Collisions between DNA replication complexes (replisomes) and impediments such as damaged DNA or proteins tightly bound to the chromosome lead to premature dissociation of replisomes at least once per cell cycle in Escherichia coli. Left unrepaired, these events produce incompletely replicated chromosomes that cannot be properly partitioned into daughter cells. DNA replication restart, the process that reloads replisomes at prematurely terminated sites, is therefore essential in E. coli and other bacteria. Three replication restart pathways have been identified in E. coli: PriA/PriB, PriA/PriC, and PriC/Rep. A limited number of genetic interactions between replication restart and other genome maintenance pathways have been defined, but a systematic study placing replication restart reactions in a broader cellular context has not been performed. We have utilized transposon insertion sequencing to identify new genetic interactions between DNA replication restart pathways and other cellular systems. Known genetic interactors with the priB replication restart gene (uniquely involved in the PriA/PriB pathway) were confirmed and several novel priB interactions were discovered. Far fewer connections were found with the PriA/PriC or PriC/Rep pathways, suggesting a primacy role for the PriA/PriB pathway in E. coli. Targeted genetic and imaging-based experiments with priB and its genetic partners revealed significant double-strand DNA break (DSB) accumulation in strains with mutations in dam, rep, rdgC, lexA, or polA. Modulating the activity of the RecA recombinase partially suppressed the detrimental effects of rdgC or lexA mutations in ΔpriB cells. Taken together, our results highlight roles for several genes in DSB homeostasis and define a genetic network that facilitates DNA repair/processing upstream of PriA/PriB-mediated DNA replication restart in E. coli.

An E. coli strain carrying MuGam-GFP (SMR14334) was derivatized to carry the sulB103 allele (wt) before P1 transduction deleted other genes of interest. Saturated cultures were diluted 100x and grown in LB for 30 min at 37 °C to enter the early exponential phase. MuGam-GFP expression was then induced at 100 ng/mL doxycycline and growth continued for an additional 2.5 hr at 37 °C. Cells were pelleted and resuspended in 1x PBS buffer (to OD₆₀₀ of 1.0) and placed on ice. About 15 min prior to imaging, cell membrane stain FM 4-64 (5 mM) was added, and 2-3 µL of cells were sandwiched between a 24x50 mM, No. 1.5 coverslip (Azer Scientific) and a 1.5% agarose pad. All cells were imaged at room temperature with a motorized inverted Nikon Ti-eclipse N-STORM microscope equipped with a 100x objective and ORCA Flash 4.0 digital CMOS C13440 (Hamatsu). Imaging was performed using NIS-Elements software with the microscope in epifluorescence mode. Cells were first imaged in the brightfield (4.5 V, 100 ms exposure). Visualization of the cell membranes was performed in the DsRed channel to ensure the focusing (4.5 V, 50 ms exposure) and then MuGam-GFP was imaged in the GFP channel (4.5 V, 50 ms exposure). Growth, preparation, and imaging were performed for each strain in biological triplicate.

Analysis of cell features was performed with Fiji software (ImageJ) equipped with plugins as described previously: Single Molecule Biophysics (https://github.com/SingleMolecule/smb-plugins) and MicrobeJ. Briefly, the nd2 raw images for each strain (4 to 8 per replicate with a maximum difference of 2 images within triplicate) were concatenated together by channels. The image processing of each channel was carried out the same way and uniformly throughout the field of view. The scale of all images was corrected to fit the Hamamatsu camera scale. The brightfield and DsRed image stacks were auto-scaled while the GFP images were processed with discoidal averaging of 1-5 and intensity scale set at 0-300. Both brightfield and DsRed channels were cleaned by running a Bandpass filter 10_2 with autoscale 5, a rolling sliding stack of 10, and an enhance contrast of 0.1. Channel stacks were converted to 8 bits before analysis in MicrobeJ. For the analysis, hyperstacks combining only the FM 4-64 and GFP channels were generated in MicrobeJ. From these hyperstacks, cell outlines were detected in the DsRed channel using the default method with a threshold of +25. Within identified cells, GFP foci were detected using the maxima features as foci with a Gaussian fit constraint. The exact setup used to identify bacteria and MuGam-GFP foci in MicrobeJ is available (Final Bacteria setup 1_5 foci 90) as a .xml file. After automatic detection, cells were manually sorted to remove poorly fitting outlines or outlines fitting to cells out of focus. Cell features analysis acquired with MicrobeJ (cell ID, cell length, number of foci per cell, foci intensity, and size) were exported as .csv files. Plots and statistical analysis were generated and performed with GraphPad Prism software. At least 650 single cells were analyzed for each condition. 

Microsoft Excel; Fiji (equipped with single molecule and MicrobeJ plugins); GraphPad Prism.