Use of mutual information theory to investigate the folding of a protein. The thioredoxin family of proteins as a model system.

CRISTINA MARINO BUSLJE; JAVIER SANTOS; MORTEN NIELSEN; JOSE MARIA DELFINO

Montevideo Uruguay

Congreso; 6th International Conference on Biological Physics and the 5th Southern Cone Biophysics Congress; 2007

Sociedad Argentina de Biofísica

Binding of the C-terminal helix to the rest of the protein in E. coli thioredoxin (trx) may signify a late event in the consolidation of the global structure of the protein, as was recently demonstrated by Santos et al. (2007). This recognition event will be investigated across the trx family with mutual information theory (MI). We have found that an identical helix is present in two other protein families endowed with different folds and functions. By applying this computational technique we intend to detect “folding blocks”, i.e. those residues in the opposing surface that interact with the helix. Furthermore, any pair of residues with mutual information will be investigated across the whole trx family (~3200 sequences, according to Pfam-). As we found shortcomings in the software available (e.g. gap treatment, background of the results, “bias” for sequence number, etc) we developed a program at home by following the specifications that we consider of interest to carry out this experiment and to render it more universally applicable. In due course we will run independent sets of proteins to measure the performance of the software and -in so doing- test its generality. By filtering residues by their exposed surface, we will evaluate if we are able to detect mutual information as a result of: (i) a folding requirement, (ii) a direct interaction (if the residues are at a distance of less than 10 Å), or (iii) an interaction through an intervening structure (if the distance is more than 10 Å).