Figure 3. Normalization of maintenance rate and recycling yield with cell volume of Escherichia coli .

• A, B
  Cell size during starvation. Length and width of the cells are measured with phase‐contrast microscopy, and the volume is computed considering cell shape as a cylinder with two semi‐spheres, as described in the graphical synopsis at the top (see also Table EV3 and Methods). Each measurement is an average of 200 cells. (A) Measured length and width of wild‐type cells starved in batch cultures and previously grown in the chemostat at different steady‐state growth rates and of GlpK22 cells starved and previously grown in batch culture at μ = 0.9 h⁻¹ (see upper color legend). Note that, as shown in Schink et al (2019), cell widths do not change from steady‐state growth to starvation, while lengths decrease. Both length and width increase exponentially with growth rate. (B) Starvation volume of the cells described in panel (A), computed as explained in the graphical synopsis (see also Methods). In agreement with literature (Schaechter et al, 1958), it increases exponentially with growth rate (black line).
• C, D
  Recycling yield and maintenance rate per cell volume. (C) Recycling yields and (D) maintenance rate measured as described in Fig 2 are normalized per cell volume and plotted versus the previous growth rate of the cultures they refer to (see color legend at the top and Table EV3). The normalized yield is constant within the uncertainty. Maintenance rate increases significantly with a slope of (0.99 ± 0.55) h, matching the increase in the death rate, (1.1 ± 0.4) h, shown in Fig 2B. Solid black lines show linear fits on log‐transformed data, using weighted least square fits.

Data information: Slopes are reported with standard error and take uncertainty in data into account (see Methods). Data shown as mean ± SD. Two replicates per condition.