Figure 4
From: A Markovian Approach towards Bacterial Size Control and Homeostasis in Anomalous Growth Processes
(a) The cell size at birth distribution, ρ(l b ), was used as a proxy for determining l0. We filter the data discarding cells with a size of birth larger than the average plus one standard deviation (i.e. cells with a Z-score larger than one): the shadow region indicates the data considered after filtering to estimate l0. The black vertical line shows the computed value of l0 = 3.6 ± 0.5 μm and its comparison with 〈l〉 = 5 ± 2 μm (40% larger than the unit division length l0). The red triangle indicates the value of 〈l b 〉 = 4 μm before filtering out filamentous cells. The inset shows additional measurements to estimate the unit division length: by computing the histograms of sizes of mother, ρ m (l), and daughter, ρ d (l), cells (244 and 488 single cell data points respectively) the unit division length l0 can be estimated as the difference between the most probable value of their sizes (dotted lines, ~3.3 μm) or by the differences of the averages (solid lines, ~3.9 μm). (b) The modulus of the averaged power spectrum 〈S(q)〉 (a.u.) of the fluorescent signal of filamentous cells (MGZ202 in LB, 165 cells) reveals a peak at q* = 0.12 μm−1 and indicates the regularity of the septum-septum distance: λ = 1/q* = 8.3 μm. In filamentous cells, the distance between consecutive Z-rings, as a proxy for the characteristic minimal size upon division, is 2 × l0 due to the doubling in size of central Min domains as cells grow longer as revealed by the kymograph (inset, adapted from35): l0 = λ/2 = 4.2 μm. The Z-ring organizes at locations where the averaged spatiotemporal concentration of Min is minimal (arrows). This phenomenon may also be the reason underlying the preferential asymmetric division of filamentous cells close to the poles during sporadic cleavage events (Fig. S9). The right-most inset shows a sample of a filamentous cell and the GFP fluorescent intensity signal (FtsZ expression).
