Developmental regulation of protein interacting with C kinase 1 (PICK1) function in hippocampal synaptic plasticity and learning.

Abstract

AMPA-type glutamate receptors (AMPARs) mediate the majority of fast excitatory neurotransmission in the mammalian central nervous system. Modulation of AMPAR trafficking supports several forms of synaptic plasticity thought to underlie learning and memory. Protein interacting with C kinase 1 (PICK1) is an AMPAR-binding protein shown to regulate both AMPAR trafficking and synaptic plasticity at many distinct synapses. However, studies examining the requirement for PICK1 in maintaining basal synaptic transmission and regulating synaptic plasticity at hippocampal Schaffer collateral-cornu ammonis 1 (SC-CA1) synapses have produced conflicting results. In addition, the effect of PICK1 manipulation on learning and memory has not been investigated. In the present study we analyzed the effect of genetic deletion of PICK1 on basal synaptic transmission and synaptic plasticity at hippocampal Schaffer collateral-CA1 synapses in adult and juvenile mice. Surprisingly, we find that loss of PICK1 has no significant effect on synaptic plasticity in juvenile mice but impairs some forms of long-term potentiation and multiple distinct forms of long-term depression in adult mice. Moreover, inhibitory avoidance learning is impaired only in adult KO mice. These results suggest that PICK1 is selectively required for hippocampal synaptic plasticity and learning in adult rodents.

Full text

Page 1

Developmental regulation of protein interacting with
C kinase 1 (PICK1) function in hippocampal synaptic
plasticity and learning
Lenora Volka,b,1, Chong-Hyun Kimc,d,1, Kogo Takamiyae, Yilin Yub, and Richard L. Huganira,b,2
aHoward Hughes Medical Institute andbDepartment of Neuroscience, The Johns Hopkins School of Medicine, Baltimore, MD 21205;cCenter for Neural
Science, Korea Institute of Science and Technology, Seoul 136-791, Korea;dDepartment of Neuroscience, University of Science and Technology, Daejeon
305-333, Korea; andeDepartment of Integrative Physiology, University of Miyazaki Faculty of Medicine, Miyazaki, Miyazaki 889-1692, Japan
Contributed by Richard L. Huganir, October 29, 2010 (sent for review September 17, 2010)
AMPA-type glutamate receptors (AMPARs) mediate the majority of
fast excitatory neurotransmission in the mammalian central ner-
vous system. Modulation of AMPAR trafficking supports several
forms of synaptic plasticity thought to underlie learning and mem-
ory. Protein interacting withC kinase1 (PICK1) is an AMPAR-binding
protein shown to regulate both AMPAR trafficking and synaptic
plasticity at many distinct synapses. However, studies examining
the requirement for PICK1 in maintaining basal synaptic transmis-
sion and regulating synaptic plasticity at hippocampal Schaffer
collateral–cornu ammonis 1 (SC–CA1) synapses have produced con-
flicting results. In addition, the effect of PICK1 manipulation on
learningandmemoryhasnotbeeninvestigated.Inthepresentstudy
weanalyzedtheeffectofgeneticdeletionofPICK1onbasalsynaptic
transmission and synaptic plasticity at hippocampal Schaffer collat-
eral–CA1 synapses in adult and juvenile mice. Surprisingly, we find
that loss of PICK1 has no significant effect on synaptic plasticity in
juvenile mice but impairs some forms of long-term potentiation and
multiple distinct forms of long-term depression in adult mice. More-
over,inhibitoryavoidancelearningisimpairedonlyinadultKOmice.
TheseresultssuggestthatPICK1isselectivelyrequiredforhippocam-
pal synaptic plasticity and learning in adult rodents.
AMPA receptor|membrane trafficking|endocytosis|receptor recycling|
metabotropic glutamate receptor
A
CNS. Modulation of AMPAR trafficking and expression at syn-
apseshasemergedasakeyregulatorofsynapticplasticity,thought
to be a cellular mechanism underlying learning and memory (re-
viewed in ref. 1). AMPAR trafficking is a highly dynamic process,
and the interaction of AMPARs with various AMPAR-binding
proteins (AMPAR BPs) plays a critical role in regulating the
movementofAMPARsinandoutofthesynapse(1–3).Ofthefour
AMPAR subunits [AMPA-type glutamate receptors 1–4 (GluA1–
4), previously glutamate receptors 1–4 (GluR1–4); see ref. 4],
GluA2, GluA3, and GluA4-short have short C-terminal tails con-
tainingatypeIIpostsynapticdensity95(PSD95)/discslarge(DlgA)/
zonula occludens-1 (Zo-1) (PDZ) ligand. This C-terminal PDZ li-
gand of GluA2/3 interacts with PDZ domains in multiple AMPAR
BPs, including protein interacting with C kinase 1 (PICK1) and
glutamatereceptor-interactingprotein1and2(GRIP1andGRIP2,
GRIP2 is also called AMPAR binding protein, or ABP) (5–8). C-
terminal phosphorylation of GluA2 at serine-880 (S880) differen-
tially regulates interaction with PICK1 and GRIP such that S880
phosphorylation disrupts GluA2 binding to GRIP1/2 but does not
affect GluA2 binding to PICK1 (9, 10).
PICK1 is an AMPAR BP (5, 8) originally identified by its in-
teraction with PKC α (11). The PICK1 domain structure includes
an N-terminal PDZ domain, a central box-dependent myc-inter-
acting protein-1 (Bin)/amphiphysin/reduced viability to nutrient
starvation-homology (Rvs) (BAR) domain, and N- and C-terminal
acidic domains (reviewed in ref. 12). Although the precise manner
MPA-type glutamate receptors (AMPARs) mediate the
majority of fast excitatory transmission in the mammalian
in which GRIP and PICK regulate basal and activity-dependent
AMPAR trafficking is still under investigation, GRIP appears to
stabilizeGluA2-containingAMPARsatthecellsurface(13)and/or
facilitate recycling ofinternalizedreceptors back to the cellsurface
(14), whereas association with PICK1 facilitates AMPAR endocy-
tosis (15–17), inhibits receptor recycling (14), or retains GluA2-
containing receptors at extrasynaptic sites (14, 18, 19). Based on
thesedata,phosphorylationofGluA2atS880hasbeenproposedto
facilitate long-term depression (LTD) by inducing a switch from
GluA2–GRIP to GluA2–PICK interaction (16, 20, 21).
A role for PICK1 in synaptic plasticity has been established
most clearly in the cerebellum, where LTD at multiple synapses
requires intact PICK1 function (18, 22, 23). PICK1 also plays an
essential role in a developmentally regulated form of LTD at
hippocampal mossy fiber–cornu ammonis 3 (CA3) synapses (24).
At hippocampal Schaffer collateral–cornu ammonis 1 (SC–CA1)
synapsesnumerousstudiesfindaroleforPICK1inregulatingLTD
(20, 25–27), although the level of impairment upon disruption of
PICK1 function varies considerably, and contradictory findings
exist (28). In addition, a few studies suggest that PICK1 may reg-
ulate activity-dependent insertion of AMPARs (29, 30), and a re-
centstudyimplicatesPICK1inmaintaininglong-termpotentiation
(LTP) at SC–CA1 synapses in the hippocampus (27).
In the present study we investigated the effect of genetic de-
letion of PICK1 on basal SC–CA1 synaptic transmission and
synaptic plasticity in adult and juvenile mice. We also determined
whetherPICK1isrequired forinhibitoryavoidance (IA)learning.
Surprisingly, we found that loss of PICK1 had no significant effect
on synaptic plasticity or IA learning in juvenile mice but resulted
in impaired synaptic plasticity and IA learning in adult mice.
Results
Basal Synaptic Transmission Is Unaffected in Adult PICK1-KO Mice. In
juvenile rodents, acute disruption of PICK1 function with a pep-
tide that mimics the phosphorylated C terminus of GluA2 or via
expression of a PICK1 mutant that is deficient for lipid binding
results in a gradual increase in synaptic responses at hippocampal
SC–CA1 synapses in most (20, 26, 31) but not all (28) studies.
Conversely, overexpression of PICK1 in hippocampal neurons
results in a decrease in surface GluA2 expression (19, 26, 32)
Author contributions: L.V., C.-H.K., K.T., and R.L.H. designed research; L.V., C.-H.K., K.T.,
and Y.Y. performed research; K.T. and Y.Y. contributed new reagents/analytic tools; L.V.,
C.-H.K., and R.L.H. analyzed data; and L.V., C.-H.K., and R.L.H. wrote the paper.
Conflict of interest statement: Under a licensing agreement between Millipore Corpora-
tion and The Johns Hopkins University, R.L.H. is entitled to a share of royalties received by
the University on sales of products described in this article. R.L.H. is a paid consultant to
Millipore Corporation. The terms of this arrangement are being managed by The Johns
Hopkins University in accordance with its conflict of interest policies.
1L.V. and C.-H.K. contributed equally to this work.
2To whom correspondence should be addressed. E-mail: rhuganir@jhmi.edu.
This article contains supporting information online at www.pnas.org/lookup/suppl/doi:10.
1073/pnas.1016103107/-/DCSupplemental.
21784–21789
| PNAS
| December 14, 2010
| vol. 107
| no. 50www.pnas.org/cgi/doi/10.1073/pnas.1016103107

Page 2

consistent with a model in which PICK1 functions to retain
GluA2-containing receptors in an intracellular or extrasynaptic
pool. However, the role for PICK1 in adult neuronal function has
not been investigated. Thus, we examined multiple measures of
basal synaptic transmission at SC–CA1 synapses using acute slices
from adult (2- to 3-mo-old) mice in which PICK1 expression had
been genetically ablated (see ref. 18).
Input–output curves, obtained by plotting the amplitude of the
fiber volley vs. the slope of the field excitatory postsynaptic po-
tential (fEPSP) at various stimulation intensities, were not dif-
ferent in adult PICK1-KO mice compared with their WT
littermates (WT: 2.97 ± 0.21 ms−1, n = 19; KO: 2.57 ± 0.22 ms−1,
n = 19; P > 0.05) (Fig. S1A). In addition to interacting with
postsynaptic AMPARs, PICK1 also binds to and regulates the
expression of a presynaptic metabotropic glutamate receptor,
mGluR7, which modulates neurotransmitter release (33–35).
However, analysis of paired-pulse facilitation (PPF) revealed no
difference between PICK1-KO and WT mice [WT, n = 19, KO,
n = 20; P > 0.05 at all interstimulus intervals (ISI)] (Fig. S1B),
suggesting that basal presynaptic release is not altered in adult
PICK1-KO mice. To assess synaptic transmission at the level of
individual synapses, we recorded miniature excitatory post-
synaptic currents (mEPSCs) resulting from action potential-in-
dependent spontaneous glutamate release. mEPSC amplitude
and frequency were unaffected in adult PICK1-KO mice [WT:
amplitude = 12.8 ± 0.40 pA, frequency = 3.34 ± 0.33 Hz, n = 25
cells;KO:amplitude = 12.2±0.44pA(P> 0.3),frequency = 2.83
± 0.27 Hz (P > 0.2), n = 22 cells] (Fig. S1C).
AMPARsconsistprimarilyofGluA1/GluA2heteromersatSC–
CA1 synapses (36). The presence of GluA2 in AMPARs not
only regulates association with AMPAR BPs, but also renders
AMPARs impermeable to Ca2+and eliminates inward rectifica-
tion measured by the current–voltage (I–V) relationship of a syn-
apse. To examine the effect of PICK1 deletion on the basal
composition of AMPA receptors, we measured the I–V relation-
ship at SC inputs to CA1 pyramidal neurons. The synaptic I–V
relationship wasnot affectedin adult PICK1-KO mice(Fig.S1D).
Additionally, using outside-out patches from somatic membranes
ofCA1pyramidalcells,wefoundnodifferenceintheamplitudeof
current induced by 20 μM AMPA at −70 mV or in the I–V re-
lationship between PICK1WT and KO mice (at −70 mV, WT: 213
± 28 pA; KO: 172 ± 20 pA; P = 0.12) (Fig. S1E), demonstrating
that the composition of extrasynaptic AMPA receptors is not al-
teredinPICK1-KOmice.Insummary,basalsynaptictransmission
at SC–CA1 synapses appears normal in adult PICK1-KO mice.
Multiple Forms of LTD Are Impaired in Adult PICK1-KO Mice. Con-
sistent with a role for PICK1 in retaining intracellular AMPARs
after activity-dependent endocytosis (14)or in regulating AMPAR
endocytosis (15–17), we found that NMDA receptor (NMDAR)-
dependentLTDinducedwithastandardlow-frequencystimulation
(LFS) protocol and measured by monitoring extracellular fEPSPs
(Fig. 1A) or with whole-cell pairing of −40 mV depolarization with
brief LFS (Fig. 1B) was significantly reduced in slices from PICK1-
KO mice. In addition to NMDAR-dependent LTD, PICK1 has
been implicated in mGluR-dependent LTD in the ventral teg-
mental area (37), perirhinal cortex (38), and cerebellum (23) of
immature rodents. Using paired pulse-LFS (PP-LFS) in the pres-
ence of the NMDAR antagonist D,L-2-amino-5-phosphonopenta-
noic acid (D,L-AP5), we found that NMDAR-independent LTD
was reduced in the hippocampus of adult PICK1-KO mice (Fig.
1C). At SC–CA1 synapses this protocol induces LTD that is de-
pendent on group I mGluRs (39–41) and/or muscarinic acetyl-
choline receptors (mAChRs) (42). These data highlight a central
role for PICK1 in multiple forms of hippocampal LTD.
LTP in Adult PICK1-KO Mice Is Reduced with a Subset of Induction
Protocols. In light of recent findings implicating PICK1 in
NMDAR-induced AMPAR insertion and hippocampal LTP in
juvenile mice (27, 30), we examined the requirement for PICK1
in LTP at mature hippocampal synapses. LTP induced by pairing
postsynaptic depolarization to 0 mV with LFS [0.66 Hz at RT
(Fig. 2A) or 2 Hz at 35 °C (Fig. S2)] was significantly reduced in
adult PICK1-KO mice. However, high-frequency stimulation
(HFS; 2× 100 Hz) (Fig. 2B) or single theta-burst stimulation
(TBS) (Fig. 2C) induced LTP that was indistinguishable between
WT and PICK1-KO mice. In addition, LTP induced by a single
HFS at RT (WT: 125 ± 5%, n = 5; KO: 124 ± 12%, n = 6) or
four TBS at 30 °C (WT: 163 ± 6%, n = 8; KO: 166 ± 13%, n = 6)
also failed to reveal a deficit in adult PICK1-KO mice. These
data suggest that PICK1 is not absolutely required for hippo-
campal LTP in adult mice, but under some conditions PICK1 is
required for normal LTP induction. Given the absolute re-
quirement for PICK1 in LTP observed in juvenile mice (27), we
wondered if the induction protocol-specific deficit seen adult
PICK1-KO mice in our study reflected a developmental differ-
ence in the induction/expression mechanism of LTP in PICK1-
KO mice. Thus, we examined basal synaptic transmission as well
as LTP and LTD in juvenile PICK1-KO mice.
Basal Synaptic Transmission Is Modestly Affected in Hippocampal CA1
Pyramidal Neurons of Juvenile PICK1-KO Mice. Input–output curves
(WT: 2.53 ± 0.13 ms−1, n = 47; KO: 2.39 ± 0.12 ms−1, n = 41;
P > 0.05) (Fig. S3A), PPF (WT, n = 47; KO, n = 41; P > 0.05 at
all ISI) (Fig. S3B), and mEPSC amplitude (WT: 16.6 ± 0.47 pA,
C
fEPSP slope, % baseline
0
25
50
75
100
Time (min)
-20020406080
PP-LFS
KO
WT
KO
WT
100µM AP5
WT
020
Time (min)
WT
KO
-40mV, 0.5-1Hz
50
100
EPSC amplitude, % baseline
B
1030
KO
25
75
60
80
100
120
-200 2040
60
WT
KO
LFS
A
fEPSP slope, % baseline
Time (min)
KO
WT
Fig. 1.
induced with LFS (1 Hz, 900 stimuli, 30 °C) is significantly reduced in PICK1-
KO mice. WT (n = 8): 80 ± 1% at 55–60 min; KO (n = 9): 92 ± 3% at 55–60 min;
P < 0.01. (B) LTD induced by pairing 200–300 pulses at 0.5–1 Hz with −40 mV
depolarization (35 °C) is reduced in PICK1-KO mice. WT (n = 10): 59 ± 5% at
31–35 min; KO (n = 11): 76 ± 4% at 31–35 min; P < 0.05. (C) NMDA receptor-
independent LTD induced by PP-LFS (50-ms ISI, 30 °C) in the NMDAR an-
tagonist D,L-AP5 (100 μM) is reduced in PICK1-KO mice. WT (n = 9): 79 ± 2%
at 75–80 min; KO (n = 8): 89 ± 2% at 75–80 min; P = 0.01. (Inset scale bars: A
and C: 0.5 mV, 5 ms; B: 100 pA, 10 ms.)
Multiple forms of LTD are impaired in adult PICK1-KO mice. (A) LTD
Volk et al. PNAS
| December 14, 2010
| vol. 107
| no. 50
| 21785
NEUROSCIENCE

Page 3

n = 35 cells; KO: 15.35 ± 0.35 pA, n = 37 cells; P > 0.1) (Fig.
S3C) in juvenile PICK1-KO mice were not different from WT.
However, mEPSC frequency was decreased in juvenile PICK1-
KO mice (WT: 5.76 ± 0.32 Hz; KO: 3.46 ± 0.23 Hz; P < 0.001)
(Fig S3C). Changes in mEPSC frequency may result from
changes in presynaptic release probability or from changes in the
number of synapses. Given that there is no evidence of altered
release probability with evoked responses (Fig. S3B), these data
suggest that either the total number of synapses is reduced or the
proportion of silent (AMPAR-lacking) synapses is enhanced in
juvenile PICK1-KO mice. Alternatively, PICK1 deletion may de-
crease spontaneous but not evoked neurotransmitter release
probability (43, 44).
LTP Is Intact in Juvenile PICK1-KO Mice. Surprisingly, we found that
LTP induced with numerous protocols and under multiple re-
cording conditions was unaffected in juvenile PICK1-KO mice
(Fig. 3 and SI Methods). LTP induced with a whole-cell pairing
protocol was indistinguishable between WT and PICK1-KO mice
(Fig. 3A). Similarly, using extracellular recordings, HFS (Fig. 3B)
or TBS (Fig. 3C) induced LTP that was indistinguishable between
WT and PICK1-KO mice.
LTD Is Unaffected in Juvenile PICK1-KO Mice. Considering the nu-
merous studies implicating PICK1 in AMPAR endocytosis or in
maintaining extrasynaptic pools of GluA2-containing AMPARs,
we were surprised to find no significant impairment of LTD in
juvenile PICK1-KO mice. LTD induced with a standard LFS
protocol and measured by monitoring extracellular fEPSPs (Fig.
4A) and LTD induced with a whole-cell pairing protocol (Fig.
4B) were not significantly affected by loss of PICK1.
IA Learning Is Selectively Impaired in Adult PICK1-KO Mice. In lightof
the developmental differences observed in hippocampal synap-
tic plasticity in PICK1-KO mice, we wanted to determine if a
hippocampal-dependent learning task would reflect this age-
dependentroleforPICK1.Wechosetouseastep-throughIAtask
because IA learning is dependent on hippocampal function (45–
47)and induces AMPAR trafficking and LTP in the hippocampus
in vivo (48). Additionally, because this task requires only a single-
trial training period, it allows reliable training in a narrow age
range in juvenile mice. Memory retention was assessed 24 h after
training. Strikingly, adult PICK1-KO mice failed to learn the IA
task, but the performance of juvenile PICK1-KO mice was in-
distinguishable from WT (Fig. 5).
Discussion
In this study we demonstrate a developmentally regulated role for
PICK1 in hippocampal synaptic plasticity and learning. We found
that the requirement for PICK1 in hippocampal LTP is limited to
a subset of LTP induction protocols in adult mice. In contrast,
PICK1 appears to be generally required for LTD in the adult hip-
pocampus. Both NMDAR-dependent and mGluR/mAChR-
C
Time (min)
0204060
50
100
150
200
WT
KO
fEPSP slope, % baseline
1x
WTKO
A
50
100
150
200
1020
30
40
WT
KO
EPSC amp, % baseline
Time (min)
0mV,
0.66Hz,
0
WTKO
B
50
100
150
200
250
-200 2040 60
KO
WT
Time (min)
fEPSP slope, % baseline
HFS
WT
KO
Fig. 2.
PICK1-KO mice. (A) LTP induced by pairing 120 pulses at 0.66 Hz with 0 mV
depolarization (RT) is reduced in PICK1-KO mice: WT (n = 10): 185 ± 13%; KO
(n = 14): 142 ± 10%; P < 0.05. (B) LTP induced by HFS (2 × 100 Hz, 20-s inter-
burst interval, RT) is unaffected in PICK1-KO mice. At 55–60 min after LTP
induction, WT (n = 6): 123 ± 9%; KO (n = 5): 123 ± 4%; P > 0.5. (C) LTP
induced by a single theta burst (35 °C) is unaffected in PICK1-KO mice. At
56–60 min after induction, WT (n = 7): 159 ± 7%; KO (n = 6): 161 ± 7%; P >
0.5. (Inset scale bars: A: 100 pA, 20 ms; B and C: 0.5 mV, 5 ms.)
LTP is impaired in an induction protocol-dependent manner in adult
Time (min)
50
100
150
200
250
-20 0 2040
60
KO
WT
HFS
fEPSP slope, % baseline
B
??
??
KO, LTP path
KO, control path
Time (min)
A
fEPSP slope, % baseline
C
-200204060
50
100
150
200
WT
KO
Time (min)
1xTBS
WT
KO
WT, LTP path
WT, control path
EPSC amplitude, % baseline
WT
KO
020
0
200
400
600
0mV-2Hz
10
30
Fig. 3.
a whole-cell pairing protocol (200 pulses, 2 Hz, at 0 mV, 35 °C) is unaffected
in PICK1-KO mice. At 26–30 min after pairing, WT (n = 5): 383 ± 70%; KO (n =
5): 365 ± 70%; P > 0.5. (B) LTP induced by HFS (2 × 100 Hz, 20-s inter-burst
interval, RT) is unaffected in PICK1-KO mice. At 55–60 min after LTP in-
duction, WT (n = 24): 126 ± 3%; KO (n = 23): 123 ± 3%; P > 0.5. (C) LTP
induced by a single theta burst (35 °C) is unaffected in PICK1-KO mice. At 56–
60 min after TBS, WT (n = 10): 133 ± 4%; KO (n = 8): 136 ± 6%; P > 0.5. (Inset
scale bars: A: 100 pA, 20 ms; B and C: 0.5 mV, 5 ms.)
LTP is unaffected in juvenile PICK1-KO mice. (A) LTP induced by
21786
| www.pnas.org/cgi/doi/10.1073/pnas.1016103107 Volk et al.

Page 4

dependent LTD were impaired in adult PICK1-KO mice. Basal
transmission was unaffected in adult PICK1-KO mice. Thus, the
observed deficits in synaptic plasticity are likely attributable to
a direct role for PICK1 in hippocampal synaptic plasticity. To our
surprise,whenweexaminedhippocampalLTPandLTDinjuvenile
mice, we observed no effect of PICK1 deletion. Additionally, IA
learning was selectively impaired in adult PICK1-KO mice, sup-
porting anage-dependent role forPICK1 innormal brainfunction.
Regulation of Basal AMPAR Trafficking and Composition by PICK1. In
neurons of juvenile mice, acute overexpression of PICK1 causes
a decrease in surface expression of GluA2 (19, 26, 32), whereas
acute disruption of PICK1 function results in an increase in basal
synaptic transmission (20, 26, 31, but see ref. 28), consistent with
a model in which, under basal conditions, PICK1 functions to
sequester and stabilize an intracellular/extrasynaptic pool of
GluA2-containing AMPARs (14) or in which PICK1 inhibits
recycling of GluA2-containing AMPARs (15–17). In the current
study we find no alterations in basal synaptic transmission in
adult PICK1-KO mice and only modest changes in juvenile mice.
This result may indicate that homeostatic changes are largely
able to compensate for the chronic loss of PICK1, resulting in
normal steady-state levels of synaptic AMPAR expression.
In juvenile mice, we observe that chronic loss of PICK1 results
in a decrease in mEPSC frequency. PICK1 is expressed pre-
synaptically and has been shown to regulate trafficking of pre-
synaptic receptors that could influence neurotransmitter release
(33–35, 49–51). However, our finding that PPF, a measure corre-
lated with presynaptic vesicle release, is unchanged in juvenile
PICK1-KO mice indicates that the observed decrease in mEPSC
frequency is unlikely to result from a general decrease in pre-
synaptic release probability. Interestingly, there is evidence that
spontaneous (action potential-independent) and evoked (action
potential-dependent)neurotransmitterreleaseusedistinctvesicle
pools (43, 44). Therefore, PICK1 could selectively affect spon-
taneous but not evoked neurotransmitter release in juvenile
mice. Alternatively, the decrease in mEPSC frequency could re-
flect a decrease in the number of total or functional (AMPAR-
containing) synapses in juvenile PICK1-KO mice.
PICK1 in LTD. Our results demonstrate that genetic ablation of
PICK1 results in significant impairment of several forms of LTD
in the mature rodent hippocampus. mGluR/mAChR-dependent
and NMDAR-dependent LTD recruit different signaling mech-
anisms to induce LTD (42, 52–55), but all are expressed via re-
moval of synaptic AMPARs (42, 56–60). The fact that both
NMDAR- and mGluR/mAChR-mediated LTD are impaired in
adult PICK1-KO mice suggests that PICK1 is an essential medi-
ator of activity-dependent AMPAR trafficking, regardless of the
upstream signal transduction pathway recruited.
A recent report showed that intracellular perfusion of a pep-
tide that mimics the S880 phosphorylated GluA2 C-terminal tail,
previously shown to inhibit PICK1 binding to endogenous GluA2,
decreased mGluR- but not NMDAR-dependent LTD in the
perirhinal cortex of immature (postnatal day 7–13) rodents (38).
This result suggests that the broad requirement for PICK1 in
distinct forms of LTD observed in mature mice in our study may
not be generalized across brain regions or developmental age.
Indeed, numerous reports find that in juvenile rodents, mGluR
LTD at SC–CA1 synapses is expressed presynaptically and does
not involve internalization of AMPARs (61–66), making it un-
likely that in immature rodents mGluR LTD at SC–CA1 synapses
requires PICK1 function. Surprisingly, we find that NMDAR-
dependent LTD is not significantly impaired in juvenile PICK1-
KO mice. It is not clear what mediates the observed de-
velopmental switch in PICK1 dependence of LTD. Induction and
expression mechanisms underlying LTD can vary dramatically by
developmental age (24, 25, 64, 67, 68); thus juvenile mice may
express an additional, PICK1-independent form of LTD. Pre-
vious studies investigating the role of PICK1 in LTD at juvenile
SC–CA1 synapses generally report some level of impairment
when PICK1 function is acutely or chronically blocked, but this
impairment ranges from a ∼40–60% reduction in the magnitude
of LTD (20, 25, 26) to complete block of LTD (27), and one study
failed to find that disrupting PICK1–GluA2 interactions had an
effect on LTD (28). These differences are not correlated with the
method of PICK1 manipulation (acute vs. chronic or inhibition of
function vs. removal of PICK1 protein). Taken together, these
data suggest that PICK1 is not absolutely required for LTD at
juvenile SC–CA1 synapses but that under some conditions PICK1
does contribute significantly to LTD induction. It has been sug-
gested that the level of protein phosphatase 1 at synapses might
determine whether PICK1 participates in LTD by affecting basal
levels of GluA2 S880 phosphorylation (see below) (25) and differ-
entslicepreparation,recovery,orrecordingconditionscanresultin
different basal levels of synaptic signaling molecules (69), which
could significantly impact whether PICK1 participates in LTD.
WT
A
KO
75
100
-2002040 60
fEPSP slope, %baseline
Time (min)
WT
KO
LFS
B
WT
KO
EPSC amplitude, % baseline
Time (min)
50
100
-100 10 2030 4050
75
WT KO
-40mV,
0.5-1Hz
Fig. 4.
with LFS (35 °C) is not significantly different in PICK1-KO mice. At 61–65 min
after beginning LTD induction, WT (n = 18): 88 ± 2%; KO (n = 18): 92 ± 2%;
P > 0.1. (B) LTD induced by pairing 200–300 pulses at 0.5–1 Hz with −40 mV
depolarization (35 °C) is unaffected in PICK1-KO mice. At 41–45 min after
beginning LTD induction, WT (n = 6): 78 ± 9%; KO (n = 9): 70 ± 6%; P > 0.4.
(Inset scale bars: A: 0.5 mV, 5 ms; B: 100 pA, 20 ms.)
LTD is modestly affected in juvenile PICK1-KO mice. (A) LTD induced
Inhibitory Avoidance
Adult
0
100
200
300
WT
KO
Step through latency (sec)
Inhibitory Avoidance
Juvenile
0
100
200
WT
KO
Pre-training24hr retention
**
**
ns
ns
**
**
Fig. 5.
not in juvenile PICK1-KO mice. Adult (2- to 3-mo-old) or juvenile (∼3-wk-old;
postnatal day 20–22) mice were trained on a step-through IA task. Latency to
cross over to the dark chamber was measured at training and 24 h later. In
adult mice, PICK1-KO results in a dramatically reduced latency 24 h after
training (WT: n = 10; KO: n = 12). Juvenile PICK1 mice acquire the IA task
normally (WT: n = 7; KO: n = 9). **P < 0.01. ns, not significant.
Hippocampal-dependent learning/memory is impaired in adult but
Volk et al.PNAS
| December 14, 2010
| vol. 107
| no. 50
| 21787
NEUROSCIENCE

Page 5

One mechanism by which PICK1 could regulate hippocampal
LTD involves an activity-dependent increase in the fraction of
GluA2 associated with PICK1 via phosphorylation of GluA2 at
S880 (9, 16, 20, 21, 25). Phosphorylation of GluA2 at this residue
disrupts binding with GRIP but not PICK1, shifting the pre-
dominant PDZ interaction to PICK1. PICK1 may facilitate LTD
by actively participating in AMPAR internalization, as suggested
by the presence of a lipid-binding BAR domain in PICK1 (70).
Alternatively, release of AMPARs from GRIP may be permis-
sive for other molecules that are actively involved in endocytosis,
followed by retention of internalized AMPARs via interaction
with PICK1. Numerous studies find that disruption of PICK1
function blocks AMPAR internalization (15–17). However, it is
difficult, using static methods for visualizing AMPARs, to dis-
tinguish between a direct role for PICK1 in internalization of
receptors vs. retention of internalized receptors. Live imaging of
AMPAR internalization induced by NMDAR stimulation in
neurons cultured from PICK1-KO mice suggests that PICK1 is
not necessary for internalization but functions by retaining in-
ternalized receptors or inhibiting receptor recycling (14).
PICK1 in LTP.Although most studies examining the role of PICK1
in activity-induced AMPAR trafficking and synaptic plasticity
find a role for PICK1 in LTD, several reports suggest that
PICK1 also may participate in AMPAR trafficking required for
LTP (27, 29, 30). Consistent with these findings, we observed
that hippocampal LTP induced with a pairing protocol is im-
paired in adult PICK1-KO mice. However, we note that several
common LTP induction protocols, including HFS and TBS,
failed to reveal a difference between adult PICK1-KO mice and
their WT littermates. Thus, our data indicate that in the mature
hippocampus PICK1 can regulate LTP but is not essential for
LTP induction.
Surprisingly, we find that LTP in juvenile PICK1-KO mice is
indistinguishable from that of WT littermates. To the best of our
knowledge, the slice preparation, LTP protocol, age of animals,
and recording conditions in Fig. 3B are identical to those used by
Terashima et al. (27), who observed a loss of LTP in juvenile
PICK1-KO. The reason for this discrepancy is not clear, because
Terashima and colleagues used the mouse line generated by our
laboratory. It is possible that subtle differences in slice prepa-
ration or recording conditions facilitate plasticity with different
induction or expression requirements. Consistent with this pos-
sibility, previous studies from this group demonstrate that LTP
at hippocampal synapses is mediated by a transient insertion
of GluA2-lacking receptors (71), whereas under the conditions
used in Fig. 3B we find that LTP does not require activity
through GluA2-lacking receptors (determined by selective an-
tagonism of GluA2-lacking receptors immediately after LTP
induction: control: 124 ± 6%, n = 5; 50 μM 1-naphthyl acetyl
spermine (NASPM): 122 ± 7%, n = 5; P > 0.5). The re-
quirement for GluA2-lacking receptors in LTP at SC–CA1 syn-
apses is controversial and may depend on the age of animals and
type of induction protocol used (71–75).
In acutely prepared cortical slices from juvenile PICK1-KO
mice, surface expression of GluA1 is decreased, whereas surface
GluA2 is elevated (76). However, basal synaptic transmission is
unaffected in this preparation, suggesting that the subunit com-
position of extrasynaptic surface pools of AMPARs may be ab-
errant in the absence of PICK1. The GluA1 subunit is thought to
be necessary for activity-dependent incorporation of AMPARs
during LTP, and recent data support the idea that the immediate
source of receptors for LTP may be lateral diffusion from ex-
trasynaptic pools of surface receptors (77, 78). Thus, the re-
distribution of extrasynaptic receptors observed in PICK1-KO
mice could impair the activity-dependent trafficking of GluA1-
containing receptors into the synapse during LTP without af-
fecting basal synaptic transmission. Interestingly, juvenile GluA1-
KOmiceexhibitLTP,buttheabilitytoinduceLTPdeclinesasthe
mice mature (79). Numerous studies report developmental dif-
ferences in the expression mechanisms of LTP (75, 79–81). If the
LTP deficit observed in adult PICK1-KO mice results from de-
pletion of the available pool of extrasynaptic GluA1, LTP in ju-
venile mice may be mediated by a GluA1-independent form of
LTP that exists selectively in young mice (79).
Developmental Regulation of PICK1 Function in Learning and/or
Memory. Our finding that IA learning and/or memory is im-
paired selectively in adult but not juvenile PICK1-KO mice cor-
roborates our electrophysiological data demonstrating develop-
mental regulation of PICK1 function in synaptic plasticity and
suggests that PICK1 function is critical for both hippocampal
synaptic plasticity and memory in adult but not in juvenile mice.
It should be noted that although IA training induces synaptic
plasticity and AMPAR trafficking in the hippocampus, and intact
hippocampal function is required for IA learning, IA learning
is not exclusively dependent on the hippocampus. These data,
however, do support a developmentally regulated requirement
for PICK1 in normal brain function.
Methods
Electrophysiology. Whole-cell or extracellular field recordings were obtained
from acute hippocampal slices prepared from juvenile (2- to 3-wk-old) or
adult (2- to 3-mo-old) PICK1-KO and WT mice.
Behavior.Astandardstep-thoughIAtaskwasusedinwhichmiceinitiallywere
placed inthelightsideofa rectangular chamber consisting ofalight chamber
and a dark chamber separated by a wall with a guillotine-style door. For
training, mice wereplaced inthelight side,andthelatencytostep through to
the dark side was measured. This value was taken as the control, pretraining
value. After crossing to the dark chamber, mice were given a 0.3-mA, 2-s foot
shock. Memory was assessed 24 h later by reintroducing mice to the light side
and measuring the latency to step through to the dark chamber. For more
details, see SI Methods.
1. Song I, Huganir RL (2002) Regulation of AMPA receptors during synaptic plasticity.
Trends Neurosci 25:578–588.
2. Bredt DS, Nicoll RA (2003) AMPA receptor trafficking at excitatory synapses. Neuron
40:361–379.
3. Shepherd JD, Huganir RL (2007) The cell biology of synaptic plasticity: AMPA receptor
trafficking. Annu Rev Cell Dev Biol 23:613–643.
4. Collingridge GL, Olsen RW, Peters J, Spedding M (2009) A nomenclature for ligand-
gated ion channels. Neuropharmacology 56:2–5.
5. Dev KK, Nishimune A, Henley JM, Nakanishi S (1999) The protein kinase C alpha
binding protein PICK1 interacts with short but not long form alternative splice
variants of AMPA receptor subunits. Neuropharmacology 38:635–644.
6. Dong H, et al. (1997) GRIP: A synaptic PDZ domain-containing protein that interacts
with AMPA receptors. Nature 386:279–284.
7. Srivastava S, et al. (1998) Novel anchorage of GluR2/3 to the postsynaptic density by
the AMPA receptor-binding protein ABP. Neuron 21:581–591.
8. Xia J, Zhang X, Staudinger J, Huganir RL (1999) Clustering of AMPA receptors by the
synaptic PDZ domain-containing protein PICK1. Neuron 22:179–187.
9. Chung HJ, Xia J, Scannevin RH, Zhang X, Huganir RL (2000) Phosphorylation of the
AMPA receptor subunit GluR2 differentially regulates its interaction with PDZ
domain-containing proteins. J Neurosci 20:7258–7267.
10. Matsuda S, Mikawa S, Hirai H (1999) Phosphorylation of serine-880 in GluR2 by
protein kinase C prevents its C terminus from binding with glutamate receptor-
interacting protein. J Neurochem 73:1765–1768.
11. Staudinger J, Zhou J, Burgess R, Elledge SJ, Olson EN (1995) PICK1: A perinuclear
binding protein and substrate for protein kinase C isolated by the yeast two-hybrid
system. J Cell Biol 128:263–271.
12. Xu J, Xia J (2006-2007) Structure and function of PICK1. Neurosignals 15:190–201.
13. Osten P, et al. (2000) Mutagenesis reveals a role for ABP/GRIP binding to GluR2 in
synaptic surface accumulation of the AMPA receptor. Neuron 27:313–325.
14. Lin D-T, Huganir RL (2007) PICK1 and phosphorylation of the glutamate receptor 2
(GluR2) AMPA receptor subunit regulates GluR2 recycling after NMDA receptor-
induced internalization. J Neurosci 27:13903–13908.
15. Hanley JG, Henley JM (2005) PICK1 is a calcium-sensor for NMDA-induced AMPA
receptor trafficking. EMBO J 24:3266–3278.
21788
| www.pnas.org/cgi/doi/10.1073/pnas.1016103107Volk et al.