- 1. Introduction
- 2. Histone demethylation
- 3. LSD family
- 4. JmjC family
- 5. Inhibitors
- 6. DNA demethylation
- 7. Potential catalytic mechanism
- 8. Sequence analysis of TET family proteins
- 9. CXXC domain
- 10. CD domain
- 11. RNA demethylase
- 12. FTO
- 13. Structural and functional comparison
- 14. Concluding remarks
- References
- 1. Introduction
- 2. Histone demethylation
- 3. LSD family
- 4. JmjC family
- 5. Inhibitors
- 6. DNA demethylation
- 7. Potential catalytic mechanism
- 8. Sequence analysis of TET family proteins
- 9. CXXC domain
- 10. CD domain
- 11. RNA demethylase
- 12. FTO
- 13. Structural and functional comparison
- 14. Concluding remarks
- References
feature articles
Structure and function of dioxygenases in histone demethylation and DNA/RNA demethylation
aStructural Genomics Consortium, University of Toronto, 101 College Street, Toronto, Ontario M5G 1L7, Canada, bOntario Cancer Institute and Department of Medical Biophysics, University of Toronto, Toronto, Ontario M5G 2M9, Canada, and cDepartment of Physiology, University of Toronto, Toronto, Ontario M5S 1A8, Canada
*Correspondence e-mail: carrow@uhnres.utoronto.ca, jr.min@utoronto.ca
Iron(II) and 2-oxoglutarate (2OG)-dependent dioxygenases involved in histone and DNA/RNA demethylation convert the cosubstrate 2OG and oxygen to succinate and carbon dioxide, resulting in hydroxylation of the methyl group of the substrates and subsequent demethylation. Recent evidence has shown that these 2OG dioxygenases play vital roles in a variety of biological processes, including transcriptional regulation and gene expression. In this review, the structure and function of these dioxygenases in histone and nucleic acid demethylation will be discussed. Given the important roles of these 2OG dioxygenases, detailed analysis and comparison of the 2OG dioxygenases will guide the design of target-specific small-molecule chemical probes and inhibitors.
1. Introduction
Histone methylation together with DNA and RNA methylation are the centerpieces in epigenetic inheritance, and play pivotal roles in the epigenetic regulation of transcriptional activation and repression (Shi, 2007; Fu & He, 2012). Histone lysine (K) methylation often occurs on H3K4, K9, K27, K36, K79 and H4K20, which can be mono-, di- and trimethylated at the ∊-amino group by the SET domain histone methyltransferase with just one exception Dot1L, which harbors a canonical open α/β like that of the of protein arginine methyltransferases and DNA methyltransferases (Martin & Zhang, 2005; Min et al., 2003). While DNA methylation usually occurs at the 5-carbon position of the cytosine ring (5mC) in a CpG dinucleotide context (Bird, 2002), the epigenetic RNA methylation typically marks the adenosine N6 position (m6A) of and long noncoding RNA (Grosjean, 2005; Wei et al., 1975).
Recent studies have indicated that histone lysine methylation as well as DNA/RNA methylation is dynamically modified in order to regulate the methylation and demethylation levels in cells (Jia et al., 2011; Zheng et al., 2013; Song & He, 2012; Mosammaparast & Shi, 2010; Shi & Tsukada, 2013). It is worth noting that some FeII and 2-oxoglutarate (2OG) dioxygenases serve as demethylases of histone and DNA/RNA. FeII- and 2OG-dependent dioxygenases require ferrous iron as a cofactor and 2OG, namely α-ketoglutarate (αKG), as a cosubstrate for their (Hausinger, 2004). These dioxygenases contain a conserved double-stranded β-helix (DSBH or jelly-roll) domain, in which the FeII and 2OG are coordinated by some invariant residues (Clifton et al., 2006). Moreover, the observed distance between an oxidized substrate atom and iron is approximately 4.5 Å in most dioxygenases (Aik et al., 2012).
Up to now, a growing number of dioxygenases have been identified that can act on histones or DNA/RNA as an oxidized substrate. Structural studies revealed that these FeII- and 2OG-dependent dioxygenases not only contain a conserved DSBH domain, but also have some auxiliary functional domains (Klose et al., 2006; Tsukada, 2012).
This review focuses on the structure and function of the three kinds of epigenetic demethylases based on different substrates namely methylated histone lysine, DNA and RNA. In addition to dioxygenase-based JmjC-domain-containing histone demethylases, a pair of amine oxidases (LSD1 and LSD2) can also demethylate histone lysine H3K4, and these two classes of histone demethylases have different catalytic mechanisms. Interestingly, a non-dioxygenase protein LOXL2 was also reported to be able to remove the methylation of H3K4 by deaminating the target lysine (Herranz et al., 2012). Ten-eleven translocation proteins TET1–3 act as 5-methylcytosine hydroxylases and were found to promote active DNA demethylation (Xu et al., 2012) and, so far, only two members, FTO and ALKBH5, have been reported to be involved in m6A RNA demethylation (Zheng et al., 2013; Jia et al., 2011).
2. Histone demethylation
Histone lysine methylation plays a critical role in regulating various biological processes, such as transcriptional regulation, DNA replication and DNA repair (Martin & Zhang, 2005; Zhang & Reinberg, 2001). The methylation status of the lysine includes mono-, di- and trimethylation. Histone methylation had been considered to be irreversible for a long time until the first histone demethylase, lysine-specific demethylase I (LSD1) was identified in 2004 by Shi et al. (Shi et al., 2004). Subsequently, a Jumonji-like domain C-terminus (JmjC) domain-containing demethylase was identified by Tsukada et al. (2006). Therefore, histone lysine methylation is a dynamic and reversible modification through the balancing of histone lysine methyltransferases and histone lysine demethylases (KDMs) (Shi & Whetstine, 2007).
3. LSD family
LSD1 is a flavin adenine dinucleotide (FAD)-dependent amine oxidase. In this catalytic reaction, LSD1 utilizes FAD as a cofactor to catalyze the amine oxidation of the methylated lysine, creating an imine intermediate. The imine intermediate is spontaneously hydrolyzed to form an unstable carbinolamine intermediate, which releases formaldehyde and generates unmethylated lysine (Shi et al., 2004). However, lysine (K) residues exit in three methylated states, namely mono-, di- or trimethylated lysine (Kme1, Kme2 and Kme3). This requires protonated nitrogen on the ∊-amino group of methylated lysine in order to form the imine intermediate, but the nitrogen of trimethylated lysine cannot be protonated. Therefore, LSD1 can demethylate Kme1 and Kme2 but not Kme3.
The LSD family includes two members: LSD1 (also known as KDM1A) and LSD2 (also known as KDM1B), and their crystal structures in complex with substrates have been determined and the demethylation mechanism has also been studied (Yang et al., 2006; Forneris et al., 2007; Zhang et al., 2013; Fang et al., 2013). Crystallographic studies show that LSD1 and LSD2 share a conserved C-terminal amine oxidase domain (AOD) and a SWIRM (Swi3p, Rsc8p and Moira) domain. In addition, LSD1 has a unique insertion (referred to as the Tower domain) that is embedded in the AOD domain (Forneris et al., 2007). On the other hand, LSD2 has a unique zinc-finger domain at the N-terminus (Figs. 1, 2a and 2b) (Zhang et al., 2013). The conserved AOD domain contains two lobes: one is used for FAD binding, and the other is responsible for substrate binding. The interface of the two lobes forms a large highly negatively charged catalytic cavity, which is large enough to accommodate a large segment of H3 tail, suggesting that the inability of LSD1 and LSD2 to demethylate trimethylated lysine is not owing to structural but rather because of the inherent limitations of its chemical mechanism. The SWIRM domain adopts an α-helical fold that packs against the FAD-binding lobe of AOD domain. The SWIRM domain is responsible for maintaining the integrity of the overall core structure and is required for enzymatic activity (Yang et al., 2006; Zhang et al., 2013). The Tower domain of LSD1 is required for binding to CoREST (co-repressor of RE1-silencing transcription factor), which enables LSD1 to demethylate nucleosomes through the interactions between CoREST and nucleosomal DNA (Yang et al., 2006). Unlike LSD1, LSD2 is able to independently work on demethylation of nucleosomal substrates and the zinc-finger domain is required for LSD2 demethylase activity (Zhang et al., 2013; Fang et al., 2010; Yang, Jiang et al., 2010).
4. JmjC family
The JmjC family contains several subfamilies, which belong to the 2OG and ferrous iron-dependent dioxygenases (Cloos et al., 2008; Nottke et al., 2009). The JmjC domain family of demethylases uses 2OG and FeII as cofactors in the presence of molecular oxygen to hydroxylate the methylated histone, producing succinate, CO2 and unstable carbinolamine; subsequently carbinolamine spontaneously releases formaldehyde and generates demethylated lysine. Unlike LSD-mediated demethylation, the JmjC family is also able to demethylate trimethylated lysine (Tsukada et al., 2006; Klose et al., 2006).
Histone lysine methylation has specific physiological functions depending on the lysine modification sites and the degree of methylation (Martin & Zhang, 2005; Mosammaparast & Shi, 2010). JmjC domain-containing proteins exhibit different substrate preferences via their specific catalytic site architectures. Recent studies show that there are more than 30 different proteins that contain a JmjC domain in the human genome, and most of them have been proved to function as histone demethylases. The modification of H3K79 is unique and its specific demethylases have not yet been discovered (Shi & Tsukada, 2013).
Structural studies revealed that the conserved JmjC domain consists of eight β-strands, which form a jelly-roll fold (also known as double-stranded β-helix or DSBH) and belong to the cupin superfamily of metalloenzymes (Aik et al., 2012; Ng et al., 2007) (Fig. 3a). The active center is buried in the interior of the JmjC domain, where Fe2+ is coordinated by three conserved residues (H…D/E…H motif) and further stabilized by the 2OG cofactor (Fig. 3b). The substrate is bound by unusual C—H⋯O hydrogen bonds, which is formed between one methyl group of methylated lysine and the oxygen atoms from active-site residues (Couture et al., 2007) (Fig. 3c).
In addition to the core structure of the JmjC domain, most JmjC-dependent demethylases contain auxiliary functional domains that contribute to substrate recognition, recruitment and catalysis (Fig. 1, e.g. KDM4A), in addition to maintaining the overall structural stability. These auxiliary domains include JmjN, PHD, Tudor, CXXC, bright/arid, FBOX, zinc finger, TPR etc. The JmjN domain, which is formed by three short helices and two β-strands, is essential for the demethylation activity of JHDM3/JMJD2 (Chen et al., 2006). In C. elegans KDM7A, the PHD domain (plant homeo domain) is required for demethylase activity towards both H3K9me2 and H3K27me2 by specifically binding to H3K4me3 (Lin et al., 2010). The PHD domain of human PHF8 and KDM7A exhibits different orientations relative to the catalytic JmjC domain when interacting with a single peptide containing both H3K4me3 and H3K9me2. In other words, PHF8 adopts a bent conformation, allowing the PHD domain to interact with H3K4me3 while the catalytic JmjC domain acts on the H3K9me2 substrate (Qi et al., 2010). On the contrary, KDM7A adopts an extended conformation, therefore the JmjC domain cannot reach H3K9me2 when the PHD domain binds to H3K4me3 (Yang, Hu et al., 2010; Horton et al., 2010; Hou & Yu, 2010). The CXXC and bright/arid domains are responsible for DNA binding (Patsialou et al., 2005; Wu et al., 2013).
5. Inhibitors
Overexpression of histone demethylases is associated with multiple kinds of human disease, especially tumorigenesis (Højfeldt et al., 2013). Therefore, inhibitors of histone demethylases have recently been actively developed. These inhibitors are often designed to mimic the preferred cofactor/substrate. The tranylcypromine (PCPA) is reported to inhibit LSD1 catalysis via forming acovalent adducts with cosubstrate FAD in the active site (Schmidt & McCafferty, 2007; Yang et al., 2007) (Fig. 2c). Lots of inhibitors have been identified based on tranylcypromine analogs, such as phenelzine and pargyline. In addition, a series of mimic and polyamine-based inhibitors have also been improved to inhibit LSD1 (Thinnes et al., 2014).
N-Oxalylglycine (NOG) (an analog of 2OG) is a common inhibitor to block the activity of JmjC family via hindering oxygen binding to iron. These 2OG derivatives (e.g. NOG, S/R-2HG) have been used to capture inactive states of these enzymes in crystallographic studies. Structural studies show that NOG is coordinated by chelating a divalent metal ion and further stabilized by some residues (Fig. 3b). Some metal ion chelator-based inhibitors have been reported as competitive inhibitors of 2OG, such as 8-hydroxyquinolines, pyridinedicarboxylic acid (PDCA), daminozide and GSK-J1 (Højfeldt et al., 2013; Thinnes et al., 2014).
6. DNA demethylation
DNA methylation often occurs at the 5-carbon position of cytosine (5mC) in a CpG dinucleotide context, which is a key epigenetic mark crucial for genomic regulation and developmental processes in mammals (Bird, 2002; Suzuki & Bird, 2008). DNA methylation is a dynamic process controlled by DNA methyltransferases and demethylases (Song & He, 2012; Hahn et al., 2013; Smith & Meissner, 2013). In 1953, 5-hydroxymethylcytosine (5hmc) was first identified in T-even bacteriophages (Wyatt & Cohen, 1953), but it was not until 2009 that the ten-eleven translocation (TET) proteins were demonstrated to be the enzymes (dioxygenases) that could catalyze the oxidation of 5mC to 5hmc (Tahiliani et al., 2009). Further research has shown that the TET family of proteins not only convert 5mC to 5hmc, but also can further oxidize 5hmc into 5-formylcytosine (5fC) and 5-carboxylcytosine (5cac) in genomic DNA (He et al., 2011; Ito et al., 2010; Pfaffeneder et al., 2014). However, the unmethylated cytosine (C) is not regenerated through the oxidative reactions performed by TET proteins. In other words, the process of 5mC in DNA demethylation cannot be completed by TET proteins themselves. Recent studies suggest that both active and passive mechanisms (Inoue & Zhang, 2011; Inoue et al., 2011) are involved in the regulation of DNA demethylation. In the active DNA demethylation pathways, 5cac can be recognized and excized by thymine DNA glycosylase (TDG) (He et al., 2011). Alternatively, 5hmc can be converted to 5-hydroxymethyluracil (5hmu) by activation-induce deaminase and further converted to unmodified cytosine by TDG or SMUG1 (single-strand selective monofunctional uracil-DNA glycosylase 1) through the base excision repair pathway (Guo et al., 2011; Cortellino et al., 2011; Gavin et al., 2013). Thus, TET-oxidized 5hmc is a prerequisite step for the active DNA demethylation pathway.
7. Potential catalytic mechanism
The TET family proteins belong to 2-oxoglutarate (2OG) and ferrous iron-dependent dioxygenases. Their catalytic reaction requires 2OG as a cosubstrate and FeII as a cofactor. In the oxidation reaction, TET protein coordinates one oxygen atom from molecular oxygen (O2) into the substrate, yielding hydroxylation of the substrate, and the other oxygen atom into 2OG, leading to a decarboxylation of 2OG and a release of succinate and carbon dioxide (CO2) (Xiao et al., 2012).
8. Sequence analysis of TET family proteins
The members of the mammalian TET family include TET1, TET2 and TET3, which are located on chromosomes 10q21, 4q24 and 2p13 coding for 2136, 2002 and 1795/1660 amino acids, respectively (Mohr et al., 2011). The TET proteins share a conserved C-terminal The or CD domain contains Cys-rich and DSBH regions that are conserved in the cupin-like dioxygenase superfamily (Fig. 1). However, the CD domain includes an unknown spacer region (a large low-complexity insert), which is supposed to be a variable and unstructured region (Tahiliani et al., 2009). In addition, there is a conserved CXXC zinc-finger domain used for DNA-binding in TET1 and TET3, but not in TET2 (Fig. 1). The CXXC domain of TET1 and TET3 both lack a `KFGG' motif compared with that of the other CXXC domain family members (Allen et al., 2006; Xu et al., 2011).
9. CXXC domain
The human TET CXXC domain is highly conserved and important for recognizing the CpG dinucleotides. The TET1 CXXC domain can bind to unmodified C as well as 5mC and 5hmC-modified CpG-rich DNA, but does not exhibit binding to CG-deleted et al., 2011). However, an unmodified cytosine is absolutely required for the TET3 CXXC domain to interact with DNA and the CXXC domain strongly binds to cytosine-containing DNA with a slight preference for CpG-rich DNA, but does not bind to the A/T-only DNA (Xu et al., 2012). The Xenopus laevis Tet3 (xlTet3) CXXC domain shares a conserved sequence with the human TET3 CXXC domain and they exhibit similar DNA binding properties. Recently, two crystal structures of xlTet3-DNA complexes (PDB codes: 4hp3 and 4hp1 ) have been determined, which provided further insight into the mechanism of DNA binding (Xu et al., 2012). In both xlTet3 CXXC domain complexes, the key residue His90 forms two hydrogen bonds, with the target cytosine (backbone carbonyl) and the complementary guanine (side chain ND1), respectively. In the structure of the xlTet3 CXXC domain with the CmCGG containing dsDNA, Ser89 of the CXXC domain flips back the peptide plane to accommodate the guanosine, which pairs with 5mC, and the Gln91 side chain becomes disordered to accommodate the complementary methylated cytosine. Therefore, this complex uncovered the mechanism of the xlTet3 CXXC domain binding to non-CpG DNAs, such as CpA and CpT. It was also found that the binding mode is conserved in the human TET3 CXXC domain by sequence alignment and ITC binding experiments (Xu et al., 2012).
(Xu10. CD domain
The Naegleria Tet-like protein 1 (NgTet1), the homolog of mammalian TET1, can use 5mC, 5hmc or 5fC as substrates to generate 5cac. Recently, the complex crystal structures of NgTet1–DNA–NOG–Mn2+ (NOG, a 2-OG analog) and the TET2–DNA–NOG–Fe2+ have been determined (Hashimoto et al., 2014; Hu et al., 2013), and they adopt similar DSBH folding with the DNA located on the basic surface (Figs. 4a and 4b). However, in the TET2 structure, the Cys-rich domain wraps around the DSBH domain to maintain the structural and catalytic integrity, and three zinc atoms further stabilize the overall structure (Fig. 4b). Both of them use a base-flipping mechanism to target the methylated cytosine which results in a distortion of the B-form DNA duplex of ∼65° and ∼40° in NgTet1 and TET2, respectively. The flipped 5mC is stabilized via hydrogen bonds formed with three residues (Asn147, His297 and Asp234 in NgTet1, and Asn1387, His1904 and Tyr1902 in TET2), but the methyl group is not involved in any interaction (Figs. 4c and 4d). In addition, the distance between the target methyl group and metal ion (∼5.2 Å in NgTet1 and ∼5.0 Å in TET2) is able to accommodate the hydroxyl (Figs. 4c and 4d), formyl or carboxylate group other than methyl, consistent with the TET family being active on 5mC-oxidized derivatives (5hmc, 5fmc or 5cac) as a substrate or product. Surprisingly, these 5mC-interacting residues are crucial for the of NgTet1 and TET2, but not important for DNA-binding affinity in TET2 (Hu et al., 2013; Hashimoto et al., 2014).
11. RNA demethylase
RNAs serve vital roles in a wide variety of cellular processes (Sharp, 2009). There is increasing evidence that the modified are involved in RNA metabolism. To date, more than 100 distinct types of post-transcriptional modifications have been characterized in and other RNAs (Cantara et al., 2011; Machnicka et al., 2013). Methylation of the adenosine N6 position (m6A) is the most common internal modification in eukaryotic cells and viruses (Grosjean, 2005). The m6A modification is mostly found in the conserved motif Pu[G>A]m6AC[A/C/U], and the m6A sites have been identified in about 7000 transcripts and 300 noncoding RNA (Wei & Moss, 1977; Dominissini et al., 2012). Furthermore, m6A sequencing studies showed that the m6A modifications preferentially occur near the stop codon, 3′UTR and the long internal exons (Meyer et al., 2012; Dominissini et al., 2012). The methyltransferase METTL3 and METTL14 (`writer') are employed to catalyze the methylation (Liu et al., 2013). Knockdown of methyltransferase in embryonic stem (ES) cells leads to the loss of self-renewal capability, suggesting the importance of m6A modification in maintaining the properties of ES cells (Wang, Li et al., 2014; Minton, 2014). The YTH domain containing proteins YTHDF1–3 are found to serve as `readers' and bind the m6A. Recent studies showed that the `readers' influence the relocalization of to affect the stability of the in vivo (Wang, Lu et al., 2014).
The research on RNA methylation (m6A) has been stimulated by the characterization of RNA demethylases (`eraser'): fat mass and obesity-associated protein (FTO) and alkylation protein AlkB homolog 5 (ALKBH5), suggesting that the modifications are reversible and dynamic (Jia et al., 2011; Zheng et al., 2013). Both FTO and ALKBH5 are members of the AlkB subfamily of iron(II)/2-oxoglutarate-dependent dioxygenases.
12. FTO
FTO is highly expressed in the brain and is mainly localized in the nucleus; a low level of FTO can also be detected in the cytoplasm (Meyer & Jaffrey, 2014). The FTO gene is associated with the body mass index (BMI) and type 2 diabetes (Wehr et al., 2010; Frayling et al., 2007; Dina et al., 2007). Overexpression of FTO in mice leads to the increase of both food intake and body weight (Church et al., 2010). Whereas the Fto-deficient mice display lean phenotypes even with an increased food intake (Fischer et al., 2009; Church et al., 2009). Notably, the conserved residue R316Q mutant (which recognizes the 2OG) causes postnatal growth retardation, thereby demonstrating that demethylase activity is associated with growth and development (Boissel et al., 2009; Gulati & Yeo, 2013). Furthermore, more recent studies have demonstrated that FTO is also involved in glucose metabolism, Alzheimer's disease, neurological diseases, cancer, etc. (Lappalainen et al., 2011; Keller et al., 2011; Kaklamani et al., 2011; Tarabra et al., 2012). However, the molecular mechanisms underlying the role of FTO in these physiological processes remain unclear. Besides m6A, FTO also demethylases m3T of single-stranded DNA (ssDNA) and m3U of single-stranded RNA (ssRNA) (Jia et al., 2008; Gerken et al., 2007).
In 2010, Han and colleagues determined the 3T (Han et al., 2010). The overall structure can be divided into two domains, namely the N-terminal domain (NTD) and C-terminal domain (CTD) (Fig. 5a). Characteristic of the NTD is the jelly-roll motif that mediates RNA demethylase activity. The canonical HxD/E…H…R…R motif in NTD (H231–D233–H307–R316–R322) is employed to coordinate the metal ion and NOG. The nucleobase of the substrate is stacked on the conserved residues Tyr108 and His231 (Fig. 5c). The carbonyl oxygen of substrate is further stabilized through the hydrogen-bond interactions with polar residues (Arg96 and Glu234), defining the substrate specificity. The most prominent feature in FTO is the βI–βII loop that links the minor β-sheet and major β-sheet. The βI–βII loop (Fig. 5a), the unique insertion in FTO members compared with other AlkB proteins, covers the substrate binding pocket and enables FTO to discriminate between dsDNA and ssDNA (or ssRNA).
of FTO in complex with mThe CTD of FTO folds into a helix-bundle structure, and makes extensive contact with NTD, especially the nucleotide recognition lid (NRL). In vitro, neither the NTD or CTD domain alone showed the demethylase activity. The results were consistent with the FTO mutations that disrupt the NTD–CTD interactions. However, the activity was rescued by mixing CTD into NTD fractions, suggesting the key roles of CTD in FTO activity. It is likely that CTD contributes to the stability and orientation of the lid region, and thereby activates the demethylation activity.
12.1. ALKBH5
ALKBH5 is another RNA demethylase identified in vivo and in vitro (Zheng et al., 2013). ALKBH5, encoded on the chromosome 17p11, is predominantly localized in the nucleus and is highly expressed in testis (Both et al., 2012; Zheng et al., 2013). The hypoxia inducible factor-1 (HIF-1) and protein arginine methyltransferase 7 (PRMT7) have been shown to play a role in regulating the ALKBH5 gene expression (Feng et al., 2013; Thalhammer et al., 2011). Recent studies suggested that ALKBH5 influences nuclear export and RNA synthesis. In Alkbh5-deficient cells, the level of m6A in is increased. Although the Alkbh5-null mice are viable, the mouse lacking ALKBH5 displays a sterile phenotype, indicating that ALKBH5 is essential for fertility (Zheng et al., 2013).
Unlike FTO, ALKBH5 only catalyzes the demethylation of m6A, but not other RNA modifications in ssRNA or ssDNA. Most recently, four groups have independently solved the crystal structures of ALKBH5 (Xu et al., 2014; Feng et al., 2014; Chen et al., 2014; Aik et al., 2014). The overall structure of ALKBH5 displays a DSBH fold (Fig. 5b). The catalytic triad HxD/E…H in ALKBH5 (H204–D206–H266) coordinates the metal ion. The conserved residues arginine (Arg277) and aromatic residue (Tyr195) are hydrogen bonded to 2OG (Fig. 5d). Unlike other Alkb members, another conserved arginine (Arg283) appears not to form the direct interaction with 2OG. Structural alignment and modeling showed that the nucleobase might be bracketed between Tyr139 (or Tyr141) and His204, and basic residues (Arg130 and/or Lys132) further contribute to stabilize the nucleobase via hydrogen-bond interactions.
The βIV–βV loop has the function of discrimination against dsDNA. Additionally, the βIV–βV loop is anchored to the minor β-sheet via the conserved disulfide bond in ALKBH5 proteins, determining the conformation of the βIV–βV loop (Fig. 5b). In the absence of a disulfide bond, ALKBH5 showed the demethylation activity towards dsDNA, suggesting the crucial role of the disulfide bond in ssDNA (or ssRNA) preference (Feng et al., 2014). However, without a disulfide bond, the βIV–βV loop in zebrafish ALKBH5 undergoes a conformational change and would also clash with the dsDNA (Chen et al., 2014). Further experiments are needed to investigate the role of the disulfide bond.
13. Structural and functional comparison
Not surprisingly, both FTO and ALKBH5 share the DSBH fold and the conserved HxD/E…H…R…R motif. The intermediates are present in the demethylation process of FTO other than ALKBH5 (Fu et al., 2013). The substrates in FTO and ALKBH5, stacking between the aromatic and histidine rings, are involved in interactions with basic residues. Furthermore, an acidic residue from the βII–βIII loop (Glu235 in FTO) also contributes to the interaction with the nucleobase. The equivalent position to Glu235 in other Alkb proteins is a polar residue, with some exceptions, where it is a neutral residue (Pro207) in ALKBH5 (Figs. 5c and 5d). Therefore, it is tempting to speculate that the nucleobase recognition modes between ALKBH5 and FTO are distinct. These interactions with nucleobase could endow substrate preference on AlkB proteins. Interestingly, swapping the βII–βIII loop between the AlkB demethylases can lead to an exchange of demethylation activities (Zhu & Yi, 2014), indicating the key roles of the βII–βIII loop in substrate specificity.
The NRLs occur at the similar position in both FTO and ALKBH5. Despite the close similarities with the NRL2 (e.g. high flexibility), the NRL1 in both proteins is strikingly different. A β-strand and a long loop that is close to the substrate-binding pocket constitute the NRL1 in FTO. However, the NRL1 in ALKBH5 consists of two β-strands, exposing an open space over the substrate-binding pocket (Figs. 5a and 5b). The differences, taken together, are responsible for substrate preference.
Although they both employ a loop to discriminate against dsDNA, the positions of the loops are distinct. Moreover, the corresponding loop in FTO is an insertion in comparison with other AlkB members. The C-terminal region is important for FTO demethylaton activity; however, the ALKBH5 variant carrying the C-terminal region deletion had no significant effect on demethylaton activity. Of note, the
of the C-terminal region may be involved in the ALKBH5 function or stability.FTO and ALKBH5 displayed distinct inhibition profiles against the generic 2OG dioxygenase inhibitors. IOX3 is a more potent inhibitor of FTO than 2,4-PDCA. By contrast, 2,4-PDCA is the most potent ALKNH5 inhibitor (Aik et al., 2013; Chen et al., 2012; Xu et al., 2014; Feng et al., 2014; Aik et al., 2014). In the presence of a citrate buffer, the 2OG-binding pockets of both the ALKBH5 and FTO are occupied by the citrate molecule, which is a modest inhibitor of ALKBH5 and FTO. In FTO, the citrate molecule occurs at the analogous position of 2OG; however, in ALKBH5, the citrate molecule is found at a different position to that in 2OG. Moreover, the metal ion is absent in ALKBH5–citrate complex structure, but the metal ion is still present in FTO–citrate complex structure (Figs. 5e and 5f). These distinct binding modes may be helpful in developing selective inhibitors of AlkB proteins.
14. Concluding remarks
Dioxygenases consist of many subgroups, including FeII and 2OG-dependent dioxygenases and FAD-dependent amine oxidases. In eukaryotes, these two classes of dioxygenases play an important role in regulating gene expression by catalyzing the demethylation of histones, DNAs or RNAs. Much progress has been made to advance our knowledge about the biological functions of these dioxygenases and their implications in human diseases, which would contribute to better and faster identifying and validating these targets for therapeutics. Without any doubt, there is still much more to be investigated about this superfamily and its involvement in epigenetics in the future.
References
Aik, W., Demetriades, M., Hamdan, M. K., Bagg, E. A., Yeoh, K. K., Lejeune, C., Zhang, Z., McDonough, M. A. & Schofield, C. J. (2013). J. Med. Chem. 56, 3680–3688. CrossRef Google Scholar
Aik, W., McDonough, M. A., Thalhammer, A., Chowdhury, R. & Schofield, C. J. (2012). Curr. Opin. Struct. Biol. 22, 691–700. Web of Science CrossRef CAS PubMed Google Scholar
Aik, W.-S., Scotti, J. S., Choi, H., Gong, L., Demetriades, M., Schofield, C. J. & McDonough, M. A. (2014). Nucleic Acids Res. 42, 4741–4754. CrossRef Google Scholar
Allen, M. D., Grummitt, C. G., Hilcenko, C., Min, S. Y., Tonkin, L. M., Johnson, C. M., Freund, S. M., Bycroft, M. & Warren, A. J. (2006). EMBO J. 25, 4503–4512. CrossRef Google Scholar
Bird, A. (2002). Genes Dev. 16, 6–21. CrossRef Google Scholar
Boissel, S. et al. (2009). Am. J. Hum. Genet. 85, 106–111. CrossRef Google Scholar
Both, J., Wu, T., Bras, J., Schaap, G. R., Baas, F. & Hulsebos, T. J. (2012). PLoS One, 7, e30907. CrossRef Google Scholar
Cantara, W. A., Crain, P. F., Rozenski, J., McCloskey, J. A., Harris, K. A., Zhang, X., Vendeix, F. A., Fabris, D. & Agris, P. F. (2011). Nucleic Acids Res. 39, D195–D201. Web of Science CrossRef CAS PubMed Google Scholar
Chen, B., Ye, F., Yu, L., Jia, G., Huang, X., Zhang, X., Peng, S., Chen, K., Wang, M. & Gong, S. (2012). J. Am. Chem. Soc. 134, 17963–17971. CrossRef Google Scholar
Chen, W., Zhang, L., Zheng, G., Fu, Y., Ji, Q., Liu, F., Chen, H. & He, C. (2014). FEBS Lett. 588, 892–898. CrossRef Google Scholar
Chen, Z., Zang, J., Whetstine, J., Hong, X., Davrazou, F., Kutateladze, T. G., Simpson, M., Mao, Q., Pan, C. H., Dai, S., Hagman, J., Hansen, K., Shi, Y. & Zhang, G. (2006). Cell, 125, 691–702. CrossRef Google Scholar
Church, C., Lee, S., Bagg, E. A., McTaggart, J. S., Deacon, R., Gerken, T., Lee, A., Moir, L., Mecinović, J., Quwailid, M. M., Schofield, C. J., Ashcroft, F. M. & Cox, R. D. (2009). PLoS Genet. 5, e1000599. CrossRef Google Scholar
Church, C., Moir, L., McMurray, F., Girard, C., Banks, G. T., Teboul, L., Wells, S., Brüning, J. C., Nolan, P. M., Ashcroft, F. M. & Cox, R. D. (2010). Nat. Genet. 42, 1086–1092. CrossRef Google Scholar
Clifton, I. J., McDonough, M. A., Ehrismann, D., Kershaw, N. J., Granatino, N. & Schofield, C. J. (2006). J. Inorg. Biochem. 100, 644–669. Web of Science CrossRef PubMed CAS Google Scholar
Cloos, P. A., Christensen, J., Agger, K. & Helin, K. (2008). Genes Dev. 22, 1115–1140. CrossRef Google Scholar
Cortellino, S. et al. (2011). Cell, 146, 67–79. CrossRef Google Scholar
Couture, J. F., Collazo, E., Ortiz-Tello, P. A., Brunzelle, J. S. & Trievel, R. C. (2007). Nat. Struct. Mol. Biol. 14, 689–695. CrossRef Google Scholar
Dina, C. et al. (2007). Nat. Genet. 39, 724–726. CrossRef Google Scholar
Dominissini, D., Moshitch-Moshkovitz, S., Schwartz, S., Salmon-Divon, M., Ungar, L., Osenberg, S., Cesarkas, K., Jacob-Hirsch, J., Amariglio, N., Kupiec, M., Sorek, R. & Rechavi, G. (2012). Nature, 485, 201–206. CrossRef Google Scholar
Fang, R. et al. (2010). Mol. Cell, 39, 222–233. CrossRef Google Scholar
Fang, R. et al. (2013). Mol. Cell, 49, 558–570. CrossRef Google Scholar
Feng, C., Liu, Y., Wang, G., Deng, Z., Zhang, Q., Wu, W., Tong, Y., Cheng, C. & Chen, Z. (2014). J. Biol. Chem. 289, 11571–11583. CrossRef Google Scholar
Feng, Y., Maity, R., Whitelegge, J. P., Hadjikyriacou, A., Li, Z., Zurita-Lopez, C., Al-Hadid, Q., Clark, A. T., Bedford, M. T., Masson, J. Y. & Clarke, S. G. (2013). J. Biol. Chem. 288, 37010–37025. Web of Science CrossRef CAS PubMed Google Scholar
Fischer, J., Koch, L., Emmerling, C., Vierkotten, J., Peters, T., Brüning, J. C. & Rüther, U. (2009). Nature, 458, 894–898. CrossRef Google Scholar
Forneris, F., Binda, C., Adamo, A., Battaglioli, E. & Mattevi, A. (2007). J. Biol. Chem. 282, 20070–20074. CrossRef Google Scholar
Frayling, T. M. et al. (2007). Science, 316, 889–894. Web of Science CrossRef PubMed CAS Google Scholar
Fu, Y. & He, C. (2012). Curr. Opin. Chem. Biol. 16, 516–524. CrossRef Google Scholar
Fu, Y., Jia, G., Pang, X., Wang, R. N., Wang, X., Li, C. J., Smemo, S., Dai, Q., Bailey, K. A., Nobrega, M. A., Han, K. L., Cui, Q. & He, C. (2013). Nat. Commun. 4, 1798. CrossRef Google Scholar
Gavin, D. P., Chase, K. A. & Sharma, R. P. (2013). Neuropharmacology, 75, 233–245. CrossRef Google Scholar
Gerken, T. et al. (2007). Science, 318, 1469–1472. CrossRef Google Scholar
Grosjean, H. (2005). Fine-Tuning of RNA Functions by Modification and Editing. Topics in Current Genetics, pp. 1–22. Berlin, Heidelberg: Springer. Google Scholar
Gulati, P. & Yeo, G. S. (2013). Diabetologia, 56, 2113–2121. CrossRef Google Scholar
Guo, J. U., Su, Y., Zhong, C., Ming, G. L. & Song, H. (2011). Cell, 145, 423–434. CrossRef Google Scholar
Hahn, M. A., Qiu, R., Wu, X., Li, A. X., Zhang, H., Wang, J., Jui, J., Jin, S. G., Jiang, Y., Pfeifer, G. P. & Lu, Q. (2013). Cell Rep. 3, 291–300. CrossRef Google Scholar
Han, Z., Niu, T., Chang, J., Lei, X., Zhao, M., Wang, Q., Cheng, W., Wang, J., Feng, Y. & Chai, J. (2010). Nature, 464, 1205–1209. CrossRef Google Scholar
Hashimoto, H., Pais, J. E., Zhang, X., Saleh, L., Fu, Z. Q., Dai, N., Corrêa, I. R., Zheng, Y. & Cheng, X. (2014). Nature, 506, 391–395. CrossRef Google Scholar
Hausinger, R. P. (2004). Crit. Rev. Biochem. Mol. Biol. 39, 21–68. Web of Science CrossRef PubMed CAS Google Scholar
He, Y. F. et al. (2011). Science, 333, 1303–1307. CrossRef Google Scholar
Herranz, N., Dave, N., Millanes-Romero, A., Morey, L., Díaz, V. M., Lórenz-Fonfría, V., Gutierrez-Gallego, R., Jerónimo, C., Di Croce, L., García de Herreros, A. & Peiró, S. (2012). Mol. Cell, 46, 369–376. CrossRef Google Scholar
Højfeldt, J. W., Agger, K. & Helin, K. (2013). Nat. Rev. Drug Discov. 12, 917–930. Google Scholar
Horton, J. R., Upadhyay, A. K., Qi, H. H., Zhang, X., Shi, Y. & Cheng, X. (2010). Nat. Struct. Mol. Biol. 17, 38–43. CrossRef Google Scholar
Hou, H. & Yu, H. (2010). Curr. Opin. Struct. Biol. 20, 739–748. Web of Science CrossRef CAS PubMed Google Scholar
Hu, L., Li, Z., Cheng, J., Rao, Q., Gong, W., Liu, M., Shi, Y. G., Zhu, J., Wang, P. & Xu, Y. (2013). Cell, 155, 1545–1555. CrossRef Google Scholar
Inoue, A., Shen, L., Dai, Q., He, C. & Zhang, Y. (2011). Cell Res. 21, 1670–1676. CrossRef Google Scholar
Inoue, A. & Zhang, Y. (2011). Science, 334, 194. CrossRef Google Scholar
Ito, S., D'Alessio, A. C., Taranova, O. V., Hong, K., Sowers, L. C. & Zhang, Y. (2010). Nature, 466, 1129–1133. CrossRef Google Scholar
Jia, G., Fu, Y., Zhao, X., Dai, Q., Zheng, G., Yang, Y., Yi, C., Lindahl, T., Pan, T., Yang, Y. G. & He, C. (2011). Nat. Chem. Biol. 7, 885–887. CrossRef Google Scholar
Jia, G., Yang, C. G., Yang, S., Jian, X., Yi, C., Zhou, Z. & He, C. (2008). FEBS Lett. 582, 3313–3319. CrossRef Google Scholar
Kaklamani, V., Yi, N., Sadim, M., Siziopikou, K., Zhang, K., Xu, Y., Tofilon, S., Agarwal, S., Pasche, B. & Mantzoros, C. (2011). BMC Med. Genet. 12, 52. Google Scholar
Keller, L., Xu, W., Wang, H. X., Winblad, B., Fratiglioni, L. & Graff, C. (2011). J. Alzheimers Dis. 23, 461–469. Google Scholar
Klose, R. J., Kallin, E. M. & Zhang, Y. (2006). Nat. Rev. Genet. 7, 715–727. CrossRef Google Scholar
Lappalainen, T. et al. (2011). Nutr. Metab. Cardiovasc. Dis. 21, 691–698. CrossRef Google Scholar
Lin, H., Wang, Y., Wang, Y., Tian, F., Pu, P., Yu, Y., Mao, H., Yang, Y., Wang, P., Hu, L., Lin, Y., Liu, Y., Xu, Y. & Chen, C. D. (2010). Cell Res. 20, 899–907. CrossRef Google Scholar
Liu, J., Yue, Y., Han, D., Wang, X., Fu, Y., Zhang, L., Jia, G., Yu, M., Lu, Z., Deng, X. Dai, Q., Chen, W. & He, C. (2013). Nat. Chem. Biol. 10, 93–95. CrossRef Google Scholar
Machnicka, M. A., Milanowska, K., Osman Oglou, O., Purta, E., Kurkowska, M., Olchowik, A., Januszewski, W., Kalinowski, S., Dunin-Horkawicz, S., Rother, K. M., Helm, M., Bujnicki, J. M. & Grosjean, H. (2013). Nucleic Acids Res. 41, D262–D267. Web of Science CrossRef CAS PubMed Google Scholar
Martin, C. & Zhang, Y. (2005). Nat. Rev. Mol. Cell Biol. 6, 838–849. CrossRef Google Scholar
Meyer, K. D. & Jaffrey, S. R. (2014). Nature Rev. Mol Cell Biol. 15, 313–326. CrossRef Google Scholar
Meyer, K. D., Saletore, Y., Zumbo, P., Elemento, O., Mason, C. E. & Jaffrey, S. R. (2012). Cell, 149, 1635–1646. Web of Science CrossRef CAS PubMed Google Scholar
Min, J., Feng, Q., Li, Z., Zhang, Y. & Xu, R. M. (2003). Cell, 112, 711–723. CrossRef Google Scholar
Minton, K. (2014). Nat. Rev. Mol. Cell Biol. 15, 76–77. CrossRef Google Scholar
Mohr, F., Döhner, K., Buske, C. & Rawat, V. P. (2011). Exp. Hematol. 39, 272–281. CrossRef Google Scholar
Mosammaparast, N. & Shi, Y. (2010). Annu. Rev. Biochem. 79, 155–179. Web of Science CrossRef CAS PubMed Google Scholar
Ng, S. S. et al. (2007). Nature, 448, 87–91. CrossRef Google Scholar
Nottke, A., Colaiácovo, M. P. & Shi, Y. (2009). Development, 136, 879–889. CrossRef Google Scholar
Patsialou, A., Wilsker, D. & Moran, E. (2005). Nucleic Acids Res. 33, 66–80. CrossRef Google Scholar
Pfaffeneder, T. et al. (2014). Nat. Chem. Biol. 10, 574–581. CrossRef Google Scholar
Qi, H. H. et al. (2010). Nature, 466, 503–507. CrossRef Google Scholar
Schmidt, D. M. & McCafferty, D. G. (2007). Biochemistry, 46, 4408–4416. CrossRef Google Scholar
Sharp, P. A. (2009). Cell, 136, 577–580. CrossRef Google Scholar
Shi, Y. (2007). Nat. Rev. Genet. 8, 829–833. CrossRef Google Scholar
Shi, Y., Lan, F., Matson, C., Mulligan, P., Whetstine, J. R., Cole, P. A., Casero, R. A. & Shi, Y. (2004). Cell, 119, 941–953. Web of Science CrossRef PubMed CAS Google Scholar
Shi, Y. G. & Tsukada, Y. (2013). Cold Spring Harb. Perspect. Biol. 5, a022335. CrossRef Google Scholar
Shi, Y. & Whetstine, J. R. (2007). Mol. Cell, 25, 1–14. Web of Science CrossRef PubMed CAS Google Scholar
Smith, Z. D. & Meissner, A. (2013). Nat. Rev. Genet. 14, 204–220. CrossRef Google Scholar
Song, C. X. & He, C. (2012). Genome Biol. 13, 173. CrossRef Google Scholar
Suzuki, M. M. & Bird, A. (2008). Nat. Rev. Genet. 9, 465–476. CrossRef Google Scholar
Tahiliani, M., Koh, K. P., Shen, Y., Pastor, W. A., Bandukwala, H., Brudno, Y., Agarwal, S., Iyer, L. M., Liu, D. R., Aravind, L. & Rao, A. (2009). Science, 324, 930–935. Web of Science CrossRef PubMed CAS Google Scholar
Tarabra, E., Actis, G. C., Fadda, M., De Paolis, P., Comandone, A., Coda, R. & Rosina, F. (2012). Eur. J. Intern. Med. 23, 65–69. CrossRef Google Scholar
Thalhammer, A., Bencokova, Z., Poole, R., Loenarz, C., Adam, J., O'Flaherty, L., Schödel, J., Mole, D., Giaslakiotis, K., Schofield, C. J., Hammond, E. M., Ratcliffe, P. J. & Pollard, P. J. (2011). PLoS One, 6, e16210. Web of Science CrossRef PubMed Google Scholar
Thinnes, C. C., England, K. S., Kawamura, A., Chowdhury, R., Schofield, C. J. & Hopkinson, R. J. (2014). Biochim. Biophys. Acta, doi:10.1016/j.bbagrm.2014.05.009. Google Scholar
Tsukada, Y. (2012). J. Biochem. 151, 229–246. CrossRef Google Scholar
Tsukada, Y., Fang, J., Erdjument-Bromage, H., Warren, M. E., Borchers, C. H., Tempst, P. & Zhang, Y. (2006). Nature, 439, 811–816. CrossRef Google Scholar
Wang, Y., Li, Y., Toth, J. I., Petroski, M. D., Zhang, Z. & Zhao, J. C. (2014). Nat. Cell Biol. 16, 191–198. CrossRef Google Scholar
Wang, X., Lu, Z., Gomez, A., Hon, G. C., Yue, Y., Han, D., Fu, Y., Parisien, M., Dai, Q., Jia, G., Ren, B., Pan, T. & He, C. (2014). Nature, 505, 117–120. CrossRef Google Scholar
Wehr, E., Schweighofer, N., Möller, R., Giuliani, A., Pieber, T. R. & Obermayer-Pietsch, B. (2010). Metabolism, 59, 575–580. CrossRef Google Scholar
Wei, C. M., Gershowitz, A. & Moss, B. (1975). Cell, 4, 379–386. CrossRef Google Scholar
Wei, C. M. & Moss, B. (1977). Biochemistry, 16, 1672–1676. CrossRef Google Scholar
Wu, X., Johansen, J. V. & Helin, K. (2013). Mol. Cell, 49, 1134–1146. CrossRef Google Scholar
Wyatt, G. R. & Cohen, S. S. (1953). Biochem. J. 55, 774–782. Google Scholar
Xiao, M., Yang, H., Xu, W., Ma, S., Lin, H., Zhu, H., Liu, L., Liu, Y., Yang, C., Xu, Y., Zhao, S., Ye, D., Xiong, Y. & Guan, K. L. (2012). Genes Dev. 26, 1326–1338. CrossRef Google Scholar
Xu, C., Bian, C., Lam, R., Dong, A. & Min, J. (2011). Nat. Commun. 2, 227. CrossRef Google Scholar
Xu, C., Liu, K., Tempel, W., Demetriades, M., Aik, W., Schofield, C. J. & Min, J. (2014). J. Biol. Chem. 289, 17299–17311.. CrossRef Google Scholar
Xu, Y. et al. (2012). Cell, 151, 1200–1213. CrossRef Google Scholar
Yang, M., Culhane, J. C., Szewczuk, L. M., Jalili, P., Ball, H. L., Machius, M., Cole, P. A. & Yu, H. (2007). Biochemistry, 46, 8058–8065. CrossRef Google Scholar
Yang, M., Gocke, C. B., Luo, X., Borek, D., Tomchick, D. R., Machius, M., Otwinowski, Z. & Yu, H. (2006). Mol. Cell, 23, 377–387. CrossRef Google Scholar
Yang, Y., Hu, L. et al. (2010). Cell Res. 20, 886–898. CrossRef Google Scholar
Yang, Z., Jiang, J., Stewart, D. M., Qi, S., Yamane, K., Li, J., Zhang, Y. & Wong, J. (2010). Cell Res. 20, 276–287. CrossRef Google Scholar
Zhang, Q., Qi, S., Xu, M., Yu, L., Tao, Y., Deng, Z., Wu, W., Li, J., Chen, Z. & Wong, J. (2013). Cell Res. 23, 225–241. CrossRef Google Scholar
Zhang, Y. & Reinberg, D. (2001). Genes Dev. 15, 2343–2360. Web of Science CrossRef PubMed CAS Google Scholar
Zheng, G. et al. (2013). Mol. Cell, 49, 18–29. Web of Science CAS PubMed Google Scholar
Zhu, C. & Yi, C. (2014). Angew. Chem. Int. Ed. 53, 3659–3662. CrossRef Google Scholar
This is an open-access article distributed under the terms of the Creative Commons Attribution (CC-BY) Licence, which permits unrestricted use, distribution, and reproduction in any medium, provided the original authors and source are cited.