Analysis of ligation and DNA binding by Escherichia coli DNA ligase (LigA) Adam Wilkinson a , Andrew Smith b , Desmond Bullard d , Manuel Lavesa-Curto d , Heather Sayer d , Alexandra Bonner d , Andrew Hemmings c , Richard Bowater d, T a Phico Therapeutics Ltd, Babraham Hall, Babraham, Cambridge, CB2 4AT, UK b Department of Biological Chemistry, John Innes Centre, Norwich NR4 7UH, UK c Schools of Biological Sciences and Chemical Sciences and Pharmacy, University of East Anglia, Norwich NR4 7TJ, UK d School of Biological Sciences, University of East Anglia, Norwich NR4 7TJ, UK Received 8 December 2004; received in revised form 3 March 2005; accepted 3 March 2005 Available online 19 March 2005 Abstract NAD + -dependent DNA ligases are essential enzymes in bacteria, with the most widely studied of this class of enzymes being LigA from Escherichia coli . NAD + -dependent DNA ligases comprise several discrete structural domains, including a BRCT domain at the C-terminus that is highly-conserved in this group of proteins. The over-expression and purification of various fragments of E. coli LigA allowed the investigation of the different domains in DNA-binding and ligation by this enzyme. Compared to the full-length protein, the deletion of the BRCT domain from LigA reduced in vitro ligation activity by 3-fold and also reduced DNA binding. Using an E. coli strain harbouring a temperature-sensitive mutation of ligA, the over-expression of protein with its BRCT domain deleted enabled growth at the non-permissive temperature. In gel-mobility shift experiments, the isolated BRCT domain bound DNA in a stable manner and to a wider range of DNA molecules compared to full LigA. Thus, the BRCT domain of E. coli LigA can bind DNA, but it is not essential for DNA nick-joining activity in vitro or in vivo. D 2005 Elsevier B.V. All rights reserved. Keywords: BRCT domain; DNA ligase; Ligation; NAD + ; Protein–DNA binding 1. Introduction DNA ligases join gaps in the phosphodiester backbone of DNA and are essential enzymes in all cells [1–3]. One of the first DNA ligases to be purified and analysed biochemically was that from Escherichia coli , which is encoded by ligA and consists of 671 amino acids (molecular weight of 74 kDa) [1]. During ligation by LigA, a covalent enzyme– adenylate intermediate is formed, with the adenylate group (AMP) being obtained from NAD + . Thus, these bacterial enzymes are termed NAD + -dependent; note that we refer to them as NAD + -ligases for convenience. As with all DNA ligases, the AMP is linked to a conserved lysine in the catalytic motif of the enzyme [1–4]. Currently, functional NAD + -ligases have not been detected in humans, leading to speculation that they could be useful targets for broad- spectrum antibiotics [5–12]. Open reading frames predicted to encode NAD + -ligases are present in the genome of every bacterial species that has been sequenced so far [4,13]. These predicted proteins are of fairly uniform size and have extensive amino acid sequence homology: a basic BLAST alignment to E. coli ligA detects typical amino acid sequence identity of 35– 50% across all sequences. High-resolution structures have been obtained for the N-terminal adenylation domain of NAD + -ligase from Bacillus stearothermophilus [5] and Enterococcus faecalis [14] and for the full-length NAD + - ligase from Thermus filiformis [6]. The good agreement 1570-9639/$ - see front matter D 2005 Elsevier B.V. All rights reserved. doi:10.1016/j.bbapap.2005.03.003 T Corresponding author. Tel.: +44 1603 592186; fax: +44 1603 592250. E-mail address: r.bowater@uea.ac.uk (R. Bowater). Biochimica et Biophysica Acta 1749 (2005) 113 – 122 http://www.elsevier.com/locate/bba