DNA methyltransferases [MTases, methylases] recognize specific duplex DNA sequences and catalyze the transfer of a methyl group from ^-adenosylmethionine (AdoMet) to recipient adenine or cytosine bases within these sequences, with the release of S-adenosylhomocysteine (AdoHcy or SAH). DNA MTases are widespread in nature and serve several physiological roles (see Methylation, DNA). Prokaryotic DNA MTases function in mismatch repair and DNA replication and as part of restriction-modification systems. Eukaryotic DNA MTases are involved in gene regulation, development, and genome compartmentalization. MTases are classified according to the base they methylate (Fig. 1). The first class methylates the C5 position of cytosine to form 5-methylcytosine (5mC). The second class methylates exocyclic nitrogen atoms of adenine or cytosine to form N6-methyladenine (N6mA) or N4-methylcytosine (N4mC). The second class is further divided into different groups, a, b, or g on the basis of amino acid sequence differences and structural analysis (1, 2). Over 100 MTase genes have been cloned and sequenced. Analysis of several 5mC MTase amino acid sequences reveal a linear arrangement of 10 conserved motifs. One motif, a glycine-rich segment, is also found in other AdoMet-dependent MTases and is involved in cofactor binding (3). Another conserved segment, PC, is unique to the 5mC MTases and is a part of the active site . Analysis of the primary sequences of N6mA and N4mC MTases reveals two motifs, where the arrangement and distance between the motifs defines the subgroup classifications (2). The first motif is the glycine-rich segment analogous to those in 5mC MTases and the second a tetrapeptide sequence, (Asp/Asn/Ser)-Pro-Pro-(Tyr/Phe), which has been implicated as part of the active site.
Figure 1. Types of methylated bases in DNA.
Discussion here will be limited to prokaryotic type II methyltransferases. For type I, IIs, III, and eukaryotic methyltransferases see Restriction-Modification Systems or DNA methylation.
1. Crystal Structures and Sequence Comparison
The crystal structures of four prokaryotic type II MTases have been determined. Two 5mC MTases, M*HhaI (4) with AdoMet and M*HaeIII (5), without AdoMet, have been solved in association with their target duplex DNAs. One N4mC MTase, M* PvuII, and one N6mA MTase, M*TaqI, have both been solved in complexes with AdoMet (6, 7). Comparison of the crystal structures of the four enzymes, as well as the results of a proteolysis study using M*EcoRI (8), reveal that all these monomeric enzymes have a two-domain organization, having the cofactor-binding and catalytic site and DNA recognition region associated with different domains. The structures indicate similar patterns of folding and organization at the cofactor binding and active sites (5, 9). In fact, the active site and cofactor-binding regions of M* HhaI, M*PvuII, and M* TaqI are found to be very similar when the central b-sheets of their structures are superimposed on each other (3, 6). On the other hand, the DNA recognition domains of M* HhaI and M*HaeIII have very dissimilar structures, except for two regions of amino acids that provide a scaffold for residues contacting the DNA (5). The common catalytic domain structures led Malone, et al. (1) to do a structure-assisted amino acid sequence analysis of 42 N6mA and N4mC MTases. This analysis revealed nine variably conserved motifs analogous to those found in 5mC MTases (I-VIII and X.) (1). Motifs I-III and X were hypothesized to be responsible for forming a cofactor-binding pocket and interacting with AdoMet. Motifs IV, VI, and VIII (the conserved PC motif in 5mC MTases and (Asp/Asn/Ser)-Pro-Pro-(Tyr/Phe) in N6mA and N4mC MTases) were proposed to participate in catalysis. Motifs V and VII were postulated to provide a structural framework for the catalytic region. Motif IX may provide a structural framework for the DNA recognition region in 5mC MTases, but no analogous motif was found in the exocyclic amino-methylating MTases (1, 3). Furthermore, the N6mA MTases analyzed by Malone et al. fell into three groups, a, b, or g depending on the order of the motifs from the N to C terminus of the protein, supporting the earlier sequence analysis by Wilson (2). For Dam MTase, a representative N6mAa MTase, motifs involved in binding AdoMet are first, followed by sequences involved in DNA recognition, and finally by catalytic-site motifs. Group b, which contains most of the N4mC MTases (including M*PvuII), has an arrangement where catalytic-site motifs are first, followed by the DNA recognition region, and then the cofactor-binding motifs. Finally, group g, including M*TaqI, has an arrangement of cofactor-binding motifs, followed by the catalytic site motifs, and terminating with the DNA recognition region. In general, the solved structures support the proposals of Malone et al. (1).
2. DNA Recognition
MTases are monomeric enzymes that transfer one methyl group to a recipient base of one strand at the recognition sequence per binding event. Most MTases show no preference for hemimethylated or unmethylated DNA and will methylate either molecule at the same rate. MTases appear to search for their canonical recognition sites by facilitated diffusion, or sliding, along the DNA until the recognition sequence is located (10). Kinetic studies of MTases, including M.* HhaI (a C5 MTase) (11), M*EcoRI, and Escherichia coli Dam MTase (N6mA MTases) (12, 13), show that bound cofactor is necessary for specific sequence recognition. X-ray crystallography of M*HhaI (14) and footprinting analysis of M*EcoRV (15), M*Sss/, and M* HhaI (16) indicate a majority of the interactions with DNA occur with bases and phosphates in the major groove. Studies of M*EcoRI (17) and M*EcoRV (18) with base analogues indicate that some interactions may also occur in the minor groove. For some N6-methyladenine methylases, site-specific recognition may be facilitated by DNA bending. Whereas the crystal structures of M*HhaI and M*HaeIII show little bending of the DNA helix, gel-shift-mobility assays and scanning force microscopy techniques have shown that M*EcoRI (19) and M*EcoRV (20) bend DNA about 52° and 60°, respectively. A striking feature, which may be a recognition mechanism as well as a catalytic property of all MTases, was found in the X-ray crystal structures of the 5mC MTases M*HhaI and M*Hae III. The substrate cytosine base is extrahelical and flipped from the helix into a pocket in the enzyme, where it is subsequently methylated and released. This extrahelical base flipping is considered in greater detail in 5-methylcytosine. Others have used fluorescence-based assays of M* EcoRI (21) or modified base interference studies with M*EcoRV (22) to provide evidence for a base-flipping mechanism for N6mA MTases. The structures of both M* TaqI and M*PvuII, although lacking DNA, have a "pocket" into which the target base could be flipped (6, 7). The N4mC MTases have sequence similarities to the N6mA MTases. All the DNA MTases likely require AdoMet binding and utilize phosphate and base interactions in combination with base flipping and possibly DNA bending to accomplish sequence-specific recognition.
3. Catalytic Properties
The steady-state kinetic analyses performed on MTases are consistent in that methylation by MTases follows an ordered bi-bi steady-state scheme, where AdoMet binding is a prerequisite to canonical DNA sequence recognition and methylation (11, 12, 23): (1) the MTase binds to nonspecific DNA randomly, with or without AdoMet bound; and (2) in the presence of AdoMet, which confers recognition specificity, the target sequence is located and the recipient base of one strand is methylated. The product ternary complex is MTase-methylated DNA-AdoHcy. AdoHcy then dissociates, and depending on the processivity of the MTase (a measure of the ability to scan base pairs), the enzyme will dissociate from the DNA or move to the next site on the same DNA molecule (24). The overall reaction of DNA MTases is slow, with turnover rates ( kcat) generally less than 0.1s-1 (12). The average Km values for AdoMet and substrate canonical DNA are in the nanomolar range. MTases are extremely efficient catalysts, with kcat/KM values reaching diffusion-controlled limits (108-109 M-1 s 1). The rate-limiting step for MTase catalysis occurs after the methylation step, since k cat is slower than the rate of methyl group deposition, kmeth (11, 12, 23). The mechanism of the 5mC MTases has been elucidated and involves a covalent enzyme intermediate (see 5-Methylcytosine for a detailed description of the mechanism). The mechanism of N4mC and N6mA MTases likely involves the direct methylation without a covalent enzyme intermediate (6, 25). The structure of M*PvuII, an N4mC-forming MTase, revealed the structural similarity of its active site to those of M*HhaI and M* TaqI, which form 5mC and N6mA, respectively (6). The similar locations of amino acids in the active sites of these three MTases that form all three kinds of methylated base allowed the assignment of specific residues to a catalytic mechanism for M*PvuII. This mechanism, which probably does not involve a covalent enzyme-DNA complex, can likely be extended to enzymes like M*TaqI that form N6mA as well as to other N4mC-forming MTases.
4. Applications
Methyltransferases, or modification enzymes, can be used to modulate cleavage by restriction endonucleases (see Restriction Enzymes). Methylation of the DNA within or outside a specific restriction endonuclease recognition site prevents or inhibits cleavage by interfering with endonuclease binding or catalysis. For instance, MTases can be used decrease the number of possible restriction sites on a long DNA molecule. For example, R*BanII cleaves at the degenerate sequences (GAGCTC) or (GGGCCC). Premethylation of DNA with M*HaeIII, which is specific for (GG5mCC), reduces subsequent R*BanII cleavage to the single sequence, (GAGCTC), because the other sequence is rendered refractory. Also, particular restriction sites can be protected from methylation by masking the restriction site with the use of DNA-binding proteins or oligonucleotides (that form triplex DNA regions). These proteins or oligonucleotides can subsequently be removed to unmask the chosen restriction site. In another technique, methylase-limited partial digestion, larger fragments of DNA can be generated by a restriction endonuclease through prior partial methylation with the cognate MTase (26).
In summary, DNA MTases are ubiquitous and simple in composition and cofactor requirements. Thus, MTases are useful for the study of DNA-protein interactions, enzyme mechanism and kinetics, and protein-cofactor interactions. They are also useful tools for the molecular biologist.
