Scientific Report 2005
Interrelationships Between Dendritic Protein Synthesis and the Efficacy and Structure of Synaptic Connections
P.W. Vanderklish, A. Aschrafi, A. Atkins, F. Smart, G.M. Edelman
The strength and reliability of synaptic
communication between neurons (synaptic efficacy) is not a fixed property. Rather,
the ability of synapses to undergo long-term changes in efficacy in response to
particular patterns of synaptic activity (synaptic plasticity) is an essential property
of neural circuits involved in learning, memory, and various other higher order
brain functions. The goal of our research is to define the mechanisms by which long-term
forms of activity-dependent synaptic plasticity are consolidated and how these mechanisms
are altered in fragile X syndrome (FXS), the most common inherited form of mental
Three basic observations guide our hypotheses. First, translation
of dendritically localized mRNAs is required to stabilize changes in efficacy in
at least 3 forms of synaptic plasticity: long-term potentiation, long-term depression
(LTD), and synaptic enhancement induced with brain-derived neurotrophic factor.
Second, each form of plasticity can be associated with unique changes in the morphology
of dendritic spines. Third, changes in efficacy can outlast the half-lives of proteins
synthesized during the induction of the changes.
We propose that local translation plays an early and necessary
role in transforming synaptic shape and that molecular determinants of synaptic
shape, in turn, regulate the synthesis and localization of proteins (e.g., glutamate
receptors) that determine synaptic efficacy. Such regulatory interrelationships
are predicted to be unique for each form of plasticity, and those mediating LTD
consolidation are proposed to underlie synaptic malfunction in FXS.
In previous work, we found evidence for reciprocal interactions
between synaptic translation and structure. We showed that stimulation of metabotropic
glutamate receptors (mGluRs) that induce a form of translation-dependent LTD leads
to elongation of dendritic spines and that this effect is blocked by a translation
inhibitor. In related work, we observed that the translation machinery is reciprocally
influenced by determinants of synaptic structure. Treatment of neural preparations
with brain-derived neurotrophic factor resulted in translocation of the initiation
factor 4E to synaptic mRNA granules. Depolymerization of F-actin and antagonism
of integrins blocked this effect.
The changes in spines we observed after stimulation of mGluRs
resembled the abnormally long and thin spines seen in FXS.
FXS is caused by the silencing of a gene, Fmr1, which encodes
a protein (FMRP) that can function as a translational suppressor in dendrites. In
mice lacking Fmr1, mGluR-induced LTD is enhanced. Because longer, thinner
spines contain fewer glutamate receptors, our data support the notion that mGluR-induced
translation leads to changes in dendritic spines that express LTD and that this
process is not properly limited in FXS.
Our most recent work on the regulation of mRNA granules in the
brains of mice lacking Fmr1 supports the idea that mGluR-induced translation
is exaggerated in FXS. Using sucrose gradient fractionation techniques to resolve
components of the translation machinery from brain lysates, we found that the levels
of heavy mRNA granules were significantly lower in the brains of mice lacking Fmr1
than in wild-type mice. Parallel imaging and in vitro studies (and work by others)
suggested that granules are composed of translationally silent polysomes, which
are released into a less dense fraction upon stimulation of translation.
Accordingly, we observed that in vivo administration of an antagonist of mGluR5
rapidly increased granules in the brains of mice lacking Fmr1 to levels matching
those in wild-type mice injected with the antagonist. These data indicate that ongoing
mGluR activity in brain leads to translation from, and reorganization of, mRNA granules
and that FMRP normally limits this process.
The fact that local translation is involved in stabilizing forms
of potentiation and depression implies that such translation is regulated differentially.
We are investigating a number of potential mechanisms for differential translation,
including heterogeneity of mRNA granules. Antibodies to mRNA-binding proteins found
in granules label distinct particles in dendrites, raising the possibility that
classes of granules exist that are used differentially during distinct patterns
of synaptic activity. Related to this finding, we determined that the RNA-binding
motif protein 3 is present in many components of the translation machinery, including
a subset of dendritic granules. The protein is distributed in dendrites and affects
the activity of a number of translation factors. A major goal for the coming year
is to further characterize mRNA granulebased mechanisms of differential translation
and their interrelationships with spine structure in wild-type mice and mice that