Pre-existing isoniazid resistance, but not the genotype of Mycobacterium tuberculosis drives rifampicin resistance codon preference in vitro

Both the probability of a mutation occurring and the ability of the mutant to persist will influence the distribution of mutants that arise in a population. We studied the interaction of these factors for the in vitro selection of rifampicin (RIF)-resistant mutants of Mycobacterium tuberculosis. We characterised two series of spontaneous RIF-resistant in vitro mutants from isoniazid (INH)-sensitive and -resistant laboratory strains and clinical isolates, representing various M. tuberculosis genotypes. The first series were selected from multiple parallel 1 ml cultures and the second from single 10 ml cultures. RIF-resistant mutants were screened by Multiplex Ligation-dependent Probe Amplification (MLPA) or by sequencing the rpoB gene. For all strains the mutation rate for RIF resistance was determined with a fluctuation assay. The most striking observation was a shift towards rpoB-S531L (TCG→TTG) mutations in a panel of laboratory-generated INH-resistant mutants selected from the 10-ml cultures (p<0.001). All tested strains showed similar mutation rates (1.33×10^?8 to 2.49×10^?7) except one of the laboratory-generated INH mutants with a mutation rate measured at 5.71×10^?7, more than 10 times higher than that of the INH susceptible parental strain (5.46–7.44×10^?8). No significant, systematic difference in the spectrum of rpoB-mutations between strains of different genotypes was observed. The dramatic shift towards rpoB-S531L in our INH-resistant laboratory mutants suggests that the relative fitness of resistant mutants can dramatically impact the distribution of (subsequent) mutations that accumulate in a M. tuberculosis population, at least in vitro. We conclude that, against specific genetic backgrounds, certain resistance mutations are particularly likely to spread. Molecular screening for these (combinations of) mutations in clinical isolates could rapidly identify these particular pathogenic strains. We therefore recommend that isolates are screened for the distribution of resistance mutations, especially in regions that are highly endemic for (multi)drug resistant tuberculosis.