Class I HDACs

The HDAC domains of class I are homologous to the yeast protein Rpd3 (Taunton et al., 1996). In humans, the class I HDACs consist of HDAC1, 2, 3, and 8. HDAC 11 is also nominally a class I HDAC, but it is more divergent from the others and has some similarities with class II HDACs (Gao et al., 2002). Class I HDACs are significantly smaller than class II HDACs and are predominantly nuclear proteins, although a recent report has indicated that HDAC8 may be cytosolic in smooth muscle cells (Waltregny et al., 2004).

Like some HATs, the class I HDACs generally function as part of protein complexes. HDAC1 and 2

are in fact often found in the same complexes. The complexes include Sin3, NuRD, and CoREST (Heinzel et al., 1997; Xue et al., 1998; Humphrey et al., 2001). In addition to other components, the Sin3 and NuRD complexes each contain a "core complex" of HDAC 1/2 and the histone binding proteins RbAp46 and 48, which may help target HDACs to histones (Zhang et al., 1999). The NuRD complex also includes CH3 and 4 (Mi-2a and (3), which contain DNA helicase/ATPase domains found in the SWI/SNF chromatin remodeling proteins (Tong et al., 1998; Zhang et al., 1998b). This suggests that histone deacetylation and chromatin remodeling are coupled events. Further supporting this idea, the SWI/SNF complexes themselves have also been reported to be associated with class I HDACs (Zhang et al., 2000). Histone deacetylation may also be linked to DNA methylation, another epigenetic mechanism associated with gene repression, as proteins with methyl-CpG binding motifs are also a component of HDAC1 complexes (Zhang et al., 1999; Ng et al., 1999). However, the mechanistic details about the interplay of these proteins in the regulation of HDAC 1 function and in the repression of gene transcription remain unclear.

Similar to HDAC1, HDAC3 was found to reside in a large complex. In this case, endogenous HDAC3 is present in complexes with the corepressors N-CoR and SMRT that also contain several other components including transducin (beta)-like I (TBL1) protein (Yoon et al., 2003). Using RNA interference to specifically inactivate these components, it was demonstrated that the HDAC3-containing N-CoR and SMRT complexes are important for mediating the transcriptional repression activity of unliganded thyroid hormone receptor (Yoon et al., 2003). Interestingly, HDAC3 appears to be the only HDAC present in these complexes. Sin3, a co-repressor in HDAC1 complexes, is also not found in SMRT or N-CoR complexes. Therefore HDAC1 and HDAC3 clearly exist in different complexes and probably regulate chromatin and transcription through distinct mechanisms. Using RNAi knockdown approaches, it should be possible to delineate the specific functions of HDAC 1 and HDAC3 deacetylase complexes in gene transcription.

Recruitment of class I HDACs to their targets takes place through at least a couple of mechanisms. Usually, class I HDACs are recruited via other members of their repressor complex. For example, the Sin3 and N-CoR complexes can both bind unliganded nuclear hormone receptors (Heinzel et al., 1997; Horlein et al., 1995). The silencing repressors polycomb and hunchback can recruit NuRD components (Kehle et al., 1998). Other factors like Ikaros and REST can also recruit HDAC

complexes (Koipally et al., 1999; Naruse et al., 1999). Class I HDACs can also sometimes bind DNA-binding factors directly. This appears to be the case with HDAC1 and the transcription factor YY1 (Yang et al., 1996). This is more common for the class IIA HDACs, however (e.g. Miska et al., 1999 and see below).

B: Class IIHDACs

The HDAC domain of class II HDACs is homologous to the yeast protein Hdal (Grozinger et al.,

1999). In contrast to class I, class II HDACs are relatively large proteins (140kD or more) and have multiple domains. Unlike class I, which are mostly localized to the nucleus, many of the class II HDACs have a significant cytoplasmic component. Class II deacetylases can be further divided into two subfamilies, IIA and IIB.

Bl: Class IIA

The class IIA HDACs consist of HDAC4, 5, 7, and 9 (Verdin et al., 2003). This subfamily of HDACs shares several unique properties. First, they all contain the N-terminal non-catalytic MITR (MEF2-interacting transcription repressor) homology domain (Zhou et al.,

2000). This domain serves critical functions as both a protein-protein interaction domain and a regulatory domain subject to phosphorylation (Miska et al., 1999; Grozinger and Schreiber, 2000). Second, they are all critical regulators of MEF2, a family of transcription factors important in muscle differentiation and neuronal apoptosis (Miska et al., 1999). Third and most importantly, they are all regulated by phosphorylation-dependent subcellular trafficking (McKinsey et al, 2000).

All of the class Ila HDACs are subject to nuclear export via a nuclear export signal at the C-terminus (Wang and Yang, 2001; McKinsey et al., 2001). For example, HDAC4 is often found in the cytoplasm; however treatment with nuclear export inhibitor leptomycin B causes its nuclear accumulation (Zhao et al., 2001). The subcellular localization of these HDAC members is cell type-dependent and tightly regulated by specific signaling "events. For example, in C2C12 myoblasts, HDAC5 is normally a nuclear protein and it translocates from the nucleus to the cytoplasm upon differentiation, presumably to allow MEF2 to execute the transcription program for muscle differentiation (McKinsey et al., 2000). In contrast, HDAC5 and HDAC4 move from the cytoplasm into the nucleus in neurons in response to apoptotic stimuli (Bolger and Yao, 2005). Thus, the subcellular localization of HDAC4 and HDAC5 is dynamically regulated, and this principle has generally held true for HDAC7 and 9 as well.

It is now known that HDAC4 and HDAC5 intracellular trafficking is regulated by phosphorylation. Upon appropriate signaling, HDAC4 or 5 become phosphorylated. The calcium/calmodulin dependent kinases (CaMKs) and more recently, protein kinase D (PKD or PKCp) have been shown to be the kinases responsible (McKinsey et al., 2000; Zhao et al., 2001; Vega et al., 2004a). The phosphorylated HDAC then binds the phospho-binding protein 14-3-3 (Grozinger and Schreiber, 2000), and this interaction activates HDAC4 and HDAC5 nuclear export through the Crm-1 dependent machinery (Wang and Yang, 2001). It is thought that cytoplasmic HDAC4 and 5 are then sequestered by 14-3-3 until they are dephosphorylated. Thus, by regulating the nuclear export activity of HDAC4 and HDAC5, specific cell signaling events can then induce the transcriptional program that is normally controlled by HDAC4 and HDAC5.

The most well-characterized target for class IIA deacetylases is the myocyte enhancer factor 2 (MEF2) (Miska et al., 1999). MEF2 is one of the master activators of the regulatory cascade required for muscle differentiation, and HDAC4 and 5 were first shown to repress MEF2 in that context (McKinsey et al., 2000). MEF2 is also involved in several other biological pathways, such as cardiac hypertrophy and T cell apoptosis (reviewed in McKinsey et al., 2002), most of which have now been shown to involve one or more class IIA HDACs as well. The class Ila HDACs bind MEF2 directly and repress MEF2-dependent transcription (Miska et al., 1999). The transcriptional repression by HDAC4 family members on MEF2 has been naturally attributed to its ability to induce histone deacetylation. Indeed, deacetylase-deflcient HDAC4 or HDAC5 mutants cannot inhibit C2C12 muscle differentiation (Lu et al., 2000). However, recent studies reveal that HDAC4 and HDAC5 use additional mechanisms to repress transcription. For example HDAC4 and HDAC5 associate with the transcriptional co-repressor CtBP and the heterochromatin-associated protein HP1 (Zhang et al., 2001; Zhang et al., 2002b). HP1 is an essential gene for establishing transcriptional silencing that binds histone H3 methylated at lysine-9 (MethH3-K9), a common hallmark of heterochromatin (Bannister et al., 2001; Lachner et al., 2001). The interaction of HP1 and MethH3-K9 is apparently important for establishing the chromatin structure required for various type of gene silencing. The interaction of HP1 with HDAC4 as well as HDAC 11 (K. Wei and T.P.Y., unpublished result) suggests that in addition to catalyzing local histone deacetylation, HDACs can contribute to transcriptional repression through direct interaction with components of the histone methylation network (Zhang et al., 2002b). Indeed, pharmacological inhibition of HDACs by trichostatin A disrupted the spatial distribution of HP1 without affecting the total amounting of H3-K9 methylation (Maison et al., 2002).

Surprisingly, HDAC4-related deacetylases can also act to promote transcription factor sumoylation. In characterizing HDAC4, we have discovered that HDAC4 can function as a SUMO E3 ligase for MEF2 (Zhao et al., 2005). We and others have found that HDAC4 promotes MEF2 sumoylation and leads an apparent loss of MEF2 transcriptional activity (Gregoire and Yang, 2005). Thus the class IIA HDACs are able to repress transcription in at least three different ways: by deacetylating histones, by interacting with other corepressors and by inducing sumoylation of target activators.

Although they were initially identified as MEF2-interacting HDACs, the class IIA HDACs target multiple transcription activators for repression. HDAC4 and/or 5 repress and interact with Runx2, GATA-1, and SRF (serum response factor) (Vega et al., 2004b; Watamoto et al., 2003; Davis et al., 2003). It is thought that the recruitment of HDAC converts these transcriptional activators into repressors. Class IIA HDACs also interact with known DNA-binding transcriptional repressors including BCL6, and PLZF (Lemercier et al., 2002; Chauchereau et al., 2004). In these cases, HDACs function more as classic transcriptional co-repressors. Consistent with the idea of multiple mechansisms of action for class IIA HDACs, recent gene ablation and other functional studies of HDAC4, 5, 7 and 9 have revealed much broader roles for class IIA HDACs in various developmental processes as summarized below.

Mice with mutations in either HDAC5 or HDAC9 are viable and grossly normal. However, they are prone to developing cardiac hypertrophy (reviewed in Metzger, 2002). Hypertrophic growth of the heart muscle is a stress response that increases myocyte size and activates a fetal cardiac gene program dependent on MEF2 transcription factors. Loss of HDAC5 and HDAC9 lead to ectopic activation of this gene program and result in enlarged hearts that are acutely sensitive to hypertrophic stimuli (Zhang et al., 2002a; Chang et al., 2004). These observations indicate that HDAC5 and HDAC9 have overlapping functions and play a major role in preventing cardiac hypertrophy. The HDAC5/9 double knockouts are prone to lethal hemorrhages and display proportional hypertrophy as well (Chang et al., 2004). It is also of great interest to note that transgenic expression of a HDAC5 mutant that is resistant to phosphorylation-induced export causes lethality associated with severe defects in mitochondria (Czubryt et al., 2003). This result substantiates the importance of intracellular trafficking of HDAC5 and reveals a potential role for class IIA HDACs in metabolic regulation.

The HDAC4 knockout mouse has a different phenotype. This mouse displays significant skeletal defects (Vega et al., 2004b). These result from defects in chrondrocyte hypertrophy. In normal development, most bones form as a cartilage template first, then chondrocytes in the cartilage undergo hypertrophy, secreting a matrix that allows vascular invasion and subsequent ossification. HDAC4 represses the hypertrophy, at least in part by inhibiting Runx2-mediated transcription. Without HDAC4, the chondrocytes undergo early and improper hypertrophy and premature ossification, leading to the skeletal defects. HDAC4 has also been implicated in DNA damage response (Kao et al., 2003). It associates with p53-binding protein-1 (53BP1) and localizes to nuclear foci induced by DNA binding agents. HDAC4 can also be cleaved by caspases after DNA damage (Liu et al., 2004; Paroni et al., 2004); however, the physiological significance of this observation is not clear. Since the HAT Esal is also involved in DNA repair (Bird et al., 2002), these findings suggest both acetylation and deacetylation are required for chromatin remodeling during DNA repair.

While no HDAC7 knockout has been reported, HDAC7 has been implicated in T-cell maturation in the thymus (Dequiedt et al., 2003). Nur77 is a steroid receptor that induces apoptosis of T-cells during negative selection, and its expression is controlled by MEF2 (Youn et al., 1999). HDAC7 represses Nur77 transcription and inhibits the activity of MEF2. It thus plays a critical role in controlling negative selection. MEF2 has also been shown to be an important factor in neuronal survival (Mao et al., 1999). It functions to inhibit neuronal cell death, at least partly through activation of neurotrophin-3 (Shalizi et al., 2003). In fact, many of the pathways involved in muscle differentiation, like CaMK signaling, function similarly to promote neuronal survival. Therefore, it is likely that class IIA HDACs have a significant role in neuronal survival and death as well. This is supported by evidence that overexpression of HDAC5 can induce cell death in cerebellar granule neurons (Linseman et al., 2003). It will be interesting to see how HDACs are involved, given that HDAC4 is highly expressed in the brain (Grozinger et al., 1999) and that HDAC inhibitors are being explored as therapy for neurodegenerative diseases (see below).

B2 : Class lib

Class lib deacetylases include HDAC6 and HDAC10. They are characterized by a tandem repeat of complete (HDAC6) or partial (HDAC10) catalytic domains (Grozinger et al., 1999; Guardiola and Yao, 2002). HDAC6 and HDAC10 are also unique in their resistance to select deacetylase inhibitors, such as trapoxin B, which can potently inhibit the deacetylase activity of both class I and Ha HDACs (Guardiola and Yao, 2002). Unlike other HDAC members, HDAC6 does not appear to play a direct role in histone acetylation and transcriptional regulation. In fact, HDAC6 is almost always cytoplasmic and has functions other than histone acetylation and transcription (Hubbert et al., 2002; Kawaguchi et al., 2003). The HDAC6-specific inhibitor tubacin does not induce histone acetylation and has little effect on gene transcription (Haggarty et al, 2003). HDAC6, however, has potent tubulin deacetylase activity which may regulate microtubule function (Hubbert et al, 2002). HDAC6 also helps to clear misfolded protein aggregates from the cell, which may or may not be related to its tubulin deacetylase activity (Kawaguchi et al, 2003). The biological function of HDAC 10 remains unclear. However, it does have a nuclear component, and when attached to GAL4 DNA binding domain, it is capable of repressing transcription (Guardiola and Yao, 2002).

C: Class III HDACs

The prototype of the class III deacetylase is yeast SIR2 (silencing information regulator 2). Within the last ten years, Sir2 has gained notoriety as being essential for lifespan determination. In yeast, additional copies of this gene can enhance lifespan, while eliminating this gene altogether can reduce lifespan (Kaeberlein et al,

1999). Sir2, along with two other adaptor proteins, Sir3 and Sir4, can transcriptionally silence mating-type loci, telomeres, and rDNA, the last of which is thought to be the molecular basis for lifespan extension in yeast (Rine and Herskowitz, 1987; Smith and Boeke, 1997). The silencing of rDNA occurs via specific histone deacetylation events, which are thought to prevent the accumulation of extrachromosomal rDNA circles (ERCs), a molecular sign of ageing in a yeast mother cell (Sinclair and Guarente, 1997).

These processes can now be understood after SIR2 was shown to be a histone deacetylase (Imai et al,

2000). However, SIR2 and class I and II HDAC family members are structurally unrelated. Unlike HDACs,

SIR2 deacetylase activity uniquely requires NAD+, a cofactor whose levels are regulated by cellular respiration rates (Imai et al, 2000). The enzymatic reaction is completely different from that of HDAC-catalyzed deacetylation, and it produces a deacetylated peptide substrate as well as two reaction byproducts: nicotinamide, which can act to inhibit Sir2 activity through a negative feedback mechanism, and O-acetyl-ADP-ribose (OAAR), a novel metabolite whose function remains unknown (Tanner et al, 2000). The requirement of NAD+ for Sir2 activity suggests that class III HDACs could be regulated by changes in the cellular NAD+/NADH ratio (see below), providing a logical link to metabolic regulation and ageing processes.

Genome-wide searches for homologs in other species have been fruitful in identifying Sir2 homologs in bacteria, C.elegans, Drosophila, and higher eukaryotes including humans. Although Sir2-related proteins in worms and flies also seem to regulate the ageing process, the molecular mechanisms appear to be quite different and more complex than the yeast Sir2 protein (Guarente and Picard, 2005). There are seven class III HDACs in humans: SIRT1 to SIRT7. SIRT1 is the closest homolog to yeast Sir2 and is most well-characterized so far. It has been shown to repress a handful of different activators, and it interacts with several repressors including BCL6, as well as HES1 and HEY2, homologs of the Drosphila repressor Hairy (Bereshchenko et al, 2002; Takata and Ishikawa, 2003). This evidence supports the idea that SIRT1 can modulate histone acetylation directly. Surprisingly, however, SIRT1 also appears to regulate transcription factors themselves through direct deacetylation events. Most notably, SIRT1 can regulate the acetylation status of the tumor suppressor p53 (Vaziri et al, 2001; Luo et al, 2001). Acetylation of p53 correlates with p53-dependent transcriptional activity, while deacetylated p53 abrogates this transcriptional activity (Gu and Roeder, 1997). Under conditions favoring cell survival and increased lifespan, SIRT1 activity would deacetylate and inactivate p53, a master regulator of apoptosis. However, during harsh conditions or prolonged environmental stress, when p53 activity is required to mount an apoptotic response, SIRT1 may become inactivated leading to enhanced p53-dependent transcription (Smith, 2002).

SIRT1 appears to modulate a suite of additional transcription factors including the pro-survival factor, NF-kB (Yeung el al, 2004). Similar to p53, regulation of NF-kB occurs through direct acetylation of the RelA/p65 subunit of NF-kB. Deacetylation of this subunit is thought to prevent NF-KB-dependent transcription and therefore sensitize cells to apoptosis induced by TNFa. In this specific instance, it appears that SIRT1 possesses anti-proliferative capabilities and may instead act as a fine-tuning response regulator at the transcriptional level. It is unclear why SIRT1 would promote apoptosis through NF-KB-dependent pathways. The Foxo family of transcription factors are also involved in cellular decisions between survival and apoptosis during the cellular response to stress (i.e oxidative stress, heat shock) (Giannakou and Partridge,

2004). Accordingly, Foxol, Foxo3a and Foxo4 are all subject to reversible acetylation and SIRT1 appears to modulate their transcriptional activity, in this case favoring cellular survival and increased lifespan in the presence of cellular stress (van der Horst et al., 2004; Brunet et al., 2004; Yang et al., 2005). Lastly, SIRT1 also regulates transcription during muscle cell differentiation. Similarly to the class IIA HDACs, SIRT1 acts as a negative regulator of the muscle differentiation program (Fulco et al., 2003). SIRT1 has been reported to directly interact with PCAF in muscle cells and can deacetylate both PCAF and the muscle transcription activator MyoD. Furthermore, the NAD+/ NADH ratio decreases during muscle differentiation, possibly regulating SIRT1 activity.

In light of the profound effects of Sir2 on yeast lifespan, one may ask whether the mammalian Sir2 homologues have similar life-extending benefits. As stated above, SIRT1 appears to promote cell survival through direct histone interactions, or through regulation of transcription factor activity. But how might SIRT1 regulate organismal ageing in higher eukaryotes? Since SIRT1 is an NAD+-dependent deacetylase, it has the ability to couple the energy status of the cell to various cellular processes including the regulation of pro-and anti-survival factors mentioned above. As a NAD+-responsive protein, one might speculate that SIRT1 is regulated by nutritional cues at the whole-body level. Consistent with this hypothesis, calorie restriction (CR), which is known to increase longevity in many organisms, induces SIRT1 expression in mammalian cells (Cohen et al., 2004). It has been proposed that CR increases NAD+ levels and consequently enhances Sir2 activity; however, the connection between CR, Sir2, and the NAD+/NADH ratio is currently under debate (Anderson et al., 2003).

Recently, Rodgers et al. reported another link between SIRT1 and nutritional status (Rodgers et al.,

2005). SIRT1 was shown to interact with and regulate PGC-1 (PPARy co-activator 1), which is a master regulator of gluconeogenesis. As part of the response to fasting, hormonal cues impinge on the transcriptional machinery including RGC-1, and as a consequence PGC-1 transcriptionally activates the necessary gluconeogenic enzymes and precursors required for glucose production under starvation conditions. Remarkably, SIRT1 was shown to deacetylate PGC-1 and enhance its transcriptional ability under fasting conditions. These results support the idea that the Sir2 class of proteins can participate in nutritional responses and regulate whole-body metabolism and glucose homeostasis. The ability of this class of proteins to utilize NAD+ as a cofactor allows a fine-tuned response to nutritional conditions. Since energy status could potentially be a factor that dictates the ageing process, SIRT1 could represent a link to ageing in higher organisms (Rodgers et al., 2005; Nemoto et al., 2005). In light of this fact, there has been an effort to identify pharmacological tools that can activate the Sir2 class of proteins in order to harness their potential health benefits. In a search for Sir2 activators, it was shown that resveratrol, a phenol derivative found in red wine, is able to activate Sir2 and increase lifespan in yeast (Howitz et al., 2003). This study and follow-ups have shown that resveratrol and calorie restriction operate in the same pathway to promote lifespan extension, suggesting they both work through similar mechanisms involving Sir2 activation. However, more detailed structural analysis is required, however, to understand the molecular details behind this compound's potential anti-ageing effects.

Although roles for SIRT1 in ageing have begun to be elucidated, much less is known about its six mammalian homologs, SIRT2 to 7. Like HDAC6, SIRT2 has been reported to be a tubulin deacetylase, although the biological function of this is not known (North et al., 2003). SIRT2 also is regulated by the cell cycle, and the levels are dramatically increased during mitosis (Dryden et al., 2003). The SIRT3 protein is localized to the mitochondrial matrix (Onyango et al., 2002; Schwer et al., 2002). It appears to have a role in adaptive thermogenesis in brown adipocytes (Shi et al., 2005). Very little is known about the other members of this family, leaving open the possibility that other Sir2 family members can regulate transcription or other biological processes by coupling energy in the form of NAD+ to various cellular functions. Indeed, much remains unknown about this class of proteins including any potential roles in the ageing process, a field that is certain to receive more attention in the future.

Staying Young

Staying Young

Discover The Secrets To Staying Young Forever. Discover How To Hinder The Aging Process On Your Body And In Your Life. Do you feel left out when it comes to trying to look young and keeping up with other people your age? Do you feel as though your body has been run down like an old vehicle on its last legs? Those feelings that you have not only affect you physically, but they can also affect you mentally. Thats not good.

Get My Free Ebook

Post a comment