DNA Double-Strand Breaks Induced by Reactive Oxygen Species Promote

bioRxiv preprint doi: https://doi.org/10.1101/533422; this version posted January 29, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

1 DNA double-strand breaks induced by reactive oxygen 2 species promote DNA polymerase IV activity in 3 Escherichia coli

5 Sarah S. Henrikus1,2, Camille Henry3, John P. McDonald4, Yvonne Hellmich5, 6 Elizabeth A. Wood3, Roger Woodgate4, Michael M. Cox3, Antoine M. van Oijen1,2, 7 Harshad Ghodke1,2, Andrew Robinson1,2*

8 1Molecular Horizons Institute and School of Chemistry and Biomolecular Science, University 9 of Wollongong, Wollongong, Australia 10 2Illawarra Health and Medical Research Institute, Wollongong, Australia 11 3Department of Biochemistry, University of Wisconsin-Madison, United States of America 12 4Laboratory of Genomic Integrity, National Institute of Child Health and Human 13 Development, National Institutes of Health, Bethesda, Maryland, United States of America 14 5Institute of Biochemistry, Goethe Universität, Frankfurt, Germany 15 *Corresponding author. Mailing address: School of Chemistry, University of Wollongong, 16 Wollongong, NSW 2500, Australia. Email: [email protected]. 17

1 Under many conditions the killing of bacterial cells by antibiotics is potentiated by 2 DNA damage induced by reactive oxygen species (ROS)1–3. A primary cause of ROS- 3 induced cell death is the accumulation of DNA double-strand breaks (DSBs)1,4–6. DNA 4 polymerase IV (pol IV), an error-prone DNA polymerase produced at elevated levels in 5 cells experiencing DNA damage, has been implicated both in ROS-dependent killing 6 and in DSBR7–15. Here, we show using single-molecule fluorescence microscopy that 7 ROS-induced DSBs promote pol IV activity in two ways. First, exposure to the 8 antibiotics ciprofloxacin and trimethoprim triggers an SOS-mediated increase in 9 intracellular pol IV concentrations that is strongly dependent on both ROS and DSBR. 10 Second, in cells that constitutively express pol IV, treatment with an ROS scavenger 11 dramatically reduces the number of pol IV foci formed upon exposure to antibiotics, 12 indicating a role for pol IV in the repair of ROS-induced DSBs. 13 Two mechanisms for ROS-induced DSB formation have been proposed (Figure 1a). 14 The first invokes oxidisation of the cellular nucleotide pool, leading to increased 15 incorporation of oxidised nucleotide triphosphates (e.g. 8-oxo-dGTP) into the DNA by DNA 16 polymerases4,7,16. Subsequent initiation of base-excision repair (BER) creates single- 17 stranded DNA (ssDNA) gaps. In cases where BER is initiated at nearby sites, DSBs may be 18 formed1,4,7. Evidence for a second mechanism of ROS-dependent DSB formation has 19 emerged from a recent mechanistic study of thymineless death in Escherichia coli5. The 20 ROS-driven potentiation of killing by both antibiotic treatment and thymineless death can be 21 abrogated through the addition of ROS scavengers to the culture medium5,17,18. Using 22 microscopy to quantify ssDNA gaps and DSBs in cells undergoing thymineless death, Hong 23 and co-workers recently discovered that thymine starvation initially leads to the accumulation 24 of ssDNA gaps, which are subsequently converted to DSBs in an ROS-dependent process5 25 (Figure 1b). In cells treated with ROS scavengers, gaps are not converted to DSBs and 26 cells are tolerant to thymine starvation5. 27 Several lines of evidence implicate pol IV in ROS-dependent DSB formation and/or 28 processing. Pol IV efficiently incorporates 8-oxo-dGTP into DNA in vitro7. Over-expression of 29 pol IV is lethal to cells and the lethality is ROS-dependent4,7,19. Cells lacking pol IV are 30 partially protected against killing by ampicillin under conditions where the nucleotide pool 31 becomes oxidised7. These observations are suggestive of pol IV playing a role in the BER- 32 dependent mechanism for DSB formation4 (Figure 1a). Other studies indicate that pol IV has 33 a role in the repair of DSBs6,9,10,12,13,20–26. First, pol IV physically interacts with the RecA 34 recombinase; a key player in DSB repair (DSBR)24. Second, when expressed from a low- 35 copy plasmid, fluorescently labelled pol IV colocalises with RecA extensively in cells treated 36 with DNA-damaging agents10. Pol IV colocalises at sites of induced DSBs in cells10. Third, 37 genetic studies reveal that the gene encoding pol IV, dinB, is required for both induced and 38 spontaneous error-prone DSBR6,9,12,13,22,23. Fourth, intermediates of DSBR known as 39 recombination D-loops are efficiently utilised as substrates by pol IV in vitro25. Interestingly, 40 pol IV’s mutagenic potential is modulated by interaction with the recombinase RecA20,21,24,26. 41 Notably, the cellular concentration of pol IV increases significantly upon induction of the SOS 42 response27,28. Single-stranded DNA formed during DSBR (through RecBCD-dependent end- 43 resection) is a major trigger for SOS induction under some conditions29–31. 44 Here, we used single-molecule fluorescence microscopy to investigate whether ROS, 45 and the DSBs they create, influence pol IV activity in cells. We made use of two antibiotics 46 for which killing is known to involve ROS generation; ciprofloxacin and trimethoprim (Figure 47 1a)32–35. For ciprofloxacin, a DNA gyrase inhibitor, a second, ROS-independent pathway 48 exists in which gyrase-stabilised cleavage complexes dissociate, creating a DSB (Figure

1 1b)32,33,36,37. During antibiotic treatment, we would expect that some pol IV binding would be 2 dependent on ROS and/or DSB processing while others would be independent of these 3 factors. Throughout the study, we used dimethyl sulfoxide (DMSO) to create low ROS 4 conditions. DMSO is an effective scavenger of antibiotic-induced ROS5,38. In agreement with 5 the scheme in Figure 1a, co-treatment with DMSO increases survival in cells exposed to 6 both ciprofloxacin and trimethoprim5,38 (Supplementary figure 1,2). 7 To compare pol IV behaviour under high-ROS (no DMSO) and low-ROS (DMSO 8 added) conditions, we recorded time-lapse movies (180 min duration)28,39 of cells treated 9 with i) ciprofloxacin alone, ii) ciprofloxacin and DMSO in combination, iii) trimethoprim alone 10 or iv) trimethoprim and DMSO in combination (Figure 2). Following antibiotic addition, we 11 recorded time-lapse movies capturing fluorescence from a functional DinB-YPet fusion28. 12 Cells treated with ciprofloxacin became filamentous and showed a clear increase in DinB- 13 YPet intensity (Figure 2a,b; Supplementary figure 3). This is consistent with our previous 14 study in which we measured a six-fold increase in intracellular DinB-YPet (pol IV) 15 concentrations from 6 ± 1 nM prior to treatment to 34 ± 3 nM 180 min after ciprofloxacin 16 addition28. Interestingly, inclusion of the ROS scavenger DMSO led to a significant reduction 17 in ciprofloxacin-induced up-regulation of pol IV levels; 180 min after ciprofloxacin addition, 18 cellular DinB-YPet intensities were increased above basal levels by less than 3 fold (Figure 19 2a,b), corresponding to a final concentration of ~17 nM (see Materials and Methods). 20 Treatment with trimethoprim alone led to a significant increase in DinB-YPet fluorescence; 21 180 min after trimethoprim addition the mean fluorescence intensity increased by 50% of 22 ciprofloxacin-alone levels (Figure 2a,b), corresponding to a final intracellular pol IV 23 concentration of ~20 nM. Inclusion of DMSO led to a significant reduction in trimethoprim- 24 induced up-regulation; cell-associated DinB-YPet intensities increased by only 11% 25 compared with ciprofloxacin-alone treatment, corresponding to a final pol IV concentration of 26 ~9 nM. Thus, for both the ciprofloxacin and trimethoprim treatments, addition of the ROS 27 scavenger DMSO significantly reduces the up-regulation of pol IV in response to antibiotic 28 treatment. 29 Ciprofloxacin-treated cells also produced multiple DinB-YPet foci (Figure 2c). These 30 foci require the catalytic activity of pol IV and are produced as individual DinB-YPet 31 molecules bind to DNA and thus experience decreased diffusional mobility28, a phenomenon 32 known as detection by localisation40. Prior to the addition of ciprofloxacin, cells contained on 33 average 0.6 ± 0.2 foci per cell in the absence of DMSO, and 0.4 ± 0.1 foci per cell in the 34 presence of DMSO (Figure 2c). Following treatment with ciprofloxacin alone, the number of 35 foci steadily increased. By 180 min, cells had 4.2 ± 1.1 foci per cell. Upon ciprofloxacin- 36 DMSO treatment, cells contained 1.8 ± 0.4 foci per cell; a significant reduction compared to 37 ciprofloxacin-alone measurements. Addition of DMSO reduced the number of DinB-YPet foci 38 by >50%. Prior to the addition of trimethoprim, cells contained on average 0.5 ± 0.1 foci in 39 the absence of DMSO and 0.4 ± 0.1 foci in the presence of DMSO. Trimethoprim-alone 40 treatment induced a slight increase in the number of DinB-YPet foci with 0.9 ± 0.2 per cell at 41 180 min. This is lower than the number of foci observed for ciprofloxacin-DMSO treatment 42 (1.8 ± 0.4 per cell), despite the measured pol IV concentration being marginally higher after 43 trimethoprim-alone treatment (Figure 2b). Strikingly, cells treated with both trimethoprim and 44 DMSO did not show any increase in DinB-YPet foci after trimethoprim addition (0.5 ± 0.1 foci 45 per cell at 180 min; Figure 2c). Together, these studies demonstrate that treatment with 46 DMSO in addition to either ciprofloxacin or trimethoprim treatment not only supresses the 47 SOS-induced up-regulation of DinB-YPet, but also the binding of DinB-YPet to its cellular 48 substrates.

1 Reasoning that the decreased induction of DinB-YPet expression in cells co-treated 2 with DMSO likely resulted from attenuation of the SOS response, we repeated the time- 3 lapse experiments on cells that carried an SOS-reporter plasmid, in which GFP is expressed 4 from the SOS-inducible sulA promoter (pUA66 PsulA-GFP; fast-folding GFP, gfpmut241). In 5 the absence of any antibiotic treatments, cells exhibit very low fluorescence intensity, 6 consistent with the repression of the sulA promoter in the absence of exogenous sources of 7 DNA damage (Figure 3, ‘0 min’). SOS levels were similarly low for cells grown in the 8 presence of DMSO. Cells exhibited robust SOS induction upon treatment with ciprofloxacin 9 as evidenced by the increase in GFP fluorescence in the three hour time window after 10 addition of ciprofloxacin (Figure 3b). Consistent with our hypothesis, SOS induction was 11 strongly inhibited upon inclusion of DMSO during ciprofloxacin treatment (7% of 12 ciprofloxacin-alone levels at 180 min; Figure 3b). Similarly, high levels of SOS induction in 13 trimethoprim treated cells were supressed by the addition of DMSO (0.6% of ciprofloxacin- 14 alone levels; Figure 3b). Notably, the addition of a different ROS scavenger, 2,2’-bipyridine 15 (BiP, 0.35 mM, 0.5 x MIC5), similarly supressed the induction of the SOS response (Figure 16 3). 17 We further reasoned that the suppression of SOS by ROS scavengers might reflect a 18 reduction in the formation and processing of DSBs. It was previously observed that induction 19 of SOS by the antibiotic nalidixic acid was completely blocked in cells that carried a recB 20 mutation and were therefore incapable of processing DSBs through the RecBCD end- 21 resection nuclease complex30,31. This implies that SOS is primarily triggered by DSB 22 processing in nalidixic acid-treated cells. As ciprofloxacin and nalidixic acid both target DNA 23 gyrase32,33,42, we repeated the GFP reporter measurements in cells lacking recB (SSH111, 24 ΔrecB PsulA-GFP) to determine if SOS induction by ciprofloxacin is also dependent on DSB 25 processing. The deletion of recB strongly inhibited the SOS response following ciprofloxacin 26 treatment (~0.4% induction at 180 min in comparison to recB+, Figure 3, Supplementary 27 figure 4). While recB deletions are known to reduce survival in cells treated with 28 ciprofloxacin43, we observed that most cells lacking recB continued to grow and divide during 29 the 180 min time-lapse measurement (Supplementary figure 3,4), indicating that the lack of 30 SOS induction observed for ciprofloxacin-treated recB-deficient cells did not stem from gross 31 inhibition of all cellular functions. Thus, induction of the SOS response is dependent on DSB 32 processing in cells treated with ciprofloxacin. Together the results are consistent with a 33 model in which the SOS response is triggered in antibiotic-treated cells via ROS-induced 34 DSBs, leading to increased levels of pol IV in cells. 35 We next considered the possibility that pol IV plays a direct role in DSBR. To this end 36 we examined rates of DinB-YPet focus formation in cells treated with ciprofloxacin and 37 trimethoprim, comparing conditions that permitted DSB resection (recB+, no DMSO) or 38 prevented DSB resection (co-treatment with DMSO to supress ROS-induced DSB formation, 39 or ΔrecB to prevent resection). To separate effects on focus formation from effects on DinB- 40 YPet expression, these measurements were carried out in a lexA(Def) background44 (dinB- 41 YPet dnaX-mKate2 lexA[Def]). These cells constitutively express DinB-YPet at levels 42 consistent with SOS induced levels, even in the absence of DNA damage28. Consistent with 43 the results from our previous study28, close to zero DinB-YPet foci were observed in 44 lexA(Def) cells in the absence of antibiotic (0.08 ± 0.05 foci per cell, Figure 4). lexA(Def) 45 cells treated with ciprofloxacin for 60 min exhibited clear foci (1.43 ± 0.15 foci per cell, 46 Figure 4b). Co-treatment with ciprofloxacin and DMSO yielded fewer foci (1.02 ± 0.13 foci 47 per cell, Figure 4b). In ΔrecB, cells contained 0.23 ± 0.05 foci per cell. Cells treated with 48 trimethoprim for 60 min contained DinB-YPet foci (2.6 ± 0.18 foci per cell), whereas cells

1 treated with both trimethoprim and DMSO contained few foci (0.19 ± 0.06). Trimethoprim- 2 treated ΔrecB cells also contained very few foci (0.14 ± 0.05). Similar effects were observed 3 in lexA+ cells, although reductions in focus formation were conflated with reductions in DinB- 4 YPet expression levels (Figure 2c). Together, these results indicate that reducing the 5 number of ROS-induced DSBs being formed, and inhibiting DSB resection, upon antibiotic 6 treatment led to reduced focus formation by DinB-YPet. This implies that pol IV is normally 7 active at ROS-induced DSBs in cells treated with ciprofloxacin or trimethoprim. Importantly, 8 RecBCD function is critical for pol IV activity. 9 In ciprofloxacin-treated cells, 10% of pol IV foci form in the vicinity of replisomes28. 10 Here we observed that in addition to reducing the number of DinB-YPet foci, DMSO 11 treatment dramatically increased the relative colocalisation of DinB-YPet with replisomes in 12 both lexA+ and lexA(Def) cells treated with ciprofloxacin (Supplementary figure 5, 6). For 13 long-lived pol IV foci (detectable within a 10s average projection image) in the lexA(Def) 14 background, 80% of foci colocalised with replisomes under ciprofloxacin-DMSO conditions 15 (Supplementary figure 6b). This is consistent with the addition of DMSO having removed 16 the vast majority of non-replisomal substrates for pol IV-dependent DNA synthesis. For 17 lexA(Def) cells treated with trimethoprim, addition of DMSO abolished long-lived pol IV foci 18 entirely (Supplementary figure 6a,c). 19 Finally, we explored if ROS-induced DSBs promote a change in the binding activity of

20 the other major error-prone polymerase in E. coli, pol V (UmuDʹ2C). As before, a lexA(Def) 21 background (RW1286, umuC-mKate2 dnaX-YPet lexA[Def]) was used to separate effects on 22 focus formation from effects on UmuC-mKate2 expression. Few UmuC-mKate2 foci were 23 observed in the absence of antibiotic (about 0.32 ± 0.08 foci per cell, Supplementary figure 24 7). In cells treated with ciprofloxacin or trimethoprim for 60 min, foci were clearly visible 25 (ciprofloxacin: 1.24 ± 0.16 foci per cell; trimethoprim 1.39 ± 0.21 foci per cell). In both cases, 26 co-treatment with DMSO had little effect on the number of UmuC-mKate2 foci (ciprofloxacin- 27 DMSO: 0.99 ± 0.12 foci per cell; trimethoprim-DMSO 1.26 ± 0.16 foci per cell) or the overall 28 levels of UmuC-mKate2 fluorescence in the cells. Thus in contrast to the effects observed for 29 pol IV, ROS-induced DSBs had little effect on pol V activity. Interestingly, trimethoprim 30 treatment (with or without DMSO) did not lead to activation of pol V: UmuC-mKate2 failed to 31 undergo its characteristic spatial redistribution39, remaining bound to the cell membrane 32 (Supplementary figure 8a, c). Consistent with this, cleavage of UmuD to UmuD′ was far 33 less efficient in trimethoprim-treated cells than in ciprofloxacin-treated cells (compare 34 Supplementary figure 9b and d). 35 We conclude that the processing of ROS-induced DSBs induces high intracellular 36 concentrations of DinB-YPet and promotes focus formation. The observations are consistent 37 with a model in which ROS-induced DSBs promote pol IV activity by inducing the SOS 38 response and by generating substrates for pol IV in the form of recombination intermediates. 39 Few DinB-YPet foci were observed in cells treated with a combination of trimethoprim and 40 DMSO (Figure 4). Based on events that occur during the analogous process of thymineless 41 death5,45, treatment with trimethoprim should induce the formation of ssDNA gaps in the 42 wake of the replisome. In the presence of ROS these would be rapidly converted to DSBs, 43 whereas in the presence of ROS scavengers the gaps would persist. The low extent of focus 44 formation observed under trimethoprim-DMSO conditions implies that pol IV rarely acts at 45 these ssDNA gaps. In contrast, the formation of foci by the pol V was not affected by the 46 addition of DMSO (Supplementary figure 7). This implies that ssDNA gaps might represent 47 a major substrate for pol V and that DSB intermediates are rarely acted upon by pol V. In 48 cells treated with ciprofloxacin, the colocalisation of DinB-YPet with replisomes was

1 substantially increased in the presence of DMSO. Ciprofloxacin induces the formation of 2 end-stabilised DNA-gyrase complexes, which slow cell growth but are non-lethal46. It is 3 possible that replisome-proximal DinB-YPet foci, that are insensitive to ROS, reflect pol IV 4 molecules that are recruited to replisomes that have stalled at end-stabilised DNA-gyrase 5 complexes. 6

7 Methods

8 Strain construction

9 EAW102 is E. coli K-12 MG1655 ΔrecB and was constructed using λRED 10 recombination. The kanamycin resistance marker in EAW102 was removed via FLP-FRT 11 recombination47 using the plasmid pLH29 to obtain kanamycin sensitive HG356. 12 SSH091 and SSH111 (dinB+ lexA+ recB+ + pUA66-sulA-GFP and dinB+ lexA+ 13 ΔrecB::FRT + pUA66-sulA-GFP) were created by transforming MG1655 and EAW102 with 14 pUA66-sulA-GFP41. 15 RW1286 is E. coli MG1655 umuC-mKate2 dnaX-YPet sulA-::kanR lexA51(Def)::CmR 16 and was made in two steps: first the wild-type sulA+ gene of EAW282 was replaced with 17 sulA-::kan by P1 transduction from EAW1339, to create EAW282 sulA-; then lexA51(Def) 18 malB::Tn9 was transferred from DE40644 into EAW282 sulA- by P1 transduction, selecting 19 for chloramphenicol resistance. To confirm the presence of the lexA(Def) genotype, colonies 20 were then screened for high levels of RecA expression by Western blotting with anti-RecA 21 antibodies48. 22 EAW1144 is E. coli K-12 MG1655 dinB-YPet dnaX-mKate2 sulA- lexA51(Def) ΔrecB 23 and was constructed in three steps: sulA- FRT-Kan-FRT was P1 transduced in EAW643 24 (KanS) using a P1 lysate grown on EAW13 to obtain the strain EAW1134. The Kan cassette 25 was removed using pLH2947. Then, lexA51(Def) malB::Tn9 was transduced into EAW1134 26 using a P1 lysate grown on DE406 to obtain the strain EAW1141. Finally, ΔrecB FRT-KanR- 27 FRT was transduced into EAW1141using P1 lysate grown on EAW102 to obtain EAW1144. 28 All mutations introduced were confirmed by PCR. 29 30 Table 1. Strains used in this study. Strain Relevant Genotype Parent Source/technique strain

MG1655 dinB+ dnaX+ recB+ lexA+ - published49 EAW102 ΔrecB::KanR MG1655 Lambda Red recombination HG356 ΔrecB::FRT MG1655 EAW102 SSH091 dinB+ lexA+ recB+ + pUA66- MG1655 Transformation of sulA-GFP MG1655 with pUA66- sulA-GFP SSH111 dinB+ lexA+ ΔrecB::FRT + HG356 Transformation of HG356 pUA66-sulA-GFP with pUA66-sulA-GFP41 EAW18 ΔdinB::KanR MG1655 published28 RW120 recA+ sulA- RW118 published50

lexA+ ΔumuDC::CmR RW546 recA+ sulA- RW542 published51 lexA51(Def) ΔumuDC::CmR RW880 ΔumuDC::CmR MG1655 Transduction of MG1655 with P1 grown on RW12050 JJC5945 dnaX-YPet::KanR MG1655 published39 EAW642 dnaX-mKate2::KanR MG1655 published28 EAW633 dinB-YPet::KanR MG1655 published28 EAW643 dinB-YPet::FRT dnaX- EAW633 published28 mKate2::KanR EAW191 umuC-mKate2::KanR MG1655 published39 EAW282 umuC-mKate2::FRT dnaX- JJC5945 published39 YPet::KanR EAW13 sulA-::KanR MG1655 published39 EAW282 sulA- umuC-mKate2::FRT dnaX- EAW282 Transduction of EAW282 YPet::FRT sulA-::KanR with P1 grown on EAW13 RW1286 umuC-mKate2::FRT dnaX- EAW282 Transduction of EAW282 YPet::FRT sulA-::KanR sulA- sulA- with P1 grown on lexA51(Def)::CmR DE40644 RW1594 dinB-YPet dnaX-mKate2 RW1588 published28 sulA-::KanR lexA51(Def)::CmR EAW1134 dinB-YPet::FRT dnaX- EAW643 Transduction of EAW643 mKate2::FRT sulA-::KanR with P1 grown on EAW13 EAW1141 dinB-YPet::FRT dnaX- EAW1134 Transduction of EAW1134 mKate2::FRT sulA-::FRT with P1 grown on DE406 lexA51(Def)::CmR EAW1144 dinB-YPet::FRT dnaX- EAW1141 Transduction of EAW1141 mKate2::FRT sulA-::FRT with P1 grown on lexA51(Def)::CmR EAW102 ΔrecB::KanR

1 DNA damaging agent sensitivity assay 2 Cells were grown in LB overnight at 37ºC. The next day, a 1/100 dilution of each

3 culture was grown in LB medium (at 37ºC, 150 rpm) until reaching mid log phase (OD600 = 4 0.2). Cell cultures were then serially diluted in PBS by factors of ten down to 10-5. Serial 5 dilutions were spotted (spot volume 10 µL) on fresh LB plates (Difco brand) and LB plates 6 containing DNA damaging agent (which were protected from light). For Supplementary 7 figure 1, DNA damaging agents were added at the following concentrations: 0.1 µg/mL 8 trimethoprim; 12 ng/mL ciprofloxacin. For Supplementary figure 2, DNA damaging agents 9 were added at the following concentrations: 0.1 µg/mL or 0.2 µg/mL trimethoprim; 10 ng/mL 10 or 12 ng/mL ciprofloxacin. Some survival essays were performed under radical-scavenging 11 conditions. Additionally to the particular antibiotic, DMSO (2% v/v, 282 mM, 0.2 x MIC5) was 12 then also added to the plates. Plates were incubated at 37ºC in the dark.

1 Trimethoprim survival assay 2 Cells were grown in LB overnight at 37⁰C. The next day, a 1/100 dilution of each 3 culture was grown in 5 mL LB medium (at 37⁰C, 150 rpm) until reaching mid log phase

4 (OD600 = 0.3). Then, trimethoprim was added to 1 µg/mL followed by incubation at 37⁰C, 5 150 rpm. Samples of 500 µL were taken at 0, 0.5, 1, 2, 3 and 5 h; samples at 0h were taken 6 just before trimethoprim addition. Each sample was centrifuged and resuspended in 500 µL 7 LB; this was done twice to dilute out trimethoprim. Then, 50 µL of 500 µL were plated on LB 8 agar plates; three technical replicates were made for each time point and strain. Plates were 9 incubated at 37⁰C in the dark. The next day, colonies were counted and CFU/mL were 10 calculated for each time point; the average over the technical replicates was then calculated. 11 Two biological replicates were performed. Data points represent the average over the 12 biological duplicates. Error bars represent standard error of the mean over two biological 13 replicates.

14 Fluorescence microscopy 15 Wide-field fluorescence imaging was conducted on an inverted microscope (IX-81, 16 Olympus with a 1.49 NA 100x objective) in an epifluorescence configuration. Continuous 17 excitation is provided using semidiode lasers (Sapphire LP, Coherent) of the wavelength 514 18 nm (150 mW max. output) and 568 nm (200 mW max. output). DnaX-mKate2 in EAW643 19 and UmuC-mKate2 in EAW282 were imaged using yellow excitation light (λ = 568 nm) at 20 high intensity (2750 Wcm-2), collecting emitted light between 610–680 nm (ET 645/75m filter, 21 Chroma) on a 512 × 512 pixel EM-CCD camera (C9100-13, Hamamatsu). Images of UmuC- 22 mKate2 in RW1286 were recorded at 275 Wcm-2. For DinB-YPet imaging of EAW643, we 23 used green excitation (λ = 514 nm) at 160 Wcm-2 collecting light emitted between 525–555 24 nm (ET540/30m filter, Chroma). For DinB-YPet imaging of RW1594, cells were imaged at 25 51 Wcm-2. DnaX-YPet imaging (EAW282, RW1286) was performed at 51 Wcm-2. Cells 26 carrying the SOS reporter plasmid pUA66-sulA-GFP (SSH091, SSH111) were imaged at 16 27 Wcm-2. 28 Two-colour time-lapse movies were recorded to visualise if DinB-YPet foci overlap 29 with DnaX-mKate2 foci (EAW643). Sets of three images were recorded (bright-field [34 ms 30 exposure], mKate2 fluorescence [100 ms exposure], YPet fluorescence [50 ms exposure]) at 31 an interval of 10 min for 3h. To measure colocalisation between UmuC-mKate2 with the 32 replisome marker DnaX-YPet (EAW282), we recorded time-lapse movies at the same 33 intervals but different exposures for the replisome marker (bright-field [34 ms exposure], 34 mKate2 fluorescence [100 ms exposure], YPet fluorescence [500 ms exposure]). 35 Burst acquisitions of DinB-YPet (movies of 300 × 50 ms frames taken every 100 ms 36 light at 514 nm) were collected, subsequently to each burst acquisition, an image of DnaX- 37 mKate2 (568 nm) was taken (imaging sequence for RW1594). With this imaging sequence, 38 we analysed activity of DinB-YPet at replisomes. RW1286 was imaged similarly; we 39 recorded burst acquisitions of UmuC-mKate2 (568 nm) followed by a snapshot of DnaX- 40 YPet (514 nm). All images were analysed with ImageJ52.

41 Flow cell designs 42 All imaging experiments were carried out in home-built quartz-based flow cells. 43 These flow cells were assembled from a no. 1.5 coverslip (Marienfeld, REF 0102222), a 44 quartz top piece (45x20x1 mm) and PE-60 tubing (Instech Laboratories, Inc.). Prior to flow- 45 cell assembly, coverslips were silanised with (3-aminopropyl)triethoxysilane (Alfa Aeser). 46 First, coverslips were sonicated for 30 min in a 5M KOH solution to clean and activate the

1 surface. The cleaned coverslips were rinsed thoroughly with MilliQ water and then treated 2 with a 5% (v/v) solution of (3-aminopropyl)triethoxysilane (APTES) in MilliQ water. The 3 coverslips were subsequently rinsed with ethanol and sonicated in ethanol for 20 seconds.

4 Afterwards, the coverslips were rinsed with MilliQ water and dried in a jet of N2. Silanised 5 slides were stored under vacuum prior to use. 6 To assemble each flow cell, polyethylene tubing (BTPE-60, Instech Laboratories, 7 Inc.) was glued (BONDiT B-482, Reltek LLC) into two holes that were drilled into a quartz 8 piece. After the glue solidified overnight, double-sided adhesive tape was stuck on two 9 opposite sides of the quartz piece to create a channel. Then, the quartz piece was stuck to 10 an APTES-treated coverslip. The edges were sealed with epoxy glue (5 Minute Epoxy, 11 PARFIX). Each flow cell was stored in a desiccator under mild vacuum while the glue dried. 12 Typical channel dimensions were 45 mm × 5 mm × 0.1 mm (length × width × height).

13 Imaging in flow cells 14 For all imaging experiments, cells were grown at 37⁰C in EZ rich defined medium 15 (Teknova) that contained 0.2% (w/v) glucose. All strains that have a KanR cassette were 16 grown in the presence of kanamycin (20 µg/mL). Cultures used for imaging under radical- 17 scavenging conditions were grown in the presence of the particular scavenger used for the 18 experiment (DMSO [2% v/v, 282 mM, 0.2 x MIC5] or BiP [0.35 mM, 0.5 x MIC5], culture time 19 ~3h). 20 Cells were loaded into flow cells, allowed a few minutes to associate with the APTES 21 surface, then loosely associated cells were removed by pulling through fresh medium. The 22 experiment was then initiated by adding either an antibiotic alone or in combination with 23 DMSO to the medium (30 ng/ mL ciprofloxacin, 30 ng/ mL ciprofloxacin with 2% (v/v) DMSO, 24 1 µg/mL trimethoprim, 1 µg/mL trimethoprim with 2% (v/v) DMSO or 1 µg/mL trimethoprim 25 with 0.35 mM BiP). Throughout the experiment, medium was pulled through the flow cell 26 using a syringe pump, at a rate of 50 µL/min. For each condition, triplicate measurements 27 were recorded.

28 Analysis of cell filamentation, concentrations, SOS induction level and number of foci 29 We selected single cells to obtain information about SOS induction, DinB and UmuC 30 levels upon UV irradiation (>100 cells for every time point). MicrobeTracker 0.93753, a 31 MATLAB script, was used to create cell outlines as regions of interest (ROI). We manually 32 curated cell outlines designated by MicrobeTracker at t = 0 min (time point of antibiotic 33 addition) and at 30 min time intervals until 180 min. By obtaining cell outlines manually, we 34 ensure accuracy and purely select non-overlapping, in-focus cells for analysis. These ROI 35 were imported in ImageJ 1.50i. The cell outlines were then used to measure mean cell 36 intensities, cell lengths and the number of foci per cell. Parameters describing foci (number, 37 positions and intensities) were obtained using a Peak Fitter plug-in, described previously 38 28,39.

39 Analysis of colocalisation events 40 Foci were classed as colocalised if their centroid positions (determined using our 41 peak fitter tool) fell within 2.18 px (218 nm) of each other. When treating with ciprofloxacin, 42 we determined that for DinB-YPet–DnaX-mKate2 localisation the background of DinB foci 43 expected to colocalise with replisomes purely by chance is ~4% at 180 min. This was 44 calculated by taking the area of each cell occupied by replisome foci (including the 45 colocalisation search radius) and dividing by the total area of the cell. The value of 4% 46 corresponds to the mean of measurements made over 121 cells. Since the foci density of

1 replisomes stays fairly constant following ciprofloxacin treatment, the chance colocalisation 2 of DinB-YPet foci with DnaX-mKate2 is ~4% during the experiment28. Chance colocalisation 3 of DnaX-mKate2 with DinB-YPet is however not constant over time because most cells 4 contain no pol IV foci in the absence of any DNA damage. Chance colocalisation is close to 5 zero at 0 min; at 60 min, chance colocalisation is ~5%; at 120 min, chance colocalisation is 6 ~3%. Moreover, chance colocalisation of DnaX-mKate2 with DinB-YPet is overall reduced 7 under radical-scavenging conditions due to a reduced number of foci per cell (chance 8 colocalisation close to zero at 0 min; at 120 min, ~2%). Chance colocalisation of DnaX- 9 mKate2 with DinB-YPet in trimethoprim treated cells amounts to ~1% from 60-90 min (close 10 to zero before 60 min). Under radical-scavenging conditions, chance colocalisation is always 11 close to zero because the number of pol IV foci per cell does not increase post treatment as 12 well as cell size (Figure 2). 13 The chance colocalisation of UmuC-mKate2 with DnaX-YPet is similar to the chance 14 colocalisation of DinB-YPet with DnaX-mKate2 (chance colocalisation: ~4%). The expected 15 colocalisation of DnaX-YPet with UmuC-mKate2 by background is close to zero until 90 min. 16 UmuC-mKate2 is neither upregulated nor released from the membrane (Supplementary 17 figure 8a). Chance colocalisation is ~3% at 180 min after ciprofloxacin treatment and ~2% 18 after the combinational treatment of ciprofloxacin/DMSO.

19 Western blotting 20 Overnight E. coli LB cultures of RW120/pRW154 and RW546/pRW15450 were diluted 21 1 to 100 in fresh LB with appropriate antibiotics and grown to mid-log (~OD 0.5, ~3 hrs). 22 Aliquots were then taken for the untreated samples. Either ciprofloxacin (30 ng/mL) or 23 trimethoprim (1 µg/mL) was added to the remaining culture and incubated with or without the 24 addition of 2% DMSO. Samples were taken at 1, 2 and 3 hours. Whole cell extracts were 25 made by centrifuging 1.5 mL of culture and adding 90 µl of sterile deionized water and 30µL 26 of NuPAGE LDS sample buffer (4X) (Novex, Life Technologies) to the cell pellet. Five cycles 27 of freeze/thaw on dry ice and in a 37⁰C water bath were performed to lyse the cells. Extracts 28 were boiled for 5 minutes prior to loading. Samples were run on NuPAGE 4-12% Bis-Tris 29 gels (Novex Life Technologies) and transferred to Invitrolon PVDF (0.45 µm pore size) 30 membranes (Novex Life Technologies). Membranes were incubated with anti-UmuD 31 antibodies (1:5,000 dilution) at room temperature overnight. Then the membranes were 32 incubated with goat anti-rabbit IgG (H+L) alkaline phosphatase conjugate (1:10,000 dilution) 33 (BIO-RAD). Subsequently, the membranes were treated with the CDP-Star substrate 34 (Tropix). Membranes were then exposed to BioMax XAR film (Carestream) to visualise 35 UmuD protein bands.

36 37 References

38 1. Dwyer, D. J., Collins, J. J. & Walker, G. C. Unraveling the physiological complexities 39 of antibiotic lethality. Annu. Rev. Pharmacol. Toxicol. 55, 313–332 (2015).

40 2. van Acker, H. & Coenye, T. The role of reactive oxygen species in antibiotic-mediated 41 killing of bacteria. Trends Microbiol. 25, 456–466 (2017).

42 3. Dwyer, D. J., Kohanski, M. A. & Collins, J. J. Role of reactive oxygen species in 43 mechanisms of antimicrobial resistance in bacteria. Curr. Opin. Microbiol. 12, 482– 44 489 (2009).

1 4. Gruber, C. C. & Walker, G. C. Incomplete base excision repair contributes to cell 2 death from antibiotics and other stresses. DNA Repair. 71, 108–117 (2018).

3 5. Hong, Y., Li, L., Luan, G., Drlica, K. & Zhao, X. Contribution of reactive oxygen 4 species to thymineless death in Escherichia coli. Nat. Microbiol. 2, 1667–1675 (2017).

5 6. Moore, J. M., Correa, R., Rosenberg, S. M. & Hastings, P. J. Persistent damaged 6 bases in DNA allow mutagenic break repair in Escherichia coli. PLoS Genet. 13, 7 e1006733 (2017).

8 7. Foti, J. J., Devadoss, B., Winkler, J. A., Collins, J. J. & Walker, G. C. Oxidation of the 9 guinine nucleotide pool underlies cell death by bactericidial antibiotics. Science. 336, 10 315–319 (2012).

11 8. Sekiguchi, M. & Tsuzuki, T. Oxidative nucleotide damage: Consequences and 12 prevention. Oncogene 21, 8895–8904 (2002).

13 9. Ponder, R. G., Fonville, N. C. & Rosenberg, S. M. A switch from high-fidelity to error- 14 prone DNA double-strand break repair underlies stress-induced mutation. Mol. Cell 15 19, 791–804 (2005).

16 10. Mallik, S., Popodi, E. M., Hanson, A. J. & Foster, P. L. Interactions and localization of 17 Escherichia coli error-prone DNA polymerase IV after DNA damage. J. Bacteriol. 18 197, 2792–2809 (2015).

19 11. Motamedi, M. R., Szigety, S. K. & Rosenberg, S. M. Double-strand-break repair 20 recombination in Escherichia coli: physical evidence for a DNA replication 21 mechanisms in vivo. Genes Dev. 13, 2889–2903 (1999).

22 12. Shee, C., Gibson, J. L., Darrow, M. C., Gonzalez, C. & Rosenberg, S. M. Impact of a 23 stress-inducible switch to mutagenic repair of DNA breaks on mutation in Escherichia 24 coli. Proc. Natl. Acad. Sci. U. S. A. 108, 13659–13664 (2011).

25 13. Rosenberg, S. M., Shee, C., Frisch, R. L. & Hastings, P. J. Stress-induced mutation 26 via DNA breaks in Escherichia coli: A molecular mechanism with implications for 27 evolution and medicine. Bioessays 34, 885–892 (2012).

28 14. Shee, C., Gibson, J. L. & Rosenberg, S. M. Two mechanisms produce mutation 29 hotspots at DNA breaks in Escherichia coli. Cell Rep. 2, 714–721 (2012).

30 15. Frisch, R. L. et al. Separate DNA Pol II- and Pol IV-dependent pathways of stress- 31 induced mutation during double-strand-break repair in Escherichia coli are controlled 32 by RpoS. J. Bacteriol. 192, 4694–4700 (2010).

33 16. Giroux, X., Su, W.-L., Bredeche, M.-F. & Matic, I. Maladaptive DNA repair is the 34 ultimate contributor to the death of trimethoprim-treated cells under aerobic and 35 anaerobic conditions. Proc. Natl. Acad. Sci. U. S. A. 114, 11512–11517 (2017).

36 17. Zhao, X. & Drlica, K. Reactive oxygen species and the bacterial response to lethal 37 stress. Curr. Opin. Microbiol. 21, 1–6 (2014).

38 18. Kaur, P. et al. Involvement of reactive oxygen species in the action of ciprofloxacin 39 against Escherichia coli. Future Microbiol. 6, 949–954 (2014).

40 19. Yamada, M. et al. Involvement of Y-family DNA polymerases in mutagenesis caused 41 by oxidized nucleotides in Escherichia coli. J. Bacteriol. 188, 4992–4995 (2006).

42 20. Indiani, C., Patel, M., Goodman, M. F. & O’Donnell, M. E. RecA acts as a switch to

1 regulate polymerase occupancy in a moving replication fork. Proc. Natl. Acad. Sci. U. 2 S. A. 110, 5410–5415 (2013).

3 21. Cafarelli, T. M., Rands, T. J. & Godoy, V. G. The DinB·RecA complex of Escherichia 4 coli mediates an efficient and high-fidelity response to ubiquitous alkylation lesions. 5 Environ. Mol. Mutagen. Mutagen. 55, 92–102 (2014).

6 22. Shee, C., Ponder, R., Gibson, J. L. & Rosenberg, S. M. What limits the efficiency of 7 double-strand break-dependent stress-induced mutation in Escherichia coli? J. Mol. 8 Microbiol. Biotechnol. 21, 8–19 (2012).

9 23. Foster, P. L. Stress-induced mutagenesis in bacteria. Crit. Rev. Biochem. Mol. Biol. 10 42, 373–397 (2007).

11 24. Godoy, V. G., Jarosz, D. F., Simon, S. M., Abyzov, A. & Walker, G. C. UmuD and 12 RecA directly modulate the mutagenic potential of the Y-family DNA polymerase 13 DinB. Mol. Cell 28, 1058–1070 (2007).

14 25. Pomerantz, R. T., Kurth, I., Goodman, M. F. & O’Donnell, M. Preferential D-loop 15 extension by a translesion DNA polymerase underlies error-prone recombination. Nat. 16 Struct. Mol. Biol. 20, 748–755 (2013).

17 26. Pomerantz, R. T., Goodman, M. F. & O’Donnell, M. E. DNA polymerases are error- 18 prone at RecA-mediated recombination intermediates. Cell Cycle 12, 2558–2563 19 (2013).

20 27. Courcelle, J., Khodursky, A., Peter, B., Brown, P. O. & Hanawalt, P. C. Comparative 21 gene expression profiles following UV exposure in wild-type and SOS-deficient. 22 Genetics 158, 41–64 (2001).

23 28. Henrikus, S. S. et al. DNA polymerase IV primarily operates outside of DNA 24 replication forks in Escherichia coli. PLoS Genet. 14, e1007161 (2018).

25 29. Vlašić, I., Ivančić-Baće, I., Imešek, M., Mihaljević, B. & Brčić-Kostić, K. RecJ nuclease 26 is required for SOS induction after introduction of a double-strand break in a RecA 27 loading deficient recB mutant of Escherichia coli. Biochimie 90, 1347–1355 (2008).

28 30. McPartland, A., Green, L. & Echols, H. Control of recA gene RNA in E. coli: 29 Regulatory and signal genes. Cell 20, 731–737 (1980).

30 31. Newmark, K. G., O’Reilley, E. K., Pohlhaus, J. R. & Kreuzer, K. N. Genetic analysis of 31 the requirements for SOS indcution by nalidixic acid in Escherichia coli. Gene 356, 32 69–76 (2005).

33 32. Zhao, X., Xu, C., Domagala, J. & Drlica, K. DNA topoisomerase targets of the 34 fluoroquinolones: a strategy for avoiding bacterial resistance. Proc. Natl. Acad. Sci. U. 35 S. A. 94, 13991–13996 (1997).

36 33. Drlica, K. & Zhao, X. DNA gyrase, topoisomerase IV, and the 4-quinolones. Microbiol. 37 Mol. Biol. Rev. 61, 377–392 (1997).

38 34. Birdsall, B., Roberts, G. C., Feeney, J., Dann, J. & Burgen, A. Trimethoprim binding to 39 bacterial and mammalian dihydrofolate reductase: a comparison by proton and 40 carbon-13 nuclear magnetic resonance. Biochemistry 22, 5597–5604 (1983).

41 35. Zheng, Q., Jockusch, S., Zhou, Z. & Blanchard, S. C. The contribution of reactive 42 oxygen species to the photobleaching of organic fluorophores. Photochem. Photobiol. 43 90, 448–454 (2014).

1 36. Champoux, J. J. DNA topoisomerases: structure, function, and mechanism. Annu. 2 Rev. Biochem. 70, 369–413 (2001).

3 37. Drlica, K., Malik, M., Kerns, R. J. & Zhao, X. Quinolone-mediated bacterial death. 4 Antimicrob. Agents Chemother. 52, 385–392 (2008).

5 38. Mi, H. et al. Dimethyl sulfoxide protects Escherichia coli from rapid antimicrobial- 6 mediated killing. Antimicrob. Agents Chemother. 60, 5054–5058 (2016).

7 39. Robinson, A. et al. Regulation of mutagenic DNA polymerase V activation in space 8 and time. PLoS Genet. 11, e1005482 (2015).

9 40. Li, G.-W. & X. Sunney Xie. Central dogma at the single-molecule level of living cells. 10 Nature 475, 308–315 (2011).

11 41. Zaslaver, A. et al. A comprehensive library of fluorescent transcriptional reporters for 12 Escherichia coli. Nat. Methods 3, 623–628 (2006).

13 42. Shen, L. L. & Pernet, A. G. Mechanism of inhibition of DNA gyrase by analogues of 14 nalidixic acid: The target of the drugs is DNA. Proc. Natl. Acad. Sci. U. S. A. 82, 307– 15 311 (1985).

16 43. Dorr, T., Lewis, K. & Vulic, M. SOS response induces persistence to fluoroquinolones 17 in Escherichia coli. PLoS Genet. 5, e1000760 (2009).

18 44. Ennis, D. G., Amunsden, S. K. & Smith, G. R. Genetic functions promoting 19 homologous recombination in Escherichia coli: A study of inversions in phage λ. 20 Genetics 115, 11–24 (1987).

21 45. Gleckman, R., Blagg, N. & Joubert, D. W. Trimethoprim: Mechanisms of action, 22 antimicrobial activity, bacterial resistance, pharmacokinetics, adverse reactions, and 23 therapeutic indications. Pharmacotherapy 1, 14–19 (1981).

24 46. Lewin, C. S. & T, S. J. DNA breakdown by the 4-quinolones and its significance. J. 25 Med Microbiol 31, 65–70 (2018).

26 47. Huang, L. C., Wood, E. A. & Cox, M. M. Convenient and reversible site-specific 27 targeting of exogenous DNA into a bacterial chromosome by use of the FLP 28 recombinase: The FLIRT system. J. Bacteriol. 179, 6076–6083 (1997).

29 48. Frank, E. G., Hauser, J., Levine, A. S. & Woodgate, R. Targeting the UmuD, UmuDʹ, 30 and MucAʹ mutagenesis proteins to DNA by RecA protein. Proc. Natl. Acad. Sci. U. S. 31 A. 90, 8169–8173 (1993).

32 49. Blattner, F. R. et al. The complete genome sequence of Escherichia coli K-12. 33 Science. 277, 1453–1474 (1997).

34 50. Ho, C., Kulaeva, O. I., Levine, A. S. & Woodgate, R. A rapid method for cloning 35 mutagenic DNA repair genes: Isolation of umu-complementing genes from multidrug 36 resistance plasmids R391, R446b, and R471a. J. Bacteriol. 175, 5411–5419 (1993).

37 51. Rangarajan, S., Woodgate, R. & Goodman, M. F. Replication restart in UV-irradiated 38 Escherichia coli involving pols II, III, V, PriA, RecA and RecFOR proteins. Mol. 39 Microbiol. 43, 617–628 (2002).

40 52. Schneider, C. A., Rasband, W. S. & Eliceiri, K. W. NIH Image to ImageJ: 25 years of 41 image analysis. Nat. Methods 9, 671–675 (2012).

1 53. Sliusarenko, O., Heinritz, J., Emonet, T. & Jacobs-Wagner, C. High-throughput, 2 subpixel-precision analysis of bacterial morphogenesis and intracellular spatio- 3 temporal dynamics. Mol. Microbiol. 80, 612–627 (2012).

5 Acknowledgements

6 MMC was supported by grant GM32335 from the National Institute of General Medical 7 Sciences USA. RW was supported by funds from the National Institutes of Health, National 8 Institute of Child Health and Human Development Intramural Research Program. AvO was 9 supported by a Laureate Fellowship FL140100027 from the Australian Research Council.

10 Figure captions

11 Figure 1. Effects of ciprofloxacin and trimethoprim on live cells. 12 a, Ciprofloxacin inhibits DNA gyrase and forms a gyrase-stabilised cleavage complex 13 which might inhibit replication forming single-stranded gaps. b, In addition, 14 ciprofloxacin treatment leads to the accumulation of reactive oxygen species (ROS). 15 ROS oxidise the nucleotide pool; pol IV can incorporate 8-oxo-dGTP. Following base- 16 excision repair (BER), single-stranded gaps (ssDNA) are created. The excision of 17 closely spaced oxidised nucleotides leads to the formation of double-strand breaks 18 (DSB). Following trimethoprim treatment, replication is inhibited which creates single- 19 stranded regions. Similar to ciprofloxacin treatment, trimethoprim treatment interferes 20 with the cellular metabolism and leads to ROS accumulation. Overall, ROS frequently 21 convert single-stranded regions into DSBs. Ultimately, DNA breakage triggers the 22 upregulation of pol IV, consequently, potentiating the effect of ROS within cells.

23 Figure 2. Pol IV concentration and activity following ciprofloxacin or trimethoprim 24 treatment under oxidised or radical-scavenging conditions 25 a, Fluorescence images showing cells expressing DinB-YPet (Pol IV) at 0, 90 and 26 180 min (left to right) after ciprofloxacin-alone, ciprofloxacin-DMSO, trimethoprim- 27 alone or trimethoprim-DMSO treatment (top to bottom). Scale bar represents 5 µm. b, 28 Concentration of DinB-YPet during stress. Mean cell brightness is plotted against time 29 (ciprofloxacin-alone: dark grey line, ciprofloxacin-DMSO: light grey line, trimethoprim- 30 alone: magenta line, trimethoprim-DMSO: light magenta line). At each time-point, data 31 are derived from >100 cells. Error bars represent standard error of the mean. c, 32 Number of DinB-YPet foci per cell are plotted against time (ciprofloxacin-alone: dark 33 grey line, ciprofloxacin-DMSO: light grey line, trimethoprim-alone: magenta line, 34 trimethoprim-DMSO: light magenta line). At each time-point, data are derived from 35 >100 cells. Error bars represent standard error of the mean.

36 Figure 3. SulA expression levels following ciprofloxacin or trimethoprim treatment 37 under oxidised or radical-scavenging conditions in different genetic 38 backgrounds 39 a, Fluorescence images showing the expression of GFP from a SOS reporter 40 plasmid (PsulA-GFP) at 0, 60, 120 and 180 min (left to right) after ciprofloxacin-alone, 41 ciprofloxacin-DMSO, ciprofloxacin-alone in ΔrecB, trimethoprim-alone, trimethoprim-

1 DMSO or trimethoprim-BiP treatment (top to bottom). Scale bar represents 5 µm. b, 2 SulA expression levels during stress. Mean cell brightness is plotted against time 3 (ciprofloxacin-alone: dark grey line, ciprofloxacin-DMSO: light grey line, ciprofloxacin 4 in ΔrecB: purple, dotted line, trimethoprim-alone: magenta line, trimethoprim-DMSO: 5 light magenta line, trimethoprim-BiP: rose-coloured, dashed line). Error bars represent 6 standard error of the mean.

7 Figure 4. Number of pol IV foci per cell in lexA(Def) cells following ciprofloxacin or 8 trimethoprim treatment under oxidised or radical-scavenging conditions 9 a, Upper row: Average projection in time (100 ms x 10 frames) showing DinB-YPet 10 (pol IV) foci. Bottom row: Discoidal filtered projections. Cells were treated for 60 min 11 prior to imaging. b, Number of DinB-YPet foci per cell. Error bars represent standard 12 error of the mean. Number of cells included in analysis: n(ciprofloxacin) = 106, 13 n(ciprofloxacin-DMSO) = 109, n(ciprofloxacin in ΔrecB) = 106, n(trimethoprim) = 145, 14 n(trimethoprim-DMSO) = 102, n(trimethoprim in ΔrecB) = 94, n(untreated recB+) = 85, 15 n(untreated ΔrecB) = 99.

16 Supplementary Figure 1. Characterising functionality of fluorescent fusion proteins. 17 a, Plate-based survival assays using ciprofloxacin or trimethoprim under oxidised or 18 radical-scavenging conditions. Cell cultures (WT, dnaX-mKate2, dinB-YPet, dinB- 19 YPet dnaX-mKate2, ΔdinB, dnaX-YPet, umuC-mKate2, umuC-mKate2 dnaX-YPet

20 and ΔumuDC) were grown to exponential growth phase (OD600 = 0.2), serial diluted, 21 and spotted onto LB agar plates +/- DMSO, plates containing 12 ng/mL ciprofloxacin 22 +/- DMSO and plates containing 0.1 ng/mL trimethoprim +/- DMSO. Under antibiotic 23 treatment, the addition of DMSO rescues survival by one to two orders of magnitude. 24 Cell constructs used in this study (dinB-YPet dnaX-mKate2 and umuC-mKate2 dnaX- 25 YPet) exhibit a similar phenotype to WT. b, Trimethoprim survival assay. Cell cultures 26 (WT, ΔdinB, ΔumuDC, dinB-YPet dnaX-mKate2 and umuC-mKate2 dnaX-YPet) were

27 grown to exponential growth phase (OD600 = 0.3). Then, trimethoprim was added to 28 1 µg/mL. At each time point (0, 0.5, 1, 2, 3 and 5 h), 500 µL culture was taken, twice 29 centrifuged and resuspended in 500 µL LB. Finally, 50 µL were plated on LB agar 30 plates; three technical replicates were made. Plates were incubated at 37⁰C in the 31 dark. Colonies were counted and CFU/mL were calculated for each time point as an 32 average over the technical replicates. Error bars represent standard error of the mean 33 over two biological replicates.

34 35 Supplementary Figure 2. Plate-based survival assays using ciprofloxacin or 36 trimethoprim under oxidised or radical-scavenging conditions.

37 a, Strains (WT, ΔdinB and ΔumuDC) were grown to exponential growth phase (OD600 38 = 0.2), serial diluted, and spotted onto LB agar plates containing no antibiotic (LB), 12 39 ng/mL of ciprofloxacin or 12 ng/mL of ciprofloxacin plus 2% v/v DMSO. The addition 40 of DMSO rescues survival by two orders of magnitude relatively to ciprofloxacin 41 treatment. b, Replicate of a. c, Strains (WT, ΔdinB and ΔumuDC) were grown to

42 exponential growth phase (OD600 = 0.2), serial diluted, and spotted onto LB agar 43 plates containing no antibiotic (LB), 0.1 µg/mL of trimethoprim or 0.1 µg/mL of 44 trimethoprim plus 2% v/v DMSO. The addition of DMSO rescues survival by two 45 orders of magnitude relatively to ciprofloxacin treatment. Cells lacking dinB also have 46 increased survival upon trimethoprim treatment. d, Strains (WT, ΔdinB and ΔumuDC)

1 were grown to exponential growth phase (OD600 = 0.2), serial diluted, and spotted 2 onto LB agar plates containing no antibiotic (LB), 0.2 µg/mL of trimethoprim or 3 0.2 µg/mL of trimethoprim plus 2% v/v DMSO. The addition of DMSO rescues survival 4 by two to three orders of magnitude relatively to ciprofloxacin treatment. Deleting dinB 5 also increases survival upon trimethoprim treatment.

6 Supplementary Figure 3. Scatter plots of cell-size from time-lapse imaging. White 7 points indicate individual data-points, while blue-to-red contours indicate 8 frequencies of observations. Blue areas indicate regions of the plot containing 9 few data points; red areas indicate regions containing a large number of data 10 points. Frequencies were normalised at each time-point to the maximum value 11 at that time-point.

12 a, EAW643 cells (dinB-YPet dnaX-mKate2) treated with ciprofloxacin-alone. b, 13 EAW643 cells (dinB-YPet dnaX-mKate2) treated with ciprofloxacin-DMSO. c, 14 EAW643 cells (dinB-YPet dnaX-mKate2) treated with trimethoprim-alone. d, EAW643 15 cells (dinB-YPet dnaX-mKate2) treated with trimethoprim-DMSO. e, SSH091 cells 16 (MG1655 PsulA-GFP) treated with ciprofloxacin-alone. f, SSH091 cells (MG1655 17 PsulA-GFP) treated with ciprofloxacin-DMSO. g, SSH091 cells (MG1655 PsulA-GFP) 18 treated with trimethoprim-alone. h, SSH091 cells (MG1655 PsulA-GFP) treated with 19 trimethoprim-DMSO. i, SSH091 cells (MG1655 PsulA-GFP) treated with trimethoprim- 20 BiP. j, SSH111 cells (ΔrecB PsulA-GFP) treated with ciprofloxacin. We conservatively 21 estimate that >100 cells were used in each measurement.

22 Supplementary Figure 4. SulA expression levels following ciprofloxacin-alone 23 treatment in ΔrecB vs. MG1655 (wild-type) 24 a, Fluorescence images showing the expression of GFP from a SOS reporter plasmid 25 (PsulA-GFP) from 0-110 min at intervals of 10 min and 120 min after ciprofloxacin 26 addition in ΔrecB. Scale bar represents 5 µm. b, Fluorescence images showing the 27 expression of GFP from a SOS reporter plasmid (PsulA-GFP) at 120 min after 28 ciprofloxacin addition in wild-type cells, MG1655. Scale bar represents 5 µm. 29 30 Supplementary Figure 5. Measuring colocalisation of pol IV with replisomes following 31 ciprofloxacin or trimethoprim treatment under oxidised or radical-scavenging 32 conditions 33 a, Merged images showing DinB-YPet (pol IV) foci in green and DnaX-mKate2 34 (replisome) foci in magenta at 0, 90 and 180 min (left to right) for ciprofloxacin-alone 35 or ciprofloxacin-DMSO treatment (top to bottom). White arrows indicate colocalization 36 events (white foci). Scale bar represents 5 µm. b, Merged images showing DinB-YPet 37 (pol IV) foci in green and DnaX-mKate2 (replisome) foci in magenta at 0, 90 and 180 38 min (left to right) for trimethoprim-alone or trimethoprim-DMSO treatment (top to 39 bottom). White arrows indicate colocalization events (white foci). Scale bar represents 40 5 µm. c, Colocalisation measurements following ciprofloxacin-alone treatment over 41 180 min: percentage of pol IV foci that are bound at replisomes (green line), 42 percentage of replisomes that contain a pol IV focus (magenta line). Error bars 43 represent the standard error of the mean from six biological replicates together. 44 Measurements are from >300 cells per time point. d, Colocalisation measurements 45 following ciprofloxacin-DMSO treatment over 180 min: percentage of pol IV foci that 46 are bound at replisomes (green line), percentage of replisomes that contain a pol IV

1 focus (magenta line). Error bars represent the standard error of the mean from four 2 biological replicates together. Measurements are from >100 cells per time point. e, 3 Colocalisation measurements following trimethoprim-alone treatment over 180 min: 4 percentage of pol IV foci that are bound at replisomes (green line), percentage of 5 replisomes that contain a pol IV focus (magenta line). Error bars represent the 6 standard error of the mean from three biological replicates together. Measurements 7 are from >100 cells per time point. f, Colocalisation measurements following 8 trimethoprim-DMSO treatment over 180 min: percentage of pol IV foci that are bound 9 at replisomes (green line), percentage of replisomes that contain a pol IV focus 10 (magenta line). Error bars represent the standard error of the mean from three 11 biological replicates together. Measurements are from >100 cells per time point.

12 Supplementary Figure 6. Measuring the colocalisation of pol IV and replisomes 13 following ciprofloxacin or trimethoprim treatment under oxidised or radical- 14 scavenging conditions in lexA(Def) cells 15 a, DinB-YPet activity at replisomes in lexA(Def) cells. Cells were treated for 60 min 16 prior to imaging. Merged images showing DinB-YPet foci in green and DnaX-mKate2 17 foci in magenta following ciprofloxacin-alone, ciprofloxacin-DMSO, trimethoprim-alone 18 and trimethoprim-DMSO treatment (from left to right). White arrow points at 19 colocalisation event (white focus). Scale bar represents 5 µm. b, Colocalisation 20 percentages of pol IV foci that bind at replisomes (green bars) and colocalisation 21 percentages of replisomes that contain a pol IV focus (magenta bars) for cells treated 22 with ciprofloxacin-alone (left) or ciprofloxacin-DMSO (right). Colocalisation was 23 measured with sets of pol IV foci that last 1, 3, 5 and 10 s. Error bars represent the 24 standard error of the mean. c, Colocalisation percentages of pol IV foci that bind at 25 replisomes (green bars) and colocalisation percentages of replisomes that contain a 26 pol IV focus (magenta bars) for cells treated with trimethoprim-alone (left) or 27 trimehtoprim-DMSO (right). Colocalisation was measured with sets of pol IV foci that 28 last 1, 3, 5 and 10 s. Error bars represent the standard error of the mean.

29 Supplementary Figure 7. Measuring the number of pol V foci per cell following 30 ciprofloxacin or trimethoprim treatment under oxidised or radical-scavenging 31 conditions in lexA(Def) cells 32 a, UmuC-mKate2 activity at replisomes in lexA(Def) cells. Cells were treated for 60 33 min prior to imaging. Upper row: unfiltered image of an average projection showing 34 UmuC-mKate2 foci that last >1s (from left to right: ciprofloxacin, ciprofloxacin-DMSO, 35 trimethoprim, trimethoprim-DMSO). Bottom row: merged image showing UmuC- 36 mKate2 foci in magenta and DnaX-YPet foci in green (from left to right: ciprofloxacin, 37 ciprofloxacin-DMSO, trimethoprim, trimethoprim-DMSO). Scale bar represents 5 µm. 38 b, Number of UmuC-mKate2 foci per cell. Error bars represent standard error of the 39 mean. Number of cells included in analysis: n(ciprofloxacin) = 97, n(ciprofloxacin- 40 DMSO) = 109, n(untreated) = 87. c, Binding behaviour of UmuC-mKate2 at 41 replisomes after ciprofloxacin-alone or ciprofloxacin-DMSO treatment. Mean average 42 autocorrelation function (ciprofloxacin-alone: dark grey line, ciprofloxacin-DMSO: light 43 grey line). Error bar represents standard error of the mean. We conservatively 44 estimate that >400 trajectories from >400 replisomes were used in each 45 measurement. d, Number of UmuC-mKate2 foci per cell. Error bars represent 46 standard error of the mean. Number of cells included in analysis: n(trimethoprim) =

1 102, n(trimethoprim-DMSO) = 120, n(untreated) = 87. e, Binding behaviour of UmuC- 2 mKate2 at replisomes after trimethoprim-alone or trimethoprim-DMSO treatment. 3 Mean average autocorrelation function (trimethoprim-alone: magenta line, 4 trimethoprim-DMSO: light magenta line). Error bar represents standard error of the 5 mean. We conservatively estimate that >550 trajectories from >550 replisomes were 6 used in each measurement.

7 Supplementary Figure 8. UmuC concentration and activity following ciprofloxacin or 8 trimethoprim treatment under oxidised or radical-scavenging conditions 9 a, Images showing UmuC-mKate2 (pol V) signal at 0, 90 and 180 min (left to right) for 10 ciprofloxacin-alone, ciprofloxacin-DMSO treatment or trimethoprim-alone treatment 11 (top to bottom). Scale bar represents 5 µm. b, Merged images showing UmuC- 12 mKate2 (pol V) foci in magenta and DnaX-YPet (replisome) foci in magenta at 0, 90 13 and 180 min (left to right). Colocalised foci would appear as white foci. Scale bar 14 represents 5 µm. c, Concentration of UmuC-mKate2 during stress. Mean cell 15 brightness is plotted against time (ciprofloxacin-alone: dark grey line, ciprofloxacin- 16 DMSO: light grey line, trimethoprim-alone: magenta line). At each time-point, data are 17 derived from >100 cells. Error bars represent standard error of the mean. d, 18 Colocalisation measurements following ciprofloxacin-alone treatment over 180 min: 19 percentage of UmuC foci that are bound at replisomes (magenta line), percentage of 20 replisomes that contain a UmuC focus (green line). Error bars represent the standard 21 error of the mean from three biological replicates together. Measurements are from 22 >100 cells per time point. e, Colocalisation measurements following ciprofloxacin- 23 DMSO treatment over 180 min: percentage of UmuC foci that are bound at 24 replisomes (magenta line), percentage of replisomes that contain a UmuC focus 25 (green line). Error bars represent the standard error of the mean from three biological 26 replicates together. Measurements are from >100 cells per time point.

27 Supplementary Figure 9. Western blots with anti-UmuD antibodies measuring levels

28 of UmuDʹ. For each lane, 30 µL of lysate were loaded from cultures at OD600 0.5. 29 All strains used are ΔumuDC expressing UmuDC from a low-copy number 30 plasmid (pRW154). After treatment, time points were taken at 1, 2, 3h. Rabbit 31 anti-UmuD antibody 1: 32 a, Western blot of recA+ lexA+ cells (RW120): untreated, treated with ciprofloxacin or 33 ciprofloxacin + 2% DMSO. b, Western blot of recA+ lexA+ cells: untreated, treated with 34 trimethoprim or trimethoprim + 2% DMSO. c, Western blot of recA+ lexA51(Def) cells 35 (RW546): untreated, treated with ciprofloxacin or ciprofloxacin + 2% DMSO. d, 36 Western blot of recA+ lexA51(Def) cells: untreated, treated with trimethoprim or 37 trimethoprim + 2% DMSO.

38 Supplementary Movie 1. Time-lapse movie of ΔrecB cells carrying PsulA-GFP. 39 Ciprofloxacin (30 ng/mL) was added to the media ay t = 0 min. An image was 40 taken every 10 min over the period of 3 h. Upper movie: bright-field, bottom 41 movie: signal from GFP expression (level of SOS induction).

42 Supplementary Movie 2. Burst acquisition movies of DinB-YPet in recB+ cells treated 43 with 1 µg/mL trimethoprim (TMP, ± DMSO) or 30 ng/mL ciprofloxacin (± DMSO), 44 ΔrecB cells treated with TMP or ciprofloxacin. Movies were recorded 60 min 45 after antibiotic addition. Frames were taken every 0.1 s.

Figure 1

A trimethoprim trimethoprim B ciprofloxacin ciprofloxacin ciprofloxacin

pol IV ROS pol IV BER gyrase-stabilised single-strand product of cleavage complex gap 8-oxo-dGTP incorporation pol IV release ROS BER

double-strand break double-strand break

SOS death SOS death

(error-prone) (error-prone) pol IV pol IV break repair break repair bioRxiv preprint doi: https://doi.org/10.1101/533422; this version posted January 29, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Figure 2 A Fluorescence images: B DinB-YPet concentration Signal DinB-YPet (Pol IV): n(cells) > 100 2000 ciprofloxacin 0 min 90 min 180 min ciprofloxacin + DMSO 1800 trimethoprim trimethoprim + DMSO 1600

1400 ciprofloxacin

1200

Mean cell brightness (a. u.) 1000 + DMSO 800 ciprofloxacin 0 60 120 180 Time (min) C DinB-YPet activity: focus formation 6 n(cells) > 100

trimethoprim ciprofloxacin ciprofloxacin + DMSO 5 trimethoprim trimethoprim + DMSO 4 + 3 DMSO trimethoprim 2

1 Number of DinB-YPet foci per cell 0 0 60 120 180 Time (min) bioRxiv preprint doi: https://doi.org/10.1101/533422; this version posted January 29, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Figure 3 Fluorescence images: SulA expression levels: A SOS reporter plasmid (PsulA-GFP): B oxygen scavangers inhibit SOS induction 0 min 60 min 120 min 180 min 8000 ciprofloxacin ciprofloxacin + DMSO 7000 ciprofloxacin in ∆recB ciprofloxacin trimethoprim 6000 trimethoprim + DMSO + trimethoprim + BiP DMSO

ciprofloxacin 5000

4000 in ∆ recB ciprofloxacin

3000 Mean cell brightness (a. u.) trimethoprim 2000 +

DMSO 1000 trimethoprim

0 +

BiP 0 60 120 180

trimethoprim Time (min) Figure 4 A DinB-YPet activity in lexA(Def) Ciprofloxacin Trimethoprim

bioRxiv preprint doi: https://doi.org/10.1101/533422; this version posted January 29, 2019. The copyright holder for this preprint (which was not certified- byDMSO peer review) is the author/funder. All rights reserved.+ DMSONo reuse allowed without permission. ∆recB - DMSO + DMSO ∆recB 100 ms x 10 Average projection Average filtered

5 µm B Mean DinB foci per cell 3.0

2.0 + cipro + DMSO in recB in ∆ recB in ∆ recB + DMSO

1.0 in ∆ recB no damage no damage trimethoprim trimethoprim ciprofloxacin ciprofloxacin trimethoprim

DinB foci per cell 0 bioRxiv preprint doi: https://doi.org/10.1101/533422; this version posted January 29, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Supplementary Figure 1

A Plate-based survival assay B Survival assay over time No Stress Ciprofloxacin Trimethoprim Trimethoprim

No DMSO + 2% DMSO No DMSO + 2% DMSO No DMSO + 2% DMSO 1 µg/mL 0 -1 -2 -3 -4 -5 0 -1 -2 -3 -4 -5 0 -1 -2 -3 -4 -5 0 -1 -2 -3 -4 -5 0 -1 -2 -3 -4 -5 0 -1 -2 -3 -4 -5 1 WT dnaX-mKate2 dinB-YPet dinB-YPet dnaX-mKate2 0.1 ∆dinB Relative CFU/mL WT ∆dinB WT ∆umuDC Relative CFU/mL dnaX-YPet dinB-YPet dnaX-mKate2 umuC-mKate2 dnaX-YPet umuC-mKate2 0.01 umuC-mKate2 dinB-YPet 0 1 2 3 4 5 ∆umuDC Time (h) bioRxiv preprint doi: https://doi.org/10.1101/533422; this version posted January 29, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Supplementary Figure 2

LB 12 ng/mL ciprofloxacin 12 ng/mL ciprofloxacin + DMSO A 0 10-1 10-2 10-3 10-4 10-5 0 10-1 10-2 10-3 10-4 10-5 0 10-1 10-2 10-3 10-4 10-5 WT

∆dinB

∆umuDC

LB 12 ng/mL ciprofloxacin 12 ng/mL ciprofloxacin + DMSO B 0 10-1 10-2 10-3 10-4 10-5 0 10-1 10-2 10-3 10-4 10-5 0 10-1 10-2 10-3 10-4 10-5 WT

∆dinB

∆umuDC

LB 0.1 µg/mL trimethoprim 0.1 µg/mL trimethoprim + DMSO C 0 10-1 10-2 10-3 10-4 10-5 0 10-1 10-2 10-3 10-4 10-5 0 10-1 10-2 10-3 10-4 10-5 WT

∆dinB

∆umuDC

LB 0.2 µg/mL trimethoprim 0.2 µg/mL trimethoprim + DMSO D 0 10-1 10-2 10-3 10-4 10-5 0 10-1 10-2 10-3 10-4 10-5 0 10-1 10-2 10-3 10-4 10-5 WT

∆dinB

∆umuDC bioRxiv preprint doi: https://doi.org/10.1101/533422; this version posted January 29, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Supplementary Figure 3 ) ABCD ciprofloxacin ciprofloxacin + DMSO trimethoprim trimethoprim + DMSO 40 40 40 40

μ m) 30 30 30 30

20 20 20 20 EAW643

10 10 10 10

Cell length ( 0 0 0 0 0 3060 90120 150 180 0 3060 90120 150 180 0 3060 90120 150 180 0 3060 90120 150 180 ( dinB-YPet dnaX-mKate2 Time (min) Time (min) Time (min) Time (min) EFGH I ciprofloxacin ciprofloxacin + DMSO trimethoprim trimethoprim + DMSO trimethoprim + BiP 40 40 40 40 40

μ m) 30 30 30 30 30

20 20 20 20 20 SSH091

10 10 10 10 10

Cell length ( 0 0 0

(MG1655 P sulA -GFP) 0 0 0 3060 90120 150 180 0 3060 90120 150 180 0 3060 90120 150 180 0 3060 90120 150 180 0 3060 90120 150 180 J Time (min) Time (min) Time (min) Time (min) Time (min) ciprofloxacin in ΔrecB 40

μ m) 30

SSH111 10 Cell length ( 0 0 3060 90120 150 180 ( Δ recB P sulA -GFP) Time (min) bioRxiv preprint doi: https://doi.org/10.1101/533422; this version posted January 29, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Supplementary Figure 4 Fluorescence images following ciprofloxacin treatment: A SOS reporter plasmid (PsulA-GFP) in ΔrecB 0 min 10 min 20 min 30 min

40 min 50 min 60 min 70 min recB Δ 80 min 90 min 100 min 120 min

Fluorescence image following ciprofloxacin treatment: B SOS reporter plasmid (PsulA-GFP) in MG1655 120 min MG1655 bioRxiv preprint doi: https://doi.org/10.1101/533422; this version posted January 29, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Supplementary Figure 5 DinB-YPet and replisome foci DinB-YPet and replisome foci A after ciprofloxacin addition (+/- DMSO) B after trimethoprim addition (+/- DMSO) 0 min 90 min 180 min 0 min 90 min 180 min trimethoprim ciprofloxacin + + DMSO DMSO trimethoprim ciprofloxacin

DinB-YPet DnaX-mKate2 colocalized DinB-YPet DnaX-mKate2 colocalized

Colocalisation: Colocalisation: Colocalisation: Colocalisation: C ciprofloxacin D ciprofloxacin + DMSO E trimethoprim F trimethoprim + DMSO n(cells) > 300 n(cells) > 100 n(cells) > 100 n(cells) > 100 50 10 50 10 50 10 50 10 9 9 9 9 40 8 40 8 40 8 40 8 7 7 7 7 30 6 30 6 30 6 30 6 5 5 5 5 20 4 20 4 20 4 20 4 3 3 3 3 10 2 10 2 10 2 10 2 1 1 1 1 0 0 DinB foci that bind at replisomes (%) 0 0 DinB foci that bind at replisomes (%) 0 0 DinB foci that bind at replisomes (%) 0 0 DinB foci that bind at replisomes (%) Replisomes that have a DinB focus (%) 0 20 40 60 80 100 120 140 160 180 0 20 40 60 80 100 120 140 160 180 Replisomes that have a DinB focus (%) 0 20 40 60 80 100 120 140 160 180 Replisomes that have a DinB focus (%) 0 20 40 60 80 100 120 140 160 180 Replisomes that have a DinB focus (%) Time (min) Time (min) Time (min) Time (min) bioRxiv preprint doi: https://doi.org/10.1101/533422; this version posted January 29, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Supplementary Figure 6 A Merged images: DinB-YPet (pol IV) activity at replisomes ciprofloxacin ciprofloxacin + DMSO trimethoprim trimethoprim + DMSO

following 60 min treatment 5 µm

DinB-YPet DnaX-mKate2 colocalised

B Colocalisation: ciprofloxacin (+/- DMSO) C Colocalisation: trimethoprim (+/- DMSO) - DMSO + DMSO - DMSO + DMSO Pol IV/replisomes Pol IV/replisomes Pol IV/replisomes 100 100 Pol IV/replisomes 100 Replisomes/polIV Replisomes/polIV 100 Replisomes/polIV Replisomes/polIV 80 80 80 80

60 60 60 60

40 40 40 40

20 20 20 20 Colocalisation (%) Colocalisation (%)

0 0 0 0 1s 3s 5s 10s 1s 3s 5s 10s 1s 3s 5s 10s 1s 3s 5s 10s Lifetime of pol IV focus (s) Lifetime of pol IV focus (s) bioRxiv preprint doi: https://doi.org/10.1101/533422; this version posted January 29, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Supplementary Figure 7

A UmuC-mKate2 (pol V) activity at replisomes ciprofloxacin ciprofloxacin + DMSO trimethoprim trimethoprim + DMSO UmuC-mKate2 signal following 60 min treatment Merged image

5 mm

UmuC-mKate2 DnaX-YPet colocalised

Number of UmuC foci per cell: Number of UmuC foci per cell: B ciprofloxacin (+/- DMSO) C trimethoprim (+/- DMSO) 2.0 2.0

1.5 1.5

1.0 1.0 cipro no damage 0.5 0.5 no damage UmuC foci per cell UmuC foci per cell cipro + DMSO trimehtoprim trimehtoprim 0.0 0.0 + DMSO bioRxiv preprint doi: https://doi.org/10.1101/533422; this version posted January 29, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Supplementary Figure 8 Fluorescence images: UmuC-mKate2 and replisome foci A Signal UmuC-mKate2 (Pol V): B during stress

0 min 90 min 180 min 0 min 90 min 180 min ciprofloxacin ciprofloxacin + + DMSO DMSO ciprofloxacin ciprofloxacin trimethoprim trimethoprim

UmuC-mKate2 signal UmuC-mKate2 DnaX-YPet colocalized

UmuC-mKate2 concentration Colocalisation: Colocalisation: C during stress D ciprofloxacin E ciprofloxacin + DMSO

n(cells) > 100 n(cells) > 100 n(cells) > 100 2000 50 10 10 ciprofloxacin 50 ciprofloxacin + DMSO 9 9 trimethoprim 1800 40 8 40 8 7 7 1600 30 6 30 6 5 5 1400 20 4 20 4 3 3 1200 10 2 10 2

Mean cell brightness (a. u.) 1 1 1000 0 0 0 0 0 60 120 180 0 20 40 60 80 100 120 140 160 180 0 20 40 60 80 100 120 140 160 180 UmuC foci that bind at replisomes (%) UmuC foci that bind at replisomes (%)

Time (min) Time (min) Replisomes that have a UmuC focus (%) Time (min) Replisomes that have a UmuC focus (%) bioRxiv preprint doi: https://doi.org/10.1101/533422; this version posted January 29, 2019. The copyright holder for this preprint (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Supplementary Figure 9 A Anti-UmuD Western blots for recA+ lexA+: ciprofloxacin B Anti-UmuD Western blots for recA+ lexA+: trimethoprim chromosome: ∆umuDC; plasmid: UmuDC chromosome: ∆umuDC; plasmid: UmuDC ciprofloxacin trimethoprim untreated ciprofloxacin untreated trimethoprim + 2% DMSO + 2% DMSO 1 2 3 1 2 3 h 1 2 3 1 2 3 h 30 30 30 30 30 30 30 µL 30 30 30 30 30 30 30 µL UmuD UmuD′

20s exposure; OD 0.5 2min exposure; OD 0.5

C Anti-UmuD Western blots for recA+ lexA51(Def): ciprofloxacin D Anti-UmuD Western blots for recA+ lexA51(Def): trimethoprim chromosome: ∆umuDC; plasmid: UmuDC chromosome: ∆umuDC; plasmid: UmuDC ciprofloxacin trimethoprim untreated ciprofloxacin untreated trimethoprim + 2% DMSO + 2% DMSO 1 2 3 1 2 3 h 1 2 3 1 2 3 h 30 30 30 30 30 30 30 µL 30 30 30 30 30 30 30 µL

20s exposure; OD 0.5 2min exposure; OD 0.5