REVIEW
T.R. Soderling and V.A. Derkach – LTP: protein phosphorylation
Postsynaptic protein phosphorylation and LTP
Thomas R. Soderling and Victor A. Derkach
Prolonged changes in synaptic strength, such as those that occur in LTP and LTD, are thought to
contribute to learning and memory processes.These complex phenomena occur in diverse brain
structures and use multiple, temporally staged and spatially resolved mechanisms, such as changes
in neurotransmitter release,modulation of transmitter receptors,alterations in synaptic structure,
and regulation of gene expression and protein synthesis. In the CA1 region of the hippocampus,
the combined activation of SRC family tyrosine kinases, protein kinase A, protein kinase C and, in
particular, Ca21/calmodulin-dependent protein kinase II results in phosphorylation of glutamatereceptor-gated ion channels and the enhancement of subsequent postsynaptic current. Crosstalk
between these complex biochemical pathways can account for most characteristics of early-phase
LTP in this region.
Trends Neurosci. (2000) 23, 75–80
R
APID excitatory synaptic transmission in the mammalian forebrain is mediated primarily by the actions of glutamate on two types of ionotropic receptors:
AMPA and NMDA receptors (Fig. 1a). The physiological
function of the activated AMPA receptor is to mediate
postsynaptic depolarization by Na1 influx, whereas the
NMDA receptor is largely non-responsive, owing to a
voltage-dependent block by Mg21 (reviewed in Ref. 1).
Strong afferent stimulation can produce sufficient postsynaptic depolarization to overcome the Mg21 block, and
the activated NMDA receptor can then mediate both
Na1 and Ca21 influx. Both these glutamate receptors are
localized at the postsynaptic membrane in a scaffolding
organelle called the postsynaptic density (PSD). In addition to glutamate receptors, the PSD also contains many
signaling molecules that appear to modulate synaptic
transmission2.
The efficiency of transmission at these glutamatergic
synapses can either be weakened (LTD) by prolonged,
low-frequency stimulation or strengthened (LTP) by
brief, high-frequency stimulation of the afferent pathway1. Although LTP occurs in many brain structures, the
hippocampus, which is where Bliss and Lo/ mo first observed LTP (Ref. 3), has been investigated most rigorously because of its relatively simple anatomy and its
potential importance in the formation of new declarative memories. Synaptic potentiation can be divided
into several temporal stages that use different mechanisms: short-term potentiation (STP), which lasts only
15–30 min, early-phase LTP (E-LTP), which is stable
for up to 2–3 h, and late-phase LTP (L-LTP), which can
last for 6–8 h in hippocampal slices. STP and E-LTP do
not require gene expression or new protein synthesis,
whereas L-LTP does4.
Stimulation paradigms that induce LTP in the CA1
region produce elevated Ca21 levels in the dendritic spine
(Fig. 1b), largely through NMDA-receptor activation.
This process is essential for the initiation of LTP, as it can
be prevented by selective blockers of NMDA receptors or
by intracellular Ca21 chelators. Although NMDA receptors are essential for the initiation of LTP, expression of
LTP is mediated primarily by AMPA receptors. There has
been considerable controversy as to whether LTP in various regions of hippocampus is mediated by presynap0166-2236/00/$ – see front matter © 2000 Elsevier Science Ltd. All rights reserved.
tic (for example, increased neurotransmitter release) or
postsynaptic (for example, increased responsiveness of
AMPA receptors) mechanisms, and although this debate
continues5, numerous observations in recent years
implicate a predominantly postsynaptic site for E-LTP
expression at Schaffer-collateral–commissural synapses
in region CA1. For example, the glutamate-transporter
currents in hippocampal astrocytes that are located
adjacent to synapses should be sensitive to the concentration of glutamate in the synapse. Indeed, these currents can be used to detect increased glutamate levels in
the synapse that are elicited by well-characterized presynaptic mechanisms such as paired-pulse facilitation6.
However, induction of LTP does not enhance these
transporter currents, which is consistent with a postsynaptic locus of LTP expression. Postsynaptic mechanisms of E-LTP at the CA1 hippocampal synapse in rat
and mouse will, therefore, be the focus of this article.
Tyrosine kinases and E-LTP
NMDA receptors, which are heteromeric channels that
consist of NR1 and NR2A–D subunits, form complexes
in the PSD with other proteins such as PSD95, Ca21–
calmodulin, SAP102, a-actinin and the tyrosine-kinase
family members2. The NR2A and NR2B subunits are subject to tyrosine-residue phosphorylation, and infusion
of tyrosine kinases potentiates the current through
NR1–NR2A or NR1–NR2B recombinant channels by
phosphorylation of intracellular, C-terminal tyrosine
residues, thereby relieving a basal zinc inhibition of
NMDA receptors7. Induction of LTP produces a rapid
activation of SRC within 1–5 min (Ref. 8; Table 1) and
increased tyrosine-residue phosphorylation of NR2B
(Refs 15,16); prior application of SRC-specific inhibitors
prevents induction of LTP (Ref. 8). Infusion of SRC, like
LTP itself, results in an enhancement of AMPA-receptormediated current that is dependent on the influx of
Ca21 through the NMDA-receptor channel (Fig. 1b).
This potentiation of AMPA-receptor-mediated current
is probably indirect and caused by the SRC-mediated
enhancement of NMDA receptor function. SRC has been
localized to the PSD and shown to co-precipitate with
the NMDA receptor17. Infusion of the tyrosine kinase
FYN also potentiates NMDA-receptor-mediated current,
PII: S0166-2236(99)01490-3
TINS Vol. 23, No. 2, 2000
Thomas R.
Soderling and
Victor A. Derkach
are at the Vollum
Institute, Oregon
Health Sciences
University,
Portland,
OR 97201, USA.
75
REVIEW
T.R. Soderling and V.A. Derkach – LTP: protein phosphorylation
(b)
(a)
Strong afferent stimulation 1
(LTP induction)
Weak afferent stimulation
(basal transmission)
Na+
Na+
AMPA
receptor
Mg2+ NMDA
receptor
PSD
~10pS
PSD
Tyrosine
kinase
Autophosphorylation
EPSC
CaMKII*
CaMKII
Ca2+
2
Ca2+—CaM
(basal)
Removal of
Mg2+ block
~10pS
Strong 3
depolarization
4
* CaMKII*
5
7
Ca2+–CaM CaMKII*
5
SRC
6
6
Autophosphorylation
ATP
Adenylate
cyclase
cAMP
PPI
MAPK
Dendritic
spine
Dendritic
spine
(c)
Weak stimulation
(E-LTP expression)
Na+
PSD
*
10pS
Mg2+
CaMKII*
25pS
PPI
I1*
Enhanced
EPSC
CaMKII
PKA
I1
cAMP
Dendritic
spine
trends in Neurosciences
and recently it has been shown that interaction of FYN
with NR2A is mediated by the scaffold protein PSD95
(Ref. 18). These results have been used to formulate the
hypothesis19 that the induction of LTP results in rapid
activation of SRC-family kinases, which phosphorylate
and enhance NMDA-receptor-mediated current (Fig. 1b).
This, in turn, increases Ca21 influx into the dendritic
spine, which triggers the biochemical pathways leading
to potentiation of AMPA-receptor-mediated current, as
discussed below.
Although the above data are rather convincing, there
are details that need to be resolved. Is the phosphorylation of the NMDA receptor by tyrosine kinases sufficiently rapid to account for the enhancement of current
through the NMDA-receptor channel? What is the
mechanism of tyrosine-kinase activation by LTP? A
recent study suggests it might be mediated by protein
kinase C (Ref. 20). Another report21 indicates that the
SRC-family member LYN can physically associate with
AMPA receptors and be activated independently of
76
TINS Vol. 23, No. 2, 2000
Fig. 1. Postsynaptic protein phosphorylation mechanisms during LTP
in the CA1 region of hippocampus. (a) Basal synaptic transmission via
glutamate (triangles) is mediated largely by low-conductance state
(~10pS) AMPA receptors that give rise to an EPSC. NMDA receptors are
inactive because of voltage-dependent block of their channels by Mg21.
Little of the Ca21/calmodulin-dependent protein kinase II (CaMKII) is in
the activated, phosphorylated state (*), owing to low basal concentrations of Ca21–calmodulin (Ca21–CaM) and high activity of protein
phosphatase 1 (PP1). (b) LTP induction by tetanic afferent stimulation (1)
elicits enhanced glutamate exocytosis (2) and strong postsynaptic depolarization via AMPA receptors (3) to remove the Mg21 block of the NMDA
receptor (4). AMPA-receptor stimulation also activates associated SRCfamily tyrosine kinases that phosphorylate (*) and further enhance conductance of the NMDA-receptor channel, which is permeable to Ca21 (5).
The elevated levels of the Ca21–CaM complex stimulate type-1 adenylate
cyclase and also stimulates autophosphorylation (*) of CaMKII (6). This
constitutively active CaMKII translocates to the postsynaptic density
(PSD), in part through interaction with the NMDA receptor (7). (c) Earlyphase LTP (E-LTP) expression is due, in part, to CaMKII-mediated phosphorylation of the AMPA receptor GluR1 subunit (*), which converts it
largely to higher conductance states (~25 pS). The CaMKII is maintained in its autophosphorylated, constitutively active form, owing to
protein kinase A-mediated phosphorylation of inhibitor 1 (I1), which
blocks the ability of PP1 to dephosphorylate CaMKII and perhaps GluR1.
The predominant species and pathways are indicated in bold.
Na1 or Ca21 influx through the channel, and results in
activation of mitogen-activated protein kinase (MAPK)
(Table 1).
Phosphorylation and modulation of AMPA receptors:
Ca21-dependent kinases in E-LTP
Given that the induction of LTP in CA1 is blocked by
general inhibitors of Ser/Thr protein kinases or by postsynaptic chelation of Ca21, a potential role for Ca21dependent protein kinases, such as Ca21/calmodulindependent protein kinase II (CaMKII) or PKC, or both,
is reasonable. Infusion of peptide inhibitors of either
of these two protein kinases or of calmidazolium, a
calmodulin (CaM) antagonist, into pyramidal neurons
blocks induction of LTP, indicating a postsynaptic locus
for their actions. Furthermore, infusion of Ca21 and
CaM (Ref. 22) or activated CaMKII (Ref. 23) potentiates
synaptic current and occludes subsequent induction
of LTP. These results indicate that CaMKII potentiates
synaptic current using the same or overlapping
REVIEW
T.R. Soderling and V.A. Derkach – LTP: protein phosphorylation
mechanism(s) as LTP. Although it is likely that multiple
protein kinases are involved in the induction or expression of E-LTP, or both, current evidence implicates
a key role for CaMKII (summarized in Ref. 24).
CaMKII, an oligomeric protein that consists of 10–12
subunits, is a major constituent of the PSD, and it has
unusual biochemical properties that make it a unique
transducer of Ca21 signals (reviewed in Ref. 25). In its
basal state, CaMKII is inactive, owing to the presence
of an autoinhibitory domain that exerts intrasubunit,
steric block of substrate binding. Binding of Ca21–CaM
adjacent to the autoinhibitory domain alters its conformation and disrupts its inhibitory interaction, thereby
causing activation of the kinase. The activated kinase first
undergoes a rapid intersubunit autophosphorylation
at Thr286, which is within the autoinhibitory domain,
prior to phosphorylation of exogenous substrates. This
autophosphorylation of Thr286 has three crucial regulatory functions: (1) it promotes association of CaMKII
with the PSD, in part through an interaction with the
NMDA receptor26,27 (Fig. 1b), (2) it decreases the dissociation rate of Ca21–CaM by three orders of magnitude25,
and (3) even after Ca21–CaM has dissociated, the autophosphorylated kinase still retains substantial kinase
activity25. Thus, rapid autophosphorylation generates a
constitutively active CaMKII that phosphorylates exogenous substrates more slowly, as well as catalyzing additional autophosphorylation on other sites. This means
that a transient elevation of Ca21 concentration in a
dendritic spine can be transduced into a prolonged
kinase activity that persists in the absence of elevated
Ca21 levels until the appropriate protein phosphatase
dephosphorylates Thr286. Protein phosphatase 1 (PP1),
which is also found at high levels in the PSD, might primarily be responsible for dephosphorylation of Thr286
and inactivation of PSD-associated CaMKII (Ref. 28).
Studies using cultured hippocampal neurons have
shown that stimulation of the NMDA receptor with
resultant Ca21 influx leads to autophosphorylation of
CaMKII at Thr286 and the generation of its Ca21-independent activity. More importantly, induction of LTP
in hippocampal slices results in activation of CaMKII
within 1 min, and this constitutive activity is stable for
at least 1 h (Ref. 9, Table 1). Furthermore, transgenic mice
in which the crucial autophosphorylation site Thr286 in
CaMKII is mutated to Ala have normal basal synaptic
transmission but do not exhibit LTP (Ref. 29).
Numerous protein substrates can be phosphorylated
by CaMKII, so what might be its target(s) after its activation during LTP? A series of studies have shown that
one particularly relevant substrate is the AMPA receptor, which becomes potentiated during LTP. In the CA1
region of the hippocampus, the heteromeric AMPA receptor consists primarily of GluR1 and GluR2 subunits.
CAMKII can phosphorylate the native AMPA receptor
in isolated PSDs (Ref. 30) and the GluR1 AMPA-receptor
subunit in HEK293 cells31. After NMDA-receptor stimulation and subsequent Ca21 influx in cultured hippocampal neurons, CAMKII phosphorylates the native
AMPA receptor32, and, most importantly, it can phosphorylate this receptor in the CA1 region after the
induction of LTP in this area10. CaMKII phosphorylates
Ser831 in GluR1 (Refs 31,33); there is no analogous
site in GluR2, the other major AMPA-receptor subunit
in hippocampal region CA1. This implies a crucial role
for the GluR1 subunit for CaMKII-mediated LTP, and this
hypothesis has been confirmed using transgenic mice.
TABLE 1. Activation of protein kinases during LTP
Kinase
LTP stimulusa
Kinase activationb
Refs
CaMKII
PKA
PKC
PKM
p42 MAPK
SRC
Theta burst
Multiple tetanic
Multiple tetanic
Single tetanic
Multiple tetanic
Multiple tetanic
5 min → 1 h → ?
2 min → 10 min
1 min → 45 min
? → 30 min → ?
2 min → ?
1min → 5 min → ?
9,10
11
12
13
14
8
a
Activation of the indicated protein kinases was determined subsequent to
LTP inducted using the indicated stimulus.
b
The times after LTP where kinase activation was observed are noted.
Abbreviations: CaMKII, Ca21/calmodulin-dependent protein kinase II;
MAPK, mitogen-activated protein kinase; PK, protein kinase; ?, earlier or
later times were not analyzed.
LTP is slightly enhanced in GluR2-subunit knockout
mice34, perhaps owing to their increased permeability to
Ca21. However, the adult GluR1-subunit knockout mouse
shows normal basal synaptic transmission, but LTP is
absent in region CA1 (Ref. 35). This requirement for
the GluR1 subunit not only substantiates its essential
role in CA1-region LTP and is consistent with CaMKII
involvement, but it also verifies a postsynaptic locus of
LTP expression. Interestingly, although CA1-region LTP
was prevented, spatial learning in the water-maze paradigm was not impaired35, suggesting a possible dichotomy between LTP in CA1 and the acquisition of spatial
learning. The crucial test for the role of CaMKII-mediated
phosphorylation of GluR1 in CA1-region LTP will be
whether LTP is absent in transgenic mice where the
Ser831Ala mutant of GluR1 replaces wild-type GluR1.
What is the functional consequence of AMPA receptor
phosphorylation? Phosphorylation of the AMPA receptor by CaMKII, either the native receptor in cultured
hippocampal neurons30 or CA1 pyramidal neurons23, or
the expressed GluR1 subunit in HEK293 cells31, results
in potentiation of whole-cell AMPA-receptor-mediated
current. Consistent with the slow rate of AMPA-receptor
phosphorylation, potentiation of current by CaMKII
requires 15–30 min for maximal effect. The Ser831Ala
mutant of GluR1 and wild-type GluR2, which does not
contain an equivalent site, are not phosphorylated or
potentiated by CaMKII (Ref. 31). A recent detailed
analysis36 has shown that expressed GluR1 can adopt
multiple conductance states of between 9 and 28 pS.
The non-phosphorylated receptor exists predominantly
in the lower conductance states, whereas phosphorylation by CaMKII or mutation of Ser831 to Asp, which,
similar to phosphorylation, introduces negative
charge, stabilizes the higher conductances36. Although
many ion channels are regulated by phosphorylation,
the enhancement of single-channel conductance by
phosphorylation is very unusual.
Is there evidence for phosphorylation and potentiation of AMPA receptors during LTP? A number of years
ago it was reported that subsequent to induction of LTP
there is an enhancement of AMPA-receptor responsiveness that develops over approximately 30 min (Ref. 37),
which is similar to the rate of AMPA-receptor phosphorylation and potentiation by CaMKII. As stated above,
induction of LTP produces a very rapid and sustained
generation of Ca21-independent CaMKII activity9, which
is consistent with its autophosphorylation at Thr286.
This has recently been confirmed using a phosphospecific antibody for Thr286 (Ref. 10). More importantly,
TINS Vol. 23, No. 2, 2000
77
REVIEW
T.R. Soderling and V.A. Derkach – LTP: protein phosphorylation
this latter study showed that LTP could increase phosphorylation of AMPA receptors in region CA1. This
phosphorylation of the AMPA receptors is probably
catalyzed by the Ca21-independent form of CaMKII
because: (1) the timecourse of AMPA-receptor phosphorylation correlates with the slow Ca21-independent
additional autophosphorylation (that is, on sites other
than Thr286) of CaMKII itself; and (2) the phosphorylation of the AMPA receptor, as well as activation of
CaMKII and induction of LTP itself, were blocked by a
general CaMK inhibitor, KN62; and (3) the 2D tryptic
32
P-peptide map of the AMPA receptor after LTP induction was similar to that obtained by CaMKII-mediated
phosphorylation of GluR1 (Ref. 31).
These results are consistent with the following
hypothesis. Within 1 min of LTP induction there is
activation of CaMKII, owing to autophosphorylation
of Thr286, which is stable for at least 1 h and allows
translocation of CaMKII to the PSD (Fig. 1b). This Ca21independent activity of CaMKII slowly phosphorylates
GluR1 of the AMPA receptor, resulting in potentiation
of the AMPA-receptor-mediated current because of an
increase in single-channel conductance (Fig. 1c). Consistent with this interpretation is a recent report that
induction of LTP does indeed increase AMPA-receptor
channel unitary conductance in about 60% of the cells
that are potentiated38. Are these regulatory mechanisms
conserved at synapses in species other than rat and
mouse? A recent study indicates that activity-dependent
enhancement of glutamate-mediated transmission at
the goldfish Mauthner-cell synapse also requires activation of postsynaptic CaMKII (Ref. 39). It will be interesting to see whether goldfish AMPA receptors are
phosphorylated and potentiated by CaMKII.
Two other protein kinases that can phosphorylate
GluR1 could also be involved in its regulation. PKC can
phosphorylate Ser831 in GluR1 (Ref. 33), and induction
of LTP generates a prolonged activation of PKC (Refs.
12,13; Table 1). Thus, a role for PKC-mediated phosphorylation of GluR1 has to be considered. Second, protein kinase A (PKA) can phosphorylate the adjacent
Ser845 in GluR1, and this apparently also potentiates
AMPA-receptor-mediated current40. However, Ser845 appears to be phosphorylated under basal conditions, and
dephosphorylation of Ser845 occurs in LTD (Ref. 41).
This could explain an earlier observation that disrupting
the interaction between PKA and its anchoring protein
(A kinase anchoring protein or AKAP) causes rundown
of basal AMPA-receptor-mediated current in hippocampal neurons42. Phosphorylation of AMPA receptors
by PKA under basal conditions might be accomplished
by localization of PKA to the PSD, mediated by the
interaction of an AKAP with NMDA receptors43.
Regulation of protein phosphatases during E-LTP
Neurons not only have high levels of many protein
kinases, but they are also rich sources for many protein
phosphatases, including protein phosphatases 1, 2A
(PP1, PP2A) and 2B (PP2B or calcineurin). PP1 and PP2A
are very effective at dephosphorylating Thr286 in
CaMKII, thereby reversing its constitutive activity to
basal levels. They are also able to dephosphorylate
Ser831 of GluR1 in vitro (A. Barria and T.R. Soderling,
unpublished observations). There is significant PP1 and
PP2A in the PSD, and evidence suggests that PP1 is
predominantly responsible for dephosphorylation of
PSD-associated CaMKII (Ref. 28).
78
TINS Vol. 23, No. 2, 2000
How might the activity of PP1 be suppressed during
LTP in order to prevent rapid dephosphorylation of
CaMKII, and perhaps GluR1, once the concentration of
Ca21 in the dendritic spine returns to basal values? The
catalytic subunit of PP1 can interact with a number of
proteins, some of which serve to localize PP1 to subcellular organelles and others of which act as inhibitory
subunits. One of these inhibitor proteins is known as
inhibitor 1 (I1) (reviewed in Ref. 44). Only when I1 has
been phosphorylated by PKA does it inhibit PP1, and
dephosphorylation of I1 is catalyzed by PP2B, which is
also anchored to the PSD through an AKAP (Ref. 43).
Regulation of PP1 activity through PKA-mediated phosphorylation of I1 during LTP has recently been studied
in detail45. Induction of E-LTP causes a transient activation of PKA (Table 1), owing to stimulation of type-I
adenylate cyclase11. The requirement for elevated cAMP
levels or PKA activation in E-LTP can be substituted by
postsynaptic injection of a stably phosphorylated (that
is, thiophosphorylated) I1 (Ref. 45). Furthermore, it has
been shown that LTP induction results in I1 phosphorylation and inhibition of PP1 activity (Fig. 1c). As
inhibition of CaMKII activation by KN62 still prevented LTP induction when phosphorylated I1 was
injected, it was concluded that the function of PP1
inhibition was primarily to prolong the activation of
CaMKII. This would also maintain the phosphorylation
of GluR1.
Transient synaptic potentiation can be induced in
hippocampal slices by treatment with elevated extracellular Ca21 with K1. This transient (40 min) potentiation is inhibited by postsynaptic infusion of a CaMKII
inhibitor peptide, and it is converted into a prolonged
(.60 min), NMDA-receptor-dependent potentiation by
pretreatment with either activators of PKA (forskolin)
or inhibitors of protein phosphatases 1 and 2A (calyculin A) (Ref. 46). When the level of CaMKII Thr286
autophosphorylation was determined at 1 h, treatment
with elevated Ca21 or K1 alone (transient potentiation)
was without effect, whereas pretreatment with either
forskolin or calyculin (prolonged potentiation) produced
a significant increase that was prevented by an NMDAreceptor antagonist. These results are consistent with
the hypothesis that stable synaptic potentiation is associated with prolonged CaMKII activation, which requires
protein phosphatase inhibition, probably through
phosphorylation of I1 by PKA.
The ‘silent synapse’ hypothesis in E-LTP
The fact that only 60% of potentiated CA1 neurons
show an increase in AMPA-receptor channel unitary
conductance38 suggests that mechanisms in addition to
phosphorylation of AMPA receptors by CaMKII are also
operative. There is extensive evidence for the ‘silent
synapse’ hypothesis47,48, which states that prior to induction of LTP some synapses do not have functional
AMPA receptors, whereas after LTP induction they exhibit AMPA-receptor-mediated currents. This might result from recruitment of AMPA receptors to the synapse
from some intracellular pool, perhaps analogous to the
rapid insertion of glucose transporters into the plasma
membrane after insulin treatment in adipocytes, a process that appears to require protein phosphorylation
(reviewed in Ref. 49). Postsynaptic infusion of inhibitors
of the membrane-fusion pathway, such as agents that
disrupt the functions of N-ethylmaleimide-sensitive
factor (NSF) and soluble NSF attachment protein (SNAP),
REVIEW
T.R. Soderling and V.A. Derkach – LTP: protein phosphorylation
strongly suppress LTP with little or no effect on baseline
synaptic transmission or NMDA receptors50. Infusion
of SNAP enhances synaptic transmission, and this
enhancement is blocked by prior induction of LTP.
Furthermore, a complex formed between NSF, SNAP
and GluR2 has been identified51, and disruption of this
interaction causes partial rundown of AMPA-receptormediated synaptic current52,53. However, interaction between NSF and AMPA receptors is specific for the GluR2
subunit51–53, but basal synaptic transmission and LTP are
not altered in the GluR2-subunit knockout mouse34.
Cultured hippocampal neurons exhibit Ca21-elicited
exocytosis of dendritic organelles, and this process might
require CaMKII activity54. Moreover, induction of LTP
induces redistribution of epitope-tagged, transiently
expressed GluR1 within 30 min in hippocampal slices
from primarily intracellular sites in the dendritic shaft
of dendritic spine apical dendrites55. While these results
are supportive of the ‘silent synapse’ hypothesis, the
challenge now is to demonstrate that these translocated AMPA receptors are functional and located at
potentiated synapses.
Concluding remarks
Studies over the past five years have begun to map out
some of the signal-transduction pathways that are used
in modulating synaptic strength at the postsynaptic
locus in CA1 hippocampal neurons. It is clear that
phenomena such as LTP, which might contribute to
learning and memory, use multiple mechanisms, many
of which involve protein phosphorylation. These studies have emphasized roles in E-LTP for the transient
potentiation of NMDA receptors, through tyrosineresidue phosphorylation, in order to enhance Ca21
influx; prolonged activation of CaMKII through autophosphorylation; inhibition of PP1 through PKA phosphorylation of I1; and prolonged potentiation of AMPA
receptors by CaMKII-mediated phosphorylation (Fig. 1).
Recent computer modeling indicates that crosstalk and
feedback loops between multiple signaling molecules
(for example, CaMKII, PKC, PKA, MAPK, etc.) must occur
to produce prolonged activation of CaMKII (Ref. 56).
Many other interesting facets remain to be resolved,
particularly relating to mechanisms and functional consequences of activity-dependent AMPA-receptor recruitment to synapses. Another area of active investigation
is the consolidation of E-LTP into L-LTP, which appears
to require gene transcription and new protein synthesis4.
Genetic studies in Aplysia, Drosophila and mice indicate
a role for cAMP-response-element-binding protein
(CREB)-mediated transcription (reviewed in Ref. 57),
but it is unclear which protein kinase(s) [for example,
PKA, CaMKIV, MAPK or ribosomal protein kinase (RSK)]
is responsible for phosphorylation of CREB during LTP.
What is the signal that triggers new gene expression
and how does it traverse from the potentiated synapse
to the nucleus? What is the role of localized protein
synthesis in dendrites58 and how is this regulated during
LTP (Ref. 59)? Furthermore, how are newly synthesized
proteins selectively targeted to potentiated synapses60?
These are areas of very active investigation and it is likely
that protein phosphorylation is also crucial to these
aspects of synaptic plasticity.
Selected references
1 Malenka, R.C. (1994) Synaptic plasticity in the hippocampus: LTP
and LTD. Cell 78, 535–538
2 Kennedy, M.B. (1998) Signal transduction molecules at the
glutamatergic postsynaptic membrane. Brain Res. Rev. 26, 243–257
3 Bliss, T.V.P. and Lo/ mo, T. (1973) Long-lasting potentiation of
synaptic transmission in the dentate area of the anaesthetized
rabbit following stimulation of the perforant path. J. Phyiol. 232,
331–356
4 Bailey, C.H. et al. (1996) Toward a molecular definition of long-term
memory storage. Proc. Natl. Acad. Sci. U. S. A. 93, 13445–13452
5 Nayak, A.S. et al. (1996) Ca21/calmodulin-dependent protein
kinase II phosphorylation of the presynaptic protein synapsin I is
persistently increased during long-term potentiation. Proc. Natl.
Acad. Sci .U. S. A. 93, 15451–15456
6 Diamond, J.S. et al. (1998) Glutamate release monitored with
astrocyte transporter currents during LTP. Neuron 21, 425–433
7 Zheng, F. et al. (1998) Tyrosine kinase potentiates NMDA receptor
currents by reducing tonic zinc inhibition. Nat. Neurosci. 1, 185–191
8 Lu, Y.M. et al. (1998) Src activation in the induction of long-term
potentiation in CA1 hippocampal neurons. Science 279, 1363–1368
9 Fukunaga, K. et al. (1993) Long-term potentiation is associated
with an increased activity of Ca21/calmodulin-dependent protein
kinase II. J. Biol. Chem. 268, 7863–7867
10 Barria, A. et al. (1997) Regulatory phosphorylation of AMPA-type
glutamate receptors by CaM-KII during long-term potentiation.
Science 276, 2042–2045
11 Roberson, E.D. and Sweatt, J.D. (1996) Transient activation of cyclic
AMP-dependent protein kinase during hippocampal long-term
potentiation. J. Biol. Chem. 271, 30436–30441
12 Klann, E. et al. (1993) Mechanism of protein kinase C activation
during the induction and maintenance of long-term potentiation
probed using a selective peptide substrate. Proc. Natl. Acad. Sci.
U. S. A. 90, 8337–8341
13 Sactor, T.C. et al. (1993) Persistent activation of the zeta isoform
of protein kinase C in the maintenance of long-term potentiation.
Proc. Natl. Acad. Sci. U. S. A. 90, 8342–8346
14 English, J.D. and Sweatt, J.D. (1996) Activation of p42 mitogenactivated protein kinase in hippocampal long-term potentiation.
J. Biol. Chem. 271, 24329–24332
15 Rostas, J.A.P. et al. (1996) Enhanced tyrosine phosphorylation of
the 2B subunit of the N-methyl-D-aspartate receptor in long-term
potentiation. Proc. Natl. Acad. Sci. U. S. A. 93, 10452–10456
16 Rosenblum, K. et al. (1996) Long-term potentiation increases
tyrosine phosphorylation of the N-methyl-D-aspartate receptor
subunit 2B in rat dentate gyrus in vivo. Proc. Natl. Acad. Sci. U. S. A.
93, 10457–10460
17 Yu, X.M. et al. (1997) NMDA channel regulation by channelassociated protein tyrosine kinase Src. Science 275, 674–678
18 Tezuka, T. et al. (1999) PSD-95 promotes Fyn-mediated tyrosine
phosphorylation of the N-methyl-D-aspartate receptor subunit
NR2A. Proc. Natl. Acad. Sci. U. S. A. 96, 435–440
19 Salter, M.W. (1998) Src, N-methyl-D-aspartate (NMDA) receptors, and
synaptic plasticity. Biochem. Pharm. 56, 789–798
20 Lu, W.Y. et al. (1999) G-protein-coupled receptors act via protein
kinase C and Src to regulate NMDA receptors. Nat. Neurosci. 4,
331–338
21 Hayashi, T. et al. (1999) The AMPA receptor interacts with and signals
through the protein tyrosine kinase Lyn. Nature 397, 72–76
22 Wang, J.H. and Kelly, P.T. (1995) Postsynaptic injection of Ca21/CaM
induces synaptic potentiation requiring CaMKII and PKC activity.
Neuron 15, 443–452
23 Lledo, P.M. et al. (1995) Calcium/calmodulin-dependent kinase II
and long-term potentiation enhance synaptic transmission by the
same mechanism. Proc. Natl. Acad. Sci. U. S. A. 92, 11175–11179
24 Malenka, R.C. and Nicoll, R.A. (1999) Long-term potentiation – a
decade of progress? Science 285, 1870–1874
25 Braun, A.P. and Schulman, H. (1995) The multifunctional calcium/
calmodulin-dependent protein kinase: from form to function.
Annu. Rev. Physiol. 57, 417–445
26 Strack, S. and Colbran, R.J. (1998) Autophosphorylation-dependent
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
27 Leonard, A.S. et al. (1999) Calcium/calmodulin-dependent protein
kinase II is associated with the N-methyl-D-aspartate receptor. Proc.
Natl. Acad. Sci. U. S. A. 96, 3239–3244
28 Strack, S. et al. (1997) Differential inactivation of postsynaptic
density-associated and soluble Ca21/calmodulin-dependent protein
kinase II by protein phosphatases 1 and 2A. J. Neurochem. 68,
2119–2128
29 Giese, K.P. et al. (1998) Autophosphorylation at Thr286 of the alpha
calcium-calmodulin kinase II in LTP and learning. Science 279,
870–873
30 McGlade-McCulloh, E. et al. (1993) Phosphorylation and regulation
of glutamate receptors by calcium/calmodulin-dependent protein
kinase II. Nature, 362, 640–642
31 Barria, A. et al. (1997) Identification of the Ca21/calmodulindependent protein kinase II regulatory phosphorylation site in
the alpha-amino-3-hydroxyl-5-methyl-4-isoxazole-propionate-type
glutamate receptor. J. Biol. Chem. 272, 32727–32730
TINS Vol. 23, No. 2, 2000
79
REVIEW
T.R. Soderling and V.A. Derkach – LTP: protein phosphorylation
32 Tan, S.E. et al. (1994) Phosphorylation of AMPA-type glutamate
receptors by calcium/calmodulin-dependent protein kinase II and
protein kinase C in cultured hippocampal neurons. J. Neurosci. 14,
1123–1129
33 Mammen, A.L. et al. (1997) Phosphorylation of the alpha-amino3-hydroxy-5-methylisoxazole4-propionic acid receptor GluR1
subunit by calcium/calmodulin-dependent kinase II. J. Biol. Chem.
272, 32528–32533
34 Jia, Z. et al. (1996) Enhanced LTP in mice deficient in the AMPA
receptor GluR2. Neuron 17, 945–956
35 Zamanillo, D. et al. (1999) Importance of AMPA receptors for
hippocampal synaptic plasticity but not for spatial learning. Science
284, 1805–1811
36 Derkach, V. et al. (1999) Ca21/calmodulin-kinase II enhances channel
conductance of alpha-amino-3-hydroxy-5-methyl-4-isoxazolepropionate type glutamate receptors. Proc. Natl. Acad. Sci. U. S. A.
96, 3269–3274
37 Davies, S.N. et al. (1989) Temporally distinct pre- and post-synaptic
mechanisms maintain long-term potentiation. Nature 338, 500–503
38 Benke, T.A. et al. (1998) Modulation of AMPA receptor unitary
conductance by synaptic activity. Nature 393, 793–797
39 Pereda, A.E. et al. (1998) Ca21/calmodulin-dependent kinase II
mediates simultaneous enhancement of gap-junctional conductance
and glutamatergic transmission. Proc. Natl. Acad. Sci. U. S. A. 95,
13272–13277
40 Roche, K.W. et al. (1996) Characterization of multiple
phosphorylation sites on the AMPA receptor GluR1 subunit.
Neuron 16, 1179–1188
41 Kameyama, K. et al. (1998) Involvement of a postsynaptic protein
kinase A substrate in the expression of homosynaptic long-term
depression. Neuron 21, 1163–1175
42 Rosenmund, C. et al. (1994) Anchoring of protein kinase A is
required for modulation of AMPA/kainate receptors on hippocampal
neurons. Nature 368, 853–856
43 Westphal, R.S. et al. (1999) Regulation of NMDA receptors by an
associated phosphatase-kinase signaling complex. Science 285, 93–96
44 Cohen, P. (1989) The structure and regulation of protein
phosphatases. Annu. Rev. Biochem. 58, 453–508
45 Blitzer, R.D. et al. (1998) Gating of CaMKII by cAMP-regulated
protein phosphatase activity during LTP. Science 280, 1940–1942
46 Makhinson, M. et al. (1999) Adenylyl cyclase activation modulates
activity-dependent changes in synaptic strength and Ca21/
calmodulin-dependent kinase II autophosphorylation. J. Neurosci.
19, 2500–2510
47 Liao, D. et al. (1995) Activation of postsynaptically silent synapses
during pairing-induced LTP in CA1 region of hippocampal slice.
Nature 375, 400–404
48 Isaac, J.T. et al. (1995) Evidence for silent synapses: implications for
the expression of LTP. Neuron 15, 427–434
49 Pessin, J.E. et al. (1999) Molecular basis of insulin-stimulated GLUT4
vesicle trafficking. J. Biol. Chem. 274, 2593–2596
50 Lledo, P.M. et al. (1998) Postsynaptic membrane fusion and longterm potentiation. Science 279, 399–403
51 Osten, P. et al. (1998) The AMPA receptor GluR2 C terminus can
mediate a reversible, ATP-dependent interaction with NSF and
alpha- and beta-SNAPs. Neuron 21, 99–110
52 Nishimune, A. et al. (1998) NSF binding to GluR2 regulates synaptic
transmission. Neuron 21, 87–97
53 Song, I. et al. (1998) Interaction of the N-ethylmaleimide-sensitive
factor with AMPA receptors. Neuron 21, 393–400
54 Maletic-Savatic, M. et al. (1998) Calcium-evoked dendritic exocytosis
in cultured hippocampal neurons. Part II: mediation by calcium/
calmodulin-dependent protein kinase II. J. Neurosci. 18, 6814–68211
55 Shi, S.H. et al. (1999) Rapid spine delivery and redistribution of
AMPA receptors after synaptic NMDA receptor activation. Science
284, 1811–1816
56 Bhalla, U.S. and Iyengar, R. (1999) Emergent properties of networks
of biological signaling pathways. Science 283, 381–387
57 Frank, D.A. and Greenberg, M.E. (1994) CREB: a mediator of longterm memory from mollusks to mammals. Cell 79, 5–8
58 Stewart, O. (1997) mRNA localization in neurons: a multipurpose
mechanism? Neuron 18, 9–12
59 Wu, L. et al. (1998) CPEB-mediated cytoplasmic polyadenylation and
the regulation of experience-dependent translation of alpha-CaMKII
mRNA at synapses. Neuron 21, 1129–1139
60 Frey, U. and Morris, R.G.M. (1998) Synaptic tagging: implications
for late maintenance of hippocampal long-term potentiation.
Trends Neurosci. 21, 181–188
Complex interactions between mGluRs,
intracellular Ca21 stores and ion channels
in neurons
Laurent Fagni, Pascale Chavis, Fabrice Ango and Joel Bockaert
Metabotropic glutamate receptors (mGluRs) can increase intracellular Ca21 concentration via
Ins(1,4,5)P3- and ryanodine-sensitive Ca21 stores in neurons. Both types of store are coupled
functionally to Ca21-permeable channels found in the plasma membrane.The mGluR-mediated
increase in intracellular Ca21 concentration can activate Ca21-sensitive K1 channels and Ca21dependent nonselective cationic channels. These mGluR-mediated effects often result from
mobilization of Ca21 from ryanodine-sensitive, rather than Ins(1,4,5)P3-sensitive, Ca21 stores,
suggesting that close functional interactions exist between mGluRs, intracellular Ca21 stores and
Ca21-sensitive ion channels in the membrane.
Trends Neurosci. (2000) 23, 80–88
Laurent Fagni,
Pascale Chavis,
Fabrice Ango and
Joel Bockaert are
at the CNRS-UPR
9023, 34094
Montpellier
cedex 05, France.
80
N NEURONS, intracellular Ca21 concentration can be
transiently increased by two mechanisms: Ca21 influx
through Ca21-permeable channels in the plasma membrane or the mobilization of Ca21 from intracellular Ca21
stores. Indeed, most neurotransmitters can trigger both
mechanisms. Glutamate activates cationic channels,
the so-called ionotropic glutamate-receptor channels,
that induce cell depolarization and, hence, Ca21 influx
through voltage-sensitive Ca21 channels. Some of these
I
TINS Vol. 23, No. 2, 2000
channels are also Ca21 permeable1. Glutamate can also
activate G-protein- and phospholipase-C-coupled receptors, the so-called group-I metabotropic glutamate receptors (mGluRs; Refs 2,3). These receptors stimulate
Ins(1,4,5)P3 synthesis and the mobilization of Ca21 from
Ins(1,4,5)P3-sensitive intracellular Ca21 stores, which
were first discovered in injected Xenopus oocytes4 and
are now well characterized in mammalian neurons5–8.
The group-I mGluR-mediated Ca21 responses have even
0166-2236/00/$ – see front matter © 2000 Elsevier Science Ltd. All rights reserved.
PII: S0166-2236(99)01492-7