THE THREE DIMENSIONAL STRUCTURE OF AN ATP DEPENDENT DNA LIGASE FROM BACTERIOPHAGE T7. Hosahalli S. Subramanya, Aidan J. Doherty, Stephen R. Ashford and Dale B. Wigley, Laboratory of Molecular Biophysics, University of Oxford, South Parks Road, Oxford, OX1 3QU, UK

DNA ligase is a vital enzyme which is required for important cellular processes such as DNA replication, repair of damaged DNA and recombination. The enzyme mediates the formation of phosphodiester bonds between adjacent 3'-OH and 5'-phosphate termini, thereby joining the nicks in the double stranded DNA. Ligases can be classified into two groups depending on their requirement for ATP or NAD⁺ as the cofactor. All eukaryotic enzymes and virally encoded enzymes are ATP-dependent, whereas prokaryotic enzymes require NAD⁺ for their activity. DNA ligase from bacteriophage T7 is a monomer with a molecular weight of 41 kDa. Here we report the structure of this enzyme at 2.6 Å resolution.

The protein was crystallized by vapour diffusion method using hanging drops. Crystals belonged to the space group P2₁2₁2 with unit cell dimensions a=65.8 Å, b=86.3 Å, c=78.3 Å. The structure was solved by MIR using mercury and selenomethionine derivatives. Non-isomorphism between the crystals was a major problem in structure determination. Crystals were found to be grouped around three major forms. The non-isomorphism between the crystal forms was sufficient to allow density averaging between them to improve the electron density maps.

The structure consists of two distinct domains, a larger N-terminal domain (residues 2-240) and a C-terminal domain ( residues 241-349). The N-terminal domain is an [[alpha]]/[[beta]] structure and comprises of three mainly antiparallel [[beta]]-sheets surrounded by six [[alpha]]-helices. The ATP-binding site is situated in this domain in a pocket beneath one of the [[beta]]-sheets. The C-terminal domain consists of highly twisted antiparallel [[beta]]-sheet and a single [[alpha]]-helix running along one edge of the sheet. The structure of this domain is remarkably similar to the oligonucleotide binding fold, observed in a number of proteins including staphylococcal nuclease, bacterial cold shock protein and gene V single-strand DNA-binding protein. The DNA-binding site is proposed to be in a groove running between the two domains.