Figure 6. Visualization of Pol III* exchange in vivo.

(A) Left: image of τ (orange) and ε (blue) foci within a single E. coli cell, averaged over 40 s. Co-localization of the two signals is shown as a white spot. Middle: bright field image of the same cell. Right and below: fluorescence intensity of τ (orange) and ε (blue) over time. The trajectories are averaged using a 2 s moving average filter. (B) Averaged, normalized cross-correlation functions. The cross-correlation function of 1210 pairs of foci in living cells shows a clear positive peak (black line). The cross-correlation function for 297 pairs of foci in fixed cells (gray line) and the cross-correlation function of 1210 pairs of foci, randomized within the same field of view (red line) show no positive cross correlation. Cross-correlation functions have been vertically offset for clarity. (C) Exponential fit (red) to the cross-correlation function in (B). We obtained an exchange time scale of τ = 4 ± 2 s. The error represents the error of the fit.

DOI: http://dx.doi.org/10.7554/eLife.23932.022

Figure 6—figure supplement 1. Growth curves for E. coli strains: wild-type E. coli (black), cells expressing both C-terminal derivatives of τ (dnaX-YPet) and ε (dnaQ-mKate2) subunits under control from their endogenous promoters (green), and cells expressing only dnaX-YPet (blue) and dnaQ-mKate2 (orange).

Growth curves were measured for 9.5 hr. The division times were obtained from a linear fit of the exponential growth phase. They are 33 ± 8 min for wild-type, 32 ± 5 min for dnaX-YPet, 32 ± 8 min for dnaQ-mKate2, and 33 ± 4 min for the double mutant. The errors represent the errors of the fit.

DOI: http://dx.doi.org/10.7554/eLife.23932.023

Figure 6—figure supplement 2. Cross-correlation analysis of simulated intensity trajectories for pairs of ε and τ foci.

Individual intensity trajectories for 300 ε and τ foci were simulated in MATLAB 2014b. The simulation allows us to set k_on and k_off (in units of frames^–1) for ε exchanging into Pol III*, and k_on and k_off for Pol III* exchanging into the replisome. By changing these rate constants, we can simulate different exchange mechanisms. The black line represents the average cross-correlation function for Pol III* exchange (both ε and τ). Here, k_on and k_off for Pol III* were set to 0.01 and the rate constants for ε were set to k_on = 1 and k_off <<1 to simulate stable binding of core to Pol III*. A clear positive peak can be seen. The green line represents the average cross-correlation function for simulated trajectories without any exchange. Here, k_on = 1 and k_off<<1 for all rate constants, to simulate stable binding to the replisome. In this case, there is no positive cross correlation. The gray line represents the average cross-correlation function for core exchange. In this case k_on = 1 and k_off<<1 for Pol III* and k_on and k_off for ε were set to 0.01. Again, we do not observe a positive cross correlation. Cross-correlation functions have been vertically offset for clarity.

DOI: http://dx.doi.org/10.7554/eLife.23932.024

Figure 6—figure supplement 3. Measurement of concentrations of τ and ε in live cells.

(A) Mean fluorescence signal during photobleaching of wild-type MG1655. (B) Bleaching of the coverslip background signal within a single field of view. (C) Bleaching of ε-mKate2 fluorescence within a single cell, corrected for the cellular autofluorescence (A) and the background fluorescence of the coverslip (B). This was fit with a single exponential decay (black line) to determine the maximum intensity. (D) Histogram of the single-molecule intensities obtained from the change-point step-fitting algorithm (inset). This was fit to a Gaussian distribution to find the mean intensity of a single mKate2 molecule.

DOI: http://dx.doi.org/10.7554/eLife.23932.025