The 16-kDa protein of Mycobacterium tuberculosis represents an important antigenic target during bacillary latency and, consequently, should be considered as candidate subunit vaccine component. In this study, we have used CD4 T cell clones that recognize the peptide p91-110, an immunodominant and genetically permissive epitope, in the context of five different HLA-DR molecules and truncated and substituted variants of this peptide, to identify the minimal binding sequence (HLA-DR-binding core) and the minimal stimulatory sequence (TCR-binding core), as well as the residues that contact HLA-DR molecules and the TCR. We have found a common 9-mer sequence, spanning amino acids 93-101, as the binding core for HLA-DR1, -DR11, -DR13 and -DR7, but a longer (13-mer) sequence spanning amino acids 92-104 was required for binding to the HLA-DR15 molecules. F(93) was required for binding to all the tested HLA-DR molecules, hence allowing us to identify it as the N-terminal primary anchor residue (P1). Additionally, the binding requirements for other residues varied considerably between the tested alleles: A(94) for HLA-DR15, V(99) for HLA-DR1, -DR15, -DR11 and -DR7, R(100) for HLA-DR11 and -DR13, and L(104) for HLA-DR15. Concerning the residues of p91-110 peptide required for binding to the TCR, the pepscan analysis results would support the contention that P(-1) E(92), P6 F(98) would be important TCR contact sites because their substitutions led to full loss of T cell activation. Moreover, P8 R(100) is found to be critical residue in binding to HLA-DR11- and -DR13-restricted T cell clones, without influencing binding to the relevant HLA-DR molecule. Our results could be useful to design peptides with altered HLA anchor residues or TCR interaction sites to achieve remarkable increase in activity and to study their vaccine potential.