In silico discovery and virtual screening of multi-target inhibitors for proteins in Mycobacterium tuberculosis.

Mycobacterium tuberculosis (MTB) is the principal pathogen which causes tuberculosis (TB), a disease that remains as one of the most alarming health problems worldwide. An active area for the search of new anti-TB therapies is concerned with the use of computational approaches based on Chemoinformatics and/or Bioinformatics toward the discovery of new and potent anti-TB agents. These approaches consider only small series of structurally related compounds and the studies are generally realized for only one target like a protein. This fact constitutes an important limitation. The present work is an effort to overcome this problem. We introduce here the first chemo-bioinformatic approach by developing a multi-target (mt) QSAR discriminant model, for the in silico design and virtual screening of anti-TB agents against six proteins in MTB. The mt-QSAR model was developed by employing a large and heterogeneous database of compounds and substructural descriptors. The model correctly classified more than 90% of active and inactive compounds in both, training and prediction series. Some fragments were extracted from the molecules and their contributions to anti-TB activity through inhibition of the six proteins, were calculated. Several fragments were identified as responsible for anti-TB activity and new molecular entities were designed from those fragments with positive contributions, being suggested as possible anti-TB agents.