Plasmid-free cheater cells commonly evolve during laboratory growth.

It has been nearly a century since the isolation and use of penicillin, heralding the discovery of a wide range of different antibiotics. In addition to clinical applications, such antibiotics have been essential laboratory tools, allowing for selection and maintenance of laboratory plasmids that encode cognate resistance genes. However, antibiotic resistance mechanisms can additionally function as public goods. For example, secretion of beta-lactamase from resistant cells, and subsequent degradation of nearby penicillin and related antibiotics, allows neighboring plasmid-free susceptible bacteria to survive antibiotic treatment. How such cooperative mechanisms impact selection of plasmids during experiments in laboratory conditions is poorly understood. Here, we show that the use of plasmid-encoded beta-lactamases leads to significant curing of plasmids in surface grown bacteria. Furthermore, such curing was also evident for aminoglycoside phosphotransferase and tetracycline antiporter resistance mechanisms. Alternatively, antibiotic selection in liquid growth led to more robust plasmid maintenance, although plasmid loss still occurred. The net outcome of such plasmid loss is the generation of a heterogenous population of plasmid-containing and plasmid-free cells, leading to experimental confounds that are not widely appreciated.