Neuron, Vol. 32, 1133–1148, December 20, 2001, Copyright 2001 by Cell Press
Regulation of Synaptic Strength
by Protein Phosphatase 1
ical events in mammalian tissue, however, remains
How the targeting of signaling proteins to excitatory
synapses is modified to influence synaptic strength has
only recently begun to be analyzed, often in the context
of elucidating the mechanisms of synaptic plasticity.
Experiments using pharmacological inhibitors or ge-
netic disruptions have implicated a myriad of signaling
proteins. In particular, numerous protein kinases have
been implicated in the triggering of long-term potentia-
tion (LTP) (Malenka and Nicoll, 1999; Sanes and Licht-
man, 1999). These include PKA, PKC, and CaMKII, all
of which can modulate glutamate receptor function by
phosphorylation of specific AMPA receptor (AMPAR)
subunits (Soderling and Derkach, 2000). PKA appears
to be anchored adjacent to AMPARs via AKAP79 (A-kin-
ase anchoring protein) (Colledge et al., 2000), while its
ability to modulate NMDA receptor (NMDAR) function
may be due to its binding to the scaffolding protein
to the intracellular tail of NMDAR subunits (Bayer et al.,
2001; Leonard et al., 1999; Strack and Colbran, 1998),
positioning it in an ideal site to respond to the calcium
entry that is the essential trigger for LTP. Furthermore,
recent work indicates that the subcellular localization
of CaMKII at synapses can be dramatically modified
by activity, in particular, NMDAR stimulation (Shen and
Compared to LTP, much less work has been per-
formed on the signaling cascades involved in the trig-
is that the triggering of this form of LTD requires the
activation of a protein phosphatase cascade involving
man, 1989; Mulkey et al., 1993, 1994). PP1 appears to
be targeted to appropriate subcellular domains in neu-
rons by a family of targeting/anchoring proteins which
include spinophilin/neurabin II (Allen et al., 1997; Hsieh-
Wilson et al., 1999), neurabin I (MacMillan et al., 1999;
McAvoy et al., 1999), neurofilament-L (NF-L) (Terry-
Lorenzo et al., 2000), and yotiao (Westphal et al., 1999).
Whether PP1 localization at excitatory synapses, due
to its interactions with one or more of these proteins,
is important for the triggering of LTD is unknown, as
is whether synaptic activity dynamically modulates the
actions and/or location of PP1 at synapses. It is also
not known whether PP1 is constitutively active at syn-
apses and thereby functions to limit synaptic strength,
et al., 1999; Yan et al., 1999). Furthermore, recent work
has suggested that in neurons expressing NMDAR-
dependent LTD, there are two additional forms of LTD
that do not require PP1 activity: mGluR LTD, which is
triggered by activation of postsynaptic group I mGluRs
(Huber et al., 2000; Oliet et al., 1997; Palmer et al., 1997),
and chemLTD, which is triggered by bath application of
NMDA (Lee et al., 1998).
To further explore the role of PP1 in LTD and specifi-
cally evaluate the importance of its interactions with
Wade Morishita,1John H. Connor,2,5Houhui Xia,1,5
Elizabeth M. Quinlan,3Shirish Shenolikar,2
and Robert C. Malenka1,4
1Nancy Pritzker Laboratory
Department of Psychiatry and Behavioral Sciences
Stanford University School of Medicine
Palo Alto, California 94304
2Department of Pharmacology and Cancer Biology
Duke University Medical Center
Durham, North Carolina 27710
3Department of Biology, Neurobiology,
and Cognitive Sciences
University of Maryland
College Park, Maryland 20742
We investigated the role of postsynaptic protein phos-
phatase 1 (PP1) in regulating synaptic strength by
loading CA1 pyramidal cells either with peptides that
disrupt PP1 binding to synaptic targeting proteins or
with active PP1. The peptides blocked synaptically
evoked LTD but had no effect on basal synaptic cur-
rents mediated by either AMPA or NMDA receptors.
They did, however, cause an increase in synaptic
strength following the induction of LTD. Similarly, PP1
had no effect on basalsynaptic strength but enhanced
LTD. In cultured neurons, synaptic activation of NMDA
receptors increased the proportion of PP1 localized
to synapses. These results suggest that PP1 does not
significantly regulate basal synaptic strength. Appro-
priate NMDA receptoractivation, however, allows PP1
to gain access to synaptic substrates and be recruited
to synapses where its activity is necessary for sus-
The mechanisms of intracellular postsynaptic signaling
at excitatory synapses in the mammalian brain are of
great interest because of their importance in influencing
synaptic strength during various forms of synaptic plas-
signaling molecules (in particular, protein kinases and
protein phosphatases) are located at appropriate cellu-
ized targeting subunits or anchoring proteins (Fraser
and Scott, 1999; Hubbard and Cohen, 1993). Such tar-
geting facilitates the formation of signaling complexes,
which position the enzymes adjacent to the appropriate
protein targets and provides substrate specificity to
these broadly acting enzymes. The importance of such
5These authors contributed equally to this work.
Figure 1. Effects of Gm and I-1 Peptides on
the Formation of Neuronal PP1 Complexes
(A) The PP1 binding RKIXF sequences of the
Gm peptide, I-1 peptide, control peptide,
neurabin I, and neurabin II are illustrated with
the RKIXF binding motif underlined.
(B) The gels illustrate the overlay of rat brain
deoxycholate extracts with DIG-labeled PP1
in the presence or absence of 25 ?M Gm or
control peptide as described in Experimental
(C) The gels illustrate the cosedimentation of
PP1 using GST-neurabin II (354-494) alone or
with 25 ?M Gm or control peptide. The lack
of cosedimentation of PP1 with GST alone is
(D)The graphshowsthe effectson PP1activ-
ity of neurabin II alone (?), or with the control
peptide (?), the I-1 peptide (?), or the Gm
peptide (?). This graph is a representative
example of at least three separate experi-
ments carried out in duplicate.
neuronal targeting proteins, as well as whether it consti-
complementary approaches. First, we have studied the
effects of peptides that disrupt the interactions of PP1
with its cognate targeting proteins on these three dif-
ferent forms of LTD, as well as their effects on basal
AMPAR- and NMDAR-mediated synaptic currents. Sec-
ond, we have examined the synaptic effects of directly
loading CA1 pyramidal cells with active PP1. Third, we
bution of endogenous PP1 in cultured hippocampal
1997), has been shown to modify NMDAR function in
HEK293 cells by displacing PP1 from yotiao (Westphal
et al., 1999). The I-1 peptide contains an analogous
binding motif critical for PP1 inhibition by I-1 (Endo et
striatal neurons (Yan et al., 1999). Importantly, none of
these peptides had an effect on the enzymatic activity
of PP1 in vitro (data not shown).
To test the peptides’ ability to disrupt PP1 binding to
endogenous targeting proteins, we performed several
different biochemical assays. First, we used the pep-
nant PP1? to proteins presentin the rat brain deoxycho-
late extract by a Far Western or overlay assay. The
denaturation of PP1 binding proteins in SDS-containing
buffer focuses attention on a subset of PP1 binding
motifs, in particular, the RKIXF motif (Beullens et al.,
2000), and the SDS-PAGE allows us to analyze PP1
binding to different PP1 binding proteins simultane-
ously. Without competing peptide, PP1 bound at least
four major bands (Figure 1B), all of which represented
components of neuronal PP1 complexes (data not
shown). Previous work had identified several of these
proteins as NF-L (70 kDa), spinophilin/neurabin II (140
kDa), neurabin I (190 kDa), and yotiao (230 kDa) (Terry-
Lorenzo et al., 2000). When Gm peptide was added to
the overlay assay, binding of PP1 to these proteins was
severely diminished, but the control peptide had no ef-
slightly weaker competition in this assay.)
While a number of studies have used a similar Far
Western or overlay assay to analyze PP1 binding to
cellular proteins, we previously showed that recombi-
nant PP1 is modified in the RKIXF binding site (Endo et
al., 1996). Thus, we used a second assay, which ana-
lyzed the ability of the neuronal PP1 binding protein
spinophilin/neurabin II to recruit native PP1 from a rat
brain extract. Using a recombinant polypeptide encom-
passing the PP1 binding domain, we showed that GST-
Biochemical Characterization of Two PP1
beenidentified, includingseveralthat arefound atexcit-
atorysynapses, wherethey arethought toplay keyroles
in synaptic growth and function. These include: inhibi-
tor-1(I-1), neurabinI, spinophilin/neurabinII, yotiao,and
NF-L (Allen et al., 1997; Endo et al., 1996; Hsieh-Wilson
et al., 1999; Hubbard and Cohen, 1993; Oliver and Shen-
olikar, 1998; Price and Mumby, 1999; Terry-Lorenzo et
al., 2000). Several of these reside within the PSD and
share a 5 amino acid motif (R/K,K/R,I/V,X,F), which rep-
resents the core PP1 binding site (Egloff et al., 1997;
Liu et al., 2000) (Figure 1A). To test the importance of
PP1 targeting at the synapse, we used two different
RKIXF motif-containing peptides derived from known
PP1 binding proteins, Gm and I-1 (Hubbard and Cohen,
lular loading experiments described below. We also
used a control peptide based on a polymorphism in the
human I-1 gene (S. Shenolikar, unpublished results), a
frameshift within the RKIXF sequence that disrupts PP1
binding. The Gm peptide, derived from the glycogen-
targeting subunit found in skeletal muscle (Egloff et al.,
Regulation of Synaptic Strength by PP1
Figure 2. Loading CA1 Pyramidal Cells with
Gm Peptide Inhibits LTD
(A) Panels 1 and 2 show an example of LTD
during simultaneous whole-cell (A1) and ex-
tracellular field (A2) recordings. In this and
all subsequent figures, sample traces were
taken at the time indicated by the numbers
on the graph.
(B) Panels 1 and 2 show the summary (n ?
22) of control experiments in which whole-
cell recordings were made with pipettes con-
taining standard solution.
(C and D) Panels show an example (C1 and
C2) and summary (n ? 21) (D1 and D2) of
experiments inwhich whole-cellpipette solu-
tion contained Gm peptide.
neurabin II (354–494), but not control GST, bound and
sedimented rat brain PP1. The addition of Gm peptide
(25 ?M) resulted in significantly reduced PP1 binding
(Figure 1C). The I-1 peptide yielded similar results but
was slightly less effective than the Gm peptide.
Recent studies (Bollen, 2001) show that with the ex-
ception of yotiao (Westphal et al., 1999), all known PP1
binding proteins interact through multiple domains to
inhibit the activity of the PP1 catalytic subunit against
the in vitro substrate phosphorylase a. Thus, in a final
native PP1 from recombinant GST-neurabin II (354–494)
and thereby attenuate the inhibition of phosphorylase
phosphataseactivity (Hsieh-Wilsonetal., 1999;MacMil-
lan et al., 1999). (It is important to note that this assay
does not reflect the physiological action in situ of PP1
indicate that targeted PP1 is constitutively active [West-
phal et al., 1999; Yan et al., 1999], and we have found
complexescontaining highlyactivePP1.This assaywas
used because it is quantitative and is a reliable readout
tiveness of GST-neurabin II (354–494) to inhibit PP1 ac-
tivity. This assay also provides a quantitative compari-
son of the two peptides as disruptors of PP1 complexes
containing neurabin II. The I-1 peptide decreased the
IC50for PP1 inhibition by GST-neurabin II (354–494) by
3-fold, while the Gm peptide shifted the dose-response
curve for GST-neurabin II (354–494) by more than 10-
fold (Figure 1D). Together, these data demonstrate that
the Gm and I-1 peptides both compete for PP1 binding
rabin II and, therefore, should displace PP1 from appro-
priate synaptic sites, albeit with slightly differing effi-
Gm and I-1 Peptides Block LTD
To study the role of PP1 targeting proteins in synapti-
cally evoked NMDAR-dependent LTD (termed simply
LTD), we filled CA1 pyramidal cells with the Gm or I-1
peptides by adding them to the whole-cell pipette solu-
tion. To ensure that LTD was induced in the cells sur-
rounding the one from which we recorded, we simulta-
neously recorded a field EPSP by placing a pipette in
stratum radiatum adjacent to the whole-cell recording
pipette. LTD was reliably induced in both the cells re-
corded with standard whole-cell pipette solution and in
the adjacent population of cells (Figures 2A and 2B;
whole-cell EPSC, ?30% ? 5%; field EPSP ?24% ? 2%,
Figure 3. Loading CA1 Pyramidal Cells with I-1 Peptide Blocks LTD While Control I-1 Peptide Does Not
(A) Panels 1 and 2 show an example of an experiment in which LTD was blocked in a cell infused with the I-1 peptide (A1), but not in the
simultaneously recorded field EPSP (A2).
(B) Panels 1 and 2 show the summary (n ? 15) of experiments in which the whole-cell pipette solution contained I-1 peptide.
(C) Panels 1 and 2 show the summary of experiments (n ? 17) in which the whole-cell pipette solution contained the control I-1 peptide.
n ? 22). When cells were filled with the Gm peptide,
LTD was significantly reduced or blocked (?9% ? 4%,
n ? 21, p ? 0.01 compared to control cells), while LTD
monitored using thefield EPSP was normal(?19% ?2%,
n ? 21) (Figures 2C and 2D). Similarly, LTD was blocked
in cells filled with the I-1 peptide (?2% ? 8%, n ? 15),
while the simultaneously recorded field EPSP showed
normal LTD (?23% ? 2%) (Figures 3A and 3B). In con-
trast, cells filled with the control mutant I-1 peptide ex-
hibited clear LTD (?23% ? 4%, n ? 17, p ? 0.01 com-
pared to cells filled with I-1 peptide, field EPSP ?26%
? 2%; see Figure 3C). Thus, two different peptides that
interfere with the binding of PP1 to targeting proteins
blocked LTD while a control peptide did not.
bath application of NMDA (Kameyama et al., 1998; Lee
et al., 1998). The lack of effect of PP1 inhibitors on
chemLTD is surprising since chemLTD and LTD are mu-
tually occluding, which suggests a common expression
mechanism (Lee et al., 1998). Loading cells with either
the Gm or I-1 peptide had no effect on chemLTD (see
Figure 4C; Gm peptide, ?46% ? 7%, n ? 6; I-1 peptide,
?46% ? 6%, n ? 6). Thus, the peptides specifically
inhibited the form of LTD that previously was shown to
be blocked by PP1 inhibitors (Mulkey et al., 1993, 1994).
Gm and I-1 Peptides Do Not Affect Basal AMPAR
EPSCs or NMDAR EPSCs
A critical question for understanding the postsynaptic
actions of PP1 on synaptic transmission and plasticity
is whether basal synaptic strength is regulated by PP1
(Westphal et al., 1999; Yan et al., 1999). To address this,
we recorded AMPAR-mediated EPSCs (AMPAR EPSCs)
while infusing cells with the Gm or I-1 peptides. Figure
5 shows that neither peptide had a significant effect on
AMPAR EPSCs (Gm peptide, 11% ? 8%, n ? 11; I-1
peptide, 12% ? 7%, n ? 9, measured 15–20 min after
breakin) when compared to recordings made with the
pipette solution alone (15% ? 7%, n ? 9). Importantly,
theeffects ofthepeptides onAMPAREPSCs weremea-
sured over a time course that was sufficient to block
LTD (Figures 2 and 3).
Wealso examinedtheeffects ofthepeptides onbasal
NMDAR-mediated EPSCs (NMDAR EPSCs), which were
(0.1 mM). This is important since any effect on NMDAR
function might impair the inductionof LTD (Malenka and
Nicoll, 1993). Similar to the lack of effect on AMPAR
EPSCs, we found that the peptides had no significant
effect on basal NMDAR EPSCs when compared to re-
cordings made with the pipette solution alone (Figure
Gm and I-1 Peptides Do Not Block mGluR
LTD or chemLTD
In addition to expressing NMDAR-dependent LTD, CA1
pyramidal cells express a form of LTD that is dependent
on metabotropic glutamate receptor (mGluR) activation
(mGluR LTD) and does not appear to depend on the
activation of protein phosphatases (Bolshakov et al.,
LTD by bath application of the group I mGluR agonist
DHPG (100 ?M) (Huber et al., 2000; Palmer et al., 1997).
Figure 4A shows that this reliably elicited LTD in both
the whole-cell (WC) and field recordings (WC, ?39% ?
3%; field, ?34% ? 4%, n ? 6) and that the mGluR
antagonist LY341495blocked the actions ofDHPG (WC,
4% ? 1%; field, ?8% ? 1%, n ? 3). Loading cells with
either Gm or I-1 peptide, however, had no effect on
mGluR LTD (see Figure 4B; Gm peptide, ?39% ? 8%,
n ? 6; I-1 peptide, ?35% ? 7%, n ? 6).
Another form of LTD reported to be independent of
PP1 activity is termed chemLTD, which is induced by
Regulation of Synaptic Strength by PP1
Figure 4. mGluR LTD and chemLTD Are Not Affected by Either Gm or I-1 Peptides
(A) Panels 1 and 2 show the summary of experiments demonstrating that bath application of DHPG (100 ?M) induces LTD (n ? 6), which is
blocked by the mGluR antagonist LY341495 (100 ?M) (n ? 3).
(B) Panel 1 shows the summary of experiments demonstrating that loading cells with Gm (n ? 6) or I-1 (n?6) peptides does not inhibit mGluR
LTD. Field EPSP recordings from these experiments (n ? 12) were combined to construct the graph in (B2).
(C) Panel 1 shows the summary of experiments demonstrating that chemLTD elicited by bath application of NMDA is not blocked by Gm (n ?
6) or I-1 (n ? 6) peptides. Field EPSP recordings from these experiments (n ? 12) were combined to construct the graph in (C2).
6) (Gm peptide, 32% ? 16%, n ? 5; I-1 peptide, 19% ?
5%, n ? 4; control, 33% ? 7%, n ? 9). Thus, in contrast
basal synaptic transmission mediated by AMPARs or
might be required for PP1 to affect synapses. Figure 7B
shows that there was no significant effect of loading
cells with PP1 following a strong LTD induction protocol
(5 Hz, 3 min, while holding cells at ?50 mV) (Inactive
PP1, ?32% ? 4%, n ? 7; PP1, ?38% ? 5%, n ? 7).
However, PP1 caused a large enhancement of the LTD
elicited by a weaker induction protocol (5 Hz, 1.5 min;
Figure 7C) (Inactive PP1, ?22% ? 5%, n ? 7; PP1,
?48% ? 7%, n ? 7). We also observed the same effect
of PP1 when an even shorter induction protocol was
used (5 Hz, 0.5 min) (Inactive PP1, ?14% ? 6%, n ? 6;
PP1, ?32% ? 7%, n ? 7). This enhancing effect of
PP1 on LTD required NMDAR activation since an LTD
induction protocol in the presence of D-APV (100 ?M)
had no lasting effect on synaptic strength in PP1 loaded
cells (Figure 7D; ?7% ? 6%, n ? 4). These results indi-
cate that the lack of effect of PP1 on basal synaptic
strength cannot be explained by the fact that PP1 did
not reach the sampled synapses at a sufficiently high
Effects of PP1 on Basal Synaptic Strength and LTD
The lack of effect of the Gm and I-1 peptides on basal
synaptic strength suggests either that PP1 is not consti-
tutively active at synapses during low levels of synaptic
synaptic substrates. To help distinguish these possibili-
ties, we examined the effect of loading cells with active
PP1, a manipulation that we expected to decrease syn-
aptic strength. Surprisingly, over the course of 50 min,
9% change from baseline, n ? 7). We next tested
whether synaptic activity of the sort that elicits LTD
Figure 5. Gm and I-1 Peptides Do Not Affect
Basal AMPAR EPSCs
(A) Panels 1 and 2 show an example in which
AMPAR EPSCs (A1) and field EPSPs (A2)
were monitored while infusing Gm peptide.
(B and C) These graphs show the summary
of experiments in which Gm (B) (n ? 11) or
I-1 (C) (n ? 9) peptides were infused while
monitoring AMPAR EPSCs.
concentration. Instead, they suggest that for PP1 to
regulate synaptic strength, NMDARs must be activated.
induction may modify the functional architecture of syn-
aptic PP1 complexes such that PP1 can dephosphory-
late the protein substrates required for LTD. To test this
hypothesis, we performed two pathway experiments in
which we initially monitored field EPSPs and induced
LTD in one pathway (?25% ? 5%, n ? 5), with the other
pathway serving as a control (?5% ? 2%) (Figure 9A).
We then obtained a whole-cell recording in the same
region of the slice, and EPSCs were evoked in the same
two pathways using the same or lower stimulus inten-
sity. This procedure allowed us to compare the effect
of the Gm peptide at synapses in which LTD was in-
duced and control synapses on the same cell. Infusing
the Gm peptide caused a growth of the EPSC in the
pathway in which LTD was induced (38% ? 13%, n ?
5), but not in the control pathway (4% ? 8%) (Figures
9B and 9C). Importantly, LTD in the surrounding cells,
as measured with the field potential recording, was
maintained throughout the duration of the experiment
Similar results were obtained when the whole-cell re-
cordings were made using pipettes filled with the I-1
peptide (Figure 10). Again, LTD was first induced in one
pathway (?23% ? 4%, n ? 6), but not the control path-
way (?2% ? 3%) (Figure 10A), and then whole-cell re-
cordings were made. Infusing the I-1 peptide caused a
clear growth of the EPSC in the depressed pathway
(62% ? 24%, n ? 6), but not in the control pathway (5%
? 4%) (Figures 10B and 10C). These results suggest
that the prior induction of LTD results in changes in
synaptic PP1 function and/or recruitment that allow it
to play a more effective role in the triggering, as well as
maintenance, of subsequent LTD.
PP1 Modulates Extrasynaptic but Not
ruption of PP1 targeting elicited an increase in the re-
sponses to exogenously applied NMDA (Westphal et
al., 1999) or AMPA (Yan et al., 1999). This raises the
possibility that PP1 may have different effects on extra-
receptors. To test this hypothesis, we loaded cells with
PP1 while simultaneously monitoring AMPAR EPSCs
and the inward current generated by puffing kainate on
the soma to activate extrasynaptic AMPARs (Figure 8).
Loading cells with inactive PP1 had no effect on either
the kainate-induced currents (107% ? 4% of baseline)
or EPSCs (101% ? 9% of baseline, n ? 8) (Figure 8A).
In contrast, when cells were loaded with active PP1
(Figure 8B), the kainate responses were significantly re-
duced (60% ? 18% of baseline, n ? 9), while the EPSCs
were unaffected (98% ? 15%of baseline). These results
suggest thatwhile PP1 has directaccess to extrasynap-
tic AMPARs, synaptic AMPARs are protected from the
actions of PP1 during basal levels of synaptic activity.
Targeted PP1 Is Required to Maintain LTD
tic strength by PP1 requires NMDAR activation and that
disrupting PP1 targeting blocks LTD. How, then, do al-
terations in PP1 targeting result in the block of LTD yet
have no effect on basal synaptic transmission? Anala-
gous with work on the activity-dependent translocation
of CaMKII (Shen and Meyer, 1999), one possibility is
that the pattern of synaptic activation used to induce
tic sites. A mutually nonexclusive alternative is that LTD
Activity-Dependent Recruitment of PP1
tion during LTD induction may either physically recruit
Regulation of Synaptic Strength by PP1
Figure 6. Gm and I-1 Peptides Do Not Affect
Basal NMDAR EPSCs
(A) Panel shows anexample of an experiment
in which NMDAR EPSCs and NMDAR field
EPSPs were simultaneously recorded using
control recording solutions.
(B) Panelsshow anexample ofan experiment
in whichthe cellwas infusedwith Gmpeptide
while recording NMDAR EPSCs and NMDAR
(C and D) These graphs show a summary of
experiments in which cells were loaded with
Gm (C) (n ? 5) or I-1 (D) (n ? 4) peptides. For
comparison, a summary graph of experi-
ments in which standard whole-cell pipette
solution was used to monitor NMDAR EPSCs
(n ? 9) is also shown. The runup of the
NMDAR EPSCs observed in both control and
peptide-filled cells is likely due in large part
to the fact that the whole-cell solution, which
leaked out of the pipette tip, depressed syn-
depression was still occurring after break-in.
active PP1 to appropriate synaptic sites and/or modify
the molecular architecture at the synapse such that ap-
propriately targeted PP1 can act on relevant synaptic
substrates. To more directly examine the first of these
alternatives, we examined whether the location of en-
ified bysynaptic activity. Immunocytochemical staining of
nonactivated neurons with an antibodyto PP1?, a major
above background with a few isolated small puncta, pre-
sumably representing clustered PP1 (Figure 11A). Stimu-
lating the cells at 5 Hz for 3 min, a protocol previously
shown to elicit LTD in these cells (Carroll et al., 1999),
caused a significant increase in the level of PP1 immuno-
tribution and clustering of PP1. This stimulation-induced
tion as it was blocked by APV (Figure 11A).
While these results indicate that synaptic activity can
affect the subcellular distribution of PP1, the important
question for LTD mechanisms is whether PP1 is re-
cruited to synapses in an activity-dependent manner.
To address this issue, we identified synapses using a
synaptophysin antibody and calculated the percentage
of synapses with detectable levels of PP1. Synaptic
(Figure 11B), likely because of a redistribution of PP1
to synaptic locations. These changes were blocked by
D-APV (Figure 11B), indicating that they required
We have shown that two different peptides that inhibit
the binding of PP1 to proteins thought to be required for
its synaptic targeting block synaptically evoked NMDAR-
dependent LTD, but not mGluR LTD or chemLTD, two
ity. Surprisingly, these peptides had no effect on basal
synaptic responses mediated by either AMPARs or
NMDARs. However, at synapses at which LTD had been
induced, the peptides caused an increase in synaptic
strength. Similarly, loading cells with active PP1 had no
effect on basal synaptic strength but greatly enhanced
LTD. We also found that synaptic activation of NMDARs
in cultured hippocampal neurons caused a recruitment/
sistent with the hypothesis that the pattern of synaptic
activity used to induce LTD causes a modification of the
synaptic molecular architecture such that appropriately
targeted PP1 has access to the substrates relevant for
induction causes a recruitment of PP1 to the activated
synapses. Finally, they suggest that preserving active
PP1 at the appropriate site contributes to the mainte-
nance of LTD, at least for the first 30 min or so after it
Specificity of Peptide Actions
Using three different biochemical assays, we showed
that the two RKIXF-containing peptides, Gm and I-1,
Figure 7. Loading Cells with PP1 Does Not Affect Basal Synaptic Strength But Enhances LTD
(A) The graph shows a summary of experiments (n ? 7) in which AMPAR EPSCs were monitored while infusing active PP1 (n ?7).
(B and C) These graphs show a summary of experiments in which LTD was elicited by a strong (B) or weak (C) induction protocol in cells
loaded with active (n ? 7) or inactive (n ? 7) PP1.
(D) Graph shows a summary (n ? 4) of experiments demonstrating that application of D-APV blocks the enhancement of LTD elicited by
loading cells with PP1.
disrupted PP1 binding to a number of targeting proteins
previously identified as the major PP1 binding proteins
in deoxycholate extracts of rat brain (Terry-Lorenzo et
al., 2000). Focusing on the interactions of PP1 with
spinophillin/neurabin II, which is concentrated at excit-
atory synapses within the PSD (Allen et al., 1997; Feng
et al., 2000; Hsieh-Wilson et al., 1999), both peptides
prevented PP1 binding to a recombinant neurabin II
peptide. Interestingly, preformed PP1 complexes from
rat brain isolated on microcystin-LR-Sepharose were
more resistant to disruption by the Gm and I-1 peptides
vations that PP1 regulators demonstrate multiple inter-
actions with the PP1 catalytic subunit (Connor et al.,
is pivotal and sufficient for PP1 binding, the RKIXF-
the formation of new PP1 complexes than disrupting
existing PP1 holoenzymes.
NF-L and yotiao are different from the two neurabins
in that these PP1 binding proteins lack the RKIXF motif.
Earlier studies (Westphal et al., 1999) showed that the
Gm peptide could inhibit PP1 association with yotiao,
and our data suggest that PP1 association with the
RKIXF peptides prevents its recruitment by a variety of
regulators: those containing this motif, as well as those,
like NF-L, that do not. Emerging evidence suggests that
sequences flanking the RKIXF-motif contribute to the
affinity or specificity of PP1 binding peptides; thus,
longer peptides, while more effective in disrupting spe-
in modulating PP1-mediated events in cells. In this re-
gard, the Gm peptide, though a more effective reagent
for disrupting PP1 binding in vitro, was somewhat less
effective than the I-1 peptide when introduced into CA1
neurons. The reason for this difference is unknown but
may include turnover, affinity, accessibility, and speci-
ficity. Finally, numerous studies have shown that tar-
geting subunits inhibit PP1 activity against the com-
monly used in vitro substrate, phosphorylase a. In
contrast, where the physiological substrate and its rele-
units enhanced the activity of PP1 (Liu and Brautigan,
2000; Moorhead et al., 1998). The fact that peptides that
disrupt PP1 binding to spinophilin/neurabin II or yotiao
increase AMPAR and NMDAR responses (Westphal et
al., 1999; Yan et al., 1999) provides further evidence
that PP1 is active when bound to its targeting partners.
Regulation of Synaptic Strength by PP1
Figure 8. PP1 Depresses Responses of Extrasynaptic AMPARs But Not EPSCs
(A) Panels 1 and 2 show an example (A1) and summary (n ? 8) (A2) of effects of loading cells with inactive PP1 on responses to kainate
(applied to the soma) and AMPAR EPSCs.
(B) Panels 1 and 2 show an example (B1) and summary (n ? 9) (B2) of effects of loading cells with active PP1 on responses to kainate and
Importantly, none of the peptides had any effect on PP1
catalytic activity (analyzed in vitro) and thus, must have
exerted their effects via a mechanism distinct from the
standard PP1 inhibitors previously found to block LTD
(Mulkey et al., 1993, 1994).
The inability of the RKIXF-containing peptides to dis-
rupt preformed PP1 complexes in vitro, in contrast to
their ability to compete for PP1 binding to newly pre-
sented targeting proteins, suggests a dynamic associa-
tion of PP1 with targeting proteins in vivo such that the
peptides disrupt or interfere with the binding of relevant
PP1 targeting proteins. Thus, the simplest explanation
for the actions of the peptides is that they interfered
with the activity-dependent recruitment or modulation
of PP1 complexes at critical synaptic sites during the
LTD induction protocol and that the initial maintenance
of LTD requires the sustained PP1 binding to its tar-
geting protein (as well as its catalytic activity). Such a
hypothesis also explains why the peptides blocked LTD
and enhanced synaptic strength at previously de-
pressed synapses but had no effect on basal synaptic
responses. Another possible explanation for our results
is that targeted PP1 activity was inhibited by high con-
centrations of phosphorylated I-1. It seems unlikely,
however, that there could be sufficient phosphorylated
I-1 to block the actions of the exogenous PP1 that was
present in the pipette solution at high concentrations.
Furthermore, this hypothesis does not explain the activ-
ity-dependent redistribution of PP1 to synapses ob-
served in cultured neurons
do not distinguish the binding of PP1 to specific tar-
geting proteins and thus, our results do not allow us to
determine which PP1 targeting protein(s) is particularly
lin/neurabin II, the genetic deletion of which results in
the almost complete inhibition ofLTD (Feng et al., 2000).
Lack of Effect of Gm and I-1 Peptides or PP1
on Basal Synaptic Transmission
Previous work has used similar or identical peptides to
examine the role of PP1 in controlling the function of
AMPARs (Yan et al., 1999) and NMDARs (Westphal et
al., 1999). Specifically, in acutely dissociated neostriatal
neurons, the rundown of inward currents generated by
activation of AMPARs with kainate was blocked by a
peptide that interferes with the binding of PP1 to
spinophilin (Yan et al., 1999). This suggested that active
PP1 bound to spinophilin negatively regulates AMPAR
function in these cells. Similarly, in HEK293 cells ex-
Figure 9. Infusing Cells with Gm Peptide Increases AMPAR EPSCs at Synapses Expressing LTD
(A) Graphs show a summary of experiments (n ? 5) from field EPSP recordings. LTD was induced in one pathway (?) while the other pathway
(?) served as a control. Whole-cell recordings were then established using pipettes filled with solutions containing the Gm peptide. The
absence of points during WC establishment is the period in which the field was depressed when the whole-cell pipette was lowered into the
slice. Once the field recovered from the depression, WC configuration was attained. Note that after establishing the whole-cell recordings,
LTD was maintained in the test pathway.
(B) Graphs show summary of changes in the AMPAR EPSCs recorded from the two pathways.
(C) Bars show summary of changes in the amplitude of AMPAR EPSCs in the two pathways during the first 1–5 min of recording and 25–30
min after breakin (*p ? 0.05).
pressing NMDARs and yotiao, infusion of the Gm pep-
tide was found to cause a runup of the inward currents
generated by application of NMDA (Westphal et al.,
1999). These results suggested that PP1 activity, in this
case associated with yotiao, also negatively regulates
NMDARs. We were therefore surprised to find that the
Gm and I-1 peptides had minimal effect on basal synap-
tic currents mediated by either AMPARs or NMDARs.
One important difference between these previous stud-
ies and our experiments is that we examined the synap-
inward currents generated by application of exogenous
ligands to isolated cells. The regulation of PP1 localiza-
tion and activity may be significantly different at intact
synapses, compared to its properties in heterologous
cells or at extrasynaptic receptors.
The lack of effect of PP1 itself on basal synaptic
strength lends further support to the idea that synaptic
substrates such as AMPARs are relatively inaccessible
to PP1 actions during basal levels of activity. PP1 was
able, however, to strongly depress the responsiveness
of extrasynaptic AMPARs, indicating that the molecular
architecture at excitatory synapses plays a critical role
in controlling the modulation of AMPARs and perhaps
other important synaptic proteins. Together, these re-
a significant role in the regulation of AMPAR or NMDAR
function at excitatory synapses on CA1 pyramidal cells.
Consistent with this conclusion, we previously found
that infusing cells with the PP1 inhibitor microcystin LR
had no detectable effect on AMPAR EPSCs (Issac and
Malenka, unpublished observations) or on NMDAR
EPSCs (R. Mulkey and R.C.M., unpublished observa-
Activity Affects the Synaptic Actions and Location
Although loading cells with PP1 had no effect on basal
tude of LTD in an NMDAR-dependent manner. This sug-
gests that NMDAR activation during the LTD induction
protocol modified the molecular architecture of the syn-
Regulation of Synaptic Strength by PP1
Figure 10. Infusing Cells with I-1 Peptide Increases AMPAR EPSCs at Synapses Expressing LTD
(A) Graphs show summary of experiments (n ? 6) from field EPSP recordings. LTD was induced in one pathway (?) while the other pathway
(?) served as a control. Whole-cell recordings were then established using pipettes filled with solutions containing the I-1 peptide. The absence
of points during WC establishment is the period in which the field was depressed when the whole-cell pipette was lowered into the slice.
Once the field recovered from the depression, WC configuration was attained. Note that after establishing the whole-cell recordings, LTD
was maintained in the test pathway.
(B) Graphs show summary of changes in the AMPAR EPSCs recorded from the two pathways.
(C) Bars show a summary of changes in the amplitude of AMPAR EPSCs in the two pathways during the first 1–5 min of recording and 25–30
min after breakin (*p ? 0.05).
apses such that PP1 could now access the critical syn-
aptic substrates. A useful analogy can be made to the
in skeletal muscle which is due to PKA (Johnson et al.,
1994; Sculptoreanu et al., 1993). A peptide that blocks
the binding of PKA to an AKAP substantially reduced
of the Ca2?channels (Johnson et al., 1994). Similarly,
loading cells with the catalytic subunit of PKA had no
effect on the level or voltage dependence of basal Ca2?
channel activity but did rescue the depolarization-
induced potentiation in the presence of the peptide in-
hibitor. Thus, PKA anchoring is not required to maintain
the basal level of activity of the Ca2?channels, and PKA
requires depolarization of the membrane to exert its
effects, perhaps because of a voltage-dependent con-
formational change in the Ca2?channel itself (Johnson
et al., 1994).
that NMDAR activation also results in the redistribution
or recruitmentof PP1 to synapses,presumably because
PP1 can now bind to a synaptic targeting protein. A
similar activity-dependent recruitment to synapses oc-
curs to CaMKII (Shen and Meyer, 1999), likely as a con-
sequence of its binding to the intracellular tails of
NMDAR subunits (Bayer et al., 2001; Leonard et al.,
intriguing possibility that the pattern of synaptic activity
controls synaptic strength by strongly influencing the
composition of the intracellular signaling cascades
found at individual excitatory synapses.
Protein Phosphatase Activity and LTD
In contrast to LTP, the triggering of which has been
suggested to involve a number of intracellular signaling
sis involving a protein phosphatase cascade has domi-
nated the thinking about the induction of LTD (Lisman,
1989). Specifically, it was proposed that a modest rise
in calcium preferentiallyactivates calcineurin, which de-
ity of PP1 via a mechanism of disinhibition. Consistent
Figure 11. Synaptic Activation of NMDARs Recruits PP1 to Synapses
(A) Panels 1 and 2 show examples (A1) and quantitation (A2) of PP1 immunoreactivity in control, stimulated, and stimulated in the presence
of D-APV (100 ?M) cultures. (*p ? 0.01.)
(B) Panels 1 and 2 show examples (B1) and quantitation (B2) of percentage of synapses (defined by synatophysin puncta) that contain PP1
in control, stimulated, and stimulated in the presence of D-APV cultures. (*p ? 0.01).
with this hypothesis, a number of different calcineurin
and PP1 inhibitors were found to block or inhibit LTD
when loaded into CA1 pyramidal cells (Mulkey et al.,
tenance of LTD, at least over the course of 20–40 min,
required persistent phosphatase activity since applica-
tion of calyculin A caused an increase in synaptic
strength in a previously depressed pathway, but not in
the simultaneously recorded control pathway (Mulkey
et al., 1993). Our results using the Gm and I-1 peptides
confirm this result and extend it by demonstrating that
the binding of PP1 to a cognate targeting protein is
required for the maintenance ofPP1 activity during LTD.
Biochemical measurements have also shown persistent
protein phosphatase activity following the generation of
LTD in the hippocampus in vivo, although the increase
in PP1 activity lasted somewhere between 5–35 min
following the induction of LTD, while PP2A activity re-
mained elevated for over 1 hr (Thiels et al., 1998).
Two importantquestions remainabout therole ofPP1
in LTD.First, isPP1 activityabsolutely requiredfor LTD?
It isnow clear thatthere aremultiple forms ofLTD, some
of which do not require PP1 activity (Bear and Linden,
2001; Bolshakov et al., 2000), and that these may even
coexist at the same set of excitatory synapses (Oliet et
in the hippocampus, which has been reported to mutu-
ally occlude with synaptically evoked LTD (Lee et al.,
1998), is not blocked by PP1 inhibitors (Kameyama et
al., 1998) or by the Gm or I-1 peptides. This suggests
that the repetitive synaptic activation of NMDARs used
to induce LTD may modify intracellular signaling cas-
cades in a manner distinct from that which occurs fol-
ligand. Indeed, the endocytosis of AMPARs caused by
bath application of NMDA to cultured neurons (and
which is thought to contribute to the expression of LTD,
see below) was not blocked by PP1 inhibitors (Beattie
et al., 2000) (but see Ehlers, 2000). Nevertheless, both
present and previous results (Mulkey et al., 1993, 1994)
indicate that the appropriate targeting and activity of
PP1 appears to be essential for synaptically induced
NMDAR-dependent LTD. A useful analogy can be made
to the properties of CaMKII, the activation of which is
strongly influenced by temporal properties of the cal-
cium transients that normally activate it (De Koninck
and Schulman, 1998), and the translocation of which to
activation (Shen et al., 2000). This activity-dependent
modulation of CaMKII and PP1 localization may provide
additional flexibility to the intracellular signaling cas-
cades activated during LTP and LTD, and may help
in preserving the malleability of synapses which have
previously been strengthened or depressed. It could
ticity (Abraham and Bear, 1996).
A second important question is what are the critical
substrates of PP1 that contribute to the expression of
ular, the GluR1 subunit. Both chemLTD and synaptically
evoked LTD are accompanied by dephosphorylation of
serine 845 on GluR1, a dephosphorylation that is main-
tained for at least 1 hr after LTD induction (Lee et al.,
1998, 2000). Importantly, both synaptically evoked LTD
and the dephosphorylation of serine 845 were blocked
by the PP1/2A inhibitor okadic acid (Lee et al., 2000).
Regulation of Synaptic Strength by PP1
Depotentiation is also blocked by PP1 inhibitors (O’Dell
and Kandel, 1994), although this results in the dephos-
phorylation of a different site, serine 831, on GluR1 (Lee
et al., 2000).
The phosphorylation state of GluR1 influences the
single channel conductance of AMPARs (Derkach et al.,
1999; Soderling and Derkach, 2000), a mechanism that
likely contributes to LTP (Benke et al., 1998), as well as
their peak open channel probability (Banke et al., 2000).
LTD, however, does not appear to involve a change
in single channel conductance (Lu ¨thi et al., 1999), and
have not been examined. Strong evidence has been
presented indicating that LTD involves the endocytosis
of synaptic AMPARs (Carroll et al., 2001, 1999; Lu ¨scher
et al., 1999; Man et al., 2000). The intracellular signaling
cytosis following NMDAR activation have recently been
studied in cultured neurons, and the actions of pharma-
cological inhibitors support a critical role for both cal-
cineurin (Beattie et al., 2000; Ehlers, 2000) and perhaps
PP1 (Ehlers, 2000) (but see Beattie et al., 2000) in this
vations concerning the events contributing to the trig-
gering and maintenance of LTD is that NMDAR-depen-
dent activation of calcineurin initially triggers AMPAR
endocytosis via dephosphorylation of endocytic pro-
teins (Beattie et al., 2000), as has been proposed for
endocytosis of presynaptic vesicles (Lai et al., 1999).
Internalized AMPARs may then be stabilized intracellu-
larly through the PP1-dependent dephosphorylation of
specific residues on AMPAR subunits such as serine
845 on GluR1 (Ehlers, 2000) or serine 880 on GluR2, a
residue which, when dephosphorylated, greatly in-
creases the affinity of GluR2 for GRIP (Chung et al.,
2000; Matsuda et al., 1999). Disruption of GRIP binding
to AMPARs due to PKC-dependent phosphorylation of
serine 880 has been suggested to be particularly impor-
tant for the reinsertion of AMPARs into the postsynaptic
membrane following their internalization (Daw et al.,
2000). According to this hypothesis, the disruption of
PP1 function/targeting blocks LTD expression by pre-
venting the internal retention of endocytosed AMPARs,
which consequently recycle back to the membrane sur-
face. ChemLTD may not require PP1 activity because
theprolonged activationof NMDARsforces theinternal-
recently has been reported to be accompanied by in-
creased phosphorylation of serine 880 on GluR2, an
effect that was blocked by both NMDAR antagonists
and the PP1 inhibitor okadaic acid (Kim et al., 2001).
Furthermore, based on the effects of peptide inhibitors,
these authors suggested that the interaction of GluR2
with PICK1, not GRIP, was particularly important for
LTD. Clearly, further work needs to be done to clarify
the exact role of the phosphorylation state of AMPAR
subunits in mediating LTD.
a critical role for PP1 in the triggering and initial mainte-
nance of NMDAR-dependent LTD. Our results suggest
that like the extensively studied protein kinase CaMKII,
synaptic activation of NMDARs may not only influence
the catalytic activity of PP1, but also its subcellular and
perhaps subsynaptic localization, which is likely critical
forpositioning itnexttothe appropriatesubstrates.This
additional level of complexity in the control of signal
transduction at synapses, while making the study of
tant substrate for the extensive repertoire of plasticity
mechanisms that appear to exist at individual excitatory
Peptides were synthesized by the BioPolymer Analysis Laboratory
(University of Pennsylvania). Purity and concentration of the pep-
tides were assessed by HPLC. Peptides were dissolved into 10 mM
Tris·HCl (pH 7.5). PP1 overlay assays were done using standard
techniques(Connor etal.,2000).Briefly, 40?gofa 1%deoxycholate
extract from rat brain was separated on 1%–10% SDS-PAGE gel
and electrophoretically transferred to a PVDF membrane. Following
transfer, gelswere blocked withdry milk and incubatedwith digoxy-
genin-conjugated PP1 (DIG-PP1), with or without 25 ?M of the com-
peting peptide. PP1 binding was detected by Western blotting with
an anti-digoxygenin antibody (Roche Biochemicals).
neurabin II (354–494), 2 ?g of recombinant GST-neurabin II was
incubated with 10 ?l of glutathione-Sepharose beads (Pharmacia)
for 30 min at 4?C. GST beads were washed four times with TBS
(Tris·HCl [pH 7.5], 150 mM NaCl). Two hundred microliters of a 20
U/ml solution of PP1 (approximately 100 ng PP1) with or without 25
?M peptide was added and the mixture was incubated at 4?C for
30 min. Beads were washed four times with 1 ml TBS. The PP1,
which remained bound to the beads, was eluted with 25 ?l 2? SDS
sample buffer. The eluted proteins were separated on a 10% SDS-
PAGE gel and electrophoretically transferred to a PVDF membrane.
PP1 was detected by Western blotting using an anti-PP1 antibody
Protein phosphatase assays were carried out as described pre-
viously (Connor et al., 2000). Assayswere run in a 60 ?g total volume
containing 20 ?M32P-phosphorylase a as a substrate and 0.2 units
of PP1 in a 50 mM Tris·HCl (pH 7.5), 1 mM EDTA, 0.1% 2-mercapto-
ethanol reaction buffer. Peptides were added to the PP1 dilution
15–60 min prior to use. GST-neurabin II (354–494) was preincubated
with PP1 for 5 min prior to the addition of substrate. Following 10
min at 37?C, reactions were terminated using 200 ?l 20% TCA and
50 ?l 10 mg/ml BSA. The precipitated protein was sedimented by
ing of 200 ?l of the supernatant.
Hippocampal slices (400 ?m) were prepared from 2- to 4-week-old
for a minimum of 1 hr, and then transferred to a submersion-type
recording chamber mounted on an Olympus BX50WI microscope
equipped with IR DIC optics, which allowed visualization of individ-
ual CA1 pyramidal cells with a 40? objective. The slices were per-
fused at room temperature (23?C) with a standard external solution
that was bubbled continuously with 95% O2and 5% CO2and con-
taining: 119 mM NaCl, 2.5 mM KCl, 2.5 mM CaCl2, 1.3 mM MgSO4,
1.0 mM NaH2PO4, 26.2 mM NaHCO3, 11 mM glucose, and 0.1 mM
picrotoxin.Field potentialand whole-cellrecording techniqueswere
as previously described (Isaac et al., 1995; Selig et al., 1995). Whole-
cell recording pipettes (2–4 M?) were filled with a solution con-
taining: 117.5 mM CsMeSO4, 10 mM HEPES, 0.5 mM EGTA, 8 mM
NaCl, 10 mM glucose, and 2 mM Mg-ATP (pH 7.2 with CsOH, osmo-
lality adjusted to 280–290 mOsm). When experiments were per-
formed in the presence of DHPG, 0.3 mM GTP was added to the
By using peptides that disrupt the binding of PP1 to
its cognate targeting proteins, but do not inhibit PP1
catalytic activity, we have provided further evidence for
pipette solution. Peptides from a stock solution were added to the
pipette solution immediately before recording so that their final con-
centration was 100–200 ?g/ml (50–120 ?M). PP1 was prepared from
rabbit skeletal muscle (DeGuzman and Lee, 1988) and added to the
pipette solution from a stock solution at a final concentration of 400
U/ml. To inactivate PP1, it was heated to 90?C for 60 min before
adding it to the pipette solution. Cells were held at ?65 to ?75
mV during the recordings except where noted. Series and input
resistances were monitored online throughout each experiment.
with 1 M NaCl. Stimulation of Schaffer collateral/commissural affer-
ents was performed using stainless steel bipolar electrodes and
To ensure stability of the recordings when monitoring basal
AMPAR and NMDAR EPSCs, electrical stimulation was initiated be-
fore the whole-cell recording was established. After breakin, re-
cordings were started within 1–3min. During this time period, stimu-
lation strength was adjusted, after which the first 6–9 responses
wereaveraged andnormalizedto 100%forcomparisonto allsubse-
quent responses. The variable runup of the AMPAR and NMDAR
EPSCs observed in both control and peptide filled cells is likely due
in large part to the fact that the whole-cell solution, which leaks out
from this depression was still occurring during the first minutes after
breakin. For two pathway experiments (Figures 9 and 10), LTD was
generated in one pathway using 5 Hz stimulation for 3 min. After
LTD had stabilized (20–35 min after LTD induction), a whole-cell
was determined using Student’s paired and unpaired t tests.
thresholded, and Metamorph software gave two sets of puncta with
central x-y coordinates and equivalent radii. These two sets of data
were fed into a custom written algorithm (Dr. Peng Liu, Intel Inc.),
which calculated all the distances between any points from the PP1
image and any points from the synaptophysin image. If the distance
between coordinates was smaller than or equal to the sum of the
two radii of the two puncta being compared, those two puncta were
considered colocalized. The “n” value given for each experiment
refers to the number of cells analyzed.
This work was supported by grants from NIH (DK52054 to S.S.;
MH00942 to R.C.M.). J.H.C. is a Terry and Frances Fellow in Cancer
at Duke University.
Received February 22, 2001; revised November 8, 2001.
Abraham, W.C., and Bear, M.F. (1996). Metaplasticity: the plasticity
of synaptic plasticity. Trends Neurosci. 19, 126–130.
Allen, P.B., Ouimet, C.C., and Greengard, P. (1997). Spinophilin, a
novel protein phosphatase 1 binding protein localized to dendritic
spines. Proc. Natl. Acad. Sci. USA 94, 9956–9961.
Banke, T.G., Bowie, D., Lee, H.K., Huganir, R.L., Schousboe, A., and
Traynelis, S.F. (2000). Control of GluR1 AMPA receptor function by
cAMP-dependent protein kinase. J. Neurosci. 20, 89–102.
H. (2001). Interaction with the NMDA receptor locks CaMKII in an
active conformation. Nature 411, 801–805.
W.M. Cowan, T.C. Sudhof, and C.F. Stevens, eds. (Baltimore: The
Johns Hopkins University Press), pp. 455–517.
Beattie, E.C., Carroll, R.C., Yu, X., Morishita, W., Yasuda, H., von
endocytosis by a signaling mechanism shared with LTD. Nat. Neu-
rosci. 3, 1291–1300.
Benke,T.A., Luthi,A.,Isaac, J.T.,andCollingridge,G.L. (1998).Mod-
ulation of AMPA receptor unitary conductance by synaptic activity.
Nature 393, 793–797.
Beullens, M., Vulsteke, V., Van Eynde, A., Jagiello, I., Stalmans, W.,
and Bollen, M. (2000). The C-terminus of NIPP1 (nuclear inhibitor of
protein phosphatase-1) contains a novel binding site for protein
phosphatase-1 that is controlled by tyrosine phosphorylation and
RNA binding. Biochem. J. 352, 651–658.
Bollen, M. (2001). Combinatorial control of protein phosphatase-1.
Trends Biochem. Sci. 26, 426–431.
Bolshakov, V.Y., Carboni, L., Cobb, M.H., Siegelbaum, S.A., and
Belardetti, F. (2000). Dual MAP kinase pathways mediate opposing
forms of long-term plasticity at CA3-CA1 synapses. Nat. Neurosci.
R.C. (1999). Rapid redistribution of glutamate receptors contributes
to long-term depression in hippocampal cultures. Nat. Neurosci. 2,
Carroll, R.C., Beattie, E.C., Von Zastrow, M., and Malenka, R.C.
(2001). Role of AMPA receptor endocytosis in synaptic plasticity.
Nat. Rev. Neurosci. 2, 315–324.
Chung, H.J., Scannevin, R.H., Zhang, X., and Huganir, R.L. (2000).
Phosphorylation of the AMPR receptor subunit GluR2 differentially
regulates its interaction with PDZ domain-containing proteins. J.
Neurosci. 20, 7258–7267.
and Scott, J.D. (2000). Targeting of PKA to glutamate receptors
through a MAGUK-AKAP complex. Neuron 27, 107–119.
Connor, J.H., Frederick, D., Huang, H., Yang, J., Helps, N.R., Cohen,
P.T., Nairn, A.C., DePaoli-Roach, A., Tatchell, K., and Shenolikar, S.
(2000). Cellular mechanisms regulating protein phosphatase-1. A
Cell Culture and Immunocytochemistry
Hippocampal cell cultures were prepared as described previously
(Lu ¨scher et al., 1999). Briefly, hippocampi were taken from P0 rat
pups, and the dentate gyri were removed. Tissue dissociation was
facilitated by papain treatment and followed by trituration with glass
pipettes. Cells were plated on poly D-lysine coated cover slips and
grown in neurobasal medium supplemented with B27. Media was
changed on the day after plating and in part each week thereafter.
Glial growth was inhibited by FUDR after one week in culture.
at 1 ?g/ml for labeling endogenous PP1. Rabbit polyclonal anti-
synaptophysin antibody (Zymed Laboratory) was used at 1 ?g/ml.
Two-week-old primary hippocampal neurons were field stimulated
using an LTD protocol (Carroll et al., 1999) in Hepes buffered ringer
15 min. The cells were then incubated with permeabilizing blocking
buffer (P buffer: 0.1% triton X 100, 2% BSA in PBS) for at least 40
min. Theprimary antibodies(anti-PP1 andanti-synaptophysin) were
they were stained with secondary antibodies (Alexa 488 conjugated
anti-mouse secondary antibody [Molecular Probes] for anti-PP1 an-
tibody, and Alexa 568 conjugated anti-rabbit secondary antibody
[Molecular Probes] for synaptophysin).
Labeled cells were imaged using a 63? oil immersion objective
on a Zeiss Axioskop2. Images were obtained using a cooled CCD
camera (Hammamatsu) and were analyzed using Metamorph Soft-
ware (Universal Imaging). For individual experiments, cover slips
from the same culture preparation were compared, and images for
all conditions were analyzed using identical acquisition parameters.
All data acquisition and analysis was performed blindly, without
Images from each experiment were thresholded to subtract the
average background fluorescence, and cell bodies were excluded
from analysis. The total thresholded, fluorescently-labeled PP1 sig-
nal was measured automatically by the Metamorph software and
divided by the total cell area, which was determined by setting a
experiment, PP1fluorescence in allcells was normalizedby dividing
by the average fluorescence of nonstimulated control cells.
For the experiments comparing colocalization of PP1 and synap-
tophysin, corresponding PP1 and synaptophysin images were
Regulation of Synaptic Strength by PP1
key functional interaction between inhibitor-2 and the type 1 protein
phosphatase catalytic subunit. J. Biol. Chem. 275, 18670–18675.
Daw, M.I., Chittajallu, R., Bortolotto, Z.A., Dev, K.K., Duprat, F.,
Henley, J.M., Collingridge, G.L., and Isaac, J.T.R. (2000). PDZ pro-
teins interacting with C-terminal GluR2/3 are involved in a PKC-
dependent regulationof AMPA receptors athippocampal synapses.
Neuron 28, 873–886.
DeGuzman, A., and Lee, E.Y.C. (1988). Preparation of low-molecular
weight forms of rabbit muscle protein phosphatase. Methods Enzy-
mol. 159, 356–368.
De Koninck, P., and Schulman, H. (1998). Sensitivity of CaM kinase
II to the frequency of Ca2?oscillations. Science 279, 227–230.
Derkach, V., Barria, A., and Soderling, T.R. (1999). Ca2?/calmodulin-
5-methyl-4-isoxazolepropionate type glutamate receptors. Proc.
Natl. Acad. Sci. USA 96, 3269–3274.
Egloff, M.P., Johnson, D.F., Moorhead, G., Cohen, P.T., Cohen, P.,
and Barford, D. (1997). Structural basis for the recognition of regula-
tory subunits by the catalytic subunit of protein phosphatase 1.
EMBO J. 16, 1876–1887.
Ehlers, M.D. (2000). Reinsertion or degradation of AMPA receptors
determined by activity-dependent endocytic sorting. Neuron 28,
Endo, S., Zhou, X., Connor, J., Wang, B., and Shenolikar, S. (1996).
Multiple structural elements define the specificity of recombinant
human inhibitor-1 as a protein phosphatase-1 inhibitor. Biochemis-
try 35, 5220–5228.
Feng, J., Yan, Z., Ferreira, A., Tomizawa, K., Liauw, J.A., Zhuo,
M., Allen, P.B., Ouimet, C.C., and Greengard, P. (2000). Spinophilin
regulates the formation and function of dendritic spines. Proc. Natl.
Acad. Sci. USA 97, 9287–9292.
Fraser, I.D., and Scott, J.D. (1999). Modulation of ion channels: a
“current” view of AKAPs. Neuron 23, 423–426.
Hsieh-Wilson, L.C., Allen, P.B., Watanabe, T., Nairn, A.C., and
tein spinophilin and its interactions with protein phosphatase-1.
Biochemistry 38, 4365–4373.
Hubbard, M.J., and Cohen, P. (1993). On target with a new mecha-
nism for the regulation of protein phosphorylation. Trends Biochem.
Sci. 18, 172–177.
Huber, K.M., Kayser, M.S., and Bear, M.F. (2000). Role for rapid
dendritic protein synthesis in hippocampal mGluR-dependent long-
term depression. Science 288, 1254–1257.
Isaac, J.T., Nicoll, R.A., and Malenka, R.C. (1995). Evidence for silent
synapses: implications for the expression of LTP. Neuron 15,
Johnson, B.D., Scheuer, T., and Catterall, W.A. (1994). Voltage-
dependent potentiation of L-type Ca2?channels in skeletal muscle
Acad. Sci. USA 91, 11492–11496.
Kameyama, K., Lee, H.K., Bear, M.F., and Huganir, R.L. (1998).
Involvement of a postsynaptic protein kinase A substrate in the
expression of homosynaptic long-term depression. Neuron 21,
Kim, C.H., Chung, H.J., Lee, H.K., and Huganir, R.L. (2001). Interac-
tion of the AMPA receptor subunit GluR2/3 with PDZ domains regu-
lates hippocampal long-term depression. Proc. Natl. Acad. Sci. USA
Lai, M.M., Hong, J.J., Ruggiero, A.M., Burnett, P.E., Slepnev, V.I.,
De Camilli, P., and Snyder, S.H. (1999). The calcineurin-dynamin 1
complex as a calcium sensor for synaptic vesicle endocytosis. J.
Biol. Chem. 274, 25963–25966.
Lee, H.K., Kameyama, K., Huganir, R.L., and Bear, M.F. (1998).
NMDA induces long-term synaptic depression and dephosphoryla-
tion of the GluR1 subunit of AMPA receptors in hippocampus. Neu-
ron 21, 1151–1162.
Lee, H.K., Barbarosie, M., Kameyama, K., Bear, M.F., and Huganir,
R.L. (2000). Regulation of distinct AMPA receptor phosphorylation
sites during bidirectional synaptic plasticity. Nature 405, 955–959.
Leonard, A.S., Lim, I.A., Hemsworth, D.E., Horne, M.C., and Hell,
J.W. (1999). Calcium/calmodulin-dependent protein kinase II is as-
sociated with the N-methyl-D-aspartate receptor. Proc. Natl. Acad.
Sci. USA 96, 3239–3244.
Lisman, J. (1989). A mechanism for the Hebb and the anti-Hebb
processes underlying learning and memory. Proc. Natl. Acad. Sci.
USA 86, 9574–9578.
Liu, J., and Brautigan, D.L. (2000). Glycogen synthase association
phatase-1. Synthase activation involves scaffolding regulated by
beta-adrenergic signaling. J. Biol. Chem. 275, 26074–26081.
Liu, J., Wu, J., Oliver, C., Shenolikar, S., and Brautigan, D.L. (2000).
Mutations of the serine phosphorylated in the protein phosphatase-
1-binding motif in the skeletal muscle glycogen-targeting subunit.
Biochem. J. 346, 77–82.
Lu ¨scher, C., Xia, H., Beattie, E.C., Carroll, R.C., von Zastrow, M.,
in synaptic transmission and plasticity. Neuron 24, 649–658.
Lu ¨thi, A., Chittajallu, R., Duprat, F., Palmer, M.J., Benke, T.A., Kidd,
F.L., Henley, J.M., Isaac, J.T., and Collingridge, G.L. (1999). Hippo-
campal LTD expression involves a pool of AMPARs regulated by
the NSF-GluR2 interaction. Neuron 24, 389–399.
MacMillan, L.B., Bass, M.A., Cheng, N., Howard, E.F., Tamura, M.,
Strack, S., Wadzinski, B.E., and Colbran, R.J. (1999). Brain ac-
tin-associated protein phosphatase 1 holoenzymes containing spi-
nophilin, neurabin, and selected catalytic subunit isoforms. J. Biol.
Chem. 274, 35845–35854.
Malenka, R.C., and Nicoll, R.A. (1993). NMDA-receptor-dependent
synaptic plasticity: multiple forms and mechanisms. Trends Neu-
rosci. 16, 521–527.
Malenka, R.C., and Nicoll, R.A. (1999). Long-term potentiation—a
decade of progress? Science 285, 1870–1874.
Man, H.Y., Lin, J.W., Ju, W.H., Ahmadian, G., Liu, L., Becker, L.E.,
Sheng, M., and Wang, Y.T. (2000). Regulation of AMPA receptor-
mediated synaptic transmission by clathrin-dependent receptor in-
ternalization. Neuron 25, 649–662.
Matsuda, S., Mikawa, S., and 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. Neuro-
chem. 73, 1765–1768.
McAvoy, T., Allen, P.B., Obaishi, H., Nakanishi, H., Takai, Y.,
Greengard, P., Nairn, A.C., and Hemmings, H.C., Jr. (1999). Regula-
tion of neurabin I interaction with protein phosphatase 1 by phos-
phorylation. Biochemistry 38, 12943–12949.
Moorhead, G., Johnson, D., Morrice, N., and Cohen, P. (1998). The
major myosin phosphatase in skeletal muscle is a complex between
the beta-isoform of protein phosphatase 1 and the MYPT2 gene
product. FEBS Lett. 438, 141–144.
Mulkey, R.M., Herron, C.E., and Malenka, R.C. (1993). An essential
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.
O’Dell, T.J., and Kandel, E.R. (1994). Low-frequency stimulation
phatases. Learn. Mem. 1, 129–139.
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.
Oliver, C.J., and Shenolikar, S. (1998). Physiologic importance of
protein phosphatase inhibitors. Front. Biosci. 3, D961–D972.
Palmer, M.J., Irving, A.J., Seabrook, G.R., Jane, D.E., and Colling-
ridge, G.L. (1997). The group I mGlu receptor agonist DHPG induces
a novel form of LTD in the CA1 region of the hippocampus. Neuro-
pharmacology 36, 1517–1532.
Price, N.E., and Mumby, M.C. (1999). Brain protein serine/threonine
phosphatases. Curr. Opin. Neurobiol. 9, 336–342.
Neuron Download full-text
Sanes, J.R., and Lichtman, J.W. (1999). Can molecules explain long-
term potentiation? Nat. Neurosci. 2, 597–604.
Sculptoreanu, A., Scheuer, T., and Catterall, W.A. (1993). Voltage-
dependent potentiation of L-type Ca2?channels due to phosphory-
lation by cAMP-dependent protein kinase. Nature 364, 240–243.
Selig, D.K., Hjelmstad, G.O., Herron, C., Nicoll, R.A., and Malenka,
R.C. (1995). Independent mechanisms for long-term depression of
AMPA and NMDA responses. Neuron 15, 417–426.
tion and localization in hippocampal neurons by NMDA receptor
stimulation. Science 284, 162–166.
Shen, K., Teruel, M.N., Connor, J.H., Shenolikar, S., and Meyer, T.
(2000). Molecular memory by reversible translocation of calcium/
calmodulin-dependent protein kinase II. Nat. Neurosci. 3, 881–886.
Soderling, T.R., and Derkach, V.A. (2000). Postsynaptic protein
phosphorylation and LTP. Trends Neurosci. 23, 75–80.
Strack, S., and Colbran, R.J. (1998). Autophosphorylation-depen-
dent targeting of calcium/ calmodulin-dependent protein kinase II
by the NR2B subunit of the N-methyl- D-aspartate receptor. J. Biol.
Chem. 273, 20689–20692.
Terry-Lorenzo, R.T., Inoue, M., Connor, J.H., Haystead, T.A., Arm-
bruster, B.N., Gupta, R.P., Oliver, C.J., and Shenolikar, S. (2000).
Neurofilament-L is a protein phosphatase-1-binding protein associ-
ated with neuronal plasma membrane and post-synaptic density.
J. Biol. Chem. 275, 2439–2446.
Thiels, E., Norman, E.D., Barrionuevo, G., and Klann, E. (1998). Tran-
long-term depression in the adult hippocampus in vivo. Neurosci-
ence 86, 1023–1029.
Westphal, R.S., Tavalin, S.J., Lin, J.W., Alto, N.M., Fraser, I.D.,
Langeberg, L.K., Sheng, M., and Scott, J.D. (1999). Regulation of
NMDA receptors by an associated phosphatase-kinase signaling
complex. Science 285, 93–96.
Yan, Z., Hsieh-Wilson, L., Feng, J., Tomizawa, K., Allen, P.B., Fien-
berg, A.A., Nairn, A.C., and Greengard, P. (1999). Protein phospha-
tase 1 modulation of neostriatal AMPA channels: regulation by
DARPP-32 and spinophilin. Nat. Neurosci. 2, 13–17.