Membrane Partitioning by Flotillin-1 Facilitates Amphetamine-Induced Dopamine Transporter Activity

Membrane partitioning by Flotillin-1 facilitates amphetamine-induced dopamine transporter activity

Wendy Mei Fong

Submitted in partial fulfillment of the requirements for the degree of Doctor of Philosophy under the Executive Committee of the Graduate School of Arts and Sciences

COLUMBIA UNIVERSITY

2017

ABSTRACT

Membrane partitioning by Flotillin-1 facilitates amphetamine-induced dopamine transporter activity

Wendy Mei Fong

Cellular membranes were once considered static and passive structures, but are now appreciated as a fluidic and dynamic assembly of macromolecules that play an active role in cellular function. Membrane composition has been proposed to play a critical role in modulating protein function by affecting everything from post-translational modifications to conformation, but the physiologic relevance of the relationship between protein and membrane has been difficult to establish. For example, membrane-associated proteins such as Flotillin-1 (Flot1) have been implicated to scaffold proteins into cholesterol-rich membranes, as well as play a role in a wide array of functions such as endocytosis and axon pathfinding; however, genetic elimination of Flot1 expression had little to no reported consequence, leaving to question the physiologic importance of scaffolding proteins to membrane microdomains. Using genetic and biochemical approaches, I sought to understand how the immediate lipid environment can influence the function of a transmembrane protein, and how this might impact brain function. Specifically, I examined how a cholesterol-rich environment can affect the function of the cell surface neurotransmitter transporter for dopamine, the dopamine transporter (DAT), and how this interaction may influence the ability of an organism to respond to the psychostimulant amphetamine (AMPH).

Although neurotransmitter transporters (NTTs) such as DAT and the serotonin transporter

(SERT), have been predicted to reside in membrane rafts, it has been difficult to establish the role of microdomain localization in transporter function. DAT localizes to the plasma membrane, where it modulates the strength and duration of neurotransmission by clearing dopamine (DA)

from the perisynaptic space. Defects in DAT have been implicated in a range of psychiatric and neurological disorders, from schizophrenia to Parkinson’s disease. Additionally, as a target of psychostimulants, such as AMPH and cocaine (COC), the role of DAT in addiction is of societal interest.

Given that Flot1 was required for scaffolding heterologously expressed DAT to cholesterol-rich membranes in cell-based systems, and was selectively necessary for the non- exocytic release of DA through DAT in response to AMPH, I sought to test the hypothesis that the

Flot1-mediated membrane localization of DAT was significant for the ability of mice to respond to AMPH. To this end, I created a series of genetic models to determine how the presence of Flot1 impacts DAT function in the brain. I found that Flot1 is not only important for scaffolding DAT into cholesterol-rich membranes, but that the ability of DAT to partition into these membranes was necessary for DAergic neurons, DAT, and ultimately mice, to respond to AMPH. Given that the other parameters of DA neuron function, as well as the ability of the animals to respond to COC was unaffected by DAT partitioning, my findings demonstrate that AMPH and COC exert different mechanisms of action in vivo. Moreover, I found that the cholesterol-rich membrane environment promoted a conformation of DAT that was favorable for reverse transport of DA through DAT, namely increasing the ability of its N-terminus to bind to the phospholipid, PIP2. This dissertation provides the first glimpse into not only how membrane localization can affect protein conformation and function but also the physiologic relevance of these Flot1-dependent membrane microdomains in brain.

Table of Contents

List of Figures and Tables ...... iv

Acknowledgements ...... vi

Chapter 1: The Fluid Mosaic Model and Beyond ...... 1 1.1 Introduction ...... 1 1.2 The plasma membrane ...... 2 1.3 Lipid rafts ...... 8 1.4 Functions of lipid rafts ...... 11 1.4.1 Signal transduction ...... 11 1.4.2 Viral entry in cells ...... 14 1.4.3 Controversies of lipid rafts ...... 14 1.5 Cholesterol and protein activity ...... 16 1.6 Concluding remarks ...... 17 Chapter 2: Flotillin/Reggie Proteins ...... 18 2.1 Introduction ...... 18 2.2 Discovery of the flotillins ...... 19 2.3 Structure of Flotillin-1 ...... 20 2.4 PHB/SPFH family members ...... 20 2.4.1 Stomatin ...... 21 2.4.2 Podocin ...... 22 2.4.3 Prohibitin ...... 23 2.4.4 HflK and HflC ...... 23 2.5 Functions of flotillin ...... 24 2.5.1 Neuronal regeneration ...... 24 2.5.2 Neuronal development ...... 25 2.5.3 Cell-cell adhesion ...... 26 2.5.4 Endocytic trafficking ...... 27 2.5.5 Signal transduction ...... 30 2.5.6 Transporter activity ...... 31 2.6 Concluding remarks ...... 32 Chapter 3: The Dopamine Transporter (DAT) ...... 33 3.1 Introduction ...... 33 3.2 Structure ...... 34 3.3 Function ...... 37 3.4 Localization and distribution in the brain ...... 37 3.5 DA, DAT, and disease ...... 39 3.5.1 Parkinson’s disease...... 39 3.5.2 Schizophrenia ...... 40 3.5.3 Obsessive compulsive disorder ...... 41

3.5.4 Attention defecit hyperactivity disorder ...... 42 3.5.5 Autism spectrum disorder ...... 42 3.6 DAT and psychostimulants: cocaine and amphetamine ...... 43 3.6.1 Cocaine ...... 44 3.6.2 Amphetamine ...... 45 3.7 Concluding remarks ...... 50 Chapter 4: Determining the Role of Flot1-Mediated Partitioning of DAT on AMPH-induced DA Release ...... 51 4.1 Introduction ...... 51 4.2 The conditional loss of Flot1 in DAergic neurons leads to a diminished response to AMPH in mice ...... 53 4.3 The loss of Flot1 in DA neuron does not affect basal DA neuron function ...... 59 4.4 Flot1 is required in DA neurons for DAT partitioning into cholesterol-rich membranes ...... 60 4.5 Genetic approaches reveal protective compensatory events during development can mask the role of Flot1 in AMPH-induced non-exocytic release of DA and in scaffolding DAT to DRMs ...... 62 4.6 DAT activity influences the partitioning of DAT into detergent-resistant, cholesterol-rich membranes ...... 69 4.7 Partitioning of DAT into cholesterol-rich membranes and Flot1 is not required for DAT phosphorylation ...... 70 4.8 DAT localized to cholesterol-rich membranes are found in a conformation distinct from DAT localized in cholesterol-poor membranes ...... 75 4.9 The conformation of DAT in DRMs depends on its N-terminal interaction with PIP2 .... 78 4.10 The working model of membrane-DAT interaction for AMPH-induced efflux of DA through DAT ...... 80 Chapter 5: General Discussion ...... 82 5.1 Overview ...... 82 5.2 The membrane environment of DAT and AMPH-induced efflux ...... 84 5.2.1 The membrane environment influences DAT function ...... 84 5.2.2 Phosphorylation of DAT ...... 87 5.2.3 AMPH-induced internalization of DAT ...... 87 5.2.4 DRMs versus membrane microdomains ...... 88 5.3 The role of Flot1 ...... 90 5.3.1 The DAT-Flot1 interaction ...... 90 5.3.2 Oligomerization of Flot1 ...... 91 5.3.3 Developmental compensation in Flot1 KO mice ...... 92 5.4 The physiologic relevance of AMPH-induced efflux ...... 93 5.4.1 DA uptake ...... 93 5.4.2 DAergic consequences of drug abuse: sensitization and neurotoxicity ...... 94 5.4.3 Flot1 as a therapeutic target ...... 96 5.5 Methodological considerations regarding mouse models ...... 97 5.6 Conclusions ...... 97 Chapter 6: Materials and Methods ...... 99 6.1 Antibodies ...... 99 6.2 Plasmids ...... 99

6.3 Transfections ...... 100 6.4 Cell culture ...... 100 6.5 Animals ...... 100 6.6 Mouse lines ...... 101 6.6.1 Conditional Flotillin-1 alleles (Flot1 flox/flox) ...... 101 6.6.2 DATiresCre (DIC) ...... 101 6.6.3 Flotillin-1 conditional knockout (Flot1 cKO) ...... 102 6.6.4 HprtCre/+ ...... 102 6.6.5 Flotillin-1 constitutive knockout (Flot1 KO) ...... 102 6.6.6 Tamoxifen-inducible Cre (ActinCreERTM/+) ...... 103 6.6.7 Flotillin-1 inducible knockout (Flot1 iKO) ...... 103 6.7 PCR Genotyping ...... 103 6.7.1 DNA isolation ...... 103 6.7.2 PCR ...... 104 6.8 Drug Preparation and Administration ...... 106 6.8.1 Amphetamine ...... 106 6.8.2 Cocaine ...... 106 6.8.3 Tamoxifen ...... 106 6.9 Mouse Behavioral Testing ...... 107 6.9.1 Open Field ...... 107 6.10 Electrophysiology ...... 108 6.10.1 Fast scan cyclic voltammetry (FSCV) ...... 108 6.10.2 High speed chronoamperometry (HSCA) ...... 109 6.11 Immunohistochemistry ...... 110 6.12 Quantification of DA levels ...... 111 6.13 Sucrose density gradients ...... 111 6.13.1 Striatal tissue preparation ...... 111 6.13.2 Striatal synaptosome preparation ...... 112 6.13.3 Cell culture ...... 113 6.13.4 Western blots ...... 113 6.14 Limited proteolysis ...... 113 6.15 Immunoprecipitation ...... 114 6.15.1 Phosphorylation ...... 114 6.16 Immunofluorescence ...... 115 6.17 Dissection and immunoblotting of substantia nigra ...... 115 6.18 Statistical analyses and figure creation ...... 116 References ...... 117 Appendix ...... 137

iii

List of Figures and Tables

Chapter 1

Figure 1.1: Davison and Danielli’s Paucimolecular Model ...... 4

Figure 1.2: Singer and Nicholson’s Fluid Mosaic Model ...... 6

Figure 1.3: Simons and Ikonen: Model for the organization of rafts and caveolae in the PM ...... 8

Chapter 2

Figure 2.1: Model of flotillin microdomains ...... 18

Figure 2.2: Structural comparison of SPFH family members ...... 21

Chapter 3

Figure 3.1: Amino acid sequence of human DAT ...... 33

Figure 3.2: A cholesterol site in Drosophila DAT ...... 36

Figure 3.3: The mechanism of action for cocaine (COC) and amphetamine (AMPH) ...... 43

Figure 3.4: Chemical structures of amphetamine and its related compounds ...... 46

Chapter 4

Figure 4.1: Design of the Flot1 conditional allele and resulting recombined allele of the mice ...53

Figure 4.2: Elimination of Flot1 in DA neurons diminishes the ability of mice to respond to AMPH in vivo: The Flot1 cKO mice ...... 56

Figure 4.3: Locomotor response in female mice shows a similar Flot1-dependence as males, with a diminished response to AMPH, and no difference in basal behavior or response to COC ...... 57

Figure 4.4: Cyclic voltammetry in slice preparation from Flot1 cHet and cKO mice ...... 58

Figure 4.5: Chronoamperometry in striatal slice preparations from the Flot1(DAT) cKO mice demonstrate significantly diminished AMPH-induced release of DA through DAT ...... 58

Figure 4.6: The loss of Flot1 leads to no detectable changes in the DA system ...... 59

Figure 4.7: HPLC reveals normal DA levels in the striata of Flot1 cKO mice ...... 60

Figure 4.8: Cyclic voltammetry in striatal slice preparations also found no difference in basal DA release and DAT-mediated uptake between cHet and cKO mice ...... 60

Figure 4.9: Sucrose density gradients of DAT striatal lysates from Flot1fl/fl mice in the absence or presence of methyl-beta-cyclodextrin (MßC) ...... 61

Figure 4.10: Sucrose density gradient centrifugation reveals a differential distribution of DAT in synaptosomes from Flot1 cKO striata ...... 61

Figure 4.11: Sucrose density gradient centrifugation reveals that DAT is insoluble in Brij 58 but soluble in Triton X-100 ...... 62

Figure 4.12: Protective compensatory events during development can mask the role of Flot1 in AMPH-induced non-exocytic release of DA and in scaffolding DAT to DRMs ...... 64

Figure 4.13: Adult-inducible deletion of Flot1 ...... 66

Figure 4.14: DAT from Flot1 iKO but not Flot1 KO mice fails to partition into DRMs ...... 68

Figure 4.15: DAT activity modulates the partitioning of DAT in cholesterol-rich membranes ....70

Figure 4.16: Phosphorylation of DAT occurs independently of Flot1 ...... 72

Figure 4.17: Flot1 protects phosphorylation of DAT ...... 73

Figure 4.18: Working model 1 ...... 74

Table 4.1: Cleavage sites of specified proteases ...... 75

Figure 4.19: Limited proteolysis reveals that DAT in DRMs is in a distinct conformation ...... 76

Figure 4.20: The interaction of PIP2 with the N-terminus of DAT is important for DAT partitioning into DRMs ...... 77

Figure 4.21: PIP2 interaction with the N-terminus of DAT is necessary for the DRM-dependent conformation of DAT ...... 78

Figure 4.22: Model of how Flot1 facilitates the AMPH-dependent reverse transport of DA through DAT ...... 79

Chapter 6

Table 6.1: Primers and conditions for PCR ...... 105

Acknowledgements

First and foremost, I would like to express my deepest gratitude to my advisor, Dr. Ai Yamamoto.

She has trained me with a scientific rigor that is unparalleled, providing guidance and invaluable feedback that has shaped my development as a scientist. Thank you for viewing my weaknesses as an opportunity for improvement and committing countless hours to my scientific growth. Ai’s immense breadth of knowledge and her dedication to her work is a source of inspiration. The joy and enthusiasm she has for research has made this Ph.D. pursuit an enjoyable one. Ai is friendly and patient, but above all, she is understanding and compassionate. Thank you for being supportive of the small victories at the bench, but especially when my Ph.D. experience and life were at its most stressful.

I would like to thank my thesis committee members, Drs. Jonathan Javitch, Elizabeth Miller, and

David Sulzer for their time and constructive criticism in shaping this project. I would like to extend a special thanks to Drs. Thomas Melia and Margaret Rice for serving as outside examiners.

The electrophysiology studies discussed in this dissertation would not have been possible without the Galli and Mosharov labs. I would like to thank Dr. Aurelio Galli for always being generous with reagents and welcoming me into his lab. Members of the Galli lab have been instrumental in the progress of my thesis. Dr. Kevin Erreger not only taught me how to perform slice biotinylation, but also helped perform high speed chronoamperometry on striatal slice preparations. I would also

vi like to thank Dr. India Reddy for all of her contributions to this project. I am particularly indebted to Dr. Se Joon Choi from the Mosharov lab for performing fast scan cyclic voltammetry on slices and for patiently explaining and re-explaining to me general electrophysiology. Drs. Jonathan

Javitch and Caline Karam have also been generous in providing cell lines and antibodies.

The past and present members of the Yamamoto lab have contributed immensely to my personal and professional time at Columbia. Thank you for providing a source of friendship and collaboration. I would also like to thank all whom I have befriended on the third floor of Black.

You were all an extension of our lab and I truly appreciate all of the laughs, advice, and of course, food and snacks, that kept me going throughout my graduate pursuit. It was a pleasure working alongside all of you.

I would also like to mention my family, whom always stressed the importance of education and instilled in me to work hard and play hard. They are truly the driving force behind all of my big endeavors. A big thank you to the Chan family. They adopted me for the holidays and always made sure I returned home with a lot of leftovers. I am grateful to be surrounded by caring people no matter where I go.

Last but not least, thank you to my closest friends, including Chandler Walker, José Martínez,

Susumu Antoku, and YoungJoo Yang. You have been there for various parts of this journey and I truly appreciate all the support and laughter you have brought to my life. All of you have truly enriched my graduate experience.

vii

To my grandfather

viii

Chapter 1

The Fluid Mosaic Model and Beyond

1.1 Introduction

The plasma membrane is an essential structure that allows molecules to travel in and out of cells, creates an internal environment for proper organelle function, protects cells from the outside environment, and functions in cell signaling and communication. Closely associated to the lipid bilayer are proteins that provide function at the membrane: Transmembrane proteins span the bilayer, peripheral proteins interact with a single leaflet on the cytosolic or extracellular side, whereas other proteins associate indirectly through transient associations to other membrane proteins. How all of these proteins interact with their lipid surroundings is only recently being appreciated to have a significant impact on protein function. The original membrane studies have focused solely on the role of lipids in the membrane or the function of membrane proteins.

However, since the proposal of the fluid mosaic model, studies began to shift toward understanding protein-lipid dynamics and how their interaction as well as the role of lipids can govern protein activity.

The understanding that a heterogeneous population of lipids can form discrete rigid clusters of membrane surrounded by more fluid membrane led to the proposal of the "lipid raft” hypothesis,

1 which has opened an intense debate on the biochemical and structural definition of these membrane compartments and how they might influence protein function. The current consensus is that membrane rafts are dynamic, cholesterol- and sphingolipid-enriched domains that scaffold proteins necessary for cellular processes. Whereas a number of functions have been ascribed to these microdomains, work in this field has stalled due to limitations in detection and questions of their physiologic relevance to whether these structures indeed exist.

In this chapter, I will first provide a brief history of the plasma membrane and the key findings that led to the development of the fluid mosaic model. Next, I will describe the evidence in support of the lipid raft hypothesis, discuss several proposed functions of lipid rafts, and address the controversies. Finally, I will touch upon the important role of cholesterol in protein function.

1.2 The plasma membrane

The plasma membrane (PM) as we know it today did not begin to solidify as a concept until centuries after Robert Hooke’s classic observation described in Micrographia dating back to 1665.

Here he detailed his findings under the microscope and he first coined the term “cells” after noticing cavities, comparable to honeycombs, in a piece of cork [1]. At a time when the resolution of the microscope was limited to the certainty of plant cell walls, the term “cell” was the first suggestion of the existence of a border enclosing a biological space. Due to difficulties in visualizing membranes in animal cells in addition to other exciting scientific discoveries in the

1800s (such as the proposal of Darwinian evolution to the discovery of mitosis), the PM was considered an unessential secondary structure and efforts toward its understanding was largely neglected.

Early osmotic studies laid a foundation for the physiologic relevance of this observed structure, leading to the departure from the idea that a cell was a protoplasmic colloid, in which the cell border was simply the edge of the protoplasm, to wider acceptance of the existence of a cell membrane. In the early 1800s, Henri Dutrochet contributed to the understanding of osmosis by ascribing significance to the process. He concluded that water moved across an animal membrane from one compartment to another due to a difference in the heterogeneity of the liquids within the compartments. Based on this, he rationalized that plant cells utilized osmosis to accumulate water, resulting in turgescence [2]. Subsequent studies by Carl Nägeli and others demonstrated that plant cells placed in hypertonic media led to the shrinking of a vesicle within the cell walls whereas

Hugo de Vries and Hartog Hamburger used both plant and animal cells to show that most solutions at equal concentrations yielded equal osmostic pressures. Though osmosis draws on the premise of a semi-permeable membrane, the prevailing view at the time was that the cell “membrane” was not a distinct structure, but rather an interface formed by the interaction between the cell protoplasm and the surrounding fluid. It was this interface that Nägeli, de Vries, and Hamburger, among others, believed osmosis to be occurring.

There was a need for direct evidence to support the existence of cell membranes. In the late 1890s,

Ernest Overton placed cells in 500 different solutions of the same concentration to study cell permeability in the presence of different molecules [3]. He noticed that cells did not shrink in solutions containing apolar molecules compared to those containing polar substances, which led him to conclude that the former entered cells more easily than the latter. Overton established that the permeability of molecules depends on their polarity, suggesting that cell membranes would be at least in part composed of lipids. With the understanding that the role of the plant cell wall was independent from this observation, he proposed that there exists a cell membrane distinct from the

3 cell wall. In 1922, with the refinement of microinjection techniques, Robert Chambers applied a hydrophilic cytolysogenic substance to the surface of starfish eggs, which did not harm the cytoplasm, and then to the cell interior, which caused massive digestion of the cytoplasm [4]. This experiment established that the cell surface was distinct from the protoplasm. Rudolf Höber’s studies on the electric conductivities of membranes in muscle indirectly affirmed the presence of cell membranes [5]. He showed that high electrical frequencies caused muscle cells to have higher conductivity, indicating that cells were disrupted leading to increased permeability. These results suggest that electrolytes are bound within a cell by an impermeable boundary that differs from the protoplasm.

In 1925, Gorter and Grendel achieved a breakthrough in the structure of membranes by determining the amount of surface lipids that can occupy the total surface of erythrocytes. Given that these cells do not contain internal membrane structures, any lipid extracted should theoretically be derived only from the cell surface. They found a ratio of 2:1 and deduced that cells were bounded only by a lipid bilayer [6]. These studies clearly indicated that lipids were a primary feature of the PM.

Figure 1.1. Schematic of the Paucimolecular Model of Danielli and Davson. In this model, two protein layers (purple) were proposed to sandwich the lipid bilayer that was revealed to be a key component of the PM by Gorter and Grendel in 1925. Taken from Lombard, J. Biology Direct (2014) 9:32.

The next step leading to our current understanding of the PM composition came soon after, which revealed that lipids may not be the only component of membranes, and that proteins may also reside there [7-9]. These studies culminated in the proposal of the Paucimolecular model of the

PM in 1935, for which Danielli and Dayson proposed that the lipid bilayer was sandwiched between two layers of proteins [10]. A weakness of the Paucimolecular model, however, was that much of the conclusions were based on studies using triglyceride oils and other non-miscible lipids, which are different from natural amphipathic membrane constituents. Later studies by

Danielli and others showed that adding fatty acids, cholesterol, and phospholipids to non-miscible solutions lowered the interfacial tension between aqueous and hydrophobic media [11, 12], showing that a protein layer could not be responsible for the decrease in surface tension.

Nonetheless how polar molecules entered the cells still remained unclear. Ongoing studies in molecular transport would confirm the presence of proteins in the PM and provided insight into the membrane structure.

Ion concentrations (sodium, potassium, etc.) were known to be different between the cell interior and its environment [13], but how this asymmetry was achieved was unclear. Some thought that the ions were simply bound to the cell colloid, whereas other believed that some ions were membrane impermeable or that active transport stabilized the ion concentrations [14, 15]. In 1941,

Dean postulated that ion movement against a gradient may be due to active transport by a pump

[16], and others suggested that these pumps were proteinaceous [17-19]. The landmark work of

Hodgkin and Huxley established that the movement of ions occurred at the membrane [20]. Then

Skou and others not only elucidated the mechanism of ion transport, but were also instrumental in establishing that proteins required access to both sides of the membrane and that they have an

5 active function [21]. Overall, these findings led to the general acceptance that proteins are not only a major component of cell membranes, but they are an integral component of these structures.

The next major technical advance that revolutionized the study of cells and membranes was the advent of electron microscopy (EM). EM allowed for direct observation of membrane cross sections, and scientists noted that the membranes looked like “railroad tracks,” with two dense lines bordering a lighter center. In 1962, Mueller et al. created the first artificial lipid bilayer [22], recreating the “railroad tracks” and definitively proving that the dense tracks were the consequence of the bilayer itself and not an independent layer of protein [23]. Studies in 1963 by Moor and

Mühlethaller irrevocably elucidated the structure of the membrane to be a mosaic of lipids and proteins [24].

Figure 1.2. The fluid mosaic model. Original figure from Singer and Nicholson (1972) Science. This model depicts a membrane cross section with integral protein embedded within the phospholipid bilayer or spanning the membrane entirely.

As the composition of the PM came into view, scientists continued to address a long-standing question regarding the lipid bilayer: its fluidity [25-27]. Whether the membrane was in more solid or fluid phase was posed as early as the studies leading to the Paucimolecular model. A key study by Frye and Edidin ultimately demonstrated that the membrane was indeed fluid, permitting the movement of protein associated with it [28]. By fusing cells of two different species, they tracked the cell surface antigens associated to each cell type to determine whether they remained separate or they mixed. Frye and Edidin famously demonstrated how the cell surface antigens mixed,

6 leading to the simple proof that the membranes were indeed fluid and that lipid mixing and the diffusion of proteins could freely occur.

In 1972, Singer and Nicolson proposed the fluid mosaic model, providing a major breakthrough in our understanding of the structure of biological membranes [29]. It describes what we now know about the PM: it is composed of a fluid phospholipid bilayer so that the interspersed proteins can diffuse rather than stay in place. The physico-chemical considerations of the fluid mosaic model also make it biologically persuasive. The formation of the membrane by phospholipids results in a very stable structure, in which the amphipathic nature of the bilayer governs interactions with not only transmembrane proteins but also asymmetric components, such as peripheral proteins, glycoproteins, and other lipids. Due to the comprehensive explanation and solid evidence in support of the fluid mosaic model, it remains largely undisputed to this day.

Whereas the fluid mosaic model has ascribed a rather inert role to the membrane, it has recently come to light that membrane proteins interact dynamically with the lipids in their surroundings.

Recent evidence suggests that the membrane is more mosaic than fluid. Moreover, it is not simply two-dimensional, but based on the lipid composition, the architecture of the membrane can change, leading to membrane patches that separate regions of structure and function [30]. In fact, protein function can be affected by bilayer thickness, bending stiffness, and membrane curvature [31, 32].

The more active role of lipids in membrane protein function is highlighted by the flexible surface model (FSM), which credits the deformation of the membrane for changes in protein conformation

[33]. Additionally, the membrane does not affect one protein at a given moment. The shape of a protein confers whether it will attract or repel other membrane proteins, thereby spatially and fully utilizing the available energy of the membrane [34-37]. On the one hand, proteins that confer similar bilayer thickness will attract each other, but those with different membrane requirements

7 will repel each other [38]. Thus, the membrane and its proteins dynamically cooperate to modulate the structure and function of membrane proteins.

1.3 Lipid rafts

According to the fluid mosaic model, the plasma membrane is a milieu of proteins and lipids – proteins that provide functionality and lipids that passively form the fluid bilayer. However, we

Figure 1.3. Model for the organization of rafts and caveolae in the plasma membrane from Simons and Ikonen (1997) Nature. The rafts (red) are discrete regions within the PM (blue), that has a different organization of intercalated cholesterol from that in the rafts. a. GPI anchors, acyl tails can associate proteins to rafts, as well as transmembrane domains directly. b. asymmetry of the rafts based on the presence of sphingomyelin and glycosphingolipids (red) enriched in the outer leaflet, and glycerolipids (green) in the inner leaflet. Cholesterol (gray) can be found in both leaflets. c. caveolae (gray) self- associate to form caveolae.

8 now know that these same lipids can exist in phases with varying levels of fluidity: gel, liquid- ordered, and liquid disordered. The heterogeneity of the membrane along with the understanding of a more dynamic role for membrane lipids brought forth the concept of lipid rafts.

The idea of lipid rafts originated from studies utilizing biophysical methods to understand the effects of temperature on membrane behavior. Stier and Sackmann observed a difference in spin label kinetics between a substrate and enzyme surrounded by distinctive lipid populations depending on temperature [39]. This led them to believe that there is an arrangement of lipids that leads to the presence of rigid “clusters” within the larger, fluid membrane. These clusters were soon refined to be “quasicrystalline” structures situated next to more “freely dispersed liquid crystalline lipid molecules” [40, 41]. Further work using x-ray diffraction showed that there is a discontinuous phase separation from a fluid part of the membrane to one that is rigid, leading these

“rigid liquid crystalline” clusters to be more aptly described as “lipids in a more ordered state”

[42]. Karnovsky and colleagues saw that the probe 1,6-diphenyl-1,3,5- hexatriene varied in its decay in lipid systems and biological membranes, indicating the presence of multiple lipid phases in the membrane [43]. These studies collectively demonstrated that different lipid states exist in the membrane. This led Karnovsky et al. to formalize the concept in 1982 by calling these ordered clusters “lipid domains.”

Originally proposed in 1997 as a platform to stabilize proteins for signal transduction by Simons and Ikonen [44], the definition of lipid rafts have been refined to be, “small (10–200 nm), heterogeneous, highly dynamic, sterol- and sphingolipid-enriched domains that compartmentalize cellular processes.” [45] Compared to the surrounding bilayer, rafts tend to have an elevated amount of cholesterol and sphingolipids, allowing the raft contents to pack more tightly together.

Sphingolipids associate with one another, leaving spaces between the head groups or saturated

9 acyl chains, which are filled by cholesterol. Although sphingolipids and cholesterol are always emphasized when defining a raft, it should be noted that proteins also contribute to the liquid-order environment. Proteins that can reside in rafts often include modifications that allow them to associate to cholesterol-enriched membranes. Such modifications include GPI

(glycophosphatidylinositol)-anchored proteins, doubly acylated proteins (e.g. Src-family kinases or the α-subunits of heterotrimeric G proteins) and cholesterol-linked and palmitoylated proteins

(e.g. Hedgehog). Raft-associated proteins are not limited to membrane-associated proteins, but can also include membrane-spanning transmembrane proteins, especially once they are modified by groups such as palmitate [46-49]. Rafts are also proposed to enrich in specific enzymes, such as kinases, phosphatases, palmitoylases, and depalmitoylases. Rafts contain a combination of lipid- lipid, protein-lipid, and protein-protein interactions. The raft bilayer gains its heterogeneity not only by its protein and lipid composition, but also by the asymmetric orientation of its contents.

Sphingolipids reside in the exoplasmic leaflet whereas glycerolipids are more likely to be on the cytoplasmic leaflet [50, 51]. Along with the hydrophobic regions of proteins that are available for interactions, this could dictate how cholesterol may insert itself into the raft and contribute to the rigidity of the microdomain. This optimization of packing and organization of molecules leads to a more ordered domain that is not described by the fluid mosaic model.

Currently, the most recognized types of lipid rafts are caveolae. Caveolae are small invaginations of the membrane and are marked by the presence of caveolin, which is a palmitoylated integral protein that has an affinity for cholesterol [52, 53]. Caveolins and therefore caveolae are found widely expressed. However, within the nervous system, caveolae are present in brain microvessels, endothelial cells, astrocytes, oligodendrocytes, Schwann cells and peripheral neurons, but are absent in most neurons and neuroblastoma cells [54]. In contrast, neurons contain planar

10 membrane rafts, which, unlike caveolae, lack distinguishing structural features and fall in the same plane as the rest of the membrane, rendering them difficult to observe. Planar membrane rafts are often marked by a membrane-associated protein called Flotillin-1, which is analogous to caveolins

[55]. Flotillin will be discussed further in Chapter 2.

Although caveolins and flotillins are recognized to be canonical raft proteins, the classic observation of raft proteins were GPI-anchored proteins, proteins with post-translational glycolipids attached to the C-terminus. GPI-anchored proteins were initially found to be present in Triton X-100-insoluble membrane fractions [56]. Since rafts are tightly packed due to the presence of sphingolipids and cholesterol, they are insoluble in cold, non-ionic detergents, which is how they are biochemically defined. Along with GPI-anchored proteins, caveolins and flotillins are also present in detergent-resistant membranes. In fact, they are reliably used as raft markers because not only have they shown detergent resistance, but they have also been consistently identified in proteomic analyses of lipid rafts [57-61].

1.4 Functions of lipid rafts

Whereas a number of cellular processes have been attributed to lipid rafts, I will be discussing the two most studied functions including signal transduction and viral cell entry.

1.4.1 Signal transduction

Signal transduction is the most widely accepted function of lipid rafts. It was discovered that crosslinked GPI-anchored proteins trigger a signaling cascade across the membrane, leading to calcium flux among other downstream events [62-66]. How this occurs remains unknown since

GPI-anchored proteins do not traverse the membrane. It has been proposed that GPI-anchored proteins may interact with a transmembrane protein that consequently transmits a signal to Src- family kinases on the other membrane leaflet [63, 67]. Since these early studies hinting at the role of rafts in signal transduction, more signaling pathways have been identified by crosslinking studies, two of the most extensively studied being immunoglobulin E (IgE) and T cell antigen receptor (TCR) signaling.

The most convincing studies implicating lipid rafts in signal transduction pertain to IgE signaling during an allergic immune response [68-70]. The cascade begins with the Fc segment of IgE binding to high-affinity receptors called FcεRI localized to the plasma membrane of mast cells and basophils. The FcεRI receptor is a tetramer of one α-, one β- and two γ-chains [70]. IgE binds to the α-chain and the β- and the γ-chains contain immune receptor tyrosine-based activation motifs

(ITAMs), which are shared among multi-subunit immune recognition receptors. Crosslinking studies have shown two or more FcεRI receptors recruits Src-like tyrosine kinase Lyn, which possibly phosphorylates the ITAMs [68, 69]. This initiates subsequent phosphorylation events and recruitment of signaling proteins that ultimately lead to increased calcium levels near the membrane, triggering granules to release histamines and inducing an allergic reaction.

Though IgE signaling was originally thought to be a result of protein–protein interactions, it was discovered that rafts may be a contributing factor [71]. FcεRI is soluble in Triton X-100 under steady state conditions, but insoluble in low concentrations of Triton X-100 after crosslinking [68].

Moreover, FcεRI crosslinking causes raft proteins, gangliosides and GPI-anchored proteins, to assemble into a patch that is visible under fluorescence microscopy, indicating that raft clustering occurs after FcεRI receptor activation [72, 73]. This was confirmed by the use of methyl-β- cyclodextrin (MβC), in which cholesterol depletion abolished IgE signaling [74]. These studies

12 demonstrate that lipid rafts serve as a site for the necessary molecular components to localize in order to initiate an allergic response.

The T cell antigen receptor utilizes lipid rafts to scaffold proteins and adaptors to initiate a signaling cascade that results in the formation of an immunological synapse, an interface between antigen presenting cells and T cells, at the T cell surface [75, 76]. The TCR is composed of αβ- heterodimers, which associate complexes containing ITAM motifs, which are phosphorylated by

Src-like tyrosine kinases, Lyn and Fyn. A kinase called ZAP-70 then binds to the phosphorylated

ITAMs and phosphorylates LAT, a transmembrane protein responsible for coupling TCR activation to downstream signaling events, including the recruitment of GPI-linked proteins and accessory molecules that help amplify the signal and phosphatases that turn the pathways on or off

[77-85]. This signal is critical for raft clustering, which serves to direct actin and microtubule networks as well as membrane trafficking to where the immunological synapse is being formed

[75, 76, 86, 87].

Similar to the FcεRI receptor, TCR complexes are not localized to rafts under basal, non-activated conditions [82, 88]. Once multiple TCRs are crosslinked, they become partially insoluble in detergent. Moreover, when the ITAM-containing complexes are crosslinked together, TCR signaling is activated [84]. Many of the signaling components, such as ZAP-70, Vav, PLCγ, Grb2 and PI3K, along with a hyperphosphorylated TCR complex, become insoluble in detergent following activation, suggesting movement into lipid rafts at the initiation of TCR signaling [83].

Application of MβC causes these proteins to fall out of the rafts and prevents any signal amplification, confirming that the proteins involved reside in lipid rafts [80, 83].

1.4.2 Viral entry in cells

Viruses hijack normal cell functions for survival and evidence suggests that viruses can utilize lipid rafts for entry into cells. The simian forest virus was reported to require sphingolipid and cholesterol to fuse with endosomal membranes, and the simian virus 40 (SV40) uses caveolae to enter mammalian cells [89-91]. In 2001, Helenius and colleagues showed that SV40 colocalizes with Cav1-GFP at the permissive temperature 37°C [92]. Moreover, using EM, they show that

SV40 particles are sequestered into caveolae and use caveolar endocytosis to enter cells. However, in their follow-up paper, they found that both genetic knockout (KO) of Cav1 and cells that naturally lack Cav1 still internalize SV40 at the permissive temperature 37°C [93]. This loss of a phenotype in the Cav1KO cells brings into question the necessity for membrane rafts if SV40 can find an alternative infectious route in the absence of caveolae.

1.4.3 Controversies of lipid rafts

The weak evidence supporting the role of lipid rafts in its various implicated functions reveals some of the controversies surrounding rafts and brings into question whether they are physiologically relevant or even exist. Contributing to the controversy is the primary methodologies used to study these membrane microdomains. The most widely used assay to detect membrane rafts utilizes non-ionic detergent extraction at 4°C followed by density gradient centrifugation. Given that detergent molecules insert into the lipid bilayer to form micellar fragments, the tight order due to the presence of sterols and sphingolipids makes the raft- membranes resistant to solubilization. Thus membranes that remain intact and their associated constituents are associated to rafts. The main critique of this method is that nonphysiological rearrangments of the bilayer could occur under different detergents or temperatures, giving rise to

14 experimental artifact. An often used control is the use of methyl-beta-cyclodextrin and other compounds that disrupt or deplete cholesterol and its packing. If the protein and membrane becomes soluble following MβC treatment, then it strengthens their presence in detergent-resistant membranes (DRMs). Nonetheless, given that cholesterol depletion will alter physical properties of the bilayer, these experiments are far from conclusive.

In vivo observation of rafts would be considered more empirical evidence, but the current resolution of light microscopy cannot capture the nanometer size of these microdomains. Studies utilizing microscopy lead to variable results. Crosslinking has allowed for the observation of patches within the membrane, but the colocalization of proteins and the sizes of the observed patches fluctuated, leading to skepticism of their significance. Fluorescence recovery after photobleaching (FRAP) analysis on membrane protein diffusion and Förster resonance energy transfer (FRET) studies looking at molecular interactions also yielded inconsistent results. These problems have since been clarified with the advent of improved microscopy techniques that allows one to observe the dynamic nature of the membrane. Hetero- and homo-FRET and fluorescence polarization anisotrophy studies have shown that GPI-anchored proteins and other proteins with lipid modifications can partition into cholesterol-dependent nanoscale assemblies [94, 95].

Moreover, total internal reflection fluorescence (TIRF) microscopy with single quantum dot tracking demonstrated that GPI-anchored proteins migrate in and out of ganglioside GM1 clusters.

Additionally, fluorescence correlation spectroscopy (FCS) analysis established that this partitioning occurs within a timescale of tens to hundreds of milliseconds [96, 97]. Super- resolution imaging by use of stimulated emission depletion (STED), photoactivated localization microscopy (PALM), and stochastic optical resconstruction microscopy (STORM) have shown in live cells that GPI-anchored proteins and sphingolipids are transiently found in cholesterol-

15 dependent complexes [98-101]. Advances is microscopy have visually clarified the existence of lipid rafts and their lifespan. Although they are becoming less elusive, much work remains to be done. Methods used to study lipid rafts have not changed dramatically and the concept of lipid rafts has not significantly evolved. It continues to be accepted as a domain where biological phenomena occur, but explanations for lipid-protein interactions within lipid rafts is still lacking.

1.5 Cholesterol and protein activity

Previous studies have focused solely on lipid dynamics or membrane protein structure and function. One important lipid-protein interaction that has been recently described focuses on the role of cholesterol on the dopamine transporter (DAT), which will be discussed further in Chapter

3. The crystal structure of the Drosophila melanogaster dopamine transporter (dDAT) has a cholesterol molecule situated in a groove formed by transmembranes 1a, 5, and 7 [102], suggesting that lipids within the protein, as opposed to lipids in the surrounding milieu, also play a role in stabilizing a specific conformation. Moreover, studies have shown that cholesterol plays an important role in DAT function. Cholesterol has been reported to be necessary for the outward- facing conformation of DAT, which increases binding of cocaine analogs [103]. Moreover, MβC was reported to reduce dopamine uptake, suggesting that dopamine uptake is cholesterol- dependent [104, 105]. Together, these studies demonstrate that possibly through a direct interaction with DAT, cholesterol is important for DAT function. In Chapters 4 and 5, we will further discuss other means by which cholesterol influences DAT function, especially in the presence of the psychostimulant AMPH.

1.6 Concluding remarks

The membrane field has made significant advances from Hooke’s initial observation to Singer and

Nicolson’s proposal of the fluid mosaic model. The definitive evidence supporting the existence of a lipid bilayer was instrumental in defining the cell membrane and further investigation proved that the bilayer was not an inert, but rather, dynamic structure, laying forth the concept of lipid rafts. For decades, membrane proteins and lipids were each studied in isolation within the context of rafts. However, recent evidence has established a growing appreciation for the understanding of the biological complexity that comprises of protein-lipid interactions within the membrane.

Examining how the membrane environment of lipid rafts affects protein function will ascribe physiologic significance to these membrane microdomains. Although studies in cells have been constructive in assigning functions for these domains, their relevance to basal functions and pathology remains controversial. It is possible that the current model of lipid rafts primarily functioning as signaling platforms has narrowed our understanding of their true biological function and significance. Therefore, concrete in vivo studies are not only necessary to prove their existence, but will also help to shape and improve our understanding of lipid rafts.

Chapter 2

Flotillin/Reggie Proteins

Figure 2.1 Model of flotillin microdomains. The N-terminal SFPH domain allows the flotillins to associate to the inner membrane leaflet. The C-terminal alpha-helices, which contain alanine and glutamine acid repeats, allow for homo- or hetero-oligomerization. The flotillins have been implicated in many cellular processes and have a number of interaction partners, some of which are shown above. Original figure from Rivera-Milla et al. (2006).

2.1 Introduction

Flotillin-1 (Flot1), also known as Reggie-2, is a highly conserved protein that is widely expressed, especially in the nervous system [106, 107]. Flot1 is a membrane-associated protein containing a conserved prohibitin homology (PHB) domain at its N-terminus and an alpha-helix that may be

18 involved in either multimerization or other protein-protein interactions at its C-terminus [108].

Although Flot1 has been described in numerous cellular processes, the exact mechanism by which it exerts its function is unknown. However, Flot1, which was named for its ability to float into buoyant detergent resistant membranes (DRMs) [109], is best known as a membrane raft scaffold and is most frequently used as a membrane raft marker.

Although Flot1 is often used to delineate the presence of membrane rafts, its significance remains unclear. In the following chapter, I will open with how the flotillins were discovered followed by a description of its structure. I will then touch upon the SPFH family members to show their similarities yet how different they are based on their diverse roles. Lastly, I will describe the many functions of Flot1 and its family member, Flot2.

2.2 Discovery of the flotillins

Flotillin-1 (or Reggie-2) and Flotillin-2 (or Reggie-1) are two homologous membrane-associated proteins and were identified independently by three groups [55, 110, 111]. Although the flotillins arise from different gene products, their mRNA and protein sequences are 50% and 44% identical, respectively and both proteins are predicted to be 48 kDa [110]. The flotillins are evolutionarily conserved in both prokaryotes and eukaryotes, but are absent in S. cerevisiae and C. elegans. They are ubiquitously expressed, and found especially in the brain, heart, and lung [112, 113]. The flotillins also have a wide subcellular distribution including the plasma membrane, endosomes, lysosomes, maturing phagosomes, exosomes, and Golgi membrane microdomains, dependent on cell type [111, 114-117].

2.3 Structure of Flotillin-1

Flotillin-1 (Flot1) does not contain any transmembrane domains, but has a N-terminal PHB

(prohibitin homology) domain, also called the SPFH (stomatin/prohibitin/flotillin/HflK/C) domain, which has hydrophobic stretches that allow it to associate to the intracellular leaflet of the membrane (Figure 2.1). In addition, Flot1 becomes palmitoylated at Cysteine 34 (Cys34), which allows it to preferentially associate to cholesterol-rich regions of the membrane and has been shown to be required for membrane association [105, 118]. Alpha-helices containing alanine and glutamic acid repeats at its C-terminus may be involved in multimerization, which may lead to membrane microdomains, or other protein-protein interactions including hetero-oligomerization with Flot2 [108, 119]. Though they share no sequence homology, Flot1 and Cav1 are analogous to one another, associating to membranes, multimerizing and associating to membrane rafts similarly [107, 120]. Flot1 may be the Cav1 equivalent in non-caveolar cells.

2.4 PHB/SFPH family members

The SPFH domain is highly conserved and its presence in Flot1 places it among a family of proteins that share this sequence. Whereas the function of the SPFH domain is unknown, proteins with it tend to exist as oligomers and have an affinity for membrane microdomains [121-126]. Yet, not all SPFH family members have a clear association to lipid rafts as they have been placed in this family by virtue of the PHB domain. Nonetheless, the similar structural characterization suggests a common, basic function for these proteins. Their propensity for rafts and involvement in many cellular processes suggest that they may be membrane raft scaffolds. Other members of

20 this family include stomatin, podocin, prohibitin, and the bacterial HflK/HflC proteins.

Figure 2.2 Structural comparison of SFPH family members. The gray boxes represent the protein backbone and the colored boxes and bars represent protein motifs and domains, respectively. Original figure from Rivera-Milla et al. (2006).

2.4.1 Stomatin

Stomatin is a 31 kDa integral membrane protein that is highly expressed in erythrocytes though its expression is not limited to the vascular system [127]. It is marked with a hairpin loop structure with a highly charged N-terminus and similarly to Flot1, both termini face the cytosol [127-129].

Stomatin localizes to the plasma membrane and has been reported to be a lipid raft protein in various cell types [57, 130, 131]. The N-terminus of stomatin is believed to mediate the oligomerization of 9-12 monomers leading to clusters or puncta believed to be microdomains [124,

132, 133]. It was originally thought to be absent in a form of hemolytic anemia called overhydrated

21 hereditary stomatocytosis (OHSt), but was later shown that the loss of stomatin does not cause

OHSt and erythrocyte function is normal in stomatin KO mice [134-137].

Stomatin has been reported to regulate channels and transporters. Stomatin has been shown to traffic and interact with the glucose transporter, GLUT-1, and its overexpression leads to reduced glucose transport [138, 139]. In both the mechanosensory cells of C. elegans and vertebrate neurons, stomatin was found to interact and regulate members of the degenerin/epithelial Na+ channel family [140, 141]. Another possible role for stomatin includes non-specific phospholipid flipping between the bilayer leaflets [142].

2.4.2 Podocin

Podocin is a 42 kDa integral membrane protein that is 47% identical to stomatin. Despite their similarity, podocin is much more specialized. It is only localized to the plasma membrane of a type of kidney epithelial cells called podocytes, which is important for filtering plasma as an initial step to urine formation [143]. It has been found as oligomers and is detergent-insoluble along with other podocyte proteins, including the adaptor CD2-associated protein (CD2AP) and nephrin [125,

144, 145]. The localization of the nephrin/CD2AP/podocin complex to membrane rafts has been proposed to augment nephrin signaling, which is important for maintaining proper kidney function

[146]. Mutations in the podocin gene prevents podocin from associating to the plasma membrane and disrupts nephrin targeting to lipid rafts [125], leading to autosomal recessive steroid-resistant nephrotic syndrome, which rapidly progresses to end-stage renal disease, requiring organ transplantation or proteinuria [147-149]. The severity of podocin mutations highlights the importance of podocin-dependent microdomains in normal physiological function.

2.4.3 Prohibitin

Prohibitin-1 and -2 are proteins that oligomerize and form ring-like complexes on the inner mitochondrial membrane [126, 150]. The exact role of the prohibitin complexes is unclear, however it is proposed to regulate the turnover of mitochondrial membrane proteins [151, 152].

Moreover, studies indicate that they served as chaperones that prevented membrane proteins of the respiratory chain from degradation [151, 152]. More recently, prohibitin-2 has been implicated in mitophagy as a receptor for targeting mitochondria for autophagic degradation [153]. The prohibitins are also found to modulate transcription by translocating to the nucleus, interacting with the retinoblastoma tumor suppressor protein, Rb, and repressing E2F-mediated transcription.

Aside from mitochondrial functions and transcription regulation, the prohibitins have not only been reported to associate to B cell receptors but were found to be in lipid rafts in B cells [154,

155]. They were also found to regulate MAP kinase signaling and cell adhesion [156].

2.4.4 HflK and HflC

HflK and HflC (high frequency of lysogenization) are E. coli proteins that are related to mouse

Flot2 and Flot1, respectively. HflK is 46 kDa and HflC is 37 kDa. Similar to the flotillins, they form hetero-oligomers and the expression of one relies on the other [157]. Although prokaryotic plasma membranes do not contain cholesterol [158-160], they have a similar phospholipid composition to that of eukaryotes, it cannot be ruled out that bacterial membrane lipids may contribute to its own version of lipid rafts [161]. Therefore, HflK and HflC may be integral membrane proteins in bacterial lipid rafts. HflK and HflC regulates whether a virus enters the lysogenic or lytic cycle during bacteriophage infection [157]. They modulate FtsH, an AAA ATP-

23 dependent protease, which affects the stability of a protein called cII. High cII levels favors lysogeny and low cII levels leads to lysis [162].

2.5 Functions of flotillin

2.5.1 Neuronal regeneration

Since the discovery that the flotillins in ganglion retinal cells (GRC) are expressed during axon regeneration in lesioned optic nerves of goldfish, the role of the flotillins in neuronal regeneration and development has been more extensively pursued. Because flotillin mRNA was present during both embryonic axon growth and neuron regeneration, but not in mature axons, it was predicted that the flotillins were important for axonal growth [110]. Though inherently rare, this has also been shown in a mammalian system. When a peripheral nerve with regenerating neurons was grafted on lesioned optic nerves of rats, a small population of rat RGCs that were able to extend an axon onto the graft expressed increased flotillin mRNA levels [55]. The retinal axon regeneration studies were extended to zebrafish optic nerves with the same results [163].

The role of the flotillins in regeneration was further dissected. Punctate flotillin structures were found on the plasma membrane of regenerating axons, including growth cones and filipodia, in both goldfish and rat, suggesting that flotillin microdomains play a role in regeneration [55]. The flotillins were reported to regulate neurite outgrowth and cytoskeletal remodeling during neuronal differentiation in Neuro2A (N2A) cells [164]. IGF-1-induced N2a cells were transfected with a dominant negative Flot2 deletion mutant, called R1EA, which interferes with flotillin multimerization and therefores its microdomain formation, leading to defects in the formation of

24 neurites along with organizational defects of the actin and tubulin cytoskeletal network due to changes in the signaling pathway involved in cytoskeletal remodeling, including the activation of small GTPases Rac1 and cdc42 and recruitment of cytoskeletal adaptor proteins CAP/ponsin. This was further examined when siRNA-mediated knockdown of Flot2 in N2a cells affected Rho

GTPases and downstream actin regulators, such as cofilin, N-WASP, cortactin, Arp2/3, as well as activation of p38, a process formation regulator, and focal adhesion kinase (FAK) [163].

2.5.2 Neuronal development

Flot1 may play a role in neuronal development, from neurite growth to synapse formation, which is critical for neuronal circuitry and synaptic plasticity. Flot1 was shown to interact with ArgBP-

2, a protein containing a SoHo domain and multiple SH3 domains [165]. In cultured PC12 cells, growth factor stimulation led to the recruitment of a protein complex comprised of tyrosine kinase

Pyk2, Cbl and ArgBP-2, into DRMs. Moreover, the coexpression of the three proteins under growth factor-induced conditions led to lamellipodia formation in neurites, which was reduced by cholesterol depletion. This study implicated Flot1-dependent membrane rafts in early stages of neuronal development.

Additionally, it has been reported that transcription factors present in different types of cortical projection neurons respond to the presence of Sema3A, the levels of which is controlled by Flot1- dependent internalization, leading to growth cone collapse and the maintenance of cell-type identity [166]. In addition to its role in differentiation, Flot1 may contribute to establishing neuronal connections. Flot1 was found to preferentially colocalize to glutamatergic synapses, as opposed to GABAergic ones. Moreover, the overexpression of Flot1 in cultured hippocampal neurons led to higher frequency of miniature excitatory postsynaptic currents (mEPSCs),

25 indicating neurotransmitter release only from glutamatergic synapses [167]. Taken together, this suggests that Flot1 may be involved in glutamatergic synaptogenesis.

2.5.3 Cell-cell adhesion

Cell-cell junctions (CCJs) are complexes of proteins that facilitate contact between adjacent cells or cells with the extracellular matrix. Constantly formed and remodeled, CCJs provide the foundation for tissue integrity and enables cell movement. Its disruption correlates with cancer metastasis [168]. The assembly of cadherin complexes at these junctions and their proper functioning require a cholesterol-rich environment, which coincides with whether the flotillins are involved in forming the necessary environment since they are found to localize to CCJs in various cell-types as well as in vivo [112, 169-175] The flotillins (Flo) were first reported to regulate cell- cell adhesion in the eye imaginal disc of D. melanogaster. Although the expression of Flotillin2

(Flo2)-null mutant led to no phenotypic defect overall, the overexpression of Flo2 alone or both

Flotillin1 (Flo1) and Flo2 caused cell adhesion molecules of the immunoglobulin superfamily

(IgCAMs), Rst and Kirre, to mislocalize to intracellular vesicular bodies [176]. It was confirmed that the flotillins were important for stabilization of cell adhesion proteins when they normally colocalized with N-cadherin, E-cadherin, and γ-catenin and knockdown of either flotillin led to the disruption of CCJ formation [177-179]. Furthermore, crosslinking studies showed distinct patches for N-cadherin with ganglioside GM1 in the presence of Flot1, indicating that Flot1 may scaffold cadherin complexes into cholesterol-rich membrane rafts. siRNA knockdown of Flot1 and cholesterol disruption by either methyl-β-cyclodextrin or N-octylglucoside absolved the patches seen by immunofluorescence and disrupted the interaction of the flotillins and N-cadherin by immunoprecipitation [178]. Since cadherin has low binding affinity to itself, it has been postulated that Flot1clustering may effectively strengthen cadherin-mediated adhesion [180, 181].

The importance of proper formation of CCJs in the presence of flotillins has been shown in the context of zebrafish development, in which the loss of the flotillins severely affected embryo survival [182]. Two different morpholinos (Mo) were used to target Flot2, resulting in efficient knockdown in zebrafish embryos and the phenotype was independently shown using

CRISPR/Cas9. The loss of the flotillins impeded zebrafish epiboly, the intercalation of several different cell layers required for morphogenesis, to the extent where embryos are either arrested or dead. Further analysis showed that E-cadherin and β-catenin were unable to localize to a type of cells called deep cells, impairing its migration and affecting epiboly as a whole. The severity of the phenotype caused by the loss of the flotillins is consistent with that seen in cells, suggesting that they may be necessary in the early stages of development.

2.5.4 Endocytic trafficking

Cells utilize endocytosis to regulate protein activity at the cell surface. Endocytosis involves the recruitment of cargo and endocytic machinery, followed by membrane deformation, and finally scission to release the vesicle. The best understood form of endocytosis is clathrin-mediated endocytosis (CME), but proteins residing in membrane rafts largely depend on clathrin- independent endocytosis (CIE). Unlike CME, the mechanisms and machinery involved in CIE has not been fully determined. The flotillins may be a player in CIE.

The flotillins were first shown to be involved in endocytosis in studies concerning the internalization of CD59, a GPI-anchored protein, and cholera toxin B (CTB), a subunit recognized by the glycosphingolipid receptor GM1 [183]. The concomitant depletion of Flot1 and inhibition of dynamin, primarily studied in the context of CME, decreased CD59 and CTB internalization.

Flot1 was shown to be present on structures that bud into the cell and was present on endosomes lacking caveolin-1 and transferrin, the classic cargo for CME. This may be explained by the membrane microdomains formed by the hetero-oligomerization of Flot1 and Flot2, as shown by colocalization with CD59, and their presence in membrane invaginations in Cav1-knockout MEFs

[120, 184]. Furthermore, it was discovered that the EGF-induced internalization of the flotillins and trafficking to late endosomes and lysosomes required their phosphorylation by Fyn, a member of the Src family kinases [185].

The membrane raft localization of proteins may be important for their internalization. Flotillin- mediated endocytosis may help regulate protein activity at the cell surface, such as controlling the neurotransmitter concentration in the perisynaptic space [186]. Flot1 was found to be essential for the membrane raft localization and PKC-triggered internalization of the dopamine transporter

(DAT) [105]. Flot1 knockdown in cells eliminated the intracellular accumulation of DAT.

However, its overexpression rescued the inhibition of PKC-triggered endocytosis by Gö6850.

Endogenous DAT from mouse brain slices was also internalized in the presence of phorbol 12- myristate 13-acetate (PMA), an activator of protein kinase C. In another instance, the flotillins were found to reside in the same buoyant fractions as Niemann-Pick C1-like 1 (NPC1L1), a protein important for cholesterol absorption [187]. Flot2 knockdown results in no internalization of either

NPC1L1-EGFP or cholesterol, suggesting Flot1-dependent microdomains are crucial sites of endocytosis.

It is still unclear how the flotillins participate in endocytosis. They may be scaffolding endocytic proteins, directing endocytic machinery and acting as an independent trafficking pathway, or could be a Flot-dependent form of CME. As previously mentioned, Flot1-mediated endocytosis may be independent of both clathrin and caveolin [183]. However, other possibilities cannot be

28 disregarded since it has also been reported that DAT is internalized through CME [188]. Moreover, ubiquitinated DAT has been found to internalize and enter the degradative pathway in a PKC- mediated, clathrin- and epsin-dependent manner [189]. One possibility is that Flot1 may differentially sort cargo into a separate trafficking pathway from clathrin. For instance, CME of the TGF-β receptor results in its recycling and signaling whereas caveolin-mediated endocytosis leads to its degradation [190]. Similarly, clathrin may sort DAT into the endosomal recycling pathway whereas Flot1 sends it to the lysosomal degradative pathway or vice versa. It is also possible that Flot1 acts as an adaptor protein like epsin. Epsin normally links cargo to clathrin, but it has been found to act independently from the CME of a tetanus neurotoxin fragment (TeNt Hc) initiated from membrane rafts [191, 192]. Hence, Flot1 may act analogously to epsin in the context of membrane rafts. The role of Flot1 in regulating membrane raft proteins can be elucidated with further studies.

2.5.5 Signal transduction

Consistent with the most popular function assigned to lipid rafts, the flotillins have been found to be involved in signal transduction. Microdomain partitioning may help spatially gather the necessary protein components for signal transduction. In Jurkat and human T cells, the flotillins and proteins involved in T cell activation and signaling are assembled into scaffolds in a process called capping. Crosslinking results in the colocalization of a protein called PrPc and the flotillins in the cap [193]. This results in T cell activation as Fyn, Lck, LAT (linker for activation of T cells), and F-actin, all components involved in signal transduction, also migrate to the capping site. The scaffolding of signaling proteins to enhance their interaction has similarly been shown in CD4+ T

29 cells. Stimulated primary human CD4+ T cells led to increased surface colocalization with GM1 and Flot1 concomitant with the accumulation of Lck and LAT, T cell receptor (TCR)-associated signaling components, in rafts [194].

The flotillins have also been shown to be involved in MAPK signaling via growth factor receptors.

Flot1 helps scaffold EGFR into DRMs at the cell surface and affects proteins, including ERK1/2, that participate in the signaling pathway [195]. Flot1 knockdown impairs tyrosine phosphorylation of the EGFR, leading to decreased phosphorylation of ERK1/2 and overall poor signal. This suggests a role for Flot1 at both the cell surface where receptor activation takes place and intracellular events downstream of activation where signaling is carried out. Separately Flot1 was found to scaffold and interact with fibroblast growth factor receptor substrate 2 (FRS2a), a signaling adaptor protein involved in the downstream signaling of various receptor tyrosine kinases, and participate in the negative feedback loop of the pathway. Flot1 binds to the phosphotyrosine-binding domain (PTB) of FRS2 [196]. Flot1 knockdown leads to increased Tyr phosphorylation of FRS2, inhibiting ERK and preventing the phosphorylation of Thr on FRS2 to stop the Tyr phosphorylation. This is consistent with the observation that Flot1 depletion leads to

ERK inhibition. Flot1’s involvement in MAPK signaling was further corroborated in studies on signal induction via transactivation of EGFR by muscarinic acetylcholine receptors (mAChRs) in human keratinocyte-like cell line HaCaT [197]. Flot1 expression not only reduced cholinergic

ERK activation in HaCaT cells, but also reduced the transcription of target genes of the cholinergic MAPK response, including HB-EGF, TGFα, AREG and EREG (all EGFR ligands) and matrix metalloproteinase (MMP-3) [198]. Thus, the flotillins can play a role in various steps of a signaling cascade, from cell surface scaffolding of necessary components to transcriptional regulation of signaling target proteins.

2.5.6. Transporter activity

Flotillin scaffolding into membrane rafts may also encourage protein activity. Recently, this has been demonstrated by the dopamine transporter (DAT). Although neurotransmitter transporters, such as DAT, are found to reside in these microdomains, how this partitioning affects transporter function has been difficult to establish [199-201]. Recently, it has been shown that Flot1 is required to localize heterologously expressed DAT into membranes rafts in cells [105]. Flot1 depletion led to greatly impaired AMPH-induced DA efflux, but did not affect DA reuptake. This was further validated in D. melanogaster larvae. Larvae expressing a mutant Flo gene respond with increased crawling speed to methylphenidate (MPH), which competitively inhibits DA uptake but does not affect efflux [202]. However, mutant Flo larvae did not respond to amphetamine (AMPH) and this phenotype was rescued by expression of human Flot1 in DA neurons of the mutant larvae. Larvae with a palmitoylation-deficient mutant of human Flot1 also did not respond to AMPH, suggesting that the lack of response to AMPH was due to DAT’s loss of membrane raft association. These data suggest that Flot1 modulates DAT function by influencing its membrane localization.

2.6 Concluding remarks

Flotillins, similarly to other SPFH family members, appear to perform various cellular functions.

The membrane rafts demarcated by the flotillins serve the traditional roles assigned to these microdomains, including endocytosis and signal transduction. However, they also contribute to other cellular processes. They are protein scaffolds for transporters and help strengthen cell-cell

31 junctions. Most of these functions begin at the cell membrane, where flotillin seems to delineate a particular membrane environment for proteins that require it. The specific impact of these specialized membrane milieu remains to be uncovered and the molecular mechanism that occur within the confines of these Flot-dependent rafts still needs to be examined.

Chapter 3

The Dopamine Transporter (DAT)

Figure 3.1 Amino acid sequence of human DAT. The N- and C-termini face the cytosol and contain consensus sites for kinases. There are twelve transmembrane domains embedded in the membrane, represented by a rectangular gray box. The Y-shaped symbols represent potential N-glycosylation sites. Original figure from Norregaard et al. (1998).

3.1 Introduction

Dopamine (DA) is an ancient neurotransmitter that modulates attention, motor activity, reward processing and goal-directed behavior in vertebrates [203, 204]. Altered DA signaling has been implicated in psychosis, movement disorders, and drug addiction. One means by which DA

33 signaling can be regulated is via the cell surface dopamine transporter (DAT), which mediates the reuptake of DA from the perisynaptic space, thereby modulating the strength and duration of neurotransmission. Psychostimulants, such as cocaine (COC) and amphetamine (AMPH), affect

DAT function, resulting in an increase in the extracellular concentration of DA and eliciting behaviors linked to hyper-DAergic signaling [205, 206]. Even though the mechanism by which

COC inhibits DA reuptake by DAT is well understood, how AMPH drives increased synaptic DA levels is less certain. AMPH leads to increased cytosolic DA, which is reverse transported

(effluxed) out of the cell by DAT in a process that is greatly enhanced by AMPH-induced phosphorylation of the DAT N-terminus [207]. Although there has been significant study of DAT protein structure and function, little is understood how it interacts with its surrounding membrane.

Therefore, how membrane composition influences DAT’s structure and function is unknown.

In my thesis work, I used the actions of AMPH to dissect how membrane localization of DAT affects its function. In this chapter, I will give an overview of DAT structure and function, followed by where it is expressed in the brain. I will address a few disorders in which abnormal DA signaling is implicated. Finally, I will touch upon how psychostimulants affect DAT, with an emphasis on

AMPH.

3.2 Structure

Neurotransmitter transporters (NTTs) are responsible for controlling the neurotransmitter concentration in the perisynaptic space [186]. The NTTs consist of the glutamate/neutral amino acid transporters (SLC1) and the neurotransmitter:sodium symporters (SLC6). Similar to its SLC6 gene family members, the dopamine transporter (DAT) has 12 transmembrane domains, including

34 five intracellular loops and six extracellular loops [208]. DAT contains various structural elements that ensure proper localization and function. For instance, the second extracellular loop contains

N-linked glycosylation sites that play a vital role in stabilizing the surface expression of DAT as well maintaining efficient uptake [209]. Additionally, mutational analysis has shown that leucine- repeat motifs on transmembrane domain 2 are required for the trafficking of DAT to the plasma membrane [210]. Both the N- and C-termini face the cytosol and have been found to interact with protein kinase A (PKA), protein kinase C (PKC), and Ca2+/calmodulin-dependent protein kinase

II (CaMKII), all of which modulate DAT function in the presence of AMPH [207, 211, 212].

The structure of DAT has contributed to our current understanding of its molecular mechanism.

Initial mechanistic understanding was drawn from the crystal structure of LeuT, a bacterial homolog of sodium-/chloride-dependent neurotransmitter transporters [213]. LeuT and SLC6

NTTs have 55-67% sequence homology in the regions that comprise of the fundamental structural components required for transport, validating its use in the study of SLC6 family members.

Molecular models of DAT based on LeuT have shed light on the DAT residues that bind to and interact with DA as well as highlight residues that may stabilize a DAT conformation for efficient

DA transport [214-216]. Although LeuT has contributed to our overall understanding of transport mechanisms and substrate and ion interactions within the transporter, there are still limitations.

Because the sequences are not identical, LeuT cannot be used to address all regions present in eukaryotic NTTs, which may be important for conferring substrate or ion specificity, underscore the NTTs relationship with antidepressants and psychostimulants, and understanding of how posttranslational modifications or the surrounding milieu affects transporter structure and function.

Recently, the crystal structure of DAT from D. melanogaster (dDAT) was elucidated. Unlike

LeuT, which has an overall ~20% sequence similarity to its eukaryotic equivalents, dDAT

35 possesses more than 50% sequence similarity [102]. The crystal structure of dDAT revealed that membrane composition might be critical to DAT function: A cholesterol molecule is wedged in one of its transmembrane grooves, which suggested why cholesterol extraction disrupts DAT- mediated uptake [105]. In addition, dDAT was crystallized bound to the antidepressant nortriptyline, giving much needed insight into how dDAT interacts with key substrates. The elucidation of dDAT will expand our comprehension of DAT structure and function.

Figure 3.2 A cholesterol site in Drosophila DAT. Different views of cholesterol modeled with dDAT. The presence of cholesterol suggests that lipids play a role in maintaining a certain conformation for DAT. Orignal figure from Penmatsa et al. (2013).

3.3 Function

DAT serves to modulate dopamine (DA) signaling by clearing DA from the perisynaptic space.

The reuptake of DA by DAT functions to modulate the strength and duration of neurotransmission in order to maintain normal dopaminergic (DAergic) signaling. Because DA reuptake is thermodynamically unfavorable, DAT relies on a gradient of sodium ions to move DA back across

36 the cell membrane. For every one dopamine that gets taken up, two sodium ions and one chloride ion must be co-transported [217]. The ion gradient is generated by the sodium/potassium ATPase pump [214]. Although it is still unclear what role the chloride ion plays in DA reuptake, it has been postulated that it may help balance the positive charge of sodium [218]. Upon entering the cytosol,

DA is then repackaged into synaptic vesicles through vesicular monoamine transporter 2

(VMAT2), readying it for subsequent release.

It should be noted that similarly to the other monoamine transporters, DAT does not necessarily only transport DA [219]. DAT can also transport norepinephrine (NE) and epinephrine [220]. This selectivity appears to depend on the localization and expression pattern of DAT. DAT is highly expressed in the striatum and can transport DA efficiently there. In contrast, in the prefrontal cortex where DAT expression is particularly low, the moderately expressed norepinephrine transporter

(NET) functions to take up both NE and DA [219].

3.4 Localization and distribution in the brain

DAT through its localization to dopaminergic (DAergic) neurons is expressed in the striatum and nucleus accumbens in the basal ganglia of the forebrain, substantia nigra/ventral tegmental area

(SN/VTA) in the midbrain, and to a lesser extent, the prefrontal cortex, medulla oblangata, hypothalamus, retina and the olfactory bulb [221, 222]. Within the different brain regions, DAT distribution is highly variable. For example, focusing on the VTA in rat brain, DAT is localized to intracellular membranes in the perikarya and large proximal dendrites, but were on the plasma membrane in small to medium diameter dendrites and unmyelinated axons [223, 224]. The differential distribution of DAT within DAergic neurons suggests varying functional necessities.

Moreover, DAT expression levels are dictated by the intensity of DA signaling. Increased levels of synaptic DA correspond to more plasmalemmal DAT to facilitate uptake and vice versa [225].

Accordingly, DAT expression coincides with tyrosine hydroxylase (TH) expression, which is responsible for the production of DA. The spatial proximity of DAT and TH mRNA translates to efficient release and reuptake of DA.

There are four major DAergic pathways: the mesolimbic, mesocortical nigrostriatal, and tuberoinfundibular pathways. Each pathway innervates a different part of the DAergic system and regulates various behavioral and physiological processes and aberrations may lead to disorders.

Both the mesolimbic and mesocortical pathways projects from the VTA into the nucleus accumbens and prefrontal cortex, respectively. Whereas the former pathway is implicated in reward and pleasure, the latter is involved in cognition and emotion. The nigrostriatal pathway projects from the substantia nigra to the dorsal striatum, where the caudate nucleus and putamen reside, and influences motor behaviors. Finally, DAergic neurons in the tuberoinfundibular pathway extend from the arcuate nucleus in the hypothalamus to the median eminence, which controls prolactin secretion from the pituitary gland [226]. The cellular localization and expression level of DAT and different projections of the DAergic neurons play critical roles in maintaining

DAergic homeostasis.

3.5 DA, DAT and disease

Given the wide-ranging importance of DA signaling, aberrations can lead to an array of disorders ranging from movement disorders to neuropsychiatric disease. Depending upon the complexity of the disorder, the exact role of DA and DAT in many of these disorders has not been fully dissected.

This section aims to highlight the broad impact of the DAergic system, including direct mutations of DAT, such as in autism, to psychostimulant interactions with DAT.

3.5.1 Parkinson’s disease

Parkinson’s disease (PD) is a chronic and progressive movement disorder of the central nervous system characterized by tremors, bradykinesia, rigidity, and postural instability. This may be accompanied by memory deficits, language impairment, hallucinations, and sleep disturbances

[227]. The most dominant form of PD involves the degeneration of DAergic neurons of the basal ganglia, especially those of the substantia nigra pars compacta and the presence of Lewy bodies, aggregates of a protein called α-synuclein, in the remaining neurons. An in vitro study has shown that interaction between α-synuclein and the C-terminus of DAT increases the plasma membrane localization of DAT and therefore, DA uptake, ultimately resulting in DA-induced apoptosis [228].

DAergic neuronal degeneration is accompanied by the localization of reactive astrocytes and activated microglia, which are important for the inflammatory response. Moreover, the recruitment of astrocytes may further contribute to loss of DA in the DAergic neurons as they have been reported to express DAT and NET as well as regulate extracellular DA levels [229-231].

Neuronal deterioration in PD correlates with the loss of DAT. It has been reported that there is up to 50% loss of DAergic nerve terminals in early PD [232]. Moreover, there was significantly less

DAT in the putamen, caudate, nucleus accumbens obtained from PD brains, compared to control

[233]. This study found that the reduction in DAT levels correlated to how long the patients have had PD. Accordingly, positron emission tomography (PET) studies have shown that within PD patients, increased DAT expression correlates with more balanced levels of synaptic DA,

39 indicating that a decrease in DAT levels, as seen in PD, most likely results in fluctuations of synaptic DA concentrations, and therefore, contributes to the motor deficits seen in disease progression [234]. Aging may be compounded with PD progression as DAT expression drops 5.6–

6.1% per decade in non-PD subjects [235]. However, although the rate was similar for the case study, the patient already had a more rapid decline in DAT levels in certain brain regions, suggesting that the dual effect of aging and PD progression may lead to a more rapid loss of DAT

[236].

3.5.2 Schizophrenia

Schizophrenia is a chronic mental disorder characterized by a disconnect between thoughts and reality, disorganized speech and behavior, and abnormal social behavior. It generally manifests in early adulthood and its cause is a combination of genetic and environmental factors. Schizophrenia was originally thought to be due to hyper-dopaminergic function since phenothiazine (PTZ) drugs, which inhibited DAT function, were able to reduce psychotic symptoms whereas amphetamine, which causes DA to accumulate in the perisynaptic space, may worsen the psychotic symptoms

[237]. Nonetheless, studies have shown that DAT levels do not change between control subjects and schizophrenics, whether or not they were undergoing antipsychotic treatment [238, 239]. This was further supported by meta-analysis, which did not find any relationship between schizophrenia and DAergic nerve terminal density, regardless of antipsychotic treatment or time since schizophrenia onset [240]. Since PTZ also has an affinity for the D2 receptor, this might offer an alternative explanation to the actions of PTZ. PET scans have also revealed that excess DA

40 synthesis rather than DAT-mediated uptake in the striata of schizophrenia patients may play a role

[241].

3.5.3 Obsessive compulsive disorder

Obsessive compulsive disorder (OCD) is a chronic psychiatric anxiety disorder characterized by uncontrollable, obsessive thoughts and repetitive behaviors. For instance, one common OCD obsession pertains to anxiety caused by contamination, leading to compulsive and often excessive hand-washing. The cause of OCD is unidentified but it is believed to be a combination of genetic and environmental factors. OCD was originally thought to be caused by abnormalities in serotonergic signaling because selective-serotonin reuptake inhibitors (SSRIs) would help manage

OCD symptoms, but 40% of afflicted individuals are unresponsive to this class of antidepressants

[242, 243]. The DAergic system was implicated when pharmacological studies showed that patients given a combination therapy of SSRIs and atypical antipsychotics, which tend to be

DAergic antagonists, responded to this regimen [244, 245]. One study found that there was increased DAT density in the caudate and putamen in OCD patients, compared to controls when measured by single-photon emission computed tomography (SPECT) [246]. Others observe that

OCD patients who have not been exposed to drugs have less striatal DAT [245]. How DAT is contributing to the symptoms of OCD remains to be studied.

3.5.4 Attention defecit hyperactivity disorder

Attention deficit hyperactivity disorder (ADHD) is a neuropsychiatric disorder typically diagnosed in children, who display impulsivity, inattentiveness, and hyperactivity. A genetic component is

41 thought to contribute to ADHD as identical twins have a 81% chance of developing it compared to a 29% chance for fraternal twins [247]. Genetic differences in the dopamine transporter gene

(DAT1), most notably a 10-repeat allele of the DAT1 variable number tandem repeat (VNTR), have been correlated with ADHD [248]. When this VNTR is present with other VNTRs, such as the 7-repeat allele of the D4 dopamine receptor or other intronic and exonic variants, there is a significant correlation to ADHD [248, 249]. Given that the typical drugs prescribed to people with

ADHD are amphetamine and methylphenidate, both of which increase extracellular DA levels, it is not surprising that the genes involved in the catecholamine pathways, including DAT, NET, D4 and D5 dopamine receptors, are believed to lead to ADHD onset [250, 251]. The familial association brought forth the hypo-dopaminergic hypothesis which claims that ADHD is due to low dopamine function [248, 252].

3.5.5 Autism spectrum disorder

Autism spectrum disorder (ASD) is a neurodevelopmental disorder characterized by difficulty with social skills, speech and nonverbal communication, obsessive interests, and repetitive behaviors.

Although autism has a strong genetic basis, it is complex and unknown whether it is caused by a rare mutation or a combination of common genetic variants as chromosomal deletions, duplications, and inversions have all been implicated in autism [253-255]. Additionally, its underlying mechanism is unclear. Speculations for the molecular basis of autism include the following: local neuronal overconnectivity in key brain regions during early development, abnormal formation of synapses and dendritic spines, immune dysregulation due to increased levels of pro-inflammatory cytokines and microglia activation, among other hypotheses [256-260].

Recently, a residue substitution of DAT (Ala559Val) has been identified in two unrelated ASD

subjects. This coding variant leads to altered hDAT trafficking due to hyper-phosphorylation of

the N-terminus of DAT and elevated PKCβ levels, leading to imbalances in DA neurotransmission

[261].

3.6 DAT and psychostimulants: cocaine and amphetamine

Figure 3.3 The mechanism of action for cocaine (COC) and amphetamine (AMPH). (A) Under basal conditions, dopamine (DA) that has been released binds to post-synaptic dopamine receptors and the excess is cleared from the perisynaptic space via DAT. (B) Cocaine inhibits DAT, resulting in the extracellular increase of DA. (C) AMPH increases the level of extracellular DA by competitively inhibiting DAT and causing non-exocytic reverse transport of DA. Original figure from Espana and Jones (2013).

Psychostimulants, such as cocaine (COC) and amphetamine (AMPH), affect DAT function,

resulting in an increase in the extracellular concentration of DA and eliciting behaviors linked to

hyper-DAergic signaling [205, 206]. Here, I will discuss both mechanisms of action, with an

emphasis on AMPH.

3.6.1 Cocaine

Cocaine (COC), extracted from the Erythroxylon coca plant, is a highly abused psychostimulant whose effects include euphoria, locomotor stimulation and a sense of reward [262]. COC has a high affinity for DAT and inhibits DA uptake, leading to the extracellular accumulation of DA

[263]. Given the structural similarities of SLC6 transporters, COC can also inhibit the uptake of serotonin and NE, which may also contribute to its euphorigenic effects [264, 265]. The role of

DAT inhibition by cocaine has been shown in DAT KO mice, which do not exhibit additional hyperlocomotor behavior even when administered 40 mg/kg cocaine [266]. The role of DAT in mediating all effects of COC were put in question when DAT KO mice still self-administered COC and showed motivation for COC as revealed by the condition-placed preference (CPP) test [267-

269]. However, this was rectified by creation of a knock-in (KI) model of a COC-insensitive DAT, in which COC binding to DAT is impaired but normal DA uptake is still intact [270]. With greater than 50% transport activity maintained, these mice did not exhibit cocaine-induced locomotor activity at various doses, did not have increased levels of extracellular DA in the nucleus accumbens, and did not display cocaine-induced CPP.

At the molecular level, based on a structure derived from LeuT, the binding site for cocaine and its analogs is buried in a deep crevice situated between transmembrane segments 1, 3, 6 and 8

[271]. More specifically, it has been reported that Cys90 affects cocaine binding and others have found that Cys306 together with cholesterol plays a role in affecting a more favorable conformation for cocaine binding [103, 272]. Given that Drosophila DAT has been crystallized with cholesterol, the importance for membrane lipid interaction as a determinant of COC binding

44 might warrant further study [102]. Finally, in addition to direct binding to DAT, COC also affects

DAT expression at the cell surface, by stabilizing its presence at the cell surface [273]. Although how this occurs remains unclear, it has been posited that this may serve to enhance the effects of

COC.

3.6.2 Amphetamine

Amphetamine (AMPH) was first synthesized in 1887 by Lazar Edeleanu. It has many synthetic variants as well as naturally occurring ones. It is chemically characterized as having the following groups: an unsubstituted phenyl ring, a two-carbon side chain between the phenyl ring and nitrogen, an a-methyl group, and a primary amino group (Figure 3.1) [274]. Although it is highly abused, it also has therapeutic potential, necessitating a mechanistic understanding of AMPH action.

Although AMPH leads to the same outcome as COC, its mechanism of action is distinct. First,

AMPH acts as a competitive inhibitor of DAT, and enters into the DAergic neuron. Second, AMPH leads to increased cytosolic DA, which is then reverse transported (effluxed) out of the cell by

DAT. The reverse transport of DA through DAT is one of the most widely studied mechanism by which AMPH increases extracellular DA. Although there is growing evidence suggesting that the action of AMPH is multifactorial, there are two primary models describing AMPH action. The facilitated exchange diffusion model explains that efflux is due to the movement of one AMPH into DAT along with one DA out of DAT [275-277]. This classical model suggests that the rate of efflux would increase because the movement of AMPH through DAT is causing the transporter to

Figure 3.4 Chemical structures of amphetamine and its related compounds. (A) The basic structure of AMPH: (1) an unsubstituted phenyl ring, (2) a two-carbon side chain between the phenyl ring and nitrogen, (3) an a-methyl group, and (4) a primary amino group. (B-I) The chemical structures of AMPH-related compounds. Original figure from Sitte and Freissmuth (2015). open up intracellularly more often and thus, bind more DA to efflux. However, there are many challenges to this model. Others have a found a non-linear relationship in terms of substrate uptake and release. For example, some AMPH-like substrates are better releasers than they were as substrates for uptake [278]. The weak base model proposes that AMPH causes efflux by increasing

46 the cytoplasmic DA levels and changing the gradient of DA across the plasma membrane [279-

281]. This is accomplished because AMPH enters synaptic vesicles by lipophilic diffusion, collapses the proton gradient that provides the energy source for the countertransport and accumulation of monoamines in vesicles, resulting in cytosolic DA accumulation [281, 282].

However, it was found that the vesicle depletion of DA by Ro4-1248 and reserpine-like compounds did not result in efflux even though there was increased cytosolic DA levels [283].

This suggests that increasing intracellular DA concentration was not enough for efflux and that

AMPH must interact with DAT.

Experimental evidence strongly suggests that post-translational modifications are important for

DAT function. For example, phosphorylation of DAT is important for AMPH-induced efflux.

Inhibition of PKC or CaMKII leads to diminished non-exocytic release of DA by AMPH [211,

284-289]. In PKCβ KO mice, there was less locomotor response to AMPH [290], although diminished cell surface DAT levels complicated interpretation. Using amperometry and patch- clamp, it has been shown that a 22-residue truncation at the N-terminus of DAT reduces DA efflux by 80% in HEK293 cells and similar results were shown when the residues were narrowed down to five serines and mutated to alanine [207]. It was also shown that CaMKII interacts with the C- terminus of DAT to presumably phosphorylate the aforementioned N-terminal serines [211].

Though basally hyperactive, CaMKII KO mice also have reduced locomotor behavior in response to AMPH [284]. Most recently, phosphorylation of DAT at Thr53 has also been shown to be increased in the presence of AMPH in cells [291]. Although phosphorylation events are important, why they are necessary remains unclear.

In addition to promoting efflux, DAT phosphorylation contributes to its internalization by PKC

[292-298]. Confocal imaging revealed that PKC activation led to the internalization of hDAT and

PKC inhibition led to increase membrane localization of DAT [299]. This was confirmed by

Galli’s group who also used electrophysiological techniques to show that surface DAT is internalized after 20 min of AMPH treatment and that internalization is at its maximal after 1 hour

[300]. Moreover, they showed decreased [3H]dopamine uptake by 27% in the presence of AMPH.

However, the internalization of DAT implies that there shouldn’t be much of a physiological effect by efflux. An even faster time frame was shown by cell surface biotinylation on rat striatal synaptosomes, in which DAT cell surface expression increased within 30 seconds of AMPH application and this increase lasted up to 2 minutes [301]. Perhaps, AMPH temporarily causes higher surface DAT expression to exert its effects before internalization takes over and there are lasting effects to AMPH because there is not nearly enough DAT to take the DA back in. The decrease in surface DAT expression can be explained by both endocytosis and less recycling to the plasma membrane as it has been reported that hDAT can be found in endosomes following

AMPH treatment [105, 302, 303].

Aside from causing the reverse transport of DA through DAT, AMPH also inhibits uptake. Prior to electrophysiological techniques, it was difficult to distinguish between release and uptake inhibition. The thinking was that cocaine, which serves as a DAT inhibitor, should in theory block both uptake and reverse transport, but it could not be ascertained whether cocaine was behaving as a substrate to block AMPH or whether it was physically blocking reverse transport [304]. People who tackled this problem went around the uptake issue by directly injecting DA into the giant DA neuron of the pond snail Planorbis corneus [281]. By way of amperometric recording, it was shown that efflux was prevented by nomifensine, an uptake inhibitor, and this was reversible, showing that DA release was by reverse transport. In another experiment, AMPH was directly injected into the cytosol to circumvent interaction with DAT. In this experiment, DA was released

48 and nomifensine again could inhibit this efflux. However, efflux occurred even in the absence of nomifensine, showing that reverse transport and uptake can be dissociated. This dissociation was later clarified using cyclic voltammetry, which was fast enough to show that AMPH’s effect on efflux was due to reverse transport but inhibition of uptake contributed to the overall effect [305].

More recently, it has been discovered that the surrounding membrane may play a role in promoting an efflux-friendly conformation for DAT. DAT has been found to localize to Flot1-dependent membrane rafts and this placement appears to be important for AMPH-induced efflux as primary

DAergic neurons transduced with shFlot1 have impaired AMPH-induced efflux compared to control [105]. Moreover, it appears the N-terminus of DAT is not only useful for phosphorylation, but also membrane component interaction. It was reported that two lysines (K3 and K5) interacted with phosphatidylinositol 4,5-bisphosphate, or PIP2 [306]. DAT in which the two lysines are mutated to alanine cannot interact with PIP2, resulting in a conformational change in which the N- terminal tail of DAT falls away from the membrane. Furthermore, D. melanogaster carrying this mutant DAT displayed reduced locomotion when given AMPH. Taken together, this suggests that the membrane environment is essential for DAT function in the presence of AMPH.

3.7 Concluding remarks

DAT is incredibly important for maintaining the proper concentration of DA in the perisynaptic space as abnormalities in DAergic signaling can lead to disorders that currently have no cure.

Nonetheless, as a target for addictive psychostimulants, the study of DAT has largely centered on the mechanism of action of these drugs. In part, these compounds have increased our understanding of DAT function and how it exerts its influence on DAergic signaling, but the

49 complexity of AMPH also gives rise to not only how it might work, but what, if any, the physiologic significance of DA efflux might be.

Chapter 4

Determining the Role of Flot1-Mediated Partitioning of DAT on

AMPH-induced DA Release.

4.1 Introduction

The biophysical and biochemical properties of membrane lipids can impact the conformation and function of membrane-spanning proteins, and thereby influence diverse cellular functions.

Whether this occurs through the formation of discrete protein-lipid microdomains such as membrane rafts or by other mechanical means such as membrane deformation is an on-going discussion. Fundamentally however, it has been difficult to capture the specific, physiologic consequence of changing a protein’s membrane milieu.

The work by Haase and colleagues indicated that members of the SLC6 family of cell surface transporters can partition in part to detergent-resistant membranes (DRMs) upon homogenization with 1% Brij 58 [307]. One such member for which this has been quite often examined is the dopamine transporter (DAT). Given that DAT is responsible for the uptake and clearance of extracellular dopamine (DA) from the perisynaptic space, modulation of DAT activity can have a profound influence on the wide range of DA-mediated behaviors from movement to cognition.

Nonetheless, although the partitioning of DAT into cholesterol-rich membranes represented by

DRMs have been replicated by several groups [103-105], the larger physiologic significance of this partitioning remains unclear.

Studies suggest that the membrane-associated protein Flotillin 1 (Flot1) scaffolds DAT to DRMs in stable cell lines, and the acute depletion of Flot1 causes DAT to no longer associate with these membranes [105]. Functionally, Flot1 depletion in cells had no impact on DAT mediated uptake of DA, but diminished the non-exocytic release of DA through DAT that is observed upon administration of the psychostimulant amphetamine (AMPH) [105]. These observations led to the hypothesis that mice lacking Flot1 will demonstrate a diminished ability to respond to acute

AMPH, but not to an inhibitor of DA uptake via DAT such as cocaine (COC). This hypothesis assumes that reverse transport of DA through DAT, also known as DAT efflux, contributes towards increasing perisynaptic levels of DA in response to AMPH in vivo. To what extent this mechanism contributes to the impact of acute AMPH in vivo has been questioned [308, 309].

To test the physiologic importance of the relationship between Flot1 and DAT, we created a series of genetically modified mice for which we conditionally or constitutively removed Flot1. We report that the loss of Flot1 in DAergic neurons leads to no change in basal DA neuron parameters including DA uptake by DAT. Although the ability of mice to respond to COC was intact, their ability to respond to AMPH was significantly diminished, segregating the ability of mice to respond to these two psychostimulants. Endogenous DAT partitioned into cholesterol-rich DRMs in a Flot1-dependent manner, and partitioning increased or decreased upon the administration of

AMPH or COC, respectively, suggesting that DAT activity influenced the membrane environment of DAT. Mechanistic studies revealed that DAT partitioning is required to stabilize a phosphorylated form of DAT that can interact with the membrane phospholipid PIP2 leading to a distinct conformation of DAT in the DRMs that is required for the AMPH-evoked response.

Interestingly, the blunted response to AMPH and Flot1-dependent partitioning of DAT was observed under conditional but not constitutive elimination of Flot1, indicating that developmental compensation masks the importance of Flot1 function. Our data indicates that the Flot1-dependent partitioning of DAT is important for AMPH-induced reverse transport of DA because the cholesterol-rich environment promotes a conformation of DAT favorable for DA efflux.

4.2 The conditional loss of Flot1 in DAergic neurons leads to a diminished response to

AMPH in mice

Figure 4.1. Design of the Flot1 conditional allele and resulting recombined allele of the mice. (A) LoxP sites were inserted between exons 2 and 3, and exons 8 and 9. PCR primers to detect the LoxP insertions are designated as Primers 1 and 2, and Primers A and B. (B) PCR genotyping for the different primer pairs. (C) PCR detection of the recombined allele. The recombined Flot1 allele can be detected in DNA collected from substantia nigra (SN) of Flot1(DAT) cKO mice (Flotfl/fl::DATiresCre (DIC)), but not in DNA from cortex (Ctx) or tail. Similarly, no recombination is detected in the Flot1fl/fl mice.

To determine the functional significance of Flot1-mediated membrane partitioning of DAT, we created a conditional KO mouse that genetically eliminated Flot1 in neurons that express DAT.

Mice carrying a conditional Flot1 allele (Flot1fl/fl) were crossed to mice carrying the DATiresCre allele resulting in the Flot1 conditional knockout (Flot1 cKO) mouse (Figure 4.1). Unlike DAT

KO mice [266], Flot1 cKO mice demonstrated no overt phenotype, and was able to breed and thrive comparably to their littermate controls, suggesting that basal DAT function was unaffected by the loss of Flot1.

Using these mice, we next examined how they responded to amphetamine (AMPH) and cocaine

(COC), two psychostimulants known to interact with DAT. Both AMPH and COC lead to increased levels of perisynaptic DA, leading to behaviors such as hyperlocomotion. COC and

AMPH act differently at the level of DAT. COC acts as a non-competitive inhibitor of DAT whereas AMPH competitively inhibits DAT, increases cytoplasmic DA levels thereby causing non-exocytic release of DA via its reverse transport through DAT.

Male mice habituated to the open field maze were administered saline or AMPH (2.5 or 5 mg/kg, i.p.) and monitored. Similarly to the Flot1fl/fl controls, mice heterozygous for conditional deletion of Flot1 (Flot1 cHet) demonstrated a hyperlocomotor response to both doses of AMPH (Figure

4.2A). In contrast, Flot1 cKO mice demonstrated a significantly blunted response, demonstrating little to no detectable locomotor response at 2.5 mg/kg and a significantly reduced response at 5 mg/kg (Figure 4.2A). This diminished response was specific to AMPH, because the ability of Flot1 cKO mice to respond to COC or to habituate to the open field was unaffected by the loss of Flot1 expression (Figure 4.2B, C). Similar results were obtained with Flot1 cKO female mice (Figure

4.3). These data demonstrate that the loss of Flot1 in DAergic neurons can segregate the ability of mice to respond to AMPH versus COC in vivo, thereby indicating that the mechanism of action of these compounds are distinct.

Figure 4.2. Elimination of Flot1 in DA neurons diminishes the ability of mice to respond to AMPH in vivo: The Flot1 cKO mice. (A) Flot1 cKO show a diminished response to AMPH, but show a normal response to COC. Flot1fl/fl (black), Flot1fl/+::DATiresCre+/- (Flot1 cHet, gray) and Flot1fl/fl::DATiresCre+/- (Flot1 cKO, red) littermate male mice were habituated in an open field maze for 60 min, then administered saline, or 2.5 mg/kg or 5 mg/kg of amphetamine (AMPH) i.p. and monitored for an additional 60 min. Flot1fl/fl and cHet mice demonstrate a typical, hyperlocomotor response to 2.5 and 5 mg/kg AMPH. In contrast, cKO mice show no response at 2.5 mg/kg and a significantly diminished response at 5mg/kg. (Repeated measures (RM)-ANOVA was used to determine the effect of genotype and treatment (Rx) on Pathlength: Saline: Genotype effect (F(2,34)=1.299, p=0.2849). AMPH 2.5: Genotype effect (F(2,34)=3.726, p=0.0365). Rx effect (F(2,12)=8.896, p<0.0001). Interaction in Genotype and Rx (F(2,24) = 3.461, p <0.0001). Fisher PLSD reveals a difference in Flot1fl/fl and cKO (p=0.0215), cHet and cKO (p=0.0323) but not Flot1fl/fl and cHet (p=0.8623). AMPH 5: Genotype effect (F(2,34)=5.005, p=0.0124). Rx effect (F(2,12)=25.025, p<0.0001). fl/fl Interaction in Genotype and Rx (F(2,24) = 3.190, p <0.0001). Fisher PLSD reveals a difference in Flot1 and cKO (p=0.0213), cHet and cKO (p=0.0057) but not Flot1fl/fl and cHet (p=0.6020)). (B) Locomotor response to 10 mg/kg of cocaine (COC) i.p. The hyperlocomotor response was indistinguishable across genotype. (RM-ANOVA finds no effect of genotype on Pathlength (F(2,34)=1.679, p=0.2016.)) (C) Basal locomotion and habituation to the open field was indistinguishable across genotype. (RM-ANOVA finds no effect of genotype on Pathlength on any day (Day 1: Genotype effect (F(2,34) = 2.108, p = 0.1371); Day 2: Genotype effect (F(2,34) = 3.241, p = 0.0675); Day 3: Genotype effect (F(2,34) = 0.324, p = 0.7252)).

Reminiscent of our behavioral findings is our previous report that depletion of Flot1 via RNA interference led to diminished DA release in response to AMPH in dissociated dopaminergic

(DAergic) neuronal cultures [105]. To determine if we observe similar changes in the mice, we used cyclic voltammetry to characterize release and uptake dynamics of DA in response to AMPH.

Striatal slices were generated from littermate Flot1 cHet and cKO mice (Figure 4.4). cHet mice were used to control for the presence of DATiresCre [310]. Quantification revealed that whereas the

AMPH-mediated inhibition of DA release and reuptake were similar across genotypes, non- exocytic release of DA in response to AMPH was significantly reduced in cKO mice (Figure 4.4B).

Similarly, chronoamperometry studies in slice demonstrated that release of DA in response to

AMPH was blunted in the cKO mice (Figure 4.5). Taken together, these results suggest that Flot1 is required for the ability of DAT to reverse transport DA in response to AMPH.

Figure 4.3. Locomotor response in female mice shows a similar Flot1-dependence as males, with a diminished response to AMPH, and no difference in basal behavior or response to COC. Statistics

(RM-ANOVA): Saline: Genotype effect (F(2,33)=0.584, p=0.5635). AMPH 2.5 mg/kg: Genotype effect (F(2,33)=5.171, p=0.0111). Treatment (Rx) effect (F(12,396)=6.958, p<0.0001). Interaction Genotype and Rx (F(24,396) =3.448, p<0.001). Fisher PLSD reveals a difference in Flot1 and cKO (p=0.0033), cHet and cKO (p=0.0472) but not Flot1 and cHet (p=0.2762). AMPH 5 mg/kg: Genotype effect (F(2,33)=8.031, p=0.0014). Rx effect (F(12,396)=15.236, p<0.0001). Interaction Genotype and Rx (F(24,396)= 1.969, p=0.0038). Fisher PLSD reveals a difference in Flot1 and cKO (p=0.0004), cHet and cKO (p=0.0144) but not Flot1 and cHet

(p=0.1822). COC 10mg/kg: Genotype effect (F(2,33)=2.136, p=0.1342). Basal: Day 1: Genotype effect (F(2,33)=0.820, p=0.4493); Day 2: Genotype effect (F(2,33)=1.828, p=0.1766); Day 3: Genotype effect (F(2,33)=4.446, p=0.0195). Fisher PLSD reveals a difference in cHet and Flot1 (p=0.0370) and cKO (p=0.007), but not in Flot1 and cKO (p=0.4765). n=12 mice. Data shown as mean ±S.E.

Figure 4.4. Cyclic voltammetry in slice preparation from Flot1 cHet and cKO mice. Flot1 cHet mice were used given that they were indistinguishable in behavioral tasks from Flot1fl/fl mice, but control for the presence of DATiresCre, which has previously been shown to diminish endogenous DAT levels. (A) Representative traces from slice recordings from cHet (top) or cKO (bottom). The peak of the spikes represents electrically stimulated vesicular DA release, or evoked release. Each peak, recorded every 2 min, is a single-pulse electrical stimulation. The arrow indicates the time when data was taken for the change in baseline, which represents AMPH-induced efflux. This was approximately 20 min following the perfusion of 10 µM AMPH onto slices. (B) Quantification of AMPH-induced DA efflux revealed a significantly reduced response in Flot1 cKO slices (Mann Whitney U test reveals a significant difference between cHet and cKO mice: p = 0.0381. N = 5 mice, n = 15 slices/genotype). (C, D) Quantification of evoked DA release (C) and reuptake represented as t1/2 (D) after 10 µM AMPH. AMPH inhibited both DA release and DA reuptake similarly across genotypes. Similar findings were made using chronoamperometry in slice (Figure 4.5). Data represented as mean ± S.E. histogram with individual data points shown.

Figure 4.5. Chronoamperometry in striatal slice preparations from the Flot1(DAT) cKO mice demonstrate significantly diminished AMPH-induced release of DA through DAT. 10 µM AMPH was perfused onto striatal slices at 15 min, as indicated by arrow. cHET were used as a control for the presence of DATiresCRE. Statistics (RM- ANOVA): Genotype effect (F(1,16)=19.247, p=0.0005). n=3 cHET mice, 2 slices per mouse; n=4 cKO, 3 slices per mouse.

4.3 The loss of Flot1 in DA neurons does not affect basal DA neuron function

Figure 4.6 The loss of Flot1 leads to no detectable changes in the DA system. Immunohistochemistry against DAT and tyrosine hydroxylase (TH) of (A) substantia nigra and (B) striatum from coronal brain sections of Flot1fl/fl, cHet, and cKO littermates.

The diminished response to AMPH in the absence of other behavioral abnormalities strongly suggested that Flot1 was selectively affecting a discrete function of DAT. Given the complexity of how the DAergic system responds to AMPH, we next performed a series of experiments to determine if the conditional elimination of Flot1 disrupted other basal parameters of the DAergic system. Immunohistochemistry (IHC) against DAT and tyrosine hydroxylase (TH) revealed no gross neuroanatomical differences in the striatum or substantia nigra (Figure 4.6), suggesting that levels of DAT and DA are unaffected by Flot1 expression levels. HPLC analyses of striatal lysates confirmed that DA levels were unchanged (Figure 4.7). Given that IHC does not reflect the levels of functional DAT, we also examined basal DA release and uptake dynamics in dorsolateral striatum of slice preparations by single-pulse electrical stimulation every 2 min. As shown in

Figure 4.8, no significant difference in evoked DA release or DA uptake as measured by t1/2 was observed across genotype. The latter findings strongly suggest that functional DAT levels are

59 similar across genotype. These data indicate that the altered response to AMPH in Flot1 cKO mice is not due to changes to basal parameters in the DAergic system.

Figure 4.7 HPLC reveals normal DA content in the striata of Flot1 cKO mice. DOPAC as well as norepinephrine and 5- HT are similarly unchanged (not shown). ANOVA revealed no effect of genotype on DA levels (F(2,6) = 0.868, p = 0.4667)). n = 3/genotype. Data represented as mean ± S.E. with individual data points shown.

Figure 4.8 Cyclic voltammetry in striatal slice preparations also showed no difference in evoked DA release and DAT-mediated uptake under basal conditions between cHet and cKO mice. (A) Representative traces from cHet and cKO mice. (B, C) Quantification of DA release and DA uptake as represented by t1/2. No significant difference was observed in either measure. (Mann Whitney U test revealed no effect of genotype on DA release (p = 0.137) or on t1/2 (p = 0.147)). N = 6 brains, n = 16 slices/genotype. Data represented as mean ± S.E. with individual data points shown.

4.4 Flot1 is required in DA neurons for DAT partitioning into cholesterol-rich

membranes.

In heterologous systems, SLC6 transporters such as DAT have been shown to partition into DRMs upon lysis with the non-ionic detergent Brij-58 and sucrose density gradient centrifugation at 4°C

[103-105].

Figure 4.9 Sucrose density gradients of DAT striatal lysates from Flot1fl/fl mice in the absence (top) or presence (bottom) of methyl-beta- cyclodextrin (MßC). Homogenates lysed in 1% Brij- 58 were subject to step gradient of 5% to 40% sucrose, spun at 43,200 rpm for 20 hrs at 4°C then collected as 10 equal fractions. The chelation of cholesterol by MßC leads to the loss of DAT partitioning into the more buoyant fractions that are positive for Flot1.

Figure 4.10 Sucrose density gradient centrifugation reveals a differential distribution of DAT in synaptosomes from Flot1 cKO striata. Homogenates lysed in 1% Brij-58 were subject to step gradient of 5% to 40% sucrose, spun at 43,200 rpm for 20 hrs at 4°C then collected as 10 equal fractions. (Left) DAT detected in lysates from Flot1fl/fl striata partitions into Flot1-positive detergent-resistant membrane fractions (DRMs) that are cholesterol-dependent, as revealed by methyl-beta-cyclodextrin treatment (Figure 4.9). (Right) In contrast, DAT from lysates of Flot1 cKO striata no longer partitions into DRMs. Flot1-staining is derived from the non-DAT positive cells in the striatum.

Endogenous DAT has also been reported to segregate similarly [103, 200]. To determine if

61 endogenous DAT segregates into DRMs in a Flot1-dependent manner, we next isolated 1% Brij-

58 resistant DRMs from striatal lysates generated from the Flot1 cHet and Flot1 cKO mice (Figure

4.9, 10). Immunoblotting revealed that similarly to heterologously expressed DAT, endogenous

DAT partitions into DRMs in a cholesterol- Flot1-dependent manner (Figure 4.9, 10). This shift in DAT partitioning in 1% Brij-dependent DRMs is specific to DAT as the loss of Flot1 do not affect the DRM localization of other proteins in a heterologous system (Appendix A). Despite the genetic elimination of Flot1, DRMs can be continued to be defined by the fractions in which Flot1

Figure 4.11 Sucrose density gradient centrifugation reveals that DAT is insoluble in Brij 58 but soluble in Triton X-100. (A) DAT from striatal lysate floats in buoyant fractions under 1% Brij 58 extraction, but falls out of these same fractions when extracted with both (B) 0.4% and (C) 1% Triton X- 100. is present because the cKO mice only eliminate Flot1 expression from the terminals of the DAergic neurons innervating the striatum. Although DAT partitions into 1% Brij58-resistant DRMs, the membranes are soluble to another non-ionic detergent, TritonX-100 (Figure 4.11). Based on these data, we hypothesize that the Flot1-dependent partitioning of DAT into cholesterol-rich membranes is necessary to permit non-exocytic release of DA through DAT in response to AMPH.

4.5 Genetic approaches reveal protective compensatory events during development can

mask the role of Flot1 in AMPH-induced non-exocytic release of DA and in

scaffolding DAT to DRMs.

Although a recent study indicates that the depletion of flotillins disrupts vertebrate embryogenesis through the flotillins’ role to scaffold cell adhesion molecules (CAMs) [182], prior studies that created constitutive murine KOs of flotillins showed little to no phenotype in this regard [311,

312]. Given the importance of CAMs especially during development, we speculated that protective developmental compensations were triggered due to the genetic modifications used to create the

Flot1 KO and double KO mice, and that these compensatory events masked Flot1-related function

[313]. In light of these studies, we posited that although Flot1 has been implicated across different cell-based processes [105, 106, 110, 121, 166, 312-315], the role for palmitoylated, membrane- associated Flot1 would be to scaffold substrates into cholesterol-rich membranes. Thus, if compensatory events are protecting Flot1 KO mice from lethality, the event would protect the scaffolding capacity of Flot1 of its different substrates, including DAT.

To examine if Flot1 KO mice regain the ability to respond to AMPH, we created a constitutive

Flot1 knockout (Flot1 KO) mouse and examined how it responded to AMPH (Figure 4.12).

Flot1fl/fl mice were crossed with HprtCre/+, which permits deletion of Flot1 in the oocyte thereby giving rise to a heritable Flot1-deleted allele (Figure 4.12). The resulting Flot1 KO, wildtype (WT) and heterozygous (Het) littermates were monitored for their locomotor response to AMPH and

COC as described in Figure 4.2. In contrast to the cKO mice, the constitutive loss of Flot1 had no consequence on either COC- or AMPH-induced behavior, with no significant differences observed across genotype. These data suggested that compensatory events in the Flot1 KO mice may mask

Flot1 function.

Figure 4.12 Protective compensatory events during development can mask the role of Flot1 in AMPH-induced non-exocytic release of DA and in scaffolding DAT to DRMs. Constitutive loss of Flot1 does not demonstrate a blunted response to AMPH, possibly due to compensatory events during development. (A) Breeding schema of the creation of the heritable Flot1-deleted (Δ) allele and constitutive Flot1 knockout mice (Flot1 KO). Flot1 heterozygous mice are crossed to create the three genotypes (WT, heterozygous (Het), KO) used in the experiments. (B) PCR detection of Flot1 conditional allele versus deleted null allele in the WT, Het and KO mice. (C) Immunoblotting of substantia nigra and (D) whole brain lysates WT, Het and KO littermates for Flot1. Blots were also probed for TH to determine relative levels of DA. Vinculin (vinc) serves as a loading control. (E-G) Locomotor behavior of mice in the presence or absence of AMPH or COC as performed in Figure 1. In contrast to cKO mice, KO mice show a normal response to AMPH. (RM-ANOVA was used to determine the effect of genotype and Rx on Pathlength:

Basal: Day 1: Genotype effect (F(2,30)=1.031, p=0.3689); Day 2: Genotype effect (F(2,30)=0.644, p=0.5321). Saline: Genotype effect (F(2,30)=0.257, p=0.7752). AMPH 2.5: Genotype effect (F(2,30)=0.429, p=0.6553). Rx effect (F(12,360)=14.985.896, p<0.0001). AMPH 5: Genotype effect (F(2,30)=0.189, p=0.8286). Rx effect (F(12,360)=32.954, p<0.0001). COC 10: Genotype effect (F(2,30)=.393, p=0.6783)). n=12/genotype. Data represented as mean ±S.E.

In light of the normal response to AMPH mounted by the Flot1 KO mice, we next determined if the lack of phenotype was due to events during development by creating a conditional mouse model that permits temporal regulation of Flot1 deletion. Flot1fl/fl mice were crossed with mice expressing a tamoxifen (tam)-inducible Cre driven by the chicken beta actin promotor

(ActinCreERTM), which permits global Cre-mediated gene excision upon administration of tam

(Figure 4.13B). Breeding ultimately gave rise to the Flot1 inducible knockout (Flot1 iKO), heterozygotes (Flot1 iHet) and the Cre-negative control Flot1fl/fl (Figure 4.13). As described in methods, a one-week regimen of tam was administered to all mice at 4-weeks-old (w/o) which led to a significant loss of Flot1 expression (Figure 4.13C, D). These mice were then assessed for a locomotor response to AMPH and COC four weeks later. iHet and iKO mice showed basal locomotion and habituation to the open field that was comparable to the Flot1fl/fl control mice

(Figure 4.13E). In contrast to the KO, the iKO demonstrated a similarly blunted response to

AMPH, with no response at 2.5 mg/kg and a diminished response at 5 mg/kg (Figure 4.13G).

Unlike the cHet, however, the iHet shows a similar response to that of the iKO (Figure 4.13G).

Female iHet and iKO mice also demonstrated a blunted response toAMPH, but at 5 mg/kg, iHet females demonstrated a response intermediate to Flot1fl/fl and iKO littermates (Figure 4.13J).

Again, the response to COC was independent of Flot1 (Figures 4.13F). The data from the iKO and

KO mice confirms that the genetic loss of Flot1 can be masked by compensatory changes during development, and strongly support the importance of Flot1 in AMPH-induced behaviors.

Figure 4.13 Adult-inducible deletion of Flot1 recaptures the Flot1-dependent response to AMPH. (A) Breeding schema of creation of the Flot1 inducible KO mouse (Flot1 iKO). Crossing the conditional Flot1 allele with a tamoxifen-inducible Cre line, ActinCreERTM leads to the creation of mice in which Flot1 can be deleted in a temporally regulated manner through the administration of tamoxifen (tam). For all experiments, Flot1 iKO (Flot1fl/fl::ActinCreERTM/+) mice were compared to Flot1fl/fl and Flot1fl/+::ActinCreERTM/+ (Flot1 iHet) littermates. (Continued on next page)

Figure 4.13 Adult-inducible deletion of Flot1 recaptures the Flot1-dependent response to AMPH. Continued from previous page. (B) PCR detection of Flot1 conditional allele versus WT allele (top), the presence of Actin CreERTM (middle) and the null allele (bottom) in Flot1fl/fl, iHet and iKO mice. (C) Immunoblotting of substantia nigra and (D) whole brain lysates show that administration of tam at 4 wks of age leads to efficient loss of Flot1. (E-G) Open Field paradigm performed in 9 week-old male mice. Unlike the Flot1 KO mice, but similarly to the Flot1 cKO mice, tam-treated Flot1 iKO mice showed a diminished response to AMPH, whereas basal locomotion and COC-induced hyperlocomotion was normal. Interestingly, in contrast to the Flot1 cHet mice, the Flot1 iHet mice also demonstrated a diminished response to AMPH. All mice were administered tam at age 4 weeks. (RM-ANOVA was used to determine the effect of genotype and Rx on Pathlength: Basal: Day 1: Genotype effect (F(2,27)=2.388, p=0.1109); Day 2: Genotype effect (F(2,27)=1.863, p=1.746). Saline: Genotype effect (F(2,27)=1.095, p=0.3488). AMPH 2.5: Genotype effect (F(2,27)=11.171, p=0.0003). Rx effect (F(12,324)=13.792, p<0.0001). Interaction in Genotype fl/fl and Rx (F(24,324) = 5.550, p <0.0001). Fisher PLSD reveals a difference in Flot1 and iKO (p=0.0002), fl/fl Flot1 and iHet (p=0.0009) but not iHet and iKO (p=0.5077). AMPH 5: Genotype effect (F(2,27)=4.839, p=0.0160). Rx effect (F(12,324)=38.337, p<0.0001). Interaction in Genotype and Rx (F(24,324) = 2.415, p =0.0003). Fisher PLSD reveals a difference in Flot1fl/fl and iKO (p=0.0174), Flot1fl/fl and iHet (p=0.0087) but not iHet and iKO (p=0.7703). COC 10: Genotype effect (F(2,27)=0.257, p=0.7754). n=10/genotype. Data represented as mean ±S.E. (H-J) Locomotor response in iKO female mice show similar Flot1-dependence as iKO males, with a diminished response to AMPH, and no difference in basal behavior or respose to COC. In contrast, however, female iHet mice show a partial response at 5 mg/kg AMPH. Statistics (RM-

ANOVA): Basal: Day 1: Genotype effect (F(2,25)=2.213, p=0.1304); Day 2: Genotype effect (F(2,25)=1.598, p=0.2223). Saline: Genotype effect (F(2,25)=0.680, p=0.5157). AMPH 2.5 mg/kg: Genotype effect (F(2,25)=19.139, p<0.0001). Treatment (Rx) effect (F(12,300)=16.742, p<0.0001). Interaction Genotype and Rx

(F(24,300)=10.635, p<0.0001). Fisher PLSD reveals a difference in Flot1 and iKO (p<0.0001), Flot1 and iHet

(p<0.0001) but not iKO and iHet (p=0.3580). AMPH 5 mg/kg: Genotype effect (F(2,25)=5.716, p=0.0090). Rx effect (F(12,300)=55.078, p<0.0001). Interaction Genotype and Rx (F(24,300)= 2.178, p=0.0014). Fisher PLSD reveals a difference in Flot1 and iKO (p=0.0025). iHet showed an intermediate response showing no difference with Flot1 (p=0.0708) and iKO (p=0.1622), COC 10mg/kg: Genotype effect (F(2,25)=0.112, p=0.8942). n=10,9,9 mice (Flot1, iHet, iKO). Data shown as mean ±S.E.

In light of the marked behavioral difference between the constitutive and conditional genetic deletion of Flot1, we next determined if the difference was reflected in the partitioning of DAT.

Sucrose density gradient centrifugation was performed on striatal lysates generated from Flot1 KO and iKO mice. The Flot1-positive DRM fractions and total lysates of control and experimental genotypes were run together on western blots to quantify partitioning with respect to control

(Figure 4.14). Unlike the Flot1 cKO mice (Figure 4.10), the constitutive loss of Flot1 led to only a small change in DAT partitioning in Flot1 KO tissue (Figure 4.14A), to a similar extent

67 previously described for Abeta in the Flot1 dKO mice [312]. In contrast, the loss of Flot1 in the

Flot1 iKO mice led to a significant decrease of DAT partitioning into all DRM fractions (Figure

4.14B), which was to a similar degree as in Flot1 cKO mice (Figure 4.10). Thus, the lysates of mice whose AMPH-induced response is diminished by the loss of Flot1 (Flot1 cKO and Flot1 iKO) also demonstrated diminished partitioning of DAT into DRMs, but lysates from mice that maintained their ability to respond to AMPH also maintained DAT partitioning. These data further strengthen the correlation that the ability of DAT to reverse transport DA in response to AMPH is related to the ability of DAT to partition into cholesterol-rich DRMs.

Figure 4.14 DAT from Flot1 iKO but not Flot1 KO mice fails to partition into DRMs. Sucrose density gradient (SDG) centrifugation on 1% Brij-58 striatal lysates from (A) Flot1 KO and (B) iKO mice. To quantify the relative partitioning of the genetically modified mice to their respective controls, the Flot1- positive DRM fractions and total lysates of control and experimental genotypes were run together for immunoblotting. Quantification revealed that in Flot1 KO lysates, a small change of DAT partitioning in the more buoyant fraction 4 was observed, whereas in Flot1 iKO lysates, a significant decrease in DAT partitioning into both the DRM fractions was observed. Mann Whitney U test of Ctrl (WT) vs. Flot1 KO fractions reveals a significant difference in fraction 4 (p = 0.045) but not in fraction 5 (p = 0.4867). Similar test of Ctrl (Flot1fl/fl) + tam vs. Flot1 iKO + tam reveals a significant difference in both fractions 4 (p = 0.045) and 5 (p = 0.045). n = 3/genotype. Data represented as mean ± S.E., with individual data points also shown.

4.6 DAT activity influences the partitioning of DAT into detergent-resistant, cholesterol

rich membranes.

Foster and Vaughan found that similarly to Flot1, DAT is subject to palmitoylation, a reversible post-translational modification that can increase the ability of proteins to partition into sterol and sphingolipid membranes [314]. Given that AMPH uptake by DAT would be achieved similarly to

DA [315] and there is a strong correlation between DAT partitioning and efflux, we hypothesized that AMPH increased DAT partitioning into cholesterol-rich membranes (Figure 4.15). We and others have previously shown that Flot1 palmitoylation is necessary for Flot1 function [105, 118], and therefore global inhibition of this post-translational modification would not test the significance of DAT palmitoylation alone. Instead, we examined whether modulating DAT activity via AMPH or COC would impact partitioning (Figure 4.15A).

DRMs derived from striatal homogenates from WT mice treated with a single injection of 5 mg/kg

(i.p.) of AMPH or saline were probed for DAT and Flot1. AMPH treatment led to a significant increase in the amount of DAT in the Flot1-positive DRM fraction relative to saline treatment.

This increase was Flot1-dependent, as AMPH treatment had no detectable impact on the distribution of DAT in Flot1 cKO striata (Figure 14.6A). In contrast, DAT inhibition by a single administration of COC was sufficient to evoke a measurable decrease in DAT partitioning to

DRMs (Figure 14.5B). This is consistent with the conclusion drawn from the Flot1 cKO and Flot1 iKO mice, which responded normally to COC; partitioning to DRMs is not necessary to observe the drug’s effect. These data indicate that DAT activity leads to increased DAT partitioning in a

Flot1-dependent manner, whereas the inhibition of DAT activity causes depletion of DAT in buoyant fractions. Flot1 therefore may be necessary for the ability of DAT to respond to AMPH because increased DAT activity requires localization of DAT to cholesterol-rich membranes.

Figure 4.15 DAT activity modulates the partitioning of DAT in cholesterol-rich membranes. (A) AMPH increases the detectable levels of DAT in DRMs. Striatal homogenates from WT mice treated with a single injection of 5 mg/kg (i.p.) of AMPH or saline were lysed in 1% Brij-58, and run on SDGs. AMPH treatment led to a significant increase in the amount of DAT in the Flot1-positive DRM fraction (Mann-Whitney U: Fraction 4, p=0.0209, Fraction 5, p=0.0209; n=4). AMPH treatment had no detectable impact on the distribution of DAT in Flot1(DAT) cKO striata (Figure 4.16). (B) Inhibition of DAT via COC diminishes the detectable levels of DAT in DRMs. Striatal homogenates from WT mice treated with a single injection of 10 mg/kg (i.p.) of COC or saline were lysed in 1% Brij-58, and run on SDGs. In contrast to AMPH, COC leads to a significant decrease in the amount of DAT in the DRM fractions ((Mann-Whitney U: Fraction 4, p=0.0209, Fraction 5, p=0.0209; n=4).

4.7 Partitioning of DAT into cholesterol rich membranes and Flot1 is not required for

DAT phosphorylation.

We next aimed to establish why partitioning of DAT into different membrane environments was necessary for AMPH-induced efflux. A hypothesis often posed for membrane microdomain function is that these protein-lipid structures serve to increase efficiency of post-translational modifications on target proteins [316]. In the case of DAT, studies have established that the phosphorylation of the NH3-terminus DAT is necessary for the reverse transport of DA [207]. For example, Ser7 and Ser12 via phosphorylation by protein kinase C and CaMKII α/β have been

70 shown to be two critical residues for efflux [207, 317]. We therefore examined whether phosphorylation of DAT at these residues required Flot1. Because the antibody against phospho-

Ser7 is against human DAT [105], we turned to a well-established heterologous expression system of DAT, YFP-DAT stably expressed in modified HEK293 cells (EM4 YFP-DAT) [318, 319]. We found phosphorylated Ser7 can be detected under baseline conditions which can be readily augmented by PKC activation via PMA as previously reported (Figure 4.16B) [320-322].

Depletion of Flot1 had no detectable impact on basal or PKC-mediated phosphorylation. We also examined the DRM distribution of endogenous CaMKII α and β from mouse striata and found that the distribution not only differed from DAT, but was independent of Flot1 and AMPH (Figure

4.16C). These data suggest that not only is the phosphorylation of Ser7 independent of basal or kinase-induced Flot1, in striatal lysates, the DRM profile of DAT and CaMKII kinases are distinct.

Finally, to examine phosphorylation of endogenous murine DAT, we also examined Thr53 phosphorylation in DRMs from Flot1 cKO and control striata. Thr53 phosphorylation has been suggested to be important for AMPH-induced efflux although how and why this phosphorylation event occurs is unclear [291]. Given that it is the only site for which an antibody against its phosphorylated form was available for mouse DAT, we determined its dependence on Flot1 and

AMPH. Immunoblotting of gradient fractions with the phospho-DAT(Thr53) antibody revealed that although the distribution of DAT, and therefore phospho-DAT(Thr53), was Flot1-dependent

(see also Figure 4.10) phosphorylation continued to occur in the absence of Flot1 and on DAT in the absence of its ability to partition into DRMs (Figure 4.16D). AMPH had no effect on phosphorylation as well. Taken together, although phosphorylation of DAT along its N-terminus has been shown to be a critical regulator of AMPH-induced efflux, our data suggest that it is not dependent on Flot1 or DAT partitioning.

Figure 4.16 Phosphorylation of DAT occurs independently of Flot1. (A) SDGs of Flot1fl/fl control or Flot1 cKO mice after 5 mg/kg AMPH i.p. Fractions were probed for DAT and Flot1. (B) Residue Serine 7 of hDAT can be phosphorylated (P-hDAT(Ser7)) in response to PMA despite Flot1 depletion using siRNA against human Flot1 (si hFlot1). Experiments were performed on stably expressed hDAT in EM4 cells. DAT is immunoprecipitated under the conditions shown. Western blotting of the immunoprecipitates by an antibody against phosphorylated Ser7 shows that DAT is phosphorylated under basal conditions, and this phosphorylation can be increased by the activation of PKC with PMA. Blots below show corresponding total lysates probed for Flot1 and γ-tubulin as a loading control. (C) The distribution of CaMKII α and β as revealed by SDGs of 1% Brij-58 striatal lysates. Note how the distribution is Flot1- and AMPH- independent, and different from DAT and Flot1. (D) Fractions from SDGs of 1% Brij 58 striatal lysates from Flot1fl/fl and Flot1 cKO mice. Probing with an antibody raised against DAT(Thr53) (P-DAT(Thr53)) reveals that Thr53 is constitutively phosphorylated in Flot1fl/fl and P-DAT(Thr53) is present in buoyant and dense fractions. In the presence of AMPH, more P-DAT(Thr53) is apparent in the buoyant fractions. DAT also becomes phosphorylated after AMPH in Flot1 cKO striata. The corresponding immunoblot against DAT can be found in Figure 4.16C top left for Flot1fl/fl saline, and Figure 4.16A for Flot1fl/fl and Flot1 cKO AMPH.

Although the loss of Flot1 did not inhibit phosphorylation of DAT(Thr53), it appeared diminished compared to control. Quantification of the levels of DAT phosphorylation in the presence or absence of Flot1 indicated that there is greater than two times higher levels of phosphorylated DAT in WT striata than Flot1 cKO (Figure 4.17). These data suggest that rather than being necessary for phosphorylation, membrane partitioning might protect the phosphorylated form of DAT which is essential for efflux. Further replicates will confirm whether this is indeed the case.

Figure 4.17 Flot1 protects phosphorylation of DAT. Co-immunoprecipitation from striatal lysates of Flot1fl/fl and cKO reveal that there is less phosphorylated DAT at Thr53 in the absence of Flot1. (n=2)

Given our findings thus far, we propose that the Flot1-dependent partitioning of DAT is increased in the presence of AMPH, possibly due to palmitoylation from increased DAT activity [314, 323], whereas partitioning is decreased in the presence of COC, which inhibits DAT uptake and possibly diminished palmitoylation. Given that phosphorylation of DAT can occur in a Flot1-independent manner, is detectable in the absence of AMPH, and that phosphorylated DAT can be detected in the DRMs (Figure 4.16), we further suggest that phosphorylation occurs outside of cholesterol- rich membranes, possibly in response to basal and AMPH-induced DAT activity, and this phosphorylation status is protected upon entry of DAT into DRMs (Figure 4.18).

Figure 4.18 Working model 1. Flot1-dependent partitioning of DAT is augmented upon increased DAT activity. We propose that phosphorylation occurs outside of the DRM in response to DAT activity, and is normally dephosphorylated. Upon increased activity and partitioning of DAT, the phosphorylation status of DAT is protected, which in turn helps to augment DAT Vmax to help to promote efflux. The DRM depicted here is to simply indicate the Flot1-dependent partitioning of DAT.

4.8 DAT localized to cholesterol-rich membranes are found in a conformation distinct

from DAT localized in cholesterol-poor membranes.

Lipid-protein interactions have been shown to profoundly influence membrane protein structure and function [324-326]. For example, the biophysical properties of the membrane microenvironment have been shown to impact the structure and function of ion channels to slow or accelerate their responsiveness [327]. An importance for cholesterol in DAT protein folding has already been proposed in the structure of dDAT [102]. We therefore next tested if DAT partitioning was necessary for AMPH-induced reverse transport because the cholesterol-rich membrane composition stabilized a conformation of DAT that promoted its activity.

To gain insight into the potential conformational differences present in the DRM versus non-DRM fractions, we performed limited proteolysis using different proteases to determine if different cleavage susceptibilities could be detected across fractions (Figure 4.19). We used trypsin, papain and 2-Nitro-5-thiocyanatobenzoic acid (NTCB). The protease cleavage sites are summarized in

Table 4.1. The fractions were exposed to each protease for the times indicated then processed for western blot analyses. All three enzymes revealed that the susceptibility of cleavage notably differed across the two sets of fractions: DAT present in DRMs were more susceptible to trypsin and papain cleavage than DAT present in non-DRMs, but less susceptible to NTCB cleavage. BSA controls indicated that this cleavage difference is due to a difference in DAT rather than differences in sucrose concentration across fractions (Figure 4.19). These data indicate that DAT in the Flot1- positive DRMs are conformationally distinct from DAT in non-DRM fractions.

Protease Cleavage Sites

Arg * not Pro Trypsin Lys * not Pro

Hydrophobic residue-Arg * not Val Papain Hydrophobic residue-Lys * not Val

2-Nitro-5-thiocyanatobenzoic acid (NTCB) Cys at position P1’

Table 4. 1 Cleavage sites of specified proteases. * denotes cleavage site.

Figure 4.19 Limited proteolysis reveals that DAT in DRMs is in a distinct conformation. Limited proteolysis of Flot1-positive versus Flot1-negative SDG fractions using (A, B) trypsin, (C) papain or (D) NTCB. The Flot1-positive fractions 3 and 4 were pooled and compared to the combined dense fractions 9 and 10. (A) Low and (B) high concentrations of trypsin revealed that the Flot1-positive fractions were more susceptible to proteolysis. (C) Papain, which has similar proteolytic cleavage sites to trypsin yielded similar results. (D) NTCB which has a different proteolytic capacity from trypsin and papain, also yielded differential proteolysis across fractions, but in this case, the Flot1-positive fractions showed increased resistance. (E) BSA controls indicate that the different sucrose levels do not affect proteolytic rates as BSA is proteolyzed similarly in both lower and higher concentrations of sucrose in the buoyant and dense fractions respectively. The amount of proteolysis appears not to change over time, suggesting that trypsin works primarily in the first minute.

Figure 4.20 The interaction of PIP2 with the N-terminus of DAT is important for DAT partitioning into DRMs. (A) Heterologous expression of hDAT or hDAT(K3A, K5A)(hDAT(KA)) in HEK293 cells. hDAT is stained for immunofluorescence with MAB369 (green) and Flot1 with BD#610821 (red). Mutagenesis does not grossly affect the membrane localization of DAT. (B) Co-immunoprecipitation of

PIP2 with DAT. Alanine scanning mutant of hDAT confirms the importance of Lys3 and Lys5 for PIP2 interaction (n=1) (C) hDAT(KA) does not partition into DRMs as well as hDAT. SDG analyses of 1% Brij- 58 revealed that the hDAT(KA) mutant continues to segregate to DRMs but to a lesser degree than hDAT WT (Mann-Whitney U test: Fraction 4, p=0.0495; Fraction 5, p=0.0495; n=3). The distribution of Flot1 does not appear affected (Figure 4.20C), suggesting that the diminishing the efficiency with which DAT interacts with PIP2 also affects its ability to remain within cholesterol-rich membranes.

4.9 The conformation of DAT in DRMs depends on its N-terminal interaction with PIP2

Recently, Galli and colleagues found that Lys3 and Lys5 of human DAT interact with the phospholipid PIP2 which leads to a change in conformation that favors DAT efflux in the presence of AMPH [306]. Loss of the PIP2 interaction by mutagenesis of these two residues diminished

AMPH-induced efflux and the ability of Drosophila to induce a hyperlocomotor response to

AMPH [306]. We therefore tested whether the conformational difference we detected in the DRMs

77 was due to PIP2 interaction with DAT. Wild-type hDAT (hDAT(WT)) or hDAT(K3A, K5A)

(hDAT(KA)) was transiently expressed into HEK293 cells. Alanine mutagenesis of Lys3 or 5 of

DAT has little to no impact on the ability of DAT to traffic to the plasma membrane (Figure

4.20A). Co-immunoprecipitation experiments revealed that even though hDAT(KA) is expressed at the cell surface, its interaction with PIP2 is diminished, as previously shown (Figure 4.20B)

[306]. We next examined membrane partitioning of hDAT(KA). On a SDG run with 1% Brij-58, hDAT(KA) can be detected in the DRMs, but it did not partition as efficiently as hDAT(WT)

(Figure 4.20C). This suggested that the diminished interaction between DAT and PIP2 affected the ability of hDAT to partition into DRMs.

Figure 4.21 PIP2 interaction with the N-terminus of DAT is necessary for the DRM-dependent conformation of DAT. Limited proteolysis reveals that K3 and K5 of hDAT is required for the increased susceptibility of bouyant fractions to trypsin-digestion. Similarly to endogenous DAT, buoyant fractions from hDAT(WT) lysates are more susceptible to proteolysis by trypsin than in dense fractions. In contrast, mutations within the PIP2 interaction site in the N-terminus of DAT changes the conformation of DAT in the buoyant fractions but not the dense fractions, rendering the buoyant fractions of hDAT(KA) mutants are relatively resistant to proteolysis. SDG fractions 9 and 10 of transiently expressed hDAT WT or hDAT(KA) show that mutations of Lys3 and Lys5 of hDAT does not change the relative resistance to trypsin proteolysis.

To confirm this observation and to determine if the interaction with PIP2 accounts for the conformational difference of DAT in DRMs, we next performed limited proteolysis by trypsin between DRM and non-DRM fractions from hDAT(WT) and hDAT(KA) as described in Figure

4.19. The resulting proteolytic pattern of hDAT(WT) was similar to the pattern for endogenous

DAT, for which hDAT in the DRM fraction cleaved faster than hDAT in the non-DRM fractions

(Figure 4.21). In contrast, hDAT(KA) was similarly resistant to trypsin across both fractions

(Figure 4.21), such that hDAT(KA) in fractions 3 and 4 continue to be resistant to trypsin even after 30 min. Taken together, the data demonstrate that DAT interacts with PIP2 in DRMs. This suggests that in the Flot1 cKO and iKO mice, the diminished partitioning of DAT permits less interaction with PIP2.

4.10 The working model of membrane-DAT interaction for AMPH induced efflux of DA

through DAT.

Figure 4.22 Model of how Flot1 facilitates the AMPH-dependent reverse transport of DA through DAT. Flot1 scaffolds DAT into DRMs in the presence of AMPH to aid increased activity. For the sake of simplicity, the DRM in this model denotes the 1% Brij58-resistant environment into which DAT partitions in the presence of Flot1. The DRMs of our biochemical experiments are likely a heterogeneous mixture of different DRM populations that are resistant to 1% Brij58. Our data suggests that the loss of Flot1 does not disupt DRMs, given that other proteins continue to partition in a Flot1-independent

manner. The DRM depicted in this model provides the environment to help stabilize the PIP2 interaction of the DAT N-terminus, giving rise to the different conformation in the Flot1-positive DRMs. It is of

note that PIP2 is not necessarily enriched in DRMs [328, 329]. We speculate that DRMs may also contain additional but not yet identified protein partners of either Flot1 or DAT to facilitate efflux. COC, which inhibits DA uptake, leads to a movement of DAT out of DRMs, possibly due to a decrease in activity.

Using mouse genetics and biochemistry, we have found that the partitioning of DAT occurs in a

Flot1-dependent manner in DAergic neurons of mouse brain, and this partitioning is necessary for 79

AMPH-induced reverse transport of DA through DAT. Thus building from our working model above, activity-dependent partitioning of DAT into cholesterol-rich membranes stabilizes the phosphorylated form of DAT, and the ability of DAT to interact with PIP2. Given that PIP2 is not necessarily enriched in DRMs [328, 329], it is possible that the membrane environment helps to stabilize the PIP2-bound conformation, thereby permitting AMPH-induced efflux to occur more efficiently (Figure 4.22). When DAT scaffolding to these membranes is lost, such as in the Flot1 cKO or iKO mice, due to either diminished phosphorylation or less stability with the membrane,

AMPH-induced efflux is significantly diminished, requiring higher doses of AMPH.

Chapter 5

General Discussion

5.1 Overview

Membrane studies have come a long way since the original proposal of the fluid mosaic model, but only recently has the intersection of proteins and lipids been acknowledged. We now know that the interaction of membrane proteins and their lipid environment serve to regulate a large array of cellular functions [33, 326, 330], however our understanding of the role of membrane lipids in protein function still remains incomplete. For a long time, discrete regions within the plasma membrane such as membrane rafts were modeled to be the primary platform for lipid-protein interactions. Now, this concept has been extended to encompass the plasma membrane (PM) in its entirety, given that the PM may be capable of restructuring itself in response to protein activity.

Regardless of the model, establishing the physiologic relevance of how the local lipid environment can influence individual protein function has proven to be difficult.

Studies have shown that SLC6 transporters can partition into DRMs, suggesting a role for lipids in their function. This led us to focus on DAT, which like other family members, can partition in part to cholesterol-rich membranes. DAT is important for DA uptake, which regulates many aspects of DA signaling. The wide reaching functions of the DAergic system warrants a closer look at the role of membranes in this process. In this regard, studies have been limited to the heterologous expression of transporters and their transport function measured in the presence of

81 cholesterol-depleting drugs, namely methyl-beta-cyclodextrin (MβC) [103, 104, 307, 331].

Because these findings were confined to in vitro or cell-based systems, the larger physiologic significance of the relationship between membrane rafts and protein function remained unclear.

Recently, it was discovered that Flot1 could scaffold DAT into DRMs and this DRM association was abolished by RNA-mediated knockdown of Flot1 [105]. The depletion of Flot1 also had selective, functional consequences for DAT; DA uptake was unaffected whereas AMPH-induced efflux was diminished [105]. To determine the relationship between DRM localization of DAT and DAT function, as well as to establish if we can replicate the cell-based findings with endogenous DAT in vivo, we created mice in which we genetically eliminated Flot1 expression.

To our knowledge, our findings represent the first in vivo demonstration of how affecting the local membrane environment of a membrane protein, as visualized by DRM isolation, can impact its activity in the brain.

Using conditional genetic approaches, we found that the loss of Flot1 in DAergic neurons led to a diminished AMPH-induced hyperlocomotor response due to diminished efflux despite an otherwise normal DAergic system. Our data indicate that Flot1 is required for efflux as it maintains

DRM localization of DAT, the partitioning of which occurred in a DAT-activity dependent manner, as evidenced by our data showing that more partitioning into DRMs occurred in response to AMPH, but less in response to COC. Limited proteolysis showed that DAT moves into DRMs and this stabilizes a distinct conformation that reflects the NH3-terminus of DAT interacting with

PIP2. Overall, our results demonstrate that in the presence of AMPH, DAT partitions into DRMs in a Flot1-dependent manner so that the cholesterol-rich environment can promote a conformation favorable for efflux.

An additional unexpected finding came from the different genetic approaches we used to eliminate

Flot1 expression. Whereas conditional loss of Flot1 led to a clear impact on AMPH-induced efflux and partitioning of DAT into DRMs, constitutive loss of Flot1 KO led to the contrary, demonstrating no behavioral or biochemical phenotype. These data reflect the importance of Flot1 in the developing organism, most likely through the different substrates Flot1 has been implicated to scaffold such as cell adhesion molecules [178]. Our data demonstrate how compensatory events during development can mask Flot1 function. This work gives insight into the lack of or mild phenotype of previous studies working with the constitutive KO of flotillins [311, 312].

5.2 The membrane environment of DAT and AMPH-induced efflux

5.2.1 The membrane environment influences DAT function

Our findings indicate that although Flot1 is necessary for a full response to AMPH, hyperlocomotion is not completely abolished upon its loss: A high dose of AMPH (5 mg/kg) leads to an impaired hyperlocomotor response, and cyclic voltammetry and amperometry reveal a blunted response to 10 µM AMPH. What accounts for this residual activity? One possibility is that

Flot1 might increase the efficiency of scaffolding into DRMs, but in its absence, some partitioning may still occur. Given the increased palmitoylation with greater DAT activity, it is possible that high concentrations of AMPH lead to more DAT that can overcome the need for Flot1 to accumulate it into cholesterol-rich membranes. A second possibility is that a higher dose of AMPH might simply lead to a greater contribution of the competitive inhibition of DAT by AMPH. Given that uptake is unaffected by the loss of Flot1, there would be an AMPH-dose dependent impact in this regard. A third alternative is the DRM environment may be necessary to achieve the full

83 potential of DAT transport activity. Thus in the absence of Flot1, efflux can occur but with less efficiency. This interpretation is based on studies showing that cholesterol enrichment of plasma membranes can lead to increased calcium (Ca2+) flux via the Ca2+ channel [332, 333], or can augment the exchange of sodium (Na+) and Ca2+ by the Na+/Ca2+ exchange protein on isolated cardiac sarcolemmal vesicles [334, 335] by modulating their activity. Additionally, structural modeling has shed light on the importance of lipid-protein interactions [38, 326, 336, 337], especially if cholesterol is shown as necessary for the protein to achieve optimal function. For

DAT, cholesterol-dependence might be an activity-dependent event. Given that COC inhibits DAT function, there is no need for cholesterol to aid in any structural adjustments. In contrast, AMPH causes frequent and sustained movement of DAT, potentially necessitating the help of DRM components for this structural change. These data suggest that Brij58-dependent detergent- resistant membranes (DRMs) may provide an environment that supports protein activity but is not absolutely essential.

Aside from cholesterol, phospholipids also participate in aiding protein function. The phospholipid

PIP2 was found to interact with the DAT NH3-terminus, and molecular modeling showed that this association brings the NH3-terminus closer to the membrane [306, 338]. This in turn facilitates

AMPH-induced efflux as mutations that disrupt this interaction ablate the ability of Drosophila to respond to AMPH. Our data suggest that the PIP2 interaction is stabilized in the DRMs thereby contributing to the distinct conformation of DAT that is revealed by limited proteolysis. Although hDAT(KA) cannot interact with PIP2, it continued to partition into DRMs (albeit to reduced levels). Despite its localization, the DRM-localized hDAT(KA) did not demonstrate the differential proteolysis like WT. This strongly suggests that both a cholesterol-rich environment and PIP2 interaction contribute to the specific DAT conformation in DRMs. Previous studies using

84 computational modeling indicated that PIP2 in lipid membranes would spatially orient and stabilize a NH3-terminal DAT peptide which was otherwise completely destabilized in water [338]. The orientation of the NH3-terminus is key because it may serve as a switch to promote or suppress

AMPH-induced efflux, by converting from a generally mobile state to a fixed state upon introduction of AMPH [339]. Additionally, whereas our data indicates that a PIP2-bound conformation is detected in the DRMs, it is still unclear why this occurs in these membranes, given that PIP2 is not enriched in cholesterol-rich membranes per se. It is possible that DRM association may hold the lever in the “on” position so DAT is committed to efflux via a permissive conformation. This might be further stabilized by increasing the probability of interaction with

PIP2. One caveat of the structural studies is that many are focused on a segment of the protein and although the microscale observations are important, ultimately an understanding of how all twelve transmembrane segments of DAT is affected will yield more insight.

The more recently proposed flexible surface model (FSM), which ascribes a greater and more dynamic role for lipids in protein-lipid interactions, may also explain what may be happening with

DAT when it effluxes. The FSM relays that the membrane is highly elastic, remodeling either in response to protein movement or to help facilitate movement, via curvature sensing [340-342].

This also implies that the protein can alter the bilayer curvature to help itself facilitate structural changes. Therefore, DAT may require and cooperate with cholesterol and PIP2 in the DRMs to bend the bilayer, which in turn, tugs at DAT to drive it to consistently reverse transport DA.

Additionally, cholesterol may be needed for its rigidity to stabilize the constant movement of DAT effluxing [343]. Our study implicates membrane components but does not go into the detail that only biophysical studies can provide.

5.2.2 Phosphorylation of DAT

In addition to interaction with PIP2, previous studies implicate phosphorylation of the NH3- terminus of DAT by either PKC or CaMKII to contribute to AMPH-induced efflux. Because it was possible that the kinases were directed to the same Brij58-dependent DRMs to phosphorylate

DAT, we looked at the DRM distribution of CaMKII and murine DAT phosphorylated at Thr 53.

Our data not only suggest that CaMKII may be in a discrete membrane composition than Flot1 and DAT, but its buoyancy is also independent of Flot1 or AMPH treatment, suggesting that the regulation of its localization is also distinct. Moreover, the phosphorylation event itself does not depend on Flot1. Further studies establishing whether CaMKII is present in Triton X-100 insoluble membranes, unlike DAT, will further establish the membrane differences. Nonetheless, based on the FSM model, membrane bending may be needed for CaMKII-mediated phosphorylation of

DAT, during which CaMKII (shown to be able to directly associate to a DAT COOH-terminus peptide [211], must “tip over” to phosphorylate the N-terminus. Perhaps the non-DRM environment contains the appropriate lipids for that event.

5.2.3 AMPH-induced internalization of DAT

One aspect of AMPH activity that we have not addressed is that previous studies have indicated that AMPH administration causes DAT internalization [300]. We have previously shown that

PKC-mediated internalization of heterologous DAT was Flot1-dependent [105], although this has been disputed [344]. Disruptions in DAT trafficking could potentially confound our study, but our

FSCV data indicates that this is not a concern. First, under basal conditions, DA uptake, represented by t1/2 (Figure 4.8C), is statistically insignificant between the two genotypes, and if

86 anything is on average faster in the absence of Flot1, suggesting that surface expression of DAT is not diminished (and if anything, gained) by the loss of Flot1. Thus, Flot1 does not impact basal

DAT recycling in vivo. Furthermore, under 10 µM AMPH, DA uptake also does not differ between cHet and cKO striatal slices, suggesting that AMPH is not leading to a differential internalization of DAT, at least within the first 25 minutes of AMPH administration.

In addition to serving as a control, the FSCV data may shed further insight into the AMPH-induced temporal regulation of DAT cell surface expression and subsequent internalization, given that the time frame of the latter has largely been unexplored. It is possible that DAT partitioning into DRMs may prevent it from being internalized and dampening the effects of AMPH, by protecting the phosphorylated form of DAT. This is in agreement with a study that found DAT in membrane rafts were more phosphorylated and less likely to be internalized in a PKC-dependent manner, compared to DAT in non-membrane rafts [201].

5.2.4 DRMs versus membrane microdomains

Finally in light of our study focused on Flot1, it comes into question whether the membranes that are represented in the DRMs are classic membrane rafts as defined by Simons and Ikonen [44].

We would argue that the based on its solubility, the DRMs we study here cannot be equated to the classic membrane rafts, although they certainly are a membranous entity containing very specific proteins. DAT-containing DRMs are insoluble in 1% Brij-58 but soluble in 1% and even 0.4%

Triton X-100. Nonetheless, our observation is inconsistent with others who have found that DAT is insoluble in different concentrations of Tx-100 [200, 201], but this may be a result of different temperatures, spin speeds and times, which highlights a pitfall of using this biochemical

87 approach.Although it looks like there is less DAT present in the Tx-100 gradients, it is not due to less DAT in the starting material since these preparations were done in wild-type mice, which should contain the same amount of DAT whether extracted with Brij 58 or Tx-100. Given the increased harshness of Tx-100, it is likely that DAT recovery after lysis in preparation of the sucrose gradient was diminsihed. The variability presented by the different detergents shows how complex these studies are and how difficult it is to capture the membrane components that are participating in a specific cellular event. Nonetheless, the 1% Brij58-DRMs represent a highly reproducible biochemical entity, given that it is consistent across cell types for both endogenous and heterologously expressed DAT, and is common to other SLC6 family members. This suggests that Flot1 and DAT are in the same biochemical entity that is defined by Brij 58. Thus, even though they do not represent the classically proposed membrane rafts, the DRMs represent a biochemically defined environment into which DAT must partition to uncover its full function.

Unlike DAT, Flot1 is found in DRMs insoluble in both Brij-58 and Tx-100, suggesting that it is probably dispersed throughout the membrane and may serve to define distinct DRMs with other protein partners or lipids when appropriate. Though the DRM localization of DAT strongly correlates to the behavior displayed by the different Flot1 mouse models, it would be ideal to be able to visualize the clustering of Flot1 and DAT in the presence of AMPH in MEFs derived from the mouse lines through live imaging. This could further be refined by the use of echol [345], a modified cholesterol that can be fluorescently labeled, which could be used to mark cholesterol- rich membrane domains. These experiments would reveal whether the addition of AMPH would increase Flot1 to localize with DAT or vice versa, or if Flot1 and DAT moves together to a particular membrane location.

5.3 The role of Flot1

5.3.1 The DAT-Flot1 interaction

Although our work demonstrates why a DAT-Flot1 interaction might be important for AMPH- induced efflux, why Flot1 is required for scaffolding is unclear. Flot1 has been implicated in different cellular roles, and it is likely that Flot1 works similarly with its different substrates. This raises a number of possibilities regarding how Flot1 functions to provide the environment needed by DAT for AMPH-induced efflux.

Although Flot1 can function to scaffold proteins into cholesterol-rich membranes, whether this scaffolding event is through a direct or indirect mechanism remains a question. Similar to the cell- based data, DAT falls out of striatal DRMs in the Flot1 cKO and iKO mice, suggesting that Flot1 is needed to hold DAT in the DRMs. We find that endogenous DAT and Flot1 co- immunoprecipitate in striatal lysates (not shown), indicating that they are in a larger complex, but whether the interaction is direct or indirect is unclear. Nonetheless, by itself or through an additional interaction partner, Flot1 may ‘bring’ DAT into DRM. To establish whether the interaction is direct, yeast two-hybrid or another in vitro binding assay may give some insights, although it will be difficult to do this in the context of the complete DAT protein. If it is not a direct interaction, then perhaps a yeast three-hybrid should be used to identify an interaction partner. Mass spectrometry can also be used on co-IP samples to narrow down the possibilities.

Rather than a scaffold however, Flot1 may simply define the discrete lipid compositions within the membrane. DAT, via palmitoylation, then ‘finds’ these Flot1-positive regions. Our data indicate that Flot1 does not scaffold all proteins and the loss of Flot1 does not disrupt all DRMs

(Appendix A), and therefore specificity to Flot1 must somehow be achieved. It is possible that the

89 membrane lipids are maintained at an equilibrium that is close to domain formation and only needs a small push, such as the presence of Flot1, to cause a small dynamic change to form the DRMs.

Although currently technically impossible, if Flot1, cholesterol, PIP2 and DAT could all be labeled, we could possibly watch the movement of DAT and a Flot1-dependent structure in the presence of AMPH. We would be able to observe whether PIP2 begins to enrich in the DRM or whether the

Flot1-positive structure is already enriched in PIP2. This experiment would ideally be performed in DAergic neurons derived from the iKO mice so that Flot1 expression can easily be turned on or off by tamoxifen. In the absence of tamoxifen, Flot1 and the lipids would either surround DAT or interact with DAT with AMPH. When tamoxifen is present, there should be no Flot1-lipid complex moving toward DAT.

5.3.2 Oligomerization of Flot1

The flotillins, Flot1 and Flot2, were found to organize into stable tetramers as homo- or hetero- tetramers [119]. It is difficult to conclude whether hetero-oligomerization of Flot1 is necessary.

Although structurally similar, Flot2 does not partition into Brij58-DRMs when Flot1 is absent

[311], suggesting that Flot1 is primarily responsible for DAT partitioning in a Brij58-dependent manner. Nonetheless, the protein stability of Flot1 depends on the presence of Flot2 [119], thereby strongly indicating that the hetero-oligomeric state might be the basal state in which Flot1 is found.

Although our co-IPs suggest that Flot1 and Flot2 are ‘pulled-down’ by DAT (data not shown), given the detergent insolubility of the Flot1 DRMs, it is difficult to be certain of the DAT and

Flot1 and/or Flot2 complex. Although a Flot2 KO mouse might give insight in this regard, it is complex. The concomitant loss of Flot1 makes it difficult to segregate the importance of Flot2

90 alone. Moreover, the loss of Flot2 may be sufficient to provoke developmental changes. This is likely since studies using the Flot2 KO mice suggest a similarly mild phenotype as the Flot1 KO

[346].

5.3.3 Developmental compensation in Flot1 KO mice

The observation that the Flot1 KO loses the phenotype brings into question what the compensatory event might be. Our results are consistent with previous flotillin KOs, which have been shown to be viable [108, 311, 312]. Because the Flot1 iKO was able to reproduce the phenotype seen in the cKO, it appears that the developmentally early, global deletion of Flot1 in the KO led to developmental events masking the loss of Flot1. Using RNAseq we have found that none of the

PHB family members were upregulated to compensate for the loss of Flot1, suggesting that there could be other functionally redundant proteins that have not yet been identified. It is possible that the compensation did not occur at the transcriptional level so perhaps Ribo-Seq would shed light on whether compensation occurred at the protein translation level, although it should be noted that

Western blotting against Flot2 and Cav1 revealed no change in these proteins. Finally, it is also possible that we did not capture the compensation early enough. The mice used for RNAseq were around six weeks old, when the compensation most likely already occurred, and therefore transcriptional changes may no longer be acutely visible. Therefore, it may be productive to perform either RNAseq or Ribo-Seq at different ages, including embryonic and postnatal ages, comparing both WT and KO mice.

The transcriptional and translational changes primarily focus on changes in proteins, but the lipid composition or arrangement within the plasma membrane may have changed in the Flot1 KO.

Given that we have seen no changes of other potential DRM-scaffolding protein levels with

RNAseq, the recovery of the AMPH-induced behavioral phenotype and DAT partitioning in the

Flot1 KO may be due to an alteration in total membrane composition, such that the ability of DAT to associate with 1% Brij58-resistant membranes in the absence of Flot1 is feasible. Defining the lipid composition of the DRMs using lipidomics analyses from the WT, KO and iKO may shed light on which lipids play a role in shaping DAT activity. Because rigid membranes tend to contain lipids with saturated acyl chains, once the DRM lipids have been identified, we can confirm whether these lipids are involved in the DRM localization of DAT, for instance, by using enzymes that alter saturated acyl chains to see whether DAT would effectively fall out of DRMs following that treatment. If there is a difference in lipid composition, perhaps the lipidomic analysis can also be carried over to our understanding of why DAT can partition into Brij-dependent, but not Tx- dependent DRMs [347].

Although the constitutive loss of Flot1 led to a viable animal and aging the animals did not highlight any aberrant behaviors (data not shown), it is unlikely that Flot1 is not essential considering it has been reported to be required for cell-cell adhesion [178], which leads to organ formation, and epiboly formation [182], a developmental milestone for zebrafish. Similar to the use of AMPH, the system may have to be subjected to stress in order to identify any abnormalities.

For instance, some of the RNAseq hits in the KO included endothelial genes implicated in stroke or experimental autoimmune encephalomyelitis (EAE) or blood brain barrier (BBB) formation and function [348]. One way to assess whether Flot1 may play a role in neuroepithelium is to give WT and KO mice strokes and test for BBB leakage. Moreover, we have not examined if there is any rewiring of the brain due to the constitutive loss of Flot1. This is reminiscent of a 5-HT1B receptor

KO mouse [349]. Whereas an antagonist for the receptor normally led to insensitivity to cocaine,

92 the genetic KO unexpectedly had a heightened response to cocaine, suggesting both compensation and re-wiring in the 5-HT1B receptor KO. Hence, it may be worth injecting the WT and KO mice with AMPH and then performing immunohistochemistry with c-Fos to map differences in the neuronal network. Stressing the system may bring out the subtleties masked by compensation.

5.4 The physiologic relevance of AMPH-induced efflux

We were able to use Flot1 to segregate the actions of COC and AMPH. It is clear that AMPH- induced efflux requires DRMs, but does DAT under basal conditions also require DRMs? If DRM localization is a requisite for higher DAT activity, then this may be the case if DA uptake is more demanding under basal conditions, so DAT is driven into DRMs for increased efficiency.

5.4.1 DA uptake

DRMs may facilitate uptake. Cholesterol has been reported to be required for uptake [104, 105], suggesting that perhaps DRMs may play a role in the removal of DA from the perisynpatic space.

This opens the possibility that Flot1 may help facilitate this process. However, the FSCV data shows that there is no significant difference in DA uptake between the cHet and cKO under basal conditions, indicating that uptake is independent of Flot1. One possibility is that under basal conditions, the rate at which uptake is occurring is not intense enough to require DRMs.

This may be tested by placing an excess of DA in the perisynaptic space and monitoring whether

DAT moves into DRMs in response to increased uptake activity. Additionally, DA uptake may

93 require DRMs after phasic firing by DAergic neurons. Tonic activity is defined by a steady action potential and is regarded as “background activity.” On the other hand, phasic activity leads to massive synaptic DA release, which requires removal via reuptake, implicating a greater need for

DAT function [350]. Because DA plays key roles in reinforcement learning, motivation, motor output, and reward, there are many opportunities for short bursts of activity [351-356]. Therefore, at any given time, there is a high chance that phasic firing is occurring. Consequently, the need for

DA reuptake would also be high. When this occurs, Flot1 may scaffold DAT into DRMs to expedite DA uptake. Behavioral studies that examine motivation and reward might evoke differential responses in the Flot1 cKO or iKO with respect to control mice, such that greater motivation or increased reinforcement might be necessary. Moreover, once mice are conditioned to expect a reward, phasic activity can be measured using voltammetry and we can biochemically examine their brains to see whether more DAT is present in DRMs, compared to non-conditioned control mice [357, 358].

5.4.2 DAergic consequences of drug abuse: sensitization and neurotoxicity

On the other side of the spectrum, Flot1 may also contribute to a permanent change in DAT conformation, leading to sensitization. DAT is a target of psychostimulants, in which repeated abuse can lead to sensitization, a heightened response to a drug as opposed to drug tolerance, which is characterized by a diminished response to a drug. When a drug is first introduced to a system, the initial psychostimulant exposure causes phasic DAergic firing leading to increased DA release in the synaptic cleft. At the same time, uptake inhibition leads to the diffusion of DA out of the synaptic cleft, decrease in phasic DA release and increase in tonic stimulation of autoreceptors,

94 which are responsible for curtailing further release [359, 360]. On the other hand, repeated use of psychostimulants alters DAergic homeostasis because once the psychostimulant inhibition has been released, DAT will try to restore the system back to steady-state levels, except the system has been compromised and it never recovers to baseline conditions. This ultimately leads to non- reversible changes to the regulation of the DA system and drug responsiveness. In the case of

AMPH, repeated administration caused enhanced DA release upon re-exposure to AMPH in mice

[361-365]. Perhaps, repeated AMPH administration causes DAT to stay in these Flot1-dependent

DRMs, where an AMPH-friendly conformation of DAT is readily available for the next hit of

AMPH. DAT in this case may not even be as functional for uptake and the non-DRM DAT are taxed with that function. Moreover, it is possible that AMPH abuse leads to increasing number of

DAT localizing to these DRMs, which also decreases the number responsible for uptake, altering the homeostatic DA levels. It would therefore be interesting to study sensitization in the conditional mice. Furthermore, it would be interesting to perform biochemical studies to determine whether permanently changes in the conformation or distribution of DAT is also observed.

The use of AMPH highlights whether the loss of Flot1 would affect neurotoxicity. High doses of

AMPH can damage DAergic neurons by degeneration of DAergic terminals as well as reduced transporter and receptor function [366]. Neurotoxicity may be a result of hyperpyrexia (high body temperature), autoxidation of DA, and the formation of reactive oxygen species [366-369]. Given that the Flot1 cKO mice do not efflux DA effectively, it is possible that cytosolic DA levels rise more precipitously causing more toxicity than either the control mice. One way to test this is to chronically inject these mice and observe when the mice die as well as examine their neuroanatomy, including neuronal death, at various later ages.

5.4.3 Flot1 as a therapeutic target

The euphorigenic effects and addictive properties of AMPH that can lead to toxic fatalities are well-known, but the molecular mechanisms of AMPH action on DAT function remain unclear.

This thesis shows how Flot1 can regulate where DAT is located within the cell and promotes the

AMPH-induced synaptic accumulation of dopamine, elucidating how AMPH affects DAT, which is critical to understanding AMPH action and how its detrimental effects may be therapeutically blocked. However, given its many roles and implications, Flot1 may be especially difficult to target. In general, there has not been a mutation in Flot1 that drives disease, rather, it is a change in protein expression levels. This has been linked to type 2 diabetes, prion disease, EAE, and

Alzheimer’s disease [112, 370-373]. Although Flot1 was upregulated in the cerebral cortex, where there was senile plaque formation, follow-up in vitro studies show a relatively weak role for Flot1.

Depending on the assay, changes in soluble Aβ levels in MEFs from a double KO of the flotillins were variable [312]. Studies on the role of the flotillins in disease is still incomplete so any therapeutic approaches aimed at Flot1 to curtail drug addiction should be cautionary until it is fully understood what it is doing and how it is functioning.

5.5 Methodological considerations regarding mouse models

Our use of three mouse models show that a constitutive KO, which is heavily used to correlate observations arrived via cell biology and biochemistry, is not necessarily always ideal. A constitutive KO can lead to lethality or compensation, both of which demonstrate the importance of a protein. Moreover, knockdown in an in vitro system does not equate the complete loss of a protein in vivo as the former may display a phenotype whereas the latter may not [346]. We were

96 able to observe a robust phenotype because we focused specifically on AMPH-induced efflux, which drove us to first make a conditional KO using DATiresCre. The phenotype was reproduced in the inducible knockout showing that both spatial and temporal considerations with regard to genetic deletions must be taken into account. To that extent, perhaps we could also have created an inducible DATiresCre to avoid a whole-body deletion and to see whether postponed Flot1 excision would change the phenotype. If the heterozygous inducible DATiresCre mice also do not respond to

AMPH, one explanation for the difference in phenotype we see for the cHet and iHet may be the time when Flot1 is deleted. Our study highlights the importance of choosing a proper in vivo model to fully utilize the power of mouse genetics.

5.6 Conclusions

Although in vivo studies of Flot1 has so far been limited, it is irrefutable that Flot1 plays an essential role in establishing the correct membrane environment for protein function and reveals that this can segregate functions for the same protein. The behavioral impact of Flot1 deletion, especially in the constitutive KO, suggests that this protein that is most often used as a membrane raft marker may be important from the beginning of a living organism to the maintenance of its everyday functions. According to our data, it could very well have evolved in the mammalian nervous system to maintain homeostatic DAergic signaling.

In summary, our work has expanded on our understanding of Flot1’s role in the adult brain and utilized it as a genetic tool to confer physiologic relevance to how membranes influence protein function. Further studies may elucidate not only the exact function of Flot1 but answer why it specifically regulates AMPH-induced efflux, but not cocaine activity. Flot1’s potential to

97 segregate the two psychostimulants could push the DA field, which has stalled on the complexities of the mechanisms of AMPH-induced efflux, forward. Finally, the compensation displayed by the

KO also serves as a commencement for Flot1’s role in the development of the mammalian system as well as the discovery of new genes that are functionally redundant to Flot1.

Chapter 6

Materials and Methods

6.1 Antibodies

The following antibodies were used for immunoprecipitation and/or western blot: rat anti-DAT

(1:500; EMD Millipore MAB369), mouse anti-Flot1 (1:250; BD #610821), rabbit anti-Flot1

(1:250; Abcam ab41927), rabbit anti-Vinculin (1:5000; Invitrogen), rabbit anti-tyrosine hydroxylase (1:1000; Calbiochem #657012), rabbit anti-CaMKII (1:1000; Cell Signaling

Technology #3362), rabbit anti-phosphoDAT (pSer7) (1:100; generously provided by Dr.

Jonathan Javitch), rabbit anti-phosphoDAT (pThr53) (1:1000; PhosphoSolutions p435-53), mouse anti-PIP2 (1:500; Enzo ADI-915-052-020), mouse anti-γ-tubulin (1:10,000; Abcam ab11316). Rat anti-dopamine transporter (1:1000; EMD Millipore MAB369) and rabbit anti-tyrosine hydroxylase (1:2000; Calbiochem #657012) were used for immunohistochemistry. Rat anti-DAT

(1:250; EMD Millipore MAB369) and mouse anti-Flot1 (1:100; BD #610821) were used for immunofluorescence.

6.2 Plasmids pCI-Hygro vectors expressing hDAT(WT) and hDAT(KA) (Lys3 and Lys5 both mutated to Ala)

(generously supplied by Dr. Aurelio Galli) were made as previously described [306].

6.3 Transfections siRNAs (first duplex sequence: 5’-CUU CAG UUC GUA AUC UCU CUG UGC CUU-3’ and

5’-GGC ACA GAG AGA UUA CGA ACU GAA G-3’; second duplex sequence: UUA UAG

AUC UCC UCC ACA GUC AUG UGC-3’ and 5’-ACA UGA CUG UGG AGG AGA UCU AUA

A-3’) (Integrated DNA Technology) were used as a mix against Flot1 using Lipofectamine 2000

(Life Technologies) according to manufacturer’s instructions. hDAT(WT) and hDAT(KA) were transfected using Lipofectamine 2000 (Life Technologies) according to manufacturer’s instructions.

6.4 Cell culture

HEK293T cell cultures were maintained in high-glucose Dulbecco's modified eagle medium

(DMEM; Life Technologies) supplemented with 10% fetal bovine serum (FBS; Life

Technologies) at 37°C and in a 5% CO2-containing atmosphere.

EM4-YFP-DAT were created and maintained as previously described [318, 319]. EM4 cells are human embryonic kidney 293 cells modified to increase their adherence to tissue culture plastic

[374].

6.5 Animals

100

All experiments were reviewed and approved by the Columbia University Medical Center’s

Institutional Animal Care and Use Committee (IACUC). Mice were bred and housed in facilities at the William Black Medical Research Building. Same-sex animals of mixed genotypes are housed four to five per cage in a humidity- and temperature-controlled room, and mice are provided food and water ad libitum. Mice are maintained on a 12 h light/dark cycle (lights on at

7:00 A.M.).

6.6 Mouse lines

6.6.1 Conditional Flotillin-1 alleles (Flot1 flox/flox)

Flot1flox/+ (strain F1 CBAxC57Bl6/J) mice were generated via homologous recombination at the

University of Connecticut Health Center’s Gene Targeting and Transgenic Facility. The targeting vector was designed with LoxP sites flanking Flot1 at exons 3 – 8. C57Bl/6 and CBA F1 hybrid

ES cell lines were used for targeting. Germline chimeras were crossed with ROSA-26Flpe mice to remove the PGKneo cassette used for in vitro selection. Germline-positive F1 mice were backcrossed with the chimeric mice. Pups from this backcross were genotyped in order to identify loss-of-wildtype allele and confirm proper targeting. The homozygous Flot1 flox/flox line was maintained by consistently breeding littermates every 5-6 months.

6.6.2 DATiresCre (DIC)

DATiresCre (DIC, strain Bl6/J) mice were acquired from the Jackson Laboratory (Bar Harbor,

Maine) [310]. This line, in which DAT mRNA has an internal ribosome entry site (IRES) Cre knock-in at the 3’ UTR, restricts Cre-mediated recombination events to DAergic neurons. DIC

101 mice were maintained as heterozygotes on a C57Bl/6 background strain by breeding with wild type (WT) C57Bl/6 mice every 5-6 months.

6.6.3 Flotillin-1 conditional knockout (Flot1 cKO)

The Flot1 cKO mice used in this study were derived from mating Flot1flox/flox mice with mice heterozygous for the DIC allele to create the F1 generation containing Flot1fl/+::DIC. For strain purposes, these Flot1fl/+::DIC mice were then crossed back with Flot1flox/flox mice to generate the following littermates used in the experiments: Flot1fl/fl, Flot1(DAT) cHet, and Flot1(DAT) cKO.

6.6.4 HprtCre/+

HprtCre/+ mice (strain: Bl6/J) were obtained from the Jackson Laboratory (JAX no. 004302, Bar

Harbor, Maine) [375], and backcrossed 10+ generations to create a Bl6/J line. The Hprt locus is located on the X chromosome and Cre expression results in 100% deletion of a conditional allele in oocytes. HprtCre/+ mice were maintained as heterozygotes on a C57Bl/6 background strain by breeding with WT C57Bl/6 mice every 5-6 months.

6.6.5 Flotillin-1 constitutive knockout (Flot1 KO)

Flot1flox/flox males were crossed with heterozygous HprtCre/+ females to create Flot1Δ/+ mice.

Flot1Δ/+ mice that did not carry HprtCre/+ were crossed together to generate the following experimental littermates: WT, Het, and KO. 102

6.6.6 Tamoxifen-inducible Cre (ActinCreERTM/+)

ActinCreERTM/+ mice (strain: Bl6/J) were obtained from the Jackson Laboratory (JAX no. 004682,

Bar Harbor, Maine) [376]. This line expresses a tamoxifen-inducible Cre recombinase protein driven by a chicken beta actin promotor. The Cre recombinase is fused to a synthetic estrogen receptor, in which the ligand binding domain is modified, and is restricted to the cytoplasm. Upon tamoxifen injection, the estrogen receptor translocates into the nucleus, where excision of LoxP- flanked sequences occurs. ActinCreERTM/+ mice were maintained as heterozygotes on a C57Bl/6 background strain by breeding with WT C57Bl/6 mice every 3-4 months.

6.6.7 Flotillin-1 inducible knockout (Flot1 iKO)

Flot1fl/fl mice were crossed with ActinCreERTM/+ mice to create mice heterozygous for the Flot1 floxed allele and containing the ActinCreERTM/+ allele (Flot1fl/+:: ActinCreERTM/+). These Flot1fl/+::

ActinCreERTM/+ mice were then crossed back with the original Flot1fl/fl mice to generate the following experimental littermates: Flot1fl/fl, iHet, and iKO.

6.7 PCR Genotyping

6.7.1 DNA isolation

21-day-old mice were ear punched for identification and genotyping. Mice who have been retired from experiments were also re-genotyped by tail-clipping. The ear punches or tails were incubated at 55°C overnight in 0.5 mg/ml proteinase K made into 500 µl lysis buffer (50 mM Tris-HCl pH

103

8.0, 100 mM EDTA pH 8.0, 100 mM NaCl, 1% SDS). DNA was extracted by adding 700 µl phenol/chloroform/isoamyl alochol (Fisher Scientific) to the digested sample, which was then centrifuged at 14,000 rpm for 5 min. To precipitate the DNA, 400 µl of the aqueous phase was carefully removed, mixed with 1 ml 100% ethanol by inverting tube, and spun at 14,000 rpm for

5 min. Once the supernatant has been decanted, the pellet was washed with 800 µl 80% ethanol and spun at 14,000 rpm for 5 min. The supernatant was removed and the DNA pellet was air-dried for 5 min at room temperature. Samples are resuspended in 50 µl (ear punch) or 200 µl (tail) of 10 mM Tris-HCl pH 8.0.

6.7.2 PCR

Each PCR reaction is 20 µl, consisting of 1 µl genomic DNA, 10 µM primer, 8 µl 5 PRIME

HotMasterMix (5 PRIME GmbH), and water up to volume. The HotMasterMix contains Taq DNA polymerase, MgCl2, and dNTPs. Primers and PCR conditions are described below:

104

Gene Forward Primer (5' – 3') Reverse Primer (5’ – 3') Temp. (°C) Time # Cycles

CTT CCC AGC CCT TAC GGG GTG GGG AGA ATT Flot1 LoxP 1 94 3 min 1 GTT CT CTA TG

94 30 s

53 1 min 35

72 1 min

72 2 min 1

AGC GGA TCC CAT TAC TCC CAT GAG GGG ATT Flot1 LoxP 2 (Frt) 94 3 min 1 AGA TG ACA AG

94 30 s

57.5 1 min 35

72 1 min

72 2 min 1

TGG CTG TTG GTG TAA GGA CAG GGA CAT GGT DAT (wild type) 94 2 min 1 AGT GG TGA CT

CCA AAA GAC GGC AAT DATiresCre 94 30 s ATG GT 35 62 30 s

72 1 min

72 4 min 1

CAC AGT AGC TCT TCA TTT CTA TAG GAC TGA Hprt (wild type) 94 2 min 1 GTC TGA TAA AA AAG ACT TGC TC

GCG GTC TGG CAG TAA GTG AAA CAG CAT TGC HprtCre 94 15 s AAA CTA TC TGT CAC TT 10* *-1°C per 65* 30 s cycle 72 45 s

94 15 s 25

105

55 30 s

72 45 s

72 7 min 1

GCG GTC TGG CAG TAA GTG AAA CAG CAT TGC ActinCreERTM 94 3 min 1 AAA CTA TC TGT CAC TT

CTA GGC CAC AGA ATT GTA GGT GGA AATTCT Internal control 94 30 s GAA AGA TCT AGC ATC ATC C 35 51.7 1 min

72 1 min

72 5 min 1

Flot/Frt CTT CCC AGC CCT TAC TCC CAT GAG GGG ATT 94 4 min 1 GTT CT ACA AG (Flot1 excision)

94 30 s

57 30 s 40

72 1:30

72 10 min 1

Table 6.1 Primers and conditions for PCR.

6.8 Drug Preparation and Administration

6.8.1 Amphetamine

Amphetamine (Sigma) was prepared in 0.9% saline at either 2.5 mg/kg or 5 mg/kg body weight

(BW) for the open field behavior test. 5 mg/kg AMPH was used for animals prepared for biochemical assays.

106

6.8.2 Cocaine

Cocaine (Sigma) was prepared in 0.9% saline at 10 mg/kg BW for the open field behavior test and at 20 mg/kg BW for sucrose density gradients on mouse brains.

6.8.3 Tamoxifen

For temporal deletion of Flot1, experimental littermates for the Flot1 iKO were injected with 1.5 mg/ 10 g BW tamoxifen (Sigma) at 4 weeks of age. The tamoxifen solution was prepared into corn oil and 10% ethanol, which was heated at 37°C for 4 hours with vortexing every 30 min.

Tamoxifen was administered to the iKO mice for five consecutive days. These mice were tested in open field behavior one month following their last injection when they are nine weeks old. All drugs were administered intraperitoneally (i.p.).

6.9 Mouse Behavioral Testing

6.9.1 Open Field

All mice intended for behavioral testing is relocated and housed in our Satellite Animal Facility, where the open field equipment is located, two weeks prior to testing so that they will acclimatize to a new environment. Mice continue to be maintained on a 12 h light/dark cycle, for which 7:00 to 19:00 is the light cycle.

Testing begins at 12-16 weeks of age for the Flot1 cKO, 8-12 weeks for the Flot1 KO mice, and 9 weeks for the Flot1 iKO mice. Open field was conducted between 9:00 A.M. and 7:00 P.M. during 107 the light cycle. All mice were habituated to the open field prior to drug testing, which comprised of exposure to the open field chambers (43.2 cm x 43.2 cm x 30.5 cm) (Med Associates) for at least one hour per day for two to three consecutive days. All mice were assigned a chamber and were placed in the respective chambers for every subsequent run. For testing, mice were exposed to the open field for 1 hour, then given either saline, AMPH or COC (i.p.) then monitored for an additional 60 minutes. Each animal’s pathlength was collected and the data was binned into 5 min intervals.

6.10 Electrophysiology

6.10.1 Fast scan cyclic voltammetry (FSCV)

FSCV was performed as previously described [377]. Briefly, coronal brain slices with thickness of 250-300 µm containing cortex and striatum were prepared from 8-12 week old mice using a vibratome (Leica) and ice-cold cutting solution consisting of (in mM): glucose (10), NaCl (125.2),

NaHCO3 (26), KCl (2.5), MgSO4 (3.7), NaH2PO4·6H2O (0.3), KH2PO4·6H2O (0.3). Slices were allowed to recover in the solution for 30 minutes at 34oC, and then transferred to recording ACSF

(10 mM glucose, 125.2 mM NaCl, 26 mM NaHCO3, 2.5 mM KCl, 1.3 mM MgSO4, 2.4 mM CaCl2,

0.3 mM NaH2PO4·6H2O, 0.3 KH2PO4·6H2O). The temperature of the recording chamber was maintained at 32°C (± 2°C). Electrochemical recordings were performed with freshly cut disk carbon fiber electrodes 5 µm in diameter. The electrode was inserted ∼50 µm into the dorsolateral striatum. For cyclic voltammetry, a triangular voltage wave (-400 to +900 mV at 280 V/s vs

Ag/AgCl) was applied to the electrode every 100 ms. Currents were evoked by local electrical stimulation with a tungsten electrode (World Precision Instruments) and stimuli (100-400 µA, 1 108 ms duration) were delivered every 2 min by an ISO-flex stimulus isolator (AMPI) and Master-8 pulse generator (AMPI). Evoked currents were recorded with an Axopatch 200B amplifier with a low-pass Bessel filter setting at 10 kHz, digitized at 25 kHz (ITC-18 board; InstruTech). Triangular wave generation and data acquisition were controlled by a personal computer running a locally written (Dr. E. Mosharov, Columbia University, New York, NY) Igor Pro program

(WaveMetrics). For amphetamine (AMPH) experiments, 10 µM AMPH was perfused into the bath for 20 min. The changes in evoked dopamine (DA) release were normalized by the average value of the peaks before AMPH perfusion. AMPH-induced DA efflux was detected 8-15 min after

AMPH perfusion and the maximal peak of DA efflux was obtained in the presence of AMPH for the analysis. Carbon fiber electrodes were calibrated with 1 µM DA before and after recording, and voltammetric current unit was converted into molarity by calibration value.

6.10.2 High speed chronoamperometry (HSCA)

Experiments were performed as previously described [378]. Briefly, striatal slices (300 µm) were prepared with a vibratome (Leica) in an ice cold oxygenated (95% O2/5% CO2) sucrose cutting solution consisting of 210 mM sucrose, 20 mM NaCl, 2.5 mM KCl, 1 mM MgCl2, 1.2 mM

NaH2PO4, 10 mM glucose, 26 mM NaHCO3. Slices were then transferred to oxygenated ACSF at

28 °C for a minimum of 1 h. The ACSF consisted of 125 mM NaCl, 2.5 mM KCl, 1 mM MgCl2,

2 mM CaCl2, 1.2 mM NaH2PO4, 10 mM glucose, 26 mM NaHCO3, and 0.25 mM ascorbic acid.

DA concentration was measured by chronoamperometry in slices as previously described [379,

380]. Briefly, a carbon fiber electrode (100 µm length × 30 µm O.D.) coated with nafion for DA selectivity was positioned at a depth of 75-100 µm. The voltage was stepped from 0 mV to 550

109 mV for 100 ms and then back to 0 mV and the charging current of the microelectrodes was allowed to decay for 20 ms before the signals were integrated. Data were collected at a frequency of 1Hz with an Axopatch 200B amplifier. The integrated charge was converted to DA concentration based on in vitro calibration with DA. The oxidative signal measured in the slices is attributable to DA as the reduction/oxidation charge ratio is in the range 0.6-1.0.

6.11 Immunohistochemistry

Staining was performed as previously described [381, 382]. Briefly, deeply anesthetized mice were transcardially perfused with 0.9% saline for 2 min followed by 4% paraformaldehyde (PFA) for 2 min. Brains were removed and post-fixed in 4% PFA for an hour before transfer to 30% sucrose in 1X phosphate buffer (P.B.) overnight. Post-fixed brains were fresh frozen in powdered dry ice, embedded with O.C.T. Compound (VWR), and stored at -80°C until use.

Immunohistochemistry (IHC) was performed on fixed sections. 30 µm free-floating coronal brain sections were cut into 1X P.B. containing 0.02% sodium azide. Sections were incubated in 10 mM sodium citrate, pH 9.0 for 30 min at 80°C for antigen retrieval. They were washed twice with 1X

PBS containing 0.02% Triton X-100 (Tx-100), 10 min/wash. The endogenous peroxidases on the sections were blocked by incubation in 1% hydrogen peroxide (H2O2) diluted into 1X PBS containing 0.02% Tx-100 for 15 min, then washed three times with 1X PBS containing 0.02% Tx-

100. After the washes, sections were blocked for 1 hr with 0.4% Tx-100 and incubated with antibodies overnight at room temperature.

On the following day, after three 10-min washes in 1X PBS, sections are incubated in biotinylated

110 secondaries made into 0.4% Tx-100 for 1 hr. For sections stained with DAT, the secondary was anti-rat (1:2000) and for sections stained with TH, the secondary was anti-rabbit (1:2000). After three 10-min washes in 1X PBS, the signal was amplified using the standard VECTASTAIN Elite

ABC HRP Kit (Vector Laboratories). The ABC complex was diluted into 1X PBS. Sections were placed in the diluted ABC solution for 1 hr at room temperature, followed by three 10-min washes in 1X PBS. Signal was detected using diaminobenzoate, or DAB (Sigma); 10 mg DAB was dissolved into every 20 ml 1X P.B. and activated with 0.00375% H2O2. Sections were washed three times in 1X P.B. then mounted on glass slides. The mounted sections were air-dried and dehydrated by dipping the slides in the order of the following solutions for 5 min each: 70% ethanol, 95% ethanol, 100% ethanol, and three times in xylene. The slides were coverslipped using

Permount mounting media (Fisher Scientific).

6.12 Quantification of DA levels

Quantification of DA and DOPAC levels was performed at the Neurochemistry Core at Vanderbilt

University as described [383]. Dissected striata were homogenized in 100–750 ul of 0.1M TCA, containing 10-2 M sodium acetate, 10-4 M EDTA, 5ng/ml isoproterenol (as internal standard) and

10.5 % methanol (pH 3.8). Samples were spun in a microcentrifuge at 10,000g for 20 minutes.

Supernatant were then analyzed for biogenic monoamines and/or amino acids using high pressure liquid chromatography with amperometric detection. Pellets were used to quantify protein. HPLC control and data acquisition are managed by Millennium 32 software.

111

6.13 Sucrose density gradients

6.13.1 Striatal tissue preparation

Mice striatum were prepared in 1% Surfact-Amps (Thermo Fisher), also known as 1% Brij 58 with protease inhibitors (Pierce). Samples were homogenized and run through a 19-gauge needle 15 times followed by 25-gauge needle 15 times. The lysate was incubated on ice for 30 min, and spun at 850g for 10 min at 4ºC. 1% Brij 58 with protease inhibitors was added to 100 µg of the sup to make a total of 500 µl, which was mixed with an equivolume of 80% sucrose (wt/vol) in 150 mM

NaCl. The mixture was transferred to an ultracentrifuge tube, which was then layered on with 1 ml of each of the following: 30% sucrose (wt/vol), 15% sucrose (wt/vol), and 5% sucrose (wt/vol).

The tube was placed in a chilled SW60 Ti swinging-bucket rotor. Centrifugation was carried out in a Beckman ultracentrifuge at 43,200 rpm for 20 hrs at 4ºC. Ten fractions of equal volumes were collected from top to bottom of the gradient.

For the gradients in which mice were injected with drugs, mice were sacrificed 30 min after 5 mg/kg AMPH injection and 10 min after 20 mg/kg injection; both times are when mice have reached maximal hyperlocomotor behavior. The striata were processed as described.

To see whether cholesterol depletion would affect the partitioning of DAT in detergent-resistant membranes, 10 mM methyl-beta-cyclodextrin (MβC) was incubated with 100 µg striatal lysate for

40 min before being processed as described.

6.13.2 Striatal synaptosome preparation

112 cKO mice with Flot1fl/fl and Flot1(DAT) cKO genotypes were sacrificed and their brains were removed into ice-cold 1X PBS and the striatum was dissected. The striatum was homogenized in

Homogenization Buffer (H.B.) consisting of 0.32 M sucrose, 2 mM EDTA, 20 mM HEPES pH

7.3 with protease inhibitors using a glass dounce tissue grinder. Samples were spun at 1000 rpm for 10 min at 4ºC. The sup was pipetted into a new tube, after which it was spun at 14,000 rpm for

20 minutes at 4ºC. The sup was aspirated and the pellet resuspended in 1X PBS. The resuspended pellet was subjected to sucrose density gradients.

6.13.3 Cell culture

HEK293 cells were transfected with hDAT(WT) and hDAT(KA) were collected 48 hrs later in

RIPA (50 mM Tris pH 7.5, 150 mM NaCl, 0.25% sodium deoxycholate, 1 mM EDTA, 1% NP-

40, 1% Tx-100) with protease inhibitors. The lysate was incubated on ice for 30 min then spun at

1000 rpm for 10 min at 4ºC. The sup was used for sucrose density gradients.

6.13.4 Western blots

The gradient fractions are run with 5 µg protein as lysate loading control. 42 µl fraction (maximum volume) was used. Western blot was run as described [384]. Band intensities were quantified using

ImageJ (NIH) and normalized to the lysate loading control.

6.14 Limited proteolysis

113

All steps were performed on ice. Fractions 4 and 5 from the sucrose gradients described above were combined separately from fractions 9 and 10 so that the combined fractions contain the same concentration of the protein-of-interest. Protease solutions were always made fresh on the day of experiment; trypsin and papain were made into water and 2-Nitro-5-thiocyanatobenzoic acid

(NTCB) was made into 100% ethanol. An aliquot was taken from the combined fraction and vehicle was added to the aliquot for the 0 min time point. Protease was added to the remaining combined fractions, mixed thoroughly, and quickly aliquoted into separate tubes. Protease inhibitor cocktail (for trypsin and papain) or β-mercaptoethanol (for NTCB) were added at the specified time point. SDS sample buffer was added and the samples were boiled then run on a western blot.

To ensure that the sucrose concentration of the fraction does not differentially affect protease cleavage capability, a gradient was run without any protein and 50 µg/ml BSA was spiked into the same gradient fractions mentioned above. Limited proteolysis was performed as described above and results were visualized by coomassie staining.

6.15 Immunoprecipitation

HEK293 cells transfected with the hDAT constructs were lysed with RIPA and protease inhibitors.

100 µg protein was incubated with 1µL anti-PIP2 antibody (ENZO) for 30 min at 4°C with gentle shaking, then incubated in Protein G microbeads (Miltenyi) overnight at 4°C with shaking and processed the following day as described [105].

6.15.1 Phosphorylation

114

EM4 cells were transfected with a mix of siFlot1. After 72 hours, cells were washed twice with

PBS and once with pre-warmed Kreb’s buffer (130 mM NaCl, 1.3 mM KCl, 1.2 mM MgSO4, 1.2 mM KH2PO4, 2.2 mM CaCl2, 2 mM NaH2PO4, 10 mM HEPES pH 7.4, 10 mM glucose, 0.1 %

BSA, 2 mM NaHCO3). They were treated with 1 µM PMA or vehicle (DMSO) in Kreb’s buffer with phosphatase inhibitors 1 mM sodium vanadate and 10 mM sodium fluoride for 30 min at

37ºC and collected using 1% Brij 58 with protease and phosphatase inhibitor cocktails. For pSer7 detection on western blot, an immunoprecipitation must first be carried out. Cell lysate was incubated overnight at 4ºC with anti-GFP conjugated to magnetic beads (Miltenyi) and IP was performed as previously described [105].

6.16 Immunofluorescence

Cells were fixed with methanol for 10 min at room temperature. Cells were washed two times with

PBS, incubated in 0.1% Tx-100 for 10 min, washed three times with PBS and blocked with 3%

BSA in PBS. Cells were incubated in primary diluted in 0.01% Tx-100 overnight at 4°C. They were washed three times with PBS and incubated with Alexa Fluor secondary antibodies (Thermo

Fisher) made into 1 µg/ml Hoechst solution (Life Technologies) for two hours at room temperature. Images were acquired using a Leica TCS SP2 confocal microscope at 63x magnification and the accompanying software package.

6.17 Dissection and immunoblotting of substantia nigra

115

Mouse brains were placed in a pre-chilled brain matrix (Ted Pella), which allowed slices of 1 mm to be generated. A slice was taken from the midbrain, placed in cold PBS and the substantia nigra dissected. Tissue was prepared in 1% Brij 58 with protease inhibitors. Protein concentration was determined using the DC Protein Assay (Bio-rad Laboratories) and equal amounts of protein were prepared for western blot as previously described [384].

6.18 Statistical analyses and figure creation

Statistical analysis was performed using Statview 5.0 (SAS Institute, Cary, NC). Mean differences of normally distributed data were tested using analysis of variance (ANOVA) as described in the figure legends. Should differences with greater than 90% be detected, post hoc analyses were performed using the Fisher’s protected least significant difference test. Non-normally distributed data was assessed using the nonparametric test, Mann-Whitney U. Complete F-statistics with p- values are presented in the figure legends, as well as n-values. Adobe Creative Suite 5 (Photoshop and Illustrator) were used to create all figures. ImageJ (NIH) was used for quantification.

116

References

1. Hooke, R. and Lessing J. Rosenwald Collection (Library of Congress), Micrographia: or, Some physiological descriptions of minute bodies made by magnifying glasses. With observations and inquiries thereupon. 1665, London,: Printed by J. Martyn and J. Allestry. 18 p. l., 246, 10 p. 2. Lombard, J., Once upon a Time the Cell Membranes: 175 Years of Cell Boundary Research. Biology Direct, 2014. 9: p. 32. 3. Overton, E., The probable origin and physiological significance of cellular osmotic properties. Papers on biological membrane structure., ed. P.R. Branton D. 1968, Boston: Little, Brown and Company. 4. Chambers, R., A Micro Injection Study on the Permeability of the Starfish Egg. J Gen Physiol, 1922. 5(2): p. 189-93. 5. Pethig, R. and I. Schmueser, Marking 100 years since Rudolf Höber’s discovery of the insulating envelope surrounding cells and of the beta-dispersion exhibited by tissue. Journal of Electrical Bioimpedance, 2012. 3(1). 6. Gorter, E. and F. Grendel, On Bimolecular Layers of Lipoids on the Chromocytes of the Blood. J Exp Med, 1925. 41(4): p. 439-43. 7. Cole, K.S., Surface forces of the arbacia egg. Journal of Cellular and Comparative Physiology, 1932. 1(1): p. 1-9. 8. Harvey, E.N. and H. Shapiro, The interfacial tension between oil and protoplasm within living cells. Journal of Cellular and Comparative Physiology, 1934. 5(2): p. 255-267. 9. Danielli, J.F. and E.N. Harvey, The tension at the surface of mackerel egg oil, with remarks on the nature of the cell surface. Journal of Cellular and Comparative Physiology, 1935. 5(4): p. 483-494. 10. Danielli, J.F. and H. Davson, A contribution to the theory of permeability of thin films. Journal of Cellular and Comparative Physiology, 1935. 5(4): p. 495-508. 11. Danielli, J.F., The Relations between Surface pH, Ion Concentrations and Interfacial Tension. Proceedings of the Royal Society B: Biological Sciences, 1937. 122(827): p. 155- 174. 12. Harkins, W.D., The physical chemistry of surface films. 1952, New York,: Reinhold. xvi, 413 p. 13. Macallum, A.B., On the distribution of potassium in animal and vegetable cells. J Physiol, 1905. 32(2): p. 95-198 3. 14. Robinson, J.D., Moving Questions: A History of Membrane Transport and Bioenergetics. 1997, New York: Springer-Verlag. 15. Fenn, W.O., Electrolytes in muscle. Physiol Rev., 1936. 16: p. 450–487. 16. Dean, R., Theories of electrolyte equilibrium in muscle. Biol Symp., 1941. 3: p. 331–348. 17. Danowski, T.S., The transfer of potassium across the human blood cell membrane. J Biol Chem., 1941. 139(2): p. 693-705. 18. Harris, E., The reversible nature of the potassium loss from erythrocytes during storage of blood at 2-5°C. Biol Bull., 1940. 79: p. 373. 19. Fenn, W., The role of potassium in physiological processes. . Physiol Rev., 1940. 20: p. 377–415. 20. Hodgkin, A.L. and A.F. Huxley, A quantitative description of membrane current and its application to conduction and excitation in nerve. J Physiol, 1952. 117(4): p. 500-44.

117

21. Skou, J.C., Enzymatic Basis for Active Transport of Na+ and K+ across Cell Membrane. Physiol Rev, 1965. 45: p. 596-617. 22. Mueller, P., et al., Reconstitution of cell membrane structure in vitro and its transformation into an excitable system. Nature, 1962. 194: p. 979-80. 23. Robertson, J.D., The molecular structure and contact relationships of cell membranes. Prog Biophys Mol Biol, 1960. 10: p. 343-418. 24. Moor, H. and K. Muhlethaler, Fine Structure in Frozen-Etched Yeast Cells. J Cell Biol, 1963. 17(3): p. 609-28. 25. Danielli, J.F., The Thickness of the Wall of the Red Blood Corpuscle. J Gen Physiol, 1935. 19(1): p. 19-22. 26. Kopac, M.J. and R. Chambers, The coalescence of living cells with oil drops. II. Arbacia eggs immersed in acid or alkaline calcium solutions. Journal of Cellular and Comparative Physiology, 1937. 9(3): p. 345-361. 27. Loeb, J., The dynamics of living matter. . The general physical constitution of living matter., ed. J. Loeb. 1906, New York: The Columbia University Press. 28. Frye, L.D. and M. Edidin, The rapid intermixing of cell surface antigens after formation of mouse-human heterokaryons. J Cell Sci, 1970. 7(2): p. 319-35. 29. Singer, S.J. and G.L. Nicolson, The fluid mosaic model of the structure of cell membranes. Science, 1972. 175(4023): p. 720-31. 30. Engelman, D.M., Membranes are more mosaic than fluid. Nature, 2005. 438(7068): p. 578- 80. 31. Andersen, O.S. and R.E. Koeppe, 2nd, Bilayer thickness and membrane protein function: an energetic perspective. Annu Rev Biophys Biomol Struct, 2007. 36: p. 107-30. 32. Jensen, M.O. and O.G. Mouritsen, Lipids do influence protein function-the hydrophobic matching hypothesis revisited. Biochim Biophys Acta, 2004. 1666(1-2): p. 205-26. 33. Brown, M.F., Curvature forces in membrane lipid-protein interactions. Biochemistry, 2012. 51(49): p. 9782-95. 34. Turner, M.S. and P. Sens, Inclusions on fluid membranes anchored to elastic media. Biophys J, 1999. 76(1 Pt 1): p. 564-72. 35. Gil, T., et al., Theoretical analysis of protein organization in lipid membranes. Biochim Biophys Acta, 1998. 1376(3): p. 245-66. 36. Muller, M.M., M. Deserno, and J. Guven, Interface-mediated interactions between particles: a geometrical approach. Phys Rev E Stat Nonlin Soft Matter Phys, 2005. 72(6 Pt 1): p. 061407. 37. Reynwar, B.J., et al., Aggregation and vesiculation of membrane proteins by curvature- mediated interactions. Nature, 2007. 447(7143): p. 461-4. 38. Phillips, R., et al., Emerging roles for lipids in shaping membrane-protein function. Nature, 2009. 459(7245): p. 379-85. 39. Stier, A. and E. Sackmann, Spin labels as enzyme substrates. Heterogeneous lipid distribution in liver microsomal membranes. Biochim Biophys Acta, 1973. 311(3): p. 400- 8. 40. Lee, A.G., et al., Clusters in lipid bilayers and the interpretation of thermal effects in biological membranes. Biochemistry, 1974. 13(18): p. 3699-705. 41. Wunderlich, F., et al., Thermotropic lipid clustering in tetrahymena membranes. Biochemistry, 1975. 14(17): p. 3730-5.

118

42. Wunderlich, F., et al., Thermotropic fluid goes to ordered "discontinuous" phase separation in microsomal lipids of Tetrahymena. An X-ray diffraction study. Biochemistry, 1978. 17(10): p. 2005-10. 43. Karnovsky, M.J., et al., The concept of lipid domains in membranes. J Cell Biol, 1982. 94(1): p. 1-6. 44. Simons, K. and E. Ikonen, Functional rafts in cell membranes. Nature, 1997. 387(6633): p. 569-72. 45. Pike, L.J., Rafts defined: a report on the Keystone Symposium on Lipid Rafts and Cell Function. J Lipid Res, 2006. 47(7): p. 1597-8. 46. Brown, D.A. and E. London, Functions of lipid rafts in biological membranes. Annu Rev Cell Dev Biol, 1998. 14: p. 111-36. 47. Hooper, N.M., Detergent-insoluble glycosphingolipid/cholesterol-rich membrane domains, lipid rafts and caveolae (review). Mol Membr Biol, 1999. 16(2): p. 145-56. 48. Resh, M.D., Fatty acylation of proteins: new insights into membrane targeting of myristoylated and palmitoylated proteins. Biochim Biophys Acta, 1999. 1451(1): p. 1-16. 49. Rietveld, A., et al., Association of sterol- and glycosylphosphatidylinositol-linked proteins with Drosophila raft lipid microdomains. J Biol Chem, 1999. 274(17): p. 12049-54. 50. van Meer, G., Lipid traffic in animal cells. Annu Rev Cell Biol, 1989. 5: p. 247-75. 51. Simons, K. and G. van Meer, Lipid sorting in epithelial cells. Biochemistry, 1988. 27(17): p. 6197-202. 52. Parton, R.G., Caveolae and caveolins. Curr Opin Cell Biol, 1996. 8(4): p. 542-8. 53. Smart, E.J., et al., Caveolins, liquid-ordered domains, and signal transduction. Mol Cell Biol, 1999. 19(11): p. 7289-304. 54. Trushina, E., et al., Neurological abnormalities in caveolin-1 knock out mice. Behav Brain Res, 2006. 172(1): p. 24-32. 55. Lang, D.M., et al., Identification of reggie-1 and reggie-2 as plasmamembrane-associated proteins which cocluster with activated GPI-anchored cell adhesion molecules in non- caveolar micropatches in neurons. J Neurobiol, 1998. 37(4): p. 502-23. 56. Brown, D.A. and J.K. Rose, Sorting of GPI-anchored proteins to glycolipid-enriched membrane subdomains during transport to the apical cell surface. Cell, 1992. 68(3): p. 533-44. 57. Foster, L.J., C.L. De Hoog, and M. Mann, Unbiased quantitative proteomics of lipid rafts reveals high specificity for signaling factors. Proc Natl Acad Sci U S A, 2003. 100(10): p. 5813-8. 58. Sprenger, R.R., et al., Comparative proteomics of human endothelial cell caveolae and rafts using two-dimensional gel electrophoresis and mass spectrometry. Electrophoresis, 2004. 25(1): p. 156-72. 59. Bae, T.J., et al., Lipid raft proteome reveals ATP synthase complex in the cell surface. Proteomics, 2004. 4(11): p. 3536-48. 60. Gupta, N., et al., Quantitative proteomic analysis of B cell lipid rafts reveals that ezrin regulates antigen receptor-mediated lipid raft dynamics. Nat Immunol, 2006. 7(6): p. 625- 33. 61. MacLellan, D.L., et al., A quantitative proteomic analysis of growth factor-induced compositional changes in lipid rafts of human smooth muscle cells. Proteomics, 2005. 5(18): p. 4733-42.

119

62. Robinson, P.J., Phosphatidylinositol membrane anchors and T-cell activation. Immunol Today, 1991. 12(1): p. 35-41. 63. Brown, D., The tyrosine kinase connection: how GPI-anchored proteins activate T cells. Curr Opin Immunol, 1993. 5(3): p. 349-54. 64. Draberova, L. and P. Draber, Thy-1 glycoprotein and src-like protein-tyrosine kinase p53/p56lyn are associated in large detergent-resistant complexes in rat basophilic leukemia cells. Proc Natl Acad Sci U S A, 1993. 90(8): p. 3611-5. 65. Morgan, B.P., et al., Cross-linking of CD59 and of other glycosyl phosphatidylinositol- anchored molecules on neutrophils triggers cell activation via tyrosine kinase. Eur J Immunol, 1993. 23(11): p. 2841-50. 66. van den Berg, C.W., et al., Exogenous glycosyl phosphatidylinositol-anchored CD59 associates with kinases in membrane clusters on U937 cells and becomes Ca(2+)- signaling competent. J Cell Biol, 1995. 131(3): p. 669-77. 67. Zisch, A.H., et al., The glypiated neuronal cell adhesion molecule contactin/F11 complexes with src-family protein tyrosine kinase Fyn. Mol Cell Neurosci, 1995. 6(3): p. 263-79. 68. Field, K.A., D. Holowka, and B. Baird, Fc epsilon RI-mediated recruitment of p53/56lyn to detergent-resistant membrane domains accompanies cellular signaling. Proc Natl Acad Sci U S A, 1995. 92(20): p. 9201-5. 69. Sheets, E.D., D. Holowka, and B. Baird, Membrane organization in immunoglobulin E receptor signaling. Curr Opin Chem Biol, 1999. 3(1): p. 95-9. 70. Baird, B., E.D. Sheets, and D. Holowka, How does the plasma membrane participate in cellular signaling by receptors for immunoglobulin E? Biophys Chem, 1999. 82(2-3): p. 109-19. 71. Metzger, H., It's spring, and thoughts turn to...allergies. Cell, 1999. 97(3): p. 287-90. 72. Stauffer, T.P. and T. Meyer, Compartmentalized IgE receptor-mediated signal transduction in living cells. J Cell Biol, 1997. 139(6): p. 1447-54. 73. Holowka, D., E.D. Sheets, and B. Baird, Interactions between Fc(epsilon)RI and lipid raft components are regulated by the actin cytoskeleton. J Cell Sci, 2000. 113 ( Pt 6): p. 1009- 19. 74. Sheets, E.D., D. Holowka, and B. Baird, Critical role for cholesterol in Lyn-mediated tyrosine phosphorylation of FcepsilonRI and their association with detergent-resistant membranes. J Cell Biol, 1999. 145(4): p. 877-87. 75. Janes, P.W., et al., The role of lipid rafts in T cell antigen receptor (TCR) signalling. Semin Immunol, 2000. 12(1): p. 23-34. 76. Langlet, C., et al., Membrane rafts and signaling by the multichain immune recognition receptors. Curr Opin Immunol, 2000. 12(3): p. 250-5. 77. Zhang, W., R.P. Trible, and L.E. Samelson, LAT palmitoylation: its essential role in membrane microdomain targeting and tyrosine phosphorylation during T cell activation. Immunity, 1998. 9(2): p. 239-46. 78. Brdicka, T., J. Cerny, and V. Horejsi, T cell receptor signalling results in rapid tyrosine phosphorylation of the linker protein LAT present in detergent-resistant membrane microdomains. Biochem Biophys Res Commun, 1998. 248(2): p. 356-60. 79. Lin, J., A. Weiss, and T.S. Finco, Localization of LAT in glycolipid-enriched microdomains is required for T cell activation. J Biol Chem, 1999. 274(41): p. 28861-4.

120

80. Moran, M. and M.C. Miceli, Engagement of GPI-linked CD48 contributes to TCR signals and cytoskeletal reorganization: a role for lipid rafts in T cell activation. Immunity, 1998. 9(6): p. 787-96. 81. Stefanova, I., et al., GPI-anchored cell-surface molecules complexed to protein tyrosine kinases. Science, 1991. 254(5034): p. 1016-9. 82. Montixi, C., et al., Engagement of T cell receptor triggers its recruitment to low-density detergent-insoluble membrane domains. EMBO J, 1998. 17(18): p. 5334-48. 83. Xavier, R., et al., Membrane compartmentation is required for efficient T cell activation. Immunity, 1998. 8(6): p. 723-32. 84. Viola, A., et al., T lymphocyte costimulation mediated by reorganization of membrane microdomains. Science, 1999. 283(5402): p. 680-2. 85. Cary, L.A. and J.A. Cooper, Molecular switches in lipid rafts. Nature, 2000. 404(6781): p. 945, 947. 86. Anton van der Merwe, P., et al., Cytoskeletal polarization and redistribution of cell-surface molecules during T cell antigen recognition. Semin Immunol, 2000. 12(1): p. 5-21. 87. Grakoui, A., et al., The immunological synapse: a molecular machine controlling T cell activation. Science, 1999. 285(5425): p. 221-7. 88. Janes, P.W., S.C. Ley, and A.I. Magee, Aggregation of lipid rafts accompanies signaling via the T cell antigen receptor. J Cell Biol, 1999. 147(2): p. 447-61. 89. Nieva, J.L., et al., Membrane fusion of Semliki Forest virus requires sphingolipids in the target membrane. EMBO J, 1994. 13(12): p. 2797-804. 90. Phalen, T. and M. Kielian, Cholesterol is required for infection by Semliki Forest virus. J Cell Biol, 1991. 112(4): p. 615-23. 91. Stang, E., J. Kartenbeck, and R.G. Parton, Major histocompatibility complex class I molecules mediate association of SV40 with caveolae. Mol Biol Cell, 1997. 8(1): p. 47-57. 92. Pelkmans, L., D. Puntener, and A. Helenius, Local actin polymerization and dynamin recruitment in SV40-induced internalization of caveolae. Science, 2002. 296(5567): p. 535-9. 93. Damm, E.M., et al., Clathrin- and caveolin-1-independent endocytosis: entry of simian virus 40 into cells devoid of caveolae. J Cell Biol, 2005. 168(3): p. 477-88. 94. Sharma, P., et al., Nanoscale organization of multiple GPI-anchored proteins in living cell membranes. Cell, 2004. 116(4): p. 577-89. 95. Vyas, N., et al., Nanoscale organization of hedgehog is essential for long-range signaling. Cell, 2008. 133(7): p. 1214-27. 96. Pinaud, F., et al., Dynamic partitioning of a glycosyl-phosphatidylinositol-anchored protein in glycosphingolipid-rich microdomains imaged by single-quantum dot tracking. Traffic, 2009. 10(6): p. 691-712. 97. Lenne, P.F., et al., Dynamic molecular confinement in the plasma membrane by microdomains and the cytoskeleton meshwork. EMBO J, 2006. 25(14): p. 3245-56. 98. Hell, S.W., Far-field optical nanoscopy. Science, 2007. 316(5828): p. 1153-8. 99. Shroff, H., et al., Live-cell photoactivated localization microscopy of nanoscale adhesion dynamics. Nat Methods, 2008. 5(5): p. 417-23. 100. Rust, M.J., M. Bates, and X. Zhuang, Sub-diffraction-limit imaging by stochastic optical reconstruction microscopy (STORM). Nat Methods, 2006. 3(10): p. 793-5. 101. Eggeling, C., et al., Direct observation of the nanoscale dynamics of membrane lipids in a living cell. Nature, 2009. 457(7233): p. 1159-62.

121

102. Penmatsa, A., K.H. Wang, and E. Gouaux, X-ray structure of dopamine transporter elucidates antidepressant mechanism. Nature, 2013. 503(7474): p. 85-90. 103. Hong, W.C. and S.G. Amara, Membrane cholesterol modulates the outward facing conformation of the dopamine transporter and alters cocaine binding. J Biol Chem, 2010. 285(42): p. 32616-26. 104. Jones, K.T., J. Zhen, and M.E. Reith, Importance of cholesterol in dopamine transporter function. J Neurochem, 2012. 123(5): p. 700-15. 105. Cremona, M.L., et al., Flotillin-1 is essential for PKC-triggered endocytosis and membrane microdomain localization of DAT. Nat Neurosci, 2011. 14(4): p. 469-77. 106. Galbiati, F., et al., Identification, sequence and developmental expression of invertebrate flotillins from Drosophila melanogaster. Gene, 1998. 210(2): p. 229-37. 107. Stuermer, C.A., Reggie/flotillin and the targeted delivery of cargo. J Neurochem, 2011. 116(5): p. 708-13. 108. Banning, A., A. Tomasovic, and R. Tikkanen, Functional aspects of membrane association of reggie/flotillin proteins. Curr Protein Pept Sci, 2011. 12(8): p. 725-35. 109. Bickel, P.E., et al., Flotillin and epidermal surface antigen define a new family of caveolae- associated integral membrane proteins. J Biol Chem, 1997. 272(21): p. 13793-802. 110. Schulte, T., et al., Reggie-1 and reggie-2, two cell surface proteins expressed by retinal ganglion cells during axon regeneration. Development, 1997. 124(2): p. 577-87. 111. Stuermer, C.A., et al., Glycosylphosphatidyl inositol-anchored proteins and fyn kinase assemble in noncaveolar plasma membrane microdomains defined by reggie-1 and -2. Mol Biol Cell, 2001. 12(10): p. 3031-45. 112. Morrow, I.C. and R.G. Parton, Flotillins and the PHB domain protein family: rafts, worms and anaesthetics. Traffic, 2005. 6(9): p. 725-40. 113. Volonte, D., et al., Flotillins/cavatellins are differentially expressed in cells and tissues and form a hetero-oligomeric complex with caveolins in vivo. Characterization and epitope-mapping of a novel flotillin-1 monoclonal antibody probe. J Biol Chem, 1999. 274(18): p. 12702-9. 114. Dermine, J.F., et al., Flotillin-1-enriched lipid raft domains accumulate on maturing phagosomes. J Biol Chem, 2001. 276(21): p. 18507-12. 115. de Gassart, A., et al., Lipid raft-associated protein sorting in exosomes. Blood, 2003. 102(13): p. 4336-44. 116. Strauss, K., et al., Exosome secretion ameliorates lysosomal storage of cholesterol in Niemann-Pick type C disease. J Biol Chem, 2010. 285(34): p. 26279-88. 117. Gkantiragas, I., et al., Sphingomyelin-enriched microdomains at the Golgi complex. Mol Biol Cell, 2001. 12(6): p. 1819-33. 118. Morrow, I.C., et al., Flotillin-1/reggie-2 traffics to surface raft domains via a novel golgi- independent pathway. Identification of a novel membrane targeting domain and a role for palmitoylation. J Biol Chem, 2002. 277(50): p. 48834-41. 119. Solis, G.P., et al., Reggie/flotillin proteins are organized into stable tetramers in membrane microdomains. Biochem J, 2007. 403(2): p. 313-22. 120. Frick, M., et al., Coassembly of flotillins induces formation of membrane microdomains, membrane curvature, and vesicle budding. Curr Biol, 2007. 17(13): p. 1151-6. 121. Langhorst, M.F., A. Reuter, and C.A. Stuermer, Scaffolding microdomains and beyond: the function of reggie/flotillin proteins. Cell Mol Life Sci, 2005. 62(19-20): p. 2228-40.

122

122. Hoegg, M.B., et al., Distinct regions within the erlins are required for oligomerization and association with high molecular weight complexes. J Biol Chem, 2009. 284(12): p. 7766- 76. 123. Browman, D.T., M.B. Hoegg, and S.M. Robbins, The SPFH domain-containing proteins: more than lipid raft markers. Trends Cell Biol, 2007. 17(8): p. 394-402. 124. Salzer, U. and R. Prohaska, Stomatin, flotillin-1, and flotillin-2 are major integral proteins of erythrocyte lipid rafts. Blood, 2001. 97(4): p. 1141-3. 125. Huber, T.B., et al., Molecular basis of the functional podocin-nephrin complex: mutations in the NPHS2 gene disrupt nephrin targeting to lipid raft microdomains. Hum Mol Genet, 2003. 12(24): p. 3397-405. 126. Tatsuta, T., K. Model, and T. Langer, Formation of membrane-bound ring complexes by prohibitins in mitochondria. Mol Biol Cell, 2005. 16(1): p. 248-59. 127. Hiebl-Dirschmied, C.M., G.R. Adolf, and R. Prohaska, Isolation and partial characterization of the human erythrocyte band 7 integral membrane protein. Biochim Biophys Acta, 1991. 1065(2): p. 195-202. 128. Hiebl-Dirschmied, C.M., et al., Cloning and nucleotide sequence of cDNA encoding human erythrocyte band 7 integral membrane protein. Biochim Biophys Acta, 1991. 1090(1): p. 123-4. 129. Salzer, U., H. Ahorn, and R. Prohaska, Identification of the phosphorylation site on human erythrocyte band 7 integral membrane protein: implications for a monotopic protein structure. Biochim Biophys Acta, 1993. 1151(2): p. 149-52. 130. Garin, J., et al., The phagosome proteome: insight into phagosome functions. J Cell Biol, 2001. 152(1): p. 165-80. 131. Snyers, L., E. Umlauf, and R. Prohaska, Association of stomatin with lipid-protein complexes in the plasma membrane and the endocytic compartment. Eur J Cell Biol, 1999. 78(11): p. 802-12. 132. Snyers, L., E. Umlauf, and R. Prohaska, Oligomeric nature of the integral membrane protein stomatin. J Biol Chem, 1998. 273(27): p. 17221-6. 133. Snyers, L., D. Thines-Sempoux, and R. Prohaska, Colocalization of stomatin (band 7.2b) and actin microfilaments in UAC epithelial cells. Eur J Cell Biol, 1997. 73(3): p. 281-5. 134. Stewart, G.W., A.C. Argent, and B.C. Dash, Stomatin: a putative cation transport regulator in the red cell membrane. Biochim Biophys Acta, 1993. 1225(1): p. 15-25. 135. Zhu, Y., et al., Stomatocytosis is absent in "stomatin"-deficient murine red blood cells. Blood, 1999. 93(7): p. 2404-10. 136. Delaunay, J., G. Stewart, and A. Iolascon, Hereditary dehydrated and overhydrated stomatocytosis: recent advances. Curr Opin Hematol, 1999. 6(2): p. 110-4. 137. Innes, D.S., et al., Exclusion of the stomatin, alpha-adducin and beta-adducin loci in a large kindred with dehydrated hereditary stomatocytosis. Am J Hematol, 1999. 60(1): p. 72-4. 138. Zhang, J.Z., et al., Association of stomatin (band 7.2b) with Glut1 glucose transporter. Arch Biochem Biophys, 1999. 372(1): p. 173-8. 139. Zhang, J.Z., et al., Overexpression of stomatin depresses GLUT-1 glucose transporter activity. Am J Physiol Cell Physiol, 2001. 280(5): p. C1277-83. 140. Goodman, M.B., et al., MEC-2 regulates C. elegans DEG/ENaC channels needed for mechanosensation. Nature, 2002. 415(6875): p. 1039-42.

123

141. Price, M.P., et al., Stomatin modulates gating of acid-sensing ion channels. J Biol Chem, 2004. 279(51): p. 53886-91. 142. Desneves, J., et al., Human erythrocyte band 7.2b is preferentially labeled by a photoreactive phospholipid. Biochem Biophys Res Commun, 1996. 224(1): p. 108-14. 143. Roselli, S., et al., Podocin localizes in the kidney to the slit diaphragm area. Am J Pathol, 2002. 160(1): p. 131-9. 144. Schwarz, K., et al., Podocin, a raft-associated component of the glomerular slit diaphragm, interacts with CD2AP and nephrin. J Clin Invest, 2001. 108(11): p. 1621-9. 145. Shih, N.Y., et al., Congenital nephrotic syndrome in mice lacking CD2-associated protein. Science, 1999. 286(5438): p. 312-5. 146. Huber, T.B., et al., Interaction with podocin facilitates nephrin signaling. J Biol Chem, 2001. 276(45): p. 41543-6. 147. Correction to "NPHS2, encoding the glomerular protein podocin, is mutated in autosomal recessive steroid-resistant nephrotic syndrome". Nat Genet, 2000. 25(1): p. 125. 148. Fuchshuber, A., et al., Mapping a gene (SRN1) to chromosome 1q25-q31 in idiopathic nephrotic syndrome confirms a distinct entity of autosomal recessive nephrosis. Hum Mol Genet, 1995. 4(11): p. 2155-8. 149. Miner, J.H., Focusing on the glomerular slit diaphragm: podocin enters the picture. Am J Pathol, 2002. 160(1): p. 3-5. 150. Ikonen, E., et al., Prohibitin, an antiproliferative protein, is localized to mitochondria. FEBS Lett, 1995. 358(3): p. 273-7. 151. Steglich, G., W. Neupert, and T. Langer, Prohibitins regulate membrane protein degradation by the m-AAA protease in mitochondria. Mol Cell Biol, 1999. 19(5): p. 3435- 42. 152. Nijtmans, L.G., et al., Prohibitins act as a membrane-bound chaperone for the stabilization of mitochondrial proteins. EMBO J, 2000. 19(11): p. 2444-51. 153. Wei, Y., et al., Prohibitin 2 Is an Inner Mitochondrial Membrane Mitophagy Receptor. Cell, 2017. 168(1-2): p. 224-238 e10. 154. Terashima, M., et al., The IgM antigen receptor of B lymphocytes is associated with prohibitin and a prohibitin-related protein. EMBO J, 1994. 13(16): p. 3782-92. 155. Mielenz, D., et al., Lipid rafts associate with intracellular B cell receptors and exhibit a B cell stage-specific protein composition. J Immunol, 2005. 174(6): p. 3508-17. 156. Rajalingam, K., et al., Prohibitin is required for Ras-induced Raf-MEK-ERK activation and epithelial cell migration. Nat Cell Biol, 2005. 7(8): p. 837-43. 157. Banuett, F. and I. Herskowitz, Identification of polypeptides encoded by an Escherichia coli locus (hflA) that governs the lysis-lysogeny decision of bacteriophage lambda. J Bacteriol, 1987. 169(9): p. 4076-85. 158. Zorick, T.S. and H. Echols, Membrane localization of the HflA regulatory protease of Escherichia coli by immunoelectron microscopy. J Bacteriol, 1991. 173(19): p. 6307-10. 159. Noble, J.A., et al., The Escherichia coli hflA locus encodes a putative GTP-binding protein and two membrane proteins, one of which contains a protease-like domain. Proc Natl Acad Sci U S A, 1993. 90(22): p. 10866-70. 160. Tannert, A., et al., Protein-mediated transbilayer movement of lipids in eukaryotes and prokaryotes: the relevance of ABC transporters. Int J Antimicrob Agents, 2003. 22(3): p. 177-87.

124

161. Bramkamp, M. and D. Lopez, Exploring the existence of lipid rafts in bacteria. Microbiol Mol Biol Rev, 2015. 79(1): p. 81-100. 162. Saikawa, N., Y. Akiyama, and K. Ito, FtsH exists as an exceptionally large complex containing HflKC in the plasma membrane of Escherichia coli. J Struct Biol, 2004. 146(1- 2): p. 123-9. 163. Munderloh, C., et al., Reggies/flotillins regulate retinal axon regeneration in the zebrafish optic nerve and differentiation of hippocampal and N2a neurons. J Neurosci, 2009. 29(20): p. 6607-15. 164. Langhorst, M.F., et al., Reggies/flotillins regulate cytoskeletal remodeling during neuronal differentiation via CAP/ponsin and Rho GTPases. Eur J Cell Biol, 2008. 87(12): p. 921- 31. 165. Haglund, K., et al., Recruitment of Pyk2 and Cbl to lipid rafts mediates signals important for actin reorganization in growing neurites. J Cell Sci, 2004. 117(Pt 12): p. 2557-68. 166. Carcea, I., et al., Flotillin-mediated endocytic events dictate cell type-specific responses to semaphorin 3A. J Neurosci, 2010. 30(45): p. 15317-29. 167. Swanwick, C.C., et al., Flotillin-1 promotes formation of glutamatergic synapses in hippocampal neurons. Dev Neurobiol, 2010. 70(13): p. 875-83. 168. Bodin, S., et al., Flotillins in intercellular adhesion - from cellular physiology to human diseases. J Cell Sci, 2014. 127(Pt 24): p. 5139-47. 169. Causeret, M., et al., N-cadherin association with lipid rafts regulates its dynamic assembly at cell-cell junctions in C2C12 myoblasts. Mol Biol Cell, 2005. 16(5): p. 2168-80. 170. Taulet, N., et al., N-cadherin/p120 catenin association at cell-cell contacts occurs in cholesterol-rich membrane domains and is required for RhoA activation and myogenesis. J Biol Chem, 2009. 284(34): p. 23137-45. 171. Fernow, I., A. Icking, and R. Tikkanen, Reggie-1 and reggie-2 localize in non-caveolar rafts in epithelial cells: cellular localization is not dependent on the expression of caveolin proteins. Eur J Cell Biol, 2007. 86(6): p. 345-52. 172. Liu, J., et al., The stomatin/prohibitin/flotillin/HflK/C domain of flotillin-1 contains distinct sequences that direct plasma membrane localization and protein interactions in 3T3-L1 adipocytes. J Biol Chem, 2005. 280(16): p. 16125-34. 173. Malaga-Trillo, E., et al., Regulation of embryonic cell adhesion by the prion protein. PLoS Biol, 2009. 7(3): p. e55. 174. Roitbak, T., et al., A polycystin multiprotein complex constitutes a cholesterol-containing signalling microdomain in human kidney epithelia. Biochem J, 2005. 392(Pt 1): p. 29-38. 175. von Philipsborn, A.C., et al., Restricted expression of reggie genes and proteins during early zebrafish development. J Comp Neurol, 2005. 482(3): p. 257-72. 176. Hoehne, M., et al., Loss- and gain-of-function analysis of the lipid raft proteins Reggie/Flotillin in Drosophila: they are posttranslationally regulated, and misexpression interferes with wing and eye development. Mol Cell Neurosci, 2005. 30(3): p. 326-38. 177. Solis, G.P., et al., Reggies/flotillins regulate E-cadherin-mediated cell contact formation by affecting EGFR trafficking. Mol Biol Cell, 2012. 23(10): p. 1812-25. 178. Guillaume, E., et al., Flotillin microdomains stabilize cadherins at cell-cell junctions. J Cell Sci, 2013. 126(Pt 22): p. 5293-304. 179. Kurrle, N., et al., Flotillins directly interact with gamma-catenin and regulate epithelial cell-cell adhesion. PLoS One, 2013. 8(12): p. e84393.

125

180. Hong, S., R.B. Troyanovsky, and S.M. Troyanovsky, Binding to F-actin guides cadherin cluster assembly, stability, and movement. J Cell Biol, 2013. 201(1): p. 131-43. 181. Truong Quang, B.A., et al., Principles of E-cadherin supramolecular organization in vivo. Curr Biol, 2013. 23(22): p. 2197-207. 182. Morris, E.A.R., et al., Flotillins control zebrafish epiboly through their role in cadherin- mediated cell-cell adhesion. Biol Cell, 2017. 109(5): p. 210-221. 183. Glebov, O.O., N.A. Bright, and B.J. Nichols, Flotillin-1 defines a clathrin-independent endocytic pathway in mammalian cells. Nat Cell Biol, 2006. 8(1): p. 46-54. 184. Babuke, T., et al., Hetero-oligomerization of reggie-1/flotillin-2 and reggie-2/flotillin-1 is required for their endocytosis. Cell Signal, 2009. 21(8): p. 1287-97. 185. Riento, K., et al., Endocytosis of flotillin-1 and flotillin-2 is regulated by Fyn kinase. J Cell Sci, 2009. 122(Pt 7): p. 912-8. 186. Melikian, H.E., Neurotransmitter transporter trafficking: endocytosis, recycling, and regulation. Pharmacol Ther, 2004. 104(1): p. 17-27. 187. Ge, L., et al., Flotillins play an essential role in Niemann-Pick C1-like 1-mediated cholesterol uptake. Proc Natl Acad Sci U S A, 2011. 108(2): p. 551-6. 188. Sorkina, T., et al., Constitutive and protein kinase C-induced internalization of the dopamine transporter is mediated by a clathrin-dependent mechanism. Traffic, 2005. 6(2): p. 157-70. 189. Miranda, M., et al., Three ubiquitin conjugation sites in the amino terminus of the dopamine transporter mediate protein kinase C-dependent endocytosis of the transporter. Mol Biol Cell, 2007. 18(1): p. 313-23. 190. Di Guglielmo, G.M., et al., Distinct endocytic pathways regulate TGF-beta receptor signalling and turnover. Nat Cell Biol, 2003. 5(5): p. 410-21. 191. Doherty, G.J. and H.T. McMahon, Mechanisms of endocytosis. Annu Rev Biochem, 2009. 78: p. 857-902. 192. Deinhardt, K., et al., Tetanus toxin is internalized by a sequential clathrin-dependent mechanism initiated within lipid microdomains and independent of epsin1. J Cell Biol, 2006. 174(3): p. 459-71. 193. Stuermer, C.A., et al., PrPc capping in T cells promotes its association with the lipid raft proteins reggie-1 and reggie-2 and leads to signal transduction. FASEB J, 2004. 18(14): p. 1731-3. 194. Slaughter, N., et al., The flotillins are integral membrane proteins in lipid rafts that contain TCR-associated signaling components: implications for T-cell activation. Clin Immunol, 2003. 108(2): p. 138-51. 195. Amaddii, M., et al., Flotillin-1/reggie-2 protein plays dual role in activation of receptor- tyrosine kinase/mitogen-activated protein kinase signaling. J Biol Chem, 2012. 287(10): p. 7265-78. 196. Tomasovic, A., S. Traub, and R. Tikkanen, Molecular networks in FGF signaling: flotillin- 1 and cbl-associated protein compete for the binding to fibroblast growth factor receptor substrate 2. PLoS One, 2012. 7(1): p. e29739. 197. Ockenga, W., et al., Epidermal growth factor receptor transactivation is required for mitogen-activated protein kinase activation by muscarinic acetylcholine receptors in HaCaT keratinocytes. Int J Mol Sci, 2014. 15(11): p. 21433-54.

126

198. Kuhne, S., et al., Cholinergic transactivation of the EGFR in HaCaT keratinocytes stimulates a flotillin-1 dependent MAPK-mediated transcriptional response. Int J Mol Sci, 2015. 16(3): p. 6447-63. 199. Bauer, M. and L. Pelkmans, A new paradigm for membrane-organizing and -shaping scaffolds. FEBS Lett, 2006. 580(23): p. 5559-64. 200. Adkins, E.M., et al., Membrane mobility and microdomain association of the dopamine transporter studied with fluorescence correlation spectroscopy and fluorescence recovery after photobleaching. Biochemistry, 2007. 46(37): p. 10484-97. 201. Foster, J.D., et al., Phorbol ester induced trafficking-independent regulation and enhanced phosphorylation of the dopamine transporter associated with membrane rafts and cholesterol. J Neurochem, 2008. 105(5): p. 1683-99. 202. Pizzo, A.B., et al., The membrane raft protein Flotillin-1 is essential in dopamine neurons for amphetamine-induced behavior in Drosophila. Mol Psychiatry, 2013. 18(7): p. 824-33. 203. Giros, B. and M.G. Caron, Molecular characterization of the dopamine transporter. Trends Pharmacol Sci, 1993. 14(2): p. 43-9. 204. Palmiter, R.D., Dopamine signaling in the dorsal striatum is essential for motivated behaviors: lessons from dopamine-deficient mice. Ann N Y Acad Sci, 2008. 1129: p. 35- 46. 205. Sulzer, D., et al., Mechanisms of neurotransmitter release by amphetamines: a review. Prog Neurobiol, 2005. 75(6): p. 406-33. 206. Sulzer, D., How addictive drugs disrupt presynaptic dopamine neurotransmission. Neuron, 2011. 69(4): p. 628-49. 207. Khoshbouei, H., et al., N-terminal phosphorylation of the dopamine transporter is required for amphetamine-induced efflux. PLoS Biol, 2004. 2(3): p. E78. 208. Giros, B., et al., Cloning and functional characterization of a cocaine-sensitive dopamine transporter. FEBS Lett, 1991. 295(1-3): p. 149-54. 209. Li, L.B., et al., The role of N-glycosylation in function and surface trafficking of the human dopamine transporter. J Biol Chem, 2004. 279(20): p. 21012-20. 210. Torres, G.E., et al., Oligomerization and trafficking of the human dopamine transporter. Mutational analysis identifies critical domains important for the functional expression of the transporter. J Biol Chem, 2003. 278(4): p. 2731-9. 211. Fog, J.U., et al., Calmodulin kinase II interacts with the dopamine transporter C terminus to regulate amphetamine-induced reverse transport. Neuron, 2006. 51(4): p. 417-29. 212. Batchelor, M. and J.O. Schenk, Protein kinase A activity may kinetically upregulate the striatal transporter for dopamine. J Neurosci, 1998. 18(24): p. 10304-9. 213. Yamashita, A., et al., Crystal structure of a bacterial homologue of Na+/Cl--dependent neurotransmitter transporters. Nature, 2005. 437(7056): p. 215-23. 214. Kristensen, A.S., et al., SLC6 neurotransmitter transporters: structure, function, and regulation. Pharmacol Rev, 2011. 63(3): p. 585-640. 215. Koldso, H., et al., Comparative modeling of the human monoamine transporters: similarities in substrate binding. ACS Chem Neurosci, 2013. 4(2): p. 295-309. 216. Chen, N. and J.B. Justice, Differential effect of structural modification of human dopamine transporter on the inward and outward transport of dopamine. Brain Res Mol Brain Res, 2000. 75(2): p. 208-15.

127

217. Gu, H., S.C. Wall, and G. Rudnick, Stable expression of biogenic amine transporters reveals differences in inhibitor sensitivity, kinetics, and ion dependence. J Biol Chem, 1994. 269(10): p. 7124-30. 218. Zomot, E., et al., Mechanism of chloride interaction with neurotransmitter:sodium symporters. Nature, 2007. 449(7163): p. 726-30. 219. Moron, J.A., et al., Dopamine uptake through the norepinephrine transporter in brain regions with low levels of the dopamine transporter: evidence from knock-out mouse lines. J Neurosci, 2002. 22(2): p. 389-95. 220. Gether, U., et al., Neurotransmitter transporters: molecular function of important drug targets. Trends Pharmacol Sci, 2006. 27(7): p. 375-83. 221. Richtand, N.M., et al., Regional quantification of dopamine transporter mRNA in rat brain using a ribonuclease protection assay. Neurosci Lett, 1995. 200(2): p. 73-6. 222. Ciliax, B.J., et al., Immunocytochemical localization of the dopamine transporter in human brain. J Comp Neurol, 1999. 409(1): p. 38-56. 223. Nirenberg, M.J., et al., Immunogold localization of the dopamine transporter: an ultrastructural study of the rat ventral tegmental area. J Neurosci, 1997. 17(14): p. 5255- 62. 224. Hersch, S.M., et al., Subcellular localization and molecular topology of the dopamine transporter in the striatum and substantia nigra. J Comp Neurol, 1997. 388(2): p. 211-27. 225. Shumay, E., J.S. Fowler, and N.D. Volkow, Genomic features of the human dopamine transporter gene and its potential epigenetic States: implications for phenotypic diversity. PLoS One, 2010. 5(6): p. e11067. 226. Weiner, R.I. and W.F. Ganong, Role of brain monoamines and histamine in regulation of anterior pituitary secretion. Physiol Rev, 1978. 58(4): p. 905-76. 227. Weintraub, D. and D.J. Burn, Parkinson's disease: the quintessential neuropsychiatric disorder. Mov Disord, 2011. 26(6): p. 1022-31. 228. Lee, F.J., et al., Direct binding and functional coupling of alpha-synuclein to the dopamine transporters accelerate dopamine-induced apoptosis. FASEB J, 2001. 15(6): p. 916-26. 229. Teismann, P. and J.B. Schulz, Cellular pathology of Parkinson's disease: astrocytes, microglia and inflammation. Cell Tissue Res, 2004. 318(1): p. 149-61. 230. Takeda, H., M. Inazu, and T. Matsumiya, Astroglial dopamine transport is mediated by norepinephrine transporter. Naunyn Schmiedebergs Arch Pharmacol, 2002. 366(6): p. 620-3. 231. Inazu, M., H. Takeda, and T. Matsumiya, Functional expression of the norepinephrine transporter in cultured rat astrocytes. J Neurochem, 2003. 84(1): p. 136-44. 232. Nutt, J.G., J.H. Carter, and G.J. Sexton, The dopamine transporter: importance in Parkinson's disease. Ann Neurol, 2004. 55(6): p. 766-73. 233. Miller, G.W., et al., Immunochemical analysis of dopamine transporter protein in Parkinson's disease. Ann Neurol, 1997. 41(4): p. 530-9. 234. Sossi, V., et al., Dopamine transporter relation to dopamine turnover in Parkinson's disease: a positron emission tomography study. Ann Neurol, 2007. 62(5): p. 468-74. 235. Ishibashi, K., et al., Regional analysis of age-related decline in dopamine transporters and dopamine D2-like receptors in human striatum. Synapse, 2009. 63(4): p. 282-90. 236. Ishibashi, K., et al., Comparison of dopamine transporter decline in a patient with Parkinson's disease and normal aging effect. J Neurol Sci, 2014. 339(1-2): p. 207-9.

128

237. Laruelle, M., et al., Single photon emission computerized tomography imaging of amphetamine-induced dopamine release in drug-free schizophrenic subjects. Proc Natl Acad Sci U S A, 1996. 93(17): p. 9235-40. 238. Lavalaye, J., et al., Dopamine transporter density in young patients with schizophrenia assessed with [123]FP-CIT SPECT. Schizophr Res, 2001. 47(1): p. 59-67. 239. Laruelle, M., et al., Dopamine and serotonin transporters in patients with schizophrenia: an imaging study with [(123)I]beta-CIT. Biol Psychiatry, 2000. 47(5): p. 371-9. 240. Fusar-Poli, P. and A. Meyer-Lindenberg, Striatal presynaptic dopamine in schizophrenia, Part I: meta-analysis of dopamine active transporter (DAT) density. Schizophr Bull, 2013. 39(1): p. 22-32. 241. Fusar-Poli, P. and A. Meyer-Lindenberg, Striatal presynaptic dopamine in schizophrenia, part II: meta-analysis of [(18)F/(11)C]-DOPA PET studies. Schizophr Bull, 2013. 39(1): p. 33-42. 242. Barr, L.C., W.K. Goodman, and L.H. Price, The serotonin hypothesis of obsessive compulsive disorder. Int Clin Psychopharmacol, 1993. 8 Suppl 2: p. 79-82. 243. Miguita, K., et al., Association study between genetic monoaminergic polymorphisms and OCD response to clomipramine treatment. Arq Neuropsiquiatr, 2011. 69(2B): p. 283-7. 244. Goodman, W.K., et al., Beyond the serotonin hypothesis: a role for dopamine in some forms of obsessive compulsive disorder? J Clin Psychiatry, 1990. 51 Suppl: p. 36-43; discussion 55-8. 245. Hesse, S., et al., Serotonin and dopamine transporter imaging in patients with obsessive- compulsive disorder. Psychiatry Res, 2005. 140(1): p. 63-72. 246. van der Wee, N.J., et al., Enhanced dopamine transporter density in psychotropic-naive patients with obsessive-compulsive disorder shown by [123I]{beta}-CIT SPECT. Am J Psychiatry, 2004. 161(12): p. 2201-6. 247. Gillis, J.J., et al., Attention deficit disorder in reading-disabled twins: evidence for a genetic etiology. J Abnorm Child Psychol, 1992. 20(3): p. 303-15. 248. Gill, M., et al., Confirmation of association between attention deficit hyperactivity disorder and a dopamine transporter polymorphism. Mol Psychiatry, 1997. 2(4): p. 311-3. 249. Roman, T., et al., Attention-deficit hyperactivity disorder: a study of association with both the dopamine transporter gene and the dopamine D4 receptor gene. Am J Med Genet, 2001. 105(5): p. 471-8. 250. Mazei-Robinson, M.S. and R.D. Blakely, ADHD and the dopamine transporter: are there reasons to pay attention? Handb Exp Pharmacol, 2006(175): p. 373-415. 251. Madras, B.K., G.M. Miller, and A.J. Fischman, The dopamine transporter and attention- deficit/hyperactivity disorder. Biol Psychiatry, 2005. 57(11): p. 1397-409. 252. Fusar-Poli, P., et al., Striatal dopamine transporter alterations in ADHD: pathophysiology or adaptation to psychostimulants? A meta-analysis. Am J Psychiatry, 2012. 169(3): p. 264-72. 253. Abrahams, B.S. and D.H. Geschwind, Advances in autism genetics: on the threshold of a new neurobiology. Nat Rev Genet, 2008. 9(5): p. 341-55. 254. Buxbaum, J.D., Multiple rare variants in the etiology of autism spectrum disorders. Dialogues Clin Neurosci, 2009. 11(1): p. 35-43. 255. Beaudet, A.L., Autism: highly heritable but not inherited. Nat Med, 2007. 13(5): p. 534-6. 256. Courchesne, E., et al., Mapping early brain development in autism. Neuron, 2007. 56(2): p. 399-413.

129

257. Persico, A.M. and T. Bourgeron, Searching for ways out of the autism maze: genetic, epigenetic and environmental clues. Trends Neurosci, 2006. 29(7): p. 349-358. 258. Hsiao, E.Y., Immune dysregulation in autism spectrum disorder. Int Rev Neurobiol, 2013. 113: p. 269-302. 259. Onore, C., M. Careaga, and P. Ashwood, The role of immune dysfunction in the pathophysiology of autism. Brain Behav Immun, 2012. 26(3): p. 383-92. 260. Rossignol, D.A. and R.E. Frye, Evidence linking oxidative stress, mitochondrial dysfunction, and inflammation in the brain of individuals with autism. Front Physiol, 2014. 5: p. 150. 261. Bowton, E., et al., SLC6A3 coding variant Ala559Val found in two autism probands alters dopamine transporter function and trafficking. Transl Psychiatry, 2014. 4: p. e464. 262. Tilley, M.R., et al., Cocaine reward and locomotion stimulation in mice with reduced dopamine transporter expression. BMC Neurosci, 2007. 8: p. 42. 263. Ritz, M.C., et al., Cocaine receptors on dopamine transporters are related to self- administration of cocaine. Science, 1987. 237(4819): p. 1219-23. 264. Izenwasser, S., The role of the dopamine transporter in cocaine abuse. Neurotox Res, 2004. 6(5): p. 379-83. 265. Ritz, M.C., E.J. Cone, and M.J. Kuhar, Cocaine inhibition of ligand binding at dopamine, norepinephrine and serotonin transporters: a structure-activity study. Life Sci, 1990. 46(9): p. 635-45. 266. Giros, B., et al., Hyperlocomotion and indifference to cocaine and amphetamine in mice lacking the dopamine transporter. Nature, 1996. 379(6566): p. 606-12. 267. Rocha, B.A., et al., Cocaine self-administration in dopamine-transporter knockout mice. Nat Neurosci, 1998. 1(2): p. 132-7. 268. Sora, I., et al., Cocaine reward models: conditioned place preference can be established in dopamine- and in serotonin-transporter knockout mice. Proc Natl Acad Sci U S A, 1998. 95(13): p. 7699-704. 269. Medvedev, I.O., et al., Characterization of conditioned place preference to cocaine in congenic dopamine transporter knockout female mice. Psychopharmacology (Berl), 2005. 180(3): p. 408-13. 270. Chen, R., et al., Abolished cocaine reward in mice with a cocaine-insensitive dopamine transporter. Proc Natl Acad Sci U S A, 2006. 103(24): p. 9333-8. 271. Beuming, T., et al., The binding sites for cocaine and dopamine in the dopamine transporter overlap. Nat Neurosci, 2008. 11(7): p. 780-9. 272. Ferrer, J.V. and J.A. Javitch, Cocaine alters the accessibility of endogenous cysteines in putative extracellular and intracellular loops of the human dopamine transporter. Proc Natl Acad Sci U S A, 1998. 95(16): p. 9238-43. 273. Little, K.Y., et al., Cocaine induction of dopamine transporter trafficking to the plasma membrane. Mol Pharmacol, 2002. 61(2): p. 436-45. 274. Sitte, H.H. and M. Freissmuth, Amphetamines, new psychoactive drugs and the monoamine transporter cycle. Trends Pharmacol Sci, 2015. 36(1): p. 41-50. 275. Fischer, J.F. and A.K. Cho, Chemical release of dopamine from striatal homogenates: evidence for an exchange diffusion model. J Pharmacol Exp Ther, 1979. 208(2): p. 203-9. 276. Paton, D.M., Mechanism of efflux of noradrenaline from adrenergic nerves in rabbit atria. Br J Pharmacol, 1973. 49(4): p. 614-27.

130

277. Burnette, W.B., et al., Human norepinephrine transporter kinetics using rotating disk electrode voltammetry. Anal Chem, 1996. 68(17): p. 2932-8. 278. Langeloh, A., H. Bonisch, and U. Trendelenburg, The mechanism of the 3H-noradrenaline releasing effect of various substrates of uptake1: multifactorial induction of outward transport. Naunyn Schmiedebergs Arch Pharmacol, 1987. 336(6): p. 602-10. 279. Sulzer, D. and S. Rayport, Amphetamine and other psychostimulants reduce pH gradients in midbrain dopaminergic neurons and chromaffin granules: a mechanism of action. Neuron, 1990. 5(6): p. 797-808. 280. Sulzer, D., et al., Weak base model of amphetamine action. Ann N Y Acad Sci, 1992. 654: p. 525-8. 281. Sulzer, D., et al., Amphetamine redistributes dopamine from synaptic vesicles to the cytosol and promotes reverse transport. J Neurosci, 1995. 15(5 Pt 2): p. 4102-8. 282. Liu, Y., et al., A cDNA that suppresses MPP+ toxicity encodes a vesicular amine transporter. Cell, 1992. 70(4): p. 539-51. 283. Jones, S.R., et al., Dopamine neuronal transport kinetics and effects of amphetamine. J Neurochem, 1999. 73(6): p. 2406-14. 284. Steinkellner, T., et al., Ca(2+)/calmodulin-dependent protein kinase IIalpha (alphaCaMKII) controls the activity of the dopamine transporter: implications for Angelman syndrome. J Biol Chem, 2012. 287(35): p. 29627-35. 285. Rickhag, M., et al., Membrane-permeable C-terminal dopamine transporter peptides attenuate amphetamine-evoked dopamine release. J Biol Chem, 2013. 288(38): p. 27534- 44. 286. Steinkellner, T., et al., In vivo amphetamine action is contingent on alphaCaMKII. Neuropsychopharmacology, 2014. 39(11): p. 2681-93. 287. Kantor, L., et al., Protein kinase C and intracellular calcium are required for amphetamine-mediated dopamine release via the norepinephrine transporter in undifferentiated PC12 cells. J Pharmacol Exp Ther, 2001. 297(3): p. 1016-24. 288. Johnson, L.A., et al., Regulation of amphetamine-stimulated dopamine efflux by protein kinase C beta. J Biol Chem, 2005. 280(12): p. 10914-9. 289. Giambalvo, C.T., Mechanisms underlying the effects of amphetamine on particulate PKC activity. Synapse, 2004. 51(2): p. 128-39. 290. Chen, R., et al., Protein kinase Cbeta is a critical regulator of dopamine transporter trafficking and regulates the behavioral response to amphetamine in mice. J Pharmacol Exp Ther, 2009. 328(3): p. 912-20. 291. Foster, J.D., et al., Dopamine transporter phosphorylation site threonine 53 regulates substrate reuptake and amphetamine-stimulated efflux. J Biol Chem, 2012. 287(35): p. 29702-12. 292. Kitayama, S., T. Dohi, and G.R. Uhl, Phorbol esters alter functions of the expressed dopamine transporter. Eur J Pharmacol, 1994. 268(2): p. 115-9. 293. Copeland, B.J., et al., Protein kinase C activators decrease dopamine uptake into striatal synaptosomes. J Pharmacol Exp Ther, 1996. 277(3): p. 1527-32. 294. Huff, R.A., et al., Phorbol esters increase dopamine transporter phosphorylation and decrease transport Vmax. J Neurochem, 1997. 68(1): p. 225-32. 295. Zhang, L., L.L. Coffey, and M.E. Reith, Regulation of the functional activity of the human dopamine transporter by protein kinase C. Biochem Pharmacol, 1997. 53(5): p. 677-88.

131

296. Vaughan, R.A., et al., Protein kinase C-mediated phosphorylation and functional regulation of dopamine transporters in striatal synaptosomes. J Biol Chem, 1997. 272(24): p. 15541-6. 297. Zhu, S.J., et al., Activation of protein kinase C inhibits uptake, currents and binding associated with the human dopamine transporter expressed in Xenopus oocytes. J Pharmacol Exp Ther, 1997. 282(3): p. 1358-65. 298. Pristupa, Z.B., et al., Protein kinase-mediated bidirectional trafficking and functional regulation of the human dopamine transporter. Synapse, 1998. 30(1): p. 79-87. 299. Melikian, H.E. and K.M. Buckley, Membrane trafficking regulates the activity of the human dopamine transporter. J Neurosci, 1999. 19(18): p. 7699-710. 300. Saunders, C., et al., Amphetamine-induced loss of human dopamine transporter activity: an internalization-dependent and cocaine-sensitive mechanism. Proc Natl Acad Sci U S A, 2000. 97(12): p. 6850-5. 301. Johnson, L.A., et al., Rapid delivery of the dopamine transporter to the plasmalemmal membrane upon amphetamine stimulation. Neuropharmacology, 2005. 49(6): p. 750-8. 302. Loder, M.K. and H.E. Melikian, The dopamine transporter constitutively internalizes and recycles in a protein kinase C-regulated manner in stably transfected PC12 cell lines. J Biol Chem, 2003. 278(24): p. 22168-74. 303. Sorkina, T., et al., Oligomerization of dopamine transporters visualized in living cells by fluorescence resonance energy transfer microscopy. J Biol Chem, 2003. 278(30): p. 28274-83. 304. Heikkila, R.E., et al., Amphetamine: evaluation of d- and l-isomers as releasing agents and uptake inhibitors for 3H-dopamine and 3H-norepinephrine in slices of rat neostriatum and cerebral cortex. J Pharmacol Exp Ther, 1975. 194(1): p. 47-56. 305. Schmitz, Y., et al., Amphetamine distorts stimulation-dependent dopamine overflow: effects on D2 autoreceptors, transporters, and synaptic vesicle stores. J Neurosci, 2001. 21(16): p. 5916-24. 306. Hamilton, P.J., et al., PIP2 regulates psychostimulant behaviors through its interaction with a membrane protein. Nat Chem Biol, 2014. 10(7): p. 582-589. 307. Magnani, F., et al., Partitioning of the serotonin transporter into lipid microdomains modulates transport of serotonin. J Biol Chem, 2004. 279(37): p. 38770-8. 308. Daberkow, D.P., et al., Amphetamine paradoxically augments exocytotic dopamine release and phasic dopamine signals. J Neurosci, 2013. 33(2): p. 452-63. 309. Siciliano, C.A., et al., Biphasic mechanisms of amphetamine action at the dopamine terminal. J Neurosci, 2014. 34(16): p. 5575-82. 310. Backman, C.M., et al., Characterization of a mouse strain expressing Cre recombinase from the 3' untranslated region of the dopamine transporter locus. Genesis, 2006. 44(8): p. 383-90. 311. Ludwig, A., et al., Flotillin microdomains interact with the cortical cytoskeleton to control uropod formation and neutrophil recruitment. J Cell Biol, 2010. 191(4): p. 771-81. 312. Bitsikas, V., et al., The role of flotillins in regulating abeta production, investigated using flotillin 1-/-, flotillin 2-/- double knockout mice. PLoS One, 2014. 9(1): p. e85217. 313. Rossi, A., et al., Genetic compensation induced by deleterious mutations but not gene knockdowns. Nature, 2015. 524(7564): p. 230-3.

132

314. Foster, J.D. and R.A. Vaughan, Palmitoylation controls dopamine transporter kinetics, degradation, and protein kinase C-dependent regulation. J Biol Chem, 2011. 286(7): p. 5175-86. 315. Cheng, M.H., et al., Insights into the Modulation of Dopamine Transporter Function by Amphetamine, Orphenadrine, and Cocaine Binding. Front Neurol, 2015. 6: p. 134. 316. Levental, I., M. Grzybek, and K. Simons, Greasing their way: lipid modifications determine protein association with membrane rafts. Biochemistry, 2010. 49(30): p. 6305- 16. 317. Foster, J.D., B. Pananusorn, and R.A. Vaughan, Dopamine transporters are phosphorylated on N-terminal serines in rat striatum. J Biol Chem, 2002. 277(28): p. 25178-86. 318. Kahlig, K.M., J.A. Javitch, and A. Galli, Amphetamine regulation of dopamine transport. Combined measurements of transporter currents and transporter imaging support the endocytosis of an active carrier. J Biol Chem, 2004. 279(10): p. 8966-75. 319. Sen, N., et al., A pincer-like configuration of TM2 in the human dopamine transporter is responsible for indirect effects on cocaine binding. Neuropharmacology, 2005. 49(6): p. 780-90. 320. Giambalvo, C.T., Protein kinase C and dopamine transport--2. Effects of amphetamine in vitro. Neuropharmacology, 1992. 31(12): p. 1211-22. 321. Kantor, L. and M.E. Gnegy, Protein kinase C inhibitors block amphetamine-mediated dopamine release in rat striatal slices. J Pharmacol Exp Ther, 1998. 284(2): p. 592-8. 322. Cowell, R.M., et al., Dopamine transporter antagonists block phorbol ester-induced dopamine release and dopamine transporter phosphorylation in striatal synaptosomes. Eur J Pharmacol, 2000. 389(1): p. 59-65. 323. Moritz, A.E., et al., Reciprocal Phosphorylation and Palmitoylation Control Dopamine Transporter Kinetics. J Biol Chem, 2015. 290(48): p. 29095-105. 324. Dowhan, W. and M. Bogdanov, Lipid-protein interactions as determinants of membrane protein structure and function. Biochem Soc Trans, 2011. 39(3): p. 767-74. 325. Dominguez, L., et al., Impact of membrane lipid composition on the structure and stability of the transmembrane domain of amyloid precursor protein. Proc Natl Acad Sci U S A, 2016. 113(36): p. E5281-7. 326. Lee, A.G., How lipids affect the activities of integral membrane proteins. Biochim Biophys Acta, 2004. 1666(1-2): p. 62-87. 327. Tillman, T.S. and M. Cascio, Effects of membrane lipids on ion channel structure and function. Cell Biochem Biophys, 2003. 38(2): p. 161-90. 328. Taverna, E., et al., Localization of synaptic proteins involved in neurosecretion in different membrane microdomains. J Neurochem, 2007. 100(3): p. 664-77. 329. van Rheenen, J., et al., PIP2 signaling in lipid domains: a critical re-evaluation. EMBO J, 2005. 24(9): p. 1664-73. 330. Zimmerberg, J. and M.M. Kozlov, How proteins produce cellular membrane curvature. Nat Rev Mol Cell Biol, 2006. 7(1): p. 9-19. 331. Scanlon, S.M., D.C. Williams, and P. Schloss, Membrane cholesterol modulates serotonin transporter activity. Biochemistry, 2001. 40(35): p. 10507-13. 332. Bialecki, R.A. and T.N. Tulenko, Excess membrane cholesterol alters calcium channels in arterial smooth muscle. Am J Physiol, 1989. 257(2 Pt 1): p. C306-14.

133

333. Gleason, M.M., M.S. Medow, and T.N. Tulenko, Excess membrane cholesterol alters calcium movements, cytosolic calcium levels, and membrane fluidity in arterial smooth muscle cells. Circ Res, 1991. 69(1): p. 216-27. 334. Kutryk, M.J. and G.N. Pierce, Stimulation of sodium-calcium exchange by cholesterol incorporation into isolated cardiac sarcolemmal vesicles. J Biol Chem, 1988. 263(26): p. 13167-72. 335. Kutryk, M.J., et al., Oxidation of membrane cholesterol alters active and passive transsarcolemmal calcium movement. Circ Res, 1991. 68(1): p. 18-26. 336. Epand, R.M., Lipid polymorphism and protein-lipid interactions. Biochim Biophys Acta, 1998. 1376(3): p. 353-68. 337. Perozo, E., et al., Physical principles underlying the transduction of bilayer deformation forces during mechanosensitive channel gating. Nat Struct Biol, 2002. 9(9): p. 696-703. 338. Khelashvili, G., et al., Computational modeling of the N-terminus of the human dopamine transporter and its interaction with PIP2 -containing membranes. Proteins, 2015. 83(5): p. 952-69. 339. Sucic, S., et al., The N terminus of monoamine transporters is a lever required for the action of amphetamines. J Biol Chem, 2010. 285(14): p. 10924-38. 340. Brown, M.F., Modulation of rhodopsin function by properties of the membrane bilayer. Chem Phys Lipids, 1994. 73(1-2): p. 159-80. 341. Gibson, N.J. and M.F. Brown, Membrane lipid influences on the energetics of the metarhodopsin I and metarhodopsin II conformational states of rhodopsin probed by flash photolysis. Photochem Photobiol, 1991. 54(6): p. 985-92. 342. Gibson, N.J. and M.F. Brown, Lipid headgroup and acyl chain composition modulate the MI-MII equilibrium of rhodopsin in recombinant membranes. Biochemistry, 1993. 32(9): p. 2438-54. 343. Semrau, S. and T. Schmidt, Membrane heterogeneity - from lipid domains to curvature effects. Soft Matter, 2009. 5(17): p. 3174-3186. 344. Sorkina, T., J. Caltagarone, and A. Sorkin, Flotillins regulate membrane mobility of the dopamine transporter but are not required for its protein kinase C dependent endocytosis. Traffic, 2013. 14(6): p. 709-24. 345. Jao, C.Y., et al., Bioorthogonal probes for imaging sterols in cells. Chembiochem, 2015. 16(4): p. 611-7. 346. Bodrikov, V., et al., Reggie-1 and reggie-2 (flotillins) participate in Rab11a-dependent cargo trafficking, spine synapse formation and LTP-related AMPA receptor (GluA1) surface exposure in mouse hippocampal neurons. Exp Neurol, 2017. 289: p. 31-45. 347. Harkewicz, R. and E.A. Dennis, Applications of mass spectrometry to lipids and membranes. Annu Rev Biochem, 2011. 80: p. 301-25. 348. Knowland, D., et al., Stepwise recruitment of transcellular and paracellular pathways underlies blood-brain barrier breakdown in stroke. Neuron, 2014. 82(3): p. 603-17. 349. Castanon, N., et al., Modulation of the effects of cocaine by 5-HT1B receptors: a comparison of knockouts and antagonists. Pharmacol Biochem Behav, 2000. 67(3): p. 559- 66. 350. Floresco, S.B., et al., Afferent modulation of dopamine neuron firing differentially regulates tonic and phasic dopamine transmission. Nat Neurosci, 2003. 6(9): p. 968-73. 351. Graybiel, A.M., et al., The basal ganglia and adaptive motor control. Science, 1994. 265(5180): p. 1826-31.

134

352. Cagniard, B., et al., Mice with chronically elevated dopamine exhibit enhanced motivation, but not learning, for a food reward. Neuropsychopharmacology, 2006. 31(7): p. 1362-70. 353. Schultz, W., Predictive reward signal of dopamine neurons. J Neurophysiol, 1998. 80(1): p. 1-27. 354. Redgrave, P. and K. Gurney, The short-latency dopamine signal: a role in discovering novel actions? Nat Rev Neurosci, 2006. 7(12): p. 967-75. 355. Tsai, H.C., et al., Phasic firing in dopaminergic neurons is sufficient for behavioral conditioning. Science, 2009. 324(5930): p. 1080-4. 356. Heien, M.L. and R.M. Wightman, Phasic dopamine signaling during behavior, reward, and disease states. CNS Neurol Disord Drug Targets, 2006. 5(1): p. 99-108. 357. Hazy, T.E., M.J. Frank, and R.C. O'Reilly, Neural mechanisms of acquired phasic dopamine responses in learning. Neurosci Biobehav Rev, 2010. 34(5): p. 701-20. 358. Grace, A.A., Dysregulation of the dopamine system in the pathophysiology of schizophrenia and depression. Nat Rev Neurosci, 2016. 17(8): p. 524-32. 359. Grace, A.A., The tonic/phasic model of dopamine system regulation and its implications for understanding alcohol and psychostimulant craving. Addiction, 2000. 95 Suppl 2: p. S119-28. 360. Grace, A.A., The tonic/phasic model of dopamine system regulation: its relevance for understanding how stimulant abuse can alter basal ganglia function. Drug Alcohol Depend, 1995. 37(2): p. 111-29. 361. Kalivas, P.W. and J. Stewart, Dopamine transmission in the initiation and expression of drug- and stress-induced sensitization of motor activity. Brain Res Brain Res Rev, 1991. 16(3): p. 223-44. 362. Robinson, T.E. and J.B. Becker, Enduring changes in brain and behavior produced by chronic amphetamine administration: a review and evaluation of animal models of amphetamine psychosis. Brain Res, 1986. 396(2): p. 157-98. 363. Robinson, T.E. and K.C. Berridge, The neural basis of drug craving: an incentive- sensitization theory of addiction. Brain Res Brain Res Rev, 1993. 18(3): p. 247-91. 364. Robinson, T.E., et al., Persistent sensitization of dopamine neurotransmission in ventral striatum (nucleus accumbens) produced by prior experience with (+)-amphetamine: a microdialysis study in freely moving rats. Brain Res, 1988. 462(2): p. 211-22. 365. Wise, R.A. and M.A. Bozarth, A psychomotor stimulant theory of addiction. Psychol Rev, 1987. 94(4): p. 469-92. 366. Bowyer, J.F. and J.P. Hanig, Amphetamine- and methamphetamine-induced hyperthermia: Implications of the effects produced in brain vasculature and peripheral organs to forebrain neurotoxicity. Temperature (Austin), 2014. 1(3): p. 172-82. 367. Carvalho, M., et al., Toxicity of amphetamines: an update. Arch Toxicol, 2012. 86(8): p. 1167-231. 368. Sulzer, D. and L. Zecca, Intraneuronal dopamine-quinone synthesis: a review. Neurotox Res, 2000. 1(3): p. 181-95. 369. Miyazaki, I. and M. Asanuma, Dopaminergic neuron-specific oxidative stress caused by dopamine itself. Acta Med Okayama, 2008. 62(3): p. 141-50. 370. Kim, H., et al., Immunohistochemical study of flotillin-1 in the spinal cord of Lewis rats with experimental autoimmune encephalomyelitis. Brain Res, 2006. 1114(1): p. 204-11. 371. Rajendran, L., et al., Increased Abeta production leads to intracellular accumulation of Abeta in flotillin-1-positive endosomes. Neurodegener Dis, 2007. 4(2-3): p. 164-70.

135

372. Schneider, A., et al., Flotillin-dependent clustering of the amyloid precursor protein regulates its endocytosis and amyloidogenic processing in neurons. J Neurosci, 2008. 28(11): p. 2874-82. 373. Chen, T.Y., et al., The intracellular domain of amyloid precursor protein interacts with flotillin-1, a lipid raft protein. Biochem Biophys Res Commun, 2006. 342(1): p. 266-72. 374. Robbins, A.K. and R.A. Horlick, Macrophage scavenger receptor confers an adherent phenotype to cells in culture. Biotechniques, 1998. 25(2): p. 240-4. 375. Tang, S.H., et al., A Cre/loxP-deleter transgenic line in mouse strain 129S1/SvImJ. Genesis, 2002. 32(3): p. 199-202. 376. Hayashi, S. and A.P. McMahon, Efficient recombination in diverse tissues by a tamoxifen- inducible form of Cre: a tool for temporally regulated gene activation/inactivation in the mouse. Dev Biol, 2002. 244(2): p. 305-18. 377. Aguilar, J.I., et al., Neuronal Depolarization Drives Increased Dopamine Synaptic Vesicle Loading via VGLUT. Neuron, 2017. 95(5): p. 1074-1088 e7. 378. Dadalko, O.I., et al., mTORC2/rictor signaling disrupts dopamine-dependent behaviors via defects in striatal dopamine neurotransmission. J Neurosci, 2015. 35(23): p. 8843-54. 379. Hoffman, A.F. and G.A. Gerhardt, Differences in pharmacological properties of dopamine release between the substantia nigra and striatum: an in vivo electrochemical study. J Pharmacol Exp Ther, 1999. 289(1): p. 455-63. 380. Gerhardt, G.A. and A.F. Hoffman, Effects of recording media composition on the responses of Nafion-coated carbon fiber microelectrodes measured using high-speed chronoamperometry. J Neurosci Methods, 2001. 109(1): p. 13-21. 381. Yamamoto, A., J.J. Lucas, and R. Hen, Reversal of neuropathology and motor dysfunction in a conditional model of Huntington's disease. Cell, 2000. 101(1): p. 57-66. 382. Dragich, J.M., et al., Autophagy linked FYVE (Alfy/WDFY3) is required for establishing neuronal connectivity in the mammalian brain. Elife, 2016. 5. 383. Robertson, S.D., et al., Insulin reveals Akt signaling as a novel regulator of norepinephrine transporter trafficking and norepinephrine homeostasis. J Neurosci, 2010. 30(34): p. 11305-16. 384. Eenjes, E., et al., Distinguishing aggregate formation and aggregate clearance using cell- based assays. J Cell Sci, 2016. 129(6): p. 1260-70.

136

Appendix A

Flot1 does not scaffold all proteins to DRMs

Appendix A. Sucrose density gradients reveal that Flot1 is a specific scaffold. Cells were transfected with siCTRL or siFlot1, lysed, and fractionated across a discontinuous sucrose gradient. Fractions were collected from top (2) to bottom (10) of the gradient. Four different conditions were run: (A) EM4-YFP-DAT cells lysed with 1% Brij 58; (B) EM4-YFP-DAT cells lysed with 1% Triton X-100; (C) HeLa cells lysed with 1% Brij 58; and (D) HeLa cells lysed with 1% Tx-100 (n = 2).

It has previously been shown that Flot1 is required for the localization of DAT into DRMs in EM4 cells stably expressing YFP-DAT (EM4-YFP-DAT) [105]. Endogenous DAT also localizes to similar DRMs in a Flot1-dependent manner (Figure 4.10). Flot1 may either act as a general scaffold for a specific membrane composition, or as a membrane raft adaptor for select proteins such as DAT. To test this, Flot1 was depleted in EM4 and HeLa cells and a panel of raft-associated proteins was examined. Based on the results, it appears that cell-type and detergent influence the

137 apparent distribution of proteins; DRMs can be found at the plasma membrane or on intracellular organelles; and many of the raft proteins continue to distribute to buoyant fractions in the absence of Flot1. These findings suggest that Flot1 does not act as a general scaffold.

138

Appendix B

Cholesterol affects the DRM localization of LC3-II

Appendix B. The DRM localization of LC3-II depends on cholesterol. HeLa cells treated with vehicle or 20 mM MβC were lysed with 1% Brij58 and fractionated across a discontinuous sucrose gradient. Fractions were collected from top (3) to bottom (10) of the gradient. Src was used to show that not all proteins are affected by MβC treatment.

Macroautophagy, the most evolutionarily conserved form of autophagy, is a process by which cytosolic components are degraded by the lysosome. The capture of substrates requires the formation of the autophagosome, a double membrane structure that engulfs its cargo and fuses into the endolysosomal system for degradation. Autophagosomes undergo nucleation, expansion, and maturation, a process that requires a group of highly conserved autophagy-related proteins (Atgs).

LC3, the mammalian homologue of Atg8, participates in a series of enzymatic reactions similar to ubiquitin conjugation to attach itself to the lipid phosphatidylethanolamine (PE), which is important for the growth of the autophagosomal membrane. LC3 can be biochemically visualized 139 in two forms – the cytosolic LC3-I form and the membrane-bound lipidated form (LC3-II).

Although LC3-II is often used as a marker for autophagosomes, it is unknown what exactly LC3 is doing on the autophagosome membrane and why it is there. Our data suggests that a subset of

LC3-II molecules is present in cholesterol-rich membranes, as defined by sucrose density gradient centrifugation in the presence or absence of MβC. This is the first demonstration to our knowledge that LC3-II has been linked to cholesterol-enriched membranes, perhaps opening up questions as to what type of membranes LC3 is conjugating to, whether it requires DRMs for conjugation, and what is the biological significance of LC3-II.

140

Appendix C

Atg3 and its role in autophagosome biogenesis

Supplementary Figure 6 Atg3 overexpression cannot rescue Atg5 deficiency. B) Atg3 overexpression by transient transfection cannot rescue lipidation in Atg5 KO cells. Atg5 KO MEFs were transfected with mock, Atg3 WT or Atg3 K11W constructs, starved for 2 hours in 1X HBSS containing 10 mM HEPES and collected for Immunoblot analysis. Autophagosome maturation was impeded using the lysosome inhibitor chloroquine (10 µM). C) Exogenously expressed Atg3 binds to endogenous Atg5 complexes. Overexpressed wildtype Atg3 with a COOH terminal V5 tag can pull down endogenous Atg5 and Atg16. FT= flow through, T= total, IP = immunoprecipiate. Taken from Nath et al. (2014)

I contributed Supplemental Figure 6B and C to a Nature Cell Biology paper describing the lipidation of LC3 by Atg3 (Nath et al., 2014), on which I am an author. The formation of the autophagosome requires many proteins and in its final stage, involves LC3. The proteins involved in the growing autophagosome are likely different from those on the mature autophagosome, perhaps delineated by the curved rim of the growing membrane. This paper shows that the E2-like enzyme, Atg3, uses its curvature-sensing domain to facilitate LC3 lipidation. Because most of the experiments in this paper were done in vitro, it was difficult to ascribed physiologic significance, so we performed experiments in cells to this extent. The Atg16/Atg5-Atg12 complex serves as a

141 putative E3 ligase to facilitate Atg3 lipidation by recruiting it to the correct membrane. Our cell- based experiments aimed to examine how Atg3 helix mutants can function in the absence of the

Atg16/Atg5-Atg12 complex to help target it to the membrane, presumably to the growing curved autophagosome membrane. The Atg3 mutants led to no LC3-II in Atg5 KO MEFs, suggesting that it requires the Atg16/Atg5-Atg12 complex for membrane targeting.

142