The Rpd3/Hda1 Family of Lysine Deacetylases: From Bacteria and Yeast To Mice and Men

REVIEWS
The Rpd3/Hda1 family of lysine

deacetylases: from bacteria and
yeast to mice and men
Xiang-Jiao Yang* and Edward Seto‡
Abstract | Protein lysine deacetylases have a pivotal role in numerous biological processes
and can be divided into the Rpd3/Hda1 and sirtuin families, each having members in diverse
organisms including prokaryotes. In vertebrates, the Rpd3/Hda1 family contains 11 members,
traditionally referred to as histone deacetylases (HDAC)1–11, which are further grouped
into classes I, II and IV. Whereas most class I HDACs are subunits of multiprotein nuclear
complexes that are crucial for transcriptional repression and epigenetic landscaping,
class II members regulate cytoplasmic processes or function as signal transducers that
shuttle between the cytoplasm and the nucleus. Little is known about class IV HDAC11,
although its evolutionary conservation implies a fundamental role in various organisms.
Orthologue Acetylation of the ε‑amino group of Lys residues was ini‑ to phylogenetic analysis and sequence homology14–16.
A functionally related gene in tially identified on histones four decades ago1 and is now In this classification system, the sirtuin family constitutes
two or more species that has known to occur in different organisms for the regulation class III, whereas the classical or Rpd3/Hda1 family
evolved from a common
of diverse cellular processes (BOX 1). This modification is is grouped into classes I, II and IV (FIG. 1), with class II
ancestor.
reversed by Lys deacetylases (KDACs; BOX 1). Because being further divided into two subclasses (IIa and IIb).
Nucleosome histones were first identified as the substrates in 1969 Such subclass division is also necessary to accommodate
The basic structural subunit of (ref. 2), eukaryotic KDACs are often referred to as histone classification of members from Caenorhabditis elegans,
chromatin, which consists of deacetylases (HDACs). In 1978, butyrate was discovered Drosophila melanogaster and zebrafish (Supplementary
~200 base pairs of DNA
wrapped around an octamer of
to be a nonspecific HDAC inhibitor3–5, and the fungal information S1 (figure)).
histones. antibiotic trichostatin A was identified in 1990 as the first The HDAC field has expanded at an amazing pace,
specific inhibitor6. Using a similar inhibitor as an affinity owing largely to the following three factors. First, chrom
Homologue tag, in 1996 Taunton et al.7 purified the first HDAC atin research has exploded over the past 10–15 years and
A gene (or its product) that is
(now known as HDAC1) from cow protein extracts and HDACs have a key role in deacetylating nucleosomes,
descended from a common
ancestral gene.
discovered it to be an orthologue of yeast Rpd3 (reduced interacting with other chromatin regulators and shaping
potassium dependency-3; FIG. 1a). Because yeast genetics epigenetic landscapes17,18. Second, in addition to histones,
had already established Rpd3 as a global gene regulator8, these enzymes deacetylate numerous transcription factors
the unexpected sequence similarity strengthened the and many other proteins that are important for a multi‑
*Molecular Oncology Group, link between histone acetylation and gene regulation. tude of cellular processes in eukaryotic and prokaryotic
Department of Medicine,
McGill University Health
Two independent studies confirmed the multiplicity systems (BOXes 1,2). Finally, HDAC inhibitors have great
Center; and McGill Cancer of HDACs9,10 and, together with the identification of potential for treating cancer and other diseases19,20. Most
Center, McGill University, HDAC1, led to the notion that these enzymes can func‑ HDAC inhibitory compounds that are currently under
Montréal, Québec, tion as transcriptional corepressors and spurred the clinical evaluation target the Rpd3/Hda1 family.
H3A 1A1, Canada.
search for additional members of the Rpd3/Hda1 (histone This review covers recent findings relating to the Rpd3/
‡
Molecular Oncology
Program, H. Lee Moffitt deacetylase-1) family11 (FIG. 1; see also Supplementary Hda1 family of deacetylases. We first discuss the domain
Cancer Center and Research information S1 (figure)). organization and three-dimensional structure of the mam‑
Institute, Tampa, In 2000, the Sir2 (silent information regulator-2) or malian members and compare this organization to that of
Florida 33612, USA. sirtuin (Sir2-like protein) family of NAD+-dependent homologues in yeast and other model organisms. We then
e-mails:
[email protected];
deacetylases was discovered12, which led to the desig‑ describe multisubunit complexes of class I and II HDACs,
[email protected] nation of the zinc-dependent Rpd3/Hda1 family as the roles of class IIa members in signal transduction, and
doi:10.1038/nrm2346 ‘classical family’13. HDACs can also be grouped according cytoplasmic functions of the class IIb member HDAC6.
206 | march 2008 | volume 9 www.nature.com/reviews/molcellbio

© 2008 Nature Publishing Group
REVIEWS
Box 1 | Nε-acetylation as a common post-translational modification Mammalian HDAC1 and HDAC2 are almost identi‑
cal, and in addition to homologous catalytic domains,
Unlike Nα-acetylation, Bromodomain As a functional group they share homology in the C‑terminal tail, which har‑
Nε-acetylation is reversible for catalysis, interaction bours tandem casein kinase-2 (CK2) phosphorylation
and is dynamically H O and so on
sites22. The tail of HDAC1 also contains a sumoylation
controlled by opposing +
actions of Lys
εN C CH3 NH3 motif22. There are three orthologues of HDAC1/2 in
δ CH CH2 Methylation C. elegans, but only one in D. melanogaster or zebrafish
acetyltransferases (KATs) 2 Ubiquitylation
and Lys deacetylases γ CH CH2 Sumoylation (Supplementary information S1 (figure)). Unlike
2
(KDACs) (see figure). This Other modifications HDAC1, HDAC3 is only conserved from flies to mam‑
β CH CH
modification has been 2 2 mals (FIG. 1 and Supplementary information S1 (figure)).
extensively studied in the α CH
2
KDAC
CH2 HDAC3 differs slightly from HDAC1 and HDAC2 in that
past decade. It is now CH CH it has only one CK2 phosphorylation site23. HDAC8 is
known that Nε‑acetylation KAT
Ac-Lys Lys present in the sea urchin Strongylocentrotus purpuratus,
not only modifies histones but not in C. elegans or D. melanogaster, suggesting
and transcription factors (>60 identified to date), but also targets proteins that are
Nature Reviews | Molecular Cell Biology that this deacetylase is deuterostome specific. Instead
involved in DNA repair and replication, metabolism, cytoskeletal dynamics, apoptosis,
protein folding and cellular signalling 130–132
. In particular, many mitochondrial proteins
of CK2 sites, this deacetylase has a conserved motif
appear to be acetylated124,133 (see also BOX 2). This modification even occurs in bacteria (Supplementary information S2a (figure)) for protein
(BOX 2), which, in evolutionary terms, suggests that it may precede protein kinase A phosphorylation and negative regulation of
phosphorylation. deacetylase activity24,25.
Nε-acetylation results in gain-of-function and loss-of-function effects. In the case of
gain-of-function effects, acetyl-Lys-containing regions may be recognized by Domains and motifs of class II HDACs. Hda1 from bud‑
bromodomains in a way that is conceptually similar to the recognition of ding yeast consists of an N‑terminal deacetylase domain
phosphopeptides by various protein modules134. The loss-of-function modification and a long C‑terminal extension of unknown function
prevents the use of the amino group for different functional purposes. For example, (FIG. 1b) . This domain organization is conserved in
acetylation of a key Lys residue inactivates acetyl CoA synthetase (BOX 2). Acetylation
cryptic locus regulator-3 (Clr3), a class II member from
also blocks the amino group from undergoing other modifications such as methylation,
ubiquitylation and sumoylation130. Furthermore, acetylation and neighbouring
S. pombe. By contrast, only the deacetylase domain is
modifications may counteract or cooperate with one another130. shared by metazoan homologues14,15,26 (FIG. 1b). In addi‑
tion to the deacetylase domain, a long N‑terminal exten‑
sion is conserved among the class IIa members (HDAC4,
-5, -7 and -9) in mammals. This N‑terminal extension
Finally, we discuss how these molecular and cellular contains binding sites for myocyte enhancer factor-2
studies are related to genetic analyses in mice, C. elegans (MEF2) and 14‑3‑3 proteins. The MEF2-binding motif is
and D. melanogaster, as well as the implications for conserved from C. elegans to mammals26 (Supplementary
pathology and drug development. Two themes link information S1 (figure)) and has a resemblance to sites
the various sections of this review: first, how sequence in other MEF2-binding proteins27. Only one or two of
homology among deacetylases predicts their structure, the 14‑3-3 binding sites are conserved in C. elegans and
Bromodomain
A conserved protein module function and regulation; and, second, how a simple D. melanogaster (Supplementary information S1 (figure)),
that was first described for the catalytic domain co-evolves with different motifs and but three are present in an orthologue from the sea urchin
Drosophila melanogaster modules to connect functions in diverse cellular pro (Supplementary information S2b (figure)). Consistent
homeotic gene regulator cesses from various organisms. Both themes represent with the sequence conservation, MEF2 and 14‑3-3 are
Brahma (means ‘creator’ in
Hindu). This domain is present
fundamental issues that are relevant to many other major binding partners of class IIa HDACs.
in many gene and/or chromatin protein families. HDAC3 appears to be responsible for the deacetylase
regulators and has the ability activities that are associated with HDAC4 and HDAC7
to recognize acetyl-Lys motifs. Domain organization of Rpd3/Hda1 HDACs (ref. 28). Consistent with this, mammalian class IIa mem‑
In humans, class I of the Rpd3/Hda1 family comprises bers have low deacetylase activity as a result of a Tyr→
Deuterostome
An animal in which the anus HDAC1, -2, -3 and -8, class IIa members include HDAC4, His substitution29. This substitution occurs in zebrafish,
develops from the first opening -5, -7 and -9, whereas class IIb members are HDAC6 but not in C. elegans, D. melanogaster or the sea urchin
of the embryo, with the mouth and -10 (FIG. 1). While members of class I are homologous (Supplementary information S2b (figure)), which implies
being formed later. These to yeast Rpd3 and class II HDACs are more related to that it is vertebrate specific. Interestingly, an equivalent
include echinoderms and
chordates.
yeast Hda1, HDAC11 shows similar homology to both Tyr residue is not essential for the esterase activity of an
Rpd3 and Hda1 and is thus grouped into a separate class HDAC-like bacterial protein30. It will be important to
Myocyte enhancer factor-2 (class IV). determine how the substitution is reflected with regard
(MEF2). An evolutionarily to the function and regulation of class IIa deacetylases
conserved transcription factor,
Domains and motifs of class I HDACs. As a founding in vertebrates.
the specific DNA-binding
activity of which is mediated by member of class I, Rpd3 from budding yeast comprises Unlike class IIa members, HDAC6 contains tan‑
an N‑terminal MADS box an N‑terminal deacetylase domain and a C‑terminal dem deacetylase domains and a C‑terminal zinc finger
and an adjacent MEF2-specific tail 8,10 (FIG. 1a) . Both regions are conserved in Clr6 (FIG. 1b and Supplementary information S1 (figure)), and
domain. Although initially (cryptic locus regulator-6) from Schizosaccharomyces has emerged as a major cytoplasmic deacetylase. The
identified as a muscle-specific
transcription factor, different
pombe21, but only the catalytic domain is conserved zinc finger can bind ubiquitin, and its core sequence is
MEF2 isoforms are important in the bacterial deacetylase AcuC (BOX 2) and in class I similar to ZnF-UBP (ubiquitin C-terminal hydrolase-like
in various tissues. members from higher eukaryotes13. zinc finger)26,31. An SE14-repeat domain that is found in
nature reviews | molecular cell biology volume 9 | march 2008 | 207

REVIEWS
a Class I Rpd3 DAC 433 (S. cerevisiae)
Clr6 66/81% 405 (S. pombe)
HDAC1 64/81% S S 482

421 423
HDAC2 65/82% S S 488
422 424 (H. sapiens)
HDAC3 57/76% S 428
424
HDAC8 S 43/65% 377
39
b Class II Hda1 DAC (26/47%) 706 (S. cerevisiae)
Clr3 53/71% 687 (S. pombe)
MEF2 14-3-3 14-3-3 14-3-3

HDAC4 S S S 36/52% 1,084
246 467 632
HDAC5 S S S 34/53% 1,122 (H. sapiens)
class IIa
HDAC7 S S S S 37/51% 912
HDAC9 S S S 35/51% 1,069
HDAC6 38/60% 44/61% SE14 ZnF 1,215 (H. sapiens)

class IIb
HDAC10 41/65% Leucine-rich 669
c Class IV HDAC11 24/42% (31/41%) 347 (H. sapiens)
Figure 1 | Domain organization of classical HDACs from yeast and humans. The histone deacetylases (HDACs) are
grouped into different classes according to sequence similarity to yeast prototypes. In Nature Reviews
mammals, | Molecular
class Cell Biology
I (Rpd3-like)
members include HDAC1, -2, -3 and -8 (a); class II members are HDAC4, -5, -6, -7, -9 and -10 (b), which are related to yeast
Hda1. Class IV consists of HDAC11 (c). Class II is further divided into two subclasses: IIa (HDAC4, -5, -7 and -9) and IIb
(HDAC6 and -10). The total number of amino acid residues in each deacetylase is shown on the right. Many of the
deacetylases have isoforms that result from alternative splicing; for simplicity, the number of amino acids refers to the
longest isoform. The deacetylase (DAC) domain is depicted as an orange cylinder, and the percentage amino acid
sequence identity/similarity to that of Rpd3 (for class I) or Hda1 (class II/IV) is shown. The sequence identity/similarity of
Hda1 and HDAC11 to Rpd3 is given in brackets. The C‑terminal domains (grey) of Hda1 and Clr3 are homologous
(identity/similarity: 26/57%). Thick black lines represent similar N‑terminal domains and C‑terminal tails of class IIa HDACs.
Myocyte enhancer factor-2 (MEF2)-binding motifs are depicted as short green cylinders, whereas 14‑3-3 binding motifs
are shown as short blue cylinders labelled with ‘S’ (for Ser). Clr3, cryptic locus regulator-3; Hda1, histone deacetylase-1;
H. sapiens, Homo sapiens; Rpd3, reduced potassium dependency-3; S. cerevisiae, Saccharomyces cerevisiae; SE14, Ser-Glu-
containing tetradecapeptide repeats; S. pombe, Schizosaccharomyces pombe; ZnF, ubiquitin-binding zinc finger.
human HDAC6 (ref. 26) is almost identical in macaques in bacteria16 and plants (Arabidopsis thaliana HDA2).
(GenBank accession number XP_001101619) and HDAC11 is sometimes referred to as a class I member,
highly conserved in horses (GenBank accession number but phylogenetic analysis indicates that this deacetylase
XP_001493915), but is only partial in mice, implying that and its homologues belong to a separate class16. Little is
14‑3-3 protein this domain is specific to higher mammals. known about its function or regulation, but its evolution‑
One of a family of acidic The other class IIb member, HDAC10, has an ary conservation implies that it has a fundamental role in
proteins that are conserved N‑terminal half that is highly similar to the first deacety‑ diverse organisms.
from yeast and plants to
humans and are highly
lase domain of HDAC6, whereas the C‑terminal half is
abundant. They are ~30 kDa Leu rich (FIG. 1b). Although there are no invertebrate ortho‑ Structures of Zn2+-dependent deacetylase domains. Crystal
in size and often recognize logues, a distant relative exists in zebrafish (Supplementary structures have been solved for the catalytic domains of
target proteins with the information S2c (figure)), which indicates that HDAC10 class I member HDAC8 (Refs 25,33), class II member
sequence motif RXXpS/TXP or
might have a vertebrate-specific role. However, little is HDAC7 (Protein Data Bank ID 2NVR) and two bacte‑
RXXXpS/TXP, where X is any
residue and pS/T denotes known about its actual biological function. rial deacetylases34,35. These four structures are extremely
phosphoSer or phosphoThr. similar and comprise a single α/β domain composed of
HDAC11 as a highly conserved deacetylase. The class IV an eight-stranded β‑sheet that is sandwiched between
ZnF-UBP member HDAC11 shows sequence similarity to class I and >12 α‑helices (also see Supplementary information S3
A ubiquitin-binding zinc finger
that is present in HDAC6 and
II HDACs32 (FIG. 1c). It is highly conserved from C. elegans (figure)). The catalytic centre contains a zinc ion that is
several ubiquitin-specific and D. melanogaster to humans (Supplementary infor‑ chelated by side chains of His and Asp residues. These
proteases. mation S1 (figure)), and there are also related proteins structures indicate that the deacetylase domains from

REVIEWS
Box 2 | Acetylation inhibits acetyl CoA synthetase from bacteria to humans but also contains two unique subunits, Eaf3 and Rco1
(Refs 38,39). Eaf3 contains a chromodomain for recogniz‑
Acetyl CoA synthetase (Acs) is a ubiquitous Pat ing K36-methylated histone H3 and recruits Rpd3S to
metabolic enzyme that is conserved from bacteria AcuA Ac
nucleosomes that have been methylated by the Set2 his‑
to humans, and that is responsible for the PX4GK PX4GK
CobB tone methyltransferase38,39,42. The recognition efficiency
conversion of acetate to acetyl CoA. In 2002, Acs AcuC
from Salmonella enterica was discovered to be Active Inactive depends on the plant homeodomain (PHD) finger of
acetylated on K609, a key catalytic residue . This modification inhibits the enzymatic
135 Rco1 (Ref. 43). Whereas Rpd3L deacetylates histones at
Nature Reviews | Molecular Cell Biology promoter regions, Rpd3S targets transcribed regions to
activity, thereby regulating the growth of S. enterica on propionate and at low
concentrations of acetate (these are used as carbon and energy sources). K609 is suppress intragenic transcription initiation38,39.
evolutionarily conserved in related enzymes, implying that this regulatory mechanism is S. pombe has three Sin3 orthologues (Pst1, -2 and -3),
important in different organisms. Indeed, the Acs orthologues in Bacillus subtilis and two of which have been shown to form deacetylase
mammals are similarly modified136–138. complexes (table 1). Pst1 associates with Clr6 to form
Two acetylation–deacetylation systems are known to regulate acetylation in bacteria: two slightly different complexes21, both of which share two
Pat (protein acetyltransferase) and CobB (cobalamin biosynthetic gene product B, a other subunits, Sds3 and Prw1; however, only one of the
Sir2 homologue) in S. enterica, and the acetoin utilization A (AcuA) and AcuC proteins
complexes contains the ING protein Png2. These com‑
in B. subtilis136 (see figure). AcuA and AcuC show some sequence similarity to yeast
Gcn5 and Rpd3, respectively136. In addition to AcuC, other bacterial proteins show
plexes silence mating-type genes and are required for cell
sequence similarity to members of the Hda1/Rpd3 family16. Although the functional viability. Pst2 forms a third Clr6 complex with Prw1 and
significance of this similarity remains to be addressed, it implies that this family has a two PHD-finger-containing proteins, one of which has a
role even in prokaryotes. helicase domain. This complex reverses K9 acetylation
of histone H3 and blocks antisense transcription21. It is
presently unclear whether Pst3 forms any complexes
with Clr6.
these family members share not only sequence similarity Mammals have two Sin3 homologues, Sin3A and
but also overall tertiary structure. Therefore, additional Sin3B. The Sin3A complex contains HDAC1, HDAC2,
domains and motifs, as well as sequence divergence within RbAp46 (retinoblastoma (Rb) associated protein-46)
the deacetylase domains, form the molecular basis for the and RbAp48 (Ref. 44) (TABLE 1 and Supplementary infor‑
unique subunit associations and biological function of mation S5 (table)), as well as RBP1 (Rb-binding protein-1),
each individual family member (see below). SAP180 (which is similar to RBP1), SDS3, BRMS1 (breast
cancer metastasis suppressor-1, similar to SDS3), SAP30,
Multiple class I HDAC complexes SAP18 and ING1 (or ING2). Because some of the subunits
With the exception of HDAC8, all class I members have different isoforms, there are several distinct Sin3A
can function as the catalytic subunits of multiprotein complexes. Interestingly, ING2 recognizes K4-methylated
complexes (TABLE 1 and Supplementary information S4 histone H3 and targets the Sin3A complex to deacetylate
(table); S5 (table)). In mammals, HDAC1 and HDAC2 nucleosomes in a methylation-dependent manner45.
SE14 repeat interact with one another and form the catalytic core of The Sin3B complex shares some similarity to the Sin3A
Ser-Glu-containing
several multisubunit complexes11. This core is present complex but might also contain distinct subunits46. Sin3
tetradecapeptide repeat that is
found in HDAC6 proteins from in the mammalian Sin3, nucleosome remodelling homologues are present in C. elegans, D. melanogaster
certain higher mammals. deacetylase (NuRD) and corepressor of RE1-silencing and A. thaliana (TABLE 1), which indicates that similar
transcription factor (CoREST) complexes (TABLE 1 and complexes might exist in various eukaryotes.
SANT domain Supplementary information S5 (table)). One common
A 60-residue module that was
initially identified in Swi3,
feature of these complexes is that they interact with Metazoan and plant Mi‑2/NuRD complexes. NuRD con‑
Ada2, N‑CoR and TFIIIB as a DNA sequence-specific transcription factors to repress tains a catalytic core of HDAC1, HDAC2, RbAp46 and
putative DNA-binding domain. transcription and cooperate with other chromatin modi‑ RbAp48 (Refs 44,47) (TABLE 1 and Supplementary infor‑
It is present in various fiers to shape epigenetic programmes11. Some subunits mation S5 (table)). In addition, this complex contains
transcriptional and chromatin
of the complexes have the ability to bind to chromatin or MTA (metastasis-associated) proteins, the nucleosome-
regulators and is now
considered to be a histone- remodel it (TABLE 1), which provides an important plat‑ remodelling ATPase Mi‑2 (dermatomyositis-specific
binding module. form for coordinating deacetylase functions with other autoantigen), MBD2 (methyl CpG-binding domain-2)
chromatin-regulating mechanisms. and p66. Through MBD2, this complex is recruited for
Chromodomain As discussed below, another common feature of DNA methylation-dependent gene silencing. It also
Originally identified as a 37-
amino-acid-residue chromobox
these HDAC complexes is that SANT domain-containing associates with the methylated DNA-binding proteins
shared by the heterochromatin subunits stimulate deacetylase activities36. MeCP1 (methyl CpG-binding protein-1) and MeCP2,
protein HP1 and the suggesting that this complex is more involved in DNA
polycomb protein Pc2, it has Sin3 complexes from fungi and plants to mammals. methylation than was previously thought47,48. Three
subsequently been found in
S. cerevisiae Rpd3 forms large and small Sin3 complexes, mammalian MTA isoforms (MTA1, -2 and -3) have
many other chromatin
regulators and recognizes known as Rpd3L and Rpd3S, respectively10,37–40 (TABLE 1 been identified47 that contain SANT domains, which
methyl-Lys protein motifs. and Supplementary information S4 (table)). Rpd3L are crucial for the integrity and deacetylase activity of
typically contains the subunits Ume1, Sds3, Sap30 and NuRD. Mi‑2 and p66 have α- and β-isoforms, enabling
Mating-type gene Pho23 (an inhibitor of growth (ING) protein). DNA- the formation of distinct complexes. NuRD subunits
One of a number of genes in
the yeast chromosome that
binding proteins, such as the zinc-finger repressor have been identified in C. elegans, D. melanogaster and
control the sexual fate of the Ume6, target Rpd3L to specific loci for transcriptional A. thaliana (TABLE 1). Similar complexes probably exist,
yeast cell. repression41. Rpd3S shares Sin3 and Ume1 with Rpd3L, therefore, in animal and plant kingdoms.

REVIEWS
Table 1 | Class I HDAC complexes and their composition in different eukaryotic organisms*
Complex Mammals D. melanogaster C. elegans S. cerevisiae S. pombe A. thaliana Protein domain
Rpd3L Rpd3S I II
Sin3 HDAC1, HDAC2 RPD3 HDA-1 Rpd3 Rpd3 Clr6 Clr6 RPD3 Class I deacetylase
RbAp46, RbAp48 p55 RBA-1, LIN-53 Ume1 Ume1 Prw1 Prw1 WD40 repeat
Sin3A Sin3 Sin3 Sin3 Sin3 Pst1 Pst2 Sin3 PAH motifs
Sds3, BRMS1 Sds3 Sds3 Sds3 Coiled coil
RBP1
SAP30 Sap30
SAP18 SAP18 Ubiquitin fold
ING1/2 Pho23 Png2 PHD finger
Rco1 PHD finger
Eaf3 Chromodomain
Cph1 PHD finger
Cph2 PHD finger
Alp13 Chromodomain
Mi-2/NuRD HDAC1, HDAC2 RPD3 HDA-1 RPD3 Class I deacetylase
RbAp46, RbAp48 p55 RBA-1, LIN-53 WD40 repeat
Mi-2α/β Mi-2 CHD-3/-4 PICKLE Helicase
MTA1–3 EGR-1, EGL-27 SANT domain
MBD2, MBD3 Methyl CpG binding
p66α/β p66
CoREST HDAC1, HDAC2 RPD3 HDA-1 Class I deacetylase
CoREST CoREST SPR-1 SANT domain
LSD1 Su(var)3-3 SPR-5 SWIRM,
K-demethylase
BHC80 PHD finger
CtBP1 Dehydrogenase
N-CoR/SMRT HDAC3 HDAC3 Class I deacetylase
N-CoR/SMRT SMRTER SANT domain
TBL1/TBLR1 WD40 repeat
GPS2
JMJD2A Jumonji demethylase
Kaiso Methyl CpG binding
*For the mammalian and yeast complexes, only those subjected to biochemical purification are listed here. For complexes in D. melanogaster, C. elegans and
A. thaliana, little biochemical purification has been performed, so the listed subunits are based on sequence and functional similarity with counterparts in yeast and
mammals. In the CoREST and N-CoR/SMRT complexes, there are only sub-stoichiometric amounts of CtBP1or JMJD2A and Kaiso, respectively. Alp13, altered polarity-13;
BHC80, an 80-kDa subunit of the BRAF-histone deacetylase complex; BRMS1, breast cancer metastasis suppressor-1 (similar to SDS3); CHD-3/-4, chromo-helicase
DNA-binding domain protein-3/-4; Clr6, cryptic locus regulator-6; CoREST, corepressor of REST (repressor element-1 silencing transcription factor); Cph1/2, Clr6-
associated PHD protein-1/-2; CtBP1, E1A C-terminal binding protein-1; Eaf3, Esa1-associated factor-3; EGL-27, egg laying mutant-27; EGR-1, EGL-27 related protein-1;
GPS2, G protein pathway suppressor-2; HDA-1, histone deacetylase-1; HDAC, histone deacetylase; ING1/2, inhibitor of growth-1 or -2; JMJD2A, jumonji demethylase D
group member-2A; LIN-53, lineage family member-53; LSD1, lysine-specific demethylase-1; MBD3, methyl CpG-binding-domain-containing protein-3;
MTA1–3, metastasis-associated proteins-1–3; Mi-2, dermatomyositis-specific autoantigen; N-CoR, nuclear receptor corepressor; NuRD, nucleosome remodelling
deacetylase complex; PAH motif, paired amphipathic helix motif; PHD finger, plant homeodomain-linked zinc finger; Pho23, an ING protein; PICKLE, A. thaliana mutant
with pickle-like roots; Png2, S. pombe ING protein-2; Prw1, S. pombe RbAp48-related WD40 protein-1; RBA-1, RbAp48-related protein; Pst1, S. pombe SIN three-1;
RBP1, retinoblastoma (Rb)-binding protein-1; Rco1, S. pombe ING protein-2; Rpd3, reduced potassium dependency-3; SANT domain, Swi3, Ada2, N-CoR and TFIIIB
homologous domain; SAP30, Sin3-associated polypeptide of 30 kDa; SDS3, suppressor of defective silencing-3; Sin3, SWI-independent-3 (also known as Rpd1);
SMRT, silencing mediator of retinoic acid and thyroid hormone receptors; SMRTER, D. melanogaster homologue of SMRT and N‑CoR; SPR-1/-5, suppressor of
presenilin-1/-5; Su(var)3-3, suppressor of position-effect variegation 3-3; SWIRM domain, Swi3, Rsc8 and Moira homologous domain; TBL1, transducin β-like protein-1;
Ume1, unscheduled meiotic gene expression-1; WD40 repeat, Trp-Asp 40 repeat. The forward slash ‘/’ represents ‘or’.
CoREST complex. Initially identified as a corepressor of and Supplementary information S5 (table)). One
the transcriptional repressor REST49, CoREST is a com‑ subunit, BHC110 (also known as KIAA0601), has an
ponent of an HDAC1–HDAC2 complex that is referred FAD-binding domain with sequence homology to maize
to as a neuronal corepressor complex50–52 ( TABLE 1 polyamine oxidase51. This subunit was recently shown

REVIEWS
to be a demethylase for K4-methylated histone H3 and Class IIa HDACs as signal transducers
was therefore renamed Lys-specific demethylase-1 HDAC4, -5, -7 and -9 have unique motifs for MEF2
(LSD1)53. The CoREST complex promotes nucleosomal and 14‑3-3 binding (FIG. 1b). Depending on the cellular
deacetylation through its SANT domains54. Interestingly, context, mammalian class IIa members can exist in the
histone demethylation stimulates the deacetylase activ‑ nucleus and the cytoplasm14,15,26. They contain intrinsic
ity, which, in turn, is required for optimal demethylation, nuclear import and export signals for dynamic nucleo‑
indicating that the two activities are interdependent55. cytoplasmic trafficking (FIG. 2a), so the association of
CoREST subunits have been found in C. elegans and partners might alter this equilibrium and the subsequent
D. melanogaster (TABLE 1), but not in fungi or plants. subcellular distribution. Diverse signalling stimuli func‑
This complex might, therefore, be specific to metazoans, tion through this equilibrium to control the repression
which is consistent with its primary role in repressing activity of these HDACs (for example, to inhibit MEF2-
neuronal gene expression. dependent transcription), which establishes them as novel
signal transducers (FIG. 2).
HDAC3 complexes: from flies to men. Biochemical
purification of HDAC3 uncovered an unexpected Nucleocytoplasmic trafficking mediated by 14‑3-3. MEF2
link to two homologous transcriptional corepressors, binds to class IIa HDACs and promotes their nuclear
nuclear receptor corepressor (N-CoR) and silencing localization. By contrast, 14‑3-3 proteins stimulate the
mediator of retinoic acid and thyroid hormone recep‑ cytoplasmic retention of class IIa HDACs in a phospho‑
tors (SMRT)56–58 (TABLE 1 and Supplementary informa‑ rylation-dependent manner15,26. HDAC4 contains three
tion S5 (table)). Conversely, HDAC3 was identified 14‑3-3-binding sites, S246, S467 and S632, all of which
through affinity purification of both corepressors59–61. are conserved in HDAC5, -7 and -9 (FIG. 1b). Only S246
The N‑CoR and SMRT corepressors stimulate the and S467 seem to be functional in HDAC5 and HDAC9
deacetylase activity of HDAC3 through their SANT (Refs 69,70), whereas a fourth site might be important in
domains62, which are highly conserved in SMRTER, the HDAC7 (Ref. 71) (FIG. 1b). Mutational analysis revealed that
D. melanogaster homologue of SMRT and N‑CoR. An the S246 site of HDAC4 is the most important15,26. This
HDAC3 orthologue exists in D. melanogaster, suggest‑ site is conserved in C. elegans and D. melanogaster (FIG. 1b
ing that SMRTER might form a functional deacetylase and Supplementary information S1b (figure)).
complex. Similar proteins are absent from C. elegans and Both 14‑3-3 binding and the presence of a nuclear
plants, so HDAC3 complexes might only be conserved export signal are necessary for the cytoplasmic localiza‑
from flies to mammals. tion of HDAC4 and paralogues in mammals. The mecha‑
nistic impact of 14‑3-3 binding is multifaceted (FIG. 2).
Yeast and metazoan class II HDAC complexes First, the S246 site is adjacent to the nuclear localization
Yeast Hda1 self-associates and forms a tetrameric com‑ signal (NLS; FIG. 2a), so 14‑3-3 might impede the access
plex with Hda2 and Hda3, in which both are impor‑ of the nuclear import receptor importin-α/β to the NLS.
PHD finger tant for the deacetylase activity63. Hda2 and Hda3 Second, 14‑3-3 binding might provide a signal in trans to
A plant homeodomain-linked
homologues are present in Candida albicans, but not promote the nuclear export of class IIa HDACs. Third,
(PHD) zinc finger that chelates
double zinc ions. This type of in S. pombe or higher organisms. Instead, Clr3 forms a 14‑3-3 binding affects the subnuclear distribution of
zinc finger is present in many quartet core complex in S. pombe64. One subunit, Mit1, class IIa HDACs. For example, upon binding to 14‑3-3
chromatin regulators and was contains a PHD finger and a helicase domain, thus con‑ proteins, an HDAC9 isoform lacking the deacetylase
recently shown to bind the ferring both deacetylase and chromatin-remodelling domain and nuclear export signal disperses from nuclear
N‑terminal tails of core
histones in a methylation-
activities within one complex64 (somewhat reminiscent punctae into a uniform nuclear distribution72. Similarly,
dependent or -independent of the NuRD complex in metazoans). The Clr3 com‑ 14‑3-3 binding to C. elegans HDAC4 (Supplementary
manner. plex is recruited to euchromatic and heterochromatic information S1a (figure)) does not seem to promote
loci for transcriptional gene silencing through Mit1, cytoplasmic localization73, perhaps owing to the lack
Euchromatic
a heterochromatin protein-1 (HP1) homologue and a of a nuclear export signal in this deacetylase. Fourth,
DNA that contains most of the
structural genes. It changes telomere-binding protein64. 14‑3-3 association protects HDAC7 from proteasomal
structure during the cell cycle Non-catalytic subunits of Hda1 and Clr3 complexes degradation74,75, which contributes to its cytoplasmic
and undergoes transcriptional are not conserved in metazoans, so these complexes accumulation. Finally, 14‑3-3 binding sites are subject to
regulation. might be specific to fungi. This is consistent with the reversible phosphorylation. By binding to phosphorylated
Chaperone
fact that metazoan class II HDACs have unique domains residues, 14‑3-3 binding might occlude phosphatases
A protein that mediates the and motifs that are absent in Hda1 and Clr3 (FIG. 1 and and therefore inhibit dephosphorylation, which might
assembly of another Supplementary information S1 (figure)). Indeed, affinity function as a feed-forward reinforcing mechanism71.
polypeptide-containing purification of mammalian class IIa HDACs identified
structure, but does not form
14‑3-3 proteins65, a kinase66, a phosphatase66 and HSP70 Kinases and phosphatases for class IIa HDACs. Five
part of the completed structure
or participate in its biological (E.S. and X.J.Y., unpublished observations), whereas groups of kinases have been shown to phosphorylate
function. HDAC6 was found to associate with cytoskeletal pro‑ the 14‑3-3 binding sites on HDACs: Ca2+/calmodulin-
teins and chaperones67,68. As discussed below, the kinase dependent kinases (CaMKs)69,76, protein kinase D77–79,
Paralogue and phosphatase, as well as 14‑3-3 proteins, have a key microtubule affinity-regulating kinases71,80, salt-inducible
A sequence, or gene, that has
originated from a common
role in regulating the subcellular trafficking of class IIa kinases73,81 and checkpoint kinase‑1 (CHK1)82 (FIG. 2a).
ancestral sequence, or gene, HDACs, whereas the association partners of HDAC6 Conversely, two phosphatases have been identified:
by a duplication event. shed light on its important functions in the cytoplasm. the MYPT1–PP1b (myosin phosphatase-targeting

REVIEWS
a cAMP subunit‑1–protein phosphatase-1β) complex83 and

LKB1 PP2A66,71,84,85. Available data indicate that most of the
PKA High DNA kinases phosphorylate all class IIa members in mam‑
[Ca2+] PKC ? salt damage ? mals. The S246 site is conserved from C. elegans and
? D. melanogaster to mammals (Supplementary infor‑
CaMK PKD MARK SIK CHK1 Kinase? mation S1 (figure)), and most of the kinases have
? orthologues in invertebrates, which suggests that these
MYPT1–PP1β and PP2A regulatory mechanisms might be evolutionarily con‑
served. Consistent with this, the salt-inducible kinase
NES orthologue KIN‑29 in C. elegans phosphorylates HDAC4
MEF2 P P P
(Supplementary information S1 (figure)) and promotes
HDAC4 S S S DAC 1,084
246 467 632 14‑3-3 binding73. Thus, the nuclear function of class IIa
HDACs seems to be regulated by different signalling
14-3-3 NLS pathways through reversible phosphorylation at 14‑3-3
binding sites. In this regard, these deacetylases function
HDAC4 236-KDDFPLRKTASEPNLKLRSRLKQ KVAERRSSPL
as novel signal transducers, similar to the STAT and
HDAC5 249-RDDFPLRKTASEPNLKLRSRLKQ KVAERRSSPL
145-PEHFPLRKTVSEPNLKLRYKPK- KSLERRKNPL
SMAD signal-responsive transcription factors (FIG. 2b).
HDAC7
HDAC9 210-QDDFPLRKTASEPNLKVRSRLKQ KVAERRSSPL Mapping out events that occur upstream of these kinases
232-ESDFPLRKTASEPNL-LKIRLKQ SVIERKARIG (FIG. 2a) will be necessary to determine how the regulation
dHDAC4
cHDAC4 188-ASEFQLRKVNSEPNLKMRIRAKL LSKGSSPVQH is integrated in cellular signalling networks under various
physiological and pathological contexts.
Conservation island
b Signals TGFβ Cytokine Regulation of class IIa HDACs by other methods. In addi‑
tion to the site-specific phosphorylation described above
Cytoplasm (FIG. 2a), class IIa HDACs are subject to ubiquitylation and
sumoylation74,86,87. Ubiquitylation promotes protein degra‑
dation74,75, but the functional consequence of sumoylation
14-3-3 P HDAC4 P 14-3-3 SMAD2 SMAD4 STAT STAT is less clear86,88. HDAC4 sumoylation is extremely efficient,
which makes it a potentially important modification. In
addition, this deacetylase is subject to non-reversible
Nucleus processing. For example, caspase cleavage produces an
P P HDAC4 fragment that corresponds to the N‑terminal
HDAC4 SMAD2 SMAD4 Smad2
STAT Smad4
STAT 289 residues89,90, within which there is a transcriptional
P P
repression domain26. Moreover, alternative splicing of
HDAC9 mRNA produces isoforms that also contain this
N‑terminal repression domain87. The HDAC4 fragment
Figure 2 | HDAC4 and other class IIa homologues as signal transducers. a | Multiple and some of the HDAC9 isoforms have an intact MEF2-
kinases (calcium/calmodulin-dependent protein kinase Nature(CaMK), protein
Reviews kinase
| Molecular D Biology
Cell (PKD), binding site and can still function as transcriptional
microtubule affinity-regulating kinase (MARK), salt-inducible kinase (SIK), checkpoint corepressors, which supports the hypothesis that class IIa
kinase-1 (CHK1) and other kinases) mediate specific phosphorylation of human histone HDACs can repress transcription independently of their
deacetylase-4 (HDAC4) on three 14‑3-3-binding sites. Myosin phosphatase-targeting cryptic deacetylase activity, owing to the aforementioned
subunit‑1 (MYPT1)–protein phosphatase-1β (PP1β) and PP2A can also act on these sites. Tyr→His substitution28,29. Regulation of this enzymatic
The association of 14‑3-3 proteins with HDAC4 retains it in the cytoplasm and prevents activity could be a potentially important mechanism.
its interaction with transcription factors such as myocyte enhancer factor-2 (MEF2), Furthermore, whereas HDAC4 expression is controlled by
thereby releasing these transcription factors for transcriptional activation. This model
microRNA91, transcription of HDAC9 mRNA is regulated
can also be extended to other class IIa members, although some of the kinases and
phosphatases have only been shown for HDAC4, -5, -7 and/or -9 (see main text for
by a negative-feedback loop in which HDAC9 represses its
references). The sequence surrounding the major 14‑3-3-binding site (S246, indicated own promoter92. Therefore, class IIa members are signal-
in red) of HDAC4 is compared with the corresponding regions of its paralogues responsive transcriptional corepressors, the functions of
(HDAC5, -7 and -9) as well as orthologues from Drosophila melanogaster (dHDAC4) and which are tightly controlled in multiple ways.
Caenorhabditis elegans (cHDAC4). Other residues that are crucial for phosphorylation
and 14‑3-3 binding are highlighted in green. This 14‑3-3 site constitutes a sequence HDAC6 as a key cytoplasmic regulator
conservation island139. The nuclear localization signal (NLS), which can be impeded by Whereas the nuclear function of class IIa members
14‑3-3 binding, is also shown. Grey shading denotes residues that are highly conserved. is well established, their role in the cytoplasm (if any) is
b | Signal-responsive transcriptional repression by HDAC4 (left) is analogous to unclear. By comparison, the class IIb member HDAC6
transcriptional activation by transforming growth factor‑β (TGFβ)-stimulated SMADs
is functionally important in this compartment and is
(middle) and cytokine-activated signal transducer and activator of transcription (STAT)
proteins (right). 14‑3-3 proteins bind to HDAC4 and retain it in the cytoplasm, thereby
predominantly (if not exclusively) cytoplasmic31. This
inhibiting its corepressor activity; upon dephosphorylation, HDAC4 translocates to the finding has led to the simple hypothesis that HDAC6 has
nucleus for transcriptional repression. For simplicity, only two 14‑3-3 molecules are functions that are unrelated to histone modification, and
shown. This schematic also applies to HDAC5, -7 and -9. DAC, deacetylase domain; available results convincingly support this theory. Indeed,
LKB1, Ser–Thr kinase-11; NES, nuclear export signal; P, phosphorylation; PKA, protein multiple cytoplasmic processes seem to be regulated by
kinase A; PKC, protein kinase C. HDAC6 (FIG. 3).

REVIEWS
HDAC6 DAC1 DAC2 ZnF

Ubiquitin
Protein deacetylation binding
α-tubulin Cortactin HSP90 IFNαR? Histones Aggresome formation

and others? Stress response
Cell motility Cell motility GR maturation Autophagy
Cell adhesion Kinase activation Macropinocytosis
Immune synapse Macropinocytosis
Ciliogenesis
Figure 3 | HDAC6 is a major deacetylase in the cytoplasm. Through its tandem deacetylase (DAC) domains, histone
Nature Reviews
deacetylase-6 (HDAC6) deacetylates α‑tubulin, cortactin and HSP90 (heat shock protein-90) | Molecular
to regulate Cell Biology
cell motility,
cilium assembly, cell adhesion, the immune synapse, macropinocytosis, maturation of the glucocorticoid receptor (GR)
and activation of some protein kinases. HDAC6 may also deacetylate the type I interferon receptor (IFNαR)132. Additional
cytoplasmic substrates are likely to be identified. The functional significance of HDAC6 in deacetylating histones remains
to be established, and further investigation is needed to determine the roles that HDAC6 may have in the nucleus.
Through its ubiquitin-binding zinc finger (ZnF), HDAC6 binds to ubiquitin and regulates aggresome formation, autophagy,
activation of heat-shock factor-1 (HSF1) and platelet-derived growth factor (PDGF)-induced macropinocytosis.
HDAC6 regulates cytoskeletal dynamics. HDAC6 has p23 (a co-chaperone) and GR102. Studies in a cell-free
been shown to function as an α‑tubulin deacetylase93–95, system showed that HDAC6 promotes the assembly of
and it associates with microtubules and colocalizes with a stable HSP90–GR heterocomplexes but does not affect
microtubule motor complex93. Exogenous expression of the intrinsic properties of GR104. Moreover, depletion of
HDAC6 leads to global deacetylation of α‑tubulin, and HDAC6 induces HSP90 acetylation and reduces its asso‑
microRNA
RNA-interference-mediated knockdown of HDAC6 ciation with the leukaemic tyrosine kinase Bcr-Abl103.
A small RNA of ~21
nucleotides that regulates the increases α‑tubulin acetylation. Ectopic expression of HSP90 acetylation also regulates the kinase activity of
expression of mRNAs to which HDAC6 also promotes chemotaxis93, which indicates CHK1 (ref. 105). Because a large number of other pro‑
it is complementary in that HDAC6-mediated deacetylation probably regu‑ teins (including many protein kinases that are involved
sequence. lates microtubule-dependent cell mobility. In addition, in oncogenesis) require HSP90 for their maturation and
Microtubule
HDAC6 might regulate cell adhesion96, the formation of function, it will be exciting to examine whether HSP90
A hollow tube, 25 nm in the immune synapse97 and motor-mediated cargo transport deacetylation is important for their regulation. Although
diameter, that is formed by the along microtubules98,99. HDAC6 also has an unexpected the list of cytoplasmic substrates of HDAC6 will probably
lateral association of 13 role in the reabsorption of primary cilia of mammalian continue to grow, the issue of whether this deacetylase has
protofilaments, which are
cells100. Similar to most HDACs, HDAC6 is a phospho‑ a nuclear role awaits further investigation31.
themselves polymers of α‑ and
β‑tubulin subunits. protein101. Phosphorylation by Aurora A kinase activates
the tubulin deacetylase activity of HDAC6 and promotes HDAC6 and ubiquitin-dependent pathways. HDAC6
Chemotaxis ciliary disassembly100. HDAC6 therefore has roles in from mouse testis associates with the mammalian
A type of migration that is various microtubule-dependent cytoplasmic processes. homologue of yeast Ufd3 (ubiquitin fusion degradation
stimulated by a gradient of a
chemical stimulant or
The filamentous (F)-actin-binding protein cortactin protein-3) and p97 (also known as valosin-containing
chemoattractant. also interacts with HDAC6 and functions as its substrate68. protein; VCP)31,67, a member of the AAA+ family of ATPases.
As shown for α‑tubulin, exogenous expression of HDAC6 Strikingly, the C‑terminal zinc finger of HDAC6 efficiently
Immune synapse leads to cortactin hypoacetylation, whereas inhibition interacts with ubiquitin, which causes HDAC6 to disso‑
A junction that forms at the
of the deacetylase activity results in hyperacetylation. ciate from the protein complex31. These results indicate
contact region between a T cell
and its target cells. T‑cell Independently, cortactin was identified as an acetylated an unexpected role for HDAC6 in ubiquitin-dependent
activation occurs here. protein in a proteomics study (M. Yoshida, personal pathways.
communication). Interestingly, HDAC6 alters the ability The expression of misfolded proteins in cells often
F-actin of cortactin to bind F‑actin by modulating the charge in the leads to the formation of aggresomes106, a process that is
(Filamentous actin). A flexible,
helical polymer of G‑actin
repeat region of cortactin and, in this way, modifies cell facilitated by the delivery of misfolded proteins to these
(globular actin) monomers that movement68. Therefore, in addition to its role in control‑ inclusion bodies by dynein-dependent transport on micro‑
is 5–9 nm in diameter. ling microtubule-dependent cell motility, HDAC6 also tubules106. HDAC6 is a component of the aggresome, and
seems to influence actin-dependent cell motility. treatment of cells with a proteasome inhibitor results in
AAA+ family of ATPases
the accumulation of HDAC6 in polyubiquitin-containing
A superfamily of ATPases that
are associated with various HDAC6 deacetylates HSP90. The heat shock chaperone aggresomes107.
cellular activities. They have protein HSP90 also associates with HDAC6 and is a bona Mechanistically, HDAC6 associates with dynein
one or two nucleotide-binding fide cytoplasmic substrate102,103. HDAC6 is required for motors and functions as an adaptor to link these motors
domains (‘AAA modules’), HSP90-dependent activation of the glucocorticoid recep‑ to ubiquitylated misfolded proteins for delivery to
which often form ring-like
oligomers and function as
tor (GR). Treatment of cells with an HDAC6-sensitive aggresomes and degradation by the proteasome107. p97
chaperones in diverse cellular deacetylase inhibitor or small interfering RNA-mediated disrupts HDAC6–ubiquitin complexes and counteracts
processes. knockdown of HDAC6 causes HSP90 to dissociate from the ability of HDAC6 to promote the accumulation of

REVIEWS
polyubiquitylated proteins108. Although the deacetylase is a prerequisite for subsequent methylation. This
activity of HDAC6 was found to be required for aggre‑ notion is well established in S. pombe21,64, and provides
some formation107, two recent studies showed that tubulin further support for an intimate connection between
acetylation or HDAC6 inhibition accelerates dynein- and histone deacetylation and methylation. And finally, all
kinesin-mediated transport98,99. four members of class IIa have been subjected to gene
Further support for the proposal that HDAC6 is targeting in mice, with the resultant phenotypes being
essential for efficient management of cytotoxic proteins linked to MEF2 (TABLE 2), thereby indicating that the
is provided by two recent studies suggesting that HDAC6 MEF2–HDAC axis (FIG. 2b) operates in many tissues in
increases the clearance of various substrates through addition to muscle114.
autophagy109,110, a lysosome-mediated degradation path‑
way. Impairment of the ubiquitin–proteasome system Implications for disease and therapy
induces autophagy, and this induction rescues the neuro‑ As described above, studies in cell-based and animal
degeneration that is associated with such impairment109. models have shed important light on the potential
Exogenous expression of HDAC6 alone compensates involvement of the Rpd3/Hda1 family members in the
for an impaired ubiquitin–proteasome system and res‑ pathophysiology of human diseases. Indeed, HDAC2 has
cues neurodegeneration by autophagy109. Exactly how been linked to tumorigenesis in colorectal cancers and
HDAC6 accelerates the turnover of misfolded proteins increased expression of this deacetylase has been found
by autophagy is not known. Whatever the mechanism, it in most human colon cancer explants115. In addition, a
is clear that HDAC6 has a key role in the cellular clearance frameshift mutation, which causes the loss of HDAC2 and
of misfolded proteins by different pathways and might be confers resistance to HDAC inhibition, has been detected
a potential therapeutic target for neurodegenerative dis‑ in some human cancers116.
eases. By associating with HSP90 and binding to ubiquitin, Members of the Rpd3/Hda1 family have also
HDAC6 functions as a sensor for ubiquitylated protein been implicated in neuropathological conditions.
aggregates to regulate the activation of heat-shock factor-1 Accumulating data indicate a key role for HDAC6 in
(HSF1) during stress responses111. Moreover, HDAC6 is neurodegeneration. It is required for aggresome forma‑
an essential component of stress granules112. This deacety‑ tion and is present in Lewy bodies that are associated
lase, therefore, has key roles in different levels of cellular with neurodegenerative disorders such as Parkinson’s
responses. HDAC6 might also regulate cytoplasmic disease and dementia107. The impairment of autophagy
processes such as macropinocytosis113. The mechanism by resulting from decreased HDAC6 activities might pre‑
which the deacetylase and ubiquitin-binding activities are dispose individuals to neurodegeneration109. By contrast,
coordinated remains to be addressed. HDAC6 inhibition promotes tubulin acetylation and
Together, recent studies have established HDAC6 as a facilitates microtubule-based vesicle transport, thereby
major cytoplasmic regulator (FIG. 3) with properties that are compensating for the transport deficit in Huntington’s
not shared by other deacetylases, which is consistent with disease99. Thus, functional roles of HDAC6 may vary
Inclusion body its unique domain organization (FIG. 1 and Supplementary in different pathological states. Overall, compared to
An insoluble aggregate of
information S1 (figure)). the rapidly accumulating knowledge about the cellular
misfolded proteins. Inclusion
bodies are common in and developmental roles of the Rpd3/Hda1family of
prokaryotes and in the brains Rpd3/Hda1 family members in development deacetylases, much less is known about their direct and
of patients with triplet-repeat Gene inactivation analyses of some class I and II mem‑ specific roles in human pathology. Additional studies
diseases. bers in C. elegans, D. melanogaster and mice have not are also required to extrapolate results accurately from
Dynein
only revealed important roles for HDAC1, -2 and -3 animal studies to humans and to determine how different
A microtubule-based molecular and class IIa members in different developmental pro HDACs might be altered or mutated in different types or
motor that moves towards the cesses, but have also raised the important issue of how stages of diseases.
minus end of microtubules. to correlate these developmental functions with results Classical HDAC inhibitors have the potential to be
from biochemical and cell-based assays. The findings therapeutic and chemopreventive agents for cancer19,117,
Kinesin
A microtubule-based molecular from the gene inactivation analyses are summarized neurodegenerative disorders20, heart disease70, muscular
motor that is most often in TABLE 2 and Supplementary information S6 (table), dystrophy118, spinal muscular atrophy119 and transplanta‑
directed towards the plus end but some important points are noteworthy. First, as tion intolerance120. Attesting to this potential, one such
of microtubules. described above, HDAC1, -2 and -3 are catalytic sub inhibitor was recently approved for treating cutaneous
Autophagy
units of multiprotein proteins (TABLE 1), so in addition T‑cell lymphoma19. Most of the inhibitors that are cur‑
A pathway for the recycling of to the deacetylases themselves, genetic analyses of asso‑ rently available display limited selectivity towards Rpd3/
cellular contents, in which ciated subunits are informative about the functions of Hda1 family members in humans, raising the important
materials inside the cell are different deacetylase complexes (TABLE 2). Second, global question about the identity of the deacetylase(s) that are
packaged into vesicles and
deletion experiments have revealed that the functions of actually targeted in vivo. The target might not be any of
then targeted to the vacuole or
lysosome for bulk turnover. mouse HDAC1 and HDAC2 differ during development the class IIa members because knockout studies suggest
and in cultured tumour cells, suggesting that these two that their activation may be beneficial (TABLE 2). HDAC6
Macropinocytosis deacetylases do not always remain together as a catalytic has an important role in the management of misfolded
A form of regulated core. Third, deletion of Sds3 or Sin3A (both of which proteins, so its activation may also be beneficial (FIG. 3).
endocytosis that involves the
formation of large endocytic
are subunits of class I HDAC complexes; TABLE 1) affects However, it positively regulates cell motility and the
vesicles after the closure of K9 methylation of histone H3 and the distribution of activation of some oncogenic kinases, and its inhibition
cell-surface membrane ruffles. HP1 within the nucleus, indicating that K9 deacetylation accelerates microtubule-based vesicle transport, so this

REVIEWS
Table 2 | Rpd3/Hda1-like HDACs and associated subunits in human disease and mouse development
Class Member Associated Map position Human disease Effect of mouse gene deletion Refs
subunit*
I HDAC1 1p34.1 Colon cancer Embryonic lethal by E9.5 140, 141
Proliferation defects in ES cells 140,142
I HDAC2 6q21 Perinatal death from cardiac defects, 141
cardiomyopathy from heart-specific
deletion of both HDAC1 and HDAC2
Sin3A 15q24.2 Embryonic lethal by E6.5; splenomegaly and 143,144
glomerulopathy in heterozygotes
Sds3 12q24.23 Embryonic lethal by E6.5, defects 145
in heterochromatin formation and
chromosome segregation
Mi-2β 12p13 Defects in T-cell development and CD4 gene 146
expression
MBD3 19p13.3 Pre- or perinatal death, defects in ES cell 147
pluripotency
p66α 19p13.11 Embryonic lethal by E9.5 148
LSD1 1p36.12 Embryonic lethal by E7.5 149
I HDAC3 5q31.3 Embryonic lethality with vascular defects R. Montgomery and
E. Olson, PC
Liver hypertrophy S. Hiebert, PC
N-CoR 17p11.2 Embryonic lethal by E15.5 and defects in 150
neuronal and haematopoietic differentiation
I HDAC8 Xq13 Craniofacial abnormalities M. Haberland and
E. Olson, PC
IIa HDAC4 2q37.2 Osteodystrophy? Small size, chondrocyte hypertrophy
Exencephaly 151
IIa HDAC5 17q21 Cardiac hypertrophy 152
IIa HDAC7 12q13.1 Cardiovascular defects 153
IIa HDAC9 7p12.1 Cardiac hypertrophy 154
IIb HDAC6 Xp11.22 Neurodegeneration? Viable with massive increase of α-tubulin 112
acetylation
IIb HDAC10 22q13.3 ?
IV HDAC11 3p25.2 ?
*For associated subunits, only those subjected to gene inactivation analysis in mice are listed here. E, embryonic day; ES, embryonic stem; HDAC, histone
deacetylase; LSD1, lysine-specific demethylase-1; MBD3, methyl CpG-binding-domain-containing protein-3; Mi-2, dermatomyositis-specific autoantigen;
N-CoR, nuclear receptor corepressor ; PC, personal communication; Sds3, suppressor of defective silencing-3.
deacetylase could be one in vivo target. Other candidates It is important to note that the actual in vivo target
include class I members. If the inhibitor targets HDAC1 may vary in various diseases and in different pathological
or HDAC2, then the question is which deacetylase com‑ stages. Because the inactivation of some classical HDACs
plex is the authentic target in patients? Of relevance is is detrimental under certain conditions, small-molecule
the fact that RNA-interference-mediated knockdown of activators may have therapeutic potential26. Because many
HDAC1 in D. melanogaster and mammalian cells recapit of the deacetylases are evolutionarily conserved (FIG. 1 and
ulates some effects of the HDAC inhibitor trichostatin A Supplementary information S1 (figure)), different model
on cell-cycle progression121,122. organisms provide economical alternatives for screening
In terms of designing new HDAC inhibitors, are iso‑ HDAC modulators and testing novel strategies for therapy
form-specific compounds preferable to inhibitors that and prevention.
target all classical HDACs? With regards to isoform-
specific intervention, would it be more reasonable to Conclusions and perspectives
target the signalling pathways that control HDAC func‑ Since the identification of the first HDACs in 1996,
tions? Each HDAC has multiple substrates (such as those additional members of the Rpd3/Hda1 family have been
for HDAC6; FIG. 3) or binding partners (such as MEF2 identified in various organisms (FIG. 1, Supplementary
and 14‑3-3 for class IIa HDACs; FIG. 2), so the inter information S1 (figure) and BOX 2). There are 11 family
action interface might be a suitable intervention point members in humans and, with the exception of HDAC8,
for designing selective inhibitors. all class I members form multisubunit complexes (TABLE 1).

REVIEWS
These complexes show different patterns of evolutionary A coherent and integrated picture from numerous stud‑
conservation and carry out key roles in deacetylating his‑ ies will not only provide a valuable guide for identifying
tones, interacting with other chromatin regulators and and characterizing related human diseases, but will also
shaping epigenetic programmes. Genome-wide localiza‑ form an important basis for rational drug design. Global
tion maps have been generated for classical HDACs and or tissue-specific gene knockout in mice tends to provide
components of their multisubunit complexes in yeast18,43,64 information that is related to early development, whereas
and should function as valuable guides for further identi human diseases often develop in the late stages of life.
fication and characterization of similar complexes in Therefore, time-specific gene inactivation or knockdown
multicellular organisms. Class I deacetylases also target is expected to greatly extend our understanding of human
non-histone substrates (BOX 1), but further characteriza‑ pathology and to efficiently mimic the effects of HDAC
tion is needed to address how the deacetylase complexes inhibition to identify the best in vivo targets for developing
are involved. new generations of HDAC inhibitors.
Members of class IIa function as novel signal transduc‑ Recent proteomic studies have identified a diverse
ers to repress transcription in a phosphorylation-depend‑ array of Nε-acetylated proteins123–125, but the responsible
ent manner (FIG. 2a). They are positioned at a central enzymes are largely unknown. There are also acetylation-
interface between signalling and gene regulatory networks like modifications such as formylation126 and propionyla‑
(FIG. 2b). For example, the MEF2–class IIa HDAC axis has tion127,128. Moreover, protein O‑acetylation has just been
a key role in regulating signal-dependent gene expres‑ reported129, and an HDAC-like bacterial protein displays
sion and cell differentiation in various tissues114. The intrinsic esterase activity30. These exciting discoveries
class IIb member HDAC6 is a major deacetylase in the make us wonder whether we are still at the dawn of a
cytoplasm, using its ubiquitin-binding domain as a stress new era for protein acetylation research. How the Rpd3/
sensor (FIG. 3). HDAC8 and HDAC10 appear to be specific Hda1 family is involved in deacetylating newly identified
to deuterostomes and vertebrates, respectively, whereas Nε-acetylated proteins, controlling O‑acetylation and
HDAC11 is conserved even in invertebrates and plants. lysine acetylation-like modifications, and interacting with
Compared with other mammalian HDACs, much less is other post-translational modifications, such as methyla‑
known about functions of these three deacetylases. tion, phosphorylation, ubiquitylation and sumoylation,
Classical HDACs and some of their associated sub is an exciting question that awaits further investigation.
units have been subjected to gene inactivation analyses in Answers to this question will provide rich and novel
mice (TABLE 2), as well as in C. elegans, D. melanogaster and insights into the fundamental issue of how members of
zebrafish (Supplementary information S6 (table)). One the Rpd3/Hda1 family are involved in shaping ‘histone
challenge is to correlate the resulting phenotypes with code-like’ regulatory programmes that have evolved from
data from biochemical, molecular and cellular studies. bacteria and yeast to mice and men.
1. Gershey, E. L., Vidali, G. & Allfrey, V. G. Chemical First report on the direct physical link of Rpd3 to a inhibitor as an anticancer drug. Nature Biotechnol.
studies of histone acetylation. The occurrence of sequence-specific transcriptional repressor. 25, 84–90 (2007).
ε‑N‑acetyllysine in the f2a1 histone. J. Biol. Chem. 10. Rundlett, S. E. et al. HDA1 and RPD3 are members of 20. Butler, R. & Bates, G. P. Histone deacetylase inhibitors
243, 5018–5022 (1968). distinct yeast histone deacetylase complexes that as therapeutics for polyglutamine disorders. Nature
2. Inoue, A. & Fujimoto, D. Enzymatic deacetylation of regulate silencing and transcription. Proc. Natl Acad. Rev. Neurosci. 7, 784–796 (2006).
histone. Biochem. Biophys. Res. Commun. 36, Sci. USA 93, 14503–14508 (1996). 21. Nicolas, E. et al. Distinct roles of HDAC complexes in
146–150 (1969). Discovery of Hda1 and Rpd3 as deacetylase promoter silencing, antisense suppression and DNA
3. Candido, E. P., Reeves, R. & Davie, J. R. Sodium subunits of two multiprotein complexes and damage protection. Nature Struct. Mol. Biol. 14,
butyrate inhibits histone deacetylation in cultured identification of related proteins in bacteria. 372–380 (2007).
cells. Cell 14, 105–113 (1978). 11. Grozinger, C. M. & Schreiber, S. L. Deacetylase 22. Sengupta, N. & Seto, E. Regulation of histone
4. Sealy, L. & Chalkley, R. The effect of sodium butyrate enzymes: biological functions and the use of small- deacetylase activities. J. Cell. Biochem. 93, 57–67
on histone modification. Cell 14, 115–121 (1978). molecule inhibitors. Chem. Biol. 9, 3–16 (2002). (2004).
5. Vidali, G., Boffa, L. C., Bradbury, E. M. & Allfrey, V. G. 12. Haigis, M. C. & Guarente, L. P. Mammalian sirtuins — 23. Zhang, X. et al. Histone deacetylase 3 (HDAC3)
Butyrate suppression of histone deacetylation leads to emerging roles in physiology, aging, and calorie activity is regulated by interaction with protein serine/
accumulation of multiacetylated forms of histones H3 restriction. Genes Dev. 20, 2913–2921 (2006). threonine phosphatase 4. Genes Dev. 19, 827–839
and H4 and increased DNase I sensitivity of the 13. de Ruijter, A. J., van Gennip, A. H., Caron, H. N., (2005).
associated DNA sequences. Proc. Natl Acad. Sci. USA Kemp, S. & van Kuilenburg, A. B. Histone deacetylases 24. Lee, H., Rezai-Zadeh, N. & Seto, E. Negative
75, 2239–2243 (1978). (HDACs): characterization of the classical HDAC family. regulation of histone deacetylase 8 activity by cyclic
6. Yoshida, M., Kijima, M., Akita, M. & Beppu, T. Biochem. J. 370, 737–749 (2003). AMP-dependent protein kinase A. Mol. Cell. Biol. 24,
Potent and specific inhibition of mammalian histone 14. Khochbin, S., Verdel, A., Lemercier, C. & Seigneurin- 765–773 (2004).
deacetylase both in vivo and in vitro by trichostatin A. Berny, D. Functional significance of histone 25. Vannini, A. et al. Crystal structure of a eukaryotic zinc-
J. Biol. Chem. 265, 17174–17179 (1990). deacetylase diversity. Curr. Opin. Genet. Dev. 11, dependent histone deacetylase, human HDAC8,
Identified the first specific HDAC inhibitor and 162–166 (2001). complexed with a hydroxamic acid inhibitor. Proc. Natl
linked HDAC inhibition to cell growth arrest. 15. Verdin, E., Dequiedt, F. & Kasler, H. G. Class II histone Acad. Sci. USA 101, 15064–15069 (2004).
7. Taunton, J., Hassig, C. A. & Schreiber, S. L. deacetylases: versatile regulators. Trends Genet. 19, 26. Yang, X. J. & Grégoire, S. Class II histone deacetylases:
A mammalian histone deacetylase related to the yeast 286–293 (2003). from sequence to function, regulation and clinical
transcriptional regulator Rpd3p. Science 272, 16. Gregoretti, I. V., Lee, Y. M. & Goodson, H. V. Molecular implication. Mol. Cell. Biol. 25, 2873–2884
408–411 (1996). evolution of the histone deacetylase family: functional (2005).
Affinity purification and identification of the first implications of phylogenetic analysis. J. Mol. Biol. 27. Han, A., He, J., Wu, Y., Liu, J. O. & Chen, L.
HDAC, supporting the link between gene silencing 338, 17–31 (2004). Mechanism of recruitment of class II histone
and histone deacetylation. An elegant phylogenetic analysis about different deacetylases by myocyte enhancer factor‑2. J. Mol.
8. Vidal, M. & Gaber, R. F. RPD3 encodes a second factor groups of HDACs. Biol. 345, 91–102 (2005).
required to achieve maximum positive and negative 17. Goldberg, A. D., Allis, C. D. & Bernstein, E. 28. Fischle, W. et al. Enzymatic activity associated with
transcriptional states in Saccharomyces cerevisiae. Epigenetics: a landscape takes shape. Cell 128, class II HDACs is dependent on a multiprotein complex
Mol. Cell. Biol. 11, 6317–6327 (1991). 635–638 (2007). containing HDAC3 and SMRT/N-CoR. Mol. Cell 9,
9. Yang, W. M., Inouye, C., Zeng, Y., Bearss, D. & Seto, E. 18. Millar, C. B. & Grunstein, M. Genome-wide patterns of 45–57 (2002).
Transcriptional repression by YY1 is mediated by histone modifications in yeast. Nature Rev. Mol. Cell. 29. Lamb, N. et al. Unraveling the hidden catalytic
interaction with a mammalian homolog of the yeast Biol. 7, 657–666 (2006). activity of vertebrate class IIa histone deacetylases.
global regulator RPD3. Proc. Natl Acad. Sci. USA 93, 19. Marks, P. A. & Breslow, R. Dimethyl sulfoxide to Proc. Natl Acad. Sci. USA 104, 17355–17340
12845–12850 (1996). vorinostat: development of this histone deacetylase (2007).

REVIEWS
30. Moreth, K. et al. An active site tyrosine residue is 55. Lee, M. G. et al. Functional interplay between histone 78. Parra, M., Kasler, H., McKinsey, T. A., Olson, E. N. &
essential for amidohydrolase but not for esterase demethylase and deacetylase enzymes. Mol. Cell. Biol. Verdin, E. Protein kinase D1 phosphorylates HDAC7
activity of a class 2 histone deacetylase-like bacterial 26, 6395–6402 (2006). and induces its nuclear export after TCR activation.
enzyme. Biochem. J. 401, 659–665 (2007). 56. Wen, Y. D. et al. The histone deacetylase‑3 complex J. Biol. Chem. 280, 13762–13770 (2005).
31. Boyault, C., Sadoul, K., Pabion, M. & Khochbin, S. contains nuclear receptor corepressors. Proc. Natl 79. Dequiedt, F. et al. Phosphorylation of histone
HDAC6, at the crossroads between cytoskeleton and Acad. Sci. USA 97, 7202–7207 (2000). deacetylase 7 by protein kinase D mediates T cell
cell signaling by acetylation and ubiquitination. 57. Zhang, J., Kalkum, M., Chait, B. T. & Roeder, R. G. receptor-induced Nur77 expression and apoptosis.
Oncogene 26, 5468–5476 (2007). The N‑CoR‑HDAC3 nuclear receptor corepressor J. Exp. Med. 201, 793–804 (2005).
32. Gao, L., Cueto, M. A., Asselbergs, F. & Atadja, P. complex inhibits the JNK pathway through the 80. Chang, S., Bezprozvannaya, S., Li, S. & Olson, E. N.
Cloning and functional characterization of HDAC11, a integral subunit GPS2. Mol. Cell 9, 611–623 An expression screen reveals modulators of class II
novel member of the human histone deacetylase (2002). histone deacetylase phosphorylation. Proc. Natl Acad.
family. J. Biol. Chem. 277, 25748–25755 (2002). 58. Karagianni, P. & Wong, J. HDAC3: taking the Sci. USA 102, 8120–8125 (2005).
33. Somoza, J. R. et al. Structural snapshots of human SMRT‑N‑CoRrect road to repression. Oncogene 26, 81. Berdeaux, R. et al. SIK1 is a class II HDAC kinase that
HDAC8 provide insights into the class I histone 5439–5449 (2007). promotes survival of skeletal myocytes. Nature Med.
deacetylases. Structure 12, 1325–1334 (2004). 59. Guenther, M. G. et al. A core SMRT corepressor 13, 597–603 (2007).
34. Finnin, M. S. et al. Structures of a histone deacetylase complex containing HDAC3 and TBL1, a WD40-repeat 82. Kim, M. A. et al. Identification of novel substrates for
homologue bound to the TSA and SAHA inhibitors. protein linked to deafness. Genes Dev. 14, human checkpoint kinase Chk1 and Chk2 through
Nature 401, 188–193 (1999). 1048–1057 (2000). genome-wide screening using a consensus Chk
35. Nielsen, T. K., Hildmann, C., Dickmanns, A., 60. Li, J. et al. Both corepressor proteins SMRT and phosphorylation motif. Exp. Mol. Med. 39, 205–212
Schwienhorst, A. & Ficner, R. Crystal structure of a N‑CoR exist in large protein complexes containing (2007).
bacterial class 2 histone deacetylase homologue. HDAC3. EMBO J. 19, 4342–4350 (2000). 83. Parra, M., Mahmoudi, T. & Verdin, E. Myosin
J. Mol. Biol. 354, 107–120 (2005). 61. Underhill, C., Qutob, M. S., Yee, S. P. & Torchia, J. phosphatase dephosphorylates HDAC7, controls its
36. Boyer, L. A., Latek, R. R. & Peterson, C. L. The SANT A novel nuclear receptor corepressor complex, N‑CoR, nucleocytoplasmic shuttling, and inhibits apoptosis in
domain: a unique histone‑tail‑binding module? Nature contains components of the mammalian SWI/SNF thymocytes. Genes Dev. 21, 638–643 (2007).
Rev. Mol. Cell. Biol. 5, 158–163 (2004). complex and the corepressor KAP‑1. J. Biol. Chem. Shows that a PP1 phosphatase complex is directly
37. Kasten, M. M., Dorland, S. & Stillman, D. J. A large 275, 40463–40470 (2000). involved in controlling the nucleocytoplasmic
protein complex containing the yeast Sin3p and 62. Guenther, M. G., Barak, O. & Lazar, M. A. The SMRT trafficking of HDAC7.
Rpd3p transcriptional regulators. Mol. Cell. Biol. 17, and N‑CoR corepressors are activating cofactors for 84. Sucharov, C. C., Langer, S., Bristow, M. & Leinwand, L.
4852–4858 (1997). histone deacetylase 3. Mol. Cell. Biol. 21, 6091–6101 Shuttling of HDAC5 in H9C2 cells regulates YY1
38. Keogh, M. C. et al. Cotranscriptional Set2 methylation (2001). function through CaMKIV/PKD and PP2A.
of histone H3 lysine 36 recruits a repressive Rpd3 63. Wu, J., Carmen, A. A., Kobayashi, R., Suka, N. & Am. J. Physiol. Cell Physiol. 291, C1029–C1037
complex. Cell 123, 593–605 (2005). Grunstein, M. HDA2 and HDA3 are related proteins (2006).
39. Carrozza, M. J. et al. Histone H3 methylation by Set2 that interact with and are essential for the activity of 85. Illi, B. et al. Nitric oxide modulates chromatin folding
directs deacetylation of coding regions by Rpd3S to the yeast histone deacetylase HDA1. Proc. Natl Acad. in human endothelial cells via protein phosphatase 2A
suppress spurious intragenic transcription. Cell 123, Sci. USA 98, 4391–4396 (2001). activation and class II histone deacetylases nuclear
581–592 (2005). 64. Sugiyama, T. et al. SHREC, an effector complex for shuttling. Circ. Res. 102, 51–58 (2007).
40. Colina, A. R. & Young, D. Raf60, a novel component of heterochromatic transcriptional silencing. Cell 128, 86. Kirsh, O. et al. The SUMO E3 ligase RanBP2
the Rpd3 histone deacetylase complex required for 491–504 (2007). promotes modification of the HDAC4 deacetylase.
Rpd3 activity in Saccharomyces cerevisiae. J. Biol. 65. Grozinger, C. M. & Schreiber, S. L. Regulation of EMBO J. 21, 2682–2691 (2002).
Chem. 280, 42552–42556 (2005). histone deacetylase 4 and 5 transcriptional activity by 87. Petrie, K. et al. The histone deacetylase 9 gene
41. Kadosh, D. & Struhl, K. Repression by Ume6 involves 14‑3‑3-dependent cellular localization. Proc. Natl encodes multiple protein isoforms. J. Biol. Chem. 278,
recruitment of a complex containing Sin3 corepressor Acad. Sci. USA 97, 7835–7840 (2000). 16059–16072 (2003).
and Rpd3 histone deacetylase to target promoters. 66. Paroni, G. et al. PP2A regulates HDAC4 nuclear 88. Tatham, M. H. et al. Polymeric chains of SUMO‑2 and
Cell 89, 365–371 (1997). import. Mol. Biol. Cell 19, 655–667 (2008). SUMO‑3 are conjugated to protein substrates by
42. Joshi, A. A. & Struhl, K. Eaf3 chromodomain 67. Seigneurin-Berny, D. et al. Identification of SAE1/SAE2 and Ubc9. J. Biol. Chem. 276,
interaction with methylated H3‑K36 links histone components of the murine histone deacetylase 6 35368–35374 (2001).
deacetylation to Pol II elongation. Mol. Cell 20, complex: link between acetylation and ubiquitination 89. Paroni, G. et al. Caspase-dependent regulation of
971–978 (2005). signaling pathways. Mol. Cell. Biol. 21, 8035–8044 histone deacetylase 4 nuclear-cytoplasmic shuttling
43. Li, B. et al. Combined action of PHD and (2001). promotes apoptosis. Mol. Biol. Cell 15, 2804–2818
chromodomains directs the Rpd3S HDAC to 68. Zhang, X. et al. HDAC6 modulates cell motility by (2004).
transcribed chromatin. Science 316, 1050–1054 altering the acetylation level of cortactin. Mol. Cell 27, 90. Liu, F., Dowling, M., Yang, X. J. & Kao, G. D. Caspase-
(2007). 197–213 (2007). mediated specific cleavage of human histone
44. Ahringer, J. NuRD and SIN3 histone deacetylase 69. McKinsey, T. A., Zhang, C. L., Lu, J. & Olson, E. N. deacetylase 4. J. Biol. Chem. 279, 34537–34546
complexes in development. Trends Genet. 16, Signal-dependent nuclear export of a histone (2004).
351–356 (2000). deacetylase regulates muscle differentiation. Nature 91. Chen, J. F. et al. The role of microRNA‑1 and
45. Shi, X. et al. ING2 PHD domain links histone H3 lysine 408, 106–111 (2000). microRNA‑133 in skeletal muscle proliferation and
4 methylation to active gene repression. Nature 442, First report on the kinase-mediated nuclear export differentiation. Nature Genet. 38, 228–233 (2006).
96–99 (2006). of a class IIa HDAC. 92. Haberland, M. et al. Regulation of HDAC9 gene
46. Le Guezennec, X., Vermeulen, M. & Stunnenberg, 70. McKinsey, T. A. & Olson, E. N. Cardiac histone expression by MEF2 establishes a negative-feedback
H. G. Molecular characterization of Sin3 PAH-domain acetylation — therapeutic opportunities abound. loop in the transcriptional circuitry of muscle
interactor specificity and identification of PAH Trends Genet. 20, 206–213 (2004). differentiation. Mol. Cell. Biol. 27, 518–525 (2007).
partners. Nucleic Acids Res. 34, 3929–3937 (2006). 71. Martin, M., Kettmann, R. & Dequiedt, F. Class IIa 93. Hubbert, C. et al. HDAC6 is a microtubule-associated
47. Denslow, S. A. & Wade, P. A. The human Mi‑2/NuRD histone deacetylases: regulating the regulators. deacetylase. Nature 417, 455–458 (2002).
complex and gene regulation. Oncogene 26, Oncogene 26, 5450–5467 (2007). First report on the unexpected α‑tubulin
5433–5438 (2007). 72. Zhang, C. L., McKinsey, T. A. & Olson, E. N. deacetylase activity of HDAC6.
48. Klose, R. J. & Bird, A. P. Genomic DNA methylation: The transcriptional corepressor MITR is a signal- 94. Matsuyama, A. et al. In vivo destabilization of dynamic
the mark and its mediators. Trends Biochem. Sci. 31, responsive inhibitor of myogenesis. Proc. Natl Acad. microtubules by HDAC6-mediated deacetylation.
89–97 (2006). Sci. USA 98, 7354–7359 (2001). EMBO J. 21, 6820–6831 (2002).
49. Andres, M. E. et al. CoREST: a functional corepressor 73. van der Linden, A. M., Nolan, K. M. & Sengupta, P. 95. Zhang, Y. et al. HDAC‑6 interacts with and
required for regulation of neural-specific gene KIN‑29 SIK regulates chemoreceptor gene expression deacetylates tubulin and microtubules in vivo.
expression. Proc. Natl Acad. Sci. USA 96, via an MEF2 transcription factor and a class II HDAC. EMBO J. 22, 1168–1179 (2003).
9873–9878 (1999). EMBO J. 26, 358–370 (2007). 96. Tran, A. D. et al. HDAC6 deacetylation of tubulin
50. You, A., Tong, J. K., Grozinger, C. M. & Schreiber, S. L. 74. Li, X., Song, S., Liu, Y., Ko, S. H. & Kao, H. Y. modulates dynamics of cellular adhesions. J. Cell Sci.
CoREST is an integral component of the CoREST– Phosphorylation of the histone deacetylase 7 120, 1469–1479 (2007).
human histone deacetylase complex. Proc. Natl Acad. modulates its stability and association with 14‑3‑3 97. Serrador, J. M. et al. HDAC6 deacetylase activity links
Sci. USA 98, 1454–1458 (2001). proteins. J. Biol. Chem. 279, 34201–34208 the tubulin cytoskeleton with immune synapse
51. Humphrey, G. W. et al. Stable histone deacetylase (2004). organization. Immunity 20, 417–428 (2004).
complexes distinguished by the presence of SANT 75. Potthoff, M. J. et al. Histone deacetylase degradation 98. Reed, N. A. et al. Microtubule acetylation promotes
domain proteins CoREST/kiaa0071 and Mta‑L1. and MEF2 activation promote the formation of slow- kinesin‑1 binding and transport. Curr. Biol. 16,
J. Biol. Chem. 276, 6812–6824 (2001). twitch myofibers. J. Clin. Invest. 117, 2459–2467 2166–2172 (2006).
52. Hakimi, M. A. et al. A core-BRAF35 complex (2007). 99. Dompierre, J. P. et al. Histone deacetylase 6 inhibition
containing histone deacetylase mediates repression of 76. Linseman, D. A. et al. Inactivation of the myocyte compensates for the transport deficit in Huntington’s
neuronal-specific genes. Proc. Natl Acad. Sci. USA 99, enhancer factor‑2 repressor histone deacetylase‑5 by disease by increasing tubulin acetylation. J. Neurosci.
7420–7425 (2002). endogenous Ca2+/calmodulin-dependent kinase II 27, 3571–3583 (2007).
53. Shi, Y. et al. Histone demethylation mediated by the promotes depolarization-mediated cerebellar granule 100. Pugacheva, E. N., Jablonski, S. A., Hartman, T. R.,
nuclear amine oxidase homolog LSD1. Cell 119, neuron survival. J. Biol. Chem. 278, 41472–41481 Henske, E. P. & Golemis, E. A. HEF1-dependent
941–953 (2004). (2003). Aurora A activation induces disassembly of the
54. Lee, M. G., Wynder, C., Cooch, N. & Shiekhattar, R. 77. Vega, R. B. et al. Protein kinases C and D mediate primary cilium. Cell 129, 1351–1363 (2007).
An essential role for CoREST in nucleosomal histone 3 agonist-dependent cardiac hypertrophy through 101. Linding, R. et al. Systematic discovery of in vivo
lysine 4 demethylation. Nature 437, 432–435 nuclear export of histone deacetylase 5. Mol. Cell. phosphorylation networks. Cell 129, 1415–1426
(2005). Biol. 24, 8374–8385 (2004). (2007).

REVIEWS
102. Kovacs, J. J. et al. HDAC6 regulates Hsp90 acetylation 121. Foglietti, C. et al. Dissecting the biological functions of 144. Cowley, S. M. et al. The mSin3A chromatin-modifying
and chaperone-dependent activation of glucocorticoid Drosophila histone deacetylases by RNA interference complex is essential for embryogenesis and T‑cell
receptor. Mol. Cell 18, 601–607 (2005). and transcriptional profiling. J. Biol. Chem. 281, development. Mol. Cell. Biol. 25, 6990–7004
103. Bali, P. et al. Inhibition of histone deacetylase 6 17968–17976 (2006). (2005).
acetylates and disrupts the chaperone function of heat 122. Senese, S. et al. Role for histone deacetylase 1 in 145. David, G., Turner, G. M., Yao, Y., Protopopov, A. &
shock protein 90: a novel basis for antileukemia human tumor cell proliferation. Mol. Cell. Biol. 27, DePinho, R. A. mSin3-associated protein, mSds3, is
activity of histone deacetylase inhibitors. J. Biol. 4784–4795 (2007). essential for pericentric heterochromatin formation
Chem. 280, 26729–26734 (2005). 123. Iwabata, H., Yoshida, M. & Komatsu, Y. Proteomic and chromosome segregation in mammalian cells.
104. Murphy, P. J., Morishima, Y., Kovacs, J. J., Yao, T. P. & analysis of organ-specific post-translational lysine- Genes Dev. 17, 2396–2405 (2003).
Pratt, W. B. Regulation of the dynamics of hsp90 acetylation and -methylation in mice by use of anti- 146. Williams, C. J. et al. The chromatin remodeler Mi-2β is
action on the glucocorticoid receptor by acetylation/ acetyllysine and ‑methyllysine mouse monoclonal required for CD4 expression and T cell development.
deacetylation of the chaperone. J. Biol. Chem. 280, antibodies. Proteomics 5, 4653–4664 (2005). Immunity 20, 719–733 (2004).
33792–33799 (2005). 124. Kim, S. C. et al. Substrate and functional diversity of 147. Kaji, K. et al. The NuRD component Mbd3 is required
105. Scroggins, B. T. et al. An acetylation site in the middle lysine acetylation revealed by a proteomics survey. for pluripotency of embryonic stem cells. Nature Cell
domain of Hsp90 regulates chaperone function. Mol. Cell 23, 607–618 (2006). Biol. 8, 285–292 (2006).
Mol. Cell 25, 151–159 (2007). 125. Xie, H., Bandhakavi, S., Roe, M. R. & Griffin, T. J. 148. Marino, S. & Nusse, R. Mutants in the mouse NuRD/
References 103–105 demonstrate that HDAC6 Preparative peptide isoelectric focusing as a tool for Mi2 component P66α are embryonic lethal. PLoS
binds to and deacetylates HSP90, thereby improving the identification of lysine-acetylated ONE 2, e519 (2007).
regulating its chaperone activity towards client peptides from complex mixtures. J. Proteome Res. 6, 149. Wang, J. et al. Opposing LSD1 complexes function in
proteins. 2019–2026 (2007). developmental gene activation and repression
106. Kopito, R. R. Aggresomes, inclusion bodies and 126. Jiang, T., Zhou, X., Taghizadeh, K., Dong, M. & programmes. Nature 446, 882–887 (2007).
protein aggregation. Trends Cell Biol. 10, 524–530 Dedon, P. C. N‑formylation of lysine in histone proteins 150. Hermanson, O., Jepsen, K. & Rosenfeld, M. G. N‑CoR
(2000). as a secondary modification arising from oxidative controls differentiation of neural stem cells into
107. Kawaguchi, Y. et al. The deacetylase HDAC6 regulates DNA damage. Proc. Natl Acad. Sci. USA 104, 60–65 astrocytes. Nature 419, 934–939 (2002).
aggresome formation and cell viability in response (2007). 151. Vega, R. B. et al. Histone deacetylase 4 controls
to misfolded protein stress. Cell 115, 727–738 127. Chen, Y. et al. Lysine propionylation and butyrylation chondrocyte hypertrophy during skeletogenesis.
(2003). are novel post-translational modifications in histones. Cell 119, 555–566 (2004).
Demonstrates that HDAC6 has a key role in Mol. Cell. Proteomics 6, 812–819 (2007). 152. Chang, S. et al. Histone deacetylases 5 and 9 govern
aggresome formation, thereby linking this 128. Garrity, J., Gardner, J. G., Hawse, W., Wolberger, C. & responsiveness of the heart to a subset of stress
deacetylase to the cellular management of Escalante-Semerena, J. C. N‑lysine propionylation signals and play redundant roles in heart
misfolded proteins. controls the activity of propionyl-CoA synthetase. development. Mol. Cell. Biol. 24, 8467–8476
108. Boyault, C. et al. HDAC6‑p97/VCP controlled J. Biol. Chem. 282, 30239–30245 (2007). (2004).
polyubiquitin chain turnover. EMBO J. 25, 129. Mukherjee, S., Hao, Y. H. & Orth, K. A newly 153. Chang, S. et al. Histone deacetylase 7 maintains
3357–3366 (2006). discovered post-translational modification — the vascular integrity by repressing matrix
109. Pandey, U. B. et al. HDAC6 rescues neurodegeneration acetylation of serine and threonine residues. Trends metalloproteinase 10. Cell 126, 321–334 (2006).
and provides an essential link between autophagy and Biochem. Sci. 32, 210–216 (2007). References 151 and 153 report the unexpected
the UPS. Nature 447, 859–863 (2007). 130. Yang, X. J. & Grégoire, S. Metabolism, cytoskeleton phenotypes of Hdac4- and Hdac7-null mice,
110. Iwata, A., Riley, B. E., Johnston, J. A. & Kopito, R. R. and cellular signaling in the grip of protein Nε- and thereby revealing the essential roles of these
HDAC6 and microtubules are required for autophagic O‑acetylation. EMBO Rep. 8, 556–561 (2007). deacetylases in bone development and
degradation of aggregated huntingtin. J. Biol. Chem. 131. Mattagajasingh, I. et al. SIRT1 promotes endothelium- maintenance of vascular integrity, respectively.
280, 40282–40292 (2005). dependent vascular relaxation by activating 154. Zhang, C. L. et al. Class II histone deacetylases act as
References 109 and 110 report the unexpected endothelial nitric oxide synthase. Proc. Natl Acad. Sci. signal-responsive repressors of cardiac hypertrophy.
link of HDAC6 to autophagy. USA 104, 14855–14860 (2007). Cell 110, 479–488 (2002).
111. Boyault, C. et al. HDAC6 controls major cell 132. Tang, X. et al. Acetylation-dependent signal transduction
response pathways to cytotoxic accumulation of for type I interferon receptor. Cell 31, 93–105 (2007). Acknowledgements
protein aggregates. Genes Dev. 21, 2172–2181 133. Lambard, D. B. et al. Mammalian Sir2 homolog SIRT3 Owing to strict space constraints and the wealth of literature,
(2007). regulates global mitochondrial lysine acetylation. we have used seminal reviews and supplemental figures and
An elegant study linking HDAC6 to the activation of Mol. Cell. Biol. 27, 8807–8814 (2007). tables to cover some of the primary findings in this field.
the transcription factor HSF1 in response to 134. Seet, B. T., Dikic, I., Zhou, M. M. & Pawson, T. Reading Research in our laboratories has been supported by grants
cytotoxic protein aggregates. protein modifications with interaction domains. from the National Institutes of Health, the American Heart
112. Kwon, S., Zhang, Y. & Matthias, P. The deacetylase Nature Rev. Mol. Cell. Biol. 7, 473–483 (2006). Association and the Kaul Foundation (to E.S.), and from the
HDAC6 is a novel essential component of stress 135. Starai, V. J., Celic, I., Cole, R. N., Boeke, J. D. & Canadian Institutes of Health Research, the National Cancer
granules involved in the stress response. Genes Dev. Escalante-Semerena, J. C. Sir2-dependent activation Institute of Canada, the Natural Sciences and Engineering
21, 3381–3394 (2007). of acetyl-CoA synthetase by deacetylation of active Research Council of Canada and the Canada Foundation for
Describes the discovery of an unexpected link lysine. Science 298, 2390–2392 (2002). Innovation (to X.J.Y.).
between HDAC6 and the formation of stress 136. Gardner, J. G., Grundy, F. J., Henkin, T. M. &
granules. Escalante-Semerena, J. C. Control of acetyl-coenzyme DATABASES
113. Gao, Y. S. et al. Histone deacetylase 6 regulates A synthetase (AcsA) activity by acetylation/ Entrez Protein: http://www.ncbi.nlm.nih.gov/sites/
growth factor-induced actin remodeling and deacetylation without NAD+ involvement in Bacillus entrez?db=protein
endocytosis. Mol. Cell. Biol. 27, 8637–8647 subtilis. J. Bacteriol. 188, 5460–5468 (2006). Hda1 | HDA2 | HDAC1 | HDAC2 | HDAC3 | HDAC4 | HDAC5 |
(2007). 137. Hallows, W. C., Lee, S. & Denu, J. M. Sirtuins HDAC6 | HDAC7 | HDAC8 | HDAC9 | HDAC10 | HDAC11 |
114. Arnold, M. A. et al. MEF2C transcription factor deacetylate and activate mammalian acetyl-CoA Rpd3
controls chondrocyte hypertrophy and bone synthetases. Proc. Natl Acad. Sci. USA 103, Entrez Nucleotide (GenBank): http://www.ncbi.nlm.nih.gov/
development. Dev. Cell 12, 377–389 (2007). 10230–10235 (2006). sites/entrez?db=nucleotide
115. Zhu, P. et al. Induction of HDAC2 expression upon loss 138. Schwer, B., Bunkenborg, J., Verdin, R. O., Andersen, XP_001101619 | XP_001493915
of APC in colorectal tumorigenesis. Cancer Cell 5, J. S. & Verdin, E. Reversible lysine acetylation controls OMIM: http://www.ncbi.nlm.nih.gov/entrez/query.
455–463 (2004). the activity of the mitochondrial enzyme acetyl-CoA fcgi?db=OMIM
116. Ropero, S. et al. A truncating mutation of HDAC2 in synthetase 2. Proc. Natl Acad. Sci. USA 103, Huntington’s disease | Parkinson’s disease
human cancers confers resistance to histone 10224–10229 (2006). RCSB Protein Data Bank: http://www.pdb.org/pdb/home/
deacetylase inhibition. Nature Genet. 38, 566–569 139. Yang, X. J. & Grégoire, S. A recurrent phospho-sumoyl home.do
(2006). switch in transcriptional repression and beyond. 2NVR
117. Bolden, J. E., Peart, M. J. & Johnstone, R. W. Mol. Cell 23, 779–786 (2006). UniProtKB: http://beta.uniprot.org/uniprot
Anticancer activities of histone deacetylase 140. Lagger, G. et al. Essential function of histone CHK1 | Clr3 | Clr6 | Eaf3 | HSP90 | LSD1 | MEF2 | Pho23 | Rco1 |
inhibitors. Nature Rev. Drug Discov. 5, 769–784 deacetylase 1 in proliferation control and CDK inhibitor Sap30 | Sds3 | Set2 | Sin3 | Sir2 | Ume1
(2006). repression. EMBO J. 21, 2672–2681 (2002).
118. Minetti, G. C. et al. Functional and morphological 141. Montgomery, R. L. et al. Histone deacetylases 1 and 2 FURTHER INFORMATION
recovery of dystrophic muscles in mice treated with redundantly regulate cardiac morphogenesis, growth, Xiang-Jiao Yang’s homepage: http://www.mcgill.ca/mog/
deacetylase inhibitors. Nature Med. 12, 1147–1150 and contractility. Genes Dev. 21, 1790–1802 research/yang/
(2006). (2007). Edward Seto’s homepage: http://hsc.usf.edu/COM/iop/
119. Avila, A. M. et al. Trichostatin A increases SMN 142. Zimmermann, S. et al. Reduced body size and cancer/seto.html
expression and survival in a mouse model of spinal decreased intestinal tumor rates in HDAC2-mutant
muscular atrophy. J. Clin. Invest. 117, 659–671 mice. Cancer Res. 67, 9047–9054 (2007). SUPPLEMENTARY INFORMATION
(2007). 143. Dannenberg, J. H. et al. mSin3A corepressor See online article: S1 (figure) | S2 (figure) | S3 (figure) | S4
120. Tao, R. et al. Deacetylase inhibition promotes the regulates diverse transcriptional networks governing (table) | S5 (table) | S6 (table)
generation and function of regulatory T cells. Nature normal and neoplastic growth and survival. Genes All links are active in the online pdf
Med. 13, 1299–1307 (2007). Dev. 19, 1581–1595 (2005).


The Rpd3/Hda1 Family of Lysine Deacetylases: From Bacteria and Yeast To Mice and Men

Uploaded by

Copyright:

Available Formats

The Rpd3/Hda1 Family of Lysine Deacetylases: From Bacteria and Yeast To Mice and Men

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

The Rpd3/Hda1 Family of Lysine Deacetylases: From Bacteria and Yeast To Mice and Men

Uploaded by

Copyright:

Available Formats

REVIEWS

The Rpd3/Hda1 family of lysine

206 | march 2008 | volume 9 www.nature.com/reviews/molcellbio

nature reviews | molecular cell biology volume 9 | march 2008 | 207

a Class I Rpd3 DAC 433 (S. cerevisiae)

Clr6 66/81% 405 (S. pombe)

HDAC1 64/81% S S 482

b Class II Hda1 DAC (26/47%) 706 (S. cerevisiae)

Clr3 53/71% 687 (S. pombe)

MEF2 14-3-3 14-3-3 14-3-3

HDAC9 S S S 35/51% 1,069

HDAC6 38/60% 44/61% SE14 ZnF 1,215 (H. sapiens)

c Class IV HDAC11 24/42% (31/41%) 347 (H. sapiens)

208 | march 2008 | volume 9 www.nature.com/reviews/molcellbio

nature reviews | molecular cell biology volume 9 | march 2008 | 209

210 | march 2008 | volume 9 www.nature.com/reviews/molcellbio

nature reviews | molecular cell biology volume 9 | march 2008 | 211

a cAMP subunit‑1–protein phosphatase-1β) complex83 and

212 | march 2008 | volume 9 www.nature.com/reviews/molcellbio

HDAC6 DAC1 DAC2 ZnF

α-tubulin Cortactin HSP90 IFNαR? Histones Aggresome formation

nature reviews | molecular cell biology volume 9 | march 2008 | 213

214 | march 2008 | volume 9 www.nature.com/reviews/molcellbio

nature reviews | molecular cell biology volume 9 | march 2008 | 215

216 | march 2008 | volume 9 www.nature.com/reviews/molcellbio

nature reviews | molecular cell biology volume 9 | march 2008 | 217

218 | march 2008 | volume 9 www.nature.com/reviews/molcellbio

You might also like