NMDA Receptor-Mediated PIP5K Activation to...
Transcript of NMDA Receptor-Mediated PIP5K Activation to...
Neuron
Article
NMDA Receptor-Mediated PIP5K Activationto Produce PI(4,5)P2 Is Essential for AMPAReceptor Endocytosis during LTDTakamitsu Unoki,1,5 Shinji Matsuda,2,3,5 Wataru Kakegawa,2,3,5 Ngo Thai Bich Van,1,6 Kazuhisa Kohda,2,3 Atsushi Suzuki,1
Yuji Funakoshi,1 Hiroshi Hasegawa,4 Michisuke Yuzaki,2,3,* and Yasunori Kanaho1,*1Department of Physiological Chemistry, Graduate School of Comprehensive Human Sciences and Institute of Basic Medical Sciences,
University of Tsukuba, 1-1-1 Ten-nohdai, Tsukuba 305-8571, Japan2Department of Physiology, School of Medicine, Keio University, 35 Shinanomachi, Shinjuku-ku, Tokyo 160-8582, Japan3Core Research for Evolutional Science and Technology, Japan Science and Technology Corporation, Kawaguchi, Saitama 332-0012, Japan4Laboratory of Developmental Neuroscience, Initiative for the Promotion of Young Scientists’ Independent Research, Graduate School of
Comprehensive Human Sciences, University of Tsukuba, 1-1-1 Ten-nohdai, Tsukuba 305-8575, Japan5These authors contributed equally to this work6Present address: Department of Biotechnology, Faculty of Chemistry Engineering, DanangUniversity of Technology, 54Nguyen LuongBang
Street, Lien Chieu District, Danang City, Vietnam
*Correspondence: [email protected] (M.Y.), [email protected] (Y.K.)DOI 10.1016/j.neuron.2011.09.034
SUMMARY
NMDA receptor activation leads to clathrin-depen-dent endocytosis of postsynaptic AMPA receptors.Although this process controls long-term depression(LTD) induction in the hippocampus, how it is regu-lated by neuronal activities is not completely clear.Here, we show that Ca2+ influx through the NMDAreceptor activates calcineurin and protein phos-phatase 1 to dephosphorylate phosphatidylinositol4-phosphate 5-kinaseg661 (PIP5Kg661), the majorphosphatidylinositol 4,5-bisphosphate (PI(4,5)P2)-producing enzyme in the brain. Bimolecular fluores-cence complementation analysis revealed that thedephosphorylated PIP5Kg661 became associatedwith the clathrin adaptor protein complex AP-2at postsynapses in situ. NMDA-induced AMPAreceptor endocytosis and low-frequency stimula-tion-induced LTD were completely blocked by inhib-iting the association between dephosphorylatedPIP5Kg661 and AP-2 and by overexpression of akinase-dead PIP5Kg661 mutant in hippocampalneurons. Furthermore, knockdown of PIP5Kg661inhibited the NMDA-induced AMPA receptor endo-cytosis. Therefore, NMDA receptor activationcontrols AMPA receptor endocytosis during hippo-campal LTD by regulating PIP5Kg661 activity atpostsynapses.
INTRODUCTION
AMPA receptors mainly mediate fast excitatory neurotransmis-
sion at synapses located on dendritic spines of vertebrate
neurons. Alterations in neural activities induce long-lasting
changes in synaptic transmission, such as long-term depression
(LTD), which underlies certain aspects of learning and memory
(Bredt and Nicoll, 2003). When neuronal activities increase
moderately during LTD-inducing stimuli, postsynaptic AMPA
receptors are thought to be freed from their anchoring proteins,
to diffuse laterally to the endocytic zone, and to undergo clathrin-
dependent endocytosis (Bredt and Nicoll, 2003; Cognet et al.,
2006; Wang and Linden, 2000). Adaptor protein complex-2
(AP-2) binds to the AMPA receptor and plays a crucial role in re-
cruiting clathrin (Carroll et al., 1999; Lee et al., 2002; Man et al.,
2000). However, it has been unclear whether and how neuronal
activities regulate AP-2 and other components of the endocytic
machinery at postsynapses.
At presynaptic sites, elevated neuronal activity induces the
clathrin-mediated endocytosis of synaptic vesicles (SVs). The
membrane phospholipid phosphatidylinositol (4,5)-bisphos-
phate (PI(4,5)P2) plays a key role in recruiting AP-2 and several
components of the endocytic machinery to endocytic hot spots
at the presynaptic terminal (Ford et al., 2001; Gaidarov and
Keen, 1999; Itoh et al., 2001; Rohde et al., 2002). PI(4,5)P2 is
produced predominantly from phosphatidylinositol 4-phosphate
by phosphatidylinositol 4-phosphate 5-kinase (PIP5K) (Loijens
and Anderson, 1996). Of three PIP5Ka-g isozymes, PIP5Kg is
highly and predominantly expressed in the brain (Akiba et al.,
2002; Wenk et al., 2001) and has three splicing variants,
PIP5Kg635, PIP5g661, and PIP5Kg687 (Giudici et al., 2004; Ish-
ihara et al., 1996; Ishihara et al., 1998; Loijens and Anderson,
1996). The small GTPase ARF6 activates both PIP5Kg (Krauss
et al., 2003) and PIP5Ka (Honda et al., 1999). PIP5Kg661 is
also activated by talin (Morgan et al., 2004). In addition,
increased neuronal activity induces dephosphorylation of
PIP5Kg661 (Akiba et al., 2002; Wenk et al., 2001) by calcineurin,
which is activated by Ca2+ influx through voltage-gated Ca2+
channels (VDCCs) (Lee et al., 2005; Nakano-Kobayashi et al.,
2007). The dephosphorylated PIP5Kg661 becomes enzymati-
cally active by binding AP-2 at presynaptic endocytic spots
and produces PI(4,5)P2 to further recruit AP-2 and other
Neuron 73, 135–148, January 12, 2012 ª2012 Elsevier Inc. 135
Figure 1. Expression of PIP5Kg661 in Hippocampal Neurons
(A and B) The expression of PIP5Kg splicing variants in mouse hippocampus (A) and cultured hippocampal neurons (B) at various developmental stages.
PIP5Kg635 (635), PIP5Kg661 (661), and PIP5Kg687 (687) overexpressed in HEK293T cells were loaded as controls. All samples were treated with the phos-
phatase l-PPase before SDS-PAGE to prevent the mobility shift by phosphorylation and were western blotted with anti-PIP5Kg antibody. E, embryonic day;
P, postnatal day; DIV, days in vitro.
(C) Localization of overexpressed GFP-PIP5Kg661 in cultured hippocampal neurons at 18 DIV. MAP2 was also stained. The region enclosed by the white square
is magnified in the right panels. Scale bars in the left and the right panels correspond to 10 mm and 1 mm, respectively.
(D and E) Immunocytochemical analysis of endogenous PIP5Kg661 and PSD-95 (D) or F-actin (E) in cultured hippocampal neurons at 18 DIV (D) or 20 DIV (E). The
dendritic segment enclosed by the white square in the left is magnified in the right panels. The arrows in the right panels indicate the colocalization of PIP5Kg661
(green) with PSD-95 or F-actin (red). Scale bars correspond in the left and right panels to 10 mm and 1 mm, respectively.
(F) Distribution of PIP5Kg661 in subcellular fractions of the mouse brain. After fractionation, PIP5Kg661 and the postsynaptic and presynaptic markers PSD-95
and synaptophysin, respectively, were detected by immunoblotting. H, Homogenates; P2, crude synaptosomal pellet; P3, lysed synaptosomal membrane
fraction; SV, synaptic vesicle; PSD1 and PSD2, postsynaptic density fractions unsolubilized with Triton X-100 from synaptosomal membranes and the PSD1
fraction, respectively; S1 and S2, supernatant extracted with Triton X-100 from synaptosomal membranes and the PSD1 fraction, respectively; 661, overex-
pressed PIP5Kg661 in HEK293T cells ; l-PPase, lambda protein phosphatase.
Neuron
Activation of PIP5K Is Essential for LTD
components of the early endocytic machinery (Nakano-Kobaya-
shi et al., 2007).
In this study, we examined whether and how the endocytic
machinery is regulated during low-frequency stimulation
(LFS)-induced LTD at postsynapses of pyramidal neurons in
the CA1 region of the mouse hippocampal slices. This form
of LTD depends on the activation of NMDA receptors and
protein phosphatases (Mulkey et al., 1993). We found that
Ca2+ influx through the NMDA receptor, but not through
VDCC, activated protein phosphatase 1 (PP1) and calcineurin
and dephosphorylated PIP5Kg661, which then bound to AP-2
at postsynapses. NMDA-induced AMPA receptor endocytosis
and the LFS-induced LTD were completely blocked by inhibit-
ing the association between PIP5Kg661 and AP-2 and by over-
expression of a kinase-dead PIP5Kg661 mutant in postsynaptic
neurons. These results suggest that NMDA receptor activation
dynamically controls early steps of the clathrin-mediated endo-
cytosis during hippocampal LTD by regulating the PIP5Kg661
activity.
136 Neuron 73, 135–148, January 12, 2012 ª2012 Elsevier Inc.
RESULTS
PIP5Kg661 Localizes to PostsynapsesOf the three splice variants of PIP5Kg, PIP5Kg661 was selec-
tively expressed in mouse hippocampus in vivo and in vitro
(Figures 1A and 1B). PIP5Kg661 expression was observed in
mouse hippocampus from approximately 2 weeks after birth
in vivo and in vitro, and it increased during late postnatal devel-
opment (Figures 1A and 1B).
To examine whether PIP5Kg661 is present on postsynapses,
we first analyzed cultured hippocampal neurons at 18 days
in vitro (DIV) by immunocytochemistry. When green fluorescent
protein (GFP)-tagged PIP5Kg661 (GFP-PIP5Kg661) was ex-
pressed in hippocampal neurons, the GFP signal was observed
in dendrites, which were immunopositive for microtubule-asso-
ciated protein 2 (MAP2), and in spines protruding from the
dendrites (Figure 1C). Like postsynaptic density 95 (PSD-95)
and filamentous actin (F-actin), which were concentrated in the
dendritic spines, endogenous PIP5Kg661 was enriched in
Figure 2. NMDA-Induced Dephosphorylation of PIP5Kg661 in Cultured Hippocampal Neurons
(A) Cultured hippocampal neurons were treated with (NMDA) or without (control) 50 mM NMDA for 10 min. Third lane shows a band of PIP5Kg661 dephos-
phorylated by l-PPase treatment of the lysate of hippocampal neurons.
(B and C) NMDA stimulation of hippocampal neurons induces the dephosphorylation of PIP5Kg661 in concentration- and time-dependent manners. Cultured
hippocampal neurons were treated with NMDA at the indicated concentrations for 10 min (B) or with 50 mM NMDA for the indicated times (C), and endogenous
PIP5Kg661 was detected by western blotting (top). Bottom panels show quantitative data for dephosphorylated PIP5Kg661 (mean ± SEM, n = 3). The ratio of
dephosphorylated/phosphorylated PIP5Kg661 intensity at 1 mM NMDA (B) or before NMDA stimulation (C) was defined as 1.0.
(D) Effects of pharmacological inhibitors on the dephosphorylation of PIP5Kg661 induced by 50 mM NMDA stimulation. Bottom shows the quantitative data for
dephosphorylated PIP5Kg661. The ratio of dephosphorylated/phosphorylated PIP5Kg661 intensity in control neurons was defined as 1.0. The inhibitors used
were 50 mM D-APV (NMDA receptor antagonist), 5 mM EGTA, 10 mM Cyclosporine A (CysA; Calcineurin inhibitor), 1 mM FK-520 (Calcineurin inhibitor), 1 mM
Okadaic acid (High Oka; PP1 and PP2A inhibitor), 10 nMOkadaic acid (LowOka; PP2A inhibitor), and 50 nMFostriecin (PP2A inhibitor). Student’s t test, *p < 0.05;
**p < 0.01; n = 3–7. Error bars indicate SEM.
(E) Effects of a cocktail of VDCC blockers on the KCl- and NMDA-induced dephosphorylation of PIP5Kg661. Cultured hippocampal neurons pretreated for 1 hr
with 2 mM TTX and a cocktail of VDCC blockers, which included 50 mM nimodipine, 500 nM u-conotoxin GVIA, and 1 mM u-agatoxin IVA, were stimulated with
50 mM KCl or 50 mM NMDA for 10 min. The ratio of dephosphorylated/phosphorylated PIP5Kg661 intensity in control neurons was defined as 1.0. (Student’s
t test, *p < 0.05; n = 3). Error bars indicate SEM.
Neuron
Activation of PIP5K Is Essential for LTD
dendritic spine-like protrusions (see Figures S1A–S1E available
online). Endogenous PIP5Kg661 partially colocalized with
PSD-95 and F-actin (Figures 1D and 1E). Furthermore, immuno-
blot analysis of the subcellular fractions of adult mouse brain
showed PIP5Kg661 not only in the SV fraction, which was immu-
nonegative for PSD-95, but also in the PSD fractions, which were
immunonegative for an SV marker synaptophysin (Figure 1F).
Together, these results indicate that PIP5Kg661 localizes at least
in part to postsynapses.
PIP5Kg661 Is Dephosphorylated by NMDA-InducedCa2+ InfluxThe dephosphorylation of PIP5Kg661 by calcineurin plays an
essential role in the activity-dependent production of PI(4,5)P2
at presynapses (Lee et al., 2005; Nakano-Kobayashi et al.,
2007). To examinewhether PIP5Kg661 is also dephosphorylated
at postsynapses, we treated hippocampal neurons with NMDA,
which induces AMPA receptor endocytosis and LTD (Beattie
et al., 2000; Carroll et al., 1999; Lee et al., 2002; Lin et al.,
2000). To block action potential-induced VDCC activation at pre-
synapses, we included tetrodotoxin (TTX) in the culture medium.
Immunoblot analysis of the cell lysates with an anti-PIP5Kg
antibody revealed that an additional PIP5Kg661 band, which
migrated faster on electrophoresis gels, appeared after NMDA
treatment (Figure 2A). This band likely corresponds to the
dephosphorylated form of PIP5Kg661, because PIP5Kg661
migrated to the same position when the lysates were treated
with l-phosphatase before electrophoresis (Figure 2A). NMDA
Neuron 73, 135–148, January 12, 2012 ª2012 Elsevier Inc. 137
Neuron
Activation of PIP5K Is Essential for LTD
treatment increased the dephosphorylated form of PIP5Kg661
in a dose-dependent manner, with an EC50 of approximately
30 mM (Figure 2B). The dephosphorylation of PIP5Kg661 was
observed as early as 5 min and was saturated by 20 min after
50 mM NMDA treatment (Figure 2C). These results indicate that
PIP5Kg661 is mostly phosphorylated at the basal level and is
rapidly dephosphorylated upon NMDA treatment. The concen-
tration and duration of NMDA treatment were similar to those
used previously to induce AMPA receptor endocytosis in
cultured neurons (Beattie et al., 2000; Carroll et al., 1999; Lee
et al., 2002; Lin et al., 2000).
To examine the molecular mechanism responsible for the
NMDA-induced dephosphorylation of PIP5Kg661, we treated
hippocampal neurons with various pharmacological reagents.
The NMDA antagonist D-APV or the Ca2+ chelator EGTA
completely blocked the NMDA-induced dephosphorylation of
PIP5Kg661 (Figure 2D), demonstrating that Ca2+ entry through
the NMDA receptor is essential for this process. Calcineurin
inhibitors, cyclosporine A and FK-520, partially inhibited the
NMDA-induced dephosphorylation of PIP5Kg661 (Figure 2D).
In contrast, the dephosphorylation of PIP5Kg661 was almost
completely blocked by a high concentration (1 mM) of okadaic
acid (Figure 2D), which inhibits both PP1 and protein phospha-
tase 2A (PP2A). Because a low concentration (10 nM) of
okadaic acid and fostriecin, which specifically inhibit PP2A
(Bialojan and Takai, 1988; Boritzki et al., 1988), were ineffective
(Figure 2D), the NMDA-induced dephosphorylation of
PIP5Kg661 was probably mediated by PP1 and partially by cal-
cineurin. These results were in contrast to the dominant role of
calcineurin in the high-KCl-induced dephosphorylation of
PIP5Kg661 at presynapses (Nakano-Kobayashi et al., 2007).
Indeed, the NMDA-induced dephosphorylation of PIP5Kg661
was not inhibited by a cocktail of Ca2+ channel blockers (Fig-
ures 2E and S2). Therefore, the direct calcium influx through
NMDA receptors likely activates a specific pathway that
involves PP1 and calcineurin and dephosphorylates PIP5Kg661
at postsynapses.
NMDA Receptor Activation Triggers an Interactionof PIP5Kg661 with AP-2The AP-2 subunit b2 adaptin was previously shown to interact
directly with the dephosphorylated form of PIP5Kg661 in vitro
(Nakano-Kobayashi et al., 2007). To examine whether NMDA
treatment induces the interaction of PIP5Kg661 with AP-2 in
neurons, we performed a coimmunoprecipitation assay using
cultured hippocampal neurons. AP-2 subunits (a and b adaptins)
were coimmunoprecipitated with PIP5Kg661 upon NMDA treat-
ment (Figure 3A). Furthermore, immunocytochemical analysis of
hippocampal neurons expressing hemagglutinin (HA)-tagged
PIP5Kg661 and FLAG-tagged b2 adaptin revealed that as early
as 5 min after NMDA application, colocalization of HA and
FLAG immunoreactivities was detected and saturated by
10 min in the dendrites (Figure S3).
To examine whether this interaction occurs at postsynapses,
we performed a bimolecular fluorescence complementation
(BiFC) assay using N- and C-terminal subfragments of Venus,
which were fused to the ear domain of b2 adaptin (VN-b2 ear)
and wild-type PIP5Kg661 (VC-PIP5K-WT), respectively. In this
138 Neuron 73, 135–148, January 12, 2012 ª2012 Elsevier Inc.
assay, interaction of fused proteins mediates the reconstitution
of Venus, resulting in efficient fluorescence emission (Kerppola,
2006). Time-lapse imaging of hippocampal neurons expressing
these fusion proteins showed Venus fluorescent puncta appear-
ing along the neurites, approximately 2 min after NMDA
treatment (Figure 3B and Movie S1). Changes in fluorescence
intensity were detected faster than those reported for full chro-
mophore maturation in BiFC studies (Kerppola, 2006), but
several studies of interaction between various proteins have re-
ported rapid changes in fluorescence intensity in response to
stimuli (Guo et al., 2005; MacDonald et al., 2006; Schmidt
et al., 2003). Furthermore, immunocytochemical analysis con-
firmed that most Venus puncta corresponded to spines on
MAP2-positive dendrites (Figure 3C) and colocalized with
PSD-95 and F-actin (Figures 3D and 3E). As a negative control,
we expressed N- and C-terminal fragments of Venus fused to
proteins that do not interact with each other in hippocampal
neurons. The combination of VN-tagged glutathione S-trans-
ferase (GST) and VC, VN-b2 ear, and VC, or VN-GST and
VC-PIP5K-WT gave rise to diffuse background fluorescence
throughout dendrites and never showed punctate fluorescence
(Figure S4A). These results indicate that NMDA receptor activa-
tion triggers the interaction of PIP5Kg661 with AP-2 at postsy-
napses in hippocampal neurons.
The phosphorylation of Ser at position 645 of PIP5Kg661
blocks its interaction with b2 adaptin (Nakano-Kobayashi
et al., 2007). To examine whether such a dephosphorylation-
dependent interaction occurs at postsynapses, we used the
phosphomimetic mutant of PIP5Kg661, VC-PIP5K-S645E, in
which Glu replaced Ser at 645. The BiFC assay using VN-b2
ear and VC-PIP5K-S645E revealed no Venus fluorescent
puncta after NMDA treatment in hippocampal neurons (Figures
3C and 3G). When dephosphorylation was inhibited by FK520
(1 mM) or okadaic acid (1 mM), the NMDA-induced formation of
Venus fluorescent puncta was significantly reduced in neurons
expressing VN-b2 ear and VC-PIP5K-WT (Figures 3F and 3G).
Because Ser645 of PIP5Kg661 is phosphorylated by Cdk5
(Lee et al., 2005), we performed the BiFC assay in the pres-
ence of a Cdk5 inhibitor olomoucine. The NMDA-induced
formation of Venus fluorescent puncta in neurons expressing
VN-b2 ear and VC-PIP5K-WT was significantly increased by
olomoucine (Figures S4B and S4C). Together, these results
confirm that the emergence of Venus fluorescent puncta re-
flects specific interaction between b2 adaptin and PIP5Kg661.
They also indicate that the NMDA-induced dephosphorylation
of PIP5Kg661 at Ser 645 plays an essential role in its associ-
ation with AP-2 at postsynapses. In contrast, immunoblot
analysis of the cell lysates at 5 min after NMDA application
showed that only approximately 35% of PI5Kg661 was de-
phosphorylated compared to the maximum dephosphorylation
level (Figure 2C). The rapid emergence of the BiFC signal may
be caused by the high sensitivity of fluorescence detection
(Kerppola, 2009) in spines, whereas the immunoblot assay
reflects total endogenous PIP5Kg661. The dephosphorylated
form of PIP5Kg661 was mostly observed in the membrane
fractions, such as SV and PSD (Figure 1F), probably because
it is more tightly associated with the plasma membrane
via AP-2.
Figure 3. NMDA-Induced Interaction of Dephosphorylated PIP5Kg661 with AP-2 in Hippocampal Neurons
(A) Coimmunoprecipitation of AP-2 with PIP5Kg661 from NMDA-stimulated hippocampal neurons. After cultured hippocampal neurons were treated with 50 mM
NMDA for 10 min, the endogenous PIP5Kg661 was immunoprecipitated with the anti-PIP5Kg antibody, and the coimmunoprecipitated AP-2 subunits were
detected by anti-a and -b adaptin antibodies. The same result was obtained in four independent experiments.
(B) BiFC assay for the spatiotemporal interaction of PIP5Kg661 with the b2 adaptin ear domain. Cultured hippocampal neurons expressing VN-b2 ear and
VC-PIP5K-WT, respectively, were stimulated with 50 mMNMDA and were observed for up to 3 min. The dendritic region enclosed by a white square is magnified
in the right panels. Scale bars represent 5 mm.
(C) Failure of the phosphomimetic PIP5Kg661 mutant to interact with b2 adaptin. Cultured hippocampal neurons coexpressing VN-b2 ear with VC-PIP5K-WT or
VC-PIP5K-S645E were treated with 50 mMNMDA for 5 min and were monitored for the Venus signal. Neurons were also immunostained for MAP2. The dendritic
regions enclosed by white squares are magnified in the bottom panels. Arrowheads indicate the Venus puncta. Scale bar represents 10 mm.
(D and E) Interaction of PIP5Kg661with b2 adaptin at spines. Cultured hippocampal neurons coexpressing VN-b2 ear and VC-PIP5K-WTwere treatedwith NMDA
and were monitored for the Venus signal. Neurons were also immunostained for PSD-95 (D) or F-actin (E). Scale bar represents 10 mm.
(F) Suppression of the PIP5Kg661 interaction with b2 adaptin by phosphatase inhibitors. After pretreatment with 1 mM FK-520 or 1 mM Okadaic acid for 1 hr,
cultured hippocampal neurons coexpressing VN-b2 ear and VC-PIP5K-WT were treated with NMDA, monitored for the Venus signals, and stained for MAP2.
Scale bar represents 10 mm.
(G) Quantitative data derived from the results shown in (C) and (F). The number of Venus puncta was normalized to the dendrite length. Student’s t test, *p < 0.05;
**p < 0.001; n = 11–30. Error bars indicate SEM.
Neuron
Activation of PIP5K Is Essential for LTD
Interaction of Dephosphorylated PIP5Kg661with AP-2 Is Required for NMDA-Induced AMPAReceptor EndocytosisAssociation with AP-2 activates PIP5Kg661, leading to the pro-
duction of PI(4,5)P2 in vitro (Nakano-Kobayashi et al., 2007);
PI(4,5)P2 triggers further accumulation of AP-2 and other endo-
cytic components at presynapses. Thus, we hypothesized that
the NMDA-induced association of PIP5Kg661 with AP-2 (Fig-
ure 3) plays an essential role in NMDA-induced AMPA receptor
endocytosis at postsynapses. To test this hypothesis, we ex-
pressed the GluA2 subunit tagged with HA at its extracellular N
terminus in cultured hippocampal neurons. As reported previ-
ously (Beattie et al., 2000; Carroll et al., 1999; Ehlers, 2000),
the surface HA-GluA2, which is mainly localized to the somato-
dendritic domain (Matsuda et al., 2008), rapidly decreased
following NMDA stimulation (Figures 4A and 4E). When the
GFP-tagged C-terminal fragment of PIP5Kg661 (PIP5K-CT-
WT) was expressed, the NMDA-induced reduction of surface
AMPA receptors was blocked (Figures 4B and 4E). Similarly,
the expression of the GFP-tagged dephosphomimetic mutant
Neuron 73, 135–148, January 12, 2012 ª2012 Elsevier Inc. 139
Figure 4. Interaction of PIP5Kg661 with AP-2 Is Essential for the NMDA-Induced AMPA Receptor Endocytosis
(A–D) Effects of phosphomimetic and dephosphomimetic C-terminal peptides of PIP5Kg661 on the NMDA-dependent endocytosis of surface GluA2. Cultured
hippocampal neurons expressing GFP (A), GFP-tagged C terminus of wild-type PIP5Kg661 (PIP5K-CT-WT) (B), dephosphomimetic PIP5K-CT (PIP5K-CT-
S645A) (C), and phosphomimetic PIP5K-CT (PIP5K-CT-S645E) (D) were treated with 50 mMNMDA for 10 min and stained for surface HA-GluA2 (red). After Triton
X treatment, the neurons were stained for the total HA-GluA2 (blue). The dendritic regionsmarked bywhite squares aremagnified in the bottom panels. Scale bars
represent 10 mm.
(E) Quantification of the NMDA-induced endocytosis of surface GluA2. Data are represented as the ratio of surface HA-GluA2/total HA-GluA2 staining intensity.
The ratio in control neurons was defined as 1.0. Student’s t test, *p < 0.005 compared with the control for each condition; n = 8–15 cells. Error bars indicate SEM.
Neuron
Activation of PIP5K Is Essential for LTD
of the PIP5Kg661 C-terminal fragment (PIP5K-CT-S645A), in
which Ala replaced Ser 645, blocked NMDA-induced AMPA
receptor endocytosis (Figures 4C and 4E). In contrast, expres-
sion of GFP-tagged phosphomimetic mutant of the PIP5Kg661
C-terminal fragment (PIP5K-CT-S645E) failed to do so (Figures
4D and 4E). GST pull-down assays confirmed that wild-type
and PIP5K-CT-S645A, but not PIP5K-CT-S645E, bound to b2
adaptin (Figure S5A). These results suggest that overexpression
of the wild-type and dephosphomimetic C-terminal fragment of
PIP5Kg661 interferes with the interaction between endogenous
PIP5Kg661 and AP-2, thereby inhibiting the NMDA-induced
AMPA receptor endocytosis.
The C terminus of PIP5Kg661 is also reported to bind to the m2
adaptin of AP-2 (Bairstow et al., 2006) and talin (Di Paolo et al.,
2002; Ling et al., 2002). In addition, b2 adaptin binds to clathrin
through its ear and hinge domains (Thieman et al., 2009) and
to hippocalcin, a protein that translocates to the plasma
membrane in the presence of Ca2+ during hippocampal LTD
(Palmer et al., 2005). Nevertheless, none of these interactions
are reported to be affected by the dephosphorylation of the
C-terminal domain of PIP5Kg661. To examine the specificity of
the dephosphorylation-dependent binding of PIP5Kg661 to b2
adaptin, we performed GST pull-down assays. Although b2
adaptin bound to GST-CT-WT and GST-CT-S645A with sig-
nificantly higher affinity than to GST-CT-S645E, m2 adaptin
equally interacted with all GST-fused C-terminal fragments of
PIP5Kg661 (Figure S5A). In addition, although PIP5Kg661
140 Neuron 73, 135–148, January 12, 2012 ª2012 Elsevier Inc.
binding to b2 adaptin was specifically inhibited by the depho-
phomimetic C-terminal peptide of PIP5Kg661 (pep-S645A),
the clathrin binding to b2 adaptin was not affected by either pho-
phomimetic (pep-S645E) or dephosphomimetic (pep-S645A)
peptide (Figure S5B). Furthermore, binding of CT-WT to talin
was inhibited by pep-S645A and pep-S645E to the same degree
(Figure S5C). The interaction between b2 adaptin and hippocal-
cin in the presence of Ca2+ was also similarly inhibited by de-
phosphomimetic or phosphomimetic PIP5Kg661 C-terminal
peptides and fragments (Figures S5D). These results indicate
that the dephosphorylated form-specific effect of PIP5Kg661
(Figure 4) is likely mediated by its specific interaction with b2
adaptin and that this specific interaction is required for NMDA-
induced AMPA receptor endocytosis.
PIP5Kg661 Activity Is Required for NMDA-InducedAMPA Receptor EndocytosisTo examine the requirement of the kinase activity of PIP5Kg661
for NMDA-induced AMPA receptor endocytosis, we coex-
pressed HA-GluA2 and GFP-tagged PIP5Kg661 (PIP5K-WT) or
its kinase-dead mutant (PIP5K-D316A), in which Ala replaced
Glu at 316 (Rao et al., 1998), in cultured hippocampal neurons.
We found that the NMDA-induced reduction in surface HA-
GluA2 was completely blocked by expression of PIP5K-D316A
(Figures 5B and 5C), but not by expression of PIP5K-WT (Figures
5A and 5C). There was no significant difference in surface HA-
GluA2 levels between neurons expressing PIP5K-WT and
Figure 5. PIP5Kg661 Activity Is Required for the
NMDA-Induced AMPA Receptor Endocytosis
(A andB) Effect of a kinase-dead PIP5Kg661mutant on the
NMDA-induced GluA2 endocytosis. Cultured hippo-
campal neurons coexpressing GFP-tagged wild-type
PIP5Kg661 (PIP5K-WT) (A) or the kinase-dead mutant
PIP5K-D316A (B) with HA-GluA2 were treated with 50 mM
NMDA for 10 min and were stained for the surface HA-
GluA2 and total GluA2 as in Figure 4. The dendritic regions
marked by white squares are magnified in the bottom
panels. Scale bars represent 10 mm.
(C) The ratio of surface HA-GluA2/total HA-GluA2 staining
intensity was quantified as in Figure 4. Student’s t test,
*p < 0.005 compared with the control for each condition;
n = 8–15 cells. Error bars indicate SEM.
(D) Cultured hippocampal neurons were treated with
(NMDA) or without (control) 50 mM NMDA for 10 min. After
immunoprecipitation of PIP5Kg from the neurons, the
PIP5K activity was determined. Student’s t test, *p < 0.05;
n = 4. Error bars indicate SEM.
Neuron
Activation of PIP5K Is Essential for LTD
PIP5K-D316A in the resting state. Next, we performed an in vitro
kinase assay of PIP5Kg661 after its immunorecipitation from
hippocampal neurons treated or untreated with NMDA. The
kinase activity of PIP5Kg661 from neurons treated with NMDA
was significantly increased (Figure 5D). These results suggest
that an NMDA-induced increase in PIP5Kg661’s kinase activity
is necessary for evoking AMPA receptor endocytosis.
To further confirm the role of PIP5Kg661 in NMDA-induced
AMPA receptor endocytosis, we employed a loss-of-function
approach using vectors for shRNA and GFP. Immunoblot anal-
ysis of the cell lysates with an anti-FLAG antibody revealed
that two shRNAs directed against PIP5Kg (shRNA 1 and 2)
specifically inhibited expression of FLAG-PIP5Kg661, but not
FLAG-PIP5Ka or FLAG-PIP5Kb, in HEK293T cells (Figures
S6A and S6B). Similarly, in GFP-positive neurons, endogenous
PIP5Kg661 immunoreactivity was markedly reduced by these
shRNAs, but not by a scrambled shRNA, whereas tubulin
immunoreactivity was not affected by either construct (Figures
S6C–S6E). NMDA-induced reduction in surface HA-GluA2 was
significantly inhibited by these PIP5Kg-specific shRNAs (Fig-
ures 6B, 6C, and 6E), but not by a scrambled shRNA (Figures
6A and 6E). Furthermore, the inhibitory effect of shRNA 2 on
NMDA-induced reduction in surface HA-GluA2 was rescued
by coexpression of the shRNA-resistant GFP-PIP5Kg661
(PIP5Kg661res) in hippocampal neurons (Figures 6D and 6E).
These results indicate that PIP5Kg661 plays a crucial role in
NMDA-induced AMPA receptor endocytosis in hippocampal
neurons.
Neuron 73, 13
LFS-Induced LTD Requiresthe Interaction of PIP5Kg661with AP-2 and Its Enzymatic ActivityFinally, to examine whether LTD is regulated by
similar mechanisms, we introduced the dephos-
phomimetic pep-S645A, which specifically in-
hibited the interaction between PIP5Kg661 and
AP-2 (Figures S5B–S5D), into theCA1 pyramidal
neurons via a patch pipette during LFS-induced
LTD in hippocampal slice preparations (Figure 7A). There was no
difference in the excitatory postsynaptic current (EPSC) ampli-
tude between neurons treated with decoy peptide and control
phosphomimetic peptide pep-S645E. To examine the effect of
peptides on the basal EPSC amplitude in the same neurons,
we measured the EPSC amplitudes just after breaking into
whole-cell mode and 9–10 min later, because it generally takes
at least several minutes for peptides to diffuse from patch
pipettes to synapses. We did not detect differences in EPSC
amplitudes between 0–1 min and 9–10 min after establishing
the whole-cell mode (Figures S7A–S7C; pep-S645A, 170 ±
14 pA at 0–1 min and 187 ± 18 pA at 9–10 min, n = 24, p =
0.152 by a paired t test; pep-S645E, 174 ± 11 pA at 0–1 min
and 195 ± 13 pA at 9–10 min, n = 17, p = 0.186 by a paired t
test; pep-S645A versus pep-S645E, p = 0.672 by a Mann-Whit-
ney U test). In contrast, after LFS was applied to the Schaffer
collateral, LTDwas only observed in neurons treated with control
peptides (the EPSC amplitude at 25–30min after LFSwas 69%±
5% of baseline; Figures 7C and 7D), but not in neurons treated
with a dephosphomimetic peptide pep-S645A (92% ± 7% of
baseline; Figures 7B and 7D). These results indicate that the
interaction of PIP5Kg661 with AP-2 is required for LFS-induced
LTD, but not for basal neurotransmission.
To examine requirement of the kinase activity of PIP5Kg661
for LFS-induced LTD, we infected hippocampal CA1 neurons
with Sindbis virus encoding GFP and wild-type (Sin-GFP-
PIP5K-WT) or kinase-dead PIP5Kg661 (Sin-GFP-PIP5K-
D316A) (Figure 7E). LTD was significantly inhibited in neurons
5–148, January 12, 2012 ª2012 Elsevier Inc. 141
Figure 6. Suppression of NMDA-Induced AMPA Receptor Endocytosis by Knockdown of PIP5Kg661
(A–D) Effect of PIP5Kg661 knockdown on NMDA-induced GluA2 endocytosis. Cultured hippocampal neurons coexpressing GFP, HA-GluA2, and scrambled
shRNA (A), shRNA 1 (B), shRNA 2 (C), or GFP-fused shRNA-resistant form (GFP-PIP5Kres) and shRNA #2 (D) were treated with 50 mM NMDA for 10 min and
were immunostained for the surface HA-GluA2 and total GluA2 as in Figure 4. The dendritic regions enclosed by white squares are enlarged in the bottom panels.
Scale bars represent 10 mm.
(E) Quantification of NMDA-induced endocytosis of surface GluA2 in shRNA-treated neurons. The ratio of surface HA-GluA2/total HA-GluA2 staining
intensity was quantified as in Figure 4. Error bars indicate SEM. Student’s t test, *p < 0.001 compared with the control for each condition; n = 15 cells for A–C,
n = 20 cells for D.
Neuron
Activation of PIP5K Is Essential for LTD
expressing Sin-GFP-PIP5K-D316A (the EPSC amplitude at
20–25 min after LFS was 89% ± 6% of baseline) than those ex-
pressing Sin-GFP-PIP5K-WT (64% ± 4% of baseline; p = 0.028,
Figure 7F). These results indicate that the activation of
PIP5Kg661 following its interaction with b2 adaptin is required
for LFS-induced LTD.
DISCUSSION
PIP5Kg661 and the Regulated Endocytosis of AMPAReceptors at PostsynapsesIn the present study, we demonstrated that NMDA receptor acti-
vation induced the dephosphorylation of PIP5Kg661 (Figure 2)
and its association with AP-2 at postsynapses in hippocampal
neurons (Figure 3). NMDA-induced AMPA receptor endocytosis
was blocked by inhibiting the interaction of PIP5Kg661with AP-2
(Figure 4), by inhibiting the PIP5Kg661 activity (Figure 5), or by
PIP5Kg661 knockdown (Figure 6). Furthermore, LFS-induced
LTD was also blocked by inhibiting the interaction of PIP5Kg661
with AP-2 or by inhibiting the kinase activity of PIP5Kg661 in the
CA1 pyramidal neurons (Figure 7). Binding to AP-2 activates
PIP5Kg661 to produce PI(4,5)P2 (Nakano-Kobayashi et al.,
142 Neuron 73, 135–148, January 12, 2012 ª2012 Elsevier Inc.
2007), which plays a key role in recruiting AP-2 to the plasma
membrane (Gaidarov and Keen, 1999). Indeed, NMDA treatment
increased the PIP5Kg661 activity in hippocampal neurons (Fig-
ure 5D). Based on these findings, we propose the following
model in which AMPA receptor endocytosis is upregulated at
postsynapses during NMDA receptor-dependent LTD (Figure 8):
(1) activity-induced Ca2+ influx through NMDA receptors
dephosphorylates PIP5Kg661 by activating PP1 and calcineurin;
(2) the dephosphorylated PIP5Kg661 is recruited to the postsyn-
aptic endocytic site by binding to preexisting AP-2 and (3) is
activated to produce PI(4,5)P2; and (4) produced PI(4,5)P2 further
recruits AP-2 and other components of the early endocytic
machinery (Ford et al., 2001; Gaidarov and Keen, 1999; Itoh
et al., 2001; Rohde et al., 2002) and stimulates the clathrin-
dependent AMPA receptor endocytosis.
Several other mechanisms have been proposed for the AMPA
receptor endocytosis during NMDA receptor-dependent LTD in
the hippocampus. For example, hippocalcin, a protein that
possesses a Ca2+/myristoyl switch, binds AP-2 and translocates
to the plasmamembrane (Palmer et al., 2005). However, how the
hippocalcin-AP2 complex is recruited to the endocytic zone
remains unclear. Activation of the small GTPase Rab5 has also
Figure 7. Suppression of Hippocampal CA1-LTD by Inhibiting the Interaction of PIP5Kg661 with AP-2 and by Expression of Kinase-Dead
PIP5Kg661 Mutant
(A) Schematic diagram of the electrophysiological experimental procedure for (B–D). CA1 pyramidal neurons were perfused with the phosphomimetic (pep-
S645E) or dephosphomimetic (pep-S645A) peptide of PIP5Kg661 C-terminus via a recording patch pipette (Rec.). To evoke EPSCs, we stimulated Schaffer
collaterals by a glass stimulating electrode (Stim.) placed on the stratum radiatum on the CA1 region.
(B and C) Representative CA1-LTD data recorded from CA1 pyramidal neurons perfused with pep-S645A (300 mM; B) or pep-S645E (300 mM; C). Inset traces
represent EPSCs recorded just before (a) or 30 min after (b) LFS and their superimposition (a + b).
(D) Averaged CA1-LTD data. EPSC amplitudes were normalized by the amplitudes immediately before LFS and averaged (n = 14 for pep-S645A, n = 12 for pep-
S645E). Error bars indicate SEM.
(E) A representative image showing GFP expression by infection of recombinant Sindbis virus at hippocampal CA1 region in wild-type mice (36 hr after virus
injection). Scale bar represents 100 mm.
(F) Averaged CA1-LTD data recorded from CA1 pyramidal neurons that express a GFP plus wild-type PIP5Kg661 (Sin-GFP-PIP5K-WT; open circles; n = 5) or the
kinase-dead PIP5Kg661mutant (Sin-GFP-PIP5K-D316A; filled circles; n = 7). EPSC amplitudes were normalized by the amplitudes 1min immediately before LFS
(1 Hz, 300 stimuli at –40mV) and averaged. Insets show EPSC traces recorded just before (black traces) and 25 min after (gray traces) LFS in each condition. The
error bars indicate the SEM. The p value was obtained by the Mann-Whitney U test.
Neuron
Activation of PIP5K Is Essential for LTD
been shown to be essential for AMPA receptor endocytosis
during hippocampal LTD (Brown et al., 2005). Rab5 regulates
multiple steps on the endocytic pathway, such as invagination
of the clathrin-coated pit, endosomal fusion, and downstream
signaling (Zerial and McBride, 2001). It also mediates the
uncoating of AP-2 from clathrin-coated vesicles by decreasing
the PI(4,5)P2 levels (Semerdjieva et al., 2008). Furthermore,
cpg2, a protein expressed in the hippocampus after increased
Neuron 73, 135–148, January 12, 2012 ª2012 Elsevier Inc. 143
Figure 8. ProposedMolecularMechanism for NMDA-InducedAMPA
Receptor Endocytosis and LTD Induction
Upon NMDA receptor activation, PIP5Kg661 is dephosphorylated by calci-
neurin and PP1 (1) and interacts with AP-2 at the restricted area of the
postsynaptic plasma membrane (2). The interaction with AP-2 induces
the activation of PIP5Kg661, resulting in the local production of PI(4,5)P2 (3).
The produced PI(4,5)P2 recruits endocytic components and induces AMPA
receptor endocytosis, thereby inducing LTD (4).
Neuron
Activation of PIP5K Is Essential for LTD
neuronal activity, is localized to the postsynaptic endocytic zone
and regulates AMPA receptor endocytosis during LTD (Cottrell
et al., 2004). Similarly, Arc, an immediate early gene transcribed
after increased neuronal activity, enhances AMPA receptor
endocytosis by binding to components of the later endocytic
machinery, such as dynamin and endophilin (Chowdhury et al.,
2006). These findings support the hypothesis that activity-
induced AMPA receptor endocytosis is regulated at the level of
the endocytic process itself. Although all these processes are
not mutually exclusive and are likely to work in concert, our
model proposes an important form of regulation at the initial
step of endocytosis: the recruitment of PIP5Kg661 to the plasma
membrane to induce the clathrin-dependent AMPA receptor
endocytosis.
An alternative scenario is that AMPA receptor endocytosis is
regulated by the lateral movement of AMPA receptors to the
constitutively active endocytic zone, but not by endocytic steps:
the application of NMDA does not cause an appreciable loss of
clathrin from the dendritic spines of hippocampal neurons (Blan-
pied et al., 2002). Similarly, the activity-dependent endocytosis
of epidermal growth factor receptor occurs via its lateral move-
ment to preformed clathrin-coated pits in HeLa cells (Rappoport
and Simon, 2009). PI(4,5)P2 is rapidly dephosphorylated by syn-
aptojanin 1 to uncoat AP-2 and other phosphoinositide-based
membrane associations from clathrin-coated membranes (Cre-
mona et al., 1999; Zoncu et al., 2007). Indeed, NMDA-induced
AMPA receptor endocytosis is severely impaired in synaptojanin
1 null hippocampal neurons, indicating that PI(4,5)P2metabolism
plays an essential role in the regulation of postsynaptic AMPA
receptor trafficking (Gong and De Camilli, 2008). Therefore, ac-
cording to this scenario, the NMDA-dependent activation of
144 Neuron 73, 135–148, January 12, 2012 ª2012 Elsevier Inc.
PIP5Kg661 by interaction with AP-2 may serve as a mechanism
for maintaining constant PI(4,5)P2 and AP-2 levels during
enhanced AMPA receptor endocytosis.
Comparison with Regulated Endocytosisat PresynapsesElevated neuronal activities increase SV endocytosis at presy-
napses. PI(4,5)P2 plays a major role in clathrin coat dynamics.
For example, activity-induced SV endocytosis is severely
retarded in Pip5kg null neurons, in which the PI(4,5)P2 levels
are decreased both in the basal state and after KCl stimulation
(Di Paolo et al., 2004). Thus, our findings demonstrate that in
addition to its role at the presynapse, PIP5Kg also has an impor-
tant postsynaptic function.
Several differences exist in the regulation of PIP5Kg between
the pre- and postsynaptic sides. The enhanced SV endocytosis
induced by high KCl or direct stimulation of nerve terminals is
largely mediated by Ca2+ influx through VDCC (Cousin and Rob-
inson, 2001). Indeed, high-KCl-induced dephosphorylation of
PIP5Kg is blocked by VDCC blockers (Figures 2E and S2). In
contrast, NMDA-induced Ca2+ influx (Dayanithi et al., 1995)
and AMPA receptor endocytosis (Beattie et al., 2000) are insen-
sitive to VDCC blockers. Similarly, NMDA-induced dephosphor-
ylation of PIP5Kg (in the presence of TTX) was dependent on
Ca2+ influx through NMDA receptors, but not VDCC (Figures
2E and S2). Furthermore, LFS-induced LTD in CA hippocampal
neurons was largely insensitive to a VDCC blocker nimodipine
under our experimental conditions (Figure S8), as reported
earlier (Oliet et al., 1997; Selig et al., 1995). Such dependency
on NMDA receptors as a predominant Ca2+ source may reflect
the geometrical arrangement of NMDA receptors, which are
highly expressed in spines (Corlew et al., 2008), together with
PSD-95 and A-kinase-anchoring proteins (AKAPs). Indeed,
AKAPs play a crucial role in NMDA-induced AMPA receptor
endocytosis by scaffolding specific protein kinases and calci-
neurin at postsynapses in hippocampal neurons (Bhattacharyya
et al., 2009). Although certain VDCCs are expressed in dendrites,
they may not be fully activated by the depolarization caused by
NMDA receptor activation, which may inhibit VDCC activities
(Chernevskaya et al., 1991).
The second major difference is that, whereas KCl-induced
SV endocytosis and the dephosphorylation of PIP5Kg are largely
blocked by calcineurin inhibitors (Cousin and Robinson, 2001;
Lee et al., 2005; Nakano-Kobayashi et al., 2007), the NMDA-
induced dephosphorylation of PIP5Kg was more potently in-
hibited by PP1 blockers (Figure 2D). Similarly, NMDA receptor-
dependent LTD requires the activation of PP1 and calcineurin
(Mulkey et al., 1993, 1994). PP1 activity is regulated by the calci-
neurin-dependent dephosphorylation of Inhibitor 1 during LTD
induction (Munton et al., 2004). In addition, the activity of PP1
is regulated by its rapid recruitment from the dendrites to
synapses during LTD stimulus (Morishita et al., 2001). Such
multiple regulatory pathways for PP1 activation may explain
why calcineurin only partially inhibited the NMDA-dependent
dephosphorylation of PIP5Kg661. PP1 can dephosphorylate
PIP5Kg661 in vitro (Nakano-Kobayashi et al., 2007). Thus,
although serine 845 of the GluA1 AMPA receptor is one of the
critical substrates of calcineurin and PP1, we propose that
Neuron
Activation of PIP5K Is Essential for LTD
PIP5Kg661 is another substrate that may regulate the endocytic
process during LTD.
Various Endocytic Machineries at the Postsynaptic SiteIn addition to NMDA receptor-activated AMPA receptor endocy-
tosis, there are other forms of activity-dependent AMPA receptor
endocytosis in hippocampal neurons. The application of AMPA
or (s)-3,5-dihydroxyphenylglycine (DHPG), a metabotropic
glutamate receptor (mGluR) agonist, induces AMPA receptor
endocytosis (Beattie et al., 2000; Snyder et al., 2001). Similarly,
mGluR-dependent LTD could be induced in hippocampal slices
using paired-pulse LFS. Unlike NMDA receptor-dependent LTD,
mGluR-dependent LTD is independent of an increase in cyto-
solic Ca2+ or the activation of serine/threonine phosphatases,
such as PP1 and calcineurin (Harris et al., 2004; Schnabel
et al., 2001). Similarly, AMPA-induced AMPA receptor endocy-
tosis is relatively insensitive to Ca2+ influx or calcineurin (Beattie
et al., 2000; Lin et al., 2000). Thus, in these forms of activity-
induced AMPA receptor endocytosis, the postsynaptic PI(4,5)P2
level may be regulated by a mechanism distinct from the PP1-
and calcineurin-dependent pathway. Indeed, during the prepa-
ration of this manuscript, it was reported that mGluR-dependent
AMPA receptor endocytosis is regulated by ARF6 activation
through the ARF6-specific guanine-nucleotide exchange factor
BRAG2 (Scholz et al., 2010). Interestingly, in contrast to DHPG,
NMDA did not activate endogenous ARF6 in hippocampal
neurons (Figure S9), indicating the existence of multiple path-
ways leading to PIP5Kg activation.
Other ARF6 guanine nucleotide exchange factors, such as
EFA6 and ARNO, are expressed at postsynapses (Krauss
et al., 2003; Vitale et al., 2002). Because inactive ARF6 also binds
to PIP5Kg (Nakano-Kobayashi et al., 2007), it may also serve to
concentrate PIP5Kg near the postsynaptic endocytic zone and
to work with the PP1- and calcineurin-dependent PIP5Kg activa-
tion pathway during NMDA receptor-mediated AMPA receptor
endocytosis. Some forms of activity-induced AMPA receptor
endocytosis may also be mediated by clathrin-independent
mechanisms, such as caveolar and lipid raft pathways (Frances-
coni et al., 2009; Glodowski et al., 2007). Further study on the
regulation of AMPA receptor endocytosis is warranted to better
understand the various forms of synaptic plasticity underlying
learning and memory.
EXPERIMENTAL PROCEDURES
Plasmid Construction
cDNAs for PIP5Kg splicing variants, the C-terminal fragment of PIP5Kg661
corresponding to amino acids 448–661, and their mutants were prepared as
previously reported (Nakano-Kobayashi et al., 2007). cDNA for HA-GluA2
was prepared by PCR and Pyrobest (Takara). These cDNAs were cloned
into the pCAGGS expression vector. For the BiFC assay, cDNAs for the
Myc-tagged ear domain of b2 adaptin corresponding to amino acids
714–951 and HA-PIP5Kg661 or its phosphomimetic mutant PIP5Kg661
S645E were inserted into the pVenus (1–173) N-1 and pVenus (155–238) C-1
vector, respectively. For experiments with shRNAs, we developed an expres-
sion vector containing dual promoters, CAG and H1. cDNAs encoding GFP
and shRNA were inserted downstream of CAG and H1 promoters, respec-
tively. PIP5Kg targeting sequences employed in this study are 50-GGACCTG
GACTTCATGCAG-30 (Ling et al., 2007; Sun et al., 2007) and 50-CATCAAGACTGTCATGCAC-30 (a mouse version of Mao et al., 2009) for shRNA
1 and 2, respectively. The sequence for a scrambled shRNA was 50-ATTGAC
CATCACTCACTGA-30. For rescue experiments, four nucleotides were
mutated in the region targeted by shRNA 2 without changing the amino acid
sequence to generate PIP5Kres. The cDNAs encoding GFP-PIP5Kres and
shRNA 2 were inserted downstream of CAG and H1 promoters, respectively.
Culture of Hippocampal Neurons and Transfection
Hippocampi dissected from E16/17 ICR mice were treated with 10 U ml�1
papain and 100 U ml�1 DNase in Dulbecco’s modified Eagle’s medium at
37�C for 20 min. The dissociated hippocampal neurons were plated on polye-
thyleneimine (PEI)-coated plates or glass coverslips and cultured in Neuro-
basal medium (Invitrogen) with B-27 supplement (Invitrogen) and 0.5 mM
L-glutamine. After 14–20 DIV culture, the neurons were transiently transfected
with plasmids using Lipofectamine 2,000 and used for assays of AMPA
receptor endocytosis after 24–48 hr or for assays of BiFC after 19–26 hr. For
shRNA experiments, neurons were transfected with shRNA expression
vectors with or without the HA-GluA2 expression vector at 7 DIV and used
for immunostaining or assay of AMPA receptor endocytosis at 14 DIV.
Assay for AMPA Receptor Endocytosis
Hippocampal neurons transfected with plasmids for HA-GluA2 and GFP
proteins or shRNA were stimulated with 50 mM NMDA for 10 min and fixed
in 4% paraformaldehyde without permeabilization for 10 min at room temper-
ature. After neurons were washed with PBS and incubated with a blocking
solution (2% bovine serum albumin and 2% normal goat serum in PBS), the
surface HA-GluA2 was visualized with the anti-HA antibody (1:1,000) and
Alexa 546 secondary antibody (1:1,000; Invitrogen). To label the total HA-
GluA2, we then permeabilized the neurons, blocked them with a blocking
solution containing 0.4% Triton X-100, and incubated them with the anti-HA
antibody (1:1,000) and Alexa 350 secondary antibodies (1:1,000). The expres-
sion of GFP proteins was also detected by anti-GFP (1:3,000) and Alexa 488
secondary antibody (1:1,000). Fluorescence images were captured by a fluo-
rescence microscope (BX60, Olympus) equipped with a charge-coupled
device camera (DP 70, Olympus) and analyzed using IPLab software (Scana-
lytics). Only the transfected neurons with well-developed spines were
analyzed. For statistical analysis of the surface expression level of HA-
GluA2, we measured the intensity of Alexa 546 for the surface HA-GluA2
and normalized it to the intensity of Alexa 350 for the total HA-tagged GluA2.
The fluorescence intensity on dendrites between 20 mm and 100 mm from
the soma was measured.
Immunoprecipitation
To investigate the NMDA-dependent interaction of endogenous PIP5Kg661
with AP-2, we stimulated cultured neurons with 50 mM NMDA for 10 min and
solubilized them in a lysis buffer (50 mM Tris-HCl [pH 7.5], 1 mM EDTA,
0.1 mM EGTA, 5 mM MgCl2, 10 mM KCl, 5 mM NaF, 2 mM Na3VO4, 4 mM
Na4P2O7, 1 mM PMSF, and 1% Triton X-100) supplemented with a protease
inhibitor cocktail (Nacalai Tesque). PIP5Kg661 was immunoprecipitated with
an anti-PIP5Kg661 antibody conjugated with protein A sepharose (GE Health-
care). Proteins in the immunoprecipitate were blotted with anti-a adaptin,
anti-b adaptin, and anti-PIP5Kg antibodies.
BiFC Assay
For time-lapse imaging, hippocampal neurons were plated on 35 mm
PEI-coated glass-bottom dishes (thickness = 0.12–0.19 mm; Mattek or Asahi
glass) and cultured in Neurobasal mediumwithout phenol red (Invitrogen), with
B-27 supplement (Invitrogen) and 0.5 mM L-glutamineas as described above
for 16–20 DIV. The neurons were transiently transfected with plasmids for
VN-b2 ear and VC-PIP5K-WT or VC-PIP5K-S645E and cultured for another
19–26 hr. During imaging, the glass-bottom dish was kept at 32�C by a micro-
scope incubation system (Tokai Hit). Time-lapse epifluorescent images were
acquired with a Nikon Eclipse Ti inverted microscope equipped with an
Apochromatic 603 oil immersion objective (NA 1.49) and an ORCA-II-ER
camera (Hamamatsu Photonics). Image capture and data acquisition were
performed using NIS-Elements BR3.0 software (Nikon). Image sequences
were subsequently processed with NIS-Elements and ImageJ software
(1.42q, National Institutes of Health).
Neuron 73, 135–148, January 12, 2012 ª2012 Elsevier Inc. 145
Neuron
Activation of PIP5K Is Essential for LTD
For immunocytochemical analysis, hippocampal neurons cultured on PEI-
coated glass coverslips were transfected as described above. After treatment
with 50 mM NMDA for 5 min, the neurons were fixed, permeabilized, and
blocked with a blocking solution containing 0.4% Triton X-100. PSD-95 and
MAP2 were labeled with anti-PSD-95 (1:1,000) and anti-MAP2 antibodies
(1:1,000), respectively, and visualized with Alexa 546 secondary antibodies
(1:1,000). F-actin was labeled with rhodamine phalloidin. The numbers of the
Venus punctate signals in the dendrites and the total length of the dendrites
between 20 and 100 mm from the soma were measured.
Assay of PIP5Kg661 Activity
PIP5Kg661 activity was determined as previously reported (Honda et al.,
1999). For more detail, see Supplemental Experimental Procedures.
Recombinant Sindbis Virus and In Vivo Injection
The recombinant Sindbis virus for the expression of GFP and wild-type or
kinase-dead PIP5Kg661 was constructed as described (Matsuda et al.,
2003). Under the deep anesthesia with an intraperitoneal injection of ket-
amine/xylazine (80/20mg/kg;Sigma), the recombinantSindbis virus (2.5ml; titer,
1.0 3 108�1.0 3 109 TU/ml) was stereotactically injected into the CA1 region
of dorsal hippocampus of P14–21 ICR mice (2.0–2.3 mm posterior to the
Bregma, 1.5–2.0 mm lateral to the midline, and 1.5–2.0 mm ventral from the
pial surface). After 24–36 hr, infected cells were identified by the GFP expres-
sion, and hippocampal slices were used for electrophysiological analyses.
Electrophysiology
Transverse hippocampal slices (300 mm thickness) were prepared from
P14–21 ICR mice or virus-infected mice according to the institutional guide-
lines. Whole-cell patch-clamp recordings were made from visually identified
CA1 pyramidal neurons using a 603 water-immersion objective attached to
an upright microscope (BX51WI, Olympus) at room temperature. The resis-
tance of the patch pipettes was 4–6 MU when filled with an intracellular solu-
tion consisting of 150mMCs-gluconate, 10 mMHEPES (pH 7.3), 8 mMMgCl2,
2 mM Na2ATP, 0.5 mM Na2GTP, 0.2 mM EGTA, and 5 mM N-ethyl bromide
quaternary salt (QX-314) (290 mOsm/kg). For the experiments shown in
Figures 7B and 7C, synthetic peptide, pep-S645A, or pep-S645E (300 mM
each) was added to the patch pipette solution to be perfused postsynaptically.
The solution used for slice storage and recording consisted of 125 mM NaCl,
2.5 mM KCl, 2 mM CaCl2, 1 mM MgCl2, 1.25 mM NaH2PO4, 26 mM NaHCO3,
and 10 mM D-glucose, which was bubbled continuously with a mixture of 95%
O2 and 5% CO2. Picrotoxin (100 mM; Sigma) was always included in the assay
to block inhibitory synaptic transmission. To evoke EPSCs, we stimulated the
Schaffer collaterals with a glass stimulating electrode placed on the stratum
radiatum of the CA1 region (200–300 mm from the recorded neurons). The
stimulus intensity was subsequently adjusted to 50% of the maximal EPSC
amplitude. For LTD experiments, EPSCs were recorded successively from
CA1 neurons voltage clamped at –80mV at a frequency of 0.1 Hz. After stable
EPSCs were observed for least 10 min, LFS (1 Hz, 300 stimuli at –40mV) was
applied. Access resistances were monitored every 10 s by applying hyperpo-
larizing steps (2mV, 50 ms) throughout the experiments; the measurements
were discarded if the resistance changed by more than 20% of its original
value. The current responseswere recorded using an Axopatch 200B amplifier
(Molecular Devices), and the pCLAMP system (version 9.2; Molecular Devices)
was used for data acquisition and analysis. The signals were filtered at 1 kHz
and digitized at 4 kHz.
SUPPLEMENTAL INFORMATION
Supplemental Information includes nine figures, Supplemental Experimental
Procedures, and one movie and can be found with this article online at
doi:10.1016/j.neuron.2011.09.034.
ACKNOWLEDGMENTS
We thank J.Miyazaki (Osaka University), K. Nakayama (Kyoto University), M.A.
Frohman (Stony Brook University), and J.G. Donaldson (National Institutes of
146 Neuron 73, 135–148, January 12, 2012 ª2012 Elsevier Inc.
Health) for providing the pCAGGS expression vector, human b2 adaptin
cDNA, pVenus (1–173) N-1 and pVenus (155–238) C-1 vectors, and the anti-
ARF6 antibody, respectively, and S. Narumi for technical assistance. This
work was supported by Ministry of Education, Culture, Sports, Science, and
Technology (MEXT) and/or Japan Society for the Promotion of Science, Grant
in Aid for Scientific Research to Y.K. (20247010), S.M. (22700343), W.K.
(23240053), and M.Y. (23689012), by Core Research for Evolutional Science
and Technology to M.Y., and by Special Coordination Funds for Promoting
Science and Technology to H.H. from MEXT and the Mitsubishi Research
Foundation.
Accepted: September 20, 2011
Published: January 11, 2012
REFERENCES
Akiba, Y., Suzuki, R., Saito-Saino, S., Owada, Y., Sakagami, H.,Watanabe, M.,
and Kondo, H. (2002). Localization of mRNAs for phosphatidylinositol phos-
phate kinases in the mouse brain during development. Brain Res. Gene
Expr. Patterns 1, 123–133.
Bairstow, S.F., Ling, K., Su, X., Firestone, A.J., Carbonara, C., and Anderson,
R.A. (2006). Type Igamma661 phosphatidylinositol phosphate kinase directly
interacts with AP2 and regulates endocytosis. J. Biol. Chem. 281, 20632–
20642.
Beattie, E.C., Carroll, R.C., Yu, X., Morishita, W., Yasuda, H., von Zastrow, M.,
and Malenka, R.C. (2000). Regulation of AMPA receptor endocytosis by
a signaling mechanism shared with LTD. Nat. Neurosci. 3, 1291–1300.
Bhattacharyya, S., Biou, V., Xu, W., Schluter, O., and Malenka, R.C. (2009).
A critical role for PSD-95/AKAP interactions in endocytosis of synaptic
AMPA receptors. Nat. Neurosci. 12, 172–181.
Bialojan, C., and Takai, A. (1988). Inhibitory effect of a marine-sponge toxin,
okadaic acid, on protein phosphatases. Specificity and kinetics. Biochem. J.
256, 283–290.
Blanpied, T.A., Scott, D.B., and Ehlers, M.D. (2002). Dynamics and regulation
of clathrin coats at specialized endocytic zones of dendrites and spines.
Neuron 36, 435–449.
Boritzki, T.J., Wolfard, T.S., Besserer, J.A., Jackson, R.C., and Fry, D.W.
(1988). Inhibition of type II topoisomerase by fostriecin. Biochem.
Pharmacol. 37, 4063–4068.
Bredt, D.S., and Nicoll, R.A. (2003). AMPA receptor trafficking at excitatory
synapses. Neuron 40, 361–379.
Brown, T.C., Tran, I.C., Backos, D.S., and Esteban, J.A. (2005). NMDA
receptor-dependent activation of the small GTPase Rab5 drives the removal
of synaptic AMPA receptors during hippocampal LTD. Neuron 45, 81–94.
Carroll, R.C., Beattie, E.C., Xia, H., Luscher, C., Altschuler, Y., Nicoll, R.A.,
Malenka, R.C., and von Zastrow, M. (1999). Dynamin-dependent endocytosis
of ionotropic glutamate receptors. Proc. Natl. Acad. Sci. USA 96, 14112–
14117.
Chernevskaya, N.I., Obukhov, A.G., and Krishtal, O.A. (1991). NMDA receptor
agonists selectively block N-type calcium channels in hippocampal neurons.
Nature 349, 418–420.
Chowdhury, S., Shepherd, J.D., Okuno, H., Lyford, G., Petralia, R.S., Plath, N.,
Kuhl, D., Huganir, R.L., and Worley, P.F. (2006). Arc/Arg3.1 interacts with the
endocytic machinery to regulate AMPA receptor trafficking. Neuron 52,
445–459.
Cognet, L., Groc, L., Lounis, B., and Choquet, D. (2006). Multiple routes
for glutamate receptor trafficking: surface diffusion and membrane traffic
cooperate to bring receptors to synapses. Sci. STKE 2006, pe13.
Corlew, R., Brasier, D.J., Feldman, D.E., and Philpot, B.D. (2008). Presynaptic
NMDA receptors: newly appreciated roles in cortical synaptic function and
plasticity. Neuroscientist 14, 609–625.
Cottrell, J.R., Borok, E., Horvath, T.L., and Nedivi, E. (2004). CPG2: a brain-
and synapse-specific protein that regulates the endocytosis of glutamate
receptors. Neuron 44, 677–690.
Neuron
Activation of PIP5K Is Essential for LTD
Cousin, M.A., andRobinson, P.J. (2001). The dephosphins: dephosphorylation
by calcineurin triggers synaptic vesicle endocytosis. Trends Neurosci. 24,
659–665.
Cremona, O., Di Paolo, G., Wenk, M.R., Luthi, A., Kim, W.T., Takei, K., Daniell,
L., Nemoto, Y., Shears, S.B., Flavell, R.A., et al. (1999). Essential role of
phosphoinositide metabolism in synaptic vesicle recycling. Cell 99, 179–188.
Dayanithi, G., Rage, F., Richard, P., and Tapia-Arancibia, L. (1995).
Characterization of spontaneous and N-methyl-D-aspartate-induced calcium
rise in rat cultured hypothalamic neurons. Neuroendocrinology 61, 243–255.
Di Paolo, G., Pellegrini, L., Letinic, K., Cestra, G., Zoncu, R., Voronov, S.,
Chang, S., Guo, J., Wenk, M.R., and De Camilli, P. (2002). Recruitment and
regulation of phosphatidylinositol phosphate kinase type 1 gamma by the
FERM domain of talin. Nature 420, 85–89.
Di Paolo, G., Moskowitz, H.S., Gipson, K., Wenk, M.R., Voronov, S., Obayashi,
M., Flavell, R., Fitzsimonds, R.M., Ryan, T.A., and De Camilli, P. (2004).
Impaired PtdIns(4,5)P2 synthesis in nerve terminals produces defects in
synaptic vesicle trafficking. Nature 431, 415–422.
Ehlers, M.D. (2000). Reinsertion or degradation of AMPA receptors determined
by activity-dependent endocytic sorting. Neuron 28, 511–525.
Ford, M.G., Pearse, B.M., Higgins, M.K., Vallis, Y., Owen, D.J., Gibson, A.,
Hopkins, C.R., Evans, P.R., and McMahon, H.T. (2001). Simultaneous binding
of PtdIns(4,5)P2 and clathrin by AP180 in the nucleation of clathrin lattices on
membranes. Science 291, 1051–1055.
Francesconi, A., Kumari, R., and Zukin, R.S. (2009). Regulation of group I me-
tabotropic glutamate receptor trafficking and signaling by the caveolar/lipid
raft pathway. J. Neurosci. 29, 3590–3602.
Gaidarov, I., and Keen, J.H. (1999). Phosphoinositide-AP-2 interactions
required for targeting to plasma membrane clathrin-coated pits. J. Cell Biol.
146, 755–764.
Giudici, M.L., Emson, P.C., and Irvine, R.F. (2004). A novel neuronal-specific
splice variant of Type I phosphatidylinositol 4-phosphate 5-kinase isoform
gamma. Biochem. J. 379, 489–496.
Glodowski, D.R., Chen, C.C., Schaefer, H., Grant, B.D., and Rongo, C. (2007).
RAB-10 regulates glutamate receptor recycling in a cholesterol-dependent
endocytosis pathway. Mol. Biol. Cell 18, 4387–4396.
Gong, L.W., and De Camilli, P. (2008). Regulation of postsynaptic AMPA
responses by synaptojanin 1. Proc. Natl. Acad. Sci. USA 105, 17561–17566.
Guo, Y., Rebecchi, M., and Scarlata, S. (2005). Phospholipase Cbeta2 binds to
and inhibits phospholipase Cdelta1. J. Biol. Chem. 280, 1438–1447.
Harris, S.L., Gallyas, F., Jr., and Molnar, E. (2004). Activation of metabotropic
glutamate receptors does not alter the phosphorylation state of GluR1 AMPA
receptor subunit at serine 845 in perirhinal cortical neurons. Neurosci. Lett.
372, 132–136.
Honda, A., Nogami, M., Yokozeki, T., Yamazaki, M., Nakamura, H., Watanabe,
H., Kawamoto, K., Nakayama, K., Morris, A.J., Frohman, M.A., and Kanaho, Y.
(1999). Phosphatidylinositol 4-phosphate 5-kinase alpha is a downstream
effector of the small G protein ARF6 in membrane ruffle formation. Cell 99,
521–532.
Ishihara, H., Shibasaki, Y., Kizuki, N., Katagiri, H., Yazaki, Y., Asano, T., and
Oka, Y. (1996). Cloning of cDNAs encoding two isoforms of 68-kDa type I
phosphatidylinositol-4-phosphate 5-kinase. J. Biol. Chem. 271, 23611–23614.
Ishihara, H., Shibasaki, Y., Kizuki, N., Wada, T., Yazaki, Y., Asano, T., and Oka,
Y. (1998). Type I phosphatidylinositol-4-phosphate 5-kinases. Cloning of the
third isoform and deletion/substitution analysis of members of this novel lipid
kinase family. J. Biol. Chem. 273, 8741–8748.
Itoh, T., Koshiba, S., Kigawa, T., Kikuchi, A., Yokoyama, S., and Takenawa, T.
(2001). Role of the ENTH domain in phosphatidylinositol-4,5-bisphosphate
binding and endocytosis. Science 291, 1047–1051.
Kerppola, T.K. (2006). Design and implementation of bimolecular fluorescence
complementation (BiFC) assays for the visualization of protein interactions in
living cells. Nat. Protoc. 1, 1278–1286.
Kerppola, T.K. (2009). Visualization of molecular interactions using bimolecular
fluorescence complementation analysis: characteristics of protein fragment
complementation. Chem. Soc. Rev. 38, 2876–2886.
Krauss, M., Kinuta, M., Wenk, M.R., De Camilli, P., Takei, K., and Haucke, V.
(2003). ARF6 stimulates clathrin/AP-2 recruitment to synaptic membranes
by activating phosphatidylinositol phosphate kinase type Igamma. J. Cell
Biol. 162, 113–124.
Lee, S.H., Liu, L., Wang, Y.T., and Sheng, M. (2002). Clathrin adaptor AP2 and
NSF interact with overlapping sites of GluR2 and play distinct roles in AMPA
receptor trafficking and hippocampal LTD. Neuron 36, 661–674.
Lee, S.Y., Voronov, S., Letinic, K., Nairn, A.C., Di Paolo, G., and De Camilli, P.
(2005). Regulation of the interaction between PIPKI gamma and talin by
proline-directed protein kinases. J. Cell Biol. 168, 789–799.
Lin, J.W., Ju, W., Foster, K., Lee, S.H., Ahmadian, G., Wyszynski, M., Wang,
Y.T., and Sheng, M. (2000). Distinct molecular mechanisms and divergent
endocytotic pathways of AMPA receptor internalization. Nat. Neurosci. 3,
1282–1290.
Ling, K., Doughman, R.L., Firestone, A.J., Bunce, M.W., and Anderson, R.A.
(2002). Type I gamma phosphatidylinositol phosphate kinase targets and
regulates focal adhesions. Nature 420, 89–93.
Ling, K., Bairstow, S.F., Carbonara, C., Turbin, D.A., Huntsman, D.G., and
Anderson, R.A. (2007). Type I gamma phosphatidylinositol phosphate kinase
modulates adherens junction and E-cadherin trafficking via a direct interaction
with mu 1B adaptin. J. Cell Biol. 176, 343–353.
Loijens, J.C., and Anderson, R.A. (1996). Type I phosphatidylinositol-4-phos-
phate 5-kinases are distinct members of this novel lipid kinase family. J. Biol.
Chem. 271, 32937–32943.
MacDonald, M.L., Lamerdin, J., Owens, S., Keon, B.H., Bilter, G.K., Shang, Z.,
Huang, Z., Yu, H., Dias, J., Minami, T., et al. (2006). Identifying off-target effects
and hidden phenotypes of drugs in human cells. Nat. Chem. Biol. 2, 329–337.
Man, H.Y., Lin, J.W., Ju, W.H., Ahmadian, G., Liu, L., Becker, L.E., Sheng, M.,
andWang, Y.T. (2000). Regulation of AMPA receptor-mediated synaptic trans-
mission by clathrin-dependent receptor internalization. Neuron 25, 649–662.
Mao, Y.S., Yamaga, M., Zhu, X.H., Wei, Y.J., Sun, H.Q., Wang, J., Yun, M.,
Wang, Y.F., Di Paolo, G., Bennett, M., et al. (2009). Essential and unique roles
of PIP5K-gamma and -alpha in Fcgamma receptor-mediated phagocytosis.
J. Cell Biol. 184, 281–296.
Matsuda, K., Fletcher, M., Kamiya, Y., and Yuzaki, M. (2003). Specific
assembly with the NMDA receptor 3B subunit controls surface expression
and calcium permeability of NMDA receptors. J. Neurosci. 23, 10064–10073.
Matsuda, S., Miura, E., Matsuda, K., Kakegawa, W., Kohda, K., Watanabe, M.,
and Yuzaki, M. (2008). Accumulation of AMPA receptors in autophagosomes
in neuronal axons lacking adaptor protein AP-4. Neuron 57, 730–745.
Morgan, J.R., Di Paolo, G., Werner, H., Shchedrina, V.A., Pypaert, M.,
Pieribone, V.A., and De Camilli, P. (2004). A role for talin in presynaptic func-
tion. J. Cell Biol. 167, 43–50.
Morishita, W., Connor, J.H., Xia, H., Quinlan, E.M., Shenolikar, S., and
Malenka, R.C. (2001). Regulation of synaptic strength by protein phosphatase
1. Neuron 32, 1133–1148.
Mulkey, R.M., Herron, C.E., and Malenka, R.C. (1993). An essential role for
protein phosphatases in hippocampal long-term depression. Science 261,
1051–1055.
Mulkey, R.M., Endo, S., Shenolikar, S., and Malenka, R.C. (1994). Involvement
of a calcineurin/inhibitor-1 phosphatase cascade in hippocampal long-term
depression. Nature 369, 486–488.
Munton, R.P., Vizi, S., and Mansuy, I.M. (2004). The role of protein phospha-
tase-1 in the modulation of synaptic and structural plasticity. FEBS Lett.
567, 121–128.
Nakano-Kobayashi, A., Yamazaki, M., Unoki, T., Hongu, T., Murata, C.,
Taguchi, R., Katada, T., Frohman, M.A., Yokozeki, T., and Kanaho, Y. (2007).
Role of activation of PIP5Kgamma661 by AP-2 complex in synaptic vesicle
endocytosis. EMBO J. 26, 1105–1116.
Neuron 73, 135–148, January 12, 2012 ª2012 Elsevier Inc. 147
Neuron
Activation of PIP5K Is Essential for LTD
Oliet, S.H., Malenka, R.C., and Nicoll, R.A. (1997). Two distinct forms of long-
term depression coexist in CA1 hippocampal pyramidal cells. Neuron 18,
969–982.
Palmer, C.L., Lim, W., Hastie, P.G., Toward, M., Korolchuk, V.I., Burbidge,
S.A., Banting, G., Collingridge, G.L., Isaac, J.T., and Henley, J.M. (2005).
Hippocalcin functions as a calcium sensor in hippocampal LTD. Neuron 47,
487–494.
Rao, V.D., Misra, S., Boronenkov, I.V., Anderson, R.A., and Hurley, J.H. (1998).
Structure of type IIbeta phosphatidylinositol phosphate kinase: a protein
kinase fold flattened for interfacial phosphorylation. Cell 94, 829–839.
Rappoport, J.Z., and Simon, S.M. (2009). Endocytic trafficking of activated
EGFR is AP-2 dependent and occurs through preformed clathrin spots.
J. Cell Sci. 122, 1301–1305.
Rohde, G., Wenzel, D., and Haucke, V. (2002). A phosphatidylinositol (4,5)-bi-
sphosphate binding site within mu2-adaptin regulates clathrin-mediated
endocytosis. J. Cell Biol. 158, 209–214.
Schmidt, C., Peng, B., Li, Z., Sclabas, G.M., Fujioka, S., Niu, J., Schmidt-
Supprian, M., Evans, D.B., Abbruzzese, J.L., and Chiao, P.J. (2003).
Mechanisms of proinflammatory cytokine-induced biphasic NF-kappaB
activation. Mol. Cell 12, 1287–1300.
Schnabel, R., Kilpatrick, I.C., and Collingridge, G.L. (2001). Protein phospha-
tase inhibitors facilitate DHPG-induced LTD in the CA1 region of the hippo-
campus. Br. J. Pharmacol. 132, 1095–1101.
Scholz, R., Berberich, S., Rathgeber, L., Kolleker, A., Kohr, G., and Kornau,
H.C. (2010). AMPA receptor signaling through BRAG2 and Arf6 critical for
long-term synaptic depression. Neuron 66, 768–780.
Selig, D.K., Lee, H.K., Bear, M.F., and Malenka, R.C. (1995). Reexamination of
the effects of MCPG on hippocampal LTP, LTD, and depotentiation.
J. Neurophysiol. 74, 1075–1082.
148 Neuron 73, 135–148, January 12, 2012 ª2012 Elsevier Inc.
Semerdjieva, S., Shortt, B., Maxwell, E., Singh, S., Fonarev, P., Hansen, J.,
Schiavo, G., Grant, B.D., and Smythe, E. (2008). Coordinated regulation of
AP2 uncoating from clathrin-coated vesicles by rab5 and hRME-6. J. Cell
Biol. 183, 499–511.
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.
Sun, Y., Ling, K., Wagoner, M.P., and Anderson, R.A. (2007). Type I gamma
phosphatidylinositol phosphate kinase is required for EGF-stimulated direc-
tional cell migration. J. Cell Biol. 178, 297–308.
Thieman, J.R., Mishra, S.K., Ling, K., Doray, B., Anderson, R.A., and Traub,
L.M. (2009). Clathrin regulates the association of PIPKIgamma661 with the
AP-2 adaptor beta2 appendage. J. Biol. Chem. 284, 13924–13939.
Vitale, N., Chasserot-Golaz, S., Bailly, Y., Morinaga, N., Frohman, M.A., and
Bader, M.F. (2002). Calcium-regulated exocytosis of dense-core vesicles
requires the activation of ADP-ribosylation factor (ARF)6 by ARF nucleotide
binding site opener at the plasma membrane. J. Cell Biol. 159, 79–89.
Wang, Y.T., and Linden, D.J. (2000). Expression of cerebellar long-term
depression requires postsynaptic clathrin-mediated endocytosis. Neuron 25,
635–647.
Wenk, M.R., Pellegrini, L., Klenchin, V.A., Di Paolo, G., Chang, S., Daniell, L.,
Arioka, M., Martin, T.F., and De Camilli, P. (2001). PIP kinase Igamma is the
major PI(4,5)P(2) synthesizing enzyme at the synapse. Neuron 32, 79–88.
Zerial, M., andMcBride, H. (2001). Rab proteins asmembrane organizers. Nat.
Rev. Mol. Cell Biol. 2, 107–117.
Zoncu, R., Perera, R.M., Sebastian, R., Nakatsu, F., Chen, H., Balla, T., Ayala,
G., Toomre, D., and De Camilli, P.V. (2007). Loss of endocytic clathrin-coated
pits upon acute depletion of phosphatidylinositol 4,5-bisphosphate. Proc.
Natl. Acad. Sci. USA 104, 3793–3798.