137 Genome-wide analysis in E. coli unravels homoplasy associated with cefotaxime resistance The outcome of the Pyseer analysis provided a similar number of CTX-resistanceassociated mutations and also confirmed the strong association of the −42 C>T mutation. Nonetheless, when zooming in on the identified positions by the homoplasy-based method, not all 24 mutations were significantly associated with CTX by Pyseer. An explanation for this discrepancy is that the combination of snippy and Pyseer identifies various mutation variants on the same overlapping genomic region by stratifying it as complex, indels and/or SNP. This greatly affects the power of the test on certain positions. An example of the differences in mutation variants is illustrated in Table 2. Remarkably, we observed that the −42C>T mutation never occurs in the presence of a pampC gene (0 out of 24 cases). This was even noticed in isolates with the same MLST, i.e. ST88 −42C>T (n=3) and pampC (n=1), suggesting preferred exclusivity for one of the resistance mechanisms. One study mentioned the co-occurrence of the −42C>T mutation and a pampC gene in only 1 out of 36 strains (Mulvey et al. 2005). One could speculate that the exclusivity is a matter of what arrives first, the plasmid or the mutation, after which there is no selective advantage for the second mechanism, or that there is actually a fitness cost to having both the mutation and the plasmid relative to having only the mutation or the plasmid. This hypothesis might be a start for future studies to determine the relative fitness and resistance provided by the −42C>T mutation relative to isolates harbouring a pampC gene. The study performed by Tracz et al. showed that position −32T>A on the promotor of ampC associates with AmpC hyperproduction that results in elevated MIC levels for FOX (Tracz et al. 2007). Surprisingly, in the current study, no significant association of −32T>A with CTX resistance was noticed despite its low consistency index. Only 2 out of 20 isolates with the −32T>A were CTX resistant, 4 out of 20 showed an intermediate elevated CTX MIC, and 14 were susceptible to CTX. Although we do not know under which conditions this mutation did arise, it can be speculated that the high level of homoplasy at the −32 position is associated with a different trait, e.g. resistance against another antibiotic. Prior studies discovered the importance of mutations in the promoter elements. Although an existing promoter is often copied upstream to the gene, a de novo promoter can also evolve out of an existing sequence region. Random sequences can even evolve expression comparable to the wild-type promoter elements after only a single mutation (Yona, Alm, and Gore 2018). This means that the de novo creation of a promoter region within the E. coli may be much more often the result of mutation rather than a rearrangement. Furthermore, these promotor elements evolve to only a few forms, indicating convergent evolution (Liu and Libchaber 2006), as also observed in the 7
RkJQdWJsaXNoZXIy MTk4NDMw