Antibiotic inhibition uncouples growth from horizontal gene transfer

Abstract

Conjugative plasmids often encode resistance to multiple antibiotics, allowing a single mobile genetic element to transfer across diverse selective environments. This co-selection is often invoked to explain the persistence and spread of multidrug resistance, with antibiotic resistance genes encoded on the same plasmid frequently treated as equivalent routes of plasmid maintenance. However, accumulating evidence suggests that different antibiotics can produce distinct co-selection outcomes. These differences have been interpreted through the lens of fitness and population-level selection, while the potential effects on plasmid transfer itself are largely unknown. Here, we show that different antibiotics generate significantly different transfer dynamics for the exact same multidrug-resistant plasmid. Surprisingly, much of this variation arises from continued conjugation after antibiotic exposure begins, whereby a transient residual conjugation window persists despite strong growth inhibition, allowing new transconjugants to continue forming for hours after treatment in an antibiotic dependent manner. A simple model for describing antibiotic effect on transfer competence of both donors and recipients quantitatively explains these antibiotic-specific dissemination dynamics. Moreover, different recipient genetic backgrounds exhibit distinct residual conjugation windows, demonstrating that the capacity for continued plasmid transfer during antibiotic exposure is both drug- and host-dependent. Together, our results show that antibiotics suppress plasmid dissemination on different timescales, creating drug-specific residual conjugation windows that are largely decoupled from growth inhibition. These findings identify conjugation dynamics as a previously overlooked contributor to co-selection outcomes and suggest that antibiotics that select for the same multidrug-resistant plasmid are not equivalent in their ability to promote plasmid persistence.