Dinucleotide Codon Mutations Reveal Tuberculosis’ Hidden Evolution

Tuberculosis remains a moving target because the bacterium M. tuberculosis evolves under two overlapping pressures: drug treatment and adaptation to the human host. Danila Zimenkov and colleagues set out to look for a subtle but potentially important form of mutation that is often overlooked — dinucleotide substitutions that change two DNA bases within a single codon. Unlike the more common single nucleotide changes, these paired changes can produce a broader range of alterations in the amino acids that make up bacterial proteins, and so they may be a useful sign of diversifying selection. To test this idea the team reviewed 43 studies that together included 11,730 clinical isolates with resistance to rifampicin. In those resistance-determining regions of the rpoB gene they identified 11 different dinucleotide substitutions affecting 54 codons. Overall the prevalence of such dinucleotide substitutions approached 4%. Importantly, the researchers also found dinucleotide changes in resistance determinants for newer drugs — linezolid and bedaquiline — in the genes rplC and atpE, despite far fewer reported resistant clinical isolates for those drugs.

Beyond the focused review of resistance genes, the study expanded to a genome-wide search for dinucleotide mutations. Using a dataset of 9,941 genomes studied by the CRYpTIC Consortium, the researchers scanned the full genetic record for unusually frequent dinucleotide changes. This analysis turned up three genes that carried a significantly elevated number of dinucleotide substitutions; the team suggests these genes are likely tied to virulence and host adaptation. Two specific substitutions, cyp138 P114F and L191A, are thought to have arisen early in the evolutionary history of lineage 2 and are now under strong selection for reverse substitutions. In a third gene, two amino acid changes, rv2024c N508T and C514L, can also be produced by single nucleotide changes, so the authors propose these may be preferred due to codon usage frequency rather than being unique to dinucleotide events. Together, the targeted and genome-wide approaches show dinucleotide substitutions are present in both drug resistance and possible virulence-related genes.

The findings introduce a new lens for studying how M. tuberculosis adapts and resists treatment. By flagging dinucleotide codon mutations as a signature of diversifying selection, Danila Zimenkov and colleagues offer a way to detect evolutionary changes that single-base analyses can miss. That matters because these paired changes can produce amino acid swaps that alter protein function in ways relevant to drug resistance and to interaction with the human host. The presence of dinucleotide substitutions in resistance determinants for rifampicin, and even for newer drugs such as linezolid and bedaquiline, suggests the bacterial population is exploring a wider set of adaptive options than previously appreciated. Finding three genes with concentrated dinucleotide changes that likely relate to virulence points to new biological targets that may be worth studying for antivirulence drug development. Altogether, the work highlights the complexity of evolutionary dynamics in M. tuberculosis and underscores the value of looking beyond single-nucleotide mutations when tracking resistance and adaptation.