Abstract
Microbes without plasmids divide faster than those harbouring them. Microbiologists rely on this difference in growth rate between both types of microbe to foresee whether a plasmid will be maintained, or else purged by the host to avoid extinction. However, here I report that plasmids change multiple life-history traits and show that growth rate alone can be a bad predictor for plasmid maintenance. Pair-wise competition experiments between two constructs of Escherichia coli—one of which carries a plasmid—revealed that harbouring plasmids can also increase yield and delay growth (lag). Crucially, yield engaged in a trade-off with growth rate. The plasmid borne by one construct (R), non-transmissible and with a tetracycline-resistance gene, reduced its host’s growth rate by 20%. However, given this trade-off, R outgrew its sensitive counterpart (S) in the absence of tetracycline when the competition favoured yield over growth rate. The trade-off makes unclear whether the plasmid is costly to maintain. R-mutants that acquired additional copies of the plasmid, through random segregation, exploited this trade-off and were selected with tetracycline concentrations below the ‘minimal selective concentration’—the lowest antimicrobial concentration thought to select for resistant mutants. My data suggests that plasmids interfere with multiple traits, and whether plasmids are costly to maintain will depend on the relationship between them and which is under strongest selection. Thus, concepts that rely on plasmid carriage costs must be used cautiously.