- Top of page
The Class IIa histone deacetylases (HDAC)4 and HDAC5 play a role in neuronal survival and behavioral adaptation in the CNS. Phosphorylation at 2/3 N-terminal sites promote their nuclear export. We investigated whether non-canonical signaling routes to Class IIa HDAC export exist because of their association with the co-repressor Silencing Mediator Of Retinoic And Thyroid Hormone Receptors (SMRT). We found that, while HDAC5 and HDAC4 mutants lacking their N-terminal phosphorylation sites (HDAC4MUT, HDAC5MUT) are constitutively nuclear, co-expression with SMRT renders them exportable by signals that trigger SMRT export, such as synaptic activity, HDAC inhibition, and Brain Derived Neurotrophic Factor (BDNF) signaling. We found that SMRT's repression domain 3 (RD3) is critical for co-shuttling of HDAC5MUT, consistent with the role for this domain in Class IIa HDAC association. In the context of BDNF signaling, we found that HDAC5WT, which was more cytoplasmic than HDAC5MUT, accumulated in the nucleus after BDNF treatment. However, co-expression of SMRT blocked BDNF-induced HDAC5WT import in a RD3-dependent manner. In effect, SMRT-mediated HDAC5WT export was opposing the BDNF-induced HDAC5 nuclear accumulation observed in SMRT's absence. Thus, SMRT's presence may render Class IIa HDACs exportable by a wider range of signals than those which simply promote direct phosphorylation.
Histone acetylation and deacetylation plays an important regulatory role in gene transcription, with acetylation strongly associated with increased transcriptional activity [for review (Haberland et al. 2009; Perissi et al. 2010, 2004)]. Acetylation/deacetylation of histones is a dynamic process that depends on the balance between activity of histone acetyltransferases (HAT) and histone deacetylases (HDAC). The HDACs are a large protein family named after their ability to remove acetyl groups from lysine residues located within the N-terminal tail of histones (Haberland et al. 2009; Perissi et al. 2010), but can act on non-histones as well. In mammals, there are five classes of HDACs (Class I, IIa, IIb, III, and IV) containing many different HDACs (18 in humans). Class IIa HDACs (HDAC4, 5, 7, 9), unlike other HDACs, are expressed in a tissue-specific manner with high expression levels in skeletal, cardiac, and smooth muscle, bone, the immune system, the vascular system, and the brain [reviewed in (Parra and Verdin 2010)]. Studies with knockout mice for various Class IIa HDACs show their important function in differentiation and developmental processes such as formation of skeletal muscle, cardiac hypertrophy, bone development and T-cell differentiation (Martin et al. 2007; Haberland et al. 2009; Parra and Verdin 2010). In the CNS, Class IIa HDACs controls neuronal survival (Linseman et al. 2003; Bolger and Yao 2005; Chen and Cepko 2009), differentiation (Schneider et al. 2008), long-term-memory-related synaptic plasticity (Guan et al. 2002), and behavioral responses (Tsankova et al. 2006; Renthal et al. 2007; Taniguchi et al. 2012).
A characteristic of Class IIa HDACs is that their ability to suppress transcription in the nucleus is subject to dynamic regulation. Class IIa HDACs are subject to phosphorylation on two or three conserved serine residues in their N-termini domain, and this modification exposes a nuclear export signal and creates a binding site for 14-3-3 proteins in the cytoplasm which anchor them there, thus preventing them from acting as transcriptional co-repressors (McKinsey et al. 2001; Martin et al. 2007; Parra and Verdin 2010). Mutation of these phosphorylation sites renders Class IIa HDACs ‘non-exportable’. The primary kinase family responsible for signal-dependent Class IIa phosphorylation and subsequent export are the Ca2+/calmodulin-dependent protein kinases (CaM kinases), although other kinases such as PKD and SIK are also capable of acting at these sites. In neurons, the major Class IIa HDACs expressed are HDAC4 and HDAC5, and these are subject to activity- and CaM kinase-dependent nucleo-cytoplasmic shuttling (Chawla et al. 2003). Moreover, CaM kinase-dependent nuclear exclusion of HDAC5 has been implicated in promoting the survival of neurons (Linseman et al. 2003). HDAC4, in contrast may play pro-survival or pro-death roles in the nucleus which may depend on the neuron's developmental stage (Bolger and Yao 2005; Majdzadeh et al. 2008). HDAC4 has been additionally shown to have an active pro-survival role in retinal neurons when resident in the cytoplasm (Chen and Cepko 2009). Furthermore, alterations to Class IIa activity mediate the actions of certain anti-depressants and control addiction pathways (Tsankova et al. 2006; Renthal et al. 2007; Taniguchi et al. 2012). Thus, signaling mechanisms that alter the subcellular localization of Class IIa HDACs have the capacity to influence important aspects of neuronal fate and function.
Class IIa HDACs can exist in the nucleus as part of multi-protein co-repressor complexes centred on SMRT or its close relative N-CoR, and containing other co-repressor proteins and Class I HDACs, particularly HDAC3 (Huang et al. 2000; Guenther et al. 2001; Perissi et al. 2004; Haberland et al. 2009; Watson et al. 2012). Indeed, one function of Class IIa HDACs is thought to be as a physical bridge between the SMRT/N-CoR complex and certain transcription factors such as myocyte enhancer factor 2 (MEF2), rather than contributing substantial HDAC activity as such (Fischle et al. 2002; Lahm et al. 2007). Interaction of Class IIa HDACs with SMRT is through SMRT's repression domain 3 (RD3) (Huang et al. 2000; Fischle et al. 2002). SMRT or N-CoR are known to undergo nuclear export in response to certain stimuli, including EGF-induced MEK1 signaling or Akt activation, and dissociation from chromatin in response to IκB kinase (Hong and Privalsky 2000; Hermanson et al. 2002; Hoberg et al. 2006; Perissi et al. 2004, 2010). In neurons, activity-dependent Ca2+ transients trigger SMRT export via a combination of MEK1 and CaM kinase pathways (McKenzie et al. 2005; Soriano et al. 2011). Other stimuli are also known to promote SMRT export, such as inhibition of Class I HDAC activity, specifically HDAC3 (Soriano and Hardingham 2011).
Thus, while CaM kinase signaling is capable of promoting the nuclear export of both Class IIa HDACs and SMRT, certain signaling pathways are selective for one or the other co-repressor. Given that SMRT and Class IIa HDAC family members associate with each other this raises the possibility, hitherto untested, that SMRT may be able to co-shuttle Class IIa HDACs out of the nucleus independent of the classical phosphorylation site-dependent mechanism. By studying the movement of ‘non-exportable’ phospho-site mutants of HDAC4 and 5, we find that SMRT, acting via its RD3 domain, is able to co-shuttle HDAC4/5 out of the nucleus in response to SMRT-exporting stimuli. Relevance of this pathway is shown in the context of BDNF signaling. We show that BDNF-induced MEK1 signaling promotes HDAC5 import, but that in the presence of SMRT this import is canceled out because of the promotion of SMRT export by MEK1 signaling which acts to co-shuttle HDAC5 back out of the nucleus via SMRT's RD3 domain.
- Top of page
Subcellular localization of Class IIa HDACs represents the major mechanism to regulate their function. Phosphorylation at two or three conserved Serine residues at the N-terminus leads to interaction with 14-3-3, nuclear export and derepression of the target genes. Here, we have shown that in neurons Class IIa HDACs can be exported from the nucleus in an independent manner to the classical phosphorylation mechanism, by co-shuttling with SMRT. As several signals are known to be able to influence the subcellular localization of SMRT and its close relative N-CoR, an implication of this work is that the presence of SMRT/N-CoR may render Class IIa HDACs responsive to those signals, particularly given that Class IIa HDACs can exist in SMRT/N-CoR-containing complexes in the nucleus. We have shown that BDNF-induced MEK1 signaling promotes SMRT export but has the reverse effect on HDAC5, promoting import. However, when SMRT is present BDNF-induced HDAC5 import is blocked – likely because of direct import being canceled out by SMRT-mediated export. The influence of SMRT is lost upon deletion of its HDAC4/5-interacting RD3 domain, consistent with a direct co-shuttling mechanism. Other signals may also impact on Class IIa HDAC localization in the presence of SMRT/N-CoR, such as the PI3-kinase/Akt pathway, which in response to CNTF treatment mediates the nuclear export of N-CoR in neural stem cells (Hermanson et al. 2002).
Although this is the first time that Class IIa HDAC export has been demonstrated totally independent of their signal-responsive phosphorylation sites, co-shuttling of one class member HDAC5, by another (HDAC4) has been shown (Backs et al. 2008). Unlike CaM kinases IV and I, which can promote export of both HDAC4 and 5, CaM kinase II can only promote HDAC4 export because HDAC5 lacks a CaM kinase II docking site (Backs et al. 2006). However, in the presence of HDAC4, HDAC5 can be co-shuttled out of the nucleus by CaM kinase II by interacting with HDAC4. Even ‘non-exportable’ HDAC5MUT can be co-shuttled in this way, suggestive of ‘piggy-pack’ export. However, a certain degree of export is also observed if ‘non-exportable’ HDAC4MUT is paired with wild-type HDAC5, indicating that docking of CaM kinase II on HDAC4 followed by trans-phosphorylation of HDAC5 can also lead to export (Backs et al. 2006). Importantly, when HDAC5MUT and HDAC4MUT are co-expressed there is no export in response to CaM kinase II activity: phosphorylation sites on either HDAC4 or HDAC5 are essential for the export to take place, unlike the situation we describe where export of HDAC5MUT or HDAC4MUT can be promoted by SMRT. Of note, Class IIa HDACs can themselves co-shuttle other proteins in response to CaM kinase signaling, through direct interaction. For example, both HDAC4 and HDAC5 physically associate with ankyrin-repeat proteins ANKRA2 and RFXANK. Phosphorylation of HDAC by CaM kinases promotes the nuclear export of the HDAC/ANKRA2 and HDAC/RFXANK complexes but not if the HDAC used is a non-exportable mutant (McKinsey et al. 2006).
To conclude, the prominent role that Class IIa HDACs play in controlling neuronal survival and death, and differentiation, as well as certain behavioral responses, such as addiction, make an understanding of signals that control their subcellular localization to be of direct importance. Outside of the CNS, the roles played by Class IIa HDACs in regulating transcriptional programs relating to skeletal myogenesis, cardiac hypertrophy, and thymocyte development (Lu et al. 2000b; McKinsey et al. 2000; Zhang et al. 2002; Haberland et al. 2009; Parra and Verdin 2010; Watson et al. 2012) raise the question as to whether these functions may be influenced by signals that promote SMRT-mediated co-shuttling independent of the classical export pathways.