Inhibition study of Vibrio cholerae Endonuclease I

Abstract

The Vibrio cholerae bacteria resistance against introduction of new genetic material through transformation is caused mainly by a small extracellular or periplasmic endonuclease of type I, the VcEndA of 24.7 kDa coded by the dns gene, (Focareta et al, 1987; Focareta et al, 1991; Altermark et al, 2007b). The VcEndA homologues in other bacteria, Serratia marcescens (Timmins et al, 1973), Erwinia chrysanthemi (Moulard et al, 1993), Aeromonas hydrophila (Chang et al, 1992), Vibrio vulnificus (Wu et al, 2001) and Vibrio salmonicida (Altermark et al, 2007b), are identified as the main mechanisms of preventing a successful transformation in these organisms. Of a broader commercial interest is the identification of the EndoI in Esterichia coli, that shares 60 % sequence identity with VcEndA (Jekel et al, 1995). This project aims to find a lead compound for an inhibitor that is commercially exploitable, and will be applied as an additive in a transformation kit that prevents nuclease activity. An inhibitior would increase the yield in transformation procedures, and delete the step of creating endonuclease type I negative strains prior to transformation experiments. As a starting point, the Hepes molecule known to decrease the activity of VcEndA and the homologues VsEndA from Vibrio salmonicida (Altermark, Ph.D thesis 2006), was used as a template to find more active compounds. In this thesis I report the work and results from an in vitro screening of selected compounds with similar structural features as the Hepes molecule, and their activity measured by IC50 values. I also report an X-ray crystallography study with both soaked and co-crystallized approaches, and observed changes in the active site of the catalytic important residues Arg99 and Glu113 upon binding of an inactive compound. Computational modeling experiments with molecular docking, and comparison of the performance of three different docking programs, GOLD, AutoDock and Glide are carried out. To find more novel active compounds, a virtual screening by the program GOLD was performed with two libraries of small molecules. By the activity measurements, three compounds with the consensus feature of an aminoethanesulfonic acid group followed by a hetero- or homo cyclohexane ring were identified. In the structures from data sets collected from soaked crystals, the inactive molecule cacodylate was found bound in the active site. Observation of a change in the conformations of residue Arg99 and the nearby Glu113 is shown for two data sets compared to an empty site. The formation of a salt bridge between Arg99 and Glu113 shows similarity to the findings of Arg99 conformations in dsDNA-VVn complexes of the close homologous endonuclease type I, Vvn in Vibrio vulnificus (Wu et al, 2001; Li et al, 2003). The comparison of the molecular modeling programs GOLD, AutoDock and Glide indicate that GOLD is most suited to perform modeling experiments of the VcEndA system. This program is able to differentiate between active and inactive compounds upon assigning fitness scores, as well as consistently treat active compounds by assigning similar docking poses. The program AutoDock is also considered to give satisfactory docking poses, but are penalized for not consistently differentiate between active and inactive compounds when assigning fitness scores. The results from the docking experiments and the virtual screening strengthen the interpretation of the IC50 values, the consensus structural features and the changes in Arg99 upon binding of the inactive compound cacodylate.