Fragile X syndrome is a monogenic neurodevelopmental disorder resulting from a trinucleotide repeat expansion in the 5′ untranslated region of the FMR1
gene. Subsequent hypermethylation of this region leads to silencing of the gene and loss of its protein product FMRP (fragile X mental retardation protein). FMRP has been demonstrated to be an mRNA binding protein that negatively regulates a host of mRNA substrates, likely through mechanisms involving ribosome stalling and through association with microRNAs (Wang et al., 2012
). While the complete details of its role in regulating translation are still under investigation, elegant work has identified many of the target mRNAsthat are regulated by FMRP (Darnell et al., 2011
). FMRP can bind to a large number of mRNAs, but a significant proportion of these encode for synaptic proteins. Based upon this, and the many endophenotypes that have been established in Fmr1
ko mice, the predominant view has been that FMRP is localized to dendrites,close to spines and synapses,where it can rapidly regulate translation of synaptic proteins in an activity dependent manner. Additionally, there has been a large focus on FMRP-group1 mGluR interactions at the synapse (Waung and Huber, 2009
), further emphasizing the need to understand the postsynaptic roles of FMRP in regulating translation as a way of developing targeted therapeutics.
The study in this issue of Neuron by Deng et al uncovers a novel and unconventional role for FMRP in directly regulating the function of presynaptic ion channels in axons, that can ultimately regulate transmitter release(Deng et al., 2013
). This is not the first study to propose a role for FMRP beyond the postsynaptic density or spine. FMRP has been localized to growth cones of developing axons(Antar et al., 2006
) and has been proposed to have presynaptic roles in establishing synaptic connections (Christie et al., 2009
; Hanson and Madison, 2007
).Ultrastructural analysis of hippocampal CA1 synapses in Fmr1
ko mice has revealed a reduction in the length of active zones and a reduced density of docked vesicles in the terminals, all suggesting that FMRP plays a role in the formation of mature presynaptic terminals. Moreover,prior work from the authors of this present study hasprovided evidence of a functional alteration in presynaptic neurotransmitter release in mature synapses(Deng et al., 2011
). In that previous study, trains of stimuli delivered to activate Schaffer collateral synapses in the CA1 of the hippocampus produced greatly augmented responses in Fmr1
ko mice at stimulation frequencies above 20Hz, and most significantly during stimulation using a natural spike pattern (Deng et al., 2011
).This increase in transmitter release was attributed to elevated Ca2+
influx during train stimulation in synapses of the knockout mice, although it was not apparent how Ca2+
influx through voltage gated channels on the synaptic terminals might be enhanced when FMRP is ablated (Deng et al., 2011
In the current edition of Neuron, Deng et al follow up on these previous observations by performing a comprehensive and sophisticated group of experiments to identify the mechanisms by which FMRP might regulate neurotransmitter release(Deng et al., 2013
). They again use the hippocampal Schaffer collateral synapse, formed between the axons of CA3 neurons and CA1 pyramidal neurons, as their model system for understanding presynaptic roles of FMRP. Because the axons and terminals themselves are not accessible to electrophysiological recording, most of the measurements are made by recording from the somatic compartment of the presynaptic CA3 neuron. In initial experiments the authors demonstrate the width of the action potential (AP) broadens excessively in Fmr1
ko mice when the CA3 neuron spikes at high frequency. The duration of the AP in the axon and terminal is a critical determinant of neurotransmitter release, and therefore this observation is key in demonstrating the mechanisms underlying elevated synaptic augmentation during high frequency activity in the fragile X mouse model. However, it is important to resolve how FMRP acts, as many potential mechanisms could explain AP broadening in the knockout mice. To address the question of mechanism, the authors use disruption and reintroduction strategies to determine if they can reproduce or rescue the AP broadening phenotype observed in the knockout mouse. Introduction of the recombinant protein fragment of FMRP directly into neurons through the recording pipette fully reversed the excessive AP broadening in knockout neurons; whereas introduction of an antibody directed against FMRP into wildtype CA3 neurons reproduced the excessive AP broadening (presumably by functionally uncoupling endogenous FMRP from its signaling role). Taken together these experiments demonstrate that FMRP normally acts to limit AP broadening during high frequency AP trains, and these effects are translation-independent because protein synthesis inhibitors do not block them.
AP duration is largely determined by the activation of voltage gated potassium (K+) channels and there are a large number of possible channels that might mediate this effect in CA3 neurons.To isolate these channels, the authors first use a dynamic clamp technique to “replay” the AP waveform into neurons while blocking various K+conductances. Using broadly selective blockers they isolated a TEA (tetraethylammonium) sensitive component that was more suppressed during AP trains in knockout recordings. Further pharmacological isolation of this component clearly determined that this conductance was mediated by BK (big potassium) channels,voltage and Ca2+dependent K+ channels, identifying these as the primary target of FMRP regulation.
While it is a surprising finding that FMRP directly regulatesan integral membrane protein through a translation-independent pathway,there is prior precedent for FMRPregulating K+
channels. The Kaczmarek laboratory provided evidence that FMRP can bind to, and change the open probability of, a Na+
channel, Slack (Brown et al., 2010
; Zhang et al., 2012
). In the present study the authors next asked how FMRP might dynamically regulate BK channel function. These channels are both voltage and Ca2+
dependent, therefore the authors dissected each of these properties and demonstrated a specific effect on the Ca2+
sensitivity of the channel. BK channels are tetramers of the α pore-forming subunits, with β regulatory subunits that modify the channels properties including Ca2+
sensitivity. In CA3 neurons β4 is the predominant modulatory subunit, and co-immunoprecipitation experiments confirmed a putative interaction between FMRP and β4. Further confirmation of the a functional coupling between FMRP and the β4subunit were obtained using β4 ko mice, in which it was demonstrated that the effect of FMRP disrupting antibodies on AP broadening was lost, confirming that FMRP regulates AP duration by altering Ca2+
sensitivity of BK channels via interaction with the β4 regulatory subunit ().
Figure 1 Schematic of interaction between BK channels and FMRP. FMRP associates with the β4 regulatory subunit to enhance Ca2+ dependent activation of the channel during high frequency AP trains, increasing membrane repolarization and shortening AP width. (more ...)
Presynaptic action potential properties affect the degree of activation of presynaptic voltage gated Ca2+ channels, and thus regulate transmitter release. To test whether FMRP regulation of AP width did in fact alter Ca2+ dynamics, the authors directly measured presynaptic Ca2+in synaptic terminals using multiphoton imaging. Disrupting FMRP-BK interactions with an intracellular FMRP disrupting antibody, they demonstrated that during trains of stimulation elevated AP broadening resulted in enhanced Ca2+ influx into presynaptic terminals. Interestingly, presence of the antibody had no effect on the Ca2+ currents themselves, demonstrating a specific action of FMRP on BK channels.
These observations begin to dissect a novel and provocative mode of operation for FMRP in the brain, but the more interesting question is how this might modulate information transfer in the normal orFMRP null brain.Synapses that facilitate strongly during bursts of presynaptic activity can behave as high pass filters, allowing the transfer of information carried by high frequency spike bursts. Loss of FMRP control of AP width and subsequent elevation of short term plasticity removes the specificity of this filter and might allow transmission of information at frequencies that would normally be filtered out as background noise. This elevated transfer of noise will be detrimental to circuits that are required to transmit only salient information required for cognitive processes, and could be the basis of the intellectual and learning disabilities associated with fragile X syndrome.This is an interesting proposal but it is important to consider this in the broader context of other neuronal changes that might impact the fragile X brain. Is loss of FMRP regulation of BKchannels likely to be a primary basis for some of the behavioral phenotypes associated with the disorder when there are wholesale changes in protein translation, significant alterations in spine morphology and a multitude of effects on synaptic function and plasticity also present when FMRP is absent? These questions remain to be answered but a better understanding of the roles of native FMRP is a critical first step in determining how the brain is affected in fragile X, and whether targeting specific signaling pathways can reverse or alleviate some of the symptoms of the disease.