Neto1 and Neto2 Are Auxiliary Subunits of Synaptic Kainate Receptors

Neto1 and Neto2 are Auxiliary Subunits of Synaptic Kainate Receptors

Man Tang

A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy Graduate Department of Molecular Genetics University of Toronto

Neto1 and Neto2 are Auxiliary Subunits of Synaptic Kainate Receptors

Man Tang Doctor of Philosophy Graduate Department of Molecular Genetics University of Toronto 2013

Abstract

Neto1 and Neto2 are CUB domain-containing transmembrane proteins that are expressed in the mammalian brain. Previous studies showed that Neto1 is a NMDAR-associated protein with important roles in synaptic plasticity and learning/memory (Ng et al., 2009). To establish the functions of Neto2, I first searched for its binding partners. Using yeast two-hybrid analysis,

GST pull-down and co-immunoprecipitation studies, I found that Neto2 can bind to the PDZ domain-containing protein GRIP. In the brain, GRIP regulates the synaptic trafficking and stability of AMPA and kainate receptors (KARs) (Hirbec et al., 2003). To determine whether

Neto2 is required for the synaptic expression of KARs and/or AMPARs, I examined whether

Neto2 was associated with these receptors at the postsynaptic membrane.

Coimmunoprecipitation studies showed that while Neto2 is a component of postsynaptic KAR protein complexes, it is not associated with AMPARs. In the cerebellum, Neto2-null mice showed a 44% (n=3;p<0.01) decrease in the abundance of postsynaptic KARs with no change in the level of total KARs, thus suggesting a specific deficit in KAR synaptic localization.

Unexpectedly, loss of Neto2 had no effect on the abundance of hippocampal KARs (n=3; p>0.05), or on neurotransmission by them (n=12; p>0.05). To determine whether this normal

KAR function might be due to compensation by Neto1, which also interacts with KARs, I

examined KAR abundance in Neto1-null, and Neto1/2-double null hippocampus. Loss of Neto1 resulted in a 53% decrease in postsynaptic levels of GluK2-KARs (n=3;p<0.01). However, in double null animals, the reduction was indistinguishable from Neto1 single null mice, suggesting that Neto2 is not involved in the postsynaptic localization of hippocampal KARs. In

Neto1-null mice, KAR-mediated currents showed smaller amplitude (61% of wild- type;n=14;p<0.01), and faster decay kinetics (40% of wild-type;n=14;p<0.001). Together, these findings establish both Neto1 and Neto2 as auxiliary proteins of native KARs: Neto1 regulates the synaptic abundance and kinetics of KARs in the hippocampus, while Neto2 mediates the synaptic localization of cerebellar KARs. Additionally, the results presented here, in conjunction with previous findings, reveal a unique ability of Neto1 in controlling synaptic transmission by serving as an auxiliary protein for two different classes of ionotropic glutamate receptors.

iii

Acknowledgements

Being part of the McInnes lab has been a truly fantastic experience for me. Not only did it open my eyes to a wonderful and challenging field of science, but it allowed me to work in a collaborative environment with many intelligent, kind, and fun people. As I approach the end of my graduate journey, I would like to thank all of the people who have helped me along the way.

I would not be completing my dissertation today without your guidance, support, and friendship.

First, and foremost, my deepest gratitude goes to my wonderful supervisor Roderick McInnes.

Rod is a fabulous mentor who has expertly guided me through my graduate education. His unwavering passion and enthusiasm for science have been an inspiration for the career I have chosen to follow. Thank you Rod for taking me on as an international student, and giving me the opportunity to work with you; for being a supportive advisor who continuously helped and challenged me to become a better scientist, writer, and public speaker; for believing in me and genuinely caring about my career development, and for always making sure I had a place to be during the holiday seasons. I would also like to thank my supervisory committee members Dr.

Mike Salter, and Dr. Sabine Cordes for their advice and encouragement. I feel very fortunate to have you both in my committee!! In the last two years of my PhD, Dr. Salter has also taken on a co-supervisory role, and has provided invaluable input and established critical collaborations that made the Neto studies possible.

Living away from home as a grad student, the McInnes lab has become my second family. I’d like to first thank all the past and present members of the super-awesome Neto team,

Rachel Szilard, Dave Ng, Zhenya Ivakine, Jeff Gingrich, and Vivek Mahadevan. Rachel is the most helpful, patient, and knowledgeable biochemistry “lab consultant” I’ve ever met. Her

guidance (and world’s most thorough protocols!) has been instrumental in getting the project moving during my first two years in the lab. I thank you for your continued friendship, support, and advice! Dave, the “father of Netos” has been a wonderful resource of all knowledge on

Netos. I truly enjoyed the many hours we spent brainstorming ideas and discussing new avenues to take. Zhenya has been a most wonderful mentor, scientific advisor, “archenemy”, and partner in crime. Thank you for teaching me so much about science, and about life, and for giving me the confidence to pursue my dreams. I feel truly blessed to have your friendship!

Vivek was the last member to join the Neto team but his immense enthusiasm for neuroscience research is unrivalled, and his perseverance and positive attitude, admirable. Thank you for bringing so much joy, and laughter into the lab, and for always cheering me on! I’d also like to thank Jeff, our lab’s “walking encyclopedia” for sharing his expertise and knowledge in neuroscience. I’m grateful to Irene Chau, Cynthia Jung, Coco Jiang, and Alexa Bramall for their friendship, to Lynda Ploder for her support on the project, and Dorothy Carlin for helping me set up all my meetings.

I am also extremely grateful for my phenomenal collaborators, Dr. Ken Pelkey and Dr.

Chris McBain. Ken has done a magnificent job on all the electrophysiology experiments needed for this project, and has generously devoted his time to review and provide feedback on my manuscripts. Our first paper would not have been possible without his timely contribution.

Finally, I’d like to thank my family for teaching me the value of hard-work, and the importance of integrity; and for supporting me throughout this very long journey.

Table of Contents

ABSTRACT……………………………………………………………………………………………………………………………………………….ii

ACKNOWLEDGEMENTS……………………………………………………………………………………………………………….…………iv

LIST OF TABLES………..……………………………………………………………………………….……………………………………………ix

LIST OF FIGURES…………………………….………………………………………………………………………………………………………ix

LIST OF APPENDICES………………………………………………………………………………………………………………………………xi

FREQUENTLY USED ABBREVIATIONS………………………………………………………………………………………….………….xii

Chapter 1: Introduction ...... 1 1.1. Introduction to the mammalian central nervous system ...... 2 1.1.1. The hippocampus structure ...... 2 1.1.2. Basic hippocampal neural pathways ...... 6 1.1.3. The cellular organization of the cerebellum ...... 9 1.1.4. The basic cerebellar circuitry ...... 12 1.2. Neuronal communication at chemical synapses ...... 15 1.2.1. Structure of a chemical synapse ...... 16 1.2.2. Chemical synaptic transmission in the brain ...... 17 1.2.3. Ionotropic glutamate receptors ...... 20 1.3. Kainate Receptors ...... 22 1.3.1. KAR subunits: general structure and biophysical properties ...... 23 1.3.2. KAR pharmacology ...... 26 1.3.3. KAR expression, protein distribution and trafficking ...... 27 1.3.4. Neuronal function of KARs ...... 33 1.3.5. KAR interacting proteins ...... 38 1.3.6. KARs and disease ...... 40 1.4. The Neto family of transmembrane proteins ...... 41 1.4.1. Domain structure and organization ...... 42 1.4.2. Expression of Neto1 and Neto2 in the CNS ...... 45 1.4.3. Function of Neto proteins in the nervous system ...... 48 Thesis Objectives ...... 50

Chapter 2: Neto2 is a CUB domain protein that regulates the synaptic abundance of cerebellar KARs ...... 52 2.1. Introduction ...... 53 2.2. Materials and Methods ...... 56 2.3. Results ...... 71 2.3.1. Identification of Neto2 intracellular interacting proteins from adult mouse brain ...... 71 2.3.2. Neto2 binds to GRIP through PDZ ligand:PDZ domain interactions ...... 78 2.3.3. Neto2 interacts with KARs but not AMPARs ...... 81 2.3.4. Neto2 associates with GluK2-KARs predominantly through the second CUB domain ...... 84 2.3.5. Neto2 forms a ternary complex with GluK2-KARs and GRIP ...... 88 2.3.6. Loss of Neto2 does not alter the synaptic abundance of KARs in the hippocampus ...... 89 2.3.7. KAR synaptic transmission at MF-CA3 synapses is normal in Neto2-null mice ...... 91 2.3.8. Synaptic abundance of KARs is reduced in the cerebellum of Neto2-null mice ...... 94 2.4. Discussion ...... 100 Chapter 3: Neto1 is an auxiliary subunit of native synaptic kainate receptors ...... 107 3.1. Introduction ...... 108 3.2. Materials and Methods ...... 111 3.3. Results ...... 121 3.3.1. Neto1 interacts with native KARs ...... 121 3.3.2. Synaptic KAR currents are reduced in Neto1-null mice ...... 126 3.3.3. Loss of Neto1 affects NMDAR-mediated currents at A/C-CA3 but not MF-CA3 synapses 131 3.3.4. Neto1-null mice have normal presynaptic function at MF-CA3 synapses ...... 133 3.3.5. Neto1 is required for the synaptic abundance of hippocampal KARs ...... 134 3.3.6. Neto1 binds to the synaptic scaffolding protein PICK1 ...... 137 3.4. Discussion ...... 141 Chapter 4: Future directions ...... 147 4.1. Final discussion and future directions ...... 148 4.1.1. Additional studies on the role of Netos on KAR synaptic physiology ...... 152 4.1.2. Characterization of KAR synaptic localization defects...... 158 4.1.3. Systematic analysis of the modulation of KAR biophysical properties by Neto1/2 ...... 160 4.1.4. Behavioural studies on Neto1- and Neto2-null mice ...... 162 4.1.5. Additional studies on the regulation of synaptic NMDARs by Neto1 ...... 164 vii

Appendix A: Putative Neto2 interacting molecules identified by a yeast two-hybrid screen of an adult mouse brain cDNA library ...... 179

Appendix B: Proteins present in the GST-Neto2cyto pull down of adult mouse brain membrane fraction as detected by mass spectrometry ...... 179 Appendix C: Neto2 is associated with NMDARs in vivo ...... 180

viii

List of Tables

Table 1.1. Kainate receptor gene expression in the rat brain……………………………………29

List of Figures

Figure 1.1. The structure of the hippocampus………………………………………………5

Figure 1.2. Major input and output pathways of the hippocampus……………………….8

Figure 1.3. The structure of the cerebellum…………………………………………………11

Figure 1.4. Basic circuitry of the cerebellar cortex………………………………………….14

Figure 1.5. KAR subunit topology and conformational change upon ligand binding……25

Figure 1.6. Expression and subcellular localization of KARs in hippocampal neurons…32

Figure 1.7. Pre- and postsynaptic function of KARs………………………………………..37

Figure 1.8. Neto1, Neto2, and related CUB domain proteins in mouse, C. elegans, and Drosophila………………………………………………………………………………………44

Figure 1.9. Neto1 and Neto2 expression in the mature brain………………………..………….47

Figure 2.1. Neto2 protein is distributed ubiquitously in the adult mouse brain…………..74

Figure 2.2. The cytoplasmic domain of Neto2 interacts with PDZ 4-7 of GRIP in the yeast two- hybrid system……………………………………………………………………………...77

Figure 2.3. Neto2 binds to the scaffolding protein GRIP…………………………………...80

Figure 2.4. Neto2 is associated with KARs, but not AMPARs in vivo……………………..83

Figure 2.5. Neto2 associates with GluK2 and GRIP(PDZ4-7) in a ternary complex………...…...85

Figure 2.6. Neto2 binds to GluK2 KARs through extracellular CUB domains…………...…...87

Figure 2.7. Neto2-null mice have normal levels of GluK2 and GluK5 KARs in hippocampal PSDs……………………………………………………………………………………..90

Figure 2.8. KAR synaptic transmission at MF-CA3 is normal in Neto2-null mice……….93 ix

Figure 2.9. Neto2 is localized in the cerebellar granule cell layer and associates with KARs in the cerebellum……………………………………………………………………………….96

Figure 2.10. Neto2-null mice have reduced GluK2 KAR subunits at cerebellar PSDs…...99

Figure 3.1. Neto1 associates with both KARs and NMDARs in vivo………………...………123

Figure 3.2. Neto1 binds to GluK2 KARs through extracellular CUB domains………….…...125

Figure 3.3. Neto1 is localized to the stratum lucidum……………………………………127

Figure 3.4. KAR synaptic transmission at MF-CA3 is impaired in Neto1-null, and Neto1/Neto2 double-null mice………………………………………………………………………….……130

Figure 3.5. Reduced NMDAR-mediated transmission at A/C-CA3 synapses……….……….132

Figure 3.6. KAR subunits are reduced in hippocampal PSDs of Neto1-null mice……..…….136

Figure 3.7. Neto1 associates with PICK1……………………………………………………138

Figure 3.8. Neto1, PICK1, and GluK2 are associated in a ternary complex………………….140

List of Appendices

Appendix A: Putative Neto2 interacting molecules identified from an adult mouse brain cDNA library by a yeast two-hybrid screen……………………………………………………….…..179

Appendix B: Proteins present in the GST-Neto2cyto pull down fraction as detected by mass spectrometry……………………………………………………………..…………………….179

Appendix C: Neto2 is associated with NMDARs in vivo………………………..…………………….180

Frequently used Abbreviations

A/C-CA3 Associational/Commissural – CA3 pyramidal cell synapses

AMPAR α-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid receptor

CA1 Cornu ammonis region 1

CA3 Cornu ammonis region 3

CNS Central nervous system

CUB Domain found in the proteins Clr, Cls/Uegf/BMP1

DG Dentate gyrus

EPSC Excitatory postsynaptic current

GCL Granule cell layer

GRIP Glutamate receptor-interacting protein

KAR Kainate receptor

LDLa low-density lipoprotein receptor class A

LTP Long-term potentiation

MCL Molecular cell layer

MF-CA3 Mossy fiber - CA3 pyramidal cell synapses

Neto1 Neuropilin and Tolloid-like 1

Neto2 Neuropilin and Tolloid-like 2

NMDAR N-methyl-D-aspartic acid receptor

PCL Purkinje cell layer

PDZ Domain found in PSD-95/Discs-large/ZO-1

PICK1 protein interacting with C kinase-1

xii

PSD Postsynaptic density

SC-CA1 Schaffer collateral – CA1 pyramidal cell synapse

xiii

Chapter 1: Introduction

1.1. Introduction to the mammalian central nervous system

The mammalian central nervous system (CNS) is composed of the spinal cord and the brain, which is undoubtedly the most complex organ in nature. The brain can be subdivided into several anatomically distinct areas: the medulla, pons, midbrain, diencephalon, cerebellum, and telencephalon. The telencephalon, also referred to as the cerebral hemispheres, is the largest region of the brain. It is involved in perceptual, motor, and cognitive functions. The telencephalon is composed of a thin outer layer called the cortex, underlying white matter, and three subcortical structures (basal ganglia, amygdala, and hippocampal formation). Each brain region is different from another in terms of the number and types of neurons it is composed of, as well as the different ways that these neurons are connected with each other. Of particular relevance to this thesis are the two brain regions where the kainate-type of glutamate ion channels have the most abundant expression: the hippocampus and the cerebellum. In the CNS, the hippocampus is responsible for the formation of long-term memories, whereas the cerebellum is involved in motor control. The following sections will provide an overview of the structure and neuronal connections within the hippocampus and cerebellum.

1.1.1. The hippocampus structure

The hippocampal formation is a neural structure in the medial temporal lobe of the brain composed of the hippocampus proper, the dentate gyrus, and the subiculum.

The hippocampus proper, cut in cross section, is a C-shaped structure that resembles a ram’s horn. It is, therefore, also referred to as cornu ammonis (CA), which means Ammon’s 2

horn (Ammon is the egyptian deity, who has the head of a ram). The hippocampus proper is comprised mostly of pyramidal neurons, the major excitatory neurons of the hippocampus. It is divided into three subfields or regions along its curved structure: CA1, CA2, and CA3 (Figure

1.1). All the subfields, in turn, contain a number of layers or strata: 1) the alveus, which is the deepest layer, contains the axons of pyramidal cells; 2) the stratum oriens, which contains the basal dendrites of pyramidal neurons, as well as the cell bodies of inhibitory basket cells; 3) the stratum pyramidale, where the cell bodies of pyramidal neurons are found; 4) the stratum lucidum, the thinnest hippocampal layer present only in the CA3 subfield. It receives input from mossy fibers of dentate gyrus granule cells; 5) the stratum radiatum, which has the proximal segments of the pyramidal cell apical dendrites that connect with Schaffer collateral fibers, the axon projections from CA3 pyramidal neurons to the CA1; and 6) the stratum lacunosum/moleculare, which contains the distal segments of the pyramidal cell apical dendrites that connect with perforant path fibers from the entorhinal cortex (Kandel et al., 2000; Paxinos,

2004; Andersen et al., 2007).

The dentate gyrus is one of the few neural structures with high rates of neurogenesis in the mature brain (Cameron and McKay, 2001). It is comprised of tightly packed small granule cells wrapped around the end of the hippocampus proper, and has three layers: 1) the stratum moleculare, which contains the apical dendrites of granule cells and incoming axons that synapse with them; 2) the statum granulosum, comprised of the cell bodies of densely packed granule cells. These cells are the principal excitatory neurons of the dentate gyrus, and they project to inhibitory neurons and pyramidal cells; 3) the polymorphic layer, which contains the initial segments of the granule cell axons as they bundle together to form the so-called mossy

fiber. It is the most superficial layer of the dentate gyrus, and is also referred to as the hilus

(Kandel et al., 2000; Paxinos, 2004; Andersen et al., 2007).

The subiculum is located between the entorhinal cortex and the CA1 subfield of the hippocampus proper. It receives input from the CA1, and serves as the main output of the hippocampus (Kandel et al., 2000).

1.1.2. Basic hippocampal neural pathways

The different types of hippocampal neurons are connected through a relatively simple and well-characterized neural circuitry (Figure 1.2). First, information from the visual, auditory, and somatic associative cortexes arrive at the parahippocampal region of the cortex, then passes through the entorhinal cortex (EC) and on to the hippocampus. Information enters the hippocampus via axons of layer II and III neurons of the EC. Axons from layer II neurons project on to granule cells of the dentate gyrus and to the most distal dendrites of CA3 pyramidal neurons. These axons are referred to as the perforant pathway, as they “perforate” the subiculum to reach the hippocampus. The layer III axons, called the alvear fibers, synapse directly with CA1 pyramidal neurons (Amaral and Witter, 1989; Baudry and Thompson, 1993;

Amaral and Witter, 1995; Bear et al., 2001).

Granule cells of the dentate gyrus, which receives input from the EC, project axons through the mossy fiber pathway. In the CA3 stratum lucidum, mossy fibers form giant presynaptic terminals on to large complex spines (thorny excrescences) present on proximal dendrites of CA3 pyramidal neurons (Bear et al., 2001; Andersen et al., 2007). The mossy fiber to CA3 (MF-CA3) synapses have been used extensively to study the properties and function of native kainate receptors. Postsynaptic kainate receptors at MF-CA3 synapses have been shown to contribute to membrane depolarization and to have slow decay kinetics (Castillo et al., 1997;

Vignes and Collingridge, 1997), while presynaptic kainate receptors play a significant role in synaptic plasticity (Bortolotto et al., 1999; Lauri et al., 2001b).

The axons of CA3 neurons divide into two branches, one projecting to other CA3 neurons through the recurrent commissural/associational pathway, while the other synapses with

apical dendrites of CA1 neurons by way of the Schaffer collateral pathway. Pyramidal cells of the CA1 then send their axons to the subiculum, and deep layers of the EC; this pathway constitutes the principal output from the hippocampus back to the EC, and completes the so- called trisynaptic circuit – EC to dentate gyrus to CA3 to CA1 (Amaral and Witter, 1989, 1995;

Kandel et al., 2000; Andersen et al., 2007; Daumas et al., 2009).

In addition to pyramidal and granule cells, the hippocampus has a small number of morphologically and physiologically diverse interneurons. These inhibitory neurons are usually involved in local circuitry only, but they play a crucial role in regulating the neuronal activity and complex interactions of large number of excitatory neurons (Freund and Buzsaki, 1996).

1.1.3. The cellular organization of the cerebellum

The cerebellum constitutes only 10% of the total volume of the brain, but contains more than half of all its neurons. It is composed of an outer mantle of gray matter (cerebellar cortex), internal white matter, and the deep cerebellar nuclei. The internal white matter is mostly composed of myelinated nerve fibers that carry information in and out of the cerebellum. The deep cerebellar nuclei are clusters of gray matter organized into a branched, tree-like structure embedded within the white matter. The deep nuclei constitute, with the minor exception of the vestibular nuclei, the sole sources of output from the cerebellum. The cerebellar cortex is a simple three-layered structure containing only five types of neurons: the inhibitory stellate, basket, Purkinje, and Golgi cells, and the excitatory granule cells (Kandel et al., 2000).

The three layers of the cerebellar cortex (from inner to outer layer) are the granular layer, the Purkinje layer, and the molecular layer (Figure 1.3) (Ramnani, 2006; Tanaka et al., 2008).

The granular layer contains the cell bodies of mainly two types of cells: the small and densely packed granule cells, and the much larger, but fewer in number Golgi interneurons. Granule cells are among the smallest in the brain (~5 μm cell body diameter), but amount to more than half of all neurons in the mammalian CNS (Andersen et al., 1992). A granule cell emits only four to five small dendrites, each of which ends in an enlargement called a dendritic claw.

These enlargements receive inhibitory input from Golgi interneurons, and excitatory input from mossy fibers (not to be confused with the hippocampal mossy fibers) within large synaptic complexes called the cerebellar glomeruli. Granule cells extend thin, unmyelinated, and slowly- conducting axons called parallel fibers into the molecular layer (Eccles et al., 1967; Heck, 1993).

The parallel fibers are the only excitatory fibers within the cerebellar cortex, and they synapse with the dendrites of Purkinje cells (Kandel et al., 2000). 9

The Purkinje layer contains only a single layer of large Purkinje cell bodies (50-80 μm).

The dendrites of Purkinje cells are large fan-like arbors of hundreds of spiny branches that reach up into the superficial molecular layer. The axons of Purkinje cells project into the white matter, and form inhibitory synapses with neurons of the deep cerebellar nuclei, or of the vestibular nuclei (Kandel et al., 2000).

The molecular layer primarily contains the dense dendritic arbors of Purkinje cells, the parallel fiber tracts of granule cells, and two types of inhibitory neurons, the stellate and basket cells. Within this layer, each Purkinje cell receives hundreds of thousands of excitatory synaptic inputs (Napper and Harvey, 1988) from the parallel fibers, which run perpendicular to the

Purkinje cell dendritic arbor. Additionally, it receives excitatory input from climbing fibers originating from the inferior olive nucleus. Each Purkinje cell only receives input from one climbing fiber, but as this single fiber “climbs” around the soma and proximal dendrites of the

Purkinje cells, it ends up making a large of number of synapses.

1.1.4. The basic cerebellar circuitry

The cerebellum receives two main types of afferent inputs: the climbing fibers, and the mossy fibers (Figure 1.4). The climbing fibers are the axons of neurons located in the inferior olive (IO), and convey somatosensory, visual or cerebral cortical information. As climbing fibers enter the cerebellum, they split into two branches, one that innervates neurons of the deep nuclei, and one that projects into the molecular layer and wraps around the cell bodies and dendritic arbor of Purkinje cells. The mossy fibers, on the other hand, are axons originating from neurons of the brain stem, pontine, and spinal cord, and thus carry information from the periphery and the cerebral cortex. Mossy fibers form excitatory synapses with the claw-like dendrites of granule cells within the granular layer. The granule cells, in turn, extend long parallel fibers into the molecular layer, and synapse with Purkinje cell dendrites. The Purkinje cells, therefore, receive excitatory input from two afferent fiber systems: directly through the climbing fibers, and indirectly through the mossy fibers (Herrup and Kuemerle, 1997; Kandel et al., 2000). In addition to excitatory inputs, Purkinje cells also receive inhibitory inputs from nearby stellate and basket cells, both of which are facilitated by parallel fibers of granule neurons. The activity of Purkinje cells can be further inhibited indirectly by Golgi interneurons.

As described earlier, granule cells receive inhibitory inputs from Golgi cells, and excitatory inputs from mossy fibers. The function of Golgi cells is to suppress the excitatory action of mossy fibers on granule cells, thereby reducing the overall output of granule cell axons (parallel fibers) onto Purkinje cells (Kandel et al., 2000; Evans, 2007).

The axons of Purkinje cells constitute the sole efferent outputs from the cerebellar cortex.

These cells project predominantly to the deep cerebellar nuclei and make inhibitory synaptic

connections. The deep cerebellar nuclei, in turn, target a diverse range of structures outside of the cerebellum (Herrup and Kuemerle, 1997).

1.2. Neuronal communication at chemical synapses

The mammalian brain comprises a vast network of neurons that communicate with each other through specialized cell junctions called synapses. It has been estimated that the human brain has about 1011 neurons, each one making an average of 1,000 – 10,000 synaptic connections with other neurons (Kandel et al., 2000). Synapses are critical for neuronal communication, as they allow signals to be propagated from one cell to another with high speed and spatial precision.

All neurons make use of one of two basic forms of synaptic transmission: electrical or chemical. Electrical synapses are special gap-junction channels that structurally connect two adjacent cells. Each channel is actually formed by two hemi-channels, one in each apposite cell, that match up in the gap-junction through homophilic interactions. The channels create a continuous bridge between the cytoplasm of the two cells, thus permitting a rapid propagation of electrical signals through the direct exchange of ions (Connors and Long, 2004). Chemical synapses, on the other hand, do not have structural continuity between the pre- and post-synaptic neurons. As a result, the propagation of electrical signals between two cells relies on chemical agents called neurotransmitters. During a chemical synaptic transmission, an electrical signal arriving at the axon terminal of the presynaptic cell leads to the release of neurotransmitters.

The transmitter molecules diffuse across the synaptic cleft – the region separating the pre- and post-synaptic cell- and bind to specific receptors on the postsynaptic cell membrane. This in turn activates the receptors and triggers a cellular response that converts the chemical message back into an electrical signal (Purves, 2008). The following sections will describe in more detail the structure and function of the chemical synapse.

1.2.1. Structure of a chemical synapse

Most of the synapses in the brain are chemical synapses (Greengard, 2001). These synapses are referred to as excitatory, if synaptic activity drives the postsynaptic neuron above its firing threshold, and inhibitory if activity drives the cell below the firing threshold. Both excitatory and inhibitory chemical synapses are composed of a presynaptic and postsynaptic specialization separated by a 20-40 nm synaptic cleft (Kandel et al., 2000). In general, the presynaptic compartment (synaptic bouton) is localized in axon terminals of the neuron transmitting the signal, and it contains synaptic vesicles filled with neurotransmitters. A small number of these vesicles are positioned along the presynaptic plasma membrane at neurotransmitter-release sites called ‘active zones’; others, however, are kept further away from the presynaptic membrane until needed (Rettig and Neher, 2002). Synaptic vesicles are held in place by Ca2+-sensitive vesicle membrane proteins (VAMP), which bind to various elements of the cytoskeleton. Directly opposite the synaptic bouton is the postsynaptic specialization, which contains neurotransmitter receptors, and is found on cell bodies or dendrites of the target/receiving neuron (Cowan et al., 2001). The postsynaptic specialization of excitatory synapses is composed of an elaborate complex of interlinked proteins called the postsynaptic density (PSD). The PSD appears as an electron-dense structure by electron microscopy and is found immediately underneath the postsynaptic membrane facing the synaptic bouton. Proteins in the PSD serve a variety of roles that support neurotransmission, from anchoring neurotransmitter receptors to the membrane to regulating receptor activity (Cowan et al., 2001;

Sheng and Kim, 2002).

1.2.2. Chemical synaptic transmission in the brain

The process of chemical synaptic transmission begins when an action potential reaches the presynaptic axon terminal and alters the local resting membrane potential. This change in membrane potential leads to opening of voltage-gated Ca2+ channels followed by a large influx of Ca2+ ions into the presynaptic terminal. The transient rise in intracellular Ca2+ concentration causes the fusion of synaptic vesicles with the presynaptic plasma membrane, and as a result, the release of their neurotransmitters into the synaptic cleft. Released neurotransmitters diffuse across the synaptic cleft and bind to specific receptors on the postsynaptic membrane. This in turn, activates the receptors resulting in an ionic flux that alters the membrane conductance and potential of the postsynaptic cell (Purves, 2008). At excitatory synapses, the ion flux that follows the activation of neurotransmitter receptors leads to a net increase in positively-charged ions within the postsynaptic terminal. The result is a temporary depolarization of the postsynaptic membrane potential referred to as excitatory postsynaptic potential (EPSP). At inhibitory synapses, receptor activation leads to the production of an inhibitory postsynaptic potential (IPSP). An EPSP will drive a cell toward a point above its firing threshold, whereas an

IPSP does the opposite. When firing threshold is achieved, membrane depolarization occurs and is propagated down the neuron. In this way, an action potential that was converted into a chemical message at the presynaptic axon terminal has been converted back into an electrical signal in the postsynaptic cell (Purves, 2008).

There are two broad families of neurotransmitter receptors on the postsynaptic membrane. The ionotropic, or ligand-gated receptors, are multimers composed of at least 4-5 protein subunits. These receptors contain an extracellular domain that binds to neurotransmitters, and a membrane-spanning region that forms an ion channel. Upon ligand 17

binding, the receptor undergoes a conformational change that results in opening of the channel.

Channel-permeable ions can then flow in, or out of the neuron (Kandel et al., 2000). The second family of neurotransmitter receptors is the metabotropic receptor. Unlike ionotropic receptors, the metabotropic receptors are not ion channels. Instead, they affect the opening or closing of other channels through intermediate molecules called G-proteins. For this reason, these receptors are also referred to as G-protein-coupled receptors. When a neurotransmitter binds to the receptors’ extracellular domain, it initiates the recruitment and activation of G- proteins, which may interact directly with ion channels causing them to open or close.

Activated G-proteins may also evoke a variety of other cellular responses as it can stimulate the production of intracellular second messengers (Kandel et al., 2000; Greengard, 2001; Purves,

2008).

Ionotropic and metabotropic receptors have different physiological functions. Binding of neurotransmitters to ionotropic receptors produce relatively rapid postsynaptic effects (on the order of milliseconds) because channel opening involves a change in the conformation of a single macromolecule. The role of the ionotropic receptors in synaptic transmission is to either excite a neuron to fire an action potential, or to inhibit it from firing an action potential.

Activation of metabotropic receptors, on the other hand, typically produces slower (tens of millisecond to seconds) and longer-lasting (seconds to minutes) responses because it involves an indirect gating of ion channels through a cascade of intracellular reactions. The slow synaptic actions of metabotropic receptors normally are not enough to get cells to an action potential firing threshold. As a result, these receptors are not involved in the rapid on-off behavior of the ionotropic receptors. However, metabotropic receptors, through G-proteins, can recruit and stimulate the production of freely diffusible intracellular second messengers. These molecules

can, in turn, affect the function of a variety of channels. For example, they can modulate postsynaptic ionotropic receptors to alter the size of fast postsynaptic potentials. They can also act on resting channels and voltage-gated channels in the cell soma, thus influencing a number of electrophysiological properties of the neuron, including resting potential, input resistance, and action potential duration. Metabotropic receptors on presynaptic terminals can also influence neurotransmitter release and therefore the size of the postsynaptic potential, by regulating presynaptic K+ and Ca2+ channels. Thus, metabotropic receptors serve as modulators of synaptic transmission (Kandel et al., 2000).

Neurotransmitters that mediate excitatory synaptic transmission include glutamate, acetylcholine, and serotonin, whereas those involved in inhibitory transmission include GABA and glycine. In the mammalian CNS, glutamate is the transmitter of the vast majority of fast excitatory synapses and plays an important role in a wide variety of CNS functions (Hollmann and Heinemann, 1994). At these glutamatergic excitatory synapses, there are three classes of ionotropic receptors for glutamate (iGluRs): the N-methyl-D-aspartic acid receptors (NMDARs), the α-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid receptors (AMPARs), and the kainate receptors (KARs), named after the synthetic agonists that activate them most effectively

(Dingledine et al., 1999). Additionally, there are metabotropic glutamate receptors (mGluRs), which have been classified as group I, II, and III mGluRs (Conn and Pin, 1997). Interestingly, while kainate receptors (KARs) have been typically classified as ionotropic receptors based on their topology and function, several studies in the past decade suggest that KARs can also have a metabotropic mode of action (Rodriguez-Moreno and Sihra, 2007a). The following sections will provide an overview of the biology and function of the three ionotropic glutamate receptors in excitatory synaptic transmission.

1.2.3. Ionotropic glutamate receptors

Glutamate receptors mediate the majority of the excitatory neurotransmission in the mammalian CNS, and participate in plastic changes in the efficacy of synaptic transmission.

However, during a variety of acute and chronic neurological disorders, excessive activation of glutamate receptors can also lead to excitotoxic neuronal cell death. Thus, glutamate receptors are closely involved in both the physiology and pathology of brain functions (Ozawa et al.,

1998).

The ionotropic glutamate receptors (iGluRs) consisting of AMPAR, NMDAR, and KAR, are cation channels that mediate rapid excitatory transmission. KARs and AMPARs are permeable to Na+, and K+, while NMDARs allow the flux of Na+, K+, and Ca2+. Near a membrane resting potential, the driving force for K+ is low, so activation of AMPARs and

KARs by glutamate leads to depolarization as a result of an inward Na+ resting potential current.

NMDARs, however, cannot be opened by binding of glutamate alone as their channel pore is blocked by Mg2+. The NMDAR ion channel will open only if glycine and glutamate are bound to the receptors, and if there is sufficient membrane depolarization to drive the Mg2+ out of channel. Once open, there is an influx of both Ca2+ and Na+ through the channel which further contributes to membrane depolarization. Additionally, influx of Ca2+ leads to the activation of

Ca2+-dependent enzymes and protein kinases, thus triggering signal transduction cascades that contribute to long-lasting changes in the properties of the postsynaptic neurons. These changes are thought to mediate the plasticity of chemical synapses, which underlies the processes of learning and memory (Dingledine et al., 1999; Kandel et al., 2000; Levitan and Kaczmarek,

2001).

AMPARs and KARs are closely related; they are often collectively referred to as non-

NMDARs, as initially there were neither agonists nor antagonists that could clearly distinguish between the two. However, the development of specific antagonists that differentially block either AMPARs or KARs has allowed a better characterization of each of these receptors in vivo

(Ozawa et al., 1998). KARs play diverse roles in the CNS. In addition to contributing to membrane depolarization at a subset of excitatory synapses, they can influence neuronal excitability and regulate excitatory and inhibitory neurotransmitter release (Lerma, 2006). The function and biology of KARs will be described in more detail in a separate section.

AMPARs are distributed ubiquitously throughout the CNS and are composed of four types of subunits, GluA1-4, which combine to form tetramers (Mayer, 2005). At synapses,

AMPARs are essential for basal excitatory synaptic transmission and given their rapid kinetics, are responsible for the early component of EPSPs (Herman, 2003). AMPARs are associated with a number of proteins, many of which are PDZ domain proteins that regulate receptor trafficking and synaptic stabilization or that link receptors to signaling proteins (Sheng and Sala,

2001). AMPARs also interact with a family of small transmembrane AMPAR regulatory proteins called TARPs (Jackson and Nicoll, 2011). A number of studies over the past decade have shown that TARPs are auxiliary subunits that critically regulate the function of synaptic

AMPARs. For example, TARPs are required for the surface expression and synaptic targeting of AMPARs. Moreover, they can modulate AMPAR channel properties by altering the gating kinetics and the affinity of these receptors to glutamate (Jackson and Nicoll, 2011).

NMDARs are found in many excitatory synapses in the CNS. They are composed of an obligatory GluN1 subunit and variable GluN2A-D subunits (McBain and Mayer, 1994). GluN1 is required for the active surface expression of NMDARs (Dingledine et al., 1999; Squire, 2003), 21

while GluN2 subunits are important for modulating receptor activity (McBain and Mayer, 1994;

Squire, 2003). At synapses, NMDARs function as “molecular coincidence detectors” as the ion channel opens only when glutamate is bound to the receptor and when the postsynaptic cell is depolarized. Moreover, while NMDARs have much slower kinetics than AMPARs and contribute to the late phase of EPSPs, their activation, which increases postsynaptic Ca2+ concentration, is critical for the initiation of long-term potentiation (LTP), a persistent increase in synaptic strength (Herman, 2003; Malenka and Bear, 2004). Similar to the AMPARs, the cytoplasmic domain of NMDARs bind to a number of proteins, including the membrane- associated guanylate kinase (MAGUK) proteins. Members of this family of proteins are thought to regulate NMDAR trafficking from the endoplasmic reticulum to the synaptic membrane and to cluster and anchor these receptors at the synapse (Wenthold et al., 2003; Prybylowski and

Wenthold, 2004). Additionally, NMDARs have been shown to interact with Neto1, a CUB domain-containing single pass transmembrane protein (Ng et al., 2009). Neto1 regulates the synaptic abundance of the Glu2A subunit of NMDARs either by influencing receptor delivery or stability at the synapse.

1.3. Kainate Receptors

Kainate receptors are one of the three subtypes of ionotropic receptors for the excitatory neurotransmitter glutamate (Dingledine et al., 1999). The first KAR subunit gene (GluK1) was cloned in 1990 (Bettler et al., 1990), and though KARs are still the least well-understood receptors of the glutamate-gated ion channel family, significant progress have been achieved over the past two decades in understanding their biophysical properties and function in the brain. 22

This section provides an overview of current knowledge about KARs, from their subunit composition and distribution to their function and regulation in the brain. Some of the outstanding questions in the field are also presented.

1.3.1. KAR subunits: general structure and biophysical properties

KARs are tetrameric ion channels composed of combinations of GluK1-5 subunits

(previously referred to as GluR5-7, and KA1-2) (Wisden and Seeburg, 1993; Hollmann and

Heinemann, 1994; Bettler and Mulle, 1995). All subunits have a large extracellular domain containing the ligand binding site, a transmembrane region composed of three alpha helices and a re-entrant loop, and an intracellular C-terminal (Mayer, 2006). This last region is the most variable among the subunits and in the case of KAR subunits GluK1-3, has several alternatively spliced variants which affect receptor trafficking (Jaskolski et al., 2005a). In addition to alternative splicing, the diversity of KAR isoforms is expanded further by RNA editing

(Sommer et al., 1991). In GluK1-2 subunits, editing of a conserved glutamine (Q) to an arginine

(R) significantly affects several aspects of channel function including, calcium permeability

(Egebjerg and Heinemann, 1993; Burnashev et al., 1995) and channel conductance (Swanson et al., 1996).

The GluK1-3 subunits, which bind to kainate or glutamate with low affinity (micromolar range), can form functional ion channels either as homomers (Sommer et al., 1992; Egebjerg and Heinemann, 1993; Schiffer et al., 1997) or in heteromeric combination with GluK1-3 or

GluK4-5 subunits (Herb et al., 1992; Cui and Mayer, 1999; Paternain et al., 2000; Christensen et al., 2004). The high affinity GluK4-5 subunits, in contrast, cannot form functional homomeric

receptors, but need to be part of heteromers containing GluK1-3 (Werner et al., 1991; Herb et al.,

1992; Schiffer et al., 1997; Ren et al., 2003a).

The mechanisms of KAR activation and desensitization are similar to those of AMPARs

(Mayer, 2006). Agonist binding leads to closure of the KAR ligand binding domain (LBD).

This event triggers conformational changes in the receptor’s pore-forming region, which results in opening of the channel. Receptor desensitization occurs when the strain induced by LBD closure and channel opening eventually causes a rearrangement of the LBD interface (Sun et al.,

2002) (Figure 1.5). Similar to the AMPARs, recombinant KARs desensitize rapidly (1-6 milliseconds exponential decay) and profoundly under saturating glutamate concentrations. The time course of recovery from desensitization, however, is generally slower in KARs. For example, GluK2 receptors recover with a time constant of 2-3 seconds (Heckmann et al., 1996;

Paternain et al., 1998; Bowie and Lange, 2002), while AMPARs recover within several hundred milliseconds or less (Lomeli et al., 1992). The rate of entry into and recovery from desensitization are both strongly influenced by the subunit composition of KARs (Perrais et al.,

2010).

1.3.2. KAR pharmacology

Earlier studies of KAR physiology has been hindered by the lack of selective pharmacological compounds that can discriminate between AMPARs and KARs. For instance,

AMPA can act as a low affinity agonist for certain subtypes of KARs (Herb et al., 1992;

Swanson et al., 1996; Schiffer et al., 1997), while kainate can also induce rapid desensitization of AMPARs (Patneau et al., 1993). Recent progress in our understanding of the function of native KARs has been made possible by the discovery of AMPAR selective antagonists, such as the 2,3-benzodiazepine GYKI 53655 (Paternain et al., 1995; Wilding and Huettner, 1996), and by the development of KAR-null mice (Mulle et al., 1998; Contractor et al., 2001; Contractor et al., 2003) and KAR-specific agonists and antagonists (Traynelis et al., 2010).

Agonists that activate KARs to a greater degree than AMPARs include domoic acid

(Jane et al., 2009), dysiherbaine (Sakai et al., 2001), SYM2081 (Zhou et al., 1997), and ATPA

(Clarke et al., 1997; Alt et al., 2004). Domoic acid has higher binding affinity for GluK4 and

GluK5-containing receptors, while dysiherbaine has a higher affinity for GluK1 and GluK2.

SYM2081 is a glutamate analogue that displays selectivity for GluK1 and GluK2-containing

KARs and causes pronounced KAR desensitization (Jane et al., 2009). ATPA, on the other hand, is an AMPA analog that activates homomeric and heteromeric GluK1-containing KARs with low micromolar potency and at least 100 times selectivity (Clarke et al., 1997; Alt et al.,

2004).

Currently only a small number of compounds have been described as selective antagonists of KARs. Most of the antagonists that inhibit KAR activation also inhibit AMPARs with varying degrees of selectivity (Pinheiro and Mulle, 2006). Compounds of the

quinoxalinedione family, such as CNQX, DNQX, and NBQX act as competitive antagonists of native and recombinant KARs; however, CNQX does not discriminate well between AMPARs and KARs (Egebjerg et al., 1991; Alt et al., 2004), while NBQX is actually 100 times more selective for AMPARs, and has, therefore, been used to isolate KAR currents (Bureau et al.,

1999). Other KAR antagonists include, the willardine derivate UBP-302, a potent and selective

GluK1 inhibitor that shows 200 times selectivity for GluK1-containing KARs than for

AMPARs (More et al., 2004), and MSVIII-19, also a GluK1 antagonist that is a synthetic analog of the natural KAR agonist dysiherbaine (Sanders et al., 2005).

Many of the currently known KAR agonists and antagonists show selectivity and higher affinity towards GluK1 than other KAR subunits. This is due to the ability of the larger GluK1 binding cavity to accommodate bulky ligands (Mayer, 2005, 2006). Future studies focused on the development of compounds that target receptors with a different subunit composition, such as the GluK2/GluK5 heteromers (the major KAR subtype in the brain (Petralia et al., 1994)) will help us to further elucidate the neurophysiological role of KARs.

1.3.3. KAR expression, protein distribution and trafficking

KARs are present throughout the central and peripheral nervous systems (Bahn et al.,

1994; Bischoff et al., 1997), however, the expression level of each subunit differs greatly between different brain regions and cell types (Table 1.1). GluK1, GluK2, and GluK5 are the main subunits expressed in the adult CNS, with GluK1 and GluK2 having little overlap in neuronal expression. GluK1 mRNA is mostly limited to Purkinje cells of the cerebellum, the subiculum, and CA1 interneurons (Wisden and Seeburg, 1993; Bahn et al., 1994; Paternain et al.,

2000). GluK2, on the other hand, is highly expressed in hippocampal CA3 pyramidal cells, the dentate gyrus, the cerebellar granule cell layer, caudate putamen, and the piriform cortex

(Egebjerg et al., 1991; Wisden and Seeburg, 1993; Bahn et al., 1994). GluK5 is the most ubiquitously expressed KAR subunit in the CNS. Relatively few populations of neurons do not express GluK5; the main exceptions being interneurons and cerebellar Purkinje cells (Herb et al.,

1992; Wisden and Seeburg, 1993; Bahn et al., 1994). GluK3 has a low level of expression and is located primarily in the neocortex and thalamus (Wisden and Seeburg, 1993), while GluK4 mRNA is found only in CA3 pyramidal neurons and the granule cells of the dentate gyrus

(Werner et al., 1991).

Immunocytochemical localization has been used in combination with functional studies to determine the subcellular distribution of the KAR proteins. In the hippocampus, anti-

GluK2/3 and anti-GluK5 antibodies have localized these subunits at postsynaptic membranes and in dendritic spines of pyramidal cells in CA3 and CA1 regions (Petralia et al., 1994).

Interestingly, KARs can be differentially targeted to various postsynaptic locations even within a single neuronal population. For example, electrophysiological studies showed that in CA3 pyramidal cells, which express GluK2, GluK4 and GluK5 subunits (Wisden and Seeburg, 1993;

Bureau et al., 1999), postsynaptic KARs can be found at synapses formed with mossy fiber

(MF) inputs, but not at distal synapses (on the same neuron) formed with the associational/commissural (A/C) inputs (Castillo et al., 1997; Vignes and Collingridge, 1997;

Mulle et al., 1998). This observation is in agreement with immunohistological studies which showed a preferential localization of GluK2, GluK4 and GluK5 in the stratum lucidum (Darstein et al., 2003), the region of MF synaptic contacts. Evidence for a presynaptic localization of

KARs comes primarily from electrophysiological data suggesting that presynaptic KARs modulate synaptic transmission (Vignes et al., 1998; Contractor et al., 2000; Kamiya and Ozawa,

2000; Contractor et al., 2001; Schmitz et al., 2001). In the hippocampus, presynaptic KARs containing GluK2, and perhaps GluK1 subunits, are thought to be present at axon terminals of

CA3 pyramidal cells that synapse onto CA1 pyramidal neurons (Schaffer collateral-CA1 synapses) (Vignes et al., 1998; Bortolotto et al., 1999; Clarke and Collingridge, 2004; Jaskolski et al., 2005a). Moreover receptors formed from the heteromeric combinations of GluK2, GluK4, and GluK5 subunits are localized to presynaptic boutons of MF-CA3 synapses (Contractor et al.,

2000; Contractor et al., 2001; Contractor et al., 2003; Darstein et al., 2003), while GluK1- and

GluK2-containing KARs are found in somatodendritic (Bureau et al., 1999; Mulle et al., 2000)

and presynaptic membranes of CA1 interneurons (Isaac et al., 2004). Thus, analysis of KAR distribution in the hippocampus shows that KAR subunits are selectively targeted to specific synapses and subcellular compartments (Figure 1.6). However, the molecular mechanisms that account for this complex polarized KAR localization remain largely unknown.

While the polarized targeting of KARs is not well understood, studies with recombinant receptors in cell lines and neurons have characterized the molecular determinants for the trafficking of KARs from the ER to the plasma membrane. Many of these determinants are cis- acting regulatory elements present within the C-terminal domain of KAR subunits (Coussen,

2009). For example, the GluK2a subunit isoform has a forward trafficking signal (870-

KCQRRLKHKPQ-880) that efficiently exports assembled receptors to the plasma membrane

(Jaskolski et al., 2004; Yan et al., 2004). On the other hand, GluK5 and GluK1c subunits contain an arginine rich (“RXR”) ER-retention motif, and fail to reach the plasma membrane unless assembled into a heteromeric receptor with subunits such as GluK2a (Gallyas et al., 2003;

Hayes et al., 2003; Ren et al., 2003a). All other subunits and alternative splice variants are expressed at the plasma membrane at varying degrees (Ren et al., 2003b; Jaskolski et al., 2004;

Jaskolski et al., 2005b). The subunit- and isoform-specific membrane targeting motifs are likely responsible for specific protein-protein interactions that regulate various steps of the ER to plasma membrane transport (Isaac et al., 2004). However, unlike NMDARs (Standley et al.,

2000; Scott et al., 2001; Xia et al., 2001), surface expression of KARs is not dependent on PDZ interactions involving the receptor’s C-terminal PDZ binding motif (Coussen et al., 2002; Ren et al., 2003a; Ren et al., 2003b; Jaskolski et al., 2004; Yan et al., 2004).

1.3.4. Neuronal function of KARs

KARs regulate synaptic transmission in the CNS through a variety of mechanisms. For example, some postsynaptic KARs act as ion channels to mediate membrane depolarization at a subset of excitatory synapses, while others influence neuronal excitability through effects on voltage-gated ion channels. Presynaptic KARs, on the other hand, fine tune synaptic plasticity and influence the strength of excitatory and inhibitory transmission by regulating neurotransmitter release (Lerma, 2006; Pinheiro and Mulle, 2006; Pinheiro and Mulle, 2008;

Contractor et al., 2011).

The first KAR-mediated excitatory postsynaptic currents (EPSCKA) were detected at the hippocampal mossy fiber synapse (Castillo et al., 1997; Vignes and Collingridge, 1997), which is formed between granule cell axons and the proximal dendrites of CA3 pyramidal neurons

(Figure 1.7). Subsequently, KAR-mediated synaptic responses were demonstrated in

GABAergic interneurons of the CA1 region (Cossart et al., 1998; Frerking et al., 1998), at parallel fiber-Golgi cell synapses (Bureau et al., 2000), in cerebellar Purkinje cells (Huang et al.,

2004), in basolateral amygdala neurons (Li and Rogawski, 1998), at thalamocortical connections on cortical interneurons (Miyata and Imoto, 2006), in synapses established by cones on bipolar cells in the retina (DeVries and Schwartz, 1999), and in synapses between sensory fibers and dorsal horn neurons in the spinal cord (Li et al., 1999).

KAR-mediated postsynaptic responses have comparatively small amplitude compared to

AMPAR-mediated EPSCs (~ 10% of total peak current). However, their long lasting time- course, which prolongs membrane depolarization beyond the brief temporal window mediated by AMPAR activation, allows significantly large charge transfer as well as integration of

excitatory inputs over a longer time period. In this way, KARs can contribute significantly to the generation of action potentials and to network activity (Lerma, 2003). For instance, studies on hippocampal CA1 interneurons found that activation of synaptic KARs, but not of the rapidly deactivating AMPARs, produced substantial tonic depolarization during moderate presynaptic activity, which then resulted in enhanced action potential firing in the postsynaptic neuron

(Frerking and Ohliger-Frerking, 2002).

While the remarkably slow decay kinetics (~30-150 ms time constant) is a hallmark of

KAR-mediated EPSCs at diverse sites in the CNS (Cossart et al., 1998; Frerking et al., 1998;

Bureau et al., 2000; Miyata and Imoto, 2006), in vitro expressed KARs activated by exogeneous glutamate pulses desensitize/deactivate in only a few milliseconds (Erreger et al., 2004;

Contractor et al., 2011). A number of mechanisms have been proposed over the past decade to reconcile the large kinetic difference between native and recombinant KARs, but no clear explanation has yet emerged. For example, it was initially hypothesized that KARs would be located extrasynaptically, and that the slow KAR-EPSCs were the result of activating extrasynaptic receptors by glutamate spillover. Subsequent studies, however, ruled out this possibility based on the observations that 1) increasing extrasynaptic glutamate by reducing glutamate diffusion or inhibiting reuptake did not alter the kinetics of KAR-mediated responses

(Castillo et al., 1997; Kidd and Isaac, 1999); and 2) KARs can be activated by quantal release of glutamate and the resulting miniature KAR EPSCs also have slow kinetics (Cossart et al., 2002).

It has also been proposed that the slow kinetics of native KARs was conferred by the GluK5 subunit (Barberis et al., 2008). However, while KAR-EPSCs at MF-CA3 synapses become faster in GluK5 knockout mice (Contractor et al., 2003), they were still not as fast as those observed in recombinant homomeric KARs expressed in heterologous cells, suggesting that

GluK5 is not likely to be the sole (if at all) mediator of the slow decay kinetics of native receptors. A third possibility is that KAR-associated proteins could determine the channel properties of synaptic receptors. Indeed, a number of proteins (e.g. PSD95, KRIP6, Neto2) have been recently reported to alter KAR kinetics (Garcia et al., 1998; Bowie et al., 2003; Laezza et al., 2007; Zhang et al., 2009). However, these studies have all been carried out in heterologous systems, and consequently their effect on native receptors remains to be investigated.

In addition to being present at the postsynaptic membrane, KARs are also localized to presynaptic terminals where they play a crucial role as regulators of excitatory and inhibitory neurotransmitter release. The role of presynaptic KARs in inhibitory neurotransmission has been most thoroughly explored in the hippocampus. For example, at synapses between hippocampal interneurons, KAR stimulation has been shown to enhance the release of GABA

(Mulle et al., 2000; Cossart et al., 2001), while at some other synapses (e.g. between interneuron-CA1 pyramidal cells), KAR activation inhibits GABA release (Clarke et al., 1997;

Rodriguez-Moreno et al., 1997; Rodriguez-Moreno and Lerma, 1998; Maingret et al., 2005).

The role of presynaptic KARs in excitatory transmission has been intensely examined at hippocampal MF-CA3 synapses (Figure 1.7), where early studies have suggested a preponderant presynaptic localization of these receptors. For instance, autoradiography data revealed that high-affinity binding for [3H] kainate in the stratum lucidum (where mossy fibers synapse onto

CA3 pyramidal cells) was significantly reduced after selective destruction of the afferent mossy fibers, but remained almost intact when the pyramidal cells were ablated (Represa et al., 1987).

At mossy fiber terminals, activation of presynaptic KARs by synaptically released glutamate has been shown to facilitate glutamate release in a frequency-dependent manner, thereby contributing to the characteristic short-term plasticity of mossy fiber excitatory transmission

(Contractor et al., 2001; Lauri et al., 2001b; Schmitz et al., 2001; Pinheiro et al., 2007). A similar role for presynaptic KARs has also been described at a variety of other synapses in the central and peripheral nervous systems (Pinheiro and Mulle, 2006; Contractor, 2008; Pinheiro and Mulle, 2008). However, while there is no question about the important role of presynaptic

KARs in excitatory and inhibitory neurotransmission, the mechanism(s) underlying the modulation of glutamate release by the activated receptors is still unclear.

KARs are also unique among ionotropic glutamate receptors in that they not only operate as conventional ion channels, but also mediate some of their function through metabotropic (G protein-mediated) signaling pathways (Rodriguez-Moreno and Sihra, 2007b).

For example, through a metabotropic action in CA1 pyramidal neurons, KARs reduce the K+ current-mediated after-hyperpolarization typically observed after cell firing (Melyan et al.,

2002). Given that after-hyperpolarization curtails repetitive firing, reduction of the K+ currents by KARs significantly enhances neuronal excitability (Melyan et al., 2002; Melyan et al., 2004).

However, while this and other studies have uncovered a metabotropic function by KARs, these receptors do not have conventional motifs for direct binding to G-proteins. Therefore, it seems likely that intermediate proteins may act as linkers or scaffolds in the KAR/G-protein signaling complex (Lerma, 2006; Contractor et al., 2011). However, these intermediate proteins remain to be identified as none of the currently known KAR-interacting proteins have been shown to play such a role.

1.3.5. KAR interacting proteins

KARs interact with a number of molecules, many of which are cytosolic proteins identified in proteomic and yeast two-hybrid screens. Studies have shown that some of these molecules affect KAR channel function while others affect subcellular distribution. However, most of these interactions have not been studied in neurons and synapses and as a result, their functional significance on native, synaptic KARs is still not well-understood.

The first KAR-interacting proteins identified were the PDZ domain-containing proteins

PSD95, SAP102, and SAP97 (Garcia et al., 1998). In vivo, GluK2 and GluK5-containing KARs and SAP proteins are present in the same macromolecular complexes (Coussen et al., 2002).

Interaction of PSD95 and GluK2 involves the last four amino acids of GluK2 (ETMA) and the first PDZ domain of PSD95 (Garcia et al., 1998). In heterologous cells, PSD95 does not affect the trafficking of KARs from the ER to the plasma membrane, but causes receptor clustering at the cell surface (Garcia et al., 1998). Moreover, coexpression of GluK2 and GluK5 with PSD95 significantly reduced the desensitization of glutamate-evoked currents (Garcia et al., 1998).

However, a subsequent study using outside-out patch membranes reported that PSD95 does not modulate the rate at which receptors desensitize but rather accelerates the recovery from desensitization (Bowie et al., 2003). Whether PSD95 family members regulate the channel function of synaptic KARs is unknown.

KRIP6 (kainate receptor interacting protein for GluR6/GluK2) is another protein shown to modify GluK2-KAR channel properties but not surface trafficking in heterologous cells

(Laezza et al., 2007). KRIP6 binds to the GluK2 C-terminus independent of the subunit’s PDZ

binding motif. Coexpression of GluK2 and KRIP6 reduces KAR peak current amplitude and steady-state desensitization, but does not significantly alter decay kinetics (Laezza et al., 2007).

Other PDZ domain-containing proteins that KARs (GluK1 and GluK2 subunits) interact with are GRIP (glutamate receptor interacting protein) and PICK1 (protein interacting with C kinase-1) (Hirbec et al., 2003). GRIP, PICK1 and PSD95 interact with KARs in the brain, and all three proteins are present in the PSD. Interactions with GRIP and PICK1 proteins have been proposed to stabilize KARs to the postsynaptic membrane of MF-CA3 synapses because disruption of PDZ interactions with specific peptides and recombinant proteins causes a rapid decrease in KAR-mediated synaptic transmission (Hirbec et al., 2003).

A number of KAR-associated molecules have been identified by immunoprecipitation from transgenic mice overexpressing myc-tagged GluK2 subunits. These include the cadherin/catenin complex of transmembrane adhesion molecules (Coussen et al., 2002), which are enriched in the perisynaptic region (Uchida et al., 1996). Cadherin/catenin do not interact directly with KARs, but GluK2 has been shown to colocalize with ß-catenin at cell-cell junctions in transfected COS-7 cells (Coussen et al., 2002). Moreover, activation of cadherin in

COS-7 cells caused a redistribution of GluK2-KARs to cadherin/catenin complexes, suggesting that this interaction in neurons may be important for the synaptic localization of KARs (Coussen et al., 2002). In a follow-up study by Coussen et al., a set of proteins was identified that associate with the GluK2a and GluK2b subunit isoforms (Coussen et al., 2005). GluK2a and

GluK2b can assemble into the same heteromeric complex in native receptors, and can therefore bring together different sets of interacting cytosolic proteins. Some of these proteins (e.g. dynamin-1, dynamitin, 14-3-3) are known to be involved in the assembly and trafficking of membrane receptors, while others (spectrin, profiling II) may participate in cytoskeletal 39

reorganization. GluK2b was also found to be associated with proteins that are involved in the regulation of receptors and ion channels by Ca2+, such as calcineurin, calmodulin, and neurocalcin. Additional studies will be needed, however, to determine the exact role of each of these proteins on KAR distribution and function.

Recent studies identified the CUB-domain transmembrane protein Neto2 to be a KAR- associated protein (Zhang et al., 2009). Neto2 was shown to slow the deactivation and desensitization of GluK2 homomeric KARs in heterologous systems and to accelerate recovery from desensitization (Zhang et al., 2009). Neto2 did not, however, affect the cell surface expression of KARs. In cerebellar granule neurons, cotransfection of Neto2 with a GluK2 mutant that reduces KAR desensitization significantly increased the frequency of miniature

EPSCs that could be detected. Neto2 coexpression also slowed the decay kinetics of these

KAR-mediated mEPSCs (Zhang et al., 2009). Thus, Neto2 is a novel KAR-associated protein that modulates KAR channel properties. Whether Neto2 acts as an auxiliary protein for synaptic

KARs in the brain, however, is still not known.

1.3.6. KARs and disease

KARs have long been implicated in epileptogenic activity. As described in an earlier section, GluK1 and GluK2 subunits undergo Q/R RNA editing in the KAR pore forming region.

Editing at this site reduces Ca2+ permeability (Egebjerg and Heinemann, 1993; Burnashev et al.,

1995) and the single channel conductance (Swanson et al., 1996). Studies with editing-mutant mice have found that these mice are more susceptible to seizures following systemically administered kainic acid (Vissel et al., 2001). On the other hand, mice that do not express the

GluK2 subunit have a reduced susceptibility to kainate-induced seizures (Mulle et al., 1998).

Furthermore, in a mouse model of medial temporal lobe epilepsy, application of GluK1- selective antagonists blocks seizures induced by pilocarpine (Smolders et al., 2002). Thus, these findings clearly support a central role for KARs in the induction and propagation of seizures in mice. Future studies, however, will need to establish whether KARs are involved in human epilepsies before selecting these receptors as clinical targets for antiepileptic therapies.

1.4. The Neto family of transmembrane proteins

Neto1 and Neto2 comprise a family of closely related type I transmembrane neuronal proteins. Neto1 was identified by an in silico screen of human retinal ESTs to find novel proteins that might be involved in eye development (Ng et al., 2009). Neto2 was subsequently identified through database searches using BLAST for additional Neto1 related genes.

Sequence alignments show that murine Neto1 shares high sequence identity (~51%) with Neto2

(Ng, 2006).

The Neto proteins are conserved between vertebrates and invertebrates. In the human genome, the Neto1 gene maps to 18q22, and the Neto2 gene maps to 16q12. Human and mouse

Neto1 proteins share ~95% sequence identity. In Drosophila, a Neto-like protein (dNeto) has been recently described to function at the neuromuscular junction (Kim et al., 2012). The homology between dNeto and vertebrate Netos is mostly restricted to the extracellular domains:

24% identity with Neto1, and 19% with Neto2. A predicted protein with the same domain organization as the Neto proteins has also been identified in C. elegans (K05C4.11).

1.4.1. Domain structure and organization

Both Neto1 (533 amino acids) and Neto2 (525 amino acids) contain a signal sequence, two tandemly arranged extracellular CUB domains followed by a low-density lipoprotein receptor class A (LDLa) domain, a single-pass transmembrane domain, and a cytoplasmic tail

(Figure 1.8). The extracellular domain of the two Neto proteins is relatively well conserved:

CUB1, CUB2, and LDLa of Neto1 and Neto2 share 63%, 72%, and 84% identity, respectively

(Michishita et al., 2004; Ng, 2006). CUB domains – originally identified in the complement subunits C1r/C1s, sea urchin epidermal growth factor, and bone morphogenetic protein 1

(BMP1) – are often involved in protein:protein interactions (Bork and Beckmann, 1993). They contain approximately 110 amino acid-residues that form a conserved antiparallel ß-sheet structure similar to the antigen binding region of immunoglobulins (Bork and Beckmann, 1993;

Dias et al., 1997; Romero et al., 1997). The CUB domains of the Neto proteins are most related to those in neuropilins and tolloids (Stohr et al., 2002; Michishita et al., 2003, 2004).

Neuropilins are type I transmembrane receptors for class 3 semaphorins and are involved in axon guidance during development (He and Tessier-Lavigne, 1997; Kolodkin et al., 1997), while tolloid has been shown to be important for dorsoventral patterning of the embryo in D. melanogaster (Shimell et al., 1991).

The LDLa domain was initially identified in the low-density lipoprotein (LDL) receptor

(Yamamoto et al., 1984). The LDLa sequence in both Neto1 and Neto2 contain six invariant cysteines which form three disulfide bonds required for proper folding (Koduri and Blacklow,

2001). However, the LDLa domain does not have the highly conserved SDE motif, which is present in members of the LDL receptor gene family. Given that the SDE motif is required for ligand binding (Mahley, 1988), the LDLa motif of the Neto proteins may not be directly 42

involved in protein:protein interactions. Alternatively, it might bind to molecules that are different from ligands of the LDL receptor.

The cytoplasmic tails of Neto1 and Neto2 (168, and 157 amino acids, respectively) are located at the C-terminus and constitute the most divergent region between the two proteins

(~38% identity) (Michishita et al., 2004; Ng, 2006). Sequence analysis showed that the cytoplasmic domain of Neto2, but not Neto1, has a predicted 29 amino acid coiled-coil motif

(Michishita et al., 2004). In addition, while the last three residues of the Neto1 C-terminus constitute a canonical binding motif (TRV) for class I PDZ domains (Stohr et al., 2002; Ng,

2006), the C-terminal tripeptide of Neto2 (IDF) is a putative ligand for class II PDZ domains

(Ng, 2006). The differences between the cytoplasmic domains of the Neto1 and Neto2 suggest that each protein may be associated with a different set of intracellular binding partners.

1.4.2. Expression of Neto1 and Neto2 in the CNS

Murine Neto1 and Neto2 transcripts can be detected by northern blotting at E12.0, and

E9.0, respectively, although Neto2 ESTs have been identified in cDNA libraries from earlier developmental stages (i.e. 2-cell stage) (Ng, 2006). During embryonic development, Neto1 and

Neto2 are expressed in similar regions including the neural tube, developing cerebral cortex, corpus striatum, pons, medulla oblongata (Ng, 2006), and peripheral neurons such as the trigeminal ganglia and the dorsal root ganglia (Michishita et al., 2004; Ng, 2006). However, while there is a general overlap in expression, sub-regional differences were also present. For example, a cross-section of the neural tube at E13.5 shows that while both Neto1 and Neto2 are diffusely expressed throughout this region, Neto1 has a much stronger expression in two small areas which correspond to a subset of developing motoneurons, whereas Neto2 has a higher expression in the floor plate (Ng, 2006). Moreover, in the cerebral cortex, between E15.0 and

E18.0, Neto2 is strongly expressed in the cortical plate while Neto1 expression is primarily localized to the marginal zone and the subplate (Michishita et al., 2004; Ng, 2006), with only sparsely distributed puncta in the cortical plate (Michishita et al., 2004).

In the adult, Neto1 and Neto2 are both expressed in the spinal cord (Michishita et al.,

2004), and share overlapping areas of expression in the brain such as the olfactory bulb, olfactory tubercle, cerebral cortex, hippocampus, thalamus, and pons (Michishita et al., 2003,

2004; Ng, 2006; Ng et al., 2009). In the hippocampus, Neto1 mRNA is particularly abundant in pyramidal cells of the CA3 region but it is also detected in the DG granule cell layer, CA1 interneurons and throughout the CA1-3 pyramidal neurons (Michishita et al., 2003, 2004; Ng et al., 2009). Neto2, on the other hand, shows a relatively uniform expression along the pyramidal cell layer but has very weak expression in the DG granule cells (Ng, 2006), although Michishita 45

et al. observed no expression of Neto2 in DG at P21 by in situ hybridization (Michishita et al.,

2004) (Figure 1.9). In the cerebral cortex, both Neto1 and Neto2 transcripts show a diffuse distribution throughout all the cortical layers with the exception of layer VI, where Neto1 displays much higher expression (Ng, 2006). The cerebellum is the one brain region where the expression of Neto1 and Neto2 is markedly different: Neto2 is strongly expressed in the cerebellar granule cell layer and to a lesser extent in the Purkinje cell layer (Michishita et al.,

2004; Ng, 2006), whereas Neto1 shows very weak expression that is limited to the Purkinje cell layer (Ng, 2006) (Neto1 expression was not detected in neither the Purkinje cells nor in the granule cell layer in Michishita et al., 2004).

1.4.3. Function of Neto proteins in the nervous system

-Neto proteins in axon guidance

Both Neto1-null and Neto2-null mice display axon guidance defects during development.

In the developing embryo, Neto1 is expressed in commissural neurons in the neural tube, while

Neto2 is expressed in the floor plate. In wild-type animals, axons of the commissural neurons extend towards the floor plate at ~E10.5, and migrate away after crossing the midline at ~E12.5.

In both Neto1-null and Neto2-null mice, commissural axons are stalled at the floor plate at

E12.5 suggesting a role for Netos in repelling commissural axons away from the plate after midline crossing (Ng, 2006).

Neto1 is also thought to be involved in the guidance of axon projections of the corticospinal tract (CST), and the fornix (bundle of axons projecting from the hippocampus to the hypothalamus). In post-natal Neto1-null mice, a small number of misrouted axons from the

CST have been observed, and nearly a third of fornix projections are defasciculated (Ng, 2006).

-Neto proteins in synaptic transmission

Subcellular fractionation studies have shown that in the adult mouse brain Neto1 is enriched in the PSD of excitatory synapses (Ng et al., 2009). At these synapses, Neto1 is associated with the NMDA-type of ionotropic glutamate receptors, and with the synaptic scaffolding protein PSD95 (Ng et al., 2009). NMDARs are heterotetrameric assemblies composed of the obligate GluN1 subunits and the GluN2 subunits. In vitro assays showed that

Neto1 can interact with the GluN2A and GluN2B subunits, but not with GluN1 (Ng et al., 2009).

In the hippocampus, Neto1 plays an important role in NMDAR-mediated synaptic plasticity and learning. Loss of Neto1 results in decreased NMDAR-mediated currents, reduced LTP at 48

hippocampal Schaffer collateral-CA1 synapses, and impaired spatial learning and memory in

Morris water maze tests (Ng et al., 2009). One mechanism by which Neto1 affects NMDAR synaptic function is by regulating the delivery and/or stability of NMDARs at the postsynaptic membrane. In Neto1-null hippocampus, while the cell surface expression of all NMDAR subunits are normal, basal levels of postsynaptic GluN2A subunits are reduced by ~30%, indicating that Neto1 is required to establish or maintain the normal abundance of GluN2A- containing NMDARs in the PSD (Ng et al., 2009).

In addition to NMDARs, excitatory synapses contain the AMPA-, and kainate-type of ionotropic glutamate receptors. Though the learning impairments and synaptic plasticity deficits in Neto1-null mice were pharmacologically rescued with an AMPAR agonist (CX546), biochemical studies showed that Neto1 is not directly associated with AMPAR protein complexes (Ng et al., 2009). Whether Neto1 interacts with KARs in the brain and regulates their function has not yet been tested. On the other hand, recent studies by Zhang et al. showed an association of Neto2 with KARs in rat cerebellar lysates (Zhang et al., 2009). In heterologous expression systems, Neto2 increased the glutamate-evoked currents of GluK2- homomeric KARs by altering the rate at which KARs enter and recover from desensitization

(Zhang et al., 2009). Neto2 did not, however, change the cell surface expression of these receptors. In cerebellar granule neurons, when a GluK2 mutant with reduced desensitization was coexpressed with Neto2, mEPSCs occurred at a higher frequency, and with slower decay kinetics (Zhang et al., 2009). Together these results indicate that Neto2 can modulate the channel function of GluK2-KARs. However, it remains to be determined whether and how

Neto2 regulates the function of native, synaptic KARs.

A recent study in Drosophila showed that the Drosophila Neto, which has the same domain structure as the vertebrate Netos, is an essential protein for the clustering of ionotropic glutamate receptors (iGluRs) at the neuromuscular junction (NMJ) (Kim et al., 2012).

Moreover, in C. elegans, two other CUB domain-containing proteins have been shown to be auxiliary proteins for ion channels: SOL-1 slows the desensitization of GLR-1 AMPARs

(Walker et al., 2006), and LEV-10 is involved in clustering acetylcholine receptors (Gally et al.,

2004). Altogether, the association of SOL-1, LEV-10, and vertebrate and invertebrate Netos with different classes of neurotransmitter receptors suggests that a critical interaction with a

CUB domain-containing protein may be a well-conserved mechanism for the regulation of ligand-gated ion channels. In mice, 36 of the 42 genes that encode CUB domain proteins are expressed in the central nervous system (Allen Brain Atlas). However, it is unclear at present whether most or all ion channels are associated with CUB domain proteins, or whether a one-to- one specificity exists between ion channels and CUB domain proteins.

Thesis Objectives

1) The CUB domain transmembrane proteins Neto1 and Neto2 are expressed in the mammalian central nervous system. Neto1 has been previously shown to regulate the synaptic abundance of the NMDA-type of glutamate receptors. Neto2 has been recently described as a KAR- interacting protein that modulates homomeric GluK2 receptor kinetics in vitro. In Chapter II of my thesis, I describe experiments completed to identify novel Neto2 interacting proteins and to determine the in vivo roles of Neto2 on synaptic KARs.

2) The sequence and structural similarity between the Neto proteins suggested that Neto1 may also be a KAR-interacting protein. In Chapter II, I found that loss of Neto2 affected the postsynaptic abundance of KARs in the cerebellum, but not in the hippocampus. Given that

Neto1 is minimally expressed in the cerebellar cortex but is abundant in the hippocampus, I hypothesized that Neto1 may regulate KAR function in the latter brain region. In Chapter III, I describe the experiments that were carried out to elucidate the effect of Neto1 loss on synaptic

KAR function in the hippocampus.

Chapter 2: Neto2 is a CUB domain protein that regulates the synaptic abundance of cerebellar KARs

**This work was the result of a collaborative effort with Dr. Evgueni Ivakine from the lab of Dr.

Roderick McInnes, and Dr. Kenneth Pelkey from the lab of Dr. Chris McBain. While the majority of the work presented is my own, Dr. Evgueni Ivakine assisted in the identification of

Neto2 interacting proteins by GST pull-down presented in the Appendix, and Dr. Kenneth

Pelkey assisted in generating the electrophysiology data presented in Figure 2.8.

Data presented in Figures 2.4, 2.6, 2.7, and 2.8 have been published in J Neurosci 31: 10009-

10018 52

2.1. Introduction

Neto1 and Neto2 comprise a family of structurally similar single-pass transmembrane proteins present in the developing and mature nervous system. Both proteins encode a signal sequence, two CUB domains, an LDLa motif, a transmembrane domain and a cytoplasmic tail.

The extracellular region of Neto1 and Neto2 are relatively well conserved: the CUB1 and CUB2 domains share 63%, and72% identity, respectively; the LDLa motif is 84% identical. In contrast, the intracellular region of the Neto proteins share only 39% identity with most of the divergence occurring between the last 85, and 70 amino acids of Neto1, and Neto2, respectively.

One notable difference, for example, is found within the last three amino acids at the C-terminus.

In the case of Neto1, the C-terminal tripeptide encodes a class I PDZ ligand, allowing it to bind to class I PDZ domain-containing proteins such as PSD95 (Ng et al., 2009). Neto2, on the other hand, has a putative class II PDZ ligand and no known intracellular binding partners. The differences between the cytoplasmic domains of Neto1 and Neto2 suggest that they could be associated with very different intracellular molecules.

In the mature brain, Neto1 and Neto2 have an overlapping expression pattern with some sub-regional differences. For example, both proteins are strongly expressed in the cerebral cortex, hippocampus, olfactory bulb, and pons (Michishita et al., 2003, 2004; Ng, 2006 Ph.D. thesis). However, within the hippocampus, Neto1 is present in both pyramidal cells and granule cells, with particularly high expression in the CA3 pyramidal cell layer, whereas Neto2 mRNA is evenly distributed among the pyramidal cells of the CA1-CA3 region, with no expression in the granule cells of the dentate gyrus (Michishita et al., 2003, 2004; Ng, 2006 Ph.D. thesis). In

the cerebellum, the expression of Neto1 and Neto2 is also markedly different. While in situ shows that Neto2 is abundantly expressed in both granule cells and Purkinje cells, Neto1 mRNA is only weakly detected in the Purkinje cell layer (Michishita et al., 2004; Ng, 2006 Ph.D. thesis).

The differences in the distribution of Neto1 and Neto2 also suggest that, though structurally similar, these two proteins may play different roles and/or associate with different protein complexes in the brain.

Previous studies have focused mostly on Neto1, which was identified as an interacting protein of PSD95 (Ng et al., 2009), a major scaffolding molecule of the NMDA-type of glutamate receptors (Scannevin and Huganir, 2000; Sheng and Kim, 2002). The observed

Neto1:PSD95 interaction led to the eventual discovery that Neto1 acts as an auxiliary protein of synaptic NMDARs. It was found that loss of Neto1 significantly reduces hippocampal

NMDAR-mediated synaptic currents, alters NMDAR-dependent synaptic plasticity, and impairs spatial learning and memory (Ng et al., 2009). In contrast to our knowledge on Neto1, our understanding of the roles of Neto2 in the brain is far less advanced. Most recently, however,

Neto2 has been found to interact with the kainate-type of glutamate receptors, and studies in heterologous systems showed that Neto2 can prolong the decay kinetics and increase the glutamate-evoked currents of recombinant GluK2 KARs (Zhang et al., 2009). Coexpression of

Neto1 also enhances the glutamate-evoked currents of GluK2 KARs, though to a much lesser extent than does Neto2 (Zhang et al., 2009). Whether Neto2, and perhaps Neto1, regulate the channel function of native KARs, or affect their highly polarized neuronal distribution, however, remains to be determined.

A former graduate student, David Ng, has previously generated Neto2-null mice with the aim of elucidating the function of Neto2 in the mammalian nervous system. Mice lacking

Neto2 were viable and fertile, and pups from heterozygous intercrosses were obtained at the expected Mendelian ratios (Ng, 2006 Ph. D. thesis). The expression of Neto2 transcripts in the developing and mature mouse brains has also been characterized in the lab and elsewhere

(Michishita et al., 2004; Ng, 2006 Ph.D. thesis). In this chapter, I will describe the work that I have done to further study this protein. I began with a number of biochemical approaches to identify the cellular complexes that Neto2 might be associated with. The search focused primarily on intracellular binding partners of Neto2 given that most of the divergence between the Neto1 and Neto2 sequences is found within their cytoplasmic domain. I have also characterized the Neto2-KAR interaction in vivo, and explored its role on the function of KARs in the hippocampus and the cerebellum.

2.2. Materials and Methods

Mice

Neto2-null, and wild-type mice used in this study were previously generated in the lab by David Ng (Ng, 2006 Ph.D. thesis). The Neto2 gene was disrupted by homologous recombination using a targeting vector with a a loxP-pgk-neo-loxP cassette cloned in-frame with the Neto2 start codon. All animals have been maintained at the Toronto Center for

Phenogenomics (TCP).

PCR genotyping

Neto2 genotyping was performed using PCR. Each PCR reaction contained 50-100 ng of each primer, 8 ul Qiagen Multiplex PCR mix, 1-10 ng of DNA template in a final volume of

20 ul. Samples were first heated to 95 oC for 15 min followed by 35 thermal cycles consisting of a short denaturation at 94 oC for 30 sec, 56 oC annealing for 90 sec, and 72 oC extension for 1 min. After the last cycle, the samples were subjected to a 72 oC final extension for 10 min, and were stored at 4 oC. PCR products were analyzed using 1% agarose gels. For Neto2 genotyping, the following primers were used: mRtl2-2Larm-F2 (5’ GTA GGT ATA GGT AGG ATG GTT

3’), mRtl2-intron-R (5’ GCA GAA GTA CCA GAA AGC 3’), and DTA-R2 (5’ CTA GTG

AGA CGT GCT ACT TC 3’).

Commercial antibodies

The following commercial antibodies were used: rabbit polyclonal antibodies to GluK2,

GRID2 (Abcam), GluK5, GluA2/3, GRIP, and normal rabbit IgG (Millipore); mouse monoclonal antibodies to GluN1 (BD Biosciences), PSD95 (Thermo Scientific), Homer

(Abcam), VAMP2 (Synaptic Systems), and GRIP (BD Transduction Laboratories); goat polyclonal antibodies to Neto2 (R&D systems).

Purification of Neto2 and Neto1 antibodies

Neto2 antiserum was generated from rabbits injected with a GST fusion protein containing the C-terminal 70 amino acids of Neto2. Antiserum was affinity purified against a

MBP-tagged Neto2 (C-terminal 70 amino acids) fusion protein. MBP-Neto2(C-70) was expressed in BL21 cells. For a 1 liter culture, the cell pellet was resuspended in 30 ml of lysis buffer (0.1 mg/ml lysozyme, 76 units/ml of DNAseI (New England Biolabs), and 0.01 M MgCl2, in 30 ml of B-Per® (Pierce) reagent) by vortexing for 2 min. The cell lysate was incubated at room temperature for ~20 min to achieve thorough lysis, and diluted in 140 ml of ice-cold ACB buffer (0.2 M NaCl, 20 mM NaH2PO4, 1 mM EDTA). The diluted cell lysates were centrifuged at 20,000 x g for 10 min. at 4 oC to pellet the cell debris. The resulting supernatant was passed through 10 ml of amylose resin (New England Biolabs) column. The column was washed with ice-cold ACB buffer, and bound fusion protein was eluted in 10 x 1.5 ml fractions using 10 mM maltose in ACB buffer. The fractions with the highest protein concentration were pooled and stored in aliquots at -80oC until needed for preparing the affinity column used for antibody purification.

To make the affinity column, 2 mg MBP-Neto2(C-70) protein was conjugated to 1.5 g

CNBr-activated Sepharose 4B beads (GE Healthcare) by 3 h incubation at room temperature in coupling buffer (0.1 M NaHCO3 [pH 8.3], 0.5 M NaCl), followed by an overnight incubation at

4 oC. Beads were then washed with coupling buffer, and blocked in 0.2 M glycine [pH 8.0] for

2 h at room temperature followed by 16 h at 4 oC with end-over-end mixing. Beads were then pre-stripped with acid/base washes: 4 alternating washes of acetate buffer [pH 4.0] containing

0.5 M NaCl; and 0.1 M Tris/HCl [pH 8.0] containing 0.5 M NaCl; two consecutive washes with each of the following: 0.1 M glycine [pH 2.5]; 10 mM Tris/HCl [pH 8.8]; 0.1 M triethylamine

[pH 11.5]; 10 mM Tris/HCl [pH 7.5]. After pre-stripping the affinity column with acid/base washes, the column was equilibrated with 5 washes of PBS. Neto2 antisera (2 ml) were bound to the affinity column for 3 h at room temperature followed by overnight incubation at 4 oC with rotation. The column was then washed with 20 bed volumes of PBS, followed by 20 bed volumes of PBS containing 0.5 M KCl. Bound antibodies were eluted off the column with 20 ml of 0.1 M glycine [pH 2.5], and collected in 20 x 1 ml fractions (acid elution). Each fraction was immediately neutralized by addition of 50 ul of 2 M Tris/HCl [pH 8.8]. After all the fractions were collected, the column was neutralized and antibodies that were still attached were eluted with 0.1 M triethylamine [pH 11.5], and collected in 20 x 1 ml fractions (base elution).

Each fraction was immediately neutralized with 300 ul of 2 M Tris/HCl [pH 6.8]. The fractions with the highest protein concentrations from the acid and base elutions were pooled, dialyzed against 10 mM Tris/HCl [pH 7.5], and stored at 4 oC.

The Neto1 antiserum was raised in guinea pigs against a GST-Neto1 fusion protein containing the C-terminal 86 amino acids of Neto1; it was purified following the same protocol as that described for the Neto2 antiserum, except that a MBP-Neto1 fusion protein containing 58

the C-terminal 100 amino acid residues of Neto1 was used instead of MBP-Neto2(C-70) when preparing the affinity column.

Two-hybrid interaction studies

Fragments encoding the cytoplasmic region of Neto2 (amino acids 368-525, or amino acids 403-525) and the Neto2 C-terminal mutant ∆IDF (amino acids 368-522) were amplified by PCR from mouse whole brain cDNA and fused to the yeast GAL4 DNA-binding domain in pDBLeu (Invitrogen). Mouse GRIP[PDZ4-7] cDNA was obtained from RIKEN and subcloned in- frame with the GAL4 activation domain in the yeast vector pPC86 (Invitrogen). The controls used were the cytoplasmic domain of mouse Neto1 cloned into pDBLeu, and full-length mouse

PSD-95 cloned into pPC86. The vectors were sequentially transformed into MaV203 yeast cells and the interactions were scored by growth on triple dropout media (-Trp/-Leu/-His) and by using a ß-galactosidase filter assay that tests for the activation of the lacZ reporter gene.

Mammalian expression constructs

Full length Neto1 and Neto2 cDNA (encoding amino acids 1-533, and 1-525, respectively), and Neto2∆7 (amino acids 1-518) were generated by PCR, and subcloned into pcDNA3.1mycHisA(+) (Invitrogen) with a stop codon before the myc epitope tag. Full length

Neto2 cDNA (encoding amino acids 1-525), and deletion mutants Neto2-∆CUB1, Neto2-∆CUB2,

Neto2-∆LDLa, Neto2-∆cyto were generated by PCR and subcloned into a variant of pcDNA3.1mycHisA(+) (Invitrogen) containing two copies of the influenza hemagglutinin (HA) epitope tag (tagged to the C-terminus of the protein), and sequence verified. GRIP[PDZ4-7] cDNA

was subcloned into pcDNA3.1mycHisA(+) in-frame with the myc epitope tag to generate C- terminal myc-tagged GRIP[PDZ4-7]. FLAG-GluK2 was a kind gift from Dr. Katherine Roche

(National Institutes of Health, Bethesda, Maryland, USA).

Cell culture and transfection

Human embryonic kidney (HEK) 293 cells and transformed African Green Monkey

o kidney fibroblasts (COS-7) cells were maintained at 37 C, 5% CO2 in Dulbecco’s Modification of Eagle’s Medium (DMEM) (Wisent) containing 10% fetal bovine serum (FBS) (Wisent).

Cells were passaged every 3-5 days as follows: Cells growing in a 75 cm2 tissue culture flask

(Sarstedt) were washed once with pre-warmed 10 ml PBS, and incubated with 2 ml trypsin-

EDTA (Wisent) for 2-4 min to dislodge the cells from the growth surface of the flask. After incubation, 8 ml of DMEM containing 10% FBS was added, followed by thorough resuspension of the cells. Half a millilitre of this cell suspension was transferred into a new flask containing

10 ml of DMEM +10% FBS, and incubated at 37 oC until the next subculture.

Twenty-four hours before transfection, cells were seeded onto 6-well plates (Falcon).

Cells were transfected at 70% confluency with the appropriate constructs as follows: For one transfection, 97 ul of FBS-free DMEM was incubated with 3 ul of Fugene HD reagent for 5 min, followed by incubation with 1 ug total plasmid DNA for 15 min at room temperature. The entire transfection mixture was added drop-wise to cells growing in DMEM+10% FBS, and cells were maintained for another 48 h at 37 oC to allow over-expression of the desired proteins.

Forty-eight hours after transfection, cells were washed once with ice-cold PBS and lysed in 300 ul of RIPA buffer (50 mM Tris/HCl [pH7.4], 150 mM NaCl, 1 mM EDTA, 1% Nonidet

P-40, 0.5% deoxycholic acid (DOC), and 0.1% SDS) supplemented with Complete® Protease

Inhibitor Cocktail tablets (Roche). Lysed cells were scraped off the well, and transferred into individual microfuge tubes. Samples were incubated on ice for 30 min, and centrifuged at

13,000 x g for 15 min at 4 oC. The protein concentration of the supernatant was determined using the detergent-compatible DC Protein Assay according to the manufacturer’s protocol

(Bio-Rad). Quantified samples were stored at -80 oC, or used immediately for pull-down assays or coimmunoprecipitation experiments.

In vitro binding assay (GST pull-down)

The Neto2 cytoplasmic domain (cdNeto2) or the C-terminal mutant ∆IDF (cdNeto2∆IDF) were fused to glutathione-S-transferase (GST) by subcloning into a pGEX-4T-1 vector (GE

Healthcare). E. coli strain BL21 was transformed with pGEX-4T-1-cdNeto2, or pGEX-4T-1- cdNeto2∆IDF and cultured for 18 h at 37 oC in 2X YT medium containing 50 ug/ml ampicillin.

This culture was subsequently inoculated at 1:100 dilution into fresh 2X YT medium containing

o ampicillin, and grown at 37 C. When the bacterial suspension reached an OD600 of 0.5-0.8, protein expression was induced by addition of isopropylthio-ß-galactosidase (IPTG) to a final concentration of 1 mM. Cells were grown in IPTG containing media for 3 h at 30 oC, followed by centrifugation at 3400 x g for 10 min at 4 oC. The cell pellet was lysed with B-Per Bacterial

Protein Extraction Reagent according to manufacturer’s protocol. Fusion proteins were purified on glutathione agarose beads (Sigma), and quantified using the Bio-Rad detergent-compatible

assay. The integrity and purity of the proteins were analysed by performing Coomassie Blue staining of SDS-PAGE gels.

For the in vitro pull-down assay, lysates of HEK293 cells transiently transfected with a myc-tagged GRIP[PDZ4-7] cDNA were incubated with equal amounts of purified GST fusion proteins coupled to glutathione agarose beads. After overnight incubation at 4oC, beads were washed four times with RIPA buffer minus SDS and DOC and bound proteins were eluted with

6X sample buffer (0.375 M Tris/HCl [pH 6.8], 60% (v/v) glycerol, 12% SDS, 0.06% bromophenol blue, and 0.6 M DTT) followed by SDS-PAGE gel and immunoblotting.

Brain membrane fraction

Tissue from wild-type and Neto2-null mice was homogenized in ice-cold PBS with 20 up-and-down strokes in a glass Teflon homogenizer and centrifuged at 200 x g for 5 min at 4 oC.

The pellet was resuspended in ice-cold lysis buffer (50 mM Tris/HCl [pH 7.4], 1 mM EDTA), homogenized with 20 up-and-down strokes, and centrifuged at 10,000 x g for 30 min at 4oC.

The membrane pellet was homogenized in solubilization buffer (50 mM Tris/HCl [pH 7.4], 0.05 mM EDTA, 1% Triton X-100, 1% DOC) with 5 up-down-strokes, rotated for 3 h at 4oC, and centrifuged at 10,000 x g for 1 h at 4oC. All buffers were supplemented with Complete®

Protease Inhibitor Cocktail tablets (Roche). The supernatant was used for immunoprecipitation.

Isolation of crude synaptosomes

To prepare crude synaptosomes, mouse brain tissue was homogenized on ice in a glass

Teflon homogenizer (700 rpm, 20 up and down strokes) containing sucrose buffer (320 mM 62

sucrose, 10 mM EDTA, and 10 mM Tris/HCl [pH 7.4]). The homogenate was centrifuged at

1000 x g for 15 min at 4oC. The pellet was discarded while the supernatant was centrifuged at

10,000 x g for 15 min at 4oC. The supernatant from the last centrifugation was removed and the pellet was solubilized in DOC buffer (50 mM Tris/HCl [pH 9.0], and 1% DOC) at 37 oC for 30 min. The solubilised sample was centrifuged at 100,000 x g for 15 min at 4 oC. The supernatant was carefully isolated and an equal volume of modified RIPA buffer (50 mM Tris/HCl [pH 7.4],

150 mM NaCl, 1 mM EDTA, and 1% Triton X-100) was added to it. The sample was quantified using detergent-compatible DC Protein Assay (Bio-Rad) according to protocol, and stored at -80 oC. All the buffers used in this protocol were supplemented with Complete®

Protease Inhibitor Cocktail tablets (Roche).

Co-immunoprecipitation

For coimmunoprecipitation experiments using transfected HEK-293 or COS-7 cells,

~0.25mg of cell lysates were incubated with antibodies for 2 h at 4 oC on a rotating platform.

Lysates were subsequently incubated with 20 ul GammaBind IgG beads (GE Healthcare) for 1 h at 4 oC on a rotating platform. After incubation, beads were washed twice with 1 ml of ice-cold

RIPA buffer, twice with RIPA buffer minus SDS and DOC, and once with TBS-T (100 mM

Tris/HCl [pH 7.5], 150mM NaCl, 0.1% Tween-20). All buffers were supplemented with

Complete® Protease Inhibitor Cocktail tablets (Roche) and all washes were carried out for 10 min at 4 oC on a rotating platform. Bound proteins were eluted off the beads with 6X sample buffer and subjected to SDS-PAGE and immunoblotting. For coimmunoprecipitation from crude synaptosomes or brain membrane fractions, 1-2 mg of protein was incubated with antibodies or normal rabbit IgGs overnight at 4 oC on a rotating platform, and subsequently

incubated with 30 ul of GammaBind IgG beads for 2 h with rotation at 4 oC. Beads incubated with crude synaptosome samples were washed twice with 1ml ice-cold RIPA buffer, twice with

RIPA minus SDS and DOC, and once with TBS-T. Beads incubated with brain membrane fractions were washed 5 times with solubilization buffer (50 mM Tris/HCl [pH 7.4], 0.05 mM

EDTA, 1% Triton X-100, 1% DOC). All buffers were supplemented with Complete® Protease

Inhibitor Cocktail tablets (Roche) and all washes were carried out for 10 min at 4 oC on a rotating platform. Bound proteins were eluted with 6X sample buffer and subjected to SDS-

PAGE and immunoblotting.

SDS-PAGE and immunoblot analysis

Protein samples were separated on denaturing SDS-PAGE gels using standard methods.

Samples were boiled for 5 min or incubated at 50 oC for 20 min (PSD samples only) with the appropriate volume of 6X sample buffer, loaded on the gel, and electrophoresed in SDS-PAGE running buffer (192 mM glycine, 25 mM Tris/HCl [pH 8.3], and 0.1% SDS) for 90 min at 140

V. Protein samples separated on the gel were transferred onto Hybond-C Extra nitrocellulose membranes (GE Healthcare) at 40 V in transfer buffer (192 mM glycine, 25 mM Tris/HCl [pH

8.3], and 20% methanol). Following overnight transfer at 4 oC, membranes were briefly stained with Ponceau S solution (0.1% (w/v) Ponceau S, 5% acetic acid) to confirm successful protein transfer and to locate protein bands of interest. To proceed with immunoblotting, membranes were rinsed with distilled water to remove the Ponceau stain, and were then blocked for 1 h at room temperature with 5% skim milk powder dissolved in TBS-T, followed by overnight incubation at 4 oC with primary antibody. Membranes were washed four times with TBS-T (10 minutes per wash) and incubated with the appropriate horseradish peroxidise (HRP)-conjugated

secondary antibody for 1 h at room temperature. After treatment with secondary antibody, membranes were washed four times with TBS-T, and proteins, to which the primary antibody was bound, were detected by enhanced chemiluminescence.

Hippocampal and cerebellar homogenates

One pair of hippocampi or cerebella was extracted from animals and placed in a glass

Teflon homogenizer containing RIPA buffer supplemented with Complete® Protease Inhibitor

Cocktail tablets (Roche). The tissue was homogenized on ice using 20 up and down strokes at

700 rpm. Homogenized samples were incubated on ice for 15 min and centrifuged at 13,000 x g for 15 min at 4 oC. The supernatant was isolated, quantified using the detergent-compatible DC

Protein Assay (Bio-Rad), and stored at -80 oC for further use.

Subcellular fractionation and PSD isolation

Subcellular fractionation of mouse brains were performed by the method of Huttner et al.

(Huttner et al., 1983), as described (Kalia and Salter, 2003).

To isolate the PSD fraction, wild-type and Neto2-null whole brains, and pooled hippocampi, or cerebella were homogenized in ice-cold Solution A (0.32 M sucrose, 1 mM

NaHCO3, 1 mM MgCl2, 0.5 mM CaCl2) with 12 up-and-down strokes in a glass Teflon homogenizer. Homogenates were then centrifuged at 1400 x g for 10 min at 4 oC. The supernatant was saved (sup1), the pellet was resuspended with 3 up-and-down strokes in

Solution A, and then centrifuged at 710 x g for 10 min at 4 oC. The supernatant from this second centrifugation was saved (sup2) and combined with sup1. The combined supernatant 65

(sup1 + sup 2) was centrifuged at 710 x g for 10 min at 4 oC, and the resulting pellet was discarded while the supernatant (sup3) was transferred into a new tube and centrifuged at 13,800 x g for 10 min at 4 oC. After centrifugation, the pellet was saved and resuspended with 6 up- and-down strokes in Solution B (0.32 M sucrose, 1 mM NaHCO3). The resuspended sample was layered onto a sucrose gradient of 0.85 M, 1.0 M and 1.2 M sucrose/1.0mM NaHCO3, and centrifuged at 82,500 x g for 2 h. The band between 1.0 M and 1.2 M sucrose layers was carefully transferred into a new tube and an equal volume of solution C (1% Triton X-100 (v/v) in 0.32 M sucrose, 12 mM Tris/HCl [pH 8.0]) was added. The sample was incubated at 4 oC for

15 min with end-over-end mixing and centrifuged at 32,000 x g for 20 min at 4 oC. The supernatant was discarded and the white pellet was resuspended in 0.5 ml of PBS. Samples were stored at -80 oC for later use.

Samples to be used for coimmunoprecipitation experiments were solubilized in 50 mM

Tris/HCl [pH 9.0] containing 1% DOC at 37 oC for 30 min, and centrifuged at 100,000 x g for

15 min at 4 oC. The supernatant was collected and mixed with an equal volume of modified

RIPA buffer (50 mM Tris/HCl [pH 7.4], 150 mM NaCl, 1 mM EDTA, 1% NP-40, 0.1% SDS).

Protein concentrations were determined using a detergent-compatible Bio-Rad assay.

Samples to be used directly for immunoblot analysis were resuspended in 40 mM

Tris/HCl [pH 8.0] containing 1% SDS and 10 mM DTT, and were incubated at 60 oC for 20 min.

The solubilised samples were diluted with an equal volume of modified RIPA buffer (50 mM

Tris/HCl [pH 7.4], 150 mM NaCl, 1 mM EDTA, 1% NP-40, 0.25% DOC). Protein concentrations were determined using a detergent-compatible Bio-Rad assay.

All buffers were supplemented with Complete® Protease Inhibitor Cocktail tablets

(Roche).

Whole cell recordings

Hippocampal slices for whole cell recordings were prepared as previously described

(Pelkey et al., 2005) using P15-22 wild-type, and Neto2-null mice as indicated. Briefly, animals were anaesthetized with isoflurane and decapitated allowing removal of the brain into ice-cold saline solution (130 mM NaCl, 24 mM NaHCO3, 3.5 mM KCl, 1.25 mM NaH2PO4, 0.5 mM

CaCl2, 4.5 mM MgCl2, and 10 mM glucose, saturated with 95% O2 and 5% CO2 [pH 7.4]).

After dissection of the brain, individual hemispheres were transferred to the stage of a VT-

1000S vibratome (Leica Microsystems, Bannockburn, IL) and sectioned to yield transverse hippocampal slices (300 µm) which were incubated in the above solution at 35 oC for at least a

30 minute-recovery until use. All animal procedures conformed to the National Institutes of

Health animal welfare guidelines.

All recordings were interleaved with the experimenter blind to mouse genotype.

Individual slices were transferred to a recording chamber and perfused (2-3 ml/min) with extracellular solution (130 mM NaCl, 24 mM NaHCO3, 3.5 mM KCl, 1.25 mM NaH2PO4, 2.5 mM CaCl2, 1.5 mM MgCl2, 10 mM glucose, 0.005-0.010 mM bicuculline methiodide saturated

o with 95% O2 and 5% CO2 [pH 7.4], 32-35 C). Whole-cell patch-clamp recordings using a multiclamp 700A amplifier (Axon Instruments, Foster City, CA) in voltage-clamp mode (Vh=-

70 or +40 mV as indicated) were made from individual CA3 pyramidal neurons, visually identified with infrared video microscopy and differential interference contrast optics.

Recording electrodes (4-5 MΩ) pulled from borosilicate glass (World Precision Instruments) were filled with intracellular solution (ICS) composed of: 95 mM Cs-gluconate, 5 mM CsCl; 0.6 mM EGTA, 5 mM MgCl2, 4 mM NaCl, 2 mM Na2ATP, 0.3 mM NaGTP, 40 mM HEPES, 10 mM BAPTA, 1 mM QX-314 [pH 7.2-7.3], 290-300 mOsm. Uncompensated series resistance

(8-15 MΩ) was rigorously monitored by the delivery of small voltage steps at regular intervals and recordings were discontinued following changes of >10%. Synaptic responses (paired pulses or trains of 4 pulses, both at 20 Hz) were evoked at 0.1 Hz (for train recordings) or 0.2

Hz (for paired pulse recordings) by low-intensity microstimulation (100 µsec duration; 10-30

µA intensity) via a constant-current isolation unit (A360, World Precision Instruments, Sarasota,

FL) connected to a patch electrode filled with oxygenated extracellular solution in either the dentate gyrus or stratum lucidum for MF inputs. Mossy fiber-origin of EPSCs was confirmed by a rapidly rising AMPAR-mediated component showing strong short-term frequency facilitation and in train protocols by a residual KAR-mediated component upon AMPA receptor antagonism at a holding potential Vh of -70 mV. For MF train recordings, initially, dual component KAR/AMPAR-mediated synaptic responses were monitored at Vh= -70 mV following which the KAR-mediated component was pharmacologically isolated by applying the

AMPAR-specific antagonist GYKI 53655 (50 µM, Tocris Bioscience). The GYKI resistant component at Vh = -70 mV was confirmed to be KAR-mediated by subsequent application of

DNQX (25 µM, Tocris Bioscience) in the continued presence of GYKI 53655 and the holding potential was moved to +40 mV to obtain the NMDAR-mediated component of EPSCs followed by application of the NMDAR antagonist dl-APV (100 µM, Tocris Bioscience) in the continued presence of GYKI53655 and DNQX.

To measure AMPAR and KAR-mediated EPSCs, averaged traces (10-20 individual sweeps) obtained in GYKI 53655 with DNQX at Vh = -70 mV were digitally subtracted from averaged traces obtained at the end of the control and GYKI 53655 alone conditions, respectively. Similarly, for NMDAR-mediated EPSC analysis, averaged traces obtained in dl-

APV at Vh= +40 mV were digitally subtracted from those obtained in GYKI 53655/DNQX at

Vh=+40 mV. For each recording, EPSC amplitudes were measured during a 1-2 msec window around the peak of the waveform of the averaged traces for each condition. KAR- and

NMDAR-mediated EPSC amplitudes were measured for the 4th pulse of the trains and normalized to the amplitude of the corresponding AMPA receptor-mediated EPSC to eliminate slice to slice and animal to animal variability in the number of fibers recruited by extracellular stimulation. Data are presented as means ± SEMs unless otherwise indicated. Statistical significance was assessed using parametric (paired or unpaired t-tests) or non-parametric

(Mann-Whitney U test) tests as appropriate.

Immunohistochemistry

For preparation of brain slices, mice were intracardially perfused with PBS. Brains were dissected out immediately after perfusion, cryo-protected in 30% sucrose before embedding in

OCT, and sectioned at 50 µm-thickness. Cerebellar slices with comparable anatomy from each genotype were combined onto the same glass slide and subsequently processed together under identical conditions. Slices were rinsed three times with PBS to remove OCT and fixed with 4%

PFA in PBS for 1 minute on ice. The slices were washed three times in PBS and blocked with

10% goat serum, 0.3% Triton X-100 in PBS at room temperature for 1 hour. Primary antibodies in blocking solution were then incubated with slices overnight at 4 oC followed by three washes

with PBS, and incubation with appropriate secondary antibodies for 1 hour at room temperature.

Following incubation, slices were washed three times with PBS and cured for 24 hours at room temperature with Prolong Gold antifade reagent (Invitrogen). Images of slices from the same glass slide were acquired with fixed exposure settings using a Zeiss LSM 510 confocal microscope.

2.3. Results

2.3.1. Identification of Neto2 intracellular interacting proteins from adult mouse brain

To elucidate the molecular function of Neto2 in the adult mouse brain I have performed an unbiased yeast two-hybrid screen using an adult mouse brain cDNA library, and the cytoplasmic domain of Neto2 (Neto2(CD)) as bait. Clones were selected based on their ability to activate all three reporter genes (HIS3, URA3, and lacZ) when cotransformed with Neto2(CD)

(please refer to Appendix A for a list of clones isolated from this screen). Most of the clones identified in this screen correspond to the hypothetical protein AK078147, a ubiquitously expressed 33kDa molecule which has a coiled–coil domain and a disordered domain containing five proline rich-repeats of 16 residues each ([A/V]PQ[P/T][S/R]ENPPSPPTSPA). Recent studies have shown that the protein encoded by AK078147 is the mammalian homologue of the yeast nuclear protein Sfr1 implicated in the repair of DNA strand breaks (Akamatsu and Jasin,

2010). Additional putative Neto2 binding proteins from the screen include RIM2, a presynaptic

PDZ domain-containing protein involved in neurotransmitter release and presynaptic plasticity

(Mittelstaedt et al., 2010, review), and neurochondrin, a negative regulator of CAMKII phosphorylation and an essential protein for the spatial learning process (Dateki et al., 2005).

GST-pull down was also used to identify Neto2 interacting proteins. In this approach, proteins from a detergent-solubilized mouse brain membrane fraction retained by the GST-

Neto2(CD) fusion protein were identified by mass spectrometry. One advantage of the GST-pull down approach over the yeast two-hybrid system is that it allows the identification of Neto2- associated proteins, which may not interact directly with Neto2. A number of potential Neto2

binding partners were discovered by GST-pull down, such as the potassium-chloride cotransporter 2 (KCC2) (Ivakine et al., 2012), and the sodium-potassium ATPase1α3 (please refer to Appendix B for complete list of molecules). ATPase1α3 plays a critical role in controlling neuronal excitability by maintaining the Na+ and K+ electrochemical gradient across the plasma membrane. Reduced ATPase1α3 function results in epileptic seizures in mice and neuronal hyperexcitability (Clapcote et al., 2009). KCC2, on the other hand, is a neuron- specific membrane protein that uses the Na+/K+ gradient to transport chloride ions out of the cell, ultimately creating a Cl- gradient that is essential for inhibitory synaptic transmission (Payne et al., 1996; Hubner et al., 2001). Altogether, these preliminary findings suggest that Neto2 can be associated with a variety of proteins of different functions in the brain.

In addition to an unbiased screen, I also searched for Neto2 interacting proteins by a candidate approach. In this approach, I selected putative Neto2 interacting proteins based on distributions in the brain similar to Neto2 and the presence in those proteins of domains that may bind to the cytoplasmic region of Neto2. Previous in situ studies showed that the Neto2 mRNA is widely distributed in the brain (Michishita et al., 2004; Ng, 2006). Using anti-Neto2 antibodies raised against the C-terminus of Neto2, I detected a specific immunoreactive band of

~66kDa in brain lysates of wild-type mice. This band was absent from brain lysates of Neto2- null mice (Figure 2.1A). Immunoblot analysis of various brain regions (i.e. cerebellum, olfactory bulb, hippocampus, and cortex) confirmed that, consistent with the in situ expression data, Neto2 protein was present in all of the tissues analyzed (Figure 2.1B). Next, to determine the subcellular compartments in which Neto2 is localized, I performed immunoblotting experiments on biochemically separated subcellular fractions. As shown in Figure 2.1C/D,

Neto2 is prominently expressed in crude synaptosomal fractions, where it is present on both the 72

pre-, and postsynaptic membranes, but is absent from the synaptic vesicle fraction. Moreover, on the postsynaptic side, Neto2 is present in the PSD fraction, thus showing that it is a component of the PSD of excitatory synapses.

To identify protein domains or ligands that could bind to Neto2, I analyzed the intracellular amino acid sequences of Neto2. I found that the last three C-terminal residues (-

IDF) of Neto2 fit the sequence requirements of motifs that bind to class II PDZ domains (motif of class II PDZ ligands: ɸ-X-ɸ, where X can be any amino acid, and ɸ can be any hydrophobic residue (Sheng and Sala, 2001)). Based on this observation and my previously established subcellular localization of Neto2, my search for Neto2 interacting proteins focused on synaptic class II PDZ domain-containing proteins that display an overlapping distribution with Neto2 in the brain. While a literature search revealed that none of the currently identified class II PDZ domain-containing proteins can bind to motifs identical to that of Neto2, I selected the glutamate receptor interacting protein (GRIP) for initial analysis given the similarity of some of its class II ligands (e.g. SVKI) to the C-terminus tripeptide of Neto2. Moreover, GRIP is enriched at the PSD where it acts as a scaffolding protein for ion channels, receptors involved in axon guidance, and signalling molecules. Using the yeast two-hybrid assay, I found that coexpression of Neto2(CD) with a fragment encoding PDZ 4-7 of GRIP (GRIP(PDZ4-7)) resulted in positive lacZ reporter gene activity based on ß-galactosidase assays, suggesting an interaction between the two molecules. On the other hand, I did not observe any interaction between

Neto2(CD) and the PDZ domain of PICK1, a synaptic protein that can bind to both class I and class II PDZ binding motifs (Figure 2.2). To determine whether the last three C-terminal residues of Neto2 mediate the interaction with GRIP(PDZ4-7), I generated a Neto2(CD) truncation mutant (Neto2(CD-∆IDF)). Coexpression of Neto2(CD-∆IDF) and GRIP(PDZ4-7) did not result in activation of the lacZ gene, suggesting that the deleted residues constitute a PDZ binding motif that is critical for the Neto2:GRIP(PDZ4-7) interactions (Figure 2.2). As a control, Neto2(CD) was also tested against the class I PDZ domain-containing protein PSD95. No interaction between

these two proteins was observed in the two-hybrid assays, indicating a selective binding of

Neto2 to GRIP. In contrast, the Neto1(CD), which has a class I binding motif, showed interaction with PSD95, in agreement with previously reported observations (Ng et al., 2009), but it did not bind to GRIP(PDZ4-7) (Figure 2.2).

2.3.2. Neto2 binds to GRIP through PDZ ligand:PDZ domain interactions

To determine whether the interaction between Neto2 and GRIP(PDZ4-7) can also be demonstrated using an independent approach, I performed both GST pull-down experiments and coimmunoprecipitation studies with Neto2 and GRIP(PDZ4-7) expressed in COS-7 cells. For GST pull-down, equal amounts of GST-Neto2(CD), and GST-Neto2(CD-∆IDF) fusion proteins bound to glutathione agarose beads were incubated with recombinant myc-tagged GRIP(PDZ4-7). Protein complexes recovered from the beads were analyzed by immunoblotting. As shown in Figure

2.3A, I detected an association of GRIP(PDZ4-7)-myc with GST-Neto2(CD), but not with GST-

Neto2(CD-∆IDF). This observation confirms the earlier conclusion that the C-terminal tripeptide of

Neto2 is necessary for its interaction with GRIP. For coimmunoprecipitation experiments, I expressed full length Neto2, the Neto2 C-terminal truncation mutant (Neto2∆7), or GRIP(PDZ4-7) in COS-7 cells, and incubated the cell lysates with anti-Neto2 antibodies. I found that

GRIP(PDZ4-7) coimmunoprecipitated with the anti-Neto2 antibody only when coexpressed with

Neto2, but not when it was expressed alone or together with Neto2∆7 (Figure 2.3B). This result shows that GRIP(PDZ4-7) associates with full length Neto2 through a Neto2 C-terminal mediated interaction. To determine whether there is an interaction between native Neto2 and GRIP, coimmunoprecipitation experiments were performed using whole brain lysates. Here, I found that Neto2 coimmunoprecipitated with the anti-GRIP antibody but not with the negative control

IgG, thus indicating that Neto2 and GRIP are indeed associated in vivo (Figure 2.3). In summary, I have identified an interaction between Neto2 and the scaffolding protein GRIP.

Furthermore, these studies indicate that the interaction is critically dependent on the C-terminal residues of Neto2.

To examine whether GRIP interacts specifically with Neto2 but not Neto1, as suggested by yeast two-hybrid studies, I expressed full length Neto1 and GRIP(PDZ4-7) in COS-7 cells. In this system, GRIP(PDZ4-7) did not coimmunoprecipitate with Neto1, indicating a lack of interaction between the two proteins (Figure 2.3C). Additionally, to exclude the possibility that

Neto1 functions in the same protein complex as GRIP through interactions mediated by additional proteins not present in vitro, I examined Neto1 and GRIP associations in the brain. In whole brain lysates, I did not observe a coimmunoprecipitation of Neto1 with anti-GRIP antibodies (Figure 2.3D). On the other hand, coimmunoprecipitation of Neto2 was detected under the same conditions (Figure 2.3D). Based on these results, I conclude that GRIP does not interact with Neto1 and is not likely to be associated with it within the same protein complex in vivo.

2.3.3. Neto2 interacts with KARs but not AMPARs

GRIP was initially identified as an AMPAR-interacting protein (Dong et al., 1997) and has been implicated in regulating the recycling (Braithwaite et al., 2002; Mao et al., 2010), synaptic expression (Osten et al., 2000; Lu and Ziff, 2005), and Ca2+ permeability (Liu and

Cull-Candy, 2005) of these receptors. Subsequent studies showed that it also binds to kainate- type of glutamate receptors (KARs) (Hirbec et al., 2003), and a variety of other proteins in the brain such as the huntingtin associated protein HAP1-A (Ye et al., 2000), the Fraser syndrome protein Fras1 (Long et al., 2008), the adaptor protein liprin-α (Wyszynski et al., 2002), and the ephrin receptors and ligands (Torres et al., 1998). Since GRIP is thought to function as a scaffolding protein that connects and anchors various synaptic proteins, Neto2 may be in a complex with some of these proteins through GRIP. Of these proteins, I hypothesized that

Neto2 may be associated with GRIP-linked glutamate receptors given that the Neto1 was previously found to regulate the NMDA-type of glutamate receptors (NMDARs) (Ng et al.,

2009). Indeed, a recent study has shown that Neto2 coimmunoprecipitates with KARs in cerebellar lysates and regulates the kinetics of homomeric GluK2-KARs in heterologous cells

(Zhang et al., 2009). To confirm the in vivo interaction between Neto2 and KARs and to investigate whether Neto2 also associates with AMPARs, I performed coimmunoprecipitation experiments from mouse brain crude synaptosomes. Using an antibody against the intracellular domain of Neto2, I found a robust coimmunoprecipitation of the GluK2 subunit of KARs in wild-type but not Neto2-null samples (Figure 2.4A). However, under the same conditions, I did not detect an interaction of Neto2 with the AMPAR subunits GluA2/3 (Figure 2.4A). In reciprocal assays, anti-GluK2 and anti-GluK5 antibodies, but not anti-GluA2/3 antibodies, coimmunoprecipitated Neto2 from wild-type synaptosomes (Figure 2.4B). Additional

experiments using different detergent solubilisation of brain lysates, or different anti-Neto2 antibodies, also showed no detectable coimmunoprecipitation of AMPARs (data not shown).

Together, these results indicate that Neto2 has a specific association with the kainate-, but not the AMPA-type of glutamate receptors in the brain.

The PSD is an electron dense structure of excitatory synapses, where glutamate receptors and proteins directly involved in the regulation of synaptic function are organized and concentrated. To determine whether Neto2 is a component of KARs at the PSD, I tested the association of Neto2 and KARs in whole brain PSD fractions. Here, I found that anti-Neto2 antibodies coimmunoprecipitated the GluK2 and GluK5 subunits of KARs, but not the AMPA receptor subunit GluA2/3 (Figure 2.4C). Consequently, I conclude that Neto2 is an integral part of the postsynaptic KAR protein complex.

2.3.4. Neto2 associates with GluK2-KARs predominantly through the second CUB domain

To determine whether the association of Neto2 with KARs depends on the binding of its

C-terminal PDZ ligand to GRIP, COS-7 cells were transfected with various combinations of expression plasmids FLAG-GluK2, untagged Neto2, and GRIP(PDZ4-7)-myc. Cell lysates were incubated with an antibody against the extracellular domain of GluK2 and immunoprecipitated proteins were analyzed by western blotting. As shown in Figure 2.5, Neto2 was able to coimmunoprecipitate with GluK2 in the presence or absence of GRIP(PDZ4-7) (compare lanes 4 and 6, immunoblotted for Neto2). Furthermore, in cell lysates coexpressing GRIP(PDZ4-7),

GluK2, and Neto∆7, we found that Neto∆7, which does not bind to GRIP(PDZ4-7), was also coimmunoprecipitated with GluK2. (Figure 2.5, compare lanes 6 and 7, immunoblotted for

Neto2). Altogether, these results indicate that the interaction between Neto2 and GluK2-KARs does not require the binding of the Neto2 PDZ ligand to GRIP(PDZ4-7).

To define the region of Neto2 that mediates the interaction with KARs, I examined the binding of GluK2-containing KARs to a series of Neto2 deletion proteins. The Neto2 variant lacking the entire cytoplasmic domain was still able to coimmunoprecipitate with GluK2 (Figure

2.6), suggesting that this domain is not critical for binding to KARs. Removal of the first CUB domain also failed to abolish Neto2:GluK2 interactions (Figure 2.6), whereas deletion of the second CUB domain significantly reduced the amount of GluK2 that was coimmunoprecipitated (33% ± 10% of full length Neto2, p< 0.01; mean±SEM) (Figure 2.6).

Previous studies by Zhang et al. reported that the LDLa domain of Neto2 was necessary for modulating the channel activity of GluK2-KARs based on the observation that a mutant Neto2, in which two cysteine residues in the LDLa domain were changed to serines, failed to enhance glutamate-evoked KAR currents (Zhang et al., 2009). To test whether the LDLa domain of

Neto2 is required for binding to GluK2, I generated a Neto2 construct lacking the LDLa sequence. I found that the absence of the LDLa domain did not diminish the interaction between Neto2 and GluK2 (Figure 2.6) indicating that while the LDLa domain of Neto2 may be required for modulation of GluK2 channel function, it is not necessary for Neto2 to interact with

GluK2 homomeric receptors. Taken together, these results demonstrate that Neto2 binds to

GluK2-KARs primarily through the second CUB domain.

2.3.5. Neto2 forms a ternary complex with GluK2-KARs and GRIP

Given that Neto2 is able to interact with the GluK2 subunit of KARs through its extracellular CUB domains, and that both GluK2 and Neto2 can bind to GRIP(PDZ4-7) through their C-terminal PDZ ligand, I next examined whether coexpression of all three proteins leads to a competitive interaction or to the formation of a ternary complex. COS-7 cells were transfected with various combinations of expression plasmids FLAG-GluK2, Neto2, and GRIP(PDZ4-7)-myc, and cell lysates were immunoprecipitated with an anti-GluK2 antibody. As shown in Figure 2.5, the fraction of Neto2 that could be coimmunoprecipitated with GluK2 was not altered by coexpression with GRIP(PDZ4-7). Thus, GRIP(PDZ4-7) does not compete or interfere with the

Neto2:GluK2 interaction which occurs via the Neto2 ectodomain. On the other hand, when

GluK2 and GRIP(PDZ4-7) were coexpressed with Neto2, I observed a substantial increase in the amount of GRIP(PDZ4-7) that coimmunoprecipitated with GluK2 (250%±41% of signal in the absence of Neto2, p<0.05; mean±SD) (Figure 2.5, compare lanes 5 and 6, GRIP immunoblot).

In contrast, coexpression of the C-terminal deletion mutant Neto2∆7 had no effect on the amount of GRIP(PDZ4-7) that coimmunoprecipitated with GluK2 (103%±26% of signal in the absence of Neto2, p>0.05; mean±SD), indicating that the C-terminal PDZ binding motif of

Neto2 is required for the increased association of GluK2 and GRIP (Figure 2.5, compare lanes 5 and 7, GRIP immunoblot). Given that full length Neto2 does not alter total GRIP(PDZ4-7) protein levels, it is likely that the increase in the fraction of GRIP that is complexed with GluK2 occurs either 1) by an indirect binding of GRIP to GluK2 through Neto2, or 2) by Neto2 stabilizing existing interactions between GRIP with GluK2, or both. In any case, the results obtained from these coimmunoprecipitation experiments indicate that Neto2, GRIP, and GluK2 can form a ternary protein complex when coexpressed within heterologous cells.

2.3.6. Loss of Neto2 does not alter the synaptic abundance of KARs in the hippocampus

Previous studies have proposed a role for GRIP in anchoring KARs to the postsynaptic membrane in the hippocampus (Hirbec et al., 2003). Given that Neto2 can bind to both GRIP and KARs, and that similar ion channel-associated, CUB domain proteins such as Neto1 and

LEV-10 can regulate the synaptic localization of their respective receptors (Gally et al., 2004;

Ng et al., 2009), I next asked whether Neto2 influences the overall abundance of postsynaptic

KARs. To address this question, I isolated hippocampal PSDs from wild-type and Neto2-null mice, and quantified relative protein levels by densitometry analysis of immunoblots. As shown in Figure 2.7, the abundance of KARs, or any other synaptic proteins examined, including Neto1, was not significantly different between wild-type and Neto2-null PSD samples. To determine if loss of Neto2 affected the total levels of KARs, I quantified immunoblots of synaptic proteins from hippocampal homogenates. Again, all the proteins examined were present at similar levels in wild-type compared to Neto2-null mice (Figure 2.7). Therefore, based on biochemical analysis of the hippocampus, I conclude that loss of Neto2 does not perturb the expression or synaptic localization of KARs.

2.3.7. KAR synaptic transmission at MF-CA3 synapses is normal in Neto2-null mice

Previous studies have shown that Neto2 alters the kinetics of GluK2-KAR currents in heterologous expression systems (Zhang et al., 2009). Having established that Neto2 is associated with native KARs at the PSD, we next asked whether Neto2 is involved in regulating

KAR-mediated synaptic transmission. In the hippocampus, Neto2 is expressed in pyramidal cells of the CA region but is absent from the dentate gyrus (DG) (Michishita et al., 2004).

Consistent with the in situ localization data (Michishita et al., 2004), strong Neto2- immunoreactivity was seen in the stratum lucidum (Figure 2.8A), a thin hippocampal layer occupied by MF-CA3 synapses. Given these observations, we examined KAR function at MF-

CA3 connections where the contribution of these receptors to postsynaptic currents have been well characterized by previous studies (Castillo et al., 1997; Vignes and Collingridge, 1997;

Mulle et al., 1998; Marchal and Mulle, 2004).

To reliably evoke KAR-mediated events in CA3 pyramidal neurons on acute hippocampal slices, we used brief trains (4 pulses at 20 Hz) of MF stimulation and measured the amplitudes of EPSCs associated with the 4th pulse (Castillo et al., 1997; Vignes and

Collingridge, 1997; Marchal and Mulle, 2004). We initially obtained a MF input by monitoring the combined AMPAR- and KAR-mediated EPSC while holding the postsynaptic pyramidal cell at Vh=-70 mV, then applied GYKI 53655 (50 µM) to pharmacologically isolate KAR- mediated events (Figure 2.8B). This approach allowed us to control for slice-to-slice variability in MF recruitment by normalizing the KAR-mediated response to that of the initially observed

AMPAR dominated EPSC in the same recording prior to GYKI 53655 treatment. As shown in

Figure 2.8C, the ratio of KAR-mediated EPSCs to that of AMPAR-dominated EPSCs was

indistinguishable between wild-type and Neto2-null mice, suggesting that at these synapses the abundance and/or the channel function of postsynaptic KARs are not affected by the loss of

Neto2. Moreover, despite the ability of Neto2 to slow the desensitization and deactivation of recombinant GluK2-KARs in heterologous expression systems (Zhang et al., 2009), we did not observe faster decay kinetics of KAR-mediated EPSCs in Neto2-null mice compared to wild- type mice (Figure 2.8D). Together, these results show that at MF-CA3 synapses, Neto2 is not required for regulating KAR-mediated synaptic transmission.

2.3.8. Synaptic abundance of KARs is reduced in the cerebellum of Neto2-null mice

In addition to the hippocampus, KARs are abundantly expressed in the cerebellum, and in particular in the granule cell layer (GCL) (Wisden and Seeburg, 1993; Petralia et al., 1994).

Given that in situ studies also showed prominent Neto2 expression in the GCL (Michishita et al.,

2004; Ng, 2006), I examined whether Neto2 could mediate the synaptic expression of cerebellar

KARs. To define the cellular localization of the Neto2 protein, we performed immunofluorescent staining of cerebellar sections. As shown in Figure 2.9A, the GCL, as detected by staining with the neuronal nuclear antigen, NeuN, exhibited the strongest Neto2 immunoreactivity and displayed no obvious differences in thickness or cell density between wild-type and Neto2-null sections. Higher magnification images revealed that, within this layer,

Neto2-positive structures had irregular shapes and were present in nuclear-free islets, which suggest an accumulation of Neto2 in the cerebellar glomeruli. No staining for Neto2 was observed in the Purkinje cell layer, which corresponds to the cell bodies of Purkinje cells, but a diffuse signal could be detected in the molecular layer.

To characterize the type of KARs associated with Neto2 in the cerebellum, I performed coimmunoprecipitation experiments from cerebellar membrane fractions. Granule cells of the

GCL predominantly express the GluK2 and GluK5 subunits of KARs (Bahn et al., 1994). As shown in Figure 2.9B, anti-Neto2 antibodies coimmunoprecipitated both the GluK2 and GluK5 subunits from wild-type but not Neto2-null samples. However, the fraction of total GluK2 input that coimmunoprecipitated with Neto2 was consistently greater than the fraction of total GluK5 input (compare IP vs. input lanes in Neto2+/+ samples). Given that GluK2 protein in the

cerebellum is ~10 times that of GluK5 (Ripellino et al., 1998), it can be inferred from the coimmunoprecipitation results that Neto2 is mostly associated with GluK2-homomeric KARs.

To investigate whether Neto2 plays a role in the synaptic localization of KARs, I isolated cerebellar PSDs from wild-type and Neto2-null mice and quantified relative protein levels by densitometry analysis of immunoblots. In Neto2-null PSDs I observed a significant reduction of GluK2 KAR subunits when compared to wild-type mice (56%±9% of wild-type; mean±SD, n=3; p<0.01), whereas the abundance of other synaptic proteins tested were all similar between the two genotypes (Figure 2.10A). To determine whether the expression of

Neto1 was increased in the absence of Neto2, I evaluated Neto1 protein levels in PSD samples by immunoblot analysis. Neto1 could not be detected in cerebellar PSDs of either wild-type, nor Neto2-null samples (data not shown), suggesting that there is no compensatory upregulation of Neto1 in Neto2-null mice and that Neto1 is not responsible for the synaptic localization of the remaining KARs.

I also evaluated the distribution of GluK2-KARs in the cerebellum of wild-type and

Neto2-null mice by immunofluorescence staining. Cerebellar sections were double-stained with anti-GluK2 and anti-PSD95 antibodies. The most intense GluK2 immunoreactivity in both wild-type and Neto2-null sections was found in clusters, within the granule cell layer, that were also positive for PSD95. These brightly stained structures likely correspond to the granule cell glomeruli, where mossy fiber and Golgi cell terminals synapse onto granule cell dendrites. In

Neto2-null sections, though there was no obvious change in the number or size of GluK2- immunopositive clusters, their relative fluorescence intensity (GluK2/PSD95) was reduced by

20% (n=3; p<0.05) (Figure 2.10B). This result is consistent with the observed decrease in

GluK2 protein levels in Neto2-null PSDs. To determine whether the reduction of GluK2-KARs in the PSD was the result of changes in the amount of total GluK2 protein, I compared GluK2 levels in wild-type and Neto2-null cerebellar homogenates. Immunoblot analysis of GluK2 and

other cerebellar proteins showed that their abundance was not different between wild-type and

Neto2-null mice (Figure 2.10A). Taken together, these results indicate that in the cerebellum,

Neto2 does not affect overall KAR protein levels but is required for maintaining the abundance of these receptors at the PSD.

2.4. Discussion

We have examined the role of Neto2 in the mammalian nervous system. Neto2 is a

CUB-domain containing synaptic protein expressed in the brain. Biochemical analyses revealed an interaction between Neto2 and the seven PDZ domain protein GRIP. Although GRIP serves as a scaffolding protein for synaptic AMPARs and KARs, we found that Neto2 is associated only with KARs. Neto2 can form a ternary complex with GRIP and GluK2-KARs in which the extracellular CUB domain interacts with GluK2 while the intracellular class II PDZ motif binds to GRIP. Contrary to previous in vitro studies in heterologous cells suggesting a role for Neto2 in slowing the decay kinetics of GluK2-KARs, we did not observe any change in the amplitude or decay kinetics of KAR-mediated EPSCs at MF-CA3 synapses of the Neto2-null hippocampus.

Furthermore, loss of Neto2 did not affect the overall synaptic levels of hippocampal KARs. On the other hand, in the cerebellum, where both Neto2 and KARs are abundantly expressed, we found a ~40% reduction in GluK2-KARs at the PSD without any change in the synaptic levels of AMPARs or NMDARs. Our finding that Neto2 regulates the synaptic abundance of KARs is consistent with recent observations made by Copits et al. using hippocampal cultured neurons.

In that system, coexpression of Neto2 with the GluK1 subunit of KARs greatly enhanced the accumulation of GluK1 in dendritic spines and its colocalization with the synaptic marker

PSD95 (Copits et al., 2011). Though we have not yet explored the effect of Neto2 deletion on the channel properties of cerebellar KARs, our results clearly demonstrate an essential role for

Neto2 in regulating the synaptic localization of these receptors and support our conclusion that

Neto2 is an integral component of native KAR complexes.

100

Many PDZ domain-containing molecules are involved in the trafficking, clustering and synaptic localization of ion channels and receptors. PSD95, for example, targets AMPARs to synapses by binding to the AMPAR auxiliary protein stargazin/ɤ2 (Chen et al., 2000; Schnell et al., 2002), while NHERF/EBP50 binds directly to ß2 adrenergic receptors and regulates their plasma membrane recycling (Cao et al., 1999). GRIP and PICK1 have been proposed to stabilize KARs at the postsynaptic membrane, since disruption of either GRIP or PICK1 binding leads to a reduction in the number of functional synaptic KARs (Hirbec et al., 2003). I have found through yeast two-hybrid and in vivo interaction studies that Neto2 also binds to GRIP through a C-terminal PDZ motif. Moreover, coexpression of GRIP, GluK2, and Neto2 in a mammalian system showed that while GRIP does not disrupt or enhance the interaction between

Neto2 and GluK2 KARs, the addition of Neto2 actually increases the fraction of GRIP that coimmunoprecipitates with GluK2. Based on these results, I propose a model in which Neto2 interacts simultaneously with GluK2 and GRIP in a ternary complex: extracellularly with GluK2 and intracellularly with GRIP. Furthermore, the observation that coexpression of Neto2 causes more GRIP to co-purify with GluK2 suggests that Neto2 helps to stabilize the GluK2-GRIP interaction by binding to a different PDZ domain of the GluK2-bound GRIP molecule, and/or that Neto2 brings more GRIP molecules into the Neto2-GluK2 complex. In the latter case, separate GRIP molecules may bind to Neto2 and GluK2; alternatively, given that GRIP molecules can multimerize (Dong et al., 1999), each Neto2-, and GluK2-bound GRIP could form homomers that further stabilize the entire protein complex. Molecular structure studies and finer mapping of PDZ domain interactions will be required to distinguish between these possibilities and to determine the stoichiometry of the Neto2-GRIP-GluK2 protein complex.

Either way, our results, as well as the overlapping distribution of GluK2, Neto2 and GRIP in the

101

brain provide strong evidence that the three proteins are associated within the same complex in vivo. Moreover, the disruption of this complex in the Neto2-null cerebellum may underlie the reduction of postsynaptic GluK2 levels seen in the absence of Neto2.

Overexpression studies in heterologous systems showed that Neto2 slows the desensitization and deactivation of KARs (Zhang et al., 2009; Copits et al., 2011; Straub et al.,

2011a). It is not known, however, whether Neto2 modulates the channel properties of native

KARs. Based on the 40% reduction of synaptic GluK2 levels in the cerebellum of Neto2-null mice, we propose that Neto2 is a regulatory subunit of KARs in this region of the brain. As described in Chapter 1, the cerebellar cortex is divided into three morphologically distinct layers.

The innermost granule cell layer (GCL) is packed with small cerebellar granule neurons (CGN) which, within the GCL, form excitatory synapses with mossy fibers. The axons of the CGNs, referred to as parallel fibers, extend into the outermost molecular cell layer (MCL) where they innervate Purkinje cells dendrites. Between the MCL and the GCL lies the Purkinje cell layer

(PCL), composed of a single layer of evenly spaced Purkinje cells bodies. In situ analysis shows that both Neto2 and GluK2 are strongly expressed in CGNs (Bahn et al., 1994; Ng, 2006), while Neto2 (Ng, 2006), but not GluK2 (Wisden and Seeburg, 1993; Bahn et al., 1994), expression can also be detected in Purkinje cells. In accordance with in situ data, Neto2 and

GluK2-immunoreactivity is highest in the GCL with only light to moderate staining in the MCL

(Petralia et al., 1994) (Figure 2.10). Given the enrichment of Neto2 and GluK2 in the PSD

(Hirbec et al., 2003; Zhang et al., 2009) (Figure 2.1) and their abundant expression in CGNs, it is likely that the immunofluorescent signal in the GCL arises predominantly from CGN cell bodies and dendrites. In fact, strong immunoreactive signal for GluK2 has been reported in the postsynaptic membranes of CGNs in contact with mossy fibers (Jaarsma et al., 1995). Based on

102

these observations, future studies exploring the regulatory effects of Neto2 on KARs should, in principle, examine KAR EPSCs at mossy fiber-CGN synapses. Unfortunately, despite the high levels of GluK2 mRNA in CGNs (Wisden and Seeburg, 1993; Bahn et al., 1994) and the preferential binding of [3H] kainate over the GCL (Monaghan and Cotman, 1982), there have been no reports of KAR-mediated transmission at these synapses. In the cerebellum, KAR- mediated EPSCs have only been recorded at the climbing fiber (CF)-Purkinje cell synapse

(Huang et al., 2004) and at the parallel fiber (PF)-Golgi cell synapse (Bureau et al., 2000). In

Purkinje cells, however, KAR currents display fast decay kinetics similar to those of AMPARs, suggesting that these KARs might function independently of Neto2. In Golgi cells, on the other hand, while KARs may be regulated by auxiliary subunits as suggested by the slower kinetics of their EPSCs, Neto2 does not appear to be expressed. Neto1 mRNA, in contrast, has been detected in these neurons by the translating ribosome affinity purification (TRAP) method

(Doyle et al., 2008). Whether Neto1 affects the function of these native KARs, however, remains to be determined.

The absence of an effect of Neto2 deletion on hippocampal KARs was unexpected, given that Neto2 is expressed in hippocampal pyramidal neurons and is able to alter the kinetics of recombinant KARs in heterologous systems (Zhang et al., 2009; Copits et al., 2011; Straub et al., 2011a). One explanation for this unexpected result is that hippocampal KARs may be regulated by the closely homologous protein Neto1. It has been shown previously that both

Neto1 and Neto2 can enhance the glutamate-evoked currents of recombinant KARs in vitro

(Zhang et al., 2009; Copits et al., 2011). Although the currents generated by the coexpression of

Neto1 with GluK2 were only ~1/25 of those resulting from coexpression of Neto2, in situ hybridization suggests that the Neto1 protein is expressed at much higher levels, in particular in 103

CA3 pyramidal neurons (Ng, 2006; Ng et al., 2009). Consequently, Neto1 may be the predominant KAR auxiliary subunit in the hippocampus. Alternatively, Neto1 and Neto2 may have redundant roles in the hippocampus, allowing Neto1 to compensate for the loss of Neto2.

In either case, analysis of KAR function in Neto1-null and Neto1/Neto2 double null mice is likely to distinguish between these possibilities.

Neto1 has been shown to be an NMDAR-associated protein involved in the stability and/or delivery of GluN2A-containing NMDARs in the hippocampus (Ng et al., 2009). Neto2, on the other hand, regulates the postsynaptic abundance of cerebellar KARs. Given that both

Neto1 and Neto2 interact with their respective ion channels through their well conserved extracellular CUB domains (63% identity for CUB1; 72% for CUB2), I asked whether Neto2 could also be associated with NMDARs in vivo, and Neto1 with KARs. As shown in Appendix

C, anti-Neto2 antibodies coimmunoprecipitated the GluN1, and GluN2A subunits from whole brain synaptosomes, suggesting that Neto2 can associate with NMDAR protein complexes as well as with KARs. Under similar conditions, I also observed coimmunoprecipitation of Neto1 with GluK2-containing KARs (please refer to Figure 3.1 in Chapter 3). This result is consistent with the previously reported role of Neto1 as a modulator of recombinant GluK2-KARs (Zhang et al., 2009), and suggests that native KARs may be regulated by Neto1 as well as Neto2. The functions of Neto1 on KARs in vivo will be discussed in the next chapter.

Two other CUB domain-containing proteins that associate with ligand-gated ion channels are the C. elegans proteins SOL-1 (Zheng et al., 2004; Zheng et al., 2006), and LEV-

10 (Gally et al., 2004), which are part of GLR-1 AMPA receptors and acetylcholine receptors, respectively. SOL-1 modulates the kinetics of GLR-1 AMPARs while LEV-10 controls the

104

synaptic localization of the acetylcholine receptors. Collectively, these findings support the hypothesis that CUB domain proteins have conserved roles as regulatory subunits of ion channels (Ng et al., 2009), and suggest that other CUB domain proteins may be important accessory proteins for other ligand-gated ion channels.

In addition to KARs, Neto2 has also been shown to interact, through its CUB domains, with KCC2, the neuron-specific K+Cl- cotransporter responsible for Cl- extrusion. In the mature nervous system, low levels of intracellular Cl-, maintained by KCC2, are required for inhibitory synaptic transmission as they allow the influx of Cl- during GABAR activation (Rivera et al.,

1999). Loss of Neto2 impairs GABAR-mediated synaptic inhibition, as it reduces the efficacy of KCC2 Cl- extrusion and decreases overall KCC2 protein levels (Ivakine et al., 2012, submitted). The regulation of an ion channel (KAR) and a cotransporter (KCC2), two completely diverse classes of molecules, by a single transmembrane auxiliary protein, in this case Neto2, has not been previously reported. This finding raises the interesting possibility that other CUB-domain containing proteins may also regulate multiple types of membrane proteins.

In conclusion, we have shown that Neto2 is a regulatory protein of native, synaptic

GluK2-containing KARs. Although the loss of Neto2 did not affect KAR-mediated EPSCs at hippocampal MF-CA3 synapses, or alter synaptic KAR levels in hippocampal PSDs, we found that cerebellar postsynaptic KARs were significantly reduced in the absence of Neto2. Future studies will be needed to establish whether Neto2 also modulates KAR channel properties in cerebellar synapses as it does in heterologous systems. We also discovered that Neto2 can form a ternary complex with GluK2-KARs and its scaffolding protein GRIP. Although it remains to be determined whether the synaptic reduction of KARs in Neto2-deficient cerebellum is the

105

result of decreased KAR-GRIP interactions in the absence of Neto2, our findings demonstrate a critical role for Neto2 in controlling the synaptic accumulation of KARs.

106

Chapter 3: Neto1 is an auxiliary subunit of native synaptic kainate receptors

**This work was the result of a collaborative effort with Dr. Kenneth Pelkey from the lab of Dr.

Chris McBain. While the majority of the work presented here is my own, Dr. Pelkey assisted in generating the data contained in figures 3.4 and 3.5.

Data presented in Figures 3.1-3.7 have been published in J Neurosci 31: 10009-10018

107

3.1. Introduction

Excitatory synaptic transmission in the mammalian central nervous system is primarily mediated by the neurotransmitter glutamate. Pharmacological, biophysical, and molecular studies have classified the ionotropic receptors, to which glutamate binds, into AMPA-, NMDA-, and kainate-sensitive glutamate receptors. AMPARs mediate most rapid glutamatergic transmission, while NMDARs are recruited with increased neuronal activity through relief of voltage-dependent Mg2+ blockade, allowing them to serve as coincidence detectors to gate synaptic plasticity induction. The roles of KARs are less well understood, but are thought to involve modulation of synaptic plasticity and neuronal excitability (Traynelis et al., 2010;

Contractor et al., 2011, review).

In heterologous expression systems, glutamate receptors alone can form functional ion channels at the cell membrane. In the brain, however, native receptors do not work in isolation, but are components of large, dynamic multiprotein complexes. Within such a complex, receptor-associated proteins critically influence synaptic transmission by regulating one or more properties of the receptor, including its surface expression, subcellular distribution, recycling, degradation, or gating kinetics (Jackson and Nicoll, 2011). Interestingly, the majority of these regulatory proteins are intracellular molecules that bind to the receptors’ cytoplasmic domain, and it wasn’t until recently that transmembrane auxiliary proteins have been identified for all three glutamate receptors.

The TARP family of proteins were the first transmembrane proteins found to be associated with an ionotropic glutamate receptor, the AMPA receptor. TARPs have been shown

108

to affect AMPAR trafficking and kinetics (Letts et al., 1998; Hashimoto et al., 1999; Chen et al.,

2000; Tomita et al., 2003). Subsequently, the CUB domain-containing protein Neto1 was identified as the auxiliary subunit of the NMDARs (Ng et al., 2009). Loss of Neto1 led to a selective decrease of GluN2A subunits in hippocampal PSDs leading to a reduction in

NMDAR-mediated currents and impaired LTP at Schaffer collateral-CA1 synapses (Ng et al.,

2009). More recently, studies in heterologous cells have revealed that Neto1 and Neto2 can modulate the function of GluK2-, and GluK1-homomeric KARs, though the effects exerted by both Netos can be qualitatively and quantitatively different (Zhang et al., 2009; Copits et al.,

2011). For instance, Neto1 can augment glutamate-evoked currents of GluK2-KARs, but the increase is only ~1/25 of that produced by coexpression of Neto2 (Zhang et al., 2009).

Although we do not yet know whether Neto1 affects native KARs, these results suggest that the same subtype of KARs may have different physiological profiles depending on the Neto with which they are associated.

In the brain, the most common KAR subtype is the heteromeric receptor composed of

GluK2 and GluK5 subunits. While GluK5 is expressed ubiquitously in the brain, GluK2 mRNA is largely restricted to the cerebellar granule cells, the hippocampus (CA1-3 and DG), the pyriform cortex, and the caudate-putamen of the striatum (Wisden and Seeburg, 1993; Bahn et al., 1994). A similar distribution pattern can be seen for kainate binding sites using an in vitro autoradiographic technique (Foster et al., 1981; Monaghan and Cotman, 1982). In Chapter 2, we showed that Neto2 plays a role in the synaptic localization of KARs in the cerebellum. In the hippocampus, however, loss of Neto2 had no effect on postsynaptic KAR abundance or synaptic transmission. Given that Neto1 is highly expressed in the hippocampus, in particular in the CA3 pyramidal neurons (Ng et al., 2009), we hypothesized that Neto1 may be the auxiliary

109

subunit of KARs in the hippocampus. I examined the role of the Neto1 on native KARs in vivo using Neto1-null mice generated in our lab (Ng et al., 2009). In this chapter, I will discuss the biochemical and electrophysiological experiments that have been carried out for this purpose.

Note: Following our in vivo studies of Neto1 and KARs, another report was published describing the effect of the loss of Neto1 on KARs in the hippocampus (Straub et al., 2011b).

110

3.2. Materials and Methods

Mice and PCR genotyping

Neto1-null, Neto2-null, Neto1/Neto2-null and wild-type mice used in this study were previously generated in the lab by David Ng (Ng, 2006 Ph.D. thesis). The Neto1 and Neto2 genes were disrupted by homologous recombination using a targeting vector with a tau-lacZ- loxP-pgk-neo-loxP cassette cloned in-frame with the Neto1 start codon, or a loxP-pgk-neo-loxP cassette (for Neto2). All the animals have been maintained at the Toronto Center for

Phenogenomics (TCP).

Neto1 and Neto2 genotyping was performed using PCR. Each PCR reaction contained

50-100 ng of each primer, 8 ul Qiagen Multiplex PCR mix, 1-10 ng of DNA template in a final volume of 20 ul. Samples were first heated to 95 oC for 15 min followed by 35 thermal cycles consisting of a short denaturation at 94 oC for 30 sec, 56 oC annealing for 90 sec, and 72 oC extension for 1 min. After the last cycle, the samples were subjected to a 72 oC final extension for 10 min, and were stored at 4 oC. PCR products were analyzed using 1% agarose gels. For

Neto1 genotyping, the following primers were used in the PCR reaction: mRtl1-5UTR-F (5’

AGA TCG GAG CCT CTG GTG TAA C 3’), mRtl1-Intron-R (5’ GGA TTA CGT GAA TCT

CTT AAC TG 3’), and pcDNA3tau-R (5’ TTA CTG ACC ATG CGA GCT TG 3’). For Neto2 genotyping, the following primers were used: mRtl2-2Larm-F2 (5’ GTA GGT ATA GGT AGG

ATG GTT 3’), mRtl2-intron-R (5’ GCA GAA GTA CCA GAA AGC 3’), and DTA-R2 (5’

CTA GTG AGA CGT GCT ACT TC 3’).

111

Antibodies

The following antibodies were used: rabbit polyclonal antibodies to GluK2 (Abcam),

GluK5 (Millipore), GluA2/3 (Millipore), GluN2B (Novus Biologicals), and actin (Abcam); mouse monoclonal antibodies to GluN1 (BD Biosciences), VAMP2 (Synaptic Systems), NeuN

(Millipore), and HA (Covance). Guinea pig polyclonal anti-Neto1 and rabbit polyclonal anti-

Neto2 antibodies were generated and purified in-house.

Mammalian expression constructs

Full-length Neto1 and Neto2 cDNA (encoding amino acids 1-533, and 1-525, respectively), deletion mutants Neto1-∆CUB1, Neto1-∆CUB2, Neto1-∆CUB1+2, Neto1-∆cyto, and full length mouse PICK1 were generated by PCR and subcloned into a variant of pcDNA3.1mycHisA(+) (Invitrogen) containing two copies of the influenza hemagglutinin (HA) epitope tag, and sequence verified. FLAG-GluK2 was a kind gift from Dr. Katherine Roche

(National Institutes of Health, Bethesda, Maryland, USA).

SDS-PAGE and immunoblot analysis

Protein samples were separated on denaturing SDS-PAGE gels using standard methods.

Samples were boiled for 5 min, or incubated at 50 oC for 20 min (PSD samples only) with the appropriate volume of 6X sample buffer (0.375 M Tris/HCl [pH 6.8], 60% (v/v) glycerol, 12%

SDS, 0.06% bromophenol blue, and 0.6 M DTT), loaded on the gel, and electrophoresed in

SDS-PAGE running buffer (192 mM glycine, 25 mM Tris/HCl [pH 8.3] and 0.1% SDS) for 90 min at 140 V. Protein samples separated on the gel were transferred onto Hybond-C Extra 112

nitrocellulose membranes (GE Healthcare) at 40V in transfer buffer (192 mM glycine, 25 mM

Tris/HCl [pH 8.3], and 20% methanol). Following overnight transfer at 4 oC, membranes were briefly stained with Ponceau S solution (0.1% (w/v) Ponceau S, 5% acetic acid) to confirm successful protein transfer and to locate protein bands of interest. To proceed with immunoblotting, membranes were rinsed with distilled water to remove the Ponceau stain, and were blocked for 1 h at room temperature with 5% skim milk powder dissolved in TBS-T (100 mM Tris/HCl [pH 7.5], 150mM NaCl, 0.1% Tween-20), followed by overnight incubation at

4 oC with primary antibody. Membranes were washed four times with TBS-T (10 min per wash), and incubated with the appropriate horseradish peroxidise (HRP)-conjugated secondary antibody for 1 h at room temperature. After treatment with secondary antibody, membranes were washed four times with TBS-T, and proteins to which the primary antibody was bound were detected by enhanced chemiluminescence.

Cell culture and transfection

o HEK293 or COS-7 cells were maintained at 37 C, 5% CO2 in Dulbecco’s Modification of Eagle’s Medium (Wisent) containing 10% fetal bovine serum (Wisent). For coimmunoprecipitation experiments, cells were cultured on 6-well plates and transfected with the appropriate constructs using FuGene HD (Roche) at 70% confluency. Forty-eight hours after transfection, cells were washed once with ice-cold PBS and lysed in 300 ul of RIPA buffer (50 mM Tris/HCl [pH7.4], 150 mM NaCl, 1 mM EDTA, 1% Nonidet P-40, 0.5% deoxycholic acid

(DOC), and 0.1% SDS) supplemented with Complete® Protease Inhibitor Cocktail tablets

(Roche). Lysed cells were scraped off the well and transferred into individual microfuge tubes.

Samples were incubated on ice for 30 min and centrifuged at 13,000 x g for 15 min at 4 oC. The

113

protein concentration of the supernatant was determined using the detergent-compatible DC

Protein Assay according to the manufacturer’s protocol (Bio-Rad). Quantified samples were stored at -80 oC or were used immediately for coimmunoprecipitation experiments.

Co-immunoprecipitation

For coimmunoprecipitation experiments using transfected HEK293 or COS-7 cells,

~0.25mg of cell lysates were incubated with antibodies for 2 h at 4 oC on a rotating platform.

Lysates were subsequently incubated with 20 ul GammaBind IgG beads (GE Healthcare) for 1 h at 4 oC on a rotating platform. For coimmunoprecipitation from crude synaptosomal fractions, 1 mg of synaptosomal protein was incubated with antibodies or normal rabbit IgGs overnight at 4 oC on a rotating platform, and subsequently incubated with 30 ul of GammaBind IgG beads for

2 h with rotation at 4 oC. Beads were washed twice with 1ml ice-cold RIPA buffer, twice with

RIPA minus SDS and DOC, and once with TBS-T. All buffers were supplemented with

Complete® Protease Inhibitor Cocktail tablets (Roche), and all washes were carried out for 10 min at 4 oC on a rotating platform. Bound proteins were eluted with 6X sample buffer and subjected to SDS-PAGE and immunoblotting.

Preparation of hippocampal homogenates

One pair of hippocampi was placed in a glass Teflon homogenizer containing 2 ml of ice-cold RIPA buffer supplemented with Complete® Protease Inhibitor Cocktail tablets (Roche).

The tissue was homogenized on ice using 20 up and down strokes at 700 rpm. Homogenized samples were incubated on ice for 15 min and centrifuged at 13,000 x g for 15 min. The

114

supernatant was isolated, quantified using the detergent-compatible DC Protein Assay (Bio-

Rad), and stored at -80 oC.

Purification of crude synaptosomes and PSDs

To prepare crude synaptosomes, mouse brain tissue was homogenized on ice in a glass

Teflon homogenizer (700 rpm, 20 up and down strokes) containing sucrose buffer (320 mM sucrose, 10 mM EDTA, and 10 mM Tris/HCl [pH 7.4]). The homogenates were centrifuged at

1000 x g for 15 min at 4oC. The pellet was discarded while the supernatant was centrifuged at

10,000 x g for 15 min at 4oC. The supernatant from the last centrifugation was removed, and the pellet was solubilized in DOC buffer (50 mM Tris/HCl [pH 9.0], and 1% DOC) at 37 oC for

30 min. The solubilised sample was centrifuged at 100,000 x g for 15 min at 4 oC. The supernatant was carefully isolated and an equal volume of modified RIPA buffer (50 mM

Tris/HCl [pH 7.4], 150 mM NaCl, 1 mM EDTA, and 1% Triton X-100) was added to it. The sample was quantified using detergent-compatible DC Protein Assay (Bio-Rad) according to protocol, and stored at -80 oC. All the buffers used in this protocol were supplemented with

Complete® Protease Inhibitor Cocktail tablets (Roche).

To isolate the PSD fraction, whole brain or pooled hippocampi tissue was homogenized in solution A (0.32 M sucrose, 1 mM NaHCO3, 1 mM MgCl2, 0.5 mM CaCl2) using a glass

Teflon homogenizer (20 up and down strokes, 700 rpm). The homogenate was then diluted to

10% (w/v) in Solution A and centrifuged at 1400 x g for 10 min at 4 oC. The supernatant (1st sup) is transferred into a clean tube and the pellet is resuspended in solution A (10 ml of solution A per 1 g of initial tissue) with three up and down strokes in the homogenizer. The resuspended

115

pellet is centrifuged at 710 x g for 10 min at 4 oC. The supernatant (2nd sup) is collected and mixed with the supernatant from the first centrifugation (1st sup). The combined supernatant is centrifuged at 13,800 x g for 10 min at 4 oC and the resulting pellet was resuspended in solution

B (0.32 M sucrose, 1 mM NaHCO3) with 6 up and down strokes in a homogenizer. A sucrose gradient was prepared by gently layering the following solutions in an SW-28 centrifuge tube starting with 10ml of 1.2 M sucrose/1 mM NaHCO3, followed by 10ml of 1.0 M sucrose/1 mM

NaHCO3, and 10ml of 0.85M sucrose/1 mM NaHCO3. The resuspended pellet was slowly layered onto the top of the sucrose gradient to avoid disturbing the layers, and the entire gradient was centrifuged at 82,500 x g for 2 h. After centrifugation the cloudy band, (between the 1.2 M and the 1.0 M sucrose layers) which contained the synaptic membranes, was carefully isolated and diluted with solution B. To separate the PSD from other detergent soluble membrane fractions, a Triton extraction was carried out by adding an equal volume of solution C (1%

Triton X-100, 12 mM Tris/HCl [pH 8.0], and 0.32 M sucrose) to the diluted synaptic membranes. The mixture was incubated at 4 oC for 30 min with end-over-end rotation and centrifuged at 32,000 x g for 20 min at 4 oC. The supernatant was removed without disturbing the white thin pellet (PSD), and discarded. The pellet was resuspended in PBS and stored at -80 oC.

For immunoblot analyses, the PSD pellet was solubilised with SDS. Briefly, the PSD in

PBS suspension was centrifuged at 13,000 x g for 20 min at 4 oC. The supernatant was discarded and the pellet was resuspended in 40 mM Tris/HCl [pH 8.0] containing 1% SDS and

10 mM DTT followed by incubation at 55 oC for 20 min. The solubilised pellet was diluted with modified RIPA buffer without SDS (50 mM Tris/HCl [pH 7.4], 150 mM NaCl, 1 mM EDTA, 1%

NP-40, and 0.25% DOC), and stored at -80 oC for further use. 116

For coimmunoprecipitation experiments, the PSD pellet was solubilised with DOC.

Briefly, the PSD in PBS suspension was centrifuged at 13,000 x g for 20 min at 4 oC and the resulting pellet was resuspended in 50 mM Tris/HCl [pH 9.0] containing 1% DOC. The sample was incubated at 37 oC for 30 min and centrifuged at 100,000 x g for 15 min at 4 oC to remove any insoluble particles. The supernatant was collected and diluted with an equal volume of modified RIPA buffer without DOC (50 mM Tris/HCl [pH 7.4], 150 mM NaCl, 1 mM EDTA, 1%

NP-40, and 0.1% SDS). All the buffers described in the protocol for PSD isolation contained a cocktail of protease inhibitors (Roche).

Immunofluorescence staining of the hippocampus

Immunostaining of hippocampal slices was adapted from Schneider Gasser et al.

(Schneider Gasser et al., 2006). Briefly, fresh 250-μm vibratome-cut hippocampal slices, trimmed from sagittal brain slices, were placed in 6-well plates and fixed in 2% paraformaldehyde/PBS on ice for 10 min. Slices were then washed three times with PBS, transferred into 24-well plates (1-2 slices per well), and incubated “free-floating” in blocking solution (10% goat serum, 0.1% Triton-X in PBS) for 1 h at room temperature. Slices were then incubated with primary antibodies in a humidified chamber for 16-24 h at 4 °C with gentle agitation. After overnight incubation, slices were transferred back into 6-well plates to be washed three times with PBS, 10 min for each wash. Subsequently, slices were incubated with the appropriate secondary antibodies conjugated to Alexa 488 (Molecular Probes), Cy3, or Cy5

(Cedarlane) fluorophores, for 24 h at 4 °C in the dark. Slices were washed again three times with PBS, mounted onto glass slides with Immuno-Mount (Thermo Scientific), and visualized

117

using a Zeiss LSM 510 Laser Scanning Confocal Microscope. Images were acquired using the

LSM 510 software package.

Whole cell recordings

Hippocampal slices for whole cell recordings were prepared as previously described

(Pelkey et al., 2005) using P15-22 wild-type, Neto1-null, Neto2-null, or Neto1/Neto2-null mice as indicated. Briefly, animals were anaesthetized with isoflurane and decapitated allowing removal of the brain into ice-cold saline solution (130 mM NaCl, 24 mM NaHCO3, 3.5 mM KCl,

1.25 mM NaH2PO4, 0.5 mM CaCl2, 4.5 mM MgCl2, and 10 mM glucose, saturated with 95% O2 and 5% CO2 [pH 7.4]). After dissection of the brain, individual hemispheres were transferred to the stage of a VT-1000S vibratome (Leica Microsystems, Bannockburn, IL) and sectioned to yield transverse hippocampal slices (300 µm) which were incubated in the above solution at 35 oC for at least a 30 minute-recovery until use. All animal procedures conformed to the National

Institutes of Health animal welfare guidelines.

All recordings were interleaved with the experimenter blind to mouse genotype.

o with 95% O2 and 5% CO2 [pH 7.4], 32-35 C). Whole-cell patch-clamp recordings using a multiclamp 700A amplifier (Axon Instruments, Foster City, CA) in voltage-clamp mode (Vh=-

70 or +40 mV as indicated) were made from individual CA3 pyramidal neurons, visually identified with infrared video microscopy and differential interference contrast optics.

118

Recording electrodes (4-5 MΩ) pulled from borosilicate glass (World Precision Instruments) were filled with intracellular solution (ICS) composed of 95 mM Cs-gluconate, 5 mM CsCl; 0.6 mM EGTA, 5 mM MgCl2, 4 mM NaCl, 2 mM Na2ATP, 0.3 mM NaGTP, 40 mM HEPES, 10 mM BAPTA, 1 mM QX-314 [pH 7.2-7.3], at 290-300 mOsm. Uncompensated series resistance

(8-15 MΩ was rigorously monitored by the delivery of small voltage steps at regular intervals and recordings were discontinued following changes of >10%). Synaptic responses (paired pulses or trains of 4 pulses, both at 20 Hz) were evoked at 0.1 Hz (for train recordings) or 0.2

Hz (for paired pulse recordings) by low-intensity microstimulation (100 µsec duration; 10-30

µA intensity) via a constant-current isolation unit (A360, World Precision Instruments, Sarasota,

FL) connected to a patch electrode filled with oxygenated extracellular solution in either the dentate gyrus or stratum lucidum for MF inputs, or in the stratum radiatum for associational/commissural (A/C) inputs. Mossy fiber-origin of EPSCs was confirmed by a rapidly rising AMPA receptor-mediated component showing strong short-term frequency facilitation and in train protocols by a residual KAR-mediated component upon AMPA receptor antagonism at a holding potential Vh of -70mV. For MF train recordings, initially dual component KA/AMPAR-mediated synaptic responses were monitored at Vh= -70mV following which the KAR-mediated component was pharmacologically isolated by applying the AMPA receptor-specific antagonist GYKI 53655 (50 µM, Tocris Bioscience). The GYKI resistant component at Vh = -70mV was confirmed to be KAR-mediated by subsequent application of

DNQX (25 µM, Tocris Bioscience) in the continued presence of GYKI 53655 and the holding potential was moved to +40 mV to obtain the NMDA receptor-mediated component of EPSCs followed by application of the NMDA receptor antagonist dl-APV (100 µM, Tocris Bioscience) in the continued presence of GYKI53655 and DNQX. For paired pulse experiments examining

119

just AMPA and NMDA components at MF or A/C inputs, EPSCs were first obtained and monitored at Vh=-70mV, then the NMDA component was monitored at Vh=+40mV in the presence of DNQX and confirmed by subsequent application of dl-APV.

Data analysis for whole cell recordings

To measure AMPA receptor and KAR-mediated EPSCs, averaged traces (10-20 individual sweeps) obtained in GYKI 53655 with DNQX at Vh = -70mV were digitally subtracted from averaged traces obtained at the end of the control and GYKI 53655-alone conditions, respectively. Similarly, for NMDA receptor-mediated EPSC analysis, averaged traces obtained in dl-APV at Vh= +40mV were digitally subtracted from those obtained in

GYKI 53655/DNQX at Vh=+40mV. For each recording, EPSC amplitudes were measured during a 1-2 ms window around the peak of the waveform of the averaged traces for each condition. KAR- and NMDA receptor-mediated EPSC amplitudes were measured for the 4th pulse of the trains and normalized to the amplitude of the corresponding AMPA receptor- mediated EPSC to eliminate slice to slice and animal to animal variability in the number of fibers recruited by extracellular stimulation. Short-term frequency facilitation was assessed using the AMPA or NMDA receptor-mediated EPSCs by determining the ratio of the amplitude of the 4th to the 1st EPSC in the train (P4/P1). In paired pulse recordings examining only the

AMPA and NMDA components of MF and A/C inputs, the amplitudes of the first EPSCs were used to characterize NMDA/AMPA ratios and paired pulse ratios were determined by the ratios of the amplitudes of the second peak to the first peak (P2/P1). Data are presented as means ±

SEMs unless otherwise indicated. Statistical significance was assessed using parametric (paired or unpaired t-tests) or non-parametric (Mann-Whitney U test) tests as appropriate.

120

3.3. Results

3.3.1. Neto1 interacts with native KARs

Given that Neto1 has been shown to enhance glutamate-evoked currents of recombinant

GluK2 homomeric KARs in heterologous cells (Zhang et al., 2009), I asked whether Neto1 is associated with KARs in vivo. Wild-type whole brain crude synaptosomes were incubated with anti-Neto1 antibodies and the coimmunoprecipitation of KAR subunits was detected by immunoblotting. As shown in Figure 3.1A, two of the most abundantly expressed KAR subunits in the brain, GluK2 and GluK5, coimmunoprecipitated robustly with Neto1. Similarly, anti-GluK2 and anti-GluK5 antibodies coimmunoprecipitated Neto1 (Figure 3.1B). Consistent with previously reported observations (Ng et al., 2009), Neto1 was also found to interact with the GluN1 subunit of the NMDARs but not with the GluA2/3 subunit of the AMPARs (Figure

3.1C). Together, these results show that Neto1 is associated with two of the three ionotropic glutamate receptors, NMDARs, and KARs. However, contrary to what has been observed for

KARs, the amount of NMDARs (i.e. GluN1) that coimmunoprecipitated with Neto1 is a relatively small fraction of the total GluN1 input. This finding suggests that the interaction between Neto1 and GluN1 is either easily disrupted by solubilising detergents, or that only a small fraction of total NMDARs in the brain is associated with Neto1.

[3H] kainate binding experiments suggest that KARs are highly enriched in synaptic junctions (Foster et al., 1981), and both electron microscopy and immunoblot analysis of subcellular fractions show that they can be found in the PSD (Petralia et al., 1994; Hirbec et al.,

2003). KARs present on the postsynaptic membrane are activated by synaptically released

121

glutamate and contribute to excitatory postsynaptic transmission (Huettner, 2003). To determine whether Neto1 was associated with KAR protein complexes specifically at the PSD, I examined the coimmunoprecipitation of KARs with anti-Neto1 antibodies in biochemically isolated PSD fractions. I found that anti-Neto1 antibodies coimmunoprecipitated the GluK2 and

GluK5 KAR subunits from PSDs of wild-type mice. However, neither GluK2 nor GluK5 were coimmunoprecipitated from Neto1-null samples. In contrast, no coimmunoprecipitation was observed between Neto1 and the GluA2/3 subunit of AMPARs, indicating a specific interaction of KARs with Neto1 at the PSD (Figure 3.1D).

122

123

To determine whether Neto1 interacts specifically with GluK2, a major subunit of KARs,

I coexpressed GluK2-Flag and Neto1-HA in HEK 293 cells. As shown in Figure 3.2, GluK2 coimmunoprecipitates with full length Neto1, demonstrating that these two proteins do associate.

To define the primary sequence of Neto1 that mediates the interactions with GluK2, I evaluated the binding of the GluK2-KARs to a series of Neto1 deletion proteins in HEK293 cells. A

Neto1 variant lacking the entire cytoplasmic domain was still able to coimmunoprecipitate with

GluK2 (Figure 3.2), suggesting that this domain is not critical for binding to KARs. Removal of the first CUB domain also failed to abolish Neto1:GluK2 interactions (Figure 3.2-B, lane 3), whereas deletion of the second CUB domain of Neto1 significantly reduced the amount of

GluK2 that was co-immunoprecipitated, relative to the full length protein (39% ± 6%, of full length Neto1, p<0.01; mean±SEM); (Figure 3.2-B, lane 4). Furthermore, no interaction with

GluK2 was observed when both extracellular CUB domains of Neto1 were deleted (Figure 3.2-

B, lane 5). Taken together, these results demonstrate that Neto1 binds to KARs through its extracellular CUB domains, and that this interaction is mediated primarily by the second CUB domain.

124

125

3.3.2. Synaptic KAR currents are reduced in Neto1-null mice

Having established that Neto1 is an interacting protein for KARs at PSDs, we next investigated whether Neto1 regulates synaptic KAR function in the brain. Immunofluorescence studies indicated that in the hippocampus, Neto1 is primarily localized to the CA3 stratum lucidum layer (Figure 3.3). This distribution pattern bears striking resemblance to that of hippocampal [3H] kainate binding sites identified by autoradiography (Foster et al., 1981;

Monaghan and Cotman, 1982), and overlaps with the immunostaining pattern of GluK2 and

GluK5 subunits in the brain (Ripellino et al., 1998). Given that Neto1 mRNA is strongly expressed in the hippocampal CA3 and DG region (Ng et al., 2009), the Neto1- immunoreactivity in the stratum lucidum layer suggests a postsynaptic localization of Neto1 to proximal dendrites of CA3 pyramidal cells and/or a presynaptic localization to the terminals of

DG granule cell axons (the mossy fibers). However, based on the enrichment of Neto1 in PSDs

(Ng et al., 2009; Zhang et al., 2009) and its much stronger in situ profile in CA3 pyramidal neurons over DG granule cells (Michishita et al., 2003, 2004; Ng et al., 2009), it is likely that the signal we observed for Neto1 in the stratum lucidum is predominantly postsynaptic.

126

127

The involvement of KARs in synaptic transmission has been well documented at hippocampal MF-CA3 synapses, where postsynaptic KARs composed of mostly GluK2/GluK5 heteromers (Petralia et al., 1994; Mulle et al., 1998; Bureau et al., 1999; Contractor et al., 2003;

Darstein et al., 2003) mediate a small component of the EPSC with slow rise and decay kinetics

(Castillo et al., 1997; Vignes and Collingridge, 1997). In contrast, presynaptic KARs regulate the release of glutamate (Pinheiro and Mulle, 2008, review) and contribute to short-term plasticity (Contractor et al., 2001; Lauri et al., 2001b; Schmitz et al., 2001; Pinheiro et al., 2007) and the induction of LTP (Bortolotto et al., 1999; Contractor et al., 2001; Lauri et al., 2001b;

Schmitz et al., 2003; Pinheiro et al., 2007; Scott et al., 2008).

To determine the effect of Neto1 loss on the postsynaptic function of KARs, we recorded KAR-mediated EPSCs at MF-CA3 synapses in acute hippocampal slices from wild- type and Neto1-null mice. To reliably evoke KAR-mediated events, MFs were stimulated with brief trains of 4 pulses at 20 Hz (Castillo et al., 1997; Vignes and Collingridge, 1997; Marchal and Mulle, 2004). The amplitudes of EPSCs associated with the 4th pulse were measured in

CA3 pyramidal neurons. We first recorded the combined AMPAR- and KAR-mediated EPSC while holding the postsynaptic pyramidal cell at Vh=-70 mV. Subsequently, KAR-mediated

EPSCs were isolated by application of the AMPAR-antagonist GYKI 53655 (50 µM) (Figure

3.4A). For each recording, we calculated the ratio of KAR-mediated response to the initially observed AMPAR-dominated EPSCs in order to control for slice-to-slice variability in MF recruitment and make comparisons across different animals. In mice lacking Neto1, we observed a severe deficit in KAR-mediated EPSCs compared to interleaved recordings from age-matched wild-type mice: KA/AMPA EPSC amplitude ratios for wild-type and Neto1-null mice were 0.065± 0.006, and 0.040± 0.005, respectively (mean±SD, p<0.01) (Figure 3.4B). 128

Moreover, KAR-mediated EPSCs in Neto1-null mice displayed significantly faster decay kinetics compared to wild-type mice (20±1.8 ms, and 50±4.9 ms, for Neto1-null and wild-type mice, respectively; mean±SD, p<0.001) (Figure 3.4C). Together, these results show that Neto1 regulates both the kinetics, and the amplitude of native, synaptic KARs.

In Chapter 2, we showed that loss of Neto2 does not affect KAR-mediated transmission at MF-CA3 synapses. Given that Neto1 contributes significantly to KAR function at these synapses, we next asked whether the normal KAR-mediated EPSCs previously observed in

Neto2-null mice (Figure 2.8B/C) were the result of compensation by Neto1. To address this question, we recorded from Neto1/Neto2-double null mice and compared their KA/AMPA

EPSC ratio and KAR decay kinetics to those of wild-type animals. As shown in Figure 3.4B/C, the combined loss of both Neto proteins did not further exacerbate the phenotype observed in

Neto1 single knockout mice. From this result, we can infer that Neto2 is not likely to be involved in regulating KAR-mediated synaptic transmission at MF-CA3 synapses and that normal KAR function in Neto2-null mice is not due to a compensatory action by Neto1.

129

130

3.3.3. Loss of Neto1 affects NMDAR-mediated currents at A/C-CA3 but not MF-CA3 synapses

Previous studies showed that Neto1-null mice display a preferential reduction of synaptic GluN2A subunits and impaired NMDAR-mediated EPSCs at Schaffer collateral-CA1 synapses (Ng et al., 2009). To determine if MF-CA3 synapses exhibit a similar deficit in

NMDAR-mediated transmission, we probed NMDAR-mediated events at the end of every recording by blocking AMPARs and KARs, then moving Vh to +40 mV (Figure 3.4A).

Surprisingly, all mice examined yielded similar NMDA/AMPA ratios (Figure 3.4B). The comparable NMDA/AMPA ratios observed across all mice confirm that the altered KA/AMPA

EPSC ratios observed in Neto1-null and Neto1/Neto2-double null mice result from impaired

KAR-mediated transmission rather than enhanced AMPAR function. However, the lack of effect on MF-CA3 NMDA/AMPA ratio was unexpected and prompted us to investigate whether the loss of Neto1 affected NMDAR-mediated EPSCs at associational/commissural-CA3 pyramidal cell (A/C-CA3) synapses, which more closely resemble Schaffer collateral-CA1 synapses. Consistent with prior observations in CA1 (Ng et al., 2009), A/C-CA3 synapses exhibited reduced NMDA/AMPA EPSC ratios in Neto1-null neurons (Figure 3.5A/B).

Importantly, in additional interleaved control MF-CA3 recordings, we again failed to observe any effect of Neto1 on NMDA/AMPA ratios (Figure 3.5A/B) confirming that Neto1 regulation of postsynaptic receptor function is synapse-specific.

131

132

3.3.4. Neto1-null mice have normal presynaptic function at MF-CA3 synapses

Activation of presynaptic KARs at MF-CA3 synapses has been reported to facilitate glutamate release from granule cell axon terminals and contribute to short-term synaptic plasticity of mossy fiber excitatory transmission (Contractor et al., 2001; Lauri et al., 2001b;

Schmitz et al., 2001; Pinheiro et al., 2007). Given that Neto1 is also expressed in the DG

(Michishita et al., 2004; Ng et al., 2009), we asked whether Neto1 regulated the function of presynaptic KARs. To assess any changes in presynaptic KAR activity, we calculated short- term frequency facilitation at MF-CA3 synapses by determining the ratio of the last to first

AMPAR-mediated EPSCs. We found that the magnitude of facilitation was not significantly different between wild-type and Neto1-null or Neto1/Neto2-double null mice, suggesting normal presynaptic function across all animals (Figure 3.4D). Similarly, when we examined short-term frequency facilitation using the NMDA component of transmission, we found that it was also indistinguishable between wild-type and Neto-deficient mice (Figure 3.4D). Moreover, the extent of facilitation observed with NMDAR-mediated EPSCs was similar to that of

AMPAR-mediated events (Figure 3.4D). Given that AMPAR-EPSCs were recorded with KAR transmission intact while NMDAR currents were recorded with KARs blocked, the results indicate that under our experimental conditions, presynaptic KARs did not participate in regulating transmission (see Kwon and Castillo, 2008a). This finding is surprising given the extensive literature describing presynaptic KAR-mediated facilitation of glutamate release at

MF-CA3 pyramidal cell synapses. One possible explanation for this apparent discrepancy is that presynaptic KAR-mediated regulation of MF-CA3 transmission during paired or 4-pulse protocols requires shorter interpulse intervals (Contractor et al., 2001; Marchal and Mulle, 2004) than the ones used in this study. Additionally, the investigation of presynaptic glutamate release

133

alterations based on measurements of NMDAR-mediated EPSC amplitude could be complicated by the slow NMDAR decay kinetics. Thus, while we cannot conclude whether Neto1 or Neto2 plays a role on presynaptic KARs, the normal presynaptic function observed in all animals examined under our stimulation protocols indicate that the reduced KA/AMPA EPSC ratios in

Neto1-null and Neto1/Neto2-double null mice reflect a selective deficit in postsynaptic KAR activation due to the absence of Neto1.

3.3.5. Neto1 is required for the synaptic abundance of hippocampal KARs

The reduction in KAR-mediated EPSCs at hippocampal MF-CA3 synapses of Neto1- null mice suggests that the loss of Neto1 leads to a decrease in the number and/or the function of synaptic KARs. To determine whether the absence of Neto1 was associated with any changes in synaptic expression of KARs, I isolated hippocampal PSDs from wild-type and Neto1-null mice and quantified relative protein levels by densitometry analysis of immunoblots. To control for variability between blots, pairs of wild-type and Neto1-null PSDs were run on the same blot, and the signal intensity of each protein examined in Neto1-null samples was normalized to that of the same protein in wild-type samples. As shown in Figure 3.6A/B, GluK2 and GluK5 KAR subunits were significantly reduced in PSDs of Neto1-null mice (47%±9%, and 57%±7% of wild-type mice, for GluK2 and GluK5, respectively; mean±SD, p<0.01). To determine whether the decrease in KAR subunits was part of an overall protein reduction in Neto1-null PSDs, I assessed the abundance of other synaptic proteins, such as GluN1, GluN2B, and GluA2/3, but found that they were all comparable between wild-type and Neto1-null samples (Figure 3.6A/B).

To determine whether the reduction of KARs subunits in the PSD was the result of changes in

134

the amount of total protein, I compared the levels of GluK2 in wild-type and Neto1-null hippocampal homogenates. Immunoblot analysis showed that the abundance of GluK2 and of all the other proteins examined were comparable between wild-type and Neto1-null mice

(Figure 3.6A/B). Altogether these results indicate that Neto1 does not alter total KAR protein levels but is required for maintaining the normal abundance of KARs specifically in the PSD.

The decrease in the abundance of postsynaptic KARs in Neto1-null mice is consistent with the observed reduction of KAR-mediated synaptic transmission in these mice. We therefore conclude that Neto1 serves as a critical regulatory element of native postsynaptic KAR complexes at hippocampal PSDs.

To determine if the remaining KARs were localized to the PSD by Neto2, I also examined KAR protein levels in hippocampal PSDs of Neto1/Neto2 double-null mice. In animals lacking both Neto proteins, the GluK2 and GluK5 subunits were reduced by approximately the same amount as that resulting from the loss of Neto1 alone (42%±5%, and

50%±15% of wild-type mice, for GluK2 and GluK5, respectively; mean±SD, p<0.01)

(Figure3.6C/D). This result, in conjunction with the observation that the postsynaptic abundance of KARs is normal in Neto2-null mice, shows that Neto2 does not contribute to the synaptic localization of KARs in the hippocampus. The remaining KARs at the PSDs of

Neto1/Neto2 double-null mice also demonstrate the existence of pool of KARs that are localized and/or stabilized at synapses by Neto-independent mechanisms.

135

136

3.3.6. Neto1 binds to the synaptic scaffolding protein PICK1

KARs interact with a number of PDZ domain-containing proteins such as PSD95,

SAP102, SAP97 (Garcia et al., 1998; Mehta et al., 2001), CASK (Coussen et al., 2002), GRIP,

PICK1, and syntenin (Hirbec et al., 2003). These interactions are thought to be involved in regulating channel function, synaptic localization, and the organization of receptors and other proteins into functional complexes. Given that Neto1, which has a PDZ motif at its C-terminus, regulates the abundance of synaptic KARs, we wondered whether it was associated with PDZ domain-containing proteins linked to the synaptic accumulation of KARs. In a previous unbiased yeast two-hybrid screen of adult mouse brain library (Ploder, McInnes lab, unpublished data) PICK1 was identified as a putative intracellular Neto1 interacting protein. To test the binding between Neto1 and PICK1 observed in the two-hybrid screen using an independent biochemical approach, I assessed their interactions in a heterologous mammalian cell system. In HEK293 cells, incubating cell lysates coexpressing Neto1 and PICK1 with an anti-Neto1 antibody resulted in coimmunoprecipitation of PICK1. However, removal of the

Neto1 C-terminus PDZ motif (Neto1∆TRV) severely diminished PICK1 coimmunoprecipitation, thereby suggesting a PDZ domain-dependent interaction between PICK1 and Neto1 (Figure 3.7).

137

138

Given that GluK2 also binds to PICK1 (Hirbec et al., 2003) and was shown to associate with

Neto1 earlier in this Chapter, I next asked how simultaneous coexpression of GluK2, PICK1 and

Neto1 would affect their interactions. As shown in Figure 3.8A, the amount of GluK2 that coimmunoprecipitates with Neto1 was unaffected by the presence of PICK1 (compare lanes 5 and 6). On the other hand, more PICK1 coimmunoprecipitated with Neto1 in the presence of

GluK2 (compare lanes 4 and 6 of Figure 3.8A). Similarly, increased coimmunoprecipitation of

PICK1 with GluK2 was observed in the presence of Neto1 (Figure 3.8B). The increase in the amount of PICK1 associated with GluK2 was abolished, however, when Neto1∆TRV was coexpressed instead of the full length Neto1 protein. Taken together, these results suggest that

Neto1 enhances the interaction between GluK2 and PICK1, or recruits more PICK1 into a complex containing GluK2, in a PDZ-motif dependent fashion. Based on these observations, I propose a model in Neto1 and GluK2 interact extracellularly through CUB domains as shown earlier in this Chapter, while their intracellular cytoplasmic domains bind to PICK1.

139

140

3.4. Discussion

We have investigated the role of Neto1 as a regulatory protein of KARs in vivo. We established that in the brain, Neto1 is a critical auxiliary subunit of the KAR protein complex.

Neto1 is associated with KARs in synaptosomal and PSD fractions, and interacts with GluK2- containing KARs through the extracellular CUB domains. In the hippocampus, loss of Neto1 significantly reduced the abundance of KAR subunits at PSDs. Consistent with this finding, we observed a substantial (~50%) decrease in KAR-mediated EPSCs at MF-CA3 synapses of

Neto1-null mice. Moreover, KAR-mediated EPSCs in mice lacking Neto1 displayed significantly faster decay kinetics compared with wild-type mice. Collectively, these findings indicate that Neto1 plays a crucial role in regulating postsynaptic KARs at MF-CA3 synapses.

One of the unanswered puzzles in the KAR field is the discrepancy between the slow decay kinetics displayed by synaptic KAR EPSCs, and the fast kinetics of currents mediated by recombinant KARs. One of the mechanisms proposed to explain the slow kinetics of synaptic

KARs argues that the receptors might be located extrasynaptically where they are activated by glutamate spill over. This possibility has been excluded by the observation that the prevention of glutamate re-uptake, or the reduction of glutamate diffusion has no effect on the kinetics or the amplitude of KAR EPSCs. A second possibility often discussed states that proteins that bind to the cytoplasmic domain of KAR subunits may be involved in altering receptor properties.

Although a number of KAR-interacting proteins have been identified to date, most of these studies have been performed on receptors expressed in heterologous systems. Consequently, it is not known whether these proteins contribute to the slow kinetics of synaptic KARs currents in the brain. A third possibility, which has gained wider acceptance, is that the GluK5 subunit

141

could confer the slow gating properties of native KARs, of which the GluK2/GluK5 heteromer constitutes the most abundant receptor subtype in the brain (Petralia et al., 1994). This hypothesis has been supported by studies showing that decay time constant of KAR EPSCs were reduced by ~30% in GluK5-/- hippocampal MF-CA3 synapses (Contractor et al., 2003).

Moreover, in heterologous cells, the decay kinetics of glutamate-activated GluK2/GluK5 currents were significantly slower than those of GluK2 homomeric receptors (Barberis et al.,

2008). In the current study, we discovered that loss of the single-pass transmembrane protein

Neto1 causes ~50-60% reduction in the decay time constant of native KAR currents. Given the previously described role of Neto1 and its homologue, Neto2, in slowing the kinetics of recombinant KARs in heterologous cells (Zhang et al., 2009), our finding strongly supports the idea that Neto1 is the molecule that directly contributes to the slow decay of KAR-mediated

EPSCs. One way Neto1 could alter the decay kinetics of KAR EPSCs is by increasing the affinity of KARs for glutamate and/or stabilizing the ligand binding dimer interface. However, since the loss of Neto1 also selectively decreases the abundance of postsynaptic KARs, we cannot exclude the possibility that the faster KAR kinetics in Neto1-null neurons is due at least in part to a loss of synaptic GluK5-containing KARs.

In addition to altering the kinetics of KAR-mediated EPSCs the loss of Neto1 also leads to a ~40% reduction in the amplitude of these currents. This significant change in KAR- mediated synaptic transmission could be attributed, at least partially, to a selective loss of KARs at Neto1-null hippocampal PSDs. Moreover, since the reduction is restricted to the PSDs in the absence of any change in total protein levels, we suggest that the lack of Neto1 impairs either the delivery and/or the stability of postsynaptic KARs. Previous studies have proposed that

PDZ domain-containing proteins, such as PICK1 bind to KARs and regulate their synaptic

142

stability. Our discovery that Neto1 can also bind to PICK1 suggests that these two proteins may act in concert to modulate levels of synaptic KARs.

In a recent study, Straub et al. also described faster decay kinetics and a smaller amplitude of KAR EPSCs in Neto1-null MF-CA3 synapses. Although we also observed a 50% reduction of KARs in the PSD, these authors found no such change in the absence of Neto1, and concluded that impairments in KAR-mediated synaptic transmission are due to changes in channel function rather than distribution (Straub et al., 2011b). The reason(s) for the apparent discrepancy between our data and that of Straub et al. is unclear, but one possibility could be the fact that we assessed changes in PSD KARs between wild-type and homozygous Neto1-null mice, while Straub et al. compared samples derived from heterozygous and homozygous animals. If losses of one or both copies of Neto1 have the same impact on KAR synaptic localization, then no differences would be expected between heterozygous and homozygous brain samples. Future studies that examine the extent to which KAR-mediated synaptic transmission is affected in Neto1 heterozygous mice could help to confirm or exclude this possibility. Alternatively, these incongruent results could be due to genetic differences between the two lines of Neto1-null mice used in these studies. Finally, another potential source of discrepancy could be differences in the methodology used for the extraction of the PSD fraction.

We used the protocol described by Cho et al. (Cho et al., 1992), and routinely analyze the purity of our PSD samples to ensure that they were free from contamination with non-PSD components, such as the presynaptic protein VAMP2. This quality control step is critical, particularly when examining changes in postsynaptic KARs, as these proteins are also present on the presynaptic side.

143

The role of Neto1in regulating KARs may be akin to the function of the TARP family of transmembrane proteins (ɣ-2 (stargazin), ɣ-3,ɣ -4, ɣ-5, ɣ-7, and ɣ-8), which modulate the synaptic localization and the channel properties of AMPA receptors (Tomita, 2010).

Interestingly, despite the fact that the GLR-1 AMPA receptors in the invertebrate C. elegans are regulated by both the CUB domain protein SOL-1 (Zheng et al., 2004; Zheng et al., 2006) and stargazin-like proteins, STG-1 and STG-2 (25% identity to vertebrate stargazin), neither Neto1 nor Neto2 interacts with vertebrate AMPA receptors. Conversely, TARPs are not associated with KARs or NMDARs (Chen et al., 2000; Zhang et al., 2009). These differences between vertebrate and invertebrate AMPA receptor accessory proteins suggest that different mechanisms may have evolved for the regulation of this ion channel, or that as yet unidentified

CUB domain proteins may regulate vertebrate AMPA receptors.

In addition to being a critical component of the KAR protein complex, Neto1 is also an

NMDAR-associated protein (Ng et al., 2009). Neto1-null mice displayed a preferential reduction of synaptic GluN2A subunits and impaired NMDAR-mediated EPSCs at Schaffer collateral-CA1 synapses. Moreover, at these synapses, where long-term potentiation (LTP) is

NMDAR-dependent, loss of Neto1 reduced the magnitude of the potentiation to ~50% of wild- type mice. NMDAR-dependent learning and memory as measured by Morris water maze tests are also impaired in Neto1-null mice. These results indicated that Neto1 is an important subunit of the NMDAR complex required for NMDAR-mediated synaptic plasticity and learning (Ng et al., 2009). In the present study, we found that loss of Neto1 led to a significant reduction of

NMDAR-mediated EPSCs at A/C collateral synapses of the CA3, a result that is consistent with the reduction seen at Schaffer collateral-CA1 synapses. At MF-CA3 synapses, however, we found that KAR-mediated, but not NMDAR-mediated, synaptic currents were altered in Neto1-

144

null neurons. These results suggest that the accessory proteins required for functional regulation of a particular glutamate receptor may differ among synapses, even within a single type of neuron.

The differential role of Neto1 in NMDARs at A/C versus MF synapses is reflective of the known structural and functional differences between these two synapses (Zalutsky and

Nicoll, 1990; Williams and Johnston, 1991; Ishizuka et al., 1995; Salin et al., 1996). For example, LTP induction at A/C-CA3 synapses, as well as at Schaffer collateral-CA1 and perforant path synapses in the DG, is dependent on NMDAR activation that results in a postsynaptic enhancement of AMPAR neurotransmission (Bliss and Collingridge, 1993). MF-

CA3 synapses, on the other hand, express lower levels of NMDARs (Watanabe et al., 1998) and display a presynaptic, NMDAR-independent form of LTP (Nicoll and Schmitz, 2005).

Furthermore, MF-CA3 synapses but not A/C-CA3 synapses selectively express a depolarization-induced form of LTD that is dependent on postsynaptic Ca2+ elevation (Lei et al.,

2003), as well as a type of LTP characterized by a long-lasting increase in NMDAR-mediated transmission (Kwon and Castillo, 2008b; Rebola et al., 2008).

Differences in the functional properties of MF vs. A/C-CA3 synapses are likely a result of the differential trafficking and stabilization of proteins at these two synapses. For instance,

AMPARs display an even distribution among all MF synapses, but are absent in a large number of A/C and CA1 synapses (Nusser et al., 1998) where they can be incorporated into the postsynaptic membrane during the expression of LTP (Shi et al., 2001; Kakegawa et al., 2004).

In addition, postsynaptic KARs also show synapse-specific targeting within a single neuron as

KAR-mediated EPSCs have only been observed at MF but not at AC-CA3 synapses (Castillo et al., 1997; Vignes and Collingridge, 1997). While a number of molecular and functional 145

differences have been described for the MF- CA3 and the A/C-CA3 synapses, similar characteristics have been observed between the A/C-CA3 and Schaffer collateral-CA1 synapses.

Overall, our findings of the differential dependence of NMDAR EPSCs on Neto1 at MF-CA3 vs.

A/C-CA3 synapses are consistent with the general characteristics of these synapses. In future studies it will be important to explore whether the differential Neto1 effect on NMDARs vs.

KARs results from a titration of Neto1 away from NMDARs in synapses where both ion channels are expressed. Altogether, our results demonstrate that Neto1 can be an auxiliary protein for either the NMDA or the kainate class of glutamate receptors, depending on the synapse or the region of the brain. Our findings therefore suggest that a specific auxiliary protein, namely Neto1, may regulate more than one type of ligand-gated ion-channel.

In summary, we have discovered that Neto1 is a key component of KAR protein complexes. At MF-CA3 synapses, KAR synaptic transmission is decreased in Neto1-null mice, likely as a result of the reduction of KAR protein levels at the PSD. In addition, in mice lacking

Neto1, we observed a faster decay of KAR-mediated EPSCs. Thus, our findings indicate that

Neto1 is a critical auxiliary subunit of native, synaptic KARs.

146

Chapter 4: Future directions

147

4.1. Final discussion and future directions

Ionotropic glutamate receptors are critical mediators of excitatory synaptic transmission in the mammalian CNS. They do not, however, work in isolation but are regulated by a number of accessory proteins (Jackson and Nicoll, 2011). For instance, AMPARs are associated with various cytoplasmic proteins, as well as transmembrane molecules (e.g. TARPs and cornichon homologs-2 and -3) which control receptor gating and surface expression (Jackson and Nicoll,

2009). The list of NMDAR regulatory proteins is also extensive, including many PDZ domain- containing scaffolding and trafficking proteins (Husi et al., 2000). Moreover, NMDARs also have a transmembrane auxiliary subunit, Neto1, which is involved in delivery/stability of synaptic GluN2A subunits (Ng et al., 2009). While there are many known AMPA and NMDA receptor regulatory proteins, fewer studies have focused on the identification and characterization of KAR interacting proteins. Moreover, the roles of these proteins on the functional regulation of native receptors have mostly not been tested in vivo. One of the reasons for this lack of knowledge may be our relatively limited understanding of the biological role of

KARs in the CNS; up until the late 90’s, it was difficult to discriminate between responses mediated by the AMPARs and the KARs. However, the development of increasingly well- characterized and selective pharmacological agents (Bleakman and Lodge, 1998; Chittajallu,

1999; Frerking and Nicoll, 2000, Fletcher and Lodge 1996) that distinguish AMPARs and

KARs have enabled research in the past two decades to progressively uncover important roles of

KARs in neuronal function. For example, postsynaptic KARs mediate a small but slow component of the EPSC which is thought to contribute to large charge transfer and temporal summation of synaptic currents. Meanwhile, presynaptic KARs regulate neurotransmitter

148

release at excitatory and inhibitory synapses and contribute to presynaptic forms of synaptic plasticity. Furthermore, unlike NMDARs and AMPARs, which are only known to function as ion channels, KARs can also act as metabotropic receptors to regulate neuronal excitability

(Contractor et al., 2011).

In this thesis, I have presented data demonstrating the regulation of synaptic KARs by a family of CUB domain-containing, single-pass transmembrane proteins called Neto1 and Neto2.

Results from Chapters 2 and 3 indicate that both Neto1 and Neto2 are required for the postsynaptic localization of native KARs. The two proteins, however, appear to regulate receptors expressed in different neuronal populations. This region-specific regulation may arise in part from differences in the pattern of Neto1 and Neto2 expression. For instance, in the hippocampus, Neto1 mRNA is particularly abundant in pyramidal cells of the CA3 region, whereas Neto2 shows a relatively low but uniform distribution along the CA1-3 pyramidal layer

(Michishita et al., 2004; Ng et al., 2009). In the cerebellar cortex, Neto2 is strongly expressed in the GCL, while Neto1 is conspicuously absent from this region (Michishita et al., 2004). Other areas of high Neto1 and Neto2 expression in the brain include the amygdala and the cerebral cortex (Ng, 2006; Allen Brain atlas), both of which also express KARs (Bettler et al., 1990;

Bahn et al., 1994). Whether Neto1 or Neto2 contribute to KAR abundance and function in these and other brain regions remain to be determined. The complementary, and in some cases overlapping, expression patterns of Neto1 and Neto2 throughout the brain are reminiscent of the differential distribution of members of the transmembrane AMPA receptor regulatory protein

(TARP) family (γ-2, 3, 4, 5, 7,8). TARPs control AMPAR trafficking and gating, and their distinct regional distribution is thought to contribute to the synapse-specific regulation of their associated AMPARs (Jackson and Nicoll, 2011). For example, a spontaneous mutant of the

149

cerebellar-enriched stargazin/γ2 subunit results in a selective loss of functional AMPARs in cerebellar granule neurons but has no effect on receptors present in forebrain neurons, such as the CA1 pyramidal neurons (Hashimoto et al., 1999; Chen et al., 2000). In contrast, loss of the

γ-8 subunit, which is most highly expressed in the hippocampus, but is absent from the cerebellum, severely reduced the levels of synaptic and extrasynaptic hippocampal AMPARs

(Rouach et al., 2005).

In addition to differences in neuronal distribution, Neto1/2 might also regulate the synaptic localization of KARs depending on the subunit composition of the receptor. In a recent study by Copits and coworkers (Copits et al., 2011), coexpression of Neto2 and GluK1 in hippocampal culture neurons greatly enhanced the accumulation of GluK1-containing KARs to dendritic spines and the colocalization of these receptors with the synaptic marker PSD95.

Coexpression of Neto1 and GluK1, however, did not alter KAR subcellular distribution. On the other hand, in the hippocampus, Neto1, but not Neto2, has been shown to be crucial for the abundance of endogenous KARs at the PSD. The apparent discrepancy between the roles of

Neto1 and Neto2 in the synaptic expression of KARs in culture neurons vs. hippocampal tissue may be explained by the different subunit composition of the KARs in the two studies. In the culture neuron study, pyramidal neurons were cotransfected with GluK1 and Neto2 (or Neto1) cDNA, and analysis of the synaptic accumulation of KARs focused on GluK1-containing receptors. In the hippocampus, however, GluK1 expression is restricted to inhibitory interneurons, which do not express Neto2. In hippocampal pyramidal neurons, and in particular at MF-CA3 synapses, which show the highest expression of Neto1 in the hippocampus, postsynaptic KARs are predominantly heteromers composed of GluK2/GluK5 subunits (Mulle et al., 1998; Contractor et al., 2003; Fernandes et al., 2009). Thus, the seemingly contradictory

150

results from the hippocampus, and the culture neuron studies (Copits et al., 2011) suggest the possibility that the regulation of KARs by Neto1 or Neto2 could at least partially, be attributed to the subunit composition of the specific receptor being regulated. In the cerebellum, the subunit composition of KARs may also differ from that of the hippocampus. While both GluK2 and GluK5 subunits are strongly expressed (primarily in CGNs) (Wisden and Seeburg, 1993),

GluK2 protein is ten times more abundant than GluK5 (Ripellino et al., 1998). Given this difference in relative abundance, and the fact that CGNs do not appear to express other subunits, it is likely that the majority of cerebellar GluK2-containing KARs are GluK2 homomers. Neto2 may, therefore, be responsible for the synaptic localization of GluK2 homomeric KARs, in addition to that of GluK1-containing KARs, whereas Neto1 predominantly regulates GluK5- containing KARs.

In addition to their role in regulating synaptic levels of KARs, the Neto proteins can also modify KAR channel properties. In heterologous cells, coexpression of Neto1 or Neto2 can significantly alter the rates of KAR entry and recovery from desensitization (Zhang et al., 2009;

Copits et al., 2011; Straub et al., 2011a). Similarly, at mossy fiber synapses, Neto1 is required not only for the synaptic expression of KARs (Tang et al., 2011), but also for generating the slow decay kinetics of KAR-mediated EPSCs (Straub et al., 2011b; Tang et al., 2011). The dual role played by the Neto proteins on KARs raise the possibility that these two processes -receptor gating and localization- are not mutually exclusive. Meanwhile, in vitro studies indicate that the modulatory action of Netos on KAR kinetics may also vary according to the receptor subtype, akin to the subunit-dependent regulation of KAR synaptic targeting (Copits et al., 2011). Thus, the differential regulation of KAR subtypes combined with distinct Neto1/2 expression patterns, and the potential association of Neto1/2 with diverse PDZ domain regulatory and scaffolding

151

molecules, such as PICK1 and GRIP, constitute a possible mechanism for the synapse-specific regulation of KARs.

Recently, a Neto-like protein has also been identified in Drosophila (Kim et al., 2012).

Disruption of Drosophila Neto in striated muscles of flies was found to severely compromise synaptic trafficking, and clustering of ionotropic glutamate receptors (iGluRs) at the PSD. Thus, similar to the loss of mammalian Neto1 and Neto2, Neto-deficiency in Drosophila significantly impaired the abundance of postsynaptic iGluRs. The similar function of mammalian Neto1,

Neto2, and Drosophila Neto indicate these proteins are all evolutionarily conserved regulatory elements of glutamate receptors.

In summary, our studies have uncovered a critical role of the Neto proteins as regulatory elements of synaptic KARs. These findings, however, also raise a number of interesting questions that will be described in the following sections. Future research on these questions will expand our understanding of the roles of Neto1 and Neto2 in the CNS, and provide further insight into the way KAR function and subcellular localization are regulated by these auxiliary proteins.

4.1.1. Additional studies on the role of Netos on KAR synaptic physiology

In the hippocampus

In the hippocampus, postsynaptic KARs have been shown to contribute to excitatory synaptic transmission at MF-CA3 and Schaffer collateral (SC)-CA1 interneuron synapses. At these and other synapses where KAR EPSCs have been detected, KAR currents decay with time

152

constants of 30-150 milliseconds (Castillo et al., 1997; Kidd and Isaac, 1999; Cossart et al.,

2002; Wu et al., 2007a). This relatively slow decay of synaptic KARs is not consistent with the fast deactivation/desensitization (<10 milliseconds) described for recombinant KARs (Traynelis et al., 2010). As discussed in Chapter 3, we have determined that Neto1 confers the slow kinetics of native synaptic KARs at MF-CA3 connections, as KAR EPSCs decay at a faster rate in the absence of Neto1. Moreover, we found that at MF-CA3 synapses, loss of Neto1 leads to a

~ 40% reduction in KAR-mediated currents, which can be attributed to lower KAR protein levels in Neto1-null hippocampal PSDs. Future studies should, therefore, examine whether

Neto1 is a general KAR auxiliary subunit that regulates receptor kinetics and synaptic abundance at all synapses where it is expressed. To begin, one could ask whether Neto1, which is expressed in hippocampal CA1 interneurons (Ng et al., 2009), regulates KAR-mediated synaptic function in these cells. Pharmacological studies indicate that KAR subunit composition in CA1 interneurons differs from that of MF-CA3 synapses. Postsynaptic KARs of MF-CA3 synapses are composed of GluK2/GluK5 heteromers (Petralia et al., 1994), whereas in CA1 interneurons, KAR-mediated EPSCs are largely generated by GluK1-containing KARs

(Wondolowski and Frerking, 2009), although it is not known whether these are homomeric

GluK1 receptors or are heteromeric receptors involving other subunits. Given that Neto1 has been shown to accelerate the desensitization rate of homomeric GluK1 receptors in a heterologous expression system (Copits et al., 2011), investigating the changes, if any, in the decay kinetics of synaptic KARs in Neto1-null CA1 interneurons may provide insight into the subunit composition of these receptors.

In addition to contributing to excitatory transmission at the postsynaptic side, KARs are also present on the presynaptic membrane, where they are involved in regulating

153

neurotransmitter release (Pinheiro and Mulle, 2008). Studies at MF-CA3 synapses have shown that endogeneous glutamate released during MF stimulation activates presynaptic KARs, causing a facilitation of evoked neurotransmitter release. This process contributes to the frequency facilitation of synaptic transmission (frequency-dependent enhancement of MF EPSC amplitude) observed during increased rates of stimulation (Contractor et al., 2001; Lauri et al.,

2001a). Frequency facilitation is particularly prominent at MF-CA3 synapses, where it requires the presence of GluK2-containing presynaptic KARs (Contractor et al., 2001). As discussed in

Chapter 3, however, we did not observe an involvement of KARs in frequency facilitation at mossy fiber synapses. One possible explanation for this discrepancy could be differences in the stimulation protocols used in ours vs. other studies. Therefore, additional studies should use established stimulation protocols to determine 1) whether synaptic activation of presynaptic

KARs can indeed facilitate glutamate release and contribute to the pronounced frequency facilitation characteristic of MF synapses; and if so, 2) whether the function of presynaptic

KARs, as determined by short-term frequency facilitation, is altered in Neto1-, Neto2-, and

Neto1/Neto2-double null animals. At present, preliminary results from immunogold electron microscopy (EM) studies examining KARs protein levels on presynaptic MF axon terminals showed that receptor levels are reduced in Neto1/Neto2-double null mice compared to wild-type

(Wyeth et al., 2012). This observation supports a role for the Neto proteins in regulating presynaptic KARs. Consequently, further studies on presynaptic KAR-mediated events in Neto- null animals are warranted.

At MF-CA3 synapses, long-term potentiation (LTP), a form of long-term synaptic plasticity, is maintained by a long-lasting enhancement in neurotransmitter release (Zalutsky and

Nicoll, 1990). Currently, it is widely accepted that MF LTP is independent of NMDAR

154

activation (Harris and Cotman, 1986). On the other hand, pharmacogical and genetic studies support a critical role for KARs in MF LTP, though it is not clear whether LTP induction is due to activation of presynaptic receptors, postsynaptic receptors, or both (Contractor et al., 2001).

One hypothesis is that presynaptic KARs facilitate neurotransmitter release during repetitive stimulation, thereby enhancing synaptic transmission (Nicoll and Schmitz, 2005). Given that

Neto1 is associated with KARs, and regulates receptor abundance and kinetics at MF-CA3 synapses, future studies should explore whether MF LTP is impaired in the absence of Neto1.

While loss of Neto1 significantly alters KAR-mediated EPSCs, KAR function in Neto2- null MF-CA3 synapses is indistinguishable from wild-type. This result is surprising given that 1)

Neto2 is thought to regulate the kinetics of recombinant GluK2/GluK5 heteromeric receptors

(the predominant KAR subtype on the postsynaptic membrane of MF-CA3 synapses) in heterologous cells; 2) Neto2 protein is localized to the stratum lucidum; and 3) anti-Neto2 antibodies can coimmunoprecipitate a significant fraction of GluK2-containing KARs from hippocampal synaptosome-enriched fractions. So why does the loss of Neto2 have no effect on

KAR decay kinetics, and synaptic abundance in the same way that Neto1 does? An initial hypothesis is that Neto1 compensates for the loss of Neto2. However, this seems unlikely because changes in KAR EPSCs in the double-null mice were not different from Neto1 single- null animals. Another possibility is that at MF synapses, Neto2 is localized outside the PSD.

To address this question, one could perform EM studies to look at the distribution of Neto2 at these synapses. While Neto2 had no effect on synaptic KARs at MF-CA3 synapses, and is not likely to modulate KAR function at SC-CA1 interneuron synapses (Neto2 is not expressed in interneurons), one could use the Neto2-null mice to also test whether Neto2 is associated at all with any functional KARs in the hippocampus (e.g. KARs that do not necessarily contribute to

155

EPSCs and are not localized in the postsynaptic membrane). For example, one could bath apply kainate on wild-type and Neto2-null hippocampal slices and record kainate-mediated currents from CA1 and CA3 pyramidal neurons in the presence of AMPAR antagonists.

Another aspect of KAR neuronal function where the role of Neto1 and Neto2 must be investigated is the regulation of neuronal excitability through metabotropic (G-protein-mediated) signaling pathways. For example, in CA1 pyramidal cells, KARs are located extrasynaptically, and upon activation they inhibit the slow after-hyperpolarizing potential (AHP) through a metabotropic action, resulting in a long-lasting enhancement of neuronal excitability (Melyan et al., 2004). Future work should investigate whether the Neto proteins regulate the function of metabotropic KARs by examining AHP in CA1 pyramidal cells of wild-type and Neto-null mice.

In the cerebellum

Neto2, GluK2, and GluK5 mRNAs are highly expressed in granule cells of the cerebellum. Moreover, previous studies on cerebellar slices (Smith et al., 1999) or cultured granule neurons (Savidge et al., 1997; Pemberton et al., 1998) have confirmed the presence of functional KARs in these cells. Given that the large majority of cerebellar GluK2 and GluK5 expression is found only in granule cells (Bahn et al., 1994), and that total cerebellar GluK2 protein is ten times more abundant than GluK5 (Ripellino et al., 1998), it is likely that most of the KARs in granule cells are GluK2 homomers. Future experiments could examine whether

Neto2 modulates channel properties of endogeneous KARs in granule cells by measuring kainate-evoked currents in wild-type and Neto2-null cerebellar slices; previous studies have only investigated the effect of Neto2 on cultured granule neurons transfected with recombinant

GluK2 receptors containing a mutation that slows desensitization.

156

Although Neto2 had no effect on the abundance of synaptic KARs in the hippocampus, loss of Neto2 significantly reduced the levels of GluK2 subunits in PSDs isolated from whole cerebellum. Given that GluK2 is expressed primarily in granule neurons and that immunofluorescence studies show intense GluK2 staining in nuclear-free, irregularly-shaped structures resembling the glomeruli (where granule cell dendrites receive input from mossy fiber axon terminals), it is likely that the changes in GluK2 levels in Neto2-null cerebellar PSDs results from a decrease of postsynaptic KARs in granule cells. Future studies could use post- embedding immunogold EM to examine in more detail the distribution of KARs at mossy fiber- granule cell synapses.

To determine how the postsynaptic reduction of GluK2-containing receptors affects

KAR-mediated synaptic transmission in Neto2-null mice, one could, in theory, examine granule cell KAR EPSCs. However, while there are abundant GluK2-containing receptors in cerebellar

PSD fractions, strangely no contribution of KARs to excitatory postsynaptic responses have been reported so far at mossy fiber-granule cell synapses. The only two cerebellar synapses where KAR EPSCs have been observed are at the parallel fiber (PF)-Golgi cell synapse (Bureau et al., 2000) and the climbing fiber (CF)-Purkinje cell synapse (Huang et al., 2004). KAR

EPSCs in Golgi cells have slow rise and decay kinetics (Bureau et al., 2000) similar to those observed in CA3 pyramidal cells (Castillo et al., 1997; Vignes and Collingridge, 1997). In contrast, KAR EPSCs in Purkinje cells display fast decay kinetics similar to those of AMPARs

(Huang et al., 2004), suggesting that receptors at these synapses may function independently of

Netos.

Golgi cells are large inhibitory interneurons that are sparsely scattered among the densely packed granule cells within the GCL. They extend long dendrites into the MCL where 157

they receive excitatory inputs from granule cell axons (parallel fibers). Reverse-transcription

(RT)-PCR analysis indicates that both the GluK1 and GluK2 subunits of KARs are expressed in

Golgi cells, although it is not known whether both subunits are incorporated into the PSD

(Bureau et al., 2000). The presence of Neto1, but not Neto2, mRNA in Golgi cells has been reported in a comprehensive study of translated mRNAs from defined cell populations of the

CNS using the translating ribosome affinity purification (TRAP) method (Doyle et al., 2008).

To determine whether Neto1 protein is present in Golgi cells, cerebellar slices could be immunostained with Neto1 and somatostatin, a marker of Golgi cells (Vincent et al., 1985).

Golgi cells can also be distinguished from granule cells by size: Golgi cell somata are 8-25 μm, whereas granule cell somata are only 4-7 μm. If Neto1 is present in Golgi cells, then one could use the Neto-null mice to examine whether Neto1 is essential for the regulation of synaptic

KAR function. For instance, post-embedding immunogold EM analysis could be used to characterize the abundance and distribution of KARs at PF-Golgi cell synapses in wild-type and

Neto-null mice. Additionally, electrophysiology can be used to identify any changes in the amplitude and/or kinetics of KAR EPSCs in synapses with or without Neto proteins.

4.1.2. Characterization of KAR synaptic localization defects

To characterize the role of Neto proteins in the neuronal distribution and synaptic localization of KARs, immunofluorescence studies can be carried out on hippocampal cultured neurons. KARs can be visualized by labeling with fluorescence-tagged antibodies to determine whether there are any changes in their synaptic accumulation between wild-type and Neto1-null neurons. The function of the synaptic receptors in these neurons can also be monitored by

158

recording the amplitude and frequency of KAR synaptic responses. Surface levels of KARs in

Neto1-null neurons can be investigated by treating the cells with membrane-impermeable biotin reagents and then capturing biotin-labeled receptors with avidin-linked agarose. If there are any changes in the distribution and subcellular localization of KARs in the absence of Neto1, then one can also determine whether these abnormalities can be rescued with full length Neto1.

Moreover, to determine if the C-terminal PDZ binding motif of Neto1 is important for the synaptic localization of KARs, full length Neto1 (as a control), and Neto1ΔTRV can be transfected into wild-type and Neto1-null neurons. If full length Neto1 can restore the synaptic abundance of KARs in Neto1-null neurons but Neto1ΔTRV cannot, the implication would be that the C-terminal amino acids of Neto1 is important for targeting and/or maintaining the receptors at the synapse.

Previous studies using peptides that disrupt PDZ domain-PDZ ligand interactions in hippocampal slices suggest that the synaptic scaffolding protein PICK1 is important for stabilizing KARs at the PSD (Hirbec et al., 2003). Since Neto1 was shown to bind to PICK1 through its C-terminal PDZ motif, future studies could also investigate whether overexpression of PICK1 increases synaptic accumulation of KARs in wild-type and Neto1-null neurons. If

PICK1 overexpression increases the number of synaptic KAR puncta in wild-type but not

Neto1-null cells, this would suggest that Neto1 is required for the PICK1-mediated synaptic localization of KARs.

Similar experiments as the ones described above can be done on cerebellar granule neurons to examine KARs distribution in wild-type and Neto2-null neuronal cell types as well.

159

In addition to investigating the synaptic localization of KARs, one could also use the hippocampal neurons (or cerebellar granule neurons) to characterize the cellular distribution of

Neto1 and Neto2. Here, one could compare the fraction of each Neto that colocalizes with excitatory synaptic markers and with KARs, their surface vs. intracellular localization, and their distribution in dendrites or axons.

4.1.3. Systematic analysis of the modulation of KAR biophysical properties by Neto1/2

The first study that showed the regulation of KAR channel properties by Neto1 and

Neto2 was performed on recombinant homomeric GluK2 receptors (Zhang et al., 2009). A subsequent study compared the regulation of GluK1 receptors by Neto1 and Neto2 (Copits et al.,

2011), and another examined the modulation of GluK1, GluK1/GluK5, and GluK2/GluK5 receptors by Neto2 (Straub et al., 2011a). While all these investigations agree that Neto1 and

Neto2 have a significant impact on KAR function, they also show that 1) KAR modulation by

Neto1/2 varies with the receptor’s subunit composition, and that 2) the effects exerted by Neto1 and Neto2 on a given KAR subtype can be qualitatively and quantitatively different. Thus, in the nervous system, the biophysical properties and function of native KARs at a particular synapse are determined not only by their subunit composition, but also by whether they are associated with Neto1 or Neto2. Given the critical role of Neto1/2 as KAR auxiliary subunits, future work should investigate the effect of coexpressing Neto1 or Neto2 with all possible KAR subtypes, while taking into account some of the issues not addressed in previous work. For instance,

160

1) In the study by Zhang et al., the authors concluded that Neto1 does not regulate GluK2-

KAR activity as much as Neto2, based on the observation that KAR currents were

significantly larger in receptors coexpressed with Neto2. While this is certainly a

possibility, the authors did not ask whether the two Neto proteins were expressed at

similar levels. Future studies comparing the effect of Neto1 and Neto2 on a given KAR

subtype should ensure that both proteins have a similar Neto to KAR protein ratio.

2) The characterization of Neto1/2 regulation of heteromeric channels in cells transfected

with two or more different KAR subunits can be complicated by the possible expression

of a mixture of homomeric and heteromeric channels. Given that each KAR subtype has

different properties, it is necessary to ensure that most of the functional KARs expressed

in the cells have an identical subunit composition. To determine whether one has a

homogeneous KAR population, coimmunoprecipitation experiments can be performed

on cell surface receptors. One way to coimmunoprecipitate cell surface receptors is to

biotinylate all cell surface proteins, followed by immunoprecipitation of the protein of

interest with specific antibodies, and subsequent pull-down of only the biotinylated

protein of interest with streptavidin coated beads. If the receptors are all assembled as

heteromers, then it will be possible to coimmunoprecipitate the vast majority of either

subunit protein with an antibody directed against the other subunit (ie. if cells

cotransfected with GluK2 and GluK5 subunits express only GluK2/GluK5 heteromers,

then an anti-GluK2 antibody that can immunoprecipitate all of the GluK2 will also

coimmunoprecipitate all of GluK5).

3) Given that the research on the regulation of different KARs by Neto1/2 has been

conducted in different labs using different experimental protocols, comparisons across

161

studies have been difficult. Future studies should therefore examine the effect of Netos

on the biophysical properties of all predominant KAR subtypes of the CNS under the

same experimental conditions.

Another aspect of the modulation of KAR channel function that has not yet been explored is the stoichiometry of KAR-Neto protein complexes. One could therefore ask, for example, 1) how many Neto1 or Neto2 molecules are incorporated into a functional tetrameric ion channel; 2) whether the Neto-KAR protein complex have a fixed or variable stoichiometry, and 3) whether both Neto1 and Neto2 can be part of the same KAR ion channel, if they are present in the same cell. One approach to address these questions would be to determine the protein composition of KAR protein complexes by blue native polyacrylamide gel electrophoresis (BN-PAGE). BN-PAGE preserves native protein complexes, thus allowing their composition and stoichiometry to be analyzed. If the stoichiometry of the Netos on KARs is found to be variable, or if both Neto1 and Neto2 can be incorporated into a single KAR complex, then one could examine how the number and/or composition of Neto subunits per channel impacts the electrophysiological properties of the receptors. On the other hand, if both Neto1 and Neto2 are found always to be present in different KAR complexes, then the preference of a particular KAR subtype for Neto1 or for Neto2 could be explored.

4.1.4. Behavioural studies on Neto1- and Neto2-null mice

KARs are believed to be important for learning and memory, in part, because of their contribution to synaptic plasticity in the hippocampus and amygdala (Frerking and Nicoll, 2000;

Kullmann, 2001; Huettner, 2003; Lerma, 2003). The amygdala has a primary role in the

162

formation and storage of memories associated with emotional events, such as fear. A common test for alterations in fear memory is contextual or cue fear conditioning. During conditioning, a foot shock (unconditioned stimulus) is paired with a conditioned stimulus, such as a particular context and/or an auditory cue. Contextual or cue fear memory is determined at various time points after conditioning by measuring the freezing response of the animal (due to fear), when presented with the context or the cue in the absence of any foot shock. In GluK2-/- mice, studies have shown that, while freezing responses were normal immediately after conditioning

(suggesting normal associative learning), retention of both the contextual and auditory fear memory at later time points is significantly impaired (Ko et al., 2005). Consistent with this behavioural deficit, synaptic potentiation at thalamic input synapses to the lateral amygdala is significantly reduced in the mutant mice (Ko et al., 2005). Given the abundant expression of

Neto1 and Neto2 in the amygdala and their previously described roles in modulating KAR function, future studies should examine Neto-null mice in tasks related to fear memory.

The nature of the involvement of KARs in hippocampal-dependent spatial learning and memory is not clear. In a Morris water maze protocol that tests for spatial reference memory,

GluK2-/- mice displayed comparable learning ability to wild-type mice (Mulle et al., 1998).

However, additional testing using protocols that place a greater demand on memory processes may identify more subtle or selective deficits in learning and memory. For instance, previous studies showed that in reference memory tests, Neto1-null mice perform as well as their wild- type counterparts. In the delayed-matching-to-place protocol, however, which tests working memory, the loss of Neto1 significantly impaired performance (Ng et al., 2009).

The selective learning and memory deficits in Neto1-null mice have been attributed to a decrease in hippocampal NMDAR abundance and function (Ng et al., 2009). However, the 163

discovery that Neto1 is also an auxiliary subunit of KARs raises the possibility that the impaired performance of Neto1-null mice in learning and memory tasks may also be related to changes in synaptic transmission that result from reduced KAR abundance and function. Therefore, future studies on the involvement of KARs in hippocampal-dependent spatial learning and memory and on the role of Neto1 in KAR-mediated hippocampal synaptic plasticity, will both be required to clarify the molecular mechanisms underlying the learning and memory deficits of

Neto1-null mice.

In addition to studies of spatial learning in the Morris water maze, behavioural analysis using the elevated-plus maze test has revealed that Neto2-, but not Neto1-null mice, display increased anxiety-like behaviors (Lipina, T, unpublished observations). A similar phenotype has also been reported in GluK1-/- mice, and studies suggest that the development of anxious behaviour in GluK1-/- mice is the result of reduced inhibitory transmission in the amygdala (Wu et al., 2007b). To understand whether the anxious behaviour exhibited by Neto2-null mice is caused by a reduction in GluK1 function in the amygdala, KAR-mediated GABAergic transmission should be examined in these mutant animals. Moreover, to determine if these abnormal behaviours are caused by a developmental effect or by impaired functions of other brain regions, conditional Neto2 knockout mice could be used.

4.1.5. Additional studies on the regulation of synaptic NMDARs by Neto1

Neto1 has been previously shown to regulate NMDAR function in the hippocampus. At

SC-CA1 synapses, loss of Neto1 leads to a ~50% reduction in NMDAR-mediated currents (Ng et al., 2009). Similarly, at A/C-CA3 synapses, NMDAR currents in Neto1-null mice were only

164

~60% of wild-type. To our surprise, however, at MF-CA3 synapses, NMDAR currents were not significantly different between the two genotypes. Given that KAR-mediated EPSCs have been recorded at mossy fiber synapses, but are absent from SC-CA1 and A/C-CA3 synapses, this differential effect of Neto1 on NMDARs results from KARs titrating Neto1 away from the

NMDARs in synapses, where both ion channels are expressed. To address this question, one approach would be to compare the binding affinities of Neto1 for KARs vs. NMDARs using isothermal titration calorimetry (ITC), which is one of the most quantitative methods for measuring and characterizing biomolecular interactions (Pierce et al., 1999). ITC is a thermodynamic technique that directly measures the heat released or absorbed during the association of two molecules. Measurement of this heat allows the determination of the binding constant (KB), stoichiometry (n), enthalpy (∆H), and entropy (∆S) of binding in one single experiment (Leavitt and Freire, 2001). One could also examine KAR and NMDAR function at other synapses, where both proteins and Neto1 are present, to determine whether the preferential regulation of KARs by Neto1 is a general characteristic of all synapses expressing KARs,

NMDARs and Neto1.

At SC-CA1 synapses, the reduction in GluN2A subunits at the PSD is thought to account, at least in part, for the decrease in NMDAR synaptic currents (Ng et al., 2009). Therefore, another explanation of the normal NMDAR-mediated EPSCs at MF-CA3 synapses could be compensation by other NMDAR subunits. At mossy fiber synapses, most NMDARs are composed of GluN1, the obligatory subunit, and GluN2A subunits (Fritschy et al., 1998;

Watanabe et al., 1998). However, preliminary results showed an increased sensitivity of these postsynaptic NMDARs to ifenprodil in Neto1-null neurons (Pelkey and McBain, unpublished data). Given that ifenprodil is a selective NMDAR inhibitor that is specific to GluN2B-

165

containing receptors, our results imply that there is an increase in the proportion of GluN2B subunits at these synapses. Immunoblot analysis of PSD proteins isolated solely from mossy fiber synapses could be used to identify any changes in the composition of NMDAR ion channels and associated proteins in Neto1-null mice. In addition, EM studies could be carried to determine the distribution of various NMDAR subunits at MF-CA3 synapses of wild-type and

Neto1-null mice. If immunoblot and EM analysis show a reduction in GluN2A subunits accompanied by an increase in GluN2B subunits (as suggested by the ifenprodil studies), then this would suggest that at mossy fiber synapses, Neto1 may still be necessary for the synaptic accumulation of GluN2A-containing receptors.

In addition to the role of Neto1 on NMDAR function, biochemical evidence indicates that Neto2 also interacts with NMDARs. Future studies should explore whether Neto2 also affects NMDAR abundance and synaptic transmission. Moreover, Neto1 or Neto2 and various

NMDAR subtypes, could be expressed in heterologous systems (Xenopus oocytes, HEK293 cells) to determine whether Neto1/2 can modulate NMDAR channel properties as they do for

KARs.

166

Bibliography

Akamatsu Y, Jasin M (2010) Role for the mammalian Swi5‐Sfr1 complex in DNA strand break repair through homologous recombination. PLoS Genet 6:e1001160. Alt A, Weiss B, Ogden AM, Knauss JL, Oler J, Ho K, Large TH, Bleakman D (2004) Pharmacological characterization of glutamatergic agonists and antagonists at recombinant human homomeric and heteromeric kainate receptors in vitro. Neuropharmacology 46:793‐806. Amaral DG, Witter MP (1989) The three‐dimensional organization of the hippocampal formation: a review of anatomical data. Neuroscience 31:571‐591. Amaral DG, Witter MP (1995) Hippocampal formation. In: The rat nervous system (Paxinos G, ed). San Diego: Academic Press. Andersen BB, Korbo L, Pakkenberg B (1992) A quantitative study of the human cerebellum with unbiased stereological techniques. J Comp Neurol 326:549‐560. Andersen P, Morris R, Amaral D, Bliss T, O'Keefe J, eds (2007) The hippocampus book. New York: Oxford University Press. Bahn S, Volk B, Wisden W (1994) Kainate receptor gene expression in the developing rat brain. J Neurosci 14:5525‐5547. Barberis A, Sachidhanandam S, Mulle C (2008) GluR6/KA2 kainate receptors mediate slow‐deactivating currents. J Neurosci 28:6402‐6406. Baudry M, Thompson RF (1993) Synaptic Plasticity: Molecular, Cellular, and Functional Aspects. Cambridge: MIT Press. Bear MF, Connors BW, Paradiso MA (2001) Exploring the brain, 2nd Edition. Baltimore: Lippincott Williams & Wilkins. Bettler B, Mulle C (1995) Review: neurotransmitter receptors. II. AMPA and kainate receptors. Neuropharmacology 34:123‐139. Bettler B, Boulter J, Hermans‐Borgmeyer I, O'Shea‐Greenfield A, Deneris ES, Moll C, Borgmeyer U, Hollmann M, Heinemann S (1990) Cloning of a novel glutamate receptor subunit, GluR5: expression in the nervous system during development. Neuron 5:583‐595. Bischoff S, Barhanin J, Bettler B, Mulle C, Heinemann S (1997) Spatial distribution of kainate receptor subunit mRNA in the mouse basal ganglia and ventral mesencephalon. J Comp Neurol 379:541‐ 562. Bliss TV, Collingridge GL (1993) A synaptic model of memory: long‐term potentiation in the hippocampus. Nature 361:31‐39. Bork P, Beckmann G (1993) The CUB domain. A widespread module in developmentally regulated proteins. J Mol Biol 231:539‐545. Bortolotto ZA, Clarke VR, Delany CM, Parry MC, Smolders I, Vignes M, Ho KH, Miu P, Brinton BT, Fantaske R, Ogden A, Gates M, Ornstein PL, Lodge D, Bleakman D, Collingridge GL (1999) Kainate receptors are involved in synaptic plasticity. Nature 402:297‐301. Bowie D, Lange GD (2002) Functional stoichiometry of glutamate receptor desensitization. J Neurosci 22:3392‐3403. Bowie D, Garcia EP, Marshall J, Traynelis SF, Lange GD (2003) Allosteric regulation and spatial distribution of kainate receptors bound to ancillary proteins. J Physiol 547:373‐385. Braithwaite SP, Xia H, Malenka RC (2002) Differential roles for NSF and GRIP/ABP in AMPA receptor cycling. Proc Natl Acad Sci U S A 99:7096‐7101. Bureau I, Bischoff S, Heinemann SF, Mulle C (1999) Kainate receptor‐mediated responses in the CA1 field of wild‐type and GluR6‐deficient mice. J Neurosci 19:653‐663.

167

Bureau I, Dieudonne S, Coussen F, Mulle C (2000) Kainate receptor‐mediated synaptic currents in cerebellar Golgi cells are not shaped by diffusion of glutamate. Proc Natl Acad Sci U S A 97:6838‐6843. Burnashev N, Zhou Z, Neher E, Sakmann B (1995) Fractional calcium currents through recombinant GluR channels of the NMDA, AMPA and kainate receptor subtypes. J Physiol 485 ( Pt 2):403‐418. Cameron HA, McKay RD (2001) Adult neurogenesis produces a large pool of new granule cells in the dentate gyrus. J Comp Neurol 435:406‐417. Cao TT, Deacon HW, Reczek D, Bretscher A, von Zastrow M (1999) A kinase‐regulated PDZ‐domain interaction controls endocytic sorting of the beta2‐adrenergic receptor. Nature 401:286‐290. Castillo PE, Malenka RC, Nicoll RA (1997) Kainate receptors mediate a slow postsynaptic current in hippocampal CA3 neurons. Nature 388:182‐186. Chen L, Chetkovich DM, Petralia RS, Sweeney NT, Kawasaki Y, Wenthold RJ, Bredt DS, Nicoll RA (2000) Stargazin regulates synaptic targeting of AMPA receptors by two distinct mechanisms. Nature 408:936‐943. Cho KO, Hunt CA, Kennedy MB (1992) The rat brain postsynaptic density fraction contains a homolog of the Drosophila discs‐large tumor suppressor protein. Neuron 9:929‐942. Christensen JK, Paternain AV, Selak S, Ahring PK, Lerma J (2004) A mosaic of functional kainate receptors in hippocampal interneurons. J Neurosci 24:8986‐8993. Clapcote SJ, Duffy S, Xie G, Kirshenbaum G, Bechard AR, Rodacker Schack V, Petersen J, Sinai L, Saab BJ, Lerch JP, Minassian BA, Ackerley CA, Sled JG, Cortez MA, Henderson JT, Vilsen B, Roder JC (2009) Mutation I810N in the alpha3 isoform of Na+,K+‐ATPase causes impairments in the sodium pump and hyperexcitability in the CNS. Proc Natl Acad Sci U S A 106:14085‐14090. Clarke VR, Collingridge GL (2004) Characterisation of the effects of ATPA, a GLU(K5) kainate receptor agonist, on GABAergic synaptic transmission in the CA1 region of rat hippocampal slices. Neuropharmacology 47:363‐372. Clarke VR, Ballyk BA, Hoo KH, Mandelzys A, Pellizzari A, Bath CP, Thomas J, Sharpe EF, Davies CH, Ornstein PL, Schoepp DD, Kamboj RK, Collingridge GL, Lodge D, Bleakman D (1997) A hippocampal GluR5 kainate receptor regulating inhibitory synaptic transmission. Nature 389:599‐603. Conn PJ, Pin JP (1997) Pharmacology and functions of metabotropic glutamate receptors. Annu Rev Pharmacol Toxicol 37:205‐237. Connors BW, Long MA (2004) Electrical synapses in the mammalian brain. Annu Rev Neurosci 27:393‐ 418. Contractor A, Swanson G, Heinemann SF (2001) Kainate receptors are involved in short‐ and long‐term plasticity at mossy fiber synapses in the hippocampus. Neuron 29:209‐216. Contractor A, Mulle C, Swanson GT (2011) Kainate receptors coming of age: milestones of two decades of research. Trends Neurosci 34:154‐163. Contractor A, Swanson GT, Sailer A, O'Gorman S, Heinemann SF (2000) Identification of the kainate receptor subunits underlying modulation of excitatory synaptic transmission in the CA3 region of the hippocampus. J Neurosci 20:8269‐8278. Contractor A, Sailer AW, Darstein M, Maron C, Xu J, Swanson GT, Heinemann SF (2003) Loss of kainate receptor‐mediated heterosynaptic facilitation of mossy‐fiber synapses in KA2‐/‐ mice. J Neurosci 23:422‐429. Contractor AS, G. T. (2008) Kainate Receptors. In The glutamate receptors. Totowa: Humana Press. Copits BA, Robbins JS, Frausto S, Swanson GT (2011) Synaptic targeting and functional modulation of GluK1 kainate receptors by the auxiliary neuropilin and tolloid‐like (NETO) proteins. J Neurosci 31:7334‐7340.

168

Cossart R, Esclapez M, Hirsch JC, Bernard C, Ben‐Ari Y (1998) GluR5 kainate receptor activation in interneurons increases tonic inhibition of pyramidal cells. Nat Neurosci 1:470‐478. Cossart R, Tyzio R, Dinocourt C, Esclapez M, Hirsch JC, Ben‐Ari Y, Bernard C (2001) Presynaptic kainate receptors that enhance the release of GABA on CA1 hippocampal interneurons. Neuron 29:497‐ 508. Cossart R, Epsztein J, Tyzio R, Becq H, Hirsch J, Ben‐Ari Y, Crepel V (2002) Quantal release of glutamate generates pure kainate and mixed AMPA/kainate EPSCs in hippocampal neurons. Neuron 35:147‐159. Coussen F (2009) Molecular determinants of kainate receptor trafficking. Neuroscience 158:25‐35. Coussen F, Normand E, Marchal C, Costet P, Choquet D, Lambert M, Mege RM, Mulle C (2002) Recruitment of the kainate receptor subunit glutamate receptor 6 by cadherin/catenin complexes. J Neurosci 22:6426‐6436. Coussen F, Perrais D, Jaskolski F, Sachidhanandam S, Normand E, Bockaert J, Marin P, Mulle C (2005) Co‐assembly of two GluR6 kainate receptor splice variants within a functional protein complex. Neuron 47:555‐566. Cowan WM, Sudhof TC, Stevens CF, eds (2001) Synapses. Baltimore: The Johns Hopkins University Press Cui C, Mayer ML (1999) Heteromeric kainate receptors formed by the coassembly of GluR5, GluR6, and GluR7. J Neurosci 19:8281‐8291. Darstein M, Petralia RS, Swanson GT, Wenthold RJ, Heinemann SF (2003) Distribution of kainate receptor subunits at hippocampal mossy fiber synapses. J Neurosci 23:8013‐8019. Dateki M, Horii T, Kasuya Y, Mochizuki R, Nagao Y, Ishida J, Sugiyama F, Tanimoto K, Yagami K, Imai H, Fukamizu A (2005) Neurochondrin negatively regulates CaMKII phosphorylation, and nervous system‐specific gene disruption results in epileptic seizure. J Biol Chem 280:20503‐20508. Daumas S, Ceccom J, Halley H, Frances B, Lassalle JM (2009) Activation of metabotropic glutamate receptor type 2/3 supports the involvement of the hippocampal mossy fiber pathway on contextual fear memory consolidation. Learn Mem 16:504‐507. DeVries SH, Schwartz EA (1999) Kainate receptors mediate synaptic transmission between cones and 'Off' bipolar cells in a mammalian retina. Nature 397:157‐160. Dias JM, Carvalho AL, Kolln I, Calvete JJ, Topfer‐Petersen E, Varela PF, Romero A, Urbanke C, Romao MJ (1997) Crystallization and preliminary X‐ray diffraction studies of aSFP, a bovine seminal plasma protein with a single CUB domain architecture. Protein Sci 6:725‐727. Dingledine R, Borges K, Bowie D, Traynelis SF (1999) The glutamate receptor ion channels. Pharmacol Rev 51:7‐61. Dong H, O'Brien RJ, Fung ET, Lanahan AA, Worley PF, Huganir RL (1997) GRIP: a synaptic PDZ domain‐ containing protein that interacts with AMPA receptors. Nature 386:279‐284. Dong H, Zhang P, Song I, Petralia RS, Liao D, Huganir RL (1999) Characterization of the glutamate receptor‐interacting proteins GRIP1 and GRIP2. J Neurosci 19:6930‐6941. Doyle JP, Dougherty JD, Heiman M, Schmidt EF, Stevens TR, Ma G, Bupp S, Shrestha P, Shah RD, Doughty ML, Gong S, Greengard P, Heintz N (2008) Application of a translational profiling approach for the comparative analysis of CNS cell types. Cell 135:749‐762. Eccles JC, Ito M, Szentagothai J (1967) The cerebellum as a neuronal machine. New York: Springer. Egebjerg J, Heinemann SF (1993) Ca2+ permeability of unedited and edited versions of the kainate selective glutamate receptor GluR6. Proc Natl Acad Sci U S A 90:755‐759. Egebjerg J, Bettler B, Hermans‐Borgmeyer I, Heinemann S (1991) Cloning of a cDNA for a glutamate receptor subunit activated by kainate but not AMPA. Nature 351:745‐748. Erreger K, Chen PE, Wyllie DJ, Traynelis SF (2004) Glutamate receptor gating. Crit Rev Neurobiol 16:187‐ 224. Evans GJ (2007) Synaptic signalling in cerebellar plasticity. Biol Cell 99:363‐378. 169

Fernandes HB, Catches JS, Petralia RS, Copits BA, Xu J, Russell TA, Swanson GT, Contractor A (2009) High‐affinity kainate receptor subunits are necessary for ionotropic but not metabotropic signaling. Neuron 63:818‐829. Foster AC, Mena EE, Monaghan DT, Cotman CW (1981) Synaptic localization of kainic acid binding sites. Nature 289:73‐75. Frerking M, Nicoll RA (2000) Synaptic kainate receptors. Curr Opin Neurobiol 10:342‐351. Frerking M, Ohliger‐Frerking P (2002) AMPA receptors and kainate receptors encode different features of afferent activity. J Neurosci 22:7434‐7443. Frerking M, Malenka RC, Nicoll RA (1998) Synaptic activation of kainate receptors on hippocampal interneurons. Nat Neurosci 1:479‐486. Freund TF, Buzsaki G (1996) Interneurons of the hippocampus. Hippocampus 6:347‐470. Fritschy JM, Weinmann O, Wenzel A, Benke D (1998) Synapse‐specific localization of NMDA and GABA(A) receptor subunits revealed by antigen‐retrieval immunohistochemistry. J Comp Neurol 390:194‐210. Gally C, Eimer S, Richmond JE, Bessereau JL (2004) A transmembrane protein required for acetylcholine receptor clustering in Caenorhabditis elegans. Nature 431:578‐582. Gallyas F, Jr., Ball SM, Molnar E (2003) Assembly and cell surface expression of KA‐2 subunit‐containing kainate receptors. J Neurochem 86:1414‐1427. Garcia EP, Mehta S, Blair LA, Wells DG, Shang J, Fukushima T, Fallon JR, Garner CC, Marshall J (1998) SAP90 binds and clusters kainate receptors causing incomplete desensitization. Neuron 21:727‐ 739. Greengard P (2001) The neurobiology of slow synaptic transmission. Science 294:1024‐1030. Harris EW, Cotman CW (1986) Long‐term potentiation of guinea pig mossy fiber responses is not blocked by N‐methyl D‐aspartate antagonists. Neurosci Lett 70:132‐137. Hashimoto K, Fukaya M, Qiao X, Sakimura K, Watanabe M, Kano M (1999) Impairment of AMPA receptor function in cerebellar granule cells of ataxic mutant mouse stargazer. J Neurosci 19:6027‐6036. Hayes DM, Braud S, Hurtado DE, McCallum J, Standley S, Isaac JT, Roche KW (2003) Trafficking and surface expression of the glutamate receptor subunit, KA2. Biochem Biophys Res Commun 310:8‐13. He Z, Tessier‐Lavigne M (1997) Neuropilin is a receptor for the axonal chemorepellent Semaphorin III. Cell 90:739‐751. Heck D (1993) Rat cerebellar cortex in vitro responds specifically to moving stimuli. Neurosci Lett 157:95‐98. Heckmann M, Bufler J, Franke C, Dudel J (1996) Kinetics of homomeric GluR6 glutamate receptor channels. Biophys J 71:1743‐1750. Herb A, Burnashev N, Werner P, Sakmann B, Wisden W, Seeburg PH (1992) The KA‐2 subunit of excitatory amino acid receptors shows widespread expression in brain and forms ion channels with distantly related subunits. Neuron 8:775‐785. Herman BH, ed (2003) Glutamate and addiction. Totowa: Humana Press. Herrup K, Kuemerle B (1997) The compartmentalization of the cerebellum. Annu Rev Neurosci 20:61‐90. Hirbec H, Francis JC, Lauri SE, Braithwaite SP, Coussen F, Mulle C, Dev KK, Coutinho V, Meyer G, Isaac JT, Collingridge GL, Henley JM (2003) Rapid and differential regulation of AMPA and kainate receptors at hippocampal mossy fibre synapses by PICK1 and GRIP. Neuron 37:625‐638. Hollmann M, Heinemann S (1994) Cloned glutamate receptors. Annu Rev Neurosci 17:31‐108. Huang YH, Dykes‐Hoberg M, Tanaka K, Rothstein JD, Bergles DE (2004) Climbing fiber activation of EAAT4 transporters and kainate receptors in cerebellar Purkinje cells. J Neurosci 24:103‐111.

170

Hubner CA, Stein V, Hermans‐Borgmeyer I, Meyer T, Ballanyi K, Jentsch TJ (2001) Disruption of KCC2 reveals an essential role of K‐Cl cotransport already in early synaptic inhibition. Neuron 30:515‐ 524. Huettner JE (2003) Kainate receptors and synaptic transmission. Prog Neurobiol 70:387‐407. Husi H, Ward MA, Choudhary JS, Blackstock WP, Grant SG (2000) Proteomic analysis of NMDA receptor‐ adhesion protein signaling complexes. Nat Neurosci 3:661‐669. Huttner WB, Schiebler W, Greengard P, De Camilli P (1983) Synapsin I (protein I), a nerve terminal‐ specific phosphoprotein. III. Its association with synaptic vesicles studied in a highly purified synaptic vesicle preparation. J Cell Biol 96:1374‐1388. Isaac JT, Mellor J, Hurtado D, Roche KW (2004) Kainate receptor trafficking: physiological roles and molecular mechanisms. Pharmacol Ther 104:163‐172. Ishizuka N, Cowan WM, Amaral DG (1995) A quantitative analysis of the dendritic organization of pyramidal cells in the rat hippocampus. J Comp Neurol 362:17‐45. Ivakine EA, Brooke AA, Mahadevan V, Ormond J, Tang M, Pressey J, Ng D, Delpire E, Salter MW, Woodin M, McInnes RR (2012) Neto2 is a KCC2 auxiliary protein required for neuronal Cl‐ regulation in mature hippocampal neurons. In. Jaarsma D, Wenthold RJ, Mugnaini E (1995) Glutamate receptor subunits at mossy fiber‐unipolar brush cell synapses: light and electron microscopic immunocytochemical study in cerebellar cortex of rat and cat. J Comp Neurol 357:145‐160. Jackson AC, Nicoll RA (2009) Neuroscience: AMPA receptors get 'pickled'. Nature 458:585‐586. Jackson AC, Nicoll RA (2011) The expanding social network of ionotropic glutamate receptors: TARPs and other transmembrane auxiliary subunits. Neuron 70:178‐199. Jane DE, Lodge D, Collingridge GL (2009) Kainate receptors: pharmacology, function and therapeutic potential. Neuropharmacology 56:90‐113. Jaskolski F, Coussen F, Mulle C (2005a) Subcellular localization and trafficking of kainate receptors. Trends Pharmacol Sci 26:20‐26. Jaskolski F, Normand E, Mulle C, Coussen F (2005b) Differential trafficking of GluR7 kainate receptor subunit splice variants. J Biol Chem 280:22968‐22976. Jaskolski F, Coussen F, Nagarajan N, Normand E, Rosenmund C, Mulle C (2004) Subunit composition and alternative splicing regulate membrane delivery of kainate receptors. J Neurosci 24:2506‐2515. Kakegawa W, Tsuzuki K, Yoshida Y, Kameyama K, Ozawa S (2004) Input‐ and subunit‐specific AMPA receptor trafficking underlying long‐term potentiation at hippocampal CA3 synapses. Eur J Neurosci 20:101‐110. Kalia LV, Salter MW (2003) Interactions between Src family protein tyrosine kinases and PSD‐95. Neuropharmacology 45:720‐728. Kamiya H, Ozawa S (2000) Kainate receptor‐mediated presynaptic inhibition at the mouse hippocampal mossy fibre synapse. J Physiol 523 Pt 3:653‐665. Kandel ER, Schwartz JH, Jessell TM (2000) Principles of neural science, 4th Edition. New York: McGraw‐ Hill Medical. Kidd FL, Isaac JT (1999) Developmental and activity‐dependent regulation of kainate receptors at thalamocortical synapses. Nature 400:569‐573. Kim YJ, Bao H, Bonanno L, Zhang B, Serpe M (2012) Drosophila Neto is essential for clustering glutamate receptors at the neuromuscular junction. Genes Dev. Ko S, Zhao MG, Toyoda H, Qiu CS, Zhuo M (2005) Altered behavioral responses to noxious stimuli and fear in glutamate receptor 5 (GluR5)‐ or GluR6‐deficient mice. J Neurosci 25:977‐984. Koduri V, Blacklow SC (2001) Folding determinants of LDL receptor type A modules. Biochemistry 40:12801‐12807.

171

Kolodkin AL, Levengood DV, Rowe EG, Tai YT, Giger RJ, Ginty DD (1997) Neuropilin is a semaphorin III receptor. Cell 90:753‐762. Kullmann DM (2001) Presynaptic kainate receptors in the hippocampus: slowly emerging from obscurity. Neuron 32:561‐564. Kwon HB, Castillo PE (2008a) Role of glutamate autoreceptors at hippocampal mossy fiber synapses. Neuron 60:1082‐1094. Kwon HB, Castillo PE (2008b) Long‐term potentiation selectively expressed by NMDA receptors at hippocampal mossy fiber synapses. Neuron 57:108‐120. Laezza F, Wilding TJ, Sequeira S, Coussen F, Zhang XZ, Hill‐Robinson R, Mulle C, Huettner JE, Craig AM (2007) KRIP6: a novel BTB/kelch protein regulating function of kainate receptors. Mol Cell Neurosci 34:539‐550. Lauri SE, Delany C, VR JC, Bortolotto ZA, Ornstein PL, J TRI, Collingridge GL (2001a) Synaptic activation of a presynaptic kainate receptor facilitates AMPA receptor‐mediated synaptic transmission at hippocampal mossy fibre synapses. Neuropharmacology 41:907‐915. Lauri SE, Bortolotto ZA, Bleakman D, Ornstein PL, Lodge D, Isaac JT, Collingridge GL (2001b) A critical role of a facilitatory presynaptic kainate receptor in mossy fiber LTP. Neuron 32:697‐709. Leavitt S, Freire E (2001) Direct measurement of protein binding energetics by isothermal titration calorimetry. Curr Opin Struct Biol 11:560‐566. Lei S, Pelkey KA, Topolnik L, Congar P, Lacaille JC, McBain CJ (2003) Depolarization‐induced long‐term depression at hippocampal mossy fiber‐CA3 pyramidal neuron synapses. J Neurosci 23:9786‐ 9795. Lerma J (2003) Roles and rules of kainate receptors in synaptic transmission. Nat Rev Neurosci 4:481‐ 495. Lerma J (2006) Kainate receptor physiology. Curr Opin Pharmacol 6:89‐97. Letts VA, Felix R, Biddlecome GH, Arikkath J, Mahaffey CL, Valenzuela A, Bartlett FS, 2nd, Mori Y, Campbell KP, Frankel WN (1998) The mouse stargazer gene encodes a neuronal Ca2+‐channel gamma subunit. Nat Genet 19:340‐347. Levitan IB, Kaczmarek LK (2001) The Neuron: Cell and molecular biology, 3rd Edition. Oxford: Oxford University Press. Li H, Rogawski MA (1998) GluR5 kainate receptor mediated synaptic transmission in rat basolateral amygdala in vitro. Neuropharmacology 37:1279‐1286. Li P, Wilding TJ, Kim SJ, Calejesan AA, Huettner JE, Zhuo M (1999) Kainate‐receptor‐mediated sensory synaptic transmission in mammalian spinal cord. Nature 397:161‐164. Liu SJ, Cull‐Candy SG (2005) Subunit interaction with PICK and GRIP controls Ca2+ permeability of AMPARs at cerebellar synapses. Nat Neurosci 8:768‐775. Lomeli H, Wisden W, Kohler M, Keinanen K, Sommer B, Seeburg PH (1992) High‐affinity kainate and domoate receptors in rat brain. FEBS Lett 307:139‐143. Long J, Wei Z, Feng W, Yu C, Zhao YX, Zhang M (2008) Supramodular nature of GRIP1 revealed by the structure of its PDZ12 tandem in complex with the carboxyl tail of Fras1. J Mol Biol 375:1457‐ 1468. Lu W, Ziff EB (2005) PICK1 interacts with ABP/GRIP to regulate AMPA receptor trafficking. Neuron 47:407‐421. Mahley RW (1988) Apolipoprotein E: cholesterol transport protein with expanding role in cell biology. Science 240:622‐630. Maingret F, Lauri SE, Taira T, Isaac JT (2005) Profound regulation of neonatal CA1 rat hippocampal GABAergic transmission by functionally distinct kainate receptor populations. J Physiol 567:131‐ 142. Malenka RC, Bear MF (2004) LTP and LTD: an embarrassment of riches. Neuron 44:5‐21. 172

Mao L, Takamiya K, Thomas G, Lin DT, Huganir RL (2010) GRIP1 and 2 regulate activity‐dependent AMPA receptor recycling via exocyst complex interactions. Proc Natl Acad Sci U S A 107:19038‐ 19043. Marchal C, Mulle C (2004) Postnatal maturation of mossy fibre excitatory transmission in mouse CA3 pyramidal cells: a potential role for kainate receptors. J Physiol 561:27‐37. Mayer ML (2005) Glutamate receptor ion channels. Curr Opin Neurobiol 15:282‐288. Mayer ML (2006) Glutamate receptors at atomic resolution. Nature 440:456‐462. McBain CJ, Mayer ML (1994) N‐methyl‐D‐aspartic acid receptor structure and function. Physiol Rev 74:723‐760. Mehta S, Wu H, Garner CC, Marshall J (2001) Molecular mechanisms regulating the differential association of kainate receptor subunits with SAP90/PSD‐95 and SAP97. J Biol Chem 276:16092‐ 16099. Melyan Z, Wheal HV, Lancaster B (2002) Metabotropic‐mediated kainate receptor regulation of IsAHP and excitability in pyramidal cells. Neuron 34:107‐114. Melyan Z, Lancaster B, Wheal HV (2004) Metabotropic regulation of intrinsic excitability by synaptic activation of kainate receptors. J Neurosci 24:4530‐4534. Michishita M, Ikeda T, Nakashiba T, Ogawa M, Tashiro K, Honjo T, Doi K, Itohara S, Endo S (2003) A novel gene, Btcl1, encoding CUB and LDLa domains is expressed in restricted areas of mouse brain. Biochem Biophys Res Commun 306:680‐686. Michishita M, Ikeda T, Nakashiba T, Ogawa M, Tashiro K, Honjo T, Doi K, Itohara S, Endo S (2004) Expression of Btcl2, a novel member of Btcl gene family, during development of the central nervous system. Brain Res Dev Brain Res 153:135‐142. Mittelstaedt T, Alvarez‐Baron E, Schoch S (2010) RIM proteins and their role in synapse function. Biol Chem 391:599‐606. Miyata M, Imoto K (2006) Different composition of glutamate receptors in corticothalamic and lemniscal synaptic responses and their roles in the firing responses of ventrobasal thalamic neurons in juvenile mice. J Physiol 575:161‐174. Monaghan DT, Cotman CW (1982) The distribution of [3H]kainic acid binding sites in rat CNS as determined by autoradiography. Brain Res 252:91‐100. More JC, Nistico R, Dolman NP, Clarke VR, Alt AJ, Ogden AM, Buelens FP, Troop HM, Kelland EE, Pilato F, Bleakman D, Bortolotto ZA, Collingridge GL, Jane DE (2004) Characterisation of UBP296: a novel, potent and selective kainate receptor antagonist. Neuropharmacology 47:46‐64. Mulle C, Sailer A, Swanson GT, Brana C, O'Gorman S, Bettler B, Heinemann SF (2000) Subunit composition of kainate receptors in hippocampal interneurons. Neuron 28:475‐484. Mulle C, Sailer A, Perez‐Otano I, Dickinson‐Anson H, Castillo PE, Bureau I, Maron C, Gage FH, Mann JR, Bettler B, Heinemann SF (1998) Altered synaptic physiology and reduced susceptibility to kainate‐induced seizures in GluR6‐deficient mice. Nature 392:601‐605. Napper RM, Harvey RJ (1988) Number of parallel fiber synapses on an individual Purkinje cell in the cerebellum of the rat. J Comp Neurol 274:168‐177. Ng D (2006) Wiring the brain with Neto1: A multivalent NMDA receptor interacting CUB domain protein with essential roles in axon guidance, synaptic plasticity, and hippocampal‐dependent spatial learning and memory. In. Ng D, Pitcher GM, Szilard RK, Sertie A, Kanisek M, Clapcote SJ, Lipina T, Kalia LV, Joo D, McKerlie C, Cortez M, Roder JC, Salter MW, McInnes RR (2009) Neto1 is a novel CUB‐domain NMDA receptor‐interacting protein required for synaptic plasticity and learning. PLoS Biol 7:e41. Nicoll RA, Schmitz D (2005) Synaptic plasticity at hippocampal mossy fibre synapses. Nat Rev Neurosci 6:863‐876.

173

Nusser Z, Lujan R, Laube G, Roberts JD, Molnar E, Somogyi P (1998) Cell type and pathway dependence of synaptic AMPA receptor number and variability in the hippocampus. Neuron 21:545‐559. Osten P, Khatri L, Perez JL, Kohr G, Giese G, Daly C, Schulz TW, Wensky A, Lee LM, Ziff EB (2000) Mutagenesis reveals a role for ABP/GRIP binding to GluR2 in synaptic surface accumulation of the AMPA receptor. Neuron 27:313‐325. Ozawa S, Kamiya H, Tsuzuki K (1998) Glutamate receptors in the mammalian central nervous system. Prog Neurobiol 54:581‐618. Paternain AV, Morales M, Lerma J (1995) Selective antagonism of AMPA receptors unmasks kainate receptor‐mediated responses in hippocampal neurons. Neuron 14:185‐189. Paternain AV, Rodriguez‐Moreno A, Villarroel A, Lerma J (1998) Activation and desensitization properties of native and recombinant kainate receptors. Neuropharmacology 37:1249‐1259. Paternain AV, Herrera MT, Nieto MA, Lerma J (2000) GluR5 and GluR6 kainate receptor subunits coexist in hippocampal neurons and coassemble to form functional receptors. J Neurosci 20:196‐205. Patneau DK, Vyklicky L, Jr., Mayer ML (1993) Hippocampal neurons exhibit cyclothiazide‐sensitive rapidly desensitizing responses to kainate. J Neurosci 13:3496‐3509. Paxinos G, ed (2004) The rat nervous system, 3rd Edition. New York: Elsevier. Payne JA, Stevenson TJ, Donaldson LF (1996) Molecular characterization of a putative K‐Cl cotransporter in rat brain. A neuronal‐specific isoform. J Biol Chem 271:16245‐16252. Pelkey KA, Lavezzari G, Racca C, Roche KW, McBain CJ (2005) mGluR7 is a metaplastic switch controlling bidirectional plasticity of feedforward inhibition. Neuron 46:89‐102. Pemberton KE, Belcher SM, Ripellino JA, Howe JR (1998) High‐affinity kainate‐type ion channels in rat cerebellar granule cells. J Physiol 510 ( Pt 2):401‐420. Perrais D, Veran J, Mulle C (2010) Gating and permeation of kainate receptors: differences unveiled. Trends Pharmacol Sci 31:516‐522. Petralia RS, Wang YX, Wenthold RJ (1994) Histological and ultrastructural localization of the kainate receptor subunits, KA2 and GluR6/7, in the rat nervous system using selective antipeptide antibodies. J Comp Neurol 349:85‐110. Pierce MM, Raman CS, Nall BT (1999) Isothermal titration calorimetry of protein‐protein interactions. Methods 19:213‐221. Pinheiro P, Mulle C (2006) Kainate receptors. Cell Tissue Res 326:457‐482. Pinheiro PS, Mulle C (2008) Presynaptic glutamate receptors: physiological functions and mechanisms of action. Nat Rev Neurosci 9:423‐436. Pinheiro PS, Perrais D, Coussen F, Barhanin J, Bettler B, Mann JR, Malva JO, Heinemann SF, Mulle C (2007) GluR7 is an essential subunit of presynaptic kainate autoreceptors at hippocampal mossy fiber synapses. Proc Natl Acad Sci U S A 104:12181‐12186. Prybylowski K, Wenthold RJ (2004) N‐Methyl‐D‐aspartate receptors: subunit assembly and trafficking to the synapse. J Biol Chem 279:9673‐9676. Purves D (2008) Principles of cognitive neuroscience. Sunderland, Mass.: Sinauer Associates. Ramnani N (2006) The primate cortico‐cerebellar system: anatomy and function. Nat Rev Neurosci 7:511‐522. Rebola N, Lujan R, Cunha RA, Mulle C (2008) Adenosine A2A receptors are essential for long‐term potentiation of NMDA‐EPSCs at hippocampal mossy fiber synapses. Neuron 57:121‐134. Ren Z, Riley NJ, Garcia EP, Sanders JM, Swanson GT, Marshall J (2003a) Multiple trafficking signals regulate kainate receptor KA2 subunit surface expression. J Neurosci 23:6608‐6616. Ren Z, Riley NJ, Needleman LA, Sanders JM, Swanson GT, Marshall J (2003b) Cell surface expression of GluR5 kainate receptors is regulated by an endoplasmic reticulum retention signal. J Biol Chem 278:52700‐52709.

174

Represa A, Tremblay E, Ben‐Ari Y (1987) Kainate binding sites in the hippocampal mossy fibers: localization and plasticity. Neuroscience 20:739‐748. Rettig J, Neher E (2002) Emerging roles of presynaptic proteins in Ca++‐triggered exocytosis. Science 298:781‐785. Ripellino JA, Neve RL, Howe JR (1998) Expression and heteromeric interactions of non‐N‐methyl‐D‐ aspartate glutamate receptor subunits in the developing and adult cerebellum. Neuroscience 82:485‐497. Rivera C, Voipio J, Payne JA, Ruusuvuori E, Lahtinen H, Lamsa K, Pirvola U, Saarma M, Kaila K (1999) The K+/Cl‐ co‐transporter KCC2 renders GABA hyperpolarizing during neuronal maturation. Nature 397:251‐255. Rodriguez‐Moreno A, Lerma J (1998) Kainate receptor modulation of GABA release involves a metabotropic function. Neuron 20:1211‐1218. Rodriguez‐Moreno A, Sihra TS (2007a) Kainate receptors with a metabotropic modus operandi. Trends Neurosci 30:630‐637. Rodriguez‐Moreno A, Sihra TS (2007b) Metabotropic actions of kainate receptors in the CNS. J Neurochem 103:2121‐2135. Rodriguez‐Moreno A, Herreras O, Lerma J (1997) Kainate receptors presynaptically downregulate GABAergic inhibition in the rat hippocampus. Neuron 19:893‐901. Romero A, Romao MJ, Varela PF, Kolln I, Dias JM, Carvalho AL, Sanz L, Topfer‐Petersen E, Calvete JJ (1997) The crystal structures of two spermadhesins reveal the CUB domain fold. Nat Struct Biol 4:783‐788. Rouach N, Byrd K, Petralia RS, Elias GM, Adesnik H, Tomita S, Karimzadegan S, Kealey C, Bredt DS, Nicoll RA (2005) TARP gamma‐8 controls hippocampal AMPA receptor number, distribution and synaptic plasticity. Nat Neurosci 8:1525‐1533. Sakai R, Swanson GT, Shimamoto K, Green T, Contractor A, Ghetti A, Tamura‐Horikawa Y, Oiwa C, Kamiya H (2001) Pharmacological properties of the potent epileptogenic amino acid dysiherbaine, a novel glutamate receptor agonist isolated from the marine sponge Dysidea herbacea. J Pharmacol Exp Ther 296:650‐658. Salin PA, Scanziani M, Malenka RC, Nicoll RA (1996) Distinct short‐term plasticity at two excitatory synapses in the hippocampus. Proc Natl Acad Sci U S A 93:13304‐13309. Sanders JM, Ito K, Settimo L, Pentikainen OT, Shoji M, Sasaki M, Johnson MS, Sakai R, Swanson GT (2005) Divergent pharmacological activity of novel marine‐derived excitatory amino acids on glutamate receptors. J Pharmacol Exp Ther 314:1068‐1078. Savidge JR, Bleakman D, Bristow DR (1997) Identification of kainate receptor‐mediated intracellular calcium increases in cultured rat cerebellar granule cells. J Neurochem 69:1763‐1766. Scannevin RH, Huganir RL (2000) Postsynaptic organization and regulation of excitatory synapses. Nat Rev Neurosci 1:133‐141. Schiffer HH, Swanson GT, Heinemann SF (1997) Rat GluR7 and a carboxy‐terminal splice variant, GluR7b, are functional kainate receptor subunits with a low sensitivity to glutamate. Neuron 19:1141‐ 1146. Schmitz D, Mellor J, Nicoll RA (2001) Presynaptic kainate receptor mediation of frequency facilitation at hippocampal mossy fiber synapses. Science 291:1972‐1976. Schmitz D, Mellor J, Breustedt J, Nicoll RA (2003) Presynaptic kainate receptors impart an associative property to hippocampal mossy fiber long‐term potentiation. Nat Neurosci 6:1058‐1063. Schneider Gasser EM, Straub CJ, Panzanelli P, Weinmann O, Sassoe‐Pognetto M, Fritschy JM (2006) Immunofluorescence in brain sections: simultaneous detection of presynaptic and postsynaptic proteins in identified neurons. Nat Protoc 1:1887‐1897.

175

Schnell E, Sizemore M, Karimzadegan S, Chen L, Bredt DS, Nicoll RA (2002) Direct interactions between PSD‐95 and stargazin control synaptic AMPA receptor number. Proc Natl Acad Sci U S A 99:13902‐13907. Scott DB, Blanpied TA, Swanson GT, Zhang C, Ehlers MD (2001) An NMDA receptor ER retention signal regulated by phosphorylation and alternative splicing. J Neurosci 21:3063‐3072. Scott R, Lalic T, Kullmann DM, Capogna M, Rusakov DA (2008) Target‐cell specificity of kainate autoreceptor and Ca2+‐store‐dependent short‐term plasticity at hippocampal mossy fiber synapses. J Neurosci 28:13139‐13149. Sheng M, Sala C (2001) PDZ domains and the organization of supramolecular complexes. Annu Rev Neurosci 24:1‐29. Sheng M, Kim MJ (2002) Postsynaptic signaling and plasticity mechanisms. Science 298:776‐780. Shi S, Hayashi Y, Esteban JA, Malinow R (2001) Subunit‐specific rules governing AMPA receptor trafficking to synapses in hippocampal pyramidal neurons. Cell 105:331‐343. Shimell MJ, Ferguson EL, Childs SR, O'Connor MB (1991) The Drosophila dorsal‐ventral patterning gene tolloid is related to human bone morphogenetic protein 1. Cell 67:469‐481. Smith TC, Wang LY, Howe JR (1999) Distinct kainate receptor phenotypes in immature and mature mouse cerebellar granule cells. J Physiol 517 ( Pt 1):51‐58. Smolders I, Bortolotto ZA, Clarke VR, Warre R, Khan GM, O'Neill MJ, Ornstein PL, Bleakman D, Ogden A, Weiss B, Stables JP, Ho KH, Ebinger G, Collingridge GL, Lodge D, Michotte Y (2002) Antagonists of GLU(K5)‐containing kainate receptors prevent pilocarpine‐induced limbic seizures. Nat Neurosci 5:796‐804. Sommer B, Kohler M, Sprengel R, Seeburg PH (1991) RNA editing in brain controls a determinant of ion flow in glutamate‐gated channels. Cell 67:11‐19. Sommer B, Burnashev N, Verdoorn TA, Keinanen K, Sakmann B, Seeburg PH (1992) A glutamate receptor channel with high affinity for domoate and kainate. EMBO J 11:1651‐1656. Squire LR (2003) Fundamental Neuroscience, 2nd Edition. San Diego: Academic Press. Standley S, Roche KW, McCallum J, Sans N, Wenthold RJ (2000) PDZ domain suppression of an ER retention signal in NMDA receptor NR1 splice variants. Neuron 28:887‐898. Stohr H, Berger C, Frohlich S, Weber BH (2002) A novel gene encoding a putative transmembrane protein with two extracellular CUB domains and a low‐density lipoprotein class A module: isolation of alternatively spliced isoforms in retina and brain. Gene 286:223‐231. Straub C, Zhang W, Howe JR (2011a) Neto2 modulation of kainate receptors with different subunit compositions. J Neurosci 31:8078‐8082. Straub C, Hunt DL, Yamasaki M, Kim KS, Watanabe M, Castillo PE, Tomita S (2011b) Distinct functions of kainate receptors in the brain are determined by the auxiliary subunit Neto1. Nat Neurosci 14:866‐873. Sun Y, Olson R, Horning M, Armstrong N, Mayer M, Gouaux E (2002) Mechanism of glutamate receptor desensitization. Nature 417:245‐253. Swanson GT, Feldmeyer D, Kaneda M, Cull‐Candy SG (1996) Effect of RNA editing and subunit co‐ assembly single‐channel properties of recombinant kainate receptors. J Physiol 492 ( Pt 1):129‐ 142. Tanaka M, Yamaguchi K, Tatsukawa T, Theis M, Willecke K, Itohara S (2008) Connexin43 and bergmann glial gap junctions in cerebellar function. Front Neurosci 2:225‐233. Tang M, Pelkey KA, Ng D, Ivakine E, McBain CJ, Salter MW, McInnes RR (2011) Neto1 is an auxiliary subunit of native synaptic kainate receptors. J Neurosci 31:10009‐10018. Tomita S (2010) Regulation of ionotropic glutamate receptors by their auxiliary subunits. Physiology (Bethesda) 25:41‐49.

176

Tomita S, Chen L, Kawasaki Y, Petralia RS, Wenthold RJ, Nicoll RA, Bredt DS (2003) Functional studies and distribution define a family of transmembrane AMPA receptor regulatory proteins. J Cell Biol 161:805‐816. Torres R, Firestein BL, Dong H, Staudinger J, Olson EN, Huganir RL, Bredt DS, Gale NW, Yancopoulos GD (1998) PDZ proteins bind, cluster, and synaptically colocalize with Eph receptors and their ephrin ligands. Neuron 21:1453‐1463. Traynelis SF, Wollmuth LP, McBain CJ, Menniti FS, Vance KM, Ogden KK, Hansen KB, Yuan H, Myers SJ, Dingledine R, Sibley D (2010) Glutamate receptor ion channels: structure, regulation, and function. Pharmacol Rev 62:405‐496. Uchida N, Honjo Y, Johnson KR, Wheelock MJ, Takeichi M (1996) The catenin/cadherin adhesion system is localized in synaptic junctions bordering transmitter release zones. J Cell Biol 135:767‐779. Vignes M, Collingridge GL (1997) The synaptic activation of kainate receptors. Nature 388:179‐182. Vignes M, Clarke VR, Parry MJ, Bleakman D, Lodge D, Ornstein PL, Collingridge GL (1998) The GluR5 subtype of kainate receptor regulates excitatory synaptic transmission in areas CA1 and CA3 of the rat hippocampus. Neuropharmacology 37:1269‐1277. Vincent SR, McIntosh CH, Buchan AM, Brown JC (1985) Central somatostatin systems revealed with monoclonal antibodies. J Comp Neurol 238:169‐186. Vissel B, Royle GA, Christie BR, Schiffer HH, Ghetti A, Tritto T, Perez‐Otano I, Radcliffe RA, Seamans J, Sejnowski T, Wehner JM, Collins AC, O'Gorman S, Heinemann SF (2001) The role of RNA editing of kainate receptors in synaptic plasticity and seizures. Neuron 29:217‐227. Walker CS, Francis MM, Brockie PJ, Madsen DM, Zheng Y, Maricq AV (2006) Conserved SOL‐1 proteins regulate ionotropic glutamate receptor desensitization. Proc Natl Acad Sci U S A 103:10787‐ 10792. Watanabe M, Fukaya M, Sakimura K, Manabe T, Mishina M, Inoue Y (1998) Selective scarcity of NMDA receptor channel subunits in the stratum lucidum (mossy fibre‐recipient layer) of the mouse hippocampal CA3 subfield. Eur J Neurosci 10:478‐487. Wenthold RJ, Prybylowski K, Standley S, Sans N, Petralia RS (2003) Trafficking of NMDA receptors. Annu Rev Pharmacol Toxicol 43:335‐358. Werner P, Voigt M, Keinanen K, Wisden W, Seeburg PH (1991) Cloning of a putative high‐affinity kainate receptor expressed predominantly in hippocampal CA3 cells. Nature 351:742‐744. Wilding TJ, Huettner JE (1996) Antagonist pharmacology of kainate‐ and alpha‐amino‐3‐hydroxy‐5‐ methyl‐4‐isoxazolepropionic acid‐preferring receptors. Mol Pharmacol 49:540‐546. Williams SH, Johnston D (1991) Kinetic properties of two anatomically distinct excitatory synapses in hippocampal CA3 pyramidal neurons. J Neurophysiol 66:1010‐1020. Wisden W, Seeburg PH (1993) A complex mosaic of high‐affinity kainate receptors in rat brain. J Neurosci 13:3582‐3598. Wondolowski J, Frerking M (2009) Subunit‐dependent postsynaptic expression of kainate receptors on hippocampal interneurons in area CA1. J Neurosci 29:563‐574. Wu LJ, Ko SW, Zhuo M (2007a) Kainate receptors and pain: from dorsal root ganglion to the anterior cingulate cortex. Curr Pharm Des 13:1597‐1605. Wu LJ, Ko SW, Toyoda H, Zhao MG, Xu H, Vadakkan KI, Ren M, Knifed E, Shum F, Quan J, Zhang XH, Zhuo M (2007b) Increased anxiety‐like behavior and enhanced synaptic efficacy in the amygdala of GluR5 knockout mice. PLoS One 2:e167. Wyeth MS, Pelkey KA, Petralia RS, Tang M, Salter MW, McInnes RR, McBain CJ (2012) Neto protein interactions regulate pre‐ and postsynaptic GluK2/3 localization at hippocampal mossy fiber to CA3 pyramidal cell synapses. In: Society for Neuroscience. New Orleans, LA: 2012 Neuroscience Meeting Planner.

177

Wyszynski M, Kim E, Dunah AW, Passafaro M, Valtschanoff JG, Serra‐Pages C, Streuli M, Weinberg RJ, Sheng M (2002) Interaction between GRIP and liprin‐alpha/SYD2 is required for AMPA receptor targeting. Neuron 34:39‐52. Xia H, Hornby ZD, Malenka RC (2001) An ER retention signal explains differences in surface expression of NMDA and AMPA receptor subunits. Neuropharmacology 41:714‐723. Yamamoto T, Davis CG, Brown MS, Schneider WJ, Casey ML, Goldstein JL, Russell DW (1984) The human LDL receptor: a cysteine‐rich protein with multiple Alu sequences in its mRNA. Cell 39:27‐38. Yan S, Sanders JM, Xu J, Zhu Y, Contractor A, Swanson GT (2004) A C‐terminal determinant of GluR6 kainate receptor trafficking. J Neurosci 24:679‐691. Ye B, Liao D, Zhang X, Zhang P, Dong H, Huganir RL (2000) GRASP‐1: a neuronal RasGEF associated with the AMPA receptor/GRIP complex. Neuron 26:603‐617. Zalutsky RA, Nicoll RA (1990) Comparison of two forms of long‐term potentiation in single hippocampal neurons. Science 248:1619‐1624. Zhang W, St‐Gelais F, Grabner CP, Trinidad JC, Sumioka A, Morimoto‐Tomita M, Kim KS, Straub C, Burlingame AL, Howe JR, Tomita S (2009) A transmembrane accessory subunit that modulates kainate‐type glutamate receptors. Neuron 61:385‐396. Zheng Y, Mellem JE, Brockie PJ, Madsen DM, Maricq AV (2004) SOL‐1 is a CUB‐domain protein required for GLR‐1 glutamate receptor function in C. elegans. Nature 427:451‐457. Zheng Y, Brockie PJ, Mellem JE, Madsen DM, Walker CS, Francis MM, Maricq AV (2006) SOL‐1 is an auxiliary subunit that modulates the gating of GLR‐1 glutamate receptors in Caenorhabditis elegans. Proc Natl Acad Sci U S A 103:1100‐1105. Zhou LM, Gu ZQ, Costa AM, Yamada KA, Mansson PE, Giordano T, Skolnick P, Jones KA (1997) (2S,4R)‐4‐ methylglutamic acid (SYM 2081): a selective, high‐affinity ligand for kainate receptors. J Pharmacol Exp Ther 280:422‐427.

178

Appendix A: Putative Neto2 interacting molecules identified by a yeast two-hybrid screen of an adult mouse brain cDNA library

Genbank Protein Accession # NM_026377 SWI5 dependent recombination repair 1 (Sfr1) NM_007505 ATP synthase, H+ transportin, mitochondrial F1 complex, α‐ subunit, isoform 1 (Atp5a1) NM_053271 Regulating synaptic membrane exocytosis 2 (RIM2) NM_134050 RAB15, member RAS oncogene family (Rab15) NM_011986 Neurochondrin (Ncdn) NM_021432 Nucleosome assembly protein 1‐like 5 (Nap1l5) NM_133195 Elav‐like family member 4 (Celf4) NM_008655 Growth arrest and DNA damage inducible 45 beta (Gadd45b)

Appendix B: Proteins present in the GST-Neto2cyto pull down of adult mouse brain membrane fraction as detected by mass spectrometry

Genbank Protein Accession # NM_020333 Solute carrier family 12 (potassium‐chloride transporter), member 5 (Slc12a5) (synonym: KCC2) NM_144921 ATPase, Na+/H+ transporting, alpha 3 polypeptide (Atp1a3) NM_009001 RAB3A, member RAS oncogene family (Rab3a) NM_133769 Cytoplasmic FMR1 interacting protein 2 (Cyfip2) NM_007505 ATP synthase, H+ transporting, mitochondrial F1 complex, alpha subunit 1 (Atp5a)

179

Appendix C: Neto2 is associated with NMDARs in vivo

180