The control of transcriptional corepressors by synaptic activity

Objectives
Gene expression induced by neuronal electrical activity controls a number of fundamental developmental and physiological processes in neurons, including fate determination, development, long-term synaptic plasticity and survival. The overall project objective is to characterise a novel mechanism by which gene expression can be regulated by synaptic activity: not via the standard route of regulation of the transcription factor, but by regulation of the inhibitory corepressor SMRT (silencing mediator of retinoic acid and thyroid hormone receptor), with which many transcription factors interact.

We recently showed (Mckenzie et al., 2005) that coincident synaptic activity potentiates the activation of several transcription factors that are repressed by SMRT. For example, in synaptically active neurons, gene expression mediated by the thyroid hormone receptor (TR) is far more effectively induced by low levels of thyroid hormone than in electrically silent neurons. We went on to provide evidence that this potentiation effect was due to disruption of TR-associated corepression, which is mediated via SMRT. Activity-dependent Ca2+ transients caused SMRT export from the nucleus and dissociation from TR. Other SMRT-repressed transcription factors such as CBF1 and the retinoic acid receptor respond to synaptic activity in a similar way.

SMRT-repressed transcription factors play roles in diverse processes such as neuroprotection (NF-kB, SRF) synaptic plasticity (NF-kB, SRF, CBF1) and in differentiation and development (CBF1, TR, RAR, Pit-1/Oct-1). Therefore electrical activity may influence a wide range of events via its effect on SMRT. The project objective is to understand the molecular mechanism and consequences of activity-dependent control of broad specificity corepressors, with particular focus on SMRT. Specifically, the objectives are:

1. Identify the direct activity-induced signalling events leading to SMRT export:
We will determine whether SMRT nuclear export relies on an intrinsic export signal, or if 'piggy-back' export is the more likely mechanism. We will also identify the region of SMRT that confers activity-dependent export and within this region we will determine whether direct phosphorylation of SMRT is responsible for triggering export, and which residues are targeted. Our recent work uncovered roles for CaM kinase IV and the Ras-MEK1-ERK1/2 pathway in triggering SMRT export, making these pathways prime candidates for mediating direct phosphorylation events.

2. Identify the contribution of activity-dependent class IIa HDAC export in promoting SMRT export:
In the nuclei of electrically silent neurons SMRT colocalises with the class IIa Histone Deacetylase, HDAC5, which is exported from the nucleus in response to synaptic activity in advance of SMRT's export. We have preliminary evidence that this prior export of HDAC5 plays a role in promoting the subsequent export of SMRT. We will directly address this question.

Objective 3:
Original objective: Understand the activity-dependent disruption of corepressor complexes on endogenous SMRT-repressed promoters.
Altered objective: Investigate the activity-dependent acetylation of SMRT repressed promoters, and activity-dependent activation of PGC1a.

As part of the work within this objective we have successfully studied activity-dependent histone acetylation on an endogenous SMRT-repressed promoter (the C/EBP-regulated Sestrin2), using chromatin immunoprecipitation (ChIP). However, despite intensive investigation, we have not been able to detect SMRT association with promoter elements by ChIP. Our collaborator and SMRT expert, Prof. Martin Privalsky has also had this problem. This problem could not be forseen, since the ChIP technique itself works well in the lab: we have successfully detected the presence of transcription factors (Papadia et al., 2008) as well as histone modification (this project). A possible explanation is that as corepressors do not bind directly to DNA, SMRT's association with the DNA, via a bridging transcription factor, is to weak and/or transient to withstand the ChIP protocol.
We will continue to study activity-dependent promoter derepression at the level of histone acetylation. We will not continue attempts to study SMRT interactions by ChIP, unless time allows at the end of the project. Instead we will investigate the activation of PGC-1a activation by synaptic activity (see above for background information on this). We will determine the mechanism by which PGC-1a transcription is induced, testing initially the working hypothesis that it is through activation of CREB (the PGC-1a promoter contains a CREB binding site). We will also investigate the mechanisms behind the post-translational activation of PGC-1a. Additionally we will determine the extent to which PGC-1a activation contributes to the potentiation of thyroid hormone activation of the TR, in addition to SMRT export.

Objective 4:
Original objective: Analyse the role of activity dependent SMRT export on a downstream physiological output: development of dendritic architecture:
Altered objective: Analyse the role of activity dependent SMRT export and PGC-1a activation on a downstream physiological output: regulation of antioxidant defenses.

Studies that we have performed to date have revealed that SMRT (and also PGC-1a) do not influence the development of dendritic architecture, which was the "downstream physiological output" originally selected to study. This was not something that could have been predicted at the outset of the project. Fortunately, however, we have found that SMRT and PGC-1a do regulate a different "downstream physiological output": neuronal antioxidant defenses (see above). We will change the "physiological output" that we study in this objective, from dendritic development to regulation of antioxidant defenses. We will characterise the antagonism between SMRT and PGC-1a with regard to negative and positive influence on antioxidant defenses. We will also determine the extent to which synaptic activity, in activating PGC-1a and exporting SMRT, can tip the balance in favour of antioxidant effects. We'd like to emphasize that the purpose of this objective was to study a downstream physiological output as a paradigm of the consequences of SMRT export. Despite changing the exact output under study, we remain true to the objective of the project in understanding the influence of synaptic activity on the corepressor-coactivator balance.

Nerve cells (neurons) communicate by releasing chemical messengers (neurotransmitters) onto each other at structures called synapses (synaptic activity). These messengers are detected by receptors on the cell surface, which then allow calcium and sodium ions to flow into the cell. This triggers the release of neurotransmitter onto other neurons. This means of “communication” is the way by which information flows round the brain.
However, "synaptic activity" also triggers changes inside neurons. The calcium ions which flow into the neuron trigger signals, which activate the transcription of genes, a process whereby genes (made of DNA and located in the nucleus) are “read” and decoded into new proteins. These proteins are crucial for many fundamental processes in the neuron such as learning and memory, and development from foetus, through infancy and on to adulthood. Equally importantly, these new proteins also make individual neurons healthier and more likely to survive for longer than neurons that don't experience synaptic activity. Understanding “survival signals” is important given the prevalence of neurodegenerative diseases in the aging human population. Our research has characterised new ways by which genes can be activated by synaptic activity, and the impact of this on neuronal survival.
The transcription of genes is suppressed in the nucleus by molecules called corepressors. One important corepressor is called SMRT which we found is “shipped out” of the nucleus into the cytoplasm by synaptic activity-dependent calcium influx. Once in the cytoplasm, SMRT is unable to suppress transcription because the genes are all in the nucleus and so these genes become much easier to activate. By chopping SMRT up and removing a small piece at time we have found that multiple SMRT regions are responsible for sensing calcium and responding by mediating export from the nucleus. SMRT “represses” genes by recruiting other repressor factors to the DNA called histone deacetylases (HDACs). HDACs work by changing the structure of the DNA into a more inactive, “closed” state. We found that an important class of HDACs (Class 2) are “dragged out” of the nucleus by SMRT in active neurons, representing a novel mechanism of export. We have also found that another class of HDACs (Class I) are actually needed to keep SMRT inside the nucleus: inhibiting these HDACs causes SMRT to export. Furthermore, we have shown that the action of these HDACs is important in controlling the activation of genes that we have found help protect neurons from oxidative stress due to harmful free radicals.
We have also studied the regulation by neuronal activity of a gene coactivator called PGC-1a. PGC-1a has the opposite effect to SMRT in that it activates, rather than represses, genes. It is of particular interest because it can protect neurons against oxidative stress and levels of it have been found to be reduced in the brains of patients with Huntington’s Disease, a neurodegenerative disorder. We have found that PGC-1a is controlled by synaptic activity at multiple levels. Firstly, the gene encoding PGC-1a is turned on by synaptic activity and secondly, PGC-1a protein itself is activated by a signal induced by synaptic activity. We have also found that SMRT is able to block the antioxidant, protective effect of PGC-1a in inactive neurons. However, synap

1. The transcriptional cofactors PGC-1α and SMRT are mutually antagonistic with respect not only to control of target genes, but also the regulation of neuronal antioxidant defences. Furthermore, the balance of this antagonism is controlled by neuronal activity through the export of SMRT and the phosphorylation (and activation) of PGC-1α.
2. SMRT can drive the export of Class-IIa HDACs (e.g. HDAC4/5) by a novel mechanism independent of the classical CaMKinase-dependent HDAC4/5 phosphorylation route. “Non-exportable” mutants of HDAC4/5 are exported from the nucleus in response to neuronal activity by a mechanism reliant on SMRT and dependent on SMRT’s HDAC4/5 interaction domain (RD3).
3. The nuclear localisation of SMRT is not reliant on Class-IIa HDAC activity but instead depends on Class-I HDAC activity (specifically HDAC2/3 activity). Inhibition of HDAC2/3 activity promotes SMRT export which is mediated by a region of SMRT known as the RD4 domain.