues including heart, skeletal muscle, kidney, spleen, thymus, breast, colon, liver, placenta, lung, and peripheral blood leukocyte;116,117 current study also found abundant Bif-1 expression in adipose tissues. Bif-1 localizes mainly in the cytosol, and forms a homodimer with another Bif-1 protein or a heterodimer with endophilin B2 through the

N-BAR domain.117 This dimerization is important for the membrane bending activity and may be needed for optimal Bif-1 function.

Autophagy:

Our laboratory has previously demonstrated that Bif-1 interacts with Beclin1 through UVRAG (UV Radiation Resistance Associated) to promote the activity of the

PI3KC3 (class III phosphatidylinositol 3-kinase) complex in autophagosome formation.

In response to nutrient starvation, Bif-1 accumulates in cytosolic foci positive for Atg5 and LC3.15 Given that Atg5 localizes to the phagophore only during early stages of autophagosome formation, these findings support a role for Bif-1 in the nucleation and/or the expansion of phagophores. Furthermore, Ip and collegues have recently demonstrated in starved neurons that cyclin-dependent kinase 5 (Cdk5) phosphorylates Bif-1 on

Thr145, located within the N-BAR domain, resulting in the induction of autophagy in models of Parkinson’s disease.118 Mechanistically, Thr145 phosphorylation facilitates

Bif-1 dimerization and the recruitment of UVRAG/Beclin1, therefore stimulating autophagy in starved neurons. Moreover, Bif-1 regulates the trafficking of Atg9, which is a multispanning transmembrane protein critical for autophagosome formation.37,119

During autophagy, vesicles containing Atg9 are generated and delivered to the autophagosome formation sites dependent on Bif-1. Our recent study provided more mechanistic details to the Atg9 vesicle generation with the identification of Dynamin 2

(DNM2) as a Bif-1 interacting protein during starvation. Bif-1 and DNM2 coordinately induce the generation of Atg9-containing vesicles that appear to originate from a Rab11- positive reservoir.119 These findings further support the importance of Bif-1 function in autophagy.

Apoptosis:

Bif-1 was originally discovered as a Bax-binding protein. In response to apoptotic stimuli, Bif-1 translocates from the cytosol to the mitochondria where it binds to and heterodimerizes with Bcl-2-associated X (Bax) on the mitochondria membrane.116 This results in a conformational change in Bax leading to its integration as an oligomer into the mitochondrial membrane that subsequently causes the release of cytochrome C from the mitochondrial intermembrane space and the initiation of caspase dependent apoptosis.

Bif-1 also functions in the Bax-dependent and detachment induced apoptosis, a process referred to as anoikis.120 This process is characterized as a natural cellular response to loss of proper integrin function and is vital to the prevention of metastasis as cells resistant to anoikis are able to survive in circulation.

Endocytosis:

Bif-1 also functions in endocytic trafficking in a context-dependent manner. Ip and collegues identified Bif-1as a regulator of the endocytic trafficking of neuronal growth factor (NGF) and its receptor TrkA (tropomyosin receptor kinase A) in PC12 cells, through an interaction with early endosome antigen 1(EEA1)107. In neuronal cells, suppression of Bif-1 promotes the endocytic trafficking of TrkA to late endosomes/lysosomes, increases NGF/TrkA lysosomal degradation, attenuates the downstream signaling, and thus abolishes neurite outgrowth. These findings imply a role for Bif-1 as a negative regulator of NGF induced TrkA endocytic trafficking. Conversely, recent work demonstrated by our lab and others, reveal a positive role for Bif-1 in EGF-

induced EGFR endocytosis. Previous studies in our lab demonstrated that in MDA-MB-

231 breast cancer cells, suppression of Bif-1 delays EGFR endocytosis by disrupting endosome maturation106. Moreover, studies in HeLa cervical cancer cells showed that silencing of Bif-1 with small interference RNAs delays the degradation of EGF and

EGFR potentially through interactions with the UVRAG-Beclin1-PI3KC3 complex.109

Taken together, future studies are needed to unravel the discrepancies in Bif-1 function observed in various cell culture model systems and with different receptors under investigation.

1.3.2. Bif-1 and Cancers

The Bif-1 gene locates on chromosome 1p22, a region that is frequently deleted in neoplasms121. Downregulation of Bif-1 expression has been observed in gastric122, colorectal123, prostate124, urinary bladder and gallbladder cancers125. Our lab has previously identified Bif-1 as a critical tumor suppressor in multiple studies. Knockdown of Bif-1 enhances the tumorigenicity of HeLa cells in vivo; targeted deletion of Bif-1 in mice promotes spontaneous tumorigenesis, without causing obvious developmental defect at birth; Bif-1 functions as a haploinsufficient tumor suppressor in a mouse model of Eμ-myc driven lymphomagenesis, as losing even one allele of the Bif-1gene accelerates tumorigenesis and mortality126. Moreover, a potential function for Bif-1 in breast cancer progression has been implied, as decreased Bif-1 expression was correlated with metastasis in a mouse mammary tumor model.127 Somatic mutations in Bif-1 are very rare as revealed by analysis on lung, gastric, hepatocellular, colorectal and breast carcinomas as well as acute

adulthood leukemias,128 indicating that loss of Bif-1 in various tumors is unlikely due to any

Bif-1- intrinsic mechanism such as protein truncation or misfolding.

1.4. Atg9 as an Essential Molecule in Autophagy

1.4.1. Overview of Atg9 Biology and Importance for Autophagy

Atg9, the only multispanning transmembrane protein among over 30 Atg proteins identified to date, is essential to autophagy. Atg9 was originally identified in yeast genetic screens for defects in both the Cvt and autophagic pathways, and is highly conserved across species.129 It has six conserved transmembrane domains and cytosolic

N- and C-termini that are divergent in both length and sequence.34 Deletion of the Atg9 homolog in C. elegans shortens the life span of worms and increases the lethality of embryos that require autophagy for survival.130 Genetic knockdown of Drosophila melanogaster Atg9 in S2 cells increases the infection from vesicular stomatitis virus

(VSV) that is normally eliminated by autophagy.131 Moreover, knockout of Atg9 in plants accelerates leaf senescence due to defective autophagy.132 Finally, mice depleted of Atg9 do not survive the neonatal starvation period, for which autophagy is critical to ensure energy supply.133 On the molecular level, while it is a general consense that Atg9 is critical for autophagy, it has been controversial or unclear in terms of Atg9’s exact role in autophagosome formation. Kishi-Itakura C. et al. reported that Atg9 is absolutely required for autophagosome formation and LC3 lipidation.25 Similarly, Orsi A. et al. showed that Atg9 acts at an early stage, as depletion of Atg9 inhibits the formation of

early ULK1-, DFCP1 (Double FYVE-Containing Protein 1)-, and WIPI2 (WD Repeat

Domain Phosphoinositide Interacting 2)-positive phagophores.34 However, the current study and work of others found that while Atg9 is likely important for both the autophagosome nucleation and expansion (Figure 1.4); its depletion does not completely abolish the phagophore formation or the LC3 lipidation. Precise function of Atg9 in autophagy thus warrants extensive futher studies.

1.4.2. Yeast Atg9 Trafficking and Interactions

Much of our knowledge on Atg9 trafficking and regulation comes from yeast studies. In yeast, Atg9 localizes to peripheral punctate structures scattered throughout the cytoplasm, as well as to the pre-autophagosomal structure or phagophore assembly site

(PAS).134 This is distinctive to the localization of most other Atg proteins, which either target to the PAS or stay diffused in the cytosol. Reggiori et al. and others have shown that Atg9 cycles between the PAS and peripheral pool through interacting with a cascade of binding partners.35 Under autophagy-inducing conditions, Atg17 recruits Atg9 to the

PAS and this process requires the protein kinase Atg1. Atg23 and Atg27 interact with

Atg9 to form a cycling complex, which is important but likely not essential for efficient

Atg9 traffic. Because Atg9 is not present on autophagic vesicles that accumulate in the vacuole (a lysosome like organelle in yeast) upon addition of a protease inhibitor PMSF, it is proposed and generally accepted that Atg9 is absent from the completed autophagosome and is recycled back to the peripheral pool. Retrieval of Atg9 from the

PAS requires the peripheral membrane proteins Atg2 and Atg18 that are recruited by the

autophagosome-asscociating PI3K complex.35 The recruitment of Atg2 also requires

Atg1, Atg18, and Atg9 itself. Atg2 and Atg18 interact with Atg9 and transport it back to the cytosol. Moreover, pull-down assays showed that the interaction of Atg9 with Atg18 requires both Atg1 and Atg2.35 Therefore, Atg1, Atg2, and Atg18 are all interconnected in a complex manner in Atg9 trafficking. Moreover, it is interesting that Atg9 plays a part in recruiting machinery for its own recycling.

1.4.3.Mamalian Atg9 Trafficking and Interaction Network

Subcellular distributions:

Relative to that of yeast, the trafficking, interactions, as well as roles of Atg9 in autophagy are less well understood and appear far more complex. In mammals, under nutrient-rich conditions, ATG9 localizes on the trans-Golgi network (TGN) as well as on recycling and late endosomal compartments.135 Previous reports by others and recent work of our laboratory have suggested that the Rab11-positive recycling endosome may serve as an “Atg9 reservoir” in mammalian cells.119,136,137 Upon nutrient deprivation or pharmacological inhibition of mTOR, Atg9 travels out of the juxtanuclear region resembling the TGN and/or recycling endosomes is redistributed to a pool of vesicles

(Atg9+ vesicles) through a series of membrane tubulation and fission; the Atg9+ vesicles are believed to interact transiently with the autophagosome precursors in a “kiss - and - run” manner without being incorporated into the membrane.34 As Atg9 is not detected on the autophagosome, amphisome, or lysosome surface, the dissociation of Atg9 from the

autophagosomal membrane likely precedes the autophagosome closure. It remains unclear whether Atg9 is retrieved back to the Golgi or endosomes.

Regulators of mAtg9 trafficking:

ULK1, the mammalian homologue of Atg1, was one of the first identified Atg9 trafficking regulators. Knockdown of ULK1 in HEK293 cells abolishes ATG9 redistribution from the juxtanuclear to the peripheral pool in response to starvation.138

ULK1 controls Atg9 trafficking likely through promoting the activity of myosin II that potentially act as a motor protein for Atg9+ vesicle movement.139 ULK1 was shown to phosphorylate zipper-interacting protein kinase (ZIPK), which in turn phosphorylates myosin regulatory light chain (MRLC). Phosphorylation of MRLC promotes the activity of myosin II that was shown to be required for starvation-induced autophagy and ATG9 redistribution.139 Myosin heavy chain, MRLC and ATG9 showed co-localization during nutrient deprivation in imaging experiments. Moreover, myosin heavy chain directly interacts with Atg9; this interaction is disrupted by depletion of either ULK1 or ZIPK,139 suggesting that the ULK1-ZIPK-myosin II pathway is necessary for ATG9 trafficking.

PI3K complex was also shown to be required for Atg9 trafficking, as inhibitors of

PI3K prevented Atg9 redistribution upon nutrient deprivation.37,138 Consistently, depletion of other members of the PI3K complex, such as Beclin-1 and UVRAG, had a similar effect. 37

Bif-1 is another critical regulator of Atg9 trafficking. Earlier work of our lab suggested that Atg9-containing membranes may be partly derived from TGN.37 It was

initially thought that Bif-1 is necessary for tubulating TGN, as HeLa cells lacking Bif-1 show no tubulation and fission of ATG9-positive membranes emanating from Golgi- resembling locations in response to nutrient deprivation. However, a recent study in our lab showed that during starvation-induced autophagy, it is a membrane population originated from TGN, but not the TGN membrane per se, that is tubulating to generate

Atg9+ vesicles.119 The more recent findings support a model that the majority of Atg9 are constitutively transported out of the TGN to a juxtanuclear Rab11-positive reservoir, where Atg9+ vesicles are generated during starvation; the delivery of Atg9 from TGN to

Rab11-positive compartment is independent of Bif-1, whereas the starvation-induced generation of Atg9+ vesicles from this endosomal compartment is tightly regulated by

Bif-1 and its newly identified interaction partner GTPase dynamin 2 (DNM2).

Moreover, p38-interacting protein (p38IP) was also shown to bind to Atg9 and positively influence its anterograde trafficking. While p38IP knockdown disrupts the redistribution of Atg9 and the autophagy induction by starvation, overexpression of p38IP also affects protein degradation, suggesting that a delicate balance of p38IP levels may be essential. Further, the interaction of p38IP with Atg9 is governed by p38α

MAPK, which competes with Atg9 for p38IP binding.140

Further, given that in yeast Atg18 is critical for Atg9 retrieval from PAS back to the peripheral pool, its mammalian homologue, WIPI2, has been examined for a potential role in Atg9 cycling. Depletion of WIPI2 causes Atg9 to be localized on PI3P-positive structures that are unable to mature,34 suggesting that WIPI2 may be important for retrieving Atg9 from autophagosome precursors. 33

Collectively, in both yeast and mammalian cells, a cascade of proteins regulates

Atg9 trafficking during autophagy, although the precise role of Atg9 in autophagosome formation awaits further characterization. In particular, what potential molecule(s) Atg9 may shuttle to or from the autophagosome in mammalian cells remains unclear, largely due to poor understanding of its interaction network. As will be discussed in detail in later chapters, the dissertation study identifies a putative Atg9 interaction partner – valosin containing protein (VCP). In the following section, I will briefly review its protein biology, cellular function, and disease implications.

Figure 1.4. Mammalian Atg9 membrane compartments. Under basal states, ATG9 localizes at the trans- Golgi network (TGN) and recycling endosomes (RE), as well as late endosomes (LE). Upon autophagy induction by starvation, ATG9 redistributes to a peripheral location, where it colocalizes with markers of phagophores, and interacts dynamically with forming autophagosomes. This anterograde trafficking of Atg9 is regulated by the ULK1 complex, the PI3K complex, Bif-1 and p38IP, whereas the retrograde trafficking is regulated by WIPI2. (Adapted from: Biology and trafficking of ATG9 and ATG16L1, two proteins that regulate autophagosome formation, FEBS Lett. 2013, with modifications.)

1.5. VCP as A Novel Regulator of Autophagy

1.5.1. VCP Structure and Function Overview

VCP, also called p97, is a member of the AAA ATPase (ATPases associated with diverse cellular activities) family that generally transfers the energy derived from ATP hydrolysis to structural remodeling of substrate molecules.141 VCP has two ATPase domains, D1 and D2, a long N-terminal domain interacting with many other proteins, and a short C-terminal tail.142 A collection of at least 30 cofactors of VCP has been identified and the interactions between VCP and these cofactors are key to its activity and functional diversity, although it is unclear if and how a particular type of interaction determines a specific role of VCP.143 These cofactors contain specific interaction domains or motifs that bind to VCP mostly at its N-terminal domain and some at the C- terminal tail. Some of these cofactors serve as ubiquitin adaptors or recruit VCP to intracellular membranes, and many others are orphan cofactors with unknown function.

VCP serves multiple cellular functions, with a common theme being to extract ubiquitylated proteins from membranes or cellular structures, or to segregate them from protein complexes. Its well-established role is to ensure protein quality control by facilitating the proteasomal degradation of various damaged or misfolded proteins in different compartments including the ER (termed ER-associated degradation or ERAD), the outer mitochondrial membrane, the nucleus, and the ribosome. In addition, VCP mediated degradation is also critical for signaling events.144 Typical examples include the degradation of IkBα (inhibitor of kappa B) leading to NFκB (nuclear factor κB) activation, and the degradation of HIF1α attenuating the response to hypoxia.145 36

VCP also serves non-degradative functions. For example, it activates transcription factors by extracting their precursors from membranes, such as for Spt23 that regulates the expression of the fatty acid desaturase Ole1 in yeast.146 Moreover, VCP plays critical roles in ensuring genomic stability via chromatin-associated functions. VCP dissociates proteins from chromatin, either for their degradation or for recycling, to modulate the dynamics of a growing list of chromatin regulators.144 Prominent examples are degradation of CDT1 (Chromatin Licensing and DNA Replication Factor 1) and stalled

RNA polymerase II in response to UV-induced DNA damage or degradation of DDB2

(DNA damage-binding protein 2) as part of nucleotide excision repair.147,148

Further, VCP has been associated with various membrane trafficking processes, such as Golgi reassembly following mitosis.144 Recently, it has also been shown to regulate lipid droplet biogenesis.149 Finally, emerging studies identify VCP as a new player in the regulation of the endosomal pathway as well as autophagy.150 The underlying mechanisms, particularly those in autophagy, are still unclear and controversial.151

1.5.2. Reported Roles of VCP in Autophagy

The links between VCP and autophagy were first revealed in studies of the VCP- associated disease IBMPFD (inclusion body myopathy with Paget's disease of bone and frontotemporal dementia). Patient’s muscle was characterized by the accumulation of vacuoles enriched in ubiquitin, LC3-II and p62.150 In accordance with this finding, accumulation of nondegradative autophagosomes was also observed in transgenic mice

expressing disease-associated mutants of VCP as well as in cultured cells with VCP mutation or deletion.152 Mechanistically, while some studies reported VCP is required for the fusion of autophagosomes with lysosomes,153 others concluded that VCP only acts downstream of the fusion as enlarged and acidified, Cathepsin B-positive autolysosomes were still able to form in the mutants.154 In contrast to these observations in mammals,

Krick et al. found that the yeast VCP orthologue Cdc48 and cofactor Shp1 are essential for autophagosomes formation and that the direct interaction of Shp1 with Atg8

(homologue of the mammalian LC3) is needed during macroautophagy.155 Currently, these different reports suggest various functions for VCP at different stages of the autophagy process. The discrepancies are possibly caused by lack of tools to accurately study autophagosome maturation, or due to different cell line model systems and VCP inhibition methods. Additionally, it is possible that VCP may play dual roles both upstream and downstream of autophagosome-lysosome fusion.

1.5.3. VCP and Human Diseases

In accordance with the essential role of VCP in maintaining cellular protein homeostasis, VCP- knockout is embryonic lethal to mice.156 Importantly, in humans, autosomal dominant mutations in VCP are associated with a rare multi-system degenerative disorder referred to as IBMPFD (inclusion body myopathy with Paget's disease of bone and frontotemporal dementia).157

It has been proposed that defective autophagy observed in VCP disease mutants may be responsible, at least partially, for the development of IBMPFD. VCP disease

mutations disrupt the consolidation of aggregate-prone proteins into inclusion bodies, accumulate non-degrading autophagosomes and thus impair the autophagic degradation of ubiquitylated proteins.150,153,154 Moreover, cells with mutant VCP have enlarged late endosomes with absent intraluminal vesicles (ILVs), implicating a defect in multivesicular body (MVB) biogenesis. Given that autosomal dominant mutations in the endosomal sorting complexes required for transport (ESCRT) component Chmp2B, a protein essential for MVB biogenesis, also cause frontotemporal dementia with ubiquitin inclusions (FTD-U),158,159 it is possible that impaired sorting in the endocytic pathway due to VCP mutation may also contribute to the neuronal symptoms in IBMPFD.

Chapter 2. ROLES OF BIF-1 IN OBESITY

2.1. Abstract

Bif-1 is a membrane-curvature inducing protein that is implicated in the regulation of autophagy and tumorigenesis. The dissertation study shows that Bif-1 plays a critical role in regulating lipid catabolism to control the size of lipid droplets and prevent the development of obesity and insulin resistance upon aging or dietary challenge. Our data show that Bif-1 deficiency promotes the expansion of adipose tissue mass without altering food intake or physical activities. While Bif-1 is dispensable for adipose tissue development, its deficiency reduces the basal rate of adipose tissue lipolysis and results in adipocyte hypertrophy upon aging. The importance of Bif-1 in lipid turnover is not limited to adipose tissue since the lipid droplet clearance is also attenuated by Bif-1 loss in the liver of mice starved and red-fed. Interestingly, obesity induced by a high fat-diet or Bif-1 deficiency downregulates the expression of proteins involved in the autophagy-lysosomal pathway, including Atg9a and Lamp1 in the adipose tissue. These findings thus identify Bif-1 as a novel regulator of lipid homeostasis to prevent the pathogenesis of obesity and its associated metabolic complications.

2.2. Results

2.2.1. Ablation of Bif-1 Promotes Aging- and Diet-induced Weight Gain and Fat Accumulation.

To assess whether Bif-1 is critical in regulating obesity, we placed wild-type

(WT) and Bif-1 KO mice of the C57BL/6 background on a high fat diet (HFD) or a control/standard chow diet (CD) starting at six weeks of age, at which time the mice were indistinguishable by weight. Bif-1 KO mice gained more weight upon aging than their

WT littermates (Fig. 2.1a; Fig. 2.2a), which was notably exacerbated with the HFD challenge (Female WT: 34.83±1.36 g, Bif-1 KO: 44.83±1.87 g, Figs.1a, 1b; Male WT:

40.44±2.03 g, Bif-1 KO: 51.14±0.76 g, Figs. 2.1a; Fig. 2.2a). Significant weight gain occurred as early as one week after HFD feeding in males and eight weeks in females, which is in line with the previous observation that female mice are less sensitive to HFD- induced obesity160.

To determine whether the higher body weight gain in Bif-1 KO mice correlates with increased fat accumulation, we analyzed the full body composition of these mice at study end-points (19 and 23 weeks after the initiation of dietary challenge for males and females, respectively). Bif-1 KO mice showed a significant increase in the absolute and relative amount of body fat compared to WT mice regardless of diet (Figs. 2.1c, 2.1d;

Fig. 2.2b, 2.2c). In contrast, lean mass (Fig. 2.1e, Fig.2.2d) and body water content (data not shown) were comparable between the WT and Bif-1 KO mice fed with CD. While a moderate increase in lean mass was observed in male Bif-1 KO mice after prolonged

HFD challenge (Fig. 2.2d), the increase was not nearly as significant as that observed in

fat mass. In agreement with these observations, we observed significantly larger abdominal gonadal and subcutaneous fat pads (Figs. 2.1f, 2.1g) as well as enlargement of white adipocytes (Fig. 2.1h) in Bif-1 KO mice. Similarly, the brown adipose tissue of

Bif-1 KO mice also appeared to contain more fat than their WT littermates (Fig. 2.1i).

We did not observe apparent ectopic fat accumulation in peripheral tissues such as liver and muscle within the time span of the current study. Taken together, these results indicate that Bif-1 deficiency increases adiposity upon aging and promotes obesity after a

HFD challenge.

Figure 2.1 Loss of Bif-1 induces obesity. (a) Female wild type (WT) and Bif-1 knockout (KO) mice were fed with either control chow diet (CD) or high-fat diet (HFD) for the indicated periods of time. Body weight of each mouse was monitored weekly (n=9-12). Statistical significance was determined using one- way ANOVA followed by Bonferroni’s multiple comparison test. (b) Representative images of WT and Bif-1 KO mice fed on HFD for 23 weeks. (c-e) Body composition of mice fed with the indicated diet for 23 weeks was measured using TD-NMR. Loss of Bif-1 leads to larger total (c) and percent (d) fat mass and has no effect on lean mass (e). Statistical significance was determined using one-way ANOVA followed by Bonferroni’s multiple comparison. (f) Representative images of fat deposits in WT and Bif-1 KO mice fed on HFD for 23 weeks. (g) Representative images of abdominal gonadal fat pads from WT and Bif-1 KO mice fed on either CD or HFD for 23 weeks. (h, i) Loss of Bif-1 leads to adipocyte enlargement. 44

Hematoxylin and eosin stain of the abdominal gonadal white adipose tissue (WAT) (h) and interscapular brown adipose tissue (i) sections from mice fed with CD or HFD for 23 weeks. Scale bars in (h) and (i) represent 50μm and 200μm, respectively. All values are mean ± SEM. Differences with controls were significant for *p < 0.05, **p < 0.01, and ***p < 0.001.

Figure 2.2. Metabolic characteristics of male mice fed with indicated diet. (a) Mice were fed with HFD or control CD starting at 6-week of age and body weight of mice was monitored for 15 weeks. (b-d) Loss of Bif-1 leads to larger total (b) and percent (c) fat mass and has minimal effect on lean mass (d). Body composition of mice fed with the indicated diet for 19 weeks was measured using TD-NMR. Statistical significance was determined using one-way ANOVA followed by Bonferroni’s multiple comparison test. Oral glucose tolerance test was performed in 16 h-fasted mice, n=9-11. All values are mean ± SEM. Differences with controls were significant for *p < 0.05, **p < 0.01, and ***p < 0.001.

2.2.2. Bif-1 KO Mice Develop Insulin Resistance Following Excessive Adiposity.

Obesity causes insulin resistance63,161,162; therefore higher adiposity in Bif-1 KO mice suggests that these animals may be less responsive to insulin. To investigate the possibility of such metabolic complications, we first measured the circulating insulin levels in the mice. We found that Bif-1 KO mice fed with CD had normal insulin levels at

10 weeks of age at which time their body weight had not yet diverged from WT littermates (Fig. 2.3a). However, upon weight augmentation by aging or diet challenge

(30 weeks old), the Bif-1 KO mice showed hyperinsulinemia (Figs. 2.4a, 2.4b) and similar circulating glucose levels compared to the WT mice (data not shown), indicating that the obese Bif-1 KO mice had lower insulin sensitivity. Consistently, Bif-1 KO mice displayed intensified insulin resistance after HFD challenge compared to the WT mice

(Fig. 2.4c), although glucose tolerance was not significantly affected (Fig. 2.4d).

Moreover, Bif-1 KO mice (30 weeks old, CD fed) displayed dampened insulin signaling in the liver compared to WT mice, as determined by a moderate decrease in phosphorylated-Akt (p-Akt) in response to insulin injection (Fig. 2.4e). While we were unable to detect p-Akt in the muscle lysates collected from the insulin-treated mice, we used the phosphorylation of ERK (p-ERK) as an alternative means for detecting insulin signaling. While the WT muscle displayed elevated levels of p-ERK following insulin injection, the Bif-1 KO muscle did not show such induction of signaling (Fig. 2.4f).

Notably, we also observed hyperleptinemia in Bif-1 KO mice upon the development of obesity (Fig. 2.3b, Figs. 2.4g, 2.4h), which further demonstrates that Bif-1 is important for the suppression of obesity-associated metabolic complications. 46

Figure 2.3. Plasma insulin and leptin levels in 10-week-old mice. Levels of insulin (a) and leptin (b) in plasma from 10-week-old mice fed with CD were measured by ELISA assays, n=1-2.

Figure 2.4. Obesity induced by Bif-1 loss leads to insulin resistance. (a, b) Levels of insulin in plasma from mice fed with CD (a) or HFD (b) for 23 weeks were measured using BioRad Luminex assays, n=9- 12. Statistical significance was determined using Student’s t-test. (c, d) Insulin tolerance test (c) and oral glucose tolerance test in mice fed with HFD for 23 weeks were performed as described in Methods. n=9-

11. Statistical significance was determined using Student’s t-test. (e, f) WT and Bif-1 KO mice fed with CD for 23 weeks were fasted for 16 h and then given 0.75 U/kg of insulin intraperitoneally 15 min prior to sacrifice. Liver (e) and gastric muscle (f) lysates were prepared and subjected to immunoblot analysis using the indicated antibodies. The intensities of phosphorylated-Akt (pAkt S473) or phosphorylated-ERK1/2 (p-

ERK) relative to total Akt or total ERK1/2 were normalized against β-actin. Statistical significance was determined using Student’s t-test in (e) and One-way ANOVA followed by Bonferroni’s multiple comparison test in (f). (g, h) Levels of leptin in plasma from mice fed with CD (g) or HFD (h) for 23 weeks were measured using BioRad Luminex assays, n=9-12. Statistical significance was determined using Student’s t-test. All values are mean ± SEM. Differences with controls were significant for *p < 0.05.

2.2.3. Bif-1 Does Not Regulate Food Intake, Physical Activities Or Fuel Preference.

Weight gain is a result of imbalance in energy intake versus expenditure; thus we analyzed the food intake and energy metabolism of the WT and Bif-1 KO mice using metabolic cages. In order to eliminate the confounding effects due to obesity (e.g. hyperleptinemia), we analyzed WT and Bif-1 KO mice both before and after their body weights diverged. Food intake and physical activity were comparable between WT and

Bif-1 KO mice at a young age (10 weeks old, CD fed) (Figs. 2.5a, 2.5b), suggesting that these two factors are not the cause of increased obesity in Bif-1 KO mice. Upon weight augmentation by aging (27 weeks old), the total amount of food intake in CD-fed Bif-1

KO mice was lower than that in the WT littermates (Fig. 2.5c). This is in agreement with a previous finding that augmentation of fat mass leads to a graded decrease in voluntary feeding as a feedback mechanism of the body to favor a reduction in fat stores163.

Moreover, the 27-week-old Bif-1 KO mice fed CD were slightly more physically active

during the day relative to the nighttime compared to the WT, suggesting a mild impairment of circadian rhythms (Fig. 2.5d).

As Bif-1 is involved in the maintenance of mitochondrial integrity, we next asked whether Bif-1 deficiency compromises the mitochondrial fatty acid oxidation. To answer this question, we analyzed the respiratory exchange ratio (RER) in 10-week-old WT and

Bif-1 KO mice. RER is the ratio between the amount of oxygen (O2) consumed and carbon dioxide (CO2) produced, and it is an indicator of which energetic substrate is being oxidized to supply the body with energy164. RER is 1.0 for carbohydrates and 0.7 for lipids. As shown in Figure 2.6, both the WT and Bif-1 KO mice had RER values slightly above 0.8 during the day, which equally increased to close to 1 during the night.

These results indicate that Bif-1 deficiency does not affect the animals’ fatty acid oxidation, and WT and Bif-1 KO mice can oxidize a mix of carbohydrates and fat with similar efficiency. Consequently, the WT and Bif-1 KO mice had comparable energy expenditure at 10 weeks of age (Fig. 2.5e). Finally, aging or long-term HFD feeding significantly reduced the energy expenditure of Bif-1 KO mice compared to WT HFD controls (Fig. 2.5f), thereby indicating a positive energy balance and excessive energy store in the animals at the obese stage.

Figure 2.5. Effects of Bif-1 on food intake, physical activities, and fuel preference. Indirect calorimetry analysis of WT and Bif-1 KO mice fed with CD for 4 weeks (10-week old, n=4) (a, b, e) or with the indicated diet for 19 weeks (25-week old, n=8) (c, d, f). Food intake (a, c), physical activities (b, d), and energy expenditure (e, f) were measured during a 48-h cycle using metabolic cages. Statistical significance was determined using one-way ANOVA followed by Bonferroni’s multiple comparison. All values are mean ± SEM. Differences with controls were significant for *p < 0.05 and ****p < 0.0001. 51

Figure 2.6. Respiratory exchange ratio (RER) of WT and Bif-1 KO mice. (a, b) RER values for male (a) and female (b) mice were calculated using VCO2 and VO2 measured during a 48 h cycle (n=4). All values are mean ± SEM.

2.2.4. Development Of Obesity With Bif-1 Deficiency Is Independent Of Adipogenesis.

Obesity is characterized by an expansion of adipose tissue mass. To determine the potential importance of Bif-1 in adipogenesis, we first examined the protein expression of

peroxisome-proliferator activated receptor gamma (PPARγ), a master regulator of adipogenesis, in the white adipose tissue (WAT) of 6-week-old WT and Bif-1 KO mice.

However, we did not observe any apparent difference between the two genotypes under both basal and nutrient-deprived states (Fig. 2.7a). Consistently, an in-vitro adipogenesis assay using immortalized peri-vascular cells isolated from the brown adipose tissue of the

WT and Bif-1 KO mice did not reveal a significant difference in adipogenesis, as indicated by comparable intensities of Oil Red O stain and protein expression levels of adipogenesis markers after stimulation with an adipogenic cocktail (Figs. 2.7b, 2.7c).

This similarity in the adipogenic potential between the WT and Bif-1 KO peri-vascular cells was also confirmed using 3T3-L1 cells (Figs. 2.7d, 2.7e). Interestingly, Bif-1 expression was dramatically increased upon the induction of PPARγ expression, which is in line with a recent report showing that Bif-1 is one of the target genes of PPARγ165.

Taken together, our results suggest that obesity induced by the loss of Bif-1 is not likely due to the promotion of adipogenesis.

Figure 2.7. Bif-1 deficiency does not affect adipogenesis. (a) Bif-1 does not impact adipogenesis in vivo. Relative protein expression levels of PPAR-γ in WT and Bif-1 KO WAT homogenates prepared from 6 week-old mice were determined via immunoblotting, (n=3). (b) Immortalized perivascular cells that were isolated from brown adipose tissue (iPBA) of 6-week-old mice as described in Methods were induced to differentiate with an adipogenic cocktail. Cells were subjected to Oil Red O (ORO) staining on indicated days following adipogenesis induction. Representative images were acquired, and the optical density of ORO extracted from cells was quantified by absorption at 520nm. (n=4) (c) Lysates prepared from iPBA cells harvested on indicated days during adipogenesis were subjected to immunoblotting with indicated antibodies. (d) 3T3-L1 cells transduced with lentiviruses encoding Bif-1 shRNA (ShBif-1) or scrambled shRNA (shScr) were induced to differentiate as described in Methods, and cells on indicated days during adipogenesis were stained with ORO. Representative images were acquired, and the optical density of ORO extracted from cells was quantified by absorption at 520nm. (n=4) Statistical significance was determined using Student’s t-test. All values are mean ± SEM. Differences with controls were significant for **p < 0.01. (e) Lysates prepared from 3T3-L1 shScr and shBif-1 cells harvested on indicated days during adipogenesis were subjected to immunoblotting with indicated antibodies.

2.2.5. Bif-1 Loss Impairs Lipid Turnover In Adipose Tissue And Liver.

As Bif-1 deficiency does not enhance adipogenesis, we next determined whether

Bif-1 is important for lipolysis or the mobilization of lipids from WAT. At the resting state, WAT triacylglycerol (TAG) is in a constant state of flux as a result of a largely futile cycle of lipolysis and fatty acid re-esterification166. During times of energy deprivation, the WAT enhances the rate of lipolysis to hydrolyze TAG to free fatty acids

(FFAs) and glycerol that are released into the circulation and taken up by other organs as energy substrates. The best-known mechanism of fasting-induced lipolysis involves the activation of PKA upon the binding of catecholamines to the β2-adrenergic receptors, which in turn phosphorylates and activates hormone sensitive lipase (HSL). To examine 55

whether Bif-1 regulates lipolysis, we measured the FFA levels in the plasma of the 30- week-old mice under basal and fasting states. Interestingly, loss of Bif-1 appeared to cause a slight decrease in plasma FFA levels during resting state but not after fasting or a long term HFD challenge (Fig. 2.8a). This decrease in FFA is well correlated with a reduction in the protein expression levels of phosphorylated-HSL (p-HSL) in the Bif-1

KO mice under basal state (Fig. 2.8b). This suggests that Bif-1 may contribute to basal lipolysis but is likely not important for the fasting-induced lipolysis. Therefore, Bif-1 KO mice could provide a sufficient amount of FFAs to meet the energy demand during short- term fasting. Importantly, we observed a moderate decrease in the plasma FFA levels and protein levels of p-HSL in young animals when their body weights had not diverged

(Figs. 2.8c, 2.8d, Fig. 2.9), suggesting this impairment in basal lipolysis is not caused by development of obesity.

In addition to the examination of lipolysis in WAT, we also determined whether

Bif-1 is important for lipid breakdown in the liver, which is the major site of lipid synthesis and temporary lipid storage. We subjected the mice to overnight fasting and then re-fed the animals for another day. During fasting, FFAs generated by adipose tissue lipolysis are transported to the liver where they are re-esterified to form triglycerides, which are then degraded upon re-feeding. While Bif-1 deficiency did not affect the initial accumulation of hepatic lipid droplets upon fasting, lipid droplets were significantly increased in number and size in Bif-1 KO mice following re-feeding (Figs. 2.8e, 2.8f), suggesting that Bif-1 is important for hepatic lipolysis in the presence of excessive lipids.

Taken together, these results suggest that Bif-1 regulates lipid catabolism not only in the adipose tissue, but also in other peripheral tissues.

Figure 2.8. Bif-1 regulates lipid turn over in adipose tissue and liver. (a) Plasma levels of free fatty acid (FFA) from mice fed with the indicated diet for 23 weeks in resting and 16 h-fasted states were measured (n=4). (b) Lysates prepared from white adipose tissue collected from the same animals in (a) were subjected to immunoblotting with indicated antibodies. Same lysates were also used in Figure 2.10b. (c) Plasma levels of free fatty acid (FFA) from 6-week old mice fed with normal CD were measured in the same way as in (a). Statistical significance was determined using Student’s t-test. (d) Lysates prepared from white adipose tissue collected from the same animals in (c) were subjected to immunoblotting with indicated antibodies. Same lysates were also used in Figure 2.7a. (e, f) Representative images (e) and quantification (f) of ORO stain of the liver sections from mice fed with CD for 20 weeks fasted overnight or fasted overnight and then re-fed for 24 h. Scale bars represent 100 µm.

Figure 2.9. Body weight and body fat percentage in 6-week-old mice. (a, b) Body weight (a) and fat percentage over body weight (b) in control diet-fed, 6-week-old mice.

2.2.6. Bif-1 Deficiency Reduces Autophagic Flux And Downregulates Autophagic- Lysosomal Proteins.

It has been recently discovered that autophagy can directly degrade lipid droplets in hepatocytes. Since then, a plethora of other studies have also identified lipophagy as a conserved mechanism to mobilize cellular fat stores in diverse cell types65,92,94,96.

However, the role of autophagy in adipose tissue remains unclear.

In the present study, we first evaluated the expression levels of p62 and LC3-II, which are widely used markers for autophagy167. During autophagy p62 serves to link autophagy substrates to autophagosomes and is therefore degraded also during autophagic flux167. LC3-II marks autophagosomal membrane turnover and it is generated through the conjugation of its precursor, LC3-I, to the membrane lipid phosphatidylethanolamine via a series of ubiquitin-like conjugation enzymes1. In our study we observed that compared to WT, Bif-1 deficient adipose tissue had slightly higher levels of p62 when fed CD under both basal and fasting states (Figs. 2.10a, 2.10b,

Fig. 2.11), indicating a decrease in autophagic flux. A two-fold increase in the LC3-

II/LC3 ratio was detected in WT adipose tissue upon fasting, while the Bif-1-deficient adipose tissue showed a minor elevation in LC3-II/LC3 ratio, indicating a dampened autophagic response. Interestingly, we found that in WT animals a long-term HFD challenge significantly suppressed the expression of Bif-1 and Atg9a, which is another essential autophagy protein (Fig. 2.10a, Fig. 2.11). This Atg9a reduction was phenocopied by the CD-fed Bif-1 KO mice (Figs. 2.10a, 2.10b, Fig. 2.11). Similarly, the expression of lysosomal-associated membrane protein 1 (Lamp1) was also reduced in

Bif-1 KO adipose tissue compared to WT regardless of diet or nutrition status (Figs.

2.10a, 2.10b, Fig. 2.11). As the downregulation of Atg9 and Lamp1 in Bif-1 KO adipose tissue was only observed in 30-week-old but not 6-week- or 10-week-old animals (Fig.

2.10d, Fig. 2.11), this effect appears to be consequences of the development of obesity in

Bif-1 KO mice. The expression levels of other Atg proteins, including Beclin 1 and the

Atg12-Atg5 conjugate, did not appear to be affected by Bif-1 expression (Fig. 2.10a).

Interestingly, expression of perilipin A (Plin1), which is a lipid droplet coating protein, is also slightly higher under both basal and fasting states in the WAT of Bif-1 KO mice, which correlates with the reduced TAG hydrolysis observed in these animals (Fig.

2.10a). To further assess the potential role of Bif-1 and autophagy in the regulation of

Plin1 in adipose tissue, a portion of fasted WT and Bif-1 KO mice were injected with insulin to activate mTOR signaling and inhibit autophagy168. Insulin administration led to the accumulation of p62 in WT adipose tissue, indicating rapid autophagy flux in WAT following fasting, and potent suppression of autophagy by insulin (Fig. 2.10c, Fig. 2.11).

Notably, Plin1 also accumulated upon insulin treatment in WT adipose tissue, indicating a role for autophagy in Plin1 degradation. In contrast, the level of p62 and Plin1 in fasted

Bif-1 KO mice was not further increased by insulin (Fig. 2.10c, Fig. 2.11), suggesting an impairment of autophagy under fasting conditions in these mice.

Figure 2.10. Bif-1 is important for autophagy in adipose tissue. (a, b, c) Immunoblots for indicated proteins in WAT homogenates prepared from 30-week-old mice. The mice were fed with CD or HFD for 23 weeks (a), fed with CD, at basal or 16 h-fasted states (b), or fasted for 16 h, with or without insulin injection 15 min prior to sacrifice (c). n=3-4. (d) Immunoblots for indicated proteins in WAT homogenates from 6-week old mice fed with CD or 11-week old mice fed with 5 weeks HFD. n=3.

Figure 2.11. Quantifications for immunoblots in Figure 2.10. (a-c) Quantification for immunoblots in Figure 2.10a. (d-e) Quantification for immunoblots in Figure 2.10b. (f-h) Quantification for immunoblots in Figure 2.10c. (j, k) Quantification for immunoblots in Figure 2.10d. Image J software was used in quantifications and expression of indicated proteins was normalized to β-actin. Statistical significance was determined using Student’s t-test for single comparison and One-way ANOVA for multiple comparison followed by Bonferroni’s or Dunn’s multiple comparison test. n=3-4. All values are mean ± SEM. Differences with controls were significant for *p <0.05, **p <0.01, and ***p <0.001.

2.3. Discussion

Autophagy is a critical housekeeping pathway that controls cellular and organismal metabolism by degrading proteins, lipids and organelles169. This catabolic process is suppressed in adipocytes in human and rodent obesity170. However, the role of autophagy in adipose tissue fat catabolism remains unclear. In this study, we show that

Bif-1 deficiency induces adipocyte hypertrophy without altering food intake and physical activity and promotes the development of obesity and insulin resistance upon aging or dietary challenge in mice. The observations that Bif-1 loss decreases the levels of phosphorylated HSL and plasma concentration of glycerol under fed condition suggest an important role of autophagy in the promotion of basal lipolysis in the WAT. Interestingly, unlike fed condition, Bif-1 deficiency does not block adipose tissue lipolysis and LD formation in the liver during fasting, suggesting that autophagy may be dispensable for hormone-stimulated lipolysis. This observation is consistent with a recent report showing that lysosomal inhibition has a minimal effect on the β-adrenergic agonist isoproterenol- stimulated lipolysis in vitro170. Therefore, Bif-1 regulates basal adipose tissue lipolysis through a mechanism that is independent of the catecholamine-induced canonical lipolytic pathway. Since the hepatic lipid droplet clearance in mice starved and re-fed was also suppressed by Bif-1 deficiency, the importance of Bif-1-mediated lipid turnover may not be limited to adipose tissue.

While lipid droplet-coating Plin1 limits the access of HSL to prevent triacylglycerol hydrolysis under basal conditions88, its degradation facilitates lipolysis171.

Our data show that Bif-1 deficiency impairs turnover of both Plin1 and autophagic 63

substrate p62 and accumulates poly-ubiquitinated proteins in WAT. This observation is in agreement with previous finding that autophagy inhibition compromises the ubiquitin- proteasome system172. Since the degradation of Plin1 is mediated through the ubiquitin- proteasome173 and the lysosomal171 pathways, inhibition of overall Plin1 degradation could be the mechanism by which Bif-1 loss attenuates basal lipolytic activities in adipose tissue. In addition, it is possible that Bif-1 may regulate adipose tissue lipid catabolism by promoting lipophagy; whereby portions or entire LDs are engulfed into autophagosomes and delivered to lysosomes for degradation81. Indeed, our data show that the induction of autophagy by fasting promotes the degradation of Plin1 in a Bif-1- dependent manner. However, it is unlikely that lipophagy plays a major role in adipose tissue lipid catabolism since, as described above, Bif-1 deficiency does not inhibit adipose tissue lipolysis upon fasting. Moreover, in contrast to basal condition88, Plin1 has been shown to facilitate lipolysis by recruiting HSL on the surface of LDs under the exposure to β-adrenergic agonist isoproterenol174, which mimics fasting condition.

Recently, chaperone-mediated autophagy (CMA) has been shown to selectively degrade

Plin2 and Plin3 to control the access of adipose tissue triglyceride lipase and several key autophagy-related proteins in fibroblasts175. It would be interesting to determine if the

Bif-1-mediated autophagic pathway collaborates with CMA to regulate basal lipolysis in adipose tissue.

Our data also show that the expression levels of Atg9a and Lamp1 in WAT are reduced by the loss of Bif-1 upon aging or dietary challenge. Downregulation of these two critical proteins for autophagy could in turn undermine the autophagic activity on LD 64

and promote further fat accumulation. Although further studies are required to understand the mechanism responsible for the downregulation of these autophagy-lysosomal proteins, recent studies have shown that Lamp1 and Atg9 genes are targets of the transcription factor EB (TFEB)82 that is negatively regulated by mTOR176. Therefore obesity-attributed hyperinsulinemia in Bif-1 KO animals may mediate the downregulation of Atg9 and Lamp1 through the activation of mTOR and suppression of

TFEB signaling, and this hypothesis warrants further investigation in the future.

In summary, the current work demonstrates the importance of Bif-1 in the control of lipid catabolism and whole body energy and nutrient homeostasis to prevent the development of obesity. As the expression of Bif-1 is upregulated during fat development and then declined after full adipogenesis, Bif-1-deficient mouse may be a suitable model to study the roles of adipose tissue autophagy in the progress from obesity to obesity- associated metabolic complications.

Figure 2.12. Proposed model of Bif-1 function in lipid catabolism. Bif-1 plays a critical role in regulating lipid catabolism to control the size of lipid droplets and prevent the development of obesity. Bif- 1 loss disrupts autophagic degradation of lipids in adipose tissue (lipophagy), indirectly reduces the basal rate of cytosolic lipolysis, thus impairs overall triglyceride (TAG) hydrolysis and results in adipocyte hypertrophy. Interestingly, Bif-1 deficiency following development of obesity downregulates the levels of proteins involved in the autophagy-lysosomal pathway, including Atg9a and Lamp1 in the adipose tissue. This could be mediated by hyperinsulinemia and the downstream signals of insulin - mTOR and TEFB. Reduction in Atg9a and Lamp1 could impair autophagy and promote further fat accumulation.

2.4. Material and Methods

2.4.1. Reagents

The following antibodies were used in this study: polyclonal goat anti-Bif-1

(GTX 21343, GeneTex); poly-clonal guinea pig anti-P62 (03-GP62-C, American

Research Products); monoclonal rabbit anti-Plin1 (9349, Cell Signaling); polyclonal rabbit anti-Atg9a (GTX128427, Genetex), monoclonal rat anti-Lamp1 (19992, Santa

Cruz); polyclonal rabbit anti-phospho-mTOR (S2448) (2971, Cell Signaling); monoclonal rabbit anti-mTOR (2983, Cell Signaling); monoclonal rabbit anti-phospho-

Akt (S473) (9271, Cell Signaling); monoclonal mouse anti-Akt (MAB2055, R&D

Systems); monoclonal mouse anti-β actin (A5441, Sigma). Mouse leptin ELISA kit and mouse insulin ELISA kit were purchased from Millipore. Mouse diabetes Luminex multiplex assay kit was purchased from Bio-Rad. Fatty acid kit was purchased from

Cayman Chemicals. Unless otherwise stated, all other chemicals were purchased from

Sigma.

2.4.2. Cell Lines and Culture Conditions

Human cervical cancer cell lines HeLa and HeLa-S3, human embryonic kidney cell line HEK293T, immortalized mouse embryonic fibroblasts MEF, and human hepatic adenocarcinoma cell line LH86, were maintained in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% fetal bovine serum (FBS). Immortalized 3T3-

L1 mouse fibroblasts were maintained in DMEM supplemented with 10% bovine calf serum (BCS). In all cases, cells were grown at 37°C with 5% CO2.

2.4.3. Isolation and Immortalization of Perivascular Cells from Brown Adipose Tissue

Perivascular cells were isolated from brown adipose tissue of 10-week-old WT and Bif-1 KO mice fed with CD. Briefly, interscapular brown fat pad was isolated from mice, dissected and minced in PBS. Minced tissue was then pipetted into isolation buffer containing 123.0 mM NaCl, 5.0 mM KCl, 1.3 mM CaCl2, 5.0 mM Glucose, 100 mM

HEPES, 1% pen/strep, 4% BSA, and 1.5 mg/ml type I collagenase. The mixture was then vortexed for 10 seconds, and incubated at 37 degrees for 40 min with gentle shaking.

Digested tissue was then filtered using 100 µm Nylon cell strainers into clean Eppendorf tubes, and centrifuged at 1500 rpm for 5 min. Pellet was re-suspended in culture medium

(DMEM medium supplemented with 20% fetal bovine serum (FBS), 20 mM HEPES, and

1% pen/strep), and cultured overnight to allow cells to attach. Primary cells were then immortalized with SV40 large T antigen and used for adipogenesis experiments.

2.4.4. In Vitro Adipogenesis

For immortalized peri-vascular cells, cells were seeded at 5000 cells/cm2 in culture dishes and grown to confluence in culture medium. Confluent cells were then switched to differentiation media containing 20% FBS, 20 nM insulin, and 1 nM triiodothyronine for another 48 hours. Adipogenesis was induced by treating cells for 48 h in differentiation medium that was further supplemented with 0.5 µM dexamethasone,

0.5 mM isobutylmethylxanthine (IBMX), and 0.125 mM indomethacin. After induction, cells were returned to differentiation medium, which was replenished every 2 days.

For 3T3-L1, cells were transduced with lentiviruses encoding Bif-1 shRNA

(shBif-1) or scrambled control shRNA (shScr) and induced to differentiate with a standard adipogenic cocktail. Briefly, cells were seeded at 5000 cells/cm2 in culture dishes and grown to confluence in culture medium (DMEM medium supplemented with

10% bovine calf serum and 1% pen/strep). Cells were then cultured in fresh culture medium for another 2 days. At two days post-confluence (designated day 0), cells were induced to differentiate in induction medium (DMEM supplemented with 10% FBS, 1% pen/strep, 1 μM dexamethasone, 0.5 mM IBMX, and 10 μg/ml insulin). After 3 days induction, the media were replaced with maintenance medium (induction medium minus

IBMX) that was replenished every 2 days during the rest of differentiation time course.

2.4.5. Conjugation of Oleate to BSA and Conjugates Treatment Conditions

To make 24% fatty-acid-free BSA, 12 g of FAF-BSA (2g per aliquot) was dissolved into 35mL of 150mM NaCl at room temp with stirring over 3 hours, and pH

was adjusted to 7.4 carefully with 5N NaOH. Final volume was brought o 50mL with

150mM NaCl. 95 mg of Sodium Oleate (Sigma) was dissolved in 2mL EtOH in a glass flask. 100uL of 5N NaOH was added to the flask, and mixed thoroughly. EtOH was then by evaporated under a stream of nitrogen gas. 10mL of 150mM NaCl was then to the dried sodium oleate. The solution mixture was heated at 60 degree in water bath for 10 min, and then placed on a stirring plate. With stirring while the solution is still warm,

12.5mL of ice-cold 24% FAF-BSA was added to the solution. The mixture was stirred for

10 min, and the final volume was brought to 25mL with 150mM NaCl. This final solution was aliquoted and stored at -20 degree. Final stock concentration of oleate is

12.5mM.

For oleate-BSA feeding assays, cells were seeded 24 h ahead of treatment, and then challenged with various doses of oleate-BSA for different durations. For feeding plus starvation experiments, cells were first fed with oleate-BSA for 24 h, and then incubated in DMEM supplemented with 0.1% FBS for 0, 24, 48, and 72 h.

2.4.6. Lipid Droplet Visualization and Quantifications by Oil Red O and Bodipy Stains

ORO staining solutions were prepared following standard protocols. Briefly, for staining cells, a 3.5 mg/ml stock solution was made by dissolving ORO powder in isopropanol. Immediately before stain, fresh working solution was prepared by mixing 6 parts of stock solution with 4 parts of distilled H2O. The mixture was incubated at room temperature for 20 min and then filtered through 0.2 µm filters. Cells were fixed with 4% paraformaldehyde for 10 min, and washed with 60% isopropanol for 5 min x 2 times. 69

Cells were briefly air-dried, and stained with ORO working solution for 30 min at room temperature. Cells were washed with PBS for 5 min x 3 times and pictures of stained dishes were acquired using a digital camera. To quantify stain, ORO in cells was extracted with 4% NP-40 in isopropanol and subjected to optical density measurement at

520nm. For staining liver tissue, a 2.5 mg/ml ORO stock solution was made in isopropanol. Immediately before tissue staining, working solutions were prepared by mixing 6 parts of stock solution and 4 parts of distilled H2O. The mixture was incubated at 4 degree for 5 min, and filtered through 0.45 µm filters. 12-μm-thick liver sections from frozen tissue were fixed with 4% paraformaldehyde, and stained with the ORO working solution for 30 min at room temperature. Bright field microscopic images were acquired, and quantification of ORO stain was performed using Image J software with a

Threshold Color plugin. Briefly, images were converted to 8-bit type with default color thresholding. Total area (pixel2) of ORO-positive particles was then analyzed for the entire image. Images for at least two mice per condition were used for quantification.

2.4.7. Extraction of proteins from cell culture and mice tissues

For in-vitro cultures, cells were placed on ice, rinsed once with ice-cold PBS, and scraped off culture dishes. Cell suspension was then transferred to 15 ml tubes and centrifuged at 4°C, 1000x g for 5min. The supernatant was aspirated off and the cells were resuspended in 1ml ice-cold PBS, and transferred to eppendorf tubes. Cell suspension was then centrifuged at 4°C, 3000 RPM for 5min and the supernatant was aspirated off. The cell pellets were then resuspended in 3-4 times volume of RIPA lysis

buffer supplemented with protease inhibitor (1:500) and lyzed 30 min on ice. Lysates were centrifuged at 4°C 15000 RPM for 15 min to spin down insoluble components.

Supernatant was removed and transferred to new eppendorf tubes.

To extract protein from isolated mice livers, the liver tissue was homogenized with a hand-held polytron homogenizer in 3-4 times volume of RIPA lysis buffer supplemented with protease inhibitor. Homogenates were placed on ice for 30 min with gentle tapping every 10 min to allow complete lysis. Lysates were then centrifuged at

4°C 15000 RPM for 15 min to get rid of insoluble component and to collect supernatant.

For muscle tissues, big chunks of the isolated muscles were first chopped to fine pieces, and were homogenized in the same way as in liver protein extraction. To extract protein from white adipose tissue and brown adipose tissue, 3 times volume of tissue homogenization buffer was added to the fat tissue placed in a glass dounce homogenizer, and grinded with 30 strokes. Homogenates were transferred to an eppendorf tube and lysed on ice for 30 min with gentle tapping every 10 min to allow complete lysis. WAT and BAT samples were ready to use without additional centrifugation. Protein concentrations were determined using BCA protein assay (Thermo Scientific).

2.4.8. Determination of Free Fatty Acid, Triglyceride and glycerol Concentration in Plasma

Blood was collected from mice through cardiac puncture immediately after euthanization, and centrifuged at 1300 g at room temperature to obtain plasma. Plasma levels of fatty acid were then measured using fatty acid kit purchased from Cayman

Chemicals. Triglyceride and glycerol concentrations were determined using triglyceride reagent and glycerol reagent respectively purchased from Sigma Aldrich.

2.4.9. Glucose Tolerance Test (GTT) and Insulin Tolerance Test (ITT)

GTT and ITT were performed in overnight and 9 h-fasted mice, respectively.

Glucose was delivered by oral gavage at 2.5 g/kg body weight after initial measurement of fasting blood glucose. Insulin was delivered by intraperitoneal injection (0.75 U/kg body weight; Novolin, Novo Nordisk). Blood glucose was determined at 0, 30, 60, and

120 min after the glucose or insulin administration with an Embrace blood glucose meter

(Omnis Health).

2.4.10. Body Composition Measurement

Fat and lean body mass of each mouse was analyzed on an LF90 TD-NMR

(Bruker Optics).

2.4.11. Indirect Calorimetric Measurement

For measuring metabolic parameters, mice were acclimated in the metabolic chambers (TSE Systems) for 2 days, and food and water intake, energy expenditure, and physical activities were measured for another 2 consecutive days. Constant airflow (0.2 l/min) was drawn through the chamber and the concentrations of oxygen and carbon dioxide were monitored at the inlet and outlet of the sealed chambers to calculate oxygen consumption. Each chamber was measured for 3.75 min at 30 min intervals. Physical

activity was monitored by infrared technology (OPT-M3, Columbus Instruments), as the count of 3D beam breaking (X total, Y ambulatory, and Z) was measured.

2.4.12. Mice handling and conditions

Generation of wild type Bif-1+/+ mice and whole body Bif-1-/- mice in C57BL/6 background was described previously. For dietary challenge experiments, 6-week-old mice were fed with standard chow providing 18% calories from fat (Harlan Teklad,

TD2918) or a HFD providing 55% calories from fat (Harlan Teklad, TD93075). Mouse studies were conducted in accordance with federal guidelines and were approved by the

Pennsylvania State University Animal Care and Use Committee.

2.4.13. Statistical analysis

All data are expressed as mean ± S.E.M. Student t-tests were used for single comparisons. One-Way ANOVA tests were performed followed by Bonferroni’s post- hoc tests for multiple comparisons unless specified. Statistical significance is assumed at *p < 0.05, **p < 0.01, ***p <0.001, and ****p <0.0001. Sample size was chosen based on results from pilot studies.

Chapter 3. IDENTIFICATION OF ATG9 INTERACTING PROTEINS

3.1. Abstract

Despite recent advances in understanding the molecular machinery of autophagy, the dynamics of the isolation membranes, including their nucleation, expansion, bending and closure, are of major interest and remain elusive to the field. The only transmembrane ATG protein identified so far, ATG9, is required at both early and later stages in autophagy in mammalian cells, and is therefore the best candidate for delivering lipids for the initiation and expansion of the isolation membranes. During autophagy, vesicles containing Atg9 are generated from a Rab11-positive reservoir by the Bif-1 and

Dynamin2 membrane fission machinery; these Atg9 vesicles are shuttled back and forth to the autophagosome formation site, only transiently associating with the growing isolation membrane. The precise role of Atg9 trafficking in autophagosome biogenesis is still unclear. Furthermore, the molecular interactions controlling the association and dissociation between Atg9 vesicles and the expanding autophagosomes are essentially unknown, at least partially due to the difficulty in capturing the transient interactions between Atg9 and its potential partners using conventional biochemical methods.

Using inducible protein crosslinking coupled with affinity purification and proteomics, in the dissertation study, I identify VCP/p97 as a novel interaction partner of

Atg9. Moreover, genetic knockdown of VCP in cells appears to disrupt autophagosome maturation and causes the accumulation of LC3 positive puncta during starvation.

Although further studies are needed to precisely dissect the molecular function of VCP in the autophagy machinery, I propose a potential role for VCP in the retrieval of Atg9 and other autophagy proteins from the autophagosome. Taken together, the current study 75

advances our knowledge in the Atg9 interaction network and the discovery of VCP as an

Atg9 interactor may open new doors for understanding the molecular mechanisms of autophagosome biogenesis and maturation.

3.2. Results

3.2.1. Purification of Atg9-cPTP Complexes Through Tandem Affinity Purification

As our strategy for identifying Atg9 interaction partners is based on affinity purification, we first confirmed that the exogenous Atg9 as the “bait” is functional for autophagy activity. To test this, we expressed Atg9 with a C-terminal tandem affinity tag

PTP (Protein C-TEV-Protein A, Atg9-cPTP) in Atg9-/- mouse embryonic fibroblasts

(MEFs). As shown in Figure 3.1, loss of Atg9 impairs both the autophagy induction and degradative activity, as reflected by the LC3-II levels. This autophagy defect is partially rescued by expression of Atg9a-cPTP in the Atg9-/- cells.

During autophagy, mammalian Atg9 travels to the autophagosomal membrane and associates with it transiently.34 Therefore, the interactions between Atg9 and its interaction partners at this location are potentially difficult to capture with conventional biochemical methods. In present study, I applied UV-inducible crosslinking in combination with the use of photo leucine and methionine (analogues of the natural leucine and methionine, hereafter referred to as photo amino acids) to fix cellular interactions prior to affinity purification. The photo amino acids can be taken up by cells and incorporated during protein synthesis; they form stable conjugates with adjacent 76

molecules upon UV (λ=365 nm) irradiation. Therefore, two interacting proteins that contain photo amino acids in at least one of them at the binding pocket are conjugated to form a stable complex (Figure 3.2). As shown in Figure 3.3, the UV crosslinking fixed a portion of Atg9 to complexes, as reflected by the induction of upper bands on Atg9 immunoblot. Importantly, these complexes can be recovered after the PTP tandem affinity purification process (Figure 3.3).

Figure 3.1. Atg9a-cPTP is Functional for Autophagy Activity. MEF Atg9+/+, Atg9-/-, and the Atg9-/- cells expressing exogenous Atg9a-cPTP were subjected to a 4 h- amino acid/serum starvation assay. (a) Cell lysates were analyzed by immunoblotting with the indicated antibodies. Representative immunoblots were shown. (b-d) The autophagic activity under basal state and under starvation was calculated based on

normalized LC3-II levels (normalized to actin). Autophagic activity: normalized LC3-II for Baf A1 treated sample minus that for untreated sample under basal (basal activity) or under starvation (induced activity); autophagy induction degree: normalized LC3-II for the starved and Bafilomycin A1 (Baf A1) treated sample minus that for the Baf A1 treated and un-starved sample. The activity/induction value in MEF Atg9+/+ was then used as 100%.

Figure 3.2. Illustration of UV-inducible protein crosslink using photo amino acids. L-Photo-Leucine and L-Photo-Methionine (photo amino acid) are amino acid derivatives that contain diazirine rings. They can be incorporated into the proteins during synthesis and then ultraviolet light (UV)-activated to covalently crosslink proteins within protein-protein interaction domains in their native environment within live cells.

Figure 3.3. Analysis of fractions collected in tandem affinity purification of Atg9-cPTP complex. HEK293T cells overexpressing Atg9-cPTP or Ctrl-cPTP were subjected to a 2-h amino acid/serum starvation assay and subsequently the UV-inducible protein crosslinking; Atg9-cPTP or Ctrl-cPTP complex was then purified from the cell lysates as described in Methods and Materials. Fractions were analyzed by (a). SyproRuby stain. Annotation T: TEV protease recovered in the TEV eluate; (b). Immunoblotting with anti-HPC4 (protein C tag).

3.2.2. Interaction Candidates Identified by Proteomics

The purification products were subjected to proteomic analysis, which identified a total of 140 Atg9-interacting hits, which were narrowed down to 128 after filtering out the overlapping hits derived from both the sample (Atg9-cPTP) and the negative control

(Ctrl-cPTP) (Figure 3.4). The remaining hits were then sorted based on peptide sores, subcellular localizations, as well as reported functions. A list of 20 proteins was finally considered as top candidates, which were further screened for their effects on autophagosome formation and Atg9 trafficking.

Figure 3.4. Stratification of Atg9-interacting hits. (a) Hits were compared between the Atg9-cPTP sample and the Ctrl-cPTP negative control, and the overlapping hits were filtered out. The rest of the hits were further stratified by cellular function, and contaminant-like hits (e.g. immunoglobins and keratins) were removed from the list. (b, c) Analyses of the biological functions (b) and the subcellular localization (c) of the top 20 Atg9-interacting candidates.

3.2.3. Autophagic Function Screen in Atg9 Interacting Top Candidates

To further concentrate on interaction candidates that play potentially critical roles in autophagy and Atg9 trafficking, we performed a function screen for autophagosome formation among the top 20 hits using small interfering RNAs (siRNAs). Smartpool siRNAs targeting the candidates were introduced into HeLa cells stably overexpressing a proprietary tagged LC3 (tLC3) fusion protein; the siRNA treated cells were subjected to

2 h- starvation, and then labeled with fluorophore-conjugated ligands that can specifically bind to the tLC3 either on the inner or the outer membrane of an autophagic vesicle. In addition, we also introduced the siRNAs to HeLa/Atg9-EGFP cells and examined the potential effects of the interaction candidates on the Atg9 distribution patterns during starvation-induced autophagy. We identified one of the hits, VCP (also named p97;

TERA; ALS14; CMT2Y; IBMPFD; HEL-220; IBMPFD1; HEL-S-70), as regulating autophagosome maturation. Specifically, HeLa/tLC3 cells treated with siVCP accumulated significantly higher numbers of autophagosomes in contrast to the non- targeting control during autophagy; these autophagosomes are positive for both inner and outer membrane-residing tLC3, (Figure 3.5a), suggesting these structures are closed autophagosomes with the outer membrane residing LC3 not being able to dissociate from the vesicle. Interestingly, the siVCP treated cells showed larger amount of p62 puncta, indicating a possible defect in autophagic degradation (Figure 3.5b). As p62 can also be degraded by proteasome in addition to by autophagy, future experiments using proteasome and autophagy inhibitors are needed to accurately determine the nature of this p62 accumulation. However, VCP does not appear to affect Atg9 anterograde trafficking. 82

In addition to VCP, in these function screens, some other interacting candidates showed minor differences either in the autophagosome biogenesis or Atg9 trafficking as compared to the control. Although it is possible that the minor phenotype might be due to insufficient knockdown, these candidates are not being further studied by far in this part of the dissertation.

Figure 3.5. VCP depletion disrupts autophagosome maturation and accumulates p62. (a) HeLa/tLC3 cells were treated with small interference RNA (siRNA) targeting VCP (siVCP), or treated with non- targeting ctrl. 72 h post siRNA transfection, cells were incubated in amino acids/serum free media for 2 h and then subjected to the tLC3 ligand labeling. Representative fields were shown. (b) HeLa/Atg9-EGFP Cells were transfected and starved in the same way as in (a), and stained with anti-p62 antibody. Representative fields were shown.

3.2.4. Validation of VCP and Atg9 Interaction

In order to confirm the interaction between VCP and Atg9, a Co-

Immunoprecipitation experiment was performed in HEK293T cells. VCP-EGFP or Ctrl-

EGFP (tag alone) was overexpressed in the cells, which were subjected to starvation and then immunoprecipitation using GFP-Trap. As shown in Figure 3.6, endogenous Atg9 was pull-down along with the VCP-EGFP, but not with the EGFP tag control, indicating an interaction between VCP and Atg9. Future determination on whether endogenous

VCP and Atg9 can be pull-down together is needed to further confirm this interaction.

Figure 3.6. Co-Immunoprecipitation of VCP-EGFP and Atg9. HEK293T/VCP-EGFP and HEK293T/EGFP cells were subjected to 2 h- amino acid/serum starvation, harvested and lyzed as described in Methods and Materials. Lysates were then used to pulldown VCP-EGFP with GFP-Trap matrix. Immunoprecipitates (IP) and total cell lysates (TCL) were subjected to immunoblotting with the indicated antibodies and representative image of three individual experiments was shown.

3.3. Discussion

In the past, only two proteins have been reported as Atg9-interaction partners in mammalian cells. The poor understanding of the Atg9 interaction network could be because transient interactions are difficult to preserve in conventional biochemical assays. In current study, utilizing the UV inducible protein crosslink in photo amino acids incorporating cells, I identified VCP as a novel Atg9 interacting protein.

VCP has been reported to play a role in autophagy, although the underlying mechanism is unclear and controversial. In yeast, the VCP orthologue Cdc48 and cofactor Shp1 are essential for autophagosomes formation and the direct interaction of

Shp1 with Atg8 is needed during autophagy.155 Whether the homologous VCP-p47 complex interacts with the Atg8 homolog LC3 to regulate autophagy in mammalian cells is unclear, although it has been shown recently that VCP indeed has an LC3-interacting domain177. For mammalian cells, the link between VCP and autophagy was first reported in studies of the VCP-associated degenerative disease IBMPFD. Patient’s muscle was characterized by the accumulation of vacuoles enriched in ubiquitin, LC3-II and p62.150

Consistently, accumulation of non-degrading autophagic vacuoles was also observed in transgenic mice expressing disease-associated VCP mutants as well as in cultured cells with VCP mutation or deletion.152 These autophagy defects were either interpreted as blockage in the fusion of autophagosomes with lysosomes,153 or as impairment downstream of the autolysosome formation.154 This discrepancy is possibly caused by lack of research tools to accurately study autophagosome maturation.

With the novel use of the proprietary tag - LC3 fusion protein and its ligands, we have found in present study that cells depleted of VCP accumulated closed autophagic structures containing LC3 on the outer surface. As VCP is an AAA ATPase that segregates proteins from membranes or complexes144, it is possible that VCP may interact with Atg9 to mediate its retrograde trafficking and to retrieve LC3 from autophagosome.

Further, as it has been predicted that VCP may interact with Atg5, Atg12, and Atg2

(Biological General Repository for Interaction Datasets), it will be necessary to determine whether VCP indeed interacts with these protein and mediates their dissociation from the autophagosome in the future.

Moreover, it is noteworthy that VCP is known to play a role in endosomal sorting.178 This raises the question whether VCP performs independent functions in endosomes and autophagosomes separately, or whether the defects of VCP mutation/deletion in autophagy could be indirect consequence of impaired endosomal sorting, as autophagy is functionally linked to the latter179-181. It has been shown that the interaction between multivesicular bodies (MVBs) and autophagic compartments is required for autophagosome maturation, and that several protein complexes required for fusion with lysosomes are shared between both MVBs and autophagosomes.182 It will be interesting to determine whether VCP participates in the interaction (or donation) of endosomal membranes with (to) the autophagosome to ensure its subsequent fusion with the lysosome, and whether the accumulated autophagic vacuoles observed in VCP deletion are abnormal amphisome-like hybrid structures containing trapped outer membrane LC3. 86

Taken together, the current findings broaden our knowledge in the Atg9 interaction network through the identification of VCP. Overall, this work helps to further our understanding in the molecular mechanisms of autophagosome biogenesis and maturation.

3.4 Material and Methods

3.4.1 Determination of Half-Life of Photo Amino Acids

Photo-Leucine and photo-methionine were dissolved in PBS at 1 mg/ml and transferred to a cell culture dish. A portion was removed to determine initial absorbance.

UV lamp was turned on 5 minutes before irradiating samples, and was then positioned above the dish with a 1-cm distance from the cells to irradiate sample for 30 minutes. A portion of the sample was removed every 2-5 minutes and transferred to a microplate.

The absorbance of each sample portion in the microplate was measured using a spectrophotometer at 345 nm and blanked with PBS. Half-life of the photo amino acids is the time it takes for half of the sample to be photo-activated as determined by a 50% decrease in absorbance at 345 nm in comparison to starting material.

3.4.2 Optimization of UV Irradiation Conditions

Photo-L-Met and photo-L-Leu were added to DMEM-LM supplemented with

10% dialyzed FBS to a final conc. of 2mM and 4mM respectively. Regular culture media was aspirated off cells grown to 60-70% confluency, and the cells were washed twice

with PBS. Photo amino acids-supplemented media was then added to the cells that were then incubated in a 5% CO2 tissue-culture incubator for 24 hours. The photo amino acid containing media was then removed from cells and the cells were washed twice with

PBS. 6 ml of PBS was added to the cells and then cell culture dish was then placed on ice and the UV lamp was positioned above cells at 1-cm distance. Cells were irradiated for three times, five times or 10 times the half-life determined in Section 3.3.2. Two different

UV sources of 6 W and 15 W powers respectively were tested. To determine crosslink efficiency, cells post irradiation were harvested and analyzed by Western blotting.

3.4.3 Photo-Activated Crosslinking in Atg9a-cPTP/Ctrl-cPTP Overexpressing Cells

Wild-type HEK293T cells were seeded in 10 - cm dishes so that next day the cells reached 50% confluence. Cells were then transfected with Atg9-cPTP or Ctrl-cPTP using calcium transfection. Six hours post transfection, the media was replaced with photo amino acids -containing media and cells were further incubated in this media for 24 h.

Cells were then washed with PBS and replenished with 4.5 g/L glucose supplemented minimal media lacking FBS or amino acids to induce autophagy for 2 h. Post induction of autophagy, cells were placed on ice and subjected to UV irradiation at 345 nm from a 15

W lamp with a 1 cm irradiation distance for 35 min.

3.4.4 Purification of Atg9-cPTP Complex

Atg9-cPTP and Ctrl-cPTP overexpressing cells subjected to UV crosslinking were harvested and lyzed to extract protein using buffer containing 30 mM Tris-HCl, pH 7.5,

150 mM NaCl, 10% glycerol, 1% Triton X-100, 0.5 mM DTT, and is supplemented with protease inhibitor cocktail at 1:500 dilution ratio. A Western blotting was performed to confirm the protein/peptide expression prior to purification. On day of purification, 200

μl (settled volume) of IgG beads was first equilibrated in a Poly-Prep chromatography column with 50 ml of Buffer PA-150 containing 150 mM KCl, 20 mM Tris-HCl, pH 8.0,

3 mM MgCl2, 0.5 mM DTT and 0.1 %Tween20. Cell lysates were transferred to a pre- cooled 15 ml-Falcon (Six ml of lysates expressing Atg9-cPTP and four ml of lysates expressing Ctrl-cPTP), and spiked with Tween 20 to a final concentration of 0.1%. A small portion of the lysates was then removed and saved as Input sample. The rest of the lysates were added to the equilibrated column, and the column was then closed and placed it horizontally on an end-over-end rotator to rotate at 4 °C for 3 hours at low speed

(‘16’ on the scale of our rotator). The column was then placed in upright position to let the beads settle down. The flow-through was then collected, flash-frozen and stored at -

80° C. The column was filled with 10 ml of Buffer PA-150, and rotated for 10 min at 4

°C. The beads were then settled to let the wash solution drop out (rotation-wash) and washed for a second time with 15 ml of Buffer PA by direct flow through. Washed beads were equilibrated with 15 ml of TEV buffer, and then rotated/washed for 10 min at 4 °C.

2 ml of TEV buffer and 30 µl of (300 U) of TEV protease (10 U/µl) were added to the beads and the mixture was rotated overnight at 4°C. On the second day, 150 μl (settled volume) of anti-protein C affinity matrix was first equilibrated in a second Poly-Prep chromatography column with 50 ml of Buffer PC-150 containing 150 mM KCl 20 mM

Tris-HCl, pH 8.0, 3 mM MgCl2, 1 mM CaCl2, 0.1 % Tween20. The eluate from the 89

IgG column post overnight incubation was collected by gravity flow. The IgG column was rotated at RT with 2 ml of Wash Buffer containing 500 mM KCl, 20 mM Tris-HCl, pH 8.0, 2.5 mM DTT, 1 % Triton X-100, the washing mixture was recovered and combined with the eluate, and this washing and recovering step was repeated once.1 M

CaCl2 was added to the collected mixture to have a final concentration at 2 mM in order to titrate the EDTA in the TEV protease eluate. Protease inhibitor cocktail was also spiked in at a 1:500 ratio. A small portion of the TEV protease eluate was removed, flash frozen and stored at -80° C. The rest of the eluate was then added to the column containing the equilibrated anti-protein C beads, and the mixture was rotated for 3 hours at 4° C. Flow-through was then collected, flash frozen and stored at -80° C. The anti- protein C beads were washed six times with 10 ml of Buffer PC-150, with the first and fourth washes including a 10 min rotation step. To collect the final eluate, 600 µl of the

Elution Buffer was added to the settled matrix and the buffer-matrix mixture was transferred to a silicon-coated microcentrifuge tube to incubate for 5 min at RT with gentle rotation. The mixture was then transferred back to the column and the eluate was collected by gravity flow. This step was repeated twice and all three fractions of eluates were combined. A small portion was removed, flash frozen, and stored at -80° C. The rest of the eluate was then analyzed by proteomics at the Mass Spectrometry and Proteomics

Resource Laboratory.

3.4.5 SDS-PAGE and SyproRuby Stain of Purification Fractions

For SDS-PAGE, 0.01% volume of total cell lysates, 4.5% of TEV eluate, and

16.7% of final eluate was loaded on SDS-PAGE gels. Post electrophoresis, gels were fixed with fixative containing 50% methanol and 7% acetic acid for two times, each time

15 min. Fixed gels were then merged in SyproRuby gel stain solution and stained overnight at RT with gentle rocking. On the second day, stained gels were washed for 30 min with solution containing 10% methanol and 7% acetic acid, rinsed for 10 min with dH2O, and imaged on the ChemiDoc XRS system.

3.4.6 Introduction of siRNAs to Cells, Starvation Assays

HeLa-S3 cells were seeded at 0.5 x 10 4 cells per well in 8-well chamber slides.

Next day cells reached approximately 15% confluence. Cells were rinsed once with Acell

Delivery Media (Dharmacon), and then incubated in the Delivery Media containing 1 μM

Accell siRNA for 72 h at 37 °C with 5% CO2. For starvation assays, cells post siRNA treatment were rinsed twice with 4.5 g/L glucose supplemented minimal media lacking

FBS and amino acids, and incubated in this minimal media for 2 h.

3.4.7 Co-Immunoprecipitation Experiments

HEK293T cells were transfected with VCP-EGFP or EGFP tag protein alone as negative control. 24 h post transfection, cells were subjected to starvation assay as described in section 3.4.6. Cells were then harvested and lyzed in Co-IP Buffer containing 10 mM Tris-HCl pH 7.5, 150 mM NaCl, 0.5 mM EDTA, 0.5% NP-40, and protease inhibitors. 750 μg of total protein was added to 20 μl GFP-Trap beads that were

washed three times with Dilution Buffer containing 10 mM Tris-HCl pH 7.5, 150 mM

NaCl, and 0.5 mM EDTA, and the total incubation volume was added up to 500 μl with

Dilution Buffer. The lysate and beads mixture was rotated end-over-end for 2 h at RT, and then centrifuged at 2500x g for 2 min at 4 °C. A small portion of the supernatant was removed and saved for analysis and the rest of supernatant was discarded. The beads were resuspended and rotated in 500 μl of ice-cold Dilution Buffer and centrifuged at

2500x g for 2 min at 4 °C to wash off non-specific bound proteins, and this was done for a total of four times. 25 μl of the 2x SDS-sample buffer was then added to the beads that were approximately 25 μl in volume. The total cell lysate, post IP supernatant, and the IP fraction were then analyzed via SDS-PAGE.

3.4.8 Statistical analysis

All data are expressed as mean ± S.E.M. Student t-tests were used for single comparisons. One-way ANOVA tests were performed followed by Bonferroni’s post-hoc test for multiple comparisons unless specified. Statistical significance is assumed at *p

< 0.05, **p < 0.01, ***p <0.001, and ****p <0.0001.

Chapter 4. SUMMARY, FUTURE DIRECTIONS, AND SIGNIFICANCE

4.1. Role of Bif-1 in Lipid Homeostasis and Obesity (Chapter 2)

4.1.1. Summary of Findings

In present study, we identify Bif-1 as regulating lipid catabolism to control the size of lipid droplets and to prevent the development of obesity and insulin resistance.

Overall, we have found that:

 Bif-1 deficiency promotes the expansion of adipose tissue mass without altering

food intake or physical activities. Mice lacking Bif-1 develop obesity and

hyperinsulinemia upon aging or a high fat-diet challenge.

 Bif-1 is dispensable for adipose tissue development or lipid droplet biogenesis.

 Loss of Bif-1 decreases adipose tissue lipolysis rate and causes adipocyte

hypertrophy. Adipose tissue deficient in Bif-1 displays impaired degradation of

autophagy adaptor/substrate p62 and lipid droplet coating protein PLIN1

(Perilipin A) upon aging.

 The importance of Bif-1 in lipid turnover is not limited to adipose tissue since the

lipid droplet clearance is also attenuated by Bif-1 loss in the liver of mice starved

and red-fed.

 Moreover, the development of diet-induced obesity as well as Bif-1 deficiency

reduces the proteins levels of two other proteins critical for the autophagy-

lysosomal pathway, Atg9 and Lamp1, in the adipose tissue. Downregulation of

these two proteins could in turn undermine the autophagic degradation of lipids

and thus promote further fat accumulation.

Albeit these important findings that advance our understanding of Bif-1 function in adipose tissue and in obesity, questions still stand. In the following section, I will discuss some of the open questions and future directions pertaining both current project and also the research topic more generally.

4.1.2. Open Questions and Future Directions

 Limitations of current animal model for studying autophagy and obesity

Others have shown that in mice, targeted deletion of Atg5 or Atg7 in adipocytes reduced adipose tissue (AT) mass, decreased body weight, enhanced insulin sensitivity, and attenuated adipose tissue inflammation.75,76 However, in those mice models, adipocyte-specific interference with autophagy was achieved genetically using an adipocyte promoter, thereby affecting adipocytes from their early development onwards.

Whether modulation of autophagy in fully matured adipocytes or in obese individuals would have similar effects remains to be established. The present study shows that deletion of Bif-1 in mice whole body does not affect adipogenesis but rather promotes diet-induced obesity likely through defective lipophagy in AT, suggesting the potential importance of autophagic lipid degradation in this tissue for the prevention of obesity.

However, as Bif-1 is a multifunctional protein with ubiquitous tissue expression, currently we cannot completely exclude other potential confounding factors mediated by

Bif-1 in other tissues. Future studies with an inducible and adipose tissue specific Bif-1 knockout mouse model are needed in order to consolidate our current findings and to more precisely define the molecular mechanisms of Bif-1 and lipophagy in the

pathogenesis of obesity. In addition, studies of autophagy in metabolic diseases using isolated animal or human tissue samples are difficult owing to the technical barriers in the measurement of autophagic activity. Therefore, development of novel experimental tools in the future to assess autophagic activity in vivo or ex vivo will be necessary to further investigate the role of autophagy and its crosstalk with other cellular pathways in obesity and in normal physiology.

 The molecular mechanism of Bif-1 in lipophagy: LD recognition, autophagosome

biogenesis

Over the last few years, a number of distinct selective autophagy receptors have been identified, for instance, p62, NDP52, optineurin, and NBR1 etc., and it is possible that any of these could also serve as the LD receptors. Moreover, another protein that may function as a potential LD cargo recognition receptor is Huntingtin (Htt). Htt was recently shown to act as a scaffold for selective autophagy, and mutations in Huntingtin have been shown to lead to the generation of large, empty autophagosomes that fail to sequester cargo91. Since cells expressing mutant Huntingtin have revealed remarkable lipid accumulation, it is possible that Htt may be a LD recognition receptor protein.

Further, as Bif-1 and Htt are predicted to interact by similarity analysis, it would be interesting to determine whether Bif-1 and Htt regulate LD breakdown in the same pathway. Future studies will be necessary to validate these hypotheses.

 Reduced Atg9 and Lamp1 protein level

Our data also show that the protein levels of Atg9a and Lamp1 in WAT are reduced by the loss of Bif-1 upon aging or dietary challenge. Downregulation of these two proteins that are critical for autophagy could in turn impair the autophagic activity on

LD and promote further fat accumulation. Although further studies are required to understand the mechanism responsible for the downregulation of these autophagy- lysosomal proteins, recent studies have shown that Lamp1 and Atg9 genes are targets of the transcription factor EB (TFEB) that is negatively regulated by mTOR.82 Therefore obesity-attributed hyperinsulinemia in Bif-1 KO mice may mediate the downregulation of

Atg9 and Lamp1 through the activation of mTOR and suppression of TFEB signaling, and this hypothesis warrants further investigation in the future.

 Could modulation of autophagy in adipose tissue become a therapeutic approach in

obesity?

In clinical settings, upregulation of autophagy is observed in adipose tissue in obesity and type II diabetes (relative to lean individuals without diabetes) in correlation with higher AT apoptosis and inflammation in a recent work,183 and the authors concluded that the increased autophagic acitivity leads to higher inflammation. However, such causality relationship cannot be established due to lack of solid evidence. As autophagy is frequently viewed as an adaptive response to stress, its upregulation in obesity may be a response mechanism to protect adipose tissue from further metabolic stress and inflammation. Augmented adipose tissue mass exposes adipocytes to hypoxia 97

and ER stress184,185, which may lead to accumulation of misfolded proteins that need to be eliminated by autophagy. Consistently, inhibition of autophagy led to increased ER stress and subsequently promoted adipose tissue inflammation.186 Therefore, simply inhibiting autophagy may be deleterious by removing the tissue of an essential protective mechanism. On the other hand, as excessive autophagy activity could lead to cell death,187 which in the context of adipose tissue translates to higher inflammation that can lead to insulin resistance, simply enhancing autophagy may also be unrealistic or detrimental. Taken together, the complex regulation of autophagy in obesity and diabetes emphasizes the need for additional knowledge before modulating autophagy could be considered as a therapeutic strategy. Moreover, given that autophagy is required for diverse functions in normal physiology, it would appear that targeted modulation of lipophagy rather than non-selective autophagy could be adopted to treat patients with excessively high or low lipophagy activity as part of their disease etiology, seeking to normalize, rather than to overly inhibit or fuel this process.

4.2. Atg9 and the Identification of Its Interaction Partners (Chapter 3)

4.2.1. Summary of Findings

In this Chapter of the dissertation, I explored the interaction network of Atg9 with the ultimate goal to closely study its molecular functions in autophagy. The use of a combination of multiple elegant research tools allowed me to identify a putative interaction partner of Atg9. Overall, the main findings include:

 The use of photo inducible crosslink and tandem affinity purification

successfully preserved Atg9 interactions occurred during autophagy, and

allowed me to identify a list of interaction candidates.

 Among the interaction candidate, VCP/p97 appears to regulate autophagome

maturation. As VCP is a seggregase that extracts protein from membranes

and protein complexes, a potential role for VCP in Atg protein retrieval from

autophagosome is proposed, which needs future examination.

 An interaction between VCP-EGFP and Atg9 was observed by Co-IP,

although more biochemical experiments are needed to futher confirm this

interaction and to determine whether the endogenous proteins also interact

with each other.

4.2.2. Open Questions and Future Directions

 Roles of Atg9 in Autophagy

Despite intense studies on Atg9 in recent years, its precise function in autophagy remains elusive, and only disjointed routes of the ATG9 trafficking pathway are understood. It has been proposed that Atg9 may function to deliver membrane to growing phagophores and autophagosomes, but this hypothesis has not been formally proven. A collection of studies on Atg9 has reported inconsistent functions for Atg9 in autophagy.

Some reported that Atg9 is a key regulator of autophagy induction. Kishi-Itakura C. et al. showed that Atg9 is essential for the formation of phagophores as no isolation membrane could be seen in cells lacking Atg9, and that Atg9 acts downstream of ULK1 but

upstream of other proteins in the autophagosome biogenesis machinery.25 Later, Orsi A. et al. suggested that in mammalian cells Atg9 acts at an early stage, as depletion of Atg9 inhibits the formation of early ULK1-, DFCP1-, and WIPI2-positive phagophores.34 The current study showed that loss of Atg9 drastically reduced the autophagosome formation induced by nutrient deprivation; however, a small number of immature autophagosomes does form but cannot close. Therefore, Atg9 likely has dual functions both in autophagosome nucleation and also in the delivery of lipids and/or proteins to the growing phagophores. Currently, the Atg9 interacting protein candidates are being profiled and evaluated, and in the future a lipidomic profiling of lipids that are associated with Atg9 awaits determination. Moreover, given the diverse subcellular distribution of

Atg9 and its complex trafficking pathways during autophagy, it is possible that each trafficking route (e.g. originating from plasma versus from TGN/ endosomal compartments) is correlated with a specific type of autophagy induction signal or a specific type of cargo to be sequestered. Future studies will need to more thoroughly address the relationship between the different subpopulations of Atg9 and their potentially diverse roles in the autophagosome biogenesis.

 Role(s) of VCP in the autophagy machinery

Autophagosomes fuse with the lysosome once the vesicle membranes are sealed and the Atg machinery is disassembled and dissociated from the autophagosome.

Accordingly, failure to release LC3 from the autophagosome surface by LC3-PE delipidation leads to impaired autophagy. VCP, identified as a novel Atg9 interaction

100

partner in current study, has been reported to play a role in autophagosome maturation, although the underlying mechanism is unclear or controversial. Different cell models and experimental methods may contribute to the discrepancies among various studies, although what is really limiting the investigation on autophagosome maturation mechanism and data interpretation is lack of appropriate research tools. With the novel use of tLC3 fusion protein and its ligands, we have found in present study that cells depleted of VCP accumulated closed autophagic vacuoles containing LC3 on the surface.

As discussed earlier, VCP is an AAA ATPase that segregates proteins from membranes or complexes;144 therefore, it is possible that VCP may serve a role in the retrieval of Atg proteins such as Atg9 and LC3. Further, as it has been predicted that VCP may interact with Atg5, Atg12, and Atg2 (Biological General Repository for Interaction Datasets), it will be necessary to determine whether VCP indeed interact with these protein and mediates their dissociation from the autophagosome in the future.

4.3. Overall Significance and Perspectives

Obesity has become a worldwide health problem as a consequence of excessive high-calorie intake and the adoption of sedentary lifestyles; it is associated with various health effects such as heart disease, type II diabetes, stroke, and several types of cancer.

However, controlling obesity and its associated health conditions remains a major challenge due to an incomplete understanding of its molecular machinery.

101

The roles of autophagy in lipid homeostasis suggest that impairment in this pathway may be a fundamental mechanism of obesity. Through the identification of Bif-1 as a novel regulator of lipid homeostasis, the dissertation study furthered our understanding in the potential roles of adipose tissue autophagy in lipid degradation and thus in the prevention of obesity. In Chapter 2 of the study, we demonstrated that Bif-1 deficiency promotes the expansion of adipose tissue mass without altering food intake or physical activities; Bif-1 is dispensable for adipose tissue development; depletion of Bif-

1 reduces the rate of adipose tissue lipolysis, lowers the degradation of autophagy adaptor and substrates, and results in adipocyte hypertrophy upon aging. Consequently, mice lacking Bif-1 develop diet-induced obesity accompanied by hyperinsulinemia.

Bif-1 therefore has the potential to serve as an important therapeutic target for the treatment of obesity. Whether individual variation in Bif-1 expression level contributes to the development of obesity and its progression to other metabolic complications warrants future investigation.

Many questions are standing. For instance, does Bif-1 mediated lipid degradation depend on Atg9, and what other potential interactors of the Atg9-Bif-1 complex promote the biogenesis of autophagosomes? In order to address this question, as highlighted in

Chapter 3, I performed an Atg9 interactome profiling using inducible protein crosslinking coupled with affinity purification and proteomics. The current screen identifies VCP/p97 as a novel putative interaction partner of Atg9. Moreover, genetic knockdown of VCP in cells appears to disrupt autophagosome maturation. While further 102

studies are needed to precisely dissect the function of VCP in the autophagy machinery, I propose a potential role for VCP in the retrieval of autophagy proteins from the autophagosome.

Collectively, my dissertation study identifies Bif-1 as a novel player in lipid homeostasis and obesity. Therefore targeting lipophagy may provide new avenues for the treatment of obesity and its related metabolic complications. Furthermore, the current findings broaden our knowledge in the Atg9 interaction network through the identification of VCP as a putative Atg9 interacting partner. The development of new research tools during the dissertation study will also help to advance our understanding in the molecular mechanisms of autophagy and lipophagy.

Future studies will be needed to address: what is the molecular mechanism of Bif-

1 in lipophagy? How do we more precisely study lipophagy in adipose tissue? What molecule(s) can be used as lipophagy marker? Is this process dependent on the Bif-1

Atg9 interaction? Do Bif-1 and/or Atg9 promote the autophagosome biogenesis surrounding lipid droplets in the same fashion as in canonical autophagy? Does lipophagy and macroautophagy share the same set of adaptors for substrate recognition? What is the mechanism of Bif-1 dependent reduction in the protein levels of Atg9 and Lamp1 – is it mediated through TFEB? Related to the autophagosome biogenesis, what are the exact roles of Atg9 in autophagy? What are the potential lipids, in addition to proteins, that

Atg9 may shuttle to/from the autophagosome? Does different type of stress (e.g. amino acid deprivation versus hypoxia or ROS) induce specific trafficking itinerary of Atg9, 103

potentially from a specific membrane source? Finally, whether VCP mediates the retrieval of Atg proteins also awaits future determination.

104

REFERENCES

1 Mizushima, N., Yoshimori, T. & Ohsumi, Y. The role of Atg proteins in autophagosome formation. Annual review of cell and developmental biology 27, 107-132, doi:10.1146/annurev-cellbio-092910-154005 (2011). 2 Mizushima, N. Autophagy: process and function. Genes & development 21, 2861- 2873, doi:10.1101/gad.1599207 (2007). 3 Levine, B. & Kroemer, G. Autophagy in the pathogenesis of disease. Cell 132, 27-42, doi:10.1016/j.cell.2007.12.018 (2008). 4 Mizushima, N., Levine, B., Cuervo, A. M. & Klionsky, D. J. Autophagy fights disease through cellular self-digestion. Nature 451, 1069-1075, doi:10.1038/nature06639 (2008). 5 Choi, A. M., Ryter, S. W. & Levine, B. Autophagy in human health and disease. The New England journal of medicine 368, 1845-1846, doi:10.1056/NEJMc1303158 (2013). 6 Jung, C. H. et al. ULK-Atg13-FIP200 complexes mediate mTOR signaling to the autophagy machinery. Molecular biology of the cell 20, 1992-2003, doi:10.1091/mbc.E08-12-1249 (2009). 7 Kim, J., Kundu, M., Viollet, B. & Guan, K. L. AMPK and mTOR regulate autophagy through direct phosphorylation of Ulk1. Nature cell biology 13, 132- 141, doi:10.1038/ncb2152 (2011). 8 Russell, R. C. et al. ULK1 induces autophagy by phosphorylating Beclin-1 and activating VPS34 lipid kinase. Nature cell biology 15, 741-750, doi:10.1038/ncb2757 (2013). 9 Di Bartolomeo, S. et al. The dynamic interaction of AMBRA1 with the dynein motor complex regulates mammalian autophagy. The Journal of cell biology 191, 155-168, doi:10.1083/jcb.201002100 (2010). 10 Fan, W., Nassiri, A. & Zhong, Q. Autophagosome targeting and membrane curvature sensing by Barkor/Atg14(L). Proceedings of the National Academy of Sciences of the United States of America 108, 7769-7774, doi:10.1073/pnas.1016472108 (2011). 11 Axe, E. L. et al. Autophagosome formation from membrane compartments enriched in phosphatidylinositol 3-phosphate and dynamically connected to the endoplasmic reticulum. The Journal of cell biology 182, 685-701, doi:10.1083/jcb.200803137 (2008). 12 Wilz, L., Fan, W. & Zhong, Q. Membrane curvature response in autophagy. Autophagy 7, 1249-1250, doi:10.4161/auto.7.10.16738 (2011).

105

13 Araki, Y. et al. Atg38 is required for autophagy-specific phosphatidylinositol 3- kinase complex integrity. The Journal of cell biology 203, 299-313, doi:10.1083/jcb.201304123 (2013). 14 Zhong, Y. et al. Nrbf2 protein suppresses autophagy by modulating Atg14L protein-containing Beclin 1-Vps34 complex architecture and reducing intracellular phosphatidylinositol-3 phosphate levels. The Journal of biological chemistry 289, 26021-26037, doi:10.1074/jbc.M114.561134 (2014). 15 Takahashi, Y. et al. Bif-1 interacts with Beclin 1 through UVRAG and regulates autophagy and tumorigenesis. Nature cell biology 9, 1142-1151, doi:10.1038/ncb1634 (2007). 16 Farsad, K. et al. Generation of high curvature membranes mediated by direct endophilin bilayer interactions. The Journal of cell biology 155, 193-200, doi:10.1083/jcb.200107075 (2001). 17 Geng, J. & Klionsky, D. J. The Atg8 and Atg12 ubiquitin-like conjugation systems in macroautophagy. 'Protein modifications: beyond the usual suspects' review series. EMBO reports 9, 859-864, doi:10.1038/embor.2008.163 (2008). 18 Kabeya, Y. et al. LC3, GABARAP and GATE16 localize to autophagosomal membrane depending on form-II formation. Journal of cell science 117, 2805- 2812, doi:10.1242/jcs.01131 (2004). 19 Kirisako, T. et al. The reversible modification regulates the membrane-binding state of Apg8/Aut7 essential for autophagy and the cytoplasm to vacuole targeting pathway. The Journal of cell biology 151, 263-276 (2000). 20 Ichimura, Y. et al. A ubiquitin-like system mediates protein lipidation. Nature 408, 488-492, doi:10.1038/35044114 (2000). 21 Barth, S., Glick, D. & Macleod, K. F. Autophagy: assays and artifacts. The Journal of pathology 221, 117-124, doi:10.1002/path.2694 (2010). 22 Knorr, R. L., Lipowsky, R. & Dimova, R. Autophagosome closure requires membrane scission. Autophagy 11, 2134-2137, doi:10.1080/15548627.2015.1091552 (2015). 23 Weidberg, H. et al. LC3 and GATE-16/GABARAP subfamilies are both essential yet act differently in autophagosome biogenesis. The EMBO journal 29, 1792- 1802, doi:10.1038/emboj.2010.74 (2010). 24 Velikkakath, A. K., Nishimura, T., Oita, E., Ishihara, N. & Mizushima, N. Mammalian Atg2 proteins are essential for autophagosome formation and important for regulation of size and distribution of lipid droplets. Molecular biology of the cell 23, 896-909, doi:10.1091/mbc.E11-09-0785 (2012). 25 Kishi-Itakura, C., Koyama-Honda, I., Itakura, E. & Mizushima, N. Ultrastructural analysis of autophagosome organization using mammalian autophagy-deficient cells. Journal of cell science 127, 4089-4102, doi:10.1242/jcs.156034 (2014). 26 Klionsky, D. J. The molecular machinery of autophagy: unanswered questions. Journal of cell science 118, 7-18, doi:10.1242/jcs.01620 (2005).

106

27 Lamb, C. A., Yoshimori, T. & Tooze, S. A. The autophagosome: origins unknown, biogenesis complex. Nature reviews. Molecular cell biology 14, 759- 774, doi:10.1038/nrm3696 (2013). 28 Ravikumar, B. et al. Dynein mutations impair autophagic clearance of aggregate- prone proteins. Nature genetics 37, 771-776, doi:10.1038/ng1591 (2005). 29 Kochl, R., Hu, X. W., Chan, E. Y. & Tooze, S. A. Microtubules facilitate autophagosome formation and fusion of autophagosomes with endosomes. Traffic 7, 129-145, doi:10.1111/j.1600-0854.2005.00368.x (2006). 30 Kimura, S., Noda, T. & Yoshimori, T. Dynein-dependent movement of autophagosomes mediates efficient encounters with lysosomes. Cell structure and function 33, 109-122 (2008). 31 Wijdeven, R. H. et al. Cholesterol and ORP1L-mediated ER contact sites control autophagosome transport and fusion with the endocytic pathway. Nature communications 7, 11808, doi:10.1038/ncomms11808 (2016). 32 Diao, J. et al. ATG14 promotes membrane tethering and fusion of autophagosomes to endolysosomes. Nature 520, 563-566, doi:10.1038/nature14147 (2015). 33 Itakura, E., Kishi-Itakura, C. & Mizushima, N. The hairpin-type tail-anchored SNARE syntaxin 17 targets to autophagosomes for fusion with endosomes/lysosomes. Cell 151, 1256-1269, doi:10.1016/j.cell.2012.11.001 (2012). 34 Orsi, A. et al. Dynamic and transient interactions of Atg9 with autophagosomes, but not membrane integration, are required for autophagy. Molecular biology of the cell 23, 1860-1873, doi:10.1091/mbc.E11-09-0746 (2012). 35 Reggiori, F., Tucker, K. A., Stromhaug, P. E. & Klionsky, D. J. The Atg1-Atg13 complex regulates Atg9 and Atg23 retrieval transport from the pre- autophagosomal structure. Developmental cell 6, 79-90 (2004). 36 Longatti, A. & Tooze, S. A. Vesicular trafficking and autophagosome formation. Cell death and differentiation 16, 956-965, doi:10.1038/cdd.2009.39 (2009). 37 Takahashi, Y. et al. Bif-1 regulates Atg9 trafficking by mediating the fission of Golgi membranes during autophagy. Autophagy 7, 61-73 (2011). 38 Singh, R. & Cuervo, A. M. Autophagy in the cellular energetic balance. Cell metabolism 13, 495-504, doi:10.1016/j.cmet.2011.04.004 (2011). 39 Mortimore, G. E. & Schworer, C. M. Induction of autophagy by amino-acid deprivation in perfused rat liver. Nature 270, 174-176 (1977). 40 Mortimore, G. E., Poso, A. R., Kadowaki, M. & Wert, J. J., Jr. Multiphasic control of hepatic protein degradation by regulatory amino acids. General features and hormonal modulation. The Journal of biological chemistry 262, 16322-16327 (1987). 41 Noda, T. & Ohsumi, Y. Tor, a phosphatidylinositol kinase homologue, controls autophagy in yeast. The Journal of biological chemistry 273, 3963-3966 (1998). 42 Blommaart, E. F., Luiken, J. J., Blommaart, P. J., van Woerkom, G. M. & Meijer, A. J. Phosphorylation of ribosomal protein S6 is inhibitory for autophagy in 107

isolated rat hepatocytes. The Journal of biological chemistry 270, 2320-2326 (1995). 43 Ravikumar, B. et al. Inhibition of mTOR induces autophagy and reduces toxicity of polyglutamine expansions in fly and mouse models of Huntington disease. Nature genetics 36, 585-595, doi:10.1038/ng1362 (2004). 44 Dennis, P. B. et al. Mammalian TOR: a homeostatic ATP sensor. Science 294, 1102-1105, doi:10.1126/science.1063518 (2001). 45 Meijer, A. J., Lorin, S., Blommaart, E. F. & Codogno, P. Regulation of autophagy by amino acids and MTOR-dependent signal transduction. Amino acids 47, 2037- 2063, doi:10.1007/s00726-014-1765-4 (2015). 46 Koga, H., Kaushik, S. & Cuervo, A. M. Altered lipid content inhibits autophagic vesicular fusion. FASEB journal : official publication of the Federation of American Societies for Experimental Biology 24, 3052-3065, doi:10.1096/fj.09- 144519 (2010). 47 Komiya, K. et al. Free fatty acids stimulate autophagy in pancreatic beta-cells via JNK pathway. Biochemical and biophysical research communications 401, 561- 567, doi:10.1016/j.bbrc.2010.09.101 (2010). 48 Niso-Santano, M. et al. Direct interaction between STAT3 and EIF2AK2 controls fatty acid-induced autophagy. Autophagy 9, 415-417, doi:10.4161/auto.22910 (2013). 49 Tan, S. H. et al. Induction of autophagy by palmitic acid via protein kinase C- mediated signaling pathway independent of mTOR (mammalian target of rapamycin). The Journal of biological chemistry 287, 14364-14376, doi:10.1074/jbc.M111.294157 (2012). 50 Mei, S. et al. Differential roles of unsaturated and saturated fatty acids on autophagy and apoptosis in hepatocytes. The Journal of pharmacology and experimental therapeutics 339, 487-498, doi:10.1124/jpet.111.184341 (2011). 51 Lee, S. M., Kim, J. H., Cho, E. J. & Youn, H. D. A nucleocytoplasmic malate dehydrogenase regulates p53 transcriptional activity in response to metabolic stress. Cell death and differentiation 16, 738-748, doi:10.1038/cdd.2009.5 (2009). 52 Budanov, A. V. & Karin, M. p53 target genes sestrin1 and sestrin2 connect genotoxic stress and mTOR signaling. Cell 134, 451-460, doi:10.1016/j.cell.2008.06.028 (2008). 53 Zalckvar, E. et al. DAP-kinase-mediated phosphorylation on the BH3 domain of beclin 1 promotes dissociation of beclin 1 from Bcl-XL and induction of autophagy. EMBO reports 10, 285-292, doi:10.1038/embor.2008.246 (2009). 54 Yu, T., Robotham, J. L. & Yoon, Y. Increased production of reactive oxygen species in hyperglycemic conditions requires dynamic change of mitochondrial morphology. Proceedings of the National Academy of Sciences of the United States of America 103, 2653-2658, doi:10.1073/pnas.0511154103 (2006). 55 Sedeek, M. et al. Critical role of Nox4-based NADPH oxidase in glucose-induced oxidative stress in the kidney: implications in type 2 diabetic nephropathy.

108

American journal of physiology. Renal physiology 299, F1348-1358, doi:10.1152/ajprenal.00028.2010 (2010). 56 Takeda, K. et al. Transient glucose deprivation causes upregulation of heme oxygenase-1 and cyclooxygenase-2 expression in cardiac fibroblasts. Journal of molecular and cellular cardiology 36, 821-830, doi:10.1016/j.yjmcc.2004.03.008 (2004). 57 Cao, M., Jiang, J., Du, Y. & Yan, P. Mitochondria-targeted antioxidant attenuates high glucose-induced P38 MAPK pathway activation in human neuroblastoma cells. Molecular medicine reports 5, 929-934, doi:10.3892/mmr.2012.746 (2012). 58 Tsai, K. H. et al. NADPH oxidase-derived superoxide anion-induced apoptosis is mediated via the JNK-dependent activation of NF-kappaB in cardiomyocytes exposed to high glucose. Journal of cellular physiology 227, 1347-1357, doi:10.1002/jcp.22847 (2012). 59 Song, J. J. & Lee, Y. J. Cross-talk between JIP3 and JIP1 during glucose deprivation: SEK1-JNK2 and Akt1 act as mediators. The Journal of biological chemistry 280, 26845-26855, doi:10.1074/jbc.M502318200 (2005). 60 Wong, C. H. et al. Simultaneous induction of non-canonical autophagy and apoptosis in cancer cells by ROS-dependent ERK and JNK activation. PloS one 5, e9996, doi:10.1371/journal.pone.0009996 (2010). 61 Poso, A. R., Schworer, C. M. & Mortimore, G. E. Acceleration of proteolysis in perfused rat liver by deletion of glucogenic amino acids: regulatory role of glutamine. Biochemical and biophysical research communications 107, 1433- 1439 (1982). 62 Ogden, C. L., Carroll, M. D., Kit, B. K. & Flegal, K. M. Prevalence of childhood and adult obesity in the United States, 2011-2012. JAMA : the journal of the American Medical Association 311, 806-814, doi:10.1001/jama.2014.732 (2014). 63 Kahn, B. B. & Flier, J. S. Obesity and insulin resistance. The Journal of clinical investigation 106, 473-481, doi:10.1172/JCI10842 (2000). 64 Kaushik, S. et al. Loss of autophagy in hypothalamic POMC neurons impairs lipolysis. EMBO reports 13, 258-265, doi:10.1038/embor.2011.260 (2012). 65 Kaushik, S. et al. Autophagy in hypothalamic AgRP neurons regulates food intake and energy balance. Cell metabolism 14, 173-183, doi:10.1016/j.cmet.2011.06.008 (2011). 66 Tsukamoto, S., Kuma, A. & Mizushima, N. The role of autophagy during the oocyte-to-embryo transition. Autophagy 4, 1076-1078 (2008). 67 Merz, E. A., Brinster, R. L., Brunner, S. & Chen, H. Y. Protein degradation during preimplantation development of the mouse. Journal of reproduction and fertility 61, 415-418 (1981). 68 Stitzel, M. L. & Seydoux, G. Regulation of the oocyte-to-zygote transition. Science 316, 407-408, doi:10.1126/science.1138236 (2007). 69 Qu, X. et al. Autophagy gene-dependent clearance of apoptotic cells during embryonic development. Cell 128, 931-946, doi:10.1016/j.cell.2006.12.044 (2007). 109

70 Kent, G., Minick, O. T., Volini, F. I. & Orfei, E. Autophagic vacuoles in human red cells. The American journal of pathology 48, 831-857 (1966). 71 Takano-Ohmuro, H., Mukaida, M., Kominami, E. & Morioka, K. Autophagy in embryonic erythroid cells: its role in maturation. European journal of cell biology 79, 759-764, doi:10.1078/0171-9335-00096 (2000). 72 Miller, B. C. et al. The autophagy gene ATG5 plays an essential role in B lymphocyte development. Autophagy 4, 309-314 (2008). 73 Kuma, A. et al. The role of autophagy during the early neonatal starvation period. Nature 432, 1032-1036, doi:10.1038/nature03029 (2004). 74 Schiaffino, S., Mammucari, C. & Sandri, M. The role of autophagy in neonatal tissues: just a response to amino acid starvation? Autophagy 4, 727-730 (2008). 75 Baerga, R., Zhang, Y., Chen, P. H., Goldman, S. & Jin, S. Targeted deletion of autophagy-related 5 (atg5) impairs adipogenesis in a cellular model and in mice. Autophagy 5, 1118-1130 (2009). 76 Singh, R. et al. Autophagy regulates adipose mass and differentiation in mice. The Journal of clinical investigation 119, 3329-3339, doi:10.1172/JCI39228 (2009). 77 Zhang, Y. et al. Adipose-specific deletion of autophagy-related gene 7 (atg7) in mice reveals a role in adipogenesis. Proceedings of the National Academy of Sciences of the United States of America 106, 19860-19865, doi:10.1073/pnas.0906048106 (2009). 78 Martinez-Lopez, N. et al. Autophagy in Myf5+ progenitors regulates energy and glucose homeostasis through control of brown fat and skeletal muscle development. EMBO reports 14, 795-803, doi:10.1038/embor.2013.111 (2013). 79 Kovsan, J. et al. Altered autophagy in human adipose tissues in obesity. The Journal of clinical endocrinology and metabolism 96, E268-277, doi:10.1210/jc.2010-1681 (2011). 80 Haim, Y. et al. Elevated autophagy gene expression in adipose tissue of obese humans: A potential non-cell-cycle-dependent function of E2F1. Autophagy 11, 2074-2088, doi:10.1080/15548627.2015.1094597 (2015). 81 Singh, R. et al. Autophagy regulates lipid metabolism. Nature 458, 1131-1135, doi:10.1038/nature07976 (2009). 82 Settembre, C. et al. TFEB controls cellular lipid metabolism through a starvation- induced autoregulatory loop. Nature cell biology 15, 647-658, doi:10.1038/ncb2718 (2013). 83 Ma, D. et al. Autophagy deficiency by hepatic FIP200 deletion uncouples steatosis from liver injury in NAFLD. Molecular endocrinology 27, 1643-1654, doi:10.1210/me.2013-1153 (2013). 84 Jung, H. S. et al. Loss of autophagy diminishes pancreatic beta cell mass and function with resultant hyperglycemia. Cell metabolism 8, 318-324, doi:10.1016/j.cmet.2008.08.013 (2008). 85 Quan, W. et al. Autophagy deficiency in beta cells leads to compromised unfolded protein response and progression from obesity to diabetes in mice. Diabetologia 55, 392-403, doi:10.1007/s00125-011-2350-y (2012). 110

86 Masiero, E. et al. Autophagy is required to maintain muscle mass. Cell metabolism 10, 507-515, doi:10.1016/j.cmet.2009.10.008 (2009). 87 Thiam, A. R., Farese, R. V., Jr. & Walther, T. C. The biophysics and cell biology of lipid droplets. Nature reviews. Molecular cell biology 14, 775-786, doi:10.1038/nrm3699 (2013). 88 Brasaemle, D. L. et al. Perilipin A increases triacylglycerol storage by decreasing the rate of triacylglycerol hydrolysis. The Journal of biological chemistry 275, 38486-38493, doi:10.1074/jbc.M007322200 (2000). 89 Miyoshi, H. et al. Perilipin promotes hormone-sensitive lipase-mediated adipocyte lipolysis via phosphorylation-dependent and -independent mechanisms. The Journal of biological chemistry 281, 15837-15844, doi:10.1074/jbc.M601097200 (2006). 90 Zechner, R. et al. FAT SIGNALS--lipases and lipolysis in lipid metabolism and signaling. Cell metabolism 15, 279-291, doi:10.1016/j.cmet.2011.12.018 (2012). 91 Martinez-Vicente, M. et al. Cargo recognition failure is responsible for inefficient autophagy in Huntington's disease. Nature neuroscience 13, 567-576, doi:10.1038/nn.2528 (2010). 92 Ouimet, M. et al. Autophagy regulates cholesterol efflux from macrophage foam cells via lysosomal acid lipase. Cell metabolism 13, 655-667, doi:10.1016/j.cmet.2011.03.023 (2011). 93 Khaldoun, S. A. et al. Autophagosomes contribute to intracellular lipid distribution in enterocytes. Molecular biology of the cell 25, 118-132, doi:10.1091/mbc.E13-06-0324 (2014). 94 Hubbard, V. M. et al. Macroautophagy regulates energy metabolism during effector T cell activation. Journal of immunology 185, 7349-7357, doi:10.4049/jimmunol.1000576 (2010). 95 Xu, X. et al. Obesity activates a program of lysosomal-dependent lipid metabolism in adipose tissue macrophages independently of classic activation. Cell metabolism 18, 816-830, doi:10.1016/j.cmet.2013.11.001 (2013). 96 Kaini, R. R., Sillerud, L. O., Zhaorigetu, S. & Hu, C. A. Autophagy regulates lipolysis and cell survival through lipid droplet degradation in androgen-sensitive prostate cancer cells. The Prostate 72, 1412-1422, doi:10.1002/pros.22489 (2012). 97 van Zutphen, T. et al. Lipid droplet autophagy in the yeast Saccharomyces cerevisiae. Molecular biology of the cell 25, 290-301, doi:10.1091/mbc.E13-08- 0448 (2014). 98 Lapierre, L. R., Gelino, S., Melendez, A. & Hansen, M. Autophagy and lipid metabolism coordinately modulate life span in germline-less C. elegans. Current biology : CB 21, 1507-1514, doi:10.1016/j.cub.2011.07.042 (2011). 99 Nguyen, L. N. et al. Autophagy-related lipase FgATG15 of Fusarium graminearum is important for lipid turnover and plant infection. Fungal genetics and biology : FG & B 48, 217-224, doi:10.1016/j.fgb.2010.11.004 (2011).

111

100 Spandl, J., Lohmann, D., Kuerschner, L., Moessinger, C. & Thiele, C. Ancient ubiquitous protein 1 (AUP1) localizes to lipid droplets and binds the E2 ubiquitin conjugase G2 (Ube2g2) via its G2 binding region. The Journal of biological chemistry 286, 5599-5606, doi:10.1074/jbc.M110.190785 (2011). 101 Li, Z., Johnson, M. R., Ke, Z., Chen, L. & Welte, M. A. Drosophila lipid droplets buffer the H2Av supply to protect early embryonic development. Current biology : CB 24, 1485-1491, doi:10.1016/j.cub.2014.05.022 (2014). 102 Dupont, N. et al. Neutral lipid stores and lipase PNPLA5 contribute to autophagosome biogenesis. Current biology : CB 24, 609-620, doi:10.1016/j.cub.2014.02.008 (2014). 103 Kaushik, S. & Cuervo, A. M. Degradation of lipid droplet-associated proteins by chaperone-mediated autophagy facilitates lipolysis. Nature cell biology, doi:10.1038/ncb3166 (2015). 104 Takahashi, Y., Meyerkord, C. L. & Wang, H. G. Bif-1/endophilin B1: a candidate for crescent driving force in autophagy. Cell death and differentiation 16, 947- 955, doi:10.1038/cdd.2009.19 (2009). 105 Takahashi, Y. et al. Loss of Bif-1 suppresses Bax/Bak conformational change and mitochondrial apoptosis. Molecular and cellular biology 25, 9369-9382, doi:10.1128/MCB.25.21.9369-9382.2005 (2005). 106 Runkle, K. B., Meyerkord, C. L., Desai, N. V., Takahashi, Y. & Wang, H. G. Bif- 1 suppresses breast cancer cell migration by promoting EGFR endocytic degradation. Cancer biology & therapy 13, 956-966, doi:10.4161/cbt.20951 (2012). 107 Wan, J. et al. Endophilin B1 as a novel regulator of nerve growth factor/ TrkA trafficking and neurite outgrowth. The Journal of neuroscience : the official journal of the Society for Neuroscience 28, 9002-9012, doi:10.1523/JNEUROSCI.0767-08.2008 (2008). 108 Karbowski, M., Jeong, S. Y. & Youle, R. J. Endophilin B1 is required for the maintenance of mitochondrial morphology. The Journal of cell biology 166, 1027-1039, doi:10.1083/jcb.200407046 (2004). 109 Thoresen, S. B., Pedersen, N. M., Liestol, K. & Stenmark, H. A phosphatidylinositol 3-kinase class III sub-complex containing VPS15, VPS34, Beclin 1, UVRAG and BIF-1 regulates cytokinesis and degradative endocytic traffic. Experimental cell research 316, 3368-3378, doi:10.1016/j.yexcr.2010.07.008 (2010). 110 Huttner, W. B. & Schmidt, A. Lipids, lipid modification and lipid-protein interaction in membrane budding and fission--insights from the roles of endophilin A1 and synaptophysin in synaptic vesicle endocytosis. Current opinion in neurobiology 10, 543-551 (2000). 111 Bhatia, V. K. et al. Amphipathic motifs in BAR domains are essential for membrane curvature sensing. The EMBO journal 28, 3303-3314, doi:10.1038/emboj.2009.261 (2009).

112

112 Masuda, M. et al. Endophilin BAR domain drives membrane curvature by two newly identified structure-based mechanisms. The EMBO journal 25, 2889-2897, doi:10.1038/sj.emboj.7601176 (2006). 113 Cui, H., Ayton, G. S. & Voth, G. A. Membrane binding by the endophilin N-BAR domain. Biophysical journal 97, 2746-2753, doi:10.1016/j.bpj.2009.08.043 (2009). 114 Modregger, J., Schmidt, A. A., Ritter, B., Huttner, W. B. & Plomann, M. Characterization of Endophilin B1b, a brain-specific membrane-associated lysophosphatidic acid acyl transferase with properties distinct from endophilin A1. The Journal of biological chemistry 278, 4160-4167, doi:10.1074/jbc.M208568200 (2003). 115 Hrabchak, C., Henderson, H. & Varmuza, S. A testis specific isoform of endophilin B1, endophilin B1t, interacts specifically with protein phosphatase-1c gamma2 in mouse testis and is abnormally expressed in PP1c gamma null mice. Biochemistry 46, 4635-4644, doi:10.1021/bi6025837 (2007). 116 Cuddeback, S. M. et al. Molecular cloning and characterization of Bif-1. A novel Src homology 3 domain-containing protein that associates with Bax. The Journal of biological chemistry 276, 20559-20565, doi:10.1074/jbc.M101527200 (2001). 117 Pierrat, B. et al. SH3GLB, a new endophilin-related protein family featuring an SH3 domain. Genomics 71, 222-234, doi:10.1006/geno.2000.6378 (2001). 118 Wong, A. S. et al. Cdk5-mediated phosphorylation of endophilin B1 is required for induced autophagy in models of Parkinson's disease. Nature cell biology 13, 568-579, doi:10.1038/ncb2217 (2011). 119 Takahashi, Y. et al. The Bif-1-Dynamin 2 membrane fission machinery regulates Atg9-containing vesicle generation at the Rab11-positive reservoirs. Oncotarget 7, 20855-20868, doi:10.18632/oncotarget.8028 (2016). 120 Yamaguchi, H. et al. SRC directly phosphorylates Bif-1 and prevents its interaction with Bax and the initiation of anoikis. The Journal of biological chemistry 283, 19112-19118, doi:10.1074/jbc.M709882200 (2008). 121 Knuutila, S. et al. DNA copy number losses in human neoplasms. The American journal of pathology 155, 683-694, doi:10.1016/S0002-9440(10)65166-8 (1999). 122 Lee, J. W. et al. Decreased expression of tumour suppressor Bax-interacting factor-1 (Bif-1), a Bax activator, in gastric carcinomas. Pathology 38, 312-315, doi:10.1080/00313020600820880 (2006). 123 Coppola, D. et al. Down-regulation of Bax-interacting factor-1 in colorectal adenocarcinoma. Cancer 113, 2665-2670, doi:10.1002/cncr.23892 (2008). 124 Coppola, D. et al. Bax-interacting factor-1 expression in prostate cancer. Clinical genitourinary cancer 6, 117-121, doi:10.3816/CGC.2008.n.018 (2008). 125 Kim, S. Y. et al. Decreased expression of Bax-interacting factor-1 (Bif-1) in invasive urinary bladder and gallbladder cancers. Pathology 40, 553-557, doi:10.1080/00313020802320440 (2008).

113

126 Takahashi, Y. et al. Bif-1 haploinsufficiency promotes chromosomal instability and accelerates Myc-driven lymphomagenesis via suppression of mitophagy. Blood 121, 1622-1632, doi:10.1182/blood-2012-10-459826 (2013). 127 Ho, J. et al. Novel breast cancer metastasis-associated proteins. Journal of proteome research 8, 583-594, doi:10.1021/pr8007368 (2009). 128 Kim, M. S., Yoo, N. J. & Lee, S. H. Somatic mutation of pro-cell death Bif-1 gene is rare in common human cancers. APMIS : acta pathologica, microbiologica, et immunologica Scandinavica 116, 939-940, doi:10.1111/j.1600- 0463.2008.01091.x (2008). 129 Webber, J. L., Young, A. R. & Tooze, S. A. Atg9 trafficking in Mammalian cells. Autophagy 3, 54-56 (2007). 130 Toth, M. L. et al. Longevity pathways converge on autophagy genes to regulate life span in Caenorhabditis elegans. Autophagy 4, 330-338 (2008). 131 Shelly, S., Lukinova, N., Bambina, S., Berman, A. & Cherry, S. Autophagy is an essential component of Drosophila immunity against vesicular stomatitis virus. Immunity 30, 588-598, doi:10.1016/j.immuni.2009.02.009 (2009). 132 Liu, Y. et al. Autophagy regulates programmed cell death during the plant innate immune response. Cell 121, 567-577, doi:10.1016/j.cell.2005.03.007 (2005). 133 Saitoh, T. et al. Atg9a controls dsDNA-driven dynamic translocation of STING and the innate immune response. Proceedings of the National Academy of Sciences of the United States of America 106, 20842-20846, doi:10.1073/pnas.0911267106 (2009). 134 Webber, J. L. & Tooze, S. A. New insights into the function of Atg9. FEBS letters 584, 1319-1326, doi:10.1016/j.febslet.2010.01.020 (2010). 135 Zavodszky, E., Vicinanza, M. & Rubinsztein, D. C. Biology and trafficking of ATG9 and ATG16L1, two proteins that regulate autophagosome formation. FEBS letters 587, 1988-1996, doi:10.1016/j.febslet.2013.04.025 (2013). 136 Longatti, A. et al. TBC1D14 regulates autophagosome formation via Rab11- and ULK1-positive recycling endosomes. The Journal of cell biology 197, 659-675, doi:10.1083/jcb.201111079 (2012). 137 Longatti, A. & Tooze, S. A. Recycling endosomes contribute to autophagosome formation. Autophagy 8, 1682-1683, doi:10.4161/auto.21486 (2012). 138 Young, A. R. et al. Starvation and ULK1-dependent cycling of mammalian Atg9 between the TGN and endosomes. Journal of cell science 119, 3888-3900, doi:10.1242/jcs.03172 (2006). 139 Tang, H. W. et al. Atg1-mediated myosin II activation regulates autophagosome formation during starvation-induced autophagy. The EMBO journal 30, 636-651, doi:10.1038/emboj.2010.338 (2011). 140 Webber, J. L. & Tooze, S. A. Coordinated regulation of autophagy by p38alpha MAPK through mAtg9 and p38IP. The EMBO journal 29, 27-40, doi:10.1038/emboj.2009.321 (2010).

114

141 Erzberger, J. P. & Berger, J. M. Evolutionary relationships and structural mechanisms of AAA+ proteins. Annual review of biophysics and biomolecular structure 35, 93-114, doi:10.1146/annurev.biophys.35.040405.101933 (2006). 142 DeLaBarre, B. & Brunger, A. T. Complete structure of p97/valosin-containing protein reveals communication between nucleotide domains. Nature structural biology 10, 856-863, doi:10.1038/nsb972 (2003). 143 Schuberth, C. & Buchberger, A. UBX domain proteins: major regulators of the AAA ATPase Cdc48/p97. Cellular and molecular life sciences : CMLS 65, 2360- 2371, doi:10.1007/s00018-008-8072-8 (2008). 144 Meyer, H. & Weihl, C. C. The VCP/p97 system at a glance: connecting cellular function to disease pathogenesis. Journal of cell science 127, 3877-3883, doi:10.1242/jcs.093831 (2014). 145 Alexandru, G. et al. UBXD7 binds multiple ubiquitin ligases and implicates p97 in HIF1alpha turnover. Cell 134, 804-816, doi:10.1016/j.cell.2008.06.048 (2008). 146 Jentsch, S. & Rumpf, S. Cdc48 (p97): a "molecular gearbox" in the ubiquitin pathway? Trends in biochemical sciences 32, 6-11, doi:10.1016/j.tibs.2006.11.005 (2007). 147 Dobrynin, G. et al. Cdc48/p97-Ufd1-Npl4 antagonizes Aurora B during chromosome segregation in HeLa cells. Journal of cell science 124, 1571-1580, doi:10.1242/jcs.069500 (2011). 148 Puumalainen, M. R. et al. Chromatin retention of DNA damage sensors DDB2 and XPC through loss of p97 segregase causes genotoxicity. Nature communications 5, 3695, doi:10.1038/ncomms4695 (2014). 149 Olzmann, J. A., Richter, C. M. & Kopito, R. R. Spatial regulation of UBXD8 and p97/VCP controls ATGL-mediated lipid droplet turnover. Proceedings of the National Academy of Sciences of the United States of America 110, 1345-1350, doi:10.1073/pnas.1213738110 (2013). 150 Ju, J. S. & Weihl, C. C. p97/VCP at the intersection of the autophagy and the ubiquitin proteasome system. Autophagy 6, 283-285 (2010). 151 Bug, M. & Meyer, H. Expanding into new markets--VCP/p97 in endocytosis and autophagy. Journal of structural biology 179, 78-82, doi:10.1016/j.jsb.2012.03.003 (2012). 152 Custer, S. K., Neumann, M., Lu, H., Wright, A. C. & Taylor, J. P. Transgenic mice expressing mutant forms VCP/p97 recapitulate the full spectrum of IBMPFD including degeneration in muscle, brain and bone. Human molecular genetics 19, 1741-1755, doi:10.1093/hmg/ddq050 (2010). 153 Ju, J. S. et al. Valosin-containing protein (VCP) is required for autophagy and is disrupted in VCP disease. The Journal of cell biology 187, 875-888, doi:10.1083/jcb.200908115 (2009). 154 Tresse, E. et al. VCP/p97 is essential for maturation of ubiquitin-containing autophagosomes and this function is impaired by mutations that cause IBMPFD. Autophagy 6, 217-227 (2010).

115

155 Krick, R. et al. Cdc48/p97 and Shp1/p47 regulate autophagosome biogenesis in concert with ubiquitin-like Atg8. The Journal of cell biology 190, 965-973, doi:10.1083/jcb.201002075 (2010). 156 Muller, J. M., Deinhardt, K., Rosewell, I., Warren, G. & Shima, D. T. Targeted deletion of p97 (VCP/CDC48) in mouse results in early embryonic lethality. Biochemical and biophysical research communications 354, 459-465, doi:10.1016/j.bbrc.2006.12.206 (2007). 157 Watts, G. D. et al. Inclusion body myopathy associated with Paget disease of bone and frontotemporal dementia is caused by mutant valosin-containing protein. Nature genetics 36, 377-381, doi:10.1038/ng1332 (2004). 158 Skibinski, G. et al. Mutations in the endosomal ESCRTIII-complex subunit CHMP2B in frontotemporal dementia. Nature genetics 37, 806-808, doi:10.1038/ng1609 (2005). 159 Parkinson, N. et al. ALS phenotypes with mutations in CHMP2B (charged multivesicular body protein 2B). Neurology 67, 1074-1077, doi:10.1212/01.wnl.0000231510.89311.8b (2006). 160 Nishikawa, S., Yasoshima, A., Doi, K., Nakayama, H. & Uetsuka, K. Involvement of sex, strain and age factors in high fat diet-induced obesity in C57BL/6J and BALB/cA mice. Experimental animals / Japanese Association for Laboratory Animal Science 56, 263-272 (2007). 161 Xu, H. et al. Chronic inflammation in fat plays a crucial role in the development of obesity-related insulin resistance. The Journal of clinical investigation 112, 1821-1830, doi:10.1172/JCI19451 (2003). 162 Shoelson, S. E., Lee, J. & Goldfine, A. B. Inflammation and insulin resistance. The Journal of clinical investigation 116, 1793-1801, doi:10.1172/JCI29069 (2006). 163 Cohn, C. & Joseph, D. Influence of body weight and body fat on appetite of "normal" lean and obese rats. The Yale journal of biology and medicine 34, 598- 607 (1962). 164 Even, P. C. & Nadkarni, N. A. Indirect calorimetry in laboratory mice and rats: principles, practical considerations, interpretation and perspectives. American journal of physiology. Regulatory, integrative and comparative physiology 303, R459-476, doi:10.1152/ajpregu.00137.2012 (2012). 165 Oger, F. et al. Peroxisome proliferator-activated receptor gamma regulates genes involved in insulin/insulin-like growth factor signaling and lipid metabolism during adipogenesis through functionally distinct enhancer classes. The Journal of biological chemistry 289, 708-722, doi:10.1074/jbc.M113.526996 (2014). 166 Duncan, R. E., Ahmadian, M., Jaworski, K., Sarkadi-Nagy, E. & Sul, H. S. Regulation of lipolysis in adipocytes. Annual review of nutrition 27, 79-101, doi:10.1146/annurev.nutr.27.061406.093734 (2007). 167 Klionsky, D. J. et al. Guidelines for the use and interpretation of assays for monitoring autophagy. Autophagy 8, 445-544 (2012).

116

168 Kanazawa, T. et al. Amino acids and insulin control autophagic proteolysis through different signaling pathways in relation to mTOR in isolated rat hepatocytes. The Journal of biological chemistry 279, 8452-8459, doi:10.1074/jbc.M306337200 (2004). 169 Kaur, J. & Debnath, J. Autophagy at the crossroads of catabolism and anabolism. Nature reviews. Molecular cell biology 16, 461-472, doi:10.1038/nrm4024 (2015). 170 Soussi, H. et al. DAPK2 Downregulation Associates With Attenuated Adipocyte Autophagic Clearance in Human Obesity. Diabetes 64, 3452-3463, doi:10.2337/db14-1933 (2015). 171 Kovsan, J., Ben-Romano, R., Souza, S. C., Greenberg, A. S. & Rudich, A. Regulation of adipocyte lipolysis by degradation of the perilipin protein: nelfinavir enhances lysosome-mediated perilipin proteolysis. The Journal of biological chemistry 282, 21704-21711, doi:10.1074/jbc.M702223200 (2007). 172 Korolchuk, V. I., Mansilla, A., Menzies, F. M. & Rubinsztein, D. C. Autophagy inhibition compromises degradation of ubiquitin-proteasome pathway substrates. Molecular cell 33, 517-527, doi:10.1016/j.molcel.2009.01.021 (2009). 173 Xu, G. et al. Post-translational regulation of adipose differentiation-related protein by the ubiquitin/proteasome pathway. The Journal of biological chemistry 280, 42841-42847, doi:10.1074/jbc.M506569200 (2005). 174 Sztalryd, C. et al. Perilipin A is essential for the translocation of hormone- sensitive lipase during lipolytic activation. The Journal of cell biology 161, 1093- 1103, doi:10.1083/jcb.200210169 (2003). 175 Kaushik, S. & Cuervo, A. M. Degradation of lipid droplet-associated proteins by chaperone-mediated autophagy facilitates lipolysis. Nature cell biology 17, 759- 770, doi:10.1038/ncb3166 (2015). 176 Roczniak-Ferguson, A. et al. The transcription factor TFEB links mTORC1 signaling to transcriptional control of lysosome homeostasis. Science signaling 5, ra42, doi:10.1126/scisignal.2002790 (2012). 177 Guo, X. et al. VCP recruitment to mitochondria causes mitophagy impairment and neurodegeneration in models of Huntington's disease. Nature communications 7, 12646, doi:10.1038/ncomms12646 (2016). 178 Ritz, D. et al. Endolysosomal sorting of ubiquitylated caveolin-1 is regulated by VCP and UBXD1 and impaired by VCP disease mutations. Nature cell biology 13, 1116-1123, doi:10.1038/ncb2301 (2011). 179 Jager, S. et al. Role for Rab7 in maturation of late autophagic vacuoles. Journal of cell science 117, 4837-4848, doi:10.1242/jcs.01370 (2004). 180 Filimonenko, M. et al. Functional multivesicular bodies are required for autophagic clearance of protein aggregates associated with neurodegenerative disease. The Journal of cell biology 179, 485-500, doi:10.1083/jcb.200702115 (2007).

117

181 Razi, M., Chan, E. Y. & Tooze, S. A. Early endosomes and endosomal coatomer are required for autophagy. The Journal of cell biology 185, 305-321, doi:10.1083/jcb.200810098 (2009). 182 Lamb, C. A., Dooley, H. C. & Tooze, S. A. Endocytosis and autophagy: Shared machinery for degradation. BioEssays : news and reviews in molecular, cellular and developmental biology 35, 34-45, doi:10.1002/bies.201200130 (2013). 183 Kosacka, J. et al. Autophagy in adipose tissue of patients with obesity and type 2 diabetes. Molecular and cellular endocrinology 409, 21-32, doi:10.1016/j.mce.2015.03.015 (2015). 184 Ozcan, U. et al. Endoplasmic reticulum stress links obesity, insulin action, and type 2 diabetes. Science 306, 457-461, doi:10.1126/science.1103160 (2004). 185 Miranda, M. et al. Relation between human LPIN1, hypoxia and endoplasmic reticulum stress genes in subcutaneous and visceral adipose tissue. International journal of obesity 34, 679-686, doi:10.1038/ijo.2009.290 (2010). 186 Yoshizaki, T. et al. Autophagy regulates inflammation in adipocytes. Biochemical and biophysical research communications 417, 352-357, doi:10.1016/j.bbrc.2011.11.114 (2012). 187 Cinti, S. et al. Adipocyte death defines macrophage localization and function in adipose tissue of obese mice and humans. Journal of lipid research 46, 2347- 2355, doi:10.1194/jlr.M500294-JLR200 (2005).

118 VITA

YING LIU

EDUCATION

2009-2016 Ph.D. Pharmacology The Pennsylvania State University, Hershey, PA, USA 2009 M.A. Biochemistry University of Scranton, Scranton, PA, USA 2006 B.S. Biotechnology Sichuan Agricultural University, Ya’an, Sichuan, China Outstanding Graduate 2006

PEER-REVIEWED PUBLICATIONS

 Liu Y, Takahashi Y, Desai N, Zhang J, Serfass JM, Shi YG, Lynch CJ, Wang HG. Bif-1 deficiency impairs lipid homeostasis and causes obesity accompanied by insulin resistance. Scientific Reports, 2016 Feb 9; 6:20453.  Takahashi Y, Tsotakos N, Liu Y, Young MM, Serfass J, Tang Z, Abraham T, Wang HG. The Bif-1-Dynamin 2 membrane fission machinery regulates Atg9-containing vesicle generation at the Rab11-positive reservoirs. Oncotarget, 2016 Apr 12; 7 (15): 20855-68.  Platt A, Xia Z, Liu Y, Chen G, Lazarus P. Impact of nonsynonymous single nucleotide polymorphisms on in-vitro metabolism of exemestane by hepatic cytosolic reductases. Pharmacogenetics and genomics, 2016 Aug; 26 (8): 370-80.

MEETING ABSTRACTS & PRESENTATIONS

 Liu Y, Takahashi Y, Wang HG. ‘Loss of Bif-1 leads to obesity through suppression of lipid catabolism’ Penn State Hershey College of Medicine Graduate Student Forum. March 6, 2015. Hershey, PA. (Poster)  Liu Y, Takahashi Y, Desai N, Shi YG, Lynch CJ, Wang HG. ‘Loss of Bif-1/SH3GLB1 impairs lipid catabolism and leads to obesity’. Keystone Symposium on Systems Biology of Lipid Metabolism. February 9-13, 2015. Breckenridge, CO. (Poster; awarded ‘Future of Science Keystone Symposia Scholarship’)  Liu Y, Takahashi Y, Wang HG. ‘Role of Bif-1/SH3GLB1 in obesity’. Pediatric Research Day, Penn State Hershey Children’s Hospital. May 22, 2014. Hershey, PA. (Oral)  Liu Y, Takahashi Y, Wang HG. ‘Role of Endophilin-B1 in obesity’. Penn State Hershey College of Medicine Graduate Student Forum. March 2014. Hershey, PA. (Oral)

HONORS & AWARDS

2015 Future of Science Scholarship; Keystone Symposia 2007 Tuition Scholarship; University of Scranton 2007 Certificate of Achievement; Sanofi 2006 Outstanding Graduate; Sichuan Agricultural University

PROFESSIONAL MEMBERSHIPS

American Honors Chemistry Society