5-Methylcytosine (Molecular Biology)

Cytosine bases in DNA can be modified to 5-methylcytosine (5mC) by DNA methyltransferases (MTases, methylases). DNA methyltransferases are monomeric enzymes that catalyze the transfer and covalent attachment of a methyl group from the cofactor ^-adenosylmethionine (AdoMet) to DNA with the release of S-adenosylhomocysteine (AdoHcy). Thus, methylation occurs at cytosine or adenine bases and forms C5-methylcytosine (5mC), N-4-methylcytosine (N4mC) or N-6-methyladenine (N6mA) (see Methyltransferase, DNA for structures). DNA methylation is involved in biological phenomena ranging from prokaryotic DNA replication and host DNA protection (as part of restriction-modification systems ) to eukaryotic gene regulation and embryonic development (see Methylation, DNA). The X-ray crystallography structures of two prokaryotic 5mC MTases, M*HhaI (1) and MHaeIII (2), have increased our understanding of the structural and mechanistic basis of 5mC methylation. The structure of M’Hha I (GmCGC), where mC indicates the methylated cytosine in the GCGC DNA recognition sequence, was determined in a complex with AdoMet and in a ternary complex with the cofactor and a 13-bp DNA duplex containing a suicide substrate, base analogue, 5-fluorocytosine (5FC). MHaeIII (GGmCC) was solved as a binary complex with an 18-bp DNA duplex, also containing 5FC. These complexes are consistent with a catalytic mechanism for 5mC MTases that involves a covalent DNA-protein intermediate. In a mechanism first described by Wu and Santi (3) and later modified by Erlanson, et al. (4), methylation of a cytosine base occurs in three steps (Fig. 1). The second step is prevented when 5FC is the substrate; the electronegative fluorine in place of the hydrogen at C5 cannot be eliminated, and the enzyme is locked in a deadend, covalent complex with the DNA (5, 6). In contrast, less direct evidence for the mechanism for the amino-methylating, N4mC-forming and N6mA-forming MTases exists. Interestingly, structural comparison of the enzymes M’TaqI (a N6mA MTase), and M*PvuII (a N4mC MTase), both crystallized without DNA, with M*HhaI, revealed a basic similarity in active-site architecture that allowed the identification and assignment of catalytic residues involved in a proposed catalytic mechanism for the exocylic amino-methylating MTases (7). The proposed mechanism involves a direct transfer of the methyl group from AdoMet to the amino group, rather than through a covalent intermediate. The crystal structures of M*HhaI and MHaeIII have provided snapshots of the mechanism for both the 5mC and the exocyclic amino group methyltransferases.

Figure 1. Mechanism of 5mC methylation. (a) Nucleophilic attack on C6 carbon by thiol group of cysteine residue of MTase; ( b) abstraction of methyl group from AdoMet and removal of hydrogen at C5; (c) regeneration of thiol group; ( d) 5-methylcytosine ( indicates bond attaching cytosine to the DNA, and the long straight lines indicate the active site of the protein).

Mechanism of 5mC methylation. (a) Nucleophilic attack on C6 carbon by thiol group of cysteine residue of MTase; ( b) abstraction of methyl group from AdoMet and removal of hydrogen at C5; (c) regeneration of thiol group; ( d) 5-methylcytosine ( indicates bond attaching cytosine to the DNA, and the long straight lines indicate the active site of the protein).

Both M*HhaI and M* HaeIII are two-domain proteins, with the larger domain containing catalytic and cofactor binding sites and the smaller domain involved in DNA recognition. The AdoMet binding regions of M*HhaI and M*HaeIII contain Rossman folds for nucleotide binding analogous to the structures of other AdoMet-binding proteins, such as M*TaqI and catechol O-methyltransferase (2, 6), and a comparison of Ca positions of the M* HhaI and M*HaeIII DNA MTases shows structural similarity throughout the large domain and bridging region. In contrast, the smaller DNA-binding domains have little structural similarity, except for a small conserved region that serves as a scaffold for the amino acids that interact with the DNA (2). The most remarkable feature in both crystal structures containing DNA is the extrahelical cytosine base, which is flipped out of the DNA duplex and positioned in an active-site pocket in the MTases. For M*Hha I, the cytosine is located near AdoHcy in the AdoMet binding pocket. How does the DNA duplex compensate structurally for the cavity created by the flipped base? M*HhaI shows significant distortion of the phosphate backbone of the strand in which the extrahelical cytosine is located, along with interdigitation of two amino acid residues from the recognition domain into the cavity opposite the orphaned guanine base. These residues hydrogen-bond to the unpaired guanine residue (6). With M*HaeIII, on the other hand, flipping out of the base is accompanied by a reorganization of base pairs such that the lone guanine residue hydrogen bonds with a neighboring (3′) cytosine base on the opposite strand. This produces another unpaired 3′ guanine that hydrogen-bonds to an arginine residue, but leaves a large unfilled pocket in the DNA duplex (2). The overall structural similarity of C5-cytosine methyltransferases suggests that base flipping may be a common mechanism for this family of enzymes.

The base-flipping mechanism was first found with M* HhaI, but several lines of evidence suggest that the extrusion of a base from duplex DNA (or RNA) may be a prevalent occurrence among enzymes that must perform chemistry on bases. Studies using X-ray crystallography on short, double-stranded DNA containing unpaired bases have shown that the unpaired bases can have both intra and extrahelical positions stabilized by crystal packing (2, 8). NMR studies of imino proton exchange in short, double-stranded DNA have shown the standard free-energy change for the base-pair opening to be 7-9 kcal/mol, comparable to protein-induced DNA distortion energies provided by the contacts made at the DNA-protein interface (8). Several DNA-binding enzymes crystallized without DNA, such as T4 b-glycosyltransferase, M* TaqI, and M*PvuII have been proposed to contain active-site pockets that would accommodate extrahelical bases (9, 10). The crystal structure of a DNA repair enzyme involved in pyrimidine dimer excision, T4 endonuclease V, complexed with duplex DNA containing a thymine dimer, has been found to have an extrahelical adenine base. The expelled adenine base, which is complementary to one of the bases in the thymine dimer, is trapped in a pocket on the surface of the enzyme (11). Although the expelled base in this structure is opposite the strand in which the thymine dimer resides, its exit from the helix renders the damaged bases accessible to the repair enzyme. The detailed mechanism of how the cytosine or adenine bases are extruded remains speculative.

Determination of the mechanism of 5mC methylation and of the crystal structures of M*HhaI and M* HaeIII provides insight into the structure-function relationships of these proteins. These findings also raise intriguing questions about the DNA dynamics resulting from DNA-enzyme interactions. Aberrant methylation patterns in mammals promote tumorigenesis and lead to developmental abnormalities (12, 13). Since eukaryotic C5-methylcytosine methyltransferases have amino acid sequence similarities to their prokaryotic relatives, the results from studies of the bacterial enzymes will likely contribute to understanding their eukaryotic counterparts.

Next post:

Previous post: