The antibiotic sensitivity of ∆ybeX depends on the growth history of cells
Our finding that while the ∆ybeX cells have a lengthened lag phase during outgrowth from the stationary phase, they appear to retain similar levels of metabolic activity during this lag phase to the WT cells, as well as similar exponential growth rate (Fig. 1D ), led us to hypothesize that any cellular defects conferred by the lack of YbeX may accumulate during the late growth, preceding entry into the stationary phase and/or in the stationary phase itself. Such a stochastic process could lead to the observed single-cell level growth heterogeneity (Fig. 2 ). Accordingly, we assayed whether the phenotypes of ∆ybeZ , ∆ybeY, and ∆ybeX depend on the growth phase where the cells originate. We surmised that if the∆ybeX phenotype is caused by a gradual accumulation of harm, then cells that have been given ample time to accumulate such harm, should exhibit a stronger phenotype than cells with only a few divisions.
First, we tested the antibiotic sensitivity phenotype. In this experimental setup, we start by growing a single bacterial colony for 12 hours into the early stationary phase (Fig. 4A ). Then, the experiment is divided into two. In the first arm, to assay the antibiotic sensitivity during outgrowth, stationary liquid cultures are directly dot spotted into agar plates containing sub-inhibitory concentrations of antibiotics. In the second arm, to assay the antibiotic sensitivity of exponentially growing cells, the same stationary cultures are first diluted a hundred-fold into fresh liquid media and grown at 37°C for four to five cell divisions until OD600 reaches 0.2-0.4, after which they are dot spotted.
The ∆ybeX strain exhibited very strong chloramphenicol and erythromycin sensitivity in cells originating from the early stationary phase but no sensitivity to Rifampin (Fig. 4B ). In contrast, the ∆ybeX cells plated on the antibiotic after only a few rounds of the division had WT-like sensitivity to all tested antibiotics. In comparison, ∆ybeZ cultures had similar intermediate levels of sensitivity to chloramphenicol, regardless of the growth history of cells, while they are not sensitive to erythromycin, rifampicin, and tetracycline (Fig. 4B ). Exponentially growing ∆ybeZcells in MOPS minimal medium, supplemented with 0.3% glucose as the carbon source, also exhibited sensitivity to chloramphenicol (Figure S4a ). ∆ybeY cells had a very strong sensitivity to all tested antibiotics under both growth conditions. This is not surprising, considering its strong growth phenotype.
Testing the culture growth in liquid media, after diluting the culture directly from the early stationary phase, again showed a lengthened lag phase for ∆ybeX but not for ∆ybeZ , while the exponential growth rates of both ∆ybeX and ∆ybeZ were very similar to WT (Fig. 4C ). The ∆ybeY strain behaves similarly in both experiments, exhibiting a reduced exponential growth rate and reaching a lower maximal cell density. In contrast, when the cells are outgrown from exponential phase cultures, the WT, ∆ybeZ and ∆ybeXstrains grow equally well, with no visible lag phase, while the∆ybeY strain has a reduced growth rate and a lower growth end-point, as expected (Fig. 4D ).