Characterization of histone modifying enzymes and their substrates in maize

Table of contents

Introduction
I- The structure of the chromatin
The structure of the nucleosome
The histones
II- The chromatin modifications

A- The chromatin remodeling factors

B- Covalent modification of histones
1- Histone phosphorylation
2- Histone ubiquitination
3- Glycosylation of histones
4- Histone acetylation
A- Histone acetyltransferases (HATs)
B- Histone deacetylases (HDACs)
I- The RPD3-like histone deacetylases- Class I
II- The HDA1-like histone deacetylases- Class II
III- The Sir2-like histone deacetylases- Class III
IV- The plant specific HD2-histone deacetylases
C- HDAC inhibitors (HDACi)
5- Histone methylation
A- The PRMT family of HMTs
Type I PRMTs
PRMT4/CARM1
PRMT2
PRMT3
PRMT6
PRMT8
Type II PRMTs
PRMT5/JBP1
PRMT7
B- SET domain HMTs
I- SUV39 family
II- SET1 family
III- SET2 family
IV- RIZ family
C- Non-SET domain HMT

The histone code hypothesis

The 14-3-3 proteins

Aim of the work

Materials and Methods

A- Materials
A. 1 Germination of Zea mays grains
A. 2 E. coli strains
A. 3 E. coli cultivation and storage of cultures
A. 4 Plasmids
A. 5 Primers
a. ZmHdal primers
b. GF14 primers

B- Molecular Biology Techniques
B. l RNA extraction from maize tissues
B. 2 cDNA synthesis from total RNA
B. 3 Estimating the concentration of DNA and RNA
B. 4 Polymerase Chain Reaction (PCR)
B. 5 Gel electrophoresis of DNA and RNA
a. Agarose gel electrophoresis of DNA
b. Agarose gel electrophoresis of RNA
B. 6 Purification and isolation of DNA
B. 7 Cloning of DNA fragments into plasmid vectors
B. 8 Transformation of bacterial cells
a. Bacterial transformation by electroporation
b. Transformation of chemically competent E. coli cells
B. 9 Isolation of plasmids from bacterial cells
a. Plasmid isolation- small scale preparation (mini-preparation)
b. Large scale plasmid preparation
B. 10 Restriction enzyme digestion

C- Biochemical Techniques
C. 1 Cellular fractionation of maize seedlings
C. 2 Enzymatic activity assays
a. Histone deacetylase assay (HDAC assay)
b. Histone methyltransferase assay (HMT assay)
c. ATPase assay
C. 3 Estimation of protein concentration by the Bradford assay
C. 4 Immunoprecipitation (IP) experiments
C. 5 Pull-down experiments
a. GST pull-down
b. 6xHis pull-down
C. 6 Overexpression of proteins in E. coli
C. 7 Concentrating protein samples
a. TCA precipitation
b. TCA-DOC precipitation
c. Concentration by centrifugation
C. 8 Electrophoretic separation of proteins
a. Polyacrylamide gel electrophoresis under denaturing conditions (SDS-PAGE)
b. Native polyacrylamide gel electrophoresis
C. 9 Staining of gels
a. Coomassie Brilliant Blue staining
b. Silver staining
C.10 Western blotting and immunodetection of proteins
Immuno-detection of blotted proteins with alkaline phosphatase
Immuno-detection of blotted proteins with ECL
C.11 Antibodies
a. Primary antibodies
b. Secondary antibodies
C.12 Extraction of histones
a. Extraction of histones from maize tissues
b. Extraction of histones from A20 cells
C.13 Assembly of nucleosomes
C.14 Dialysis
C.15 Chromatographic techniques
a. Ion exchange chromatography (IEC)
a.1. Q-Sepharose Big Beads (Amersham, Biosciences)
a.2. Q-Sepharose Fast Flow (Amersham, Biosciences)
a.3. Source 15Q (Amersham, Biosciences)
a.4. Mono Q HR 5/5 (Amersham, Biosciences)
b. Hydrophobic interaction chromatography (HIC)
c. Affinity chromatography
c. 1. Heparin Sepharose
c. 2. Histone Agarose
c. 3. Poly-L-lysine Agarose
d. Hydroxyapatite chromatography
e. Size exclusion chromatography
e.1. HiLoad 16/60 Superdex™ 200 column
e.2. TSK G4000 PWXL
C.16 In-gel digestion and MALDI-TOF/TOF protein identification
C.17 Protein digestion and nanospray-MS/MS protein identification

Results
I- Histone deacetylase ZmHdal of Zea mays
A- Cloning of ZmHda1
B- Expression and purification of ZmHda1
C- Characterization of pep29 and pep90 anti-ZmHda1 antibodies
D- Pull-down experiments
E- Immunoprecipitation experiments
E. 1. Partial purification of ZmHda1 protein forms
E. 2. IP experiments using recombinant ZmHda1 as input
E. 3. IP experiments using 72h maize soluble fraction as input
E. 4. IP experiments using partially purified ZmHda1-p48 form
E.5. IP experiments using partially purified ZmHda1-p84 form complex
F- Experiments in progress
II- Histone methyltransferases in germinating maize embryos
A- Changes in methylation profile of histones H3 and H4 during germination of maize embryos
B- Total methyltransferase activity in different extracts of maize embryos before
and after germination
C- Chromatographic fractionation of HMT activities of maize at 0h and 72h of germination
C.1. Cytoplasmic extract
C. 2. Chromatin extract
D- Purification and characterization of HMT activities from chromatin extract of
72h maize embryos
D. 1. Q-Sepharose Fast Flow chromatography
D. 1.1. Heparin-Sepharose chromatography of AI
D. 1.2.a.1. Heparin-Sepharose chromatography of AII
D. 1.2.a.2. Source 15Q chromatography
D. 1.2.a.3. Histone-Agarose chromatography
D. 1.2.a.4. MonoQ HR 5/5 chromatography
D. 1.2.a.5. Gel filtration chromatography
D. 1.2.b.1. Estimation of the molecular size and site specificity of AII2 HMT
activity
D. 1.2.b.2. Poly-L-Lysine Agarose chromatography
E- Purification of AII2 HMT activity from 72h maize embryos chromatin extract
E. 1. Source 15Q chromatography
E. 2. Histone-Agarose chromatography
E. 3. MonoQ chromatography
F- Purification and characterization of HMT activities from cytoplasmic extract of
72h maize embryos
F. 1. Q-Sepharose Big Beads chromatography
F. 2. Phenyl Sepharose chromatography
F. 3. Hydroxyapatite chromatography
F. 4. Source 15Q chromatography
F. 4.a.1. TSK G-4000 chromatography of BI HMT activity
F. 4.a.2. MonoQ chromatography of BI HMT activity
F. 4.b.1. TSK G-4000 chromatography of BII HMT activity
F.4.b.2. MonoQ chromatography of BII HMT activity
F.4.c. TSK G-4000 chromatography of BIII HMT activity
F.4.c.1. MonoQ chromatography of BIII1 HMT activity
F.4.c.2. MonoQ chromatography of BIII2 HMT activity
G- Identification of the BIII2 activity
G. 1. The substrate specificity of the BIII2 HMT activity
G. 2. Methylation of the 30 kDa protein requires posttranslational modifications
of H3/H4
G. 3. Identification of the 30 kDa protein
G. 4. Cloning of the GF14 proteins
G. 5. Expression and purification of GF14 constructs
G. 6. HMT assays using recombinant GF14 proteins
G. 7. Pull-down experiment
H- Purification of the BII HMT activity
H. 1. Q-Sepharose XL chromatography
H. 2. Phenyl Sepharose chromatography
H. 3. Q-Sepharose FF chromatography
H. 4. Poly-L-Lysine Agarose chromatography
H. 5. Hydroxyapatite chromatography
H. 6. Source 15Q chromatography
H. 7. Gel filtration chromatography (Superdex 200)
H.8. MonoQ chromatography
H.9. Native PAGE analysis of the BII activity
H. 10. Nano-LC/MS-MS Identification of the BII activity
I- Biochemical characterization of ZmPRMT5
I. 1. Substrate specificity
I.2. Site specificity
I.3. The relation between ZmPRMT5 HMT activity and ATP

Schematic representation of the purification steps of the 72h maize embryos HMTs
I- The chromatin bound HMTs
II- Preliminary purification of the HMTs soluble in the cytoplasmic extract
III- Purification of the BII HMT activity

Discussion

Summary and conclusion

References

Appendix

Introduction

I- The structure of chromatin:

The huge size and complexity of eukaryotic DNA has demanded stringent structural organization of the genetic material which has to match two basic requirements. First, a huge amount of DNA has to be packed into a cell nucleus of approximately 5 pm in diameter. Second, the DNA has to be accessible to various enzymatic machineries during interphase, since the whole DNA has to be replicated and specific parts of the DNA have to transcribe in each cell cycle. Moreover, parts of the DNA have to be repaired (Loidl, 1994). The first requirement is matched by winding each 146 bp of DNA around a histone octamer, each composed of two copies of each of the four core histone H2A, H2B, H3 and H4, forming the nucleosome, which is considered the basic structure of the chromatin.

illustration not visible in this excerpt

Figurel: The different hierarchal levels of DNA packaging into chromatin

(Felsenfeld and Groudine, 2003).

The nucleosomes array, which is 11 nm in diameter and also called “beads on string”, is further packaged by the aid of linker histone H1 into 30 nm fibers with six nucleosomes per turn in spiral or solenoid arrangement (Figure 1). These 30 nm fibers appear to correspond to the euchromatin, the component of the eukaryotic genome that is actively transcribed. Heterochromatin, on the contrary, is typically transcriptionally silent and is characterized by higher order packaging of the 30 nm chromatin up to several thousand-fold (Zhang and Reinberg, 2001; Bowler et al., 2004). Two types of heterochromatin can be recognized: a) the constitutive heterochromatin which represents DNA that contains no genes and so can always be retained in a compact organization. This fraction includes centromeric and telomeric DNA as well as certain regions of some chromosomes. b) the facultative heterochromatin which is not a permanent structure. It is thought to contain genes that are inactive in some cells or at some periods of the cell cycle (Brown, 2002). The second requirement if fulfilled by changing the chromatin structure to make it more accessible to proteins and enzymes required for processes like replication, transcription and repair. There are three general ways in which chromatin structure can be altered. First, nucleosome remodeling that can be induced by protein complexes utilizing the energy released by ATP hydrolysis. Second, covalent modification of histones forming the nucleosomes. Third, histone variants may replace one or more of the core histones (Felsenfeld and Groudine, 2003; Brzeski and Jerzmanowski, 2004; Chakravarthy et al., 2005).

The structure of the nucleosome:

A nucleosome, the fundamental repeating structural unit of chromatin, consists of an octamer of core histones and 147 bp of DNA wrapped around the octamer in a left-handed superhelix. The histone octamer has a tripartite structure, which is organized as a histone (H3-H4)2 tetramer flanked by two H2A-H2B dimers (Morales and Richard-Foy, 2000). Each nucleosome is connected to its neighbors by a short segment of linker DNA (10-80 bp in length) and this polynucleosome string is folded into a compact fiber with a diameter of 30 nm. The 30 nm fiber is stabilized by the binding of a fifth histone, H1, to each nucleosome and to its adjacent linker DNA (Figure 2) (Felsenfeld and Groudine, 2003).

In vitro experiments have demonstrated that the nucleosome is a very stable structure. However, in vivo experiments using green fluorescent protein (GFP)- histone fusion proteins showed that about half of the total histone H2B pool and 20% of the histone H3 and histone H4 are turned over by passive and transient mechanisms involving incoming histones. Consequently, this exchange could provide a successful way to maintain the nucleosome in a dynamic state. Nonetheless, the nucleosomes are maintained on the chromatin fiber during processes that involve DNA metabolism, including replication, transcription, etc. Therefore, this dynamic state of nucleosome might still allow important information to be passed on from one cell generation to the next. This form of
inheritance, which does not involve the DNA sequence per se, is referred to as the epigenetic inheritance (Vaquero et al., 2003).

illustration not visible in this excerpt

Figure 2: Stabilization of the 30 nm chromatin by binding of the histone H1 The histones:

The histones are abundant small basic proteins found in eukaryotes. The histones are among the most highly conserved eukaryotic proteins. Moreover, histone equivalents and simplified chromatin structures have also been found in single-cell organisms from the kingdom Archaeabacteria (Felsenfeld and Groudine, 2003). Histones are the primary proteins whose properties mediate the folding of DNA into chromatin. Aside from the compaction of DNA, histones undertake protein­protein interactions between themselves and other distinct chromosomal proteins. These interactions lead to several constraints on the properties of histones, contributing to maintaining their high degree of evolutionary conservation. Since histones can be removed from DNA by high salt extraction, the major interaction between DNA and core histones appear to be electrostatic in nature. All core histones are remarkably conserved in length and amino acid sequence through evolution. Histones H3 and H4 are the most conserved, for example calf and pea histone H4 differ at only two sites in 102 residues. Histones H2A and H2B are slightly less conserved. All histones are small basic proteins (11- 16 kDa molecular weight) containing relatively large amounts of lysine and arginine residues (more than 20% of the amino acids). Histones H2A and H2B contain more lysine, and histones H3 and H4 contain more arginine. All the four core histones contain an extended histone-fold domain at the carboxyl (C-) terminal end of the protein through which histone-histone and histone-DNA interaction occur, and charged tails at the amino (N-) terminal end which contain the bulk of the lysine residues. The C-terminal histone fold domain contains three α-helices. The histone fold domain is expected to be conserved due to its central structural role in the nucleosome (Wolffe, 1998). The N-terminal tails are flexible, protruding out of the nucleosome surface and they are the sites of many post transitional modifications, thus they are responsible for the molecular communication with the nuclear environment (Loidl, 1994). The N-terminal tails also provide contact surfaces for the interaction with other proteins that organize higher-order chromatin structures.

The eukaryotic cells contain a fifth histone called the linker histone, of which the most common is called histone H1. Another linker histone from chicken erythrocytes known as histone H5 has been extensively studied. Both histones H1 and H5 are highly basic, being particularly rich in lysine and are slightly larger than core histones. Linker histones are the least tightly bound to DNA of all the histones. The linker histones have a central structured winged-helix domain and highly charged tails at both the N- and C- termini. The central domain associates with the nucleosome and stabilizes the histone-DNA while the tails interact with the DNA between the nucleosomes (Wolffe, 1998).

The bulk of nucleosome assembly occurs during DNA replication; therefore, a large amount of histones is needed at this time of the cell cycle. To fulfill this requirement, the bulk of histone gene expression also occurs during the S-phase. However, outside the S-phase there is a low, but constant expression of histone proteins. These histones, called “replacement histones” or “histone variants” are used for nucleosome assembly independent of DNA replication. All the histones except H4 are encoded by several different genes. In some cases, the various genes encode histones that differ in only a few amino acids. For example, the histone H3.3 variant differs from the major H3 at only four amino acid positions. In other cases, the differences are more extensive, such as in the case of macro H2A, where about 60% of its sequence is unrelated to the sequences of any other histones (Vaquero et al., 2003; Takahashi et al., 2006). Although the functions of the histone variants are still largely uncharacterized, more and more evidence indicates that at least some of the variants play specific physiological roles (Vaquero et al., 2003).

II- The chromatin modifications:

The fact that the DNA is highly compacted into chromatin has profound effects on DNA metabolism. Many of the processes in which DNA acts as the substrate are repressed by chromatin. It was initially thought that this repression was simply due to the blocking of DNA by core histones so that other proteins, such as the transcription machinery, had restricted access to the DNA. However, it was shown recently that chromatin has a more participatory function in regulation rather than simply occluding the binding of other proteins (Orphanides and Reinberg, 2002). To circumvent the inhibitory effects of chromatin, cells use a variety of factors to make the structure of chromatin more dynamic. Two types of activities have been described. One is the ATP-dependent chromatin remodeling. The other activity is performed by enzymes that covalently modify the N-terminal tail residues of histones (Fransz and de Jong, 2002; Reyes et al., 2002; Vaquero et al., 2003).

A- The chromatin remodeling factors:

Chromatin remodeling is usually associated structurally with chromatin opening and functionally with transcriptional activation. But not all remodeling factors have this effect. Some alter chromatin by forming transcriptionally repressed structures, for example, NuRD, a chromatin remodeling complex that is associated with histone deacetylases (Xue et al., 1998). The ATP-dependent chromatin remodeling factors are divided into four families according to similarities in their ATPase subunit: the ISWI, SWI/SNF, INO and CHD families. These ATPases exhibit mechanistic differences; for example, the ATPase activity of the SWI/SNF complex is stimulated by naked DNA and does not require histone tails. The ATPase of Mi2 (a member of CHD family) is stimulated by oligonucleosomes, but not by naked DNA, and does not require the presence of histone tails. The ISWI ATPase activity is slightly stimulated by naked DNA and strongly stimulated by oligonucleosomes, and the histone H4 tail is essential for this activity (Vaquero et al., 2003; Varga-Weisz and Becker, 2006). These chromatin remodeling factors are playing very crucial roles in plant development. Arabidopsis has four members of the SWI/SNF protein family, mutations in one of these, SPLAYED (SYD), causes severe developmental defects. SYD is required for maintenance of the shoot apical meristem, for repression of the floral transitions (especially under non-inductive short-day conditions), for proper floral homeotic gene expression and for ovule development (Fransz and de Jong, 2002; Wagner and Meyerowitz, 2002; Sung et al., 2003). The chromatin remodeling complexes modify the chromatin structure through different mechanisms. They can disrupt the histone-DNA contacts in mononucleosomes, catalyze the mobilization of nucleosomes relative to the DNA template, mediate the transfer of histones from one DNA template to a separate one and/or generating superhelical torsion in DNA or chromatin (reviewed by Lusser and Kadonaga, 2003).

B- Covalent modification of histones:

The N-terminal tails of the nucleosome core histones are exposed on the nucleosome surface and preferentially subjected to enzyme-catalyzed posttranslational modifications. These modifications include acetylation, methylation, phosphorylation, ubiquitination, glycosylation, ADP-ribosylation, carbonylation, sumoylation, and biotinylation. These histone modifications can alter chromatin structure and serve as binding platform for proteins that influence transcriptional activity and they also establish epigenetic information that can be propagated from one generation to the next (Loidl, 2004; Janzen et al., 2006).

1- Histone phosphorylation:

All histones, including H1, have been found to be substrates for phosphorylation in vivo. Phosphorylation of H1 and H3 has been associated with chromosome condensation and segregation. This function is highly conserved among species from Tetrahymena to mammals (Vaquero et al., 2003). Phosphorylation takes place at serine and threonine residues. In eukaryotes, H3 phosphorylation occurs in two serine residues, S10 and S28. H3 phosphorylation appears early in the G2 phase of the cell cycle, first in the pericentromeric heterochromatin of all chromosomes and then spreading, by the time of metaphase, to the rest of the chromosome. Several kinases are involved in H3 phosphorylation. Among them is Aurora-B, a member of the Aurora/AIK kinase family, which participates in mitotic regulation (Hsu et al., 2000). Genetic studies have pointed out that protein phosphatase 1 is involved in the reverse reaction. Phosphorylation of S10 affects K14 acetylation. In yeast, K14 acetylation is promoted by phosphorylation of S10. In accordance with this observation, a subset of genes activated by the transcription factor Gcn5p (but not all Gcn5p-dependent genes) requires the phosphorylation of S10 for normal levels of transcription in vivo (Manzanero et al., 2000; Vaquero et al., 2003; Bottomley, 2004). In mammalian cells, up to six serines and threonines in histone H1 are phosphorylated in vivo in a cell cycle dependent manner that has long been linked with chromatin condensation. This phosphorylation is carried out by cyclin-dependent kinases (CDKs) (Swank et al., 1997). H1 phosphorylation takes place at both the C- and N-terminal tails (Contreras et al., 2003). During the cell cycle of mouse fibroblasts, phosphorylation of H1 begins in the late G1 phase and increases throughout the S and G2 phases. During late G2 and mitosis, phosphorylation of H1 subtypes becomes maximal and decreases sharply thereafter. These cell cycle-related H1 histone modifications imply that H1 modulates chromatin condensation and decondensation (Talasz et al., 1996). In plant systems, like in animals, histone H1 phosphorylation level increases during the G2/M phase (Stratton and Trewavas, 1981). For example, the induction of the mitosis by gibberellin in the intercalary meristem of rice internodes, as indicated by the decline in the number of cells in the G2 phase, is brought about by increase in the p34cdc2/CDC28-like histone H1 kinase activity (Sauter et al., 1995).

2- Histone ubiquitination:

Ubiquitin (Ub) is a 76 amino acid protein that is ubiquitously distributed and highly conserved throughout eukaryotic organisms. Whereas the four extreme C- terminal amino acids are in a random coil, its 72 N-terminal amino acids have a tightly folded globular structure (Zhang, 2003). The C-terminal tails of histones H2A, H2B and H3 are subject to ubiquitination at lysine residues (Thorn et al., 1987; Nickel and Davie, 1989). These ubiquitinations are stable in vivo, and their incorporation into nucleosomes has been proposed to alter chromatin structure locally (Robzyk et al., 2000). However, ubiquitinated H2A (uH2A) is incorporated into nucleosomes without major changes in the organization of nucleosome cores (Kleinschmidt, 1981). Histone H2A was the first protein identified to be ubiquitinated (Goldknopf et al., 1975). The ubiquitination site has been mapped to the highly conserved residue K119 (Nickel and Davie, 1989). H2A ubiquitination has been reported in a variety of higher eukaryotic organisms, but not in the budding yeast Saccharomyces cerevisiae (Robzyk et al., 2000). Ubiquitinated H2A is more abundant (~ 5-15% of the total H2A) compared to ubiquitinated H2B (~ 1­2% of the total H2B) (Zhang, 2003). The majority of H2A is in mono- ubiquitinated form; however, poly-ubiquitinated H2A has also been reported in many tissues and cell types (Nickel et al., 1989). Ubiquitinated H2B appears to be widely distributed throughout eukaryotic organisms from budding yeast to humans. It is being ubiquitinated at lysine residues located at the C-terminus, namely, K120 in humans and K123 in yeast. Thus far, only mono-ubiquitinated H2B has been reported (Thorn et al., 1987; Zhang, 2003). Ubiquitination on H3 and H1 has also been reported (Chen et al., 1998; Pham and Sauer, 2000). However, ubiquitinated H3 and H1 are not as prevalent as H2A and H2B and the ubiquitination site on H3 or H1 has not been determined (Zhang, 2003).

Histone ubiquitination is a reversible modification. Therefore, the steady-state histone ubiquitination levels depend on the availability of free ubiquitin and enzymatic activities involved in adding or removing the ubiquitin moiety from histones. Addition of ubiquitin moiety to a protein involves the sequential action of E1, E2 and E3 enzymes. Removing the ubiquitin moiety, on the other hand, is achieved through the action of enzymes called isopeptidases (Wilkinson, 2000; Zhang, 2003).

Accumulating evidence indicates that histone ubiquitination participates in gene activation. For example, it has been reported that nucleosomes of transcriptionally active heat shock protein (hsp) 70 genes contain up to 50% uH2A, whereas nucleosomes of nontranscribed satellite DNA contain only one uH2A per 25 nucleosomes (Levinger and Varshavsky, 1982). On the other hand, fractionation of micrococcal nuclease-digested myotube nuclei revealed that uH2A was not enriched in transcriptionally active or inactive chromatin, but simply in nuclease- sensitive fractions (Parlow et al., 1990). In addition, ubiquitinated histones have also been found in transcriptionally inactive compartments, such as Tetrahymena micronuclei (Nickel et al., 1989). Thus, histone ubiquitination most likely regulates gene transcription in a positive and negative fashion, depending on its genomic and gene location (Zhang, 2003).

3- Glycosylation of histones:

Non-enzymatic glycosylation (glycation) of proteins is a multistage chemical process starting as a condensation reaction between reducing sugars and primary amino groups (mainly from the side chain of Lys and Arg) and ending up with the formation of complex heterocyclic compounds called advanced end products. Enzymatic glycosylation consists of systemic addition of one or more sugar moieties to either hydroxyl groups of Ser, Tyr or Thr (O-glycosylation) or the amide group of Asn (N-glycosylation) (Stoynev et al., 2004). Many lines of evidence indicate that the non-enzymatic modification of long lived biomolecules plays an important role in the aging process and the pathophysiology of diseases whose incidence increases as a function of age (Smith et al., 1994). Both, individual histones (Liebich et al., 1993), and histones in nucleosomes (Gugliucci, 1994) have been reported to be glycated in vitro.

4- Histone acetylation:

Histone acetylation has been the most exclusively studied covalent modification of core histones (Strahl and Allis, 2000). Acetylation of the four core histones occurs in all animal (Allfrey et al., 1964; Wade et al., 1997; Kornberg and Lorch, 1999), plant (Georgieva et al., 1991; Lopez-Rodas et al., 1991; Lusser et al., 1999; Graessle et al., 2001; Pandey et al., 2002; Chua et al., 2003; David Law and Suttle, 2004; Causevic et al., 2006) and fungal species (Rundlett et al., 1996; Guo et al., 2006) so far studied. The sites of acetylation are the ε-amino group of lysine residues of the positively charged amino terminal tails, where each acetate group added to a histone reduces its net positive charge by 1 (Wolffe, 1998). A nucleosome contains a total of 26 potentially acetylated lysine residues (Lusser et al., 2001). The acetylation state of histones is dynamic depending on the equilibrium between histone acetylation catalyzed by histone acetyltransferases (HATs) and deacetylation catalyzed by histone deacetylases (HDACs) (Figure 3).

illustration not visible in this excerpt

Figure 3: Histone acetylation state is regulated by HATs and HDACs.

Two populations of acetylated histones appear to exist in a particular cell nucleus. For example, in embryonic chicken erythrocytes, 30% of core histones are stably acetylated while the acetylation status of about 2% changes rapidly (Wolffe, 1998). The pioneering studies on the acetylation of core histones, done by Allfrey and his colleagues (Allfrey et al., 1964), showed that the acetylation of histone N-termini is strongly correlated with enhanced transcription. Histone acetylation has been linked to active transcription while histone deacetylation has been linked to a repressed transcription state. This effect has been ascribed to the effect of adding or removing acetyl group on the net charge of the histone N-termini, which in turn affects the interaction between the nucleosome and the negatively charged DNA and hence the state of the chromatin, whether accessible or less accessible to the transcription machinery (Chen et al., 2001; Graessle et al., 2001). The discovery of the first HATs (GCN5, CBP and P300), and the characterization of the well-known RPD3 as HD AC, led to the discovery of novel HATs and HDACs. Since then numerous transcriptional regulators and coregulators have been shown to act as HATs or HDACs (Chen et al., 2001; Graessle et al., 2001).

A- Histone acetyltransferases (HATs):

HATs catalyze the transfer of the acetyl moiety of acetyl coenzyme A (acetyl CoA) to the ε-amino group of lysine residues. In addition to the known residues being acetylated (K9, K14, K18 and K23 on H3; K5, K8, K12, K16 and K20 on H4; K5 and K9 on H2A; K5, K12, K15 and K20 on H2B), novel acetylation sites have been discovered recently by a peptide mass fingerprint approach (K31 on H4; K13 and K15 on H2A; (Zhang et al., 2003a).

In addition to the classical view that lysine acetylation weakens the interaction between histones and DNA, two other theories explain the effects of lysine acetylation on chromatin structure and function; however, the classical theory can not be excluded. These two theories are:

I- It has been supposed that acetylation may act as a specific signal that alters the protein-protein interactions and may recruit specific proteins into specific loci

within the chromatin (Loidl, 1988; Lopez-Rodas et al., 1993; Kouzarides, 2000). This hypothesis is supported by the finding that non-histone proteins can also be acetylated and deacetylated by the same enzymes involved in histone acetylation and deacetylation which means a general mechanism to control protein-protein interactions, (reviewed by Sterner and Berger, 2000). Moreover, the acetylated lysines are recognized and bind to specific protein domains (the bromodomain). This idea leads then to the concept of “The histone code”, as it will be discussed later.

II- Histone acetylation may also interfere with the higher order packing of chromatin and hence alter the accessibility of chromatin for regulatory proteins reviewed by (Horn and Peterson, 2002).

HATs have been found in all eukaryotes (Table 1) and they are implicated in many cellular and developmental processes such as transcriptional regulation, cell cycle progression (Jasencakova et al., 2000; Carrozza et al., 2003; Chua et al., 2003), apoptosis, DNA repair (Giordano and Avantaggiati, 1999; Ikura et al., 2000) and dormancy break in potato tubers (David Law and Suttle, 2004). Recently, histone acetylation levels have been used as a biomarker of morphogenesis in sugar beet (Causevic et al., 2006).

Historically, HATs have been divided into two categories: type A, located in the nucleus and type B located in the cytoplasm. Some HATs can act in multiple complexes or locations and thus do not precisely fit this classical classification. Type B HATs are believed to have somewhat of a housekeeping role in the cell, acetylating newly synthesized free histones in the cytoplasm for transport into the nucleus and incorporation into chromatin. On the other hand, type A HATs acetylate nucleosomal histones and these HATs are potentially linked to transcription (Sterner and Berger, 2000). HATs can also be divided into 5 families (reviewed by Sterner and Berger, 2000; Carrozza et al., 2003), based on their structure and function. These are GNATs, the MYST-related HATs, p300/CBP HATs, the general transcription factor HATs which include TFIID subunit TAF250, and the nuclear hormone-related HATs SRC1 and ACTR. Table 1 shows a summery of known and putative HATs.

Whereas histone H4 is being the predominant target of acetylation in animals and fungi, histone H3 was found to be the most extensively acetylated histone in plants (Waterborg et al., 1990; Bhat et al., 2003). H4 in Medicago, Arabidopsis, tobacco and carrot was detected in five acetylated isoforms (mono- to penta-acetylated). In animals and yeast, H4 K20 is not acetylated but methylated instead (Waterborg, 1992; Strahl et al., 1999).

Table 1: Summary of putative histone acetyltransferases of various organisms (Lusser et al., 2001).

illustration not visible in this excerpt

In germinating Zea seedlings, at least 3 HAT activities can be detected in chromatographic fractions of cellular extracts. One activity, HATB, is homologous to yeast Hat1 (Lopez-Rodas et al., 1991; Lusser et al., 1999). The two other enzymes, HATA1 and HATA2, are biochemically distinct nuclear enzymes. HATA1 acetylates histones H3 and H4 while HATA2 specifically acetylates H3 (Lopez-Rodas et al., 1991; Lusser et al., 1999). These two HAT activities have been identified as homologues of the yeast GCN5 and they exist in complexes reminiscent of the yeast ADA complex (Bhat et al., 2003). The Zea mays HATB is responsible for the acetylation of newly synthesized H4 at lysines K5 and K12 before nucleosome assembly (Kölle et al., 1998). H4 acetylation is suggested to be important for the transport of newly synthesized H4 into the nucleus and/or correct assembly into the nucleosome (Grunstein, 1997).

B- Histone deacetylases (HDACs):

In general, HDACs have been correlated with transcriptional repression and gene silencing (Loidl, 2004). However, recently it has been shown that HDAC1 serves as a transcriptional coactivator for ISGF3 in response to viral infection (Nusinzon and Horvath, 2005). HDACs are generally found as part of multiprotein complexes that contain transcriptional repressors, corepressors and a variety of other proteins. In some cases, HDAC complexes contain ATP-dependent chromatin-remodelling proteins and proteins that bind to methylated DNA (Murfett et al., 2001; Verdin et al., 2003).

Histone deacetylases have been classified into distinct families depending on the homology to their founding members and cofactor dependency. Class I HDACs are related to the first described histone deacetylase, the yeast protein, RPD3. They include HDAC1, HDAC2, HDAC3 and HDAC8. Class II HDACs are similar to yeast HDA1, and include HDAC4, HDAC5, HDAC6, HDAC7, HDAC9 and HD AC 10. Class III HDACs are related to the yeast SIR2, an NAD-dependent deacetylase. A plant specific HDAC family has been described after its first member, HD2, was discovered in germinating maize seedlings (Lusser et al., 1997). A summary of histone deacetylases from various organisms is shown in table 2.

I- The RPD3-like histone deacetylases- Class I:

These HDACs are found almost exclusively in the nucleus. The nuclear localization of HDACs occurs via a nuclear localization signal (NLS) or via co­localization together with other proteins (de Ruijter et al., 2003). The localization of the human HDAC1, HDAC2 and HDAC8 is exclusively nuclear, while HDAC3 has both nuclear import signal and nuclear export signal (NES), suggesting that it can also localize to the cytoplasm. However, HDAC3 is nearly always localized in the nucleus (de Ruijter et al., 2003). In rice, three HDACs belong to the class I HDACs, namely OsHDAC1, 2 and 3. OSHDAC1 is expressed at similar levels in leaves, roots and callus cells, whereas OsHDAC2 and 3 show a tissue specific expression pattern. They are expressed in roots and callus cells but not in leaves (Jang et al., 2003). Two RPD3-type HDACs, ZmRpd3/HD1BI and HD1BII have been characterized in maize embryos and they have been shown to localize predominantly to the nucleus as chromatin bound proteins (Brosch et al., 1992; Kölle et al., 1999; Lechner et al., 2000). Genetic analysis of Aspergillus nidulans revealed the presence of two putative histone deacetylase genes, rpdA and hosA, which are the first to be analyzed from filamentous fungi. PCR analysis and southern hybridization experiments have shown that no further members of the Rpd3 family are present in the genome of the fungus (Graessle et al., 2000).

illustration not visible in this excerpt

Sequence analysis performed on the class I HDACs showed that most of them contain a large conserved domain homologous to the yeast Rpd3 covering most of the N-terminal end. The C-terminal region has more variable sequence (Khochbin and Wolffe, 1997). The Rpd3 homology domain is also shared by several prokaryotic proteins that interact with various acetylated substrates (Leipe and Landsman, 1997). Mutagenesis of this homology domain confirmed its involvement in deacetylase enzymatic activity (Wu et al., 2000a).

The overexpression of rice OsHDAC1 in transgenic calli, leaves and roots correlated with an increase of HDAC activity and an elevated acetylation level of H4. This increase in acetylation of H4 resulted in a significant increase in the growth rate and developmental abnormalities in shoot and root tissues. The transgenic rice plants were also thicker and shorter than the non-transgenic plants (Jang et al., 2003). It has been shown (Murfett et al., 2001) that mutating the Arabidopsis AtHDAC6 gene leads to transgene silencing. The antisense inhibition of At1HD1 caused dramatic reduction in endogenous AtHD1 transcription which in turn resulted in accumulation of acetylated histone, notably tetra-acetylated H4. This reduction in AtHD1 is associated with various developmental abnormalities, including early senescence, ectopic expression of silenced genes, suppression of apical dominance, flower defects and male and female sterility. Some of these phenotypes could be attributed to ectopic expression of tissue-specific genes, suggesting that AtHD1 is a global regulator controlling gene expression during development through epigenetic mechanisms (Tian and Chen, 2001).

In animals, Rb (retinoblastoma) protein is a key player in cell cycle regulation. It binds to the S-phase specific transcription factor E2F, blocking its transcriptional activation and creating a repressive chromatin environment through the recruitment of HDAC1 (Brehm et al., 1998). Immunological data indicate that the maize ZmRpd3/HD1BI and HD1BII cofractionate with a protein that is related to tomato LeMSI and Rbap46/48) proteins (Lechner et al., 2000). Later, glutathione-S- transferase (GST) pull-down experiments showed that ZmRpd3 proteins can interact with the maize retinoblastoma related protein 1 (ZmRBR1) and with the maize retinoblastoma associated protein 1 (ZmRbAp1) which is a histone binding protein that participate in forming a ZmRBR1/ZmRbAp1 complex (Varotto et al., 2003). These results suggest a general role for the Rpd3 proteins in the plant cell cycle and development. Furthermore, ZmRpd3/HD1BI can functionally complement a yeast rpd3 null mutant (Rossi et al., 1998). In maize, only the HD1B activity can deacetylate the characteristic deacetylation pattern introduced by maize HATB on H4, suggesting a role for it in the histone deposition process (Lusser et al., 1999). The Arabidopsis HDAC, AtRPD3A, is induced by jasmonate (JA) and ethylene, which are produced in response to pathogen attack. In addition, the expression of AtRPD3A:ß-glucuronsidase (GUS) was induced by wounding, jasmonate, ethylene and the pathogen Alternaia brassicicola. Overexpression of AtRPD3A in transgenic Arabidopsis resulted in increase in resistance to the pathogen A. brassicicola whereas silencing of AtRPD3A by RNA interference (RNAi) resulted in decreased resistance to the pathogen (Zhou et al., 2005).

II- The HDAl-like histone deacetylases- Class II:

Class II HDACs are able to shuttle between nucleus and cytoplasm in response to certain cellular signals (de Ruijter et al., 2003). The human HDAC6 is localized exclusively in the cytoplasm, where it associates with microtubules and localizes with the microtubule motor complex containing the p150 protein (Hubbert et al.,

2002) . HDAC 10 can localize in the cytoplasm and the nucleus, although the function of localization in both compartments has not been clarified (Fischer et al., 2002; Kao et al., 2002). The subcellular localization of HDAC9 can be cytosolic as well as nuclear, depending on the splice variant (Bertos et al., 2001; Zhou et al., 2001). The shuttling of HDACs 4, 5 and 7 is stringently regulated (de Ruijter et al.,

2003) . For example, in differentiating muscle cells, due to a pre-differentiation signal, HDAC4 is phosphorylated by Ca² +/calmodulin-dependent kinase (CaMK), resulting in export of HDAC4 together with CRM1, a cellular export factor for proteins with a leucine-rich NES, into the cytoplasm. In the cytoplasm, 14-3-3 protein, a cytosolic anchor protein, binds the phosphorylated HDAC4, thereby retaining the HDAC4 in the cytosol. After fusion of muscle cells, terminal differentiation occurs and HDAC4 is released from the 14-3-3 protein due to decrease in its phosphorylation status and consequently shuttle back to the nucleus (Bertos et al., 2001; de Ruijter et al., 2003).

HDAC4, 5, 7 and 9 contain highly conserved C-terminal catalytic domains (about 420 amino acids) homologous to yeast HDA1, and N-terminal domains with no similarity to other proteins. The N-terminal domain is used primarily as a targeting domain for distinct promoters by the MEF2 transcription factors (Grozinger et al., 1999; Verdin et al., 2003). In contrast to other HDACs, the histone deacetylase activity of HDAC6 is found to be resistant to trapoxin and to a new potent deacetylase inhibitor (CHAP1), whereas, like other HDACs, HDAC6 is inhibited efficiently by trichostatin A (Furumai et al., 2001). Interestingly, HDAC6 possesses two deacetylase domains in addition to a zinc finger motif. The two catalytic domains appear to function independently of each other as the truncated versions of HDAC6, each containing one domain, are fully functional HDACs and contribute independently to overall activity of the wild type HDAC6 protein. Homologues of this double domain protein have been found in mice, Caenorhabditis elegans and Drosophila melanogaster (Grozinger et al., 1999; Pandey et al., 2002).

Class II HDACs represent the catalytic domain of large multi-protein complexes. They do not bind directly to DNA and are thought to be recruited to specific promoters through their interaction with DNA sequence-specific transcription factors. The MEF2 family of transcription factors is one of the major targets of class II HDACs. Other interactions occur with CtBP (E1A C-terminal binding proteins), 14-3-3 proteins, calmodulin (CaM), transcriptional co-repressors, heterochromatin protein 1 (HP1a) and SUMO (Verdin et al., 2003).

The enzymatic activity of class II HDACs is regulated at several levels, including tissue specific gene expression, recruitment of distinct cofactors and nucleo- cytoplasmic shuttling (Verdin et al., 2003). The human HDAC7 has enzymatic activity in vivo when it is in the cell nucleus but not when it is in the cytoplasm. This enzymatic activity is dependent on the association of the C-terminal region of HDAC7 with class I HD AC HDAC3. This interaction is mediated by the transcriptional corepressors SMRT and NCoR which simultaneously bind class II HDACs and HDAC3 by two distinct repressor domains (Fischle et al., 2001).

HDA, a class II HDAC in A. nidulans is a main contributor to the HDAC activity in this fungus (Tribus et al., 2005). An hdaA null mutant strain did not show any phenotypic difference in comparison to wild type under normal growth conditions. However, growth of the hdaA mutant was significantly inhibited under oxidative stress conditions compared to the wild type under the same. This indicates that this HDAC activity plays a role in controlling the transcription of genes responsible for protecting the fungus against oxidative stress conditions (Tribus et al., 2005).

A loosely chromatin-bound HDAC, ZmHdal, has been partially characterized and purified to homogeneity by chromatography from maize extracts as a 48 KDa protein that is subject to phosphorylation (Brosch et al., 1992; Brosch et al., 1996a). In vitro dephosphorylation of the purified enzyme is accompanied by a significant increase of deacetylase activity and altered substrate specificity with respect to differentially acetylated isoforms of histone H4 (Kölle et al., 1999). The ZmHda1 is expressed in root, leaf and shoot tissues as demonstrated by Northern blot analysis (Pipal et al., 2003). It has been thought that this enzyme exists in 3 forms. One form is of low molecular size, 48 kDa, it is the only enzymatically active form and exists as a monomer. The other two forms are of higher molecular size, 84 and 65 KDa, which occur as high molecular weight complexes and are not enzymatically active. By in vitro limited proteolysis, these high molecular size proteins convert into the lower molecular size form with a significant increase in enzymatic activity. However, the function and composition of these high molecular weight complexes are not known (Pipal et al., 2003). The transactivation experiments using different ZmHda1 truncated versions showed that only the 48 KDa form is the biologically active protein and may have a repressive effect on transcription in vivo (Pipal et al., 2003).

III- The Sir2-like histone deacetylases- Class III:

Sir2-like enzymes (sirtuins) named after the founding member, the yeast Silent information regulator 2 (Sir2). Unlike class I and II, Sir-like proteins catalyze a unique reaction that requires the coenzyme NAD+. In this reaction, nicotinamide is liberated from NAD+ and acetyl group of the substrate is transferred to cleaved NAD+ generating the novel metabolite O-acetyl-ADP ribose (OAADPr) (Imai et al., 2000; Denu, 2003). In yeast, Sir2 is required for silencing at telomeres, mating type loci (Aparicio et al., 1991) and rDNA (Fritze et al., 1997). In contrast to class I and II enzymes, the histone deacetylase activity of Sir2 is not inhibited by the potent histone deacetylase inhibitor trichostatin A (Imai et al., 2000). The fact that Sir2 as a chromatin modifying enzyme requires a coenzyme, NAD+, derived from metabolic pathways, suggest a tight synchronization among apparently diverse cellular processes (Denu, 2003).

Sir2 is the most evolutionary conserved protein deacetylase, with homologs in all kingdoms but most of the current understanding of Sir2 cellular functions is derived from genetic studies in yeast (Furuyama et al., 2004). Now evidence is present that Drosophila Sir2 is involved in epigenetic silencing by the polycomb group proteins. Sir2 mutations enhance the phenotypes of polycomb group mutants and disrupt silencing of a mini-white reporter transgene mediated by a polycomb response element. Consistent with this, Sir2 is physically associated with components of an E(Z) histone methyltransferase complex as it binds to many euchromatic sites on the polytene chromosomes and colocalize with E(Z) at most sites (Furuyama et al., 2004). Sir2 has also been shown to have an intrinsic ADP- ribosyltransferase activity (Tanny et al., 1999; Frye, 2000).

IV- The plant specific HD2-histone deacetylase:

The founding member of this family, HD2, was first identified in germinating maize seedlings as tightly chromatin bound, acidic, high molecular size (about 400 KDa) protein. This enzyme is composed of three almost identical polypeptides of 39, 42 and 45 KDa (Brosch et al., 1996b). It has been shown that this enzyme localizes to the nucleolus. This nucleolar localization suggests a possible role in the regulation of rRNA genes expression. The enzymatic activity of HD2 is regulated by phosphorylation; dephosphorylation of the HD2 complex abolishes its enzymatic activity (Lusser et al., 1997). HD2-like HDACs form multigene families within the plant kingdom: three HD2 variants in the maize genome, four homologs in A. thaliana, and one in Oryza sativa have been identified. No closely related proteins have been identified in animals and fungi (Dangl et al., 2001). Functional analysis of the Arabidopsis homologs of HD2 showed a repressive effect on genes when targeted to promoters used to drive reporter gene expression. On the other hand, silencing the AtHD2A genes resulted in aborted seed development in transgenic Arabidopsis plants, suggesting a role in plant reproductive development (Wu et al., 2000b). Furthermore, ectopic expression of AtHD2A as a fusion protein disrupted normal development and generated many pleitropic effects in a variety of somatic and reproductive tissues. Moreover, overexpression of AtHD2A may induce silencing of the endogenous HD2 genes (Zhou et al., 2004).

C- HDAC inhibitors (HDACi):

There are many HDACi known, but the most potent discovered so far is trichostatin A (TSA). TSA is a fermentation product of Streptomyces, originally used as an antifungal agent. Later it was discovered to have potent proliferation inhibitory properties when applied to cancer cells (Yoshida et al., 1990; de Ruijter et al., 2003).

Currently, there are several structurally diverse classes of HDACi, natural and synthetic. They are classified according to their structure into:

a- short-chain fatty acids: e.g. valproic acid, sodium butyrate, phenylbutyrate and AN-9.
b- cyclic and non-cyclic hydroxamates: e.g. suberoyl anilide hydroxamic acid, TSA, m-Carboxycinnamic acid bishydroxamic acid, suberic bishydroxamic acid, oxamflatin, proxamide and PDX101.
c- cyclic peptides or tetrapeptides: e.g. Depsipeptide, Trapoxin, and apicidin. d- benzamides: e.g. MS-275 and CI-994 (p-N-acetyl dinaline). e- ketones: e.g. trifluoromethyl ketone and a-ketoamides. f- hybrids of hydroxamic acid and cyclic tetrapeptide: e.g. CHAP.

Among them, sodium phenyl-butyrate, SAHA, LAQ824, depsipeptide, MS-275, CI-994, pyroxamide, PXD101, and valproic acid are already in phase I or II clinical trials for treatment of cancer patients (de Ruijter et al., 2003; Liu et al., 2006).

HDACis are emerging as a new class of anticancer agents. HDACis have shown activity against diverse cancer types and notable effects on tumour cell proliferation, programmed cell death, differentiation and angiogenesis in vitro and in vivo. Currently, there are more than a dozen of phase I and II clinical trials involving the use of HDACis in patients with hematological and solid malignancies. Data from the clinical trials demonstrate that HDACi treatment leads to tumour regression and symptomatic improvement in some heavily pre-treated and multiply relapsed patients at an advanced stage with a surprisingly low side- effect profile, and thus a wide therapeutic index (Liu et al., 2006).

Although the effects of HDACi in inhibiting HDAC activity are considered to be the same, there are two exceptions with regard to the mechanism of inhibition. All HDACi inhibit HDAC in a reversible fashion, except for trapoxin and depudesin, which inhibit the enzyme irreversibly through a different mechanism, namely via covalent binding to the epoxyketone group (de Ruijter et al., 2003).

HDACis are also implicated in host-pathogen interactions. For example, HC-toxin, a cyclic tetrapeptide, is a determinant of virulence and host specificity in the interaction between the producing fungus, Cochliobolus carbonum, and its host maize plant (Baidyaroy et al., 2002). When maize becomes infected with C. carbonum, the fungus secretes HC-toxin which causes the maize core histones to become hyperacetylated in vivo in susceptible, but not in resistant maize lines. The HC-toxin is believed to either interfere with the expression of maize defence genes or to inhibit protein synthesis by interference with ribosomal gene expression through action on the nucleolar deacetylase HD2 (Brosch et al., 1995; Ransom and Walton, 1997; Brosch et al., 2001). Interestingly, the HDAC activity in the toxin producing C. carbonum, is toxin resistant. This activity was partially purified and it has a native molecular size of 60 KDa, it does not react with anti-Cochliobolus HDAC1 and HDAC2 antibodies suggesting that it may be a yet novel HDAC (Brosch et al., 2001).

5- Histone méthylation:

Histone methylation is lagging behind acetylation in its characterization, but it is rapidly catching up (Kouzarides, 2002). Protein methylation is a covalent modification commonly occurring on carboxyl groups of glutamate, leucine and isoprenylated cysteine or on the side chain nitrogen atoms of lysine, arginine and histidine residues (Clarke, 1993). But histones are known to be methylated only on arginine and lysine residues. Methylation of core histones was reported almost 40 years ago (Murray, 1964), but its function remains unclear. Arginine residues can be either mono- or di-methylated at the guanidino nitrogens with the latter in symmetric or asymmetric configuration. Lysine can be methylated at its ε-amino group with up to three methyl groups per residue (figure 4). Histone methyltransferases (HMTs) catalyze the transfer of methyl group from S-adenosyl- L-methionine (SAM) to either arginine or lysine residues. Two major groups of HMTs can be distinguished, depending on their target amino acid. Arginine methylation is catalyzed by protein arginine methyltransferases (PRMTs) while lysine methylation is catalyzed by SET-domain containing HMTs, although a non­SET protein (DOT1) has been reported to methylate lysine 79 of histone H3 (Feng et al., 2002), thus potentially defining a third group.

One major obstacle in studying the function of histone methylation was the lack of information regarding the responsible enzymes, but the demonstration that a nuclear receptor coactivator-associated protein, CARM1 (also known as PRMT4), is an H3 specific arginine methyltransferase and that the human homologue of the Drosophila heterochromatin protein Su(var) 3-9 is an H3 specific lysine methyltransferase, provided substantial evidence for the involvement of histone methylation in transcriptional regulation (Chen et al., 1999; Rea et al., 2000).

While histone acetylation and phosphorylation are highly dynamic, histone methylation appears more stable. In particular, lysine methylation is thermodynamically very stable and the amino-methyl group is resistant to direct cleavage of the N-CH3 bond. This stability is thought to contribute to the epigenetic programs, which are responsible for the non-Mendelian inheritance of phenotypic alterations. The recent discovery of histone lysine demethylases that reversibly remove methyl marks appear to challenge the epigenetic potential of histone lysine methylation (reviewed by Kubicek and Jenuwein, 2004; Trojer and Reinberg, 2006).

illustration not visible in this excerpt

Figure 4: A) méthylation of arginine residues. B) méthylation of lysine residues (Zhang and Reinberg, 2001).

A- The PRMT family of HMTs:

The méthylation of arginine residues has been reported in the late sixtieths last century (Paik and Kim, 1968) and the first protein arginine methyltransferase was described in 1977 (Lee et al., 1977). The sites of arginine methylation on the N- terminal tails of histones are R2, 17 and 26 on H3 and R3 on H4. The PRMTs can be classified into two types with type I catalyzing the formation of monomethylarginine and asymmetric dimethylarginine residues, whereas type II catalyzes the formation of monomethylarginine and symmetric dimethylarginine residues (figure 4) (Gary and Clarke, 1998). Several RNA-associated proteins including hnRNP A1, fibrillarin and nucleolin have been identified as substrates for the type I PRMTs, whereas the only substrate identified being symmetrically dimethylated for type II PRMTs is the myelin basic protein (Gary and Clarke, 1998; Zhang and Reinberg, 2001).

The PRMTs share a conserved catalytic core, but have little similarity outside the core domain. Sequence alignment revealed several highly conserved regions involved in SAM binding and catalysis (figure 5), therefore it is likely that these enzymes use a similar reaction mechanism but differ with regard to substrate specificity (Zhang and Reinberg, 2001). Moreover, the suggested site for arginine

Figure 5: Schematic representation of five mammalian PRMT proteins (Zhang and Reinberg, 2001).

binding contains several amino acids that appear to be conserved throughout the PRMT family (Weiss et al., 2000; Zhang et al., 2000).

Many proteins have been identified as being methylated at arginine residues using site specific antibodies and mass spectrometry (MS) and they are involved in a broad range of biological activities such as pre-mRNA processing, translation, transcription, signaling, apoptosis and cytoskeleton structure (Boisvert et al., 2003).

Type I PRMTs

PRMT1. It is the founding member of the PRMT group of proteins (Lin et al., 1996). PRMT1 and its yeast homologue RMT1/HMT1 are well characterized proteins. Recombinant PRMT1 has intrinsic protein arginine methyltransferase activity toward arginine residues in the RGG and RXR motifs of many RNA binding proteins (Lin et al., 1996; Zhang and Reinberg, 2001). In a study designed to isolate enzymes responsible for the methylation of H4 resulted in the purification of PRMT1 as a major H4-specific enzymatic activity. PRMT1 methylates H4 at R3 in vivo (Wang et al., 2001). It can also methylate H2A at R3 since H2A has the extreme N-terminal sequence "SGRGK" as that of H4 (Wang et al., 2001). The PRMT1 enzymatic activity resides in a single polypeptide of 43 KDa which functions as a 350 KDa oligomer (probably homo-oligomer) (Wang et al., 2001). PRMT1 is implicated in multiple cellular processes through methylating proteins involved in nuclear-cytoplasm transport, signal transduction and transcription.

Nuclear-cytoplasmic transport: the yeast Np13p protein is an abundant hnRNP implicated in pre-rRNA processing and RNA transport (Russell and Tollervey, 1992; Lee et al., 1996). In a genetic screen aimed at isolating proteins that interact with Np13p, the gene coding for the yeast homologue of PRMT1 was isolated (Henry and Silver, 1996). The gene product (HMT1/RMT1) is able to methylate
the Np13p protein in vitro and in vivo (Henry and Silver, 1996). Mutations in the SAM binding domain of PRMT1 where shown to block methylation of Np13p in vivo, providing evidence that Np13p is an in vivo substrate (McBride et al., 2000). Importantly, these mutations affected the nuclear-cytoplasm transport of Np13p. In addition, PRMT1-mediated methylation of another hnRNP protein, Hrp1p, was also shown to facilitate its nuclear export (Shen et al., 1998). Therefore, one function of HMT1/RMT1 in yeast is to regulate the subcellular localization of Np13p which is necessary for pre-rRNA processing and RNA transport.

Interferon-mediated signal transduction: PRMT1 has been shown to interact with the intracellular domain of the interferon-α/β receptor (Abramovich et al., 1997). The cytoplasmic domain of this receptor is also a docking site for many signaling proteins such as tyrosine kinases, serine/threonine kinases, and STAT (signal transducers and activators of transcription) transcription factor (Leaman et al., 1996). More recent studies indicate that PRMT1 regulates the transcription activity of STAT1 by methylating the protein at R31. Methylation of STAT1 enhances its DNA binding activity and therefore ensures interferon responsiveness of cells, as it could be shown by inhibition of PRMT1 methyltransferase activity (Mowen et al., 2001).

Histone methylation and transcriptional activation: in addition to methylating transcriptional factors such as STAT1, PRMT1 also participates in transcriptional regulation through methylation of core histones. PRMT1 is specifically responsible for methylating H4 at R3 as the methylation on this residue is not detected in H4 from PRMT1 null embryonic stem (ES) cells (Wang et al., 2001). Consistent with the potential role of PRMT1 in transcription, PRMT1-mediated methylation of H4 R3 facilitates its subsequent acetylation by p300 (Wang et al., 2001). This observation provides the molecular explanation why PRMT1 functions as a transcriptional coactivator with members of the p160 family of coactivators (Koh et al., 2001). In accordance is the observation that mutation of the SAM-binding site of PRMT1 crippled PRMT1 enzymatic activity as well as its coactivator function (Wang et al., 2001).

PRMT4/CARM1. The cofactor associated arginine methyltransferase 1 (CARM1) was identified in a yeast two hybrid screen using the C-terminus of GRIP1, a member of the p160 family of coactivator, indicating that it has a role in transcriptional regulation (Chen et al., 1999). CARM1 preferentially methylates H3 in vitro (Chen et al., 1999) and mapping of residues demonstrated specificity for R2, R17 and R26 residues; it also methylates the C-terminus of H3 at R128, R129, R131 and R134 residues (Schurter et al., 2001). However, which residues are methylated by CARM1 in vivo remains to be determined. CARM1/PRMT1 is a nuclear protein as shown by GFP-fusion protein localization experiments (Frankel et al., 2002). It has been found that CARM1 is able to methylate the histone acetylase CBP/p300 even in the presence of a high ratio of histones to p300 in the reaction mixture. This methylation takes place near the KIX domain which is necessary for interaction between p300 and CREB, causing inhibition of this interaction and thus inhibits CREB signaling (Xu et al., 2001). This provides a new model for the crosstalk between arginine methylation and lysine acetylation. Both PRMT1 and CARM1 have been shown to interact with the p160 coactivator family and synergistically stimulated transcription by nuclear receptors (Koh et al., 2001).

PRMT2. PRMT2 was isolated based on its similarity with PRMT1 by screening EST databases (Katsanis et al., 1997). PRMT2 has a unique SH3 domain in its N- terminal region; this domain may interact with polyproline-regions in signaling proteins (Katsanis et al., 1997; Scott et al., 1998). This may indicate a role for PRMT2 in transcriptional regulation. In accordance with this, in a yeast two hybrid screen, PRMT2 was found to interact with ERa (estrogen receptor alpha), a member of the nuclear receptors family (Qi et al., 2002). Further, it has been shown that the SAM- binding domain of PRMT2 is crucial for its coactivator activity (Qi et al., 2002). Thus, although the enzymatic activity of PRMT2 has not been described in vitro, a methyltransferase activity is likely.

PRMT3. PRMT3 was isolated in a yeast two-hybrid screen to identify proteins that interact with PRMT1 (Tang et al., 1998). The identified protein showed substantial sequence similarity to PRMT1 and had methyltransferase activity (Tang et al., 1998). However, there are differences between PRMT1 and PRMT3 regarding the oligomerization properties, substrate specificity and subcellular localization. PRMT3 function as a monomer and is predominantly localized in the cytoplasm (Tang et al., 1998). In contrast to the recombinant PRMT1, which was shown to methylate numerous substrates in cell extracts, methylation of cell extract proteins by recombinant PRMT3 occurs only following RNase treatment (Frankel and Clarke, 1999), indicating a possible regulation of protein methylation by RNA binding in vivo. A unique feature of PRMT3 is the presence of a C2H2 zinc finger at its amino terminus. This zinc finger is required for recognition of RNA- associated substrates in cell extracts and therefore, appears to play an important role in determining substrate specificity of PRMT3. In addition, the PRMT3 amino terminal domain contains tyrosine-phosphorylation consensus sequence, this may be a site for specific regulation of this enzyme (Frankel and Clarke, 2000).

PRMT6. In searching the human genome for protein arginine N-methyltransferase (PRMT) family members, a novel gene has been found on chromosome 1 that encodes for an apparent methyltransferase, PRMT6. The polypeptide chain of PRMT6 is 41.9 kDa consisting of a catalytic core sequence common to other PRMT enzymes. Expressed as a GST-fusion protein, PRMT6 demonstrates type I PRMT activity. A comparison of substrate specificity reveals that PRMT6 is functionally distinct from the previously characterized type I enzymes, PRMT1 and PRMT4. In addition, PRMT6 displays automethylation activity; it is the first PRMT to do so. This novel human PRMT resides solely in the nucleus when fused to the green fluorescent protein (Frankel et al., 2002). The human immunodeficiency virus (HIV) transactivator protein, TAT, is specifically associated with and methylated by PRMT6 within cells. Overexpression of wild type PRMT6, but not the methylase inactive PRMT6 mutant, decreased TAT transactivation of an HIV-1 long terminal repeat luciferase reporter plasmid in a dose dependent manner. Knocking down PRMT6 consistently increased HIV-1 production and led to increased viral infectiousness (Boulanger et al., 2005). It has also been found that PRMT6 specifically methylates the non-histone protein HMGA1a (Miranda et al., 2005).

PRMT8. It is the most recently characterized PRMT family member (Lee et al., 2005a). This novel enzyme is most closely related to PRMT1, although it has a distinctive N-terminal region. The unique N-terminal end harbors a myristoylation motif and a patch of basic residues. PRMT8 is indeed modified by the attachment of a myristate to the glycine residue after the initiator methionine. The myristoylation of PRMT8 results in its unique association with the plasma membrane (Lee et al., 2005a). A second unique property of PRMT8 is its tissue- specific expression pattern; it is largely expressed in the brain (Lee et al., 2005a). A GST fusion protein of PRMT8 has type I PRMT activity, catalyzing the formation of monomethylated and asymmetrically dimethylated arginine residues on a recombinant glycine- and arginine-rich substrate (Lee et al., 2005a).

Type II PRMTs

PRMT5/JBP1. It was first identified as a kinase binding protein in a yeast two- hybrid screen (Gilbreth et al., 1998; Pollack et al., 1999). Amino acid similarity has been found between this protein and other members of the PRMT family. The enzymatic activity of this protein has been later demonstrated using the recombinant form and it has been named PRMT5 (Pollack et al., 1999; Lee et al., 2000; Rho et al., 2001). PRMT5 is localized predominantly in the cytoplasm (Rho et al., 2001) and it is able to methylate myelin basic protein, fibrillarin and histones H4 and H2A in vitro (Pollack et al., 1999; Lee et al., 2000; Rho et al., 2001). Whether these proteins are the in vivo substrates for PRMT5 remains to be determined. PRMT5 can catalyze the symmetric methylation of guanidino nitrogen of the arginine residues, representing the first type II PRMT member to be identified (Branscombe et al., 2001). PRMT5 was found to be a negative regulator of mitosis as the protein associates with the cdc2 complex in Schizosaccharomyces pombe and its null mutants exhibited phenotypes characteristic of mitotic inhibitors (Gilbreth et al., 1998). Moreover, PRMT5 has been found to be a negative regulator of the Swe1 kinase, an important cell cycle regulator (Ma et al., 1996). Furthermore, chromatin immunoprecipitation (CHIP) experiments revealed the recruitment of PRMT5 to the cyclin E1 promoter in G0 arrested cells, in which cyclin El expression is down regulated. Overexpression of recombinant PRMT5 abolishes cyclin E1 expression (Fabbrizio et al., 2002).

PRMT7. It has been discovered through searching the human genome for novel PRMTs (Miranda et al., 2004). When expressed as full length human cDNA construct in Escherichia coli as a GST-tagged protein, it is able to catalyze the SAM-dependent monomethylation of the synthetic peptide GGPGGRGGPGG- NH2 (Miranda et al., 2004). Later, the immunopurified native PRMT7 has been shown to have the ability to methylate histones, myelin basic protein, a fragment of human fibrillarin and splicesomal protein SmB. Furthermore, amino acid analysis showed that the modifications produced were predominantly symmetric dimethylarginine and monomethylarginine in addition to small amounts of asymmetric dimethylarginine. These data demonstrate that PRMT7 is a type II PRMT (Lee et al., 2005b). A unique feature of PRMT7 is that it seems to have arisen from a gene duplication event as it contains two putative SAM-binding domains. However, when each domain was separately expressed as a GST-tagged fusion protein and tested for activity using synthetic peptide, none of them showed enzymatic activity. This indicates that both domains are required for functionality (Miranda et al., 2004).

Very little is known about protein arginine methylation in plants, especially histones methylation (Tuck and Paik, 1984). Arginine methylation of many pea chloroplast polypeptides has been identified and it has been shown that this methylation is stimulated by light (Niemi et al., 1990). In vitro arginine methylation of calf histone H4 was found to be catalyzed by partially purified protein methylase I, from wheat germ (Gupta et al., 1982). This enzyme is also able to methylate H4, H2B and H2A extracted from wheat germ indicating that these histones may be the in vivo substrates (Gupta et al., 1982). The proteolytic digestion followed by HPLC separation of the H4 peptides showed that the R35 residue could be the methylation site (Disa et al., 1986). The function of this histone arginine methylation in wheat and generally in plants is not known, mainly due to the lack of any information about the enzymes responsible for methylation.

B- SET domain HMTs:

Methylation of lysine residues is known to occur on histone H3 at lysines K4, K9, K27, K36 and K79, on histone H4 at K20 (Kouzarides, 2002). Recently, K14, K18 and K29 on H3 and K59 on H4, and K13, K15 and K99 on H2A have been identified as methylation sites (Zhang et al., 2003a), but no enzymes have been identified that methylate these residues. Lysine methylation is also present at the histone H1 amino terminus (Syed et al., 1992; Lachner and Jenuwein, 2002; Kuzmichev et al., 2004). The SUV39 protein was the first histone lysine methyltransferase to be discovered (Rea et al., 2000). The methyltransferase activity of SUV39 is directed against lysine 9 of histone H3 and its catalytic domain resides within a highly conserved structure, the SET domain. The SET domain was initially identified in Drosophila position effect variegation (PEV) suppressor SU(VAR)3-9, the Polycomb group protein Enhancer of zeste (EZ) and the Trithorax group protein Trithorax, hence it was given the name SET. The SET domain is found in all lysine methyltransferases except for the DOT1 protein (Feng et al., 2002). The SET domain consists of 130 amino acids and is crucial for the catalytic activity but requires adjacent cysteine-rich domains (Zhang and Reinberg, 2001; Lachner and Jenuwein, 2002).

Over 200 proteins of diverse functions, ranging from mammals to bacteria and viruses have been identified to contain this SET domain (http://smart.embl- heidelberg.de/smart/do annotation.pl?DOMAIN=SET&BLAST). A major function of SET-domain containing proteins is to modulate gene activity (Jenuwein et al., 1998). Methylation of lysine residues on histones creates a binding site for chromo (chromatin organization modifier)-domain containing proteins such as the HP1 protein (Jacobs et al., 2001; Jenuwein and Allis, 2001; Lachner et al., 2001b). The SET domain proteins can be classified into four classes (figure 6) depending on the similarity to the human SET domains and their relationship to SET domains in yeast (Kouzarides, 2002).

I- SUV39 family:

The most striking feature in this group of proteins is the presence of a PRE-SET domain in all members; this domain is not present in any protein outside the SUV39 family (Kouzarides, 2002). The PRE-SET domain seems to provide the specificity necessary for the SET domain to methylate K9 of histone H3, rather than any other lysine residue (Rea et al., 2000). In addition to the PRE-SET domain, another cysteine rich domain called POST-SET domain is present in this family (figure 6).

The family is named after its founding member, the SUV39H1, which is also the first HMT to be characterized (Rea et al., 2000). Originally, the Drosophila Su(var) 3-9 gene was identified in genetic screens aimed at isolating suppressors of PEV (Tschiersch et al., 1994). Later, the sequence similarity between the SET domain of SUV391 and some plant SET-domain containing methyltransferases (Klein and Houtz, 1995; Zheng et al., 1998) prompted Jenuwein and colleagues to test for possible methyltransferase activity associated with SUV39H1. This study revealed that SUV39H1 and its S. pombe homologue Clr4 contain intrinsic HMT activity specifically methylating H3 at K9 (Rea et al., 2000). The SUV39H1 and SUV39H2 proteins contain a chromodomain in the N-terminal region. This chromodomain is perhaps not present in any other SET domain protein (Kouzarides, 2002). The function of the chromodomain is still unclear. Although it does not bind methylated K9, it is still possible that it may recognize other methylated lysines in histones or other proteins (Kouzarides, 2002).

illustration not visible in this excerpt

Figure 6: Structural features within the SET-domain families (Kouzarides, 2002).

Posttranscriptional modifications have long been known to affect protein-protein interactions. The covalent histone modifications have been proposed to serve as markers, which are recognized by other proteins (Strahl and Allis, 2000). One piece of evidence that supports this hypothesis is the demonstration that the chromodomain present in the HP1 binds specifically to methylated H3K9 (Bannister et al., 2001; Lachner et al., 2001a; Nakayama et al., 2001). Studies have shown that the chromodomain present in HP1 and its S. pombe homolog Swi6 are required for heterochromatin formation. Importantly, HP1 colocalizes with SUV39H1 (Aagaard et al., 1999). In addition, heterochromatin localization of Swi6 requires functional Clr4 (Ekwall et al., 1996). These observations collectively suggest that the chromodomain of HP1/Swi6 recognizes the H3K9 methylated by SUV39H1/Clr4 (Zhang and Reinberg, 2001). Interestingly, one HP1 chromodomain mutant (V23M) that disrupts the methylated K9-specific binding activity also disrupts the gene silencing function of Drosophila HP1 (Platero et al., 1995). It has been found that mutations in yeast clr3, a gene encoding for an H3K14 specific deacetylase, impaired H3K9 methylation and heterochromatin Swi6 localization (Nakayama et al., 2001). This indicates that deacetylation of H3K14 functions upstream of H3K9 methylation and Swi6 localization. On the other hand, Swi6 mutations do not cause any detectable change in methylation of H3K9 (Nakayama et al., 2001). Therefore, SWi6 likely functions downstream of Clr4. On the basis of the above observations, a time course for heterochromatin formation can be inferred. After removal of the acetyl groups from K14 of H3 by Clr3, SUV39H1/Clr4 methylates H3K9. Methylated H3K9 serves as the binding site for the recruitment of HP1/Swi6. HP1/Swi6 can oligomerize to form heterochromatin through the shadowdomain localized at its carboxyl terminus (Brasher et al., 2000; Zhang and Reinberg, 2001). It is currently unknown, whether oligomerization of HP1/Swi6 requires H3K9 methylation at each nucleosome.

The existence of a SET-domain and cysteine rich regions in G9a protein prompted Tachibana and colleagues to test this protein for HMT activity (Tachibana et al., 2001). G9a is able to methylate H3 at K9 in addition to K27, a site which has been reported to be methylated in vivo (Strahl et al., 1999). Interestingly, these two residues are present in strikingly similar amino acid sequences, suggesting that G9a might recognize lysine residues within a motif, TKXXARKS, (Tachibana et al., 2001). In addition to SET, PRE-SET and POST-SET domains, G9a also contains a polyglutamic acid stretch and six ankyrin repeats at its amino terminus (Zhang and Reinberg, 2001; Kouzarides, 2002). The ankyrin repeats are present in other signalling molecules. These repeats are thought to represent a protein-protein interactions surface (Kouzarides, 2002). The floral repressor gene FLOWERING LOCUS C (FLC) in Arabidopsis acquires methylation at H3K27 and H3K9 during vernalization, a cold treatment which is required for floral induction. The proteins that are necessary for the maintenance phase of vernalization include the polycomb group protein (PcG) VERNALIZATION2 (VEN2) but the protein that catalyze this methylation is not yet known (Bastow et al., 2004; Sung and Amasino, 2004a). Two proteins, CLLD8 and ESET, each has a split SET domain and the adjacent PRE-SET and POST-SET domains as well (Mabuchi et al., 2001). Preliminary studies showed that ESET possesses H3 specific HMT activity (Blackburn et al., 2003) indicating that the SET domain can be separated into two subdomains without destroying its enzymatic activity. Interestingly, both ESET and CLLD8 contain MBD domain, a structure that is potentially capable of recognizing methylated DNA (Ng and Bird, 1999). This structure may direct the lysine methylase activity of ESET and CLLD8 to DNA-methylated promoters. In the fungus Neurospora crassa, DIM5 was shown to encode an H3K9 specific methylase and dim5 mutants showed abnormal growth and complete loss of DNA methylation (Tamaru et al., 2003). This suggests that H3K9 trimethylation provides a master epigenetic mark that controls growth and DNA methylation in Neurospora. Two independent forward genetic screens of Arabidopsis for the release of transcriptional gene silencing (TGS) led to the identification and characterization of the KRYPTONITE gene, the first H3K9-specific methylase in plants (Jackson et al., 2002). The Arabidopsis genome encodes nine SU(VAR)3-9- related genes, named SUVH1-SUVH9 (Baumbusch et al., 2001). In contrast to SU(VAR)3-9 and CLR4 proteins, the Arabidopsis SUVH proteins lack a chromo­domain at their N-terminus. Instead, the Arabidopsis SUVH proteins share a novel central sequence motif named the YDG domain, with divergent N-termini.