2 switch

Cells were loaded into the Mother Machine as above. Cells were allowed to grow in fresh media for 10 h, then exposed to 35 mM H₂O₂ for 35 min and then supplied with fresh media again for at least 12 h. The media was switched with a Fluigent 2-switch or M-switch (Fluigent, France). Two 35 min pulses of 3 to 12 mM propidium iodide were supplied with the second round of fresh media and the cells were imaged in the RFP channel to observe DNA chelation of dead cells. This approach was not robust for identifying survivors and dead cells. Thus the movies for each mother cell were manually curated to determine survival using solely the phase contrast channel. If the cell began growing post-H₂O₂ treatment and before the movie ended, it was counted as a survivor. Ambiguous cases were excluded from the tally (WT, 14% of cells excluded, ΔrpoS, 5%), however including these cells in the survival fraction calculation did not change the results.

The flow experiments were performed in Ibidi chambers (µ-Slide VI 0.4, Ibidi). Two 50 ml Falcon tube reservoirs containing SD-full +0.5 µg/ml fluorescein-dextran (D3305, ThermoFischer) and SD-full +0.6 M NaCl were put under a pressure of 30 mbar (FlowEZ, Fluigent). The media coming from each reservoir were connected using FEP tubing (1/16″ OD × 0.020″ ID, Fluigent) to a 3-way valve (2-switch, Fluigent). The concentration of NaCl in the medium was controlled using a Pulse-Width Modulation strategy^{66 ,67} . Periods of 4 s were used and within this time, the valve controlled the fraction of time when SD-full versus SD-full + NaCl was flowing. TTL signals generated by an Arduino Uno board and dedicated scripts were used to control precisely the switching of the valve. The fluorescein present in the SD-full medium quantified outside the Cell object provided an estimate of the NaCl concentration in the medium. Some strong fluctuations in this signal were probably generated by dust particles in the imaging oil or FLSN-dextran aggregates in the flow chamber. Following 24 h log-phase growth, cells bearing the pSTL1-PP7sl reporter, Hog1-mCherry, and Hta2-tdiRFP tags were diluted to OD 0.2, briefly sonicated and loaded in the ibidi channel previously coated by Concanavalin A. Cells were left to settle in the channel for 10 min before SD-full flow was started.