Báo cáo khoa học: How mammalian transcriptional repressors work Gerald Thiel, Michael Lietz1 and Mathias Hohl Department of Medical Biochemistry and Molecular

doi:10.1111/j.1432-1033.2004.04174.x

Eur. J. Biochem. 271, 2855–2862 (2004) (cid:1) FEBS 2004

R E V I E W A R T I C L E

How mammalian transcriptional repressors work

Gerald Thiel, Michael Lietz1 and Mathias Hohl

Department of Medical Biochemistry and Molecular Biology, University of the Saarland Medical Center, Homburg, Germany

human forms of cancer, are connected with the activities of human repressor proteins indicates that transcriptional repression and gene silencing is essential for maintenance of the cellular integrity of a multicellular organism. The wide range of diseases caused by aberration in transcriptional repression sheds light on the importance of understanding how mammalian transcriptional repressor proteins work.

Keywords: heterochromatin; histone deacetylation; MeCP2; REST; retinoblastoma.

Research on the regulation of transcription in mammals initially focused on the mechanism of transcriptional acti- vation and (cid:1)positive control(cid:2) of gene regulation. In contrast, transcriptional repression and (cid:1)negative control(cid:2) of gene transcription was viewed rather as part of the (cid:1)prokaryotic book of biology(cid:2). However, results obtained in recent years have shown convincingly that transcriptional repression mediated by repressor proteins is a common regulatory mechanism in mammals and may play a key role in many biological processes. In particular, the fact that human diseases, such as Rett and ICF syndromes as well as some

Introduction

restrictive ground state, occluding proteins such as RNA polymerases and transcriptional regulators from binding to DNA [2]. Thus, eukaryotic gene activation is usually understood as a relief of repression by the nucleosomal structure of the chromatin [3] – transcriptional repressors therefore seem to provide a redundant biological activity. Positive regulation also seems more efficient than negative regulation in a multicellular organism and is probably the logical solution. However, economical design principles do not always define cellular regulation mechanisms. Today, we know that transcriptional repression is quite common in eukaryotes, indicating that transcriptional repressors play an important role in the regulation of eukaryotic genes. (cid:1)Repressors may not make sense, but they certainly are common(cid:2) [4]. Here, we will illustrate the biological princi- ples of mammalian transcriptional repressors with exam- ples given for each broad classification without aiming to provide an exhaustive account of known mammalian transcriptional repressors.

importantly,

Passive transcriptional repressors

Transcriptional repressor proteins such as the lac and tryptophan repressors were first discovered in prokaryotes. The DNA-tethered form of the repressor turns genes off by blocking RNA polymerase binding to the promoter or its movement along the DNA. Thus, competition between a gene-specific repressor protein and RNA polymerase is the most common form of transcriptional regulation in Escherichia coli. In contrast, many eukaryotic transcrip- tional activator proteins have been characterized in the last two decades, suggesting that positive regulation via activa- tor proteins is the predominant mode of gene control in eukaryotes. Accordingly, the question of how mammalian transcriptional activators work has been extensively deba- ted [1]. Gene regulation in prokaryotes and eukaryotes differs substantially in many aspects, including the function of genetic elements, the coupling or separation of tran- the scription and translation, and, most structure of the genetic material. The RNA polymerase in bacteria generally has access to the promoters, thus defining a nonrestrictive transcriptional ground state in bacteria. In mammals, however, the chromatin structure causes a

Mammalian transcriptional repressors can be classified as passive or active repressor proteins [5]. Passive repressor proteins do not have intrinsic repressing activity or a portable repression domain. Rather, these proteins repress RNA synthesis by competing with transcriptional activators for DNA binding, by forming inactive heterodimers with transcriptional activators rendering them incapable of interaction with DNA, or by binding to coactivators required for the transcriptional activator proteins. Thus, passive repressor proteins transmit their biological function either via DNA- or protein–protein interactions.

Inducible cAMP early repressor

The (cid:1)inducible cAMP early repressor(cid:2) (ICER) is an example of a passive repressor protein [6]. ICER is one of several

Correspondence to G. Thiel, Department of Medical Biochemistry and Molecular Biology, Building 44, University of Saarland Medical Center, D-66421 Homburg, Germany. Fax: + 49 6841 1626500, Tel.: + 49 6841 1626506, E-mail: gerald.thiel@uniklinik-saarland.de Abbreviations: BDNF, brain-derived neurotrophic factor; CREB, cAMP response element binding protein; CREM, cAMP response element modifier; ICER, inducible cAMP early repressor; heterochromatin, protein 1 (HP1). 1Present address: Department of Molecular Cardiovascular Research, University Hospital, Rheinisch-Westfa¨ lische Technische Hochschule, D-52074 Aachen, Germany. (Received 12 February 2004, revised 25 March 2004, accepted 20 April 2004)

2856 G. Thiel et al. (Eur. J. Biochem. 271)

(cid:1) FEBS 2004

ICER counteracts the biological activity of CREB to protect neurons against cytotoxic challenges. Accordingly, adenovirus-evoked expression of ICER induced pro- grammed cell death, probably via impairment of CREB- mediated transcription of antiapoptotic genes [8].

bZIP transcription factor family

It is not surprising that nature’s method of repression of bZIP proteins by dominant-negative mutants has been widely copied in research laboratories to investigate the the bZIP function of distinct members of biological transcription factor family [9,10]. Remarkable tools for the study of basic region leucine zipper protein composition and activity in living cells are the A-ZIP mutants (Fig. 1B) – amphipathic molecules that contain an acidic region instead of the natural basic domain N-terminal to the leucine zipper domain. This acidic extension of the leucine zipper forms a heterodimeric coiled-coil structure with the basic region of its target that is more stable than the bZIP dimer bound to DNA. These molecules inhibit DNA-binding of the wild- type bZIP proteins in a leucine zipper-dependent fashion. The use of these molecules in molecular medical research was elegantly demonstrated by the generation of transgenic mice expressing an A-ZIP mutant in adipose tissue that inhibits both C/EBP and AP-1 transcriptional activity [11]. These transgenic mice had no white fat tissue throughout life and may be used as a mouse model for the human disease lipoatrophic diabetes (Seip–Berardinelli syndrome).

Sp1-like transcriptional repressors

Nature’s design of passive repressors is not restricted to the bZIP family of transcription factors, but is rather a common scheme in the control of transcriptional activation. The Sp1-like zinc finger proteins, for instance, compete for a common GC-rich DNA-binding site [12]. The balance between Sp1, a transcriptional activator, and Sp1-like transcriptional repressors decides which discrete sets of genes are turned on or off.

products encoded by the cAMP response element modifier (CREM) gene. Additional splice variants encode both transcriptional activators and repressors. The modular structure of ICER includes a basic region leucine zipper domain necessary for dimerization and DNA binding, but no transcriptional activation domain (Fig. 1A). ICER represses transcription mediated by the cAMP response element via blocking of the DNA-binding site using its basic region. Additionally, ICER may induce the formation of nonfunctional heterodimers with wild-type basic region leucine zipper proteins. A robust induction of ICER has been reported after partial hepatectomy, indicating that inhibition of cAMP response element binding protein (CREB)/CREM-mediated gene transcription is part of the tissue repair program of liver [7]. In the nervous system,

Fig. 1. Passive transcriptional repressors. (A) Modular structure of CREM and ICER. The basic region leucine zipper (bZIP) domains, the two glutamine-rich activation domains, and the kinase-inducible domain (KID) are depicted. The phosphorylation site for cAMP- dependent protein kinase within the CREM protein is indicated. ICER shares with CREM the bZIP domain, but not the constitutive and phoshorylation-mediated activation domains. (B) Mechanism of transcriptional repression by A-ZIP proteins. These mutants retain the leucine zipper required for dimerization. The basic DNA-binding domain, however, is exchanged for a domain containing acidic resi- dues. Dimerization between a wild-type bZIP protein with an A-ZIP repressor molecule inhibits DNA-binding of the wild-type bZIP proteins due to the blockage of the DNA-binding domain. (C) IjB repressed NFjB activity by preventing nuclear translocation of NFjB. NFjB, shown here as a p50–p65 dimer, is immobilized in the cytosol by IjB. To relieve this repression, IjB needs to be phosphorylated by the IjB kinase complex. Ubiquitination of phosphorylated IjB signals proteolytic degradation by the proteasome. NFjB translocates to the nucleus and binds to consensus jB sequences.

Mammalian transcriptional repressors (Eur. J. Biochem. 271) 2857

(cid:1) FEBS 2004

IjB

In a broad sense, IjB can also be classified as a passive repressor because it sequesters the NFjB transcription factor complex in the cytosol (Fig. 1C). Upon cellular including inflammatory cytokines, bacterial stimulation, lipopolysaccharides or phorbol esters, IjB becomes phos- phorylated by the IjB kinase complex. Phosphorylated IjB serves as the target for ubiquitination, resulting in proteo- lytic degradation of IjB by the proteasome [13]. As a result, NFjB translocates into the nucleus and activates NFjB- responsive genes.

Active transcriptional repressors

Active mammalian transcriptional repressor proteins have an intrinsic repression activity that targets the chromatin organization of the genome. This type of transcriptional repression is activator-independent and functions over long distances. Two types of active transcriptional repression can be distinguished: transcriptional repression via histone deacetylation, and gene silencing via histone methylation and heterochromatin formation.

Transcriptional repression via recruitment of histone deacetylases

recruitment of histone deacetylases promoter-specific [15,16]. Histone deacetylases may bind directly to the repressor protein or via adaptor proteins such as mSin3A or the nuclear corepressors N-CoR or SMRT. The mamma- lian proteins REST, Rb, and MeCP2 are examples of active transcriptional repressors that target histone deacetylases as a primary means of transcriptional repression, showing that a wide range of biological functions such as neuron-specific gene transcription (via REST), control of the cell cycle (via Rb) or the functional implications of DNA methylation (via MeCP2) are controlled by the same molecular mechanism. The zinc finger protein RE-1 silencing transcription factor (REST) is mainly expressed in non-neuronal cell types and functions as a negative regulator of neuron- specific gene transcription [17]. REST recruits histone deacetylases to neuronal genes, resulting in removal of acetyl groups from the core histones [18]. Consequently, the neuronal genes are embedded into more tightly packed chromatin that is inaccessible for transcriptional activators (Fig. 2, bottom). In neurons, REST is expressed at extremely low concentrations. As a result, the chromatin has an open configuration, allowing transcriptional activa- tors to bind and initiate transcription of neuronal genes (Fig. 2, top). Therefore, gene transcription of neuronal genes is the result of a relief of repression. Active control of neuronal gene expression can be attained by prevention of gene transcription in non-neuronal cells via the continuous presence of a transcriptional repressor. Accordingly, the REST-responsive synapsin I promoter directs tissue-unspe- cific transcription following deletion of the REST binding site [19].

The retinoblastoma protein Rb plays a key role in the regulation of cell cycle progression from G1 to S phase. Accordingly, mutations altering the functional properties of Rb are connected with a dysregulation of cell cycle control. The regulatory role of Rb is based on its interaction with E2F, a transcription factor that controls several genes encoding essential proteins for the progression from G1 to the S phase of the cell cycle. Rb is able to mask the transactivation domain of E2F (passive repression) and additionally recruits the histone deacetylase HDAC1 to the E2F-site-containing transcription units via protein–protein interactions (active repression) [20–22]. Thus, Rb transmits active repression to E2F-responsive genes by modifying the chromatin architecture. The balance between histone acetylation and deacetylation is therefore a key element in the control of the cell cycle.

A critical determinant in the regulation of eukaryotic genes is the structural organization of DNA in chromatin. The fundamental unit of chromatin is the nucleosome with two molecules each of the histones H2A, H2B, H3 and H4 building the core histone and (cid:1) 146 base pairs of DNA wrapped 1.65 turns around the histone octamer. The single nucleosomes are linked by short stretches of DNA and the linker histone H1. The N-terminal regions of the core histones are often modified by acetylation, methylation or phosphorylation. The acetylation of histones involves the transfer of an acetyl group from acetyl coenzyme A to the e-amino group of a lysine residue. Histone acetylation is of major importance for the regulation of gene transcription because this modification reduces the net positive charges of the core histones, leading to a decrease in their binding affinity for DNA. The termini are displaced subsequently from the nucleosome, the nucleosome unfolds and provides access for transcription factors. Thus, transcriptional acti- vation occurs only after the repressive histone–DNA interaction has been destabilized by histone acetylases [14]. in Deacetylation of histones by histone deacetylases, contrast, removes the acetyl group from the e-amino group of lysine residues of histones allowing ionic interactions between the negatively charged DNA phosphate backbone and the positively charged amino termini of the core histones. This results in a more compact chromatin structure that is not easily accessible for the transcriptional machinery. While histone acetylation and hyperacetylation has been correlated with transcriptionally active chromatin, histone deacetylation is thought to be involved in repression of transcription. Histone acetylation and deacetylation are major regulatory mechanisms of transcription that function by modulating the accessibility of transcription factors to their binding site on DNA. Consequently, a common feature of mammalian transcriptional repressors is the

DNA in vertebrates is commonly subjected to methyla- tion of the C-5 position of cytosine nucleotides within the dinucleotide CmpG. This modification of DNA is associated primarily with transcriptional repression. DNA methylation may affect gene expression by interfering with binding of sequence-specific transcription factors leading to an impair- ment of gene transcription. For instance, methylation of the promoter region of the p21WAF1/Cip1 gene, that encodes a cyclin-dependent protein kinase inhibitor, blocks DNA damage-induced p21WAF1/Cip1 synthesis [23]. DNA methy- lation may also influence the chromatin architecture directly by transforming the methylated DNA sequences into a condensed state. Methylated DNA interacts with methy- lation-specific transcriptional repressor proteins, most prominently the MeCP2 protein that binds to single,

2858 G. Thiel et al. (Eur. J. Biochem. 271)

(cid:1) FEBS 2004

methyl-CmpG base pair regardless of the sequence context via a methyl-CpG-binding domain [24]. The fact that CpG methylation occurs throughout the mammalian genome led to the hypothesis that MeCP2 may function as a transcrip- tional noise reduction device filtering out transcriptional noise coming from inappropriate or spurious promoters [25]. MeCP2 represses transcription via recruitment of histone deactylases to methylated DNA regions [26,27], leading to a denser packing of the chromatin and generating an unfavorable environment for transcription. Therefore, a direct causal relationship exists between DNA methylation- mediated transcriptional repression and the histone acety- lation status [28]. Recently, the gene encoding brain-derived neurotrophic factor (BDNF), a neurotrophin, has been identified as a genuine target gene of MeCP2 [29,30]. BDNF belongs to the neurotrophin family of proteins that are important regulatory molecules for the development and differentiation, survival, structure and function of neurons. Inhibitors of histone deacetylases such as trichostatin A are, however, only able to partially relieve MeCP2-mediated transcriptional repression, indicating that MeCP2 may have a second way to repress transcription aside from the recruitment of histone deacetylases.

Formation of facultative heterochromatin

genome is affected by histone deacetylation induced chro- matin compaction at a specific time point. In contrast, gene silencing makes a larger portion of the genome inaccessible for transcriptional activation, thus acting in a regional manner. The eukaryotic genome is divided into euchro- matin and heterochromatin, the latter representing a relatively inaccessible form of chromatin. Heterochromatic regions exhibit hypoacetylation in comparison to the bulk chromatin, and distinct, heterochromatin-specific binding proteins have been identified. Euchromatic regions that are packed into a compact heterochromatic-like architecture are defined as facultative heterochromatin. These regions are not characterized by large stretches of repetitive sequences characteristic for constitutive heterochromatin. How are facultative heterochromatic domains established and maintained? A stepwise model has been proposed as illustrated in Fig. 3. First, the amino termini of the H3 histones are deacetylated by histone deacetylases and methylated subsequently on Lys9 by methyltransferases. Second, the H3-mLys9 modification functions as a high affinity binding site for the silencing protein, heterochro- matin protein 1 (HP1). Third, HP1 proteins homo- and heterodimerize [31] and promote the formation of a higher- order structure that leads to an extension of the hetero- chromatic region into neighboring chromatin [32]. Two protein families play key roles in the biochemical events leading to gene silencing via heterochromatin formation, the histone methyltransferases and the HP1 proteins. Although methylation of the N-termini of histones H3 and H4 can occur at Arg or Lys residues, the lysine-specific methylases are important in the context of heterochromatin formation

Transcriptional repression via histone deacetylation is gene- specific, i.e. histone deacetylases have to be recruited to selected transcription units either by sequence-specific repressor proteins such as REST or by methylation-specific transcription factors. Therefore, only a portion of the

Fig. 2. The transcriptional repressor REST controls neuron-specific gene transcription via recruitment of histone deacetylases. In neurons, only very low levels of REST can be detected. In fact, a decrease in the REST concentration during neuronal differentiation is most probably the essential prerequisite for transcription of neuronal genes. The chromatin configuration of neuronal genes is open, due to an extensive acetylation of the core histones. Transcriptional activators, as well as the RNA polymerase II complex, can gain access to the DNA and trigger gene transcription. In non- neuronal cells, however, REST is present and binds in a sequence-specific manner to several neuronal genes. Using the mSin3A or CoREST corepressor proteins as adaptors, REST recruits histone deacetylases to the transcription unit. The deacetylation of histones results in an enhanced binding between the histones and the DNA, thus impairing the binding of transcription factors to their specific DNA sequences. Moreover, further compaction of the chromatin is accomplished by inducing interactions between neighboring nucleosomes.

Mammalian transcriptional repressors (Eur. J. Biochem. 271) 2859

(cid:1) FEBS 2004

Fig. 4. Modular structure of heterochromatin protein 1 (HP1) and SUV39H1 histone methyltransferase. The mammalian HP1 proteins have a tripartite domaine structure, consisting of an aminoterminal chromatin organization modifier, or (cid:1)chromo domain(cid:2), a structurally related carboxyterminal (cid:1)chromo shadow domain(cid:2), and a central (cid:1)hinge(cid:2) region. The evolutionarily conserved 60-amino acid chromo domain is a structural component of nuclear proteins involved in remodelling of the chromatin structure and serves as a versatile protein interaction module. The chromo domain of HP1 proteins is required for the high-affinity binding to histone H3 tails methylated on Lys9 (H3mLys9). The chromo shadow domain directs intra- and inter- molecular interactions of HP1 proteins triggering an oligomerization of HP1 proteins. An intact, dimeric chromo shadow domain is necessary for complex formation [52]. SUV39H1 posseses intrinsic histone methyltransferase activity. Mutagenesis studies revealed that the 130-amino acid SET domain and two adjacent cysteine-rich regions are necessary for enzyme activity [34]. The SET domain, named after the proteins Su(var)3–9, Enhancer of zeste and Trithorax, in which this domain was orginally found, is evolutionary conserved from yeast to mammals and is found in proteins that function as (cid:1)chromatin modulators(cid:2).

formation of facultative heterochromatin.

HP1 and SUV39H1 is depicted in Fig. 4. HP1 binds to the H3mLys9 epitope through the (cid:1)chromo domain(cid:2) [35,36]. Subsequent oligomerization of HP1 causes the formation of a higher ordered chromatin compaction. Moreover, SUV39H1 and HP1 are complexed together, allowing the methylation to spread to adjacent nucleosomes. In trans- genic mice lacking both Suv39h1 and Suv39h2-encoding genes, HP1 is no longer able to associate with heterochro- matin although the re-introduction of SUV39H1 methyl- transferase restores this association [36]. Taken together, these data indicate that the (cid:1)SUV39H1-HP1 methylation system(cid:2) [33] represents an essential regulatory mechanism for the establishment and propagation of heterochromatin in mammals. Recent experiments performed with fission yeast or Drosophila indicated a direct role for the RNA interference (RNAi) machinery in heterochromatin assem- bly [37,38]. The fact that formation of heterochromatin is highly conserved between yeast, flies and human suggests that RNAi is most probably involved in the initiation of mammalian heterochromatin formation.

Dual-specific repressors

Transcriptional repression via recruitment of histone deacetylases and gene silencing via HP1-induced hetero- chromatin formation represent two different modes of this sequence of transcriptional repression. However,

[33]. One of the most specific histone modifying enzymes is the histone methyltransferase, SUV39H1, with an exquisite site specificity for Lys9 of histone H3 [34]. Methylation of H3Lys9 provides a high-affinity binding site for the HP1 family of proteins. HP1 plays an essential role in the establishment and maintenance of the transcriptionally silent state of heterochromatin. The modular structure of

(A) Fig. 3. Sequential Recruitment of histone deacetylases to the transcription unit by tran- scriptional repressors such as REST, Rb or by methyl-CpG binding proteins induces deacetylation of histone tails, thus making them suitable substrates for histone methyltransferases such as SUV39H1. These enzymes transfer methyl groups to the e-nitrogen of Lys9 of histone H3 using S-adenosyl-L-methionine as methyl donor. (B) The methylated Lys9 of histone H3 provides a high affinity binding site for HP1. (C) Dimerization and oligomerization of HP1 proteins spreads the compaction of nucleosomes, forming facultative heterochromatin.

2860 G. Thiel et al. (Eur. J. Biochem. 271)

(cid:1) FEBS 2004

biochemical reactions are connected, and, in fact, a dynamic link exists between histone deacetylation and heterochro- matin-mediated gene silencing. Here, we term transcrip- tional repressor proteins that block transcription by using both mechanisms (cid:1)dual-specific repressors(cid:2). One of these proteins is the methyl-CpG-binding protein MeCP2. We have already illustrated that MeCP2 is able to repress transcription by triggering histone deacetylation via recruit- ment of histone deacetylases. However, transcriptional repression by MeCP2 does not exclusively rely on local recruitment of histone deacetylases. Rather, deacetylation of histone (cid:1)tails(cid:2) sets the stage for histone methylation to occur. Connecting both biochemical reactions, MeCP2 was found to associate in vivo with histone methyltransferases that are able to methylate the Lys9 residue of histone H3 [39]. This H3mLys9 modification subsequently recruits the hetero- chromatin inducing protein HP1, the starting point for the establishment of facultative heterochromatin. Thus, MeCP2 links a repressive modification on DNA (CpG methylation) with repressive modifications on histones (removal of acetyl groups followed by methylation) to accomplish transcrip- tional repression and gene silencing in mammals.

Active transcriptional repression by Rb via the recruit- ment of histone deacetylases is accomplished by the (cid:1)pocket(cid:2) domain of Rb (amino acids 379–772) as shown by the fact that mutations within this (cid:1)pocket(cid:2) domain or binding of the human papilloma virus oncoprotein E7 reduces the associ- ation of Rb with histone deacetylases [20–22]. It has been suggested that E2F target genes are regulated during the cell cycle by a cycling recruitment of histone acetylases and deacetylases [40] (Fig. 5, left side). In differentiating cells, however, an irreversible repression occurs, due to the recruitment of the histone methyltransferase SUV39H1 by Rb (Fig. 5, right side). The differentiation of muscle cells into myotubes, for instance, requires Rb, to block prolif- eration of the cells, the prerequisite for the initiation of the differentiation process. As a result, E2F and other prolif- eration associated genes become unresponsive to mitogenic signals. The (cid:1)pocket(cid:2) domain of Rb acts as a recruitment site for the histone methyltransferase SUV39H1 and HP1 [41]. Given the fact that methylation of Lys9 residue of histone H3 cannot take place when this residue is acetylated, the prior recruitment of a histone deacetylase seems necessary to allow SUV39H1-mediated methylation of histones. Thus, histone deacetylases and methyltransferases work in concert in the silencing of euchromatic genes. In fact, a physical interaction between the SUV39H1 histone methyltrans- ferase and histone deacetylases has been demonstrated [42]. As shown for the cyclin E promoter, a natural target for E2F-mediated transcriptional activation and Rb-mediated transcriptional repression, the presence of Rb is required to direct histone H3 methylation and HP1 binding to the promoter. These data support the view that the sequential recruitment of a histone deacetylase, a H3Lys9 histone methyltransferase, and the subsequent binding of HP1 is used to repress euchromatic genes via the formation of facultative heterochromatin.

HP1 – in addition to the ability of forming complexes with histone deacetylases 1 and 2 – suggests that REST may induce gene silencing of adjacent transcriptional units that lack REST binding sites. The involvement of the methyl- CpG binding protein MeCP2 in REST-mediated gene silencing indicates that both the presence of a REST binding site and a specific DNA methylation pattern are essential to silence neuronal genes via the corepressor protein CoREST. It is important to emphasize that the CpG methylation status decides whether the CoREST-mediated silencing machinery is employed on a neuronal gene and its neighbourhood. Lack of CpG methylation impairs binding of MeCP2 to the DNA and the subsequent interaction with CoREST. In this case, active repression of neuronal genes is accomplished by histone deacetylase mediated compaction of chromatin as outlined in Fig. 2.

Collectively,

the biological activity of dual-specific the conclusion that while gene

supports

repressors

Likewise, the neuron-specific gene regulator REST has been proposed to induce gene silencing in addition to the recruitment of histone deacetylases [43]. The observation that the corepressor protein CoREST, that binds to the C-terminal repression domain of REST, is able to attract

Fig. 5. Modus operandi of the dual-specific repressor Rb. Transcrip- tional repression mediated by Rb. Rb physically interacts with the E2F transcription factor, a heterodimer composed of a subunit derived from the E2F family of transcription factors and a subunit belonging to the DP family of proteins. E2F regulates genes encoding proteins that are essential for cell cycle progression into the S phase. Tran- scriptional repression mediated by Rb involves the recruitment of histone deacetylases to transcription units having binding sites for E2F. The acetylation status of histones of E2F-regulated genes is important as histone deacetylase-mediated compaction of the chro- matin leads to transcriptional repression. Cell differentiation requires halting of the cell cycle before the genetic differentiation program is switched on. The fact that Rb plays an essential role in the process of cell differentiation may be explained by the recruitment of SUV39H1 histone methyltransferase to Rb that triggers a methylation of Lys9 of histone H3. Thus, high affinity binding sites for the heterochromatin inducing protein HP1 are generated.

Mammalian transcriptional repressors (Eur. J. Biochem. 271) 2861

(cid:1) FEBS 2004

diminished methylation-associated gene silencing. The histone deacetylase inhibitor valproic acid is also used as an established drug in the long term therapy of epilepsy [51]. Most probably, specific histone methyl- transferase inhibitors will complement the repertoire of (cid:1)transcriptional repressor drugs(cid:2). Collectively, the prelim- inary results obtained with these compounds in treatment of diseases caused by aberration in transcriptional repres- sion shed light on the importance of understanding of the molecular mechanisms of negative gene regulation in mammals.

silencing has been defined as a regional, rather than a promoter-specific event, the binding of those repressors to a distinct sequence or to methylated CpG nucleotides of a transcription unit functions as starting point to recruit the molecular machinery that is able to impose gene silencing over a chromosomal region that is inaccessible for DNA binding proteins. Thus, gene silencing may start as a promoter-specific event. Moreover, HP1, the inducer and maintainer of the heterochromatin structure, is in a very dynamic equilibrium with unbound HP1, suggesting that silenced genes may be released from the heterochomatin and may be silenced again after reception of the appro- priate signal.

Acknowledgements

Concluding remarks

We apologise that we could not cite all the important publications of this field, due to length limitations. We thank Libby Guethlein and Oliver Ro¨ ssler for critical reading of the manuscript. Work on transcriptional repression has been supported by the Deutsche Forschungsgemeinschaft (TH 377/6–3).

References

1. Ptashne, M. & Gann, A.A.F. (1990) Activators and targets. Nature 346, 329–331.

2. Workman, J.L. & Kingston, R.E. (1998) Alterations of nucleo- some structure as a mechanism of transcriptional regulation. Annu. Rev. Biochem. 67, 545–579. 3. Kornberg, R.D. (1999) Eukaryotic transcriptional control. Trends Cell Biol. 9, M46–M49.

4. Johnson, A.D. (1995) The price of repression. Cell 81, 655–658. 5. Cowell, I.G. (1994) Repression versus activation in the control of gene transcription. Trends Biochem. Sci. 19, 38–42.

6. Molina, C.A., Foulkes, N.S., Lalli, E. & Sassone-Corsi, P. (1993) Inducibility and negative autoregulation of CREM: An alternative promoter directs the expression of ICER, an early response repressor. Cell 75, 875–886.

7. Servillo, G., Della Fazia, M.A. & Sassone-Corsi, P. (2002) Cou- pling cAMP signaling to transcription in the liver: pivotal role of CREB and CREM. Exp. Cell Res. 275, 143–164.

8. Jaworski, J., Mioduszewska, B., Sa´ nchez-Capelo, A., Figiel, I., Habas, A., Gozdz, A., Prosynski, T., Hetman, M., Mallet, J. & Kaczmarek, L. (2003) Inducible cAMP early repressor, an endogenous antagonist of cAMP responsive element-binding protein, evokes neuronal apoptosis in vitro. J. Neurosci. 23, 4519– 4526.

9. Brown, P.H., Alani, R., Preis, L.H., Szabo, E. & Birrer, M.J. (1993) Suppression of oncogene-induced transformation by a deletion mutant of c-jun. Oncogene 8, 877–886.

Mammalian transcriptional repressors use different strat- egies to impair transcriptional activation. The most straight- forward way is to block the DNA-binding site for activators or, to immobilize the activator outside of the nucleus. These activities are accomplished by passive repressors that are not actively involved in the biochemical reactions required for transcriptional activation. Active mammalian transcrip- tional repressors, in contrast, target the chromatin structure. The importance of active transcriptional repression is highlighted by the fact that aberrant gene silencing may be the cause of a range of diseases. In a wide variety of human cancers, for instance, a genome-wide DNA hypomethylation is observed, suggesting that this modification is the molecular basis for chromosomal instability and tumor formation. Accordingly, reduction of Dnmt1 DNA methyltransferase expression in transgenic mice to 10% of wild-type levels led to the development of aggressive T cell lymphomas [44]. Mutations in the gene DNMT3B encoding a DNA methyl- transferase are found in patients suffering from human ICF syndrome (immunodeficiency, centromeric heterochromatin instability, facial anomalies) and cells derived from patients with this disease show an extensive loss of methylation from the pericentromeric heterochromatin. Mutations in the X-linked MeCP2 gene are the major cause of Rett syndrome [45], a neurological disorder characterized by the loss of speech and hand skills, microcephaly and seizures. Accord- ingly, a Rett-like phenotype is induced in transgenic mice lacking MeCP2 in neurons of the central nervous system [46,47]. Deregulation of gene transcription, i.e. the gene encoding brain-derived neurotrophic factor, that occurs when the biological function of MeCP2 is impaired, may affect synaptic development and maturation and cause the pathology of Rett syndrome. Future research will certainly identify further neuronal genes whose misregulation may contribute to the pathology of Rett syndrome.

10. Steinmu¨ ller, L. & Thiel, G. (2003) Regulation of gene transcription by a constitutively active mutant of activating transcription factor 2 (ATF2). Biol. Chem. 384, 667–672.

11. Moitra, J., Mason, M.M., Olive, M., Krylov, D., Gavrilova, O., Marcus-Samuels, B., Feigenbaum, L., Lee, E., Aoyama, T., Eckhaus, M., Reitman, M.L. & Vinson, C. (1998) Life without white fat: a transgenic mouse. Genes Dev. 12, 3168–3181.

New developments in pharmacological research have successfully targeted mammalian transcriptional repres- sors. Hypermethylation of tumor suppressor genes can be reversed by powerful DNA methyltransferase inhibitors such as 5-aza-2¢-deoxycytidine used in the treatment of cancer [48]. However, while genomic demethylation may protect against some forms of cancer, it may increase the risk of other forms of cancer due to the promotion of genomic instability [44]. Additionally, histone deacetylase inhibitors have been tested as a new class of anticancer drugs [49,50], causing a hyperacetylation of histones and a

12. Kaczynski, J., Zhang, J.-S., Ellenrieder, V., Conley, A., Duenes, T., Kester, H., van der Burg, B. & Urrutia, R. (2001) The Sp1-like protein BTEB3 inhibits transcription via the basic transcription element box by interacting with mSin3A and HDAC-1 co-repressors and competing with Sp1. J. Biol. Chem. 276, 36749– 36756. 13. Karin, M. (1999) How NF-jB is activated: the role of the IjB kinase (IKK) complex. Oncogene 18, 6867–6874. 14. Wade, P.A., Pruss, D. & Wolffe, A.P. (1997) Histone acetylation: chromatin in action. Trends Biochem. Sci. 22, 128–132.

2862 G. Thiel et al. (Eur. J. Biochem. 271)

(cid:1) FEBS 2004

15. Ng, H.H. & Bird, A. (2000) Histone deacetylases: silencers to hire. Trends Biochem. Sci. 25, 121–126.

35. Bannister, A.J., Zegerman, P., Partridge, J.F., Miska, E.A., Thomas, J.O., Allshire, R.C. & Kouzarides, T. (2001) Selective recognition of methylated lysine 9 on histone H3 by the HP1 chromo domain. Nature 410, 120–124.

16. Free, A., Grunstein, M., Bird, A. & Vogelauer, M. (2001) Histone deacetylation: mechanisms of repression. In Chromatin Structure and Gene Expression (Elgin S. C. R. & Workman, J. L. eds), pp. 156–181, Oxford University Press, New York. 36. Lachner, M., O’Carroll, D., Rea, S., Mechtler, K. & Jenuwein, T. (2001) Methylation of histone H3 lysine 9 creates a binding site for HP1 proteins. Nature 410, 116–120.

17. Thiel, G., Lietz, M. & Cramer, M. (1998) Biological activity and modular structure of RE-1 silencing transcription factor (REST), a repressor of neuronal genes. J. Biol. Chem. 273, 26891–26899.

37. Pal-Bhadra, M., Leibovitch, B.A., Gandhi, S.G., Rao, M., Bhadra, U., Birchler, J.A. & Elgin, S.C.R. (2004) Heterochromatic silencing and HP1 localization in Drosphila are dependent on the RNAi machinery. Science 303, 669–672. 18. Huang, Y., Myers, S.J. & Dingledine, R. (1999) Transcriptional repression by REST: recruitment of Sin3A and histone deacetylase to neuronal genes. Nat. Neurosci. 2, 867–872.

38. Verdel, A., Jia, S., Gerber, S., Sugiyama, T., Gygi, S., Grewal, S.I.S. & Moazed, D. (2004) RNAi-mediated targeting of hetero- chromatin by the RITS complex. Science 303, 672–676. 19. Schoch, S., Cibelli, G. & Thiel, G. (1996) Neuron-specific gene expression of synapsin I: Major role of a negative regulatory mechanism. J. Biol. Chem. 271, 3317–3323.

20. Brehm, A., Miska, E.A., McCance, D.J., Reid, J.L., Bannister, A.J. & Kouzarides, T. (1998) Retinoblastoma protein recruits histone deacetylase to repress transcription. Nature 391, 597–601.

39. Fuks, F., Hurd, P.J., Wolf, D., Nan, X., Bird, A.P. & Kouzarides, T. (2003) The methyl-CpG-binding protein MeCP2 links DNA methylation to histone methylation. J. Biol. Chem. 278, 4035–4040. 40. Ferreira, R., Naguibneva, I., Pritchard, L.L., Ait-Si-Ali, S. & Harel-Bellan, A. (2001) The Rb/chromatin connection and epi- genetic control: opinion. Oncogene 20, 3128–3133.

41. Nielsen, S.J., Schneider, R., Bauer, U.-M., Bannister, A.J., Morrison, A., O’Carroll, D., Firestein, R., Cleary, M., Jenuwein, T., Herrera, R.E. & Kouzarides, T. (2001) Rb targets histon H3 methylation and HP1 to promoters. Nature 412, 561–565. 21. Luo, R.X., Postigo, A.A. & Dean, D.C. (1998) Rb interacts with histone deacetylase to repress transcription. Cell 92, 463–473. 22. Magnaghi-Jaulin, L., Groisman, R., Naguibneva, I., Robin, P., Lorain, S., Le Villain, J.P., Troalen, F., Trouche, D. & Harel- Bellan, A. (1998) Retinoblastoma protein represses transcription by recruiting a histone deacetylase. Nature 391, 601–605.

42. Vaute, O., Nicolas, E., Vandel, L. & Trouche, D. (2002) Func- tional and physical interaction between the histone methyl trans- ferase Suv39H1 and histone deacetylases. Nucleic Acids Res. 30, 475–481. 23. Allan, L.A., Duhig, T., Read, M. & Fried, M. (2000) The p21WAF1/CIP1 promoter is methylated in rat-1 cells: stable restor- ation of p53-dependent p21WAF1/CIP1 expression after transfection of a genomic clone containing the p21WAF1/CIP1gene. Mol. Cell. Biol. 20, 1291–1298.

24. Nan, X., Meehan, R.R. & Bird, A. (1993) Dissection of the methyl-CpG binding domain from the chromosomal protein MeCP2. Nucleic Acids Res. 21, 4886–4892. 43. Lunyak, V.V., Burgess, R., Prefontaine, G.G., Nelson, C., Sze, S.-H., Chenoweth, J., Schwartz, P., Pevzner, P.A., Glass, C., Mandel, G. & Rosenfeld, M.G. (2002) Corepressor-dependent silencing of chromosomal regions encoding neuronal genes. Science 298, 1747–1752.

25. Nan, X., Campoy, F.J. & Bird, A. (1997) MeCP2 is a transcrip- tional repressor with abundant binding sites in genomic chroma- tin. Cell 88, 471–481.

44. Gaudet, F., Hodgson, J.G., Eden, A., Jackson-Grusby, L., Dausman, J., Gray, J.W., Leonhardt, H. & Jaenisch, R. (2003) Induction of tumors in mice by genomic hypomethylation. Science 300, 489–492.

26. Jones, P.L., Veenstra, G.J.C., Wade, P.A., Vermaak, D., Kass, S.U., Landsberger, N., Strouboulis, J. & Wolffe, A.P. (1998) Methylated DNA and MeCP2 recruit histone deacetylase to repress tran- scription. Nat. Genet. 19, 187–191.

45. Amir, R.E., Van den Veyver, I.B., Wan, M., Tran, C.Q., Francke, U. & Zoghbi, H.Y. (1999) Rett syndrome is caused by mutations in X-linked MECP2, encoding methyl-CpG-binding protein. Nat. Genet. 23, 185–188.

27. Nan, X., Ng, H.-H., Johnson, C.A., Laherty, C.D., Turner, B.M., Eisenman, R.N. & Bird, A. (1998) Transcriptional repression by methyl-CpG-binding protein MeCP2 involves a histone deacety- lase complex. Nature 393, 386–389. 28. Bestor, T.H. (1998) Methylation meets acetylation. Nature 393, 311–312.

46. Chen, R.Z., Akbarian, S., Tudor, M. & Jaenisch, R. (2001) Deficiency of methyl-CpG binding protein-2 in CNS neurons results in a Rett-like phenotype in mice. Nat. Genet. 27, 327–331. 47. Guy, J., Hendrich, B., Holmes, M., Martin, J.E. & Bird, A. (2001) A mouse Mecp2-null mutation causes neurological symptoms that mimic Rett syndrome. Nat. Genet. 27, 322–326.

29. Chen, W.G., Chang, Q., Lin, Y., Meissner, A., West, A.E., Griffith, E.C., Jaenisch, R. & Greenberg, M.E. (2003) Derepres- sion of BDNF transcription involves calcium-dependent phos- phorylation of MeCP2. Science 302, 885–889. 48. Karpf, A.R. & Jones, D.A. (2002) Reactivating the expression of methylation silenced genes in human cancer. Oncogene 21, 5496– 5503.

49. Marks, P.A., Rifkind, R.A., Richon, V.M., Breslow, R., Miller, T. & Kelly, W.K. (2001) Histone deacetylases and cancer: causes and therapies. Nat. Rev. Cancer 1, 194–202. 30. Martinowich, K., Hattori, D., Wu, H., Fouse, S., He, F., Hu, Y., Fan, G. & Sun, Y.E. (2003) DNA-methylation-related chromatin remodeling in activity-dependent Bdnf gene regulation. Science 302, 890–893.

50. Melnick, A. & Licht, J.D. (2002) Histone deacetylases as thera- peutic targets in hematologic malignancies. Curr. Opin. Hematol. 9, 322–332. 31. Nielsen, A.L., Oulad-Abdelghani, M., Ortiz, J.A., Remboutsika, E., Chambon, P. & Losson, R. (2001) Heterochromatin formation im mammalian cells: interaction between histones and HP1 pro- teins. Mol. Cell 7, 729–739.

32. Grewal, S.I. & Elgin, S.C. (2002) Heterochromatin: new possibi- lities for the inheritance of structure. Curr. Opin. Genet. Dev. 12, 178–187. 33. Jenuwein, T. (2001) Re-SET-ting heterochromatin by histone 51. Go¨ ttlicher, M., Minucci, S., Zhu, P., Kra¨ mer, O.H., Schimpf, A., Giavara, S., Sleeman, J.P., Lo Coco, F., Nervi, C., Pelicci, P.G. & Heinzel, T. (2001) Valproic acid defines a novel class of HDAC inhibitors inducing differentiation of transformed cells. EMBO J. 20, 6969–6978. methyltransferases. Trends Cell. Biol. 11, 266–273.

52. Brasher, S.V., Smith, B.O., Fogh, R.H., Nietlispach, D., Thiru, A., Nielsen, P.R., Broadhurst, R.W., Ball, L.J., Murzina, N.V. & Laue, E.D. (2000) The structure of mouse HP1 suggests a unique mode of single peptide recognition by the shadow chromo domain dimer. EMBO J. 19, 1587–1597. 34. Rea, S., Eisenhaber, F., O’Carroll, D., Strahl, B.D., Sun, Z.-W., Schmid, M., Opravil, S., Mechtler, K., Ponting, C.P., Allis, C.D. & Jenuwein, T. (2000) Regulation of chromatin structure by site- specific histone H3 methyltransferases. Nature 406, 593–599.