Annu. Rev. Neurosci. 2001. 24:299–325
Copyright c ? 2001 by Annual Reviews. All rights reserved
PROTEIN SYNTHESIS AT SYNAPTIC SITES
Oswald Steward1and Erin M. Schuman2
1Reeve-Irvine Research Center and Departments of Anatomy/Neurobiology and
Neurobiology and Behavior, College of Medicine, University of California at Irvine,
Irvine, California 92697; e-mail: email@example.com
2Division of Biology, 216-76 Howard Hughes Medical Institute, California Institute of
Technology, Pasadena, California 91125; e-mail: firstname.lastname@example.org
synaptic plasticity, long-term potentiation, gene expression,
I Abstract Studies over the past 20 years have revealed that gene expression in
neurons is carried out by a distributed network of translational machinery. One com-
ponent of this network is localized in dendrites, where polyribosomes and associated
membranous elements are positioned beneath synapses and translate a particular pop-
ulation of dendritic mRNAs. The localization of translation machinery and mRNAs
at synapses endows individual synapses with the capability to independently control
synaptic strength through the local synthesis of proteins. The present review discusses
recent studies linking synaptic plasticity to dendritic protein synthesis and mRNA
trafficking and considers how these processes are regulated. We summarize recent
information about how synaptic signaling is coupled to local translation and to the de-
may play a role in activity-dependent synaptic modification.
Synapses can undergo long-lasting changes in strength that likely contribute to
learning and memory. Many studies that have used protein synthesis inhibitors
have shown that long-lasting forms of behavioral and synaptic plasticity require
protein synthesis (Davis & Squire 1984, Bailey et al 1996, Kang & Schuman
1996, Mayford et al 1996, Nguyen & Kandel 1996, Schuman 1999). According
to this idea, behavioral experience or electrical activity at synapses induces the
synthesis of particular proteins that are critical for establishing enduring modifi-
cations. Possible roles for newly synthesized proteins include replacing degraded
proteins, increasing the levels of existing proteins, or expressing novel or alterna-
tively spliced forms of proteins.
STEWARD ? SCHUMAN
then there must be a mechanism for regulating translation via synaptic activity.
As discussed below, there exist signal transduction pathways that are activated by
synaptic events, resulting in the stimulation of translation. If new gene expres-
sion is required, there must be signaling from the synapse to the transcriptional
machinery of the neuron.
Recently synthesized proteins also need to be made available, selectively, to
those synapses undergoing changes in strength. The observation that individual
neurons are endowed with so many synapses suggests the capability of inde-
pendently processing and storing many bits of information. Electrophysiological
experiments have confirmed this to some extent; anatomically isolated groups of
strength at one set of afferents does not spread to enhance synaptic strength at
the other set of afferents (e.g. Andersen et al 1977). This property of synaptic
plasticity is often referred to as “input specificity.”
of afferents are spatially remote. If they are close (<∼50µm), there may be some
heterosynaptic enhancement (Bonhoeffer et al 1989, Engert & Bonhoeffer 1996,
potentiation (LTP) in the CA1 region as “site” rather than “input specific.” In the
dentate gyrus, the situation is somewhat different in that activation of one set
of afferents to a particular dendritic segment produces LTD at inactive synapses
that terminate nearby (a form of heterosynaptic depression; see Levy & Steward
1979, 1983). Again, this heterosynaptic interaction does not occur when inputs
are spatially separate (White et al 1988, White et al 1990).
Given these issues of specificity and protein synthesis dependence, there must
strength. If the source of new proteins is the soma, neurons could utilize known
protein transport and targeting mechanisms coupled with a “tag” or “marker” gen-
erated by synaptic activity (Martin et al 1997, Frey & Morris 1997). According
to this idea, newly synthesized proteins are selectively deposited or stabilized at
sites containing the tag. Another solution involves the local, rather than somatic,
synthesis of proteins at synaptic sites. This idea circumvents the need for a pro-
tein trafficking/capture mechanism, but does potentially require a mechanism for
targeting the mRNA to activated synapses.
The notion of dendritic protein synthesis had its roots in the discovery of
synapse-associated polyribosome complexes (SPRCs)—polyribosomes and as-
sociated membranous cisterns that are selectively localized beneath postsynaptic
sites on the dendrites of central nervous system (CNS) neurons (Steward 1983,
Steward & Fass 1983, Steward & Levy 1982). Based on the assumption that form
implies function, the highly selective localization of SPRCs beneath synapses
PROTEIN SYNTHESIS AT SYNAPSES
ular constituents of the synapse, and (b) translation might be regulated by activity
at the individual postsynaptic sites. As discussed below, there is now substantial
evidence to support these ideas. In addition, many of the mRNAs that are present
in dendrites have been identified; there is now considerable evidence that these
mRNAs enable a local synthesis of the encoded proteins (Steward et al 1996,
Steward & Singer 1997).
What has been missing until recently is a link between synaptic activation
and either the transport of mRNAs into dendrites or the local translation of these
this link and also begin to address the mechanisms underlying mRNA trafficking
and local translation. We begin with a brief review of the nature of the protein-
mRNAs that are localized in dendrites (and thus potentially are present at SPRCs).
We summarize recent evidence that local translation of mRNA at the synapse may
be regulated by synaptic activity and then summarize new information regarding
how afferent activity regulates the translation and trafficking of dendritic mRNAs.
We review the data suggesting that dendritic protein synthesis plays a key role
in long-term synaptic modifications. Finally, we consider how these pieces might
fit together to suggest a mechanism by which protein synthesis at the synapse
mediates long-term synaptic modifications induced by activity.
THE MACHINERY FOR PROTEIN SYNTHESIS
AT SYNAPTIC SITES
Synapse-associated polyribosome complexes are precisely localized in the post-
synaptic cytoplasm. One very important feature of SPRCs is the selectivity of
their localization. Quantitative electron-microscopic analyses have revealed that
the vast majority of the polyribosomes that are present in dendrites are precisely
positioned beneath postsynaptic sites and are absent from other parts of the den-
drite (Steward & Levy 1982). SPRCs are most often localized at the base of the
Thus, SPRCs are located within or near the portal between the spine neck and the
shaft of the dendrite—the route through which current must flow when spine
synapses are activated. In this location, SPRCs are ideally situated to be influ-
enced by electrical and/or chemical signals from the synapse as well as by events
within the dendrite proper. An important implication of this selective localization
is that there be some mechanism that causes ribosomes, mRNA, and other compo-
nents of the translational machinery to dock selectively in the postsynaptic cyto-
plasm. The mechanisms underlying this highly selective localization remain to be
STEWARD ? SCHUMAN
Although most dendritic polyribosomes are localized beneath synapses, a few
the synapse-associated polyribosomes. One possibility is that the polyribosomes
in the dendritic core are associated with mRNAs that encode proteins that are not
destined for synaptic sites but play some other role in dendritic function. This
speculation is of particular interest given the functional diversity of the mRNAs
that have been identified in dendrites (see below). Alternatively, the clusters of
polyribosomes in the core of the dendrite may represent packets of mRNAs and
ribosomes that are in transit from the cell body.
SPRCs are often associated with membranous organelles in a rough endoplas-
mic reticulum (RER)-like configuration. Serial section reconstructions of mid-
proximodistal dendrites of dentate granule cells and hippocampal pyramidal cells
revealed that ∼50% of the polyribosomes are found in association with tubular
cisterns (Steward & Reeves 1988). A common configuration is one in which the
ribosomes seem to surround a blind end of a cistern. Thus, the SPRC/cisternal
complex may be a form of RER that could allow the synthesis of integral mem-
brane proteins or soluble proteins destined for release. Evidence in support of this
hypothesis is considered in more detail below.
Interestingly, the cisterns with which SPRCs are associated are sometimes con-
nected with a spine apparatus (Steward & Reeves 1988). The significance of these
connections is not known. One interesting hypothesis is that the spine apparatus
may be involved in some aspect of posttranslational processing of proteins that
are synthesized at the SPRCs (more on this below).
Synapse-Associated Polyribosome Complexes at Different
Types of Synapses
SPRCs are present at spine synapses on different neuron types. Quantitative anal-
pyramidal cells, cortical neurons, and cerebellar Purkinje cells. These analyses
reveal that SPRCs are present in a roughly similar configuration in all of the spine-
bearing neurons that have been evaluated.
Estimates of the incidence of polyribosomes at spine synapses vary depending
on the quantitative methods used. In evaluations of single sections, ∼11%–15%
of the identified spines have underlying polyribosomes (Steward & Levy 1982).
However, this is clearly an underestimate because not all of the area under a spine
is contained within a single section. Serial section reconstructions of dendrites in
the dentate gyrus reveal that the actual incidence of polyribosomes in spines on
mid-proximodistal dendrites is ∼25% (Steward & Levy 1982). The estimates of
reconstruction techniques to evaluate the distribution of individual ribosomes (not
polyribosomes) yield higher estimates of incidence (Spacek & Hartmann 1983).
PROTEIN SYNTHESIS AT SYNAPSES
spines had ribosomes in the head, 42% had ribosomes in the neck, and 62% had
in the head, and 22% had ribosomes at the base. It is likely that an important
reason for the higher incidence values in this study is that single ribosomes were
counted rather than polyribosomes. In any case, it is clear that polyribosomes
are a ubiquitous component of the postsynaptic cytoplasm in a variety of neuron
SPRCs are also present at nonspine synapses. There have been no detailed
quantitative evaluations of polyribosome distribution in the dendrites of nonspiny
neurons, but it is clear that the same basic relationships exist as in spiny dendrites.
with submembranous cisterns and are found beneath both asymmetric (presumed
to yet another type of postsynaptic location; second, most (perhaps all) synapses
to spine apparatuses, known cisternal organelles. Based on the localization of sub-
synaptic cisternal organelles beneath both excitatory and inhibitory synapses on
axon initial segments, it may be worthwhile to reconsider the possible functions
of these enigmatic organelles. This is especially true because previous hypothe-
ses have focused on functions that would be especially important at excitatory
synapses and perhaps of minimal importance at inhibitory synapses (e.g. Ca2+
If protein synthetic machinery is localized at synapses in order to synthesize
some of the components of the synaptic junction, one would expect SPRCs to be
especially prominent at synapses during periods of synapse growth. This is the
case. Polyribosomes are very abundant in the dendrites of developing neurons and
again appear to be preferentially localized beneath postsynaptic sites, although
the degree of selectivity has not been evaluated quantitatively (Steward & Falk
PROTEINS SYNTHESIZED AT SYNAPSES
The discovery of polyribosomes beneath synapses focused attention on the ques-
tion of what proteins were synthesized in the postsynaptic cytoplasm. The ap-
proaches that have been used to address this question include the following:
(a) biochemical studies of proteins synthesized by subcellular fractions enriched
ment of mRNAs in isolated dendrites, most often from immature neurons grown
STEWARD ? SCHUMAN
Proteins Synthesized in Subcellular Fractions
Biochemical approaches take advantage of subcellular fractionation techniques
that allow the isolation of synaptosomes with attached fragments of dendrites
that retain their cytoplasmic constituents, including polyribosomes and associated
mRNAs. We have called these “synaptodendrosomes” (Rao & Steward 1991a).
toneurosomes” (Weiler & Greenough 1991, 1993; Weiler et al 1997).
The major limitation in using synaptodendrosomes or synaptoneurosomes to
study dendritic protein synthesis is that the fractions are contaminated with frag-
ments of neuronal and glial cell bodies. For example, high levels of the mRNA
encoding glial fibrillary acidic protein are present (Chicurel et al 1990, Rao &
Steward 1991b), and it is likely that there are also fragments of neuronal cell bod-
ies containing mRNAs that are normally not present in dendrites. The problem
of contamination can be partially circumvented by focusing on proteins that are
synthesized in synaptosomes and then assembled into synaptic structures. For ex-
ample, pulse-labeling techniques have been used to label the proteins synthesized
within synaptosomes and then subcellular fractionation and detergent extraction
tional complex were then characterized using polyacrylamide gel electrophoresis
combined with fluorography (Leski & Steward 1996, Rao & Steward 1991a). This
strategy has revealed characteristics of the labeled bands, but so far, the approach
has not provided definitive identification of the labeled proteins. This combined
strategy also has the limitation that it is useful only for proteins that are assembled
into the synaptic membrane or synaptic junctional complex. Thus, proteins that
are not assembled into the synapse are not detected.
Studies using synaptoneurosome fractions without the secondary purification
step of subcellular fractionation have provided evidence for the dendritic synthe-
sis of one novel protein that had not previously been identified—fragile X mental
retardation protein (FMRP) (Weiler et al 1997). FMRP is encoded by the fmr1
gene, whichisaffectedinhumanfragileXsyndrome. Treatmentofsynaptoneuro-
somes with agonists for metabotropic glutamate receptors caused a rapid increase
in the amount of FMRP in the synaptoneurosome fractions, as determined by
Western blot analysis. These data suggested that FMRP was being synthesized
within the fractions and that the synthesis was enhanced by metabotropic glu-
tamate receptor (mGluR) activation. This evidence has led to the idea that the
neuronal dysfunction that is part of fragile X syndrome may result from a dis-
ruption of local synthesis of protein at synapses (Comery et al 1997, Weiler et al
that FMRP mRNA is not evident in dendrites of neurons in vivo (Hinds et al 1993,
PROTEIN SYNTHESIS AT SYNAPSES
Valentine et al 2000). One can conceive of reasons why an mRNA might not
be detected in dendrites by in situ hybridization. For example, the mRNA could
be present at levels that are below the threshold for detection by standard in situ
data, however, which indicate an almost twofold increase in the amount of FMRP
detectable by Western blots within 5 min after treatment with mGluR agonists
(Weiler et al 1997). Presumably such a large change in protein concentration
could be achieved only if the levels of the mRNA were substantial.
It has also been suggested that FMRP plays a role in the regulation of trans-
lation of mRNAs at synapses (Feng et al 1997). This hypothesis is based on two
facts: (a) FMRP is an RNA-binding protein, and (b) immunocytochemical studies
reveal that the protein is localized at polyribosome clusters in neuronal dendrites.
Localization at dendritic polyribosomes is also consistent with the hypothesis that
FMRP is synthesized within dendrites. It is clear that the story regarding FMRP is
an evolving one.
The only other identified protein that has been shown to be synthesized in
synaptoneurosomes is the alpha subunit of calcium-/calmodulin-dependent pro-
tein kinase II (CAMKII) (Sheetz et al 2000). These studies involved metabolic
fluorography or by immunoprecipitation of metabolically labeled protein using an
antibody against CAMKII. Synthesis by fragments of neuronal-cell bodies or glia
observation that the labeling was regulated by treatment with neurotransmitters.
Although it was already known that the CAMKII message is present in dendrites,
this study documents that this mRNA is in actuality translated in synaptoneuro-
so as to yield fractions of greater purity. In this regard, one recent study reported
a fractionation approach that yields synaptosomes in which glial-fibrillary-acidic-
protein mRNA is not detected by reverse transcription-polymerase chain reaction
(Bagni et al 2000), suggesting a lack of contamination by glial fragments. Pre-
viously identified dendritic mRNAs such as those encoding CAMKII, Arc, and
an inositol 1,4,5 triphosphate (InsP3) receptor InsP3R1were detected in the same
by neurotransmitter activation. Interestingly, the mRNA for FMRP was also de-
tected in these fractions by reverse transcription-polymerase chain reaction. Thus,
these fractions may represent a purer population of synaptodendrosomes than has
been available previously, which could provide a means to identify novel dendritic
mRNAs in Dendrites in Vivo
Important clues about the identity of the proteins that may be synthesized at
STEWARD ? SCHUMAN
presence of particular mRNAs in dendrites. In most studies, dendritic localization
has been inferred by the pattern of labeling in brain regions where neuronal cell
bodies are concentrated in discrete layers and where there are distinct neuropil
layers that contain dendrites and axons but few neuronal cell bodies (e.g. cortical
regions including the hippocampus and the cerebellar cortex). Definitive evidence
using nonisotopic in situ hybridization techniques. Dendritic localization can also
be confirmed by studies of neurons in culture, although one must consider the
possibility that neurons in culture may express an unusual complement of mRNAs
or sort mRNAs in different ways than neurons in vivo.
not establish that the mRNA is translated at synapses. There are polyribosomes in
dendrites that are not localized beneath synapses, and these polyribosomes could
be associated with a different set of mRNAs than are translated in the postsynaptic
cytoplasm (Steward & Reeves 1988). Nevertheless, identification of mRNAs that
are present in dendrites provides candidates for synapse-associated mRNAs that
can be further evaluated in other ways.
Table 1 lists the mRNAs for which the evidence for dendritic localization in
the cell body. Certain other mRNAs that are localized primarily in cell bodies may
extend slightly into proximal dendrites. For example, it has been reported that
the mRNAs for two protein kinase C substrates (F1/GAP43 and RC3) extend
somewhat further into the proximal dendrites of forebrain neurons than other “cell
body” mRNAs. (Laudry et al 1993). The differences in the distribution of mRNAs
encoding F1/GAP43 and RC3 vs other cell body mRNAs are slight and indeed
were not evident in studies using nonisotopic in situ hybridization techniques that
produced heavy labeling over cell bodies (Paradies & Steward 1997).
Taken together, the information on mRNAs in dendrites allows several gener-
1. A number of different mRNAs that encode unrelated proteins are present in
dendrites. The proteins encoded by mRNAs that are present in dendrites
include a variety of different classes of protein (Table 1), including
cytoplasmic, cytoskeletal, integral-membrane, and membrane-associated
proteins. The proteins also have very different functions. Thus, it is likely
that the translation of these mRNAs subserves different aspects of cellular
and synaptic function [for additional discussion of this point, see Steward
2. All of the dendritic mRNAs that have been identified so far are expressed
differentially by different types of neurons. This is especially evident when
considering the mRNAs that are present in the dendrites of forebrain
neurons vs cerebellar Purkinje cells. For example, the mRNAs for MAP2,
CAMKII, dendrin, and Arc are found in forebrain neurons, but are not
expressed at high levels by Purkinje cells. Purkinje cells, on the other
PROTEIN SYNTHESIS AT SYNAPSES
hand, express a different complement of mRNAs including the InsP3
receptor and other Ca2+-interacting proteins, for example L7 and PEP19
(Bian et al 1996). These mRNAs are expressed by Purkinje cells and a few
other neuron types. The fact that different mixtures of mRNAs are present
in the dendrites of different cell types suggests that dendritic protein
synthesis may have different purposes in different cell types.
3. Although a number of mRNAs are present in the dendrites of forebrain
neurons, the patterns of expression and subcellular distributions of the
mRNAs are different. The mRNAs encoding CAMKII, dendrin, and Arc
(when induced) are localized throughout the dendrites. In contrast, the
mRNA for microtubule-associated protein 2 (MAP2) is found at high levels
in the proximal one third to one half of the total dendritic length, but it is
not detectable in distal dendrites of most neurons. The mRNAs that are
present in the dendrites of Purkinje cells also exhibit different localization
patterns. For example, the mRNA for the InsP3 receptor is present
throughout dendrites but is concentrated in the proximal one third of the
total dendritic length. The mRNA for L7 appears to be more uniformly
distributed throughout the dendrites (Bian et al 1996). Studies of the
trafficking of the immediate early gene Arc also reveal that the complement
of dendritic mRNAs can vary over time in an activity-dependent fashion
(see below). These findings indicate that the capability exists for a
different mixture of proteins to be synthesized locally at different times in
different neuron types and different dendritic domains, providing a
considerable complexity in the mechanisms underlying protein synthesis at
synapses. Also, the variety of subcellular distributions implies that there
must be multiple signals mediating mRNA localization within dendrites.
4. The presence of mRNAs encoding different classes of proteins implies the
existence of different types of translational machinery in dendrites. MAP2,
CAMKII, and Arc are nonmembrane proteins and thus would presumably
be synthesized by free polysomes. In contrast, the InsP3 receptor is an
integral membrane protein that presumably must be synthesized by
membrane-bound ribosomes (RER). As noted above, electron-microscopic
studies have revealed subsynaptic polyribosomes closely associated with
membranous cisterns that may represent a form of RER (Steward & Reeves
1988). However, the InsP3 receptor is also a glycoprotein; thus, there is the
question of how the newly synthesized protein is glycosylated (see below).
5. The dendritic localization of some mRNAs may be developmentally
regulated. For example, the mRNA for calmodulin can be detected by in
situ hybridization in dendritic laminae of developing, but not mature
animals (Berry & Brown 1996). This is consistent with the idea that local
dendritic protein synthesis is especially important during periods of
synaptogenesis (Palacios-Pru et al 1981, 1988; Steward & Falk 1986).
Several other mRNAs have been shown to be present in the dendrites
mRNAs that have been shown to be localized within dendrites of neurons in vivo by in situ hybridization
Localization in dendrites
Class of protein
Cortex, hippocampus, dentate gyrus
Proximal 1/3 − 1/2
Cortex, hippocampus, dentate gyrus
Cortex, hippocampus, dentate gyrus
Throughout (when induced)
depending on inducing stimulus
Hippocampus, dentate gyrus,
Cortex, hippocampus, dentate gyrus,
Ca2+signaling in conjunction
with CAMII kinase
Homology to PDGF
Proximal one third
Not shown are mRNAs that are localized only in the most proximal segments.
aGarner et al 1988.
bBurgin et al 1990.
cLink et al 1995, Lyford et al 1995.
dHerb et al 1997.
eWatson et al 1994.
fBerry & Brown 1996.
gGazzaley et al 1997.
hRacca et al 1997.
iPrakash et al 1997.
jParadies & Steward 1997.
kFuruichi et al 1993.
lBian et al 1996.
STEWARD ? SCHUMAN
of young neurons developing in vitro. For example, the mRNAs for
brain-derived neurotrophic factor (BDNF) and tyrosine kinase (trkB)
receptors extend into the proximal 30% of the total dendritic length of
hippocampal neurons in culture (Tongiorgi et al 1997). Potassium-induced
depolarization (Tongiorgi et al 1997) or BDNF treatment (Righi et al 2000)
increases the extent of dendritic labeling, so that the mRNAs extend to an
average of 60%–70% of the total dendritic length. Despite the easily
detectable dendritic labeling in neurons in vivo, the mRNAs for BDNF and
trkB receptors appear to be largely restricted to the region of the cell body
in young neurons in vivo (Dugich et al 1992). It remains to be seen
whether a dendritic localization can be induced in neurons in vivo by
manipulating neuronal activity.
Another mRNA that has been demonstrated in the dendrites of hippocampal
neurons in vitro encodes the fatty acylated membrane-bound protein ligatin. Flu-
orescent in situ hybridization analyses indicate that the mRNA for ligatin extends
for >100 µm into the dendrites of hippocampal neurons in culture; indeed, the
dendritic labeling produced by ligatin probes is nearly as extensive as that pro-
duced by probes for CAMKII (Severt et al 2000). Nevertheless, previous studies
of ligatin mRNA distribution in neurons in vivo reveal that the mRNA is largely
restricted to the cell body region of hippocampal neurons in vivo (Perlin et al
1993). It is possible that the radioisotopic in situ hybridization techniques used in
mRNAs in Dendrites of Neurons in Vitro
One approach to identifying dendritic mRNAs has been to use patch pipettes to
aspirate the cytoplasmic contents of individual dendrites of neurons grown in cul-
ture and then use RNA amplification techniques to clone the mRNAs (Miyashiro
et al 1994). This study provided intriguing evidence that there may be a substan-
tial number of mRNAs in dendrites, many of which remain to be characterized.
However, there are certain inconsistencies between these and other findings. For
example, the mRNAs for GluR1 receptors were detected, although the mRNAs
for these receptors have not been detected by in situ hybridization in dendrites of
neurons in vivo or in vitro. The reason for the disparity of results is not clear.
There are several possible explanations for the inconsistencies. One possibility
is that Miyashiro et al analyzed cytoplasm from dendrites of very young neurons
developing in culture. These might contain a different complement of mRNAs
than the dendrites of mature neurons. Another possibility is that the amplification
techniques detect mRNAs that are present in such low abundance that they are not
that some of the mRNAs in dendrites are in a form that somehow interferes with
hybridization by complementary probes.
PROTEIN SYNTHESIS AT SYNAPSES
New Dendritic mRNAs
It is almost certain that there are more dendritic mRNAs yet to be found. For
example, biochemical studies of proteins synthesized within synaptodendrosomes
prominent of these are not in a molecular-weight range that would be consistent
with their being the translation products of known dendritic mRNAs. Systematic
searches for new members of the family of dendritic mRNAs must deal with the
problem of how to obtain sufficient quantities of mRNA from dendrites that are
not contaminated by mRNA from neuronal cell bodies or supporting cells. So far,
systematic searches have not yet identified new members of the family of den-
dritic mRNAs whose presence in dendrites in vivo was later confirmed by in situ
Posttranslational Processing Within Dendrites
The presence of mRNAs encoding integral membrane proteins raises the question
of whether dendrites contain the machinery for posttranslational processing of
recently synthesized proteins (specifically, components of the RER and Golgi
apparatus [GA]). This question has been evaluated by assessing the distribution of
Initial studies evaluated the distribution of glycosyltransferase activities by
pulse-labeling neurons with the sugar precursors that are the substrates of various
added to nascent glycoproteins in the RER. Thus, when neurons are pulse-labeled
with [3H]mannose under conditions in which transport of recently synthesized
proteins is blocked, the sites of mannose incorporation can be revealed autoradio-
that are characteristic of the GA, for example fucosyltransferase and galactosyl-
transferase. Studies of this type provided the initial evidence for the presence of
both mannosyltransferase and higher-order (Golgi-like) glycosyltransferase activ-
ity in the dendrites of hippocampal neurons grown in culture (Torre & Steward
Additional evidence regarding the localization of the RER and GA in dendrites
has come from immunocytochemical studies of the subcellular distribution of
proteins that are considered markers of the two endomembrane systems (Torre &
RER marker) reveal staining that extends well into dendrites. In general, however,
immunostaining for Golgi markers extended only into proximal dendrites. The
immunocytochemical data were generally consistent with the autoradiographic
evidence regarding the intracellular distribution of glycosyltransferase activity
characteristic of the RER and GA.
STEWARD ? SCHUMAN
These data raise the question of what membranous organelles are actually re-
sponsible for the activities characteristic of the RER and GA. Recent electron-
microscopic immunocytochemical studies indicate that the membranous cisterns
present near spine synapses stain for Sec6Ialpha protein complex, which is part of
(Pierce et al 2000). The cisterns exhibiting labeling have the same appearance
as the cisterns that represent the membranous component of SPRCs. Moreover,
immunostaining for ribosomal protein S3 revealed labeling over the same mem-
branous cisterns that were labeled for Sec6Ialpha.
The organelle responsible for higher-order glycosyltransferase (Golgi-like) ac-
tivity remains to be identified. The studies of RER and GA distribution in den-
drites imply that RER is present throughout dendrites whereas machinery for
higher-order glycosylation may be present only in proximal dendrites, at least
in hippocampal neurons in culture (Torre & Steward 1996). Thus, if any in-
tegral membrane proteins are synthesized in distal dendrites of forebrain neu-
rons, their glycosylation may be incomplete. In this regard, nonglycosylated
membrane receptors can still function, although they may have different prop-
erties than fully glycosylated versions of the receptor (Giovannelli et al 1991). It
will be of interest to determine whether the story is different in Purkinje cells,
where the mRNA for at least one integral membrane protein is present throughout
REGULATION OF mRNA TRAFFICKING IN DENDRITES
Evidence that synaptic activation triggers the transport of new mRNA transcripts
to the synapse has come from studies of the immediate early gene Arc. Arc was
discovered in screens for novel immediate early genes, defined as genes that are
induced by activity in a protein synthesis-independent fashion (Link et al 1995,
Lyford et al 1995). In both studies, the inducing stimulus was a single electrocon-
vulsive seizure, and the protein synthesis independence was ensured by treating
animals with cycloheximide to block protein synthesis (and hence the synthesis
of secondary response genes). Arc was one of a number of novel immediate early
genes that were identified using this paradigm.
Recent studies have revealed an interesting feature of the trafficking of Arc
mRNA (Steward et al 1998). Patterned synaptic activation both induces Arc and
causes the newly synthesized mRNA to localize selectively to activated dendritic
stimulating the entorhinal cortical projections to the dentate gyrus in anesthetized
rats. The projection from the entorhinal cortex to the dentate gyrus (the perforant
path) terminates in a topographically organized fashion along the dendrites of
middle molecular layer of the dentate gyrus, whereas projections from the lateral
entorhinal cortex terminate in the outer molecular layer. Hence, by positioning a
PROTEIN SYNTHESIS AT SYNAPSES
stimulating electrode in the medial entorhinal cortex, it is possible to selectively
activate a band of synapses that terminate on mid-proximodistal dendrites.
When the medial perforant path was activated using a stimulation paradigm
that is commonly used to induce LTP (400-Hz trains, 8 pulses per train, deliv-
ered at a rate of 1 pulse/10 s), Arc expression was strongly induced. The newly
synthesized mRNA migrated into dendrites and accumulated selectively in the
middle molecular layer in exactly the location of the band of synapses that had
been activated. Similarly, activation of other afferent systems that terminate at
different locations in the molecular layer caused the newly synthesized mRNA
to localize selectively in other dendritic laminae. For example, simulation of the
lateral entorhinal cortex produced a band of labeling for Arc mRNA in the outer
molecular layer; stimulation of the commissural projection produced a band of la-
beling in the inner molecular layer (Steward et al 1998). Thus, these experiments
revealed that synaptic activation generated a signal that caused Arc mRNA to lo-
calize near the active synapses. The nature of this docking signal remains to be
Localization of Arc mRNA in activated dendritic laminae is associated with a
local accumulation of Arc protein. Immunostaining of tissue sections from stim-
ulated animals using an Arc-specific antibody revealed a band of newly synthe-
sized protein in the same dendritic laminae in which Arc mRNA was concentrated
(Steward et al 1998). The fact that synaptic activation leads to the selective tar-
geting of both recently synthesized mRNA and protein suggests that the targeting
of the mRNA underlies a local synthesis of the protein. Given that Arc protein
is usually highly localized in postsynaptic junctions, it is likely that the newly
synthesized Arc protein is targeted to the postsynaptic sites in the activated region.
This remains to be established, however.
A number of pieces of the puzzle are still missing. First, it remains to be
established whether Arc is directly involved in activity-induced synaptic modi-
fication. Additional clues to Arc’s function will likely come from studies of the
protein itself and its interactions with other functional molecules of the synaptic
junctional region. But even if Arc does turn out to be a red herring, these studies
have delineated RNA trafficking mechanisms that could be used for sorting other
mRNAs that participate in activity-dependent modifications.
constitutively. These include the mRNAs for molecules that have already been
strongly implicated in activity-dependent synaptic modification (e.g. the alpha
subunit of CAMII kinase). The mRNAs that are present constitutively provide
an opportunity for local regulation of the synthesis of key signaling molecules
via translational regulation (see below). Hence, protein synthesis at individual
synapses may be regulated in a complex fashion, first through the regulation of the
this is coordinated and how all of these molecules actually fit in to the molecular
consolidation process remain to be established.
STEWARD ? SCHUMAN
mRNA Binding Proteins
It is very likely that mRNAs are transported into dendrites in RNA-protein com-
plexes, which contain trans-acting factors responsible for RNA transportation as
identification of the minimal cis-elements sufficient for dendritic RNA targeting
should facilitate the characterization of the trans-acting factors that are involved
in this process. Several proteins that are involved in RNA targeting and asymmet-
ric distribution have been described in Drosophila embryos and Xenopus oocytes.
specific mRNAs in Drosophila oocytes and neuroblasts. Recently multiple mam-
malian staufen-related genes have been identified (Kiebler et al 1999, Marion
et al 1999, Tang et al 1999, Wickham et al 1999). At least two of them are present
in hippocampal neurons. Immunostaining experiments indicate somatodendritic
distribution patterns of the staufen protein in hippocampal neurons, with higher
pal slices. The staufen distribution pattern overlaps with that of dendritic RNA in
cultured neurons (Tang et al 1999). In addition, RNA-containing staufen protein
particles move along microtubules in the dendrites (Kohrmann et al 1999, Tang
et al 1999). Expression of a truncated staufen, lacking the microtubule-binding
domain, decreases the amount of RNA detected in dendrites (Tang et al 1999).
In addition, overexpression of staufen can increase the amount of clustering of
RNA in cultured hippocampal dendrites (Tang et al 1999). These observations are
that staufen binds in vivo remain to be determined, as does the potential regulation
of staufen trafficking during synaptic plasticity.
REGULATION OF mRNA TRANSLATION AT SYNAPSES
The selective localization of ribosomes at synapses provides a mechanism for
involve mRNAs already in place and/or mRNAs that are induced by synaptic
activity and delivered into dendrites. Until recently, there was no experimental
link between synaptic activity and either the transport of mRNAs into dendrites or
the local translation of these mRNAs at synapses. Recent studies, however, have
shown that synaptic activity can trigger the transport of new mRNA transcripts to
synapticsites, asdiscussedabove, andmodulatethetranslationofmRNAsalready
in place, as discussed below.
Initial evidence for synaptic modulation of protein synthesis within dendrites
came from studies of protein precursor incorporation by hippocampal slices in
vitro (Feig & Lipton 1993). When the Schaffer collaterals were activated in the
PROTEIN SYNTHESIS AT SYNAPSES
Interestingly, neither Schaffer collateral stimulation nor carbachol alone produced
caused a clear increase in protein synthesis, there was no detectable change in
Given the abundance of the mRNAs for the alpha-subunit of CAMKII and
MAP2 in the dendrites of hippocampal neurons (Burgin et al 1990), one obvious
experiment is to evaluate whether the dendritic synthesis of CAMKII or MAP2
can be modulated by synaptic activity. Two studies have addressed this ques-
tion by stimulating afferent projections to the hippocampus and dentate gyrus
and evaluating CAMKII levels using immunocytochemistry. Ouyang et al (1997,
1999) approached the question by delivering high-frequency stimulation to the
Schaffer collateral system in hippocampal slices in vitro. They observed increased
immunostaining for both phosphorylated and nonphosphorylated CAMKII in the
dendritic laminae. The increase in nonphosphorylated CAMKII occurred within
5 min of high-frequency stimulation and was blocked by anisomycin, strongly
suggesting that the kinase was synthesized in the dendrites. A fast transport of
CAMKII from the cell bodies seemed unlikely given the distance of the CAMKII
increases from the cell bodies (e.g. ∼150–200µm) as well as the fact that these
experiments were conducted at room temperature. In a related study, Steward and
path projections to the dentate gyrus in vivo also causes increases in immunos-
taining for CAMKII in the activated dendritic laminae (Steward & Halpain 1999).
There were also alterations in the pattern of immunostaining for MAP2, but the
nature of the changes was different than was the case for CAMKII. Specifically,
the activated lamina. The changes in immunostaining for MAP2 were diminished
but not eliminated by systemic or local application of protein synthesis inhibitors.
Surprisingly, however, the increases in immunostaining for CAMKII were not
affected by inhibiting protein synthesis. Thus, high-frequency synaptic activity
can cause domain-specific alterations in the molecular composition of dendrites,
but only a portion of the change may be attributable to local protein synthesis. In
comparing these results to those of Ouyang et al (1997, 1999), it is possible that
different stimulation frequencies (e.g. 100 vs 400 Hz) differentially invoke local
synthesis vs trafficking of CAMKII. In addition, these studies were conducted in
different areas of the hippocampus.
Although these combined results suggest synaptic regulation of dendritic pro-
tein synthesis, the possibility that newly synthesized proteins were actually pro-
ruled out. In this regard, parallel studies in subcellular fractions provide important
Studies in Subcellular Fractions
In addition to the FMRP studies discussed above, recent studies of synapto-
neurosomes prepared from frog tectum have revealed a novel mechanism for
STEWARD ? SCHUMAN
synaptic regulation of the translation of the alpha subunit of CAMKII. There is
activity and that the N-methyl-D-aspartate (NMDA) receptors play an important
role in this process (Sheetz et al 1997). Based on these findings and the fact that
receptor activation of synaptoneurosomes from the tectum modulates CAMKII
synthesis. Their results revealed a surprising and complex translation regulation
mechanism. NMDA receptor activation did enhance CAMKII synthesis within
time, however, there was an increase in the phosphorylation of the initiation factor
ing the overall protein synthesis rate). This apparent paradox can be explained
by the fact that studies of mRNA competition in translation assays reveal that de-
creases in elongation favor the translation of weakly initiated mRNAs. CAMKII
is one of the mRNAs for which initiation is inefficient, and so general decreases in
elongation consequent to IF2 phosphorylation could lead to increases in CAMKII
synthesis. Sheetz et al (2000) also provided evidence in support of this idea by
showing that low to moderate concentrations of cycloheximide (which partially
inhibited overall protein synthesis) caused increases in the synthesis of CAMKII
at the same time that overall levels of protein synthesis were diminished. These
results thus provided evidence for regulation of CAMKII mRNA translation via
NMDA receptor activation.
Other recent studies using synaptoneurosomes have revealed another novel
mechanism for the control of the translation of CAMKII mRNA at synapses (Wu
et al 1998). An unusual feature of CAMKII mRNA is that it contains a sequence
in its 3?-untranslated region that is a consensus sequence for the binding of a
cytoplasmic polyadenylation element (CPE). CPEs are known to play a key role in
regulating the translation of maternal mRNAs in oocytes. These mRNAs, which
regulate the development of polarity in the embryo (i.e. defining dorsal vs ventral,
anterior vs posterior, and particular body regions).
Maternal mRNAs are translationally repressed until fertilization. Upon fertil-
ization, translation repression is relieved through the action of the CPE. Trans-
lationally repressed maternal mRNAs have very short poly-A tails. At fertiliza-
tion, CPE is activated and triggers an elongation of the poly-A tail, resulting
in translation induction. In an interesting experiment, Wu et al (1998) demon-
strated that the translation of CAMKII kinase was regulated in a similar way
in brain. In particular, NMDA receptor activation triggered polyadenylation of
the mRNA for CAMII kinase, which in turn increased the synthesis
of CAMKII protein. They further showed that this activation could be triggered
by behavioral experience (light exposure for animals raised in the dark). This
study thus revealed a second mechanism through which CAMKII protein
synthesis could be regulated at the synapse. It remains to be seen how this
PROTEIN SYNTHESIS AT SYNAPSES
mechanism interacts with the mechanism suggested in the experiments by Sheetz
et al (2000).
mRNAs at synapses. Activation of particular neurotransmitter receptors appears
to play a role in regulating translation, and both metabotropic and NMDA-type
different mRNAs are controlled in different ways or whether different control
mechanisms are present at different types of synapses.
Coupling Synaptic Activity to Translation
machinery are not yet well understood. One signaling pathway that is stimulated
by growth factors and results in the translation of several mRNAs includes the
rapamycin-sensitive kinase mammalian target of rapamycin [mTOR, also known
BP-2, and eIF-4E, are present in the rat hippocampus, as shown by Western blot
analysis and immunostaining studies (Tang et al 1998). In cultured hippocampal
neurons, the distribution of these factors overlaps substantially with a synaptic
protein, synapsin-I, suggesting synaptic localization. Disruption of mTOR ac-
tivity by rapamycin results in deficits in late-phase LTP expression induced by
high-frequency stimulation, while the early phase of LTP is unaffected (Tang
et al 1998). Rapamycin also blocks the synaptic enhancement induced by BDNF
in hippocampal slices (Tang et al 1998) as well as the long-term synaptic facilita-
tion induced by repeated presentations of 5-hydroxytryptamine (5-HT) at Aplysia
synapses in culture (Casadio et al 1999). These results imply an essential role for
the rapamycin-sensitive signaling in three different forms of synaptic plasticity
that require new protein synthesis. The localization of this translational signaling
pathway at synaptic sites may provide a mechanism that controls local protein
synthesis at potentiated synapses.
ROLE OF LOCAL SYNTHESIS IN SYNAPTIC PLASTICITY
It is now well accepted that the late phase of LTP requires both transcription and
translation. Indeed, the term late-phase LTP is operationally defined as the tem-
poral phase of potentiation that can be blocked by inhibitors of protein synthesis.
Initial evidence came from studies of hippocampal LTP in vitro, which reported
thatthreedifferenttranslationinhibitors, emetine, cycloheximide, andpuromycin,
reduced the proportion of slices exhibiting LTP (Stanton & Sarvey 1984). A
fourthinhibitor, anisomycin, reducedonlythemagnitudeofLTPofthepopulation
spike. In subsequent studies of LTP in the dentate gyrus in vivo, it was found that
STEWARD ? SCHUMAN
pretreatment with anisomycin did not block LTP induction, but did result in the
Subsequent studies of LTP in hippocampal slices yielded similar results (protein
synthesis inhibitors block the late phase of LTP but not the early phase). (Nguyen
& Kandel 1997, Osten et al 1996). These studies and others provide the basis for
the widely held conclusion that the synthesis of new proteins is required for long
lasting (e.g. ≥∼1 h) potentiation at both the Schaffer collateral-CA1 synapses in
the hippocampus and perforant path synapses in the dentate gyrus.
A recent study (Raymond et al 2000) has suggested that the “priming” of LTP
by mGluRs is also mediated by stimulation of local protein synthesis. Priming
refers to the transformation of a small, decaying potentiation into a long-lasting
mGluRs (by a selective agonist) 20 min before a tetanus promotes LTP induction
and persistence (Cohen & Abraham 1996, Cohen et al 1998). The application of
a protein synthesis inhibitor during mGluR activation prevents the priming effect
(Raymond et al 2000). Given the rapidity of the priming effect (<20 min), these
data are consistent with a mechanism involving mGluR stimulation of translation
(Weiler et al 1997).
Although the above studies support the conclusion that the late phase of LTP
are synthesized in the cell body or in the dendrites or, for that matter, whether the
required protein synthesis occurs pre- or postsynaptically.
Dissecting Sites of the Protein Synthesis Required for LTP
One of the earliest studies linking dendritic protein synthesis to synaptic plasticity
examined the protein synthesis dependence of potentiation induced by the growth
factors BDNF and neurotrophin-3 (Kang & Schuman 1995, 1996). Application of
either BDNF or neurotrophin-3 to CA1 synapses caused a large and long lasting
(2- to 3-h) enhancement of synaptic transmission (Kang & Schuman 1995). Sur-
prisingly, pretreatment with a protein translation inhibitor blocked both the early
and late phases of the enhancement (Kang & Schuman 1996). This early (e.g.
within 10-min) requirement for protein synthesis is not consistent with a somatic
origin, given the distance between the recording site (distal dendrites) and the
cell bodies. Microlesion experiments, in which the synaptic neuropil was isolated
from the cell bodies, showed directly that the protein synthesis source was not the
cell bodies. Dendrites isolated from their cell bodies still exhibited growth-factor-
induced synaptic potentiation that was sensitive to translation inhibitors. More
recent experiments have shown that BDNF can stimulate the translation of a den-
dritically localized green fluorescent protein mRNA in isolated dendrites (Smith
et al 1999).
PROTEIN SYNTHESIS AT SYNAPSES
Proteins synthesized in dendrites may also contribute to long-term decreases in
synaptic strength or long-term depression (LTD). There are two forms of LTD,
one that requires the activation of NMDA receptors and the other that requires
the activation of mGluRs. Capitalizing on the link between mGluR activity and
stimulation of dendritic protein synthesis, Bear and colleagues examined whether
mGluR-dependent LTD requires protein synthesis that is local (Huber et al 2000).
somycin or the postsynaptic infusion of mRNA cap analog m7GpppG (Huber et al
examined. Taken together with the results above, these data indicate that locally
synthesized proteins can contribute to bidirectional changes in synaptic strength.
Long-Term Facilitation in Aplysia Species
There is abundant evidence that long-term synaptic modifications in invertebrates
also require local protein synthesis at or near synapses. An entire review could be
written on this topic, so here we summarize the evidence only briefly. In Aplysia
the molecular bases of behavioral sensitization have been studied in a variety of
reduced preparations, including the synapses formed between isolated sensory
and motoneurons in culture (Montarolo et al 1986). In this system, repeated ap-
plications of 5-HT result in a protein synthesis-dependent long-term facilitation
of synaptic transmission between the sensory and motoneurons (Clark & Kandel
1993, Martin et al 1997). When a single sensory neuron with a bifurcating axon
(Martin et al 1997). This long-term facilitation requires local protein synthesis:
restricted application of a translation inhibitor to the 5-HT-treated synapse blocks
the facilitation (Martin et al 1997). The long-term facilitation can be “captured”
by the neighboring “naive” synapse if it is treated with single application of 5-HT
within a few hours (Casadio et al 1999, Martin et al 1997). The long-term (e.g.
72-h) expression of synaptic capture also requires local protein synthesis. If a
translation inhibitor is coapplied with the single 5-HT pulse to the “naive,” this
synapse will no longer exhibit long-term facilitation (Casadio et al 1999).
The presence of polyribosomes, translation machinery, and mRNAs in dendrites
endows individual synapses with the capability to independently control synaptic
strength through the local synthesis of proteins. Studies in the past few years have
provided strong evidence linking synaptic plasticity to dendritic protein synthesis
and mRNA trafficking. Relatively little is know about how these processes are
STEWARD ? SCHUMAN
regulated including the coupling of synaptic signaling to translation machinery,
the selective translation of specific mRNAs present at synapses, and the specific
delivery of newly transcribed mRNAs to activated synaptic sites. Future studies
will no doubt continue to strengthen the idea that local control of synaptic mRNAs
and proteins is essential for maintaining the complexity of synaptic connections
in the nervous system.
and HHMI (EMS).
Visit the Annual Reviews home page at www.AnnualReviews.org
Andersen P, Sundberg SH, Sveen O, Wigstrom
synaptic transmission in hippocampal slices.
Bagni C, Mannucci L, Dotti CG, Amaldi
F. 2000. Chemical stimulation of synap-
tosomes modulates alpha-Ca2+/calmodulin-
dependent protein kinase II mRNA associ-
ation to polysomes. J. Neurosci. 20(RC76):
Bailey CH, Bartsch D, Kandel ER. 1996. To-
ward a molecular definition of long-term
memory storage. Proc. Natl. Acad. Sci. USA
Berry FB, Brown IR. 1996. CaM I mRNA is
localized to apical dendrites during postnatal
development of neurons in the rat brain. J.
Neurosci. Res. 43:565–75
Bian F, Chu T, Schilling K, Oberdick J. 1996.
Differential mRNA transport and the regula-
of Purkinje cell dendritic mRNAs to trans-
lational inhibition. Mol. Cell. Neurosci. 7:
Bonhoeffer T, Staiger V, Aertsen A. 1989.
Synaptic plasticity in rat hippocampal slice
cultures: Local “Hebbian” conjunction of
pre- and postsynaptic stimulation leads
to distributed synaptic enhancement. Proc.
Natl. Acad. Sci. USA 86:8113–17
Brown EJ, Schreiber SL. 1996. A signaling
pathway to translational control. Cell 86:
SA, Mobley WC, Kelly PT. 1990. In situ hy-
bridization histochemisty of Ca/calmodulin
dependent protein kinase in developing rat
brain. J. Neurosci. 10:1788–89
Casadio A, Martin KC, Giustetto M, Zhu H,
form of CREB-mediated long-term facilita-
tion can be stabilized at specific synapses by
local protein synthesis. Cell 99:221–37
cellular localization of mRNA: isolation and
characterization of mRNA from an enriched
preparation of hippocampal dendritic spines.
Soc. Neurosci. Abstr. 16:353
Clark GA, Kandel ER. 1993. Induction of
long-term facilitation in Aplysia sensory
neurons by local application of serotonin to
remote synapses. Proc. Natl. Acad. Sci. USA
Cohen AS, Abraham WC. 1996. Facilitation of
long-term potentiation by prior activation of
metabotropic glutamate receptors. J. Neuro-
Cohen AS, Raymond CR, Abraham WC. 1998.
Priming of long-term potentiation induced
by activation of metabotropic glutamate re-
ceptors coupled to phospholipase C. Hip-
PROTEIN SYNTHESIS AT SYNAPSES
Irwin SA, et al. 1997. Abnormal dendritic
spines in fragile X knockout mice: matura-
tion and pruning deficits. Proc. Natl. Acad.
Sci. USA 94:5401–4
Davis HP, Squire LR. 1984. Protein synthe-
sis and memory: a review. Psychol. Bull.
Dugich MM, Tocco G, Willoughby DA, Najm
I, Pasinetti G, et al. 1992. BDNF mRNA ex-
kainic acid-induced seizure activity. Neuron
Engert F, Bonhoeffer T. 1996. Pairing-induced
LTP in hippocampal slice cultures is not
strictly input-specific. Soc. Neurosci. Abstr.
Feig S, Lipton P. 1993. Pairing the cholinergic
agonist carbachol with patterned scaffer col-
lateral stimulation initiates protein synthesis
via a muscarinic, NMDA-dependent mecha-
nism. J. Neurosci. 13:1010–21
Feng Y, Gutekunst C-A, Eberhart DE, Yi
H, Warren ST, Hersch SM. 1997. Fragile
X mental retardation protein: nucleocyto-
plasmic shuttling and association with so-
matodendritic ribosomes. J. Neurosci. 17:
Frey U, Krug M, Reymann KG, Matthies H.
1988. Anisomycin, an inhibitor of protein
ena in the hippocampal CA1 region in vitro.
Brain Res. 452:57–65
Frey U, Morris RGM. 1997. Synaptic tagging
Furuichi T, Simon-Chazottes D, Fujino I,
Yamada N, Hasegawa M, et al. 1993.
Widespread expression of inositol 1,4,5-
triphosphate receptor type 1 gene (Insp3r1)
in the mouse central nervous system. Recept.
Garner CC, Tucker RP, Matus A. 1998. Selec-
tive localization of messenger RNA for cy-
Gazzaley AH, Benson DL, Huntley GW,
Morrison JH. 1997. Differential subcellular
regulation of NMDAR1 protein and mRNA
in dendrites of dentate gyrus granule cells
after perforant path transection. J. Neursci.
Giovannelli A, Grassi F, Eusebi F, Miledi R.
of acetylcholine receptors in cultured mouse
muscle cells. Proc. Natl. Acad. Sci. USA
Herb A, Wisden W, Catania DMV, Marechal
D, Dresse A, Seeberg PH. 1997. Prominent
dendritic localization in forebrain neurons of
Cell Neurosci. 8:367–74
Hinds HL, Ashley CT, Sutcliffe JS, Nelson DL,
Warren ST, et al. 1993. Tissue specific ex-
pression of FMR-1 provides evidence for a
functional role in fragile X syndrome. Nat.
Huber KM, Kayser MS, Bear MF. 2000. Role
for rapid dendritic protein synthesis in hip-
pocampal mGluR-dependent long-term de-
pression. Science 288:1254–56
Kang H, Schuman EM. 1995. Long-lasting
neurotrophin-induced enhancement of syn-
aptic transmission in the adult hippocampus.
Kang H, Schuman EM. 1996. A requirement
for local protein synthesis in neurotrophin-
induced synaptic plasticity. Science 273:
Kiebler MA, Hemraj I, Verkade P, Kohrmann
M, Fortes P, et al. 1999. The mammalian
Staufen protein localizes to the somatoden-
dritic domain of cultured hippocampal neu-
rons: implications for its involvement in
mRNA transport. J. Neurosci. 19:288–97
TJ, Ellisman MH, et al. 1996. Translocation
of RNA granules in living neurons. J. Neu-
Kohrmann M, Luo M, Kaether C, Des-
Groseiller L, Dotti CG, Kiebler MA.
1999. Microtubule-dependent recruitment
of Staufen-green fluorecent protein into
STEWARD ? SCHUMAN
RNA-containing granules and subsequent
dendritic transport in living hippocampal
neurons. Mol. Biol. Cell 10:2945–53
Laudry CF, Watson JB, Handley VW, Cam-
pagnoni AT. 1993. Distribution of neuronal
and processes. Soc. Neurosci. Abstr. 19:1745
Leski ML, Steward O. 1996. Synthesis of pro-
teins within dendrites: ionic and neurotrans-
mitter modulation of synthesis of particu-
lar polypeptides characterized by gel elec-
trophoresis. Neurochem. Res. 21:681–90
Levy WB, Steward O. 1979. Synapses as asso-
ciative memory elements in the hippocampal
formation. Brain Res. 175:233–45
Levy WB, Steward O. 1983. Temporal conti-
guity requirements for long-term associative
potentiation/depression in the hippocampus.
Link W, Konietzko G, Kauselmann G, Krug M,
pression of an immediate early gene is regu-
lated by synaptic activity. Proc. Natl. Acad.
Sci. USA 92:5734–38
Lyford G, Yamagata K, Kaufmann W, Barnes
and activity-regulated gene, encodes a novel
cytoskeleton-associated protein that is en-
Marion RM, Fortes P, Beloso A, Dotti C, Ortin
J. 1999. A human sequence homology of
staufen is an RNA-binding protein that is
associated with polysomes and localizes to
the endoplasmic reticulum. Mol. Cell. Biol.
Martin KC, Casadio A, Zhu H, E Y, Rose JC,
ory storage. Cell 91:927–38
Mayford M, Bach ME, Huang Y-Y, Wang L,
Hawkins RD, Kandel ER. 1996. Control
of memory formation through regulated ex-
pression of a CaMKII transgene. Science
Miyashiro K, Dichter M, Eberwine J. 1994.
On the nature and differential distribution of
mRNAs in hippocampal neurites: implica-
tions for neuronal functioning. Proc. Natl.
Acad. Sci. USA 91:10800–4
J, Kandel ER, Schacher S. 1986. A critical
term heterosynaptic facilitation in aplysia.
Nguyen PV, Kandel ER. 1996. A macromoleu-
lar synthesis-dependent late phase of long-
term potentiation requiring cAMP in the me-
dial perforant pathway of rat hippocampal
slices. J. Neurosci. 16:3189–98
Nguyen PV, Kandel ER. 1997. Brief theta
burst stimulation induced a transcription-
in area CA1 of the mouse hippocampus.
Learn. Mem. 4:230–43
Osten P, Valsamis L, Harris A. 1996. Protein
synthesis-dependent formation of protein
kinase M zeta in long-term potentiation. J.
tein synthesis in the dentate gyrus, but not
the entorhinal cortex, blocks maintenance of
Otani S, Marshall CJ, Tate WP, Goddard GV,
Abraham WC. 1989. Maintenance of long-
protein synthesis but not messenger RNA
synthesis immediate post-tetanization. Neu-
Ouyang Y, Kantor DB, Harris KM, Schu-
man EM, Kennedy MB. 1997. Visualiza-
tion of the distribution of autophosphory-
lated calcium/calmodulin-dependent protein
kinase II after tetanic stimulation in the
CA1 area of the hippocampus. J. Neurosci.
Ouyang Y, Rosenstein A, Kreiman G, Schu-
man EM, Kennedy MB. 1999. Tetanic stim-
ulation leads to increased accumulation
of Ca(2+)/calmodulin-dependent protein ki-
nase II via dendritic protein synthesis in hip-
pocampal neurons. J. Neurosci. 19:7823–33
PROTEIN SYNTHESIS AT SYNAPSES
Palacios-Pru EL, Miranda-Contreras L, Men-
doza RV, Zambrano E. 1988. Dendritic
RNA and postsynaptic density formation
in chick cerebellar synaptogenesis. Neuro-
Palacios-Pru EL, Palacios L, Mendoza RV.
1981. Synaptogenetic mechanisms during
chick cerebellar cortex development. J. Sub-
microsc. Cytol. 13:145–67
Paradies MA, Steward O. 1997. Multiple sub-
cellular mRNA distribution patterns in neu-
ysis. J. Neurobiol. 33:473–93
Perlin JB, Gerwin CM, Panchision DM, Vick
produces long-lasting and selective changes
in gene expression in hippocampal neurons.
Proc. Natl. Acad. Sci. USA 90:1741–45
Pierce JP, van Leyen K, McCarthy JB. 2000.
Translocation machinery for synthesis of in-
tegral membrane and secretory proteins in
dendritic spines. Nat. Neurosci. 3:311–13
Prakash N, Fehr S. Mohr E. Richter D. 1997.
Dendritic localization of rat vasopressin
mRNA: ultrastructural analysis and map-
ping of targeting elements. Eur. J. Neurosci.
Racca C, Gardiol A, Triller A. 1997. Dendritic
and postsynaptic localizations of glycine re-
ceptor alpha subunit mRNAs. J. Neurosci.
Rao A, Steward O. 1991a. Evidence that pro-
tein constituents of postsynaptic membrane
ysis of proteins synthesized within synapto-
somes. J. Neurosci. 11:2881–95
assessment of contamination by glia and
comparison with total RNA. Soc. Neurosci.
Raymond CR, Thompson VA, Tate WR, Abra-
ham WC. 2000. Metabotropic glutamate
receptors trigger homosynaptic protein syn-
thesis to prolong long-term potentiation. J.
Righi M, Tongiorgi E, Cattaneo A. 2000.
Brain-derived neurotrophic factor (BDNF)
induces dendritic targeting of BDNF and
tyrosine kinase B mRNAs in hippocam-
pal neurons through a phosphatidylinositol-
3 kinase-dependent pathway. J. Neurosci.
Schuman EM. 1999. mRNA trafficking and lo-
cal protein synthesis at the synapse. Neuron
Schuman EM, Madison DV. 1994. Locally dis-
tributed synaptic potentiation in the hip-
pocampus. Science 263:532–36
Severt WL, Biber TUL, Wu X-Q, Hecht NB,
sion of testis-brain RNA binding protein and
kinesin heavy chain disrupts mRNA sorting
in dendrites. J. Cell. Sci. 112:3691–702
Sheetz AJ, Nairn AC, Constantine-Paton M.
1997. N-methyl-D-aspartate receptor activa-
tion and visual activity induce elongation
factor-2 phosphorylation in amphibian tecta:
a role for N-methyl-D-aspartate receptors
in controlling protein synthesis. Proc. Natl.
Acad. Sci. USA 94:14770–75
Sheetz AJ, Nairn AC, Constantine-Paton M.
2000. NMDA receptor-mediated control of
protein synthesis at developing synapses.
Nat. Neurosci. 3:211–16
Smith WB, Aakalu GN, Tsung MI., Reis GF,
Schuman EM. 1999. Neurotrophin-induced
hippocampal neurons. Soc. Neurosci. Abstr.
dimensional analysis of dendritic spines. I.
spines and synaptic morphology in cerebral
and cerebellar cortices. Anat. Embryol.
Stanton PK, Sarvey JM. 1984. Blockade of
long-term potentiation in rat hippocampal
CA1 region by inhibitors of protein synthe-
sis. J. Neurosci. 4:3080–84
Steward O. 1983. Polyribosomes at the base of
dendritic spines of CNS neurons: their pos-
sible role in synapse construction and mod-
ification. Cold Spring Harbor Symp. Quant.
STEWARD ? SCHUMAN
Steward O. 1997. mRNA localization in neu-
Steward O, Falk PM. 1986. Protein syn-
thetic machinery at postsynaptic sites dur-
ing synaptogenesis; a quantitative study
of the association between polyribosomes
and developing synapses. J. Neurosci. 6:
Steward O, Falk PM, Torre ER. 1996. Ultra-
structural basis for gene expression at the
synapse: synapse-associated polyribosome
complexes. J. Neurocytol. 25:717–34
Steward O, Fass B. 1983. Polyribosomes as-
sociated with dendritic spines in the dener-
vated dentate gyrus: evidence for local regu-
lation of protein synthesis during reinnerva-
tion. Prog. Brain Res. 58:131–36
Steward O, Halpain S. 1999. Lamina-specific
synaptic activation causes domain-specific
alterations in dendritic immunostaining for
MAP2 and CAM kinase II. J. Neurosci.
Steward O, Levy WB. 1982. Preferential local-
ization of polyribosomes under the base of
dendritic spines in granule cells of the den-
tate gyrus. J. Neurosci. 2:284–91
Steward O, Pollack A, Rao A. 1991. Evi-
dence that protein constituents of postsy-
naptic membrane specializations are locally
synthesized: time course of appearance of
recently synthesized proteins in synaptic
junctions. J. Neurosci. Res. 30:649–60
Steward O, Reeves TM. 1988. Protein syn-
thetic machinery beneath postsynaptic sites
bosomes and other organelles at the synaptic
site. J. Neurosci. 8:176–84
Steward O,RibakCE. 1986.Polyribosomesas-
sociated with synaptic sites on axon initial
segments: localization of protein synthetic
machinery at inhibitory synapses. J. Neu-
Steward O, Singer RH. 1997. The intracellu-
lar mRNA sorting system: postal zones, zip
codes, mail bags and mail boxes. In mRNA
Metabolism and Post-Transcriptional Gene
Regulation, ed. JB Hartford, DR Morris, pp.
127–46. New York: Wiley-Liss
Steward O, Wallace CS, Lyford GL, Wor-
ley PF. 1998. Synaptic activation causes the
Tang SJ, Meulemans D, Vasquez L, Schuman
EM. 1999. The role of a rat staufen-like pro-
pal neurons. Proc. Annu. Meet. Soc. Neu-
Tang SJ, Smith WB, Schuman ME. 1998. Iden-
tification of components of a translational
pocampus. Proc. Annu. Meet. Soc. Neurosci.
Tongiorgi E, Righi M, Cattaneo A. 1997.
Activity-dependent dendritic targeting of
BDNF and TrkB mRNAs in hippocampal
neurons. J. Neurosci. 17:9492–505
Torre ER, Steward O. 1996. Protein synthesis
within dendrites: distribution of the endo-
dendrites of hippocampal neurons in culture.
J. Neurosci. 16:5967–78
C, Herkenham M. 2000. Fragile X (fmr1)
mRNA expression is differentially regulated
in two adult models of activity-dependent
gene expression. Mol. Brain Res. 75:337–41
Watson JB, Coulter PM, Margulies JE, de
gamma7 subunit is selectively expressed in
medium-sized neurons and dendrites of the
rat neostriatum. J. Neurosci. Res. 39:108–16
Weiler IJ, Greenough WT. 1991. Potassium
ion stimulation triggers protein translation
in synaptoneuronsomal polyribosomes. Mol.
Cell. Neurosci. 2:305–14
Weiler IJ, Greenough WT. 1993. Metabotropic
tein synthesis. Proc. Natl. Acad. Sci. USA
Weiler IJ, Irwin SA, Klintsova AY, Spencer
CM, Brazelton AD, et al. 1997. Fragile X
mental retardation protein is translated near
PROTEIN SYNTHESIS AT SYNAPSES Download full-text
synapses in response to neurotransmitter ac-
White G, Levy WB, Steward O. 1988. Evi-
dence that associative interactions between
afferents during the induction of long-term
potentiation occur within local dendritic
domains. Proc. Natl. Acad. Sci. USA 85:
White G, Levy WB, Steward O. 1990. Spatial
overlap between populations of synapses de-
termines the extent of their associative in-
teraction during the induction of long term
Groseillers L. 1999. Mammalian staufen is a
double-stranded RNA- and tubulin-binding
protein which localizes to the rough endo-
plasmic reticulum. Mol. Cell. Biol. 19:2220–
Wu L, Wells D, Tay J, Mendis D, Abborr
M-A, et al. 1998. CPEB-mediated cytoplas-
mic polyadenylation and the regulation of
experience-dependent translation of alpha-
CaMKII at synapses. Neuron 21:1129–39