Biocides are used to control microorganisms across different applications, but emerging resistance may pose risks for those applications. Resistance to biocides has commonly been studied using adaptive laboratory evolution (ALE) experiments with growth at subinhibitory concentrations linked to serial subculturing. It has been shown recently that E. coli adapts to repeated lethal stress imposed by the biocide benzalkonium chloride (BAC) by increased survival (i.e. tolerance) and not by evolving the ability to grow at increased concentrations (i.e. resistance). Here, we investigate the contributions of evolution for tolerance as opposed to resistance for the outcome of ALE experiments with E. coli exposed to BAC. We find that BAC concentrations close to the half maximal effective concentration (EC50, 4.36 µg mL-1) show initial killing (~40%) before the population resumes growth. This indicates that cells face a two-fold selection pressure: for increased survival and for increased growth. To disentangle the effects of both selection pressures, we conducted two ALE experiments: (i) one with initial killing and continued stress close to the EC50 during growth and (ii) another with initial killing and no stress during growth. Phenotypic characterization of adapted populations showed that growth at higher BAC concentrations was only selected for when BAC was present during growth. Whole genome sequencing revealed distinct differences in mutated genes across treatments. Treatments selecting solely for survival led to mutations in genes for metabolic regulation (cyaA) and cellular structure (flagella fliJ), while treatments selecting for growth and survival led to mutations in genes related to stress response (hslO and tufA). Our results demonstrate that serial subculture ALE experiments with an antimicrobial at sub-inhibitory concentrations can select for increased growth and survival. This finding has implications for the design of ALE experiments to assess resistance risks of antimicrobials in different scenarios such as disinfection, preservation, and environmental pollution.

The dataset contains microbial growth data in the presence and absence of the biocide benzalkonium chloride. The data was measured as increase of OD over time in a microtiter plate reader (Fig. 2A, Fig. 3B & C) or as counts of colony forming units (cfu) (Fig. 2B & C and Fig. 3A). In addition, antimicrobial susceptibility data regarding different antibiotics and benzalkonium chloride for E. coli populations that evolved under benzalkonium chloride exposure are provided. The details of the underlying experiments and data processing are described in the Methods section of the linked manuscript. The provided data constitutes results of processed growth data, which was used to create the graphs shown in Figure 2 and Figure 3 in the manuscript.