Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Neuroprotection by peptides designed to block interactions between glutamate receptors and neuronal activity… Smith, Carlo 2012

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2012_fall_smith_carlo.pdf [ 2.1MB ]
Metadata
JSON: 24-1.0073128.json
JSON-LD: 24-1.0073128-ld.json
RDF/XML (Pretty): 24-1.0073128-rdf.xml
RDF/JSON: 24-1.0073128-rdf.json
Turtle: 24-1.0073128-turtle.txt
N-Triples: 24-1.0073128-rdf-ntriples.txt
Original Record: 24-1.0073128-source.json
Full Text
24-1.0073128-fulltext.txt
Citation
24-1.0073128.ris

Full Text

NEUROPROTECTION BY PEPTIDES DESIGNED TO BLOCK DIRECT INTERACTIONS BETWEEN GLUTAMATE RECEPTORS AND NEURONAL ACTIVITY REGULATED PENTRAXIN by  Carlo Smith  B.Sc., Brock University, 2007  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE  in THE FACULTY OF GRADUATE STUDIES (Neuroscience)  THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver)  August 2012 ©Carlo Smith, 2012  ABSTRACT Excitotoxicity is caused by prolonged stimulation of N-methyl-D-aspartate type glutamate receptors (NMDARs) resulting in internalization of α-amino-3-hydroxy-5-methyl-4isoxazolepropionic acid type glutamate receptors (AMPARs) and long-term depression (LTD) of post-synaptic response, and this process has been causally linked to neuronal cell death. Indeed, NMDAR/AMPAR excitotoxicity is believed to underlie neurodegenerative diseases including Alzheimer’s disease, Parkinson’s disease and ischaemic brain injury. Neuronal activity regulated pentraxin (NARP) is a secreted immediate early gene product that binds to and clusters AMPAR subunits GluR1-4 in response to normal and pathological synaptic activity. To test whether NARP could potentiate excitotoxic neuronal death by clustering AMPARs at cortical synapses, we used peptide arrays to develop four peptides that mimic sites on GluR1 that might bind NARP. Three of the four peptides correspond to sites on the N-terminal domain of GluR1, a region previously implicated in facilitating NARP-mediated GluR1 clustering at synapses. We show that NARP is up-regulated 4-8 hours after excitotoxic NMDA treatment of primary cortical neurons. We found that a mixture of all four peptides inhibited NMDA-induced GluR1 internalization and was neuroprotective in a dose-dependent manner. Also, three of peptides were individually neuroprotective. We conclude that the peptides inhibit NARP’s ability to form synaptic GluR1 clusters which may be required for coordinated and sustained GluR1 internalization 4-8 hours after NMDA stimulation. This is an essential step in NMDA-induced cell death since blocking it is neuroprotective. These peptides suggest new approaches to treatment of neurodegenerative diseases caused by excitotoxicity.  ii  PREFACE This work involved primary cortical tissue cultures from embryonic rats and was given ethics approval by the UBC Committee on Animal Care. Certificate Number A10-0169  iii  Table of Contents  ABSTRACT .................................................................................................................................... ii PREFACE ...................................................................................................................................... iii Table of Contents ........................................................................................................................... iv List of Figures ............................................................................................................................... vii ACKNOWLEDGEMENTS ......................................................................................................... viii CHAPTER 1.  INTRODUCTION ............................................................................................. 1  1.1  General structure of neuronal activity regulated pentraxin protein.................................. 1  1.2  Known functions of neuronal activity regulated pentraxin .............................................. 6  1.2.1  Uptake of extracellular synaptic material ................................................................. 6  1.2.2  Excitatory synaptogenesis ......................................................................................... 7  1.2.3  Synaptic plasticity ................................................................................................... 10  1.3  Neuronal activity regulated pentraxin and neurodegenerative diseases......................... 17  1.3.1  Alzheimer’s disease ................................................................................................ 17  1.3.2  Parkinson’s disease ................................................................................................. 18  1.4  Aims of the thesis ........................................................................................................... 20  CHAPTER 2.  IDENTIFICATION OF PUTATIVE BINDING POCKETS FOR  NEURONAL ACTIVITY REGULATED PENTRAXIN ON GLUR1 ....................................... 21 2.1  Introduction .................................................................................................................... 21 iv  2.2  Methods .......................................................................................................................... 25  2.2.1  High density peptide array of the extracellular part of GluR1................................ 25  2.2.2  Synthesis of GluR1-derived cell-penetrating peptides ........................................... 26  2.2.3  Blocking assay for NARP on the membrane .......................................................... 27  2.3  Results ............................................................................................................................ 27  2.3.1  Peptide array reveals that NARP binds the NTD and one other region on the  extracellular part of GluR1 ................................................................................................... 27 2.3.2  Synthetic peptides can block the interaction between recombinant NARP and  GluR1 peptide array .............................................................................................................. 28 2.4  Discussion ...................................................................................................................... 29  2.4.1  NARP-binding pockets on the NTD of GluR1 ....................................................... 30  2.4.2  NARP binds GluR1 at a sequence linking the S1 glutamate binding module to the  first transmembrane domain. ................................................................................................ 32 CHAPTER 3.  EFFECTS OF NARP-BINDING PEPTIDES ON NMDA-INDUCED AMPA  RECEPTOR INTERNALIZATION AND CELL DEATH .......................................................... 41 3.1  Introduction .................................................................................................................... 41  3.2  Methods .......................................................................................................................... 43  3.2.1  Rat primary cortical neuron culture ........................................................................ 43  3.2.2  Dc protein assay ...................................................................................................... 44  3.2.3  SDS-PAGE and Western blot ................................................................................. 45  3.3  Cytotoxicity treatment with N-methyl-D-aspartic acid (NMDA) .................................. 46 v  3.3.1  Testing the of effect cytotoxic NMDA treatment on NARP expression ................ 46  3.3.2  Peptide treatment in cultured neurons..................................................................... 47  3.3.3  Cell surface biotinylation ........................................................................................ 47  3.3.4  Assessment of cell death: Lactate dehydrogenase assay ........................................ 48  3.4  Statistical Analyses ........................................................................................................ 49  3.5  Results ............................................................................................................................ 50  3.5.1  Excitotoxic NMDA treatment induces NARP protein expression in primary cortical  neurons 50 3.5.2  The peptides block NMDA-induced AMPA receptor internalization .................... 50  3.5.3  The peptides rescued neurons from NMDA-induced death ................................... 51  3.6  Discussion ...................................................................................................................... 54  3.6.1  Time course for NMDA-induced NARP expression in vitro ................................. 54  3.6.2  Up-regulation of NARP is linked to NMDA-induced GluR1 internalization ........ 56  3.6.3  Mechanism of inhibition of NARP-GluR1 interactions ......................................... 57  3.6.4  Optimal neuroprotection requires blockade of NARP-binding on the NTD and  glutamate-binding domain of GluR1 .................................................................................... 57 3.6.5  NARP-GluR1 interactions facilitate NMDAR-LTD and cell death ....................... 63  CHAPTER 4.  CONCLUSION & PERSPECTIVE ................................................................ 74  Literature Cited ............................................................................................................................. 79  vi  List of Figures Figure 2.1. Schematic of high density peptide array synthesis. .................................................... 36 Figure 2.2. Identification of peptides that may block the interaction between NARP and GluR1 N-terminus. ................................................................................................................................... 37 Figure 2.3. A mixture of all four peptides blocks the interaction of NARP with GluR1 Nterminus on the membrane. ........................................................................................................... 38 Figure 2.4. Location of NARP interacting peptides on the folded extracellular portion of GluR1 ....................................................................................................................................................... 39 Figure 2.5. Schematic showing location of synthetic peptides relative to functional domains of GluR1 ............................................................................................................................................ 40 Figure 3.1. NARP protein is up-regulated after excitotoxic NMDA treatment. ........................... 70 Figure 3.2. Peptides block NMDA- induced AMPA receptor internalization. ............................. 71 Figure 3.3. Synthetic GluR1 N-terminal derived peptides are neuroprotective. .......................... 72 Figure 3.4. Identification of the most potent peptides. ................................................................. 73  vii  ACKNOWLEDGEMENTS I would like to thank my supervisor, Dr. Max Cynader for giving me the privilege of working under his guidance and for encouraging me with his enthusiasm, knowledge and support.  I would also like to thank the members of my supervisory committee: Dr. Yu Tian Wang, Dr. Neil Cashman and Dr. Hakima Moukhles for their helpful discussion and advice.  I would like to thank Peter Axerio, PhD Candidate for his assistance with the 3-D model of the extracellular surface of GluR1. Thank you also to Dr. Loren Oschipok for synthesizing the peptides at the UBC Peptide Array Facility and to Wendy Wen for preparing all the primary neurons.  Many thanks my colleagues, Dr. Kaiyun Yang, Dr. Ainsley Coquinco, Dr. Guang Yang, Dr. Luba Kojic, Eddie Pokrishevsky, Guoyu Liu, Rui Liu, Wendy Wen, Dong Qiang, Bryson Armstrong, Kimberly Girling, Shelly Fan, Shanshan Zhu etc… for their mentorship and support.  My special thanks to my family for their continued confidence in me and support in all my endeavours.  This work was supported by NSERC (PGS M 2010-2011).  viii  CHAPTER 1.  INTRODUCTION  1.1 General structure of neuronal activity regulated pentraxin protein Neuronal activity regulated pentraxin (NARP) is a 46.2-kDa secreted synaptic protein that was originally identified by subtractive cloning from stimulated rat hippocampus in the 1990s [1]. The DNA for NARP is situated on the distal portion of mouse chromosome 5 [1] which has homology to the locus on human chromosome 7q21.3-q22.1 that encodes NPTX2, the human version of NARP [2]. NARP is a member of a family of proteins called pentraxins, which are secreted innate immunity proteins that are highly conserved across vertebrates [3]. These proteins self-multimerize into radially symmetric five-member rings and can further dimerize into decamers [1, 4-6]. They are functional ancestors of antibodies and function to cluster bacteria, toxins, chromatin, carbohydrates and cellular debris, marking them for macrophagemediated phagocytosis and degradation by opsonisation [4, 5, 7-9]. Pentraxins have a conserved 200 amino acid C-terminal called the pentraxin domain, and it contains the “pentraxin family signature” of eight amino acids: Histidine-X- Cysteine-X-Serine/Threonine-Tryptophan-XSerine (where “X” represents any amino acid) arranged into a β-sheet [7, 8]. Classical pentraxins include C-reactive protein (CRP) and serum amyloid protein (SAP), both found on chromosome 1q21-23. CRP is the canonical human acute phase protein recognized for its binding to pneumococcal C-polysaccharide. CRP specifically recognizes phosphocholine that is found in many micro-organisms and plays a role in exacerbating ischemic injury [10, 11] and nonantibody-mediated immunity by binding and aggregating bacteria and other pathogens in 1  preparation for phagocytosis [8]. SAP is a constitute plasma protein that coats molecules in the basement membrane protecting them from proteolytic degradation [12]. Interestingly, SAP is also found in all known forms of amyloid plaques including those of Alzheimer’s disease and could therefore contribute to some disease processes by inhibiting cleavage of amyloid fibrils [13, 14]. Both CRP and SAP have been shown to function in the brain to regulate phagocytic activity of microglia and complement C activation [3]. NARP is 24% and 26% homologous to CRP and SAP, respectively; over a 210 span of amino acids that includes the full sequence of mature CRP and SAP and the 216 amino acid C-terminal of NARP [1].  Comparative modelling based on the crystal structure of SAP [12] and CRP [11] suggests that NARP monomers contain at least 15 antiparallel β-strands that form two β-sheets arranged around a compact hydrophobic core of aromatic amino acids. In the pentraxin domain, this arrangement is essential to the ability of these proteins to bind complex polysaccharides/agar in a calcium-dependent manner. As a result, pentraxins can also be classified a calcium-dependent lectins (sugar binding molecules). The sugar binding motif of SAP, CRP and NARP is closely homologous to that of the plant lectin concanavalin A, with which they share secondary and tertiary structure [1, 12]. A physiological role for calcium-dependent sugar-binding by NARP has yet to be determined.  Pentraxins are divided into two subfamilies based on size; the classical short pentraxins (~25kDa) and novel long pentraxins (~40-55-kDa) [7, 15]. CRP and SAP are considered short pentraxins. Long pentraxins include NARP, neuronal pentraxin 1 (NP1), neuronal pentraxin receptor (NPR), tumour necrosis inducible acute phase reactant (PTX3) and guinea pig sperm  2  acrosome protein p50/apexin [6, 16]. NARP is 50% homologous to NP1 and NPR [6, 7, 9, 17] and its C-terminus is 20-30% homologous to that of other pentraxins [9, 18, 19]. NARP, NP1 and NPR are synaptic proteins, enriched in neurons [18, 20] and are collectively referred to as neuronal pentraxins. NARP and NP1 are also expressed at lower levels various tissues, most notably the pancreas and liver where NARP has been shown to be protective against tumorigenesis [21] and NP1 to contribute to β-cell dysfunction causing impairment of insulin secretion [22], respectively.  Neuronal pentraxins have a higher molecular weight than short pentraxins because their Nterminal is extended [1], containing twice as many amino acids [6]. While NARP and NP1 are secreted from neurons [18], NPR has a single N-terminal transmembrane domain anchoring it synaptic membranes [17]. NP1 and NPR are constitutively expressed whereas NARP expression is regulated by synaptic activity [18] and dopamine-dependent signalling [23, 24]. Thus, NARP may have distinct activity-dependent functions in the brain.  NARP has an N-terminal signal sequence with a putative cleavage site between glycine and glutamine. It also has three N-glycosylation sites at amino acids 133, 174 and 378. The Nterminal of NARP has a predicted secondary structure that is 61% alpha helical containing two amphipathic coiled domains (called coiled-coil domains) spanning amino acids 40-82 and 105132 [1]. These contain heptad repeats of hydrophobic amino acids and are implicated in anchoring NARP to synapses via association with proteins in the synaptic cleft, where there is an abundance of proteins with coiled-coil domains [17, 25]. Furthermore, Xu and colleagues [26] identified N-terminal cysteines which are essential for NARP multimerization. Cysteine 14 and  3  26 precede the first coiled domain and form position-specific disulphide bonds between NARP monomers leading to the formation of hexameric NARP rings. Arrangement of NARP monomers into six member rings was inferred from the molecular weight of multimeric NARP [1] and homology to a crystal structure of SAP [27]. Cysteine 79 is located C-terminal to the first coiled domain and disulphide bonds between these residues link NARP hexamers.  In addition to mediating parallel binding between NARP monomers, alignment of cysteine 14 and 79 in the N-termini of NARP, NP1 and NPR allow them to form heterocomplexes stabilized by both disulphide linkages and protein-protein interactions [17, 19, 26]. Cysteine 26 specifically links NARP monomers and does not interact with NP1 or NPR. In this way NARP can be secreted as a homo-multimer or co-processed with NP1 and/or NPR and secreted as a heteromultimer [19]. When co-processed with NPR, NARP is immobilized to synaptic membranes (both pre- and post-synaptic), another factor contributing to its localization at the synapse [28, 29].  The C-terminus of NARP contains the calcium binding pocket which essential to the proper folding of the pentraxin domain and mediates its sugar binding activity [12]. Molecular studies of endogenous NARP in developing spinal neurons revealed that the pentraxin domain (amino acids 234-432) is essential for axonal transport and secretion [25]. In addition to this, the pentraxin domains of both NARP and NP1 co-immunoprecipitate with GluR1-4 AMPA-type glutamate receptor subunits [30-32] but chimeric molecules containing the N-terminal of NARP and pentraxin domain of NP1 are far more successful in assays of AMPA receptor surface clustering than those with the inverse arrangement [26]. Therefore, NARP’s N-terminal mediates  4  polymerization of NARP into disulphide-linked NARP-NARP and NARP-NP1 multimers moreso than that of NP1. This causes enhanced secondary clustering of AMPA receptors due to combined NARP-GluR and NP1-GluR interactions [25, 26, 30].  The observation that NARP only interacts with GluR1-4 and not GluR6, NR1 or the neuronal glutamate transporter (EAA1) suggests that the binding is specific and not simply mediated by its lectin-binding capability. This was solidified by a study showing that co-immunoprecipitation of NARP by GluR1-4 is maintained despite the presence of tunicamycin, an antibiotic that blocks the glycosylation of surface proteins [30]. Moreover, NARP lacks PDZ (postsynaptic density 95/Discs large/zona occuldens-1)-binding domains [1, 33] which would allow it to cluster AMPA receptors through interactions with the C-terminal of GluRs in the cytoplasm.  Interestingly, Sia and colleagues [32] demonstrated that NARP/NP1 require the N-terminal domain (NTD) of GluRs to associate with AMPA receptors. This was observed at sites of contact between co-cultured glia and primary hippocampal neurons and as well as at synapse in which mutant NTD-deficient GluR4 receptors was expressed. Thus, NARP may directly interact with the extracellular surface of GluRs, a concept that is repeatedly proposed in the literature [3, 25, 30]. Moreover, endogenous NARP is often translated in distal dendrites of primary neurons before it is eventually secretion via presynaptic elements (axon to pre-synaptic terminal) [25]. This initial processing in the cytoplasm of dendrites coupled with its ability to bind glycosylated proteins suggests that NARP could be co-processed with other secreted or transmembrane proteins known to be in fast turnover at the synapse, including AMPA receptors. Although co-  5  immunoprecipitation does not prove that NARP directly interacts with the NTD of GluRs, the specificity of the interaction suggests a very close association between the two proteins.  Notably, O’Brien et al. [30] proposed that the N-terminal coiled-coil domain of NARP may also facilitate interactions with GluRs. Indeed, this domain contains hydrophobic and hydrophilic properties making it well suited to mediate interactions with transmembrane proteins.  1.2 Known functions of neuronal activity regulated pentraxin 1.2.1 Uptake of extracellular synaptic material Apart from their ability to bind AMPARs, neuronal pentraxins (NARP, NP1 and NPR) contain a calcium binding motif that this remarkably homologous to that of acute phase immunity pentraxins CRP and SAP [1]. This motif allows them to interact with molecules in calciumdependent manner. One such molecule is the presynaptic snake venom neurotoxin taipoxin. Taipoxin works by inhibiting presynaptic vesicle recycling in association with taipoxinassociated calcium binding protein (TCBP-49) which has six EF hand calcium-binding domains and a His-Asp-Glu-Leu endoplasmic reticulum luminal retention sequence [34]. NARP and NP1 bind to taipoxin affinity columns along with TCBP-49 in a calcium dependent manner [19, 29, 34]. In the presence of calcium, NARP and NP1 can also bind TCBP-49 in primary hippocampal neurons and transfected Chinese hamster ovary (CHO) cells [19]. NPR only binds to taipoxin binding columns and TCBP-49 when coexpressed with NARP and/or NP1. Furthermore, disulphide-linked heterocomplexes tethering NARP to the plasma membrane via NPR or NPRNP1 complexes in primary hippocampal neurons or CHO cells can be released from the membrane by proteolysis [29], and a later study showed that these complexes are endocytosed in 6  several types of primary neuron cultures [35]. Together these studies suggest that the calciumdependent ligand binding properties of neuronal pentraxins may be involved in presynaptic targeting of taipoxin, and thus could normally play a role in uptake of synaptic materials such as debris or proteins.  Missing from these studies is a demonstration that neuronal pentraxins are required for uptake of synaptic material related to synapse formation and remodeling in animal models or post-natal neuron cultures. Mice lacking NARP and NP1 would be ideal to test this possibility but no such studies have been reported.  Notwithstanding the possibility that neuronal pentraxins may be involved in uptake of synaptic material, the body of literature tends to favour the idea that they are physiologically relevant in the brain because they can cluster AMPARs [25, 26, 28, 30, 36]. In this view, NARP is of particular interest because it most potently aggregates AMPARs at sites of cell contact [25, 26, 30, 37] and its expression is regulated by synaptic activity thought to underlie information storage, synaptic plasticity and neurodegenerative disease [1, 38].  1.2.2 Excitatory synaptogenesis Synapses allow communication between neurons and are formed at sites of contact between presynaptic and postsynaptic neurons. The presynaptic terminal is filled with neurotransmitter vesicles docked to the membrane by protein complexes primed for the signal that triggers release. The postsynaptic membrane contains neurotransmitter receptors clustered there by scaffolding and signaling proteins, ready to respond to neurotransmitter release. Molecules that  7  cluster neurotransmitter receptors at synapses are essential to synaptic maturation and confer functionality to synapses [28].  Glutamate is the major neurotransmitter at excitatory synapses in the brain and activates current through AMPA and NMDA receptors (AMPAR and NMDAR), required for synapse maturation and modulation of synaptic strength [28]. Cytosolic proteins containing PDZ-binding domains and modulators of endosomal trafficking both regulate clustering of glutamate receptors, however extracellular adhesion molecules and secreted proteins also contribute to this processes [28, 39].  Neuronal activity regulated pentraxin (NARP) is secreted from presynaptic terminals upon synaptic activation and co-localizes with clusters of AMPAR subunits GluR1-4 on the surface of cultured neurons and heterologously transfected HEK 293T cells [25, 26, 30]. NARP is transported in axons [30, 40] and immuno-electron microscopy of the molecular layer in the dentate gyrus of rat hippocampus revealed that it is present in presynaptic vesicle-like structures [30]. The mechanism of release of these NARP-vesicles during activity-dependent synaptogenesis is not completely understood.  A role for NARP in excitatory synaptogenesis was first proposed by O’Brien and colleagues [30] who used co-cultures of GluR1 expressing HEK cells and primary spinal neurons overexpressing NARP to show that secreted NARP clusters GluR1 at contact sites between the two cell types. This was solidified by a follow-up study showing that a dominant negative version of NARP that specifically inhibits presynaptic NARP secretion also prevents postsynaptic GluR1 clustering [25]. Dominant negative NARP can still multimerize with endogenous NARP produced in 8  primary spinal neurons but cannot be secreted because of broad deletions of the C-terminal pentraxin domain; effectively sequestering NARP inside neurons. In addition to this, Bjartmar and colleagues [9] showed that mice lacking NARP and NP1 exhibit defects in activitydependent segregation and refinement of eye-specific retinal ganglion cell projections to the dorsal lateral geniculate nucleus. This process is dependent on maturation of glutamatergic synapses. While the structure of these synapses is normal in young NARP/NP1 knockout mice, glutamatergic synaptic transmission is aberrant, as demonstrated by reductions in both the amplitude and frequency of spontaneous mEPSCs (miniature excitatory postsynaptic currents). However, glutamate transmission in older NARP/NP1 knockout mice was normal suggesting that NARP’s contribution to synaptogenesis is developmentally regulated and cell type specific. Together these studies suggest that NARP plays a role in maturation of excitatory synapses via recruiting AMPARs to postsynaptic membranes.  Worley and colleagues [26] showed that the synaptogenic effect of NARP is attributed to the ability of its N-terminal domain to form large surface aggregates of disulphide-linked multimers containing NARP and NP1, both of which also bind AMPARs via their homologous pentraxin domains. The N-terminal of NP1 allows it to multimerize via disulphide bonds as well but the number and size of the resulting AMPAR co-clusters is far less than that of NARP [26]. Interestingly, this work also revealed that NP1 expression is high during early postnatal activityindependent synaptogenesis in rat forebrain and levels drop tenfold between postnatal days 6 and 11, when activity-dependent synaptogenesis ensues. Coincident with the initiation of activitydependent synaptogenesis, there is a steep increase in NARP expression. Together, this suggests that NP1 secretion is more relevant to early, activity-independent synapse formation while  9  NARP is more relevant to later activity-dependent synapse formation. Thus, upon induction by synaptic activity, NARP is able to recruit that nascent synaptogenic ability of NP1 by polymerizing with it via N-terminal mediated disulphide bonds. Hence, the secondary coclustering of AMPARs at developing excitatory synapses is enhanced rendering functionality to these synapses. Indeed, synaptic potentiation by NARP induction has been shown to drive the conversion of silent synapses into functional ones in the developing visual system [41]. NARP/NP1 deficient mice show reduced AMPAR-mediated retinogeniculate transmission that prevents silent synapse conversion at these synapses [41].  1.2.3 Synaptic plasticity The two major forms of enduring plasticity in the mammalian brain are termed long-term potentiation (LTP) and long-term depression (LTD), and are characterized by a long-lasting increase or decrease in synaptic strength, respectively. There are several forms of both of these processes but ultimately they require modulation of postsynaptic AMPAR trafficking and surface expression to exert a lasting effect on synaptic transmission [42]. LTP and LTD are thought to underlie information storage and so to be involved in learning and memory and other physiological or pathological processes involving AMPARs [42-47]. Due to its regulation as an activity-dependent immediate early gene and ability to interact with surface AMPARs, NARP is poised to contribute to long-term synaptic plasticity.  1.2.3.1  In vivo long-term potentiation  NARP is rapidly regulated by physiological synaptic activity. The first demonstration of this was done by blocking natural synaptic activity in the rat visual cortex with monocular TTX 10  (tetrodoxin) injection [1]. This led to decreased NARP expression in the deafferented cortex. Furthermore, Worley and colleagues [1, 30] demonstrated that bilateral high frequency stimulation (HFS) of the perforant path and hilus of the dentate gyrus, induces LTP of excitatory inputs and NARP mRNA expression in the hippocampus. NARP expression is dependent on LTP expression since low frequency stimulation (LFS) did not induce NARP or LTP. Both LTP and NARP expression is blocked by the non-competitive NMDAR antagonist MK-801. HFSinduced LTP requires activation of NMDARs and synaptic recruitment of AMPARs, so this work provides a clear link between NARP expression and NMDAR activation in vivo.  Related experiments showed that NARP mRNA is induced in the hippocampus (dentate gyrus and hilus) and cortical layers 2/3 and 5/6 after maximal electroconvulsive shock (MECS) in rat brain [1, 30], a process that may be related to pathological forms of LTP that later induce seizures in these animals [48]. This was corroborated in model designed to examine the consequences of chronic neuronal stimulation in rats where Reti and colleagues [36] showed that NARP protein levels are sustained for up to 48 hours after repeated MECS (five or six given every other day) in rats. They proposed that through clustering AMPARs at synaptic sites, NARP elicits an enduring increase in the strength of excitatory synaptic transmission. This is supported by the fact that repeated MECS also increases GluR1 expression [49]; thus sustained NARP expression may be necessary to localize newly inserted AMPARs to synaptic sites.  1.2.3.2  Type I/V Metabotrophic glutamate receptor-dependent  Long-term depression Cho and colleagues [35] showed the neuronal pentraxins are required for mGluR1/5-dependent LTD. This team outlined a mechanism whereby the matrix metalloprotease, TACE, cleaves NPR 11  in an mGluR1/5 activation-dependent manner, and this is blocked by TACE inhibitors. Cleaved NPR interacts with surface NARP-AMPAR and NP1-AMPAR complexes and immuno-electron microscopy confirmed that NPR is rapidly internalized into vesicle-like structures with NARP/NP1 and their associated pool of AMPARs. Direct pharmacological activation of mGluR1/5, a known activator of TACE, potentiates AMPAR endocytosis and decreases surface AMPAR levels below steady-state. Both of these are blocked by TACE inhibitors and do not occur in neuronal pentraxin triple knockout animals. This process seems to be generalizable to all excitatory synapses since these findings were repeated in primary hippocampal, cortical and cerebellar Purkinje neuronal cell cultures, as well as, hippocampal Schaeffer collateral CA1 synapses in acute brain slices. Similar to NARP’s role in potentiating AMPAR clustering at synapses during synapse formation, this work implicates that it may contribute to mGluR1/5dependent LTD by clustering AMPARs at sites of regulated endocytosis. The strongest support for this idea is that the presence of NARP/NP1 during the internalization phase of mGluR1/5LTD decreases surface AMPARs levels below baseline. Coupled with the fact that NARP enhances the AMPAR clustering ability of NP1, forming surface clusters that are 2.6 times larger [25], it is possible that NARP directly increases the efficiency of LTD in this paradigm by creating conditions where most surface AMPAR are aggregated with it and would be internalized once NPR is cleaved. Unfortunately, the question of whether NARP knockout alone decreases the efficiency of mGluR1/5-dependent LTD has not been addressed.  Enhanced AMPAR clustering by NARP during mGluR1/5-LTD may be facilitated by the ability of ADAM 17 to catalyze both the release of the ectodomain of pro-TNFα (tumor necrosis factor α) and its transition to TNFα via activation of protein kinase C (PKC). TNFα signalling induces  12  expression of PTX3 [18] which may be co-transcribed with NARP via a shared transcription factor NFκB (nuclear factor kappa-light-chain-enhancer of activated B cells) [50]. Therefore, NARP upregulation would increase the number of NARP-associated AMPAR clusters ready for internalization by the ectodomain of NPR.  1.2.3.3  Homeostatic plasticity  Homeostatic plasticity is the process of maintaining synaptic strength within an optimal range to ensure neural networks retain their ability to undergo long-lasting increases (LTP) and decreases (LTD) in activity. Using NARP knockout mice, Chang and colleagues [31] showed that NARP is required for homeostatic scaling of excitatory inputs onto parvalbumin-expressing inhibitory interneurons (PV-Ins) in primary hippocampal neurons and acute hippocampal slices. NARP is highly enriched at synapse onto PV-Ins and colocalizes with GluR4. This requires the presence of peri-neuronal nets associated with PV-In synapses, which may contain molecules with moieties NARP can bind in a calcium-dependent manner. Furthermore, in primary hippocampal cultures NARP secretion is regulated by network activity and increasing NARP concentration in the hippocampus by administering MECS to wild type mice generates a robust increase in NARP at excitatory synapses onto PV-In. This occurs with a concomitant increase in the surface accumulation of GluR4. The presumptive increase in inhibitory transmission in the brain under these conditions has been reported as a homeostatic mechanism against the development of epilepsy [48]. NARP may indeed play a role in chronic and long-term seizure plasticity as 3-4 month old mice lacking NARP develop severe seizures (class V), more rapidly and at a lower stimulation intensity after repeated stimulation of the basolateral amygdala than their wild type counterparts.  13  NARP also modulates circadian homeostatic plasticity at synapses of hypocretin/orexin neurons in the hypothalamus and pineal gland of living zebra fish. NARP is highly expressed in rat hypothalamic hypocretin/orexin neurons as well [51] and may regulate the rhythmic synaptic activity of these neurons that ultimately controls the sleep and wake cycle. Using time-lapse two photon imaging of these neurons in living zebra fish, Appelbaum and colleagues [52] showed that overexpression of NARP increases the strength and frequency of synaptic transmission of hypocretin/orexin neurons projecting to the pineal gland; which abolishes their rhythmicity. This is associated with increased postsynaptic clustering of AMPAR by overexpressed NARP protein. Therefore, in addition to counteracting excessive network activity in the hippocampus through modulating excitatory input onto GABAergic interneurons [31], NARP can contribute to the homeostasis of structural plasticity in neurons under circadian regulation by maintaining synaptic activity within an optimal range to support rhythmic cycles of activity.  1.2.3.4  Drug-induced behavioural plasticity  In the mammalian brain AMPA receptor trafficking at limbic system synapses underlies neuroadaptations to drugs of abuse [53]. NARP is prominently expression in projection pathways of the limbic system and has been shown to be secreted from pre-synaptic elements and act as an extracellular regulator of AMPA receptor trafficking in response to various drugs of abuse [40]. Studies using NARP knockout mice (KO) have revealed functions for NARP in drug-induced synaptic plasticity. For instance, NARP KO mice display enhanced conditioned place preference to the location of cocaine injection during withdrawal from this drug, as well as, reduced psychomotor response to infusions of cocaine directly to the nucleus accumbens [54]. These phenotypes were found to be due to reduced surface expression and postsynaptic clustering of  14  GluR1 containing AMPA receptors in nucleus accumbens which normally receives input from NARP-positive neurons [54, 55].  NARP is induced in the nucleus accumbens, amygdala and extended amygdala by both natural and imposed withdrawal from morphine, nicotine and 9-delta tetrahydrocannabinol [56-58]. Of these three, the role of NARP in mechanisms of withdrawal from morphine has studied using NARP KO mice [56, 59]. These mice have greater aversion to the environment conditioned with morphine withdrawal and also have faster extinction of this aversion. This suggests that NARP is involved in both the acquisition and extinction of aversive behavioural responses to opiate withdrawal. The mechanisms underlying acquisition of conditioned place aversion to morphine are unknown but extinction is thought to be dependent on inhibitory control of fear-based processes in the amygdala by input from the prefrontal cortex [60]. Interestingly, there are many NARP positive neurons in the prefrontal cortex [1, 61-63] and they have been shown to project to the amygdala by retrograde labeling [63]. In the amygdala itself, extinction may depend on an initial increase AMPAR signalling [64] that is then inhibited by the prefrontal cortex in a process that may rely on NARP-mediated AMPAR endocytosis [35], a form of NARP-LTD. Notably, NARP KO mice have blunted surface GluR1 expression in the amygdala and therefore it is possible that they have faster extinction of conditioned place aversion to opiates because NARP depletion precludes the expression of surface GluR1 during extinction. This has strong implications for vulnerability to drugs of abuse since alternations in the intensity and stability of aversive responses to withdrawal may be key determinants for relapse [53, 58].  15  Alcohol (ethanol) also induces NARP in the nucleus accumbens 24 hours after injection or ingestion, suggesting that it is involved in ethanol induced AMPA receptor plasticity [65]. Alcohol-experienced NARP KO mice tend to choose low alcohol concentrations over high ones as assessed by free-choice alcohol drinking and alcohol-conditioned place preference. The accumulation of GluR1 at synapses in the nucleus accumbens as a compensatory response to the inhibitory effects of alcohol was found to be concomitant with NARP induction [65-67]. Accordingly, NARP gene KO prevents both NARP and GluR1 upregulation after alcohol exposure. This deficit in synaptic accumulation of de novo GluR1 by NARP is believed underlie the lack of response of NARP KO mice to the rewarding properties of alcohol with alcohol experience [65].  Finally, in terms of reward learning, mice with localized NARP deficits in medial prefrontal cortex projections to the amygdala by local injection of dominant negative NARP were unable to use the changing value of a reward to drive future behaviour in a reinforcer devaluation task [62, 63]. In this task, the reward associated with pressing different levers is regularly changed and animals must update their representation of where they get rewards in order to be successful; NARP depleted animals fail to do so. This is analogous to the inability of drug addicts to change their behaviour, despite evidence of the difficulties that arise from engaging in that behaviour (reduction in reward) [53]. Overall, these studies each demonstrate that NARP contributes to drug-induced neuroadaptations in a way that is consistent with its role in AMPA receptor clustering and/or trafficking.  16  1.3 Neuronal activity regulated pentraxin and neurodegenerative diseases 1.3.1 Alzheimer’s disease Recently, NARP protein was identified a biomarker for familial Alzheimer’s disease in the cerebrospinal fluid (CSF) of pre-symptomatic individuals [68]. It is not understood how NARP is involved but its presence in the CSF, along with other synaptic proteins including the neuronal pentraxin receptor [68, 69] and amyloid precursor protein [68], suggest that its loss is involved in synaptic dysfunction early in the disease. On the other hand, NARP homologues NP1 and serum amyloid protein are known to be components of amyloid plaques in Alzheimer’s disease brain [3, 13, 14, 70, 71]. These pentraxins form proteolysis resistant complexes that co-cluster with amyloid fibrils, protecting them proteolytic cleavage [3, 14, 29, 71]. Although, NARP and NP1 can form hetero-multimers [26], suggesting that NARP can also associate with amyloid fibrils as well, this has yet to be demonstrated.  Hints to the possible role of NARP in Alzheimer’s disease come from work done with NP1 with which it shares 50% amino acid identity and some degree of functional redundancy [3]. NP1 is upregulated in primary cortical neurons exposed to amyloid β1-42 before neurotoxicity occurs. This suggests that amyloid β1-42 contributes to the pathology of Alzheimer’s disease by regulating NP1. Indeed, endogenous upregulation and overexpression of NP1 has been shown to cause apoptotic neuronal death in both cortical and cerebellar neurons after hypoxic ischemia [72, 73] and during periods of prolonged synaptic depression [74]. NP1 translocates to the outer mitochondrial membrane where it interacts with the pro-apoptotic transmembrane protein BAX, 17  facilitating the formation of pores that release pro-apoptotic factors from the mitochondria into the cytosol [74].  Interestingly, during hypoxic ischemia-induced apoptosis of cortical neurons NP1 is upregulated after deactivation of protein kinase B (also called Akt), which removes its tonic inhibition of GSK3β (glycogen synthase kinase 3 beta) [73]. In Alzheimer’s disease, hyper-activation of GSK3β has also been proposed to underlie synaptic loss through stimulating AMPAR endocytosis leading to prolonged synaptic depression and elimination of these quiescent synapses [75]. During this form to synaptic depression, it is possible that NP1 potentiates cell death in ways described in the preceding paragraph. Along these same line, one study showed that a chimeric molecule consisting of a portion of the N-terminal of NARP (including the signal peptide) and the pentraxin domain of NP1 resides mainly in the cytosol but can translocate to the mitochondria in a similar fashion as NP1 during periods of synaptic depression, and this process may facilitate efflux of pro-apoptotic factors [74]. That being said, it is possible that the signal sequence of NARP may target it to the intracellular compartments rather than secretion during periods of synaptic depression and this may be relevant in neurodegeneration caused by hypoxic ischemia or Alzheimer’s disease. Work is underway to further elucidate the role of NARP in Alzheimer’s disease [68].  1.3.2 Parkinson’s disease Excitotoxicity is caused by excessive neuronal membrane depolarization by AMPA receptor currents leading to the activation of calcium current through NMDA receptors [76]. This process is thought to underlie neuronal death in various neurodegenerative diseases, including Parkinson’s disease [24, 77-80]. Excitotoxicity caused by kainic acid-induced seizures in rats 18  leads to NMDA receptor hyperactivity that results in programmed cell death (apoptosis) of neurons in the hippocampal CA1 region [81]. Recently, it was shown that NMDAR excitotoxicity is mediated by the activation of the cytosolic lipid second messenger ceramide [50]. Ceramide has been shown to cause apoptosis is primary cortical neurons and neuronally differentiated PC 12 cells by activating cytokine signalling via a pathway involving TNFα (tumor necrosis factor alpha) and evolution of mitochondrial reactive oxygen species [50]. Also, excitotoxicity mediated by AMPA receptor signalling was shown to underlie selective programmed cell death of dopamine producing neurons in embryonic substantia nigra pars compacta [82] and this is mediated by ceramide activity [78].  Interestingly, through activation of cytokine signalling, ceramide has also been shown to cause the upregulation of NARP in primary cortical neurons by facilitating the translocation of the transcription factor NFκB to the nucleus [50]. This transcription factor regulates the expression of the long pentraxin PTX3, and is thought to also transcribe NARP during to ceramide signalling [16, 18, 83]. This may be one mechanism by which NARP is induced in neurons after NMDAR-mediated synaptic activity. Furthermore, NARP is selectively expressed in dopamine producing neurons of the substantia nigra and gene expression analysis of this region and cerebral cortex of Parkinson’s disease patients shows that NARP is >800% upregulated [78]. In the same study, NARP protein was identified as a novel component of Lewy bodies and Lewy neurites in Parkinsonian the substantia nigra. NARP is suspected to play a role in Lewy body growth because of its appearance on the crenulated surface of Lewy bodies as well as the presence of immunoreactive NARP bridges between alpha-synuclein aggregates. Moreover, the presence of NARP immunoreactivity in a subset of cortical dendrites of Parkinson’s disease  19  patients [78] further suggests that, in addition to contributing to Lewy formation, excess NARP protein may also be involved in excitotoxicity through clustering synaptic AMPARs.  1.4 Aims of the thesis Although NARP is known to facilitate synaptogenesis in developing neuronal networks and synaptic plasticity in mature ones, its role in activity-dependent neurodegeneration is incompletely understood. NARP plays a role in long-term depression of synaptic transmission via facilitating AMPA receptor endocytosis, particularly through its association with the extracellular surface of the GluR1 subunit. AMPAR endocytosis has already been shown to be an essential step in excitotoxic neuronal death [84]. Thus, inhibition of NARP-AMPAR interactions may be a therapeutic target for excitotoxic neurodegeneration.  High density peptide array is a powerful tool to study protein-protein interactions. We hypothesize that the high density peptide array technology is capable of identifying interaction domains between NARP and GluR1.  Aim 1: To identify peptide candidates that may block the interaction of NARP and GluR1. Aim2: To identify peptide candidates that may block excitotoxicity mediated GluR1 internalization and cell death.  20  CHAPTER 2.  IDENTIFICATION OF PUTATIVE BINDING  POCKETS FOR NEURONAL ACTIVITY REGULATED PENTRAXIN ON GLUR1 2.1 Introduction Ionotrophic glutamate receptors play a major role in excitatory neurotransmission in the central nervous system. Dysfunctions of AMPARs are associated with neurological and psychological disorders including: amyotrophic lateral sclerosis, Alzheimer’s disease, Parkinson’s Disease, Xlinked mental retardation, limbic encephalitis, Rasmussen’s encephalitis, and ischemic brain injury [77, 80, 85, 86]. They are divided into three classes based on molecular and pharmacological criteria: AMPA-, NMDA- and kainate-type glutamate receptors [46, 87, 88]. AMPA receptors (AMPARs) are the major charge carriers during fast synaptic transmission, while NMDA receptors (NMDARs) conduct calcium currents thought modulate signal transduction pathways.  AMPARs are heteromeric type 1 proteins composed of four homologous subunits GluR1-4 and are localized to dendritic spines and shafts at excitatory synapses [39, 85, 89, 90]. AMPARs play vital roles in normal neural function and neurological disease, and hence are being pursued as therapeutic targets for the treatment of major depression, amyotrophic lateral sclerosis, Alzheimer’s disease and Parkinson’s disease. The intracellular C-terminal domain contains a PDZ-domain that interacts with PDZ-domains of cytoskeletal proteins that are known to regulate AMPAR trafficking and clustering at excitatory synapses. These proteins link AMPARs to the  21  cytoplasm and include: PICK1 (protein kinase C-interacting protein), GRIP/ABP (glutamate receptor interacting protein, also called AMPAR-binding protein), SAP-97 (synapse associated protein 97) [28]. PSD-95 is known to bind to PDZ-domains of the NMDAR but also has an indirect effect on AMPAR recruitment to synapses via associations with TARPs (transmembrane AMPAR-binding proteins) [39]. Furthermore, recent reports suggest that AMPARs can be clustered at synapse by several C-terminal-interacting adhesion molecules that work via PDZdependent and PDZ-independent mechanisms [28].  The extracellular region of eukaryotic GluRs makes up more than half of the full length protein and consists of three modular domains: a ~400 amino acid N-terminal domain (NTD), a glutamate binding domain including two discontinuous poly-peptide chains called S1 and S2, and a transmembrane ion channel [91-93]. When linked together S1 and S2 fold into a clamshelllike structure with two domains that dimerize in the tetramer, forming the glutamate binding core [39, 42, 85, 92]. Glutamate binding gates current through the cation channel resulting depolarization of the postsynaptic neurons. The amount of time it takes the channel to become inactivated when glutamate is bound is referred to as the desensitization rate and is thought to play a neuroprotective role [91, 94-97]. Desensitization is governed by S2-associated splice variants of GluRs called flip and flop [98]. The region preceding the flip/flop sequence is vulnerable to RNA editing and results in AMPARs with faster recover times from desensitization [97]. Flip and flop variants of GluR1 desensitize at the same rate [96, 98] unless there is a perturbation of serine 750 in the flip/flop domain [95] . However, flop variants of GluR2-4 desensitize faster, closing the channel sooner at given concentrations of glutamate. Moreover, S1-associated regions can influence desensitization independent of the flip/flop region [97, 99].  22  The current gating properties of endogenous heteromeric AMPARs [100] are governed by the flip/flop isoforms and RNA editing of its constituent GluRs and this is believed to underlie synaptic transmission and plasticity.  The NTD of GluRs have few known physiological function and studies to identify them are ongoing [28, 39, 42, 85, 92, 101]. The NTD of GluRs is similar to bacterial amino acid binding proteins and the glutamate binding domain of metabotrophic glutamate receptors [102, 103], but does not bind glutamate or AMPA. It is absent from the prokaryotic glutamate receptor GluR0 [104] and removing it from eukaryotic GluR1, GluR2 or GluR4 does not affect the central gating function or desensitization of AMPARs [39, 91, 92]. However, the NTD of GluRs is not obsolete because it mediates subfamily specific heteromeric assembly of AMPARs [105], dendritic spine formation and presynaptic stability [32, 100, 101], and participates in trans-synaptic signalling by binding to neuronal pentraxins (NARP, NP1 and NPR) and N-cadherin [93].  Sia et al. [31] are the first to describe a role for the NTD in binding NARP and NP1 in a process where these secreted pentraxins mediate recruitment and clustering of AMPARs at excitatory synapses in primary hippocampal cultures and reconstituted synapses in neuron-glia co-cultures. NARP is dynamically regulated by synaptic activity and secreted from presynaptic terminals [1, 38]. NARP co-immunoprecipitates with GluR1-4 in rat hippocampus and co-localizes with aggregates of these subunits on the surface of cultured neurons and transfected mammalian cell lines, associations facilitated by the NTD of GluR1 and GluR4 [30, 38]. The C-terminal pentraxin domain and N-terminal coiled-coil domains of NARP cooperatively mediate interactions with the NTD of cell surface GluR1-4 in a manner independent of NARP’s general sugar-binding capabilities [25, 30]. Preventing presynaptic NARP secretion greatly diminishes 23  the number of postsynaptic GluR1 and GluR4 containing AMPAR clusters at excitatory synapses on dendritic spines and shafts [25, 26, 30, 32] and can delay the development of glutamatergic synaptic transmission [9]. Although NARP preferentially immunoprecipitates with GluR1 in mature neurons [25, 30], it has yet to be confirmed whether its interactions with GluR1’s NTD are direct. A recent 2.5 Å resolution crystal structure of the NTD of GluR1 revealed proteinprotein interactions regions that can mediate both inter-AMPAR clustering and interactions with trans-synaptic proteins [93]. NARP is indeed an excellent candidate to bind to such a region given its well documented role in AMPAR clustering. Also, the structure-function interactions between NARP and GluR1 are incompletely understood so we thought it worthwhile to investigate whether it interacts with other functional domains on GluR1, such as S1 and S2. Thus, we set out to identify putative binding sites between NARP and the entire extracellular portion of GluR1.  A high density peptide array of the extracellular surface of GluR1 including the NTD (residues 22-384) and S1 (residues 392-506) and S2 (residues 632-755) domains was incubated with recombinant NARP to identify the putative binding regions. This technology can detect domainmediated protein-protein interactions and be utilized to scan hundreds of immobilized peptides. The peptide array was generated by SPOT synthesis. The technique was originally developed by Ronald Frank in the 1990s for generating short peptide sequences to map antibody epitopes, but has since become an extremely powerful tool for various proteomic applications [106]. Peptides are synthesized and immobilized onto a cellulose membrane by sequential spotting of activated amino acids, and are kept within 6-18 amino acids long to yield optimal performance. A major advantage of this is that immobilized peptides adopt a uniform conformation on the membrane  24  that simulates their native and active state as they interact with other proteins. Using this technique, we aim to identify putative binding sites for NARP on the extracellular surface of GluR1. High density peptide array technology is a convenient, economic and powerful tool for predicting binding regions for soluble ligands on GluR1 and is likely to closely reflect physiological and pathophysiological binding of secreted NARP to AMPARs.  2.2 Methods 2.2.1 High density peptide array of the extracellular part of GluR1 The procedure for peptide array construction consists of three phases: the preparation of a cellulose membrane for peptide coupling, stepwise coupling of the amino acids and cleavage of the side chain protection groups. The membranes and peptides were prepared by the Peptide Synthesis & Purification Core Facility (Brain Research Centre, Vancouver, BC) using the previously described SPOT synthesis protocol [107]. Derivatized cellulose-based membranes were purchased from Invatis AG (Köln, Germany). Peptides spanning the entire 599 amino acid (AA) sequence of the extracellular part of GluR1 were built up from C-terminus to N-terminus. The peptide scans were performed by synthesizing 10-mer peptides spanning the GluR1 sequence with a frame shift of 2 amino acids between subsequent spots (see Figure 2.1). The C-terminal amino acid of each peptide was bound to the cellulose membrane by an amide bond. Amino acids were preactivated with pentafluorophenyl ester and delivered to corresponding positions on the membrane at 0.11µl per spot. Non-spot areas were blocked by acetylation. The Fmoc groups were then removed by washing the membrane four times with DMF for 30s, twice with 20% piperidine/DMF for 5 min, four times with DMF for 30s and twice with methanol. The spots were visualized by treatment with a methanol based solution containing 0.02% bromophenol blue. After several washes in methanol and  25  drying in the air stream of a fume hood, the membranes were ready for coupling of the next amino acid. Finally, the side chain protecting groups were cleaved by trifluoroacetic acid (TFA) treatment in the presence of 3% triisopropylsilane, 2% water and 1% phenol.  Full-length recombinant NARP protein was obtained from Abnova (Jhongli City, Taiwan). Membranes were blocked with 5% sucrose and 4% non-fat dry milk in Tris-buffered saline tween-20 (TBST) buffer for 4h and then incubated with NARP protein (5-15µg/ml) overnight in 4°C. Membranes were washed three times in TBST for 15 minutes followed by incubation in polyclonal rabbit NARP antibody (1:1000, Santa Cruz Biotechnology) overnight at 4°C. After another three 15 minute washes in TBST, membranes were incubated with goat anti-rabbit secondary antibody conjugated to horseradish peroxidase (1:1000, PerkinElmer Life Sciences) for 3 hours at room  temperature. Finally, membranes were washed three times in TBST for 15 minutes and protein interactions were visualized by enhanced chemiluminescence reaction assay (PerkinElmer Life Sciences).  2.2.2 Synthesis of GluR1-derived cell-penetrating peptides Cell-penetrating peptides derived from the extracellular fragments of GluR1 were synthesized with the addition of a truncated TAT domain at the N-terminal to permit cell penetration. Peptide synthesis was carried out in the Peptide Synthesis & Purification Core Facility (Brain Research Centre, Vancouver, BC) and was purified by HPLC. These synthetic peptides were more than 90% pure and verified by mass spectrometry. Working solutions of each synthetic peptide were prepared by dissolving them in double distilled water followed by storage in a freezer (-20°C).  26  2.2.3 Blocking assay for NARP on the membrane Membranes were blocked as above and then incubated with a mixture of NARP protein (5µg/ml) and/or a mixture of synthetic peptides (100ug/ml) overnight at 4°C. The membranes were then washed in TBST and incubated with primary and secondary antibodies as described earlier. Protein interaction was visualized by enhanced chemiluminescence reaction assay (PerkinElmer Life Sciences).  2.3 Results 2.3.1 Peptide array reveals that NARP binds the NTD and one other region on the extracellular part of GluR1 We used the GluR1 peptide array to identify binding regions between NARP and the extracellular portion of GluR1. Each experiment was repeated three times, twice on the same membrane after stripping and reprobing and once on a newly synthesized GluR1 peptide array. The membrane was incubated with NARP protein and sites of interaction were identified by ECL reaction following incubation in primary antibody against NARP and the appropriate HRPlinked secondary antibody (Figure 2.1). Positive sites show up as dark spots, while noninteracting sites remain blank (Figure 2.1). Probing the membrane with NARP revealed four NARP binding domains on GluR1 (Figure 2.2B): after accounting for non-specific interactions between the NARP antibody and the membrane (Figure 2.2A). Because each consecutive spot contains 12 amino acids shifted by 2 amino acids, the common sequence in a row of dark spots was considered a putative binding site (Figure 2.1). Peptides based on these possible binding areas in the GluR1 NTD and a region outside of the NTD were designed and synthesized.  27  The names of the synthetic peptides are Pep NT1 (9 amino acids), Pep F (9 amino acids), Pep B (8 amino acids) and Pep O (5 amino acids). The location of each of these peptides on the globular GluR1 protein was determined using software from Molecular Operating Environment protein modelling (Chemical Computing Group, www.chemcomp.com/software.htm) and is represented using a surface contour model of the crystallized GluR1 extracellular domain (Figure 2.4). The model reveals that each peptide is in distinct pockets where NARP is likely to interact. Indeed, the series of bumps and holes that comprise these kinds of pockets commonly facilitate protein-protein interactions [108, 109]. Furthermore, each synthetic peptide was more than 40% hydrophilic according to the Hopp-Woods hyphophilicity index [110] and the surface contour model confirms that they are mostly hydrophilic and thus are very good candidates to interact with secreted NARP.  Using a publically available structure analysis tool from the National Center for Biotechnology Information (http://www.ncbi.nlm.nih.gov/Structure/) we noted that each synthetic peptide has the following secondary structure: Pep F alpha helix and short loop; Pep B alpha helix; Pep NT1 alpha helix; Pep O loop. Also, Figure 2.5 shows that Pep NT1, F and B are located on the NTD of GluR1, while Pep O comprises part of a linker region between the S1 glutamate binding module and the first transmembrane domain (pre-TM1 region).  2.3.2 Synthetic peptides can block the interaction between recombinant NARP and GluR1 peptide array A mixture of the four synthetic peptides (PepNT, PepF, PepB and PepO) was tested for their ability to block the interaction of NARP with the membrane. After the membrane was incubated 28  with NARP and a mixture of synthetic peptides, the NARP antibody was unable to detect any NARP-binding regions on the GluR1 peptide array (Figure 2.3 A and B). This experiment was repeated three times (twice on the same membrane after stripping and reprobing and once a new membrane). The lack of binding after the membrane is co-treated with recombinant NARP and synthetic peptides is not due to peptide treatment per se because a scrambled version of Pep B does not block interactions between NARP and the Pep B site or any of the three remaining domains on the GluR1 peptide array (Figure 2.3 C).  2.4 Discussion We identified putative binding sites for NARP on the N-terminus of the GluR1 subunit. The Nterminal domain (NTD) of GluR1 is required for co-immunoprecipitation with NARP [25, 26, 30, 32], but until now the actual binding surfaces have never been investigated. Immobilizing the N-terminal of GluR1 on the peptide array membrane and probing with recombinant NARP was advantageous because it recapitulates the conditions in neurons where GluR1 is tethered to the postsynaptic membrane and NARP is targeted towards it by secretion from presynaptic terminals. Furthermore, the synaptic cleft is small containing a marginal amount of basement membrane and therefore secreted NARP can readily diffuse across it to facilitate trans-synaptic effects on AMPARs via interactions with extracellular portions of GluR1 [28]. We also note that the calcium-dependent lectin-like properties that NARP shares with canavalin A do not contribute to its association with GluR1 as tunicamycin treatment does not block its interaction with GluR1-4 [25, 26, 30, 99]. Thus, the binding between NARP and GluR1 discussed here is likely based on direct protein-protein interactions.  29  2.4.1 NARP-binding pockets on the NTD of GluR1 Using peptide arrays, we report that NARP directly binds the NTD (N-terminal domain) of GluR1 at three binding pockets comprising the peptide sequences of Pep NT1, F and B. This can be blocked when recombinant NARP is pre-incubated with synthetic peptides mimicking the binding regions before being exposed to GluR1 on the peptide array membrane. The presence of three binding regions suggests that NARP can recognize several binding sites on the NTD and more than one molecule may bind at time. This would be convenient because NARP monomers become closely associated when they multimerize into symmetric rings [26, 30]. In addition, NARP binding to multiple interactions sites on GluR1 could reflect an ability to bind the NTD of GluR1-4 in a manner specific to the AMPAR subfamily per se but perhaps indiscriminately between individual GluR subunits. This is supported by the observation that NARP cannot interact with any other ionotrophic glutamate receptor subfamily (i.e. kainate receptors or NMDA receptors) but can immunoprecipate GluR1-4 [26, 30]. Finally, this finding raises the possibility that NP1 may not have as many binding sites on the NTD of GluRs and this could explain the superior AMPAR clustering ability of NARP [26].  Since NARP can directly cluster AMPARs in HEK 293 cells and neurons it is not likely that other proteins are involved in the interaction [25, 26, 30]. Moreover, studies have shown that removal of NTD of GluR1 and GluR4 [26, 32], as well as perturbations of the pentraxin domain of NARP [26] prevent the interaction. In conjunction with our findings, this suggests a stable and direct interaction between NARP and GluR1 NTD which has long been implicated in the literature [25, 26, 32]. A possible structural reason for this is that NARP multimers are highly resistant to proteolysis due to ligand binding-induced conformational changes in the pentraxin 30  domain [4, 11, 19]. This likely underlies its ability to remain elevated and steadily associated with AMPARs for up to 24 hours after induction in vivo [25, 36, 65].  A crystal structure of NTD revealed that it is comprised of two lobes: a rigid L1 lobe and flexible L2 lobe [93]. These lobes arrange to form a partially closed clamshell cleft that is stabilized by four pairs of hydrogen bonds and salt bridges involving serine 188, aspartic acid 219 and glutamine 236. This clamshell arrangement is predicted to mediate protein-protein interactions with trans-synaptic proteins, such as N-cadherin and NARP and to facilitate subfamily-specific heteromeric assembly among GluR1-4 at the synapse [93, 111]. Interestingly, Pep F and NT1 are located on the rigid L1 lobe near a specificity loop predicted to limit NTD interactions to those amongst the NTD of GluRs and not NMDA- or kainate-type glutamate receptor subunits [93]. Moreover, the amino acid sequence of Pep F includes aspartic acid 219 and is flanked by the two other residues that stabilize the partially closed clamshell conformation of the NTD. This agrees with the predictions of Yao et al. [93] for the location of binding of trans-synaptic proteins on the NTD of GluR1. Furthermore, Pep B is located on the flexible L2 lobe which is predicted to modulate heteromeric interactions with the NTD of other GluRs [93], [111]. This is particularly important during formation of GluR1/2 containing AMPARs which limit the amount of calcium that is gated by the channel as the synapse matures.  In addition to promoting subfamily specific binding and heteromeric assembly of GluRs, the crystal lattice structures of both L1 and L2 lobes of the NTD of GluR1 are predicted to mediate inter-tetramer interactions leading to high density AMPAR clustering [93]. Taken together, our work is the first demonstration that NARP binds to putative functional domains on NTD of GluR1 that are predicted to regulate AMPAR tetramer formation and clustering. This provides a 31  molecular basis for the established roles for NARP in excitatory synaptogenesis and enhancement of glutamatergic synaptic transmission in mature neural networks [9, 36] and may be a potential target for therapies against neurological illnesses affecting AMPAR-mediated synaptic transmission.  2.4.2 NARP binds GluR1 at a sequence linking the S1 glutamate binding module to the first transmembrane domain. One NARP-binding region on the extracellular domain of GluR1, corresponding to Pep O, is not located on the NTD. This sequence is part of a segment linking the S1 module to the first transmembrane domain. S1 and S2 build the glutamate-binding core and are composed of the sequences adjacent to TM1 and M2-TM4, respectively (Figure 2.5). That NARP may interact with GluR1 at sites on the NTD (Pep NT1, F and B) as well as on this distal fragment (Pep O) is reasonable because they are all exposed on the surface of GluR1. Looking at the static crystal structure of GluR1 (Figure 2.4), it is apparent that Pep O is distant from and not as overtly exposed as the other three peptides. However, Pep O forms a loop in its secondary structure and it is closely associated with S1 suggesting that it has some degree of flexibility whereby it could be more exposed to NARP during ligand-induced conformational changes of the glutamate binding core. This inference is supported by the close proximity of Pep O to the S1 glutamate binding module which is flexible and regularly undergoes conformational changes related to ligand-binding. Moreover, NARP is a soluble protein and thus may freely shift between conformational states of similar energies as well. Together, this may cause fluctuations in how NARP and the extracellular surface of GluR1 interact.  32  There is a 38 amino acid flip/flop sequence cassette linking the S2 domain to the last transmembrane unit (TM4) in GluR1-4 that is known to regulate the desensitization rate by governing transitions between open and closed ion channel states [98, 112]. While GluR2 flop desensitizes faster, closing the channel faster after opening when glutamate binds, the flip/flop isoforms of GluR1 open and close at the same rate at any given glutamate concentration. However, several studies have proposed a role for the S1 subunit in controlling channel gating in GluR1 receptors. Desensitization rates between flip/flop variants of GluR1 can be altered by mutagenesis of serine 750 in the flip/flop cassette, leucine 646 in S2 or leucine 497 in the region linking S1 to the first transmembrane domain (TM1) [97, 99]. The effect of this later point mutation of the pre-M1 region at leucine 497 results in AMPARs that do not desensitize after binding glutamate [97]. Similar studies of the NR2A receptors (NMDARs) also demonstrate that the pre-TM1 region downstream from S1 is involved in desensitization [99]. Interestingly, Pep O is part of the pre-TM1 region in GluR1 and is less than 20 amino acids downstream of leucine 497. This raises the possibility that NARP-binding in this region could modulate GluR1 desensitization. A possible consequence of this, in the context of known functions of NARP, is that its attachment the pre-TM1 segment imposes a conformational restraint on GluR1 receptor channel movement that prohibits access to the desensitized state. In support of this, NARP is upregulated during activity-dependent synaptic development (between postnatal days 6 and 11 in rats) [26] and is essential for the normal progression of synaptogenesis [25, 26, 30] and silent synapse conversion [41, 113], processes that depend on coupling presynaptic with postsynaptic depolarization. The developmental induction of NARP is coincident with when AMPAR desensitize slowly due to the predominance of flip subunit isoforms (before postnatal day 8 in rats) [114]. Considered with our results, this data suggests 33  that NARP may also contribute to slower desensitization rates during the activity dependent phase of synaptic development. Furthermore, in mature brain flip variants of GluRs comprise the majority of AMPARs and NARP becomes a stoichiometric partner during episodes of neuronal activity [26, 114]. In this context, the association of NARP with the pre-TM1 region in GluR1 could mediate its role in synaptic plasticity. For example, NARP may promote in vivo LTP [1, 30] by binding to regions near the glutamate-binding region and slowing down AMPAR desensitization during periods of activity, and perhaps for some time afterward. In summary, we identified on GluR1 expected NARP binding regions on the NTD and an unexpected binding region between S1 and TM1. The positions of the NARP-binding regions of GluR1 could be explained by considering which part of NARP interacts with them. The Cterminal pentraxin domain is folded such that only hydrophilic residues are exposed on its surface [1, 11] and may therefore be better suited to interact with the NTD of GluR1. On the other hand the N-terminal of NARP is amphipathic, able to interact with both hydrophobic and hydrophilic surfaces, and may therefore facilitate interactions with the pre-TM1 region [1, 25, 30], which is more closely associated with the plasma membrane [99]. NARP binding a these sites may function to coordinate synaptic recruitment of GluR1 receptors and modify channel gating kinetics in these newly recruited receptors. We believe these sites represent specific NARP binding pockets based on examination of the solved crystal structure of the extracellular domain of GluR1 [93]. In addition, despite their small size (5-9 amino acids), these peptides could block the interaction of NARP with GluR1 at corresponding sites on a peptide array. This suggests that they are “hot spots” for the interaction, meaning that they contain residues crucial for protein-protein interaction [108]. Site-directed mutagenesis of GluR1 NTD and pre-TM1 region may further elucidate which residues are required for interaction. Also, a demonstration 34  that each individual peptide blocks interaction with the surface of GluR1 on the peptide array or in vitro will be informative in future experiments. Clearly, further study of the consequences of NARP-GluR1 interactions on synaptic plasticity and pathophysiology is warranted.  35  Figure 2.1. Schematic of high density peptide array synthesis. The peptides are built up from C-terminus to N-terminus on a commercially available derivatized cellulose membrane (Intavis AG; Köln, Germany). An amide bond links the C-terminal amino acid to the membrane. Each peptide consists of 10 amino acids with a frame shift of 2 amino acids between subsequent spots. Interaction between the peptide chains and a recombinant bait protein is visualized using ECL sensitive immuno-labeling against the bait protein.  36  Figure 2.2. Identification of peptides that may block the interaction between NARP and GluR1 N-terminus. NARP protein binds to specific regions on the peptide array of the N-terminus of GluR1. A. GluR1 N-terminus membrane probed with NARP antibody only; B. Two separate experiments showing GluR1 N-terminus membrane probed with NARP then visualized with NARP antibody; n=3 total repeats. Four peptides named NT1, F, B and O were synthesized based on the binding regions denoted by the red rectangles.  37  Figure 2.3. A mixture of all four peptides blocks the interaction of NARP with GluR1 Nterminus on the membrane. The GluR1 N-terminus high density peptide array membrane was incubated with a mixture of all four peptides (NT, F, B and O; 100µg/ml) and NARP protein (5µg/ml). A and B are repeats showing a membrane incubated with NARP only (1) and the same membrane stripped and reprobed with a mixture of NARP and all four peptides (2); n=3. C shows that a scrambled version of Pep B containing the same amount and type of amino acids in random order does not block NARP-GluR1 interactions. Red rectangles denote interaction domains.  38  Figure 2.4. Location of NARP interacting peptides on the folded extracellular portion of GluR1 A surface contour model of the extracellular 528 amino acid portion of GluR1 shows the location of NARP interaction sites identified by peptide array. Pep F, NT1 and B are all located on the same face in binding pockets. Pep O is in a binding region perpendicular to the location of the other three peptides.  39  Figure 2.5. Schematic showing location of synthetic peptides relative to functional domains of GluR1 A. A schematic of the topology of GluR1 showing the N-terminal domain (NTD), S1 and S2 moieties of the glutamate (Glu) binding core, transmembrane domains (TM), re-entrant loop between S1 and S2 (M3) and intracellular C-terminal (C). Pep F, NT1 and B are on the NTD whereas Pep O is part of a linker sequence between S1 and the last transmembrane domain. B. A schematic of full length GluR1. Pep O binds to a region linking S1 with the first transmembrane domain A. The flip/flop domain and relative locations of Pep NT1, F and B on the N-terminal domain are also shown. Numbers associated with certain domains denote their sequence location as noted in [39, 96, 99].  40  CHAPTER 3.  EFFECTS OF NARP-BINDING PEPTIDES ON  NMDA-INDUCED AMPA RECEPTOR INTERNALIZATION AND CELL DEATH 3.1 Introduction Prolonged activation of α-amino-3-hydroxy-5-methylisoxale-4-propionic acid-type glutamate receptors (AMPARs) and N-methyl-D-aspartate-type glutamate receptors (NMDARs) is the basis of excitotoxic neuronal death and underlies the loss of neurons and neuronal functions in many neurodegenerative diseases and acute brain insults [76, 77, 84, 115, 116]. While AMPARs mediate fast excitatory neurotransmission, activation of NMDARs elevates intracellular calcium levels and this has been implicated as the primary cause of neuronal apoptosis, a form of programmed cell death [76]. However, activation of NMDARs also triggers long term depression (LTD) of glutamatergic synaptic transmission by faciliatating clathrin-mediated endocytosis of α-amino-3-hydroxy-5-methylisoxale-4-propionic acid-type glutamate receptors (AMPARs), termed NMDAR-LTD [43, 117-119]. NMDAR-LTD has been causally linked to apoptosis through a signalling pathway involving caspase-3 and 9, both of which degrade their substrates during the execution phase of apoptosis [120] . Li and colleagues [120] provide the first evidence that short-term activation of NMDARs facilitates LTD via moderate activation of caspases 3 and 9 such that apoptosis is prevented, whereas prolonged NMDAR stimulation increase caspase activity to lethal levels shifting the balance from synaptic depression toward apoptosis.  41  Treating primary hippocampal neurons with the NMDAR agonists NMDA and glycine for 1 hour incudes apoptosis 24 hours afterward which is hallmarked by activation of cysteine proteases (caspases), cellular shrinkage, DNA degradation and margination of the nuclear membrane into apoptotic bodies [120, 121]. Using this NMDA treatment paradigm, Wang and colleagues [84] demonstrated that blocking stimulated clathrin-mediated AMPAR endocytosis with the GluR2 C terminal-derived peptide, GluR23Y, prevents NMDA-induced apoptosis. GluR23Y specifically inhibits AMPAR endocytosis by preventing the phosphorylation of tyrosine 879 on the GluR2 subunit, which is required for this to occur [122] during the early proteinsynthesis independent phase of NMDAR-LTD. Their results showed that NMDA-induced clathrin-mediated endocytosis of AMPARs occurred after intracellular calcium levels rose but before activation caspase 3 activation and inhibition of protein kinase B (Akt). These molecules are key regulators of neuronal apoptosis because inhibition of Akt deregulates its substrates which are pro-apoptotic proteins when active and they work to prompt the release of cytochrome c from the mitochondria leading to caspase activation [115, 123, 124]. Therefore, in the case of prolonged activation of NMDARs, calcium influx may trigger clathrin-mediated AMPAR endocytosis which then contributes to a mechanism that initiates downstream apoptotic (cell death) cascades [84].  However, AMPAR endocytosis occurs very soon after NMDAR-LTD is triggered and not much is known about the identity of proteins whose synthesis is required for LTD to be sustained [125]. Since secreted NARP associates with synaptic AMPARs and is regulated as an immediate early gene by NMDAR-mediated synaptic activity, we postulated that it may contribute to sustained AMPAR endocytosis and possibly excitotoxic neuronal death. There is evidence that  42  NARP increases the efficiency of mGluR1/5-dependent LTD in primary neurons from hippocampus, cortex and cerebellum by clustering AMPARs at sites of regulated endocytosis [35]. Furthermore, NARP may facilitate a form of LTD imposed on neurons in the central amygdala by NARP-expressing terminals from the prefrontal cortex [59]. However, the role of NARP in NMDAR-LTD remains elusive. Notwithstanding the need for further clarification of NARP in LTD, it is involved in excitotoxicity-based neurodegenerative diseases including Parkinson’s disease [78] and Alzheimer’s disease [68] which may have a requisite NMDARLTD component [43]. Here, we investigated the effect of prolonged NMDAR activation on NARP expression and used synthetic peptides that block the interaction between NARP and the GluR1 AMPAR subunit to explore the effect this has on AMPAR endocytosis and neuronal death.  3.2 Methods 3.2.1 Rat primary cortical neuron culture All animals used in this research were handled in accordance with the Institutional Animal Care and Use Committee (IACUC) guidelines. Primary cortical neuron cultures were prepared from Wistar rats (UBC Animal Care Centre, Vancouver, BC, Canada) at embryonic day 18-19. After pregnant rats were anesthetized and sacrificed, the embryos were transferred to cold dissection buffer containing Hanks Balanced solution (Gibco-BRL, Grand Island, NY) and 10mM HEPES (Sigma, Saint Louis, MO), pH 7.4 and osmolarity 310-320 mOsm. The entire brain was removed and the meninges were carefully pealed from each cortical lobe. The cortices were separated and put into 0.25% trypsin (Gibco-BRL, Grand Island, NY) for 10-20 minutes at 37°C. To remove trypsin, cortical tissues were washed twice in DMEM (Gibco-BRL, Grand Island, NY) 43  containing 10% Fetal Bovine Serum (Gibco-BRL, Grand Island, NY) and 1% penicillin/streptomycin (Gibco-BRL, Grand Island, NY). The Cortical tissues were homogenized using a 10ml pipette and centrifuged at 14000 x g for 3 minutes. The supernatant was removed and the pellet suspended in plating medium containing Neurobasal (Gibco-BRL, Grand Island, NY), 2% B-27 supplement with AO (Gibco-BRL, Grand Island, NY), 2mM L-glutamine (Sigma, Saint Louis, MO), 25μM glutamic acid (Sigma, Saint Louis, MO), 10mM β-mercaptoethanol (Gibco-BRL, Grand Island, NY) and 1% penicillin/streptomycin (Sigma, Saint Louis, MO). The cortical cells were then plated and cultured onto poly D-lysine coated tissue culture plates at 7.5 x 105 cells per well for 6 well plates (for biochemical assays) and 8 X 104 cell per well in 24 well plates (for neuroprotection assays). Cortical cultures were maintained in a humidified incubator with 5% CO2 at 37°C for 2 to 3 days. Then the entire medium was replaced with fresh medium containing Neurobasal, 2% B-27 with antioxidant supplement, 2mM L-glutamine and 1% penicillin/streptomycin. Every 3 or 4 days half of the medium volume was replaced with fresh maintenance medium.  3.2.2 Dc protein assay The Bio-Rad Dc protein assay protocol was used to determine protein concentrations. Serial dilutions of bovine serum albumin were used to generate a linear standard curve of known protein concentrations including 0.3125mg/ml, 0.625mg/ml, 1.25mg/ml, 2.5mg/ml and 5mg/ml. Following the manufacturer’s instructions, cortical cell lysates were incubated with Dc reagents in a 96-well plate. The protein concentration was determined by measuring the optical density of each sample at 560nm using the “uQuant” microplate spectrophotometer (Bio-Tek Instruments, USA).  44  3.2.3 SDS-PAGE and Western blot Cultured neurons or streptavidin beads (see section 2.2.3) were collected in 2x loading buffer containing 10% glycerol, 50mM Tris-HCl, 2% SDS, 5% β-mercaptoethanol and 0.01mg/ml bromophenol blue. The samples were boiled for 5 minutes and the polypeptides separated by SDS-Polyacrylamide Gel Electrophoresis using the Bio-Rad Electrophoresis system (Bio-Rad, Hercules, CA). For each sample, 50ug of protein was loaded in 10% SDS-polyacrylamide resolving gels and 5% stacking gels. The PageRulerTM Plus Prestained Protein ladder (Fermentas, CA) was used to identify the molecular weight of proteins in each sample. Proteins were transferred from SDS-PAGE gels to nitrocellulose membranes using the Bio-Rad Wet Transfer system (Bio-Rad, Hercules, CA) running at 100 volts for 2 hours in a cold room (4°C). The nitrocellulose membrane was then blocked in 5% non-fat milk in TBST for 1 hour at room temperature. After a quick wash in TBST the membrane was incubated in primary antibody dissolved in 3% bovine serum albumin-TBST (Sigma, Saint Louis, MO) overnight at 4°C. The primary antibodies used included: rabbit polyclonal NP2 (NARP) antibody (1:1000, Santa Cruz biotechnology, Santa Cruz, CA), mouse monoclonal NP2 antibody (1:500, Santa Cruz biotechnology, Santa Cruz, CA), mouse monoclonal NR1 antibody (1:500, Millipore, Billerica, MA), rabbit polyclonal GluR1 (1:500, Calbiochem, La Jolla, CA), mouse monoclonal GluR1 Nterminus antibody (1:1000, Millipore, Temecula, CA) and rabbit polyclonal Actin antibody (1:1000, Cell Signaling, Danvers, MA). Membranes were washed three times in TBST for 15 minutes and incubated with HRP-conjugated goat anti-rabbit or goat anti-mouse secondary antibody (1:1000, PerkinElmer Life Sciences) for 1 hour at room temperature. After another three 15 minute washes in TBST proteins were visualized using the enhanced chemiluminescence reaction assay (ECL, PerkinElmer Life Sciences). Calculation and 45  normalization of protein intensity against actin (or other internal controls) was done using Quantity One software (Bio-Rad, Hercules, CA).  3.3 Cytotoxicity treatment with N-methyl-D-aspartic acid (NMDA) Treating cultured neurons with NMDA and Glycine for 1 hour at 37°C is known to produce substantial cell death 24 hours later [121]. Here, we dissolved NMDA and Glycine (Sigma, Saint Louis, MO) in double distilled water to 50mM and 10mM, respectively. Then the NMDA and Glycine solutions were added to the cortical culture medium at 1:1000 to reach a respective final concentration of 50µM and 10µM. NMDA/Glycine was left in the medium for 1 hour at 37°C and then removed by replacing the total medium with fresh maintenance medium (see section 2.2). Control cultures (without NMDA/Glycine added) underwent the same number of media changes.  3.3.1 Testing the of effect cytotoxic NMDA treatment on NARP expression To examine the effect of cytotoxic NMDA treatment on NARP expression we treated cultured neurons as described above and collected them for Western immunoblot analysis 4, 8, 12 and 24 hours later. Quantity One software (Bio-Rad, Hercules, CA) was used to calculate and normalize NARP protein intensity against actin.  46  3.3.2 Peptide treatment in cultured neurons We examined the effect of synthetic peptides on NMDA-induced cell death in cultured primary cortical neurons. Cells were pretreated with 0.5µM, 5µM and 10µM of synthetic peptides for 30 minutes followed by NMDA/Glycine treatment (50µM NMDA plus 10µM Glycine) for 1 hour at 37°C. The entire medium was then replaced with fresh medium and synthetic peptides were added again. The cultures were kept in a 37°C humidified chamber for an additional 24 hours before being assessed for viability.  3.3.3 Cell surface biotinylation We examined the effect of synthetic peptides on NMDA-induced AMPA receptor internalization in primary cortical neuron cultures. Cortical neurons cultures were prepared in 6 well plates and used after 13 days in vitro (DIV13). The experimental groups included: NMDA treatment only; Peptide treatment only; NMDA and peptide treatment; No treatment. All groups were treated under the same conditions and underwent the same number of medium changes. Cells were pretreated with 10µM of synthetic peptides for 30 minutes and then treated with NMDA/Glycine for 1 hour at 37°C. The medium was replaced with fresh medium containing 10µM of synthetic peptides. After 4 hours, the cell surface protein isolated by biotinylation using previously describe methods [126]. Cells were washed once with ice-cold phosphate buffered saline (PBS) and incubated for 30 minutes at 4°C in 500µl of biotin reagent (0.5mg/ml EZ-Link Sulfo-NHSLC-Biotin in cold PBS); Pierce, Rockford, IL). Cells were washed once with ice-cold PBS and then washed twice with 1ml of ice-cold 50mM Tri-PBS (pH 7.4) was for 2 minutes in order to quench the biotin reaction. Cell were lysed and solubilized with 500µl of RIPA buffer containing 150mM NaCl, 50mM Tris-base, 1mM EDTA, 1% Triton X-100 and protease inhibitor cocktail, 47  pH 7.4 (Pierce, Rockford, IL). Cells were then triturated using a 281/2 gauge needle and 50µl of lysate was taken as the input control. The input controls were mixed with 2x loading buffer and then boiled for 5 minutes before storage at -80°C. To isolate biotinylated proteins, streptavidinagarose beads (Sigma, Saint Louis, MO) were washed three times with RIPA buffer and added to the remaining lysates (50µl per sample) for 16 hours at 4°C with rotation. Unspecific binding was  removed by washing the beads 5 times with RIPA buffer. Excess RIPA buffer was removed with 281/2 gauge needle. Biotinylated proteins were eluted by adding 50µl of 2x loading buffer to beads, followed by boiling for 4 minutes at 103°C and centrifugation at 17000 x g for 1 minute.  Biotinylated proteins were resolved by 10% SDS-PAGE and transferred to nitrocellulose membranes (Bio-Rad, Hercules, CA). The same membrane was sequentially probed, stripped, and re-probed for GluR1, NR1 and Gsk3β. NR1 is a subunit of another ionotrophic glutamate receptor (NMDA receptor) and serves as specificity control relative to GluR1. Gsk3β is an intracellular protein and serves as a negative control for surface biotinylation. The stripping buffer used here contained 2% SDS, 62.5mM Tris-HCl and 100mM β-mercaptoethanol. Biotinylated cell surface proteins were visualized using ECL and quantified relative to the amount of proteins detected in the total lysate. For example, the proportion of GluR1 at the cell surface was assessed by measuring the band densities of biotinylated GluR1 and total GluR1 (combined surface and intracellular) in the input control.  3.3.4 Assessment of cell death: Lactate dehydrogenase assay Cell death in cortical cultures 24 hours after excitotoxic NMDA treatment was assessed using the Lactic Dehydrogenase (LDH) based In Vitro Toxicology Assay Kit (Sigma, Saint Louis, MO). LDH is a cytosolic enzyme that is released in the medium when the cell membrane is compromised. The medium was incubated with reagents for 30 minutes in a 96-well plate according to the  48  manufacturer’s instructions and LDH was detected based on the fact that it catalyzes the oxidation of lactate to pyruvate with concomitant reduction of NAD+ to NADH and H+. This causes the conversion of a tetrazolium dye to a soluble, coloured formazan derivative. Since the amount of colour change produced by the LDH reaction is proportional to the number of lysed cells we quantified this by measuring the wavelength absorbance at 490nm using the “uQuant” microplate spectrophotometer (Bio-Tek Instruments, USA). The wavelength absorbance at 750nm was used as a reference filter. We determined the percentage of cell death using the formula % cytotoxicity = NMDA-LDH release (OD490)/maximum LDH release (OD490) after correcting for baseline LDH  release. Maximum LDH release was determined in each independent experiment by deliberately lysing neurons in one well with 0.3% Triton X-100 for 10 minutes before treating the medium with LDH reagents and measuring the wavelength absorbance.  3.4 Statistical Analyses The results are expressed as the mean + standard error of the mean (SEM). The n value refers to independent cortical neuron cultures. The statistical significance of the differences was analysed using independent t tests for comparison between two groups or one-way ANOVA to compare multiple groups; a value of p < 0.05 or p < 0.001 is statistically significant.  49  3.5 Results 3.5.1 Excitotoxic NMDA treatment induces NARP protein expression in primary cortical neurons The NARP antibody detected a 55kDa band which is the predicted molecular weight of glycosylated NARP [1]. Quantification of this band by Western blot from NMDA-treated primary cortical neuron lysates revealed a three-fold increase in NARP expression 8 hours after treatment relative to untreated controls (Figure 3.1A; n=3, p<0.05). The elevation in NARP 4 hours after NMDA treatment approached significance relative to controls (n=3; p = 0.07, t-test) and was not significantly different from NARP levels at 8 hours (n=3; p = 0.7, t-test) or 12 hours (n=3; p = 0.1, t-test). NARP levels drop back to baseline levels 24 hours after NMDA treatment and this is significantly less than its levels at both 4 and 8 hours post-NMDA (Figure 3.1A; n=3, p<0.05). The amount of GluR1 protein detected at each time point in NMDA treated cultures varied but was never significantly different from non-treated control cultures by t-test statistics (Figure 3.1B). The induction and sustained expression of NARP 8 hours after excitotoxic NMDA treatment suggests that it may be involved in a mechanism contributing to NMDAinduced AMPAR internalization.  3.5.2 The peptides block NMDA-induced AMPA receptor internalization NMDA treatment decreases surface GluR1 expression by four-fold (~233%) compared to untreated controls (n=5, p<0.05, one-way ANOVA). Incubation with synthetic peptides prevents the NMDA-induced surface GluR1 reduction such that the amount of surface GluR1 is not different from that in non-treated control cultures (Figure 3.2B). Importantly, NMDA treatment 50  does not significantly alter total GluR1 expression levels (Figure 3.1B). This suggests that NMDA treatment specifically alters the ratio of cell surface to internal GluR1 without effecting overall GluR1 expression. We are confident the biotinylation procedure specifically targeted surface proteins because GSK 3β, an intracellular protein, can only be detected in the input sample in which biotinylated proteins were not isolated by incubation with streptavidin beads (Figure 3.2A; for detailed protocol see section 3.3.3). In addition, the effect on ionotrophic glutamate receptors is specific to AMPARs since the surface expression of NR1, a subunit of NMDARs, is not changed as a result of NMDA or synthetic peptide treatment (Figure3.2B). Because AMPARs in mature neurons are hetero-tetrameric, consisting of GluR1, GluR2 and GluR3 subunits [127, 128], internalization of a single subunit is coincident with removal of the entire AMPAR. Taken together our data suggests that the GluR1 N-terminal derived peptides specifically block NMDA-induced AMPAR internalization.  3.5.3 The peptides rescued neurons from NMDA-induced death After 14 days in vitro, primary cortical neurons were treated with 0.5µM, 5µM and 10µM of synthetic peptides 30 minutes prior to and immediately after excitotoxic NMDA treatment. Another subset of neurons was treated with NMDA alone or a scrambled synthetic peptide, Scram B. Scram B is a scrambled version of Pep B containing the same amount and type of amino acids in random order; this peptide served as a negative control to rule out the possibility that treatment with synthetic peptides has any effect on NMDA-induced cell death per se. After NMDA treatment the medium was replaced with fresh medium containing synthetic peptides and cell viability was assessed 24 hours later using the LDH assay.  51  We compared the effect of individual GluR1 N-terminal derived peptides and a mixture of all four peptides (Pepmix) on NMDA-induced cell death (Figure 3.3A). The LDH cell death assay shows that 32% of the neurons in NMDA treated cultures die 24 hours afterwards, compared to only 13% cell death in untreated controls (n=3, p<0.001, one-way ANOVA). Pepmix and Pep F, Pep B and Pep O all reduced NMDA-induced cell death (n=3, p < 0.001, one-way ANOVA) and there was no difference in cell death between any of these groups. Compared to cultures treated with NMDA only, percent cell death was reduced in cultures also treated with Pepmix in a dosedependent manner with only 19% death at 0.5µM, 12% death at 5µM and 9% death at 10µM dose (n=3, p < 0.05, t-test). Pep F and Pep O also reduce cell death after NMDA treatment at 0.5µM (Pep F, 19%; Pep O, 16%, p < 0.001 one-way ANOVA) and further reduced cell death at 5µM and 10µM (Pep F, 8% and 13% respectively; Pep O, 9% for both doses). There was no significant difference in cell death percentage between the 5µM and 10µM for either peptide. Pep B reduced the percentage of NMDA-induced cell death equally well at all doses (10%, 6% and 11%; n=3, p<0.001). Cell death in NMDA treated neurons and neurons treated with both NMDA and either PepNT1 or Scram B was both elevated compared to untreated controls. Specifically, there was 29% and 25% cell death in PepNT1 and Scram B treated cultures, respectively, compared to 13% cell death in controls (n=3, p<0.001 one-way ANOVA). The percentage of cell death in Pep NT1 and Scram B treated cultures was comparable with that in NMDA treated cultures where the death rate was 32%. Together, these results suggest that PepNT1 does not contribute to the protective effect of Pep mix and that reduction in NMDAinduced cell death is not caused by synthetic peptide treatment per se. Furthermore, treating neuronal cultures with 10µM Pep NT1 or 0.5µM and 10µM Pep F, B or O without the addition of NMDA did not increase cell death with respect to untreated cultures (Figure 3.3B). However,  52  a 0.5µM dose of Pep NT1 did increase cell death with respect to controls (n=3, p<0.001 one-way ANOVA), further confirming the deleterious effects of this peptide on neuronal viability. Therefore, with the exception of a low dose of Pep NT1, the GluR1 N-terminal derived peptides are non-toxic and mostly protective against NMDA-induced cell death.  Next, we sought to narrow down which peptides most effectively prevented NMDA-induced cell death in primary cortical neuron cultures (Figure 3.4A). Since Pep NT1 did not reduce NMDAinduced cell death when used alone and may in fact be toxic to neurons (Figure 3.3), we removed it from the synthetic peptide mixture (Pepmix) and tested the effect of different combinations of the remaining three peptides on NMDA-induced cell death. Pep FBO, FB and BO all reduced cell death at 0.5µM, 5µM and 10µM. Pep FO reduced NMDA-induced cell death at 5µM and 10µM (n=3, p<0.05) but not at 0.5µM. Between 0.5µM and 5µM Pep FB provides a dosedependent decrease in percentage cell death (14% at 0.5µM, 7% at 5µM; n=3 p< 0.05, t-test), which is not significantly different from that at the 10µM dose for this peptide (6%). The reduction in cell death percentage provided by each dose of Pep FB is not significantly different from that provided by corresponding doses of Pep mix. Pep BO also reduces cell death percentage at each dose (11%, 5% and 12%; n=3, p<0.001) and the 0.5µM dose of Pep BO is more potent than the same dose of Pepmix or Pep FB (p<0.05, t-test). Neurons treated with Pep FO are not protected from NMDA-induced cell death at the 0.5µM dose (20% death) but are protected when the dose is increased to 5µM and 10µM (only 7% cell death at both doses; n=3, p<0.001). Pep FBO reduces cell death percentage at all doses (5%, 7% and 18%; n=3, p<0.001). Since only 5% of neurons die 24 hours after NMDA treatment in neuronal cultures treated with 0.5µM Pep FBO, this makes it the most potent peptide against NMDA-induced cell death  53  (Figure 3.4B). Notably at the 10µM dose, Pep FBO treated neuronal cultures have significantly more cell death than the same dose of Pep mix, Pep FB or Pep FO (Figure 3.4A and B; n=3, p<0.001).  3.6 Discussion Our data showed for the first time that activation of NMDARs with an excitotoxic dose of NMDA and glycine can induce NARP expression in primary cortical neuron cultures. This increase in NARP protein is associated with both internalization of surface GluR1 and neuronal cell death 4 and 24 hours after NMDA treatment, respectively. NARP-binding peptides derived from the extracellular surface of GluR1 (TAT-Pep NT1, TAT-Pep F, TAT-Pep B and TAT-Pep O) prevents both NMDA-induced GluR1 internalization and cell death. We previously demonstrated using peptide array that a 10µM mixture of these peptides block the interaction between full-length recombinant NARP and the extracellular surface of GluR1 at three binding pockets on the NTD of GluR1 and a short sequence on the pre-TM1 region. The specific inhibitory effect of these synthetic peptides on GluR1 internalization suggests that they have a similar function in vitro. From these data, it is hypothesized that the interaction between NARP and the extracellular surface of GluR1 contributes to NMDA-induced AMPAR internalization and cell death.  3.6.1 Time course for NMDA-induced NARP expression in vitro Consistent with NARP being an immediate early gene product specifically induced by NMDARmediated synaptic activity [1], we found that excitotoxic NMDA treatment up-regulated this protein in primary cortical neurons. The time course of in vitro NARP expression after excitotoxic NMDAR activation studied here matches that in vivo in that the level of NARP 54  protein tends to dramatically rise 4 hours after NMDAR activation and is definitively maximal 8 hours after treatment [1, 30, 36]. However, unlike its expression in rat cortex, hippocampal CA1 and dentate gyrus following high frequency stimulation of the perforant pathway [25, 26, 30] or shock-induced seizure [1, 36], the rise in NARP level at 4 hours was not statistically significant; although we did detect a consistent trend for NARP to increase at this time point by Western blot. This may be attributed to subtle variations in NARP level between biological repeats or may be a particular feature of the time course of NARP induction in primary neuron cultures that distinguished it from that in vivo.  That NARP induction in vitro may have a different time course than in vivo is supported by the fact that in addition to NARP becoming maximal 8 hours after NMDA treatment, the level of NARP drops back to baseline at 24 hours rather than remaining elevated as in previous reports [1, 25, 26, 30, 36, 40, 52]. The stability of NARP protein on the cell surface is promoted by addition of N-linked glycans [1, 129] and formation of covalent disulphide bonds during multimerization [26, 30]. The conformation conferred by these two factors may protect NARP multimers from proteolytic cleavage by tucking residues susceptible to cleavage deep within the hydrophobic core of its pentraxin domain [11]. Thus, given the similarities we noted between in vitro and in vivo NARP induction, it seems paradoxical that in vitro NARP induction defers in this parameter. This can be reconciled by the fact that excitotoxic NMDA treatment causes neuronal cell death 24 hours later [43, 84, 121] likely explaining the lack in NARP expression at this time. Cell death assays confirmed that on average the excitotoxic NMDA treatment caused 32% neuronal cell death 24 hours afterward in the primary cortical neurons used here. A reduction in NARP due to poor cell viability could be compounded by the fact that NARP  55  mRNA is only present during the first 4 hours post-induction [25, 30], limiting de novo NARP protein synthesis.  Overall, we find that excitotoxic NMDA treatment tends to induce NARP 4 hours afterward and is maximal by 8 hours before falling back to baseline between 12 and 24 hours. The lack of sustained NARP levels for up to 24 hours distinguishes NARP induction in primary neuron cultures from that in brain and may be related to higher susceptibility of cultured neurons to NMDA-induced cell death.  3.6.2 Up-regulation of NARP is linked to NMDA-induced GluR1 internalization We found that excitotoxic NMDA treatment increased NARP expression and reduced GluR1 surface expression after 4 hours in primary cortical neurons without changing overall expression levels of GluR1. This was associated with significant neuronal cell death 24 hours later. This is consistent with evidence that NARP is upregulated during excitotoxic NMDAR stimulation through a mechanism that activates the cytosolic pro-apoptotic lipid ceramide [50, 81]. Active Ceramide induces NARP expression in primary cortical neurons by activating TNFα resulting in the translocation of the transcription factor NFκB to the nucleus, where it is known to transcribe long pentraxins, including NARP [18, 50, 83]. Pep mix inhibited NMDA-induced AMPAR internalization and protected against NMDA-induced cell death in a dose-dependent manner. Three of the peptides in Pep mix (Pep NT1, F and B) correspond to NARP-binding pockets on the NTD of GluR1 and one (Pep O) is from a region linking the S1 glutamate-binding module to the first transmembrane domain (pre-TM1 region) in GluR1. We previously showed that Pep mix can block NARP-GluR1 binding on a peptide array membrane. Therefore, each synthetic peptide 56  likely interferes with NARP-GluR1 interactions required for NMDA-induced AMPAR internalization and cell death. We will discuss our data in the context of the possible functional outcomes of NARP-GluR1 interactions in vitro.  3.6.3 Mechanism of inhibition of NARP-GluR1 interactions Each of the synthetic peptides was cell permeable due to the attachment of an N-terminal TAT sequence. Our data suggests that the peptides prevent NARP-GluR1 interactions by mimicking extracellular binding sites on GluR1 that can bind NARP inside the cell. This may block NARP secretion or abolish its ability to interact with GluR1 once it is secreted. Indeed, the C-terminal pentraxin domain of NARP is essential for axonal transport and secretion [25] so peptide binding here could inhibit this function. In this way, peptide-bound NARP would behave like dominant NARP that can still multimerize with endogenous NARP, but cannot be secreted due to broad deletions in the pentraxin domain [25, 63]. Dominant negative NARP prevents aggregation of postsynaptic AMPARs at nascent synapses [25] and alters synaptic plasticity related to behavioural flexibility when injected into NARP-rich neurons in the medial prefrontal cortex of mature rodents [63]. Thus, the synthetic peptides may block NARP-AMPAR interactions in a similar manner after prolonged NMDAR stimulation and prevent AMPAR internalization and cell death.  3.6.4 Optimal neuroprotection requires blockade of NARP-binding on the NTD and glutamate-binding domain of GluR1 GluR1 contains three structurally and functionally distinct modules: the N-terminal domain (NTD), glutamate-binding domain (S1 and S2 modules) and the transmembrane ion channel  57  [93]. A crystal structure of NTD revealed that it is comprised of two lobes: a rigid L1 lobe and flexible L2 lobe [93]. These lobes organize into a partially closed clamshell conformation. The NTD is of GluR1 is in the synaptic cleft and is known to engage in protein-protein interactions with N-cadherin and all neuronal pentraxins, including NARP [93]. The functions of the NTD are not completely understood but it seems to be an important anchor for trans-synaptic signalling molecules and for mediating subfamily-specific heteromeric assembly among GluR14 at the synapse [111]. This is particularly important during formation of GluR1/2 containing AMPARs which limit the amount of calcium that is gated by the channel as the synapse matures. On the other hand, the glutamate-binding domain mediates AMPAR activation/deactivation and ion channel gating [99]. Since the synthetic peptides could bind both the NTD and a region associated with the glutamate-binding module, we attempted to identify which peptide or combination of peptides best protected primary cortical neurons against NMDA-induced cell death.  While Pep F, B and O each reduced NMDA-induced cytotoxicity at all doses, Pep NT1 had no effect at any dose and a low dose was cytotoxic even without NMDA treatment. Therefore, to test the possibility that Pep mix is more potent without Pep NT1, we removed it from the mixture leaving Pep F, B, and O (Pep FBO). This combination did prove to be the most potent protectant against NMDA-induced cell death when compared to the same dose (0.5µM) of Pep FB, BO, FO or Pepmix. Pep FBO was also more protective against cell death than Pep mix at the 5µM dose, confirming the antagonistic role of Pep NT1 in Pepmix. This could be due to the fact that Pep NT1 mimics a region on the NTD of GluR1 that is relatively richer in glutamine and charged residues compared to the other three peptides. It is known that proteins containing glutamine  58  repeats can precipitate protein aggregation in the cytoplasm and/or nucleus causing cytotoxicity and neuronal cell death [130, 131]. Therefore, Pep NT1 may have many off-target associations limiting its ability to specifically block NARP-AMPAR interactions, as demonstrated by its potent toxicity in NMDA-naive neurons. This result may in fact indicate that this site on GluR1 can bind to more proteins than just NARP. As a result, Pep NT1 may block NTD interactions with another trans-synaptic protein leading to cytotoxicity and lack of protection from NMDAinduced neuronal cell death. One such protein may be N-cadherin, whose interactions with the NTD of GluR2 (which is often associated with GluR1) are necessary and sufficient to promote the formation and growth of dendritic spines in cultured hippocampal neurons [132].  The varying levels of neuroprotection provided by each peptide and peptide combination may be related to the interactions they inhibit between NARP and GluR1. Pep F and B mimic two NARP-binding regions on the NTD of GluR1 and are both neuroprotective. The protective effect is greater at each dose when these peptides are used together in Pep FB, suggesting increased blockade of NARP interactions with the NTD of GluR1. Because Pep mix prevents cell death apparently by inhibiting AMPAR internalization, it is likely that Pep FB is neuroprotective for the same reason. Since the NTD of GluR1 resides in the synaptic cleft and can engage in proteinprotein interactions with NARP that mediates clustering of these subunits [25, 26, 30, 32, 93], our data suggest that this process is essential to AMPAR internalization resulting in cell death.  Interestingly, Pep F is located on the L1 lobe (see Chapter 2 Discussion) in a region near the specificity loop that dictates subfamily-specific interactions amongst GluRs whereas Pep B resides on the flexible L2 lobe that may mediate heteromeric interactions with GluR2 [111]. This  59  suggests that by clustering GluR1 subunits at high density, NARP may facilitate AMPAR assembly via NTDs of neighboring subunits as well as clustering of fully formed AMPARs. This idea is supported by the fact that dominant negative NARP or NARP gene knockout decreases the number of postsynaptic AMPAR [25, 113].  Lastly, Pep O mimics a region linking the S1 glutamate-binding module to the first transmembrane domain in GluR1. This segment is referred to as the pre-TM1 region and is an important regulatory site for GluR1-containing AMPAR channel gating function [99]. Pep O is neuroprotective on its own but is more protective in combination with Pep FB in Pep FBO. Although Pep FBO is neuroprotective at all doses, there is an apparent dose-dependent decrease in the efficiency of this protection that contrasts with the trend for dose-dependent increase in neuroprotection by other peptide combinations. This could reflect non-specific binding by Pep FBO at concentrations over 0.5µM which would limit its ability to bind NARP or changes in the fine-tuning of NARP-mediate changes to AMPAR function and clustering via the Pep O site and Pep F/B site, respectively. Notably, Pep FO is not protective at 0.5µM but is protective at higher doses, suggesting that this peptide has a shorter half-life relative to the others and requires a higher dose to be effective. Also, this is the only peptide mixture that does not include Pep B, perhaps indicating that Pep B enhances the stability of the synthetic peptide mixtures used here.  That Pep FBO is the most potent protectant against NMDA excitotoxicity suggests that NARP binding to both the NTD and pre-TM1 region plays a role in NMDA-induced GluR1 internalization and cell death. Specifically, clustering of AMPAR by NARP-binding at Pep F and B regions on the NTD can cluster AMPAR at the synapse while interactions at the Pep O  60  sequence could mediate changes in AMPAR function, such as increased sensitization to glutamate binding, that drive AMPAR internalization. A decrease in AMPAR desensitization has previously been linked to AMPAR internalization and LTD-like phenomenon, and there is some evidence that synaptic activity-inducible NARP influences AMPAR function [38, 57].  Pacchioni et al. [54, 133] demonstrated that NARP/NP1 knockout mice have aberrant AMPAR function which causes a decrease in cocaine-withdrawal induced locomotion. This behaviour is normally potentiated by elevated glutamate release in the nucleus accumbens, which is able to induce NMDAR-LTD via activation of extrasynaptic NR2B-containing NMDARs [43, 134]. Paradoxically, these mutants also have enhanced sensitization to cocaine administration [54]. Since these animals certainly have less postsynaptic surface AMPARs, the conflicting AMPARmediated phenotypes implicates dissociation between the ability of NARP/NP1 to cluster AMPARs at synapses and regulate current flow through these receptors. Therefore, increased sensitization to cocaine and decreased locomotion after cocaine withdrawal in NARP/NP1 knockout mice could mean that NARP/NP1 is needed to decrease ion flow through AMPARs in the nucleus accumbens that are endogenously highly sensitized to glutamate during instatement of cocaine-seeking behaviour when synaptic AMPARs are being recruited. This may serve to prevent over-depolarization of the postsynaptic neuron and confer neuroprotection. However, during cocaine withdrawal NARP/NP1 may potentiate cocaine-seeking because of the increase in synaptic AMPARs [133]. Our finding that NARP can bind a region in close proximity to the glutamate-binding core (Pep O on pre-TM1 region) supports a role for this protein in AMPAR channel gating. Furthermore, that we found binding sites for NARP on the NTD of GluR1,  61  which is quite distant from the pre-TM1 region, agrees with the observation that the ability NARP to influence AMPAR function is dissociated from its ability to cluster these receptors.  Where NARP binds the pre-TM1 segment is less than 60 amino acids upstream of glutamine 582 in the pore forming region. Overexpressing GluR1 with a single point mutation at glutamine 582 in mice decreases current through heteromeric AMPARs and these animals display the same abnormalities in cocaine-induced behaviour as NARP/NP1 mutants (increased sensitization to cocaine and blunted locomotor response to withdrawal) [134]. Recall also that this NARPbinding region is less than 20 amino acids downstream of leucine 497 (see Chapter 2 Discussion) in GluR1, which when mutated results in AMPARs that do not desensitize to glutamate binding [97]. Moreover, a role for NARP in AMPAR sensitization is suggested by Li et al. [135] who showed that NARP is upregulated in the inferior colliculus after a single episode of audiogenic seizure. GluR1 and GluR2 expression levels remained the same [135] suggesting that NARP may sensitize GluR1/2-conatining AMPARs already present at the synapse.  Hence, we suspect that NARP binding in the pre-TM1 region of GluR1 could bi-directionally modulate AMPAR channel gating function. For instance, after the initial transient AMPAR internalization event when NMDAR-LTD is triggered by NMDA plus glycine treatment delayed NARP expression (4 hours later) may cluster reinserted GluR1-containing AMPARs at the synapse and possibly increase current through these receptors. In conjunction with elevated intracellular calcium levels due to prolonged NMDAR stimulation [111], this would provide the stimulus to evoke enhanced glutamate release by further activating NMDARs which may trigger an enduring form of AMPAR endocytosis.  62  Together these data show that all the peptides, with the exception of Pep NT1 and low dose of Pep FO, are protective against NMDA- induced cell death. The differences in neuroprotection between peptides and doses of the same peptide suggest that they interact with NARP over a wide range of concentrations and have different in vitro stabilities. This could change depending on the conformation of the peptide and its affinity for NARP in the context of prolonged NMDAR stimulation. For example, changes in pH and overactive intracellular calcium pathways may influence the affinity of each peptide for NARP. Overall, our findings suggest that NARP induced by excitotoxic NMDA promotes AMPAR internalization and cell death through interactions with Pep, F, B and O sites on the extracellular surface of GluR1.  3.6.5 NARP-GluR1 interactions facilitate NMDAR-LTD and cell death Excitotoxic NMDAR stimulation causes apoptotic cell death in neurons [84, 121]. Apoptosis is carried out by cysteine protease (caspases) that degrade their substrates and NMDAR stimulation activates caspase 9 and 3 by stimulating cytochrome c release from mitochondria. This is caused by calcium influx that activates protein phosphatase 1 (PP1) and protein phosphatase 2B (PP2B or calcineurin) that dephosphorylate the pro-apoptotic protein BAD leading to its activation [111]. BAD translocates to the mitochondrial outer membrane where it facilitates cytochrome c efflux though pores formed by other pro-apoptotic membrane proteins. This process also requires rapid clathrin-mediated endocytosis of AMPARs at stimulated synapses, a phenomenon referred to as NMDAR-LTD [43, 117]. Notably, NMDAR-LTD also relies on the activity of PP1 and PP2B/calcineurin. In particular, PP1 dephosphorylates serine 845 at the C-terminal of GluR1 leading to internalization of GluR1-containing AMPAR and LTD [43]. In addition, elevated intracellular calcium due to NMDAR activation also activate protein tyrosine kinases resulting in 63  phosphorylation of tyrosine 879 on the C terminal of GluR2 leading to clathrin-mediated endocytosis of GluR2-containing AMPARs [122, 136]. Recent work by Li et al [111, 120] shows that mild NMDAR stimulation (30µM NMDA for 5 minutes) results in LTD only but strong NMDAR stimulation (100µM NMDA or repeated 30µM NMDA) shifts the balance toward apoptosis.  Our results suggest that GluR1 internalization hours after NMDA treatment is necessary to induce cell death 24 hours later. We showed that this occurs without changes in surface NMDAR, which is the characteristic feature of NMDAR-LTD [43]. These findings are consistent with those of Wang et al [84] who demonstrated that NMDA-induced clathrinmediated endocytosis of AMPARs is an essential step for NMDA-induced apoptosis in primary hippocampal neurons. They showed that blocking this process with a peptide derived from the tyrosine rich portion of the C-terminal of GluR2 prevent caspase 3 activation and neuronal cell death. Therefore, this study blocked the early phase of LTD that involves rapid AMPAR endocytosis after calcium-mediated activation of protein tyrosine kinases. On the other hand, our study seems to block the late, protein synthesis-dependent phase of LTD that is required for LTD maintenance. This is because NARP is not highly expressed until 4 hours after NMDA treatment and this time coincides with when GluR1 surface expression is dramatically reduced. Together, these data solidify that NMDAR-LTD per se (either early or late phases) is essential to NMDAinduced excitotoxicity and cell death.  How then may NARP contribute to NMDAR-LTD? The established roles of NARP in excitatory synaptogenesis and synaptic plasticity may provide some insight. NARP is known to cluster  64  synaptic GluR1 containing AMPARs at both immature and mature synapses by interacting with the NTD of GluR1, facilitating synaptogenesis or synaptic long-term potentiation (LTP) [25, 26, 32]. The role of NARP expression in LTD is less clear, clustering of AMPARs by NARP enhances the efficiency of LTD mediated by type I and V metabotrophic glutamate receptor (mGluR1/5) activation in primary hippocampal, cortical and cerebellar Purkinje neurons [35] and the nucleus accumbens in vivo [54, 133]. Cho et al. [35] showed that this type of LTD is dependent on neuronal pentraxin expression and cleavage of the neuronal pentraxin receptor (NPR) from its membrane tether by the matrix metalloprotease ADAM 17. The soluble NPR fragment is biologically active as an ectodomain and induces AMPAR internalization to endosomes by binding to surface NARP/NP1-AMPAR clusters, the largest of which always include with NARP multimers [26]. Accordingly, in our study, NMDA-induced NARP upregulation likely increases the number NARP-associated AMPAR clusters as it does in assays of NARP overexpression in hippocampal and spinal neuron cultures, GluR1 clustering in HEK 293T cells and seizure induction in vivo [1, 25, 26, 30, 135]. Taken together, this raises the possibility that by interacting with the NTD of GluR1 at the binding pockets we identified in this study, NARP could increase the efficacy of NMDAR-LTD by clustering AMPARs at sites of regulated endocytosis. Our data support this idea since peptide mixtures including those that mimic NARP-binding sites on the NTD of GluR1 (Pep F and B) are neuroprotective and can prevent NMDAR-LTD of GluR1.  Interestingly, at synapses that have experienced LTD, phosphorylation of the scaffold protein PICK1 (protein interacting with C kinase) by PKC (protein kinase C) may allow dissociation of ABP (AMPA receptor-binding protein) from AMPARs outside of the synapse and facilitate  65  synaptic recruitment of these receptors [88]. This process enables synaptic re-potentiation. Notably, PICK1 phosphorylation by PKC is also involved in mGluR-dependent LTD, which seems to also require AMPAR clustering by NARP to be efficient [35, 43]. Therefore in NMDAR-LTD, PICK1 activation by PKC leading to liberation of AMPARs from ABP intracellularly may facilitate their synaptic recruitment by NARP extracellularly. Once at the synapse, additional LTD may occur due to sustained elevation of intracellular calcium after prolonged exposure to NMDA, thereby enabling the internalization of both synaptic and formerly extrasynaptic AMPARs. Internalization of synaptic and extrasynaptic AMPAR could be a mechanism of NMDAR-LTD maintenance. This may explain the almost total of loss of surface AMPARs we observed when NMDA treated neurons were not co-treated with NARP blocking synthetic peptides.  Along these same lines, NMDAR-LTD may be involved in permanent removal of synaptic AMPARs resulting in elimination of these “silent” (inactive) synapses during development. NARP has been shown to regulate this process in nascent excitatory neuronal networks in vivo [113, 137] and may do so through a mechanism that activates glycogen synthase kinase (GSK3β) during periods of synaptic depression in primary cortical neurons [74]. GSK3β is also activated by induction of NMDAR-LTD when PP1 directly activates it or deactivates its inhibitor, Akt. Upon activation, GSK3β causes AMPAR endocytosis through an unknown mechanism [138] and has been proposed to underlie synaptic silencing and synapse elimination in developing neural networks [43]. The coordinate function of NARP and GSK3β during synaptic pruning has not been studied, but based on the known functions of these proteins a possible mechanism may  66  involve NARP setting up AMPAR surface clusters for coordinated endocytosis that is then carried out via a GSK3β-dependent pathway.  Interestingly, in Alzheimer’s disease brains excessive NMDAR-LTD by GSK3β activity can be stimulated by neurotoxic amyloid β1-42 [75] and homologues of NARP namely, NP1 and serum amyloid protein, can be induced by this amyloid fragment and are known become components of amyloid plaques [13, 14, 70]. The proteolysis stable pentraxin polymers protect amyloid fibrils from proteolytic degradation contributing to disease progression, possibly by facilitating GSK3β signalling and aberrant synapse elimination [29, 70]. While NARP has not been identified in amyloid plaques, it is closely associated with NP1 and highly present in the cerebrospinal fluid of patients with familial Alzheimer’s disease and is therefore thought to be a biomarker of synaptic loss or dysfunction early in this disease [68].  The requirement for GluR1 internalization in NMDA-induced cell death in our study and others [84] suggests that this process not only decreases AMPAR-mediated synaptic transmission, but also that internalized AMPARs may play a role in intracellular cell death signalling pathways. NARP has been shown to be endocytosed with its pool of AMPARs during mGluR-LTD [35] and our results show that NARP is upregulated during a time of enhanced NMDA-induced GluR1 internalization; this may also occur during NMDAR-LTD. The intracellular function of NARP is not well studied but one report indicates that it promotes neuronal migration and neurite outgrowth in cortical explant cultures within the concentration range of other known growth factors (nM) [1]. Also, induction of NARP after prolonged NMDAR activation has been associated with cytokine signalling in neuronally differentiated PC-12 cells via the activation of  67  ceramide [50, 81]. That being said, Akt can be regulated by both growth factors and cytokines [139], providing a possible link between internalized NARP-GluR1 clusters and reduced Akt activity resulting in caspase 3 activation and cell death. Furthermore, a less detailed study suggest that the signal sequence of NARP may target it to the mitochondrial outer membrane during periods of synaptic depression where it facilitates release of the pro-apoptotic molecule cytochrome c [74]. Therefore, AMPARs internalized during NMDAR-LTD could become associated with Akt deactivation-dependent cell death signalling cascades via their association with NARP.  An intracellular role for NARP in neurodegenerative disease is further suggested by the observation that it is in close association with α-synuclein in Lewy bodies and Lewy neurites in Parkinsonian substantia nigra pars compacta [78]. Programmed cell death of dopaminergic neurons was shown to occur through an excitotoxic AMPAR and NMDAR-dependent mechanism [82]. Incidentally, NARP is regulated by synaptic activity in dopaminergic neurons where it bi-directionally regulates AMPAR-mediated synaptic plasticity [23, 40, 57, 58, 62, 65, 140]. Whole genome expression profiling identified NARP as the most highly up-regulated gene cohort the substantia nigra pars compacta and cerebral cortex in Parkinsonian brain, and this process is concomitant with GluR3 upregulation and deregulation of NSF (N-ethylalimide sensitive fusion protein) [78]. Interestingly, NSF binds the C-terminal of GluR2 and disruption of this interaction has been shown to cause AMPAR internalization, mimicking NMDAR-LTD [43, 88]. These data suggests that NARP is a structural component of these cytoplasmic protein aggregates and further implicates it in synaptic dysfunction resulting death of dopamine producing neurons in Parkinson’s disease patients. Notably, NARP expression may be dependent  68  of NMDAR-induced activation of ceramide which is normally suppressed by the action of Akt [50, 81], a kinase deactivated during NMDAR-LTD [84, 120]. Thus, induction of NMDAR-LTD may potentiate the activation of pro-apoptotic ceramide signalling and maintenance of NMDARLTD by NARP could sustain this effect and therefore contribute to excitotoxic cell death. Indeed, the abnormal presence of NARP inside cortical dendrites of Parkinsonian brain [78] may hint at prolonged AMPAR internalization events that cause lethal synaptic dysfunction.  In conclusion, our results suggest that NARP plays a role in synaptic changes involved in excitotoxic neuronal death in primary cortical neurons and may therefore be a novel target for the development of neuroprotective therapeutics in neurodegenerative disease.  69  Figure 3.1. NARP protein is up-regulated after excitotoxic NMDA treatment. DIV 10-13 days primary cortical neurons were treated with 50µM NMDA and 10µM glycine for 1 hour. The medium was changed to remove NMDA and neurons were collected for Western blot analysis at the indicated time points. A. NARP expression increase 4-8 hours after NMDA treatment. The mean + SEM is plotted; n = 3 independent cultures; *p < 0.05 compared to Control and # p < 0.05 compared to the 24 hour time point, t-test. B. Total GluR1 levels fluctuate but are not significantly altered as a result of NMDA treatment. C. Example of a Western blot from which quantified data are derived.  70  Figure 3.2. Peptides block NMDA- induced AMPA receptor internalization. Primary cortical neurons were treated with a 10µM mixture of all four synthetic peptides 30 minutes prior to and immediately following 1 hour treatment with NMDA. A. Surface proteins were biotinylated 4 hours after NMDA treatment and identified by Western blot. B. Western blot quantification showing specific effect on AMPA receptor internalization. *p<0.05 compared to control group; **p<0.05 compared to NMDA group. The mean + SEM is plotted; n=5 biological repeats. 71  Figure 3.3. Synthetic GluR1 N-terminal derived peptides are neuroprotective. Neurons were treated with NMDA alone in combination with 0.5µM, 5µM or 10µM peptides. Cell death was assessed 24 hours after NMDA treatment. A. Each GluR1 N-terminal derived peptide, except for PepNT1 and control peptide ScramB, reduce the amount of cell death 24 hours after NMDA treatment. *p < 0.001 with respect to controls; **p <0.001with respect to NMDA and NMDA-ScramB group, one-way ANOVA. # p < 0.05 with respect to 0.5µM dose of that peptide group, t-test. B. The peptides are not toxic, except for 0.5µM PepNT1. * p < 0.001 with respect to control, one-way ANOVA. The means + SEMs are plotted, n=3 repeats.  72  Figure 3.4. Identification of the most potent peptides. Different combinations of PepF, B and O at 0.5µM, 5µM and 10µM were added to NMDA treated neurons and cell death was assessed 24 hours later. A. Comparison of all combinations of PepF, B and O. *p<0.001 compared to controls; **p<0.001compared to NMDA and NMDAScramB; ^ p<0.001 compared to the same dose of Pepmix, one-way ANOVA. # p < 0.05 compared to 0.5µM dose of that peptide group; ##p<0.05 compared to 10µM dose of that peptide group, t-test. B. Pep FBO is the most potent against NMDA-induced cell death. *p<0.001 compared to NMDA group; **p<0.001 compared to the NMDA and Pepmix group, one-way ANOVA. The means + SEMs are plotted, n=3 biological repeats. 73  CHAPTER 4.  CONCLUSION & PERSPECTIVE  Using surface receptor biotinylation and LDH cell death assays, we showed that NARP-binding synthetic peptides derived from extracellular surface of GluR1 prevent NMDAR-induced AMPAR internalization and cell death. We hypothesize that the peptides accomplish this by blocking NARP-GluR1 interactions on the surface primary cortical neurons. This may prevent NARP-mediated AMPAR clustering which reduces the efficiency of NMDAR-LTD. The requirement for NMDAR-LTD for NMDA-induced cell death suggests that endocytosed NARPassociated AMPARs may contribute to an intracellular signalling cascade that promotes neuronal cell death. In addition, because NARP is an immediate early gene that is upregulated between 4 and 8 hours after NMDA treatment it may be considered as a novel protein required for maintained expression of NMDAR-LTD.  NARP has been shown to directly interact with the extracellular surface of GluR1 in vivo in stimulated rat hippocampus [1, 26, 30]. We did not observe direct interaction between NARP and GluR1 by co-immunoprecipitation from lysates of NMDA stimulated primary cortical neurons, but overcame this limitation by showing that the NARP binds to specific regions on GluR1 on a peptide array. However, parts of GluR1 could have denatured or unfolded on the high density peptide array leading to non-specific binding by recombinant NARP. To mitigate this possibility we repeated each peptide array experiment at least three times, accounting for background staining and discounting interaction sites that had hydrophilicity scores below 40%. In the end, all four peptides we selected were located in distinct binding pockets on the surface of GluR1, ideal to facilitate protein-protein interaction. In fact, we identified three NARP-binding pockets on the NTD of GluR1, a region previously implicated in mediating interactions with 74  NARP [25, 26, 30]. We then used surface receptor biotinylation to show that synthetic peptide treatment prevents NMDA-induce GluR1 internalization in primary cortical neurons. Therefore, in lieu of co-immunoprecipitation, the combined use of high density peptide array, protein modelling and surface biotinylation suggest that NARP and GluR1 directly interact and that this interaction contributes to NMDA-induced GluR1 internalization.  However, surface biotinylation assesses total surface GluR1, not only those that are specifically synaptic. In future, the peptides can be used in hippocampal slices to assess their effect on low frequency induced (LFS) NMDAR-LTD in the CA1 region, which shares a common mechanism with chemical NMDAR-LTD elicited in the present study [117]. In this paradigm, surface GluR1 expression in synaptic and extrasynaptic fractions of synaptoneurosomes could elucidate whether the peptides specifically prevent synaptic GluR1 internalization. Whether peptide treatment blocks the interaction of NARP with GluR1 can also be assessed using established detergent conditions for co-immunoprecipitation of NARP and GluR1 from rodent brain [30, 54]. Also, although it has been shown that NARP knockout [31] or NARP/NP1 double knockout [113] does not alter NMDAR-LTD in the CA1 region of the hippocampus by LFS, these studies did not rule out possible functional compensation by the residual pentraxin(s) during synaptogenesis or synaptic signalling when the brains of these animals mature. NARP, NP1 and NPR are highly homologous in their AMPAR binding regions and can all facilitate surface AMPAR clustering to some extent [3]. Therefore, an investigation of NMDAR-LTD in the CA1 region of mice with conditional NARP gene knockout rather than chronic NARP gene knockout is warranted.  75  An interesting finding in this study is that NARP can bind to a region closely associated with the glutamate-binding core of GluR1. Point mutations in this region or a peptide designed to chronically bind to it may clarify the putative role of NARP in modulating AMPAR channel gating properties. Similar manipulations of the three putative NARP-binding regions on the NTD of GluR1 may suggest novel roles for NARP in facilitating homo- and heteromeric subunit assembly. This also may have implications in synapse maturation and remodelling as the NTD of GluR1 partially regulates synapse maturation via protein-protein interactions with the NTD of GluR2 which facilitate incorporation of this calcium impermeable subunit into AMPAR tetramers at mature synapses [93].  Notwithstanding these interesting findings, we note that our data does not address the issue of whether the synthetic peptides directly interact with endogenous NARP or interfere with its secretion. There are possibly off-target effects or peptide binding to NARP could inhibit a previously unknown intracellular function of NARP and either or both of these may indirectly prevent GluR1 internalization. With regard to the later possibility, there is a large consensus in the literature that the function of NARP is completely extracellular NARP [3, 28]. The only known exceptions to this is the apparent ability of NARP to facilitate formation of α-synuclein based Lewy bodies in the substantia nigra pars compacta [78], and the ability of the first 80 amino acids of the N-terminal of NARP to direct a chimeric molecule to the cytosol and mitochondria during periods of prolonged synaptic depression [74]. Furthermore, the peptides are short sequences so if NARP binds them specifically then other ligand binding domains (such as the lectin-binding domain) on NARP may still be available to interact with any putative partners. This idea is based on the historical nature of pentraxins to bind multi-ligands  76  simultaneously leading to formation of multifunctional protein complexes [1, 4, 11, 16, 141]. For example, the acute phase short pentraxins CRP and SAP can activate the classic complement cascade by forming a complex in which it directly binds its ligands as well as the globular domain of C1q, the first component of the classic complement pathway [142]. If NARP can behave in a similar manner, then any effects of peptide treatment would be specifically due to inhibition of NARP-GluR1 interactions rather a general blockade of its ligand-binding abilities. A peptide array of full length NARP probed with recombinant GluR1 may provide further details about the domains on NARP that bind GluR1. In future, we aim to repeat our experiment with the same peptides not conjugated to TAT in an attempt to only target secreted NARP in the synaptic cleft.  However, since we also plan on using these peptides in vivo, they must be able to be absorbed into neurons safely. We have already demonstrated here that the peptides are not neurotoxic up to 10µM, with the exception of a low dose of Pep NT1. In fact, in preliminary experiments we found that the peptides were non-toxic to primary cortical neurons up to 100µM (data not shown). For further in vitro characterization of the specificity of our peptides for NARP we will capitalize on the fact that NARP is secreted into the medium of transfected HEK 293, CHO and COS-1 mammalian cell lines [25, 26, 30]. Thus, we may verify whether NARP is still secreted when these cell lines are treated with cell-penetrating synthetic peptides. In addition, whether the synthetic peptides prevent NARP-GluR1 interactions can be tested by co-culturing GluR1 expressing HEK 293 cells and NARP overexpressing primary neurons. This method has been used before to show that NARP co-clusters with AMPARs at sites of contact between HEK cells  77  and axon terminals [30, 32], and may therefore be used in conjunction with synthetic peptide treatment to test whether it prevents the association between secreted NARP and AMPARs.  Overall, our data show that NARP is upregulated by excitotoxic NMDAR stimulation and is associated with the late, protein synthesis-dependent phase of NMDAR-LTD. Since NARP secretion is dynamically regulated by synaptic activity and it can bind, cluster and sustain associations with AMPARs for long periods of time, NARP may well be involved in mechanisms regulating enduring synaptic depression. This has already been suggested in at least two published works [35, 57]. Furthermore, NARP may provide a link between AMPAR internalization and activation of intracellular cell death pathways. Given its involvement in excitotoxic neurodegenerative disease, particularly Parkinson’s disease and Alzheimer’s disease, peptides that can block the association of NARP with its intracellular or extracellular partners may become valuable therapeutics. Clearly, more work must be done to clarify the still elusive role of NARP in excitotoxic neurodegeneration.  78  Literature Cited 1.  Tsui, C.C., et al., Narp, a novel member of the pentraxin family, promotes neurite outgrowth and is dynamically regulated by neuronal activity. J Neurosci, 1996. 16(8): p. 2463-78.  2.  Hsu, Y.C. and M.S. Perin, Human neuronal pentraxin II (NPTX2): conservation, genomic structure, and chromosomal localization. Genomics, 1995. 28(2): p. 220-7.  3.  Osera, C., et al., Pentraxins and Alzheimer's disease: at the interface between biomarkers and pharmacological targets. Ageing Res Rev, 2012. 11(2): p. 189-98.  4.  Gewurz, H., X.H. Zhang, and T.F. Lint, Structure and function of the pentraxins. Curr Opin Immunol, 1995. 7(1): p. 54-64.  5.  Martinez de la Torre, Y., et al., Evolution of the Pentraxin Family: The New Entry PTX4. The Journal of Immunology, 2010. 184(9): p. 5055-5064.  6.  A. K. Schlimgen, J.A.H. and H.V.a.M.S. Perin, Neuronal Pentraxin, a Secreted Protein with Homology to Acute Phase Proteins of the Immune System. Neuron, 1995. 14: p. 519-526.  7.  Adam R. Goodman, T.C., Ruben Abagyan, Anne Altmeyer, and H.-G.W.a.J. Vilcek, Long Pentraxins: an Emerging Group of Proteins with Diverse Functions. Cytokine & Growth Factor Reviews, 1996. 7(2): p. 191-202.  8.  A P Osmand, B.F., H Gewurz, R H Painter, T Hofmann, and E Shelton, <Characterization of C-reactive protein and the complement subcomponent C1t as homologous proteins displaying cyclic pentameric symmetry (pentraxins)>. Proc Natl Acad, 1977. 74(2): p. 739-743.  79  9.  Bjartmar, L., et al., Neuronal pentraxins mediate synaptic refinement in the developing visual system. J Neurosci, 2006. 26(23): p. 6269-81.  10.  Pepys, M.B., et al., Targeting C-reactive protein for the treatment of cardiovascular disease. Nature, 2006. 440(7088): p. 1217-21.  11.  Mikolajek, H., et al., Structural basis of ligand specificity in the human pentraxins, Creactive protein and serum amyloid P component. J Mol Recognit, 2011. 24(2): p. 371-7.  12.  Emsley, J., et al., Structure of pentameric human serum amyloid P component. Nature, 1994. 367(6461): p. 338-45.  13.  Coria, F., et al., Isolation and characterization of amyloid P component from Alzheimer's disease and other types of cerebral amyloidosis. Lab Invest, 1988. 58(4): p. 454-8.  14.  Tennent, G.A., L.B. Lovat, and M.B. Pepys, Serum amyloid P component prevents proteolysis of the amyloid fibrils of Alzheimer disease and systemic amyloidosis. Proc Natl Acad Sci U S A, 1995. 92(10): p. 4299-303.  15.  M. B. PEPYS, A.C.D., THELMA C. FLETCHER, N. RICHARDSON, E. A. MUNN A. FEINSTEIN, Analogues in other mammals and in fish of human plasma proteins, Creactive protein and amyloid P component. Nature, 1978. 273: p. 168 - 170.  16.  Goodman, A.R., et al., Long pentraxins: an emerging group of proteins with diverse functions. Cytokine Growth Factor Rev, 1996. 7(2): p. 191-202.  17.  D’Nette C. Dodds, I.A.O., Susan J. Cushman, Jill A. Helms, and Mark S. Perin, <Neuronal Pentraxin Receptor, a Novel Putative Integral Membrane Pentraxin That Interacts with Neuronal Pentraxin 1 and 2 and Taipoxin-associated Calcium-binding Protein 49>. THE JOURNAL OF BIOLOGICAL CHEMISTRY, 1997. 272(34): p. 21388-21494.  80  18.  Polentarutti, N., et al., Inducible expression of the long pentraxin PTX3 in the central nervous system. J Neuroimmunol, 2000. 106(1-2): p. 87-94.  19.  Kirkpatrick, L.L., Biochemical Interactions of the Neuronal Pentraxins. NEURONAL PENTRAXIN (NP) RECEPTOR BINDS TO TAIPOXIN AND TAIPOXIN-ASSOCIATED CALCIUM-BINDING PROTEIN 49 VIA NP1 AND NP2. Journal of Biological Chemistry, 2000. 275(23): p. 17786-17792.  20.  Schlimgen, A.K., et al., Neuronal pentraxin, a secreted protein with homology to acute phase proteins of the immune system. Neuron, 1995. 14(3): p. 519-26.  21.  Zhang, L., et al., The neuronal pentraxin II gene (NPTX2) inhibit proliferation and invasion of pancreatic cancer cells in vitro. Mol Biol Rep, 2011. 38(8): p. 4903-11.  22.  Schvartz, D., et al., Modulation of Neuronal Pentraxin 1 expression in rat pancreatic beta-cells submitted to chronic glucotoxic stress. Mol Cell Proteomics, 2012.  23.  Reti, I.M. and J.M. Baraban, Opiate Withdrawal Induces Narp in the Extended Amygdala. Neuropsychopharmacology, 2003. 28(9): p. 1606-1613.  24.  Berke, J.D., et al., A complex program of striatal gene expression induced by dopaminergic stimulation. J Neurosci, 1998. 18(14): p. 5301-10.  25.  O'Brien, R., et al., Synaptically targeted narp plays an essential role in the aggregation of AMPA receptors at excitatory synapses in cultured spinal neurons. J Neurosci, 2002. 22(11): p. 4487-98.  26.  Xu, D., et al., Narp and NP1 Form Heterocomplexes that Function in Developmental and Activity-Dependent Synaptic Plasticity. Neuron, 2003. 39(3): p. 513-528.  27.  Hohenester, E., et al., Crystal structure of a decameric complex of human serum amyloid P component with bound dAMP. J Mol Biol, 1997. 269(4): p. 570-8.  81  28.  Gerrow, K. and A. El-Husseini, Receptors Look Outward: Revealing Signals That Bring Excitation to Synapses. Science's STKE, 2007. 2007(408): p. pe56-pe56.  29.  Kirkpatrick, L.L., et al., Biochemical interactions of the neuronal pentraxins. Neuronal pentraxin (NP) receptor binds to taipoxin and taipoxin-associated calcium-binding protein 49 via NP1 and NP2. J Biol Chem, 2000. 275(23): p. 17786-92.  30.  O'Brien, R.J., et al., Synaptic clustering of AMPA receptors by the extracellular immediate-early gene product Narp. Neuron, 1999. 23(2): p. 309-23.  31.  Chang, M.C., et al., Narp regulates homeostatic scaling of excitatory synapses on parvalbumin-expressing interneurons. Nature Neuroscience, 2010. 13(9): p. 1090-1097.  32.  Sia, G.M., et al., Interaction of the N-terminal domain of the AMPA receptor GluR4 subunit with the neuronal pentraxin NP1 mediates GluR4 synaptic recruitment. Neuron, 2007. 55(1): p. 87-102.  33.  Daw, M.I., et al., PDZ proteins interacting with C-terminal GluR2/3 are involved in a PKC-dependent regulation of AMPA receptors at hippocampal synapses. Neuron, 2000. 28(3): p. 873-86.  34.  Dodds, D., et al., Novel reticular calcium binding protein is purified on taipoxin columns. J Neurochem, 1995. 64(5): p. 2339-44.  35.  Cho, R.W., et al., mGluR1/5-Dependent Long-Term Depression Requires the Regulated Ectodomain Cleavage of Neuronal Pentraxin NPR by TACE. Neuron, 2008. 57(6): p. 858-871.  36.  Reti, I.M. and J.M. Baraban, Sustained increase in Narp protein expression following repeated electroconvulsive seizure. Neuropsychopharmacology, 2000. 23(4): p. 439-43.  37.  <Obrien Glur1 antibody.pdf>.  82  38.  Chang, M.C., et al., Narp regulates homeostatic scaling of excitatory synapses on parvalbumin-expressing interneurons. Nat Neurosci, 2010. 13(9): p. 1090-7.  39.  Tomita, S., et al., Stargazin interacts functionally with the AMPA receptor glutamatebinding module. Neuropharmacology, 2007. 52(1): p. 87-91.  40.  Reti, I.M., et al., Prominent Narp expression in projection pathways and terminal fields. J Neurochem, 2002. 82(4): p. 935-44.  41.  Koch, S.M. and E.M. Ullian, Neuronal Pentraxins Mediate Silent Synapse Conversion in the Developing Visual System. Journal of Neuroscience, 2010. 30(15): p. 5404-5414.  42.  Malinow, R. and R.C. Malenka, AMPA receptor trafficking and synaptic plasticity. Annu Rev Neurosci, 2002. 25: p. 103-26.  43.  Collingridge, G.L., et al., Long-term depression in the CNS. Nat Rev Neurosci, 2010. 11(7): p. 459-73.  44.  Appleby, V.J., et al., LTP in hippocampal neurons is associated with a CaMKII-mediated increase in GluA1 surface expression. J Neurochem, 2011. 116(4): p. 530-43.  45.  Bortolotto, Z.A. and G.L. Collingridge, Characterisation of LTP induced by the activation of glutamate metabotropic receptors in area CA1 of the hippocampus. Neuropharmacology, 1993. 32(1): p. 1-9.  46.  Collingridge, G.L., et al., The synaptic activation of NMDA receptors and Ca2+ signalling in neurons. Ciba Found Symp, 1992. 164: p. 162-71; discussion 172-5.  47.  Fitzjohn, S.M., et al., A study of long-term potentiation in transgenic mice overexpressing mutant forms of both amyloid precursor protein and presenilin-1. Mol Brain, 2010. 3(1): p. 21.  83  48.  Morimoto, K., M. Fahnestock, and R.J. Racine, Kindling and status epilepticus models of epilepsy: rewiring the brain. Prog Neurobiol, 2004. 73(1): p. 1-60.  49.  Naylor, P., et al., Repeated ECS induces GluR1 mRNA but not NMDAR1A-G mRNA in the rat hippocampus. Brain Res Mol Brain Res, 1996. 35(1-2): p. 349-53.  50.  Decraene, C., et al., Identification of genes involved in ceramide-dependent neuronal apoptosis using cDNA arrays. Genome Biol, 2002. 3(8): p. RESEARCH0042.  51.  Reti, I.M., et al., Selective expression of Narp, a secreted neuronal pentraxin, in orexin neurons. J Neurochem, 2002. 82(6): p. 1561-5.  52.  Appelbaum, L., et al., Circadian and Homeostatic Regulation of Structural Synaptic Plasticity in Hypocretin Neurons. Neuron, 2010. 68(1): p. 87-98.  53.  Tiffany, S.T., A cognitive model of drug urges and drug-use behavior: role of automatic and nonautomatic processes. Psychol Rev, 1990. 97(2): p. 147-68.  54.  Pacchioni, A.M., et al., Neuronal Pentraxins Modulate Cocaine-Induced Neuroadaptations. Journal of Pharmacology and Experimental Therapeutics, 2008. 328(1): p. 183-192.  55.  Pacchioni, A.M., et al., Neuronal pentraxins modulate cocaine-induced neuroadaptations. J Pharmacol Exp Ther, 2009. 328(1): p. 183-92.  56.  Reti, I.M. and J.M. Baraban, Opiate withdrawal induces Narp in the extended amygdala. Neuropsychopharmacology, 2003. 28(9): p. 1606-13.  57.  Reti, I.M., et al., Narp regulates long-term aversive effects of morphine withdrawal. Behavioral Neuroscience, 2008. 122(4): p. 760-768.  58.  Reti, I.M., et al., Nicotine and Δ9-tetrahydrocannabinol withdrawal induce Narp in the central nucleus of the amygdala. Synapse, 2009. 63(3): p. 252-255.  84  59.  Reti, I.M., et al., Narp regulates long-term aversive effects of morphine withdrawal. Behav Neurosci, 2008. 122(4): p. 760-8.  60.  Sotres-Bayon, F., C.K. Cain, and J.E. LeDoux, Brain mechanisms of fear extinction: historical perspectives on the contribution of prefrontal cortex. Biol Psychiatry, 2006. 60(4): p. 329-36.  61.  Lu, W., et al., Amphetamine and cocaine do not increase Narp expression in rat ventral tegmental area, nucleus accumbens or prefrontal cortex, but Narp may contribute to individual differences in responding to a novel environment. Eur J Neurosci, 2002. 15(12): p. 2027-36.  62.  Johnson, A.W., et al., A selective role for neuronal activity regulated pentraxin in the processing of sensory-specific incentive value. J Neurosci, 2007. 27(49): p. 13430-5.  63.  Johnson, A.W., et al., Localized disruption of Narp in medial prefrontal cortex blocks reinforcer devaluation performance. Learning & Memory, 2010. 17(12): p. 620-626.  64.  Sutton, M.A., et al., Extinction-induced upregulation in AMPA receptors reduces cocaine-seeking behaviour. Nature, 2003. 421(6918): p. 70-5.  65.  Ary, A.W., et al., Ethanol up-regulates nucleus accumbens neuronal activity dependent pentraxin (Narp): implications for alcohol-induced behavioral plasticity. Alcohol, 2012. 46(4): p. 377-87.  66.  Marty, V.N. and I. Spigelman, Long-lasting alterations in membrane properties, k(+) currents, and glutamatergic synaptic currents of nucleus accumbens medium spiny neurons in a rat model of alcohol dependence. Front Neurosci, 2012. 6: p. 86.  67.  Stuber, G.D., et al., Voluntary ethanol intake enhances excitatory synaptic strength in the ventral tegmental area. Alcohol Clin Exp Res, 2008. 32(10): p. 1714-20.  85  68.  Ringman, J.M., et al., Proteomic changes in cerebrospinal fluid of presymptomatic and affected persons carrying familial Alzheimer disease mutations. Arch Neurol, 2012. 69(1): p. 96-104.  69.  Yin, G.N., et al., Neuronal pentraxin receptor in cerebrospinal fluid as a potential biomarker for neurodegenerative diseases. Brain Research, 2009. 1265: p. 158-170.  70.  Duong, T., P.J. Acton, and R.A. Johnson, The in vitro neuronal toxicity of pentraxins associated with Alzheimer's disease brain lesions. Brain Res, 1998. 813(2): p. 303-12.  71.  Pepys, M.B., et al., Targeted pharmacological depletion of serum amyloid P component for treatment of human amyloidosis. Nature, 2002. 417(6886): p. 254-9.  72.  Hossain, M.A., Neuronal Pentraxin 1: A Novel Mediator of Hypoxic-Ischemic Injury in Neonatal Brain. Journal of Neuroscience, 2004. 24(17): p. 4187-4196.  73.  Russell, J.C., et al., Neuronal pentraxin 1 induction in hypoxic-ischemic neuronal death is regulated via a glycogen synthase kinase-3alpha/beta dependent mechanism. Cell Signal, 2011. 23(4): p. 673-82.  74.  Clayton, K.B., et al., NP1 Regulates Neuronal Activity-Dependent Accumulation of BAX in Mitochondria and Mitochondrial Dynamics. Journal of Neuroscience, 2012. 32(4): p. 1453-1466.  75.  Hooper, C., R. Killick, and S. Lovestone, The GSK3 hypothesis of Alzheimer's disease. J Neurochem, 2008. 104(6): p. 1433-9.  76.  Nicotera, P. and S.A. Lipton, Excitotoxins in neuronal apoptosis and necrosis. J Cereb Blood Flow Metab, 1999. 19(6): p. 583-91.  77.  Mattson, M.P., Apoptosis in neurodegenerative disorders. Nat Rev Mol Cell Biol, 2000. 1(2): p. 120-9.  86  78.  Moran, L.B., et al., Neuronal pentraxin II is highly upregulated in Parkinson's disease and a novel component of Lewy bodies. Acta Neuropathol, 2008. 115(4): p. 471-8.  79.  Nicoletti, F., et al., Metabotropic glutamate receptors: from the workbench to the bedside. Neuropharmacology, 2011. 60(7-8): p. 1017-41.  80.  Schapira, A.H., et al., Novel pharmacological targets for the treatment of Parkinson's disease. Nat Rev Drug Discov, 2006. 5(10): p. 845-54.  81.  Mikati, M.A., et al., Hippocampal programmed cell death after status epilepticus: evidence for NMDA-receptor and ceramide-mediated mechanisms. Epilepsia, 2003. 44(3): p. 282-91.  82.  Dorsey, D.A., et al., Ultrastructural characterization of alpha-amino-3-hydroxy-5methyl-4-isoxazolepropionic acid-induced cell death in embryonic dopaminergic neurons. Apoptosis, 2006. 11(4): p. 535-544.  83.  Basile, A., et al., Characterization of the promoter for the human long pentraxin PTX3. Role of NF-kappaB in tumor necrosis factor-alpha and interleukin-1beta regulation. J Biol Chem, 1997. 272(13): p. 8172-8.  84.  Wang, Y., et al., alpha-Amino-3-hydroxy-5-methylisoxazole-4-propionic acid subtype glutamate receptor (AMPAR) endocytosis is essential for N-methyl-D-aspartate-induced neuronal apoptosis. J Biol Chem, 2004. 279(40): p. 41267-70.  85.  Nakagawa, T., The biochemistry, ultrastructure, and subunit assembly mechanism of AMPA receptors. Mol Neurobiol, 2010. 42(3): p. 161-84.  86.  Moran, L.B., et al., Neuronal pentraxin II is highly upregulated in Parkinson’s disease and a novel component of Lewy bodies. Acta Neuropathologica, 2007. 115(4): p. 471478.  87  87.  Affaticati, P., et al., Sustained calcium signalling and caspase-3 activation involve NMDA receptors in thymocytes in contact with dendritic cells. Cell Death Differ, 2011. 18(1): p. 99-108.  88.  Collingridge, G.L., J.T. Isaac, and Y.T. Wang, Receptor trafficking and synaptic plasticity. Nat Rev Neurosci, 2004. 5(12): p. 952-62.  89.  Molnar, E., et al., Membrane topology of the GluR1 glutamate receptor subunit: epitope mapping by site-directed antipeptide antibodies. J Neurochem, 1994. 63(2): p. 683-93.  90.  Seal, A.J., G.L. Collingridge, and J.M. Henley, An investigation of the membrane topology of the ionotropic glutamate receptor subunit GluR1 in a cell-free system. Biochem J, 1995. 312 ( Pt 2): p. 451-6.  91.  Horning, M.S. and M.L. Mayer, Regulation of AMPA receptor gating by ligand binding core dimers. Neuron, 2004. 41(3): p. 379-88.  92.  Pasternack, A., et al., Alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) receptor channels lacking the N-terminal domain. J Biol Chem, 2002. 277(51): p. 49662-7.  93.  Yao, G., et al., Crystal structure of the glutamate receptor GluA1 N-terminal domain. Biochem J, 2011. 438(2): p. 255-63.  94.  Mano, I., Y. Lamed, and V.I. Teichberg, A venus flytrap mechanism for activation and desensitization of alpha-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid receptors. J Biol Chem, 1996. 271(26): p. 15299-302.  95.  Mosbacher, J., et al., A molecular determinant for submillisecond desensitization in glutamate receptors. Science, 1994. 266(5187): p. 1059-62.  88  96.  Partin, K.M., M.W. Fleck, and M.L. Mayer, AMPA receptor flip/flop mutants affecting deactivation, desensitization, and modulation by cyclothiazide, aniracetam, and thiocyanate. J Neurosci, 1996. 16(21): p. 6634-47.  97.  Stern-Bach, Y., et al., A point mutation in the glutamate binding site blocks desensitization of AMPA receptors. Neuron, 1998. 21(4): p. 907-18.  98.  Quirk, J.C., E.R. Siuda, and E.S. Nisenbaum, Molecular determinants responsible for differences in desensitization kinetics of AMPA receptor splice variants. J Neurosci, 2004. 24(50): p. 11416-20.  99.  Dingledine, R., et al., The glutamate receptor ion channels. Pharmacol Rev, 1999. 51(1): p. 7-61.  100.  Nakagawa, T., et al., Structure and different conformational states of native AMPA receptor complexes. Nature, 2005. 433(7025): p. 545-9.  101.  Passafaro, M., et al., Induction of dendritic spines by an extracellular domain of AMPA receptor subunit GluR2. Nature, 2003. 424(6949): p. 677-81.  102.  Gregor, P., et al., Expression and novel subunit isoforms of glutamate receptor genes GluR5 and GluR6. Neuroreport, 1993. 4(12): p. 1343-6.  103.  Kuusinen, A., et al., Oligomerization and ligand-binding properties of the ectodomain of the alpha-amino-3-hydroxy-5-methyl-4-isoxazole propionic acid receptor subunit GluRD. J Biol Chem, 1999. 274(41): p. 28937-43.  104.  Chen, G.Q., et al., Functional characterization of a potassium-selective prokaryotic glutamate receptor. Nature, 1999. 402(6763): p. 817-21.  89  105.  Ayalon, G. and Y. Stern-Bach, Functional assembly of AMPA and kainate receptors is mediated by several discrete protein-protein interactions. Neuron, 2001. 31(1): p. 10313.  106.  Schneider-Mergener, J., Synthetic Peptide and Protein Domain Arrays Prepared by the SPOT Technology. Comp Funct Genomics, 2001. 2(5): p. 307-9.  107.  Hilpert, K., D.F.H. Winkler, and R.E.W. Hancock, Peptide arrays on cellulose support: SPOT synthesis, a time and cost efficient method for synthesis of large numbers of peptides in a parallel and addressable fashion. Nature Protocols, 2007. 2(6): p. 13331349.  108.  Erlanson, D.A., et al., In situ assembly of enzyme inhibitors using extended tethering. Nat Biotechnol, 2003. 21(3): p. 308-14.  109.  Arkin, M.R. and J.A. Wells, Small-molecule inhibitors of protein-protein interactions: progressing towards the dream. Nat Rev Drug Discov, 2004. 3(4): p. 301-17.  110.  Hopp, T.P. and K.R. Woods, Prediction of protein antigenic determinants from amino acid sequences. Proc Natl Acad Sci U S A, 1981. 78(6): p. 3824-8.  111.  Jiao, S. and Z. Li, Nonapoptotic function of BAD and BAX in long-term depression of synaptic transmission. Neuron, 2011. 70(4): p. 758-72.  112.  Pei, W., et al., Flip and flop: a molecular determinant for AMPA receptor channel opening. Biochemistry, 2009. 48(17): p. 3767-77.  113.  Bjartmar, L., Neuronal Pentraxins Mediate Synaptic Refinement in the Developing Visual System. Journal of Neuroscience, 2006. 26(23): p. 6269-6281.  90  114.  Monyer, H., P.H. Seeburg, and W. Wisden, Glutamate-operated channels: developmentally early and mature forms arise by alternative splicing. Neuron, 1991. 6(5): p. 799-810.  115.  Graham, S.H. and J. Chen, Programmed cell death in cerebral ischemia. J Cereb Blood Flow Metab, 2001. 21(2): p. 99-109.  116.  Hardingham, G.E., Y. Fukunaga, and H. Bading, Extrasynaptic NMDARs oppose synaptic NMDARs by triggering CREB shut-off and cell death pathways. Nat Neurosci, 2002. 5(5): p. 405-14.  117.  Beattie, E.C., et al., Regulation of AMPA receptor endocytosis by a signaling mechanism shared with LTD. Nat Neurosci, 2000. 3(12): p. 1291-300.  118.  Ehlers, M.D., Reinsertion or degradation of AMPA receptors determined by activitydependent endocytic sorting. Neuron, 2000. 28(2): p. 511-25.  119.  Lin, J.W., et al., Distinct molecular mechanisms and divergent endocytotic pathways of AMPA receptor internalization. Nat Neurosci, 2000. 3(12): p. 1282-90.  120.  Li, Z., et al., Caspase-3 activation via mitochondria is required for long-term depression and AMPA receptor internalization. Cell, 2010. 141(5): p. 859-71.  121.  Collingridge, G.L., et al., Effects of kainic and other amino acids on synaptic excitation in rat hippocampal slices: 1. Extracellular analysis. Exp Brain Res, 1983. 52(2): p. 1708.  122.  Ahmadian, G., et al., Tyrosine phosphorylation of GluR2 is required for insulinstimulated AMPA receptor endocytosis and LTD. EMBO J, 2004. 23(5): p. 1040-50.  123.  Pollack, M. and C. Leeuwenburgh, Apoptosis and aging: role of the mitochondria. J Gerontol A Biol Sci Med Sci, 2001. 56(11): p. B475-82.  91  124.  Yu, S.P., L.M. Canzoniero, and D.W. Choi, Ion homeostasis and apoptosis. Curr Opin Cell Biol, 2001. 13(4): p. 405-11.  125.  Manahan-Vaughan, D., A. Kulla, and J.U. Frey, Requirement of translation but not transcription for the maintenance of long-term depression in the CA1 region of freely moving rats. J Neurosci, 2000. 20(22): p. 8572-6.  126.  Sarker, R., et al., Casein kinase 2 binds to the C terminus of Na+/H+ exchanger 3 (NHE3) and stimulates NHE3 basal activity by phosphorylating a separate site in NHE3. Mol Biol Cell, 2008. 19(9): p. 3859-70.  127.  Seeburg, P.H., The TINS/TiPS Lecture. The molecular biology of mammalian glutamate receptor channels. Trends Neurosci, 1993. 16(9): p. 359-65.  128.  Hollmann, M. and S. Heinemann, Cloned glutamate receptors. Annu Rev Neurosci, 1994. 17: p. 31-108.  129.  Watanabe, I., et al., Glycosylation affects the protein stability and cell surface expression of Kv1.4 but Not Kv1.1 potassium channels. A pore region determinant dictates the effect of glycosylation on trafficking. J Biol Chem, 2004. 279(10): p. 8879-85.  130.  Kim, T.W. and R.E. Tanzi, Neuronal intranuclear inclusions in polyglutamine diseases: nuclear weapons or nuclear fallout? Neuron, 1998. 21(4): p. 657-9.  131.  Trottier, Y., et al., Polyglutamine expansion as a pathological epitope in Huntington's disease and four dominant cerebellar ataxias. Nature, 1995. 378(6555): p. 403-6.  132.  Saglietti, L., et al., Extracellular interactions between GluR2 and N-cadherin in spine regulation. Neuron, 2007. 54(3): p. 461-77.  92  133.  Pacchioni, A., The Role of AMPAR Trafficking Mediated by Neuronal Pentraxins in Cocaine-induced Neuroadaptations. Molecular and Cellular Pharmacology, 2009. 1(2): p. 71-75.  134.  Bachtell, R.K., et al., Role of GluR1 expression in nucleus accumbens neurons in cocaine sensitization and cocaine-seeking behavior. Eur J Neurosci, 2008. 27(9): p. 2229-40.  135.  Li, S.Y., D.S. Xu, and H.T. Jia, AGS-induced expression of Narp is concomitant with expression of AMPA receptor subunits GluR1 and GluR2 in hippocampus but not inferior colliculus of P77PMC rats. Neurobiol Dis, 2003. 14(3): p. 328-35.  136.  Fox, C.J., et al., Tyrosine phosphorylation of the GluR2 subunit is required for long-term depression of synaptic efficacy in young animals in vivo. Hippocampus, 2007. 17(8): p. 600-5.  137.  Koch, S.M. and E.M. Ullian, Neuronal pentraxins mediate silent synapse conversion in the developing visual system. J Neurosci, 2010. 30(15): p. 5404-14.  138.  Peineau, S., et al., LTP inhibits LTD in the hippocampus via regulation of GSK3beta. Neuron, 2007. 53(5): p. 703-17.  139.  Manning, B.D. and L.C. Cantley, AKT/PKB Signaling: Navigating Downstream. Cell, 2007. 129(7): p. 1261-74.  140.  Crombag, H.S., et al., Narp deletion blocks extinction of morphine place preference conditioning. Neuropsychopharmacology, 2009. 34(4): p. 857-66.  141.  Lu, J., et al., Structural recognition and functional activation of FcgammaR by innate pentraxins. Nature, 2008. 456(7224): p. 989-92.  142.  Bottazzi, B., et al., An integrated view of humoral innate immunity: pentraxins as a paradigm. Annu Rev Immunol, 2010. 28: p. 157-83.  93  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0073128/manifest

Comment

Related Items