Methods

1) Sequences, sequence alignment and template detection. Sequences of Bacterial Nucleoside Monophosphate Kinases were extracted from Expasy web site (www.expasy.ch/) and were subjected to position specific PSI-Blast search. Each sequence was used a s a query in 5 iterations PSI-Blast search against non redundant database of sequences (ncbi.nih.gov) to generate a checkpoint file (position specific scoring matrix ? PSSM). Then the checkpoint file was used in sequential search against non redundant database of sequences corresponding to existing PDB files. The template with the highest similarity (smallest e-value) and smallest number of gaps was selected form the hits. In several cases, we couldn?t identify appropriate template and thus we created an artificial templates by merging two PDB files. Then, the sequence alignment was obtained with the H-map software (trantor.bioc.columbia.edu).

2) Models building. The 3D models of all Bacterial Nuleoside Monophosphate Kinase proteins were built using the sequence alignment and template detection described above. The models were built with NEST program (trantor.bioc.columbia.edu).

3) pKa calculations. The 3D structures generated in the previous section were inputted to Multi Conformation Continuum Electrostatics (MCCE) method (sci.ccny.cuny.edu/mcce) to calculate the pKa?s of ionizable groups. The MCCE output generates a file with predicted pKa?s as well as a file with the desolvation and polar energies. In addition, the package outputs a file that reports the net charge of both folded and unfolded states as a function of pH. Default parameters were used in the calculations as internal dielectric constant equal to 4 and external equal to 80. The ionic strength was in physiological range of 0.15M.