University of Southern California Dissertations and Theses

An orchestra of glutamate receptors and transporters in synaptic transmission, plasticity and excitotoxicity

(USC Thesis Other)

An orchestra of glutamate receptors and transporters in synaptic transmission, plasticity and excitotoxicity

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

AN ORCHESTRA OF GLUTAMATE RECEPTORS AND TRANSPORTERS IN
SYNAPTIC TRANSMISSION, PLASTICITY AND EXCITOTOXICITY

by

Wei Xu

______________________________________________________

A Dissertation Presented to the
FACUTY OF THE GRADUATE SCHOOL
UNIVERISTY OF SOUTHERN CALIFORNIA
In Partial Fulfillment of the
Rquirements for the Degree
DOCTOR OF PHILOSOPHY
(NEUROSCIENCE)

May 2008

Copyright 2008 Wei Xu
Acknowledgements

My late grandfather took care of me since I was a baby and supported my education
till my days in college. Without his insistence, I might not have a chance to come out
of the small farm where I was born. My young aunt was my first teacher and her
expectations and encouragement might be the earliest motivation I had to pursue
science. My parents, brother, sister, elder aunt and other relatives also gave me
tremendous support in these years. My wife, Ying Li, since our first meeting in 2000,
has encouraged me to pursue my true passion and helped me in basically everything
ever since. I am so grateful that I have such a supportive family.

My advisor, Michel, gave me valuable directions in scientific research. I am always
indebted to him for his warmhearted support. I am also very thankful to Dr. Long-
Chuan Yu at Peking University and Dr. Yu Tian Wang at UBC. I feel fortunate to
have worked with these nice and considerate mentors. I also enjoyed working with
wonderful colleagues since I first started research. They made the labs a friendly and
cooperative environment. Among them Tan Pan Wong, Miou Zhou and some others
also contributed data for the work presented here.

I am especially grateful to my committee, Dr. Ko, Dr. Thompson, Dr. Johnson and
Dr. Chow for their support and guidance.
ii
Table of Contents

Acknowledgements......................................................................................................ii
Table of Contents ........................................................................................................iii
List of Figures ..............................................................................................................v
Abstract ......................................................................................................................vii
Introduction.................................................................................................................ix
Chapter 1. Interaction between VGLUT1 and EAATs................................................1
1.1. Summary ...........................................................................................................1
1.2. Background and Introduction............................................................................2
1.2.1. The Organization of Glutamatergic Synapses............................................2
1.2.2. Glutamate Transporters..............................................................................5
1.2.3. Glutamate Transporters in Synaptic Transmission and Brain Diseases.....9
1.3. Methods and Materials....................................................................................11
1.4. Results.............................................................................................................12
1.4.1. Physical Interactions between VGLUT1 and EAATS.............................12
1.4.2. The Localizations of Glutamate Transporters..........................................15
1.4.3. The Model of VGLUT1-EAATs Interaction ...........................................16
1.4.4. Identification of VGLUT1-EAAT Interaction Site..................................19
1.4.5. VGLUT1-EAATs Interaction Occurs in Neurons and Glial Cells ..........22
1.4.6. The Functional Implications of VGLUT1-EAAT Interaction .................23
1.5. Discussion .......................................................................................................25
1.5.1. The Controversy Regarding Presynaptic EAATs ....................................25
1.5.2. VGLUT1-EAATs Interaction and Synaptic Function .............................28
1.5.3. VGLUT1-EAAT Mediates a New Model of Glutamate Uptake? ...........30
Chapter 2. Homeostasis in Synaptic Plasticity...........................................................33
2.1. Summary .........................................................................................................33
2.2. Background and Introduction..........................................................................34
2.2.1. Synaptic Plasticity and Memory ..............................................................34
2.2.2. Expression Mechanisms of Synaptic Plasticity .......................................36
2.3. Methods and Materials....................................................................................38
iii
2.4. Results.............................................................................................................42
2.4.1. Measurement of Spontaneous Synaptic Release......................................42
2.4.2. Measurement of Evoked Release.............................................................47
2.4.3. DHPG Facilitates Glutamate Release ......................................................50
2.4.4. DHPG Induces LTD.................................................................................53
2.5. Discussion .......................................................................................................59
2.5.1. Assessing Presynaptic Activity................................................................59
2.5.2. Expression Mechanism of DHPG-induced LTD .....................................62
2.5.3. Pre- and Postsynaptic Homeostasis..........................................................64
Chapter 3. mGluR1 α and Calpain in Excitotoxicity ..................................................66
3.1. Summary .........................................................................................................66
3.2. Background and Introduction..........................................................................67
3.2.1. mGluR1 Signaling in Excitotoxicity........................................................67
3.2.2. Calpain and Its Substrates ........................................................................70
3.3. Methods and Materials....................................................................................72
3.4. Results.............................................................................................................77
3.4.1. NMDA Receptor Activation Induces Truncation of mGluR1 α...............77
3.4.2. mGluR1 α Is Truncated by Calpain at Ser
936
............................................81
3.4.3. C-terminal Truncated mGluR1 α Remains Functional .............................87
3.4.4. C-terminal Truncation Alters mGluR1 α Signaling..................................89
3.4.5. Truncation Alters mGluR1 α Targeting....................................................92
3.4.6. Full-length and Truncated mGluR1 α in Excitotoxicity ...........................94
3.4.7. Truncation of mGluR1 α in vivo...............................................................99
3.5. Discussion .....................................................................................................103
3.5.1. Regulation of mGluR1 α by Calpain-mediated Truncation....................103
3.5.2. NMDA Receptor/mGluR1 Interactions in Excitotoxicity .....................106
Summary and Conclusions.......................................................................................110
References................................................................................................................115

iv
List of Figures

Figure 1. Physical Interactions between VGLUT1 and EAATS ...............................14
Figure 2. The Localizations of Glutamate Transporters ............................................16
Figure 3. The Model of VGLUT1-EAATs Interaction..............................................19
Figure 4. Identification of Interaction Site on VGLUT1 ...........................................21
Figure 5. VGLUT1-EAATs Iinteraction Occurs in Neurons and Glial Cells............22
Figure 6. The Functional Implications of VGLUT1-EAAT Interaction....................25
Figure 7. VGLUT1/VGAT Specifies Excitatory/Inhibitory Synapses ......................45
Figure 8. Measurement of Spontaneous Release .......................................................46
Figure 9. Measurement of Evoked Release................................................................50
Figure 10. DHPG Iinduces Long-Term Facilitation of Glutamate Release...............52
Figure 11. DHPG Induces LTD through AMPA Receptor Internalization ...............54
Figure 12. GluR2
3y
Peptide Reduces DHPG-induced LTD in Hippocampal Slices..55
Figure 13. The Presynaptic Facilitation Is Downstream of AMPAR Internalization 58
Figure 14. Glutamate Induces Carboxyl-terminal Truncation of mGluR1α ............79
Figure 15. NMDA Receptor Induced Truncation of mGluR1α................................80
Figure 16. Dosage-effect of NMDA-induced Truncation of mGluR1α ....................81
Figure 17. mGluR1α Is Truncated by Calpain ..........................................................83
Figure 18. mGluR1α Is Truncated at Ser
936
...............................................................85
Figure 19. TAT-mGluR1 Peptide Protected mGluR1 α from Truncation ..................87
Figure 20. Truncated mGluR1α Remains Functional................................................88
v
Figure 21. C-terminal Truncation Alters mGluR1α Signaling..................................90
Figure 22. Altered mGluR1-Akt Signaling by Calpain-mediated Truncation...........91
Figure 23. NMDA Treatment Alters mGluR1α Targeting........................................93
Figure 24. C-terminal truncation alters mGluR1α targeting......................................94
Figure 25. Distinct Roles of Wild-type and Truncated mGluR1α in Excitotoxicity.97
Figure 26. TAT-mGluR1 Peptide Protected Neurons from NMDA Toxicity ...........98
Figure 27. Truncation of mGluR1α in KA-induced Epilepsy .................................100
Figure 28. Truncation of mGluR1α in Stroke..........................................................101
Figure 29. Neuroprotection of TAT-mGluR1 in vivo..............................................103
Figure 30. NMDA Receptor, Calpain and mGluR1 in Excitotoxicity.....................109

vi
Abstract

Glutamate is the major excitatory neurotransmitter in the mammalian central nervous
system. Synaptic transmission at glutamatergic synapses mediates and regulates
basically all aspects of brain functions. The strength of these synapses is subjected to
potentiation and depression and these plastic modifications are plausible candidates
for information storage in the brain. Excessive activity at glutamatergic synapses,
namely excitotoxicity, occurs in many brain diseases and is a critical factor for
neuronal death or degeneration. Therefore, to elucidate the molecular mechanisms
underlying glutamatergic transmission, plasticity and excitotoxicity constitutes the
first step to understand neuronal information processing, learning, memory and brain
diseases. In my dissertation, I report the results of three independent but closely
related studies on synaptic transmission, plasticity and excitotoxicity, respectively. In
the first part, I provide evidence that two different glutamate transporters, one
involved in storing glutamate into synaptic vesicles for synaptic release and the other
one responsible for re-uptake of released glutamate inside cells, physically interact
with each other and may cooperate for regulating synaptic transmission. In the
second study, I first developed a technique to simultaneously measure presynaptic
release of glutamate (by targeting vesicular glutamate transporters) and postsynaptic
expression of glutamate receptors. Using this technique, I was able to show the
existence of a homeostatic relationship between presynaptic and postsynaptic
activities. In the third study, I studied interactions between NMDA type ionotropic
vii
glutamate receptors and the Group 1 metabotropic glutamate receptors through
calpain-mediated proteolytic processing. I showed that this interaction is a critical
step for acute neuronal degeneration in excitotxicity and I therefore developed a
neuroprotective method by selectively blocking this process. Together, these studies
indicate that coordinated functional interactions between glutamate receptors,
transporters and different synaptic structures is critical for normal brain functions,
that alterations in these processes might participate in brain diseases and that
selective targeting of critical elements may provide therapeutic approaches for brain
diseases.

viii
Introduction

A major purpose of molecular and cellular neuroscience is to provide enough details
to reconstruct a complete model of synapses, neurons and eventually of the whole
brain, and at the same time, to provide tools for manipulating the nervous system and
for intervening into the processes of brain diseases. Despite significant progress, we
are still far away from this goal and need to reveal more details of the working
mechanisms of neurons and synapses. Through a few decades of extensive study,
most of the major proteins constituting neurons and synapses have been identified
and characterized; however, their interactions and functional regulations in normal
brain function and in brain diseases need to be further explored. In the current study,
we focused on the functional interactions between transporters, receptors and
different synaptic structures at glutamatergic synapses, which represent the majority
of synapses in the mammalian brain.

The release and recycling of the neurotransmitter glutamate is the first step for these
synapses to transmit information between neurons. The proteins that transport
glutamate into synaptic vesicles (vesicular glutamate transporters, VGLUTs), and the
proteins that remove released glutamate by its reuptake into cells (EAAT) have been
identified and characterized. Yet several questions remain. First, it is not clear how
the release of glutamate is coordinated with its uptake to maintain synaptic
transmission at a proper level. Second, the relative contributions of neuronal
ix
terminals vs. glial processes in the uptake of glutamate remain a controversial issue.
In addition, it is not known whether glutamate can be directly taken up into synaptic
vesicles from extracellular space, which would be a more efficient way to maintain
synaptic transmission. In Chapter 1, we studied the interactions between VGLUT1
and EAATs in order to develop insights for further exploring these questions.

The strength of synaptic transmission at glutamatergic synapses is subject to plastic
regulations, which, at least in part, contribute to the adaptive modifications of animal
behaviors. The aims in studying synaptic plasticity include revealing the mechanisms
underlying its induction, expression and maintenance. While extensive knowledge
regarding the induction part has been obtained, fewer questions have been solved for
expression and maintenance. One long-lasting controversy concerns the locus (pre-
synaptic vs. post-synaptic) of the expression of synaptic plasticity. In recent years,
accumulating evidence has led to the generally accepted view that alterations in the
number of post-synaptic glutamate receptors represent the primary mechanism for
the expression of synaptic plasticity; yet, the roles of pre-synaptic sites in plasticity
remain unclear. In Chapter 2, we developed a technique to measure pre-synaptic
activity by targeting the synaptic specific vesicular transporters and studied
alterations in pre-synaptic functions in a model of synaptic plasticity.

Dysregulation, especially over-activation of glutamatergic synapses leads to the
degeneration and death of neurons, which is a critical part of the pathogenesis of
x
xi
many brain diseases. It is well known that the toxic effects are mediated by
glutamate receptors, which are a large family of proteins comprised of ion channels
(ionotropic glutamate receptors) and G-protein coupled receptors (mGluRs). It is
well accepted that the NMDA-type ionotropic glutamate receptors is one of the
major mediator of glutamate toxicity, but the roles of the mGluRs are less clear. In
Chapter 3, we studied the interaction between NMDA receptors and mGluR1 in
excitotoxic neuronal degeneration.

Chapter 1. Interaction between VGLUT1 and EAATs
1.1. Summary
Glutamatergic synaptic transmission is initiated by active transport of glutamate into
synaptic vesicles mediated by vesicular glutamate transporters (VGLUTs) and
release of vesicular glutamate into synaptic cleft by exocytosis; and it is terminated
by removal of extracellular glutamate by membrane excitatory amino acid
transporters (EAATs) located in glial cells and neurons. VGLUT1 and VGLUT2 are
complementarily expressed in the brain and are responsible for glutamate packaging
in most glutamatergic terminals. They are predominantly expressed in neurons but
are also present in glial cells to a lesser degree. EAAT1 and EAAT2 are expressed
predominantly in glial cells and may also present in neuronal terminals to a much
lesser degree. By using co-immunoprecipitation, we found that VGLUT1, but not
VGLUT2 physically interacts with EAAT1 and EAAT2 in the brain. The
interactions take place between vesicles and cell membranes in individual neuronal
terminals or glial processes. When expressed in HEK cells, VGLUT1 did not
significantly alter the uptake activity of EAATs, but EAATs changed the trafficking
properties of VGLUT1-containg vesicular structures. Based on these observations,
we proposed that VGLUT1-EAATs interaction may regulate the vesicular release in
synaptic terminals. In addition, VGLUT1-EAATs may form a protein complex
mediating direct uptake of glutamate from extracellular space into synaptic vesicles.
1
1.2. Background and Introduction
1.2.1. The Organization of Glutamatergic Synapses
Glutamatergic synapses: Glutamate, an amino acid, is not only one of the building
blocks of proteins, but also one of the key molecules in metabolism as an
intermediate in the TCA cycle. It had been known as a neurotoxin for a long time.
But it was not until the 1960s-70s when its major function as a neuronal transmitter
was discovered. Follow-up studies revealed that it is not just a neurotransmitter, but
the major excitatory neurotransmitter in the mammalian CNS. Moreover, it is also
the precursor to the major inhibitory neurotransmitter, GABA, in the brain.
Furthermore, glutamate may also be involved in some intracellular signaling
mechanisms. Anatomically, over 75% of the total synapses in the brain are
glutamatergic. In certain brain regions, such as hippocampus and cortex, this
percentage is even higher. Functionally, brain metabolism is coupled with the
activity of these synapses (Shulman RG et al., 2004).

Pre-synaptic site: Under electron microscopy, a typical glutamatergic synapse in the
brain is asymmetric in terms of the different electron densities across the synaptic
cleft. The post-synaptic site contains a high density region (hence named
postsynaptic density, PSD). The presynaptic site is low in electron density and
contains ~50-100 synaptic vesicles. The synaptic vesicles have a diameter of about
40 nanometers. A recent proteonomic study aiming to generate a comprehensive and
2
quantitative molecular description of synaptic vesicles revealed that the vesicles
contain a diversity of proteins, such as the membrane trafficking proteins, transmitter
transporters, etc (Takamori S et al., 2006). Most of these proteins have numerous
copies. One exception is the V-ATPase, with only one or two copies on each vesicle.
Inside the pre-synaptic membrane the cytoskeletal proteins are organized into active
zones which contain the machinery for the release and recycle of synaptic vesicles.
On the presynaptic membrane there are a few kinds of receptors, including the group
2 and group 3 mGluRs. These receptors mainly work as auto-inhibitory receptors.
There are also certain receptors like CB1 receptor for endocannabinoid, which is
suggested to be a retrograde messenger from post-synaptic site.

Post-synaptic site and glutamate receptors: The post-synaptic density can be
formed on dentritic shafts or spines (the small protrusions on neuronal dendrites).
The spines are relatively independent signaling compartments since the diffusion of
signaling molecules are restricted by the narrow neck of these spines. Over 90% of
the glutamatergic synapses in typical pyramidal cells in the hippocampus or cortex
are formed on the top of spines. The glutamatergic synapses formed between
glutamatergic neurons and GABAergic interneurons are largely located on dentritic
shafts. The postsynaptic site of glutamatergic synapses houses a large family of
receptors for glutamate, including the glutamate-gated ion channels (iontropic
glutamate receptors, can be further divided into NMDA-, AMPA- and KA- types of
glutamate receptors according to their pharmacological properties) and G-protein
3
coupled metabotropic glutamate receptors (mGluRs, which consist of 8 subtypes and
are divided into three groups according to the distinct signaling mechanisms, sub-
cellular distribution and pharmacological properties). Glutamate receptors, especially
the AMPA type ionotropic glutamate receptors, are concentrated in the PSD. A few
studies have been done to quantify the number of receptors on the PSD, utilizing
electron microscopy, fluorescence measurements or proteonomic methods. These
studies indicate that there are around 15-20 copies of AMPA and/or NMDA type of
glutamate receptors in each synapse (Sheng M and Hoogenraad CC, 2007). The
groups 1 mGluRs, including mGluR1 and mGlur5, are mainly located at the post-
synaptic sites and are concentrated at the edge of the PSD.

Peri-synaptic glial processes: The glutamatergic synapses are enveloped by the
processes arising from glial cells. In the brain glial cells provide mechanical and
nutrient support to the neurons and synapses and also seem to directly provide
neurons with lactate as the major energy supply (though controversial, see Bonvento
G et al., 2005). In addition glia appears directly involved in the processes of synaptic
transmission. Some recent studies indicate that the glial cells can form synapse-like
structure with neurons and that they also contain the machinery for release D-serine
and glutamate as signaling molecules (Paukert M, Bergles DE, 2006). For
glutamatergic synapses, the glial cells play an additional critical role - uptaking
glutamate from synaptic cleft and recycling it back to neurons. The protein-protein
interaction between glial cells and neurons has also been shown to be important for
4
the formation and function of synapses (eg. EphA4 - ephrinA3 interaction, Murai
KK et al., 2003).

Synaptic cleft: The ~20 nm distance between the pre- and post-synaptic site is the
synaptic cleft. EM observations showed that the synaptic cleft is composed by some
electronically dense materials (Rostaing P et al., 2006). In fact, numerous proteins
have been localized in the synaptic cleft, among which are trans-synaptic coupling
proteins like the neurexin (presynaptic)-neuroligin (post-synaptic), EphB and
EphrinB, NCAM, intergrin, etc. These proteins and their interactions may be
involved in the functional coupling between the pre- and post- synaptic sites.

1.2.2. Glutamate Transporters
Glutamate/glutamine cycle: Glutamate is a molecule with multiple functions. In
addition to mediating synaptic transmission, it is one of the 20 amino acids used for
making proteins, an intermediate of energy metabolism and a possible intracellular
signaling molecule. These functions require different optimal concentrations of
glutamate. To this end, the brain has developed a sophisticated machinery to
maintain glutamate at different concentrations within different compartments. The
extracellular glutamate must be kept very low (less then a few µM) to avoid the
excitotoxic effects. The cytosolic glutamate is estimated to be at 1-10 mM, which is
around a hundred times higher than that outside the cells (Carlson MD et al., 1989).
The concentration inside the synaptic vesicles is in the range of 100 mM to 1 M
5
(Maycox PR et al., 1988, Carlson MD et al., 1989). When synaptic vesicles are fused
with presynaptic membrane the release of vesicular glutamate can transiently
increase the concentration of glutamate in the synaptic cleft to mM level. Soon the
released glutamate gets sequestered into the cells by glutamate transporters (see
review by Danbolt NC, 2001).

The classical view of the glutamate cycle during synaptic activity includes the
following steps: 1) glutamate is packed into synaptic vesicles by vesicular glutamate
transporters; 2) vesicular glutamate gets released into the synaptic cleft; 3) glutamate
is then uptaken by high affinity glutamate transporters on glial cells (the majority)
and neurons (the minority); 4) glutamate is transformed into glutamine in glial cells
and returned to neurons through the transporters of glutamine; 5) In neurons
glutamine is converted to glutamate. All these steps consume energy and some
researchers even estimated that the energy spent on this cycle accounts for 70-80%
of the total energy consumed by the brain (Shulman RG et al., 2004).

Vesicular glutamate transporters: The existence of vesicular glutamate
transporters (VGLUTs) and their functional properties had been extensively studied
with biochemical and pharmacological approaches long before they were cloned.
These studies indicate that these transporters have low affinity for glutamate (Km = 2
mM, Carlson MD et al., 1989) and are dependent on the electrical gradient across the
membrane of synaptic vesicles for transporting activity. VGLUT1 was cloned in
6
1994 and identified as a glutamate transporter in 2000. Then VGLUT2 and 3 were
soon identified for their high homology with VGLUT1. Scanning of mammalian
genome showed that the VGLUT1-3 may be all the subtypes of VGLUTs. Among
them VGLUT1 and VGLUT2 are complementarily expressed in the brain and
account for the majority of excitatory transmission in the brain, while VGLUT3 is
sparsely expressed and may have functions other than mediating excitatory
transmission (Gras C et al., 2002). VGLUT1 is more concentrated in the forebrain
structures such as hippocampus and cortex, while VGLUT2 is more concentrated in
the thalamus and other lower brain structures (Fremeau RT Jr et al., 2001). Some
VGLUT1-expressing neurons seem to express also VGLUT2 in the early stages in
the development. The implications of this developmental switch remains unclear. It
is still controversial if the two transporters can co-exist in the same vesicle (Fremeau
RT Jr et al., 2004A). But it looks like that even if they could co-localize in the same
vesicles it would be a rare event. No reports have shown any difference in the
transporting activity of these two transporters.

Other than packaging glutamate into synaptic vesicles VGLUT1 may play other roles
as well. Actually these transporters were initially cloned as transporters for inorganic
phosphate. It is not clear if this function is related to synaptic transmission. Some
researchers hypothesized that the uptake of phosphate into synaptic terminals might
help to activate the enzymes for converting glutamine to glutamate but no
experimental evidence has been published to support this idea (Bellocchio EE et al.,
7
1998). Some recent reports also showed that VGLUT1 may directly interact with
endophilin to regulate the trafficking of synaptic vesicles (Voglmaier SM et al.,
2006).

Membrane glutamate transporters: Distinct from VGLUTs, glutamate
transporters are located on the cell membrane in charge of removing glutamate from
extracellular space and into cells. Five membrane transporters have been cloned. In
humans they are named excitatory amino acid transporter 1-5 (EAAT1-5). In mice or
rats EAAT1 is also known as glutamate–aspartate transporters

(GLAST), EAAT2 as
glutamate transporter subtype 1 (GLT-1) and EAAT3 as excitatory amino-acid
carrier 1 (EAAc1) (in the current study we refer to them as EAAT). These
transporters have a very high affinity for glutamate (at µM level) and this property is
critical to ensure that the extracellualr glutamate concentration may be kept at a very
low level. These transporters are secondary-active transporters which do not directly
use ATP but are dependent on the sodium gradient across the cell membrane for their
uptake activity. The stoichiometry of EAATs is movement of 1 glutamate, 3 Na+ and
1 H+ into the cell coupled with 1 K+ out of the cell. Therefore, these transporters are
electrogenic and generate a net inward current. The EAATs have distinct distribution
patterns. EAAT2 is enriched in the forebrain while EAAT1 is enriched in the
cerebellum. EAAT1 and EAAT2 are predominantly expressed in glial cells. But
increasing evidence suggests that they are also expressed in synaptic terminals.
EAAT3 is predominantly expressed in the post-synaptic site. Studies implicate that
8
EAAT3 may not be one of the major contributors of glutamate uptake activity
compared to EAAT1 and EAAT2; instead, it might work as a transporter for cysteine
(Aoyama K et al., 2006). EAAT4 is also a neuronal transporter, but its distribution is
limited to Purkinje cells in the cerebellum of adult brain. So far, EAAT5 is detected
only in the retina (see review by Danbolt NC, 2001).
1.2.3. Glutamate Transporters in Synaptic Transmission and Brain Diseases
Recent quantitative studies indicate that there are about 10-12 copies of vesicular
glutamate transporters (VGLUT1 or VGLUT2) on each vesicle (Takamori S et al.,
2006). The copy number of VGLUTs may be related to the quantal size since over-
expression of VGLUT1 in cultured neurons can increase the amplitude of miniature
EPSC (Wilson NR et al., 2005). A similar observation has been made on VGLUT2
(Moechars D et al., 2006). Loss of VGLUTs may contribute to the development of
schizophrenia (Reynolds GP and Harte MK, 2007). VGLUTs may also be involved
in neuropathic pain and other brain diseases (Moechars D et al., 2006).

Deletion of EAAT1 can elongate the duration of EPSC, indicating that the activity of
glutamate transporters is an active part of fast synaptic transmission (Takayasu Y et
al., 2005). When the glutamate transporters were pharmacologically inhibited, the
mGluR-mediated current was found to be enhanced. At the same time, the threshold
for induction of synaptic plasticity was altered (Brasnjo G and Otis TS, 2001). This
study indicated that the regulation of transporter activity can directly regulate
synaptic plasticity. Beside directly affecting synaptic transmission by altering
9
extracellular glutamate concentration, the electronic effect of the activity of the
transporters can also affect the electric property of the presynaptic terminals and
affect the release of transmitters (Veruki ML et al., 2006). The activity of EAATs is
closely coupled with the energy metabolism in the brain. The sodium influx through
these transporters seems to serve as the functional link for the metabolic crosstalk
between neurons and astrocytes (Voutsinos-Porche B et al., 2003). It is reported that
the sensory-evoked intrinsic optical signals in the olfactory bulb was reduced by
inhibition of glutamate transporters (Gurden H et al., 2006).

The dysfunction of EAATs has been found to be involved in numerous diseases of
the central nerves system. Deletion of EAATs increased the extracelluar
concentration of glutamate, induced epilepsy, excitoxic neuronal degeneration and
progressive paralysis (Rothstein JD et al., 1996). The reverse of the transporting
direction of GLT-1 is reported to be the major source of glutamate release in the
early phase of stroke which is critical for following neuronal death (Rossi DJ et al.,
2000). The involvement of EAATs in the development of amyotrophic lateral
sclerosis (ALS) has been extensively studied. The protein levels of EAATs,
especially, of EAAT2, in the spinal cords of ALS patient or animal models were
down-regulated. The mutated SODs found in familiar ALS, were found to trigger the
internalization and degradation of glutamate transporters (Vanoni C et al., 2004). In
addition, increasing the expression level of GLT-1 with Beta-lactam antibiotics was
found to be helpful for ALS mice (Rothstein JD et al., 2005).
10
1.3. Methods and Materials
Immunoprecipitation: Lysates of brain tissues or cultured cells were collected with
modified RIPA buffer. The lysates were cleared by centrifugation at 100, 000 g for 1
h and pre-cleared by incubating with Protein A beads for 30 min. The pre-cleared
lysate containing ~300 µg total protein was incubated with ~5 µg primary antibodies
for 1 h first and then together with protein A agarose beads for overnight. After three
washes with RIPA buffer the next day, the antibody-bound proteins were eluted from
the beads with 50 µl loading buffer and subjected to SDS-PAGE and immunoblots.

Immunohistochemistry and immunocytochemistry: Cultured cells were fixed
with 4% formaldehyde in PBS (RT, 10 min) and were further permeabilized with 1%
Triton X-100 for 5 min. After blocking with 5% goat serum (60 min), cells were
incubated in primary antibodies dissolved in 0.5% goat serum in PBS for 1 h. After
several washes, cells were incubated with corresponding fluorescence-tagged
secondary antibody for 30 min. After three washes, neurons were then mounted for
observation. For immunohistochemistry, adult rats were deeply anesthetized with
Ketamine and Xylazine and intracardially perfused with 20 ml chilled PBS (pH 7.4)
followed by 100 ml 4% formaldehyde. Brains were removed and post-fixed for 2 h at
room temperature. After being incubated in PBS with 30% sucrose until sinking,
brains were cut into 30 µm thick sections with a cryostat. Brain sections were
incubated in primary antibodies for overnight at 4 °C followed by incubation with
11
fluorescence-tagged secondary antibody for 30 min. Sections were then mounted to
slides with antifade mounting medium.

Brain tissue fractionation: Synaptosomes and Glial Plasmalemmal Vesicles (GPV)
were purified from rat forebrain according to a protocol described in Nakamura T et
al., 1993.
1.4. Results
1.4.1. Physical Interactions between VGLUT1 and EAATS
To study the function of VGLUT1 in synaptic transmission we first checked if it can
interact with any other synaptic proteins. We conducted immunoprecipitation on
cortical tissue with an antibody against VGLUT1 (Synaptic Systems, Cat# 135 302)
and the proteins co-immunoprecipitated (co-IP) with VGLUT1 was then analysed by
mass spectrometry (MS). To our surprise, MS analysis showed two masses matching
the molecular weights of EAAT2, a membrane glutamate transporter. To verify this
finding, we repeated the immunoprecipitation and followed with immunoblots. As
shown in Fig 1, the antibodies against VGLUT1 (Cat# 135 302), but not that of
VGLUT2 (Cat# 135 402), pulled down both EAAT1 and EAAT2. We also probed
EAAT3 (EAAc1, the neuronal type of membrane glutamate transporter), though the
antibody did not generate convincing signal for immunoblot or co-IP experiment. To
further confirm this result, we tested another antibody of VGLUT1 (Synaptic
Systems, monoclonal, Cat#135 311), and found that this antibody could also pull
12
down EAAT1 or 2. When this antibody was pre-incubated with its antigen peptide
(Synaptic Systems, 135-3P, composed of the residues 456 to 560 of VGLUT1) for
one hour and then used for the co-immunoprecipitation experiment, it lost its ability
to pull down EAATs, confirming the specificity of this antibody. The co-IP
experiment was also performed in a reverse way and we found that the antibodies
against EAAT2 were able to pull down VGLUT1, but not VGLUT2. Since VGLUT1
is enriched in the hippocampus while VGLUT2 in thalamus, we conducted the co-IP
with both tissues. The results indicate that only VGLUT1 but not VGLUT2 can
interact with EAATs in both tissues.

To test if this VGLUT1-EAATs interaction is a specific binding, we also tested other
antibodies, including the vesicle associated protein synaptotagmin1, the cytoskeletal
protein spectrin, the glutamate receptors GluR2/3 and mGluR1a. Among these
antibodies, only that against EAAT2 pulled down VGLUT1, indicating that the
VGLUT1-EAATs interaction is specific.

We went on to test if the VGLUT1-EAATs association was formed before or after
the brain tissue was lysed. We added purified His-tagged EAAT2 protein (a generous
gift from Max Plank) into brain lysate and inferred that if VGLUT1 protein could
interact with EAATs in the brain lysate the His-tagged EAAT2 would also co-IP
with VGLUT1. Interestingly, we found that although VGLUT can co-IP with the
endogenous EAAT2, it did not pull down his-tagged EAAT2. The data suggest that
13
the VGLUT1-EAAT association is formed in the brain before the brain tissue was
lysed. Together these results indicate that VGLUT1 but not VGLUT2 can
specifically physically interact with the membrane glutamate transporters EAT1 and
EAAT2.

Figure 1. Physical Interactions between VGLUT1 and EAATS
(A) Lysates of cortical immunoprecipitated with antibodies against VGLUT1 or VGLUT2
respectively and immunoblotted with EAAT1 or EAAT2 respectively. VGLUT1, but not VGLUT2,
was co-immunoprecipitated with both EAAT1 and EAAT2. (B) Lysates of hippocampus (enriched
with VGLUT1) or thalamus (enriched with VGLUT2) were immunoprecipitated with antibodies
against VGLUT1 or VGLUT2 respectively and immunoblotted with EAAT1. (C) The VGLUT1
antibody was pre-incubated with its antigen peptide for 1 h before it was used for
immunoprecipitation (VGLUT1-ab). This pre-absorption blocked VGLUT1-EAAT2 co-
immunoprecipitation. (D) EAAT2, but not mGluR1, synaptotagmin1, spectrin or GluR2 co-
immunoprecipitated VGLUT1 from lysate of cortical tissue.
14
1.4.2. The Localizations of Glutamate Transporters
To understand how VGLUT1 interacts with EAATs, we first conducted
immunohistochemistry to examine the localization of these transporters in the brain.
As shown in Fig 2, in the three brain regions checked, including hippocampus,
cortex and cerebellum, VGLUT1 displayed a punctuate distribution, presumably
synaptic terminals. The immunoreactivity of EAATs was concentrated in tructures
surrounding the terminals and the processes of glial cells. The immunoreactivity of
VGLUT1 and EAATs was largely complementary, but some overlap was observed
in a subset of neuronal terminals. The overlap is apparent in the mossy fibers of
cerebellum. Consistent with previous studies, this result indicates that VGLUT1 is
predominantly distributed in the presynaptic terminals while EAATs are
predominantly distributed in glial cells, but they can also co-localize in certain
neuronal terminals.

Based on this distribution pattern, we proposed three models for the VGLUT1-
EAATs interaction. In the first model, VGLUT1 and EAATs interact with each other
between neurons and glial cells. When synaptic vesicles fuse with presynaptic
membrane during synaptic release VGLUT1 is transiently expressed on the
presynaptic membrane. In this situation, VGLUT1 may interact with EAATs which
are distributed on the adjacent glial membrane. In the second model, VGLUT1
interact with EAATs between synaptic vesicles and presynaptic membranes since a
small fraction of EAATs is located in the presynaptic terminals. In the third model,
15
VGLUT1 move to cell membrane during synaptic release and interact with EAATs
directly on cell membrane.

Figure 2. The Localizations of Glutamate Transporters
(A) Brain sections were immunostained with VGLUT1 and EAATs. VGLUT1 was predominantly
localized in synaptic terminals while EAATs were predominantly in the processes of glial cells. Co-
localization of VGLUT1 and EAATs was observed in a subset of nerve terminals, especially in the
mossy fibers of the cerebellum. (B) The localization of VGLUT1 and EAATs suggests three
possibilities for the sites of VGLUT1-EAATs interactions: between neurons and glial cells (a),
between synaptic vesicles and synaptic membranes in synaptic terminals (b) or on synaptic
membranes (c).

1.4.3. The Model of VGLUT1-EAATs Interaction
Each of the above mentioned models has its own properties and provides clues for
confirmation. In the first model, but not in the second or the third, the interaction
should occur between two adjacent cells and the interaction site on each protein
should be their extracellular domains. In the second model, VGLUT1 must be
16
localized on synaptic vesicles and the interaction site must be on intracellular
domains. In the third model, VGLUT1 and EAATs must be present in the cell
membrane of the same cells. Following these clues we designed the following
experiments. First we tested if VGLUT1 could interact with EAATs between two
individual cells. In one group, HEK 293 cells were co-transfected with two plasmids
encoding VGLUT1 and EAAT2 respectively so that each transfected cell expressed
both proteins. In another group, HEK cells were transfected with VGLUT1 or
EAAT2 respectively and cultured together for two additional days to mimic the
second model mentioned above. In the third group, the cells were transfected with
VGLUT1 or EAAT2 respectively and cell lysates were collected separately and
mixed together. As shown in Fig 3 VGLUT1 could be co-immunoprecipitated with
EAAT2 in the first group while it failed to do so in the second and third groups,
indicating that VGLUT1 only interacts with EAATs in individual cells.

In a second experiment, we tried to localize VGLUT1 and EAATs when they were
interacting with each other. Synaptosome was prepared and subjected to surface
biotinlynation (the biotinylation reagent is membrane impermeable and therefore
only labels the surface proteins). The syanptosome was further purified after the
surface biotinylation to make sure that broken synaptosome (since their internal
proteins may be biotinylated) were removed from later analysis. Then the
synaptosome was lysed and used for immunoprecipitation. The product of IP was
then divided into two equal parts: one part to be mixed with avidin-beads to remove
17
the biotin-tagged proteins, the other to be mixed with same quantity of glutathione-
beads from the same company as a control. As shown in Fig 3, when the lysates were
pre-absorbed with the beads the majority of EAAT2 was removed by avidin-beads,
while VGLUT1 only slightly reduced, suggesting that VGLUT1 is distributed
predominantly intracellularly and EAAT2 predominantly surface. VGLUT1 could be
co-immunoprecipitated by an antibody against EAAT2. When pre-absorbed with
avidin-beads, the quantity of immunoprecipitated VGLUT1 was not significantly
changed. On the other hand, EAAT2 can be immunoprecipitated with an antibody
against VGLUT1 and this immunoprecipitated EAAT2 was almost completely
removed by avidin-beads. This result indicates that EAAT2 is on the cell membrane
when it is interacting with VGLUT1, while VGLUT1 is inside of the cells
(presumably on synaptic vesicles) when it is interacting with EAAT2. Together,
these results support the second model, namely VGLUT1-EAATs interaction occurs
between synaptic vesicles and cell membrane.

18

Figure 3. The Model of VGLUT1-EAATs Interaction
(A) HEK cells were co-transfected with VGLUT1 and EAAT2 (co-transfection), or were transfected
with VGLUT1 or EAAT2 respectively and then cultured together (co-culture), or were transfected
with VGLUT1 or EAAT2 respectively and their lysates were mixed together after their lysates were
collected separately (mixed lysate). VGLUT1 co-immunoprecipitated with EAAT2 only in the co-
transfection system. (B, C) Synaptosomes were purified from rat brain and subjected to surface
biotinylation. Synaptosomes were then lysed for immunoprecipitation. Tissue lysates or
immunoprecipitation products were absorbed with avidin beads to removed biotin-tagged proteins
before SDS-PAGE and immunoblots. The data indicate that during their interaction VGLUT1 resides
inside the cells while EAAT2 is localized on the cell membranes.

1.4.4. Identification of VGLUT1-EAAT Interaction Site
Identification of the interaction site(s) for VGLUT1-EAAT interaction may provide
a useful tool to perturb this interaction and study its functional implications. To this
end, we did serial truncations and mutations on VGLUT1. Firstly, 7 stop codons
were introduced at residues 492, 456, 335, 200, 104 and 33, respectively. The
truncated VGLUT1 were co-transfected with EAAT2 into HEK cells. The VGLUT1-
EAAT2 interaction was then checked by co-IP with the antibody against the N-
terminal domain of VGLUT1. We found that when VGLUT1 was truncated at
residue 104 the interaction remained indicating that the amino-terminus of VGLUT1
may contain an interaction site. The interaction disappeared when VGLUT1 was
truncated at residues 33, suggesting that the sequence between residues 33 and 104 is
19
critical for the VGLUT1-EAAT interaction (caution ought to be exercised here due
to the possibility that the N-terminal 33 amino acids may not be able to bind to the
VGLUT1 antibody). Next, we made a series of plasmids encoding fusion proteins
composed by short sequences of VGLUT1 amino-terminal region and EYFP (EYFP
allowed us to use a GFP antibody for immunopecipitation). These plasmids were co-
transfected with EAAT2 and VGLUT1-EAAT2 interaction was checked by co-IP
with an antibody against GFP. The results are summarized in Fig 4. It appeared that
the short sequence at the amino-terminal intracellular region from residue 63-68 near
the first trans-membrane domain of VGLUT1 is critical for the VGLUT1-EAAT
interaction. Then we made a mutant of VGLUT1 which lacked this sequence and
found that this mutation significantly reduced but not abolished the interaction,
suggesting the existence of additional interaction site. It should be noted that the
region from residue 63 to residue 68 is conserved between VGLUT1 and VGLUT2
while only VGLUT1 could interact with EAATs. One possibility is that in the native
VGLUT2 this interaction site is not exposed. It is also possible that this sequence is
critical for keeping certain other structures of VGLUT1 in the right conformation for
the interaction but not the interaction site by itself. We also made a construct by
fusing EYFP with the sequence from 63 to 68 and found that it did not pull down
EAAT2 co-transfected with it. These results suggested that this sequence may be
necessary but not sufficient for the VGLUT1-EAAT interaction. Taken together; the
amino-terminus of VGLUT1 contains an interaction site for EAATs. The sequences
20
among this region between 63 and 68 appeared to be critical for this interaction and
may be one of the interaction sites.

Figure 4. Identification of Interaction Site on VGLUT1
(A) A serial of truncations of VGLUT1 were made at indicated residues. VGLUT1-EAAT2
interaction remained when VGLUT1 was truncated at residue 104, but not at residue 33. (B) A serial
fusion proteins were made by fusing the indicated N-terminal fragments of VGLUT1 with EYFPc1.
When the VGLUT1 sequence from residue 63 to 68 was deleted the fusion protein was no longer able
to pull down co-expressed EAAT2. The red segment indicates the putative first transmembrane
domain.

21
1.4.5. VGLUT1-EAATs Interaction Occurs in Neurons and Glial Cells
Although VGLUT1 is predominantly distributed in neuronal terminals, accumulating
evidence indicats that glial cell can also release glutamate through VGLUT1-
containing vesicles. So we tested if the VGLUT1-EAATs interaction can also occur
in glial cells. With sucrose gradient centrifugation rat forebrain tissue was separated
into fractions which are enriched with GPV (the reseal of the processes of glial cells)
or enriched with synaptosome. As shown in Fig 5, the GPV fraction contained more
EAAT2 and less VGLUT1 than synaptosome. Then the lysates of these two fractions
were used for immunoprecipitation with an antibody against VGLUT1, respectively.
Interestingly, VGLUT1 antibody pulled down almost equal amount of EAAT2 from
both fractions. This result suggests that the VGLUT1-EAATs interaction may
happen in both neurons and glial cells.

Figure 5. VGLUT1-EAATs Iinteraction Occurs in Neurons and Glial Cells
(A) Synaptosomes and Glial Plasmalemmal Vesicles (GPV) were purified from rat hippocampus and
lysed respectively for immunoprecipitation. EAAT2 was enriched in GPV fraction while VGLUT1
was enriched in synaptosomes. VGLUT1 co-immunoprecipitated nearly equal amount of EAAT2
from syanptosome and GPV, suggesting that the VGLUT1-EAAT2 interaction could occur in both
neuronal terminals and in the processes of glial cells. (B) Schematic showing that the VGLUT1-
EAATs interaction occurs between vesicles and cytoplasmatic membrane, in both neurons and glial
cells.
22
1.4.6. The Functional Implications of VGLUT1-EAAT Interaction
Next we investigated the functional implications of the VGLUT1-EAATs
interaction. We first studied if VGLUT1 alters the function of EAATs. HEK cells
were transfected with EAAT2 and control vectors or plasmids encoding VGLUT1 or
VGLUT2. 48 hours after transfection cells were incubated with different
concentrations of glutamate (10 to 500 µM) containing H
3
-labeled glutamate (the
ratio of H
3
-glutamate to unlabeled glutamate remains the same) for 5 min. Then the
cells were washed with cold solutions to remove the extracellular glutamate. The
glutamate uptaken by the cells were then measured by counting H
3
. As shown in Fig
6, the expression of EAAT2 increased the uptake of glutamate by around 10-fold
compared to that of non-transfected cells. The transfections with VGLUT1 or
VGLUT2 by themselves did not increase the uptake of HEK cells. Co-expression of
VGLUT1 or VGLUT2 with EAAT2 did not significantly alter the uptake activity
mediated by EAAT2, suggesting that the VGLUT1-EAAT interaction does not
change the function of EAATs.

Since EAATs are eletrogenenic and produce a net inward current when they
transport glutamate we also tested if this current can be altered by VGLUT1.
EAAT2 were co-transfected into HEK cells with a control vector or VGLUT1.
EAAT2 mediated inward current was elicited with a fast perfusion of 100 μM of
glutamate and recorded with a whole cell recording model. We found that glutamate
23
induced a strong inward current in the cells expressing EAAT2, but the current was
not altered by co-expression of VGLUT1.

Next we checked if EAAT can change the function of VGLUT1. HEK cells were
divided into two groups. In one group, cells were transfected with EAAT2 or
VGLUT1 respectively and then cultured together for additional 48 hours (co-
culture). In the other group, cells were co-transfected with EAAT2 and VGLUT1
(co-transfection). Then the cells were fixed for immunocytochemistry. As shown in
Fig 6 EAAT2 immunoreactivty was mainly on cell membrane while the
immunoreactivity of VGLUT1 mainly on vesicle-like intracellular structures. In the
co-culture system, the VGLUT1-positive vesicles were evenly distributed in the
cytosol of cells. By contrast, in the co-transfected cells, the immunoreactivity of
VGLUT1 was concentrated in the peri-membrane region. This result suggests that
the presence of EAAT on the cell membrane altered the trafficking properties of the
VGLUT1-containing vesicles. This finding is consistent with previous reports
showing that the VGLUT1-, but not VGLUT2-containing vesicles are accumulated
in the peri-memebrane region in PC12 cells where EAAT is endogenously expressed
(Kobayashi S and Millhorn DE., 2001).
24

Figure 6. The Functional Implications of VGLUT1-EAAT Interaction
(A) HEK cells were transfected with indicated plasmids and incubated with H
3
-labeled glutamate at
indicated concentrations for 5 min. The uptake of glutamate was measured with scintillation counting.
Co-expression of VGLUT1 or VGLUT2 did not alter glutamate-uptake activity of EAAT2 alone. (B)
HEK cells were transfected with VGLUT1 or EAAT2 separately and then cultured together (co-
culture), or co-transfected with VGLUT1 and EAAT2 (co-transfection). VGLUT1 immunoreactivity
was evenly distributed in the cytosol in the co-culture condition, whereas it was mainly restricted to
the perimembrane region in the co-transfection condition.

1.5. Discussion
1.5.1. The Controversy Regarding Presynaptic EAATs
It is still controversial as to whether the two major membrane glutamate transporters,
EAAT1 and EAAT2, which are thought to be glia-specific, also exist on the
presynaptic membrane and if so, what their contributions to the uptake of glutamate
are. Initially it was believed that the two major membrane transporters, GLAST and
GLT-1, were exclusively expressed in glial cells and absent from neurons (see
25
review by Danbolt NC., 2001). Later it was suggested that EAAT1 and EAAT2 were
expressed in neurons in certain pathology conditions or induced by some
experimental manipulations (Mennerick S et al., 1998). But accumulating data hint at
the presence of EAAT1 or EAAT2 in presynaptic terminals under normal conditions.
An early study showed that the metabolically inactive glutamate analog, the D-
aspartate, which can be uptaken by EAATs but can not be transferred from glia to
neurons, can be directly uptaken into synaptic terminals in hippocampal slices. This
presynaptic uptake was so robust that it accounted for moe than half of the total
uptake (Gundersen V et al., 1993). Interestingly, the same group of researchers did
not observe the uptake of D-aspartate in synaptic terminals of spinal cord, indicating
that the presynaptic uptake may only happen in the brain (Gundersen V et al., 1995).
More recently, a biochemical study purified synaptosome and GPV and detected
EAAT2 in the synaptosome preparation. The authors found that the synaptosome
could uptake glutamate which is sensitive to the inhibitors of EAAT2 (Suchak SK et
al., 2003). With the EAAT2-knockout mice as control, a recent
immunohistochemistry study also detected EAAT2 in a subset of synaptic terminals
in the hippocampus (Chen W et al., 2004). As the high protein level of EAAT1 or
EAAT2 in the processes of glial cells may mask their presence in neighboring
neuronal terminals, Dr.Danbolt conducted an quantities EM study with a preparation
where the cell membrane of neurons is more separated from glial cells and he found
that ~20% EAAT2 was actually distributed in synaptic terminals and axons (Danbolt
et al., 2006, SFN annual meeting, A quantitative assessment of glutamate uptake into
26
hippocampal synaptic terminals and astrocytes). In the current study we confirmed
the presence of EAAT2 in the synaptosome preparation. More importantly, we found
that VGLUT1 can interact with EAAT2 in this preparation, which lends strong
support for the presynaptic distribution of EAAT2 (and EAAT1). A study on retina
indicates that the presynaptic EAATs are actually the major transporters for the
uptake and recycle of glutamate (Hasegawa J et al., 2006) although no similar work
has been conducted on the synapses in the brain. Some studies suggest that EAAT2
changes its stoichiometric ratio of transport and transitions to a high throughput
mode when the extracellular glutamate is elevated to mM level, which is a
concentration can be reached in synaptic cleft during vesicular release (A.Y.Kabakov
and P.A.Rosenberg, 2005, 2007, SFN annual meeting). These studies suggested that
the presynaptic EAATs may be able to uptake a large amount of glutamate despite
their lower expression level. Consistant with these studies, it was found that
impairment of glutamate-glutamine cycle only partly depressed the amplitude of
EPSP and this effect was rapidly reversed. More importantly, synaptic transmission
persisted under this situation (Kam K and Nicoll R, 2007), indicating that the
glutamate-glutamine transformation between glia and neurons may not be as crucial
for the excitatory transmission as previously expected.
On the other hand, VGLUTs were thought be exclusively localized in presynaptic
terminals. But increasing evidence indicates their presence in glial cells. EM studies
showed that glial cells contain VGLUT1 and VGLUT2-positive vesicular structures
(Bezzi P et al., 2004). Functional studies also indicate the release of vesicular
27
glutamate from glial cells in response to stimuli (see review by Montana V et al.,
2005). Consistent with these studies we detected VGLUT1 in the purified GPV
preparation (although less than that in synaptosome) in the current study.
Interestingly, we found that VGLUT1 can also interact with EAATs in this
preparation, suggesting that VGLUT1-EAATs interaction may play some general
roles for the vesicular release of glutamate in both neurons and glial cells.
1.5.2. VGLUT1-EAATs Interaction and Synaptic Function
In the current study we provided data establishing the physical interaction between
VGLUT1 and EAATs and revealing the model for this interaction. We provided only
some preliminary data obtained from non-neuronal cells as to the functional
implications of this interaction. Since the interaction occurs only between VGLUT1,
but not VGLUT2, and EAATs, the VGLUT1-EAAT interaction may account for the
functional differences between VGLUT1 and VGLUT2. The VGLUT1 and
VGLUT2 are complementarily distributed in the adult brain. And their expression is
developmentally regulated in a way that some VGLUT1-only neurons in adult
express VGLUT2 in early development. The distinct distribution and developmental
switch indicate that some functional differences lie between these two major
transporters. But the uptake properties of these two VGLUTs appear to be very
similar (Takamori S et al., 2001), suggesting some other functional difference
between them. Interestingly, the expression of VGLUT1 vs. VGLUT2 is correlated
with the release probabilities of the synapses. The typical glutamatergic syanpses in
the brain does not reliably convert action potential into release events of glutamate.
28
The probability of each action potential to trigger a release (release probability) is
about 0.1 to 0.2 (see review by Sudhof TC 2004). VGLUT1 is densely expressed in
the hippocampus, cerebral cortex and cerebellum cortex where the synapses
generally have lower releasable probabilities, while VGLUT2 is enriched in the
lower structures such as the thalamus/hypothalamus nuclei, brainstem and deep
cerebellar nuclei where synapses possess a relatively higher release probability.
Interestingly, the VGLUT1-expressing synapses are also more susceptible to plastic
modifications (Fremeau RT Jr et al., 2001, 2004B). In HEK cells, expressing EAAT2
can recruit the VGLUT1-positive vesicles into the peri-cell membrane region,
suggesting that the presynaptic EAAT2 may alter the trafficking properties of the
VGLUT1-containing synaptic vesicles in the brain and therefore change the release
probability. To further confirm this, the release probabilities of certain synapses (e.g.
the pyramidal cell in CA1 region of hippocampus, where only VGLUT1 is
expressed) should be compared between wild type and EAAT2 and/or EAAT1-null
mice. It can be expected that deletion of the EAATs will increase the release
probability of these synapses.

Considering the importance of the release and uptake of glutamate in the brain
functions, some mechanism should be in place to couple these two processes to
maintain the transmission at the right level and to avoid the possibility of excitotoxic
effects. An early study showed that the extracellular glutamate can increase the
surface expression of EAAT1 (Duan S, et al., 1999). This effect happened on a
29
relatively slow time scale with the peak effect occuring after 15 minutes. This study
was conducted with cultured glial cells and it is unknown if the same mechanism is
also present in the brain. In addition, the surface ratio of EAATs in the brain is very
high and therefore does not have a large reserve for upregulation. Can there be other
mechanisms to couple the release and uptake of glutamate? Since the VGLUT1 is the
molecule packing glutamate for vesicular release and EAATs uptaking released
glutamate, the physical interaction between these two groups of transporters may
provide a mechanism to couple release and uptake. One hypothesis is that VGLUT1
on the vesicles interacts with and tonicly inhibits the uptake activity of EAATs on
the presynaptic membrane. When the vesicle is released, the VGLUT1-EAAT
interaction is broken unleashing its inhibition on EAAT. Under this situation, the
EAATs increase its uptaking to remove the released glutamate. To test this
hypothesis, we measured the uptake of glutamate by EAAT2 in HEK cells. To our
surprise, the co-expression of VGLUT1 or VGLUT2 barely altered the activity of
EAAT2. But this result did not disprove the hypothesis since HEK cells lack the
structures comparable to synaptic vesicles. In the future, similar work needs to be
conducted in the neuronal system.
1.5.3. VGLUT1-EAAT Mediates a New Model of Glutamate Uptake?
The glutamate/glutamine cycle involving both glial cells and neurons was believed to
be the major mechanism for recycling glutamate in the brain. But as mentioned
above, the uptake by presynaptic terminals may be underestimated in previous
30
studies. In addition, disruption of the glutamate/glutamine cycle caused only a minor
impact on excitatory transmission. These studies indicate that the current model of
the glutamate transmission may need to be updated. In the current study, we showed
that the presynaptic EAATs can physically interact with the VGLUT1 on synaptic
vesicles. Can these two transporters together form a transporting complex which
uptakes glutamate directly from extracellular space into synaptic vesicles? The
advantage of this model is obvious. Firstly, glutamate can be more efficiently reused
without the energy-consuming glutamate-glutamine cycle. Secondly, glutamate can
be directly packed into synaptic vesicles without changing the cytosolic
concentration of glutamate in synaptic terminals. As some previously studies showed
that glutamate can also function as an intracellular signaling molecule (Maechler P,
Wollheim CB, 2001), it would be advantageous avoid the fluctuation of the
intracellular concentration of glutamate. Another attractive aspect of this hypothesis
is that it may help to explain the source of extracellular glutamate in ischemic stroke.
Extracellular glutamate concentration is elevated in the early phase of ischemia and
is considered to be one of the critical steps for neuronal death. More and more
researchers accept the hypothesis that the source of this increased extracellular
glutamate is the reverse uptake by EAATs (namely release glutamate from inside of
cells, Rossi DJ et al., 2000). Since EAATs are predominantly distributed in glial cells,
one can imagine that glutamate is released from glial cells under this situation. But
surprisingly, semi-quantitative measurement of glutamate in ischemia indicated that
the majority of the glutamate actually comes from neurons (Torp R et al., 1991;
31
Ottersen OP et al., 1996). If what we proposed is true, VGLUT1-EAAT complex
directly connects extracellular space with synaptic vesicles, this complex may also
reverse its activity like EAATs was described in ischemia. This reversed activity of
VGLUT1-EAAT complex may allow the direct release of glutamate from synaptic
vesicles. Since the concentration of glutamate in vesicles is much higher than that in
the cytosol of glial cells, this will reconcile the facts that glutamate is released by
reversed EAATs in neurons while they are mainly expressed in glial cells. Although
this hypothesis is attractive, it is technically difficult to directly prove due to the size
of synaptic terminals and vesicles. Electron microscopic analysis of the uptake of the
isotope-tagged glutamate may be conducted for this purpose in the future.

32
Chapter 2. Homeostasis in Synaptic Plasticity

2.1. Summary
Despite continuous advances in our understanding of the mechanisms of synaptic
plasticity the roles of presynaptic modifications in long term synaptic plasticity
remain unclear partly due to a shortage in techniques to directly measure the release
of glutamate from synaptic terminals. Here we developed a method to measure
presynaptic release through the biotinylation of vesicular transporters in vesicles
fused with presynaptic membranes during neurotransmitter release. With this method,
which allowed us to selectively measure the spontaneous or evoked release of
glutamate or GABA, we investigated the expression mechanism of DHPG induced
long-term depression in cultured hippocampal neurons, which manifests as a
depression of miniature excitatory postsynaptic currents (mEPSCs). Surprisingly,
this LTD was associated with an increased presynaptic release of glutamate, which
argues against the possibility of a presynaptic expression mechanism. In addition,
DHPG-induced internalization of AMPA receptors, along with the LTD of mEPSCs,
was attenuated by a GluR2-derived peptide that specifically inhibits clathrin-
mediated AMPAR endocytosis. Interestingly, we found that the blockade of
AMPAR endocytosis also reduced DHPG-induced increase in presynaptic release.
Thus, our work demonstrates that DHPG-induced postsynaptic endocytosis of
AMPA receptors and synaptic depression in turn results in a homeostatic
modification at the presynaptic site. This study also indicates that the biotinylation of
33
vesicular transporters in live cultured hippocampal neurons is a valuable approach
for studying presynaptic functions.

2.2. Background and Introduction
2.2.1. Synaptic Plasticity and Memory
Learning and memory is a critical ability which enables animals to adapt to their
environments. Since the brain is a network of neurons (glial cell may also be
included) connected through synapses, memory may be stored in the brain through
modifying certain properties of the neurons (eg. neuronal excitability) or through
modifying of the synaptic connections between neurons (eg. synaptic plasticity,
which is the plastic modifications of the strength of synaptic connections). An
advantage of synaptic plasticity over modification of neuronal excitability for
memory storage is its much larger capacity (considering that each neuron may form
1,000 to 10,000 synapses with other neurons). In the last a few decades, a huge
amount evidences were accumulated supporting the notion that synaptic plasticity is
critical for memory although it is still not enough to make a conclusion (see review
by Martin SJ et al., 2000).

In living animals or in brain slice preparations, the plastic changes of synaptic
strength can be reliably induced by applying a train of electric stimuli with specific
frequencies to the presynaptic terminals. Long-term potentiation (LTP) and long-
term depression (LTD) are the prominent samples of these kinds of synaptic
34
plasticity. LTP and LTD can be induced in basically all the brain regions checked so
far. They exhibit some special properties which make them sound candidates for
information storage in the brain. The induction process sometimes can be very fast
(eg. 1 second for LTP). The potentiation or depression can last long after the
induction stimuli. In some reports, LTP can even last for over 1 year in living
animals. The plasticity is synaptic specific while it exhibits associability between
different synapses. All these properties to some degree mimicked the processes of
memory (see review by Bliss TV and Collingridge GL, 1993).

In the last a few years, since the increasing understanding of the induction and
expression mechanisms of LTP and LTD, numerous experiments have been
conducted in living animals to test the roles of synaptic plasticity in learning and
memory. These studies provide strong evidences for two issues: (1) learning and
memory is accompanied by synaptic plasticity in the brain (eg. Whitlock JR et al.,
2006); (2) interfering the induction or expression of synaptic plasticity can impair
memory (eg. Rumpel S et al., 2005). In the future, the following two lines of
evidences may be required to conclude that synaptic plasticity is indeed memory: (1)
to selectively tag the synaptic plasticity associated with specific memory and to
check if reversing this plastic modification can erase this specific memory; (2) to
create memory trace of non-experienced experience in the brain by inducing synaptic
plasticity in a specific pattern. To fulfill these aims, we need to gain deeper
understandings of the molecular mechanisms underlying synaptic plasticity.
35

2.2.2. Expression Mechanisms of Synaptic Plasticity
Since the first demonstration of long-term potentiation (LTP) of synapses in the
central nervous system it has been debated for three decades as to where, presynaptic
versus postsynaptic, is the locus for the expression of this persistent plastic change.
In recent years, with accumulating evidence, it has been generally accepted that the
postsynaptic trafficking of glutamate receptors is the major mechanism for long term
plasticity (Malinow R and Malenka RC, 2002, Collingridge GL et al., 2004).

But it remains unclear what roles the presynaptic site plays in this process, which is,
at least in part, due to the lack of direct and quantifiable methods to measure
presynaptic release. In most previous studies, the presynaptic activity is inferred
indirectly through electrophysiological analysis of EPSC or EPSP. For example,
changes in miniature EPSC frequency, in failure rate and in paired pulse facilitation
ratio were used as indicators of presynaptic function. But the study of silent synapses,
and the rapid changes in the number of postsynaptic receptors in particular, put the
validity of these parameters to question (Malinow R and Malenka RC, 2002, Moult
PR et al., 2006). Imaging of synaptic vesicular turnover provides a valuable
alternative for the direct measurement of transmitter release. Early work utilized the
vesicular uptake of antibody against the lumenal domain of synaptotagmin-1
(Malgaroli A et al., 1995). More recently, FM1-43 and its analogs have been used to
label the release of synaptic vesicles (Zakharenko SS et al., 2001; Zakharenko SS et
36
al., 2002; Stanton PK et al., 2003; Zhang XL et al., 2006). However, these
techniques have their limitations. Firstly, release from excitatory and inhibitory
synapses cannot be distinguished. In addition, the efficiency of vesicle labeling is
limited by the large size of antibodies, especially in the kiss-and-run mode.
Furthermore, quantification of images can be greatly affected by subjectivity and
high variability of image quality between samples. With these shortcomings, it is not
surprising to find conflicting results (Zakharenko SS et al., 2002; Zhang XL et al.,
2006).
Here we designed a biochemical assay of transmitter release by measuring the
amount of vesicular transporters exposed on the extracellular side of the presynaptic
membrane during transmitter release. Glutamatergic and gamma-aminobutyric acid
(GABA)-ergic synaptic vesicles can now be distinguished due to the presence of
unique vesicular transporters in these two types of terminals (McIntire SL et al.,
1997; Takamori S et al., 2000; Fremeau RT Jr et al., 2001). The small size of the
biotinylating reagents greatly increases the efficiency of labeling. Moreover,
quantification of data from this biochemical assay is less prone to subjectiveness or
sampling error compared to imaging techniques. Finally, this method allows
simultaneous measurement of the surface expression of transmitter receptors and
therefore provides an opportunity to compare the pre- and post-synaptic activity
under the same conditions. With this method we studied a chemically induced long-
term depression (LTD) by DHPG in neuronal culture and found an elevation of
37
presynaptic activity in this model, suggesting a homeostatic relationship between the
presynaptic and postsynaptic functions.
2.3. Methods and Materials
Materials: the Biotinylation reagents including Sulfo-NHS-LC-Biotin, Biotin-PEO
4
-
Hydrazide and Maleimide PEO2-Biotin and NeutraAvidin beads are from Pierce.
The following antibodies have been used: VGLUT1 (Synaptic Systems, Cat
#
135
302); VGAT (Synaptic Systems, Cat
#
131 002); synaptotagmin-1 (Synaptic Systems,
Cat
#
105 011); transferrin receptor (Zymed, Cat
#
13-6890); α-1 subunit of GABA (A)
Receptor (Chemicon, Cat
#
AB5946), GluR1 (Chemicon, Cat
#
AB1504); GluR
2

(Chemicon, Cat
#
MAB397). DHPG was from Tocris, TTX, PMA and CHA were
from Sigma.
Hippocampal cultures: Culture dishes and plates and coverslips were coated in 10
μg/ml poly-D-lysine dissolved in Borax buffer (0.15 M, PH 8.4) overnight. Neurons
from E18 rat embryonic hippocampi were dissociated and plated in neurobasal
medium supplemented with B27, 0.5 mM of glutamine and 12.5 μM of glutamate in
6-well plates at 0.5X10
6
cell/well or in 10-cm dishes at 3 x 10
6
cell/dish. Neurons
were switch to maintenance medium (Neurobasal medium supplemented with B27
and 0.5 mM glutamine) the next day and fed twice per week until they were ready
for experiments at 14-18 DIV.
Immunocytochemistry : Neurons were fixed with 4% paraformaldehyde, 4%
sucrose in PBS (on ice, 10 min) and methanol (–20°C, 30 min) successively.
38
Neurons were further permeabilized with 1% Triton X-100 for 5 min. After blocking
with 2% goat serum (30 min), neurons were incubated in primary antibodies
dissolved in 0.5% goat serum in PBS for 1 hour. After several washings, neurons
were incubated with corresponding fluorescence-tagged secondary antibody for 30
minutes. After three washes, neurons were then mounted for observation. The
primary antibodies used in current study including: anti-GluR
2
(Chemicon,
MAB397); anti-GABA
A
R α-1 (Chemicon, AB5946); anti-VGLUT1 (Synaptic
Systems); guinea pig anti-VGAT (Chemicon, AB5855) and rabbit anti-VGAT
(Synaptic Systems). The fluorescence-tagged secondary antibodies (Molecular
Probes) including: Alexa Fluor 488 goat anti–mouse IgG (Green); Alexa Fluor 546
goat anti–rabbit IgG (Orange); Alexa Fluor 350 goat anti–rabbit IgG (Blue) and
Alexa Fluor 546 goat anti–guinea pig IgG (Orange).
Biotinylation assay: Neurons were firstly switched from culture medium to normal
extracellular solution (ECS, containing (mM): NaCl, 140; KCl, 2.7; CaCl
2
, 1.7;
MgCl
2
, 1.0; HEPES, 10; glucose, 33). After being incubated in ECS at room
temperature for 45-60 min, neurons were subjected to various treatments as
described in the results. Then the neurons were incubated with Sulfo-NHS-LC-Biotin
or other biotinylation reagents for indicated Labeling Time (5-30 seconds; 30
seconds was used in most assays unless otherwise noted) to allow the vesicular
transporters exposed on cell membrane during synaptic release to react with the
biotinylation reagents and also to allow released synaptic vesicles to uptake
biotinylation reagents. The biotinylation reagents were dissolved in normal ECS for
39
spontaneous release (in this case, the neurons were pre-treated with 1 μM of TTX for
5 min before the onset of Labeling Time), in ECS containing high concentration of
potassium (high K
+
, containing (mM): NaCl, 31.5; KCl, 90; CaCl
2
, 1.7; MgCl
2
, 1.0;
HEPES, 10; glucose, 33) for depolarization-evoked release, or in hypertonic ECS
with an addition of 500 mM of sucrose (to achieve a total osmolarity of ~810 mOsm)
for measurement of readily releasable pool. After the Labeling Time, biotinylation
reagent was removed and pre-chilled ECS was poured promptly into the culture
dishes or plates to stop synaptic release. After being washed three times with pre-
chilled ECS, the neurons were incubated in ECS on ice for 30-45 min and then
homogenized in boiling lysis buffer (1% SDS, 10 mM Tris, 0.2 mM sodium ortho-
vanadate; pH 7.4. For each well in 6-well plate, 125 μl of lysis buffer was used).
Cell lysates were shaken on thermomixer (95 ºC) for 5 -10 min, sonicated for 30
seconds and then centrifuged at 14,000 rpm for 30 min. The supernatants containing
equal amounts of total protein were then diluted into RIPA buffer at 1:10 ratio and
incubated with NeutraAvidin beads at 4 ºC for overnight to precipitate biotinylated
proteins. After being washed in RIPA Buffer, biotinylated proteins were eluted from
the beads by boiling in SDS-Sample buffer. Finally, the biotinylated proteins were
loaded for SDS-PAGE and immunoblot. The densities of immunoblot bands were
digitized and quantified with NIH image. All band intensities were normalized to
that of the control samples.
40
Peptide treatment: The GluR2
3y
peptide was synthesized according to the following
sequence: Tyr-Lys-Glu-Gly-Tyr-Asn-Val-Tyr-Gly-Ile-Glu-Ser-Val-Lys-Ile. The
transmembrane peptide carrier, pep-1, was synthesized according to a previous
report (Morris MC et al., 2001) with the following sequence: Lys-Glu-Thr-Trp-Trp-
Glu-Thr-Trp-Trp-Thr-Glu-Trp-Ser-Gln-Pro-Lys-Lys-Lys-Arg-Lys-Val. GluR2
3y

peptide was first mixed with pep-1 at a ratio of 1:20 and incubated for 30 min to let
the GluR2
3y
- pep-1 complex to form (all peptides were dissolved in ECS). Then the
neurons were pre-treated with this peptide complex (with the final concentration of
GluR2
3y
to be 1 μM) for 1 hour to allow the complex to enter neurons.
Electrophysiology: Whole-cell voltage clamped recordings were made from
cultured hippocampal neurons 14-18 days after plating or CA1 neurons in 400 µm
hippocampal slices acutely dissected from 19-25 day old rats. The patch electrode
solution contained the following (mM): Cs Gluconate, 110; CsCl 17.5; EGTA, 0.5;
MgCl2, 2; HEPES, 10; QX314, 5; Tris-GTP, 0.4; K2ATP, 4 (pH 7.2; osmolarity
290-300 mosmol-1). The extracellular (perfusion or bathing) solution (ECS) for
cultured neurons contained (mM): NaCl, 140; CaCl2, 1.3; KCl, 5.0; MgCl2, 1.0;
HEPES, 10; glucose, 33; TTX, 0.0005; bicuculline methiodide, 0.01 (pH 7.4; 325-
335 mosmol-1). All slice recordings were done in bicarbonate buffered extracellular
solution (NaCl, 125; CaCl2, 2.0; KCl, 2.5; MgCl2, 1.0; NaHCO3, 26; NaH2PO4, 1.3;
bicuculline methiodide, 0.01) continously bubbled by 5%/95% CO2/O2 carbogen.
Recordings were performed at room temperature and lasted from 40 to 80 minutes.
41
mEPSCs from cultured neurons were recorded using an Axopatch 1-D amplifier
(Axon Instruments, Inc.), sampled at 10 kHz after being filtered at 2 kHz, and stored
in computer using Clampex 8.0 (Axon instrument). EPSCs from CA1 neurons in
slices were evoked every 20 seconds and recorded in a similar manner using a
MultiClamp 700A amplifier (Axon Instruments, Inc.). After 10 minutes of stable
baseline recording, 100 µM DHPG was bath applied for 5 minutes to induce LTD.
Statistics: Data are presented as mean ± S.E.M. Student's t-tests were used for
statistical comparison between two groups. ANOVA was employed for comparisons
between more than 2 groups.
2.4. Results
2.4.1. Measurement of Spontaneous Synaptic Release
The vesicular neurotransmitter transporters are molecules that assign identity to the
synaptic terminals. To validate vesicular glutamate transporter 1 (VGLUT1) as the
presynaptic marker of excitatory synapses in hippocampal neuronal culture we first
examined the cellular distribution of VGLUT1 and AMPA receptors (Fig. 7).
Immunocytochemistry double-labeling showed that over 90% of the clusters of
GluR2 subunit of AMPA receptor were juxtaposed to VGLUT1-containng
presynaptic boutons. Less than 10% of GluR2 clusters were not contacted by
VGLUT1-positive boutons and those might be VGLUT2-positive synapses (Fremeau
RT Jr et al., 2001). Next we double-labeled the vesicular transporter of GABA
(VGAT) and the α1 subunit of GABA (A) receptors. Over 95 % of α1-positive
42
clusters were contacted with VGAT-positive boutons. Double-labeling of the two
transporters, VGLUT1 and VGAT, showed almost no co-localization indicating that
they are expressed in two distinct populations of terminals. These results indicate
that VGLUT1 and VGAT respectively specify the excitatory and inhibitory synapses
in hippocampal cultures.
Vesicular transporters (VGLUT1 and VGAT) are separated from extracellular matrix
by presynaptic membrane (Fig. 8). Only when the synaptic vesicles fuse with cell
membrane are the lumenal domains of these proteins exposed to extracellular fluid
and subject to biotinylation by membrane-impermeable reagents. We first tested if
biotinylation can be used to monitor the spontaneous release of synaptic vesicles.
Hippocampal neuronal culture was incubated in extracellular solution (ECS)
supplemented with tetrodotoxin (TTX) for 5 min to block the action potentials (This
procedure was used for all the following experiments involving spontaneous release
and will not be reiterated hereafter). The neurons were then switched into ECS
containing a biotinylation reagent (Sulfo-NHS-LC-Biotin, 1 mg/ml) for different
durations (5–30 seconds) at room temperature (23°C). After this short incubation
time (referred to as “Labeling Time” henceforth), the Sulfo-NHS-LC-Biotin was
washed away with pre-chilled ECS (the purpose of using pre-chilled ECS is to stop
active membrane trafficking). The vesicular transporter in some vesicles can be
labeled by biotin right away when the vesicles were released and exposed their
transporters to the reagent. Some vesicles can also uptake the biotinylation reagent
when they retreat to the inside of the cells and allow the biotinylation reaction to
43
continue inside the vesicles (Fig. 8). After the biotinylation reagent was washed
away from extracellular solution the cells were kept on ice for 30 minutes (min) to
allow the incorporated biotinylation reagent to react completely with vesicle-
associated proteins. Then the neurons were lysed and the biotin-tagged proteins were
affinity precipitated with avidin beads and analyzed with SDS-PAGE and
immunoblotting. We discovered that the amount of biotin-tagged VGLUT1
increased in proportion to the length of Labeling Time From 5 seconds (s) to 30 s,
the amount of biotin-labeled VGLUT1 increased to 3.28 ± 0.29 fold (n = 4). As a
control, we measured the biotinylation of transferrin receptor (TfR), a dendritic
protein (Silverman MA et al., 2001) whose surface expression is not directly affected
by presynaptic release. The biotinylation of TfR increased to 1.85 ± 0.31 fold (n = 4)
by the elongating the Labeling Time from 5 s to 30 s. This increase suggests that the
biotinylation reaction with surface proteins could not complete within this amount of
time and that increasing the reaction time could allow more proteins to be labeled.
The increase of the labeling of VGLUT1 was significantly higher than that of TfR
(two-way ANOVA, P < 0.001, n = 4). The faster increase in the biotinylation of
VGLUT1 should be attributed to the accumulation of spontaneous release events
during the Labeling Time which increased the amount of the transporters accessible
to biotinylation reagent. For further confirmation, we conducted the biotinylation at 4
°C whereas the active vesicular release was stopped. Under this condition, the
biotinylation of VGLUT1 but not that of TfR decreased significantly compared to
that at room temperature. The results indicate that biotinylation with membrane-
44
impermeable reagents can be used to monitor the spontaneous turnover of synaptic
vesicles.

Figure 7. VGLUT1/VGAT Specifies Excitatory/Inhibitory Synapses
Immunocytochemistry with cultured hippocampal neurons. Upper panel: the VGLUT1-positive
presyanptic boutons are juxtaposed to GluR
2
-positive clusters; middle panel: VGAT-positive boutons
are juxtaposed to the immunoreactivity of the α1 subunit of GABAA receptors. Bottom panel: No
colocalization of VGLUT1 and VGAT in presynaptic boutons, indicating that they are two distinct
groups of synapses.

Next we tested if biotinylation of transporters is sensitive enough to detect the
modifications of the spontaneous release. We treated neurons with Phorbol 12-
myristate 13-acetate (PMA) (1 μM, 10 min), a phorbol ester known to increase
glutamate release (Lonart G and Sudhof TC, 2000; Rhee JS et al., 2002). We found
45
that the biotinylation of VGLUT1 (30 s Labeling Time) increased to 1.67 ± 0.20 fold
by treatment with PMA (t-test, P < 0.05, n = 4). The biotinylation of VGAT was not
significantly changed (0.93 ± 0.06 fold, t-test, P > 0.05, n = 4). We also treated cells
for 10 min with 10 μM of N6-cyclohexyladenosine (CHA), an adenosine A1
receptor agonist that has been reported to reduce synaptic release (Manzoni OJ et al.,
1994; Jeong HJ et al., 2003). This treatment significantly decreased the biotinylation
of both VGLUT1 (0.62 ± 0.08, t-test, P < 0.05, n = 4) and VGAT (0.59 ± 0.12, t-test,
P < 0.05, n = 4). Together, these results indicate that the spontaneous release of
glutamate and GABA can be selectively monitored by biotinylation of VGLUT1 and
VGAT, respectively, with cell- impermeable reagents.

Figure 8. Measurement of Spontaneous Release
(A) Schematic showing the biotinylation method. (B, C) Representative blots and quantification
showing the biotinylation of VGLUT1 increasing with elongation of Labeling Time (duration of
incubation with biotinylation reagents). (D, E) Representative blots and quantification showing the
facilitation and inhibition of vesicular release by PMA and CHA, respectively.
46
2.4.2. Measurement of Evoked Release
We then tested if biotinylation can be used to monitor the evoked release.
Spontaneous release occurs with low frequency in cultured neurons (Murthy VN and
Stevens CF, 1999), whereas depolarization of neurons can elicit massive
synchronized vesicular release events and therefore increase the number of vesicular
transporters accessible to biotinylation reagent. We first used ECS containing high
concentrations of potassium (high K
+
) to depolarize neurons (Fig.9). Neurons were
incubated with biotinylation reagent (Sulfo-NHS-LC-Biotin, 1mg/ml) diluted in
normal ECS or high K
+
for 30 s Labeling Time to monitor spontaneous and evoked
release respectively. The high K
+
group had dramatically higher biotinylation of
VGLUT1, but not that of other membrane proteins such as TfR and GluR2/3,
indicating that the increase is due to the vesicular release. Omission of calcium from
the high K
+
significantly curbed the increase of VGLUT1-biotinylation, indicating
that depolarization-induced release is calcium-dependent. But interestingly, omission
of calcium did not completely erase the effect of high K
+
. It is possible that some
residual calcium left during the switch of solutions was sufficient to support part of
the release, or the depolarization-induced release has a calcium-independent
component. Pre-treating the neurons with TTX (1 μM, 5 min) only slightly reduced
the increase of VGLUT1 biotinylation induced by high K
+
, indicating that action
potential is not critically involved in this type of evoked release (since the synaptic
terminals were depolarized directly).
47
Sulfo-NHS-LC-Biotin labels proteins through reacting with the primary amines of
these proteins while the neurotransmitter, glutamate or GABA, also contains primary
amines and may compete with proteins for this reagent and reduce its labeling
efficiency. To optimize the biotinylation method, we tested two other types of
biotinylation reagents including Maleimide PEO2-Biotin reacting with sulfhydryl
groups and Biotin-PEO
4
-Hydrazide reacting with carbohydrate. VGLUT1 contains
two cysteine residues and a putative glycosylation site in its putative lumenal
domains (Fremeau RT Jr et al., 2001) and therefore may be able to react with the two
reagents. The biotinylation reagents were diluted in normal ECS (for spontaneous
release) or in high K
+
ECS (for evoked release) at a concentration of 1 mg/ml. The
“Labeling Time” was 30 second for all three reagents. As shown in Fig. 9B,
Maleimide PEO2-Biotin (-SH) showed similar efficiency in labeling spontaneous
release as indicated by the generation of biotin-labeled VGLUT1 to that of Sulfo-
NHS-LC-Biotin (-NH2), while it was much less efficient in tagging TfR and
GluR2/3. Biotin-PEO
4
-Hydrazide (=CO) did not generate detectable signal, possibly
because that the presence of glucose in the ECS quenched the reagent or that
oxidization may be required to create carbonyls for the biotinylation reaction. When
biotinylation was conducted in high K
+
, the Maleimide PEO2-Biotin generated
stronger signals compared with that of Sulfo-NHS-LC-Biotin, suggesting that the
massive release of glutamate in this situation reduced the efficiency of the latter
reagent. Together, these data indicate that the biotinylation reagents targeting the –
SH groups are best suited for measuring presynaptic release while the reagents
48
targeting –NH2 groups provide more information by labeling more membrane
proteins simultaneously. We employed Sulfo-NHS-LC-Biotin for the rest of the
study.

We further tested if the biotinylation method can sensitively detect the modifications
of evoked release. Neurons were treated with PMA (1 μM, 10 min) or CHA (10 μM,
10 min), respectively, and the synaptic release was elicited by incubating the neurons
for 30 s in high K
+
supplemented with Sulfo-NHS-LC-Biotin (Fig. 9C and Fig. 9D).
We found that PMA increased biotinylation of VGLUT1 to 1.43 ± 0.12-fold (t-test, P
< 0.05, n = 4) while it showed no effect on the biotinylation of VGAT (1.07 ± 0.18;
t-test, P > 0.05, n = 4). Treatment with CHA resulted in a significant decrease in the
release of both glutamatergic (0.47 ± 0.02, t-test, P < 0.001, n = 4) and GABAergic
(0.62 ± 0.02, t-test, P < 0.001, n = 4) vesicles. We also used hypertonic ECS
(containing 500 mM sucrose, the total osmolarity was 810 mOsm) to elicit release. It
is known that hypertonic sucrose solution can selectively deplete the readily
releasable pool (Rosenmund C and Stevens CF, 1996). After treatment with PMA (1
μM, 10 min), the biotinylation of VGLUT1 increase significantly (2.04 ± 0.21, t-test,
P < 0.05, n = 4) but not that of VGAT (0.89 ± 0.07, t-test, P > 0.05, n = 5). After
treatment with CHA (10 μM, 10 min), the hypertonic ECS-induced release of both
VGLUT1- (0.81 ± 0.04, t-test, P < 0.05, n = 4) and VGAT- (0.83 ± 0.06, t-test, P <
0.05, n = 6) containing vesicles decreased slightly. Theses results suggest that
49
evoked vesicular release and the readily releasable pool can be estimated with
biotinylation.

Figure 9. Measurement of Evoked Release
(A) Monitoring of vesicular release evoked by depolarization. Neurons were labeled with Sulfo-NHS-
LC-Biotin diluted in normal ECS (spontaneous) or ECS containing high concentration of potassium
(high K
+
) respectively. High K
+
significantly increased the immunoreactivty of biot-VGLUT1 but not
that of biot-TfR or biot-GluR2/3. This effect was reduced by omitting calcium from the ECS (high K
+
w/o Ca
2+
), but not by pretreatment with 1 μM of TTX for 5 minutes (high K
+
w/ TTX). (B)
Comparison of different biotinylation reagents. Sulfo-NHS-LC-Biotin (-NH2, reacting with primary
amines), Biotin-PEO
4
-Hydrazide (=CO, reacting with carbohydrate) and Maleimide PEO2-Biotin (-
SH, reacting with sulfhydryl groups) were used to label spontaneous release or release evoked by high
K
+
. Compared to Sulfo-NHS-LC-Biotin, Maleimide PEO2-Biotin was more efficient in labeling
evoked release but not in labeling spontaneous release and it is less efficient in labeling TfR and
GluR2/3. (C, D, E) Representative blots and quantification showing the facilitation and inhibition of
evoked vesicular release by PMA and CHA, respectively.

2.4.3. DHPG Facilitates Glutamate Release
Having established this biotinylation assay we went on to use it to examine the
presynaptic release in a synaptic plasticity model. We followed well-established
protocols to induce long term depression (LTD) in hippocampal neuronal cultures by
50
using a group 1 mGluR agonist, DHPG. Neurons were treated with DHPG (100 μM)
for 5 min and then DHPG was washed away with ECS (the DHPG treatment
remained the same throughout the current study). After being incubated in DHPG-
free ECS for 45 min, neurons were switched to ECS containing Sulfo-NHS-LC-
Biotin for 5-20 s Labeling Time. As shown in Fig. 10A, the biotinylation of
VGLUT1 but not that of VGAT or TfR increased in the groups treated with DHPG
compared to control groups. We also blotted another synaptic vesicle associated
protein, synaptotagmin 1. Similar to VGLUT1, the biotinylation of synaptotagmin 1
was increased by DHPG. Since in cultured hippocampal neurons there are much
more glutamatergic (~80% of total synapses, Ma L et al., 1999) than GABAergic
synapses the increase of the biotinylation of synaptotagmin 1 further proved that the
increased biotinylation of VGLUT1 was due to enhanced presynaptic release. In
another experiment 45 min after treatment with DHPG, neurons were switched to
Sulfo-NHS-LC-Biotin diluted in normal ECS (for spontaneous release) or high K
+
(for evoked release) for 30 seconds. In the DHPG treated group, the spontaneous
release was increase to 1.72±0.23 fold (P<0.05, t-test, n=5), while the evoked release
was increased to 1.56±0.03 fold (P<0.001, t-test, n=5). Together, these data indicate
that DHPG enhances the presyanptic activity in glutamatergic synapses.
51

Figure 10. DHPG Iinduces Long-Term Facilitation of Glutamate Release
(A) Representative blots showing the effect of DHPG on the spontaneous release of VGLUT1-
containing vesicles. Neurons were treated with 100 µM of DHPG for 5 min followed by incubation in
normal ECS for 45 min. Then the cells were switched into Sulfo-NHS-LC-Biotin diluted in normal
ECS for indicated Labeling Time. DHPG increased the biotinylation of VGLUT1 and synaptotagmin
1, but not that of VGAT and TfR. (B, C) Representative blots and quantification showing the effect of
DHPG on the spontaneous and evoked release of VGLUT1-containing vesicles. Neurons were treated
with DHPG. 45 min later, neurons were switch to Sulfo-NHS-LC-Biotin diluted in normal ECS or
high K
+
for 30 s Labeling Time. DHPG increased the biotinylation of VGLUT1 but not that of TfR in
both situations (* P< 0.05; ** P < 0.001, n=5). (D, E) Representative blots and quantification showing
the effect of DHPG on the size of readily releasable pool. Neurons were treated with DHPG. 30 or 60
min later, neurons were switched to Sulfo-NHS-LC-Biotin diluted in hypertonic ECS for 30 s
Labeling Time. DHPG increased the biotinylation of VGLUT1 and synaptotagmin 1, but not that of
VGAT at both time points (* P < 0.05; ** P < 0.001, n= 4-6).

Next we tested if the DHPG facilitates the presynaptic function through the
augmentation of readily releasable pool as suggested by a previous study on
synaptosome (Lonart G and Sudhof TC, 2000). Neurons were treated with DHPG
and then incubated in DHPG-free ECS for 30 or 60 min, respectively. Then they
were switched to Sulfo-NHS-LC-Biotin diluted in the hypertonic ECS and incubated
for 30 seconds. 30 min after treatment of DHPG, biotinylation of both VGLUT1
52
(1.68 ± 0.16, t-test, P < 0.05, n = 4) and synaptotagmin1 (1.45 ± 0.13, t-test, P < 0.05,
n = 4) increased while biotinylation of VGAT did not change significantly (0.94 ±
0.16, t-test, P > 0.05, n = 4). A similar result was achieved 60 min after DHPG
treatment: biotinylation of VGLUT1 (1.67 ± 0.07, t-test, P < 0.01, n = 4) and
synaptotagmin1 (1.62 ± 0.17, t-test, P < 0.05, n = 4) increased while VGAT was not
significant altered (1.02 ± 0.08, t-test, P > 0.05, n = 4). These results suggested that
DHPG selectively increases the release of glutamate through augmentation of the
readily releasable pool of excitatory synapses.
2.4.4. DHPG Induces LTD
The long-term facilitated glutamate release after DHPG treatment was unexpected,
so we conducted electrophysiological recording to confirm if DHPG induced LTD
under our condition. Cultured hippocampal neurons were treated with DHPG (100
μM) for 5 min. This treatment induced a significant decrease in the frequency of
miniature EPSC and a slight decrease in the amplitude 15 min after DHPG was
washed away (Fig 11A, B, C). Similarly, on the hippocampal slices, DHPG induced
an acute depression of the amplitude of evoked EPSC followed by a long term
depression phase (Fig. 12). These results indicated that LTD indeed occurred under
our conditions and the expression of DHPG-LTD was unlikely due to a presynaptic
mechanism because the presynaptic release had been facilitated. Previous reports
suggested that DHPG-LTD could induce the internalization of ionotropic glutamate
receptors (Snyder EM et al., 2001; Xiao MY et al., 2001; Huang CC et al., 2004,
53
Moult PR et al., 2006, Hsieh H, et al., 2006). We repeated the surface biotinylation
assay of AMPA receptors described in some of these studies and obtained similar
results. The total protein expression levels of GluR1 (0.93±0.05, P>0.05, t-test, n=5)
and GluR2 (0.96±0.09, P>0.05, t-test, n=4) units of AMPA receptors remained
unaltered. But the surface expression of both GluR1 (0.62±0.04, P<0.001, t-test, n=5)
and GluR2 (0.65± 0.06, P<0.01, t-test, n=4) was significantly reduced (Fig 4D, E).
These indicated that the expression of DHPG-induced LTD is due to the
internalization of AMPA receptors at the postsynaptic sites.

Figure 11. DHPG Induces LTD through AMPA Receptor Internalization
(A) Raw traces of mEPSCs recorded from cultured hippocampal neurons before (upper panel) and 30
minutes after (lower panel) DHPG application. Note decrease in the incident of mEPSC after DHPG
application. (B, C) Histograms and commutative plots showed significant decrease in mEPSC
frequency at 30 minutes after DHPG treatment. (D, E) Representative blots and quantification
showing the effect of DHPG on the surface expression of AMAP receptors. Neurons were treated
with DHPG, and 45 min later switched to pre-chilled ECS containing Sulfo-NHS-LC-Biotin and
maintained on ice for 30 min. DHPG did not significantly alter the total protein level of GluR1 and
GluR2, but reduced the surface expression of both subunits (* P < 0.01; ** P < 0.001, n = 4-5).
54
Since previous work demonstrated that the tyrosine phosphorylation sites on the C-
terminus of GluR2 (between the residue 869 to 879) is critical for the clathrin-
dependent AMPA receptor internalization (Ahmadian G et al., 2004) and a peptide
(GluR2
3y
) containing this sequence can competitively block the endocytosis of
AMPA receptors (Wang Y et al., 2004; Brebner K et al., 2005). Here we checked if
GluR2
3y
could also be used to block DHPG-induced depression of synaptic
transmission. LTD was induced in hippocampal slices. Intracellular delivery of the
GluR2
3y
through recording pipette significantly reduced the LTD induced by DHPG
in hippocampal slices but left the acute depression phase unaltered (Fig. 12).
Together, these results indicate that DHPG-induced LTD is dependent on the
internalization of AMPA receptors.

Figure 12. GluR2
3y
Peptide Reduces DHPG-induced LTD in Hippocampal
Slices
EPSCs were recorded in CA1 neurons from hippocampal slices using whole-cell recordings under the
voltage-clamp mode at a holding potential of -60 mV. Normalized EPSCs are plotted from neurons
recorded with pipettes containing standard intracellular solution (control, n=6) or intracellular solution
supplemented with GluR2
3Y
(100 µg/ml; n=7). Time zero is defined as the time point at which the
first EPSC was evoked (typically within 1–2 min of the initiation of whole-cell recording). Evoked
EPSCs obtained from control neurons were significant smaller than that from GluR2
3Y
–loaded
neurons at 30 min after DHPG treatment (P<0.05).
55
The spontaneous release of synaptic vesicles was then measured for the neurons
receiving the same DHPG and GluR2
3y
treatments (Fig 13A, B). The biotinylation of
VGLUT1 was increased by DHPG treatment (1.74 ± 0.15, One-Way ANOVA, P <
0.05 compared with control, n = 6). The increase was significantly attenuated in the
presence of GluR2
3y
although it was not completely blocked (1.26 ± 0.14, One-Way
ANOVA, P < 0.05 compared with DHPG treatment group, P > 0.05 compared with
control, n = 6). GluR2
3y
alone did not change the amount of biotinylation of
VGLUT1 significantly (0.85 ± 0.10, One-Way ANOVA, P > 0.05 compared with
control, n = 6). The biotinylation of VGAT was not affect by either DHPG (1.05 ±
0.15, One-Way ANOVA, P > 0.05 compared with control, n = 6) or GluR2
3y
(0.93 ±
0.15, One-Way ANOVA, P > 0.05 compared with control, n = 6). These results
suggested that the presynaptic facilitation of glutamate release was, at least in part, a
downstream event to the post-synaptic internalization of AMAP receptors.
To test the possibility that the GluR2
3y
peptide complex attenuated the DHPG-
induced enhancement of glutamate release through unknown direct effects in
presynaptic sites, we examined the effect of GluR2
3y
on the PMA-induced
presynaptic facilitation, which was not associated with changes in postsynaptic
AMPA receptor surface expression. As shown in Fig 13C, PMA (1.05 ± 0.26, One-
Way ANOVA, P > 0.05 compared with control, n = 4) or GluR2
3y
alone (0.88 ± 0.19,
One-Way ANOVA, P > 0.05 compared with control, n = 6) did not change the
surface expression of GluR1 significantly. The biotinylation of VGLUT1 increased
56
after PMA treatment (1.71 ± 0.21, One-Way ANOVA, P < 0.05 compared with
control, n = 4). The increase was not significantly changed in the presence of
GluR2
3y
(1.59 ± 0.20, One-Way ANOVA, P > 0.05 compared with DHPG treatment
group; P < 0.05 compared with control, n = 4). The GluR2
3y
alone didn’t change the
amount of biotinylation of VGLUT1 significantly either (0.90 ± 0.04, One-Way
ANOVA, P > 0.05 compared with control, n = 4). These results indicated that the
GluR2
3y
attenuates DHPG-induced presynaptic facilitation through blocking the
postsynaptic internalization of AMPA receptors rather than some direct presynaptic
mechanism.
Previous studies indicate that DHPG can trigger the release of retrograde messengers,
such as endocannabinoids (Chevaleyre V and Castillo PE, 2003; Rouach N and
Nicoll RA, 2003), arachidonic acid and its metabolites (Feinmark SJ et al., 2003).
Here we tested if these messengers took part in DHPG-induced presynaptic
facilitation. Bovine serum albumin (BSA) is known to able to bind to these
molecules and block their action. Hippocampal neurons were incubated with 1%
BSA for 10 minutes followed by treatment with DHPG (in the presence of 1% BSA).
30 min after washing away DHPG (in the presence of BSA in this 30 min incubation
time), the vesicular release was measured. As shown in Fig 5E and 5F, BSA by itself
did not alter the biotinylation of VGLUT1 nor did it change DHPG-induced increase
in the biotinylation of VGLUT1. Together these data indicate that the release of
57
endocannibinoids or arachidonic acid may not be a critical mechanism for the
DHPG-induced facilitation of presynaptic activity.

Figure 13. The Presynaptic Facilitation Is Downstream of AMPAR
Internalization
(A, B) Representative blots and quantification showing that DHPG-induced increase in the
biotinylation of VGLUT1 was reduced by blocking the endocytosis of AMPA receptors. Neurons
were pre-treated with GluR2
3y
peptide complex for 1 hour followed by DHPG (in the presence of the
peptide complex). This peptide blocked the DHPG-induced reduction of AMPA receptor surface
expression and, at the same time, reduced the increase in the biotinylation of VGLUT1. The
biotinylation of VGAT and TfR were not affected by DHPG or the peptide complex. (C, D)
Representative blots and quantification showing the effect of GluR2
3y
peptide complex on PMA-
induced increase in the biotinylation of VGLUT1. Neurons were treated with PMA or GluR2
3y

peptide complex. The peptide complex did not significantly alter the effect of DHPG. (E, F)
Representative blots and quantification showing the effect of BSA on DHPG-induced increase of the
biotinylation of VGLUT1. Neurons were pre-treated with 1% BSA for 10 min followed by DHPG (5
min, in the presence of BSA) and incubated in ECS containing BSA for another 45 min before the
biotinylation assay was conducted. BSA did not alter the effect of DHPG (* P < 0.05; n = 4-6).
58
2.5. Discussion
2.5.1. Assessing Presynaptic Activity
Monitoring the vesicular release from glutamatergic or GABAergic synapses, the
major excitatory and inhibitory synapses in CNS, is technically challenging. The
extracellular concentration of glutamate or GABA in the brain is mainly determined
by the re-uptake by membrane transporters and therefore is not accurately related to
the release. The neutral properties of glutamate and GABA make them unsuitable for
amperometry which is sensitive for direct detection of reductive/oxidative
transmitters. The tiny size of the central synapses also makes it difficult to measure
membrane capacitance of synaptic terminals which is also a direct indicator of
release event (Angleson JK and Betz WJ, 1997). As mentioned before, the analysis
of EPSC is always compromised by postsynaptic modifications while the imaging
techniques lack the ability to recognize the transmitters. Here we provide evidence
that the biotinylation assay can be used to monitor the vesicular turnover of
glutamatergic and GABAergic synapses selectively.
The biotinylation reagents used in current study are small molecules with molecular
weights range from 500 to 600. This is comparable to the size of the commonly used
styryl dyes for labeling of vesicular turnover, such as FM1-43 (molecular weight
610). These reagents were designed to contain a hydrophobic tail and a charged
hydrophilic base (to prevent them from entering the cells). Such a structure also
59
resembles that of the styryl dyes. Therefore, it can be predicted that these reagent can
label the vesicular turnover with efficiency similar to that of styryl dyes.
The selection of vesicular transporters as the targets for labeling enables us to
simultaneously detect the release of different transmitter as it is these transporters
that render the synapses identity (McIntire SL et al., 1997; Takamori S et al., 2000).
To date, three vesicular glutamate transporters have been identified and among them,
VGLUT1 and VGLUT2 are complementarily expressed by excitatory neurons and
are responsible for most of the excitatory transmission (Fremeau RT Jr et al., 2001).
In this study, we chose VGLUT1 because of its predominance in hippocampus. In
addition, VGLUT1 is localized in the terminals of the Schaffer collateral system
from CA3 pyramidal cells, whereas VGluT2 is barely distributed there (Fremeau RT
Jr et al., 2001). This may make the results achieved on cultures more comparable to
the electrophysiological studies conducted in Schaffer collateral pathway in
hippocampal slices. In addition, the biotinylation method allows the simultaneous
detection of post-synaptic proteins, such as AMPA receptors, which is directly
related to the synaptic strength and make it possible to compare pre- and post-
synaptic activity under the same conditions.
As a preliminary work, the biotinylation method has its limitations and needs
improvements. Firstly, this method can only be used in neuronal cultures. The
application in brain slices is compromised by the fact that it takes time for the
biotinylating reagents to diffuse into the inner layers of the slices. The damaged
60
neurons on the surface of slices may also cause problems if the reagents enter those
cells and label the intracellular vesicular transporters. Secondly, this technique can
only measure the population activity and can not be used for observations on
individual synapses. Therefore, it is difficult to tell whether the changes are due to
the increases/decrease in release probabilities of each synapse or to the
increase/decrease of the number of release sites. Thirdly, this technique targets
vesicular turnover instead of transmitters directly. If the amount of transmitters in
each vesicle is somehow altered, this change may not be detected in this method.
Besides this method may not be sensitive enough for certain vesicular release. For
example, some kiss-and-run releases are so fast that even small molecules like
HEPES (MW 121) can not get into the vesicles (Gandhi SP and Stevens CF, 2003).
This type of kiss-and-run events may not be detected in the biotinylation method. In
another situation, if vesicles are reused during the “Labeling Time”, they will be
counted as only one release in this method. To minimize this possibility we limited
the “Labeling Time” to not more than 30 second, in which the reuse may not be a
large component. Lastly, the vesicular proteins may have a low background
expression level on cell membrane, they may also exist on cell membrane during
development of synapses or assembling of synaptic vesicles. The membrane
expression may contaminate the biotinylation assay in measuring the spontaneous
release. Cleavable biotinylation reagents might be helpful in this situation since they
allow erasure of the biotinylation of surface static proteins later on. Despite these
61
shortcomings, this method is a valuable approach in addition to current available
techniques and may provide new insights into mechanisms of synaptic release.
2.5.2. Expression Mechanism of DHPG-induced LTD
In the last three decades, the expression locus (pre- or post-synsptic) of LTP /LTD is
one of the most controversial issues in neuroscience. In recent years some important
progress has been achieved through studying the postsynaptic trafficking of
glutamate receptors: the insertion and removal of AMPA receptors correspond to the
strengthening and weakening of synaptic efficacy respectively (Malinow R and
Malenka RC, 2002; Collingridge GL et al., 2004). The development of new
techniques to directly measure and manipulate the postsynaptic functions contributes
significantly to these progresses.
On the other hand, however, due to a shortage in reliable and quantifiable techniques
to study the presynaptic function directly, understanding of the role of presynaptic
sites in plasticity lags behind. After establishing the sensitivity of the biotinylation
method in measuring presynaptic release we applied it to studying long-term
plasticity. DHPG-induced LTD (DHPG-LTD) was selected for the current study for
several reasons. First, DHPG-LTD can be easily and reliably induced on both
hippocampal culture and slices (Fitzjohn SM et al., 2001; Snyder EM et al., 2001).
Second, it is still not clear how the depression is expressed. Some evidence points to
a presynaptic expression mechanism because mEPSC frequency and success rate is
decreased, while the paired-pulse facilitation ratio is increased (Fitzjohn SM et al.,
62
2001). By contrast, evidence is also accumulating supporting a postsynaptic
mechanism as DHPG induces rapid internalization of AMPA receptors (Snyder EM
et al., 2001; Huang CC et al., 2004, Moult PR et al., 2006).
The results from the biotinylation assay were unexpected: the LTD of excitatory
synaptic transmission was accompanied by a long lasting enhancement of glutamate
release in DHPG-LTD. This facilitating effect of DHPG on glutamate release is
consistent with early studies on synaptosomes (Reid ME, et al., 1999, Lonart G and
Sudhof TC, 2000), indicating that DHPG-LTD is unlikely presynaptically expressed
and should have a post-synaptic explanation. This notion is supported by the
observation of internalization of AMPA receptors and is further strengthened by the
effect of GluR2
3y
which significantly reduced DHPG-induced LTD.
How the internalization of AMPA receptors at postsynaptic sites induced the
presynaptic changes remains to be elucidated. The AMPA receptor internalization
may induce the release of some diffusible retrograde messengers which in turn
potentiate the glutamate release. The endocannabinoids (Chevaleyre V and Castillo
PE, 2003; Rouach N and Nicoll RA, 2003), arachidonic acid and its metabolites
(Feinmark SJ et al., 2003) and NO (Stanton PK et al., 2003) are the candidates
(although our results with BSA did not support this possibility). In addition, AMPA
receptor internalization may modify the function of some trans-synaptic interacting
molecules such as neuroligin-neurexin (Futai K et al., 2007), EphrinB-EphB
63
(Grunwald IC et al., 2004), NCAM (Bozdagi O et al., 2000) and intergrin (Chavis P
and Westbrook G, 2001), etc.
2.5.3. Pre- and Postsynaptic Homeostasis
The pre- and post-synaptic coordination has drawn increasing attention. Studies on
Drosophila (Haghighi AP et al., 2003) and mouse (Sandrock AW Jr et al., 1997)
neuromuscular junction have pointed to a homeostatic interaction between
presynaptic release and postsynaptic activities. In CNS synapses, similar patterns of
regulation have been revealed in development or in plasticity with large timescales
(hours or days) (Turrigiano GG et al., 1998; Burrone J et al., 2002; Turrigiano GG
and Nelson SB, 2004, Thiagarajan TC et al., 2005). Here we showed a homeostatic
relationship between pre- and post-synaptic sites in LTD.
The implication of the presynaptic facilitation in long-term depression is unknown.
In view of the neuronal network, this homeostatic modification will help to maintain
the balance between excitation and inhibition and stabilize the circuits (Turrigiano
GG and Nelson SB, 2004). At the level of individual synapses, it can be postulated
that this facilitation may help to maintain the synaptic transmission. When AMPA
receptors are removed from a synapse, the synapses become silent in terms of EPSC.
But other glutamate receptors, eg. NMDA receptor and mGluRs, may remain on the
postsynaptic sites to detect glutamate and mediate some transmission. Another
postulation is that the enhancement of presyanptic release may be an active
component for maintenance of LTD. One striking property of DHPG-induced LTD
64
is that it can be reversed by mGluR antagonist when the LTD has been established
indicating a persistent activation of mGluR (Schnabel R et al., 1999; Rouach N and
Nicoll RA, 2003). The potentiated glutamate release may therefore contribute to the
constant activation of mGluRs and maintenance of the LTD. It is also possible that
the enhanced synaptic release may signal the formation of new synapses. If
depressed synapses are destined to be eliminated, then the facilitated presynaptic
function may serve as a signaling mechanism for the remaining presynaptic boutons
to find another post-synaptic partner and form a new synapse.

65
Chapter 3. mGluR1 α and Calpain in Excitotoxicity

3.1. Summary
Overactivated glutamatergic synapses under certain conditions leads to a state
referred to as excitotoxicity which eventually induces neuronal degeneration or cell
death. It is well accepted that overactivation of ionotropic glutamate receptors,
especially NMDA receptors, is neurotoxic, however, the roles of metabotropic
glutamate receptors (mGluRs), and especially mGluR1 in these situations remain
controversial. Here we report that the calcium-dependent protease calpain can
mediate truncation of the C-terminal domain of mGluR1α at residue Ser
936
. This
truncation can be induced in vitro by treating neurons with NMDA or in vivo in
animal models of stroke and epilepsy. The truncated mGluR1α maintains its ability
to increase cytosolic calcium while it no longer activates the neuroprotective PI
3
K-
Akt signaling pathways. The full-length and truncated form of mGluR1α play
distinct roles in excitotoxic neuronal degeneration. A fusion peptide derived from the
calpain-cleavage site of mGluR1α efficiently blocks NMDA-induced truncation of
mGluR1α and exhibits neuroprotection against excitotoxicity both in vitro and in
vivo. These findings shed new light on the relationship between NMDA and
mGluR1α and indicate the existence of a positive feed-back regulation in
excitotoxicity involving calpain and mGluR1α.
66
3.2. Background and Introduction
3.2.1. mGluR1 Signaling in Excitotoxicity
Glutamatergic synapses are the major type of excitatory synapses in the brain. Thus,
synaptic transmission and its modulation at these synapses mediate or regulate the
majority of brain functions. Not surprisingly, activity at these synapses is subject to
tight regulation in a normal, healthy brain while their deregulation is associated with
many brain diseases. In particular, these synapses can become overactivated under
certain conditions, leading to a state referred to as excitotoxicity, eventually resulting
in neuronal death. Excitotoxicity is well documented to be a critical mechanism in
ischemic stroke, traumatic brain injury and epilepsy. It is also an important part of
the pathogenesis of several brain diseases such as Alzheimer's disease (AD),
Parkinson’s disease (PD) and amyotrophic lateral sclerosis (see review by Hara and
Snyder 2007).

Under normal conditions, extracellular glutamate concentration is maintained at very
low levels through the activity of high affinity glutamate transporters. However,
under disease conditions, this safe-guarding mechanism is deregulated. For example,
in ischemic stroke, decreased energy supply leads to alterations in ionic gradients
across cell membranes that can reverse the direction of the uptake activity of
glutamate transporters. In addition, alterations in ionic gradients increase neuronal
excitability, resulting in increased neuronal firing and glutamate release. Together,
67
these changes contribute to glutamate accumulation in extracellular space. Increased
glutamate concentrations produce prolonged activation of glutamate receptors. These
receptors include ionotropic glutamate receptors, namely AMPA, NMDA and
kainate receptors, as well as G-protein coupled metabotropic glutamate receptors
(mGluRs). Prolonged NMDA receptor activation has been repeatedly shown to play
a central role in neuronal toxicity. NMDA receptors are permeable to calcium, and
their activation leads to massive calcium influx in neurons. At the same time,
depolarization of neurons can also activate other calcium-permeable channels
including Ca
2+
-permeable acid-sensing ion channels (Xiong ZG and others 2004),
gap junction hemi-channels (Thompson and others 2006) and certain types of AMPA
receptors and further contribute to the elevation of calcium concentration in neurons.

The group I mGluRs (including mGluR1 and mGluR5) are also activated in this
process and their involvement (especially mGluR1) in pathogenesis has also been
proposed in view of the protective or toxic effects of antagonists or agonists.
However, unlike for NMDA receptors, the roles of mGluRs in excitotoxicity appear
to be more complicated. Such complexity resides in the fact that mGluR1s are
coupled with numerous signaling pathways. One signaling pathway consists in the
activation of PLC-ß through the Gq/11 family of G-proteins. PLC activation leads to
IP
3
and DAG production. In turn, IP
3
gates the release of calcium from internal
stores, while DAG activates PKC. mGluR1s can also activate some membrane
channels, such as TRPC1 and GIRK. Furthermore, mGluR1s were also found to
68
activate phosphoinositide 3-kinase (Rong and others 2003), MAP kinases (Thandi
and others 2002), mTOR signaling (Page and others 2006) and increase cAMP level
(Tateyama and Kubo 2006). In addition to the multiple signaling pathways coupled
to mGluR1s, the receptors themselves are also subjected to complex regulations: for
instance, their activity can be directly regulated by the level of intracellular calcium
(Batchelor and Garthwaite 1997), and they can also be desensitized (Rodríguez-
Moreno and others 1998) or remain constitutively activated.

This diversity of signaling mechanism likely accounts for the difficulty of evaluating
the precise role(s) of mGluR1s in excitotoxicity. Activation of group I mGluRs
appears to be neuroprotective under various conditions. The agonist of group I
mGluRs, DHPG, prevented nitric oxide, hydrogen peroxide, or platelet-activating
factor-induced neurotoxicity in neuronal cultures (Vincent AM and Maiese K., 2000;
Zhu P et al., 2004). Activation of mGluRs also protected neurons from oxidative
stress (Sagara Y and Schubert D, 1998). In organotypic hippocampal slice cultures,
mGluR1 activation protected against NMDA-induced excitotoxicity (Blaabjerg M et
al., 2003). Selective mGluRI blockade exacerbated Aß toxicity (Allen JW et al.,
1999). Recent studies indicated that the neuroprotective effects of mGluRI were
mediated by activation of PI3K-Akt signaling through the formation of an mGluRI-
Homer-PIKE-L signaling complex (Rong R et al., 2003). Activation of Akt and
neuroprotection by mGluRI were also reported in other studies (Hou L and Klann E,
2004; Chong ZZ et al., 2006).
69

However, numerous experiments demonstrate neurotoxic effects of mGluRI
activation. In models of cerebral ischemia, activation of mGluRI, especially of
mGluR1, is neurotoxic while antagonists of mGluR1 are neuroprotective. The
neurotoxic effects of mGluR1 activity in ischemia might be due to their effects on
cytosolic free Ca
2+
and their stimulation of glutamate release (Pellegrini-Giampietro
DE, 2003).

Since multiple types of glutamate receptors are simultaneously activated in
excitotoxic conditions, interactions between receptors might play important roles in
determining the outcome of toxic insults. The present study was therefore directed at
evaluating interactions between NMDA and mGluR1α receptors. Our results
indicate that overactivation of NMDA receptors alters signaling mechanisms and
roles in excitotoxicity of mGluR1α through calpain-mediated truncation.
3.2.2. Calpain and Its Substrates

Calpain is a family of proteases which are activated by calcium. Among the 14
calpain isoforms present in the genome, calpain-1 and calpain-2 are ubiquitously
expressed in brain (Goll and others 2003). While calpain-1 and calpain-2 exhibit
different sensitivity to calcium, with calpain-1 requiring micromolar calcium
concentrations and calpain-2 millimolar concentrations, different subcellular
localizations and might be activated in different situations, they share similar
70
sensitivity to calpain inhibitors and similar cleavage sites for their protein substrates.
Therefore, it has been difficult to discriminate the specific contributions of calpain-1
or calpain-2 in various calpain-mediated mechanisms. This difficulty has recently
been further expanded by reports indicating that the sensitivity of calpain to calcium
can be regulated by other signaling pathways, such as MAP kinase (Glading and
others 2004).

It has generally been assumed that localized calpain activation is induced in dendritic
spines by physiological levels of neuronal activity, while massive and generalized
calpain activation is elicited by excitotoxic insults. Using specific calpain-mediated
cleavage of the cytoskeletal protein spectrin as a maker of calpain activation
indicated that calpain is stimulated in different brain structures following epileptic
activity and ischemia (Hong and others 1994; Bi et al., 1996). Calpain can also be
acutely activated by MPTP, a neurotoxin that elicits PD-like symptoms via selective
degeneration of dopaminergic neurons (Crocker and others 2003). Chronic activation
of calpain has also been observed in neurodegenerative diseases such as AD (Lee
and others 2000).

The consequences of calpain activation in brain are critically dependent on the nature
and functions of the proteins this protease cleaves following its activation. In the
absence of a consensus sequence for calpain-mediated cleavage, calpain protein
substrates still need to be identified experimentally. The ever-expanding pool of
71
calpain substrates contains basically all protein categories, including receptors, ion
channels, calcium pumps, cytoskeletal proteins, kinases, phosphatases, transcription
factors and other signaling proteins. Interestingly, most of these proteins have been
found to be cleaved by calpain under pathological conditions, indicating that massive
calpain activation is likely to initiate a cell disruption process. However, the specific
nature of the substrates as well as the temporal sequence for their cleavage under
various experimental conditions remain unknown. This is especially important for
excitotoxic neuronal death since the nature of the proteins that are cleaved at early
stages are more likely to be critical for eliciting neuronal death, while those cleaved
at late stages might simply be a result of cell damaging process. In addition, early
cleaved proteins might be better targets for developing therapeutic approaches.
3.3. Methods and Materials

Materials: Monoclonal anti-mGluR1α
1142-1160
antibody (cat# 610965) was from BD
Pharmingen (San Diego, CA), anti-myc (Cat# ab32) from Abcam (Cambridge, MA),
anti-mGluR5 (cat# 06-451) from Upstate, anti-VGluT1 (Cat# 135 302) from
Synaptic Systems (Goettingen, Germany), anti-spectrin (Cat#1622) and anti-GluR2/3
(Cat# 1506) from Chemicon (Temecula, CA), anti-phospho-Akt (S473) (cat# AF887)
and total Akt (cat# MAB2055) from R&D systems (Minneapolis, MN). Human µ-
calpain was obtained from Sigma (Cat# C6108), calpain inhibitor III (Cat# 208722)
and cpm-VAD-CHO (cat# 218830) from Calbiochem (San Diego, CA), APV,
CNQX, DHPG, LY367385, MPEP from Tocris (Ellisville, Missouri), and NVP-
AAM077 was a gift from Dr. Yves P. Auberson. TAT peptide was from Anaspec
72
(San Jose, CA). TAT-mGluR1 peptide was synthesized by USC/Norris
Comprehensive Cancer Center Core Facilities. The myc-mGluR1α construct was
from Dr. Anna Francesconi. GST-mGluR1α

fusion proteins were made by
subcloning mGluR1α sequences 812-943 or 889-1058 into PEGX 4T-1. The
mGluR1 αS
936
∆ was made by introducing a stop codon right after Ser
936
of myc-
mGluR1 α. The constructs were verified by sequencing.

Cell cultures and transfection: Neurons from E18 rat cortex were dissociated and
plated in Neurobasal medium supplemented with B27, 0.5 mM glutamine and 12.5
µM glutamate in 6-well plates at 10
5
cell/well for immunostaining and in 60-mm
dishes at 1x10
6
cell/dish for immunoblotting. Neurons were switched to maintenance
medium (Neurobasal medium supplemented with B27 and 0.5 mM glutamine) the
next day and fed twice per week until ready for experiments at 14-18 DIV. Neurons
were transfected with a modified calcium phosphate precipitation method (Jiang M
et al., 2004). HEK293 cells were transfected with Lipofectamine 2000 Reagent.

SDS-PAGE and immunoblot: Cultured cells or brain tissues were homogenized in
boiling lysis buffer (1% SDS, 10 mM Tris, 0.2 mM sodium ortho-vanadate; pH 7.4).
The lysate was denatured at 95 ºC for 5 -10 min, sonicated and centrifuged at 14,000
rpm for 30 min. Supernatants containing equal amounts of total proteins were
processed for SDS-PAGE and immunoblotted with standard chemiluminescence
73
protocol. Blots were digitized and quantified with NIH image. All band intensities
were normalized to that of control samples. For immunoprecipitation, cells were
lysed with modified RIPA buffer and lysates incubated with anti-myc or anti-
mGluR1α
1142-1160
antibodies. Proteins were pulled down with protein A agarose.
Silver stain of SDS gel was conducted with EMBL silver stain protocol.

Immunocytochemistry, immunohistochemistry and silver staining: For
immunocytochemistry, neurons were fixed with 4% formaldehyde in PBS (RT, 10
min) and permeabilized with 1% Triton X-100 for 5 min. After blocking with 5%
goat serum (60 min), neurons were incubated in primary antibodies dissolved in
0.5% goat serum in PBS for 1 h. Neurons were then incubated with corresponding
fluorescence-tagged secondary antibody for 30 min. For immunohistochemistry,
mice were deeply anesthetized with Ketamine and Xylazine and intracardially
perfused with 20 ml chilled PBS (pH 7.4) followed by 100 ml 4% formaldehyde.
Brains were removed and post-fixed for 2 h at room temperature. After incubating in
PBS with 30% sucrose until sinking, brains were cut into 30-µm thick sections on a
cryostat. Free-floating brain sections were immunostained with ABC Kit. Brain
section silver staining was conducted with FD NeuroSilver kit1 from FD
NeuroTechnologies (Ellicott City, MD) with the manufacturer provided protocol.

Electrophysiological recordings: HEK293 cells plated on poly-D-lysine coated
coverslips were transfected with mGluR1α (either wild-type or trunctated
74
form)+TRPC1+red fluorescent protein (DsRED), or DsRED alone, using
lipofectamine (Lipofectamine
TM
2000; Invitrogen). For each well of a 24-well plate,
0.5 μg lipofectamine and cDNAs (0.4 μg mGluR1α + 0.4 μg TRPC1+0.1 μg DsRED;
DsRED alone: 0.1 μg) were used for transfection. Electrophysiological recording
was conducted 48 h after transfection. In some experiments, TRPC1 was replaced
with G protein-coupled inwardly rectifying potassium channel cDNAs (GIRK1 and
GIRK2 – 0.4 μg each). HEK293 cells on coverslips were transferred to a recording
chamber continuously perfused with extracellular solution (ECS, pH 7.4) containing
(in mM): 140 NaCl, 5.4 KCl, 1 MgCl
2
, 1.3 CaCl
2
, 25 HEPES, 33 glucose.
Transfected cells were identified from their DsRED signal under a fluorescence
upright microscope. Patch pipettes were pulled from borosilicate glass capillaries
(Sutter Instrument) and filled with an intracellular solution (pH 7.2; 300-310 mOsm),
composed of (in mM): 115 Cs gluconate, 17.5 CsCl, 10 HEPES, 2 MgCl
2
, 10 EGTA,
4 ATP, 0.1 GTP. For GIRK current recordings, the Cs in the intracellular solution
was replaced with K
+
. Recording of whole-cell currents were obtained following
application of DHPG (100 μM) using a fast perfusion system. An Axopatch 200B
amplifier (Axon Instruments) was used for recording. Access resistance was
monitored throughout each experiment. Recordings with a series resistance variation
of more than 10% were rejected. No electronic compensation for series resistance
was used. Whole-cell current recordings were performed in voltage-clamp mode and
the membrane potential was maintained at -60 mV. Recorded currents were blocked
by the mGluR1 antagonist LY367345 (50 μM; Sigma) (data not shown). Recordings
75
were low-pass filtered at 2 kHz, sampled at 10 kHz, and stored in a PC using
Clampex 8.2 (Axon).

Calcium imaging: HEK293 cells plated on poly-D-lysine-coated 16 mm coverslips
were transfected with mGluR1α (either wild-type or truncated form) + DsRED.
Two days after transfection, cells were incubated with a calcium reporter Oregon
green 488 BAPTA-1 AM (0.63% in ECS) for 30 min. After washes with ECS,
intracellular calcium concentration changes in Oregon green loaded cells were
monitored with a cool CCD camera mounted on a Leica microscope under a 63x
objective with the help of the Openlab 3.7.5 software running in a PowerMac
computer. Although all HEK cells were loaded with Oregon green, only 10-20% of
them were mGluR1-transfected so that non-transfected HEK cells in the same visual
field could be used as controls. To induce mGluR1-dependent intracellular calcium
release, cells were perfused with an extracellular solution containing 100 µM DHPG
for 15 sec. Images were acquired continuously before and after DHPG application at
5 Hz for 30 sec and analyzed offline using Image J software.

Cell toxicity assay: Cell live/dead condition of cultured neurons was assessed with
LIVE/DEAD Viability/Cytotoxicity Kit (L-3224, molecular probes). In this kit,
EthD-1 (Ethidium homodimer-1, red fluorescent when binding to DNA) is used to
stain the nucleus of dead cells with disrupted cell membrane integrity; Calcein AM
(become green fluorescent once it is converted by the esterase of live cells) is used to
76
stain live cells. Cultured cortical neurons were incubated with 0.5 μM EthD-1 and/or
2 μM Calcein AM (diluted with neurobasal culture medium) for 10-20 min and
washed with neurobasal medium twice before observation under microscope.

Behavioral assessment of KA-induced seizure activity: Seizure activity was
assessed with methods described by Holcik M et al. (Holcik M et al., 2000) with
some modifications. Mice were assigned a score each 5 min after KA injection
according to a 7-stage scale: 0 represents normal behavior, 1 immobility, 2 rigid
posture, 3 repetitive scratching, circling, or head bobbing, 4 forelimb clonus, rearing,
and falling, 5 repeated episodes of level four behaviors, 6 severe tonic-clonic
behavior and 7 death. The 12 scores obtained each hour following KA injection were
summed to generate the seizure score for that hour. All animal experiments were
conducted in accordance with NIH guidelines and protocols approved by the
Institutional Animal Care and Use Committee with care to minimize distress to the
animals.
3.4. Results
3.4.1. NMDA Receptor Activation Induces Truncation of mGluR1 α
To detect possible mGluR1α modifications following glutamate-induced
excitotoxicity, cultured cortical neurons (14-18 DIV) were incubated with glutamate
(100 µM) for 1 to 60 min. Total cell lysate was collected immediately after treatment
and subjected to SDS-PAGE and immunoblotting with an antibody against the
77
carboxyl terminus (residues 1142-1160) of mGluR1α (Fig. 14Α). Levels of full-
length mGluR1α decreased with increasing incubation time, while levels of a low-
molecular weight band at about 38 kD increased, suggesting that truncation of
mGluR1α occurred at the carboxyl terminus. To probe the amino-terminus of
mGluR1α after truncation, cortical neurons were transfected with an mGluR1α
construct tagged with a myc-epitope at the N-terminus (myc-mGluR1α, the
EQKLISEEDL epitope was inserted in frame after Ala
30
of mGluR1α (Francesconi
A and Duvoisin RM, 2002) ) after 3 days in vitro. Neurons were transfected with a
calcium phosphate precipitation method modified for high transfection rate (Jiang M
et al., 2004). One week later, transfected neurons were treated with 100 µM
glutamate for 1 h and total lysates processed for western blots. When probed with an
anti-myc antibody, blots revealed a low-molecular weight band at about 100 kD after
glutamate treatment (Fig. 14B), which corresponds to the N-terminal fragment after
truncation. The optical density of the original myc-mGluR1α did not decrease
significantly, suggesting that only a small fraction of overexpressed myc-mGluR1α
was cleaved. Since the apparent molecular weight of mGluR1α on SDS-PAGE is
about 140 kD, our data suggested that glutamate induced truncation of mGluR1α in
the carboxyl terminus, most likely at a single cleavage site.

To test for the specificity of glutamate-mediated truncation of mGluR1α,
immunoblots were also probed with an antibody against mGluR5, another member
78
of group I mGluRs, which shares a high similarity with mGluR1α. As shown in Fig.
14C, levels of mGluR5 were not significantly altered by glutamate. Similarly, no
significant changes occurred to the GluR2/3 subunits of AMPA receptors, or to the
vesicular glutamate transporter (VGluT1), a presynaptic protein in glutamatergic
synapses.

Figure 14. Glutamate Induces Carboxyl-terminal Truncation of mGluR1α
(A) Neurons were treated with glutamate for indicated periods of time. Total cell lysates were blotted
with an antibody against the carboxyl terminus of mGluR1α (mGluR1α
1142-1160
). (B) Neurons were
transfected with myc-mGluR1α and treated with glutamate. Total cell lysates were probed with anti-
myc antibody (the schematic shows locations of epitopes, the arrow indicates the putative cleavage
site). (C) Neurons were treated with glutamate and whole lysates blotted with antibodies against
mGluR5, GluR2/3 and VGluT1 respectively.

Selective glutamate receptor antagonists were used to identify the glutamate
receptor(s) involved in mGluR1α truncation. Glutamate (100 µM, 1 h) induced a
significant reduction in the levels of mGluR1α. An NMDA receptor antagonist, MK-
801 (10 µM), completely blocked this effect, whereas the non-NMDA glutamate
receptor antagonist, DNQX (100 µM), had no effect. Similarly, mGluR1 blockade
with LY367385 (100 µM) or mGluR5 blockade with MPEP (100 µM) did not
prevent truncation (Fig1. D, F). Furthermore, when cortical neurons were treated
79
with NMDA (50 µM, 1 h), a similar degree of mGluR1α truncation was produced
(Fig. 15).

Figure 15. NMDA Receptor Induced Truncation of mGluR1α
(A, C) Representative blot and quantification of the effects of an NMDA receptor antagonist on
glutamate-induced truncation of mGluR1α. Neurons were treated with glutamate alone (n=7) or
glutamate plus MK 801 (n=4), DNQX (n=5), LY367385 (n=4) or MPEP (n=4) respectively. Total cell
lysates were blotted with anti-mGluR1α
1142-1160
. (B, C) NMDA induced a similar truncation of
mGluR1α. Neurons were treated with glutamate or NMDA (n=5) and whole lysates blotted with anti-
mGluR1α
1142-1160
. Results are means ± s.e.m. * P < 0.001, student’s t test.

To characterize NMDA-induced mGluR1α truncation, concentration- and time-
dependencies of NMDA effects were studied. Cortical neurons were first incubated
with 1-100 µM of NMDA for a fixed period of time (1 h) (Fig. 16). The minimum
concentration of NMDA required to induce significant mGluR1α truncation was 10
µM (p <0.001, n=4, Student’s t-test for Control vs. 10 µM). In a following
experiment, cortical neurons were incubated with 10 µM of NMDA for 1 to 60 min.
The minimum time required for 10 µM NMDA to induce truncation was 5 min (p
<0.01, n=4, Student’s t-test for Control vs. 5 min). In summary, these data indicate
that activation of NMDA receptors but not non-NMDA glutamate receptors results in
80
mGluR1α truncation and that mGluR1α truncation requires prolonged activation of
NMDA receptors.

Figure 16. Dosage-effect of NMDA-induced Truncation of mGluR1α
Cortical neurons were treated for 1 h with the indicated concentrations of NMDA (G) or were treated
with 10 µM NMDA for the indicated periods of time (H). Results are means ± s.e.m. of 4 experiments.

3.4.2. mGluR1 α Is Truncated by Calpain at Ser
936

We then determined which protease(s) mediated NMDA-induced
mGluR1α truncation. Previous studies have indicated that both the calcium-
dependent neutral protease calpain and caspases could be activated by neurotoxic
concentrations of NMDA. Cortical neurons were pretreated with a calpain inhibitor,
the cell-permeable calpain-inhibitor III (10 µM), or the caspase inhibitor cpm-VAD-
CHO for 2 h followed by 50 µM NMDA for 1 h. Pretreatment with calpain-inhibitor
III significantly blocked NMDA-induced truncation (Fig. 17A), while cpm-VAD-
CHO had no effect. To confirm the activation of calpain by NMDA under our
experimental conditions, we determined the levels of spectrin degradation fragments.
The 145 kD spectrin degradation band (spectrin DB), which is widely used as a
81
marker of calpain activation, could not be detected under control conditions but
appeared after treatment with NMDA (Fig. 17B). In addition, NMDA-induced
spectrin degradation could be blocked by ifenprodil (10 µM), a selective antagonist
of NR2B subunit-containing NMDA receptors, but not by NVP-AAM077 (0.4 µM),
a selective antagonist for NR2A subunit-containing NMDA receptors. Similarly,
NMDA-induced mGluR1α truncation was selectively blocked by ifenprodil but not
by NVP-AAM077. Note that under these conditions, the 120 kD spectrin degradation
fragment generated by caspase-induced spectrin truncation (Newcomb et al. 2000)
was never observed, further strengthening the conclusion that calpain but not caspase
was responsible for mGluR1α truncation.

The involvement of calpain in mGluR1α truncation did not necessarily imply that
calpain could directly cleave the C-terminus of mGluR1α. To confirm that mGluR1α
truncation is directly mediated by calpain, we first transfected HEK293 cells with the
N-terminus myc-epitope tagged mGluR1α. Two days later, mGluR1α was
immunoprecipitated with anti-mGluR1α C-terminus antibody and incubated with
different concentrations of µ-calpain for 30 min. Aliquots of precipitated proteins
were processed for SDS-PAGE and silver staining, and the rest of the samples was
used for western blots with antibodies against myc-epitope or mGluR1α C-terminus,
respectively (Fig. 17C). With silver staining, the density of the 145 kD band that
represents full-length mGluR1α decreased dose-dependently with calpain treatment.
In parallel, two additional bands appeared after calpain treatment with apparent
82
molecular weights of 100 kD and 38 kD, respectively. The 100 KD proved to be the
N-terminus of mGluR1α derived from truncation since it reacted with the anti-myc
antibody. Likewise, the 38 KD protein proved to be the C-terminus of mGluR1α as it
was labeled with the anti-mGluR1α C-terminus antibody. These data indicated that
calpain could directly cleave mGluR1α at the C-terminus.

Figure 17. mGluR1α Is Truncated by Calpain
(A) NMDA-induced truncation of mGluR1α was significantly blocked by calpain inhibitor III (Calpi3)
but not by caspase inhibitor cpm-VAD-CHO (Caspi). (B) Calpain was activated by NMDA. Neurons
were treated as indicated and total cell lysates were blotted with anti-spectrin or anti-mGluR1α
1142-1160,
respectively. Only the degradation bands of spectrin (spectrin DB, at ~145 KD) are shown. NMDA-
induced spectrin degradation and mGluR1α truncation were blocked by ifenprodil but not by NVP-
AAM077. (C) Immunoprecipitated myc-mGluR1α was digested with µ-calpain at the indicated
concentrations (U/ml) for 30 min. The digested material was processed for SDS-PAGE and subjected
to silver staining or immunoblots. With silver staining, myc-mGluR1α (~140 kD) was cleaved by
calpain into two bands of molecular weight of 100 kD (corresponding to the amino-terminal fragment,
as it was detected with anti-myc antibody) and 38 kD (corresponding to the carboxyl-terminal
fragment, as it was detected with anti-mGluR1α
1142-1160)
).

From the size of mGluR1α fragments after cleavage, we deducted that calpain
cleavage site in mGluR1α should be between residue I
812
(the molecular weight of
83
the sequence from I
812
to the C-terminus is 38.25 kD) and S
943
(the molecular weight
from N-terminus to Ser
943
is 105.62 kD). We therefore made a GST fusion protein
containing the sequence of mGluR1 α from I
812
to S
943
. The GST fusion protein was
expressed in and purified from BL21 E. Coli and digested with calpain. This fusion
protein became about 2 kD smaller after digestion, suggesting that the cleavage site
was close to S
943
. To get enough C-terminal fragment for Edman protein sequencing,
another construct was made by fusing the mGluR1α sequence from N
889
to L
1058
to
the C-terminus of GST. As expected, after digestion with calpain, this fusion protein
generated a 10 kD fragment. The N-terminus of this fragment was sequenced to be
YQGS with Edman degradation, indicating that the calpain cleavage site in
mGluR1α is between S
936
and Y
937
(Fig. 18). The molecular weight (MW) of
mGluR1α from residue 1 to 936 is 104.87 kD, which matches well with the apparent
MW of the N-terminal fragment after truncation. The MW of mGluR1α 937-1199 is
28.41 kD, which is lower than the 38 kD apparent MW of the C-terminal fragment.
However, the C-terminal domain of mGluR1α has a high percentage of proline (45
out of the 263 amino acids) with some proline-rich domains composed solely of
prolines. Proline-rich proteins normally show higher apparent MW in SDS-PAGE,
which might explain the seemingly different molecular weights of the C-terminal
fragment. To obtain further confirmation of this truncation site, a stop codon was
introduced into the myc-mGluR1α plasmid immediately after Ser
936
to generate a
construct for truncated mGluR1α (myc-mGluR1αS
936
∆). Transfected in HEK293
cells, this construct generated a protein with the same apparent molecular weight
84
(~100 kD) as the mGluR1α N-terminal fragment generated after calpain-mediated
truncation (Fig. 18).

Figure 18. mGluR1α Is Truncated at Ser
936

(A) Alignment of mGluR1 (902-941) and mGluR5 (889-927) sequences surrounding the calpain
cleavage site and sequence of the TAT-mGluR1 peptide. The calpain cleavage site was identified by
sequencing calpain-digested GST-mGluR1
889-1058
fusion protein. (B) Immunoblot probed with anti-
myc antibody showing that myc-mGluR1αS
936
Δ exhibited the same molecular weight as the amino-
terminal fragment of calpain-cleaved myc-mGluR1α. Lane 1 represents myc-
mGluR1α immunoprecipitated from HEK-293 cells, lanes 2 and 3 myc-mGluR1α digested with
increasing calpain concentrations, and lane 4 myc-mGluR1αS
936
Δ. (C) Point mutations of mGluR1 α
at the cleavage site reduced but did not abolish its sensitivity to calpain. Amino acids at the calpain-
cleavage site, namely P932, L933, T934, K935, S936, Y937 were mutated to alanine respectively.
This result is consistent with previous reports suggesting that calpain recognizes the conformations
rather than the primary sequences of its substrates (Goll DE et al., 2003).
85

To selectively block calpain-mediated truncation of mGluR1α, a peptide was
constructed by fusing the mGluR1α sequence spanning the calpain cleavage site
with the TAT protein transduction domain, a procedure that has been used to transfer
material across cell membranes (Wadia JS et al., 2004). We reasoned that this
peptide would compete with endogenous mGluR1α for calpain and therefore protect
it from truncation. Incubation of brain membrane fractions with calpain resulted in
truncation of both mGluR1α and spectrin. The TAT-mGluR1 fusion peptide dose-
dependently reduced calpain-mediated truncation of mGluR1α but not of spectrin.
The TAT peptide itself showed no significant effect on calpain-mediated truncation
of mGluR1α or spectrin (Fig 19). We then tested the effect of the peptide in cultured
neurons. Cortical cultures were incubated with different concentrations of TAT-
mGluR1 for 90 min followed by 50 µM NMDA for 1 h. As shown in Fig. 19C,
treatment with the peptide resulted in a dose-dependent inhibition of NMDA-induced
mGluR1α truncation At the lower concentrations (1-2 µM), the peptide significantly
reduced mGluR1α truncation with little effects on spectrin degradation, while at
higher concentrations (4-8 µM), spectrin degradation was also blocked.
86

Figure 19. TAT-mGluR1 Peptide Protected mGluR1 α from Truncation
(A) The sequence of TAT-mGluR1 α peptide, which is made by fusing the TAT protein transducing
domain with mGluR1 α sequence spanning the calpain-cleavage site. (B) Brain membranes were
purified as described in Methods and digested with calpain in the absence or presence of TAT-
mGluR1 α or TAT peptides. Digested brain membranes were blotted with anti-mGluR1 α1142-1160
and anti-spectrin respectively. (C) TAT-mGluR1 peptide blocked NMDA-induced mGluR1α
truncation. Neurons were pretreated with the TAT-mGluR1 peptide or vehicle followed by NMDA
and total cell lysates were blotted with anti-mGluR1α
1142-1160
or

anti-spectrin antibodies.

3.4.3. C-terminal Truncated mGluR1 α Remains Functional
The main signaling mechanism activated by mGluR1α consists in PI hydrolysis
through G-protein and phospholipase C, which eventually leads to calcium release
from internal calcium stores. Therefore, we first performed calcium imaging to
determine whether mGluR1α remains functional following C-terminal truncation.
Wild-type mGluR1 (myc-mGluR1α) or the truncated form (myc-mGluR1αS
936
∆)
was cotransfected with DsRED (red fluorescent protein) into HEK293 cells. Two
days after transfection, cells were loaded with a calcium reporter, Oregon green 488
BAPTA-1 AM (0.63% in extracellular solution). As shown in Fig. 20, significant
increase in intracellular calcium concentration could be detected after treatment of
cells transfected with myc-mGluR1α with 100 µM DHPG , a selective agonist for
87
group 1 mGluRs (max. ΔF/F=30.6 ± 4.2%; n=8). This calcium response requires
mGluR1α activation because no change in fluorescent signals could be observed in
non-transfected cells. In cells transfected with myc-mGluR1αS
936
∆, DHPG could
induce qualitatively similar, although significantly smaller, calcium transients (max.
ΔF/F=20.1 ± 2.1%; n=10; p <0.05, Student’s t-test for myc-mGluR1α vs. myc-
mGluR1αS
936
∆). These data indicated that mGluR1α remains functional following
C-terminal truncation at Ser
936
.

Figure 20. Truncated mGluR1α Remains Functional
HEK293 cells were co-transfected with DsRED and myc-mGluR1α or myc-mGluR1αS
936
∆. DHPG
was used to induce mGluR1α-mediated calcium transient. (A) Photos in the left show expression of
DsRED, which labels transfected cells; photos in the middle illustrate the fluorescent signal obtained
with the calcium indicator Oregon green 488 BAPTA-1 AM before and after DHPG application;
photos in the right are obtained by subtracting “before DHPG” from “after DHPG”, thus representing
DHPG-induced calcium transients. (B) Quantification of DHPG-induced calcium transients in myc-
mGluR1α or myc-mGluR1αS
936
∆-transfected cells; results are means ±S.E.M of 8-10 experiments.
88
3.4.4. C-terminal Truncation Alters mGluR1 α Signaling
Activation of mGluR1α can also stimulate non-selective cation excitatory
postsynaptic conductance. We therefore analyzed mGluR1α-dependent currents
before and after calpain-mediated truncation. We transfected HEK293 cells with
either myc-mGluR1α or myc-mGluR1αS
936
∆ plus GIRK1, GIRK2 or TRPC1.
Application of DHPG evoked an inward current in HEK293 cells transfected with
myc-mGluR1α and either TRPC1 or GIRK1,2. Mean amplitude of DHPG-induced
current in myc-mGluR1α/TRPC1-transfected cells was -36.7 ± 9.6 pA (mean ±
s.e.m., n=7 cells), while in myc-mGluR1α/GIRK1,2-transfected cells, average
amplitude of whole-cell current was -24.2 ± 6.3 pA (n=5 cells). In contrast, whole-
cell currents were markedly reduced or absent in cells transfected with myc-
mGluR1αS
936
∆, with a mean amplitude of -9.7 ± 6 .8 pA in myc-
mGluR1αS
936
∆/TRPC1-transfected cells (n=5) and -1.4 ± 0.8 pA in myc-
mGluR1αS
936
∆/GIRK1,2-transfected cells (n=3). As illustrated in Fig. 4A, whole-
cell response induced by wild-type mGluR1α was significantly greater than that
mediated by truncated mGluR1α (p<0.05, t-test for myc-mGluR1α vs. myc-
mGluR1αS
936
∆ in both TRPC1- and GIRK1,2-transfected cells).
89

Figure 21. C-terminal Truncation Alters mGluR1α Signaling
Recordings of mGluR1-induced whole-cell currents in HEK293 cells. Representative current traces
recorded from HEK293 cells transfected with either myc-mGluR1α (left panel) or myc-
mGluR1αS
936
Δ (right panel) and GIRK1, GIRK2 (top) or TRPC1 (bottom). DHPG (100 μM in
extracellular solution) was applied using a fast perfusion system for the period indicated by the scale
bars. Each trace is the average of 5 continuous sweeps. Bar graphs represent means ± S.E.M. of
DHPG-induced current amplitude recorded in cells transfected with myc-mGluR1α or myc-
mGluR1αS
936
Δ plus GIRK1, GIRK2 (top right) or TRPC1 (bottom right) (n=3-7).

As reported in previous studies, mGluR1 activation can stimulate the PI
3
K-Akt
signaling pathway through the mGluRI-Homer-PIEL-PI
3
K signaling complex.
Therefore, we tested whether this signaling mechanism remained functional after
mGluR1α truncation. We first induced mGluR1α truncation by treating cortical
neurons with 50 µM NMDA for 1 h. Two h after washing out NMDA, neurons were
incubated with DHPG (50 µM) for 10 min. Neurons were then lysed and levels of
Akt phosphorylation were determined with western blots by calculating the ratio of
phosphorylated Akt to total Akt. DHPG induced a moderate increase in Akt
phosphorylation levels in neurons (p<0.05, n=6, Student’s t-test for Cont1 vs.
90
DHPG1) (Fig. 4B). Pre-treatment with NMDA reduced basal levels of
phosphorylated Akt (p <0.001, n=6, Student’s t-test for Cont1 vs. cont2). Under this
condition, DHPG failed to increase Akt phosphorylation levels (p=0.27, n=6;
Student’s t-test for cont2 vs. DHPG2). When neurons were pretreated with 8 µM
TAT-mGluR1 for 90 min to block mGluR1α truncation, NMDA-induced reduction
of Akt phosphorylation was partially reversed (p <0.05, n=6, Student’s t-test for
cont3 vs. cont2). More importantly, DHPG-induced increase of Akt phosphorylation
was restored (p <0.05, n=6, Student’s t-test for cont3 vs. DHPG3). Together, the data
indicated that mGluR1α-PI
3
K-Akt signaling pathway was disrupted by calpain-
mediated mGluR1α truncation and that preventing truncation with the TAT-mGluR1
peptide restored this signaling mechanism.

Figure 22. Altered mGluR1-Akt Signaling by Calpain-mediated Truncation
DHPG induced a moderate increase in Akt phosphorylation levels in neurons (p <0.05, Cont1 vs.
DHPG1, n =6). Pre-treatment with NMDA reduced levels of phosphorylated Akt (p <0.001, Cont1 vs.
cont2, n=6) and abolished DHPG effect (p =0.27, cont2 vs. DHPG2, n =6). Pretreatment with TAT-
mGluR1 partially blocked NMDA-induced reduction in levels of phosphorylated Akt (p <0.05, cont3
vs. cont2, n =6) and restored the effect of DHPG in activating Akt (p <0.05, cont3 vs. DHPG3, n =6).
91
3.4.5. Truncation Alters mGluR1 α Targeting
Previous studies have shown that the C-terminus domain of mGluR1α is crucial for
its dendritic localization. Therefore, it was interesting to determine whether
mGluR1α targeting was modified following calpain-mediated truncation. Cortical
neurons were transfected with myc-mGluR1α. After 48 h, neurons were treated with
50 µM of NMDA or vehicle for 1 h. After 3 h of washing out NMDA, neurons were
fixed and stained with antibodies against the N-terminal myc-epitope or the C-
terminus of mGluR1α respectively. Full-length mGluR1α was selectively targeted to
dendrites and almost excluded from axons (Fig. 23A). But in ~ 34% of neurons
treated with NMDA (124 out of 367 neurons counted), myc-immunoreactivity could
be detected in axons, especially in the proximal segment, whereas immunoreactivity
for the C-terminus of mGluR1α was still restricted to dendrites (Fig. 23B). These
results indicated that NMDA-induced C-terminal truncation altered mGluR1α
targeting.

To test whether NMDA-induced translocation of mGluR1α from dendrites to axons
was an active process or passive diffusion after truncation, we transfected cortical
neurons with myc-mGluR1αS
936
∆. Targeting of mGluR1αS
936
∆ 48 h after
transfection was dramatically different from that of full length mGluR1α ( Fig.
24 ). In the majority of transfected neurons (~73%, 500 out of 681 neurons counted),
immunoreactivity for myc-tag was strictly restricted to cell bodies (upper panel). The
92
cell-body restriction was the same when immunostaining was performed 6 days after
transfection, suggesting that it was not the result of a delay in expression or delivery
but was mediated by targeting signals. In a smaller fraction of neurons (~22%, 150
out of 681 neurons counted), mGluR1α was selectively targeted to axons (middle
panel). There were also a few neurons (~4.5%, 31 out of 681 neurons counted) where
immunostaining appeared in dendrites. However, in contrast to the even distribution
of full-length mGluR1α in dendrites, truncated mGluR1α formed large clusters
(bottom panel).

Figure 23. NMDA Treatment Alters mGluR1α Targeting
Neurons were co-transfected with DsRED and myc-mGluR1α. After 48 h, neurons were fixed and
stained with anti-myc (A) or anti-mGluR1α
1142-1160
(B) and anti-DsRED antibodies. Myc-mGluR1α
(upper panel) was not present in axons (arrowhead) under control conditions. After treatment with
NMDA (lower panel), myc-tag staining appeared in the proximal segment of axons (A), while
staining of mGluR1α C-terminus was still absent in axons (B).
93

Figure 24. C-terminal truncation alters mGluR1α targeting
Neurons were co-transfected with DsRED and myc-mGluR1αS
936
Δ. Neurons were stained with anti-
myc and anti-DsRED antibodies 48 h later. In ~73% of neurons, myc-immunoreactivity was restricted
to cell bodies (upper panel; arrowheads point to cell bodies); in ~20% of neurons, myc-
mGluR1αS
936
Δ was preferentially targeted to axons (middle panel; arrowheads point to axon). In ~5%
of neurons, myc-mGluR1αS
936
Δ was preferentially targeted to dendrites and formed large clusters
(lower panel; arrowheads point to clusters in dendrites).
3.4.6. Full-length and Truncated mGluR1 α in Excitotoxicity
Calpain-mediated mGluR1α truncation was induced mainly by toxic concentrations
of NMDA. The downstream signaling pathways of mGluR1α, including intracellular
calcium release and activation of PI
3
K-Akt, are both important for excitotoxicity.
Therefore, we postulated that NMDA-induced truncation would alter the role of
mGluR1α in neuronal toxicity. To test this possibility, we first co-transfected
neurons with green fluorescent protein (GFP) and a control vector, wild-type
mGluR1α or truncated mGluR1α respectively. Two days later, neurons were treated
with 25 µM glutamate for 1 h. Twelve h after treatment, neurons were stained with
94
0.5 µM EthD-1 for 10 min to label nuclei of dead cells. Numbers of GFP-expressing
neurons and EthD-1 positive GFP-expressing neurons on each 18x18 mm coverslip
were counted to calculate the percentage of dead cells. Toxic effects of glutamate
treatment could readily be observed in most GFP-expressing neurons, which
exhibited significant neurite retraction (Fig. 25A,B). In neurons co-transfected with
control vector, 35% (n=9, an average of 405 GFP-expressing neurons were counted
on each coverslip) of GFP-expressing neurons were positive for EthD-1 staining and
therefore had died. In neurons co-transfected with wild-type mGluR1α, only 15% of
neurons (n=8, an average of 428 GFP-expressing neurons were counted on each
coverslip) were EthD-1 positive, an effect that was statistically different from that in
neurons co-transfected with control vector (p<0.01, Student’s t-test). In contrast,
59% of neurons co-transfected with mGluR1αS
936
∆ (n=9, an average of 439 GFP-
expressing neurons were counted on each coverslip) were EthD-1 positive, an effect
significantly higher than observed in neurons co-transfected with control vector
(p<0.05, Student’s t-test). The opposite effects of wild-type and truncated mGluR1α
indicated that wild-type and truncated mGluR1α have distinct roles in excitotoxicity.

We then tested whether calpain-mediated truncation of endogenous mGluR1α also
alters its role in excitotoxicity. We used NMDA to elicit mGluR1α truncation and
evaluated the role of mGluR1 in neuronal toxicity by applying DHPG before and
after truncation. Cortical neurons cultured on 18x18 mm coverslips were divided into
6 groups, which received the following treatments: 1) “Control”, vehicle; 2)
95
“NMDA”: 100 µM NMDA for 1 h; 3) “DHPG before NMDA”: DHPG, 100 µM for 1
h, followed by 100 µM NMDA for 1 h; 4) “DHPG after NMDA”: NMDA, 100 µM
for 1 h, followed by 100 µM DHPG for 1 h; 5) “DHPG+LY before NMDA”: 100 µM
DHPG was co-applied with 100 µM LY367385 for 1 h followed by 100 µM NMDA
for 1 h; 6) “DHPG+LY after NMDA”: NMDA, 100 µM for 1 h, followed by 100 µM
DHPG co-applied with 100 µM LY367385. One day after treatments, all neurons
were stained for 20 min with 0.5 µM EthD-1 to label dead cells and 2 µM Calcein
AM to label live cells. After washes, coverslips were mounted and one microscopy
photo was immediately taken from the center area of each coverslip. EthD-1 and
Calcein AM positive cells on each photo were then counted. The percentage of dead
cells was calculated as the number of EthD-1 positive cell/ (number of EthD-1
positive cell+number of Calcein AM positive cell) *100. As shown in Fig. 25C and
D, “control” had 8 % cell death (n=12). In “NMDA”, cell death rate increased to 37%
(n=12), which was significantly higher than that in “control” ( p<0.001, Student’s t-
test) indicating toxicity of NMDA; 29% cell death was found in “DHPG before
NMDA” (n=9), which was significantly lower than that of “NMDA” (p<0.001,
Student’s t-test) suggesting a neuroprotective effect of DHPG under this condition.
This effect was blocked by co-applying LY367385 (“DHPG+LY before NMDA”,
35.1% cell death, n=11; p<0.05 compared with “DHPG before NMDA”; p=0.38
compared with “NMDA”, Student’s t-test) indicating that it was mediated by
mGluR1. “DHPG after NMDA” had 54% cell death (n=12), which was significantly
higher than that of “NMDA” (p<0.001, Student’s t-test), indicating a neurotoxic
96
effect of DHPG treatment under this condition. This toxic effect was blocked by co-
applying LY367385 (“DHPG+LY after NMDA”, 39.9% cell death, n=9; p<0.001
compared with “DHPG after NMDA”; p=0.27 compared with “NMDA”, Student’s t-
test). The opposite effects of DHPG before and after NMDA-induced truncation thus
demonstrate the distinct roles of full-length and truncated endogenous mGluR1α.

Figure 25. Distinct Roles of Wild-type and Truncated mGluR1α in
Excitotoxicity
(A, B) Cortical neurons were co-transfected with GFP and vector DNA, myc-mGluR1α or myc-
mGluR1αS
936
Δ, respectively. Two days later, neurons were treated with glutamate and after another
12 h stained with EthD-1 to label dead cells. Arrowheads in 7A indicate EthD-1 positive GFP-
expressing neurons. * p <0.05. (C, D) Representative photos and bar graphs showing differential roles
for endogenous mGluR1 in neuronal toxicity before and after NMDA-induced truncation. Cortical
neurons were treated as indicated (see “results” for detailed description) and were stained with EthD-1
and calcein AM (for live cells) 24 h later. * p <0.001.

Based on these data, we reasoned that blockade of mGluR1α truncation might have
neuroprotective effects. To test this possibility, cortical neurons were incubated with
2, 4 or 8 µM TAT-mGluR1 peptide or vehicle for 90 min followed by 100 µM
97
NMDA for 1 h. Neurons were stained with EthD-1 and Calcein AM to determine the
percent dead cells 24 h after NMDA treatment. NMDA treatment significantly
increased the proportion of dead cells (p <0.001, n=12-15, Student’s t-test, NMDA
vs. cont) (Fig. 26D&F). Pretreatment with TAT-mGluR1 peptide dose-dependently
reduced NMDA-induced cell death (p <0.001, n=9-15, Student’s t-test for NMDA vs.
NMAD+TAT-mGluR1 8 µM, NMDA vs. NMAD+TAT-mGluR1 4 µM and NMDA
vs. NMAD+TAT-mGluR1 2 µM respectively). The protective effect of TAT-
mGluR1 was significant at a concentration as low as 2 µM, where the fusion peptide
showed no significant effect on spectrin degradation. These data suggested that
neuroprotection can be achieved by blocking mGluR1α truncation.

Figure 26. TAT-mGluR1 Peptide Protected Neurons from NMDA Toxicity
Neurons were treated with vehicle or TAT-mGluR1 at the indicated concentrations followed by
NMDA. Neurons were stained with EthD-1 and calcein AM 24 h later. * p <0.001. Results are
presented as means ±S.E.M, Student’s t-test (G) Schematic representation of the alterations in
signaling and roles in excitotoxicity of mGluR1α resulting from NMDA receptor-induced calpain-
mediated truncation.
98
3.4.7. Truncation of mGluR1 α in vivo
To test whether calpain-mediated truncation of mGluR1α occurs with excitotoxicity
in vivo, we first used kainic acid-induced seizure activity as a model, since neuronal
damage resulting from seizures has been shown to be NMDA receptor dependent
and calpain activation has also been well documented in this animal model (Siman R
and Noszek JC, 1988). Two-month old male SD rats received systemic injection of
kainic acid (KA) (i.p., 12 mg/kg). Hippocampi were collected 3, 6 or 12 h after
injection, homogenized and processed for immunoblots. As shown in Fig. 27A, the
38 KD C-terminal fragment of mGluR1α was detectable as early as 3 h after KA
injection. At 6 or 12 h after injection, levels of full-length mGluR1α were
significantly reduced. Similarly, the 145 KD spectrin degradation band was present
after KA injection indicating that calpain was activated. A similar result was
obtained in FVB/N mice 12 h after subcutaneous injection of KA (30 mg/kg), (Fig.
29). Compared with that in cultured neurons, immunoreactivity of mGluR1α C-
terminal fragment after in vivo truncation was relatively weaker. The weaker
immunoreactivity of the CT fragment made it possible to determine the location of
mGluR1α truncation in vivo with immunohistochemistry. Immunoreactivity for
mGluR1α in stratum radiatum of hippocampus, especially in CA1, was significantly
reduced following KA injection (Fig. 27B). Interestingly, mGluR1α immunostaining
in stratum oriens was not reduced or even slightly enhanced, suggesting that
mGluR1α truncation might take place mainly in pyramidal cells and not in
99
interneurons located in stratum oriens. GluR2/3 immunoreactivity was not
significantly changed except in CA3 pyramidal cell layer, where immunoreactivity
largely disappeared.

Figure 27. Truncation of mGluR1α in KA-induced Epilepsy
(A) SD rats were injected with kainic acid. Hippocampi were collected at indicated times after
injection and homogenates were blotted with anti-mGluR1α
1142-1160
and anti-spectrin respectively. (B)
FVB/N mice were injected with KA and brains were fixed 18 h after injection. Brain sections were
stained with anti-mGluR1α
1142-1160
and anti-GluR2/3 respectively.

Next we tested an animal model of ischemic stroke (Fig 28). In this in vivo model of
stroke, mice received a unilateral occlusion of common carotid artery combined with
a 45-minute hypoxia (Adhami F et al., 2006). This protocol reliably induces
widespread neuronal death in the occluded side, with no significant damage in the
contralateral side. Twenty-four hours after the onset of ischemia, levels of mGluR1a
were determined with western blotting using an antibody against the carboxyl
terminus of mGluR1; levels of native mGluR1 in the ipsilateral side were decreased
by ~70% as compared to levels found in the contralateral side, and the 37-kDa band
corresponding to calpain-mediated truncated C-terminal domain of mGluR1
appeared in the lesion side.

100

Figure 28. Truncation of mGluR1α in Stroke
(A) Nissl staining showing the neuronal loss in the lesion side. (B) mGLUR1a immunoreactivity is
reduced in the lesion side. (C) Immunoblots showing that activation of calpain and truncation of
mGluR1a in lesion side.

To determine whether in vivo truncation of mGluR1α contributes to neuronal
degeneration, we used the TAT-mGluR1 fusion peptide to block truncation. FVB/N
mice were injected with different concentration of TAT-mGluR1 fusion peptide (i.p,
25, 50 or 100 mg/kg) or vehicle 90 min before systemic KA injection (s.c, 30 mg/kg).
mGluR1α truncation was significantly reduced following treatment with TAT-
mGluR1 fusion peptide (50 and 100 mg/kg) (Fig. 29). As was observed in cultured
neurons, KA-mediated spectrin degradation was also reduced by peptide treatment.
To eliminate the possibility that blockade of calpain activation was due to an effect
of the peptide on seizure susceptibility, we analyzed the effects of the fusion peptide
on intensity and duration of KA-induced seizure activity. FVB/N mice were injected
with 50 mg/kg TAT-mGluR1 peptide or vehicle followed by KA injection 90 min
later. We did not observe any behavioral abnormality in mice that received TAT-
mGluR1 peptide. Peptide pretreatment did not significantly alter KA-induced seizure
101
(p >0.05, n=10-14, Student’s t-test for KA alone vs. KA+TAT-mGluR1 for each
hour, respectively). We then tested whether blockade of mGluR1α truncation could
protect hippocampal neurons from KA-induced neurodegeneration. Since KA-
induced neuron degeneration is closely related to seizure severity, only mice
exhibiting 5- or higher stages of seizures were used for the cell toxicity experiment.
Mice were perfused and fixed 7 days after KA injection. Brains were sectioned and
processed for Nissl and silver staining. Representative photos from KA alone or
peptide-pretreated (50 mg/kg) mice are shown in Fig. 29. As previously reported in
this mouse strain (Schauwecker PE and Steward O, 1997), KA induced severe
neurodegeneration in CA1 and CA3, as illustrated by the reduction of Nissl staining
and appearance of degenerated neurons in silver staining (degenerated neurons were
stained black). Pretreatment with the TAT-mGluR1 peptide significantly reduced
KA-induced neurodegeneration, especially in CA1. Among the 8 mice injected with
KA and vehicle, 5 exhibited severe neurodegeneration in CA1 and CA3, 2 mice
exhibit moderate neurodegeneration, and only 1 mouse appeared relatively normal.
In contrast, among the 8 mice pretreated with TAT-mGluR1 peptide, only 2 showed
moderate neurodegeneration, 1 low level of neurodegeneration while the remaining 5
appeared normal. Together, the data indicate that blockade of mGluR1α truncation
in vivo protects neurons from KA-induced degeneration.
102

Figure 29. Neuroprotection of TAT-mGluR1 in vivo
(A) TAT-mGluR1 blocked truncation of mGluR1α in vivo. FVB/N mice were injected with vehicle
or TAT-mGluR1 at the indicated doses followed by KA injection. Hippocampi were collected 12 h
after KA injection and total homogenates were blotted with anti-mGluR1α
1142-1160
and anti-spectrin
respectively. (B) TAT-mGluR1 did not affect KA-induced seizures. FVB/N mice were injected (i.p)
with vehicle or TAT-mGluR1 followed by KA. Seizure activity was scored as described in Methods.
Results are presented as means ± S.E.M., n=10-14. (C) Representative photos showing the
neuroprotective effect of TAT-mGluR1 in vivo. Mice were first treated with KA and sacrificed 7 days
later. The brains were sectioned and processed for Nissl and silver staining.

3.5. Discussion
3.5.1. Regulation of mGluR1 α by Calpain-mediated Truncation
In central nervous system, calpain is important for neural development, synaptic
plasticity and neural degeneration. Our laboratory previously demonstrated calpain-
mediated truncation of both AMPA (Bi X et al., 1996) and NMDA receptors (Bi X et
103
al., 1998). The present study demonstrates that calpain also truncates the C-terminal
domain of mGluR1α. mGluR1α signaling mechanism involves the Gq family of
heterotrimeric

G-proteins. The major signal transduction mechanism for mGluR1
consists in activation of phospholipase C (PLC), which hydrolyses membrane
phosphoinositides and leads to IP
3
-mediated Ca
2+
release from intracellular stores
(Masu M et al., 1991). The mGluR1α-G protein coupling involves mainly the second
intracellular loop of the receptor (Gomeza J et al., 1996); therefore it is not surprising
that mGluR1α remains functional after calpain-mediated truncation, which removes
a long fragment from the intracellular C-terminus domain. Besides interacting with
G-proteins, mGluR1α also directly interacts via its large intracellular C-terminal
domain with several other proteins, including Homer family proteins (Fagni L et al.,
2002), calmodulin and Siah1A (Ishikawa K et al., 1999) and membrane channels
(Saugstad JA et al., 1996; Kitano J et al., 2003; Kim SJ et al., 2003). The Homer-
binding motif is located in the distal C-terminus of mGluR1α and downstream of the
calpain cleavage site. The binding of the receptor to Homer proteins is essential for
coupling mGluR1α with the PI
3
K-Akt system (Rong R et al., 2003). Homer proteins
also serve as a bridge linking mGluR1α in cell plasma membranes with IP
3
receptors
in ER membrane and facilitate mGluR-dependent calcium release from intracellular
stores (Fagni L et al., 2002). As expected, mGluR1-PI
3
K-Akt signaling was
disrupted by calpain-mediated truncation. The mGluR1α-induced calcium transient
was also moderately reduced. mGluR1α also gates some membrane channels, such
as TRPC1 (Kim SJ et al., 2003) and GIRK1,2 (Saugstad JA et al., 1996), which
104
transduce mGluR1α-dependent EPSC. The detailed mechanism responsible for
mGluR1α/membrane channel interactions is largely unknown, and our data indicated
that the C-terminal sequence located downstream of calpain cleavage site is critical
for those interactions.

Using immunogold electromicroscopy, Lujan et al. located mGluR1α to the peri-
synaptic area of postsynaptic sites (Lujan R et al., 1996). Consistent with this finding,
mGluR1α was found to be selectively targeted to dendrites in cultured neurons
(Stowell JN and Craig AM, 1999). A detailed analysis of mGluR1α targeting
revealed two distinct targeting signals in the C-terminus domain, RRK
877-879
for
axonal targeting and another one located in more distal sequences. It has been
suggested that the axonal targeting signal is masked by the proximal C-terminus
(1012-1071) of mGluR1α, resulting in selective delivery of mGluR1α to dendrites
(Francesconi A and Duvoisin RM, 2002). In agreement with this idea, we found that
mGluR1αS
936
Δ was preferentially targeted to axons in about 20% of neurons. But
unlike the 100% axonal targeting of the short mGluR1 splice variant mGluR1b,
mGluR1αS
936
Δ was strictly restricted to cell bodies in the majority of neurons. This
finding suggests that, although mGluR1αS
936
Δ is very similar in length to mGluR1b
and mGluR1c, it might have different targeting and signaling properties. What
governs the differential targeting of mGluR1αS
936
Δ (cell body or axons) in
individual neurons is unknown, but such targeting might be due to specific
105
characteristics of neurons (eg. interneurons vs. pyramidal neurons, or neurons
derived from different layers or different regions of the cortex).

3.5.2. NMDA Receptor/mGluR1 Interactions in Excitotoxicity

NMDA receptors and mGluR1 are colocalized on the postsynaptic membrane and
co-activated by glutamate release from presynaptic sites. Functional interactions
between these two types of receptors have long been noticed (Lan JY et al., 2001).
Physical interactions between the two receptors have also been proposed. It was
suggested that mGluR1 could be linked to NMDA receptors through the Homer-
Shank-PSD95 complex (Tu JC et al., 1999) or by Ephrin (Calo L et al., 2005). These
physical interactions would keep mGluR1 close to NMDA receptor channels and
target it for cleavage by locally activated calpain.

Interactions between NMDA and mGluR1 receptors in excitotoxicity appear
complicated. Activation of mGluR1 was reported to attenuate NMDA-induced
neurotoxicity in cortical neuronal cultures (Koh JY et al., 1991) and in organotypic
hippocampal slice cultures (Baskys A et al., 2005). Similarly, pretreatment with
mGluRI agonists reduced the excitotoxic effects induced in the retina by intraocular
NMDA injection (Siliprandi R et al., 1992). However, many studies indicate that
mGluR1 activation exacerbates NMDA receptor-mediated neurotoxicity (Calo L et
al., 2005, Bruno V et al., 1995). Interestingly, activation of mGluR1 before NMDA
application tends to be neuroprotective (Blaabjerg M et al., 2003), while activation of
106
mGluR1 after NMDA application tends to enhance NMDA toxicity (Bruno V et al.,
1995). Similarly, the protective effects of mGluRI agonists in ischemia are time-
dependent: they are neuroprotective only when administered before the onset of
ischemia (Schroder UH et al., 1999).

Our findings regarding NMDA-induced mGluR1α truncation provide a possible
explanation for these phenomena: before NMDA application or onset of ischemia,
mGluR1α receptors are coupled to the PI3K-Akt signaling and their activation is
neuroprotective. Although mGluR1α activation leads to calcium release from
internal stores, the extent of calcium release might be too low and transient to
produce significant toxic effects. Following NMDA application or onset of ischemia,
NMDA receptor activation would induce calpain-mediated truncation of mGluR1α.
As a result, the neuroprotective effect of the mGluR1α-PI3K-Akt signaling cascade
would be disrupted. In addition, mGluR1α-dependent calcium release from
intracellular stores would further contribute to calcium overload due to calcium
influx through NMDA receptors and thus enhance neurotoxicity. Finally, since
mGluR1 also exhibits presynaptic localization and function (Herrero I et al., 1998;
Moroni F et al., 1998), truncation-mediated mGluR1α translocation to axons might
further enhance glutamate release, thereby exacerbating excitotoxicity.

Accumulating evidence suggests that calpain also plays a critical role in
excitotoxicity. Overactivation of NMDA receptors leads to calpain activation and
107
calpain blockade protects neurons from excitotoxicity (Higuchi M et al., 2005). By
cleaving different substrates, calpain triggers multiple neurotoxicity mechanisms.
First, by partial truncation of NMDA receptors (Simpkins KL et al., 2003) and
Na+/Ca exchanger (Bano D et al., 2005), calpain enhances calcium overload; second,
through degradation of intracellular anti-apoptotic proteins, eg. NF-κB (Scholzke
MN et al., 2003), calpain promotes cell death signaling. Third, by degrading
cytoskeleton proteins, such as spectrin, calpain disrupts cell structure. Here we
showed the cleavage of mGluR1 by calpain. This cleavge disrupts the
neuroprotective signaling of mGluR1 but maintains its function in increasing
calcium release from internal stores. Without the balance from the protective
signaling, internal calcium release combined with calcium influx from activation of
NMDA receptor and/or other calcium channels on cell surface further activate
calpain. In addition, mGluR1 is shown to be able to activate MAP kinase, which
might directly activate calpain (Glading A, et al., 2004). Furthermore, increased
intracellular calcium facilitates mGluR1-mediated release of calcium from internal
stores (Batchelor AM and Garthwaite J, 1997). Therefore a positive feedback loop is
established that has the potential to produce uncontrollable and massive calpain
activation and ultimately neuronal death. In the current study we showed that this
feedback loop may play critical roles in stroke and epilepsy. In the future, it will be
interesting to test the involvement of this positive feedback loop in other models of
neurodegeneration where excitotoxicity has been implicated, such as Parkinson's
disease and Alzheimer’s disease. Moreover, further optimization of the TAT-
108
mGluR1 peptide might provide a very useful tool not only to test this mechanism in
these diseases, but also to develop new therapeutic treatments.

Figure 30. NMDA Receptor, Calpain and mGluR1 in Excitotoxicity
Activation of NMDA receptors produces calcium influx (1), while activation of mGluR1 induces the
release of calcium from internal store (2). Both events result in the accumulation of intracellular
calcium. The accumulation of calcium in turn activates calpain (3) which disrupts the neuroprotective
mGluR1-PI3K signaling that depends on the interaction between the carboxyl terminus of mGluR1
and Homer family proteins. The activation of MAP kinase by mGluR1 may facilitate calpain activity
(4) and the intracellular calcium may also facilitate mGluR1 activity (5). Together, these interactions
constitute a positive feedback loop leading to excessive activation of calpain and cell toxicity.
109
Summary and Conclusions

In my dissertation studies I attempted to find general features regarding the
regulation of synaptic transmission by analyzing several characteristics of
glutamatergic synapses. I first focused on evaluating potential interactions between
VGLUT1 and EAATs in order to reveal new regulatory principles between
transmitter release and transmitter reuptake. I then determined potential interactions
between two types of glutamate receptors, an ionotropic receptor, the NMDA
receptor, and a metabotropic receptor, the mGluR1, and the role of this interaction in
excitotoxicity. Finally, I analyzed potential homeostatic interaction between pre- and
post-synaptic components in synaptic plasticity.

I first showed that VGLUT1 physically interacts with EAATs. This interaction
occurs between intracellular vesicles and cell membranes and in both neurons and
glial cells. The expression of VGLUT1 does not alter glutamate uptake activity of
EAATs, but EAATs alter the trafficking properties of VGLUT1-containing vesicles.
This interaction represents a rare example of protein-protein interactions between
vesicles and cell membranes distinct from the SNARE (soluble N-ethylmaleimide-
sensitive-factor attachment protein receptor) proteins, which are the major proteins
bridging vesicular and cell membranes and regulating vesicular trafficking. The
finding that VGLUT1 interacts with EAAT2 in synaptosome preparations lends
support to the notion that EAATs exist also in the pre-synaptic terminals and play
110
important roles there. Similarly, the existence of VGLUT1-EAAT2 interactions in
Glial Plasmalemmal Vesicle (GPV) preparations supports the recent observations
that glial cells also contain the machinery for vesicular release of glutamate as a
signal molecule in a way similar to synaptic terminals. The observations that the
interaction occurs between only VGLUT1, but not VGLUT2, and EAATs, and that
EAATs alter the trafficking property of VGLUT1-containing vesicles suggest that
pre-synaptic EAATs may regulate the release or recycling of VGLUT1-containg
synaptic vesicles and therefore provide a potential explanation for the different
release probabilities observed between VGLUT1- and VGLUT2- containing synaptic
terminals. Based on the model of VGLUT1-EAATs interactions, we propose that a
transporter complex may be formed to directly uptake glutamate from extracellular
space into synaptic vesicles, a mechanism that would increase the efficiency for
recycling glutamate and save energy for glutamate/glutamine cycle. This model is
consistent with recent findings that excitatory transmission persists upon disruption
of the glutamate/glutamine cycle. In addition, the special functions of these two
groups of proteins, one located at the very beginning of synaptic transmission
involved in storing glutamate into synaptic vesicles and the other at the very end of
excitatory transmission by recapturing glutamate inside cells, lead us to hypothesize
that VGLUT1-EAATs interactions might provide a mechanism for coupling and
regulating release and uptake of glutamate in synaptic transmission. We also
narrowed down the interaction site on VGLUT1 to a small region of its N-terminal
domain. Further identification of the exact interaction site will provide tools for
111
specifically perturbing VGLUT1-EAAT interactions and offer conclusive answers to
the above hypotheses. Together, our investigation of VGLUT1-EAATs interaction
provides a valuable clue for further studies to elucidate some remaining questions
related to the release and recycle of glutamate.

We then showed that calcium influx through NMDA receptor/channels under
excitotoxic conditions activated the calcium-dependent protease calpain, which in
turn, truncated metabotropic glutamate receptor mGluR1a in its C-terminal domain.
This truncation altered mGluR1a signal transduction by disrupting its function of
activating the neuroprotetive PI3K-Akt pathway while maintaining its ability to
trigger calcium release from internal stores. This truncation thus switched the role of
mGluR1a from “neuroprotective” to “neurotoxic”. The functional interactions
between NMDA receptors, calpain and mGluR1a in this situation therefore constitute
a positive feedback loop which leads to neuronal degeneration and death. Calpain-
mediated truncation of mGluR1a is likely to be induced under various disease
conditions, such as epilepsy and stroke (in Preliminary Experiments, we did observe
such truncation in both in vitro and in vivo stroke models). We showed that
intervention in this process directed at disrupting this positive feedback loop
provided neuroprotection, at least in the case of excitotoxicity. This study indicates
that glutamate receptors contribute to cell toxicity not only through their individual
activities, but also through their functional interactions. It also suggests that these
functional interactions could represent new and interesting targets for designing
112
therapeutic approaches. Furthermore, we showed that calpain-mediated truncation of
mGluR1a altered its subcellular location from exclusive dendritic location to both
dendrites and axons. This observation suggests that alterations in subcellular
locations may represent an important mechanism for regulating the functions of
receptors and other proteins, especially in highly polarized cells like neurons. In
addition, mGluR1a is the first G-protein coupled receptor to be identified as a
substrate of calpain. This result suggests that calpain- or other protease-mediated
proteolytic processing might be an important mechanism for the regulation of the
signal transduction of G-protein coupled receptors.

We developed a novel biochemical approach to study pre-synaptic vesicular release
by targeting synapse-specific vesicular transporters. Probing activity at tiny synaptic
terminals is particularly challenging and the task becomes even more difficult in
order to study different terminals (e.g. excitatory vs. inhibitory) at the same time.
This new method provides a solution to this problem. In addition, it allows
simultaneously monitoring and comparing both pre-synaptic and post-synaptic
functions. Using this technique, we demonstrated the existence of a homeostatic
relationship between pre-synaptic and post-synaptic activity in a particular model of
synaptic plasticity, and showed the critical roles of post-synaptic trafficking of
glutamate receptors in this process. This homeostatic modification at presynaptic
sites lends further evidence to the notion that the primary site for the expression of
long-term synaptic plasticity is at the postsynaptic site. In addition, the
113
corresponding presynaptic modification clearly illustrates the existence of a close
relationship between the activities of the pre- and post-synaptic sites. Combined with
previous studies demonstrating homeostatic regulations in neuronal circuits and
those showing homeostatic regulations at synapses on a longer time scale (hours to
days), our studies indicate that homeostasis may represent a general rule governing
plastic regulations in the brain both spatially at the synaptic, neuronal and, circuit
levels as well as possibly the whole brain level and temporally from seconds up to
days.

Together, our studies indicate that interactions and regulatory mechanisms between
glutamate transporters, glutamate receptors and between pre- and post-synaptic sites
are essential components of normal synaptic transmission, synaptic plasticity and
could play critical roles in the development of brain diseases.
114
References

Adhami, F., Liao, G., Morozov, Y.M., Schloemer, A., Schmithorst, V.J., Lorenz,
J.N., Dunn, R.S., Vorhees, C.V., Wills-Karp, M., Degen, J.L., et al. (2006). Cerebral
ischemia-hypoxia induces intravascular coagulation and autophagy. Am J Pathol 169,
566-583.

Ahmadian, G., Ju, W., Liu, L., Wyszynski, M., Lee, S.H., Dunah, A.W., Taghibiglou,
C., Wang, Y., Lu, J., Wong, T.P., et al. (2004). Tyrosine phosphorylation of GluR2
is required for insulin-stimulated AMPA receptor endocytosis and LTD. EMBO J 23,
1040-1050.

Allen, J.W., Eldadah, B.A., and Faden, A.I. (1999). Beta-amyloid-induced apoptosis
of cerebellar granule cells and cortical neurons: exacerbation by selective inhibition
of group I metabotropic glutamate receptors. Neuropharmacology 38, 1243-1252.

Angleson, J.K., and Betz, W.J. (1997). Monitoring secretion in real time: capacitance,
amperometry and fluorescence compared. Trends Neurosci 20, 281-287.

Aoyama, K., Suh, S.W., Hamby, A.M., Liu, J., Chan, W.Y., Chen, Y., and Swanson,
R.A. (2006). Neuronal glutathione deficiency and age-dependent neurodegeneration
in the EAAC1 deficient mouse. Nat Neurosci 9, 119-126.

Bano, D., Young, K.W., Guerin, C.J., Lefeuvre, R., Rothwell, N.J., Naldini, L.,
Rizzuto, R., Carafoli, E., and Nicotera, P. (2005). Cleavage of the plasma membrane
Na+/Ca2+ exchanger in excitotoxicity. Cell 120, 275-285.

Baskys, A., Bayazitov, I., Fang, L., Blaabjerg, M., Poulsen, F.R., and Zimmer, J.
(2005). Group I metabotropic glutamate receptors reduce excitotoxic injury and may
facilitate neurogenesis. Neuropharmacology 49 Suppl 1, 146-156.

Batchelor, A.M., and Garthwaite, J. (1997). Frequency detection and temporally
dispersed synaptic signal association through a metabotropic glutamate receptor
pathway. Nature 385, 74-77.

Bellocchio, E.E., Hu, H., Pohorille, A., Chan, J., Pickel, V.M., and Edwards, R.H.
(1998). The localization of the brain-specific inorganic phosphate transporter
suggests a specific presynaptic role in glutamatergic transmission. J Neurosci 18,
8648-8659.

Bezzi, P., Gundersen, V., Galbete, J.L., Seifert, G., Steinhauser, C., Pilati, E., and
Volterra, A. (2004). Astrocytes contain a vesicular compartment that is competent
for regulated exocytosis of glutamate. Nat Neurosci 7, 613-620.
115

Bi, X., Chang, V., Siman, R., Tocco, G., and Baudry, M. (1996). Regional
distribution and time-course of calpain activation following kainate-induced seizure
activity in adult rat brain. Brain Res 726, 98-108.

Blaabjerg, M., Fang, L., Zimmer, J., and Baskys, A. (2003). Neuroprotection against
NMDA excitotoxicity by group I metabotropic glutamate receptors is associated with
reduction of NMDA stimulated currents. Exp Neurol 183, 573-580.

Bliss, T.V., and Collingridge, G.L. (1993). A synaptic model of memory: long-term
potentiation in the hippocampus. Nature 361, 31-39.

Bonvento, G., Herard, A.S., and Voutsinos-Porche, B. (2005). The astrocyte--neuron
lactate shuttle: a debated but still valuable hypothesis for brain imaging. J Cereb
Blood Flow Metab 25, 1394-1399.

Bozdagi, O., Shan, W., Tanaka, H., Benson, D.L., and Huntley, G.W. (2000).
Increasing numbers of synaptic puncta during late-phase LTP: N-cadherin is
synthesized, recruited to synaptic sites, and required for potentiation. Neuron 28,
245-259.

Brasnjo, G., and Otis, T.S. (2001). Neuronal glutamate transporters control activation
of postsynaptic metabotropic glutamate receptors and influence cerebellar long-term
depression. Neuron 31, 607-616.

Brebner, K., Wong, T.P., Liu, L., Liu, Y., Campsall, P., Gray, S., Phelps, L., Phillips,
A.G., and Wang, Y.T. (2005). Nucleus accumbens long-term depression and the
expression of behavioral sensitization. Science 310, 1340-1343.

Bruno, V., Copani, A., Knopfel, T., Kuhn, R., Casabona, G., Dell'Albani, P.,
Condorelli, D.F., and Nicoletti, F. (1995). Activation of metabotropic glutamate
receptors coupled to inositol phospholipid hydrolysis amplifies NMDA-induced
neuronal degeneration in cultured cortical cells. Neuropharmacology 34, 1089-1098.

Burrone, J., O'Byrne, M., and Murthy, V.N. (2002). Multiple forms of synaptic
plasticity triggered by selective suppression of activity in individual neurons. Nature
420, 414-418.

Calo, L., Bruno, V., Spinsanti, P., Molinari, G., Korkhov, V., Esposito, Z., Patane,
M., Melchiorri, D., Freissmuth, M., and Nicoletti, F. (2005). Interactions between
ephrin-B and metabotropic glutamate 1 receptors in brain tissue and cultured neurons.
J Neurosci 25, 2245-2254.

116
Carlson, M.D., Kish, P.E., and Ueda, T. (1989). Characterization of the solubilized
and reconstituted ATP-dependent vesicular glutamate uptake system. J Biol Chem
264, 7369-7376.

Chavis, P., and Westbrook, G. (2001). Integrins mediate functional pre- and
postsynaptic maturation at a hippocampal synapse. Nature 411, 317-321.

Chen, W., Mahadomrongkul, V., Berger, U.V., Bassan, M., DeSilva, T., Tanaka, K.,
Irwin, N., Aoki, C., and Rosenberg, P.A. (2004). The glutamate transporter GLT1a is
expressed in excitatory axon terminals of mature hippocampal neurons. J Neurosci
24, 1136-1148.

Chevaleyre, V., and Castillo, P.E. (2003). Heterosynaptic LTD of hippocampal
GABAergic synapses: a novel role of endocannabinoids in regulating excitability.
Neuron 38, 461-472.

Chong, Z.Z., Li, F., and Maiese, K. (2006). Group I metabotropic receptor
neuroprotection requires Akt and its substrates that govern FOXO3a, Bim, and beta-
catenin during oxidative stress. Curr Neurovasc Res 3, 107-117.

Collingridge, G.L., Isaac, J.T., and Wang, Y.T. (2004). Receptor trafficking and
synaptic plasticity. Nat Rev Neurosci 5, 952-962.

Crocker, S.J., Smith, P.D., Jackson-Lewis, V., Lamba, W.R., Hayley, S.P., Grimm,
E., Callaghan, S.M., Slack, R.S., Melloni, E., Przedborski, S., et al. (2003).
Inhibition of calpains prevents neuronal and behavioral deficits in an MPTP mouse
model of Parkinson's disease. J Neurosci 23, 4081-4091.

Danbolt, N.C. (2001). Glutamate uptake. Prog Neurobiol 65, 1-105.

Duan, S., Anderson, C.M., Stein, B.A., and Swanson, R.A. (1999). Glutamate
induces rapid upregulation of astrocyte glutamate transport and cell-surface
expression of GLAST. J Neurosci 19, 10193-10200.

Fagni, L., Worley, P.F., and Ango, F. (2002). Homer as both a scaffold and
transduction molecule. Sci STKE 2002, RE8.

Feinmark, S.J., Begum, R., Tsvetkov, E., Goussakov, I., Funk, C.D., Siegelbaum,
S.A., and Bolshakov, V.Y. (2003). 12-lipoxygenase metabolites of arachidonic acid
mediate metabotropic glutamate receptor-dependent long-term depression at
hippocampal CA3-CA1 synapses. J Neurosci 23, 11427-11435.

117
Fitzjohn, S.M., Palmer, M.J., May, J.E., Neeson, A., Morris, S.A., and Collingridge,
G.L. (2001). A characterisation of long-term depression induced by metabotropic
glutamate receptor activation in the rat hippocampus in vitro. J Physiol 537, 421-430.

Francesconi, A., and Duvoisin, R.M. (2002). Alternative splicing unmasks dendritic
and axonal targeting signals in metabotropic glutamate receptor 1. J Neurosci 22,
2196-2205.

Fremeau, R.T., Jr., Burman, J., Qureshi, T., Tran, C.H., Proctor, J., Johnson, J.,
Zhang, H., Sulzer, D., Copenhagen, D.R., Storm-Mathisen, J., et al. (2002). The
identification of vesicular glutamate transporter 3 suggests novel modes of signaling
by glutamate. Proc Natl Acad Sci U S A 99, 14488-14493.

Fremeau, R.T., Jr., Kam, K., Qureshi, T., Johnson, J., Copenhagen, D.R., Storm-
Mathisen, J., Chaudhry, F.A., Nicoll, R.A., and Edwards, R.H. (2004). Vesicular
glutamate transporters 1 and 2 target to functionally distinct synaptic release sites.
Science 304, 1815-1819.

Fremeau, R.T., Jr., Troyer, M.D., Pahner, I., Nygaard, G.O., Tran, C.H., Reimer, R.J.,
Bellocchio, E.E., Fortin, D., Storm-Mathisen, J., and Edwards, R.H. (2001). The
expression of vesicular glutamate transporters defines two classes of excitatory
synapse. Neuron 31, 247-260.

Fremeau, R.T., Jr., Voglmaier, S., Seal, R.P., and Edwards, R.H. (2004). VGLUTs
define subsets of excitatory neurons and suggest novel roles for glutamate. Trends
Neurosci 27, 98-103.

Futai, K., Kim, M.J., Hashikawa, T., Scheiffele, P., Sheng, M., and Hayashi, Y.
(2007). Retrograde modulation of presynaptic release probability through signaling
mediated by PSD-95-neuroligin. Nat Neurosci 10, 186-195.

Gandhi, S.P., and Stevens, C.F. (2003). Three modes of synaptic vesicular recycling
revealed by single-vesicle imaging. Nature 423, 607-613.

Glading, A., Bodnar, R.J., Reynolds, I.J., Shiraha, H., Satish, L., Potter, D.A., Blair,
H.C., and Wells, A. (2004). Epidermal growth factor activates m-calpain (calpain II),
at least in part, by extracellular signal-regulated kinase-mediated phosphorylation.
Mol Cell Biol 24, 2499-2512.

Goll, D.E., Thompson, V.F., Li, H., Wei, W., and Cong, J. (2003). The calpain
system. Physiol Rev 83, 731-801.

Gomeza, J., Joly, C., Kuhn, R., Knopfel, T., Bockaert, J., and Pin, J.P. (1996). The
second intracellular loop of metabotropic glutamate receptor 1 cooperates with the
118
other intracellular domains to control coupling to G-proteins. J Biol Chem 271,
2199-2205.

Gras, C., Herzog, E., Bellenchi, G.C., Bernard, V., Ravassard, P., Pohl, M., Gasnier,
B., Giros, B., and El Mestikawy, S. (2002). A third vesicular glutamate transporter
expressed by cholinergic and serotoninergic neurons. J Neurosci 22, 5442-5451.

Grunwald, I.C., Korte, M., Adelmann, G., Plueck, A., Kullander, K., Adams, R.H.,
Frotscher, M., Bonhoeffer, T., and Klein, R. (2004). Hippocampal plasticity requires
postsynaptic ephrinBs. Nat Neurosci 7, 33-40.

Gundersen, V., Danbolt, N.C., Ottersen, O.P., and Storm-Mathisen, J. (1993).
Demonstration of glutamate/aspartate uptake activity in nerve endings by use of
antibodies recognizing exogenous D-aspartate. Neuroscience 57, 97-111.

Gundersen, V., Shupliakov, O., Brodin, L., Ottersen, O.P., and Storm-Mathisen, J.
(1995). Quantification of excitatory amino acid uptake at intact glutamatergic
synapses by immunocytochemistry of exogenous D-aspartate. J Neurosci 15, 4417-
4428.

Gurden, H., Uchida, N., and Mainen, Z.F. (2006). Sensory-evoked intrinsic optical
signals in the olfactory bulb are coupled to glutamate release and uptake. Neuron 52,
335-345.

Haghighi, A.P., McCabe, B.D., Fetter, R.D., Palmer, J.E., Hom, S., and Goodman,
C.S. (2003). Retrograde control of synaptic transmission by postsynaptic CaMKII at
the Drosophila neuromuscular junction. Neuron 39, 255-267.

Hara, M.R., and Snyder, S.H. (2007). Cell signaling and neuronal death. Annu Rev
Pharmacol Toxicol 47, 117-141.

Hasegawa, J., Obara, T., Tanaka, K., and Tachibana, M. (2006). High-density
presynaptic transporters are required for glutamate removal from the first visual
synapse. Neuron 50, 63-74.

Herrero, I., Miras-Portugal, M.T., and Sanchez-Prieto, J. (1998). Functional switch
from facilitation to inhibition in the control of glutamate release by metabotropic
glutamate receptors. J Biol Chem 273, 1951-1958.

Higuchi, M., Tomioka, M., Takano, J., Shirotani, K., Iwata, N., Masumoto, H., Maki,
M., Itohara, S., and Saido, T.C. (2005). Distinct mechanistic roles of calpain and
caspase activation in neurodegeneration as revealed in mice overexpressing their
specific inhibitors. J Biol Chem 280, 15229-15237.

119
Holcik, M., Thompson, C.S., Yaraghi, Z., Lefebvre, C.A., MacKenzie, A.E., and
Korneluk, R.G. (2000). The hippocampal neurons of neuronal apoptosis inhibitory
protein 1 (NAIP1)-deleted mice display increased vulnerability to kainic acid-
induced injury. Proc Natl Acad Sci U S A 97, 2286-2290.

Hong, S.C., Lanzino, G., Goto, Y., Kang, S.K., Schottler, F., Kassell, N.F., and Lee,
K.S. (1994). Calcium-activated proteolysis in rat neocortex induced by transient
focal ischemia. Brain Res 661, 43-50.

Hou, L., and Klann, E. (2004). Activation of the phosphoinositide 3-kinase-Akt-
mammalian target of rapamycin signaling pathway is required for metabotropic
glutamate receptor-dependent long-term depression. J Neurosci 24, 6352-6361.

Hsieh, H., Boehm, J., Sato, C., Iwatsubo, T., Tomita, T., Sisodia, S., and Malinow, R.
(2006). AMPAR removal underlies Abeta-induced synaptic depression and dendritic
spine loss. Neuron 52, 831-843.

Huang, C.C., You, J.L., Wu, M.Y., and Hsu, K.S. (2004). Rap1-induced p38
mitogen-activated protein kinase activation facilitates AMPA receptor trafficking via
the GDI.Rab5 complex. Potential role in (S)-3,5-dihydroxyphenylglycene-induced
long term depression. J Biol Chem 279, 12286-12292.

Ishikawa, K., Nash, S.R., Nishimune, A., Neki, A., Kaneko, S., and Nakanishi, S.
(1999). Competitive interaction of seven in absentia homolog-1A and
Ca2+/calmodulin with the cytoplasmic tail of group 1 metabotropic glutamate
receptors. Genes Cells 4, 381-390.

Jeong, H.J., Jang, I.S., Nabekura, J., and Akaike, N. (2003). Adenosine A1 receptor-
mediated presynaptic inhibition of GABAergic transmission in immature rat
hippocampal CA1 neurons. J Neurophysiol 89, 1214-1222.

Jiang, M., Deng, L., and Chen, G. (2004). High Ca(2+)-phosphate transfection
efficiency enables single neuron gene analysis. Gene Ther 11, 1303-1311.

Kam, K., and Nicoll, R. (2007). Excitatory synaptic transmission persists
independently of the glutamate-glutamine cycle. J Neurosci 27, 9192-9200.

Kim, S.J., Kim, Y.S., Yuan, J.P., Petralia, R.S., Worley, P.F., and Linden, D.J.
(2003). Activation of the TRPC1 cation channel by metabotropic glutamate receptor
mGluR1. Nature 426, 285-291.

Kitano, J., Nishida, M., Itsukaichi, Y., Minami, I., Ogawa, M., Hirano, T., Mori, Y.,
and Nakanishi, S. (2003). Direct interaction and functional coupling between
120
metabotropic glutamate receptor subtype 1 and voltage-sensitive Cav2.1 Ca2+
channel. J Biol Chem 278, 25101-25108.

Kobayashi, S., and Millhorn, D.E. (2001). Hypoxia regulates glutamate metabolism
and membrane transport in rat PC12 cells. J Neurochem 76, 1935-1948.

Koh, J.Y., Palmer, E., and Cotman, C.W. (1991). Activation of the metabotropic
glutamate receptor attenuates N-methyl-D-aspartate neurotoxicity in cortical cultures.
Proc Natl Acad Sci U S A 88, 9431-9435.

Lan, J.Y., Skeberdis, V.A., Jover, T., Zheng, X., Bennett, M.V., and Zukin, R.S.
(2001). Activation of metabotropic glutamate receptor 1 accelerates NMDA receptor
trafficking. J Neurosci 21, 6058-6068.

Lee, M.S., Kwon, Y.T., Li, M., Peng, J., Friedlander, R.M., and Tsai, L.H. (2000).
Neurotoxicity induces cleavage of p35 to p25 by calpain. Nature 405, 360-364.

Lonart, G., and Sudhof, T.C. (2000). Assembly of SNARE core complexes prior to
neurotransmitter release sets the readily releasable pool of synaptic vesicles. J Biol
Chem 275, 27703-27707.

Lujan, R., Nusser, Z., Roberts, J.D., Shigemoto, R., and Somogyi, P. (1996).
Perisynaptic location of metabotropic glutamate receptors mGluR1 and mGluR5 on
dendrites and dendritic spines in the rat hippocampus. Eur J Neurosci 8, 1488-1500.

Ma, L., Zablow, L., Kandel, E.R., and Siegelbaum, S.A. (1999). Cyclic AMP
induces functional presynaptic boutons in hippocampal CA3-CA1 neuronal cultures.
Nat Neurosci 2, 24-30.

Maechler, P., and Wollheim, C.B. (2001). Mitochondrial function in normal and
diabetic beta-cells. Nature 414, 807-812.

Malgaroli, A., Ting, A.E., Wendland, B., Bergamaschi, A., Villa, A., Tsien, R.W.,
and Scheller, R.H. (1995). Presynaptic component of long-term potentiation
visualized at individual hippocampal synapses. Science 268, 1624-1628.

Malinow, R., and Malenka, R.C. (2002). AMPA receptor trafficking and synaptic
plasticity. Annu Rev Neurosci 25, 103-126.

Manzoni, O.J., Manabe, T., and Nicoll, R.A. (1994). Release of adenosine by
activation of NMDA receptors in the hippocampus. Science 265, 2098-2101.

Martin, S.J., Grimwood, P.D., and Morris, R.G. (2000). Synaptic plasticity and
memory: an evaluation of the hypothesis. Annu Rev Neurosci 23, 649-711.
121

Masu, M., Tanabe, Y., Tsuchida, K., Shigemoto, R., and Nakanishi, S. (1991).
Sequence and expression of a metabotropic glutamate receptor. Nature 349, 760-765.

Maycox, P.R., Deckwerth, T., Hell, J.W., and Jahn, R. (1988). Glutamate uptake by
brain synaptic vesicles. Energy dependence of transport and functional reconstitution
in proteoliposomes. J Biol Chem 263, 15423-15428.

McIntire, S.L., Reimer, R.J., Schuske, K., Edwards, R.H., and Jorgensen, E.M.
(1997). Identification and characterization of the vesicular GABA transporter. Nature
389, 870-876.

Mennerick, S., Dhond, R.P., Benz, A., Xu, W., Rothstein, J.D., Danbolt, N.C.,
Isenberg, K.E., and Zorumski, C.F. (1998). Neuronal expression of the glutamate
transporter GLT-1 in hippocampal microcultures. J Neurosci 18, 4490-4499.

Moechars, D., Weston, M.C., Leo, S., Callaerts-Vegh, Z., Goris, I., Daneels, G.,
Buist, A., Cik, M., van der Spek, P., Kass, S., et al. (2006). Vesicular glutamate
transporter VGLUT2 expression levels control quantal size and neuropathic pain. J
Neurosci 26, 12055-12066.

Montana, V., Malarkey, E.B., Verderio, C., Matteoli, M., and Parpura, V. (2006).
Vesicular transmitter release from astrocytes. Glia 54, 700-715.

Moroni, F., Cozzi, A., Lombardi, G., Sourtcheva, S., Leonardi, P., Carfi, M., and
Pellicciari, R. (1998). Presynaptic mGlu1 type receptors potentiate transmitter output
in the rat cortex. Eur J Pharmacol 347, 189-195.

Morris, M.C., Depollier, J., Mery, J., Heitz, F., and Divita, G. (2001). A peptide
carrier for the delivery of biologically active proteins into mammalian cells. Nat
Biotechnol 19, 1173-1176.

Moult, P.R., Gladding, C.M., Sanderson, T.M., Fitzjohn, S.M., Bashir, Z.I., Molnar,
E., and Collingridge, G.L. (2006). Tyrosine phosphatases regulate AMPA receptor
trafficking during metabotropic glutamate receptor-mediated long-term depression. J
Neurosci 26, 2544-2554.

Murai, K.K., Nguyen, L.N., Irie, F., Yamaguchi, Y., and Pasquale, E.B. (2003).
Control of hippocampal dendritic spine morphology through ephrin-A3/EphA4
signaling. Nat Neurosci 6, 153-160.

Murthy, V.N., and Stevens, C.F. (1999). Reversal of synaptic vesicle docking at
central synapses. Nat Neurosci 2, 503-507.

122
Nakamura, Y., Iga, K., Shibata, T., Shudo, M., and Kataoka, K. (1993). Glial
plasmalemmal vesicles: a subcellular fraction from rat hippocampal homogenate
distinct from synaptosomes. Glia 9, 48-56.

Newcomb, J.K., Pike, B.R., Zhao, X., and Hayes, R.L. (2000). Concurrent
assessment of calpain and caspase-3 activity by means of western blots of protease-
specific spectrin breakdown products. Methods Mol Biol 144, 219-223.

Ottersen, O.P., Laake, J.H., Reichelt, W., Haug, F.M., and Torp, R. (1996). Ischemic
disruption of glutamate homeostasis in brain: quantitative immunocytochemical
analyses. J Chem Neuroanat 12, 1-14.

Page, G., Khidir, F.A., Pain, S., Barrier, L., Fauconneau, B., Guillard, O., Piriou, A.,
and Hugon, J. (2006). Group I metabotropic glutamate receptors activate the p70S6
kinase via both mammalian target of rapamycin (mTOR) and extracellular signal-
regulated kinase (ERK 1/2) signaling pathways in rat striatal and hippocampal
synaptoneurosomes. Neurochem Int 49, 413-421.

Paukert, M., and Bergles, D.E. (2006). Synaptic communication between neurons
and NG2+ cells. Curr Opin Neurobiol 16, 515-521.

Pellegrini-Giampietro, D.E. (2003). The distinct role of mGlu1 receptors in post-
ischemic neuronal death. Trends Pharmacol Sci 24, 461-470.

Reid, M.E., Toms, N.J., Bedingfield, J.S., and Roberts, P.J. (1999). Group I mGlu
receptors potentiate synaptosomal [3H]glutamate release independently of
exogenously applied arachidonic acid. Neuropharmacology 38, 477-485.

Reynolds, G.P., and Harte, M.K. (2007). The neuronal pathology of schizophrenia:
molecules and mechanisms. Biochem Soc Trans 35, 433-436.

Rhee, J.S., Betz, A., Pyott, S., Reim, K., Varoqueaux, F., Augustin, I., Hesse, D.,
Sudhof, T.C., Takahashi, M., Rosenmund, C., and Brose, N. (2002). Beta phorbol
ester- and diacylglycerol-induced augmentation of transmitter release is mediated by
Munc13s and not by PKCs. Cell 108, 121-133.

Rodriguez-Moreno, A., Sistiaga, A., Lerma, J., and Sanchez-Prieto, J. (1998). Switch
from facilitation to inhibition of excitatory synaptic transmission by group I mGluR
desensitization. Neuron 21, 1477-1486.

Rong, R., Ahn, J.Y., Huang, H., Nagata, E., Kalman, D., Kapp, J.A., Tu, J., Worley,
P.F., Snyder, S.H., and Ye, K. (2003). PI3 kinase enhancer-Homer complex couples
mGluRI to PI3 kinase, preventing neuronal apoptosis. Nat Neurosci 6, 1153-1161.

123
Rosenmund, C., and Stevens, C.F. (1996). Definition of the readily releasable pool of
vesicles at hippocampal synapses. Neuron 16, 1197-1207.

Rossi, D.J., Oshima, T., and Attwell, D. (2000). Glutamate release in severe brain
ischaemia is mainly by reversed uptake. Nature 403, 316-321.

Rostaing, P., Real, E., Siksou, L., Lechaire, J.P., Boudier, T., Boeckers, T.M.,
Gertler, F., Gundelfinger, E.D., Triller, A., and Marty, S. (2006). Analysis of
synaptic ultrastructure without fixative using high-pressure freezing and tomography.
Eur J Neurosci 24, 3463-3474.

Rothstein, J.D., Dykes-Hoberg, M., Pardo, C.A., Bristol, L.A., Jin, L., Kuncl, R.W.,
Kanai, Y., Hediger, M.A., Wang, Y., Schielke, J.P., and Welty, D.F. (1996).
Knockout of glutamate transporters reveals a major role for astroglial transport in
excitotoxicity and clearance of glutamate. Neuron 16, 675-686.

Rothstein, J.D., Patel, S., Regan, M.R., Haenggeli, C., Huang, Y.H., Bergles, D.E.,
Jin, L., Dykes Hoberg, M., Vidensky, S., Chung, D.S., et al. (2005). Beta-lactam
antibiotics offer neuroprotection by increasing glutamate transporter expression.
Nature 433, 73-77.

Rouach, N., and Nicoll, R.A. (2003). Endocannabinoids contribute to short-term but
not long-term mGluR-induced depression in the hippocampus. Eur J Neurosci 18,
1017-1020.

Rumpel, S., LeDoux, J., Zador, A., and Malinow, R. (2005). Postsynaptic receptor
trafficking underlying a form of associative learning. Science 308, 83-88.

Sagara, Y., and Schubert, D. (1998). The activation of metabotropic glutamate
receptors protects nerve cells from oxidative stress. J Neurosci 18, 6662-6671.

Sandrock, A.W., Jr., Dryer, S.E., Rosen, K.M., Gozani, S.N., Kramer, R., Theill,
L.E., and Fischbach, G.D. (1997). Maintenance of acetylcholine receptor number by
neuregulins at the neuromuscular junction in vivo. Science 276, 599-603.

Saugstad, J.A., Segerson, T.P., and Westbrook, G.L. (1996). Metabotropic glutamate
receptors activate G-protein-coupled inwardly rectifying potassium channels in
Xenopus oocytes. J Neurosci 16, 5979-5985.

Schauwecker, P.E., and Steward, O. (1997). Genetic determinants of susceptibility to
excitotoxic cell death: implications for gene targeting approaches. Proc Natl Acad
Sci U S A 94, 4103-4108.
124

Schnabel, R., Kilpatrick, I.C., and Collingridge, G.L. (1999). An investigation into
signal transduction mechanisms involved in DHPG-induced LTD in the CA1 region
of the hippocampus. Neuropharmacology 38, 1585-1596.

Scholzke, M.N., Potrovita, I., Subramaniam, S., Prinz, S., and Schwaninger, M.
(2003). Glutamate activates NF-kappaB through calpain in neurons. Eur J Neurosci
18, 3305-3310.

Schroder, U.H., Opitz, T., Jager, T., Sabelhaus, C.F., Breder, J., and Reymann, K.G.
(1999). Protective effect of group I metabotropic glutamate receptor activation
against hypoxic/hypoglycemic injury in rat hippocampal slices: timing and
involvement of protein kinase C. Neuropharmacology 38, 209-216.

Sheng, M., and Hoogenraad, C.C. (2007). The postsynaptic architecture of excitatory
synapses: a more quantitative view. Annu Rev Biochem 76, 823-847.

Shulman, R.G., Rothman, D.L., Behar, K.L., and Hyder, F. (2004). Energetic basis
of brain activity: implications for neuroimaging. Trends Neurosci 27, 489-495.

Siliprandi, R., Lipartiti, M., Fadda, E., Sautter, J., and Manev, H. (1992). Activation
of the glutamate metabotropic receptor protects retina against N-methyl-D-aspartate
toxicity. Eur J Pharmacol 219, 173-174.

Silverman, M.A., Kaech, S., Jareb, M., Burack, M.A., Vogt, L., Sonderegger, P., and
Banker, G. (2001). Sorting and directed transport of membrane proteins during
development of hippocampal neurons in culture. Proc Natl Acad Sci U S A 98, 7051-
7057.

Siman, R., and Noszek, J.C. (1988). Excitatory amino acids activate calpain I and
induce structural protein breakdown in vivo. Neuron 1, 279-287.

Simpkins, K.L., Guttmann, R.P., Dong, Y., Chen, Z., Sokol, S., Neumar, R.W., and
Lynch, D.R. (2003). Selective activation induced cleavage of the NR2B subunit by
calpain. J Neurosci 23, 11322-11331.

Snyder, E.M., Philpot, B.D., Huber, K.M., Dong, X., Fallon, J.R., and Bear, M.F.
(2001). Internalization of ionotropic glutamate receptors in response to mGluR
activation. Nat Neurosci 4, 1079-1085.

Stanton, P.K., Winterer, J., Bailey, C.P., Kyrozis, A., Raginov, I., Laube, G., Veh,
R.W., Nguyen, C.Q., and Muller, W. (2003). Long-term depression of presynaptic
release from the readily releasable vesicle pool induced by NMDA receptor-
dependent retrograde nitric oxide. J Neurosci 23, 5936-5944.
125

Stowell, J.N., and Craig, A.M. (1999). Axon/dendrite targeting of metabotropic
glutamate receptors by their cytoplasmic carboxy-terminal domains. Neuron 22, 525-
536.

Suchak, S.K., Baloyianni, N.V., Perkinton, M.S., Williams, R.J., Meldrum, B.S., and
Rattray, M. (2003). The 'glial' glutamate transporter, EAAT2 (Glt-1) accounts for
high affinity glutamate uptake into adult rodent nerve endings. J Neurochem 84, 522-
532.

Sudhof, T.C. (2004). The synaptic vesicle cycle. Annu Rev Neurosci 27, 509-547.

Takamori, S., Holt, M., Stenius, K., Lemke, E.A., Gronborg, M., Riedel, D., Urlaub,
H., Schenck, S., Brugger, B., Ringler, P., et al. (2006). Molecular anatomy of a
trafficking organelle. Cell 127, 831-846.

Takamori, S., Rhee, J.S., Rosenmund, C., and Jahn, R. (2000). Identification of a
vesicular glutamate transporter that defines a glutamatergic phenotype in neurons.
Nature 407, 189-194.

Takamori, S., Rhee, J.S., Rosenmund, C., and Jahn, R. (2001). Identification of
differentiation-associated brain-specific phosphate transporter as a second vesicular
glutamate transporter (VGLUT2). J Neurosci 21, RC182.

Takayasu, Y., Iino, M., Kakegawa, W., Maeno, H., Watase, K., Wada, K.,
Yanagihara, D., Miyazaki, T., Komine, O., Watanabe, M., et al. (2005). Differential
roles of glial and neuronal glutamate transporters in Purkinje cell synapses. J
Neurosci 25, 8788-8793.

Tateyama, M., and Kubo, Y. (2006). Dual signaling is differentially activated by
different active states of the metabotropic glutamate receptor 1alpha. Proc Natl Acad
Sci U S A 103, 1124-1128.

Thandi, S., Blank, J.L., and Challiss, R.A. (2002). Group-I metabotropic glutamate
receptors, mGlu1a and mGlu5a, couple to extracellular signal-regulated kinase (ERK)
activation via distinct, but overlapping, signalling pathways. J Neurochem 83, 1139-
1153.

Thiagarajan, T.C., Lindskog, M., and Tsien, R.W. (2005). Adaptation to synaptic
inactivity in hippocampal neurons. Neuron 47, 725-737.

Thompson, R.J., Zhou, N., and MacVicar, B.A. (2006). Ischemia opens neuronal gap
junction hemichannels. Science 312, 924-927.

126
Torp, R., Andine, P., Hagberg, H., Karagulle, T., Blackstad, T.W., and Ottersen, O.P.
(1991). Cellular and subcellular redistribution of glutamate-, glutamine- and taurine-
like immunoreactivities during forebrain ischemia: a semiquantitative electron
microscopic study in rat hippocampus. Neuroscience 41, 433-447.

Tu, J.C., Xiao, B., Naisbitt, S., Yuan, J.P., Petralia, R.S., Brakeman, P., Doan, A.,
Aakalu, V.K., Lanahan, A.A., Sheng, M., and Worley, P.F. (1999). Coupling of
mGluR/Homer and PSD-95 complexes by the Shank family of postsynaptic density
proteins. Neuron 23, 583-592.

Turrigiano, G.G., Leslie, K.R., Desai, N.S., Rutherford, L.C., and Nelson, S.B.
(1998). Activity-dependent scaling of quantal amplitude in neocortical neurons.
Nature 391, 892-896.

Turrigiano, G.G., and Nelson, S.B. (2004). Homeostatic plasticity in the developing
nervous system. Nat Rev Neurosci 5, 97-107.

Vanoni, C., Massari, S., Losa, M., Carrega, P., Perego, C., Conforti, L., and Pietrini,
G. (2004). Increased internalisation and degradation of GLT-1 glial glutamate
transporter in a cell model for familial amyotrophic lateral sclerosis (ALS). J Cell Sci
117, 5417-5426.

Veruki, M.L., Morkve, S.H., and Hartveit, E. (2006). Activation of a presynaptic
glutamate transporter regulates synaptic transmission through electrical signaling.
Nat Neurosci 9, 1388-1396.

Vincent, A.M., and Maiese, K. (2000). The metabotropic glutamate system promotes
neuronal survival through distinct pathways of programmed cell death. Exp Neurol
166, 65-82.

Voglmaier, S.M., Kam, K., Yang, H., Fortin, D.L., Hua, Z., Nicoll, R.A., and
Edwards, R.H. (2006). Distinct endocytic pathways control the rate and extent of
synaptic vesicle protein recycling. Neuron 51, 71-84.

Voutsinos-Porche, B., Bonvento, G., Tanaka, K., Steiner, P., Welker, E., Chatton,
J.Y., Magistretti, P.J., and Pellerin, L. (2003). Glial glutamate transporters mediate a
functional metabolic crosstalk between neurons and astrocytes in the mouse
developing cortex. Neuron 37, 275-286.

Wadia, J.S., Stan, R.V., and Dowdy, S.F. (2004). Transducible TAT-HA fusogenic
peptide enhances escape of TAT-fusion proteins after lipid raft macropinocytosis.
Nat Med 10, 310-315.

127
128
Wang, Y., Ju, W., Liu, L., Fam, S., D'Souza, S., Taghibiglou, C., Salter, M., and
Wang, Y.T. (2004). alpha-Amino-3-hydroxy-5-methylisoxazole-4-propionic acid
subtype glutamate receptor (AMPAR) endocytosis is essential for N-methyl-D-
aspartate-induced neuronal apoptosis. J Biol Chem 279, 41267-41270.

Whitlock, J.R., Heynen, A.J., Shuler, M.G., and Bear, M.F. (2006). Learning induces
long-term potentiation in the hippocampus. Science 313, 1093-1097.

Wilson, N.R., Kang, J., Hueske, E.V., Leung, T., Varoqui, H., Murnick, J.G.,
Erickson, J.D., and Liu, G. (2005). Presynaptic regulation of quantal size by the
vesicular glutamate transporter VGLUT1. J Neurosci 25, 6221-6234.

Xiao, M.Y., Zhou, Q., and Nicoll, R.A. (2001). Metabotropic glutamate receptor
activation causes a rapid redistribution of AMPA receptors. Neuropharmacology 41,
664-671.

Xiong, Z.G., Zhu, X.M., Chu, X.P., Minami, M., Hey, J., Wei, W.L., MacDonald,
J.F., Wemmie, J.A., Price, M.P., Welsh, M.J., and Simon, R.P. (2004).
Neuroprotection in ischemia: blocking calcium-permeable acid-sensing ion channels.
Cell 118, 687-698.

Zakharenko, S.S., Zablow, L., and Siegelbaum, S.A. (2001). Visualization of
changes in presynaptic function during long-term synaptic plasticity. Nat Neurosci 4,
711-717.

Zakharenko, S.S., Zablow, L., and Siegelbaum, S.A. (2002). Altered presynaptic
vesicle release and cycling during mGluR-dependent LTD. Neuron 35, 1099-1110.

Zhang, X.L., Zhou, Z.Y., Winterer, J., Muller, W., and Stanton, P.K. (2006).
NMDA-dependent, but not group I metabotropic glutamate receptor-dependent,
long-term depression at Schaffer collateral-CA1 synapses is associated with long-
term reduction of release from the rapidly recycling presynaptic vesicle pool. J
Neurosci 26, 10270-10280.

Zhu, P., DeCoster, M.A., and Bazan, N.G. (2004). Interplay among platelet-
activating factor, oxidative stress, and group I metabotropic glutamate receptors
modulates neuronal survival. J Neurosci Res 77, 525-531.

Abstract (if available)

Abstract Glutamate is the major excitatory neurotransmitter in the mammalian central nervous system. Synaptic transmission at glutamatergic synapses mediates and regulates basically all aspects of brain functions. The strength of these synapses is subjected to potentiation and depression and these plastic modifications are plausible candidates for information storage in the brain. Excessive activity at glutamatergic synapses, namely excitotoxicity, occurs in many brain diseases and is a critical factor for neuronal death or degeneration. Therefore, to elucidate the molecular mechanisms underlying glutamatergic transmission, plasticity and excitotoxicity constitutes the first step to understand neuronal information processing, learning, memory and brain diseases.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Studies on plasticity and neurodegeneration in rat hippocampus

PDF

Regulatory mechanisms of gamma-aminobutyric acid transporter 1 in the rat brain

PDF

Brain-derived neurotrophic factor (BDNF)-dependent activation of calpain: involvement of ERK pathway and regulation of AMPA receptor function and stability

PDF

Dual genetic screens for mutants in synaptic homeostatic plasticity and a characterization of insomniac as a regulator for retrograde homeostatic signaling

PDF

Computational investigation of glutamatergic synaptic dynamics: role of ionotropic receptor distribution and astrocytic modulation of neuronal spike timing

PDF

Estrogen and progesterone interaction on synaptic transmission and LTP in rodent hippocampus

PDF

Presynaptic glutamate receptors and auxiliary subunits in neurotransmission and homeostatic potentiation

PDF

Engineering genetic tools to illustrate new insights into the homeostatic control of synaptic strength

PDF

Neuroplasticity of the basal ganglia in the 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine mouse model of Parkinson's disease

PDF

Synapse maintenance and function at the mouse neuromuscular junction: implications in diseases

PDF

The role of Schwann cells in the development of the neuromuscular junction

PDF

Translational regulation and endosomal trafficking in synaptic adaptation to stress

PDF

Exploring alternative roles of visual arrestin 1 in photoreceptor synaptic regulation and deciphering the molecular pathway of retinal degeneration using mouse knockout technology

PDF

Uncovering synapse-specific roles of proteins implicated in complex brain disorders via novel and targeted approaches

PDF

New understanding of RIC-3, a regulator of alpha7 nicotinic acetylcholine receptor expression

PDF

Uncovering cell type-specific roles of proteins involved in glutamatergic synapse regulation

PDF

The structural and functional configurations of glutamate, GABA, and catecholamine pre-synaptic terminals in the parvicellular neuroendocrine part of the paraventricular nucleus of the hypothalamus

PDF

Roles of SIRT1 in neuronal oxidative damage and brain function

PDF

Memory abnormalities in Alzheimer's disease and anxiety models

PDF

Studies of intracellular cascades mediating neuronal damage in two animal models of neurodegeneration

Asset Metadata

Creator Xu, Wei (author)

Core Title An orchestra of glutamate receptors and transporters in synaptic transmission, plasticity and excitotoxicity

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Neuroscience

Publication Date 03/24/2008

Defense Date 11/30/2007

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag glutamate receptor,OAI-PMH Harvest,synaptic plasticity,synaptic transmission,transporter

Language English

Advisor Baudry, Michel (committee chair), Johnson, Deborah L. (committee member), Ko, Chien-Ping (committee member), Thompson, Richard F. (committee member)

Creator Email weixu@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-m1054

Unique identifier UC192252

Identifier etd-Xu-20080324 (filename),usctheses-m40 (legacy collection record id),usctheses-c127-41520 (legacy record id),usctheses-m1054 (legacy record id)

Legacy Identifier etd-Xu-20080324.pdf

Dmrecord 41520

Document Type Dissertation

Rights Xu, Wei

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Repository Name Libraries, University of Southern California

Repository Location Los Angeles, California

Repository Email cisadmin@lib.usc.edu