All organisms survive by adapting to their environment, but certain adaptations appearing deleterious in a permissive environment may prove beneficial in specific restrictive conditions—a phenomenon we term extreme antagonism. But how can extreme antagonism be identified given the complexity of the cellular network? Our answer was to adapt the model organism E. coli to the antibiotic rifampicin and identify mutations that reduce growth in the absence of rifampicin as a proof-of-concept for identifying extreme antagonism more generally.

The cover image for the November 17 issue of Biophysical Journal is a network that summarizes the statistically significant transcriptional reprogramming induced as a result of the rifampicin-adaptive mutations. Each rectangle represents a different strain harboring a rifampicin-resistant mutation, and the colors blue, purple, and orange indicate the strain’s slower, equal, and faster growth in the absence of rifampicin, respectively. Thus, blue strains harbor extreme antagonistic mutations. Each circle represents a genetic module controlled by a master regulator and the circle size indicates the number of genes in the module. Lines connecting rectangles with circles denote the upregulation (pointed arrow) or downregulation (flat-head arrow) of a module in the presence (red) or absence (green) of rifampicin. Gray lines between circles indicate larger modules, whose downstream nodes subsume all nodes in a smaller module other than that module’s regulator.

Besides serving as a compact representation of our research findings, the network suggests modules in the regulatory network that are responsible for extreme antagonism. Modules that have incoming green arrows from both blue and orange nodes with opposite arrow heads are correlated with the growth phenotype and may serve as targets for drug-combination treatments that make the acquisition of resistance more costly. Laying out the results in this manner prompted us to hypothesize that changes in the transcription of ribosomal RNA may account for the observed growth phenotypes. Indeed, an alignment analysis of the RNA sequencing data shows, surprisingly, a substantial enrichment in the 23S rRNA gene that correlated with the growth phenotypes in the absence of rifampicin.

This study demonstrates how small changes—such as a single base pair mutation—propagate through the cellular network to the whole cell and help inform the development of strategies that can select against antibiotic resistance.

- Thomas P. Wytock, Manjing Zhang, Adrian Jinich, Aretha Fiebig, Sean Crosson, and Adilson E. Motter.