Fast, multiphase volume adaptation to hyperosmotic shock by Escherichia coli

Abstract

All living cells employ an array of different mechanisms to help them survive changes in extra cellular osmotic pressure. The difference in the concentration of chemicals in a bacterium's cytoplasm and the external environment generates an osmotic pressure that inflates the cell. It is thought that the bacterium Escherichia coli use a number of interconnected systems to adapt to changes in external pressure, allowing them to maintain turgor and live in surroundings that range more than two-hundred-fold in external osmolality. Here, we use fluorescence imaging to make the first measurements of cell volume changes over time during hyperosmotic shock and subsequent adaptation on a single cell level in vivo with a time resolution on the order of seconds. We directly observe two previously unseen phases of the cytoplasmic water efflux upon hyperosmotic shock. Furthermore, we monitor cell volume changes during the post-shock recovery and observe a two-phase response that depends on the shock magnitude. The initial phase of recovery is fast, on the order of 15-20 min and shows little cell-to-cell variation. For large sucrose shocks, a secondary phase that lasts several hours adds to the recovery. We find that cells are able to recover fully from shocks as high as 1 Osmol/kg using existing systems, but that for larger shocks, protein synthesis is required for full recovery.

Figure 1. Passive response to hyperosmotic shock involves two phases of cell shrinking.

(A) A sequence of fluorescent images of E.coli cells prior and during the hyperosmotic shock. Top: cytoplasmic volume marked with the green protein EGFP; Middle: total cell volume marked by the red outer membrane dye FM4-64; Bottom: overlay if EGFP and FM4-64. A cell is transferred from isoosmotic buffer (10 mM Tris-HCl buffer supplemented with 150 mM sucrose) to isoosmotic buffer and 620 mM sucrose. Images were acquired at a frame every 1.6 seconds alternating between the cytoplasmic volume and total cell volume. Image brightness was adjusted due to photo bleaching and the difference in fluorescence intensity between the FM4-64 and EGFP. Videos are given in Video S1 and S2. (B) Volume recovery versus time obtained using cell area analysis. Inset: cell length and cell radius versus time. Total cell volume, length and radius are given in red and cytoplasmic cell volume, length and radius in green. An example contour of the cytoplasmic cell volume and total cell volume obtained after cell area analysis for the initial frame is shown on the bottom left. The inset shows a cartoon illustration of the post sucrose shock events described in the text.

Figure 2. Cell shrinking is fast and doesn't require AqpZ.

(A) Averaged volume traces shown for different shock magnitudes (V_n(t)). For each cell in a given data set, the recovery trace was normalized with initial volume and aligned at the time of hyperosmotic shock. The average trace was computed from normalized and aligned data sets as described in the Materials and Methods. For display, the averaged trace for each shock is shifted 10 seconds in time compared to the last one. (B) Rate of initial fast volume reduction dV_n(t)/dt for different shock magnitudes. Error bars are standard error of the mean. (C)–(E) Averaged volume traces (V_n(t)) aquired at 1 Hz resolution for AqpZ+ strain (black) and AqpZ− strain (red). (C) Cells were grown in LB (0.44 Osmol/kg) to OD of 0.45–0.65 and transferred to LB with 620 mM sucrose. Approximately 100 cells were used for the AqpZ+ strain and 14 cells for the AqpZ− strain. (D) Cells were grown in low osmolarity LB (prepared with no NaCl to a final osmolarity of 0.08 Osmol/kg) to OD = 0.8 and transferred into low osmolarity LB and 620 mM sucrose. Averaged traces were obtained from 23 cells for ApqZ+ and 24 for AqpZ−. (D) Cells were grown in LB (0.44 Osmol/kg) for 15 hours and transferred into LB supplemented with 620 mM sucrose. Averaged traces were obtained from 12 cells for AqpZ+ and 15 cells for AqpZ−.

Figure 3. Volume recovery is complex and required for growth.

Averaged recovery traces shown for different shock magnitudes (V_n(t)). The initial 10 minutes of data is recorded at 1 Hz frame rate, subsequent data is recorded at a frame every 30 seconds. For each cell in a given data set, the recovery trace was normalized by the initial volume, aligned at the time of hyperosmotic shock and resampled every 30 seconds. The average trace was computed from these normalized and aligned data sets. The following conditions are shown: 0.44 Osmol/kg data (N = 9), 0.5 Osmol/kg (N = 4), 0.6 Osmol/kg (N = 22), 0.8 Osmol/kg (N = 6), 1.15 Osmol/kg (N = 28), 1.45 Osmol/kg (N = 16), 2 Osmol/kg (N = 12) and 2.5 Osmol/kg (N = 5).

Figure 4. 2D Probability density function (PDF) of the recovery traces for a given shock magnitude.

Approximately 100 cells per shock magnitude are includes. PDFs are calculated from normalized and aligned data sets as in Figure 3. Cells were shocked by transferring them from LB into LB and a given amount of sucrose. Post shock osmolality is given in each panel.

Figure 5. Slow recovery requires protein synthesis.

(A)–(E) Averaged recovery traces shown for different shock magnitudes (V_n(t)). For each shock, the recovery trace in the presence (red) and absence (black) of the drug chloramphenicol is shown. Data was recorded and analyzed as in Figure 3. 0.44 Osmol/kg (no shock, N = 10), 0.6 Osmol/kg (N = 5), 0.8 Osmol/kg (N = 5), 1.15 Osmol/kg (N = 20) and 1.45 Osmol/kg (N = 10). (F) Final volume, calculated as the average of the last 15 minutes of data in (A–E), versus shock magnitude. Error bars are standard error of the mean.