Sequence conservation analysis and in silico human leukocyte antigen-peptide binding predictions for the Mtb72F and M72 tuberculosis candidate vaccine antigens

Abstract

Background: Requisites for an efficacious tuberculosis (TB) vaccine are a minimal genomic diversity among infectious Mycobacterium tuberculosis strains for the selected antigen, and the capability to induce robust T-cell responses in the majority of human populations. A tool in the identification of putative T-cell epitopes is in silico prediction of major histocompatibility complex (MHC)-peptide binding. Candidate TB vaccine antigen Mtb72F and its successor M72 are recombinant fusion proteins derived from Mtb32A and Mtb39A (encoded by Rv0125 and Rv1196, respectively). Adjuvanted Mtb72F and M72 candidate vaccines were shown to induce CD4+ T-cell responses in European, US, African and Asian populations. Methods: Sequence conservation of Mtb32A, Mtb39A, Mtb72F and M72 among 46 strains (prevalent Mycobacterium strains causing human TB disease, and H37Ra) was assessed by multiple alignments using ClustalX. For Mtb32A, Mtb39A and Mtb72F, 15-mer human leukocyte antigen (HLA)-class II-binding peptides were predicted for 158 DRB1 alleles prevailing in populations with high TB burden, 6 DRB3/4/5, 8 DQ and 6 DP alleles, using NetMHCII-pan-3.0. Results for 3 DRB1 alleles were compared with previously published allele-matched in vitro binding data. Additional analyses were done for M72. Nonameric MHC class I-binding peptides in Mtb72F were predicted for three alleles representative of class I supertypes A02, A03 and B07, using seven prediction algorithms. Results: Sequence identity among strains was ≥98 % for each protein. Residue changes in Mtb39A comprised primarily single residue or nucleotide insertions and/or deletions in repeat regions, and were observed in 67 % of strains. For Mtb72F, 156 DRB1, 6 DRB3/4/5, 7 DQ and 5 DP alleles were predicted to contain at least one MHC class II-binding peptide, and class I-binding peptides were predicted for each HLA-A/B allele. Comparison of predicted MHC-II-binding peptides with experimental data indicated that the algorithm's sensitivity and specificity were variable among alleles. Conclusions: The sequences from which Mtb72F and M72 are derived are highly conserved among representative Mycobacterium strains. Predicted putative T-cell epitopes in M72 and/or Mtb72F covered a wide array of HLA alleles. In silico binding predictions for class I- and II-binding putative epitopes can be complemented with biochemical verification of HLA binding capacity, processing and immunogenicity of the predicted peptides.