Mechanisms of dendritic mRNA transport and its role in synaptic tagging

Abstract

The localization of RNAs critically contributes to many important cellular processes in an organism, such as the establishment of polarity, asymmetric division and migration during development. Moreover, in the central nervous system, the local translation of mRNAs is thought to induce plastic changes that occur at synapses triggered by learning and memory. Here, we will critically review the physiological functions of well-established dendritically localized mRNAs and their associated factors, which together form ribonucleoprotein particles (RNPs). Second, we will discuss the life of a localized transcript from transcription in the nucleus to translation at the synapse and introduce the concept of the 'RNA signature' that is characteristic for each transcript. Finally, we present the 'sushi belt model' of how localized RNAs within neuronal RNPs may dynamically patrol multiple synapses rather than being anchored at a single synapse. This new model integrates our current understanding of synaptic function ranging from synaptic tagging and capture to functional and structural reorganization of the synapse upon learning and memory.

Figure 1

RNA signature. mRNAs—as depicted in the centre of the figure—contain multiple regulatory elements at both the primary sequence and structural levels that are recognized by certain trans-acting factors or RBPs. These can bind either single-stranded (simple line) or double-stranded (hairpin structures) RNA. In addition, some of them can also interact with other proteins as shown on the right. Consequently, each transcript has its own unique ‘RNA signature’, which determines the fate and function of an mRNA including its stability, localization, translational control and whether it undergoes local protein synthesis (functions listed in the outer panels). As every mRNA can contain various combinations of regulatory elements and bind different RBPs, this not only increases the functionality of a transcript, but also offers unique ways to regulate the fate of the RNA. ARE, AU-rich element; Cap, 7-methylguanosine; CPE, cytoplasmic polyadenylation element; CPEB, cytoplasmic polyadenylation element-binding protein; EJC, exon junction complex; LE, localization element; miRNA, microRNA bound to its miRNA-binding site; mRNPs, messenger RNA-containing RNPs; TLC, translation control element.

Figure 2

Sushi belt model. RNAs are transcribed in the nucleus of a neuron where they are processed. Next, the EJC is deposited on some intron–exon boundaries before the RNA is exported to the cytoplasm (see inset 1). Once in the cytoplasm—termed cell body of the neuron—the RNP (labelled G for granule) undergoes remodelling and is then transported along the MT cytoskeleton (depicted as a running sushi conveyor belt in red) bidirectionally within dendrites. Such a neuronal RNA granule or RNP typically consists of the transcripts, bound RBPs, adaptor proteins (A) and a molecular motor (M) (see inset 2). In this new model of RNA localization to the synapse termed the ‘sushi belt model’, we propose that neuronal RNPs routinely patrol a group of synapses in dendrites like a circling conveyor belt. If a particular synapse becomes activated (represented by inset 3), it may recruit dynamic MTs that now extend into dendritic spines allowing specific delivery of RNPs. This may contribute to a process termed synaptic tagging and capture, initially described by Frey and Morris (1997). Please note that the membrane structure in the neck of the dendritic spine is the spine apparatus, a specialized form of endoplasmic reticulum (inset 3). Ultimately, those synapses upon repetitive activation protocols undergo structural and functional rearrangement. Once delivered to those activated synapses, the RNA is released from the RNP and translated. In many cases, only one subunit of an active protein complex is synthesized locally, whereas the others are made in the cell body and delivered to the synapse by conventional protein transport. One prominent example is the active enzyme CaMKII: the α-subunit (blue semicircle) is made locally where it associates with the β-subunit (green semicircle) that is made and transported from the cell body. Finally, the active enzyme is anchored at the cytoskeleton and/or at the postsynaptic density (grey semiovals underneath the postsynaptic membrane). It should be noted that, while CaMKII is perhaps the best-studied example, other factors may operate similarly in synaptic tagging, and other mechanisms may exist to tag a synapse.