Fragile X syndrome: (What's) lost in translation?

The changing nature of synaptic function in neuronal dendrites during development is one of a myriad of phenomenological observations that have fascinated neurobiologists for some time. In the case of fragile X syndrome (FXS), neurons exhibit long, spindly dendritic spines that appear to be morphologically immature (1). As with all dendritic spines, these FXR spines typically house elements of the postsynaptic density juxtaposed with components of the translation apparatus (2). A dominant mutation in the FMR1 gene gives rise to FXS through the elimination of FMR1 RNA. This gene encodes the fragile X mental retardation protein (FMRP), which is a RNA-binding protein. By using in vitro assays, FMRP has been proposed to be involved in modulating translation through its binding to cellular RNAs and ribosomes (3). In early studies, Weiler et al. (4) showed that the FMR1 RNA was localized in dendrites and that its translation could be modulated by group 1 metabotropic glutamate receptor (Gp1 mGluR) agonists. Subsequent studies have extended, and begun to clarify, the potential aftereffects of such events and their impact on normal cellular functioning, as well as FXS.

In this issue of PNAS, Weiler et al. (5) have taken an important step in showing the complexity of FMRP-mediated translational control. By using cortical synaptoneurosome preparations, Weiler et al. (5) demonstrate that Gp1 mGluR (i.e., mGluR1 and mGluR5) activation has no effect on the assembly of polyribosomes in Fmr1 knockout (KO) mice. Whereas the presence of FMRP promotes polyribosome formation in this article, there has been some debate as to whether FMRP directly associates with polyribosome messenger ribonucleoproteins (RNPs) or slower sedimenting, translationally repressed RNPs in mice (6). More recent reports (7, 8) have confirmed a direct interaction that is highly sensitive to the types of detergents used in the assay. In the Weiler et al. article (5), Gp1 mGluR activation then results in a rapid initiation of translation only in WT mice. The lack of response seen in Fmr1 KO mice is not due to changes in the number of Gp1 mGluRs or the abundance of PKC, one of the downstream effectors of Gp1 mGluR-mediated Gq activation. Previous reports (9) have noted that the same Gp1 mGluR agonist used in this article regulates the trafficking of FMRP granules into the dendritic arbor. The Weiler et al. (5) data suggest that a unitary signal distal to PKC can mobilize different sets of responses (i.e., trafficking vs. polyribosome formation and translation initiation). The “distal-to-PKC” response is important in the context of specificity of the functioning of FMRP. PKC activation serves as a nexus for multiple signaling pathways, and not just for the Gp1 mGluRs. Consequently, the Weiler et al. (5) data suggest that receptors other than Gp1 mGluRs that use PKC as a signaling molecule may likewise use FMRP to modulate mRNA translation.

Current models of the Gp1 mGluR impact on FXS suggest that the cognitive impairments and developmental delays are the result of slowed synaptic maturation by means of FMRP translational repression. Bear et al. (10), in a recent review, highlighted a number of the FXS behavioral responses that are coincident with Gp1 mGluR activity. It should be noted that the context of synaptic or nonsynaptic Gp1 mGluR activation is likely to be an important issue. In the NMDA receptor systems, different signaling cascades and outcomes have been reported, depending on the subcellular origin of the signal (11). Results from our laboratory (12) and from the Warren (13) and Darnell (14) laboratories have uncovered a large set of mRNA cargoes that interact with FMRP. Interactions with microRNAs (15) and nontranslatable RNA polymerase III transcripts (6) have been identified as well. Given that FMRP modulates translation, it is reasonable to ask whether it does so through altering the translation of these mRNA cargoes or through a general action at the level of ribosome functioning. The in vitro translation studies show translational inhibition by using high levels of FMRP protein (16, 17). However, it is unclear whether in vivo FMRP works only as a translational repressor. Translation in dendrites is quite complex, with translational suppression or enhancement of a reporter mRNA being dictated by subcellular localization. Some subdendritic translational hotspots result in a decrease in reporter RNA translation, whereas others in the same dendrite for the same RNA result in an increase in reporter RNA translation (18). It is noteworthy that Weiler et al. (5), Todd et al. (19), and Miyashiro et al. (12) see either enhancement and/or suppression of translation of selected RNAs in response to the presence or absence of FMRP protein. These data suggest a complex regulatory role for FMRP in modulating synaptic RNA localization and translation that is likely spatially dictated. Such a model would incorporate the differences in subcellular localization of effectors that modulate PKC activity as well as PKC itself. Is it possible that FMRP may act in a manner similar to some of the Y-box proteins in their ability to stimulate or inhibit translation based on differing concentrations of and affinities for mRNAs that bind?

FMRP modulation of synaptic RNA localization and translation is likely spatially regulated.

Weiler et al. (5) take us beyond a “simple” Gp1 mGluR model for FXS. Whereas dysfunction of the Gp1 mGluR system is likely an important consequence of FMRP loss, the most recent data suggest that other receptor systems that modulate PKC activity warrant investigation of the role FMRP may play in modulating activity (Fig. 1). It is through these types of detailed molecular analysis that the symphony of FMRP-mediated biologies, including local protein synthesis, will be heard and appreciated.

Fig. 1.

A schematic of the Gp1 mGluR/FMRP pathway. Depicted is the Gp1 mGluR receptor coupled to the Gq-mediated signaling cascade. Gq signaling liberates DAG and IP3 through phospholipase C. G protein-coupled receptor kinase (GRK) 2/3 is known to antagonize this signaling pathway (20), and at least one GRK, GRK 4 mRNA, is a known FMRP cargo and is localized to dendrites (12). FMRP modulation of activity-dependent translation is downstream of PKC activation. The black box represents the unknown intermediates in this step. FMRP directly interacts with polyribosomes, and phosphorylation is thought to be one requirement in this step (3, 21). FMRP, in a messenger ribonucleoprotein complex containing FXR1P among other RNA-binding proteins, is known to bind a large number of mRNA cargoes where it may inhibit the translation of some or enhance the translation of others. One complicating factor at the synapse is that a number of receptor systems are known to modulate production of PKC, suggesting that multiple neuromodulatory receptors may be impacted by FMRP activity in a manner analogous to the Gp1 mGluR system.