Formation of Long-Term Memories Requires Gene Expression
The mechanisms described above are especially important for short-term forms of plasticity that underlie short-term memories. While the formation of short-term plasticity, and short-term memory, have been shown to rely on modifications of preexisting proteins at the synapse, the formation of long-term memories differs in that it depends upon new gene expression. This can be thought of in the context of the different effects of extracellular stimulation that were discussed in Chapter 15: stimuli can produce short-term changes by altering the activity of preexisting enzymes and proteins in the cell, or long-term functional changes by altering the expression of genes in the cell (see Figure 15-1). Studies in many systems and species, including in rodent hippocampus, have demonstrated that LTP and LTD can be divided into transient forms of plasticity that do not require gene expression and Long-lasting forms (L-LTP and L-LTD), that require both mRNA and protein synthesis.
The extreme morphological polarity and compartmentalization of neurons adds significant challenges to stimulus-induced changes in gene expression. First, to turn on transcription, signals must be relayed to the nucleus from the synapse, which in many cases is located at great distances from the cell body. Neurons are specialized for rapid signaling between compartments by electrical signaling, and indeed, action potentials can trigger opening of voltage-gated Ca2+ channels in the cell body, and rapid signaling from the somatic plasma membrane to the nucleus. However, many studies have also shown that signaling molecules, including kinases, phosphatases, and transcriptional regulators, are actively transported from stimulated synapses to the nucleus to regulate transcription. In most cases, this long-distance retrograde transport has been shown to involve dynein motor protein–mediated transport along microtubules (as described in Chapter 18). How signaling is faithfully maintained during this long-distance transport in order to couple synaptic stimulation with gene expression is an area of active research.
The second great challenge in understanding the mechanism of stimulus-induced gene expression during synaptic plasticity derives from the fact that each neuron has a single nucleus and yet can form thousands of synapses. Long-lasting forms of synaptic plasticity are often “synapse-specific,” that is, they involve changes in synaptic strength at some but not all synapses formed by a single neuron. Since long-term plasticity requires transcription, synapse specificity begs the question of how gene expression can be spatially regulated in such a highly compartmentalized cell. One important mechanism involves the localization of mRNAs and their local translation in response to synaptic stimulation, as was discussed in Chapter 10. Indeed, L-LTP of hippocampal synapses has been shown to require the translation of mRNAs that are localized in dendrites and at synapses. Electron micrographic studies have identified polyribosomes, actively translating ribosomes, at the base of spines in hippocampal neurons, and have further shown that the number of spines controlling polyribosomes greatly increases after induction of L-LTP. Together, these studies have focused attention on the importance of post-transcriptional gene regulation in neurons during synaptic plasticity, and on a host of questions about mRNA localization and regulated translation: What mRNAs are localized to synapses? How are they localized? How is their translation regulated by synaptic activity? What is the specific function of the locally translated protein? Why are some mRNAs translated into protein in the neuronal cell body and then transported to synapses, and others translated directly at synapses?
Indicative of the importance of post-transcriptional gene regulation in the proper functioning of the nervous system, mutations in an RNA-binding protein, the fragile X mental retardation protein, FMRP, cause a common form of mental retardation, fragile X syndrome (FXS), and also constitute the most common single-gene cause of autism. The most common mutations that lead to FXS are expansions of CGG repeats in the FMRP gene that leads to gene methylation and silencing, as was described for Huntington’s disease in Chapter 6. FMRP is a translational repressor that binds target mRNA and prevents its translation. A population of FMRP localizes to the base of dendritic spines, where it is thought to maintain mRNAs in a dormant state until synaptic stimulation triggers their translation. Genetically modified mice that lack FMRP serve as remarkably good models for the human disease. The mice show deficits in learning, reflective of the intellectual disability in FXS patients. Both mice and humans have abnormalities in the structure of their synaptic spines, which are elongated, like immature spines, rather than stubby, like mature spines. Studies in the mouse have revealed excessive basal translation of mRNAs at synapses and have further shown alterations in protein-synthesis-dependent forms of hippocampal LTD. Together, these findings indicate that synaptic translation of localized mRNAs is critical to the formation and the experience-dependent plasticity of neural circuits, and that alterations in this process are a cause of neurodevelopmental and cognitive disorders.