AMPA Receptor Trafficking and Synaptic Plasticity
by Hengye Man
A Thesis Submitted in Conformity with Requirements
For the Degree of Doctor of Philosophy
Department of Lab Medicine and Pathobiology
University of Toronto
Copyright O by Hengye Man 2001 National Library Bibliotheque nationale du Canada Acquisitions and Acquisitions et Bibliographic Services se wices bibliographiques 395 WeIlington Street 395, rue Wellington Ottawa ON KIA ON4 Ottawa ON KIA ON4 Canada Canada
The author has granted a non- L'auteur a accorde me Licence non exclusive licence allowing the exclusive pennettant a la National Library of Canada to Bibliotheque nationale du Canada de reproduce, loan, distribute or sell reproduke, prgter, distribuer ou copies of this thesis in microform, vendre des copies de cette these sous paper or electronic formats. la forme de microfiche/- de reproduction sur papier ou sur format electronique.
The author retains ownership of the L'auteur conserve la propriete du copyright in this thesis. Neither the droit d'auteur qui protege cette these. thesis nor substantial. extracts fiom it Ni la these ni des extraits substantiels may be printed or othemise de celle-ci ne doivent heimprimes reproduced without the author's ou autrement reproduits sans son permission. autorisation. AMPA Receptor Trafficking and Synaptic Plasticity
Hengye Man Ph.D. 2001
Department of Lab Medicine and Pathobiology
University of Toronto
Abstract
AMPA receptors are a subtype of glutamate receptors which are responsible for most of the fast excitatory synaptic transmissions in the mammalian central nervous system. The redistribution of postsynaptic AMPA receptors has been proposed as a mechanism for synaptic plasticities, including
long-term potentiation (LTP) and long-term depression (LTD), two cellular models for learning and memory, However, direct evidence to support this hypothesis is still lacking and the cellular mechanisms mediating AMPA receptor trafficking are poorly understood. We therefore set up to investigate the involvement of AMPA receptor tracking in both LTD and LTP.
We found that AMPA receptors undergo endocytosis via a clathrin-coated-pit pathway, and that the internalization can be accelerated by insulin in a GluR.2 subunit-dependent manner. The insulin-stimulated endocytosis rapidly decreased the number of AMPA receptors in the plasma membrane, resulting in an LTD of AMPA receptor-mediated synaptic fxansmission in hippocampal
CAI neurons. Moreover, insulin-induced LTD and low-fiequency-stimulation (LFS)-induced hippocampal CA1 homosynaptic LTD were found to mutually occlude each other and both were blocked by inhibiting postsynaptic clathrin-mediated endocytosis. These data indicate that controlling the number of postsynaptic receptors by endocytosis may be an important mechanism underlying LTD in the mammalian CNS.
On the other hand, although AMPA receptor insertion is an attractive cellular model for LTP expression and is gaining increasing support in recent years, the underlying mechanisms for AMPA receptor translocation are still &own. We found here that the phosphoinositide kinase P13-kinase
(PI3K) directly associates and co-localizes with AMPA receptors. Selective activation of synaptic
NMDA receptors by the NMDA receptor co-agonist glycine activates the AMPA receptor-associated
PUK and induces long-term potentiation (LTP) of AMPA mini-EPSCs. Furthermore, brief glycine treatment induces increases in cell-surface expression of AMPA receptors and the increase is due to an enhancement in receptor plasma membrane insertion. Consistently, the glycine-induced LTP is blocked by inhibiting receptor insertion. Moreover, PI3K activation is also shown to be necessary in the expression of homosynaptic LTP in hippocampal CA1 neurons. These data provide evidence for an NMDA receptor-PUK-AMPA receptor translocation mechanism in LTP generation. This work has been done under the excellent supervision of my supervisors Dr- Yutian Wang and Dr. Laurence Becker. I would like to express my gratitude to them for their guidance, support and encouragement. I especially enjoyed the daily discussions with Dr. Wang £?om whom I learned a great deal about experimental neuroscience. I am very gratehl to Mr. Wiiliarn Ju, Dr. Lidong Liu,
Dr. Qinghua Wang and Dr. GhoIamreza Ahmadian in our iab, as well as Dr. Wei-Yang Lu in Dr.
John MacDonald' s lab, Mr. Jerry Lin in Dr. Morgan Sheng's lab (Harvard University) and Dr. You-
Ming Lu in Dr. John Roder's lab, for collaborating on some of the studies described in this thesis.
I would also like to thank Ms Jodi Braunton for her technical assistance, Mr. William Ju, Ms. Sandra
D'Souza and Ms. Adrienne Chen for critical reading of the thesis draft-
I would like to thank my supervisory committee members Drs H Yeger, E Cutz, and M Salter for their continuous support, helpful suggestions and critical review of my thesis.
Finally, this work could never be done without the love and support fiom my family. I can never give enough thanks to my parents, parents-in-law, my wife Pingping, my son Steven and my daughter Laura.
This work was supported by the Clinician-Scientist Award fiom the Training Center, Research
Institute of the Hospital for Sick Children, and the Ontario Neurotrauma Foundation. TABLE OF CONTENTS
ABSTRACT ...... -...... I1
ACKNOWLEDGEMENT ...... -...... IV
TABLE OF CONTENTS ...... V
LIST OF FIGURES ...... XI
LIST OF ABBREVIATIONS ...... XIV
CHAPTER ONE
Introduction and Background Review
1.1 Glutamate receptors ...... 2
1.1.1 AMPA receptors...... ,...... 4
1.1.2 NMDA receptors ...... 6
1.2 Regulation of excitatory synaptic transmissions ...... 7
1.2.1 Phosphorylation of postsynaptic channel receptors ...... 9
1.2.1A AMPA receptor phosphorylation ...... -...... -9
1.2.1B NMDA receptor phosphorylation ...... 12
1.2.2 Receptor trafficking in regulation of synaptic transnnission...... 15
1.2.2A Background on plasma membrane protein- trafficking ...... ,., 16
1.2.2B AMPA receptor trafficking and synaptic plasticity ...... -23 1.3 Glutamate receptor-interacting proteins ...... 27
1-3-1 NMDA receptors interacting proteins .....~~.~...... -28
1 -3-2 AMPA receptor interacting proteins ...... 33
1-4 phosphatidylinositol 3'-kinase (P13K) ...... 36
1.4.1 Classification of PI3Ks...... 38
1 .4.2 PI3K activity and its downstream signalling pathways ...... 39
1.4.3 P13K regulates protein trafficking and neuron& hction...... 41
CHAPTER TWO
Experimental Methods
2.1 cDNA plasrnids and new constructs ...... -48
2.2 Transient transfection on HEK 293 cells ...... 48
2.3 Primary cultures of hippocampal neurons ...... 49
2.4 irnmunostaining ...... ,...... 49
2.4.1 Selective labeling of cell surface vs total AMPA receptors and internalization staining
...... 49
2.4.2 Double staining of AMPA receptor and synaptophysin or NMDAFU ...... SO
2.4.3 Pre-blocking immunostaining (AMPA receptor insertion staining) ...... 51
2.5 ~unoprecipitationand western blotting ...... -54
2.5.1 Co-immunoprecipitation of GluR2 with adaptin P2 ...... -54
2.5.2 Co-immunoprecipitation of GluR2 with PI3K ...... -54 VI 2.6 Colorimetric assay of AMPA receptor cell surface expression ...... 55
2.7 PI3-base assay ...... 56
2.8 Whole-cell recording of EPSCs in hippocampal brain slices...... 57
2.8.1 Rat brain slice preparation ...... 57
2.8.2 Whole-cell recording of EPSC in CAI neurons ...... 57
2.8.3 LTD induction ...... ,...... -61
2.9 Miniature EPSC recordings in cultured hippocampal neruon ...... ,...... 61
CHAPTER THIWE
Long-term Depression of Excitatory Synaptic Transmission Through AMPA
Receptor Internalization
Abstract ...... -64
Introduction ...... -64
Results ...... 72
3.3.1 Insulin produces a rapid decrease in cell-surface AMPA receptors via a GluR.2 subunit
dependent mechanism ...... -72
3 -3.2The carboxyl tail of GluR2 determines subunit-specificity of insulin-induced reduction
in cell-surface AMPA receptors ...... ,...... 79
Insulin selectivelyfacilitates the rate of GldUbut not GluRl receptor endocytosis ...80
Both constitutive and insulin-regulated GluR.2 endocytosis are mediated by clathrin-
VII coated pits ...... -...... 81
3 -3-5 Facilitated ciathrin-dependent endocytosis is fully responsible for insulin-induced
reduction in cell-surface expression of GluR2 receptors ...... 88
3.3.6 Insulin reduces the surface expression of AMPA receptors in cultured hippocampd
neurons in a clathrin-dependent manner ...... 89
3 -3.7 Insulin stimulates the association of native AMPA receptors with the AP2 complex in
hippocampal slices ...... 90
3.3.8 Insulin produces a long-lasting depression of the AMPA component of the excitatory
postsynaptic currents in hippocampal CAI cells ...... ,...... 9 1
3 -3-9 The induction of the insulin-induced depression of AMPA EPSCs is postsynaptic.... 91
3.3 -10 Post-synaptic expression of insulin-induced reduction in AMPA EPSCs ...... 98
3 -3.11 Insulin-induced depression of AMPA EPSCs is blocked by inhibiting clathrin-
mediated endocytosis in postsynaptic neurons ...... 99
3.3-12 Inhibiting clathrin-mediated endocytosis in postsynaptic neurons prevents CA1
homosynaptic LTD ...... ~..~...... ~...... 100
3 .3.13 Insdin induced LTD and LFS-induced LTD mutually occlude each other ...... 101
3 -4Discussion ...... 106 CHAPTER FOUR
Activation of AMPA Receptor-associated PU-kinase Induces LTP Through
AMPA Receptor Plasma Membrane Insertion
4.1 Abstract ...... 15
4.2 Introduction ...... 115
4.3 Results ...... 123
4.3.1 PUK directly interacts with AMPA receptors ...... 123
4.3 -2 PI3K colocalizes with AMPA receptors at the synaptic site ...... 124
4.3.3 Activation of synaptic NMDA receptors potentiates the activity of AMPA receptor-
associated PI3K ...... 125
4.3 -4NMDA receptor-mediated activation of AMPAR-associated PI3K leads to translocation
and activation of Akt ...... 132
4.3 -5 Selective activation of synaptic NMDA receptors induces long-term potentiation (LTP)
of AMPA-miniEPSCs through activation of PI3K in cultured neuron ...... 133
4-3.6 Synaptic NMDA receptor activation increases cell surface expression of AMPA
receptors in cultured neuron ...... -135
4.3.7 Glyche treatment increases AMPA receptor synaptic localization ...... 142
4.3.8 Glycine promotes AMPA receptor insertion in cultured neurons ...... 143
4.3 -9Glycine-induced LTP is abolished by blocking AMPA receptor exocytosis ...... 144 4.3.1 0 PUK is required in the expression of homosynaptic LTP in hippocampal slice CAI
neuron ...... 150
4.4 Discussion ...... 150
CHAPTER FIVE
Summary, Conclusions and Future Directions
5.1 Research summary ...... 156
5.2 Future directions ...... 16 1
Appendix
Publications ...... 164
Reference list
...... 176 List of figures
Figure 1.I Gl~tamatereceptors ...... 3
Figure 1.2 A schematic diagram illustrating the common steps involved in the plasma membrane
insertion and endocytosis of integral plasma membrane receptors ...... 19
Figure 1.3 Mechanisms underlying membrane insertion and endocytosis of plasma membrane
proteins ...... -21
Figure 2.1 Selective detection of cell-surface and internalized AMPA receptors by immunostaining
or colorimetric assay ...... -33
Figure 2.2 Experimental configuration for recording of evoked post-synaptic currents in hippocampal
CA 1neurons ...... -58
Figure 2.3 Recording of AMPA receptor-mediated EPSCs in hippocampal CA 1 neurons ...... 59
Figure 3.1 Selective imrnunostaining of AMPA receptors on the cell surface or its total pool ...... 67
Figure 3.2 Insulin treatment affects the cell surface expression of AMPA receptor subunits ...... 68
Figure 3.3 Quantification of insulin effect on AMPA receptor cell-surface expression rate by
colorimetric assays ...... 70
Figure 3 -4 The GWspecificity of insulin effect is determined by the intracellular C-terminus of the subunit ...... -74
Figure 3.5 AMPA receptors are subject to endocytosis which is facilitated by insulin ...... 76
Figure 3 -6 hmunostxining showing that both constitutive and insulin-stimulated GluR2 endocytosis
are mediated by clathrin-coated pits ...... 78
Figure 3 -7 Insulin-induced reduction in cell-surface AMPA receptors is blocked by inhibiting
clathrin-dependent receptor endocytosis ...... -83
Figure 3 -8 Clatbrin-mediated endocytosis of AMPA receptors in cultured hippocampal neurons..85
Figure 3 -9 Insulin increases the association of native AMPA receptors and the AP2 adaptor protein
complex in hippocampal tissue ...... 87
Figure 3.10 Insulin induces long-lasting depression of AMPA receptor-mediated excitatory
postsynaptic currents in hippocampal CAI neurons ...... -93
Figure 3 -11 Insulin selectively inhibits minimum stimulus-evoked AMPA EPSCs ...... 95
Figure 3.12 Insulin-induced depression of AMPA EPSCs is blocked by inhibiting clathrin-mediated
endocytosis in postsynaptic neurons ...... -97
Figure 3.13 Inhibition of clathrin-dependent endocytosis in postsynaptic neurons prevents the
expression of low-frequency-stimulus (LFS) induced LTD in CAl neurons ...... 103
Figure 3.14 Homosynaptic LTD and insulin-induced LTD mutually occludes each other in
hippocampal CAI neurons...... 105
Figure 4.1 Functional P13-kinase (P13K) binds to AMPA receptors ...... 1 18
Figure 4.2 PI3-kinase colocalizes with AMPA receptors in cultured hippocampal neuron...... 120
Figure 4.3 Activation ofsynaptic NMDA receptors activates the AMPA receptor-associated PDK. 122 Figure 4.4 Activation of synaptic NMDA receptors induces translocation and phosphorylation of
Akt ...... ,...... ~...... 127
Figure 4.5 Glycine induces long-term potentiation (LTP) of AMPA-miniEPSCs through activation
ofPI3K incultured neurons ...... 129
Figure 4.6 Intracellular application of active PI3Kpotentiates AMPA-miniEPSCs...... 131
Figure 4.7 Glycine treatment induces a NMDA receptor-dependent increase in cell-surface
expression of AMPA receptors in cultured hippocampa1neurons ...... 1 37
Figure 4.8 Glycine increases AMPA receptor clusters at synapses...... 139
Figure 4.9 Glycine facilitates AMPA receptor plasma membrane insertion in cultured hippocampal
neurons ...... 141
Figure 4.10 Intracellular injection oftetanus toxin blocks glycine-induced increase in AMPA receptor
cell-surface expression ...... , .... 146
Figure 4.1 1 Intracellular application of active PI3K potentiates AMPA receptor-mediated EPSCs in
hippocampal CA1 neurons ...... 148
Figure 5.1 ...... 158
Figure 5.2 ...... 159 List of Abbreviations
ABP :AMPA receptor-binding protein
AMPA: a-amino-3-hydroxy-5-methylisoxazoIe-4-proocacid
AP2: adaptor protein 2
APV: D-amino-phosphonovaleric acid, an NMDA receptor antagonist
CaMKlI: calmoddin-dependent protein kinase I1
DNQX: 6,7-dinitroquinoxaline, an AMPA receptor antagonist
EPSC : excitatory post-synaptic current
GFP: green fluorescent protein
GluRs: glutamate receptor subunits
GRIP: glutamate receptor-interacting protein
IP: immunoprecipitation
LTD: long-term depression
LTP: long-term potentiation
MAGUK: membrane associated guanylate kinase
NMDA: N-Methyl-D-Aspartate
NR1: NMDA receptor subunit R1
NSF: N-ethylmalehide-sensitive factor
PDK: pbosphoinositide-dependent base
PDZ domain: a protein-protein interaction domain, shared by PSD-95, Dlg (discs large) and ZO-1 PUK: phosphatidylinositol-3' kinase
PICK1: protein interacting with C-kinase
PKA: protein kinase A
PKC: protein kinase C
PSD-95:a postsynaptic density protein with a molecular weight of 95KD
SNAP: soluble NSF-attachment proteins
SNARE: soluble NSF-attachment protein receptor
TeTx: tetanus toxin
TTX:tetrodo toxin
VAMP: vesicle associated membrane protein CHAPTER ONE
Introduction and Background Review Synaptic transmission is one of the fhdamental processes by which neurons function, In the CNS,
an action potential propagates along the axon and triggers neurotransmitter release when it reaches
the presynaptic terminal. Neurotransmitters will then d-e across the synaptic cleft and bind to and
activate the neurotransmitter receptors on the postsynaptic membrane, and flow of ions through the
channel receptors will finally depolarize or hyperpolarize the membrane potential. As a consequence,
ion channel receptors act as transducers converting chemical signals into electrical ones. In the
mammalian CNS, fast synaptic transmission is mediated mainly through two types of
neurotransmitter receptors: GABA, receptors for inhibitory fast synaptic transmission (Thompson,
1994); and glutamate receptors for excitatory synaptic transmissions (Bekkers and Stevens, 1995).
In order to adapt the brain to a constantly changing environment and to perform sophisticated brain
functions such as learning and memory, constant regulation of the efficacy of the synaptic
transmission is of extreme importance, and therefore, understanding the mechanisms underlying the
regulation of synaptic transmission has been one of the most extensively studied areas in neuroscience.
1.1Glutamate receptors
40 years ago, Curtis et al. applied glutamate onto single neurons in the brain and spinal cord by using microelectrophoresis technique, and the resultant Wgof action potentials were found following glutamate application (Curtis et al, 1959). This was the first evidence to show the excitatory effect of glutamate at postsynaptic neurons and initiated extensive investigations aiming to confirm whether glutamate acts as a neurotransmitter and the nature of its receptors.
It is now clear that glutamate and glutamate receptors are the mediators of most fast excitatory Fig 1.1 Glutamate receptors
A. Classification of glutamate receptors
GluR2 AMPA receptor { GluR3
NRI NR2A * Ionotropic glutamate receptor NMDA receptor (Ion channel) NR2D
GluRS GluR6 Glutamate receptor Kainate receptort ~la7
Metabotropic glutamate receptor Group I (mGluR1, mGluR5) (G-protein coupled receptor) Group II (mGld3.2, mGluR3) Group In (rnGluR4, mGIuR6, rnGluR7, mGluR8)
B. Plasma membrane topology of ionotropic glutamate receptor subunits
An ionotropic glutamate receptor subunit contains an extracellular N-terminus, an intracellular C- terminus and four membrane domains (I, 11, ID, Iv). The splice variants Flip/Flop and the Q/R editing site are shown in red. synaptic transmission in the mammalian centraI nervous system (von Kitzing et al., 1994). The
glutamatergic system is important for synaptic pIasticity (Asztely and Gustdsson, 1996; Bear amd
Abraham, 1996; Larkman and Jack, 1995; Voronin, 1994), which underlie many advanced braim
functions like learning and memory. Also, excessive glutamate release or overstimulation od
glutamate receptors are believed to be the main reason in causing neuronal death in certaim
pathological conditions, such as stroke and other forms of neurotrauma (Chen et al., 1995; ColweLl
and Levine, 1996; Doble, 1995; Fujisawa et ale, 1993)-
Glutamate receptors are divided into two major groups: ionotropic glutamate receptors (ion
channels) including the NMDA receptor, the AMPA receptor and the kainate receptor; and
metabotropic glutamate receptors which are G-protein coupled receptors mediating slow and
prolonged responses (Doble, 1995;Swopeet d., 1999). As AMPA and NMDA receptors are the main
. mediators for fast synaptic transmission, only these two types of glutamate receptors will be discussed
in more detail here.
1.1.1 AMPA receptors
To date, four AMPA receptor subunits, GLuRl-4, have been identified, which all share the same
plasma membrane topology: an extracellular N-terminus, an intracellular C-terminus, three
transmembrane domains and a re-entrant loop between the fist and the second transmembrane
domains. AMPA receptor subunits are diversified by a variety of modifications at the RNA level-
First, all AMPA receptor subunits have alternatively spliced flip/flop variants (Sommer et a,., 1WO),
a fragment just prior to the 3rd transmembrane domain which is expressed by mutually exclusive
exons named flip and flop. Second, the c-termini of GluR2 and GluR4 subunits are also alternatively- spliced to form long and short forms of the corresponding subunit (Dingledine et d., 1999), and the
short forms of GlWand GluR4 share homology with Glmin the c-terminal region. Although the
function of this splicing is not clear, the specific binding of AMPA receptor-interacting proteins with
different splice variants suggests a possible role in subcellular targeting and localization for different
subtypes of AMPA receptors. Third, an RNA editing process converts a glutarnine (Q) in the re-
entrant site into arginine (R) (Seeburg, 1996)- It is specific for GluR2 in AMPA receptor subunits, but is also seen in GLUM and GluR6 of kainate receptors (Seeburg 2996). Edited GIuR2, when it forms heteromeric AMPA receptors with other subunits, confers calcium impermeability and a linear current-voltage relationship to the receptor (Washburn et al., 1997). This R/Q editing is extremely important in protecting neurons fiom excitotoxicity by blocking calcium influx during frequent synaptic transmissions. In supporting its crucial role, mice ~3.ha deficiency in GLuR.2 editing have epilepsy and die during perinatal period (Tellegrini-Giarnpietro et al., 1997; Prince et al., 1995).
AMPA receptors are heteromultimers, presumably either heteropentamers or heterotetramers, but the exact combination and stoichiorneby is still a controversy (Rosenmund et al., 1998). In heterologous expression systems, it has been demonstrated that GluRs can form both homomeric and heteromeric hctional receptor channels with distinct properties (Boulter et al., 1990; Keinanen et d., 1990). In native neurons, most AMPA receptors contain GluR2 subunits and so confer caIcium impermeability. In hippocampal CAI region, AMPA receptors contain GluR2 together with either
GluRl or GluR3, but very few receptors appear to contain both GluRl and GluR3 (Wenthold et al.,
1996). In addition, even in a single cell, the subunit combinations of AMPA receptors are not uniform (Iino et al., 1996), and the subunit composition could be determined by the different origins of the afferent inputs onto the postsynaptic neuron (Toth and McBain, 1998). 1.1.2 NMDA receptors
To date, molecular cloning has revealed two groups ofNMDA receptor subunits. The first is the NR1 subunit consisting of different splice variants. The second group contains fourNR2 subunits, termed
NR2A-2D, which share 1520% homology with MU (Monyer et al., 1992). It is believed that the
NMDA receptor is tetrameric or pentameric and its subunits have a transmembrane topology similar to that of AMPA receptor subunits. In Xenopus oocytes, NR1 subunits have smalI responses to agonist and the co-expression with MU subunits greatly enhances theses responses(Monyer et al.,
1992;Moriyoshi et al., 1991). However, in mammalian systems, expression of NRl alone fails to produce any NMDA current (Boeckman and Aizenman, 1993), a discrepancy which can probably be explained by the fmding of low level endogenous NR2 subunits in oocytes (Soloviev and Barnard,
1997). In fact, it has been reported that NR1 subunits will be retained intracellulady when expressed alone, and can reach to the plasma membrane and form fbnctional channels only when co-expressed with any NR2 subunits (Mcllhinney et al., 1996a). Studies of recombinant NMDA receptors have shown that different subunit compositions have distinct pharmacological and biophysical properties
(Sucher et al., 1996) and the subunit composition changes during development (Farrant et al., 1993;
Farrant et al., 1994). For example, cerebellar Purkinje cells express functional NMDA receptors composed of NMNR2D during early postnatal stages. However, due to a loss of MUD subunits during the subsequent periods of development, fhctional NMDA receptors are lacking in mature
Purkinje cells (Farrant et al., 1993; Akazawa et al., 1994; Akazawa et al., 1994). NMDA receptors co-localize with AMPA receptors at postsynaptic sites and as aresult, during synaptic transmissions, most EPSCs contain both AMPA and NMDA components. However, the two components can be distinguished by their distinct kinetics. The AMPA component rises quickly (1 00-200 ps) and decays rapidly with a time constant of 1-8 ms. The fast time course of the AMPA component makes it possible to mediate high frequency synaptic transmission without the loss of the temporal information (Feldrneyer and Cull-Candy, 1996). On the other hand, the NMDA component shows slow kinetics with a rise time of about 10 ms and a decay time of 100 ms. Furthermore, the NMDA receptor has several unique properties when compared with AMPA receptors. First, NMDA receptors are highly permeable to calcium, which is important both in generating synaptic plasticity and in neuronal excitotoxicity. Second, NMDA channels are blocked by M~Z'at resting conditions. As the blocking Mg" can be displaced fkom the channei by membrane depolarization, this property makes the NMDA receptor both agonist- and voltage-dependent. As a consequence, at a given synapse, expression of NMDA receptors without the co-existence of AMPA receptors will actually make the synapse hctionless, or ccsilent",in synaptic transmission (Gornperts et d., 1998; Isaac et d., 1995;
Liao et al., 1995). Third, NMDA receptor activation requires the obligatory co-agonist glycine
(Bashir et d., 1990). This characteristic provides a useful means for selective activation of synaptic, rather than non-synaptic, NMDA receptors by applying additional glycine during synaptic recording
(see chapter 4).
1.2 Regulation of excitatory synaptic transmission
Following glutamate release during synaptic transmission, AMPA receptors will be activated and allow ions (mainly Na' ) to enter the cell leading to membrane depolarization. At normal resting conditions, the membrane potential is -60 - -7OmV and NMDA receptors are mostly blocked by Mg" at this negative potential. The MgZ' blockade of NMDA channels will be removed when the membrane potential becomes less negative by depolarization, which normally occurs following activation of AMPA receptors- Due to the very limited role of kainate receptors, fast excitatory synaptic transmissions are mediated mainly through AMPA receptors while NMDA receptors are important in modulating neurotransmission and synaptic plasticity.
Synaptic activity is under precise and complex control, and any change at any step of the synaptic transmission process may alter synaptic strength. In simple terms, synaptic regulation can be classified into two categories: (1) to change the presynaptic transmitter release, and (2) to modulate the postsynaptic receptor function (Malenka and Nicoll, 1999). While neurotransmitter release is an important target in the regulation of synaptic transmission, studies indicate that alteration of the function of neurotransmitter receptors at the postsynaptic site is a more efficient way in producing long-term synaptic plasticity (Malenka and Nicoll, 1999). This review will therefore focus on the postsynaptic mechanisms in the regulation of synaptic strength.
Fast synaptic transmissions are mediated by ion channel receptors, such as GABA, receptors and AMPA type glutamate receptors. Their functions are reflected by the channel characteristics, i.e., channel conductance, channel open time and open probability . Factors which can modulate receptor channeI characteristics will change synaptic transmission strength. Over the last ten years, extensive studies have revealed that protein phosphorylation is a universd and very powerful means in regulating functions of neurotransmitter receptors (Courtneidge et al., 1993; Huganir and Greengard,
1990; Levenes et al., 1998; Roche et al., 1994; Sigel, 1995). This biochemical modulation, accomplished by kinases for phosphorylation and phosphatases for dephosphorylation, is very efficient and reversible. However, recent studies support a newly proposed mechanism in regulation of synaptic transmission, that is, in addition to modulation of the properties of pre-existing receptors in the plasma membrane, the number of synaptic receptors can also be subject to regulation (Alconada et al., 1996; Edwardson and Szekeres, 1999; Shi et al., 1999; Tehrani and Barnes, 1993). Obviously, this type of regulation of synaptic transmission requires receptor traflicking (receptor membrane insertion or receptor internalization), but the mechanisms underlying glutamate receptor trafficking are still poorly understood.
1.2.lPhosphorylation of postsynaptic channeI receptors
1.2.lA AMPA receptor phosphorylation
Of the many possible mechanisms for modulating the efficiency of ion channeIs, receptor protein phosphorylation may be the primary one. AMPA receptors, like many other proteins, are subject to phosphorylation by various kinases, including PKA, PKC and CaMKII (Levenes et al., 199 8; Swope et al., 1999; Wang et aI., 1991). As protein phosphorylation changes the channel properties and the time course and the extent of phosphorylation is under regulation, it is believed that phosphory lation is a maj or mechanism underlying the induction and maintenance of LTP and LTD (Malenka,R-C. and
Nicoll,R.A, 1999; Mulkey et al., 1993; Lu et al., 1998; Lee et al., 2000).
GluRl can be phosphorylated on multiple sites that are all located on its C-terminus. PKA specifically phosphorylates S845 of GluRl in transfected HEK cells and in cultured neurons.
Mutation of this site (S845A) blocks GluRl phosphorylation by PKA (Roche et al., 1996). In brain slices, using an antibody directed against specific phosphorylated residues in GluRl c-terminus,
Western blotting showed phosphorylation of both S845 and S831 (Mamrnen et al., 1997).
Phosphorylation of S84S results in a 40% potentiation of the peak current through homomeric GluRl channels (Roche et al., 1996), and this potentiation is likely due to an increase in receptor channel open probability, and not the channel conductance and open time (Banke et al., 2000). In addition to PKA, GluRl can also be phosphorylated by PKC at S831 of the GluRl c-terminus. This phosphoryIation site is unique to GluRl and no corresponding site is found in GluR2-4 (Roche et al.,
1996). Also, GluR.2 subunits appear to be phosphorylated at S880 by PKC (Matsuda et al., 1999).
Tmmunocytochemistry using a polyclonal antibody recognizing the phosphorylated S696 in GluR2 showed that AMPA treatment induces immunostaining in cerebellar slices (Nakazawa et al., 1995).
The kinases involved in this phophoryIation have not been identified, but they are unlikely to include either PKC or PKG, as inhibitors for either kinase do not block the immunostaining (Nakazawa et al.,
1995). In brain slices, activation of PKC increases both the frequency and amplitude of AMPA receptor miniature EPSCs in CAI neurons (Carroll et al., 1W8), indicating that PKC is involved in the regulation of synaptic transmission, presumably through the phosphorylation of AMPA receptors.
Recently, Matsuda et al found that phosphorylation of GluR2 by PKC disrupted its binding with glutamate receptor-interacting protein GRIP 1 (Matsuda et al., 1999)- Because the interacting proteins have been implicated in AMPA receptor intracellular trafficking and membrane targeting, alteration of the protein interactions of AMPA receptors with intracellular proteins by phosphorylation could be a novel means in modulating synaptic transmissions.
While AMPA receptor phosphoryIation potentiates synaptic function, other studies showed that AMPA receptor dephosphoryIation may cause downregulation of synaptic transmissions
(Kameyama et al., 1998; Lee et al., 1998; Lee et al., 2000) . It was demonstrated that NMDA treatment of hippocampal slices reliably induced long-term depression which share common features with electricd stimulation-induced homosynaptic LTD. This chemical LTD was seen to be accompanied by a persistent dephosphorylation of the GluRl subunit of AMPA receptors at serine
845, a CAMP-dependentprotein kinase (PKA) substrate, but not at serine 83 1, a substrate of protein kinase C @KC) and calcium/calmodulin-dependentprotein kinase II (CaMKII) ( Lee et al., 1998;
Karneyama et al., 1998). The dephosphorylation of GluRL Ser845 was also observed in low frequency-stimulation &FS)-induced homosynaptic LTD Ceeet al., 2000). Interestingly, if LFS was applied following LTP to depotentiate the synaptic strength back to the baseline, a dephosphorylation was still observed except that it was not on Ser845 but on Ser83 1, indicating that LTD induces dephosphorylationof different sites on GluRl depending on the previous experience of the synapse
(Lee et d,2000).
CaMKII is a major protein in the postsynaptic density (PSD). In its basal state, CaMKII is inactive due to an autoinhibitory domain which blocks the binding with its substrate. Binding with calcium/calmodulin removes the autoinhibition and activates the base following its autophosphorylation (Colbran, 1992; Soderhg et al., 1991). By using specific antibodies against phosphorylated peptides and site-directed mutagenesis, it was found that CaMKII phosphorylates
GluRl at S831, a site unique in the GIuRlc-terminus which can also be phosphorylated by PKC
(Barria et al., 1997; Roche et al., 1996; McGlade-McCulloh et al., 1993). The autophosphorylated
CaMKII retains high activity even after calcium/calmodulin disassociation, and thus a transient elevation of calcium concentration in a dendritic spine can be converted into prolonged kinase activity even in the absence of elevated calcium concentration, These unusual biochemical properties make
CaMKU a unique transducer of postsynaptic NMDA receptor-dependent calcium signalling and an ideal molecule for inducing LTP. In line with this idea, intrace1luIa.r application or expression of active CaMKII enhance both the AMPA receptor-mediated currents in cultured hippocampal neurons
(McGlade-McCulloh et d., 1993) and the synaptic current in CAI neurons of hippocampal slices pettit et al., 1994). Application of a phosphatase inhibitor to enhance the endogenous kinases in CAI neurons potentiates the AMPA, but not the NMDA, component of EPSCs, and the effect can
be blocked by CaMKII inhibitors (Figurov et al., 1993). To confirxn that the functional change of
AMPA receptors were indeed due to direct phosphorylation by CaMKII, it was found that the S83 1A
mutant of GluRlwas not phosphorylated and not potentiated by CaMKII (Barria et al., 1997).
Furthermore, the CaMKII-induced enhancement can occlude subsequent LTP induction, indicating that phosphorylation of GluRlby CaMKII may be the mechanism underlying LTP expression.
Moreover, single-channel recording demonstrated multiple conductance states for GluR1, and coexpression of CaMKI[I increased the proportion with higher conductance states (Derkach et aI.,
1999), indicating that CaMKII mediates LTP of AMPA currents either by increasing single-channel conductance of existing AMPA receptors or by recruiting new AMPA receptors with high- conductance.
1.2.1B NMDA receptor phosphorylation
NMDA receptors, like AMPA receptors, can be phosphorylated by dierent kinases, and tyrosine phosphorylation of NMDA receptors as well as the hctional significance have been most extensively studied. Using an anti-phosphotyrosineantibody, it was found that the NR2B subunit is the most prominently tyrosine-phosphorylated protein in the PSD (Moon et al., 1994), although it seems likely that bothNR2A and NR2B subunits can be tyrosine phosphorylated (Lau and Huganir,
1995). Tyrosine phoshorylation of NMDA receptors can be induced by growth factors. It was shown that brain-derived neurotrophic factor (BDNF) treatment, which can enhance excitatory postsynaptic currents in cultured dissociated hippocampal neurons, increased tyrosine phosphorylation of NR2B, but not the NR2A, subunit in cortical or hippocampal postsynaptic densities (Lin et al., 1 998a). Similar effects were also observed following insulin treatment, in which both NR2A and
NR2B were tyrosine phosphoryIated (Christie et al., 1999). While endogenous tyrosine kinases for
NMDA receptor phosphorylation are still elusive, cytosolic tyrosine kinase Src might be a good candidate, because it potentiates NMDA receptor hctionand is associated with Nh.1DA receptor in the brain (Wmg and Salter, 1994; Yu et al,, 1998; Yu et al., 1997). In cultured spinal cord dorsd horn neurons, it was found that intracellular application of the tyrosine kinase Src or a tyrosine phosphatase inhibitor lavendustin A potentiatedNMDA currents (Wang and Salter, 1994). Similarly, intracelhlar administration of Src or Fyn kinases also enhanced the glutamate-induced currents in
HEK cells expressing NRl/NR2A (Kohr and Seeburg, 1996). Furthermore, activating the endogenous
Src kinase by a Src-activating peptide, or inhibiting the endogenous Src by an anti-Src antibody, caused either an increase or decrease, respectively, of NMDA receptor single channel activities (Yu et al., 1997). Tyrosine phosphorylation of NMDA receptors appears to be important in NMDA receptor-dependent long-term plasticity of synaptic transmission. Studies showed that tyrosine phosphorylation of NR2B was enhanced during LTP (Rostas et al., 1996) and blocking the intracellular tyrosine kinases blocked LTP expression (O'Dell et al., 1991). Consistent with this idea, intracellular application of tyrosine kinase induced LTP (Lu et al., 1998). Among a variety oftyrosine kinases, Src is a particularly suitable mediator of the LTP-induction process. Src is activated quickly by tetanus stimulation; blocking endogenous Src blocks LTP and asionof active, or activating the endogenous, Src kinase induces LTP. As the Src effect is blocked by NMDA receptor antagonists, it has been proposed that the Src-dependent LTP is mediated by potentiating the ongoing NMDA receptor activity (Lu et al., 1998)-
In addition to tyrosine phosphorylation, NMDA receptors are substrates for serinekhreonine
13 phosphorylation. In cultured hippocarnpal neurons, phosphatases 1 and 2A modulates NMDA
receptor functions (Wang et al., 1994). It is estimated that about 1O-700/0 ofNMDA receptor subunits
are phosphorylated by PKA or PKC (Leonard and Hell, 1997), a percentage much higher than that
by tyrosine phosphorylation which is around 2-4% of NR2A and NR2B. Using site-specific anti- phosphoserine antibodies, Tingley et a1 (1997) demonstrated that PKC phosphorylates serine
residues 890 and 896 and PKA phosphorylates serine residue 897 of the NR1 subunit. A series of
studies showed that activation of different of G-protein coupled receptors, such as rnetabotropic glutamate receptors, opioid receptors, and muscarink acetylcholine receptors, potentiates NMDA currents in a PKC-dependent manner, presumably through phosphorylation of NMDA receptors by
PKC ( Aniksztejn E, Ben-Ari, 1992; Chen and Huang, 1992; Dildy-Mayfield and Harris, 1994). In addition to modulating NMDA receptor hction, it was found that phosphorylation of serine 890 of
NR1 by PKC results in the dispersion of surface-associated clusters of the NRl subunit expressed in
fibroblasts, while phosphorylation of serine 897 by PKA has no effect on the subcellular distribution of NR1 (Ehlers et al., 1995; Tingley et al., 1997), suggesting that serine/threonine phosphorylation may also play a role in NMDA receptor localization. However, fhther studies on native neuron are needed to clarify this possibility.
NMDA receptors can also be phosphorylated by CaMKTI. S1303 in the NUB c-terminus,
and possibly the corresponding site in NR2A, was found to be the substrate residue of CaMKfI in vitro and phosphorylation on this site was also confirmed in hippocampal neurons (Ornkumar et al.,
1996). CaMKII is also found to be able to associate with NMDA receptors. CaMKn co-cocalizes with NMDA receptors in spines and can be coirnmunoprecipitated from brain tissue. In vibo,
autophosphorylated CaMKII directly binds to NR2B and NR1. In transfected cells, colocalization
14 of CaMKII and NR2B-containing NMDA receptors was observed following NMDA receptor activation and calcium influx. Furthermore, NMDA receptor activation enhances phosphorylation of GIuRl by Cam(Strack and Colbran, 1998). These data suggest that binding of CaMKPI to
NMDA receptors will place the kinase close to a major source of calcium influx and in the proximity of AMPA receptors that can be phosphorylated and hctionally upregulated by CaMKII following
NMDA receptor activation. This cellular signdling circuit may constitute the beworkfor LTP generation.
1.2.2 Receptor traffrcking in regulation of synaptic transmission
As the modification of postsynaptic ionotropic receptors is one of the primary means for regulating the efficacy of synaptic transmission, understanding the mechanisms by which the properties and hence functions of the ionotropic receptors are modulated has been a subject of intensive research in the field of neuroscience. Traditionally, functional changes in ionotropic receptors have been thought to be mainly achieved by altering the properties of existing receptors through, for example, protein phosphorylation (Levitan, 1994; Raymond et al., 1993). However, an idea that has recently gained prominence is that rapid changes in the number of ionotropic receptors in the postsynaptic plasma membrane is also an efficient means of regulating the hctions of these receptors and hence synaptic efficacy (Malenka and Nicoll, 1997; Craig, 1998; Malinow, 1998; Wan et al., 1997). Most integral plasma membrane proteins are constitutively trafficked between the plasma membrane and the intracellular compartments via vesicle-mediated membrane fusion (insertion) and endocytosis
(internalization). Regulation of these processes has been shown to be an important means of controlling cell-surface expression, and hence function, of these proteins (Pessin et al., 1999; Karoor et d., 1998; Karoor et d., 1998). By analogy, it would not be unreasonable to speculate that the plasma membrane expression of ionotropic neurotransmitter receptors is also subject to a similar mode of regulation. Indeed, evidence is emerging to suggest that ionotropic receptors are subject to relatively rapid constitutive plasma membrane insertion and internalization, and that regdating either of these processes can lead to a rapid alteration in the number of the receptors expressed on the postsynaptic plasma membrane, and hence influence the synapticefficacy mediated by these receptors
(Carroll et al., 1999a; CarroIl et al., 1999b; Shi et al., 1999).
1.2.2A Background on plasma membrane protein trafficking
Plasma membrane proteins, including integral plasma membrane receptors (such as G protein-coupled and Ligand-gated receptors), are constitutively trafEcked between the plasma membrane and the intraceLldar compartments via vesicle-mediated membrane fusion (insertion) and endocytosis
(internalization). The current model of vesicle-mediated plasma membrane insertion is the SNARE hypothesis which is largely derived fiom studies on synaptic vesicle exocytosis. Based on this hypothesis, vesicle-mediated transport consists of several steps, schematically illustrated in Figure
1. I. Vesicles are formed at the level of the endoplasmic reticulum (ER) and the Golgi complex through complex coat protein-mediated membrane budding and fusion steps (Schekman and Orci,
1996). After being released fiom the Golgi, the secretory vesicles are trafficked to the plasma membrane, presumably through interactions with microtubules and microtubule-based motor proteins
(Vallee and Sheetz, 1996). Analogous to the established processes involved in synaptic vesicle
16 targeting and transmitter release, the secretory vesicles will then hewith the plasma membrane
through specific binding between proteins on the vesicle membrane (v-SNARES) and on the target
membrane (t-SNARES), as well as some cytosolic proteins such as NSF (N-ethylmaleimide-sensitive
factor) and SNAPS (soluble NSF-attachment proteins) (Fig 1.2A) (Rothman, 1994; Sudhof, 1995).
Although vesicles for delivering different membrane proteins and secretory cargoes may utilize
proteins that differ from those in synaptic vesicles for specific targeting, the basic processes are
believed to be the same or very similar. Accordingly, several neurotoxins which block synaptic
vesicle exocytosis by cleaving specific SNARE proteins could also be invaluable tools in studying
pIasma membrane receptor insertions. Tetanus toxin (TeTx) cleaves vesicle-associated membrane protein (VAMP)/synaptobrevin, and botulinm neurotoxin cleaves SNAP-25, syntaxin and VAMP.
Thus both toxins are capable of blocking vesicIe-mediated exocytosis (Sudhof, 1995).
On the other hand, most, if not all, plasma membrane proteins can then be removed fiom the pIasma membrane by clathrin-mediated constitutive and/or regulated endocytosis involving calthrin and several clathrin adaptor proteins. As illustrated in Figure 1.2B, clathrin-mediated endocytosis
involves a number of key steps (Goodman et al., 1998; Schmid, 1997). First, clathrin coats are recruited to the plasma membrane and linked to the receptors by adaptor proteins. For most G protein-coupled receptors, this is accomplished by a family of monomeric adaptor proteins known as p-arrestins (Goodman et d., 1998; Ferguson et al., 1998), and the process can be severely impaired by several dominant-negative p-arrestin mutants (Krupnick et al., 1997). However, most plasma membrane receptors other than G protein-coupled receptors, including the ligand-gated receptors(Nesterov et al., 1999), seem to be recruited to clathrin-coated pits by the adaptor protein Figure 1.2 A schematic diagram illustrating the common steps involved in plasma membrane insertion and endocytosis of integral plasma membrane receptors.
Following their synthesis in the ER [I] and Mermaturation in the Golgi complex [2], plasma membrane receptors are transported in membrane vesicles towards the target plasma membrane, presumably with the help of some cytoskeletal structures [3]. Through plasma membrane targeting and fusion mediated by SNARE protein-protein interactions [4,5], the receptors are finally inserted into the plasma membrane [a. Most, but not all, plasma membrane receptors can then be recycled back to the intracellular compartment via clathrin-coated-pit dependent endocytosis 181. The endocytosed receptors in the endosome [9] will then be either recycled back to the plasma membrane
PO] or targeted to the lysosome for degradation [l 11. Figure 1.2 A schematic diagram illustrating the common steps involved in the plasma membrane insertion and endocytosis of integral plasma membrane receptors.
plash membrane
mGolgi
I Receptor 0 Cytoskeleton n t-SNARE 7 V-SNARE - Unfolded receptor molecule Adaptor protein Figure 1.3 Mechanisms underlying membrane insertion and endocytosis of plasma membrane proteins
(A) SNARE protein components involved in exocytosis- Vesicle-mediated plasma membrane insertion is mediated by protein-protein interactions. SNARE proteins on the plasma membrane
(target or t-SNARE)inciuding SNAP-25 and syntaxin interact with SNARE proteins on the receptor- bearing cargo vesicles (vesicle or V-SNARE),mainly VAMP, and thus anchor the vesicle close to and fuse with plasma membrane with the help of some cytosolic SNARESsuch as NSF and SNAP.
(B) Plasma membrane receptor internalization. The intracelular part of a receptor binds to adaptor proteins [2,2]which will then recruit clathrin to form clathrin-coated pits [3,4]. A GTPase protein dynamin is brought to the complex through amphiphysin [3,4]to pinch off the pit to form vesicles r4,q Figure 1.3 Mechanisms underlying membrane insertion and endocytosis of plasma membrane proteins
Synaptotag
Synapsin
I Receptor ) Clathrin rAmphiphysin Adaptor proteins Dynarnin AP2, a heterotetrameric complex containing or/3,kq subunits (Schmid, 1997). It is thought that the p2 subunit of AP2 binds to the receptors via specific internalization sequences (e-g., the YXX+ motif) present in the carboxyl terminal region of the receptors (Schmid, 1997). However, recent studies have reported that p-arrestin may also interact with the P2 subunit of AP2, and it has been suggested that both P-arrestin and AP2 are required for the endocytosis of P2-adrenergic receptors (Laporte et al.,
1999; Laporte et al., 2000). Secondly, in addition to the binding of adaptor proteins, these adaptor proteins can also bind and recruit clathrin to the cargo-protein concentrated plasma membrane regions, thereby initiating the clathrin-coat assembly. Clathrin is organized as a complex consisting of three pairs of heavy and light chain proteins that can Merpolymerize into cages capable of capturing cargo protein-containing membrane during vesicle formation, leading to membrane invagination, a process that can be inhibited by hypertonic sucrose (Hansen et al., 1993a). Finally, the vesicles pinch-off via a dynamin GTPase-dependent process, a process can be blocked by either
GTP y s or GTPase-defective dominant negative mutant dynamin (Damke et al., 1994). Additionally, as dynamin is recruited to the clathrin-coated pits through the binding of its proline-rich region to the amphiphysin SH3 domain, the pinching off of clathrin-coated-pits can also be blocked by overexpression of the arnphiphysin-SH3 domain peptide, presumably due to its ability to competitively disrupt the binding between endogenous amphiphysin and dynamin (Herskovits et al.,
1993; Owen et al., 1998). The end- product of these processes, the cIathrin-coated vesicle, is internalized from the plasma membrane to the intracellular compartment. The internalized proteins can then either recycle back to plasma membrane through an endosome-dependent pathway or are targeted to lysosomes for degradation (Mellman, 1996; Schmid, 2 997).
It is becoming increasingly clear that both the plasma membrane insertion and internalization
22 of proteins are regulated processes and such regulation has been shown to be an important means of controlling cell-surface expression, and hence function, of these proteins (Pessin et al., 1999). Thus, understanding the mechanisms regulating receptor trafficking and plasma membrane expression of postsynaptic ligand-gated receptors, and their role in the control of synaptic efficacy has quickly become one of the major subjects of current neuroscience research. .
1.2.2B AMPA Receptor Trafficking and Synaptic Plasticity
In hippocampal CAI neurons, tetanic stimulation of the Schaffer collateral and commissural inputs elicits homosynaptic long-term potentiation (LTP) of AMPA receptor-mediated synaptic transmission, while stimulation of the same pathway at a low frequency induces homosynaptic long- term depression (LTD) in hippocampal CAI neurons (Bear and Malenka, 1994). Both LTP and LTD have been proposed as primary molecular and cellular substrates for the formation of learning and memory (Bear, 1999; Malenka and Nicoll, 1999). While it is generally accepted that, in their most commonly studied forms, the induction of both LTP and LTD is postsynaptic and depends upon Ca" influx through activated NMDA receptors, the mechanisms underlying their expression remain hotly debated, and likely involve both a presynaptic component via alteration of transmitter release and a postsynaptic one through the modification of AMPA receptors (Bliss and Collingridge, 1993;
Malinow, 1998; Malenka and Nicoll, 1999). In addition to the alteration of channel gating and conductance @erkach et d., 1 999; Greengard et al., 199 I), recent evidence suggests that the activity- dependent redistribution of MAreceptors to and fiom postsynaptic domains plays an important role in the postsynaptic modification of synaptic efficacy (Malenka and Nicoll, 1997; Malenka and
Nicoll, 1999). This has led to the "silent synapse" hypothesis to explain the postsynaptic expression of LTP and LTD (Durand et al., 1996; MaLenka and Nicoll, 1999; MaLinow, 1998). Silent synapses
contains hctional NMDA receptors but lack functional AMPA receptors, and as NMDA receptors
are inactive at normal resting potentials due to the voltage-dependent ~g2'blockade (Mayer et al.,
1984), these synapses are silent and will basically not be involved in synaptic transmissions.
However, silent synapses become activated during the induction of LTP by recruiting functional
AMPA receptors to the postsynaptic membrane (Durand et al., 1996; Isaac et al., 1995; Shi et al.,
1999). By extrapolation, an active synapse may be silenced during LTD by the loss of functional
AMPA receptors (Carroll et al., 1999b). A central question to this silent synapse hypothesis is bow synaptic AMPA receptors can be rapidly recruited to and removed fiom the postsynaptic sites, thereby resulting in switching of the synapses between active and silent states during the induction of LTP
and LTD. As discussed above, the cell-surface expression of plasma membrane proteins including
ligand-gated receptors can be rapidly altered by regulating their membrane insertion and
internalization. Therefore, regulated exocytosis and endocytosis of AMPA receptors offers an attractive mechanism for activating and silencing synapses, and hence for the expression of LTP and
LTD. Indeed, recent evidence suggests that facilitated postsynaptic membrane insertion and internalization of AMPA receptors underlies certain forms of activity-dependent redistributions of postsynaptic AMPA receptors and the expression of LTP/LTD
Using the fluorescent dye FM1-43, Maletic-Savatic et a1 first reported that there was a calcium-dependent dentritic exocytosis in cultured neurons and that this dentritic exocytosis was mediated by calcium/calmodulin-dependentprotein kinase I1 (Maletic-Savatic et al., 1996), a kinase that has been implicated in LTP (Malenka and Nicoll, 1999). Consistent with an involvement of membrane hion events in LTP, recent studies have shown that some putative inhibitors that are believed to block membrane hion at a number of different steps, when injected postsynaptically, impaired LTP (Lledo et al., 1998; Luscher et al., 1999) and when recombinant SNAP (which is required for membrane fusion) was introduced into the postsynaptic cells, enhanced AMPA receptor- mediated synaptic transmission and occluded the subsequent LTP (Lledo et al., 1998). While these data clearly implied that membrane fusion at the postsynaptic neuron is required for the production of LTP, it does not demonstrate directly the involvement of membrane fusion-dependent exocytosis of AMPA receptors during LTP. To demonstrate the dynamic trafficking of intracellular pools of
AMPA receptors during LTP, two recent studies (Hayashi et al., 2000; Shi et al., 1999) demonstrated that, in cultured hippocampal slices expressing green fluorescent protein (GFP) tagged GluRl subunits, the LTP-inducing protocol or co-expression of active CaMKIT caused a rapid translocation of GFP-GluRlfiom intracellular domain to the cell-sdace including synaptic sites. Interestingly,
LTP or CaMKII-induced receptor translocation was not blocked by mutation of the CaMKII phosphorylation site on the GFP-GluR1, but was blocked by mutating a putative PDZ domain interaction site, suggesting that although CaMKII is required for LTP, its function is probably not to directly phosphorylate GluRl itself. These studies provided the first evidence for the addition of postsynaptic AMPA receptors to the synaptic surface during LTP (Shi et al., 1999). However, those experimentd results are derived from studies using recombinant GFP-GluR1, and the trafficking of native AMPA receptors during LTP still needs to be investigated.
Another line of evidence for AMPA receptor trafficking and its role in synaptic plasticity comes from recent work showing that NSF specifically binds to and regulates the cell-surface expression of AMPA receptors (Nishimune et al., 1998; Song et al., 1998; Lin and Sheng, 1998).
NSF was reported to bind directly and specifically to a defined region on the C-termind domain of
the GluR2 subunit (Nishimune et al., 1998; Osten et al-, 1998). Viral expression of peptides
corresponding to the binding domain of GhR2 (pep2m) was found to remove most surface-expressed
GluR2-containing AMPA receptors (Noel et al., 1999). Furthermore, ihsingthe pep2m or anti-NSF
antibody into the postsynaptic neuron resulted in a rapid decrease in AMPA receptor-mediated
synaptic transmission which can occlude the electrostimulation-induced LTD in hippocampal CAI
neurons (Lin and Sheng, 1998; Nishirnue et al., 1998; Noel et al., 1999; Osten et al., 1998; Song et
al-, 1998). Because NSF is a crucial component involved in vesicIe membrane fixion, it is naturally
- assumed that the function ofNSF is to promote AMPA receptor membrane insertion. However, there
has been no direct evidence to date which demonstrates that NSF-GluR2 interaction affects the
plasma membrane insertion of AMFA receptors. The downregulation of AMPA receptor cell surface
expression, as well as the inhibition of synaptic transmission, could also be possibly due to an
enhancement of receptor endocytosis by disruption of the NSF-GIuR2 interaction without affecting
receptor insertion.
While evidence is accumulating to support an important role of receptor cycling in regulating
ligand-gated receptor function and hence synaptic plasticity, many critical questions remain. First,
while the current evidence is consistent with the mediation of Iigand-gated receptor recycling by
transport and recycling vesicle-mediated mechanisms, there have been no demonstration of the
presence of AMPA receptor-carrying vesicles in the postsynaptic domains. Second, while we now
know, &om pharmacological evidence using neurotoxins that block synaptic exocytosis or inhibitors of c1athri.n-mediated endocytosis, that mechanisms mediating recycling of Ligand-gated receptors may be very similar to those underlying the recycling of synaptic vesicles at the presynaptic terminals, we know little about how the presumed ligand-gated receptor carrying vesicles are specifically targeted to the postsynaptic density. Finally, mechanisms by which the insertion and endocytosis of ligand- gated receptors are regulated are still poorly understood. Consequently, we do not know how LTP or LTD inducing protocols that after ligand-gated receptor recycling result in a rapid change in the number of the receptors at the postsynaptic domains. Answering dl these questions will undoubtedly be facilitated by the identification and characterization of both the exocytotic transport vesicles and clathrin-coated endocytotic vesicles that cany the ligand-gated receptors. These studies should be considered priorities along with further studies of the physiological significance of intracellular trafficking and plasma membrane expression of ligand-gated receptors.
1.3 Glutamate receptor-interacting proteins
Neurotransmitter receptors are enriched at the postsynaptic site after their targeting to the plasma membrane. This is extremely important for synaptic transmission because neurotransmitters released from presynaptic terminal will only activate receptors on the postsynaptic membrane. Therefore a fimdarnental question is how receptors are concentrated on the postsynaptic site, rather than distribute evenly on the plasma membrane. Recent studies have indicated that the synaptic localization of neurotransmitter receptors is probably achieved through interactions of receptors with intracellular proteins in the postsynaptic domain. In these experiments, two techniques are commonly used: one is co-immunoprecipitation, by which one protein could be found in the immunoprecipitated receptor complexes, and in vitro pulldown assay wilI Merdetermine whether the interaction is direct. The second method frequently applied to fishing for new interacting proteins is the yeast two-hybrid system (Niethammer et al., 1996). Using a part of the glutamate receptors, usually the c-terminus, as a bait, any proteins interacting with glutamate receptors will be identified.
A large number of glutamate receptor-interacting proteins have been identified. Basically, glutamate receptor-interacting proteins can be categorized into two groups: proteins that only interact with NMDA receptors and those that only interact with AMPA receptors. According to their function, the glutamate receptor interacting protiens may also be divided into two groups: those acting as scaffolds to localize, anchor, or tracreceptors, such as PSD-95, GRIP,actinin, PICK1 (Dong et al., 1997; Gomperts, 1996; Xia et al., 1999); and those belonging to signalling molecules, such as tyrosine kinases Lyn and Src, synGAP (Hayashi et al., 1999; Kim et at., 1998; Yu et al., 1997).
1.3.1 NMDA receptor-interacting proteins
PSD-95 (or SAP90) was fust identified as an abundant cytoskeletal protein existing in the postsynaptic density (Cho et al., 1992;Kistner et al., 1993), and is a member of the membrane- associated guanylate kinase (MAGUK) family. Its developmental expression seems to parallel synaptogenesis. PSD-95 contains three PDZ domains, an SH3 domain and a guanylate kinase (GK)
-like domain, and each of these domains interact with different proteins. It was found that the second
PDZ domain in PSD-95 interacts with the seven amino acids in the c-terminus of NMDA receptor subunits (Kornau, et al. 1995). Kim et aI demonstrated that Shaker-type potassium channel subunits can also bind both the first and the second PDZ domains of PSD-95. They demonstrated that the last
28 four amino acids are critical for the binding (Kim, et al. 1995 ). PSD-95 was first proposed to be involved in the targeting and clustering NMDA receptors at the synaptic sites. However, mutation of the PSD-95 gene which abolished the expression of PSD-95 protein did not alter the synaptic localization ofNMDA receptors. Although this may reflect the functional redundancy of PSD-95 and other NMDA receptor-interacting proteins, an alternative explanation is that PSD-95 is not responsible for synaptic targeting NMDA receptors, but rather may possibly be involved in providing a platform for other protein-protein interactions and signallings. In line with this notion, PSD-95 mutant mice showed dramatically altered synaptic transmissions. LTP induced by stimulations with different frequencies were enhanced in the mutant mice, and even some low frequency-stimulations which normally inducing LTD induced LTP in the mutants. PSD-95 clusters bothNMDA receptors and K" channels. A consequence of clustering K" channels by PSD-95 is probably to regulate K' channel's trafficking. By using cell-surface biotinylation in HEK cells transfected with potassium channel KV1.4, it was found that the constitutive K+channel internalization was completely blocked when PSD-95 was co-transfected with KV1.4 (Jugloff et al., 2000), indicating a role of PSD-95 in stabilizing K' channels on the plasma membrane.
The PSD-95 family also contain several other proteins, including SAP97, PSD-93khapsyn-
110 and SAP Z 02 (Kennedy, 1997), which all share the same modules in their structure. SAP97 is found predominantly in axons and in glutamatergic terminals (Muller et al., 1995). The postsynaptic protein SAP102 coimmu11oprecipitates with NR1 subunits fi-om rat brain synaptosomes, and recombinant proteins containing the c-terminal of NMDA receptor NR2B sub~tinteract with all three PDZ domains of SAP 102 (Muller et al., 1996). All the members of the PSD-95 family proteins interact with NMDA receptors through PDZ domains. The PDZ domain is a 90 amino acid motif,
29 named after three proteins that contain this motif: PSD-95, Dlg (discs large), and 20-1.
Characterization of the interactions betweenNMDA receptor subunits, Shaker potassium channel and
PSD-95 found that NRl c-terminal T/SXV (where X stands for any amino acid) sequence is the common motif for PDZ domain binding (Kim et al., 1995 ). However, it was found that PDZ domains in other proteins have different binding specificities (Songyang et al., 1997).
In addition to the PDZ domain-containing proteins, several PDZ domain-lacking proteins are also found to interact with NMDA receptors, most of them are cytoskeletal components, such as actinin, spectrin and Yotiao. Actinin, an actin-binding protein enriched in the PSD, binds to the in.tracellular c-terminals of both NR1 and NlR2B (Wyszynski et al., 1997). As NMDA receptors can associate with F-actin through actinin, it is believed that this interaction may contribute to the anchoring of theNMDA receptor to the cytoskeleton. Indeed, one study found that depolymerization of F-actin and the disruption of actinin reduced the number of NRl clusters in cultured hippocampal neurons (AUison et al., 1998). Also, as the binding of actinin to NMDA receptors undergoes competition by Ca2+/calmodulin,the interplay between those two molecules is a possible mechanism involved in NMDA receptor desensitization. It was fomd that calcium influx through NMDA receptors depolyrnerizes actin and reduces NMDA receptor activity (Rosenmund and Westbrook,
1993) and this inactivation is due to the competitive displacement of actinin from NR1 subunits by
Ca2'//calrnodulin(Zhang et al., 1998a). In addition to actinin, three other cytoskeletal proteins were found to interact with NMDA receptors, mainly through NRl subunits. Spectrin is an actin-binding protein and was found to bind to NR1, NR2A and NR2B subunits. The interaction of NR2B and spectrin is inhibited by calcium and phoshorylation by Fyn kinase, and the association between NRl and spectrin is inhibited by calmodulin and PKA/PKC phosphorylation (Wechsler and Teichberg,
30 1998). However, the physiological significance of this interaction is still unknown. Yotiao, another cytoskeletal protein found In the PSD, associates with and colocalizes with NR1, and its function was originally unknown (Lin et al., 1998b) . However, it was recently demonstrated that Yotiao also binds to PKA and protein phosphatase 1, which contribute to the modulation of NMDA receptor activity
(Westphal et al., 1999), indicating that Yotiao may act to position the kinase and phosphatase nearby
NMDA receptors and regulate NMDA receptor function.
Association of tyrosine kinases with NMDA receptors probably has more importance in
NMDA receptor hctions, especially for synaptic piasticity. It was found that Src kinase associates with NMDA receptors and regulates the receptor activity (Yu et al., 2997; Yuand Salter, 1998; Yu and Salter, 1999), and this modulation seems play an important role in LTP (Lu et al., 1998). Another
Src family base, Fyn, can also interact with NMDA receptors and the association is up-redated during brain ischemia, and may therefore contribute to altered signal transduction in the postischemic hippocampus (Takagi et al., 1999)-
For all the NMDA receptor-interacting proteins, PSD-95 is unique in that it further associates with a variety of other proteins, making PSD-95 a major platform for anchoring different proteins, including some signalling molecules, to the proximity of NMDA receptors. For example, a synaptic
GTPase-activating protein, synGAP, interacts with all three PDZ domains of PSD-95 (Chen et al.,
1998; Kim et al., 1998). SynGAP is enriched in excitatory synapses and is co-localized with NMDA receptors. SynGAP stimulates the GTPase activity and thus can downregulate Ras under normal condition (Kim et al., 1998). Meanwhile, given that the zctivating ability of synGAP on Ras-GTPase is inhibited by CaMKII kinase, activation of NMDA receptors will remove the inhibition on Ras through activating CaMKII by Ca ", and thus allowing for downstream signding such as the MAP
kinase pathway to occur (Chen et al., 1998). Another enzyme associated with PSD-95 is neuronal
nitric acid synthase (nNOS). nNOS is a PDZ containing protein and interacts with PSD-95via PDZ-
PDZ interactions (Brenman etd., 1996)- Although the function of this association is not yet clear,
it is reasonable to postdate that the close localization maybe involved in NMDA receptor-dependent production of nitric oxide.
Besides proteins involved in signalling, PSD-95 also interacts with proteins which may act to localize it to synaptic sites. CRIl?T (cysteine-rich interactor of PDZ three), a small polypeptide that associates with rnicrotubules, binds to the third PDZ domain of PSD-95 (Niethammer et al., 1998), and thus Iinks PSD-95, and NMDA receptors indirectly, to the tubulin-based cytoskeleton. The third
PDZ domain of PSD-95 can also bind to neuroligin (Irie et al., 1997). This is especially interesting because neuroligin is a membrane-spanning cell adhesion molecuie that interacts with neurexin, another adhesion molecule sparuing the presynaptic membrane (Rao et al., 2000). This interaction will form intercellularjunctions and might play critical roles in co-ordinating the pre- and postynaptic alignment during synaptogenesis and synaptic localization of NMDA receptors.
The only protein found to date that interacts with the GK domain of PSD-95 is GKAP
(guanylate kinase-associated protein). GKAP is enriched in the PSD and can also interact with the
GK domains of SAP97, SAP 102and chapsynl10 (Kim et al-, 1997). However, its function remains to be elucidated.
Considering the large rnumbers of proteins with very diverse firnctions, which interact, directly or indirectly, with NMDA receptors, it is safe to say that an NMDA receptor is not just an ion channel, but an extremely complex communication base. Once NMDA receptors are activated,
a series of events will be initiated, simultaneously or sequentially, triggered by calcium influx or
conformational changes of NMDA receptors. Those events include changes in: ions concentration
and membrane potential; kinases and phosphatases activities and thus proteinphosphoryfation levels;
affinity of protein-protein interaction and reorganization of the infrastructure of the PSD. As such,
the information transduced into the cell is not through a single or several signalling pathways, but
through a huge signalling wave, spreading forward from NMDA channels. The subsequent
responses of aneuron will dependent on the overall integration of all the information, or the signalling pattern, following NMDA receptor activation.
1.3.2 AMPA receptor-interacting proteins
AMPA receptors are responsible for the majority of the fast excitatory synaptic transmissions in the mammalian CNS. AMPA receptors are heterotetramers or heteropentamers formed from four subunits, GluRl -GluR4, where each subunit has three transmembrane domains and an extracellular
N-terminus and an intracellular C-terminus.
The first AMPA receptor-interacting protein, GRIP, was identified by two-hybrid system in
1997 (Dong et al., 1997; Ye et al., 2000). GRIP is a 1,112 amino-acid protein containing seven PDZ domains- Unlike PSD-95 family proteins, GRIP shows no other known domains. GRIP coimmunoprecipitates with AMPA receptor subunits and colocalizes with AMPA receptors. In vitro pull down assays demonstrated that the binding is between the fourth and fifth PDZ domains of GRIP and the last seven amino acids of the c-terminal on GfuR2 or GluR.3 subunits. It was shown that
33 overexpression of the c-terminus of GluR2, but not that of GluRl, in cultured neurons disrupted the clustering of AMPA receptors. However, as it was later found that the GluR2 c-terminus also binds to several other proteins in addition to GRIP, the disruption experiments are not definitive for assigning GRXP the AMPA receptor-clustering fhction, and thus the exact role of GRIP has yet to be determined.
ABP (AMPA receptor-binding protein) is a GRIP-related protein which shares 64-93% homology to GRIP in their PDZ domains (Srivastavaet al., 1998). ABP has two isoforms: one is 130 kD containing seven PDZ domains (which is also called GRIP2), and the other short form contains six PDZ domains. ABP is enriched in synaptosomes, especially in the PSD, and can dimerize with itself or with GRIP, but not with PSD-95, through PDZ-PDZ interactions. However, coexpression of GBP and GluRs failed to show AMPA receptor clustering, suggesting that AEiP/GRIP may form scaffolds for localizing other proteins close to AMPA receptors rather than directly anchoring and clustering the receptors. This is supported by the finding showing that GRIP and ABP associate with
Ephrin receptors (Eph) and their ligands (Ephrin) (Torres et al., 1998). The Ephrin system is an important signalling pathway mediated by tyrosine phosphorylation, and the interaction of GRIP/ABP with Ephrin/Eph will bring AMPA receptors close to those signalling molecules, which may be involved in synaptic information transduction after AMPA receptor activation.
Another AMPA receptor-interacing protein is PICKl (protein interacting with C-kinase; Xia et al., 1999). PICKl is also a PDZ domain-containing protein, which, in addition to binding to the c- termini of GluR2/3/4, also binds to PKC (Staudinger et al., 1995). Unlike GRIP and ABP, PICKl is not neuronal specific, but ubiquitously distributed in different tissues. In the CNS, PICKl is enriched in the PSD and coIocalizes with AMPA receptors and PKC. Also unique to AMPA receptor-interacting proteins, PICK1, when coexpressed with GlWin HEK cells, clusters the AMPA receptor subunits, and this effect can be blocked by deleting the last seven stmino acids on GluR2 c- terminus or by a mutation of a PDZ domain in PICKZ, indicating a specific interaction is required for the receptor clustering @Ciaet al., 1999).
Interestingly, AMPA receptors also interact with signalling molecules directly. Lyn, a Src family nonreceptor tyrosine kinase, interacts with GIuR2/3 through its SH3 domain. Lyn is activated by AMPA treatment and is required for AMPA receptor activation-dependent stimulation of MAP kinase and BDNF gene expression (Hayashi et al., 1999). The activation of Lyn does not requires calcium or sodium influx and is suggested to be activated through conformational changes of the receptors. This is in line with another finding showing that the AMPA receptor activation-induced association oWAreceptors with Gaiis also independent of calcium and sodium influx (Wang et d., 1997).
In 1998, a surprising finding fiom several groups demonstrated that, NSF, a protein involved in membrane fusion, interacts with AMPA receptor GluR2 subunits (Lin and Sheng, 1998; Nishirnune et al., 1998; Osten et al., 1998; Song et al., 1998). The NSF interaction site in the GluR c-terminus is not within the last amino acids, but a site more proximal to the plasma membrane. While it is well known that NSF is crucial in presynaptic sites for neurotransmitter release, it has been also shown to be enriched in postsynaptic fractions (Walsh and Kuruc, 1992). By immunostaining, several groups demonstrated colocalization of NSF and AMPA receptors in cultured neurons. However, as NSF is localized at the presynaptic site, the localization can not provide any real support for an in vivo interaction. Nevertheless, intracellular application of a peptide encoding the NSF-interacting region
of GluR2, presumably to block NSF-AMPA receptor interactions, inhibited the AhPA receptor- mediated EPSCs in cultured hippocampal neurons (Song et al., 1998). In addition, application of the peptide (pep2m) which disrupts the NSF-AMPA interaction in hippocampal slices induces LTD and occludes eletrostimulation-induced LTD (Luthi et al., 1999). This pep2m-induced LTD is likely due to the removd of AMPA receptors from the postsynaptic sites, because overexpressionof the peptide
(pep2m) in cultured neurons dramatically reduced the cell surface expression of AMPA receptors without changing either the surface expression of NMDA receptors or the total amount of AMPA receptors (Noel et al., 1999). These studies provide strong evidence indicating that receptor trafficking, or receptor removal from the plasma membrane in this case, is the underlying mechanism for long-term synaptic plasticities.
1.4 phosphatidylinositol-3' kinase (PI3-kinase)
Functioning of the nervous system depends on chemical communications, achieved by binding of extracellularmolecules, such as neurotransmitters and growth factors, to their receptors on the plasma membrane and activating downsbeam signaling pathways. The retease of these molecules to the extracellular space and the cellular responses elicited by them all involve integrated molecular events that are membrane based. Due to the importance of the membrane in transferring signals fiom external to intraceUular domains or between different compartments inside a cell, research has focused on the identification and characterizationof key proteins that fimction within or on membrane leaflets. However, the membrane itself is usually treated as an inert platform upon which proteins interact. Recent studies have revealed that Lipid components, especially phosphorylated lipids, in the
membrane are actually highly dynamic. They hction not only as precursors for intracellular
messengers but also directly regulate cellular processes such as protein tracking and kinase
activities.
There are several different types of phospholipids in the membrane, such as
phosphatidylcholine (PC), phosphatidylethanolarnine (PE), phosphatidylserine (PS),
phosphatidylinositol (PI) and sphingomyelin (SM). PI is a relatively abundant lipid, accounting for
4% of the total lipid in the plasma membrane. PI is very important because it is not only a structural
component in the membrane, but also a crucial precursor for a variety of signalling molecules. The
inositol headgroup of PI contains five hydroxyls and those at position 3, 4 and 5 can be
phosphorylated by PI 3-kinase (PI3K), PI 4-kinase (PI4K) and PI 5-base (PISK), respectively
(Frurnan et al., 1998; Harder et d.,1998; Martin, 1997). Using PI as a substrate, these lipid kinases produce a variety of lipid products, such as PI(3)P, PI(4)P, PI(3,4)P,, PI(3,5)P2, PI(4,5)P2 and
PI(3,4,5)P3- Among these lipid kinases, PI3K is especially important and has been the most extensively studied. PI3K, through its lipid products and the wide range of downstream signalling pathways, participates in a myriad of cellular processes including cell growth and transformation, differentiation, motility, cell survival (anti-apoptosis) and protein trafEicking (Alessi and Downes,
1998; Blackstone et al., 1989; Corvera and Czech, 1998; De Camilli et al., 1996; Derman et al., 1997;
Fnunan et al., 1998; Jones and Howell, 1997).
1.4.1 Classification of PI3-kinase PI3K phosphorylates the inositol ring of PI at the 3 position and form different phosphoinositides-
Its activity can be specifically blocked by two cell-permeabk PI3K inhibitors, wortmannin and
LY294002 and antagonized by the lipid phophatase PTEN (Vazquez and Sellers, 2000). There are multiple forms of PI3Ks which can be divided into three classes, class I, class 11 and class III.
Class I PI3fi: were the first to be purified and characterized (Hiles et al., 1992). They are heterodimers consisting of a 11 OkD catalytic domain (pl10) and a reedatory domain (p85), and use
PI(4)P and P1(4,5)P2 as substrates to produce PI(3,4)P2 and PI(3,4,5)P3.
Class I PI3P can be fuaher divided into two subclasses, class IA and class IB. Class IA P13K is the first characterized PI3K which is composed of a pl10 and a 85kD regulatory domain @85).
There are three p 110 isoforms, i-e., P 11 Oa, p 1 10P and p 110 6,encoded by three different genes. Each p 1 10 subunit contains a base core domain, a p85 binding domain and a ras-binding domain. For the p85 regulatory domain, seven isoforms have been identified so far, which are derived from three genes (p85a,p859 and p85y) by alternative splicing. A p85 subunit contains several modules for protein-protein interactions, including a SH3 domain, two SH2 domains and a proline-rich domain.
Class I PI3Ks are normally activated by tyrosine-phosphorylated proteins by binding to them via the
SH2 domains of PI3Ks (Leevers et al., 1999; Alessi and Downes, 1998). As such, any factors or signalling pathways which Iead to tyrosine phosphorylation, such as growth factors or Src kinases, may activate PI3K. The class IB contains only one PI3K, a dimer consisting of a p l10y catalytic domain and a 101 kD (p 10 1) regulatory domain. It has limited tissue distribution and is abundant only in white blood cells. Different from that for class 1A, the p 110/p 10 1 PDK is activated by the G, complex of G proteins . CiasslTPI3fi: there are three forms of class lI PI3Ks, PI3K-C2q PI3K-C2P and PI3K-C2y.
They are characterized by a greater than l7Ok.D catalytic subunit which lacks ras-binding domain but has an extra C2 domain, which can bind to phospholipids independently of CaZ+. It is not clear whether class U PI3 Ks have regulatory subunits. Class I1 PI3 Ks can be activated by insulin, epidermal growth factor, platelet-derived growth factor and integrin (Bachelot et al., 1996; Alessi and Downes,
1998; Berridge, 1986; Mosthafet al., 1996), using PI, PI(4)P and P1(4,5)P2 as substrates. However, because the regulatory subunits are unknown, wbether they use the same mechanism as that for class
I PI3K for activation is not clear yet.
CCass iIIPI31Ys: are the honologues of the yeast Vps34p. Unlike PI3Ks in class T and class
II, which can phosphorylate PI, PI4P and P14,5Pz, class 111 PI3K can only phosphorylate PIS
(Carpenter and Cantley, 1996a). Their regulation is not clear, but appears to be constitutively active.
1.4.2 Pf3K activity and its downstream signalling pathways
The activities of P85/pllO PI3Ks are regulated in different ways. First, binding of the SH2 domain of PI3K enhances kinase activity. For example, in the insdin receptor signalling process, autophosphorylation and activation of insulin receptors will recruit insulin receptor substrate 1 (IRS 1) to the pLasma membrane via protein-protein interactions, followed by tyrosine phosphorylation of
IRS 1(Haring et d., 1996). The phosphorylated tyrosine residues in IRS 1 then act as docking sites for
SH2 domain-containing proteins such as P13K (Freund et al., 1995;Alessi and Downes, 1998), where further downstream signalling following insulin receptor activation can be initiated. Second, P13K can be activated by Ras or G protein py subunits (Kodaki et al., 1994; Rodriguez-Viciana et al.,
1996). Third, it was recently found that the p85 regulatory subunit can directly modulate the
function of the pl10 catalytic domain by heterodimerization. The binding with p85 decreases pl 10
activity by 80% in in viho experiments which also indicates that the phosphotyrosyl peptides-induced
activation of PIXis due to a transition from the inhibited to the disinhibited state (Yu et al., 1998a).
In addition, the dimerization of p85 with p 110 increases the resistance of p 1 10 in high temperature.
Recombinant p 110 monomers lose activity rapidly when incubated at 3 7 OC, whereas p8Yp 110 dimers are stable and show a much longer half-life in mammalian cells (Yu et al., 1998a).
Interestingly, PIXactivity can be regulated by Ca"/calmodulin. It was found that Caz+/calmodulin
associates with the SH2 domain of the PI3K p85 subunit and significantly enhances PI3K activity
both in viho and in intact cells (Joyal et al., 1997). Since Ca"/calmodulin signalling is crucial for a variety of neuronal fiuzctions, especially in LTP expression, the finding that CaZ+/calmodulin modulates PI3K activity bridges the two fundamental signalling pathways and may implicate an involvement of PIXin the regulation of synaptic fimction and neuronal plasticity.
As a lipid kinase, PI3Kaccomplishes its multipIe cellular functions through its lipid products.
PI3P and P13,4P2, produced by PUK, can be specifically recognized by two lipid binding domains,
FYVE domain (for PI3P only) (Gaullier et al., 1999; Stenmark and Aasland, 1999; Wurmser et al.,
1999) and the PH domain (for both P13P and PI3,4P2) (Antonetti et al., 1996; Bachelot et al,, 1996).
Proteins which contain either domain will be recruited to the plasma membrane following PI3K activation and phosphorylation of its substrate. Among the variety of FYVE or PH domain- containing proteins, the serine/threonine base AktRKB is the best studied and probably the most important molecule in mediating PI3K downstream signalling. It is known that Akt can be activated
40 after its translocation to the plasma membrane through its PH domain following PI3K activation
(Banfi et al., 1998;BeUacosaet al., 1998), but mechanisms for Akt activation was not clear until PDK
(phosphoinositide-dependentkinase) was identified recently (Chou et d., 1998; Toker and Newton,
2000). PDK, a PH domain-containing kinase, after being recruited to the plasma membrane by PI3K lipid products, phosphorylates Akt at Thr308 and Ser473 to make Akt fidly activated (Chou et al.,
1998). After its activation, Akt will then leave the plasma membrane to phosphorylate the intracellular substrates and initiate a wide range of cellular responses (Kandel and Hay, 1999).
In addition to Akt, several other mediators for PI3 K signallingare also found, including PLCy,
PKC, MAPK and ribosomal S6-kinase @70S6K) (Kandel and Hay, 1999).
1.4.3 P13K regulates protein trafficking and neuronai function
The involvement of PI3K in protein trafficking was first indicated when Vps34p, first and only identified PI3K in yeast, was found to be essential for the traf5cking of newly formed proteins f?om
Golgi to the vacuole, the equivalent ofthe mammalian lysosome @e Camilli et al., 1996; De Carnilli and Takei, 1996). Consistently, in mammalian cells inhibition of PIXby wortmannin blocked the trafficking of lysosomal enzymes fiom Golgi to lysosome (Brown et al., 1995). This effect is not limited only to newly synthesized proteins, but existing integral membrane proteins also seem regulated by PI3K on their lysosomal targeting. It was demonstrated that wortmannin treatment, or mutation of PDGF receptors to abolish their binding with PI3K, blocked the Iysosomal degradation of PDGF receptors following endocytosis (Joly et al., 1995). In a study involving the receptor tyrosine kioase c-Kit, blocking PI3K activity did not dramatically inhibit receptor internalization, but rather prohibited the internalized receptors fiom moving far away from the plasma membrane
(Gomerman et al., 1997).
Except for regulating the intracellular targeting in protein trafficking by PI3K, increasing amount of data demonstrated that PI3Ks are also crucial in plasma membrane protein endocytosis or exocytosis. In CHO cells, the PI3K inhibitor wortmannin decreased cell-surface transferrin receptor by 50%, mainly due to an increase in receptor internalization, but slowing down the recycling was also a reason contributing to the down-regulation of transferrin surface expression (Martys et al.,
1996). Transporters, another kind of plasma membrane protein, are also regulated by PI3Ks in terms of their cellular distribution. The best example is probably the glucose transporter 4 (GLUT4).
GLUT4 is normally localized in intracellular compartments and insulin treatment can induce a rapid translocation of GLUT4 fiom cytosolic compartment to the plasma membrane (Holman and
Cushman, 1994;Pessin et al., 1999). This process is PDK dependent because wortmannin or a dominant negative p85 subunit of PI3K can block the plasma membrane recruitment of GLUT4
(Pessin et al., 1999; Yang et al., 1996). Interestingly, PI3K also regulates the trafficking of the neuronal transporter excitatory amino acids carrier-1 (EAAC- 1). Using a combination of cell surface biotinyiation combined with western blotting, it was found that PDGF increased cell surface expression of EAAC- I without changing the total amount of the transporter. This translocation was completely blocked by PI3K inhibitors wortmannin and LY294002, indicating an involvement of
PI3K in this PDGF-induced translocation of EAAC-1 (Sims et al., 2000). Due to the fact that EAAC-
1 is a neuron& glutamate transporter, this finding suggest that, through altering EAACl efficacy,
PI3K may play a role in regulation of excitatory synaptic transmission.
42 Indeed, there is an increasing amount of evidence that indicates the importance of PI3 K in the regulation of neuronal hction. First, PI3K or Akt kinase activities can be altered by glutamate receptors. Zhang et al (1998b) found that in cerebellar granule neurons, activation of NMDA receptors enhanced PI3K activity and thus protected neurons fiom cell death. However, using the same culture neuron preparation, another group showed an inhibition of Akt activity by glutamate
(Chalecka-Franaszekand Chuang, 1999). This discrepancy may be due to that all different glutamate receptors were activated in the latter case while only NMDA receptors were specifically stimulated in the first one. Furthermore, PI3K and its downstream kinase Akt can modulate K" channel and Ca"' channels, respectively (Blair and Marshall, 1997) marvey et al., 2000), and the Akt dependent activation of Caz+channelwas required for neuronal survival (Blair et al., 1999). Moreover, recent studies showed that PI3K can directly bind to NMDA receptors in a tyrosine phosphorylation- dependent manner (Hisatsune et al., 1999). PI3 K binds to tyrosine phosphorylated NR2B subunit through its SH2 domain, indicating that PI3K may directly modulate NMDA channel function or be involved in mediating glutarnate receptor-dependent cellular functions. Consistently, in cultured striatal neurons, Ca2f-permeableAMPA receptor-induced activation of MAPK cascade was blocked by inhibiting PI3K activity (Perkinton et al., 1999). Also, PI3K was recently found to be involved in LTP expression in dentate gyms, via a presynaptic mechanism (Kelly and Lynch, 2000), which strongly indicated that PIXmay be a critical molecule in the expression of newonal plasticity. Since
PIXis a major molecule that participates in protein trafficking and seems to have important interplay with glutamate receptors, fiu-ther investigation of PIXeffects on fast excitatory synaptic transmission may deepen our understanding in the cellular mechanisms of synaptic regulation and neuronal plasticities. 1.5 Working hypothesis and objectives
Evidence fiom recent studies (see Section 1.2.2) regarding the localization, redistribution and
trafficking of neurotransmitter receptors allows us to formulate a working hypothesis concerning the
regulation of synaptic transmission through modulation of receptor intracellular trafficking. Like
most plasma membrane proteins, ligand-gated receptors, after their synthesis and oligomerization in
the ER, are constitutively recycled between the plasma membrane and intracellular compartments via
membrane insertion and internalization. As both processes can be rapidly regulated, the number of
ligand-gated receptors at the cell surface can be controlled dynamically. Such regulation may be a
common and efficient means of modulating the hction of ligand-gated receptors and hence
generating synaptic plasticity.
In the central nervous system, AMPA receptors arc the mediator for most of the fast excitatory
synaptic transmission; therefore, my work will focus on AMPA receptor trafficking, its underlying
mechanisms as well as its physiological significance in the context of synaptic transrnission and synaptic plasticity. The main objectives include:
1. Whether AWA receptor internalize and if so, what are the cellular mechanisms for receptor
internalization? It is known that for plasma membrane receptor internalization, the most commonly used pathway is via clathrin-coated-pit (Bonifacino et al., 1996). The intracellular part of plasma membrane receptors associates with clathrin via some adaptor proteins. The local plasma membrane together with the receptors fold inwardly to form a clathrin-coated pit and finally internalizes into the cytosol in the form of clathrin-coated vesicles. I postulate that AMPA receptor may use the same or
similar mechanism for endocytosis. 2. Does AMPAreceptor internalization play a role in modulation of excitatory synaptictrammission?
This depends on whether receptor internalizationreduces the receptor number on the cell surface and whether the reduction of cell surface receptors happens at the postsynaptic domain. The stable amount of receptors expressed on the plasma membrane is normally maintained by a balance between receptor internalization and receptor plasma membrane insertion. If facilitated receptor internalization is balanced by enhanced receptor insertion, there will be no net change in cell surface- receptor number. Also, receptors on the plasma membrane localize on both synaptic and non-synaptic sites. If only the non-synaptic receptors are internalized, the strength of synaptic transmission may not be affected.
3. Does LTP express through AMPA receptor plasma membrane insertion? Mechanisms for LTP have been under extensive investigation for decades and recent evidence indicates that exocytosis is involved in LTP (Isaac et al., 1995; Shi et al., 1999; Lledo et al-, 1998; Luscher et al., 1999).
Although AMPA receptor translocation to the plasma membrane has been observed by an LTP induction protocol (Shi et al., 1999; Hayashi ea al., 2000), convincing evidence to directly demonstrate native AMPA receptor insertion in LTP expression is still lacking. For this purpose, specific immunostaining method should be developed in order to directly visualize the redistribution/plasmamembrane insertion of native AMPA receptors. To have functional significance of receptor exocytosis in terms of synaptic transmission, the recruited AMPA receptors on the cell surface are expected to be localized at synaptic sites. Experiments to determine the synaptic localization (synaptichon-synaptic) of newly inserted AMPA receptors will be carried out.
4. What is the mechanism underlying AMPA receptor insertion in LTP? Proteins including receptors are transported to the plasma membrane through vesicle-mediated membrane fusion, which is
accomplished via SNARE protein interactions, i.e., proteins on the receptor-bearing vesicles (v-
SNARE) interact with proteins on the plasma membrane (t-SNARE) (Bonifacino et al., 1996;
Schekman and Orci, 1996; Schmid, 1997). Experiments will be performed to determine whether
AMPA receptors are recruited to the cell surface via a SNARE-dependent mechanism. Furthermore,
the signaling pathway leading to AMPA receptor redistribution following LTP-inducing stimulation
will be investigated.
By completion of this work, I expect to establish the notion that AMPA receptors exist on the pIasma membrane (or the postsynaptic site) with a high degree of mobility. Any factors (chemical, electrical) which interfere with the dynamics ofthe receptor tr&cking processes will change the cell
surface-receptor number, and thus alter the strength of excitatory synaptic transmission. This probably is a general mechanism in the regulation of synaptic transmission, including long-term synaptic plasticity, such as LTP and LTD. CHAPTER TWO
Experimental Methods 2.1 cDNA plasmids and new constructs
Rat GluRl and GluR2 cDNAs were amplified by PCR and cloned into the XbaVEcoRI and
HindIIIISalT sites, respectively, of the mammalian expression vector GW1 (British Biotechnology).
For extracellular HA-tagging of GluRl and GluR2 subunits, site-directed mutagenesis was performed to insert an AscI restriction site after A374 in GluRl and after G384 in GluEC2, An oligonucleotide cassette encoding the HA epitope was inserted into these AscI restriction sites. For exchanges of the carboxyl tails between GluRl and GluR2 subunits, EcoRl restriction sites were generated by silent mutation in the carboxyl termini of HA-GluRl and HA-GluR.2 immediately after the fourth transmembrane domain using site-directed mutagenesis. For the construction of GluRl,, the HA-
GluRl plasmid was digested with EcoRI to remove its cytoplasmic tail and the resulting plasmid was blunt-ended using the Klenow enzyme and dephosphorylated to prevent self-ligation. The carboxyl- terminal tail of the GluR2 was amplified using PCR, and then phosphory lated using T4 Kinase and ligated into the backbone of GluR1. A similar strategy was used to construct the chimeric GluR2,,.
Both chimeras were confirmed by sequencing. Rat GABA, receptor subunit cDNAs have been described previously wan et al., 1997). Wild-type GluR1 and GluR2 cDNAs were expressed in pCIS2 pIasmid vectors. pcDNA3 expression vectors containing HA-tagged wild-type and K44E mutant dynamin I cDNAs were gifts of Dr. R.B. Vallee (Worcester Institute, Shrewsbury, MA).
2.2 Transient transfection of HEK 293 cells
HEK 293 cells were transfected in 10 cm dishes at 80% confluency using the Cap-phosphate precipitation method. 36 to 48 hours after transfection, cells were washed with extracellular recording solution (ECS in mM : NaCl 140, CaCl, 1.3, KC1 5.4, HEPES 25, glucose 33; pH7.4, 320 mOsm) and incubated in ECS (serum starvation) for at least one hour, For insulin treatment, cells were incubated in ECS supplemented with 0.5 pM human recombinant insulin (Sigma) for 10-1 5 min and then processed for immunocytochemistry or colorimetric assays as described below.
2.3 Primary cultures of hippocampal neurons: Pooled cerebral hemispheres were collected in
Hank's Balanced Salt Solution from embryonic Wistar rats (E 18 to El9; Charles River). The entire hippocampus was micro-dissected away fiom cortical tissue, collected and then dissociated using both chemical and mechanical means (l3rewer et al. 1993). Cell suspensions were grown in Neurobasal media supplemented with B27 in either 12 or 6-well chambers (on coverslips for immunostaining).
Cells were processed for either imrnunofluorescence or colorimetric assays following 2-3 weeks of growth and maintenance in vitro as described below.
2.4 Immunostaining
2.4.1 Selective labeling of cell-surface vs total AMPA receptors and internalization staining:
HEK 293 cells were plated onto collagen-coated 22mm glass cover slips set in standard 35mm culture dishes, and transfected with 2 pg of each plasmid described above. For cell-surface receptor expression assays, cells at 48 h post-transkction were fixed with 4% pardormaldehyde in PBS for
10 min. Surface AMPA receptors were fist labelled with monoclonal anti-HA antibody (1 :5000,
Babco, Berkeley, CA) and visualized with an FITC-conjugated anti-mouse antibody (1 500,Sigma).
Following permeabilization of the cells with 0.25% Triton X-100 in PBS for 10 rnin, total cellular
AMPA receptors were then stained with polyclonal anti-GluR1 (Cedarlane Laboratories) or GLuR2
(Chemicon) primary antibodies, and Cy3-labelled anti-rabbit secondary antibodies (Fig 2.1). As a control for non-permeant staining, antibody raised against the GluR2/3 cytoplasmic terminus (Anti- GluR2/3, PharMingen) was used as the primary antibody in some experiments, and this antibody produced spec5c staining under permeant, but not non-permeant, conditions. For the surface AMPA receptor internalization assay, HEK 293 cells transfected with HA-tagged GluR2 constructs were incubated live at 4°C with 10 pg/mL monoclonal anti-HA antibody fmr one hour to label surface
AMPA receptors. Cells were then incubated at 37OC in ECS for various time periods to allow for constitutive internalization of labelled receptors. Following a 10 minute fixation with 4% paraformaldehyde without permeation, receptors remaining on the cell surface were stained with a
FITC-conjugated anti-mouse antibodies. In order to visualize internalized receptors, cells were subsequently treated for 1 rnin with 100% methanol and stained with a Cy3-conjugated anti-mouse antibodies. Colocalization of HA-tagged GluR2 receptors with EPS 15 was studied by staining first with monoclonal anti-HA and FITC-conjugated anti-mouse antibodies, and sequentially with polyclonal anti-EPSIS and Cy3-conjugated anti-rabbit antibodies under permeant conditions.
Subcellular localization of fluorescently-labelled receptors was exarnked with a Leica TCS-4D confocal microscope.
2.4.2 Double staining of AMPA receptor and synaptophysin or NMDAR1: Cultured hippocampal neurons were treated with glycine (200 pM) in the bathing solution described above for 3 min and then transferred to this solution without any added glycine for 15-20 rnin. Cells were then fixed with
2% paraformddehyde for 10 min. Neurons were fist labeled with a polydonal antibody against the
N-terminal extracellular domain of the rat GluRl receptor (Oncogene Research, 1: 100) and a Cy3- conjugated anti-rabbit secondary antibody (1 :300) under non-permeant canditions. Cells were then stained with monoclonal anti-synaptophysin (Chemicon, 1 pgM) and FITC-conjugated secondary antibodies (1:300) following permeation of the cells with 0.25% Triton X-100 in PBS for 10 min.
For colocalization of cell-surface GluR2 and NMDARI, cells were sequentially stained with goat
anti-GluR2 primary antibody (Santa Cruz, 1:200) ,FITC-conjugated anti-goat secondary antibody
(1:300), polyclonal antibody zgainst the N-terminal extracellular domain of the rat NMDARl
(Sigma, 1: 1 00) and Cy3-conjugated anti-rabbit secondary antibody (1 :3 00) under non-permeant
conditions.
2.4.3 Pre-bIocking immunostaining (staining for AMPA receptor insertion): Cells were first incubated with the polyclonal anti-GluR1 antibody for 1 hour and a cold (non-conjugated) secondary antibody for another 30 rnin at 4°C. Following treatment with or without glycine (200 fl,3 min) at room temperature, the cells were fixed at diEerent time points as indicated in the text, then stained with the same anti-GluR1 primary antibody and a Cy3-conjugated anti-rabbit secondary antibody to detect newly inserted AMPA receptors on the plasma membrane under non-permeant conditions.
The synaptophysin was then stained with the monoclonal anti-synaptophysin antibody and the FITC- conjugated anti-mouse secondary antibody under permeant conditions. In some experiments, a glass micropipette and Eppendorf microinjection system were used following the pre-blocking step to inject hippocampal neurons with extracellular recording solution supplemented with TeTx light chain (200 nM) and Lucifer Yellow (0 -5%). Glycine treatment was then performed one hour after the injection.
Optical images were collected by confocal scanning with dual channels for Cy3 and FITC fluorescence, simultaneously. For quantification of receptor clusters and colocalization, pseudo-color density mapping and profile plots of randomly selected areas fiom at least 10 individual neurons fiom more than two independent cultures were generated using Scion ImagePC software, and receptor Figure 2.1 Selective detection of cell-surface and internalized AMPA receptors by
immunostaining or colorimetric assay
HEK cells transfected with HA-tagged AMPA receptor subunits (A) were pre-labelled with an anti-
HA antibody at 4OC (to block endocytosis) and then transferred to 37OC to allow internalization to
occur. Immunostaining was performed using FITC (green) or Cy3-conjugated (red) secondary antibodies under non-permeant or permeant condition, respectively, to label the cell-surface or the
internalized AMPA receptors (imrnunostaining). Or after the pre-labeling, an HRP-conjugated
secondary antibody as well as KRP substrates were used to measure the relative amount of
surface/total pre-labelled AMPA receptors (colorimetric assay). Figure 2.1 Selective detection of cell-surface and internalized AMPA receptors by immunostaining or colorimetric assay
Prelabelling with HA epitope anti-HA antibody I at 4°C
Or Permeant
+HRP substrate lrnmunostaining Colorimetric assay clusters were defined by fluorescent peaks above the threshold level.
2.5 Immunoprecipitation and western blotting
2.5.1 Co-immunoprecipitation ofGlW with adaptin P2: Hippocampal slices (300 pm thickness) were prepared fiom adult male Sprague-Dawley rats (150-200g, Charles River) as previously described (Wan et al., 1997). For irnmunoprecipitation and immunoblotting, homogenate
(approximately 500 pg protein) from control slices or slices treated with 0.5 pM indin for 10 min was incubated with anti432 adaptin monoclonal antibody (Sigma) in 500 ml of SO rnM Tris-HCI, I50 mM MI,0.1 % Triton X- 100 for 4 hrs at 4"C. The antibody-protein complexes were then pelleted with Protein A-Sepharose beads. Proteins eluted from the beads were subjected to SDS-PAGE and immunoblotting for anti-P2 adaptin (1 :5000), GluR2/4 (Pharmingen 1 pg/mL), GluR 1 (Oncogene
Science, 1 pg/mL) or monoclonal anti-NMDAR1 (Pharmingen; 1:800), respectively. For sequential probing of the same membrane for P2 adaptin and GluR2/4, GluRl or NMDARI, the membranes were stripped of antibody and reprobed. Blots were developed using the Enhanced
Chemiluminescence (ECL) western blot detection system (Arnersharn). Band intensities were quantified using Scion ImagePC software.
2.5.2 Co-immunoprecipitation of GluR2 with PI3K: Hippocampal slices (300 pm thickness) were treated with 50 pMNMDA, 50 pM AMPA, 50 pM MK-80 l(NMDA receptor open channel blocker), or 300 pM glycine in the presence of glycine receptor antagonist strychnine (10 pM), respectively, following a 1.5 hour recovery in ACSF at 35OC. Brain slice homogenate (500 pg) was incubated with anti-P85 or anti-GluR2 antibodies in 500 yl RIPA buffer (50 mM Tris-HC1, 150 mM NaCl, 0-1 % Triton X-100) for 4 hrs at 4OC. Protein A-Sepharose beads were subsequently added and the entire mixture was Mer incubated for 2 hours (see above). Proteins eluted from the anti-P85 immunoprecipitations were subjected to SDS-PAGE and immunoblotting for GluR2. Precipitates by anti-GluR2 antibodies were used for PI3K kinase assays.
2.6 Colorimetric assay of AMPA receptor cell surface expression
Colorimetric assays were performed as illustrated in Fig 2.1. Briefly, HEK 293 cells were transfected with 10 or 15 pg of each plasmid in 10 cm or 15 cm dishes respectively as described above. At 12 hours post-transfection, cells were washed once with PBS, briefly shocked with 2.5% DMSO in complete media, rinsed with PB S several times, then returned to normal media Twenty-four hours after transfection the cells f?om each large plate were detached using ~ypsin/0.25%EDTA and then transferred to a 12-well plate at a seeding density of 2.5~10'cells/well. 36 to 48 hours after transfection, cells were washed and treated with 0SpM insulin for 10 rnins. To minimize transfection efficiency-related variation, treatment and control studies, as well as time course studies, were performed on the same sets of 12-well plates derived from the same population of transfected cells.
After treatment with insulin, cells were fixed using paraformaldehyde (2% in PBS for 3 minutes for non-permeant and 4% in PBS for 15 minutes for permeant conditions). Cells were then washed 3 times under either non-permeant (PBS alone) or permeant (PBS-0.2% Triton X-100) conditions, blocked for 1 hour at room temperature with 3% BSA, then incubated overnight at 4°C in anti-HA
(1 :2000) to detect HA-tagged receptors; polyclonal antibodies raised against the N-terminal region of GluRl (1 :1000; Oncogene Science) to detect wild type GluRl ; or monoclonal antibody raised against an N-terminal sequence of GluR2/4 for wild type GLuR2 (1 :1000; Pharmingen). As a control, specific staining of cell-surface and cellular GABA, receptors in transfected cells was performed as previously described (Wan et al., 1997). The wells were then washed with PBS and incubated for 1 hour at room temperature with the appropriate secondary antibody conjugated to horseradish peroxidase (HN?)(1 :8OO, Amersham). Cells were washed 5 times with PBS to minimize non- specific reactivity, then incubated with 1 volume of the chromogenic HRP substrate OPD (Sigma) for approximately 2 minutes. Reactions were stopped with 0.2 volume of 3N HC1, and the optical density of one mL of supernatant was read on a spectrophotometer at 492 nm.
For quantification of cell-surface expression of native AMPA receptors, colorimetric assays were performed on cultured hippocampal neurons cultured in 12-well plates. For quantification of
AMPA receptor insertion using colorimetric assay, cell-surface AMPA receptors were first blocked by the anti-GluR2 antibody and a non-HRP conjugated secondary antibody as described in Section
2.4.3, Newly inserted AMPA receptors were then labeled with the same primary antibody and an
HRP-conjugated secondary antibody under non-permeabilized conditions followed by detection with
HRP-OPD reactions,
2.7 PI3-kinase assay
AMPA receptors or PI3K were immunoprecipitated from brain slice lysates by anti-GluR2 or anti- p85 antibodies ( see above). The immunoprecipitates were washed three times on ice by each of
Washing Buffer I (PBS containing 1% NP-40), 11 (0.5 M LiCl, 0.1 M Tris, pH 7.5) and III (10 mM
Tris pH 7.5, 100 mM MI),sequentially. The washed pellets were then mixed with PI solution (50 pI washing buffer m,10 p1 of 100 mM MgCl,, 2 0 p1 PI suspension) and the reaction was started by adding 5 pI 3'~-~~~.The reaction was stopped 10 rnin later by 10 p1 HCl and 160 p1 CHClJMeOH mixture (1 A). The resulting lipid products were separated on a TLC plate and examined following autoradiography.
2.8 WhoIe-cell recording of EPSCs in hippocampal brain slices
2.8.1 Rat brain sfice preparation: Hippocampal slices (400pm)were prepared fiom 4-6-week-old
(for LTD recording, 16-26-day-old) SD rats. Following anesthetization with 20% Urethane, rat brains were immediately removed into a dish containing artificial cerebrospinal fluid (ACSF) for slicing (25.2 mM sucrose, 3 mM KCl, 1 mM MgCl,, 1 mM CaCl,, 1-2 mM KH,PO,, 26 mM NaKCO, and10 mM glucose) at 4OC. The brain was cut into two blocks, each containing one hemisphere of hippocampus, and blocks were mounted in a slicing chamber filled with ice-cold slicing ACSF.
Transverse slices cut with a vibratome were transferred into a beaker containing ACSF for recording
(same as slicing ACSF except that the sucrose is replaced by 126 mM NaCl) and were allowed to recover for at least 1 hour at room temperature before recording. All the buffers described above were kept bubbled with 95% o2/5%CO, throughout the procedure.
2.8.2 Whole-cell recording of EPSCs in CAI neurons: For electrophysiological recordings, slices were perfbed at room temperature (22 to 24°C) with recording ACSF, bubbled with %%O JS%CO,.
An incision was made between CA3 and CAI to reduce epileptiform activity. Whole-cell recordings of CAI neurons were performed using the "blind" method with an Axopatch- t D amplifier (Axon
Instruments, Foster City, CA). The recording pipettes (4-5 MQ) were filled with solutions which contained (mM): 135 CsCL, 0.05 EGTA, 10 HEPES, 4 Mg-ATP, 0.2 GTP, and 5 QX-3 14, pH 7.4, Figure 2.2 Experimental configuration for recording of evoked post-synaptic currents in hippocampal CAI neurons
Recording
Stimulating electrode
col
In a hippocampal slice, axons from CA3 neurons (Schaffer collateral fibers) form synaptic connections with CAI neurons. Synaptic transmissions are induced by electrical stimulation of the SchafXer collateral fibers and the postsynaptic cunrents are monitored at CAI neurons by whole-cell patch-clamp recording. Figure 2.3 Recording of AMPA receptor-mediated EPSCs in hippocampal CAI neurons
AMPA component
DNQX Control Bicucuhe
The electrical stimulation-induced postsynaptic current is a mixture of both inhibitory and excitatory components and the AMPA receptor-mediated excitatory postsynaptic current (EPSCs) is isolated by using bicuculline to block the GABA, component (A), and the AMPA-EPSCs is further confirmed by the AMPA receptor antagonist DNQX to completely abolish the response (El). 310 mOsm. In some experiments the intracellular solution was supplemented with one of the following: 10 mM BAPTA; anti-insulin receptor kinase domain antibody (60 pg/mL) (1 7A3, Morgan and Roth, 1987), a gift of Dr. R.A. Roth at Stanford University); mouse IgG (100 pg/mL); amphiphysin-SH3 domain (1 00 pg/mL); mutant amphiphysin-SH3 domain (1 00 pg/rnL); active PI3K
(1 500) or boiled inactive PI3K (1 500). Series and input resistance were monitored throughout each experiment and cells were excluded fiom data analysis if a greater than 20% change in the series or input resistance occurred during the course of the experiment, EPSCs were evoked by stimulation of the Schaffer collateral-commissural pathway with a bipolar tungsten electrode (0.05 ms duration at a rate of 0.07 Hz) (Figure 2.2), and the AMPA component was isolated by adding the GABPLA receptor antagonist bicuculline (20 pM) to block the GABA, receptor-mediated inhibitory synaptic currents (Figure 2.3).
The minimum stimulus-evoked EPSCs were recorded essentially following a modified protocol (Liao et at., 1995). Briefly, stimulus was delivered through a micro glass pipette that was filled with regular ACSF solution and positioned in the stratum radiatum 50-100 pM &om the recorded cell. The noise amplitude distribution was estimated by performing a similar measurement on traces with no stimulation. Noise s.d. calculated from such measurements was 2.3 *0.19 pA and
2.4*0.12pA at holding potentials of -60 and +40 respectively. Synaptic failure was determined for each epoch of responses with amplitudes within the noise levels. Stimulus intensity was adjusted to a level that produced a 20-30% failure rate. 40 responses before and after insulin (0.5 pM; 10 min) bath application were recorded. The AMPA component was recorded at a holding membrane potential of -60 mV unless otherwise indicated, and the NMDA component was recorded in the presence of the non-NMDA receptor blocker 6,7-dinitroquinoxali~e(DNQX, 10 pM) at a holding membrane potential of +40 mV (to remove the voltage-dependent MgZ' blockade). AMPA and
NMDA components were pharmacologicaliy confirmed by their blockades with DNQX and D-amino- phosphonovaIeric acid (APV,50 pM), respectively.
2.8.3 LTD induction: Homosynaptic CAI LTD was recorded in siices prepared fiom rats aged 16-26 post natal days old. Foilowing a 20 min recording of base line EPSCs evoked every 30 s, the recording was switched to current-clamp mode, and 15 minutes of train stimulation at 1Hz (900 pulses in total) was delivered fiom the same stimulating electrode. The recording was then switched back to voltage-clamp mode, and EPSC recordings at the base line stimulus rate were then recorded for more than 1 hour thereafter.
2.9 Miniature EPSC recordings in cultured hippocampal neurons
Whole cell recordings were made fiom these cultures 12 - 17 days after plating. Patch electrodes (3
- 5 MQ) were coated with Sylgard to improve signal-to-noise ratios. Recordings were performed at room temperature (20-22°C)- Recordings fiom each neuron lasted for at least 40 to 80 minutes. The series resistance in these recordings varied between 6 to 8 MQ, and recordings where the series resistance varied by more than 10% were rejected. No electronic compensation for series resistance was employed.
The patch electrode solution contained (mM): CsCi2, 140; EGTA, 2.5 mM or BAPTA, 25 mM; MgC12, 2; HEPES, 10; TEA, 2; K2ATP, 4; with pH 7.3 and osmolarity between 300 to 310 mosmol-I. The perfusion or bathing solution was of the following composition (mM): NaCl 140; CaCl, 1.3; KC1 5.0; HEPES 25; glucose 3 3; TTX 0.0005; strychnine 0.00 1; bicuculline methiodide
0.02; with pH 7.4 and osmolarity between 325-3 3 5 mosmol-l. Each cell was continuously supefised
(1 .O dmin)with this solution from a single barrel of a computer-controlled multi-barreled pefision system. Solutions supplemented with glycine or glycine and NMDA were applied fkom an alternative barrel.
Miniature EPSCs were recorded using an Axopatch 1-B amplifier (Axon Instruments Inc.) and records filtered at 2 kHz, stored on tape and subsequently acquired "off-line" with an event detection program (SCAN, Strathclyde software). The trigger level for detection of events was set approximately three times higher than the baseline noise. Inspection of the raw data was used to eliminate any false events and 80 to 300 mepscs were averaged for display purposes. The same number of events was used when averaged rnepscs were compared. All population data were expressed as mean* S.E.M. The Student's paired r test or the ANOVA test (two-way) was employed when appropriate to examine the statistical significance of differences between groups of data. CHAPTER THREE
Long-term Depression of Excitatory Synaptic Transmission
Through AMPA Receptor Internalization
This work has been published in Neuron as:
Man H Y, Lin J W, Ju W H, Ahmadian G, Liu L, Becker L E, Sheng M, Wang Y T. Regulation of
AMJ?A receptor-mediated synaptic transmission by clathrin-dependent receptor internalization.
Neuron 2000 Mar;25(3):649-62
The work in fig 3.1, 3.2,3.6,3.9,3.10, 3.12 (B) was done by me; work in fig 3.3, 3.4, 3.5, 3.7,3+8 was done by W Ju (Dr. Y T Wang's lab) and J Lin (Dr. M Sheng's lab); work in fig 3.1 1,3.13 and
3.14 was done by Dr. L Liu (Dr. Y T Wang's lab); work in fig 3.12 (A) was done by Dr. G Ahrnadian
(Dr. Y T Wang's lab). 3.1 Abstract
AMPA receptors are a subtype of glutamate receptors which are responsible for most of fast excitatory synaptic transmissions. The redistribution of postsynaptic AMPA receptors has been proposed as a means of regulating synaptic strength at glutamatergic synapses. However, mechanisms mediating these receptor redistributions are poorly understood. The present study shows that AMPA receptors undergo endocytosis via a clathrin-mediated pathway, and that the internalization can be accelerated by insulin in a GluR2 subunit-dependent manner- The insulin- stimulated endocytosis rapidly decreased the number of AMFA receptors in the plasma membrane, resulting in a long-term depression (LTD) of AMPA receptor-mediated synaptic transmission in hippocampal CAI neurons. Moreover, insulin-induced LTD and low-fiequency-stimulation (LFS)- induced hippocampal CA1 homosynaptic LTD were found to mutually occlude each other and both were blocked by inhibiting postsynaptic clathrin-mediated endocytosis, These data indicate that controlling the number of postsynaptic receptors by endocytosis may be an important mechanism underlying synaptic plasticity in the mammalian CNS-
3.2 Introduction
The AMPA subtype excitatory amino acid receptor is the principd receptor mediating fast synaptic transmission at most excitatory synapses in the CNS (Hol1rnan.n and Heinemam, 1994). Plastic changes in the strength of AMPA receptor-mediated synaptic transmission are intimately involved with various normal physiological brain functions, incl~dinglearning and memory, as well as with many neuropathological disorders, such as the neurotoxicity associated with cerebral ischemia.
Although the mechanisms by which the modification of synaptic strength is achieved remain intensely debated (Barria et aI., 1997; Bliss and Collingridge, 1993; Malenka and Nicoll, 1997), it has recently been proposed that a rapid change in the number of bctional postsynaptic AMPA receptors may be an important means of controlling synaptic efficacy (Craig, 1998; Malinow, 1998) - This has gained support fiom studies of synaptic plasticity in several experimental preparations (Li and Zhuo,
1998;Liao et al., 1995; Durand et al., 1996)- However, the mechanisms through which receptor numbers in the postsynaptic plasma membrane surface can be rapidIy regulated are poorly understood.
Most integral plasma membrane proteins are trafficked between the plasma membrane and the intracehlar compartments via vesicle-mediated membrane hsion (insertion) and endocytosis
(internalization). Regulation of these processes has been shown to be an important means of controlling the cell-surface expression, and hence fimction, of many proteins, such as the opioid receptor (Chu et d., 1997), the P-adrenergic receptor (Karoor et al., 1998), and the GLUT4 glucose transporter (Pessin et al., 1999). By analogy, it is not unreasonable to speculate that plasma membrane expression of AMPA receptors is also subject to similar modes of regdation, and such regulation at the postsynaptic membrane may be a powerfd means of controlling synaptic efficacy
(Carroll et al., 1999b)
Insulin is expressed in discrete regions throughout the brain. Neurons can synthesize and release insulin in response to membrane depolarization (Clarke et al., 1986; Raizada, 1983). In addition, peripheral insulin can penetrate the blood brain barrier and enter the brain (Wozniak et al.,
1993). Insulin receptors are also highly expressed in CNS neurons and localized to synapses (Abbott Figure 3.1 Selective immunostaining of AMPA receptors on the cell surface or AMPA receptors in the whole cell
HEK cells were transiently transfected with AMPA receptor subunits GluRl or GluR2 which were
HA-tagged on their extracellular N-termini. Under non-penneabilized condition (Non-perm), only the cell-surface AMPA receptors were stained with anti-HA primary antibodies and FITC- conjugated secondary antibodies (left panel). After permeabilization, cells were then stained again with an anti-GluR1 (or anti-GluR2) polyclonal antibody, and a Cy3-conjugated secondary antibody to label the total pool of the receptors, both those on the plasma membrane and those inside of the cytosol (middle panel). The superimposed images (right panel) show the surface receptors in yeIlow and the cytosolic receptors in red. Rgure 3.1 Selective irnmunolabeling of AMPA receptors on the cell surface or AMPA receptors in the whole cell
Non-perm Perm Overlap Figure 3.2 Insulin treatment affects the cell surface expression of AMPA receptor subunits
Control Insulin
HEK cells transiently expressing AMPA receptor subunits were treated with insulin (0.5 p.M) for10 rnins and the cell-surface receptors were then imrnunostained under non-permeabilized conditions. Insulin treatment showed no obvious effect on the cell-surface expression of GluRl subunits, but dramatically reduced the surface expression of GWsubunits. Figure 3.3 Quantitation of insulin effects on AMPA receptor cell-surface
expression rate by colorimetric assays
(A) Insulin induces down-regulation of AMPA receptor cell surface expression level in a GlW
subunit-dependent manner. Changes in cell-surface receptor expression in transfected HEK cells were determined using the ratio of absorbance readings obtained after antibody labelling under non- permeable (surface) versus permeable (total) conditions. In a control experiment in HEK cells transfected with GABA, receptor a lRAGp2y2cDNAs, GABAn receptors were translocated to the plasma membrane in response to insulin.
@) Time course of the insulin-induced reduction in cell-surface GluR2 receptors. Insulin incubation
(0.5 pM, 10 min; black bar) induces a rapid and long-lasting decrease in celI-surface GluR2 receptors (surface) without affecting the number of GlWreceptors present in the cells (total), resulting in a lower proportion of receptor on the surface (surface/total), indicative of translocation of the receptors fiom plasma membrane to intraceIIuIar compartments. Data are expressed as absorbance readings taken at the indicated time point (TJinitial absorbance (To).(n=6 for each time point), Figure 3.3 Quantitation of insulin effects on AMPA receptor cell-surface expression rate by colorimetric assays
0 Control Insulin
1.2 1 Insulin
Time (min) et al., 1999). Since glucose utilization in neurons is largely insulin-independent, insulin in the CNS may be involved in activities other than the regulation of glucose homeostasis. Indeed, recent evidence is consistent with a wide range of neuronal functions for brain insulin, including neuromodulation, neuronal growth and maturation, neuronal protection, and learning and memory
(Wickelgren, 1998). The detailed mechanisms by which brain insulin is able to modify neurond function, however, remain to be determined. Analogous to its fimction in peripheral tissues, where insulin produces rapid translocation of -GLUT4 to the plasma membrane thereby increasing glucose uptake in these cells (Pessin et al., 1999), we have recently demonstrated that insulin causes a rapid recruitment of-functional GABA, (type A y-arninobutyric acid) receptors to the postsynaptic domain in mature CNS neurons, resulting in a long-lasting enhancement of GABA, receptor-mediated synaptic transmission (Wan et al., 1997). Others have also reported that insulin can regulate the cell- surface expression and hence the function of various other ion channels and neurotransmitter receptors (Kanzaki et al., 1999; Karoor et al., 1998). These observations suggest that insulin may function as an important neuromodulator in the CNS by regulating the intracellular trafficking and plasma membrane expression of ion channels and neurotransmitter receptors in neurons.
We therefore set out to determine if the cell-surface expression of AMPA receptors can be rapidly regdated by insulin, and if so, whether such regulation can lead to an alteration of synaptic strength. Our data indicates that AMPA receptors are subject to clathrin-mediated constitutive endocytosis. By stimulating the rate ofthis clathrin-mediated endocytic removal of AMPA receptors, insulin produces long-lasting depression (LTD) of the AMPA receptor-mediated synaptic transmission in hippocampal CAI cells. The insulin-induced LTD and low-fequency-stimulation (LFS) evoked homosynaptic LTD mutually occlude each other and blocking postsynaptic endocytosis also prevents LFS-evoked LTD. Thus, our data suggests that regulation of receptor endocytosis may be a common and important mechanism of synaptic plasticity in the mammalian CNS.
3.3 Results
3.3.1 Insulin produces a rapid decrease in cell-surface AMPA receptors via a GluR2 subunit- dependent mechanism.
Native AMPA receptors are predominantly assembled eomheteromeric combinationsof four known subunits (GluR1-4). However, when expressed in mammalian cells, individual GIuR subunits can form hctional homomeric AMPA channels (Burnashev et al., 1996). To study AMPA receptor surface expression, GluRl and GluR2 cDNAs were tagged with the hemagglutinin (HA) epitope in their N-terminal extracellular domains. After transient transfection of these constructs into human embryonic kidney (HEK) 293 cells, we were able to distinguish between receptors expressed on the cell surface and intracellular receptors by antibody labelling of the HA epitope under permeant and non-permeant conditions. Using confocal immunofluorescent microscopy, we found that homomeric
HA-GluR1 and HA-GluR2 AMPA receptors were expressed both on the plasma membrane and in intracellular compartments (Figure 3 - 1). Cell-surface receptor expression (quantitative colorimetric) assays revealed that the proportion of cell-surface receptors was approximately 50% of the total expressed receptors for both GluRl and GluR2 (Figure 3.3A).
Similarly to other integral membrane proteins, AMPA receptors are likely constitutively h-a£6cked between the plasma membrane and intracellular compartments via plasma membrane
insertion and internalization. Presumably, the number of AMPA receptors on the cell surface reflects
a balance between these two processes, and factors affecting the rate of insertion or internalization
should change the density of surface AMPA receptors. We investigated this by acutely treating HA-
GIuRl or HA-GluR2 transfected cells with insulin, as brief insulin stimulation affects the membrane
insertion (Pessin et al., 1999) and endocytosis (Karoor et al., 1998) of a number of plasma membrane proteins. As a control, we transfected HEK cells with the GABA* receptor, as we have previously
found that insulin produces a rapid increase in the number of cell-surface GABA, receptors in the same expression system (Wan et al., 1997). As shown in Figure 3.2 and Figure 3.3, brief insulin exposure (0.5 pM, 10-1 5 min) resulted in a rapid and persistent reduction in GluR2 receptors on the plasma membrane. Cell-surface GluR2 receptors were depleted to 32.1&2.4% within 10 min, and the effect persisted longer than one hour after insulin washout (Figure 3.3B). Interestingly, insulin treatment did not cause any changes in the cell-suface expression of the GluRl subunit. Conversely, sdaceGABA, receptors increased in the same experiments, consistent with previous findings (Wan et al., 1997) (Figure 3 -3A). Thus, the reduction of cell-surface GluR2 by insulin is subunit-specific, and opposite to the effect of insulin on GABA, receptors.
Because native AMPA receptors are largely heteromeric complexes often containing GluR2
(Hollmann and Heinemam, 1994), it was important to determine if GluR2 could confer insulin- regulated surface expression on heteromeric receptors. Insulin stimulation of cells co-transfected with both HA-GluRI and GluR2 receptors decreased cell-surface expression of HA-GLUM to a level comparable to that seen in cells expressing HA-GluR2 alone (Figure 3.3A). Furthermore, we Figure 3.4 The GluRZ specificity of insulin effects is determined by the intracellular C-terminus of the subunit
El Control Insulin
Insulin-induced reduction in cell-surface expression of AMPA receptors is dependent on the presence of the GluR2 carboxyl tail (CT). Exchanging the carboxyl tails between the HA-GluRt and HA-GluR2 subunits switched the insulin effect fiom GluR2 to GluRl/,,, indicating that the GluR2 carboxyl tail is necessary and sufficient for insulin-induced decreases in the cell-surface expression of AMPA receptors. Figure 3.5 AMPA receptors are subject to endocytosis which is facilitated by
insulin.
(A) AMPA receptors show constitutive endocytosis. Following pre-labelling of live cells with anti-
HA antibody at 4"C, cells were returned to 37°C to allow constitutive internalization, and fixed at
the indicated time points. Cell surface and internalized receptors were sequentially stained with
secondary antibodies under non-permeant (green) and permeant (red) conditions. Superimposed
confocal images of imrnunofluorescent staining of cell surface GluR2 receptors (green) and
internalized receptors (red) showing constitutive endocytosis of HA-GluR.2.
(B) Time course of GluR2 endocytosis determined by colorimetric assays. Time-dependent loss of surface-bound antibody was observed at 37OC. It is unlikely that this loss is due to a time- dependent dissociation of antibody fkom the receptors as there is no detectable antibody loss following incubation for various periods of time at 4OC (4OC non-perm) or due to antibody instability since no detectable change in antibody absorbance readings were observed following a one-hour incubation period at 37°C under permeant conditions (37 OC perm). Thus, this time- dependent loss of surface-bound antibody is due to an inability to access the antibody under non- permeant conditions after being constitutively internalized together with the receptors. Brief insulin treatment (black bar) enhanced both the rate and magnitude of the GluR2 receptor internalization.
The green and red lines represent the best single exponential fit to the mean data of GluR receptor endocytosis without and with insulin treatment respectively (n=6 for each time point).
(C) Time course of GluRl subunit endocytosis. GluRl showed constitutive internalization in a rate
(~39.3min) similar to that of GlW(~41.7 min). However, not like for GluR2, insulin failed to enhance the receptor endocytosis in GluRl. Figure 3.5 AMPA receptors are subject to endocytosis which is facilitated by insulin.
0 min 10 min 15 min
0 37OC perm 0.2 - El 37OC non-perm W 37OC non-perm+insulin
0.0 I I I I I I 1 0 10 20 30 40 50 60 (min
0 37OC nun-perm 37OC non-perm+insulin
0 10 20 30 40 50 60 (min)
-76- Figure 3.6 Immunostaining showing that both constitutive and insulin- stimulated GluR2 endocytosis are mediated by clathrin-coated pits.
Cell-surface GluR2 receptors were prelabelled with primary antibody in live transfected HEK cells at 4 "C,and, following 10 min incubation at 37°C with or without treatments, cells were sequentially stained for GluR2 and EPS L 5 under permeant conditions. Insulin treatment enhanced GluR2 subunit internalization and most internalized GluR2 in the presence (+Insulin) and absence (-Insulin) of insulin co-localized with EPS 15, an adaptor protein participating in clathrin-coated-pit-mediated endocytosis. GluR2 endocytosis was blocked by incubation of the cells with hypertonic sucrose solution (0.45 M; 10 min) prior to insdin application (Insulin+Sucrose). Figure 3.6 Immunostaining showing that both constitutive and insulin-stimulated GluR2 endocytosis are mediated by clathrin-coated pits.
GluR2 Eps-I 5 Overlay confirmed that the reduction in surface GluR2 subunit is related neither to epitope tagging nor to a cell line-specific phenomenon, as we observed similar reductions in cell-surface receptors in response to insulin in Chinese hamster ovary (CHO) cells expressing human insulin receptors and wild-type
GLuR2 or GluRl +GluR2 cDNAs (using antibodies directed against the extracellular domain of these subunits; data not shown).
3.3.2 The carboxyl tail of GluR2 determines subunit-specificity of insulin-induced reduction in cell-surface AMPA receptors
GluR subunits display sequence divergence within the carboxyl terminal (CT) cytoplasmic tail region and this region has been shown to mediate subunit-specific interactions with various cytoplasmic proteins (Dong et ai., 1997; Lin and Sheng, 1998; Nishimune et al., 1998; Song et al., 1998; Xia et al., 1999). We tested whether the signals that confer the subunit specificity of insulin eEects may reside within the GluR2 CT region, by exchanging the CT domains of GluRl and GLuR2, creating chimeric HA-GluRlI,, and HA- G~UR~/,~AMPA receptors- Replacing the GluR2 CT with that of
GluR 1 (GluRZ, ,) completely blocked the insulin-induced decrease in the number of cell surface receptors (Figure 3.4). In contrast, insulin treatment decreased the cell-surface expression of
GluRl/,, to an extent comparable to that of wild type GluR2 (Figure 3 -4). These results demonstrate that the GLuR2 CT region contains a signal that is necessary and sufficient for the insulin-induced decrease in the cell-surface expression of the GluR2-containing AMPA receptors. 3.33 Insulin selectively facilitates the rate of GluR2 but not GIuRl receptor endocytosis.
Insulin treatment reduced the -AMPA receptors expressed on the plasma membrane surface without altering the total number of receptors expressed in these traosfected cells (Figure 3.3B), indicating that the relative rates of protein synthesis and degradation were not affected. Alterations in AMPA receptor surface expression must therefore be the result of insulin-induced inhibition of membrane insertion and/or facilitation of endocytosis. To discriminate among these possibilities, we employed an internalization assay previously used in measuring endocytosis of other plasma membrane receptors (Chu et al., 1997). Surface HA-GluR1 and HA-GluR2 receptors in live cells were pre- labelled with anti-HA antibody at 4°C (a temperature at which endocytosis is blocked). Cells were returned to 37"C to allow endocytosis to resume, and internalization of the labelled surface receptors at merent time points was then visualized using fluorescent confocal microscopy and quantified by colorimetricassays. GluR2 receptors showed a constitutive time-dependent endocytosis fiom the cell surface in untreated cells (Figure 3 SA). After 10 minutes, most internalized receptors were localized to small puncta closely associated with the plasma membrane, which may represent early endosornes.
At later time points, internalized receptors accumulated in larger punctae farther away fkom the plasma membrane, possibly indicative of vesicle hionor localization in late endosomes. The time course of the constitutive internalization determined by quantitative colorimetric assays could be described by a single exponential with a time constant of 42 min (Figure 3 SB). Fifteen minutes after the start of endocytosis, approximately 1S+2.l% of the labelled cell-surface GluR.2 receptors were internalized, and this proportion increased to 48*4.7% at one hour (Figure 3SB, n=6). When the cells were incubated with insulin (0.5 pM) during the course of the endocytosis assay, the labelled cell-surface GluR.2 receptors were found to be internalized at an acceIerated rate, following an exponential time course with a time constant of 15 minutes, approximately 3 times faster than that observed in non insulin-stimulated cells (Figure 3 33)- Moreover, insulii stimulation also increased the degree of GLUE interndization. At the 15 minute and one hour time points, 39&1,2% and
58.%1.0% of receptors were internalized, respectively. As shown in Figure 3SC, pre-labelled
GluRl receptors were also subject to constitutive endocytosis at a rate and degree comparable to that of GluR2. However, in marked contrast to its actions on GluR2, insulin failed to alter either the rate or the level of GluRl receptor endocytosis, mersupporting the dependence of the selective insulin actions on the presence of GluR2. Thus, while constitutive receptor endocytosis may be a common feature of different AMPA receptor subunits, the insulin induced response is GlW-specific.
3.3.4 Both constitutive and insulin-reguiated GluR2 endocytosis are mediated by clathria- coated pits.
Most plasma membrane proteins are internalized by clathrin-mediated endocytosis (Bonifacino et al.,
1996). Consistent with GluR2 endocytosis by clathrin-dependent mechanisms, we found that, under both constitutive and insulin-stimulated conditions, the internalized GluR2 receptors were extensively co-localized with EPS 15, an integral component of clathrin-coated pits (van et al., 1997) (Figure 3 -6,
Control and Insulin). Moreover, pretreatment of cells with hypertonic sucrose (0.45 M; 10-30 min), a blocker of clathrin-mediated endocytosis (Hansen et al., 1993b), prevented both constitutive and insulin-enhanced internalization of GluR2 receptors (Figure 3 -6,Sucrose and Insulin+sucrose). These results indicate that AMPA receptors undergo clathrin-mediated constitutive endocytosis, and that Figure 3.7 Insulin-induced reduction in rell-surface AMPA receptors is blocked by inhibiting clathrin-dependent receptwr endocytosis.
(A) Quantitative colorimetric assays show that insulin-induced depletion of cell-surface GIuR2 receptors is blocked by treatment of cells with hypentonic sucrose solution, a commonly used method to block the clathrin-dependent endocytosis. Note rthat sucrose treatment did not alter the basal level of cell-surface AMPA receptors. (n=6 in each group).
(B) Co-expression of dominant negative K44E mutant dynarnin I @yn-mut) blocks the insulin effect in reducing GluR2 cell-surface expression- Cells were co-transfected with GIuR2 and wild type or mutant dynamin. Neither wild type (Dyn) nor mutlnnt dynamin (Dyn-mut) changed the basal surface expression rate of GluR2, but the mutant dynamin Mocked insulin-induced reduction. (n=6 in each grow)- Figure 3.7 Insulin-induced reduction in cell-surface AMPA receptors is blocked by inhibiting clathrin- dependent receptor endocytosis. Figure 3.8 Clathrin-mediated endocytosis of AMPA receptors in cultured
hippocampal neurons.
(A) Confocal images of internalization of cell-surface AMPA receptors at time 0 (AI) and 10 min in the absence (A2) or the presence (A3 of insulin treatment. The receptors were prelabelled with an antibody raised against the N-terminal extracelldarepitope of GluRl in live neurons and allowed to endoctyose for varying lengths of time at 37OC. AMPA receptors remaining on the cell surface and those internalized into the interior of the cells were then sequentially stained with secondary antibodies under non-permeant (FITC green) and permeant (Cy3 red) conditions. Insulin treatment dramatically increased AMPA receptor internalization. As a control, staining of insulin treated cells under non-permeant condition showed no red signals (A4).
@) Inhibition of clathrin-mediated endocytosis by hypertonic sucrose abolishes both constitutive and insulin-stimulated AMPA receptor internalization. 10 min sucrose (0.45 M) treatment completely blocked AMPA receptor internalization even in the presence of insulin stimulation (B2), compared with the non-treated control (B 1). Figure 3.8 Clathrin-mediated endocytosis of AMPA receptors in cultured hippocampal neurons Figure 3.9 Insulin increases the association of native AMPA receptors and the
AP2 adaptor protein complex in hippocampal tissue.
(A) Antibody to adaptin P2 co-immunoprecipitates AMPA receptor GluR.2 subunits with adaptin
P2 &om hippocampal slice homogenate, but a control immunoprecipitationwith non-specific mouse
IgG precipitates neither. Immunoprecipitates were first immunoblotted with anti-GluR2/4 antibody
(the antibody recognizes only the GhR2 subunit in Western blotting), and then stripped and reprobed with anti-adaptin /32 antibody (adaptin P2).
@) Insulin treatment increases the amount of AMPA receptor associated with the AP2 complex.
Co-immunoprecipitation of protein samples f?om control and insulin-treated (0.5 pM; 10 min) hippocampal slices was performed using anti-adaptin P2 antibody. The precipitates and homogenate were then sequentially immunoblotted for GluR2 and adaptin P2.
(C) Densitometric quantification of GluR2 imrnunoblots fiom five separate experiments. * p<0.05.
@) The NMDA receptor does not appear to associate with the AP2 adaptor complex.
Immunoprecipitations of protein samples from control and insulin-treated (0.5 pM; 10 mi@ hippocampal slices were performed with anti-adaptin P2 antibody or control IgG, and the immunoprecipitates were then sequentially probed with anti-NMDARI and anti-adaptin P2 antibody. Figure 3.9 Insulin increases the association of native AMPA receptors and the AP2 adaptor protein complex in hippocampal tissue.
Adaptin P2 I &
0Control Insulin 150 -. c.V) .-c - 3 -m 100 -. 0 .-c. 0" so -.
0 - insuiin accelerates this process-
3.3.5 Facilitated clathrin-dependent endocytosis is fully responsible for insulin-induced reduction in cell-surface expression of GIuR2 receptors
As mentioned previously, the reduction of cell-surface AMPA receptors by insulin codd be either the result of increased endocytosis, decreased insertion, or combination of the two processes. As shown in Figure 3.7A, when Ievels of AMPA receptor surface expression were quantified using the cell-surface receptor expression assay, we found that the insulin-induced reduction was completely abolished by blockade of clathrin-mediated endocytosis with hypertonic sucrose, consistent with the loss of cell-surface AMPA receptors being mediated primarily by internalization.
Internalization via the clathrin pathway is dependent on the GTPase dynamin. Dominant negative mutants of dynamin (K44E or K44A) block clathrin-mediated endocytosis without affecting vesicle exocytosis (Damke et al., 1994) or insulin signalling (Kao et al., 1998). In HEK 293 cells transfected with HA-GluR.2 and co-expressing the dynamin K44E mutant, the basal level of surface receptors was maintained, but the insulin-induced decrease in cell-surface GluR2 was blocked (Figure 3 -7B).Over- expression of wild-type dynamin, on the other hand, did not affect either the basal cell-surface GluR.2 levels or insulin-induced reduction of cell-surface GLuR2, consistent with previous studies which showed that excess wild-type dynarnin does not significantly affect constitutive endocytosis
(Herskovits et al., 1993). Taken together, these results indicate that insulin can reduce the number of GIuR2-containing AMPA receptors on the plasma membrane over a time course of minutes, primarily by stimulating clathrin-mediated endocytosis.
-88- 3.3.6 InsuIin reduces the SU~C~expression of AMPA receptors in cultured hippocampal neurons in a clathrin-dependent manner
In order to determine if the results observed with recombinant HA-tagged AMPA receptor subunits were consistent for native neuronal AMPA receptors, we pre-labelled living hippocampal neurons with antibodies specific to the extracellular N-terminus of GluRl and allowed the receptors to be interndized at 37OC for 10 minute. The receptors that remained on the plasma membrane surface were stained with aFITC-conjugated secondary antibody (green) under non-permeant conditions, and the internalized receptors in the same neurons were then stained with a Cy3 conjugated secondary antibody (red) under permeant conditions. As shown in Figure 3.8, there was a modest, but clear, amount of AMPA receptor internalization 10 rnin following prelabelling under control untreated conditions (Figure 3.8, A2), suggesting a constitutive endocytosis of native AMPA receptors in these neurons. Insulin treatment greatly enhanced the internalization of AMPA receptors compared to the controls (Figure 3.8, A3). The internalized receptors under both control and insulin-stimulated conditions were characterized by small red puncta in dendritic shafts, typically lying within the boundaries of the plasma membrane delineated by surface-localized (green) receptors. The internalized red staining was virtually abolished if the permeabilization step between the FITC- and
Cy3-conjugated secondary antibody detection procedures was eliminated (Figure 3 -8,A4), indicating that the red punctate labeling truly represents a popdation of internalized AMPA receptors. Thus, native AMPA receptors in cultured hippocampal neurons also undergo both constitutive and regulated endocytosis at the rates similar to the recombinant GluR.2-containing AMPA receptors in transiently transfected HEK cells. Both the constitutive and insulin-stimulated AMPA receptor internalization in neurons were also found to be dependent on clathrin-mediated endocytosis as they were completely inhibited following hypertonic sucrose treatment (Figure 3.8, B).
3.3.7 Insulin stimulates the association of native AMPA receptors with the AP2 complex in hippocampal slices
To understand the role of AMPA receptor endocytosis in a more physiological context, we investigated whether the clathrin-mediated constitutive and insulin-induced AMPA internalization occurs in mature neurons in situ using hippocampal slices prepared from adult rats. Endocytosis of plasma membrane receptors is thought to be initiated by the recruitment and concentration of receptors in clathrin-coated pits, accomplished by binding of the receptors to clathrin adaptor proteins such as the AP2 adaptor protein complex (Schmid, 1997). Thus, we examined the association of native AMPA receptors with the AP2 complex in control or insulin-treated (0 -5 pM; 10 rnin) slices.
Under control conditions (Figure 3.9A), a small amount of AMPA receptor subunits were co- immunoprecipitated with P2-adaptin, a key component of the AP2 complex, but not with the IgG control antibody, perhaps reflecting a constitutive level of AMPA receptor internalization. Insulin treatment of the slices increased the amount of AMPA receptors co-immunoprecipitated with P2 adaptin (Figure 3.9B and C). In contrast, under the same experimental conditions, anti-adaptin P2 antibody did not imrnunoprecipitate the NMDA (N-methyl-D-aspartate) subtype glutamate receptors fiom either control or insulin-treated slice homogenates (Figure 3.9D). These results suggest that, in mature neurons in situ, AMPA receptors are constitutively internalized via clathrin-mediated
-90- mechanisms, and that insulin expedites this process by enhancing the recruitment of receptor subunits to the AP2 complex-
3.3.8 Insulin produces a long-lasting depression of the AMPA component of the excitatory postsynaptic currents in hippocampai CAI cells.
To determine the hctional effects of insulin-induced AMPA receptor internalization, we investigated insulin's effects on AMPA receptor-mediated synaptic transmission in hippocampal slices. Electrical stimulation of Schaffer collateral-commissural fibers evoked excitatory postsynaptic currents (EPSCs) in CAI cells which consist oftwo pharmacologically-distinct components mediated by AMPA and NMDA receptors (Figure 3.10A). Bath application of insulin (0.5 pM, 10 min) reduced the amplitude of the AMPA component of EPSCs, but had little effect on the NMDA component Figure 3.10A). The reduction in the AMPA EPSC amplitude was not associated with changes in either the voltage-current relationship or the reversal potential (Figure 3.1 OA, right panel).
The insulin-induced inhibition of AMPA EPSCs is long-lasting, as no recovery was observed after insulin wash out (Figure 3. IOB, Control and Insulin).
3.3.9 The induction of the insulin-induced depression of AMPA EPSCs is postsynaptic
Insulin's effects on AMPA receptor EPSCs may be exerted either pre- or postsynaptically, since insdin receptor vosine kinases are expressed at both presynaptic and postsynaptic membranes
(Abbott et al., 1999;Dore et d.,1997). The differential effects on the AMPA and NMDA component Figure 3.10 Insulin induces long-lasting depression of AMPA receptor-mediated excitatory postsynaptic currents in hippocampal CAI neurons.
(A) Bath application of insulin selectively inhibits the AMPA, but not the NMDA, component of the EPSCs. Representative EPSCs averaged from 4 individual recordings before (Control) or during application of insulin (0.5 pM). The NMDA component was recorded at a holding membrane potential of +40 mV. I-V curves of AMPA EPSCs recorded before (Control) and after a 10-min application of insulin (Insulin) show linear characteristic with the reversal potential at OmV.
(B) Insulin-induced depression of AMPA EPSCs is dependent on the activation of postsynaptic insulin receptor tyrosine kinase. Normalized EPSCs were plotted fkom neurons recorded with regular intracellular solution in the absence (Control, n=10) or presence of insulin (Insulin, n=10); or from neurons recorded in the presence of insulin with intracellular solution supplemented with
60 pg/d insulin receptor-neutralizing antibody (Anti-IR, n=6) or 100 pg/mLcontrol mouse IgG
(IgG, n=5). Insulin-induced inhibition was blocked by anti-IR neutralizing antibody, indicating that the insulin effect is through activating insdin receptors on the postsynaptic neurons. Bar indicates bath application of insulin (0.5 pM, 10 mins) for Insulin, Anti-IR and IgG groups. Figure 3.10 Insulin induces long-lasting depression of AMPA receptor-mediated excitatory postsynaptic currents in hippocampal CAI neurons.
A Control Insulin
o/ 1. Insulin
L Insulin A IgG 0.0 - I I I I I 0 5 10 15 20 25 30 Time (rnin) Figure 3.1 1 Insulin selectively inhibits minimum stimulus-evoked AMPA EPSCs.
(A) A plot of the minimum stimulus-evoked AMPA EPSC amplitudes fi-om hippocampal CA1 neurons (n=8). Individual traces before (Left) and after (Right) bath application of insulin (0.5 pM) are shown above the EPSC plot.
(B) Insulin has little effect on the NMDA component of minimum stimulus-induced EPSCs. The
NMDA EPSCs were recorded in the presence of CNQX (10 pM) at a holding membrane potential of+40 mV. (Top) Representativetraces before and after insulin. (Bottom) a plot of EPSC amplitudes taken fkom 6 cells.
(C) Bar graphs showing the selective effects of insulin on both the amplitude (left) and failure rate
(right) of AMPA (open bars; n=8), but not NMDA (filled bars; n=6), component of the minimum stimulus-evoked EPSCs. Figure 3.11 Insulin selectively inhibits minimum stimulus-evoked AMPA EPSCs
Insulin Treatment ~efore4l~After
Stimulus Number
LY < -5- n Before 41 b After Insulin Treatment z-10 1 I r 0 20 40 60 80 Stimulus Number 21
AMPA NMDA AMPA NMDA Figure 3.12 Insulin-induced depression of AMPA EPSCs is blocked by inhibiting elathrin-mediated endocytosis in postsynaptic neurons-
(A) Wild type but not mutant, amphiphysin-SH3 domain GST firsion proteins specifically interact with dynamin. Purified GST alone (GST), GST-amphiphysinSH3 (amphiSH3) and mutant GST- amphiphysinSH3 (amphiSH3m) fusion proteins were quantified using a Coomassie Blue-stained gel
(Top) and were added to hippocampal homogenates in a pull-down assay. Precipitated proteins were then immunoblotted with anti-dynamin I antibody (Bottom).
(33) Normalized AMPA-receptor-mediated EPSCs in CAI hippocampal neurons recorded with intracellular solution containing 10 mM BAPTA or 100 pg/mL of arnphiSH3 (n=7) or 100 pg/mL amphiSH3m (n=6) GST hionproteins. Note that the applied amphiSH3 competes for and interacts with endogenous dynamin, thereby inhibiting clathrin-mediated endocytosis of MAreceptors produced by insuLin application. Mutant amphiphysin cannot interact with endogencmus dynamin and does not inhibit the normal formation of the endocytic machinery and therefore has no effect on the internalization of AMPA receptors fo 110 wing insulin application. Figure 3.12 Insulin-induced depression of AMPA EPSCs is blocked by inhibiting clathrin-mediated endocytosis in postsynaptic neurons.
1 Insulin
A BAPTA AmphiSH3
Time (min) EPSCs are more consistent with a postsynaptic action of insulin-To Mertest insulin's site of action we postsynaptically applied a membrane-impermeant inhibitor of insulin receptors, a monoclonal antibody raised against the tyrosine kinase domain of the insulin receptor. This antibody has been shown to inhibit the activity of insulin receptors specifically both in in vifro kinase assays and in situ when injected directly into cells (Morgan and Roth, 1987). Postsynaptic application of this anti- insdin receptor antibody, but not control mouse IgG, eliminated insulin's ability to inhibit AMPA
EPSCs (Figure 3.10B). These results demonstrate that insulin acts postsynaptically via its receptor tyrosine kinase, supporting a postsynaptic locus for the induction -of insulin-induced depression of
AMPA EPSCs.
3.3.10 Post-synaptic expression of insulin-induced reduction in AMPA EPSCs
To determine the site(@of expression of the insulin-induced depression of AMPA EPSCs, we evoked unitary EPSCs by lowering the stimulus intensity to a level that produced a failure rate of greater than
20%. If the expression of depression is presynaptic, perhaps due to some retrograde messenger produced in the postsynaptic neuron, insulin should only affect the failure rate, but not the amplitude, of the weak stimulus-induced EPSCs. If, on the other hand, the expression is mainly postsynaptic as a result of removal of postsynaptic AMPA receptors or of altering channel conductance or open probability, insulin would be expected to alter both the amplitude and the failure rate of the EPSCs, with the latter due to postsynaptic failure in detecting the released transmitter (Liao et al., 1995).
Consistent with a postsynaptic site of expression, insulin reduced the amplitude of AMPA EPSCs in
8 out of 11 cells tested, and as predicted, this amplitude reduction was associated with a dramatic increase in failure rate (Figure 3.1 1A and C). In marked contrast, insdin failed to alter either the ampLitude or the failure rate of the minimum stimulus-evoked NMDA component EPSCs (Figure
3.1 1B and C). Together, the results strongly suggest that the locus of expression of insulin-induced= depression of AMPA responses is postsynaptic.
3.3.11 Insulin-induced depression of AMPA EPSCs is blocked by inhibiting clathrin-mediated endocytosis in postsynaptic neurons
The postsynaptic locus of induction and expression are both consistent with the effect of insulin being mediated by the clathrin-dependent endocytosis ofAMPA receptors. To address this possibility more specifically, we chelated intraceUular Ca2+ with the postsynaptic application of the membrane- impermeant Ca2' chelator BAPTA (10 mM), since clathrin-mediated endocytosis is dependent on intracellular Ca2' (Marks and McMahon, 1998). Following postsynaptic BAPTA application, insulin treatment failed to depress the AMPA EPSC amplitude (Figure 3.1223). More direct evidence for the involvement of postsynaptic clathrin-mediated endocytosis was obtained by the blockade of the insulin effect with a GST hsion protein of the amphiphysin SH3 domain (GST-amphiSH3). It is currently believed that binding of the amphiphysin SH3 domain to the proline-rich region of dynamin recruits dynamin to clathrin-coated pits, thereby initiating the process of clathrinldynamin- dependent endocytosis (Shupliakov et al., 1997). When incubated with hippocampal slice homogenates, GST-amphiSH3 was able to pull-down a protein which had a molecular weight of approximately 100 kD and was recognized by dynamin antibodies on Western blots (Figure 3.12A).
GST alone, and a GST fusion protein containing the mutated amphiphysin SH3 domain (GST- amphiSH3m), which cames two mutations in its dynamin-binding domain that render it incapable
of binding to dynamin (Shupliakov et al., 1997), failed to precipitate endogenous dynamin (Figure
3.12A). Thus GST-amphiSH3 specifically interacts with endogenous rat brain dynamin, and was
therefore used as a specific inhibitor to disrupt the binding between endogenous amphiphysin and
dynamin in rat hippocampal neurons. GST-amphiSH3 substantially reduced the ability of insulin to
depress AMPA EPSCs (Figure 3.12B). In contrast, the GST-amphiSH3m had little effect on insulin-
induced depression of AMPA EPSCs.
Collectively,our biochemical and electrophysiologicalresults in hippocampal slices and work fiom cultured hippocampal cells suggest that native AMPA receptors, similar to their recombinant counterparts in HEK cells, are subject to clathrin-mediated endocytosis. Insulin stimulates this process and causes the rapid removal of postsynaptic AMPA receptors, leading to long-term depression or even silencing of AMPA receptor-mediated synaptic transmission.
3.3.12 Inhibiting clathrin-mediated endocytosis in postsynaptic neurons prevents CAI homosynaptic LTD
Homosynaptic long-term depression (LTD) of AMPA receptor-mediated synaptic transmission in hippocampal CAI neurons is a well-characterized in vitro model of synaptic plasticity (Bear and
Malenka, 1994). While it is agreed that the induction of the most common form of LTD at this synapse is postsynaptic, mechanisms underlying its expression remain under debate (Boishakov and
Siegelbaum, 1995;Kameyama et al., 1998). Since our preceding experiments suggested that enhanced clathrin-mediated AMPA receptor endocytosis can profoundly depress AMPA receptor-mediated
-1 00- transmission, we investigated if endocytosis of AMPA receptors in the postsynaptic cell could contribute to homosynaptic LTD and whether LTD and insulin-induced reductions in AMPA EPSCs shared common pathways. Delivering 900 pulses at 1 Hz to the Schaffer collaterals, a standard low- frequency-stimulation (LFS) protocol for inducing homosynaptic CAI LTD However, when 100 pM GST-amphiSH3 was included in the recording pipette, LTD failed to be induced in every neuron tested (EPSC 9533% of the control after 60 min post-LTD induction; n=6; Figure 3.13). In contrast, inclusion of the GST-amphiSH3m altered neither the time course nor the extent of the LTD (64*4%; n=6; Fig 9A). These results indicate that a dynamin-dependent, clathrin- mediated endocytotic process, is required for the production of this particular form of LTD, and thereby suggest that LFS-induced LTD may use a similar mechanism underlying the insulin-induced LTD, which involves clathrin-mediated postssynaptic AMPA receptor internalization. 3.3.13 Insulin induced LTD and LFS-induced LTD mutually occlude each other If homosynaptic LTD shares common mechanisms with insulin-caused LTD, it is expected to see a mutual occlusion of these two forms of LTD. Indeed, as illustrated in Figure 3.14, after the LFS- induced LTD was fully established, application of insulin failed to further reduce the EPSC amplitudes (Figure 3.14A). Conversely, after insulin induced a rapid and persistent decrease in AMPA EPSCs, LFS was no longer able to produce LTD (Figure 3.14B). Thus, LFS and insulin- Figure 3.13 Inhibition of clathrin-dependent endocytosis in postsynaptic neurons prevents the expression of low-frequency-stimulus (LFS) induced LTD in CAI neurons. (Top) Superimposed individual traces recorded before (a) and 60 min after (b) the LTD induction with the LFS protocol. (Bottom) Time course of averaged AMPA EPSCs recorded with pipettes containing standard recording solution (controI; n=6) or solution supplemented with 100 pg/mL GST-amphiSH3 (AmphiSH3; n=6) or 100 pg/dGST-amphiSH3m (AmphiSH3m; n=6). The wild type amphiphysin -SH3 domain, but not the mutant, blocked the LFS-induced LTD, indicating an involvement of receptor endocytosis in homosynaptic LTD. Figure 3.13 Inhibition of clathrin-dependent endocytosis in postsynaptic neurons prevents the expression of low- frequency-stimulus (LFS) induced LTD in CAI neurons. Control AmphiSH3 AmphiSH3m s - 100 Q] U LFS 3 CI .-- 80 az- 0 60 Cr) n 40 'CJ o Control -.-A A AmphiSH3 cu 20 !E r ArnphiSH3m Time (min) Figure 3.14 Homosynaptic LTD and insulin-induced LTD mutually occlude each other in hippocampal CAI neurons. Low frequency stirnufation (LFS)-induced LTD blocks the subsequent insulin-caused long-term inhibition (A, n=Q, and conversely, after the establishmentof insulin-induced depression of AMPA EPSCs, the LFS protocol failed to induce LTD (B, n=6). This mutual occlusion indicates that the insulin-caused chemical LTD and the electrical stimulation-induced LTD may share common pathways in their expression. Figure 3.14 Homosynaptic LTD and insulin- induced LTD mutually occlude each other in hippocampal CAI neurons. LFS Insulin 0 10 20 30 40 50 60 70 Time (min) z- Insulin LFS Time (min) induced paradigms mutually occlude each other, strongly suggesting that the two systems share a common final pathway, namely the clathrin-mediated endocytosis of postsynaptic AMPA receptors. 3.4 Discussion This work presents severa.1 new findings concerning the intracellular trafficking and plasma membrane expression of the AMPA receptor. First, we show that AMPA receptors undergo endocytosis via a claWdynamin-dependent mechanism, which can be stimulated by insulin. Second, the insulin-regulated internalization of AMPA receptors is subunit-specific, and dependent on the presence of the GluR2 cytoplasmic tail. Third, stimulating AMPA receptor endocytosis can lead to a rapid change in the number of cell surface AMPA receptors. Fourth, insulin inhibits synaptic efficacy in hippocampal CAI cells, and this depression is dependent on the postsynaptic endocytosis of AMPA receptors. Finally, a similar endocytotic mechanism may contribute to the expression of low-frequency stimulation-induced homosynaptic LTD in CA1 neurons. These results, coupled with recent work suggesting AMPA receptor recruitment to synapses in long-term potentiation0;TP) (Lledo et al., 1998; Shi et al., 1999), strengthen the idea that regulated postsynaptic trafficking of AMPA receptors is an important mechanism for controlling synaptic efficacy. Constitutive versus regulated AMPA receptor endocytosis It is becoming increasingly clear that the clathrin-mediated internalization of plasma membrane proteins is a tightly regulated process and; such regulation is an important means of controlling the cell-surface expression, and hence function, of these proteins ( Karoor et al., 1998; Schmid, 1997). However, despite increasing evidence suggesting the involvement of activity-dependent redistribution of AMPA receptors in LTP and LTD (Carroll et al-, 1999; Malenka and Nicoll, 1999; Shi et al., 1999), the potential role of clathrin-mediated AMPA receptor endocytosis in these processes has not been examined. In the present study, using antibodies against extracellular epitopes on AMPA receptors, we were able to selectively visualize the trafficking of AMPA receptors expressed on the plasma membrane surface in live cells, and found that both GluRl and GluR2 receptors undergo clathrin-dependent constitutive endocytosis with a time constant of approximately 40 min, One would expect any perturbation of this constitutive endocytosis should lead to altered AMPA receptor expression on the plasma membrane. Surprisingly, we found that while hypertonic sucrose treatment and overexpression of the dominant negative dynamin mutant effectively blocked constitutive AMPA receptor endocytosis, both manipulations failed to alter the overall numbers of AMPA receptors on the cell surface. Among many possible explanations, the simplest one would be that there is a yet-to be determined mechanism tightly coup ling constitutiveendocytosis and insertion of AMPA receptors. Thus, factors affecting one pathway would also af3ect the other pathway through a feedback mechanism so as to ensure a balance between receptor insertion and removal. The result would produce a constant number of receptors expressed on the cell membrane, and hence a stable baseline synaptic transmission under conditions even where either constitutive exocytosis (Lledo et al., 1998) or endocytosis has been selectively altered. It should, however, be noted that Luscher et al have recently reported a run-down and a run-up of AMPA receptor-mediated EPSCs following postsynaptic injection of putative inhibitors for exocytosis and endocytosis, respectively, in hippocampd slices (Luscher et al, 1999)- Thus, some discrepancy between this and our own studies remain to be explained. In addition to the constitutive endocytosis, in the present work we also observed an insulin- stimulated, clathrin-dependent endocytosis of AMPA receptors in both transiently trdected HEK cells and cultured hippocampal neurons. There are several features associated with this insulin- regulated pathway that distinguish it firom constitutive endocytosis, the most prominent being its ability to rapidly reduce the number of AMPA receptors on the cell surface, thereby suppressing AMPA receptor-mediated responses. This indicates that insulin may be able to selectively facilitate the receptor endocytosis pathway, without affecting receptor insertion, thereby resetting the equilibrium between receptor endocytosis and insertion. This feature makes the regulated endocytosis a very attractive mechanism for regulating synaptic plasticity in models such as LTD. Another feature of the regulated endocytosis is its apparent subunit-speciticity. Unlike the constitutive pathway, which seems to be common across at least the GluRl and GluR2 receptor subunits, the insulin-stimulated endocytosis appears to be a GluR2-specific phenomenon, contingent on unique sequences present in the GluR2 intracellular carboxyl tail. A number of intracellular proteins have recently been revealed to be AMPA receptor-interacting proteins (e-g., the vesicle trafficking protein NSF, as well as PICK and GRIPS), and it is noteworthy that the majority of these proteins bind specifically to the GluR2 cytoplasmic-tail (Dong et al., 1997; Lin and Sheng, 1998; Nishimune et al., 1998;Xia et al., 1999) and some of these interactions seem to have important consequences on the stability ofthe cell-surface AMPA receptors (Noel et al, 1999; Luthi et al, 1999). For example, Noel et a1 have recently reported that inhibition of the NSF-GluR2 interaction by infusing either peptides corresponding to the binding domain of GluR2 (pep2m), or an anti-NSF antibody, into the postsynaptic neuron resulted in a rapid decrease in AMPA receptor-mediated synaptic transmission and a reduction in the response of cultured neurons to local AMPA application. Viral expression of pep2m was fomd to remove the majority of dace-expressed, GluR2-containing AMPA receptors (Noel et al. 2999), suggesting an important role for the NSF-GIuR2 specific interaction in regulating cell-surface expression of AMPA receptors. Since NSF is a protein that is known to be critically important for the hsion of synaptic vesicles to the plasma membrane, one obvious possibility is that through its binding to the GlWsubunit, NSF might exert a chaperone-like role in vesicle-mediated p lama membrane insertion of AMPA receptors. However, direct evidence for such a chaperone-like role of NSF in the plasma membrane insertion of AMPA receptors has yet to be adequately described. Interestingly, two recent studies (Luthi et al, 1999; Luscher et al, 1999) have found that postsynaptic application of pep2m, while producing LTD of AMPA EPSCs on its own, prevented the induction of LFS-induced LTD and conversely, the prior establishment of LFS LTD also reduced the ability of pep2m to induce EPSC depression in hippocampal CAI neurons (Liithi et al. 1999). As demonstrated in the present work, regulated clathrin-dependent endocytosis of AMPA receptors is also GluR2 subunit speciiic, and such enhanced endocytotic removal of AMPA receptors clearly contributes to the LTD. Thus, it might be equally plausible that the main bction of the NSF-GluR2 interaction is to prevent the regulated endocytosis of AMPA receptors by intedering with the interaction of the GluR2 tail with protein assemblies associated with clathrin- coated pits. It therefore would be very interesting to determine if the pep2rn-induced depression of AMPA EPSCs can be occluded by insulin-induced LTD and blocked by inhibiting postsynaptic clathrin-mediated endocytosis. Such experiments may provide fkther evidence for implicating the NSF-GluR2 interaction in preventing the initiation of regulated endocytosis, thereby stabilizing AMPA receptors on the postsynaptic plasma membrane. Thus, &om the evidence presented in this work, we speculate that there are two distinct clathrin--mediated endocytotic pathways involved in the removal of cell-surface AMPA receptors. The constitutive endocytosis pathway may be a common mechanism shared by all GluRs, and its major function would be to counteract the constitutive receptor insertion pathway, thereby ensuring that a constant number of AMPA receptors are expressed on the cell surface. In contrast, the regulated pathway is GluR2-specific, and is effective in rapidly regulating the cell-surface expression of the AMPA receptors. By interfering with this regulated pathway via their binding to the GluR2 carboxyl tail, a number of GluR2-specific interacting proteins may play an important role in controlling AMPA receptor density in the postsynaptic membrane. Therefore the regulated AMPA receptor endocytosis pathway may be important in the expression of certain forms of synaptic plasticity involving GluR2-containing AMPA receptors. It remains to be determined to what extent the constitutive and regulated pathways are distinct from each other in terms of their molecular mechanisms and how they converge on clathrin-mediated internalization as a common final step. CIathrin-mediated endocytosis in synaptic plasticity In hippocampal slice preparations, tetanic stimulation of the Schaf3er collateral and commissural inputs elicits homosynaptic long-lasting potentiation (LTP) of AMPA receptor-mediated synaptic transmission while stimulation of the same pathway at a low frequency induces homosynaptic LTD in hippocampal CAl neurons (Bear and Malenka, 1994). Both LTP and LTD have been proposed as -1 10- primary molecular and cellular substrates for the formation of learning and memory (Malenka and Nicoll, 1999). While it is generally accepted that, in their most commonly studied forms, the induction of both LTP and LTD is postsynaptic and dependent upon Ca" influx through activated NMDA receptors, the mechanisms underlying their expression remain hotly debated, and likely involve both a presynaptic component via alteration of transmitter release and a postsynaptic one through the modification of AMPA receptors (Malenka and Nicoll, 1999; Malinow, 1998). The "silent synapse" is a recent hypothesis that has been proposed to explain the postsynaptic locus of expression of LTP and LTD ciao et al., 1995; Malenka and Nicoll, 1999; Malinow, 1998). A silent synapse contains functional NMDA receptors but lacks functional AMPA receptors and becomes activated during the induction of LTP by the recruitment of functional postsynaptic AMPA receptors (Durand et al-,1996; Liao et al., 1995). By extrapolation, an active synapse may be silenced during LTD by the loss of f'unctional AMPA receptors (Carroll et al., 1999)- One key question that remains to be resolved is how the synapses are switched between active and silent states. Given that plasma membrane insertion and internalization of proteins are highly controlled, regulated exocytosis and endocytosis of AMPA receptors offer an attractive mechanism for activating and silencing synapses, and hence for the expression of LTP and LTD. Indeed, evidence is emerging to support a critical role for postsynaptic membrane fusion (Lledo et al., 1998) and the translocation of AMPA receptors (Shi et al., 1999) in the induction of LTP. In the present study, we report several observations that suggest the involvement of clathrin- mediated endocytotic removal of postsynaptic AMPA receptors in long-lasting synaptic depression. First, rapid endocytosis of AMPA receptors occurs in mature hippocampal neurons as implied by the increased association of AMPA receptors with the clathrin adaptor complex AP-2 following insulin -1 11- stimulation. Second, facilitation of clathin-mediated endocytosis of postsynaptic AMPA receptors by insulin produces long-term depression, and in some cases silencing, of AMPA receptor-mediated synaptic transmission. Third, insulin- and LFS-induced CA 1 LTDs mutually occlude each other, suggesting a common mechanism mediating the two forms of LTD. Finally, LFS-induced LTD is blocked by postsynaptic microinjection of the clathrin-mediated endocytosis inhibitor amphiphysin SH3 domain. Thus, together with the recent work from Luscher et al (Luscher et al, 1999), our results strongly suggest that a rapid, clathrin-mediated endocytotic removal of postsynaptic AMPA receptors may play an important role in the production of the LFS-induced hippocampal homosynaptic CAI LTD. The cIathrin-mediated endocytosis of postsynaptic AMPA receptors may not be limited to this well-characterized homosynaptic CA 1LTD. In a related study, Wang and Linden (Wang and Linden, 2000) found that insulMGF-I also produces a rapid and long-term depression of AMPA responses mediated by postsynaptic clathrin dependent endocytosis in cultured cerebellar neurons. The insulin/IGF-I induced depression of AMPA currents occludes a weIl-characterized cerebellar LTD, and the expression of this cerebella.LTD is blocked by inhibition of postsynaptic clathrin-dependent endocytosis (Wang and Linden, 1999). Additionally, Carroll et al have also recently reported a rapid activity-dependent reduction of postsynaptic AMPA receptors in a culture model of LTD induced by field stimulation (Carroll et al., 1999). Together, these data strongly suggest that the rapid removal of postsynaptic AMPA receptors mediated by a clathrin-dependent pathway may be a common final step in the expression of certain forms of LTD. How is the clathrin-dependent endocytosis of AMPA receptors stimulated by LTD-inducing protocols? As insulin is present in neurons and can be released from neurons in an activity-dependent manner, and aIso because insulin receptors are particularly concentrated in the postsynaptic density (Abbott et al,, 19991, one potential mechanism may involve the release of insulin from presynaptic terminals in response to LFS during LTD induction. Insulin may in turn activate its receptors on the postsynaptic neuron to facilitate clathrin-dependent endocytosis of AMPA receptors. However, postsynaptic injection of the insulin recep tor-neutralizing antibody, while blocking the insulin- induced depression of AMPA EPSCs, had little effect on either hippocampal homosynaptic LTD (Liu LD, et al, 2000 Neurosci. Soc. Abstract) or on cerebellar LTD (Wang and Linden, 1999). These results would suggest that insulin is not, by itself, directly involved in mediating the expression of these forms of LTD, but rather indicate that insulin and LTD-inducing stimuli may converge via distinct pathways to cause clathrin-mediated endocytosis of AMPA receptors. It seems likely that multiple signal transduction pathways exist for the regulation of AMPA receptor trafficking and the elucidation of these pathways promises to provide further insight into the molecular mechanisms of synaptic plasticity, and may ultimately provide mechanistic clues into the action of insulin in learning and memory. CWTERFOUR Activation of AMPA Receptor-associated PI3-kinase Induces LTP Through AMPA Receptor Plasma Membrane Insertion Part of the work is in press in Neuron and another paper derived fiom this is in preparation. H Y Man*, W Y Lu*, W Ju, W Trirnble, Y T Wang and John F MacDonald. Activation of synaptic NMDA receptors induces membrane insertion of new AMPA receptors and LTP in cultured hippocampal neurons. (* Equal contribution) Neuron, 200 1, in press. The work in fig 4.1,4.2,4.3,4.4,4.7,4.8,4.9,4.10,4.11 was mainly done by me, in which W Ju did some immunostaining in fig 4.8 and 4.9; Dr- Q H Wang did part of PI3K assays in fig 4.3 and 4.4; Dr Y M Lu did 4.1 1(B); Dr G Ahmadian (Dr Y T Wang7s lab) did 4.1 (E). Dr. W Y Lu @r J MacDonald's Iab) did the work in fig 4.5,4.6 and part of fig 4.10 (A, B). 4.1 Abstract AMPA receptors mediate the majority of excitatory synaptic transmission and are the final targets for the expression of long-term potentiation (LTP). Although membrane insertion of AMPA receptors is an attractive cellular model for LTP expression has been intensively investigated and is gaining increasing support in recent years, the underlying mechanisms for AMPA receptor translocation are still unknown. We show that phosphoinositide 3-kinase (PI3K) directly associates with -AMPA receptors and they co-localize in hippocampal neurons. Selective activation of synaptic NMDA receptors by the NMDA receptor co-agonist glycine activated the AMPA receptor-associated PI3K and induced long-term potentiation (LTP) of AMPA mini-EPSCs. Furthermore, brief treatment with glycine induced an increase in cell-surface expression of AMPA receptors and this increase was due to an enhancement in receptor plasma membrane insertion, which is also the mechanism underIying the glycine-induced LTP. Moreover, PUK activation was also shown to be necessary in the expression of homosynaptic LTP in hippocampal brain slices. These data provide evidence for an NMDA receptor--PUK-AMPA receptor translocation pathway in LTP generation. 4.2 Introduction Long-term potentiation (LTP) of synaptic transmission as a cellular model for learning and memory has been extensively investigated (Collingridge and Bliss, 1995; Larkman and Jack, 1995). Protein phosphorylation of AMPA receptors by kinases such as CaMKII is believed to be crucial in LTP (Barria et al., 1997; Blitzer et al., 1998; Pettit et al., 1993; Stevens et al., 1994), but new findings have shown that CaMKa-dependent AMPA receptor phosphorylation may be not necessary in LTP expression (Hayashi et al., 2000). A recently proposed AMPA receptor insertion hypothesis in LTP expression has gained support from several studies (Shi et al., 1999)- Studies showed that postsynaptic injection of some putative inhibitors that are believed to block membrane fusion, impaired LTP, while application of a membrane firsion protein SNAP into the postsynaptic cell enhanced AMPA receptor-mediated synaptic transmission and occluded LTP (Lledo et al., 1998; Luscher et d., 1999). More direct evidence of AMPA receptor insertion showed that in cultured hippocampal slices, transiently transfected with green fluorescent protein (GFP)-tagged GluRl subunits, an LTP-inducing protocol translocated GFP-GluRI to synaptic sites and inserted the subunit onto postsynaptic membranes (Shi et al., 1999). Studies have shown that AMPA receptors have a relatively fast constitutive internalization and presumably the same rate of exocytosis (see chapter 3), and alterations of these processes will change the cell surface expression of AMPA receptors. A striking feature is that, while AMPA receptors on the plasma membrane are quite mobile, NMDA receptors seem very stable (Carroll et al., 1999b; Carroll et al., 1999a; Lissin et al., 1999) (chapter 3). This is consistent with the finding that during LTP induction, tetanus-stimulation leads to long-term potentiation of only the AMPA component but not the NMDA component. The specificity in activity-dependent trafficking of AMPA, but not NMDA, receptors could be due to the nature of the receptor protein molecules. However, an alternative plausible explanation is that molecules involved in protein trafficking could directly associate with AMPA receptors and affect AMPA receptor trafficking. This would not only provide the receptor specificity but also greatly enhance the efficiency of receptor trafficking by shortening the signaling pathway. -1l6- Figure 4.1 Functional PI3-kinase (PI3K) binds to AMPA receptors. (A) AMPA receptors associate with PI3K in vivo. Using rat brain hoppocampal extracts, PI3K was immunoprecipitated with a polyclonal antibody against the p85 submit of PI3K and the resulting immunoprecipitates were probed with an anti-GluR2 antibody. Note that P85 co- immunoprecipitated with AMPA receptor GluR2 subunit while a rabbit IgG control did not pull down GluR2. (B) P13K specifically associates with AMPA but not NMDA receptors. Immunoprecipitates by anti- P85 antibodies were first probed with anti-GlW antibody. The membrane was stripped and re- probed with an antibody against NMDA receptor NRl subunit. (C) The AMPA receptor-associated PI3K kinase is functional. PI3K activity assays performed on GluR2 imrnunoprecipitates showed that anti-GlW antibody pulled down lipid kinase activities which could be blocked by a PI3K specific inhibitor wortrnannin, indicating an association of active, hctional PI3K with AMPA receptors. @) The C-termini of AMPA receptor subunits directly bind to PI3K in vib-o. GluRl and GluR2 c- terminus GST fusion proteins were incubated with purified active P13K and the GST fkion protein precipitates were used for PIXkinase assays. Assays showed that both GluRl and GluR2 c-termini directly interact with PI3K. (E) PI3K interacts with AMPA receptors through its p85 regulatory subunits. The p85 subunit of PDK was translated in viho, labeled with 35s-methionine. The reaction produced a single protein band with a molecular weight of 8SkD (left). The in viho translated p85 was incubated with GluRl and GluR.2 c-tenninus GST fhion proteins and both hion proteins pulled down radioactively- labeled p85 protein. Figure 4.1 Functional PB-kinase (PI3K) binds to AMPA receptors Figure 4.2 PI3-kinase colocalizes with AMPA receptors in cultured hippocampal neurons Cultured hippocampal neurons were immunostained with an anti-P85 primary and Cy3-conjugated secondary antibodies, with or without a subsequent staining of AMPA receptors by an anti-Glum primary and an FITC conjugated secondary antibody. Confocal images showed that P13K distributed in the cytosolic domain in the soma (A, left) and in the dendrites (A, right) where PI3K displayed a clustered pattern. Double staining of MAreceptors and PI3K showed a high degree of colocalization, consistent with the biochemical data showing an in vivo association between AMPA receptors and PI3K (B). Figure 4.2 PI3-kinase colocalizes with AMPA receptors in cultured hippocampal neuron Overlap Figure 4.3 Activation of synaptic NMDA receptors activates the AMPA receptor- associated P13K. Rat brain hippocampal slices were treated with 50 pM NMDA or 50 pM NMDA receptor open channel blocker MK-801 and the slice lysates were used for immunoprecipitating GluR2 or p85. (A) Blockade of synaptic NMDA receptors inhibit AMPA receptor (AMPAR)-associated PI3K activities. PI3K kinase assays were done on protein complexes immunoprecipitated by anti-GluR.2 (left panel) or anti-PI3K (right panel) antibodies. NMDA treatment only induced slight inhibition but MK-80 1greatly reduced the AMPA receptor-associated PI3K activity (left). No obvious change in PI3K activity was observed when the general cellular PI3K was pulled down by anti-P85 antibodies (right). The complete block of baseactivity by wortmaninn confirmed the specificity of the PI3K assay. (B) Activities of AMPAR-associated PI3K are potentiated by selective activation of the synaptic NMDA receptors. Cells were treated with NMDA receptor co-agonist glycine (200pM,3 mins) to activate the synaptic NMDA receptors and the subsequent PI3K assay showed a dramatic increase in AMPAR-associated (IP: GluR2), but not the total (IP: PIX), PI3K activity. The NMDA receptor dependence of this glycine effect was confinned by a complete abolish of the enhancement by NMDA receptor antagonist APV. (C)Treatments do not change the amount of protein in the precipitates. A western blot showing that the anti-P8S immoprecipitates fkom cultured neuron lysates, pre-treated with NMDA, MK-80 1 and glycine, contained similar amount of AMPA receptor GluR2 subunits. @) A bar graph summarizes the PI3K assays (~2to 4)- Figure 4.3 Activation of synaptic NMDA receptors activates the AMPA receptor-associated PI3K IP: GluR2 IP: P85 - -- IP: GluR2 IP: P13K IP: P85 IB: GluR2 P13-kinase (P13K) is a lipid kinase composed of two subunits: an 85kD regulatory domain (p85) which interacts with proteins through its diverse motifs, and a 1lOkD catalytic domain (p 1 10) which can phosphorylate phosphoinositide at the 3 position of the inositol ring (Vanhaesebroeck et al., 1996; Yu et al., l998b). PI3K was fmst characterized as an important molecule for protein sorting and transportation and now PDK as well as its lipid products have been found to be involved in vesicle formation and traEicking of a variety of proteins (l3rown et al., 1995; Carpenter and CantIey, 1996b; Cheatham et al., 1994; Chibalin et al., 1998). It is well known that PDK is required in the translocation of glucose transport 4 &om the cytosol to the plasma membrane during insulin signaling(Ho1man aad Cushman, 1994; Kotani et al., 1995; Martin et al., 1996; Quon et al., 1995a; Quon et al., 1995b). It was also found that PDGF, through activation of PI3K, increases the cell surface expression of a glutamate transporter EAA (McIlhinney et al., 1996b). 4.3 Results 4.3.1 PI3K directly interacts with AMPA receptors To investigate whether AMPA receptors bind to PI3K, PI3K was immunoprecipitated from rat hippocampal brain slice lysate and the resulting protein complex was probed with an anti-GluR2 subunit antibody. PI3K co-immunoprecipited with AMPA receptors, indicating an association between AMPA receptor and PI3K (Figure 4.1A). The interaction of PI3K appears to be specific to AMPA receptors, as PDK failed to co-immunoprecipitate with NMDA receptors (Figure 4.1B). To test whether the AMPA receptor-associated PI3K is functional, PI3K kinase assays were done on GluR2 precipitated samples. The AMPA receptor-associated protein complexes produced lipid -123- products in the assay, and this reaction was completely blocked when the sample was pre-incubated with a specific PPUK inhibitor wortmannin, indicating that AMPA receptors associated with a hctional PI3K complex (Figure 4.1 C). For the AMPA receptor iriteracting proteins found so far, all were shown to interact with the intracellula. c-termini of different AMPA receptor subunits (Dong et al., 1997; Lin & Sheng 1998; Xia et al., 1999). To examine whether the interaction of AMPA receptor with PI3K is direct, possibly through its c-terminus, we incubated the purified constitutively active PI3K with GST fusion proteins of the c-termini of GIuRl and GluR2 subunits. Both GluRl and GluR2 subunit c-termini pulled down PI3K activity (Figure 4.1D). The p85 regulatory subunit of PI3K contains multiple protein-protein interaction motifs and thus most, if not all, protein interactions with PUK are through this subunit. To test whether it is P85 that binds to GluR c-termini, the in vih-o translated radioactively-labeled pP85 was incubated with GluR c-terminus-GST fusion proteins. As shown in Figure 4.lE, the in vitro translation produced a single protein with the molecular weight corresponding to that of p8S (Figure 4.1E, Left). Furthermore, GluRlct and GLuR2ct pulled down similar amount of p85, indicating that both AMPA receptor subunits can bind to p85 directly with similar affinity (Figure 4.1E right). 4.3.2 PI3K colocalizes with AMPA receptors at the synaptic site If PI3K indeed directly binds to AMPA receptors in vivo, we expect that a co-localization of the two components should be observed in native neurons. In cultured hippocampal neurons the normal distribution of P13K was first investigated. Irnrnunostaining showed that PI3K distributed in the - 124- whole hippocampal neuron, including the soma and the dendrites (Figure 4.2A). Interestingly, PI3K on the dendrites was not uniform, but in a clustered or dotted pattern7 suggesting a synaptic localization (Figure 4.2A, right), To demonstrate the relationship of AMPA receptors with PI3K in their cellular locdization, GLuR2 and PI3K subunit P85 were sequentially imrnunostained under permeabilized conditions. In support of an interaction between AMPA receptors and PI3K, confocal images showed that the majority of dendritic PI3K co-localized with AMPA receptor GluR2 subunits Fig 4.2B). 4.3.3 Activation of synaptic NMDA receptors potentiates the activity of AMPA receptor- associated PI3K We next explored ways by which activities of AMPA receptor (AMPAR)-associated PI3K were regulated, Because an enrichment of PI3K in glutarnatergic synaptic sites was observed from irnmunostaining, effects of AMPA and M4DA receptor activities on the AMPAR-associated PI3K was examined, GluR2 subunit was imrnunoprecipitated from rat brain slice extracts treated with AMPA, NMDA or MK-80 1, and PI3K activity in the precipitated GluR2 complexes was tested. We found that AMPA and NMDA treatments showed little effect on PI3K activity. However, MK-80 1, an NMDA receptor open channel blocker, dramatically reduced the activity of AMPAR-associated PI3K (Figure 4.3A). Under physiological conditions, only the synaptic NMDA receptors could be activated by presynaptically released glutamate, or blocked by MK-80 1. Therefore, the inhibition of PI3K by MK-801 was due to a blockade of synaptic NMDA receptors. In an opposite way, we selectively activated the synaptic NMDA receptors by applying a NMDA receptor co-agonist glycine. Figure 4.4 Activation of synaptic NMDA receptors induces translocation and phosphorylation of Akt. (A) Glycine treatment causes translocation of GFP-Akt- Confocal images were collected in live hippocampal neurons which were transiently transfected with GFP-Akt. The GFP-Akt was distributed throughout the cell with clustered localization in the peripheral dendrites (A, Control). 200 pM Glycine treatment induced a rapid redistribution of GFP-Akt 1min after glycine application and the maximal change was reached within 5-10 mins. As shown in A (Glycine), the arrows and circles indicate sites or areas where changes in the intensity or localization of GFP-Akt can be observed. (B) Glycine induces activation (phosphorylation) of endogenous Akt in cultured neurons. Cultured hippocampal neurons were stained with antibodies against the phosphorylated Ser473 of Akt. Under control conditions, phosphorylated Akt was low and mainly in the somatic area (B, Control). Glycine treatment markedly increased the immmosignal for phospho-Ser473 Akt staining in both the soma and the dendrites (B, Glycine). This glycine-induced changes in Akt phosphorylation were totally blocked by the NMDA receptor antagonist APV (B, Gly+APV), indicating that the activation of Akt by glycine is NMDA receptor-dependent. Figure 4.4 Activation of synaptic NMDA receptors induces translocation and phosphorylation of Akt Control Glycine Control Glycine APV+g l ycine Figure 4.5 Glycine induces long-term potentiation (LTP) of AMPA-miniEPSCs through activation of PI3K in cultured neurons (A and B) Glycine causes LTP in cultured neurons. AMPA receptor-mediated miniatures (AMPA- mPSCs)were recorded in cultured hippocampal neurons in the presence of TIX and the GABA, receptor antagonist bicuculline. Bath application of glycine (200 pM) for 3 mins increased both the amplitude and fkequency of AMPA-rniniEPSCs which Lasted for at least 30 mins. Typical traces are shown in A and the time course is shown in B . (C and D) PI3K is required for glycine-induced LTP. Glycine treatment was given during AMPA- miniEPSCs recording with or without the PI3K specific inbibitor wortma.(100nM) in the perfusion solution. While glycine still induced LTP under control conditions,, wormaninn completely abolished the glycine effect. Figure 4.5 Glycine induces long-term potentiation (LTP) of AMPA-miniEPSCs through PI3K activation in cultured neurons Control Glycine Control g Iycine Control Glycine+wort Frequency Amplitude Z glycine 0.8 ...... , Time (min) Figure 4.6 Intracellular application of active PUK potentiates AMPA- Constitutively active PI3K was applied intracellularly through the recording pipette and AMPA- minis were recorded immediately after breakthrough in cultured hippocampal neurons. Mer 1min, AMPA-miniEPSCs began to show increases in both amplitude and fiequency. In contrast, no changes were observed when the boiled inactive PUK was applied. (A) An sample recording 3 and 15 mins after whole-cell codigurattion demonstrates changes in AMPA-miniEPSCs by supplement of active PUK in the recording solution. (B) Average traces showing that the active PI3K, but not the boiled one, increased the amplitude of the AMPA-miniEPS Cs (C)Statistical bar graphs demonstratingthat active PUK increases both the amplitude and fiequency of AMPA-miniEPSCs, while the boiled inactive PI3K has no effect. Figure 4.6 IntraceIIular application of active PDK potentiates AMPA-miniEPSCs 3c min 15 rnin 1 rnin I5min Y Active P 13K T Dead P13K In the presence of glycine receptor antagonist strychnine (5 pM), 3-5 min glycine treatment markedly enhanced the AMPAR-associated PUK activity. The NMDA receptor-dependence in the glycine- induced activation of PUK was conlirmed by a complete blockade of the glycine effect with NMDA receptor antagonist APV (Figure 4.3B). Notably, neither MK-801 nor glycine changed the total activity of PDK, indicating a specific modulation on only the AMPAR-associated PI3K (figure 4.3A and B). Also, changes of PI3K activities were not due to alterations in the amount of associated protein, for similar amounts of GluR2 were co-precipitated with PI3K following different treatments (Figure 4.3C). 4.3.4 NMDA receptor-mediated activation of AMPAR-associated PI3M leads to translocation and activation of Akt Akt is a downstream effector of PI3K. binds specifically with the membrane-localized Lipid products of Pf3K through its PH domain and is then phophorylated and activated purgering and Coffer, I99S;Eves et al., 1998;Kohn et al., 1996;Marte and Downward, 1997). Therefore, a direct consequence of PI3K activation will be a translocation of Akt £tom the cytosol to the plasma membrane, a phenomenon which was used as an index of PUK activation. In order to further study the synaptic NMDA receptor-dependent activation of PI3K in native neurona at a subcellular resolution, we transiently transfected cultured hippocampal neuron with green fluorescent protein (GFP)-tagged Akt, so that its distributionand translocation could be directly visualized by fluorescent microscopy. As shown in Figure 4.4A, GFP-Akt localized throughout the transfected neuron, with some patchy or clustered pattern in the dendrites, probably reflecting some basal activities of synaptic PI3K (Figure 4.4A). After activating the synaptic NMDA receptors by glycine, a quick redistribution of GFP-Akt was observed as soon as lmin after glycine application. Most existing Akt clusters became brighter in intensity and sharper in edges, suggesting a synaptic enrichment of Akt due to increases in local lipid products of PIX. Meanwhile some areas became dimmer, presumably places which contain no synapses or non-glutamatergic synapses (Figure 4.4A and B). The overall change reached a maximum 5 mins after treatment and some recovery was observed after 10 rnins. Activation of Akt requires phosporylation on its Ser473 and/or -308 sites (Burgering and Coffer, 1995; Kandel and Hay, 1999), which usually happens after translocation of Akt to the sub- plasma membrane domain through its interaction with PDK products. To investigate whether glycine treatment can phosphorylate and activate Akt in native neurons, the active Akt was detected by an antibody specifically against the phosphorylated Akt at Ser473. Immunostainings showed that the basal level of phopho-Akt was very low and only in the somatic area (Figure 4.4B, Control). However, after 3-5 mins glycine treatment (200 pM), all the dendrites were clearly immunolabeled, in addition to a stronger staining in the cell body, suggesting a dramatic increase in Akt phosphorytation at Ser473 and in its activity (Figure 4.4B, Glycine). Consistent with an NMDA receptor dependence, the glycine-induced enhancement in Akt phosphorylation was completely abolished by APV (Figure 4.4B, APV+glycine). 4.3-5 Selective activation of synaptic NMDA receptors induces long-term potentiation (LTP) of AMPA receptor-mediated miniEPSCs through activation of PI3K in cultured neurons To investigate the physiological significance of the association of PUK and AMPA receptors, we investigated the effect of activating the AMPAR-associated PI3K on AMPAR-mediated miniature EPSCs (AMP A-miniEPSCs). In cultured hippocampal neurons, AMPA-miniEPSCs were recorded in the presence of tetrodotoxin (T'TX) to block action potentials and bicuculline to block GABA, receptor-mediated inhibitory synaptic transmission. Because activation of synaptic NMDA receptors by the NMDA receptor co-agonist glycine specifically activated the AMPAR-associated PUK, we studied whether glycine treatment affect the AMPA receptor-mediated synaptic transmission. We found that 200pM glycine treatment for 3 mins in the presence of strychnine (glycine receptor antagonist) increased both the amplitude and the frequency of AMPA-miniEPSCs, which lasted for at least one hour (Figure 4.5A and B). Similar to the standard tetanus stimulation-induced CAI LTP, this glycine-induced long-term potentiation (LTP) was also NMDA receptor- and calcium-dependent (data not shown). To test the involvement of PI3K in this glycine-induced LTP, a PI3K inhibitor wortmannin (100 nM) was applied in the pehsion solution. At this concentration, wortmannin is specific for PI3K and completely blocked the glycine-induced LTP, indicating that the potentiation ofAMPA-miniEPSCs by glycine was through activation of PI3K (Figure 4.5C and D). To fhh confirm the involvement of P13K in the long-term potentiation of the AMPA receptor-mediated synaptic transmission, purified active PI3K was applied intracellularly through the recording pipette and recordings began 0.5 min after the whole-cell configuration- AMPA-rniniEPSCs started to increase 1 min after recording and reached its plateau at 10 min. Both the amplitude and the frequency were potentiated and the effect lasted for the time of recording, usually 30 mins but 1 hour in some cells (Figure 4.6). As a negative control, the boiled inactive PI3 K changed neither the amplitude nor the frequency of the miniEPSCs (Figure 4.6, B,C). 4.3.6 Synaptic NMDA receptor activation increases cell surface expression of AMPA receptors in cultured neurons Evidence has shown that receptor trafficking, presumably through receptor plasma membrane insertion, contributes to LTP expression (Luscher et al., 1999;Shi et al., 1999). However, so far evidence to show recruitment of AMPA receptors to the plasma membrane during LTP is either indirect or has come from studies of non-native AMPA receptors. Because it is extremely difficult to study receptor traflicking in brain slice preparations in terms of LTP, we took advantage of our glycine-induced-LTP model in cultured neurons and investigated the involvement and the mechanisms of AMPA receptor trafficking in LTP expression. In cultured hippocampal neurons, colorimetric assays were performed to quantifL the AMPA receptor cell-surface expression rate (Figure 4-7A). Assays showed that under normal conditions, 60% of AMPA receptors were on the plasma membrane and the other 40% were localized intracellularly. 3-5 rnin glycine treatment (200 pM) increased AMPA receptor cell-surface expression fkom 58.20.1% to 73&0.04% (n=10) (Figure 4.7B). The increase was blocked if APV was added during glycine treatment, indicating the glycine- induced translocation of AMPA receptors was through the activation of NMDA receptors (Figure 4.7B). Because glycine can activate AMPA receptor-associated PUK, and glycine-induced LTP can be blocked by inhibition of PI3K, we expected a requirement of PI3K activity in this glycine-induced AMPA receptor translocation. Indeed, colorimetric assays showed that glycine-induced increases in AMPA receptor cell-surface expression was markedly reduced when the PI3K activity was inhibited Figure 4.7 Glycine treatment induces a NMDA receptor-dependent increase in cell-surface expression of AMPA receptors in cultured hippocampal neurons. (A) The cartoon illustrates the colorimetric assay used to quantify the proportion of AMPA receptors expressed on the cell plasma membrane surface. Receptors expressed on the cell-surface or in the entire cell were respectively labeied with a primary antibody against the extracellularamino terminus of GluR2 subunit and an HRP-conjugated secondary antibody under non-permeant (surface) and permeant conditions. The proportion of cell-surface AMPA receptors was then determined using the ratio of absorbance readings obtained under non-permeant versus permeant conditions. (B) Colorimetric assays were performed on cultured hippocampal neurons treated with glycine (200 pM; 3 min) with or without NMDA receptor antagonist APV (50pM). Under control conditions, about 60% of AMPA receptors are on the plasma membrane. Glycine treatment significantly increases the proportion of cell-swface AMPA receptors and this effect is blocked by AP5, and this glycine effect was markedly inhibited by wortmannin, indicating an involvement of PI3K in mediating the glycine-induced increase in AMPA receptor cell-surface expression. (* p<0.05). Figure 4.7 Glycine treatment induces an NMDA receptor- and PI3K-dependent increase in cell-surface expression of AMPA receptor in cultured hippocampal neurons A Antibody Receptor Non-permeabilized Permeabilized Con Gly APV APV Con Gly +sly Figure 4.8 Glycine increases AMPA receptor clusters at synapses. (A) Control and glycine treated cultured mouse hippocampal neurons were sequentially stained for GluRl (red) and synaptophysin (green) under non-permeant and permeant conditions, respectively. Individual (GluR2, green; Synaptophy sin, red) and superimposed (Overlay) confocal images show that GlW receptor clusters are concentrated at a subfkaction of synapses identified with synaptophysin (Control). Glycine treatment increased receptor clusters at the synapses (Glycine). @) Pseudo-color intensity mapping of the randomly selected areas in the confocal images was used to quantify GIuR2 clusters and their relationship with synaptophysin labeled synapses. (C)Bar graph showing that glycine treatments increased the GluR2/Synaptophysin co-locabtion ratio. *** p>0.00 1, paired Student's &-test. Figure 4.8 Glycine increases AMPA receptor clusters at synapses A Synaptophysin GIuR2 Overlay control Glycine Figure 4.9 Glycine facilitates AMPA receptor plasma membrane insertion in cultured hippocampal neurons. In cultured hippocampal neurons, the existing cell-surface AMPA receptors were fist blocked by an anti-GluR1 antibody and a cold secondary antibody. The newly inserted AMPA receptors (red) and synaptophysin (green) were then sequentially stained under non-permeant and permeant conditions respectively. After 15 mins at room temperature, cells show clear AMPA receptor expression on the cell surface (A, Control), suggesting a rapid constitutive insertion of AMPA receptors into the plasma membrane. Glycine (200 pM; 3 mins) treatment facilitated AMPA receptor insertion (A, treatment) and thereby increased the GluR/synaptophysin colocdization rate (C). Double staining of cells immediately after pre-blockade shows that the pre-blocking protocal essentially eliminated GluRl (GluRl) iabeling, while having no effect on synaptophysin staining (Syn), codinning the specific blocking of the pre-existing cell-surface AMPA receptors (E3). Quantification of the glycine-induced increase in AMPA receptor insertion using colorimetric assay is shown in (D). Figure 4.9 Glycine facilitates AMPA receptor plasma membrane insertion in cultured hippocampal neurons A Synaptophysin GluR2 Overlay with wortmannin (Figure 4.7C), indicating that the glycine-induced increase in AMPA receptor cell- surface expression is mainly mediated via PI3K. 4.3.7 Glycine treatment increases AMPA receptor synaptic localization Glutamate receptors, like other ionotropic neurotransmitter receptors, are enriched in postsynaptic membranes. The increased amount of cell-surface AMPA receptors induced by glycine may occur at synaptic or non-synaptic sites on the plasma membrane, though potentiation of AMPA-miniEPSCs by glycine suggests an increase at synaptic sites. To investigate these possibilities, native AMPA receptors expressed on neuronal plasma membrane surfaces were labeled with a polyclond antibody against the amino-tenninal extracellular epitope of GluRl under non-permeant conditions. Subsequent staining of the same neurons for the presynaptic marker protein synaptophysin under permeant conditions identified synapses. With the use of immunofluorescent confocal microscopy, we found that cell-surface AMPA receptors formed numerous small clusters that co-localized with a subpopulation of anti-synaptophysin labeled synapses in he control cultures (Figure 4.8A). However, treatment of cultures with glycine (200 pM for 3 rnins) increased both intensity and number of AMPA receptor clusters at synapses (Figure 4.8A), and increased GluRlIsynaptophysin colocalizationrate from 48.4*2.0% (n=39) to 65.8*1.9% (n=61,p active ones, we also double-stained cell-sdce WAR1and GluR2 subunits. As expected, glycine treatment did not alter the number of NMDA receptor dusters (7.Wl.7 vs 7.6H -8 clusters110 pm - 142- dendritic membrane in control and glycine treated neurons, respectively, P<0.05; n=36), but it increased the GluR2/NMDARl colocalization rate from 67.6*2,7% (~39)in untreated controls to 7722.3% (1145,~60.05) in glycine-treated neurons (data not shown). Thus, glycine treatment is capable of recruiting AMPA receptors into NMDA receptor-only synapses, thereby switching some silent synapses into active ones which likely contributes both to the increase in frequency and amplitude of AMP A-rniniEP SCs. 4.3.8 Glycine promotes AMPA receptor insertion in cultured neurons Although the increase in cell surface expression of AMPA receptors is presumably through receptor insertion, the same effect could also be achieved through slowing down the constitutive receptor internalization. To confirm the role of receptor insertion in the glycine-induced increase in AMPA receptor surface expression, pre-blocking experiments were performed. Briefly, cultured neurons were incubated with an N-terminus specific anti-AMPA receptor antibody and a cold secondary IgG to block the cell surface GluRs, Cells were treated with glycine at room temperature and the newly inserted cell-surface AMPA receptors were then labeled with the same primary antibody together with the staining of synaptophysin, a commonly used synaptic marker protein. Under this condition, the immunostaining will only label the newly inserted AMPA receptors on the plasma membrane. 15 mins after transferring cells to room temperature, a clear, though still weak, immunostaining of the new AMPA receptors was observed (Figure 4.9A). Glycine treatment (200 pM, 3 mins) increased both the cluster number and its intensity. As a result of increased receptor insertion at the postsynaptic membrane, the AMPA receptorkynaptophysin colocalization rate was increased from 32.6*2.6% to 59.3*2.5% 12 mins after glycine treatment (n=3 8, p Consistent with imrnunofluorescent staining, the assay revealed an increase in AMPA receptor insertion by glycine of about 10% (Figure 4.9D, n=12). These results suggest that there is a rapid constitutive synaptic plasma membrane insertion of AMPA receptors, and that facilitation of this receptor insertion contributes to glycine-induced LTP 4.3.9 Glycine-induced LTP is abolished by blocking AMPA receptor exocytosis Protein plasma membrane insertion is presumably by vesicle-mediated membrane fusion through SNARE protein interactions, that is, SNAP-25 and syntaxin on the plasma membrane bind to VAMP on the cargo-bearing vesicles (Bonifacino et al., 1996; Martin, 1997;Schekrnan and Orci, 1996; Schmid, 1997), and so, plasma membrane insertion will be prevented by disrupting SNARE protein interactions. It is known that some toxins can specifically destroy different SNARE proteins and thus affect vesicle membrane fusion, and tetanus toxin (TeTx), which specifically cleaves VAMP, has been widely used to block protein exocytosis (Gaisano et al., 1994; Hajduch et al., 1997). For example, at the pre-synaptic terminal, TeTx has been shown to block synaptic vesicle hion and neurotransmitter release, and therefore inhibits synaptic transmission (Hua et al., 1998; Pellizari et al., 1998). To investigate whether the glycine-induced LTP is through receptor plasma membrane insertion, a purified TeTx light chain, which is responsible for destroying VAMP, was applied intracellularly (1 50nM) through the recording pipette in cultured hippocampal neurons during AMPA- - 144- Figure 4.10 Intracellular injection of tetanus toxin blocks both glycine-induced LTP and glycine-induced increase in AMPA receptor cell-surface expression (A, B) TeTx blocks the glycine-induced LTP. TeTx was applied into the postsynaptic neuron in hippocampal culture through the recording pipette. An example recording shows that the glycine- induced potentiation of AMPA-rniniEPSCs was completely abolished when TeTx was added in the recording solution (A). After glycine treatment, neither the amplitude nor the frequency of the AMPA-miniEPSCs was changed (B). (C,D) TeTx blocks glycine-induced increases in AMPA receptor insertion. Cultured hippocampal neurons were microinjected with active TeTx (C, a and b) or boiled inactive TeTx (C, c), along with Lucifer yellow as a marker (green). The pre-existing cell-surface AMPA receptors were first blocked by an anti-GluR1 antibody and a cold secondary antibody. After blocking, the newly expressed cell-surface AMPA receptors were imrnunostained (red) with or without glycine treatment (C). TeTx did not block the constitutive AMPA receptor insertion, but blocked the glycine-induced facilitation of AMPA receptor cell-surface insertion (C). Quantification of cell-surface receptor clusters is shown in the histogram in 0). Figure 4.10 Intracellular injection of tetanus toxin blocks both glycine-induced LTP and glycine-induced increase in AMPA receptor cell-surface expression Control 7 C Glycine 'P E IGlycine 6- T 0 T F \ E? 4- Q) CI 2 2- 0 Vn Amplitude Frequency TeTx Boiled TeTx Figure 4.11 Intracellular application of active P13K potentiates AMPA receptor- mediated EPSCs in hippocampal CAI neurons (A) AMPA receptor-mediated synaptic transmission induced by electrical stimulation of the ShaiXer/collateral fibers were recorded in CAI neurons in brain slices. The amplitude of AMPA- EPSCs was stable under basal condition. Application of active PUK into the postsynaptic neuron through the recording pipette increased the EPSCs by about 80% while addition of the boiled inactive PDK in the recording solution failed to change AMPA-EPSCs, indicating an important role of PI3 K activity in potentiating AMPA receptor-mediated excitatory synaptic transmissions in CAI neurons. (£3) Homosynaptic CAI LTP is blocked by inhibition of PI3K activity. In hippocarnpal slices, LTP is consistently induced by standard high fiequency stimulation (1 OOHz, 1s, two trains) (Control). Bath application of a specific P13K inhibitor LY294002 (10 pM) dramatically inhibited LTP amplitude, indicating a critical role of endogenous PI3K in hippocarnpal LTP expression. Figure 4.11 PDK is required in LTP expression in hippocampal CAI neurons A A Control 2-51 Boiled PI3 base Time (min) Control O LY Time (min) miniEPSC recording. Interestingly, TeTx had no obvious effect on basal synaptic transmission for at Iease 3 0 mins, but completely blocked glycine-induced LTP (Figure 4.1 0A and B), indicating that membrane insertion is involved in LTP expression and different mechanisms may be applied in the regulation of receptor membrane trafEcking under basal vs LTP-inducing conditions. Application of TeTx will non-specifically affect plasma membrane protein insertion, so the blocking of glycine-induced LTP by TeTx might not be a direct consequence of blocking AMPA receptor insertion, but rather is the consequence of a block of exocytosis of other proteins which are critical for up-regulation of AMPA receptor functions. To confirm that TeTx indeed blocks AMPA receptor cell-surface expression, immunostaining experiments were performed to directly address this issue. Cultured hippocampal neurons were microinjected with TeTx light chain (200 nM) along with Lucifer Yellow (as a marker). Each injection took about 10 mins until the cells were brightly filled with Lucifer Yellow. The existing cell-surface AMPA receptors were first pre-blocked at 4°C with anti-GluR2 N-terminus antibody (see Methods), cells were then transferred to room temperature, treated with glycine (200 pM, 3 min) and then incubated in glycine-fiee ECS for 25 mins at room temperature. The subsequent fluorescent imunostaining using the same anti-GluR2 antibody will only label the newly inserted AMPA receptors. Surprisingly, TeTx injection did not block basal AMPA receptor insertion (Figure 4.1 OC,a). However, glycine induced enhancement of cell-surface receptor insertion was completely blocked by TeTx (Figure 4.10Cb and D). This observation is consistent with the electrophysiological data showing that intracellular application of TeTx did not affect basal synaptic transmission but blocked glycine-induced LTP. As acontrol, cells injected with boiled inactive TeTx still showed glycine-induced increases in AMPA receptor cell-surface insertion (Figure 4. l OCc and D). 4.3.10 PI3K is required in the expression of homosynaptic LTP in hippocampal slice CAI neurons In cultured hippocampal neurons, we found that glycine, in an NMDA receptor- and PDK-dependent manner, induces LTP through AMPA receptor plasma membrane insertion. LTP as a cellular model for leaming and memory is mainly studied in brain slice, a preparation closer to intact brain. To test whether PI3K is involved in LTP in brain slices, constitutivelyactive PI3K was applied intracellularly in CAI neurons of hippocampal slices through the recording pipette, and AMPA receptor-mediated EPSCs in Shaffedcollateral-CAI sycapses were recorded. Under this condition, the amplitudes of AMPA-EPSCs gradually increased fiom the start of recording and then maintained at an elevated level after about 8 mins with an average increase of 80% (Figure 4.1 lA, n=17). As a control, application of boiled PI3K which lost its kinase activity failed to change EPSCs (Figure 4.1 1A, n=7). To test whether PI3K is required in homosynaptic CAI LTP in hippocampal slices, field recordings were performed in CA1 neurons. Under control condition, 100Hz tetanus stimulations reliably induced LTP which lasted for at least 1 hour. Intracellular application of the PIXinhibitor wortmanuin (200 nM) dramatically inhibited LTE' fiom 150% to 50%, indicating that activation of PI3K is a necessary step in the cellular signalling pathway leading to LTP (Figure 4.11B). 4.4 Discussion Taken together, this work showed that the AMPA receptor-associated PDK, when activated by activities of synaptic NMDA receptors, translocated AMPA receptors fiom the cytosolic compartment to the plasma membrane and thus led to long-term potentiation of AMPA receptor- mediated synaptic transmissions. It has been shown that many molecules are involved in LTP, particularly calmodulin-kinwe I1 (Cam(Barria et al., 1997;Derkach et al., 1999;Strack et al-, 1997). CaMKII was activated during LTP and blocking its activity abolished LTP generation. Furthermore, CaMKII redistributes fiom the cytosol to the proximity of the plasma membrane during LTP (Strack et al., 1997). Although some work showed that CaMKII can phosphorylate AMPA receptor subunits in vitro and enhance AMPA receptor channel conductance, a recent study demonstrated that mutation of the CaMKII phosphorylation site on AMPA receptor GluRl subunits did not affect the synaptic recruitment of AMPA receptor during LTP, indicating that the AMPA receptor translocation is not due to a direct phosphorylation of AMPA receptors by Cam1(Hayashi et al., 2000). Possibly, the hction of the Ca27calmodulin system is to activate the AMPAR- associated PI3K thereby leading to LTP expression. Indeed, studies have demonstrated that calmodulin enhances PI3K activities through its binding with the SH2 domain of P8S (Joyal et al., 1997). So a plausible mechanism in LTP generation might be that Ca" that fluxes in fiom NMDA receptor channels, combines with calmodulin and activates local AMPAR-associated PI3K, which will then cause pIasma membrane insertion of AMPA receptors. How does PI3K facilitate AMPA receptor membrane insertion? As a lipid kinase, PI3K may affect receptor trafticking through its lipid products. A great accumulation of data showing that lipids including PI3K lipid products regulate protein trafficking (Chaudhary et al., 1998;Leevers et al., 1999;Martin, 1997). Also, Akt, the downstream substrate of PIX, is a serinelthreonine kinase which may regulate AMPAreceptor trafficking through protein phosphorylation, including a tentative direct phospboryIation of AMPA receptors. In either scenario, synaptic localization and close association -151- of these molecules with AMPA receptors enable an eEcient, selective and economic way in regulating AMPA receptor cellular distribution. The regulatory subunit p85 contains different modules, such as SH2,SH3 and proline-rich domains, for protein-protein interactions, and interactions of its SH2 domains with phosphotyrosines on other proteins are a common way to localize and/or activate PI3K. AMPA receptor subunits contain several tyrosine residues on the intracellular sites, but studies have indicated that under basal conditions tyrosine phosphorylation of AMPA receptors is very low, if at all- Also, our in vitm pull down experiments using non-phosphorylated GluR c-termini suggest a negligible role of GluRs ' tyrosine residues in their interaction with PI3K Because both GluRl and GluR2 c-termini are capable of binding PI3K and the fkagment immediately following the N transmembrane domain in GluRl and GluR2 c-termini show high homology, an unknown motif for PI3K binding may reside in this fragment. In this study, intracellular application of TeTx blocked glycine-induced LTP but had no effect on basal synaptic transmission. Consistently, intracellular microinjection of TeTx did not block constitutive AMPA receptor insertion, though it abolished the glycine-induced facilitation of receptor insertion. This is surprising because TeTx, through cleaving the V-SNAREprotein VAMP for protein exocytosis, is supposed to block AMPA receptor plasma membrane insertion and inhibit synaptic transmissions. In line with this notion, a recent study showed that, in contrast to our finding, TeTx did cause long-term depression FTD) of excitatory synaptic transmissions (Luscher et d., 1999). Although we are so far not aware of what causes this discrepancy, the toxin-induced inhibition is not likely the consequence of the supposed blockade of receptor insertion, because in our TeTx injection experiments, a normal basal AMPA receptor insertion was observed. A direct assumption is that different mechanisms may be utilized in constitutive and facilitated (during LTP) receptor insertion, with the constitutive pathway VAMP independent and the facilitated one VAMP- dependent. Or, another possibility is that TeTx only cleaves a portion of VAMP and thus makes the remaining VAMP a rate-limiting component in receptor exocytosis. In this case, there will still be enough VAMP to maintain the relatively slow receptor insertion under basal condition. However, when the insertion process is accelerated during LTP, the limited amount of VAMP prohibits an increase in receptor exocytosis and LTP expression. Enhanced plasma membrane insertion of AMPA receptors during LTP will increase synaptic AMPA receptors and strengthen synaptic transmission. Since normally a portion of glutamate synapses is composed solely of NMDA receptors and are therefore functionally "silent", LTP is achieved in part by switching silent synapses into active ones- This so-called "silent synapse" hypothesis for LTP has gained experimental support fiom different preparations (Gomperts et al., 1998; Isaac et al,, 1995; Li & Zhuo 1998; Liao et d., 1995). Consistently, in our cultured neuron LTP model, we found that while under basal condition 67% of NMDA receptor clusters co-localized with AMPA receptors, the AMPA/NMDA co-localization rate increased to 77% after LTP induction by glycine. Because there is no change in the number of NMDA receptor cluster, the increase in the AMPA/NMDA receptor co-localization rate indicate a transformation of silent synapses into active ones. The AMPAR-associated PI3K may have implications other than the role in AMPA receptor translocation and LTP formation,. During the early developmental stage of the central nervous system, massive synaptic connections are first built up, and the connections are then fine tuned to form specific, precise synaptic communications in an activity-dependent manner, while apoptosis in those "redundant" neurons occurs concomitantly in this maturation process. However, the mechanisms for activity-dependent survival is not clear. PI3K is known to be a key player in maintaining neuronal survival (Ng et al., 2000), so the AMPAR-associated P13K is placed at a strategic site. Neurond activities sensed by synaptic NMDA receptors will subsequently activate AMPAR-associated PI3K and thus prevent the cell fiom apoptotic neuronal death during development, while those receiving less or no synaptic inputs only have low PI3K activity and will eventually die out. Furthermore, as a major cellular signaling molecule, activation of synaptic PI3K will trigger a variety of downstream signaling pathways and initiate multiple cellular functions. Moreover, our irnmunostaining showed that most dendritic PI3K was localized in glutamatergic synapses. However, only a small portion of PI3K activity was co-irnmunoprecipited with AMPA receptors. This indicates that different mechanisms or other protein-protein interactions may be involved in synaptic localization of P13K. CHAPTER FIVE Summary, Conclusions and Future Directions 5.1 Research. summary Long-term synaptic plasticities including LTP and LTD, are believed to be the foundation for higher brain hctions such as learning and memory. The cellular mechanisms underlying LTP/LTD has been one ofthe most extensively studied areas in neuroscience, but the findings are still controversial. A recently praposed hypothesis for LTPLTD is that the long-term potentiation or depression of excitatory synaptic transmission is due to the rapid traff~ckingof AMPA receptors to either increase or decrease its cell-surface expression level. In support for this hypothesis, the present work found that: (1) Insulin induces AMPA receptor internaIization in both transfected HEK cells and in cultured hippocampal neuron. The insulin-induced internalization is GluR2-dependent and is mediated through a clathrin-coated-pit pathway, since the internalization can be blocked by either hypertonic sucrose or dominant negative mutant dynamin, and the internalized AWA receptors show co- localization with a clathrin adaptor protein EPS 15. Furthermore, in hippocampal CAI neurons, insulin induces a long-term depression of AMPAR-mediated synaptic transmission, which can be completely abolishzd by inhibiting protein endocytosis, strongly indicate an involvement of receptor endocytosis in the insulin-induced inhibition. Moreover, similar to insulin-induced LTD, low fkequency stirnulation (LFS)-induced homosynaptic hippocampal CAI LTD can also be blocked by inhibiting the clathrin-mediated internalization process, indicating that homosynaptic LTD is also utilizing receptor internalization as its underlying mechanism. Finally, the insulin-induced and the LFS-induced LTDs occlude each other, suggesting that some signalling pathways are shared by both, with the converging common pathways yet to be determined. (2) Another part of the thesis investigated mechanisms in the expression of LTP. By irnmunoprecipitation and imrnunostaining, we found that AMPA receptors associate and co-localize with PI3K in native neurons, and the binding - 156- appears to be directly between c-termini of GluRs and the P85 subunit of PI3K. The AMPA receptor (AMPAR)-associated P13K activity is enhanced by the glycine-induced activation of synaptic NMDA receptors. To investigate the fimctional significanceof PUK in synaptic transmission, we found that, through activation of AMPAR-associated PUK, glycine induces long-term potentiation &TP) of AMPA receptor-mediated miniature EPSCs (AMPA-minEPSCs) in cultured hippocampal neurons. This glycine-induced LTP seems to be due to a recruitment of AMPA receptors fiom cytosolic compartment to plasma membrane because: Firstly, glycine treatment increases the cell-sdace expression and synaptic localization of AMPA receptors through increasing AMPA receptor membrane insertion. Secondly, intracellular microinjection of tetanus toxin (TeTx) blocks glycine- induced AMPA receptor insertion and also glycine-induced LTP. To determine whether P13K is also involved in homosynaptic LTP in hippocampal slice, application of active PI3K into CAI neuron induces a dramatic potentiation of AMPA-EPSCs, while tetanus stimulation-induced LTP is blocked by inhibiting PI3K activity, strongly indicating that PUK, probably through facilitating AMPA receptor membrane insertion, plays an critical role in LTP expression. These data collectively indicate that AMPA receptor trafficking is the main mechanism underlying LTPLTD, two major cellular models for learning and memory. Specifically, clathrin- coated-pit-mediated AMPA receptor internalization is responsible for LTD, while the PI3K- dependent AMPA receptor insertion is responsible for LTP. This explanation for long-term synaptic plasticities (more synaptic receptors mediating stronger synaptic transmission and vice versa) is different fiom the large amount of data fkom previous studies which have demonstrated that LTPLTD is due to protein phosphorylation/dephosphorylation. These seemingly controversial conclusions, however, may just reflect different aspects of a unified story. It is well known that protein -157- Figure 5.1 A model for AMPA receptor internalization and LTD -,J -,J .'2,Z L-y;i-. **'-.I. -*,z-. Interaction of AMPA receptors with GRIP stabilizes the receptor on the postsynaptic membrane Activation of NMDA receptors or insulin receptors (IR) leads to activation of a kinase, which wiU phosphorylate AMPA receptors and dk~pt AMPAR-GRIP ,,,y:.: ,.t 4 L'&? soc cia ti^. p,,- . I.. _-- 3 Phosphorylated AMPA receptors internalize through clathrin- coated pits. Figure 5.2 Proposed models of LTP expression via AMPA receptor plasma membrane insertion AlvlPA receptor NMDA receptor AMPAR-bearing vesicle for membrane insertion Pre-synaptic Phosphorylated protein Trafficking machinery AMPAR-containing vesicle phosphorylation plays a critical role in the regulation ofboth exocytosis and endocytosis (Liu, 1997). In the case of LTP/LTD, phosphorylation might be a prerequisite for AMPA receptor plasma membrane insertion or internalization. So it is understandable that while receptor trafficking is the cellular mechanism for LTP/LTD expression, protein phosphorylation may also be necessary. How does insulin or low Eeq~encystimulation (LFS) lead to AMPA receptor internalization? There is no conclusive answer yet, but some related observations fiom recent studies allow us to outline a reasonable signaling cascade. The cytosolic PDZ domain-containing protein GRIP (see Section 1.3.2) is an AMPA receptor interacting protein. Through its binding with the end of the GluR2 c-terminus, GRIP is believed to be an anchoring molecule to stabilize AMPA receptors at the postsynaptic site (Dong et al., 1997). This is confirmed by recent findings showing that phosphorylation of S880 at the GluR2 c-terminus, which specifically disrupts the association of GluR2 and GRIP, dramatically reduces the AMPA receptor cell surface expression due to an facilitated receptor internalization (Chung et al, 2000; Matsuda et al, 1999; Matsuda et al, 2000). Therefore, a possibility is that insulin and LFS, through insulin receptor and NMDA receptor-gated calcium influx, respectively, activate unidentified protein kinase(s) which eventudly lead to GluR.2 phosphorylation and AMPA receptor internalization (Fig. 5.1). To verifL this postulation, the participating kinase(s), which should show activation following insulin and LFS, and use GluR.2 subunit as its substrate, need to be identified. Combining our findings and data fiom other groups, a whole picture for LTP expression can be also proposed. Studies demonstrated that Ca2+/calmodulin,which is a crucial factor in LTP generation, activates PI3K (Joyal et al., 1997). Because PI3K is well known for promoting protein exocytosis (Pessin et al., 1999; Yang et al., 1996), we can propose a working model for LTP expression. As shown in Fig.5.2, accumulation of intracellular calcium following NMDA receptor activation by glycine or high frequency stimulation WS) will activate the AMPAR-associated PDK, which then helps to translocate AMPA receptors £?om cytosol to plasma membrane (Fig. 5.2A). Further details about the way by which PUKpromotes AMPA receptor insertion are not clear yet, one possibility is through its lipid products. If the receptor-bearing vesicles are already close to plasma membrane and ready to fuse, the PUK-produced phospholipids on the AMPAR-containing vesicles might optimize the membrane conditions and facilitate the membrane-membrane fusion. Certainly, in addition to this possibility, other scenarios exist. For example, the rate-limiting step for AMPA receptor plasma membrane insertion may not be the final membrane firsion event, but the intracelldar receptor transportation. AMPA receptor-containing vesicles are probably localized away from the plasma membrane and have no interaction with the transportation system. Activation of the AMPAR- associated PI3K will recruit Akt to the AMPA receptor-containing vesicle via the vesicle membrane- bound P13K lipid products. Since Akt is a serine/threonine kinase and has the potential to phosphorylate GluRs or other proteins on the vesicle membrane, it is reasonable to postulate that phosphorylation of proteins on the vesicle by Akt dlows the AMPAR-bearing vesicles to contact the delivery machineries and to be transported to the plasma membrane for insertion (Fig. 5.2B). 5.2 Future directions This work demonstrates that receptor trafficking is probably a common mechanism underlying synaptic plasticity, many questions are yet to be answered. Future work will include: 1. Examination and dissection of the signaling pathway in the insdin-induced LTD. Insulin receptor is a receptor tyrosine kinase, and its activation has been shown to be required in the insulin-induced LTD. However, the downstream signallingpathway following insulin receptor activation is not clear. It could be through PI3K or PKC, two main kinases mediating insulin effects, or through other molecules. 2, Do AMPA receptors co-trafEc with insulin receptors? It is knownthat insulin receptors internalize by agonist binding, and we showed a insulin-receptor-activation-dependent AMPA receptor endocytosis. Given the synaptic localization of insulin receptors which put them close to synaptic AMPA receptors, a possibility is that insulin receptors may bind to and internalize together with AMPA receptors. 3. Whether the homosynaptic hippocampal CAI LTD is induced through insulin receptor activation? Insulin-induced chemical LTD and the Iow frequency stimulation (LFS)-induced homosynaptic hippocampal CAI LTD share some similarities. Both can be blocked by inhibiting receptor internalization and mutually occlude each other. Since insulin is synthesized in the neuron and can be released at the nerve terminal in the neuronal activity-dependent manner, it is not unreasonable to hypothesize that the homosynaptic LTD is induced by insulin, released from the presynaptic terminal by LFS. 4. Does insulin protect neurons against ischemic apoptosis? During ischemia in the brain, the massively released glutamate binds to glutarnate receptors and causes overexcitation, which will trigger neuronal apoptosis and cell death. As insulin can induce AMPA receptor internalization and thus reduce AMPA receptor activity, insulin is expected to have a protective function under ischemic conditions in the brain- 5. What is the downstream effector in the PDK-dependent LTP? PUK is a lipid kinase executing its cellular functions through its Lipid products or other downstream signalling molecules. PIXproducts have been shown to be involved in protein trafficking (Leevers et d., 1999;Vanhaesebroeck et al., 1996) and regulation of synaptic transmission (Hawthorne, 1W6), suggesting that PI3K may induce LTP through the phospholipids such as PI3P, PI3,4P2 or P13,4,5P3. Alternatively, PI3K might regdate synaptic transmission through other downstream substrates such as Akt. Akt is a serine/threonine kinase and can thus regulate AMPA receptor trafficking by protein phosphorylation. 6. Identification and characterization of the machineries for AMPA receptor trafficking. To study receptor tra£€iclcing in neuronal plasticity, a key issue is to understand the related machineries. Although a motor protein specifically for NMDA receptor delivery has just been reported (Setou et al, 2000), machineries for AMPA receptor trafficking are still unknown, Studies should be done including investigation of the nature of the cargo vesicles, the motor proteins and the mechanisms for the specificity in both receptor delivery and site of targeting for AMPA receptors. Appendix: Publications I. Published papers I. Man YH, Lin JW, Ju WHYAhmadian G, Liu L, Becker LE, Sheng M, Wang YT Regulation of AMPA receptor-mediated synaptic transmission by clathrin-dependent receptor internalization. Neuron 2000,25(3):649-62 Redistribution of postsynaptic AMPA- (alpha-amino-3-hydroxy-5-methylisoxazole-4-propionicacid-) subtype glutamate receptors may regulate synaptic strength at glutamatergic synapses, but the mediation of the redistribution is poorly understood- We show that AMPA receptors underwent clathrin-dependent endocytosis, which was accelerated by insulin in a GluR2 subunit-dependent manner. Insulin-stimulated endocytosis rapidly decreased AMPA receptor numbers in the plasma membrane, resulting in long-term depression (LTD) of AMPA receptor-mediated synaptic transmission in hippocampal CA I neurons. Moreover, insulin-induced LTD and low-fkequency stimulation-(LFS-) induced homosynaptic CAI LTD were found to be mutually occlusive and were both blocked by inhibiting postsynaptic c1ath-h-mediated endocytosis. Thus, controlling postsynaptic receptor numbers through endocytosis may be an important mechanism underlying synaptic plasticity in the mammalian CNS. 2. H Y Man*, W Y Lu*, W Ju, W Trimble, Y T Wang and John F MacDonald (* Equal contribution) Activation of synaptic NMDA receptors induces membrane insertion of new AMPA receptors and LTP in cultured hippocampal neurons. Neuron, 2001, in press. Long-term potentiation (LTP) of excitatory transmission in the hippocampus likely contributes to learning and memory. The mechanisms underlying LTP at these synapses are not well understood, although phosphorylation and re-distribution of AMPA receptors may be responsible for this form of synaptic plasticity. We show here that miniature excitatory synaptic currents (mepscs) in cultured hippocampal neurons reliably demonstrate LTP when postsynaptic NMDA receptors are briefly stimulated with glycine. LTP of these synapses is accompanied by a rapid insertion of native AMPA receptors and by increased clustering of AMPA receptors at the surface of dendritic membraaes. Both LTP and gly cine-facilitated AMPA receptor insertion are blocked by intracellular tetanus toxin (TeTx) providing evidence that AMPA receptors are inserted into excitatory synapses via a SNARE- dependent exocytosis during LTP. 3. H Y Man, W Ju, G Ahmadian and Y T Wang Regulation of intraceUular receptor trafficking - a novel means in functional modulation of ligand-gated Receptors Beview]. Molecular and Cellular L$e Science, 2000, in press 4. R Q Hu, M A Cortez, H Y Man, J Roder, Y T Wang and 0 C Snead III Alteration of GluR2 expression in the rat brain following absence seizures induced by y-hydrorrybutyric acid, Epilepsy Research, 2000, in press. 5. R Q Hy M A Cortez, H Y Man, J Roder, Z P Jia, Y T Wang and 0 C Snead III. GAB-induced absence seizures in GluIU null mutant mice. Brain Research, 2000, in press. 6. McRitchie DI, Isowa N, Edelson JD, Xavier AM, Cai L, Man HY, Wang YT, Keshavjee SH, Slutsky AS, Liu M. Production of tumour necrosis factor alpha by primary cultured rat alveolar epithelial cells. Cytokine 2000, 12(6):644-54 Tumour necrosis factor alpha(TNF-alpha) is one of the most important pro-inflammatory cytokines, which plays an important role in host defense and acute inflammation related to tissue injury. The major source of TNF-alpha has been shown to be immune cells such as macrophages and neutrophils. In the present study, we demonstrated that LPS-treatment on alveolar epithelial cells isoIated from addt rat lungs also induced a dose- and time-dependent release of TNF-alpha. The purity and identity of these celIs were examined by immunofluorescent staining and confocal microscopy with antibodies for cytokeratin and pro-surfactant protein C, markers for epithelial cells and type II pneumocytes respectively- Positive staining of TNF-alpha was observed throughout the cell layer and localized intracellularly. LPS-induced TNF-alpha production fiom alveolar epithelial cells was blocked not only by cycloheximide, an inhibitor of protein translation, but also by actinomycin D, an inhibitor of gene transcription, The rnRNA of TNF-alpha rapidly increased within 1 h of LPS stimulation. These data suggest that LPS-induced TNF-dpha production from alveolar epithelial cells is primarily regulated at the transcriptional level, which is different fiom that of macrophages and neutrophils. TNF-alpha produced by alveokir epithelial cells may function as an alert signal in host defense to induce production of other inflammatory mediators. 7. Wan Q, Man HY, Liu F, Braunton J, Nimik HB, Pang SF, Brown GM, Wang YT. Differential modulation of GABAA receptor function by Mella and Mellb receptors. Nat Neurosci 1999, 2(S):4O 1-3. Melatonin, a hormone principally produced and released by the pined gland, has been shown to regulate a variety of biological fimctions including circadian rhythms, sleep-wake cycles and reproduction, presumably through activating high-a£finity G-protein-coupled receptors. We report here that these subtypes can differentially modulate the function of type A aminobutyric acid (GABAJ receptor, the principal neurotransmitter receptor mediating synaptic inhibition in the CNS. This work demonstrates that melatonin, through activation of different receptor subtypes, can exert opposite effects on the same substrate, suggesting that receptor subtype is the primary molecular basis for the diversity of melatonin effects. 8. Man HY, Erclik T, Becker LE, Wang YT. Modulation of baroreflex sensitivity by the state of protein tyrosine phosphorylation in the brainstem of the rat. Brain Res 1998,792(1): 141-148 Evidence accumulated recently suggests that protein tyrosine phosphorylation may play an important role in regulating neuronai hctions. In the present study, we investigated if the state of protein tyrosine phosphorylation in the brainstem regulates baroreflex sensitivity. Anti-phosphotyrosine irnmunoblots of brainstem tissue revealed that several phosphotyrosine-containing proteins were present in the brainstem and their level of tyrosine phosphorylation was decreased by treatment of the slices with the protein tyrosine kinase (PTK) inhibitor genistein, and increased by treatment with the protein tyrosine phosphatase (PTP) inhibitor pervanadate. In urethane-anaesthetized rats, we found that inhibiting PTK activity by topical application of genistein to the dorsal surface of the medulla reduced the phenylephrine-induced baroreflex bradycardiac response, Conversely, the baroreflex response was potentiated by activating endogenous PTK activity with insulin or by inhibiting PTP activity with pervanadate. Thus these results suggest that the state of cellular tyrosine phosphorylation within the dorsal medulla of the brainstem may regulate the baroreflex control of heart rate, thereby providing the first evidence for a role for protein tyrosine phosphorylation, a key process involved in diverse intracellular signalling pathways, in modulating baroreflex sensitivity. 9. Pristupa ZB, McConkey F, Liu F, Man HY, Lee FJ, Wang YT, Niznik HB. Protein kinase- mediated bidirectional trafficking and functional regulation of the human dopamine transporter. Synapse l998,30(1):79-87 Modification of the transport velocity of both the native neuronal and cloned presynaptic dopamine transporter @AT) has been reported following activatiodinhibition of second messenger system pathways. In order to identify the mechanism by which the hctional activity of human DAT @AT) is regulated, we assessed the [3H]dopamine uptake kinetics, [3H] CFT binding characteristics, and, via immunofluorescent confocal microscopy, the cellular localization profiles ofthe hDAT expressed in both Sf9 and COS-7 cells following modulation of protein kinase C (PKC)- and protein kinase A (EKA)-dependent pathways. As with both native neuronal and cloned DATs, acute exposure of hDAT expressing Sf9 cells to the PKC activator PMA (1 microhii), but not alphaPDD, reduced the Vmax (approximately 1 pmoVmid1 O(5) cells) for [3H]DA uptake by approximately 40%, an effect which was blocked by the protein kinase inhibitor staurosporine. Pretreatment of cells with staurosporine (500 nM) alone, however, increased [3H]DA uptake velocity by approximately 30%, an effect mimicked by the potent PKA inhibitor Rp-CAMPS.Activation of PKA-dependent pathways with Sp- CAMPSdid not significantly modify DA uptake. Neither the Km of [3H]DA uptake (approximately 200 nM) nor the &ty of various substrates and transport inhibitors was altered by either PMA or staurosporine treatment. Despite changes in functional doparnine uptake velocity by PKCIPKA- dependent mechanisms, the estimated density of hDAT as indexed by whole-cell [3H] CFT binding was unchanged. Immuno£luorescent confocal microscopy demonstrated that the observed functional consequence of PKC activation on [3H]DA uptake is associated with the rapid sequestration/internalizationof hDAT protein fkom the cell surface, while the increase in DA uptake following PKCIPKA inhibition is the result of the recruitment of internalized or intracellular transporters to the plasma membrane. Identical rapid translocation patterns were observed insirniIarly treated COS-7 ceIls transiently expressing hDAT. These data suggest that the differential regulation of DAT transport capacity by both PKC- and PKA-dependent pathways are not a result of modifications in DAT catalytic activity. Moreover, the rapid shuttling of DATs between the plasma membrane and intracellular compartments provides an efficient means by which native DAT function may be regulated by second messenger systems, possibly following activation of presynaptic dopaminergic receptors, adsuggests a roIe for cytoskeletal components in the dynamic regulation of DAT function. 10. Wan Q, Xiong ZG, Man HY, Ackerley CA, Braunton J, Lu WY, Becker LE, MacDonald JF, Wang YT. Recruitment of functional GABA(A) receptors to postsynaptic domains by insuIin. Nature 199 7, 388(6643) :686-90 Modification of synaptic strength in the mammalian central nervous system (CNS) occurs at both pre- - 16% and postsynaptic sites. However, because postsynaptic receptors are likely to be saturated by released transmitter, an increase in the number of active postsynaptic receptors may be a more efficient way of strengthening synaptic efficacy. But there has been no evidence for a rapid recruitment of neurotransmitter receptors to the postsynaptic membrane in the CNS. Here we report that insulin causes the type A gamma-aminobutyric acid (GABA[A]) receptor, the principal receptor that mediates synaptic inhibition in the CNS, to translocate rapidly from the intracellulsu compartment to the plasma membrane in transfected HEK 293 cells, and that this relocation requires the beta2 subunit of the GABA(A) receptor. In CNS neurons, insulin increases the expression of GABA(A) receptors on the postsynaptic and dendritic membranes. We found that insulin increases the number of fictional postsynaptic GABA(A) receptors, thereby iccreasing the amplitude of the GAE3A(A)- receptor-mediated miniature inhibitory postsynaptic currents (rnIPSCs) without altering their time course. These results provide evidence for a rapid recruitment of functional receptors to the postsynaptic plasma membrane, suggesting a fundamental mechanism for the generation of synaptic plasticity. I I. Wan Q, Man HY, Braunton J, Wang W, Salter MW, Becker L, Wang YT. Modulation of GABA, receptor function by tyrosine phosphorylation of beta subunits. Protein tyrosine phosphory lation is a key event in diverse intracehlar signaling pathways and has been implicated in modification of neuronai functioning. We investigated the role of tyrosine - 170- phosphorylation in regulating type A GABA (GABAA) receptors in cultured CNS neurons. Extracellular application of genistein (50 pM), a membrane-permeable inhibitor of protein tyrosine kinases (PTKs), produced a reversible reduction in the amplitude of GABAA receptor-mediated whole-cell currents, and this effect was not reproduced by daidzein (50 pM), an inactive analog of genistein. In contrast, intracellular application of the PTK pp60(c-src) (30 Uhl) resulted in a progressive increase in current amplitude, and this potentiation was prevented by pretreatment of the neurons with genistein. Immunoprecipitation and immunoblotting of cultured neuronal homogenates indicated that the P2@3 subunit(s) of the GABAA receptor are tyrosine phosphorylated in situ. Moreover, genistein (50 pM) was found to be capable of decreasing GABAA currents in human embryonic kidney 293 cells transiently expressing functional GABAA receptors containing the beta2 subunit. Thus, the present work provides the first evidence that native GABAA receptors are phosphorylated and modulated in situ by endogenous PTKs in cultured CNS neurons and that phosphorylation of the beta subunits may be sufficient to support such a modulation. Given the prominent role of GABAA receptors in mediating many brain functions and dysfunctions, modulation of these receptors by PTKs may be important in a wide range of physiological and pathological processes in the CNS. 11. Papers in preparation 1. H Y Man, Q H Wang, W Y Lu, G Ahmadian, L E Becker, J F MacDonald and Y T Wang. Long- term potentiation of excitatory synaptic transmission through activation of AMPA receptor-associated PI3 -kinase. In preparation. 2. L D Liy H Y Man, H J Chung, Q H Wag, M Sheng, R Huganir and Y T Wang. Calcineurin and Caesin kinase I1 mediate insulin- and low frequency stimulation-induced LTD. In preparation. 3. G Ahmadian, W JUH Y Man and Y T Wang. Constitutive and regulated endocytosis of glutamate receptor 2 depends upon different and specific tyrosine residues in its carboxyl terminus. In preparation. III. Abstracts 1. H Y Man, Q H Wang, W Y Lu, G Ahmadian, L E Becker, M P Wymann, J F MacDonald and Y T Wang (2000) Activation of AMPA receptor-associated PI3-kinase mediates postsynaptic membrane insertion of AMPA receptors during LTP expression. 30LhAnnual meeting, Societv ofNeurosci Absrvact 2. L D Liu, IX Y Man and Y T Wang (2000) Convergent signalling pathway in insdin- and LFS- induced long-term depression (LTD). 30LhAnnual meeting, Society qf Neurosci Abshacr 3. W Y Lu, H Y Man, W Trirnble, J F MacDonald and Y T Wang (2000) Synaptic and not extrasynaptic NMDA receptors induces LTP through insertion of AMPA receptors at excitatory synapses in cultured hippocampal neurons. 30" Annual meeting, Societv qfNeurosci Abs~act 4. G Ahmadian, W Ju, H Y Man and Y T Wang (2000). Constitutive and regulated endocytosis of glutamate receptor 2 depends upon different and specific tyrosine residues in its c-terminus. Societv ofNeurosci Abstract 5. H Y Man, F Liu, L E Becker, H B Niznik and Y T Wang (1999) Doparnine D5 receptors bind to and regulates plasma membrane expression of GABAA receptors. 29th Annual meeting, Societv for Neuroscience Abstract. 6. W Ju, H Y Man, J Lin, M Sheng and Y T Wang (1999) Rapid regulation of surface expression of AMPA receptors by clatbrin-mediated receptor endocytosis. 29th Annual meeting. Sociev fir Neuroscience Abstract. 7- S M D'Souza, G Ahmadian, H Y Man, M Wymann and Y T Wang (1 999) PI3-kinase binds to and modulates the GABAA receptor. 29th Annual meeting, Societv for Neuroscience Abstract. 8. H Y Man, L E Becker and Y T Wang (1 998) Long-term depression of AMPA receptor-mediated synaptic transmission by insulin in hippocampal CA 1 neurons. 28th Annual meeting. Sociehr -for Neuroscience Abstract, 9. H Y Man, Q Wan, Z G Xiong, C A Ackerley, J Brauton, L E Becker, J F MacDonald and Y T Wang (1997) Recruitment of functional GABAA receptors to postsynaptic domains by insulin. 27 Annual Meeting, Societv for Nezrroscience Abstract. 10.Yutian Wang, H Y Man, ErEc Ted and Laurence E Becker (1997) Modulation of baroreflex sensitivity by the state of tyrosine phosphorylation in the brainstem of rat. 27 Annual Meeting, Society for Neuroscience Abstract. 11. F Liu, Z B Pristupa, H Y Man, F McConkey, F J S Lee, Y T Wang and H B Niznik (1997) Rapid bidirectional translocation of the human dopamine transporter and the regulation of dopamine transport. 27 Annual Meeting, Society fir Neuroscience Abstract 12. McRichie, D I, J Edelson, A S Slutsky, H Y Man, Y T Wag, S Keshavjee and M Liu (1997) Lipopolysaccaride-induced tumor necrosis factor a release fiom rat lung type U pneumocytes. American Luna Association and American Thoracic Societv Conference, San Francisco 13. Q Wan, H Y Man, J L Braunton, G M Brown and Y T Wang (1 996) GABAA receptor P2 subunit is sufficient for modulation of the receptor by tyrosine phosphorylation. 26 Annual Meeting, Societv fir Neuroscience Abstract. Reference List Abbott,M.A., Wells,D.G., and Fal10n~J.R.(1999). The insulin receptor tyrosine kinase substrate p58/53 and the insulin receptor are components of CNS synapses. J.Neurosci. 19,7300-7308, Akaz~wa,C-,Shigemoto,R., Bessho,Y., Nakanishi,S ., and MzunoN. (1994). Differential expression of five N-methyl-D-aspartate receptor subunit mRNAs in the cerebellum of developing and adult rats. J.Comp.Neurol.347, 15O-I6O. Mconada,A., Bauer,U., and Hoflack,B. (1996). A tyrosine-based motif and a casein kinase II phosphorylation site regulate the in~acellulartrafficking of the varicella-zoster virus glycoprotein I, a protein localized in the trans-Golgi network, EMBO J 15,6096-6 100. ALessi,D.R. and Downes,C.P. (1 998). The role of PI 3-kinase in insulin action. Biochim.Biophys.Acta 1436, 151-164. Allison,D.W., Gelfand,V.I., Spector,I., and Craig,A.M. (1998). Role of actin in anchoring postsynaptic receptors in cultured hippocampal neurons: differential attachment of NMDA versus AMPA receptors. J-Neurosci. 18,2423-2436. Antonetti,D.A., Algenstaedt,P., and Kahn,C.R. (1996). Insulin receptor substrate 1 binds two novel splice variants of the regulatory subunit of p hosphatidy linositol 3- kinase in muscle and brain. Mol.Cel1 Biol. 16,2195-2203. Asztely,F. and Gustafsson,B- (1996). lonotropic glutamate receptors - Their possible role in the expression ofhippocampal synaptic plasticity, Molecular Neurobiology 12, 1- 11. Bachelof C., RamehL., Parsons,T., and Cantley,L.C. (1 996). Association of phosphatidylinositol3 - kinase, via the SH2 domains of p85, with focal adhesion kinase in polyoma middle t-transformed fibroblasts. Biochim.Biophys.Acta 1311,45-52, Banfi, Downes,C.P., and Rittenhouse,S.E. (1 998). Biphasic Activation of PKBalphdAkt In Platelets. Evidence for stimulation both by phosphatZdylinositol 3,4-bisphosphate, produced via a novel pathway, and by phosphatidylinositol3,4,5-trisphosphate. J-Biol-Chem. 2 73, 1 1630- 11637. Banke,T.G., Bowie,D., Lee,H., Huganir,R.L., Schousboe,A., and TrayneLis,S.F. (2000). Control of GluRl AMPA receptor function by CAMP-dependentprotein kinase. J.Neurosci. 20, 89- 102. Barria,A., Muller,D., Derkach,V., Griffith,L.C., and S0der1ing~T.R. (1997). Regulatory p hosphory lation of AMPA-type glutamate receptors by CaM-KII during long-term potentiation. Science 276,2042-2045- Bashi.r,Z.I., Tarn,B., and C01lingridge~G.L.( 1990). Activation of the glycine site in the NMDA receptor is necessary for the induction of LTP. Neutosci.Lett. 108, 261-266. Bear,M.F. (1 999). Homosynaptic long-term depression: a mechanism for memory? Proc.Nat1.Acad.Sci.U.S.A 96,9457-9458. Bear,M.F. and Abraham, W.C. (1996). Long-term depression in hippocampus. Annu.Rev.Neurosci. 19,437-462. Bear,M.F. and Malenka,R.C. (1 994). Synaptic plasticity: LTP and LTD. Curr.Opin.Neurobiol.4,3 89- 399. Bekkers,J.M. and Stevens,C.F. (1995). Quanta1 analysis of EPSCs recorded fiom small numbers of synapses in hippocampal cultures. J.Neurophysio1. 73, L 145- 1156. Bellacosa,A., Chan,T.O., Ahmed,N.N., Datta,K., MaIstrom,S ., Stokoe,D., McCormick,F., Feng,J-, and TsichIis,P. (1998). Akt activation by growth factors is a multiple-step process: the role ofthe PH domain- Oncogene 17,3 13-325- Berridge,M. J. (1986). Growth factors, oncogenes and inositol lipids. Cancer Surveys 5,41 3 -43 0. Blackstone,C.D., Supattapone,S., and Snyder,S.H. (1989). lnositolphospholipid-Linked Glutamate Receptors Mediate Cerebellar Parallel-Fiber Pukinje-Cell Synaptic Transmission. PNAS-US- 86, 43 16-43 20. Blair,L.A., Bence-Hanulec,K.K., Mehta,S., Franke,T-, Kaplan,D., and Marshall,J. (1999)- Akt- dependent potentiation of L channels by insulin-like growth factor- 1 is required for neuronal survival. J.Neurosci. 19, 1940- 1951, Blair,L.A. and Marshal1,J. (1997). IGF-2 moddates N and L calcium channels in a PI 3-kinase- dependent manner. Neuron 19,421-429. Bliss,T.V.P. and Collingridge,G.L. (1993). A synaptic model of memory: Long-term potentiation in the hippocampus. Nature 3 61, 3 1-3 9. Blitzer,R.D., Comor,J.H., Brown,G.P., Wong,T., Shendikar,S., IyengarJZ., and Landau,E.M. (1998). Gating of CaMKII by CAMP-regulatedprotein phosphatase activity during LTP. Science 280,1940- 1942. Boeckman,F.A. and Aizenrnan,E. (1993). Stable transfection of NMDARl subunit in Chinese hamster ovary (CHO-K1) cells fail to produce a functional receptor. Soc.Neurosci.Abstr. 19. Bolshakov,V.Y. and Siegelbaum,S.A. (1995). Regulation of hippocampal transmitter release during development and long-term potentiation. Science 269, 1730- 1734. Bonifacino,J.S ., Marks,M.S ., 0hno,H., and Kirchhausen,T. (1996). Mechanisms of signal-mediated protein sorting in the endocytic and secretory pathways. Proc Assoc Am Physicians 108,285-295. BodteqJ., Hollmann,M., O'Shea-Greenfield& Hartley,M., Deneris,E., Maron,C., and Heinemann,S. (1 990). Molecular cloning and functional expression of glutamate receptor subunit genes. Science 249, 1033-1037. Brenrnan,J.E., Chao,D.S., Gee,S.H., McGee,A. W., Craven,S.E., SantiIlano,D.R., Wu,Z., Huang,F., Xia,H., Peters,M.F., Froehner,S.C., and Bredt,D.S. (1996). Interaction of nitric oxide synthase with the postsynaptic density protein PSD-95 and alphal-syntrophin mediated by PDZ domains. Cell 84, 757-767. Brown,W.J., DeWald,D.B., Emr,S.D., PlutneqH., and Balch,W.E. (1995). Role for phosphatidylinositol3-kinase in the sorting and transport of newly synthesized lysosomal enzymes in mammalian cells- J-Cell Biol- 130, 78 1-796. Burgering,B,M. and Coffer,P.J-(1 995). Protein kinase B (c-Akt) inphosphatidylinositol-3-OH kinase signal transduction. Nature 3 76,599-602. Burnashev,N., Villarroel,A., and Sakmann,B. (1996). Dimensions and ion selectivity of recombinant AMPA and kainate receptor channels and their dependence on Q/R site residues. J.Physiol.(Lond.) 496 (Pt I), 165-173. Carpenter,C.L. and Cantley,L.C. (1 996a). P hosphoinositide 3 -kinase and the regulation of cell growth. Biochim-Biophys-Acta 1288, M 11 -M16. Carpenter,C.L. and Cantley,L.C. (1 996b). Phosphoinositide kinases. Cum-Opin-CellBiol. 8,153-1 58. Carroll,R.C., Beattie,E.C., Xia,H., Luscher,C ., Altschuler,Y ., Nicoll,R.A., MaIenka,R.C., and von Zastrow,M. (1999a). Dynamh-dependent endocytosis of ionotropic glutamate receptors. Proc.Nat1.Acad.Sci.U-S.A 96, 141 12- I41 17. Carroll,R.C., Lissin,D.V., von Zastrow,M., Nicoll,R.A., and Malenka,R.C. (1999b). Rapid redistribution of glutamate receptors contributes to long-term depression in hippocampal cultures. Nat.Neurosci. 2,454-460, Carr011,R.C.~Nicoll,R.A., and Malenka,R.C. (1 998). Effects of PKAand PKC on miniature excitatory postsynaptic currents in CAI pyramidal cells. J.Neurophysio1. 80,2797-2800. Chalecka-Franaszek,E. and Chuang,D.M. (I 999). Lithium activates the serine/threonine baseAkt- I and suppresses glutamate-induced inhibition of Akt-1 activity in neurons. Proc.Nat1.Acad.Sci.U.S.A 96,8745-8750. Chaudhary,A., Gy Q.M., Thum,O., Profit,A.A., Qi,Y ., Jeyakumar,L., FIeischer,S., and Prestwich,G.D. (1 998). SpecSc interaction of Golgi coatomer protein alpha-COP with phosphatidylinositol3,4,5- trisphosphate. J-Biol-Chem.273, 8344-8350. Cheatharn,B ., Vlahos,C.J., Cheatharn,L., Wang,L., Blenis,J., and Kahn,C.R. (1 994). PhosphatidylinositoI 3-kinase activation is required for insulin stimulation of pp7O S6 kinase, DNA synthesis, and glucose Qansporter translocation. Mol.Cell Biol. 14,4902-491 1. Chen,H. J., Rojas-Soto,M., Oguni,A., and Kennedy,M.B. (1 998). A synaptic Ras-GTPase activating protein (p135 SynGAP) inhibited by CaM kinase II. Neuron 20, 895-904. Chen,Q., Harris,C., Brown,C.S., Howe,A., Surmeier,D.J., and Reiner,A. (1 995). Glutamate-mediated excitotoxic death of cultured striatal neurons is mediated by non-NMDA receptors. Exp-Neurol. 136, 2 12-224. Chibalin,A.V., Zierath,J.R., Katz,A.I., Berggren,P.O., and Bertorel10,A.M. (1 998). Phosphatidylinositol3-Kinase-mediated Endocytosis of Renal Na+,K+- ATPase alpha Subunit in Response to Dopamine. Mo1.Biol.Cel.I 9, 1209-1220. Cho,K.O., Hunt,C.A., and Kemedy,M-B . (1 992). The rat brain postsynaptic density fiaction contains a homofog of the Drosophila discs-large tumor suppressor protein. Neuron 9,929-942. Chou,M.M., Hou,W., Johnson,J., Graharn,L.K., Lee,M.H., Chen,C.S., Newton,A.C., SchafThausen,B.S., and Toker,A. (1998). Regulation of protein kinase C zeta by PI 3-kinase and PDK- 1, Cur-Biol. 8, 1069-1077. Christie,J.M., Wenthold,R.J., and Monaghan,D.T. (1999). Insulin causes a transient tyrosine phosphorylation of NR2A and NR2B NMDA receptor subunits in rat hippocampus. JNeurochem. 72, 1523-1528. Chu,P., Murray,S., Lissin,D., and von Zastrow,M. (1997). Delta and kappa opioid receptors are differentially regulated by dynamin-dependent endocytosis when activated by the same alkaloid agonist- J-Biol-Chem.272,27 124-27130- Chung H J, Xia J, Scannevin R H, Zhang X, Huganir R L (2000). Phosphorylation of the AMPA receptor subunit GluR2 differentially regulates its interaction with PDZ domain-containing proteins. J Neurosci 20(19):7258-67 CIarke,D.W., Mudd,L., Boyd,F.T.J., Fields,M., and Raizada,M.K. (1986). Insulin is released from rat brain neuronal cells in culture. J Newochem 4 7, 83 1-83 6. Colbran,R.J. (1992). Regulation and role of brain calcium/calmodulin-dependentprotein kinase 11. Neurochem.Int. 21,469-497. Collingridge,G.L- and Bliss,T.V. (1995). Memories of NMDA receptors and LTP. Trends in Neurosciences I8,54-56. Colwel1,C. S. and Levine,M.S . (1996). Glutamate receptor-induced toxicity in neostriatal cells. Brain Research 724,205-2 12. Corvera,S. and Czech,M.P. (1998). Direct targets ofphosphoinositide 3-kinase products in membrane traffic and signal transduction, Trends Cell Biol. 8,442-446, Courtneidge,S.A., Fumagalli,S., Koegl,M., Superti-FurgqG., and Twarnley-SteinGM. (1993). The Src family of protein tyrosine kinases : regulation and functions. Development - Supplement 5 7-64. Craig,A.M- (1 998). Activity and synaptic receptor targeting: the long view. Neuron 21,459-462. Damke,H., Baba,T., Warnock,D.E., and Schmid,S.L. (1994). Induction of mutant dynarnin specifically blocks endocytic coated vesicle formation. J.Cell Biol. 127, 9 15-934. De Camilli,P., Emr,S .D., McPherson,P.S., and Novick,P. (1 996). Phosphoinositides as regulators in membrane traEc. Science 2 71, 1533-1539. De Camilli,P. and Takei,K. (1 996). Molecular mechanisms in synaptic vesicle endocytosis and recycling. Neuron 16,481-486. Derkach,V., Barria,A., and S0der1ing~T.R.(1999). Ca2+/calmodulin-kinase II enhances channel conductance of alpha-amino-0-3-hydroxy-5-methyl-4-isoxazolepropionatetype glutamate receptors- ProcNat1.Acad.Sci.U.S .A 96, 3269-3274. Dennan,M.P., Toker,A., Hartwig,J.H., Spokes,K., Falck, J-R-, Chen,C.S., Cantley,L.C., and Cantley,L.G. (1 997). The lipid products of phosphoinositide 3-kinase increase cell motility through protein kinase C. J.Biol.Chem. 2 72,64656470. Dingledine,R., Borges,K., Bowie,D., and Trayne1i~~S.F.(1999). The glutamate receptor ion channels. Pharmacol.Rev. 51,7-6 1. Doble,A. (1995). Excitatory amino acid receptors and neurodegeneration. Therapie 50, 3 19-3 3 7. Dong,H., OtBrien,R.J., Fung,E.T., Lanahan,A.A., Worley,P.F., and Huganir,R.L. (1 997). GRIP: a synaptic PDZ domain-containing protein that interacts with AMPA receptors. Nature 386,279-284. Dore,S., Kar,S., Rowe,W., and Quirion,R. (1997). Distributionand levels of [12SI]IGF-I, [125I]IGF- II and [125l-Jinsulinreceptor binding sites in the hippocampus of aged memory-unimpaired and - impaired rats- Neuroscience 80, 1033- 1040. Durand,G.M., Kovalchuk,Y ., and Komerth,A. (1 996). Long-term potentiation and functional synapse induction in developing hippocampus. Nature 381, 71-75. Edwardson,J.M. and Szekeres,P.G. (1 999). Endocytosis and recycling of muscarinic receptors. Life Sciences. 64,487-494. Ehlers,M.D., Tingley, W.G., and Huganir,R.L. (1995). Regulated subcellular distribution of the NR1 subunit of the NMDA receptor. Science 269, 1734-1737. Eves,E.M., Xiong,W., Bellacosa,A., Kennedy,S.G., Tsichlis,P.N., Rosner,M.R., and Hay,N. (1998). Akt, a target ofphosphatidylinositol3-kinase,inhibits apoptosis ina differentiatingneuronal cell line. Mol-Cell Biol, 18,2143-2152. Farrant,M., Feldmeyer,D., Takahashi,T., and Cull-Candy,S.G. (1993). Changes in expression of NMDA-receptor channel types during cerebellar granule cell development. Soc.Neurosci.Abstr. 19. Farrant,M., Feldmeyer,D., Takahashi,T., and Cull-Candy,S.G. (1994). NMDA-receptor channel diversity in the developing cerebellum. Nature 368,335-339. Feldmeyer,D. and Cull-Candy,S. (1996). Functional consequences of changes in NMDA receptor subunit expression during development. JNeurocytol. 25, 857-867. Ferguson,S.S ., Zhang,J., Barak,L.S ., and Caron,M.G. (1 998). Role ofbeta-arrestins in the intracelldar trafficking of G-protein- coupled receptors. Adv.Pharrnaco1. 42,420-424. Figurov,A., Boddeke,H., and Mdler,D. (1993). Enhancement of AMPA-mediated synaptic transmission by the protein phosphatase inhibitor caiyculin A in rat hippocampal slices. Eur J Neurosci 5, 1035- 1041. Freund,G.G., Wittig,J.G., and Mooney,RA. (1995). The PD-kinase serine kinase phosphorylates its p85 subunit and IRS-1 in PI3-kinaselIRS-1 complexes. Biochem.Biophys.Res.Co~unun206,272- 278. Fruman,D.A., Meyers,R.E., and Cantley,L.C. (1998). Phosphoinositide kinases. Annu-Rev-Biochem. 67,48 1-507. Fujisawa,H., Dawson,D., Browne,S .E., Mackay,K.B ., Bullock,R., and McCulIoch,J. (1 993). Pharmacological modification of glutamate neurotoxicity in vivo. Brain Research 629,73-78. Gaisano,H.Y., Sheu,L., Foskett,J.K., and Trimble,W.S. (1994). Tetanus toxin light chain cleaves a vesicle-associated membrane protein (VAMP) isoform 2 in rat pancreatic zymogen granules and inhibits enzyme secretion. J.Bio1.Chem. 269, 17062-17066. Ga~l1ier~J.M.~Simonsen,A., D'ArrigoA., Bremnes,B., and Stenmark,H. (1999). FYVE finger proteins as effectors of phosphatidylinositol3-phosphate. Chem.Phys.Lipids 98, 87-94. Gomrnerman,J.L., Rottapel,R., and Berger,S.A. (1997). Phosphatidylinositol 3-kinase and Ca2+ idux dependence for ligand- stimulated internalization of the c-Kit receptor. J-Bid-Chem. 272, 305 19-30525. Gomperts,S.N. (1996). Clustering membrane proteins: It's all coming together with the PSD- 95/SAP90protein family. Cell 84,659-662. Gomperts,S.N., Rao,A., Craig,A.M., Malenka,R.C., and Nicol1,R.A. (1998). Postsynaptically silent synapses in single neuron cultures. Neuron 21, 1443- 1451. Goodman,O.B.J., Krupnick,J.G., Santini,F., Gurevich,V.V., Penn,R.B ., Gagnon,A. W., Keen,J.H., and Benovic, J.L. (1 998). Role of arrestins in G-protein-coupled receptor endocytosis. Adv-Pharmacol. 42,429-43 3, Greengard,P., Jen,J., Nairn,A.C., and Stevens,C.F. (199 1). Enhancement of the glutamate response by CAMP-dependentprotein kinase in hippocampal neurons. Science 253, 1135-1 13 8. HAJDUCH,E., ALEDO,J.C., WATTS,C., and HUNDALJ3.S. (1 997). Proteolytic cleavage of cellubrevin and vesicle-associated membrane protein (VAMP) by tetanus toxin does not impair insulin- stimulated glucose transport or GLUT4 translocation in rat adipocytes. Biochem. J. 321,233- 238- Hansen,S.H., Sandvig,K., and van Deurs,B. (1 993b). Clathrin and HA2 adaptors: effects ofpotassium depletion, hypertonic medium, and cytosol acidification. J.Ce11 Biol, 121,61-72. Hansen,S.H., Sandvig,K., and van,D.B. (1993a). Clathrin and HA2 adaptors: effects of potassium depletion, hypertonic medium, and cytosol acidification. J Cell Biol121,61-72, Harder,T., ScheiEele,P., Verkade,P., and Sirnons,K. (1998). Lipid domain structure of the plasma membrane revealed by patching of membrane components. J-Cell Biol. 141,929-942. Haring,H,U., Tippmer,S., Kellerer,M., Mosthaf,L., Kroder,G., Bossenmaier,B., and Berti,L. (1996). Modulation of insulin receptor signaling. Potential mechanisms of a cross talk between bradykinin and the insulin receptor. Diabetes 45,S 1 15-S 119. Harvey,J., McKay,N.G., Walker,K.S., Van der,K.J., Downes,C.P., and Ashford,M.L. (2000). Essential role of phosphoinositide 3-kinase in leptin-induced K(ATP) channel activation in the rat CRI-Gl insulinoma cell line. J-BioLChem, 2 75,46604669- Hawtborne,J.N. (1996). Phosphoinositides and synaptic transmission. Subcell.Biochem. 26,43-57. Hayashi,T., Umemori,H., MishingM., and Yamamoto,T. (1999). The AMPA receptor interacts with and signals through the protein tyrosine kinase Lyn. Nature 397,72-76. Hayashi,Y., Shi,S.H., Esteban,J.A., Piccini,A., Poncer,J.C., and Malinow,R. (2000). Driving AMPA receptors into synapses by LTP and CaMKII: requirement for GluRl and PDZ domain interaction, Science 287,2262-2267. Herskovits,J.S., Burgess,C.C-, Obar,R.A., and Vallee,R.B. (1993). Effects of mutant rat dynarnin on endocytosis. J.Celf Biol. 122, 565-578. Hiles,I.D., Otsu,M., VoliniqS., Fry,M.J., Gout,I., Dhand,R., Panayotou,G., Ruiz-Larrea,F., Thompson,A., and Totty,N.F. (1 992). Phosphatidylinositol3-kinase:structure and expression of the 110 kd catalytic subunit. Cell 70,4 19-429. Hisatsune,C., Umemori,K., Mishina,M., and Yamarnoto,T. (1 999). Phosphorylation-dependent interaction of the N-methyl-D-aspartate receptor epsilon 2 subunit with phosphatidylinositol3-kinase. Genes Cells 4,657-666. HollmannJA. and Heinemann,S. (1 994). Cloned glutamate receptors. Annu.Rev.Neurosci. I 7,3 1- 108. Holman,G.D. and Cushman,S+W. (1994). Subcelldar localization and trafficking of the GLUT4 glucose transporter isoform in insulin-responsive cells. BioEssays 16,753-759. Hua,S.Y., Raciborska,D.A., Trirnble,W.S., and Char1t0n~M.P.(1998). Different VAMPkynaptobrevin complexes for spontaneous and evoked transmitter release at the crayfish neuromuscular junction. J.Neurophysiol.80,323 3-3246. Ruganir,R.L. and Greengard,P. (1 990). Regulation of neurotransmitter receptor desensitization by protein phosphorylation, Neuron 5,555-567. Iino,M., Koike,M., Isa,T., and Ozawa,S. (1996). Voltage-dependent blockage of Ca(2+)-permeable AMPA receptors by joro spider toxin in cultured rat hippocampal neurones. J-Physiol (Lond) 496 ( ~r 2), 43 1-437. Irie,M., Hata,Y., Takeuchi,M., Ichtchenko,K., ToyodqA., Hirao,K., Takai,Y ., Rosah1,T. W., and Sudhof,T.C. (1 997). Binding of neuroligins to PSD-95. Science 2 77, 15 1 1-15 15. Isaac,J.T., Nicoll,R.A., and Malenka,R.C. (1995). Evidence for silent synapses: implications for the expression of LTP. Neuron 15,427-434. Joly,M., Kazlauskas,A., and Corvera,S. (1 995). PhosphatidylinositoI 3-kinase activity is required at a postendocytic step in platelet-derived growth factor receptor traffTcking. J.Bio1.Chem. 2 70,13225- 13230. Jones,S.M. and Howel1,K.E. (1997). Phosphatidylinositol3-kinase is required for the formation of constitutive transport vesicles fi-om the TGN. J.Cell Biol, 139,339-349. Joyal,J.L., Burks,D.J., Pons,S., Matter, W.F., VIahos,C.J., White,M.F., and Sacks,D.B. (1 997). Calmoddin activates phosphatidylinositol3-kinase. J-Biol-Chem.2 72,28 183-28 186- Jugloff,D.G., Kha~a,R,Schlichter,L.C., and Jones,O.T. (2000). Internalization of the Kv1.4 potassium channel is suppressed by clustering interactions with PSD-95. J-Biol-Chem.2 75, 1357- 1364. Kameyama,K., Lee,H.K., Bear,M.F., andHuganir,R.L. (1 998). Involvement of apostsynaptic protein kinase A substrate in the expression of homosynaptic long-term depression. Neuron 21, 1163-1 175. Kande1,E.S. and Hay,N. (1999). The regdation and activities of the multifhnctional serine/threonine kinase Akt/PKB- Exp-Cell Res. 253,2 10-229. Kanzaki,M., Zhang,Y.Q., Masjedi,H., Li,L., Shibata,H., and KojirnaJ. (1999). Translocation of a cdcium-permeable cation channel induced by insulin-like growth factor--I. Nature Cell Biol. I, 165- 170. Kao,A. W., Ceresa,B.P., Santeler,S.R., and Pessin,J.E. (1998). Expression of a dominant interfering dynaminmutant in 3T3L 1 adipocytes inhibits GLUT4 endocytosis without affecting insulin signaling. .J.Biol.Chem. 2 73,25450-25457. Karoor,V., Wang,L., Wang,H.Y., and Malbon,C.C. (1998). Insulin stimulates sequestration of beta- adrenergic receptors and enhanced association of beta-adrenergic receptors with Grb2 via tyrosine 350. J.Biol.Chem. 273,33035-33041. Keinanen,K., Wisden, W., Sornmer,B., Werner,P., Herb,A., Verdoorn,T.A., Sakrnann,B., and Seeburg,P .H. (1 990). A family of AMFA-selective glutamate receptors. Science 249, 556-560. Kelly,A. and Lynch,M.A. (2000). Long-term potentiation in dentate gyms of the rat is inhibited by the phosphoinositide 3-kinase inhibitor, wortmannin. Neuropharmacology 39,643-65 1. Kennedy,M.B. (1997). The postsynaptic density at glutamatergic synapses. Trends Neurosci. 20,264- 268. Kim,E., NaisbiqS., Hsueh,Y.P., Rao,A., Rothschild,A., Craig,A.M., and Sheng,M. (1997). GKAP, a novel synaptic protein that interacts with the guanylate base- like domain of the PSD-95/SAP90 family of channel clustering molecules. J.Cel1 Biol. 136,669-678. Kim,J.H., Liao,D., Lau,L.F., and Huganir,R.L. (1998). SynGAP: a synaptic RasGAP that associates with the PSD-95/SAP90 protein family. Neuron 20, 683-691. Kim,E.; Niethammer,M.; Rothschild,A.; Jan,Y.N.; Sheng,M. (1995) Clustering of Shaker-type Kt channeIs by interaction with a family of membrane-associated guanylate kinases. Nature, 375,85-88 Kistner,U., Wenze1,B .M., Veh,R-W., Cases-Langhoff,C., Garner,A.M., Appeltauer,U., Voss,B., Gundelfinger,E.D., and Garner,C.C. (1993). SAP90, a rat presynaptic protein related to the product of the Drosophila tumor suppressor gene dlg-A- J-Biol-Chem.268,45804583- Kodaki,T., Woscholski,R., Hallberg,B., Rodriguez-VicianqP., DownwardJ., and Parker,P.J. (1 994). The activation of phosphatidyIinositol3-kinaseby Ras. Cur-Biol. 4,798-806. Kohn,A.D., Takeuchi,F., and Roth,R.A. (1996). Akt, a pleckstrin homology domain containing kinase, is activated primarily by phosphorylation. J.Biol.Chem. 271,21920-2 1926. Kohr,G. and Seeburg,P.H. (1996). Subtype-specific regulation of recombinant NMDA receptor- channels by protein tyrosine kinases of the src family. J.Physio1 (Lond) 492 (Pt 2), 445-452. Kornau,H.-C.; Schenker,L.T.; Kennedy,M.B,; Seeburg,P.H (1995). Domain interaction between NMDA receptor subunits and the postsynaptic density protein PSD-95. Science, 469, 1737-1740 Kotani,K., Carozzi,A.J., Sakaue,H., Hara,K., Robinson,L.J., CIark,S.F.Y.K., James,D.E., and Kasuga,M. (1 995). Requirement for phosphoinositide 3-kinase in insulin-stimulated GLUT4 translocation in 3T3-L 1 adipocytes. Biochemical & Biophysical Research Communications 209,343- 348. Krupnick,J.G., Santini,F., Gagnon,A. W., Keen,J.H., and Benovic,J.L. (1 997). Modulation of the arrestin-clae interaction in cells. Characterization of beta-arrestin dominant-negative mutants. J-Biol-Chem. 2 72,32507-32512. Laporte,S.A., Oakley,R.H-, Zhang,J., Holt,J.A,, Ferguson,S.S., Caron,M.G., and Barak,L.S. (1999). The beta2-adrenergic receptor/betaarrestin complex recruits the clathrin adaptor AP-2 during endocytosis. Proc.Nat1.Acad.Sci.U.S.A. 96,37 12-37 17. Laporte SAYOakley RH, Holt .JAYBarak LS, Caron MG (2000) The interaction of beta-arrestin with the AP-2 adaptor is required for the clustering of beta 2-adrenergic receptor into ciabin-coated pits. J Bid Chem 275 :23 120-6 Larkman,A.U. and Jack,J.J. (1995). Synaptic plasticity: hippocampal LTP. Current Opinion in Neurobiology 5, 324-334. Lau,L,-F. and Huganir,R.L. (1995). Differential tyrosine phosphorylation of N-Methyl-D-asparate receptor subunits, J.Biol.Chem. 270,20036-20041. Lau,L.F. and Hugank,R.L. (1994). Tyrosine phosphorylation of NMDA receptor subunits. Soc.Neurosci.Abst. 20 Suppl. PT 1&2. Lee,H.K., Barbarosie,M., Karneyama,K., Bear,M.F., and Huganir,R.L. (2000). Regulation of distinct AMPA receptor phosphorylation sites during bidirectional synaptic plasticity. Nature 405,955-959. Lee,H.K., Karneyama,K., Huganir,R.L., and Bear,M.F. (1 998). NMDA induces long-term synaptic depression and dephosphorylation ofthe GluRl subunit of AMPA receptors in hippocampus. Neuron 21, 1151-1162. Leevers,S.J., Vanhaesebroeck,B., and Waterfield,M.D. (1999). Signalling through phosphoinositide 3 -kinases: the lipids take centre stage- Curr.Opin.Cel1 Biol. 11,219-225. Leonard,A.S. and Hel1,J.W. (1997). Cyclic AMP-dependent protein kinase and protein kinase C phosphorylate N-methyl-D-aspartate receptors at difikrent sites. J.Biol.Chem. 2 72, 12107- 121 15- Levenes,C., Daniel,H., and Crepe1,F. (1998). Long-term depression of synaptic transmission in the cerebellum: cellular and molecular mechanisms revisited. Progress in NeurobioIogy 55,79-91. Levitan,I.B. (1994). ModuIation of ion channels by protein phosphorylation and dephosphorylation. Annu.Rev.Physiol.56,193-212. Li,P, and Zhuo,M. (1998). Silent glutamatergic synapses and nociception in mammalian spinal cord. Nature 393,695-698. Liao,D., Hessler,N.A., and Malinow,R. (1995). Activation of postsynaptically silent synapses during pairing- induced LTP in CAI region of hippocampal slice. Nature 3 75,400-404. Lin,J.W. and Sheng,M. (1998). NSF and AMPA receptors get physical. Neuron 21,267-270. Lin,J.W., Wyszynski,M., Madhavan,R., Sealock,R., Kim,J.U., and Sheng,M. (1 99%). Yotiao, anovel protein of neuromuscular junction and brain that interacts with specific splice variants of NMDA receptor subunit NRl . J.Neurosci. 18,20 17-2027. Lin,S.Y., Wu,K., Levine,E.S., Mount,H.T., Suen,P.C., and Black,I.B. (1998a). BDNF acutely increases tyrosine phosphorylation of the NMDA receptor subunit 2B in cortical and hippocampal postsynaptic densities. Brain Res.Mol.Brain Res. j.5,20-27. Lissin,D.V., CarroIl,R.C., Nicoll,R.A., Malenka,R.C., and von Zastrow,M. (1999). Rapid, activation- induced redistribution of ionotropic glutamate receptors in cultured hippocampal neurons. J.Neurosci. 19,1263-1272. Liu, J.P- (1997). Protein phosphorylation events in exocytosis and endocytosis. Clin Exp Pharmacol Physiol. 24 (8): 61 1-8. Lledo,P.M., Zhang,X., Sudhof,T.C., MalenkgR-C., andNicol1,R.A. (1998). Postsynaptic membrane fusion and long-term potentiation. Science 2 79,3 99-403. Lu,Y.M., Roder,J.C., Davidow,J., and Salter,M.W. (1998). Src activation in the induction of long- term potentiation in CA1 hippocampal neurons. Science 279, 1363- 1367. Luscher,C., Xia,H., Beattie,E.C., Carroll,R.C,, von Zastrow,M., Malenka,R-C., and Nicol1,R.A. (I 999). Role of AMPA receptor cycling in synaptic transmission and plasticity. Neuron 24, 649-658. Luthi,A., Chittajallu,R., Duprat,F., Palmer,M.J., Benke,T.A., Kidd,F.L., Henley,J.M., Esaac,J.T., and ColIingridge,G.L. (1999). Hippocampal LTD expression invokes a pool of AMPARS regulated by the NSF-GluR2 interaction. Neuron 24,3 89-3 99. MaIenka,R.C. and Nicoll,R.A. (1997). Silent synapses speak up. Neuron 19,473-476. Malenka,R.C. and Nicol1,R.A. (1999). Long-term potentiation--a decade of progress? Science 285, 1870-1874. Maletic-Savatic,M., Koothan,T., and Mdinow,R. CaMKZI action on microtubules promotes calcium- evoked dendritic exocytosis. Soc.Neuosci.Abstr. 22, 2 10.10. 1996. Malinow,R. (1998). Silencing the controversy in LTP?. Neuron 2 I, 1226-1327. Marnrnen,A.L., Kameyama,K., Roche,K. W., and Huganir,R.L. (1 997). Phosphorylation ofthe alpha- amino-3-hydroxy-5-methylisoxazole4- propionic acid receptor GluRl subunit by Calcium/ calmoddin-dependent base 11. J-Biol-Chem. 272,32528-32533. Man,Y.H., Lin,J.W., Ju,W.H., Ahmadian,G., Liu,L-, Becker,L.E., Sheng,M., and Wang,Y.T. (2000). Regulation of AMPA receptor-mediated synaptic transmission by clathrin-dependent receptor internalization. Neuron 25,649-662. Marks,B. and McMahon,H.T. (1 998). Calcium triggers calcineurin-dependent synaptic vesicle recycling in mammalian nerve terminals. Curr-Biol. 8,740-749. Marte,B.M. and Downward,J. (1997). PKBW:connecting phosphoinositide 3-kinase to cell survival and beyond. Trends.Biochem.Sci. 22, 355-358. Martin,S.S., Haruta,T., Morris,A.J., Klippel,A., Williams,L.T., and Olefsky,J.M. (1996). Activated phosphatidylinositol3-kinase is sufEcient to mediate actin rearrangement and GLUT4 translocation in 3T3-L 1 adipocytes. J-Biol-Chern,271, 17605-17608. Martin,T.F. (I 997). Phosphoinositides as spatial regulators of membrane traffic. Curr.Opin.Neurobio1. Martys,J.L., Wjasow,C., Gangi,D.M., Kielian,M.C., McGraw,T.E., and BackerJ-M. (1996). Wortmannin-sensitive tracking pathways in Chinese hamster ovary cells. Differential effects on endocytosis and IysosomaI sorting. J-Biol-Chem-271, 10953-10962. MatsudqS., Mikawa,S., and Hirai,H. (1999). Phosphorylation of serine-880 in GluR2 by protein base C prevents its C terminus from binding with glutamate receptor-interacting protein. J-Neurochem, 73, 1765- 1768, Matsuda,S., Launey T, Mikawa,S., and Hlrai,H. (2000). Disruption of AMPA 'eceptor GluR2 clusters following long-term depression induction in cerebellar Purkinje neurons. EMBO J. 19(12):2765-74. Mayer,M.L., Westbrook,G.L., and Guthrie,P.B- (1984). Voltage-dependent block by Mg of NMDA responses in spinal cord neurones. Nature 309,261-263. McGfade-McCulloh,E., Yamamoto,H., Tan,S.-E., Brickey,D.A., and Soderling,T.R. (1 993). Phosphorylation and regulation of glutarnate receptors by calcium/calmodulin-dependent protein kinase 11. Nature 362,640-642. McIlhinney,R.A., Molnar,E., Atack,J.R., and Whiting,P.J. (1 996b). Cell surface expression of the human N-methyl-D-aspartate receptor subunit I a requires the co-expression of the NR2A subunit in transfected cells. Neuroscience 70,989-997. McII.hinney,R.A.J., MolnBr,E., AtackJR., and Whiting,P.J. (1996a). Cell surface expression of the human N-Methyl-D-Asparate receptor subunit la requires the co-expression of the NR2A subunit in transfected cells, Neuroscience 70, 989-997. MelLmanJ. (1996). Endocytosis and molecular sorting. Ann.Rev.Cel1 Dev-Biol. 12, 575-625. Monyer,H., Sprengel,R., Schoepfer,R., Herb,A., Higuchi,M., Lomeli,H., Burnashev,N., SakmanqB., and Seeburg,P.H. (1992). Heteromeric NMDA receptors: Molecular and fimctional distinction of subtypes. Science 256, 1217-122 1. MoonJ-S., Apperson,M.L., and Kennedy,M.B. (1 994). The major tyrosine-phosphorylated protein in the postsynaptic density fiaction is N-methyl-D-aspartate receptor subunit 2B. Proc.Natl,Acad.Sci.USA 91, 3954-3 958. M0rgan~D.0.and Roth,R.A. (1987). Acute insulin action requires insulin receptor kinase activity: introduction of an inhibitory monoclonal antibody into mammalian cells blocks the rapid effects of insulin. Proc Natl Acad Sci U S A 84,4 1-45. Moriyoshi,K., Masu,M., Ishii,T., Shigemoto,R., Mimo,N., and Nakanishi,S. (199 1). Molecular cloning and characterization of the rat NMDA receptor. Nature 354, 3 1-37. Mosthaf,L., Kellerer,M., Muhllzofer,A., Mushack,J., Seffer,E., and Haring,H.U. (1996). Insulin leads to aparallel translocation of PI-3-kinase and protein kinase C zeta. Exp Clin Endocrinol Diabetes 104, 19-24, M&ey,R.M.; Herron,C.E.; Malenka,R.C. (1993) An essential role for protein phosphatases in hippocampal long-term depression. Science 26 1, 1051 - 1055 Muller,B .M., Kistner,U., Kindler,S., Chung, W.J., Kuhlendah1,S ., Fenster,S .D., Lau,L.F., Veh,R. W., Huganir,R.L., G~nde1finger~E.D.~and Garner,C.C. (1996). SAP 102, a novel postsynaptic protein that interacts with NMDA receptor complexes in vivo. Neuron 17,255-265. Muller,B.M., Kistner,U., Veh,R.W., Cases-Langhoff,C., Becker,B., Gundelfinger,E.D., and Garner,C.C. (1 995). Molecular characterization and spatial distributionof SAPanovel presynaptic protein homologous to SAP90 and the Drosophila discs-large tumor suppressor protein. J.Neurosci. 15,2354-2366, Nakazawa,K., Mikawa,S., Hashikawa,T., and Ito,M. (1 995). Transient and persistent phosphorylation of AMPA-type glutamate receptor subunits in cerebellar purkinje cells. Neuron 15,697-709. NesterovA., Carter,R.E., Sorkina,T., Gill,G.N., and Sorkin,A. (1 999). Inhibition of the receptor- binding hction of clathrin adaptor protein AP-2 by dominant-negative mutant mu2 subunit and its effects on endocytosis. EMBO J. 18,2489-2499. Ng SSW, Tsao MS, Chow S, Hedky DW (2000) Inhibition of phosphatidylinositide 3-kinase enhances gemcitabine-induced apoptosis in human pancreatic cancer cells-Cancer Res 60(19):545 1- 5 Niethammer,M., Kim,E., and Sheng,M. (1996). Interaction between the C terminus of NMDA receptor subunits and multiple members of the PSD-95 family of membrane-associated guanylate kinases. JNeurosci. 16,2 157-2163. Niethammer,M., Valtschanoff,J.G., Kapoor,T.M., Allison,D.W., Weinberg,R.J., Craig,A.M., and Sheng,M. (1998). CRIPT, a novel postsynaptic protein that binds to the third PDZ domain of PSD- 95/SAP90. Neuron 20,693-707, Nishimune,A., Isaac,J.T., Molnar,E., Noel,J., Nash,S .R., Tagaya,M., CoUingridge,G.L., Nakanishi,S., and Henley, J.M. (1998). NSF binding to GluR2 regulates synaptic transmission. Neuron 21, 87-97. NoelJ., Ralph,G.S ., Pickard,L., Williarns,J., Molnar,E., Uney,J.B., Collingridge,G.L., and Henley,J.M. (1 999). Surface expression of AMPA receptors in hippocampal neurons is regulated by an NSF-dependent mechanism. Neuron 23,365-376, OIDell,T.J., Kandel,E.R., and Grant,S .G. (1 99 1). Long-term potentiation in the hippocampus is blocked by tyrosine kinase inhibitors- Nature 353, 558-560. Ornkumar,R,V., Kiely,M.J., Rosenstein,A.J., Min,K.T., and Kennedy,M.B . (1996). Identification of a phosphorylation site for calcium/cahnodulindependentprotein kinase I1 in the NR2B subunit of the N-methyl-D-aspartate receptor. J.Biol.Chem. 2 71,3 1670-3 1678. Osten,P., Srivastava,S., Inman,G.J., Vilim,F.S., Khatri,L., Lee,L.M., States,B.A., Einheber,S., Milner,T.A., Hanson,P.I., and Ziff,E.B. (I 998). The AMPA receptor GluR2 C terminus can mediate a reversible, ATP- dependent interaction with NSF and alpha- and beta-SNAPS. Neuron 2 1, 99-2 10. Owen,D.J., Wigge,P., Vallis,Y., Moore,J.D., Evans,P.R., and McMahon,H.T. (1998). Crystal structure of the amphiphysin-2 SH3 domain and its role in the prevention of dynamin ring formation. EMBO J. 17,5273-5285. PeUegrini-GiampietroYDDEEyGorter,J.A., [email protected]., andZukin,R.S. (1 997). The GluR.2 (GluR-B) hypothesis: Ca(2+)-permeable AMPA receptors in neurological disorders. Trends-Neurosci20,464- 470. Pellizzari,R., Rossetto,O., Washbourne,P., Tonello,F., Nicotera,P.L., and Montecucco,C. (1998). In vitro biological activity and toxicity of tetanus and botulinurn neurotoxins. Toxicol-Lett. 102-103, 191-197. Perkinton,M.S., Sihra,T.S., and Williams,R.J. (1 999). Ca(2-t)-permeable AMPA receptors induce phosphorylation of CAMPresponse element-binding protein through a phosphatidylinositol3-kinase- dependent stimulation of the mitogen-activated protein kinase signaling cascade in neurons. JNeurosci. 19, 5861-5874. Pessin,J.E., Thurrnond,D.C., ELmendorfJ-S., Coker,K.J., and Okada,S. (1999). Molecular basis of insulin-stimulated GLUT4 vesicle tr&cking. Location! Location! Location! J-Biol-Chem.2 74,2593- 2596. Pettit,D.L., Perlman,S., and Malinow,R. (1993). Postsynaptic expression of constitutively active CaMKII occludes LTP in the hippocampal slice. Soc.Neurosci.Abstr- 19. Pettit,D.L., Perlrnan,S ., and Malinow,R- (1994). Potentiated transmission and prevention of Wer LTP by increased CaMKII activity in postsynaptic hippocampal slice neurons. Science 266, 188 1- 1885. Prince,H.K., Conn,P.J., BIackstone,C.D., Huganir,R.L., and Levey,A.I. - Down-regulation of AMPA receptor subunit GluR2 in amygdaloid kindling. - J Neurochem 1995 Jan;64(1):462-5 462-55. Quon,M.J., Chen,H., Ing,B.L., Liu,M.L., Zarnowski,M.J., Yonezawa,K., Kasuga,M., Cushman,S. W., and Tayl0r~S.1.(1995a). Roles of 1-phosphatidylinositoI 3-kinase and ras in regulating translocation of GLUT4 in transfected rat adipose cells. Mol.Cell Biol. IS, 5403-541 1. Quon,MJ., Chen,H., Ing,B.L., Liu,M-L., Zarnowski,M.J., Yonenwa,K., Kasuga,M., Cushman,S .W., and Tay10r~S.I.(1 995b). Roles of I-phosphatidylinositol3-base and ras in regulating translocation of GLUT4 in tcansfected rat adipose cells. Mol Cell Biol15,5403-5411. Raizada,M.K. (1983). Localization of insulin-like immunoreactivity in the neurons from primary cultures of rat brain, Experimental Cell Research 143,351-357- Rao,A., Harms,K.J., and Craig,A.M. (2000). Neuroligation: building synapses around the neurexin- neuroligin link. NatNeurosci. 3,747-749. Raymond,L.A., Blackstone,C.D,, and Huganir,R.L. (1993). Phosphorylation of amino acid neurotransmitter receptors in synaptic plasticity, Trends Neurosci. Id, 147-153. Roche,K.W., OIBrien,R.J., Marnmen,A.L., Bernhardt,J., and Huganir,R.L. (1996)- Characterization of multiple phosphorylation sites on the AMPA receptor GluRl subunit. Neuron 16, 1 179-1 188. Roche,K. W., Tingley, W.G., and Huganir,R.L. (1994). Glutamate receptor phosphory lation and synaptic plasticity. Current Opinion in Neurobiology 4, 383-388. Rodriguez-Viciana,P., Marte,B.M., Warne,P.H., and Downward,J. (1996). PhosphatidyiinositoI 3' base: one of the effectors of Ras. Philos.Trans.R.Soc.Lond B BiolSci. 351,225-23 1. Rosenmund,C., Stem-Bach,Y., and Stevens,C.F. (1998). The tetrameric structure of a glutamate receptor channel. Science 280, 1596-1599. Rosenrnund,C. and Westbrook,G.L- ( 1993). Calcium-induced actin depolymerizationreduces NMDA channel activity. Neuron 10,805-8 14. Rostas,J.A., Brent,V.A., Voss,K,, Errington,M.L., Bliss,T.V., and Gurd,J.W. (1 996). Enhanced tyrosine phosphorylation of the 2B subunit of the N-methyl-D- aspartate receptor in long-term potentiation. Proc Natl Acad Sci U S A 93, 10452-10456. Rothman,J.E. (1994). Mechanisms of intracellular protein transport. Nature 3 72, 55-63. Schekman,R and Orci,L. (1996). Coat proteins and vesicle budding. Science 271, 1526-1533. Schmid,S.L. (1997). Clathrin-coated vesicle formation and protein sorting: an integrated process. Annu.Rev.Biochem. 66,s1 1-548. Seeburg,P.H. (1996). The role of RNA editing in controlling glutamate receptor channel properties. J.Neurochem. 66, 1-5. Setou M, Nakagawa T, Seog DH, Hirokawa N. (2000). Kinesin superfamily motor protein KIF 17 and rnLin- 10 in NMDA receptor-containing vesicle transport. Science. 28 8(5472): 1796-802. Shi,S.H., Hayashi,Y., PetraLiqRS., Zaman,S.H., WenthoId,RJ., Svoboda,K., andMalinow,R. (1999). Rapid spine delivery and redistribution ofAMPA receptors after synaptic NMDA receptor activation. Science 284, 1811-1816. Shupliakov,O., Low,P., Grabs,D., Gad,H., Chen,H., David$., Takei,K., De Carnilli,P., and Brodin,L. (1997). Synaptic vesicle endocytosis impaired by disruption of dynamin-SH3 domain interactions. Science 2 76,259-263. Sige1,E. (1995). Functional modulation of ligand-gated GABAA and NMDA receptor channels by phosphorylation. J.Recept.Signa1.Transduct.Res. 15,325-332. Sims,K.D., Stre,D.J., and Robinson,M.B. (2000). Platelet-derived growth factor rapidly increases activity and cell surface expression of the EAAC1 subtype of glutamate transporter through activation of phosphatidylinositoi 3-kinase. J-Biol-Chem. 275, 5228-5237. Soderling,T.R., Fukunaga,K., Brickey,D.A., Fong,Y.L., Rich,D.P., and Srnith,K.C.RJ. (1 99 1). Molecular and cellular studies on brain calcium/calmodulin- dependent protein kinase 11. Progress in Brain Research 89, 169-1 83. Soloviev,M.M. and Barnard,E.A. (1997). Xenopus oocytes express a unitary glutamate receptor endogenously. J.Mol.Biol. 273, 14- 18. Sornmer,B ., Keinanen,K., Verdoom,T.A-, Wisden, W., Bumashev,N., Herb,A., Kohler,M., Takagi,T., Sakmann,B ., and SeeburgP-H. (1 990)- Flip and flop: a cell-specific functional switch in glutamate- operated channels of the CNS. Science 249,1580-1585. SongJ., Kamboj,S., Xia,J., Dong,H., Liao,D., and Huganir,R.L. (1998). Interaction of the N- ethylmaleimide-sensitive factor with AMPA receptors. Neuron 21,393400. Songyang,Z., Fanniog,A.S ., Fu,C., XuJ., Marfatia,S.M., Chishti,A.H., Crompton,A., Chan,A.C., Anderson,J.M., and Cantley,L.C. (1 997). Recognition of unique carboxyl-terminal motifs by distinct PDZ domains. Science 275,73-77. Srivastava,S., Osten,P., Vilim,F.S., Khatri,L., InrnqG., States,B., Daly,C., DeSouza,S., Abagyan,R., Valtschanoff,J.G., Weinberg,R.J., and Ziff,E.B. (1998). Novel anchorage of GluR2/3 to the postsynaptic density by the AMPA receptor-binding protein ABP. Neuron 21,58 1-59 1. Stanton,P.K. - LTD, LTP, and the sliding threshold for long-term synaptic plasticity. - Hippocampus l996;6(1):3 5-42 35-422. Staudinger,J., Zhou,J., Burgess,R., Elledge,S.J., and Olson,E.N. (1995). PICK1 :a perinuclear binding protein and substrate for protein kinase C isolated by the yeast two-hybrid system. J.Cell Biol. 128, 263-27 1. Stenmark,H. and Aasland,R. (1999). FYVE-finger proteins--effectors of an inositol lipid. J.Cell Sci. I12 (Pt 23), 4175-4183. Stevens,C.F., Tonegawa,S ., and Wang,Y- (1 994). The role of calcium-calmoddin kinase II in three forms of synaptic plasticity. Current Biology 4, 687-693. Strack,S., Choi,S., Lovinger,D.M., and C01bran~R.J.(1997). Translocation of autophosphorylated calcium/calmoduIin-dependent protein kinase I1 to the postsynaptic density. J.Biol.Chem. 272,13467- 13470, Strack,S. and Colbran,R-J. (1 998). Autophosphorylation-dependent targeting of Calcium/ calmodulin- dependent protein kinase II by the NR2B subunit ofthe N-methyl- D- aspartate receptor. J.Bio1.Chem. 273,20689-20692. Sucher?N.J., Awobuluyi,M., Choi,Y.B., and Lipton,S.A. (1996). NMDA receptors: &om genes to channels. Trends PhannacolSci. 17,348-355. Sudhof,T.C. (1 995). The synaptic vesicle cycle: a cascade of protein-protein interactions. Nature 3 75, 645-653. Swope,S.L., Moss,S.I., Raymond,L.A., and Huganir,R.L. (1999). Regulation of ligand-gated ion channels by protein phosphorylation. Adv-Second Messenger Phosphoprotein Res. 33,49-78. Takagi,N., Cheung,H.K., Bissoon,N., Teves,L., Wallace,M.C., and Gurd,J,W. (1 999). The effect of transient global ischemia on the interaction of Src and Fyn with the N-methyl-D-aspartate receptor and postsynaptic densities: possible involvement of Src homology 2 domains. J-Cereb-BloodFlow Metab 19,880-888. Tehrani,M.H. and Barnes,E.M.J. (1993). Identification of GABAA/benzodiazepine receptors on clathrin- coated vesicles fiom rat brain. J-Neurochern-60, 1755-176 1. Thompson,S.M. (1994). Modulation of inhibitory synaptic transmission in the hippocampus. Progress in Neurobiology 42, 575-609. Tingley, W.G., Ehlers,M.D., Kameyama,K., Doherty,C., Ptak,J.B ., Riley,C.T., and Huganir,R.L. - Characterization of protein kinase A and protein kinase C phosphorylation of the N-methyl-D- aspartate receptor NR1 subunit using phosphorylation site-specific antibodies. - J Bioi Chem 1997 Feb 21;272(8):S 157-66 5 157-5666. Toker,A. and Newton,A.C. (2000). Aktlprotein kinase B is regulated by autophosphorylation at the hypothetical PDK-2 site, J-Biol-Chem. 275, 827 1-8274. Torres,R., Firestein,B.L., Dong,H., Staudinger,J., Olson,E.N., Huganir,R.L., Bredt,D,S., Gale,N.W., and Yancopoulos,G.D. (1998). PDZ proteins bind, cluster, and synaptically colocalize with Eph receptors and their ephrin ligands. Neuron 2 I, 1453- 1463. Toth,K. and McBain,C.J. (1 998). Merent-specific innervation of two distinct AMPA receptor subtypes on single hippocampal interneurons. Nat.Neurosci. 1, 572-578. Vallee,R.B. and Sheet2;M.P. (1996). Targeting of motor proteins. Science 2 71, 1539-1544. van,D.S., Schumaches,C., Hage,W., Verkleij,A.J., and van Bergen en Henegouwen PM (1997). Associationand colocalization of Eps 15 with adaptor protein-2 and clathrin. J Cell Bio1136,S 11-82 1. Vanhaesebroeck,B., Stein,R.C., and Waterfield,M.D. (1996). The study ofphosphoinositide3-kinase function- Cancer Surv. 27,249-270. Vazquez,F, and Sellers,W.R. (2000). The PTEN tumor suppressor protein: an antagonist of phosphoinositide 3-kinase signaling. Biochim.Biophys.Acta 1470, M21-M35. von Kitzing,E., Jonas,P ., and Sakmann,B. (I 994). Quantal analysis of excitatory postsynaptic currents at the hippocampal mossy fiber-CA3 pyramidal cell synapse. Advances in Second Messenger & Phosphoprotein Research 29,235-260. Voronin,L.L. (1994)- Quantal analysis of hippocampal long-term potentiation. meview]. Reviews in the Neurosciences 5, 14 1- 170. Walsh,M.J. and Kuruc,N. (1992). The postsynaptic density: constituent and associated proteins characterized by electrophoresis, immuno blotting, and peptide sequencing. J-Neurochem. 59,667- 678 - Wan,Q., Xiong,Z.G., Man,H.Y., Ackerley,C.A., Braunton,J., Lu, WX,Becker,L.E., MacDonald,J.F., and Wang,Y.T. (1997). Recruitment of fbnctional GAElA(A) receptors to postsynaptic domains by insdin. Nature 388,686-690. Wmg LY, Orser BAYBrautigan DL, MacDonald JF (1994) Regulation of NMDA receptors in cultured hippocampal neurons by protein phosphatases 1 and 2A. Nature 369,230-2 Wang,L.Y.; Salter,M.W.; MacDonald,J.F.(1991). Regulation of kainate receptors by CAMP- dependent protein baseand phosphatase. Science 253, 1132-1 135 Wang,Y., Small,D.L., Stanimirovic,D.B., Morley,P., and DwkinJ.P. (1997). AMPA receptor- mediated regulation of a Gi-protein in cortical neurons. Nature 389, 502-504, Wmg YT and Linden DJ (2000). Expression of cerebellar long-term depression requires postsynaptic clathrin-mediated endocytosis. Neuron 25,635-647 Wang,Y.T. and Salter,M.W. (1 994). Regulation ofNMDA receptors by protein-tyrosine kinases and phosphatases. Nature 369,233-235. Washbum,M.S., Numberger,M., Zhang,S., and Dingledine,R. (1997). Differential dependence on GluR2 expression of three characteristic features of AMPA receptors. J.Neurosci. 17,9393-9406. Wechsler,A. and Tei~hberg~V.1.(1998). Brain spectrin binding to the NMDA receptor is regulated by phosphorylation, calcium and calmoduIin. EMl30 J. 17,3 93 1-3 93 9. Wenthold,R. J., Petralia,R. S., BIahos,J., II, and Niedzielski,A. S .(1 996). Evidence for multiple AMPA receptor complexes in hippocampal CA UCA2 neurons. Journal of Neuroscience Id, 1982-1989. We~tpha1,R.S.~Tavalin,S.J., Lin,J.W., Alto,N.M., Fraser,I.D., Langeberg,L.K., Sheng,M., and Scott,J.D. (1999). Regulation of NMDA receptors by an associated phosphatase-kinase signaling complex. Science 285, 93-96. Wickelgren,I. (1998). Tracking insulin to the mind. Science 280,5 17-5 19. Wozniak,M., Rydzewski,B ., Baker,S .P., and Raizada,M.K. (1993). The cellular and physiological actions of insulin in the central nervous system. Neurochernistry International 22, 1-10. Wurmser,A.E., Gary,J.D., and Ernr,S.D. (1 999). Phosphoinositide 3-kInases and their FYVE domain- containing effectors as regulators of vacuolar/lysosornal membrane trafficking pathways. J.Biol.Chem. 2 74,9 129-9 132. Wyszyaski,M., Lin,J., Rao,A., Nigh,E., Beggs,A.H., Craig,A.M., and Sheng,M- (1997). Competitive binding of alpha-actinin and calmoddin to the NMDA receptor. Nature 385,439-442. Xia,J., Zhang,X., Staudinger,J., and Huganir,R.L. (1999). Clustering of AMPA receptors by the synaptic PDZ domain-containing protein PICK1 .Neuron 22, 179-1 87. Yang,J., Clarke,J.F., Ester,C.J., Young,P.W., Kasuga,M., and HoIman,G.D. (1996)- PhosptzatidyIinosito13-kinase acts at an intracellular membrane site to enhance GLUT4 exocytosis in 3T3-Ll cells. Bi0chem.J. 313 (Pt I), 125-13 1. Ye,B., Liao,D., Zhang,X., Zhang,P., Dong,H., and Huganir,R.L. (2000). GRASP-1 : a neuronal RasGEF associated with the AMPA receptor/GRIP complex. Neuron 26,603 -6 17. Yu,J., Wjasow,C., and Backer,J.M. (1998b). Regulation of the p8Yp 11 Oalpha phosphatidylinositol 3'-kinase. Distinct roles for the n-tenninal and c-terminal SH2 domains. J.Biol.Chem. 2 73, 30 199- 30203. Yu,J., Zhang,Y., McIlroy,J., Rordorf-Nikolic,T., Orr,G.A-, and Backer, J.M. (1998a). Regulation of the p8S/p 1 10 phosphatidylinositol3'-kinase: stabilization and inhibition of the p 11 Oalpha catalytic subunit by the p85 regulatory subunit. Mol-Cell Biol. 18, 1379-1387. Yu,X.-M., Askalan,R., Keil II,G.J., and Salter,M.W. (! 997). NMDA channel regulation by channel- associated protein tyrosine kinase Src. Science 275,674-678. Yu,X.M. and Salter,M.W. (1998). Gain control ofNMDA-receptor currents by intracellular sodium. Nature 396,469474, Yu,X.M. and Salter,M.W. (1999). Src, a molecular switch governing gain control of synaptic transmission mediated by N-methyl-D-aspartate receptors. Proc.Natl.Acad.Sci.U.S.A 96,7697-7704. Zhang,F.X., Rubin,R., and Rooney,T.A- (1 998b). N-Methyl-D-aspartate Inhibits Apoptosis through Activation of Phosphatidylinositol 3 -Kinase in Cerebellar Granule Neurons. A role for insulin receptor substrate4 in the neurotrophic action of n-methyl-d- aspartate and its inhibition by ethanol- J-BiolChem, 2 73,26596-26602. Zhang,S., EhIers,M.D., BernhardtJP., Su,C.T,, and Huganir,R.L. (1998a). CaImodulin mediates calcium-dependent inactivation of N-methyl-D- aspartate receptors. Neuron 21,443-453.