Mycobacterium tuberculosis Complex Exhibits Lineage-Specific Variations Affecting Protein Ductility and Epitope Recognition

Abstract

The advent of whole-genome sequencing has provided an unprecedented detail about the evolution and genetic significance of species-specific variations across the whole Mycobacterium tuberculosis Complex. However, little attention has been focused on understanding the functional roles of these variations in the protein coding sequences. In this work, we compare the coding sequences from 74 sequenced mycobacterial species including M. africanum, M. bovis, M. canettii, M. caprae, M. orygis, and M. tuberculosis. Results show that albeit protein variations affect all functional classes, those proteins involved in lipid and intermediary metabolism and respiration have accumulated mutations during evolution. To understand the impact of these mutations on protein functionality, we explored their implications on protein ductility/disorder, a yet unexplored feature of mycobacterial proteomes. In agreement with previous studies, we found that a Gly71Ile substitution in the PhoPR virulence system severely affects the ductility of its nearby region in M. africanum and animal-adapted species. In the same line of evidence, the SmtB transcriptional regulator shows amino acid variations specific to the Beijing lineage, which affects the flexibility of the N-terminal trans-activation domain. Furthermore, despite the fact that MTBC epitopes are evolutionary hyperconserved, we identify strain- and lineage-specific amino acid mutations affecting previously known T-cell epitopes such as EsxH and FbpA (Ag85A). Interestingly, in silico studies reveal that these variations result in differential interaction of epitopes with the main HLA haplogroups.

Keywords: Mycobacterium; coding sequences; epitope polymorphisms; epitope-HLA binding; lineages; protein ductility.

Fig. 1.—

Schematic phylogenetic relationships of the MTBC showing the number of genome sequences analyzed in this work. The figure also shows the phylogenetic distribution of M. canettii, the L1–L7 lineages and the animal-adapted species, as well as the geographic distribution of each lineage and the preferred host.

Fig. 2.—

(A) Overall distribution of proteins encoded by the core-genome of M. tuberculosis H37Rv (right panel) and the average mutated proteins found in 74 species of the MTBC analyzed in this work (left panel). Both diagrams are divided according to functional classes. (B) Distribution according to functional classes of mutated proteins with experimental evidence annotated in BioCyc of M. canettii (dark blue), M. africanum (red), M. bovis AF2122 (green), M. tuberculosis Beijiing (violet), M. orygis (ligth blue), and M. caprae (orange). Cluster 0, virulence, detoxification and adaptation; cluster 1, lipid metabolism; cluster 2, information pathways; cluster 3, cell wall and cell processes; cluster 5, insertion sequences and phages; cluster 6, proline–glutamate (PE) and proline–proline–glutamate (PPE) proteins; cluster 7, intermediary metabolism and respiration; cluster 8, unknown proteins; cluster 9, regulatory proteins; cluster 10, conserved hypothetical proteins; cluster 16, conserved hypothetical proteins with an orthologue in M. bovis. The percentages are normalized taking into account the total proteins associated to each functional group (A) and the sequence identity averages (B).

Fig. 3.—

(A) Schematic representation of the topology of PhoR. Note the presence of two transmembrane helices spanning the periplasmic loop which contains the Gly71Ile mutation present in M. africanum L6 and animal-adapted species. (B) Disorder and molecular recognition motif predictions in the 61–121 amino acid segments of the periplasmic sensor loop of PhoR from M. tuberculosis and M. bovis/M.africanum L6. Y-axis values ≥ 0.5 means high probability of disorder. Note that the Gly71Ile mutation severely affects the ductility of the sensor domain. Predictions were carried out with IUPred and ANCHOR (left panel, red and blue lines, respectively), MoRFPred (middle panel, red line), and RONN (right panel, red line) (for details see “Materials and Methods” section).

Fig. 4.—

(A) Average distribution of disordered residues in the proteins encoded by the core-genome of M. tuberculosis, essential proteins, nonessential proteins, antigens, and epitopes. (B) Percentage of intrinsically disordered/ductile proteins (IDPs) encoded by the core-genome of M. tuberculosis versus the length of disordered segments. (C) Frequency of amino acids within the disordered/ductile segments (L > 30).

Fig. 5.—

Structural models of SmtB transcription factor of M. tuberculosis H37Rv (magenta) and M. tuberculosis Beijing (light blue) using the NMR structure of NmtR transcription factor of M. tuberculosis (pdb 2lkp) as template. The residues of the non-conserved motif Phe30, Ala31, Glu32, Cys33, Thr35, Phe36, Pro37 in M. tuberculosis H37Rv and Phe30, Ser31, Thr32, Ala33, Gly35, Gly36, Pro37 in M. tuberculosis Beijing are shown in sticks. The three-dimensional cartoons were drawn using PyMol 1.4.1 (Schrodinger LLC).

Fig. 6.—

(A) dN/dS in MTBC species calculated for nonessential genes, essential genes and antigens. Note that antigens present dN/dS ratios comparable to essential and nonessential genes, indicative of antigen hyperconservation in the whole MTBC. (B) Localization in the M. tuberculosis H37Rv chromosome of T-cell antigens showing one (black dots) or more than one (red dots) polymorphism in the MTBC species analyzed in this work. The inner circle indicates the protein name and gene number according to H37Rv nomenclature. The number of epitopes contained in every antigen is provided in parentheses. The outer circumference contains information about those MTBC species carrying epitope mutations. For detailed information about every polymorphism see supplementary table S3, Supplementary Material online.

Fig. 7.—

(A) Native and (B) mutated structures of the EsxG/EsxH (Rv0287/Rv0288) antigens (pdb 2kg7; Ilghari et al. 2011). The EsxG (yellow) and EsxH (blue) proteins are shown in ribbon. The amino acids Arg 57 of EsxG and Ala71/Ser71 and Met72 of EsxH are shown as spheres. The three-dimensional cartoons were drawn using PyMol 1.4.1 (Schrodinger LLC).

Fig. 8.—

In silico prediction of the interaction of epitopes with the main HLA class II haplotypes. Each graph represents the population coverage of HLA class II alleles predicted to bind specific epitope variants (Y-axis) across 15 geographic regions (X-axis). Population coverages for the M. tuberculosis H37Rv epitope are indicated by orange lines and population coverages for the mutant variants in other MTBC species are shown by blue lines.