tuberculosis strain H37Rv (http://genolist.pasteur.fr/tuberculist) and M. bovis BCG Pasteur 1173P2 (http://genolist.pasteur.fr/BCGList/). Identified proteins showed a pI variation between 3-8 and a molecular mass (M r) range between 9 and 120 kDa. The comparison of experimentally determined and theoretical M r and pI values of the identified protein spots from BCG Moreau against the predicted values for M. tuberculosis strain H37Rv proteins, obtained from the search with Mascot
version 2.2, showed a positive correlation according to the Spearman coefficient (Figure 2A and 2B) Considering the fact that the proteins identified in this study were obtained from the culture filtrate, we analyzed the presence of possible signals that could direct these proteins to the extracellular fraction (Additional file 3, Table S2), using Signal P (for sec-dependent secretion; [30]), Selleck 17DMAG LipoP (find more lipoproteins; [31]), TatP (for secretion through the twin-arginine translocation system; [32]) and SecretomeP (for non-classical secretion of leaderless proteins; [33]). Of the 101 proteins,
67 (66%) have no extracellular prediction. However, when we compare our data to 2 previous reports on the culture filtrate proteome of M. tuberculosis H37Rv – the 2DE database at the Max Planck Institute (http://web.mpiib-berlin.mpg.de/) and a recent CB-5083 in vivo work by de Souza and collaborators [34] – 93 proteins (92%) have been previously reported in one or both studies, including 60 of the proteins with no extracellular prediction.
We also evaluated the number of potential transmembrane (TM) domains using TMHMM ([35]; Additional file 3, Table S2). Thirteen proteins were found to contain 1 predicted TM domain Farnesyltransferase which, although coinciding in all cases with the signal peptide region predicted by SignalP, does not exclude a possible membrane localization for some of these proteins [36]. For the 22 proteins with a predicted signal peptide, the theoretical pI and Mr were calculated for the full protein and for the mature protein, after removal of the signal peptide region predicted by SignalP (Additional file 4, Table S3). Figure 1 2DE proteomic profile of CFPs from M. bovis BCG Moreau. Proteins (500 ug) were applied to 17 cm IPG strips in the pH intervals of 3 – 6 (panels A and C) and 5 – 8 (panels B and D) and separated in the second dimension across 12% (panels A and B) and 15% (panels C and D) SDS-PAGE. The images were merged to obtain a composite map in the pH range 3 – 8 (panel E). Protein spots were visualized by colloidal CBB-G250 staining. Identified proteins are numbered in panel E and detailed in Additional file 2, Table S1. Molecular weight standards indicated in kDa. Figure 2 Correlation between experimentally determined and theoretical pI and M r and distribution of predicted cellular localization of the identified proteins. The experimental and theoretical pI (panel A) and M r (panel B) values for the identified protein were compared.