A major problem in assessing the vaccine and diagnostic potential of various proteins encoded by Mycobacterium tuberculosis genome is the inability to produce large quantities of these proteins, even when Escherichia coli or other heterologous systems are employed for recombinant protein production. To overcome these barriers, we have constructed a modified expression vector, using pGEX-4T-1 vector as the backbone. In addition to the features offered by the pGEX-4T vectors, the new vector allowed easy purification of recombinant proteins on the highly versatile Ni-NTA-agarose affinity matrix. The utility of the new vector was demonstrated by expressing and purifying, to near homogeneity, two M. tuberculosis proteins, i.e., Rv3872 (a member of the multigene PE subfamily) and Rv3873 (a member of the multi-gene PPE subfamily), which are encoded by the RD1 region of M. tuberculosis. The proteins encoded by rv3872 and rv3873 were expressed at high levels as fusion proteins with glutathione-S-transferase in E. coli. The recombinant Rv3872 and Rv3873 proteins were purified and isolated free of the fusion partner (GST) by affinity purification on glutathione-Sepharose and/or Ni-NTA-agarose affinity matrix and cleavage of the purified fusion proteins by thrombin protease. The recombinant Rv3872 protein was nearly homogeneous (more than 95% pure) while Rv3873 preparation was more than 90% pure. The recombinant Rv3872 and Rv3873 proteins were immunologically active and reacted with antibodies in sera from TB patients. Our results demonstrate the utility of the newly constructed expression vector with two affinity tags for efficient expression and purification of recombinant M. tuberculosis proteins expressed in E. coli, which could be used for further diagnostic and immunological studies. (C) 2003 Elsevier Science (USA). All rights reserved.