AMPA receptor desensitization mutation results in
severe developmental phenotypes and early
postnatal lethality
Louisa A. Christie, Theron A. Russell, Jian Xu, Lydia Wood, Gordon M. G. Shepherd, and Anis Contractor1
Department of Physiology, Northwestern University Feinberg School of Medicine, Chicago, IL 60611
Edited by Richard L. Huganir, Johns Hopkins University School of Medicine, Baltimore, MD, and approved April 13, 2010 (received for review July 21, 2009)
GluA2 (GluR2)
| hippocampus | seizures | knock-in mouse
A
MPA receptors are tetramers assembled from the four receptor subunits GluA1–GluA4 (GluR1–4) (1). These receptors are activated by their endogenous ligand glutamate, and
rapidly undergo desensitization within milliseconds of glutamate
binding. Desensitization involves a conformational change of the
receptor complex that allows closure of the channel gate while
glutamate remains bound to the receptor (2). Synaptic currents are
predominantly mediated by AMPA receptors at most excitatory
synapses; therefore there has been interest in the development of
pharmacological agents that enhance AMPA receptor function by
limiting receptor deactivation and desensitization (3). There are
many clear examples of synapses at which postsynaptic receptor
desensitization plays a major role in synaptic depression (4–6).
Many of these synapses are specialized structures in which glutamate remains in the synaptic cleft for prolonged periods of time
during normal operation of the synapse (7). In contrast, at synapses where cleft glutamate is cleared rapidly or where AMPA
receptor stoichiometry has become specialized to support high
frequency transmission, there is little evidence that synaptic receptor desensitization has much influence on shaping the kinetics
of transmission, and it is likely that receptor deactivation is the
primary determinant of EPSC time-course (8–10).
To determine the importance of AMPA receptor desensitization
in vivo, we introduced the nondesensitizing L483Y mutation into
the mouse gene encoding GluA2 (Gria2) (11). This mutation
turned out to be homozygous lethal; however, heterozygous mice
(GluA2L483Y/wt) were viable despite a severe and progressive neuwww.pnas.org/cgi/doi/10.1073/pnas.0908206107
rological and developmental phenotype that included significant
runting, abnormal gait, development of progressively severe seizures, and early mortality in the third postnatal weeks. Overall the
very severe phenotype observed by a single amino acid alteration
in the GluA2 receptor subunit indicates that AMPA receptor desensitization is critical for the viability of the animal and function
of the CNS.
Results
Generation and Phenotype of GluA2L483Y/wt Mice. A single amino acid
exchange in the S1 domain of the AMPA receptor subunits eliminates desensitization of recombinant receptors expressed in heterologous systems (11) (Fig. 1A). To introduce this mutation into the
mouse genome, we generated a targeting construct containing exon
11 and the surrounding region of Gria2 with several mutated
nucleotides to code for a tyrosine residue at position 483. In addition, a loxP flanked neomycin (neo) selection cassette was incorporated into the intronic region downstream of this exon (Fig.
1B). The targeting construct was integrated into mouse embryonic
stem cells by standard techniques of homologous recombination to
generate mice carrying the mutated allele GluA2L483Yneo. Heterozygous GluA2L483Yneo/wt mice, which were viable and fertile, were
intercrossed with the intention of producing homozygote mutants;
however, no offspring were generated that carried two mutant
alleles. To remove the neo cassette from the intronic region of
Gria2, GluA2L483Yneo/wt mice were crossed with a “Cre-deleter” line
that expressed Cre recombinase in all tissues including the germline (CMV-Cre) (12). This cross produced offspring carrying a single mutated allele without the neo cassette (GluA2L483Y/wt).
Surprisingly, removing the neo cassette uncovered a dramatic
phenotype in heterozygote animals, suggesting that the presence of
the neo cassette had caused unequal expression of the mutant allele;
this was supported by Western analysis, which demonstrated that
GluA2 expression is decreased in GluA2L483Yneo/wt mice (Fig. S1).
Similar reductions in allele expression by intronic insertion of
a neomycin cassette have been reported previously (13, 14).
GluA2L483Y/wt offspring were runted in comparison with their
wild-type (WT) littermates with an approximate 45% decrease in
body weight at postnatal days (P) 15–17 (Fig. 1 D and E).
Moreover, GluA2L483Y/wt animals were prone to progressively
severe spontaneous seizures. At P14, we observed no seizures
when animals were observed for a 1-h period. At P16 and beyond,
GluA2L483Y/wt mice exhibited multiple spontaneous generalized
clonic/tonic convulsions when observed over a similar time period. Examination of c-fos expression in P16-18 mice demon-
Author contributions: L.A.C., J.X., G.M.G.S., and A.C. designed research; L.A.C., T.A.R., J.X.,
L.W., G.M.G.S., and A.C. performed research; L.A.C., T.A.R., J.X., and A.C. analyzed data;
and L.A.C. and A.C. wrote the paper.
The authors declare no conflict of interest.
This article is a PNAS Direct Submission.
1
To whom correspondence should be addressed. E-mail: [email protected].
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.
1073/pnas.0908206107/-/DCSupplemental.
PNAS Early Edition | 1 of 6
NEUROSCIENCE
AMPA (α-amino-3-hydroxy-5-methyl-4-isoxazole-propionate) receptors desensitize rapidly and completely in the continued presence of
their endogenous ligand glutamate; however, it is not clear what role
AMPA receptor desensitization plays in the brain. We generated
a knock-in mouse in which a single amino acid residue, which controls
desensitization, was mutated in the GluA2 (GluR2) receptor subunit
(GluA2L483Y). This mutation was homozygous lethal. However, mice
carrying a single mutated allele, GluA2L483Y/wt, survived past birth,
but displayed severe and progressive neurological deficits including
seizures and, ultimately, increased mortality. The expression of the
AMPA receptor subunits GluA1 and GluA2 was decreased, whereas
NMDA receptor protein expression was increased in GluA2L483Y/wt
mice. Despite this, basal synaptic transmission and plasticity in the
hippocampus were largely unaffected, suggesting that neurons preferentially target receptors to synapses to normalize synaptic weight.
We found no gross neuroanatomical alterations in GluA2L483Y/wt
mice. Moreover, there was no accumulation of AMPA receptor subunits in intracellular compartments, suggesting that folding and assembly of AMPA receptors are not affected by this mutation.
Interestingly, EPSC paired pulse ratios in the CA1 were enhanced
without a change in synaptic release probability, demonstrating that
postsynaptic receptor properties can contribute to facilitation. The
dramatic phenotype observed in this study by the introduction of
a single amino acid change demonstrates an essential role in vivo
for AMPA receptor desensitization.
(24 ± 4.6%, n = 6, P < 0.01) in GluA2L483Y/wt (Fig. 2A). Membrane
receptors were also reduced in the isolated synaptoneurosome
fraction. In this case, we observed a clear reduction in GluA2 receptor protein (36 ± 5.8%, n = 6, P < 0.05) and a smaller decrease in
GluA1 protein (15 ± 7.3%, n = 6, P < 0.05) (Fig. 2B). Because
AMPA and NMDA receptors are colocalized at synapses and their
relative contributions to synaptic signaling and expression are tightly
linked, we also quantified the relative amount of GluN1 protein.
Surprisingly, we observed an up-regulation of GluN1 expression
in whole hippocampus (130 ± 18%, n = 6, P < 0.05), but again only a
small alteration in the synaptoneurosome fraction (93 ± 3.2%, n = 6,
P < 0.05) (Fig. 2B). These data suggest that multiple compensatory
alterations in glutamate receptor expression occur in GluA2L483Y/wt
mice. To validate these changes in receptor expression observed
with Western blot analysis, we performed immunohistochemical
analysis on sections from GluA2L483Y/wt and GluA2wt/wt. Using
quantitative measurements of labeling density in sections from WT
and mutant animals we compared expression of GluA1, GluA2, and
GluN1 in the hippocampal regions stratum oriens (s.ori.), stratum
pyramidale (s.pyr.), and stratum radiatum (s. rad.) (Fig. 2 C–H).
Although we did not see as clear alterations in antibody density in
sections as we had in protein blots, there was a reduction in hippocampal GluA1 and GluA2 and a small increase in GluN1 consistent
with our preliminary finding. Overall these results demonstrate that
introduction of the mutant allele causes a drastic alteration in glutamate receptor expression in GluA2L483Y/wt mice.
Glutamate Receptors Are Not Sequestered in the ER in GluA2L483Y/wt
Mice. The expression of glutamate receptors is controlled by
mechanisms that regulate biogenesis and assembly of membrane
Fig. 1. Generation and characterization of GluA2L483Y mice. (A) Cartoon
representation of the AMPA receptor subunit GluA2 depicting the position of
the L483Y. (B) Schematic representation of the Gria2 genomic locus, targeting
construct and targeted allele with position of codon for leucine 483 in exon 11.
(C) Mortality curve for GluA2L483Y/wt mice. (D) Body weights comparisons in
P15–P17 mice. (E) Comparison of size of GluA2wt/wt and GluA2L483Y/wt littermate
mice at P15 (Left). Altered gait of GluA2L483Y/wt mice (Right). (F) Nissl stained
coronal sections. (G) C-fos staining in section from vehicle treated GluA2wt/wt,
kainic acid–treated GluA2wt/wt, and nontreated (P17) GluA2L483Y/wt mice. (H)
Analysis of PSD density (number of PSDs per 32.5 μm2) and size (length and
thickness) in GluA2L483Y/wt mice. (I) Example electron micrographs from hippocampal CA1 area sections.
strated activation of neurons throughout the brain. C-fos reactivity was more widespread in the brains of GluA2L483Y/wt mice,
which had been observed to have multiple seizures, than in WT
animals that had undergone acute seizures induced by kainic acid
injection (Fig. 1G). GluA2L483Y/wt mice were monitored from
birth and it was found that the LT50 (the time to 50% lethality)
was 17.5 days. Most mice died in the third postnatal week, with
very few surviving past P30 (Fig. 1C). In Nissl-stained sections
we observed no obvious alterations in cell layers or density of
GluA2L483Y/wt mice (Fig. 1F), and analysis of synaptic structure at
the electron microscopic level did not reveal any alterations in the
density or size of asymmetric excitatory synapses in area CA1 of
the hippocampus (Fig. 1 H and I).
Glutamate Receptor Expression Is Altered in GluA2L483Y/wt Animals.
Western blot analysis of whole hippocampal homogenate demonstrated a clear reduction in the amount of GluA1 (44 ± 6.4%, n = 6,
P < 0.001), and to a lesser degree GluA2 receptor subunit protein
2 of 6 | www.pnas.org/cgi/doi/10.1073/pnas.0908206107
Fig. 2. Analysis of glutamate receptor expression in GluA2L483Y/wt mice. (A)
Western blots for GluA1, GluA2, and GluN1 from hippocampal homogenates. (B) Western blots for glutamate receptor expression in the synaptoneurosome fraction. (C) Immunohistochemical analysis of GluA1 expression
in coronal sections from GluA2wt/wt and GluA2L483Y/wt mice. (Lower) Hippocampal sections. (D) Expression of GluA2 in coronal section. (E) Expression of
GluN1 (F–H) Densitometric analysis of antibody staining in hippocampal
sections. Density analysis was carried out by defining three regions of interest in hippocampal section in the stratum oriens (st. oriens), stratum
pyramidale (st. pyr.), and stratum radiatum (st. rad.) in the CA1 region.
Christie et al.
Glutamate Receptors Redistribute to Synaptic Sites in GluA2L483Y/wt
Mice. The overall expression of glutamate receptor subunits is al-
tered in GluA2L483Y/wt mice, with a less dramatic change in expression observed in the synaptoneurosome fraction. To determine
whether there were any alterations in basal levels of synaptic transmission in GluA2L483Y/wt mice, we performed several electrophysiological experiments in acute hippocampal slices. We chose to focus
on CA1 synapses, as they are the canonical excitatory synapse and
have been extensively characterized. Input output curves (I/O) were
constructed by measuring the slope of the field excitatory postsynaptic potential (fEPSP) and the amplitude of the presynaptic fiber volley. We found no difference in the I/O curves between WT
mice and mutant littermates, suggesting there are no gross alter-
Fig. 3. Intracellular trafficking of GluA2 receptor subunits is not disrupted in
GluA2L483Y/wt mice. (A) Representative example gel from Endo H assay. (B)
Representative Western blot analysis of the expression of the ER chaperone
protein Grp78/BiP (Left) and quantification of grouped data (Right). (C)
Representative gel of interaction of GluA2 with chaperone molecule calnexin
(Left). Quantified grouped data from three separate experiments (Right).
Christie et al.
ations in synaptic transmission in area CA1 of the hippocampus (Fig.
4A). We also found that the mean amplitude of miniature excitatory
postsynaptic currents (mEPSCs) in recordings from GluA2wt/wt mice
was 14 ± 1 pA (n = 7), and was not different from the amplitude of
mEPSCs recorded in GluA2L483Y/wt mice of the same age (15 ± 0.8
pA, n = 10, P > 0.05) (Fig. 4 B and C). These results suggest that
the density of AMPA receptors at hippocampal synapses is largely
unaltered despite a significant decrease in total expression of the
two primary hippocampal AMPA receptor subunits.
The relative proportion of the two major glutamate receptor
types, NMDA and AMPA, is strongly correlated with the developmental maturity of excitatory synapses, and the potential ability of
synapses to increase or decrease their efficacy (16). By measuring the
AMPA component at hyperpolarized membrane potentials (−60
mV) and the NMDA component at depolarized membrane potentials (+40 mV) (Materials and Methods), we determined a mean
NMDA/AMPA (N/A) ratio of 0.33 ± 0.03 (n = 10) in GluA2wt/wt
mice (Fig. 4D). In recordings from GluA2L483Y/wt there was a small
but significant reduction in the N/A ratio of CA1 synapses, (0.25 ±
0.03, n = 10, P < 0.05) (Fig. 4D). Viewed in light of the biochemical
analysis and the mEPSC data, it seems likely that there is little alteration in synaptic AMPA receptor distribution at hippocampal
synapses, but there is a small reduction in NMDA receptors.
The presence of edited GluA2 subunits in a heteromeric AMPA
receptor complex confers a reduction in Ca2+ permeability and
single-channel conductance upon AMPA receptors. GluA2-lacking
receptors exhibit inwardly rectifying current–voltage (I/V) relationships (17) because outward current flow at depolarized membrane potentials is blocked by intracellular polyamines. GluA2
protein is decreased in GluA2L483Y/wt mice; therefore we sought to
determine if there might be an abundance of synaptic receptors
lacking the GluA2 subunit. AMPA receptor mediated EPSCs in
WT mice exhibited linear I/V curves (Fig. 4E). To quantify the
amount of rectification, we calculated the rectification index (RI)
(Materials and Methods) of AMPA EPSCs in GluA2wt/wt as 1.0 ±
0.08 (n = 15) (Fig. 4 E and F). In interleaved recordings from littermate GluA2L483Y/wt mice the calculated RI was significantly
reduced (0.7 ± 0.07, n = 12, P < 0.001) (Fig. 4 E and F). A closer
look at the grouped data revealed a subset of recordings in which
Fig. 4. Basal synaptic transmission in hippocampal CA1 synapses in GluA2L483Y/
wt
mice. (A) Input/output curve for synaptic transmission in CA1. (B) Representative traces of mEPSCs recorded from CA1 pyramidal neurons in GluA2wt/wt
(top) and GluA2L483Y/wt mice (bottom). Calibration: 20pA, 500 ms. (C) Cumulative
probability distribution of mEPSC amplitudes from all recordings (Left)
and mean mEPSC amplitude (Right). (D) Representative EPSC traces recorded at
−60 mV and +40mV in GluA2wt/wt (top) and GluA2L483Y/wt mice (bottom). The
NMDA component was measured at +40 mV, 60 ms after the onset of the evoked
outward current (shaded area). Calibration: 50 pA, 100 ms. (E) Current/voltage
curves for AMPA EPSCs in GluA2wt/wt and GluA2L483Y/wt mice. Representative
AMPA EPSC traces from WT and mutant animals recorded at −60 mV and +40 mV
(in D-APV). Calibration: 50 pA, 10 ms. (F) Rectification index of AMPA EPSCs.
PNAS Early Edition | 3 of 6
NEUROSCIENCE
proteins in the endoplasmic reticulum (ER). Misfolded or improperly assembled receptors are not further trafficked into the
secretory pathway, becoming “trapped” in the ER. A previous study
demonstrated that when GluA2(L483Y) was exogenously expressed in cultured neurons, receptor subunits assembled normally
in tetrameric complexes but ER exit of this mutant receptor was
reduced (15). Using an EndoH assay to determine the glycosylation
state of GluA2 receptor subunits, we found that AMPA receptors
did not appear to be accumulating in intracellular compartments in
GluA2L483Y/wt mice. There was no increase in the immature ER
resident GluA2 protein, and in fact we observed less immature
protein, which is likely due to a decrease in the overall abundance of
GluA2 (Fig. 3A). As an alternative approach to examine whether
the intracellular trafficking of glutamate receptor subunits was
disrupted in GluA2L483Y/wt mice, we examined ER stress proteins.
The accumulation of misfolded proteins in the ER lumen induces
prolonged ER stress, resulting in the activation of an adaptive response known as the unfolded protein response (UPR). This is
commonly detected by an up-regulation of the ER chaperone
protein Grp78/BiP (78-kDa glucose regulated protein). In quantitative Western blots for Grp78/BiP, we found no evidence of
Grp78/BiP up-regulation in GluA2L483Y/wt mice (91 ± 5%, n = 6,
P > 0.05) (Fig. 3B). In addition, we found no alteration in the
amount of calnexin (a chaperone molecule that assists in folding
and quality control of proteins as they pass along the secretory
pathway), that coimmunoprecipitated with GluA2 in GluA2L483Y/wt
mice (103 ± 14%, n = 3, P > 0.05) (Fig. 3C). Therefore, there is
little evidence for the accumulation of glutamate receptor subunits
in intracellular compartments in GluA2L483Y/wt mice.
the RIs were closer to 0.5 (and well below the group mean). In these
five recordings, the RI of AMPA EPSCs was 0.4 ± 0.02 (Fig. 4F).
Thus it seems likely that there is an increase in the proportion of
Ca2+-permeable AMPA receptors in GluA2L483Y/wt mice at some
hippocampal CA1 synapses.
Extrasynaptic AMPA Receptor Density Is Reduced in GluA2L483Y/wt
Mice. The electrophysiological analysis of hippocampal synaptic
transmission found moderate alterations in synaptic glutamate
receptors in GluA2L483Y/wt mice. In prior studies, it was noted that
disrupting glutamate receptor expression by knockout of one of the
AMPA receptor subunits (18), or by ablation of one of the accessory proteins associated with AMPA receptors (19), did not drastically alter synaptic AMPA receptor localization, but reduced the
extrasynaptic pool of receptors. Although our biochemical analyses
was consistent with a preferential redistribution of glutamate
receptors to synaptic sites, we wanted to determine whether there
was an overall reduction in the surface expression of AMPA
receptors that would also support this model for a normalization of
synaptic receptors. Application of the agonist AMPA (100 nM)
elicited a current of amplitude 480 ± 44 pA (n = 15) in GluA2wt/wt
mice (Fig. 5). In similar recordings from GluA2L483Y/wt mice the
amplitude of the elicited current was smaller by ≈30% (340 ± 26
pA, n = 14, P < 0.001). Therefore, although the density of synaptic
receptors is largely unaltered, there is a reduction in extrasynaptic
AMPA receptors in GluA2L483Y/wt mice.
Synaptic Plasticity in GluA2L483Y/wt Mice. Previous work demon-
strated that when the extrasynaptic pool of AMPA receptors was
depleted in knockout mice, LTP in the CA1 region of the hippocampus was impaired (19, 18). This is likely due to the expression mechanisms of LTP, which involve the insertion of new
receptors into synapses either by lateral diffusion along the
membrane, or from intracellular compartments. Because of the
reduced extrasynaptic receptor pool in GluA2L483Y/wt we tested
whether the expression of LTP might be reduced in mutant mice.
We recorded fEPSP in the CA1 and induced LTP using a tetanic
stimulation (two trains at 100 Hz for 1 s, each train separated by
20 s). In GluA2wt/wt mice, the slope of the fEPSPs was potentiated
on average by 240 ± 20%, n = 9, between 50 and 60 min posttetanus (Fig. 6 A and B). As expected, in interleaved experiments,
inclusion of the NMDA receptor antagonist D-APV in the ACSF
significantly blocked LTP (130 ± 6.3%, n = 7).
Surprisingly, in recordings from littermate GluA2L483Y/wt, the
same tetanic induction protocol resulted in LTP, which was not
different in magnitude from that observed in WT recordings (210 ±
16%, n = 7, P > 0.05) (Fig. 6 D and E). When GluA2 is ablated in
knockout mice, LTP is enhanced and a small NMDA receptor independent form of plasticity is observed in CA1 (20). To determine
whether a similar LTP, presumably triggered by Ca2+-permeable
receptors, was present in GluA2L483Y/wt mice, we performed further
Fig. 5. Agonist-induced AMPA receptor currents are reduced in GluA2L483Y/wt
mice. (A) Representative traces from whole-cell recordings of CA1 pyramidal
neurons. AMPA receptors are activated by bath application of 100 nM AMPA
in the presence of other receptor antagonist and cyclothiazide. Calibration:
100 pA, 200 s. (B) Grouped data from all recordings.
4 of 6 | www.pnas.org/cgi/doi/10.1073/pnas.0908206107
LTP experiments in the presence of D-APV. At 50–60 min posttetanus, the fEPSP was 120 ± 10% (n = 4) of control, which is not
different from what we observed when NMDA receptors were
inhibited in WT animals (n = 4, P > 0.05) (Fig. 6F). Therefore,
NMDA receptor dependent LTP is unaffected in GluA2L483Y/wt
mice despite a reduction in the extrasynaptic pool of AMPA
receptors. Similarly, the small increase in Ca2+-permeable AMPA
receptors in hippocampal synapses had no effect on NMDA
receptor–independent synaptic plasticity.
Paired Pulse Facilitation Is Enhanced in GluA2L483Y/wt Mice. Repetitive
sequential presynaptic activity at several synapses results in desensitization of postsynaptic AMPA receptors and contributes to
short-term plasticity (4, 21, 5, 7). In hippocampal synapses prior
work has demonstrated either a clear effect (22), or little evidence
for postsynaptic desensitization during paired presynaptic stimuli
(10). We performed paired pulse recordings at several interstimulus intervals in CA1 neurons to readdress this issue. In
recordings from GluA2L483Y/wt mice, we found that the paired
pulse ratio (PPR) was higher at all of the intervals tested (up to 200
ms) (Fig. 7A). In a subset of recordings, PPR (40 ms interstimulus
interval) measured under conditions of increased release probability (4 mM Ca2+) was also higher in GluA2L483Y/wt (GluA2wt/wt:
1.3 ± 0.06, n = 9; GluA2L483Y/wt: 1.5 ± 0.1, n = 17, P < 0.05) (Fig.
7B). An alteration in PPR is generally interpreted as an altered
initial release probability; however, postsynaptic receptor desensitization could also play a role in determining the degree of
paired pulse facilitation (23). To distinguish between these two
possibilities, we made comparison of the rate of block of synaptic
NMDA receptors by the open channel blocker MK801, a common
proxy for determining changes in glutamate release (24). In interleaved experiments, we found no difference in the progressive
block of synaptic NMDA receptors in the CA1 of GluA2L483Y/wt
mice and littermate controls (Fig. 7 C–E). Therefore, from this
analysis, it appears that there is no evidence for altered release
probability of excitatory synapses in the CA1 region of the hippocampus of mutant mice.
Fig. 6. No impairment of hippocampal LTP in GluA2L483Y/wt mice. (A) Timecourse of LTP in GluA2wt/wt mice. LTP is induced at time 0 using tetanic
stimulation (shaded area). In interleaved experiments, LTP was blocked by
addition of NMDA receptor antagonist D-APV. (B) Representative fEPSP
traces before and 50 min after LTP induction. Calibration: 0.2 mV, 20 ms. (C)
Cumulative probability distribution of LTP in all recordings from GluA2wt/wt
and GluA2L483Y/wt mice. (D) Time-course of LTP in GluA2L483Y/wt. (E) Representative fEPSP traces from one experiment before and 50 min after LTP
induction. (F) Cumulative probability distribution of all LTP recording in the
presence of NMDA receptor antagonist D-APV.
Christie et al.
Fig. 7. Receptor desensitization contributes to paired pulse ratio of synaptic and extrasynaptic receptors. (A) Paired pulse ratio of AMPA receptor
EPSCs in hippocampus from GluA2wt/wt and GluA2L483Y/wt recordings. (B)
Paired pulse ratio at 40 ms interval in low (2 mM) and high (4 mM) external
Ca2+. (C) Progressive block of NMDA receptors by MK801 is not different in
GluA2L483Y/wt mice. (D and E) Representative NMDA receptors EPSCs before
and after block by MK801. Calibration: GluA2wt/wt: 50 pA, 100 ms;
GluA2L483Y/wt: 25 pA, 100 ms. (F) Representative traces of direct responses
from CA1 pyramidal neurons elicited by paired UV photolysis of caged glutamate. Calibration: GluA2wt/wt: 50 pA, 20 ms; GluA2L483Y/wt: 100 pA, 20 ms.
(G) Summary of all experiments.
To directly test for alterations in desensitization of postsynaptic
receptors without the complicating variable of synaptic release,
we probed AMPA receptor depression during activation by UV
photolysis of caged glutamate. We used pairs of flashes from an
UV laser to uncage glutamate over the same area of a neuron. We
found that, at the shortest intervals (<40 ms), there was a clear
difference in the paired photolysis ratio in GluA2L483Y/wt mice
(Fig. 7 F and G). At both 20 ms and 30 ms intervals, the AMPA
receptor response in WT littermate mice demonstrated depression, whereas little depression was observed in GluA2L483Y/wt
(n = 6), suggesting that the presence of nondesensitizing AMPA
receptors increased this ratio when receptors were activated repetitively over a short time window. However, at intervals of ≥40
ms, there was no difference in paired photolysis ratios, suggesting
that receptor desensitization plays a significant role only when
AMPA receptors are activated at the shortest intervals.
Christie et al.
Preferential Distribution of Receptors to Synaptic Sites. Both GluA1
and GluA2 expression was decreased in hippocampal homogenates, whereas GluN1 expression was elevated. Despite this, we
found only small differences in basal synaptic transmission in
GluA2L483Y/wt mice. I/O curves in the CA1 of the hippocampus
were not altered, and mEPSC amplitudes were unaffected, suggesting that AMPA receptors are preferentially targeted to synaptic sites. In agreement with this, we observed a significant reduction in extrasynaptic receptors on CA1 neurons. Previous
studies in GluA1 knockout mice reported similar effects on the
distribution of AMPA receptors (18); when GluA1 was ablated
synaptic AMPA receptors are not significantly altered, but extrasynaptic receptor density is reduced. Similarly, knockout of the
primary hippocampal TARP (transmembrane AMPA receptor
regulatory protein) γ-8 resulted in a relatively small reduction in
the synaptic distribution of AMPA receptors, but a significant alteration in extrasynaptic receptors (19). Therefore, our data are
consistent with a preferential targeting of AMPA receptors to
synapses at the expense of extrasynaptic receptor density.
AMPA Receptors Do Not Accumulate in the ER. The L483Y mutation
lies at the dimer interface between adjacent subunits in the receptor complex. Stabilization of this dimer interface caused by the
mutation at this site eliminates the ability of the receptor to desensitize (28). Expression studies have determined that GluA2
(L483Y) mutant receptors can assemble efficiently, yet their exit
from the ER is considerably reduced, suggesting that conformational changes are used by ER quality control mechanisms for
further processing of AMPA receptors (15). We postulated that
a similar retention of nondesensitizing GluA2(L483Y) receptor
subunits could cause retention of AMPA receptors in the ER in
the knock-in mice. We found there was no increase in the immature glycosylated form of the receptor subunit and no enhancement of the UPR in GluA2L483Y/wt, which might be expected to be
engaged if misfolded proteins were stressing the ER. Furthermore,
we found no enhancement of the interaction between GluA2
AMPA receptor subunits and the ER resident chaperone molecule calnexin, an interaction that we might also expect to be enhanced if misfolded GluA2 receptors were being improperly
processed in neurons. This suggests that introduction of this
mutation in vivo does not cause accumulation of AMPA receptors
in intracellular compartments, unlike when GluA2(L483Y) is
overexpressed in neurons. This is likely because heterotetrameric
assemblies of mutant and WT receptors assemble and traffic differently from homomeric GluA2(L483Y) receptors, which are in
large abundance when introduced exogenously (29).
PNAS Early Edition | 5 of 6
NEUROSCIENCE
Discussion
In this study, we generated a mutant mouse in which a single codon
mutation produced an amino acid switch in the S1 domain of the
GluA2 AMPA receptor subunit. Although heterozygous mice survived past birth, they displayed developmental deficits, a progressive
proclivity for seizures, and early postnatal mortality. The overall
effect of this single amino acid change was greater than that observed when GluA2 was completely ablated in GluA2 knockout
mice (20) or even when two of the major AMPA receptor subunits
were ablated in GluA2/3 double knockout mice (25). Interestingly,
a superficially similar gross phenotype was observed in mutant mice
with a deletion of the intronic editing complementary sequence in
the Gria2 gene (GluA2ΔECS) (26), although the cellular and synaptic
phenotype seemed to differ in this case (13). A recent study reported
that a novel polypeptide snail toxin that inhibits AMPA receptor
desensitization caused profound excitotoxicity, highlighting the
importance of desensitization for neuronal viability (27). The
striking phenotype engendered in GluA2L483Y/wt mice clearly demonstrates that AMPA receptor desensitization is critical for viability of the animal.
Synaptic AMPA Receptor Desensitization. The extent of desensitization
of AMPA receptors at synapses and the role that this plays in the
fidelity of synaptic transmission and plasticity has been studied in
various regions of the brain. Several studies have determined that
postsynaptic receptor desensitization plays a role in shaping the
response to repeated stimuli in the retinogeniculate synapse (4),
the mossy fiber to granule cell synapse in the cerebellum (7), and
calyceal synapses in the auditory brainstem (30). These synapses
are all specialized structures with multiple release sites and delayed glutamate clearance mechanisms, which undoubtedly contribute significantly to AMPA receptor desensitization. In hippocampal synapses early work had suggested that here, too, AMPA
receptors desensitized significantly during bursts of presynaptic
activity (22). However, a later study found that under conditions of
high-release probability, cyclothiazide, which blocks AMPA receptor desensitization, had no effect on paired pulse responses of
CA1 synapses. We found that paired pulse ratios were increased in
GluA2L483Y/wt mice. An independent measure of release probability suggested that there was no measurable difference in glutamate release at CA1 synapses, leaving us to conclude that the
enhanced ratio resulted from postsynaptic mechanisms. A recent
interesting study demonstrated that rapid lateral diffusion of
AMPA receptors out of synapses played a significant role in
responding to multiple closely timed presynaptic release events. In
particular exchange of desensitized receptors could contribute to
the paired pulse ratio (23). Our findings suggest that the amount of
receptor desensitization contributes to the changes in paired pulse
ratio in GluR2L483Y/wt CA1 synapses. Therefore, given that in vivo,
particularly during early development in the neonate, CA1 synapses receive bursts of synaptic activity, it is likely that repetitive
activation of receptors will result in an enhancement of the postsynaptic response (31). In complementary experiments, the paired
ratio to uncaged glutamate over a small area of the somatodendritic region of pyramidal neurons was also larger in mutant mice.
This demonstrated that reduced receptor desensitization was ap1. Hollmann M, Heinemann S (1994) Cloned glutamate receptors. Annu Rev Neurosci 17:
31–108.
2. Hansen KB, Yuan H, Traynelis SF (2007) Structural aspects of AMPA receptor activation,
desensitization and deactivation. Curr Opin Neurobiol 17:281–288.
3. Arai AC, Kessler M (2007) Pharmacology of ampakine modulators: From AMPA
receptors to synapses and behavior. Curr Drug Targets 8:583–602.
4. Chen C, Blitz DM, Regehr WG (2002) Contributions of receptor desensitization and
saturation to plasticity at the retinogeniculate synapse. Neuron 33:779–788.
5. Rozov A, Jerecic J, Sakmann B, Burnashev N (2001) AMPA receptor channels with
long-lasting desensitization in bipolar interneurons contribute to synaptic depression
in a novel feedback circuit in layer 2/3 of rat neocortex. J Neurosci 21:8062–8071.
6. Trussell LO, Zhang S, Raman IM (1993) Desensitization of AMPA receptors upon multiquantal neurotransmitter release. Neuron 10:1185–1196.
7. Xu-Friedman MA, Regehr WG (2003) Ultrastructural contributions to desensitization
at cerebellar mossy fiber to granule cell synapses. J Neurosci 23:2182–2192.
8. Debanne D, Guérineau NC, Gähwiler BH, Thompson SM (1996) Paired-pulse facilitation
and depression at unitary synapses in rat hippocampus: Quantal fluctuation affects
subsequent release. J Physiol 491:163–176.
9. DiGregorio DA, Rothman JS, Nielsen TA, Silver RA (2007) Desensitization properties of
AMPA receptors at the cerebellar mossy fiber granule cell synapse. J Neurosci 27:
8344–8357.
10. Hjelmstad GO, Isaac JT, Nicoll RA, Malenka RC (1999) Lack of AMPA receptor
desensitization during basal synaptic transmission in the hippocampal slice. J Neurophysiol
81:3096–3099.
11. Stern-Bach Y, Russo S, Neuman M, Rosenmund C (1998) A point mutation in the
glutamate binding site blocks desensitization of AMPA receptors. Neuron 21:907–918.
12. Schwenk F, Baron U, Rajewsky K (1995) A cre-transgenic mouse strain for the ubiquitous
deletion of loxP-flanked gene segments including deletion in germ cells. Nucleic Acids
Res 23:5080–5081.
13. Feldmeyer D, et al. (1999) Neurological dysfunctions in mice expressing different
levels of the Q/R site-unedited AMPAR subunit GluR-B. Nat Neurosci 2:57–64.
14. Mohn AR, Gainetdinov RR, Caron MG, Koller BH (1999) Mice with reduced NMDA
receptor expression display behaviors related to schizophrenia. Cell 98:427–436.
15. Greger IH, Akamine P, Khatri L, Ziff EB (2006) Developmentally regulated, combinatorial
RNA processing modulates AMPA receptor biogenesis. Neuron 51:85–97.
16. Crair MC, Malenka RC (1995) A critical period for long-term potentiation at thalamocortical
synapses. Nature 375:325–328.
17. Kamboj SK, Swanson GT, Cull-Candy SG (1995) Intracellular spermine confers rectification
on rat calcium-permeable AMPA and kainate receptors. J Physiol 486:297–303.
6 of 6 | www.pnas.org/cgi/doi/10.1073/pnas.0908206107
parent in GluA2L483Y/wt mice. We also expected to see changes in
receptor deactivation that might be apparent in the slowing of
quantal EPSCs. Surprisingly, we did not see any change in the
kinetics of mEPSCs in the CA1 region. Analysis of these events can
be complicated by dendritic filtering; therefore we also recorded
desynchronized quantal events at mossy fiber synapses (Fig. S2).
However, again there was no observed difference in deactivation
kinetics. The degree of deactivation introduced by the L483Y
mutation will be dependent on the stoichiometry of heteromeric
channels (32), possible association with auxiliary proteins (33, 34),
and the number of receptors containing mutant subunits; therefore, it is possible that the lack of observed effect on deactivation
reflects the fact that there are likely many receptors containing
only a single or no mutant subunits in GluR2L483Y/wt mice. In
conclusion, we demonstrate here that AMPA receptor desensitization is a critical mechanism for proper development and, ultimately, for the survival of the organism.
Materials and Methods
Slice Preparation and Electrophysiology. Transverse hippocampal slices (350
μM) were prepared from postnatal day 14 (P14) to P21 GluR2L483Y/wt (detailed in SI Text). Voltage clamp recordings were made from visually identified CA1 pyramidal neurons and synaptic currents were evoked in the
Schaffer collateral pathway. For UV photolysis of caged glutamate, direct
current responses were measured by uncaging glutamate directly over the
pyramidal cell body UV power was calibrated to give an initial current amplitude of between 150 and 200 pA.
Biochemical Analysis. Biochemical analysis is detailed in SI Materials and Methods.
ACKNOWLEDGMENTS. We thank Stephanie Bennett and Diya Zhang for
mouse ES cell culture, Lynn Doglio for blastocyst injections, Wensheng Liu
and Lennell Reynolds for help with EM, Herman Fernandes for comments,
and Geoffrey Swanson and Yael Stern-Bach for valuable discussion. Support
was provided by the Fragile X research foundation (to A.C.) and the National
Institutes of Health, National Institute of Neurological Disorders and Stroke
(R01NS058894 to A.C.).
18. Zamanillo D, et al. (1999) Importance of AMPA receptors for hippocampal synaptic
plasticity but not for spatial learning. Science 284:1805–1811.
19. Rouach N, et al. (2005) TARP gamma-8 controls hippocampal AMPA receptor number,
distribution and synaptic plasticity. Nat Neurosci 8:1525–1533.
20. Jia Z, et al. (1996) Enhanced LTP in mice deficient in the AMPA receptor GluR2.
Neuron 17:945–956.
21. Koike-Tani M, Kanda T, Saitoh N, Yamashita T, Takahashi T (2008) Involvement of
AMPA receptor desensitization in short-term synaptic depression at the calyx of Held
in developing rats. J Physiol 586:2263–2275.
22. Arai A, Lynch G (1998) AMPA receptor desensitization modulates synaptic responses
induced by repetitive afferent stimulation in hippocampal slices. Brain Res 799:
235–242.
23. Heine M, et al. (2008) Surface mobility of postsynaptic AMPARs tunes synaptic
transmission. Science 320:201–205.
24. Rosenmund C, Clements JD, Westbrook GL (1993) Nonuniform probability of
glutamate release at a hippocampal synapse. Science 262:754–757.
25. Meng Y, Zhang Y, Jia Z (2003) Synaptic transmission and plasticity in the absence of
AMPA glutamate receptor GluR2 and GluR3. Neuron 39:163–176.
26. Brusa R, et al. (1995) Early-onset epilepsy and postnatal lethality associated with an
editing-deficient GluR-B allele in mice. Science 270:1677–1680.
27. Walker CS, et al. (2009) A novel Conus snail polypeptide causes excitotoxicity by
blocking desensitization of AMPA receptors. Curr Biol 19:900–908.
28. Sun Y, et al. (2002) Mechanism of glutamate receptor desensitization. Nature 417:
245–253.
29. Shanks NF, Maruo T, Farina AN, Ellisman MH, Nakagawa T (2010) Contribution of the
global subunit structure and stargazin on the maturation of AMPA receptors. J
Neurosci 30:2728–2740.
30. Otis T, Zhang S, Trussell LO (1996) Direct measurement of AMPA receptor desensitization
induced by glutamatergic synaptic transmission. J Neurosci 16:7496–7504.
31. Menendez de la Prida L, Sanchez-Andres JV (2000) Heterogeneous populations of
cells mediate spontaneous synchronous bursting in the developing hippocampus
through a frequency-dependent mechanism. Neuroscience 97:227–241.
32. Robert A, Irizarry SN, Hughes TE, Howe JR (2001) Subunit interactions and AMPA
receptor desensitization. J Neurosci 21:5574–5586.
33. Priel A, et al. (2005) Stargazin reduces desensitization and slows deactivation of the
AMPA-type glutamate receptors. J Neurosci 25:2682–2686.
34. Schwenk J, et al. (2009) Functional proteomics identify cornichon proteins as auxiliary
subunits of AMPA receptors. Science 323:1313–1319.
Christie et al.