1. Cell Biology
  2. Chromosomes and Gene Expression
Download icon

Coordination between nucleotide excision repair and specialized polymerase DnaE2 action enables DNA damage survival in non-replicating bacteria

  1. Asha Mary Joseph
  2. Saheli Daw
  3. Ismath Sadhir
  4. Anjana Badrinarayanan  Is a corresponding author
  1. National Centre for Biological Sciences - Tata Institute of Fundamental Research, India
  2. Max Planck Institute for Terrestrial Microbiology, LOEWE Centre for Synthetic Microbiology (SYNMIKRO), Germany
Research Article
  • Cited 0
  • Views 528
  • Annotations
Cite this article as: eLife 2021;10:e67552 doi: 10.7554/eLife.67552
Voice your concerns about research culture and research communication: Have your say in our 7th annual survey.

Abstract

Translesion synthesis (TLS) is a highly conserved mutagenic DNA lesion tolerance pathway, which employs specialized, low-fidelity DNA polymerases to synthesize across lesions. Current models suggest that activity of these polymerases is predominantly associated with ongoing replication, functioning either at or behind the replication fork. Here we provide evidence for DNA damage-dependent function of a specialized polymerase, DnaE2, in replication-independent conditions. We develop an assay to follow lesion repair in non-replicating Caulobacter and observe that components of the replication machinery localize on DNA in response to damage. These localizations persist in the absence of DnaE2 or if catalytic activity of this polymerase is mutated. Single-stranded DNA gaps for SSB binding and low-fidelity polymerase-mediated synthesis are generated by nucleotide excision repair (NER), as replisome components fail to localize in the absence of NER. This mechanism of gap-filling facilitates cell cycle restoration when cells are released into replication-permissive conditions. Thus, such cross-talk (between activity of NER and specialized polymerases in subsequent gap-filling) helps preserve genome integrity and enhances survival in a replication-independent manner.

Introduction

DNA damage is a threat to genome integrity and can lead to perturbations to processes of replication and transcription. In all domains of life, bulky lesions such as those caused by UV light (cyclobutane pyrimidine dimers, CPD and to a lesser extent 6,4 photoproducts, 6-4PP) are predominantly repaired by nucleotide excision repair (NER) (Boyce and Howard-Flanders, 1964; Chatterjee and Walker, 2017; Kisker et al., 2013). This pathway can function in global genomic repair (GGR) via surveilling the DNA double-helix for distortions or more specifically via transcription-coupled repair (TCR) (Kisker et al., 2013). The main steps of NER involve lesion detection followed by incision of few bases upstream and downstream of the lesion, resulting in removal of a short stretch of single-stranded DNA (ssDNA). In some cases, NER can also result in generation of longer patches of ssDNA (Cooper, 1982). ssDNA gaps are then filled by synthesis from a DNA polymerase (Kisker et al., 2013; Sancar and Rupp, 1983). While the NER-mediated damage removal pathway is largely error-free, lesions encountered by the replication machinery (e.g., CPDs, 6-4PPs, and cross-links such as those generated by antibiotics including mitomycin C [MMC]) can also be dealt with via error-prone translesion synthesis (TLS) (Chatterjee and Walker, 2017; Fuchs and Fujii, 2013; Fujii and Fuchs, 2004).

TLS employs low-fidelity polymerases to synthesize across DNA lesions, with increased likelihood of mutagenesis during this process (Fuchs and Fujii, 2013; Galhardo et al., 2005; Kato and Shinoura, 1977; Nohmi et al., 1988; Warner et al., 2010). In most bacteria, expression of these polymerases is regulated by the SOS response, which is activated by the RecA-nucleoprotein filament under DNA damage (Baharoglu and Mazel, 2014). Currently, most of our understanding about TLS comes from studies on specialized Y-family polymerases of E. coli, DinB (PolIV) and UmuDC (PolV), both of which function in DNA lesion tolerance and contribute to mutagenesis in several bacterial systems (Kato and Shinoura, 1977; Nohmi et al., 1988; Steinborn, 1978; Sung et al., 2003; Wagner et al., 1999). In addition, PolV has also been implicated in RecA-dependent post-replicative gap-filling activity (Isogawa et al., 2018). In contrast to E. coli, Caulobacter crescentus as well as other bacteria including Mycobacterium sp. and Pseudomonas sp. encode an alternate, SOS-inducible error-prone polymerase, DnaE2 (Boshoff et al., 2003; Galhardo et al., 2005; Jatsenko et al., 2017; Warner et al., 2010). DnaE2 is highly conserved and is mutually exclusive with PolV in occurrence. In the limited organisms where DnaE2 has been studied so far, it is the primary TLS polymerase and the only contributor to damage-induced mutagenesis (Alves et al., 2017; Galhardo et al., 2005; Warner et al., 2010). In contrast to PolV, DnaE2 is thought to preferentially act on MMC-induced damage, where it contributes to all induced mutagenesis observed (Galhardo et al., 2005). In case of UV, there are still uncharacterized mechanisms that can contribute to damage tolerance and mutagenesis that are independent of DnaE2 (Galhardo et al., 2005). DnaE2 co-occurs with ImuB, a protein that carries a β-clamp binding motif, and is thought to act as a bridge between DnaE2 and the replisome (Warner et al., 2010). Unlike E. coli, where activities of PolIV and PolV are well-studied, in vivo investigations of DnaE2 function in damage tolerance and cellular survival are limited. This becomes particularly important, given the emerging evidences across domains of life ascribing diverse functions to these low-fidelity polymerases beyond their canonical function of replication-associated lesion bypass (Joseph and Badrinarayanan, 2020). Indeed, such polymerases are also referred to as ‘specialized polymerases’ (Fujii and Fuchs, 2020) so as to consider these broader functions.

Since error-prone polymerases can synthesize DNA and their activity is mediated by interaction with the β-clamp of the replisome (Bunting et al., 2003; Chang et al., 2019; Fujii and Fuchs, 2004; Thrall et al., 2017; Wagner et al., 2009; Warner et al., 2010), action of these polymerases has mostly been studied in the context of replicating cells, as a mechanism that facilitates continued DNA synthesis by acting at or behind the replication fork (Chang et al., 2019; Chang et al., 2020; Indiani et al., 2005; Jeiranian et al., 2013; Marians, 2018). In addition to replication-associated lesion tolerance, some studies have proposed the possibility of error-prone synthesis in a manner that is replication-independent (Janel-Bintz et al., 2017; Kozmin and Jinks-Robertson, 2013). This is supported by observations that cells can undergo stationary phase mutagenesis that is dependent on action of error-prone polymerases (Bull et al., 2001; Corzett et al., 2013; Janel-Bintz et al., 2017; Sung et al., 2003; Yeiser et al., 2002). Microscopy-based approaches have also provided evidence in line with the idea that tolerance or gap-filling could occur outside the context of the replication fork in E. coli, as replisome components, such as the β-clamp, as well as specialized polymerases (PolIV and PolV) were found to localize away from the fork in response to DNA damage (Henrikus et al., 2018; Robinson et al., 2015; Soubry et al., 2019; Thrall et al., 2017). Furthermore, while originally considered as distinct mechanisms of repair (damage tolerance vs. damage removal), recent studies also suggest cross-talk between specialized polymerases and NER in E. coli, yeast, and human cells (Giannattasio et al., 2010; Janel-Bintz et al., 2017; Kozmin and Jinks-Robertson, 2013; Sertic et al., 2018). Indeed, long-standing observations suggest that NER can be mutagenic under certain conditions in E. coli, in a manner that is dependent on RecA (Bridges and Mottershead, 1971; Cohen-Fix and Livneh, 1994; Nishioka and Doudney, 1969). However, the mechanistic basis of this process in replication-independent conditions and conservation of the same across bacteria that encode diverse specialized polymerases remains to be elucidated. For example, unlike E. coli, several bacterial systems undergo nonoverlapping cycles of DNA replication and have distinct cell cycle phases with no ongoing DNA synthesis. The relevance of lesion correction or gap-filling for genome integrity maintenance in the absence of an active replication fork (such as in non-replicating swarming cells) is relatively less explored, especially in bacterial contexts.

To probe the in vivo mechanism and understand the impact of error-prone polymerase function in non-replicating bacteria, we investigated lesion repair in Caulobacter crescentus swarmer cells. Caulobacter is well-suited to study activity of these specialized polymerases due to its distinct cell cycle. Every cell division gives rise to two different cell types: a stalked and a swarmer cell. While the stalked cell initiates replication soon after division, a swarmer cell must differentiate into a stalked cell before replication reinitiation (Schrader and Shapiro, 2015) and hence swarmers can represent a pool of naturally occurring non-replicating cells in the environment. Under laboratory conditions, these swarmer cells can be isolated via density-gradient centrifugation and replication initiation can be inhibited, resulting in a population of non-replicating cells with a single chromosome (Badrinarayanan et al., 2015; Schrader and Shapiro, 2015). Using this non-replicating system, we followed DNA damage repair with lesion-inducing agents via live-cell fluorescence microscopy. We show that low-fidelity polymerase DnaE2 functions in gap-filling damaged DNA in non-replicating cells. This is facilitated by de novo loading of replisome components (SSB, HolB [part of the clamp loader complex], β-clamp, and replicative polymerase) at long ssDNA gaps likely generated by a subset of NER events. We find that this form of gap-filling in non-replicating cells promotes cell cycle restoration and cell division, upon release into replication-permissive conditions. Our study provides in vivo evidence for a novel function of DnaE2 that is spatially and temporally separated from the active replication fork. Given that DNA damage can occur in any cell type whether actively replicating or not, coordinated activity of NER and low-fidelity polymerases can serve as a potential mechanism through which non-replicating cells such as bacteria in stationary phase or cells in other differentiated phases increase their chances of survival under damage.

Results

Monitoring mechanisms of DNA lesion repair in non-replicating bacteria

To test whether non-replicating cells can indeed engage in lesion repair, and understand the in vivo mechanism of such activity, we used Caulobacter crescentus swarmer cells as our model system. We regulated the state of replication so as to ensure that swarmer cells, with a single chromosome, do not initiate replication (and hence prevent possibility of recombination-based repair) by utilizing a previously described system to control the expression of the replication initiation regulator, dnaA, from an isopropyl β-D-1-thiogalactopyranoside (IPTG) inducible promoter (Badrinarayanan et al., 2015). In our experimental setup, we first depleted cells of DnaA for one generation of growth, followed by synchronization to isolate non-replicating swarmer cells (Figure 1A, top panel). Flow cytometry profiles of cells confirmed the presence of a single chromosome during the course of the entire experiment (Figure 1A, bottom panel).

Figure 1 with 1 supplement see all
Monitoring mechanisms of DNA lesion repair in non-replicating bacteria.

(A) Above: Schematic of experimental setup used to isolate non-replicating Caulobacter swarmer cells to monitor DNA lesion repair and tolerance independent of ongoing replication. Cells were treated with DNA damage (30 min mitomycin C [MMC] or UV at specified doses), after which damage was removed and cells were allowed to grow in fresh media (damage recovery), without ongoing replication. Below: Flow cytometry profiles show DNA content in an asynchronous population (i), synchronized non-replicating swarmer cells before (ii) and after DNA damage recovery (iii). (B) Representative images of Caulobacter cells with fluorescently-tagged replisome components (SSB-YFP, HolB-YFP, DnaN-YFP, or DnaE-mNG) in replicating or non-replicating conditions, without DNA damage (scale bar is 2 µm here and in all other images).

Given the requirement of the β-clamp for activity of specialized polymerases and evidence for damage-dependent changes in localization of replisome components such as SSB in actively replicating E. coli (Chang et al., 2019; Henrikus et al., 2018; Soubry et al., 2019; Thrall et al., 2017), we generated fluorescent fusions to the Caulobacter β-clamp (DnaN), component of the clamp-loader complex (HolB), the replicative polymerase PolIII (DnaE), and single-strand DNA binding protein (SSB) (using previously described approaches in Caulobacter [Aakre et al., 2013; Collier and Shapiro, 2009]; and 'Materials and methods') in order to visualize them in non-replicating swarmers. These fusions did not perturb the function of the proteins as cells displayed wild type growth dynamics in steady-state conditions (Figure 1—figure supplement 1A and B ‘control’). They also did not show increased sensitivity to DNA damage treatment via MMC or UV (Figure 1—figure supplement 1A and B). The fusion proteins localized on DNA in actively replicating cells (Figure 1B, +replication), and as anticipated, their localizations gradually shifted from one pole to the other within one cycle of DNA replication (Figure 1—figure supplement 1C). These observations are in line with previous reports of replisome dynamics in several bacterial systems including Caulobacter crescentus, Bacillus subtilis, and E. coli (Aakre et al., 2013; Collier and Shapiro, 2009; Jensen et al., 2001; Lemon and Grossman, 1998; Mangiameli et al., 2017; Reyes-Lamothe et al., 2008). In contrast to actively replicating cells, replication-inhibited swarmer cells were devoid of replisome foci (Figure 1B), consistent with the idea that the localization of replisome components is indicative of active DNA replication.

Replisome components are recruited to damaged DNA in non-replicating Caulobacter swarmer cells

Using the above described system, we treated non-replicating Caulobacter swarmer cells with mitomycin C (MMC) to induce DNA lesions and followed DNA damage recovery via live-cell imaging to track dynamics of the β-clamp and other replisome components (Figure 1A). MMC is a naturally produced antibiotic that acts predominantly on the guanine residue of DNA, making three major forms of damage: mono-adducts, intra-strand cross-links, and inter-strand cross-links (Bargonetti et al., 2010). In case of Caulobacter, it is thought that DnaE2 preferentially acts on MMC-induced damage as all mutagenesis associated with MMC treatment is mediated via action of this specialized polymerase; in absence of the polymerase, cells show high sensitivity to MMC treatment (Galhardo et al., 2005). To determine the range of MMC concentrations for this study, we first assessed the viable cell count for a steady-state population of wild type and ∆dnaE2 cells across increasing concentrations of MMC treatment (0.125–2 µg/ml). We focused on a treatment range where DnaE2 essentiality was observed (Figure 2—figure supplement 1A) and TLS-dependent mutagenesis has been previously reported (Galhardo et al., 2005).

We then treated non-replicating swarmer cells with the specified doses of MMC. We found that DNA damage treatment resulted in the formation of β-clamp foci in non-replicating cells (Figure 2A–B). This was found to be the case for other replisome components as well (Figure 2A–B). The percentage of cells with damage-induced β-clamp foci increased with increasing doses of MMC. At 0.125 µg/ml MMC treatment, 9% cells had β-clamp foci, while at higher doses of 0.75 µg/ml MMC, foci were observed in 59% cells (Figure 4—figure supplement 1C). To further characterize the dynamics of these localizations during the course of damage recovery, we released MMC-treated non-replicating swarmers into fresh media without damage and followed the localization of replisome components over time. We maintained the block on replication initiation, thus ensuring that cells carried only a single non-replicating chromosome during the course of the experiment (Figure 1A). Consistent with the possibility of dissociation during recovery, we found that percentage of cells with DnaN localizations gradually decreased with time (Figure 2C) and across all doses of damage tested (Figure 4—figure supplement 1C). For example, after 30 min of 0.5 µg/ml MMC treatment, 52% cells on average had DnaN localization and at 90 min after damage removal, the number reduced to 30%. This pattern of localization after damage treatment, followed by reduction in percentage of cells with foci during recovery was also observed in the case of SSB, HolB, and DnaE (Figure 2D). Interestingly, we noticed that cells had more SSB localizations on average than DnaN. 14% cells had ≥2 DnaN foci after MMC treatment, while 37% cells harbored ≥2 SSB localizations. These numbers reduced with increasing time of recovery (Figure 2D). Assessment of the extent of colocalization between DnaN and SSB further showed that 90% of DnaN foci colocalized with SSB (with distance of a DnaN focus from the nearest SSB localization being within 300 nm), while only 51% of SSB foci colocalized with DnaN (Figure 2—figure supplement 1B and C), suggesting that not all SSB may be associated with the β-clamp or that SSB could precede β-clamp localization.

Figure 2 with 1 supplement see all
Replisome components are recruited to damaged DNA in non-replicating Caulobacter swarmer cells.

(A) Representative images of non-replicating swarmer cells with fluorescently tagged replisome components (SSB-YFP, HolB-YFP, DnaN-YFP, or DnaE-mNG) with (+MMC) or without (no damage) 30 min of treatment with MMC. (B) Percentage cells with SSB, HolB, DnaN, or DnaE localization (foci) in non-replicating swarmers with (+) or without (-) MMC treatment (n ≥ 324 cells, three independent repeats). Dashed line represents median here and in all other graphs. (C) Percentage swarmer cells with 0, 1, or ≥2 DnaN foci at 0, 30, 60, and 90 min after damage removal (recovery) (n ≥ 476 cells, three independent repeats). (D) Percentage swarmer cells with 0, 1, or ≥2 foci of SSB, HolB, or DnaE at 0 and 90 min after damage removal (recovery) (n ≥ 324 cells, three independent repeats).

We asked whether similar dynamics of replication machinery components were observed in the presence of a different lesion-inducing agent as well. For this, we treated cells with sub-inhibitory doses of UV radiation (Galhardo et al., 2005 and Figure 2—figure supplement 1D). Exposure of cells to two doses of UV damage (75 and 150 J/m2) also resulted localization and subsequent reduction in percentage of cells with replisome foci during recovery (Figure 2—figure supplement 1E,F,G). Taken together, these data support the idea that SSB, along with components of the PolIIIHE, including the clamp-loader, β-clamp, and the replicative polymerase, associates with DNA during damage even in the absence of ongoing replication. Decrease in localizations over time could be indicative of potential repair in non-replicating cells.

Nucleotide excision repair (NER) generates long ssDNA gaps for localization of replisome components in non-replicating cells

How do replisome components localize in non-replicating cells? SSB foci under these conditions indicate the presence of ssDNA stretches long enough to accommodate SSB tetramers (30 nt or more) (Bell et al., 2015; Lohman and Ferrari, 1994). In replicating cells, ssDNA tracts are thought to be generated as a result of helicase activity that continues to unwind double-stranded DNA ahead of the replisome that has encountered a lesion (Belle et al., 2007). It is unclear how such tracts are formed in non-replicating cells. We wondered whether this could be mediated via pathways involved in DNA damage repair and tolerance. Given that several repair pathways are regulated under the SOS response (Baharoglu and Mazel, 2014), we first assessed the induction of the response in non-replicating cells under DNA damage. For this, we measured the induction of yfp from an SOS-inducible promoter (PsidA) integrated on the Caulobacter chromosome at the xyl locus (Badrinarayanan et al., 2015Figure 3A). We found that non-replicating cells activated the DNA damage response after MMC exposure, providing further evidence for the formation of ssDNA gaps in such conditions (Figure 3A). We thus asked whether the SOS response is essential for the formation of such gaps or if the activation of this response is a consequence of gap generation. Deletion of the SOS activator, recA, did not perturb localization of DnaN under damage. However, RecA was essential for dissociation during damage recovery as DnaN foci persisted in non-replicating cells lacking recA (Figure 3B). These observations suggest that a RecA-independent pathway is required for regulating the association of replisome components with DNA in cells that are not undergoing active DNA synthesis.

Figure 3 with 1 supplement see all
Nucleotide excision repair (NER) generates long ssDNA gaps for localization of replisome components in non-replicating cells.

(A) SOS induction was measured by assessing the expression of yfp from an SOS-inducible promoter (PsidA-yfp). On the left are representative images of cells expressing the reporter at 0 or 90 min after MMC removal and control cells (no damage). On the right, total fluorescence intensity normalized to cell area is plotted for both time points for cells with or without damage treatment. Each dot represents a single cell. Mean and SD are shown in black (n ≥ 219). (B) Percentage wild type, ∆recA, or ∆uvrA swarmer cells with DnaN foci at 0, 30, 60, and 90 min after DNA damage recovery (n ≥ 308 cells, three independent repeats). (C) Representative images of wild type or ∆uvrA swarmer cells with SSB-YFP or DnaN-YFP, treated with MMC or UV. (D) As (A) for cells lacking uvrA (n ≥ 325).

In most organisms, helix distorting lesions are recognized and excised by nucleotide excision repair (NER) (Kisker et al., 2013). A small proportion of the short gaps generated during this process could also be converted into longer stretches of ssDNA tracts under certain conditions, such as under high doses of DNA damage (Cooper, 1982; Giannattasio et al., 2010). This would require extensive DNA synthesis outside the active replication fork (Figure 3—figure supplement 1A). To test if this could be the mechanism by which replisome components associate with DNA in cells that are not replicating, we assessed the involvement of NER in orchestrating the same in Caulobacter swarmer cells. We observed that non-replicating cells with deletion of uvrA (part of the NER pathway) did not form DnaN foci under MMC or UV damage (Figure 3B–C, Figure 3—figure supplement 1C). In contrast, percentage of cells with DnaN foci in a ∆mutL background (deficient in mismatch repair; Marinus, 2012) was similar to wild type, indicating that mismatch repair did not contribute to loading of the β-clamp in non-replicating cells (Figure 3—figure supplement 1D).

Thus, our data suggest that lesion processing by NER alone results in the formation of ssDNA gaps on which replisome components can localize in non-replicating cells. Consistent with this, we observed lack of SSB localization in ∆uvrA cells both under MMC and UV damage (Figure 3C, Figure 3—figure supplement 1B–C). Furthermore, cells without NER were deficient in SOS induction (Figure 3D), suggesting that NER-mediated gap generation serves two functions: (a) providing ssDNA substrate for recruitment of SSB and other replisome components to these regions and (b) induction of the SOS response. Together, this facilitates ssDNA gap-filling in non-replicating Caulobacter.

SOS-induced low-fidelity polymerase, DnaE2, is essential for subsequent dissociation of replisome components

As stated above, we observed that ∆recA cells were not deficient in DnaN recruitment to ssDNA gaps. However, given that these cells had persistent β-clamp foci, we wondered what would be the requirement for RecA or the SOS response in ssDNA gap-filling. We ruled out a role for homologous recombination in this process as our experimental setup of non-replicating swarmer cells (with a single chromosome) does not permit gap-filling by recombination, due to absence of a homologous template for repair (Figure 1A, bottom panel). In addition, we also conducted our damage recovery experiments in cells lacking the recombination protein RecN (Vickridge et al., 2017), an essential component of recombination-based repair in Caulobacter (Badrinarayanan et al., 2015). We observed similar dynamics of β-clamp foci to that seen in wild type cells in this case as well (Figure 4—figure supplement 1A).

Reports in E. coli as well as eukaryotic systems (including yeast and human cells) have suggested that ssDNA gaps generated by NER can sometimes be filled by specialized polymerases like Polκ (Janel-Bintz et al., 2017; Kozmin and Jinks-Robertson, 2013; Sertic et al., 2018). Given that the SOS response is activated in non-replicating cells (Figure 3A), it is possible that gap-filling in Caulobacter swarmer cells is mediated via such specialized polymerases expressed under this regulon (Galhardo et al., 2005). Although we were unable to generate a functional fluorescent fusion to Caulobacter low-fidelity polymerase DnaE2, we confirmed that DnaE2 is expressed in our experimental conditions (Figure 4—figure supplement 1B) and that deletion of dnaE2 resulted in severe sensitivity of a steady-state population of cells to MMC-treatment (Figure 2—figure supplement 1A, Figure 4—figure supplement 1F). To test the involvement of DnaE2 in gap-filling, we conducted our damage recovery experiments in cells deleted for the same. Similar to ∆recA cells, we found that non-replicating cells lacking dnaE2 had persistent DnaN foci during damage recovery (Figure 4A–B). For example, in case of wild type, 52% cells had foci after 30 min of 0.5 µg/ml MMC treatment and this number reduced to 30% 90 min post-MMC removal. In contrast, in the case of ∆dnaE2 cells, 61% cells had foci after 30 min of damage treatment and this number remained constant even after removal of MMC from the growth media. DnaN foci in ∆dnaE2 cells was significantly higher than wild type after 90 min of damage recovery in the case of UV damage as well, at the two doses of damage tested (Figure 4—figure supplement 1D).

Figure 4 with 1 supplement see all
SOS-induced low-fidelity polymerase, DnaE2, is essential for subsequent dissociation of replisome components.

(A) Representative images of wild type or ∆dnaE2 swarmer cells with SSB-YFP, DnaN-YFP, or DnaE-YFP after MMC treatment. (B) Percentage wild type or ∆dnaE2 swarmer cells with 0, 1, or ≥2 DnaN foci at 0, 30, 60, and 90 min of DNA damage recovery (n ≥ 467 cells, three independent repeats, wild type data from Figure 2C). (C) Percentage wild type or ∆dnaE2 swarmer cells with SSB or DnaE foci at 0 and 90 min of DNA damage recovery (n ≥ 325 cells, mean and SD from three independent repeats). (D) Percentage wild type, dnaE2 catalytic mutant (dnaE2*) or ∆imuB swarmer cells with DnaN foci at 0, 30, 60, and 90 min of mitomycin C (MMC) damage recovery (n ≥ 342 cells, three independent repeats, wild type data from Figure 3B).

Replisome persistence in the absence of dnaE2 appeared to be a dose-dependent phenomenon (Figure 4—figure supplement 1C). At low dose of MMC treatment (0.125 µg/ml), fewer cells had DnaN foci post-DNA damage exposure (14.5% cells). The number further reduced to 9.5% during recovery in a DnaE2-independent manner. However, the percentage of cells with persistent β-clamp foci increased with increasing concentrations of damage in the absence of dnaE2, with minimal recovery observed at 0.5–0.75 µg/ml of MMC treatment (Figure 4—figure supplement 1C). The following observations in our study lend additional support to the proposed idea that a specialized polymerase is required for gap-filling across long ssDNA tracts generated by NER at higher doses of DNA damage: a. Persistence of components of PolIIIHE (DnaE and DnaN) in the absence of DnaE2. Apart from β-clamp foci, we found that the replicative polymerase, DnaE, was also unable to dissociate during damage recovery in cells lacking dnaE2 (Figure 4C), suggesting that the replicative polymerase alone cannot complete synthesis across these NER-generated ssDNA tracts. Such lack of dissociation after localization was found to be the case for SSB as well, again suggesting that long ssDNA gaps persisted in the absence of DnaE2 (Figure 4C). b. Requirement for DnaE2-mediated synthesis. To test whether synthesis by DnaE2 contributed to gap-filling in non-replicating cells, we mutated two residues known to be essential for DnaE-mediated synthesis (Lamers et al., 2006; Pritchard and McHenry, 1999). These residues have been mutated previously in M. smegmatis DnaE2, where it was shown to inhibit DnaE2-dependent mutagenesis (Warner et al., 2010Figure 4—figure supplement 1E). In the case of Caulobacter as well, catalytic mutant dnaE2* showed similar growth defects as ∆dnaE2 under MMC damage (Figure 4—figure supplement 1F). In our experimental regime, we found that cells expressing catalytically inactive DnaE2 also had persistent DnaN foci during damage recovery, as seen in the case of cells lacking the specialized polymerase (Figure 4D).

To assess the contribution of DnaE2 in damage-induced mutagenesis, we conducted mutagenesis assays by measuring the frequency of rifampicin resistance generation in a population of cells subject to damage, either with or without recovery in non-replicating conditions. We observed that this polymerase was responsible for all damage-induced mutagenesis in our experimental regimen (Figure 4—figure supplement 1G). However, the genetic complexity of this experiment and the confounding effects of replication during the outgrowth period preclude us from conclusively interpreting if this mutagenesis mediated by DnaE2 occurred in non-replicating, replicating or both phases of the cell cycle.

Finally, we also assessed the requirement for the accessory protein ImuB in DnaE2 function. ImuB is an inactive Y-family polymerase and carries a β-clamp binding motif. It is thought to act as a bridge between DnaE2 and the clamp (Warner et al., 2010). In Caulobacter, it is co-operonic with dnaE2 and is expressed in response to SOS activation (Galhardo et al., 2005). When we conducted our recovery experiments in cells lacking imuB, we observed that these cells also exhibited persistent DnaN foci, as seen for cells lacking dnaE2 (Figure 4D). These results are consistent with the idea that DnaE2-mediated synthesis contributes to gap-filling and subsequent dissociation of replisome components in non-replicating cells.

DnaE2 activity on NER-generated long ssDNA gaps enhances survival of non-replicating cells under DNA damage

Taken together, our data provide in vivo support for cross-talk between NER and specialized, low-fidelity polymerases during gap-filling in non-replicating bacteria. What could be the relevance of this in the context of damage recovery and survival of bacteria that are not actively replicating? To investigate the impact of NER-mediated DnaE2 activity in Caulobacter swarmer cells, we assessed the growth dynamics of these cells once released into replication-permissive conditions after damage recovery with three parameters: (a). Time to division and percentage of cells with successful division events after release in replication-permissive conditions (as a read-out for division restoration post-DNA damage clearance) (b). Cell length restoration (as a read-out for SOS deactivation following DNA damage clearance). (c). Cell survival measured via viable cell count assays.

To measure division restoration, we released replication-blocked swarmer cells into media containing IPTG (to allow for replication initiation via induction of dnaA) either immediately after damage treatment or after 90 min of damage recovery. We followed single cells via time-lapse imaging to assess the time taken to first division after replication initiation (Figure 5A–B). Control cells without damage treatment and with or without an additional 90 min arrest in swarmer stage were able to robustly resume cell growth and division, with >94% cells undergoing their first division within 240 min of release into replication-permissive conditions. Based on this, we followed cell division dynamics for cells treated with damage during this time window, wherein control cells (without damage) were successfully able to restore cell division. In MMC-treated conditions, we found that cells released into replication-permissive conditions immediately after damage treatment did not recover efficiently, with only 5% cells undergoing their first division within 240 min (Figure 5C). In contrast, wild type cells that were provided time for damage recovery before reinitiating replication showed restoration of cell division in the same time period, with 30% cells undergoing at least one division and 9% cells undergoing ≥2 divisions within 240 min (Figure 5B–C). These recovery dynamics were dependent on DnaE2 as only 7% cells lacking dnaE2 underwent divisions even when they were provided the same time duration as wild type for damage recovery before replication reinitiation (Figure 5B–C). Thus, DnaE2-mediated gap-filling provided significant survival advantage to non-replicating cells as measured by their ability to robustly restore cell cycle progression and cell division.

Figure 5 with 1 supplement see all
DnaE2 activity on nucleotide excision repair (NER)-generated long single-stranded DNA (ssDNA) gaps enhances survival of non-replicating cells under DNA damage.

(A) Schematic of experimental setup used to assess the impact of lesion repair/ tolerance in non-replicating cells. After mitomycin C (MMC) treatment for 30 min, cells were either released into replication-permissive media (i: no recovery) or allowed to grow for 90 min without damage and then released into replication-permissive media (ii: damage recovery). Cells were followed via time-lapse microscopy and time to division was estimated. Control cells were taken through the same growth regimes; however, no damage is added to the culture. (B) Representative time-lapse montage of wild type or ∆dnaE2 cells in replication-permissive media after DNA damage recovery. Cell divisions are marked with white asterisk. In the panel shown here, three divisions were scored in wild type, while none were observed in ∆dnaE2 cells. (C) Percentage cell division over time after replication reinitiation for wild type and ∆dnaE2 cells either without (i: no recovery) or with (ii: recovery) damage recovery time in replication-blocked conditions (n ≥ 368 cells). Inset: Percentage cells divided at 240 min in each of these conditions is summarized. (D) Survival of wild type and ∆dnaE2 cells either without (i: no recovery) or with (ii: recovery) damage recovery time in replication-blocked conditions was measured via estimation of viable cell count (three independent repeats). Fraction survival was calculated by normalizing viable cell count under DNA damage to that without DNA damage (mean with SD from three independent experiments).

To further assess the consequence of gap-filling, we measured the cell length distributions for cells released into replication-permissive conditions with or without 90 min of DNA damage recovery (Figure 5—figure supplement 1A). Continued cell length elongation would be reflective of a continued division block, a hallmark of the SOS response. On the other hand, cell length restoration would be expected only for those cells where damage has been repaired (Raghunathan et al., 2020). We found that cells that did not face damage (with or without dnaE2) had a median cell length of 4.6 µm after 90 min incubation in swarmer conditions. At 240 min after reinitiation of replication, the cell length distribution was restored close to a wild type-like pattern (control) with the median cell length dropping to 2.9 µm (Figure 5—figure supplement 1A, ‘no damage’). Length restoration was also observed in wild type cells able to engage in DnaE2-mediated gap-filling in the 90 min recovery window (Figure 5—figure supplement 1A, ‘+ damage, recovery, wild type’). This restoration in cell length was dependent on the time provided for damage recovery as well as presence of DnaE2. In the absence of recovery or dnaE2, cells continued to elongate after release into IPTG-containing media (Figure 5—figure supplement 1A, ‘+damage, no recovery’ and ‘+damage, recovery, ∆dnaE2’).

To lend support to these cell biological observations, we modified our recovery setup to measure viable cell counts (Figure 5—figure supplement 1B). For this, we assessed the ‘fraction survival’ as defined by the viable cell count obtained for cultures with damage treatment normalized to the viable cell count for cultures without damage treatment. We observed that wild type cells that were released into replication-permissive conditions without the 90 min window of damage recovery were significantly compromised in growth, with fraction survival reducing to 0.19 at higher doses of damage in the absence of recovery. On the other hand, in case of cells grown with the possibility of undergoing 90 min of damage recovery, the fraction survival increased to 0.45 at the highest dose of damage used (Figure 5D). We then asked whether the survival advantage observed during replication-independent damage recovery required DnaE2 action. Consistent with a dose-dependent effect on replisome persistence in the absence of DnaE2, we also observed that DnaE2 had a significant impact on the replication-independent survival advantage at higher doses of DNA damage. We found that cells deleted for dnaE2 were severely compromised for survival at all doses of damage used (Figure 5D). However, at higher doses of damage, cells lacking dnaE2 had similar reduction in viable cell counts whether or not they were given a 90 min window of recovery. For example, after treatment with 0.5 µg/ml of MMC, only 0.01 fraction survival was observed for cells lacking dnaE2 (with or without damage recovery). On the other hand, wild type cells which had 90 min of damage recovery showed a fraction survival of 0.45 (Figure 5D). Thus, there was a significant component of enhanced survival in cells that could undergo repair in non-replicating conditions and this survival advantage was dependent on DnaE2.

In summary, our cell biological and genetic read-outs suggest that DnaE2-mediated gap-filling enables cell cycle restoration and cell division licensing when non-replicating cells are allowed to reinitiate DNA replication. In the absence of such recovery (either dnaE2 deletion or cells grown without time for recovery), cell division is compromised and cells continue to elongate, a hallmark of persistent DNA damage and hence continuously active SOS response. The impact of delayed cell division and subsequent cell length elongation is directly observed when viable cell count of the population is measured, with a dose-dependent effect on survival in cells compromised for recovery due to deletion of dnaE2.

Discussion

DNA lesion repair and tolerance have been well studied in a replication-centric paradigm (Gabbai et al., 2014; Indiani et al., 2005; Marians, 2018). Characterization of error-prone polymerases in E. coli has informed us about mechanisms of tolerance that could occur at the replication fork or behind it, in gaps generated due to replisome skipping over the lesion, followed by repriming downstream of it (Chang et al., 2019; Gabbai et al., 2014; Indiani et al., 2005). However, DNA damage is a universal event that can occur across all stages of the cell cycle, including in non-replicating conditions. This can have effects on transcription and could also perturb replication progression upon reinitiation (Jeiranian et al., 2013; Lang and Merrikh, 2018; Rudolph et al., 2007). For example, bacteria such as Caulobacter have distinct cell cycle phases including a non-replicating swarmer state, with a single copy of its chromosome. Hence, it is imperative that DNA damage is repaired efficiently even in these conditions. Here, we provide in vivo evidence for NER-coupled DnaE2 function that is active in non-replicating bacteria. This study complements a growing body of work that supports the possibility of low-fidelity polymerase-mediated synthesis (including mutagenesis) in replication-independent conditions (such as in stationary phase cells) across domains of life (Bull et al., 2001; Corzett et al., 2013; Janel-Bintz et al., 2017; Sung et al., 2003; Yeiser et al., 2002) and underscores the need to reconsider function of such polymerases outside canonical, isolated roles of lesion bypass during replication.

DNA damage repair and tolerance in non-replicating cells: requirement for DnaE2

Here, we develop a system to specifically assess mechanisms of damage repair and tolerance employed in cells that are not undergoing active DNA synthesis. Using replication initiation-inhibited Caulobacter swarmer cells, we show that lesions can be dealt with in two main steps: a. damage processing by NER to reveal SSB-bound long ssDNA gaps and b. gap-filling by SOS-induced specialized polymerase, DnaE2. Due to absence of a second copy of the chromosome in our assay (all cells are non-replicating and have a single chromosome), role for homologous recombination in this process is unlikely. Hence, our observations are consistent with a scenario where the low-fidelity polymerase alone is sufficient to synthesize across these long ssDNA gaps generated by NER action. Why is there a need for a specialized polymerase during gap-filling of NER-generated substrates? We explore two possible scenarios here:

(1) Conventionally NER is thought to generate gaps of approximately 12 nucleotides during lesion repair, which can be gap-filled by DNA PolI (Kisker et al., 2013). However, localization of SSB in our experiments suggests the presence of gaps > 30 nucleotides, enabling SSB tetramerization and binding (Bell et al., 2015; Lohman and Ferrari, 1994). How are longer ssDNA tracts generated? Previous reports in E. coli as well as yeast and human cells have implicated a role for exonuclease activity in generating longer ssDNA tracts on some NER substrates. In these studies, it was proposed that such activity would occur on problematic intermediates generated during NER activity, including closely-spaced opposing lesions that are generated under high doses of DNA damage (Janel-Bintz et al., 2017; Kozmin and Jinks-Robertson, 2013; Sertic et al., 2018). Indeed, our observations on lack of dissociation of replicative polymerase (PolIII) in the absence of DnaE2 as well as the dose-dependent impact on cell survival would both be consistent with a speculative model where NER-mediated excision results in the production of lesion-containing ssDNA that requires synthesis by a specialized polymerase.

(2) It is equally plausible that DnaE2 contributes to gap-filling independent of the presence or absence of a DNA lesion. Earlier studies in E. coli indicated that a minor component of NER-mediated removal of UV lesions can result in long ssDNA gaps that are gap-filled in a process referred to a long patch excision repair (LPER) (Cooper, 1982). Though the molecular players of this long patch synthesis are unidentified, this process did not result in detectable mutations. Furthermore, studies in yeast have implicated a role for exonuclease activity (via Exo1) in generating long gaps during a subset of NER events which drive checkpoint activation and are eventually filled via long patch repair synthesis (Giannattasio et al., 2010). DnaE2 could function similarly in gap-filling on these long ssDNA gaps formed as a consequence extensive NER in the context of severe DNA damage. Gap-filling activity has been suggested previously for specialized polymerases such as eukaryotic Polκ (Ogi and Lehmann, 2006). In addition, recent studies on post-replicative gap-filling have proposed a scenario where long patches requiring synthesis are accessed by both replicative and TLS polymerases (PolIV and PolV) in E. coli (Isogawa et al., 2018). In the case of non-replicating Caulobacter cells, it is possible that DnaE2 can access the β-clamp and hence participate in such gap-filling, given the observed increase in DnaE2 levels via SOS induction.

A limitation of our current study is that we do not observe all NER events, a significant proportion of which could be mediated via gap-filling by PolI on short ssDNA stretches. The relative contribution of these two arms of NER (long vs. short patch repair) could vary with increasing doses of damage and subsequently impact the requirement for DnaE2 action in gap-filling. Unfortunately, using our mutagenesis assays (measuring generation of rifampicin resistant mutations during damage), we were unable to satisfactorily disentangle the individual contributions of DnaE2-mediated mutagenesis in non-replicating vs. replicating conditions (Figure 4—figure supplement 1G). Hence, we cannot reliably distinguish between the ‘gap-filling alone’ or ‘gap-filling associated with lesion bypass’ activities of this polymerase in our present study. It must be noted though, that a role for DnaE2 in gap-filling alone has not been reported before. In addition, unlike E. coli, it is the only polymerase implicated in TLS-associated functions (mutagenesis) in the bacteria that encode it. However, irrespective of the specific nature of DnaE2 activity, our work underscores a novel and necessary function for this highly conserved specialized polymerase in conjunction with NER in replication-independent conditions.

Long ssDNA gaps generated by NER serve two functions

Previous studies in E. coli have found that NER activity in GGR is dependent on the activation of the SOS response (Crowley and Hanawalt, 1998). In contrast, our results suggest that NER functions upstream of the SOS response in non-replicating Caulobacter in the context of long patch repair. Although uvr genes are SOS-induced even in Caulobacter (da Rocha et al., 2008), it is possible that basal levels of Uvr proteins are sufficient to carry out damage scanning and subsequent lesion processing. Indeed, in E. coli, basal UvrA levels are variable, but range from 9 to 43 copies in minimal media and more than 120 copies in rich media (Ghodke et al., 2020). Long ssDNA gaps generated by NER serve two purposes:

(a) Activation of the SOS response for specialized polymerase expression; it is likely that in case of Caulobacter, RecA is essential only for turning on the SOS regulon as DnaE2-mediated synthesis has been previously shown to function independent of RecA (Alves et al., 2017; Galhardo et al., 2005), unlike E. coli PolV (Goodman, 2014; Nohmi et al., 1988).

(b) Providing substrate for SSB and PolIIIHE localization and specialized polymerase-mediated gap-filling. SSB localization on ssDNA could further facilitate recruitment and loading of the PolIIIHE. While PolIII activity could directly contribute to gap-filling (Isogawa et al., 2018; Sedgwick and Bridges, 1974; Soubry et al., 2019), it is also likely that it is the loading of the β-clamp that is essential for DnaE2 activity (Bunting et al., 2003; Chang et al., 2019; Fujii and Fuchs, 2004; Wagner et al., 2009). Additionally, recent studies have highlighted a role for SSB as well in enriching the local pool of PolIV at a lesion, thus enabling polymerase switching (Chang et al., 2020). The lack of a significant percentage of cells with multiple replisome foci under damage would suggest that only a limited number of long ssDNA gaps are generated per cell or that some repair/replisome component involved in gap processing or gap-filling is limiting.

It would be interesting now to ask how these additional components (such as ImuB and other accessory components to DnaE2) contribute to the regulation of the ‘specialized replisome’ outside the realms of active replication and whether the properties of the ssDNA gaps generated vary under damaging conditions that result in different types of lesions (CPDs in UV vs. monoadducts and cross-links in MMC) (Bargonetti et al., 2010; Chatterjee and Walker, 2017; Mitchell and Nairn, 1989). Indeed, although discussed in the context of non-replicating cells, it is plausible that, under high doses of damage, this mechanism can occur spatially and temporally disconnected from the active replication fork in replicating cells as well, in support of observations in E. coli that have reported localization of PolIIIHE and specialized polymerases away from the active replication fork (Henrikus et al., 2018; Soubry et al., 2019).

Relevance of NER-mediated specialized polymerase activity in non-replicating cells

Our study provides comprehensive insights into a mechanism of lesion repair and gap-filling in non-replicating bacteria, which relies on a coordinated action between NER and low-fidelity polymerases. Our data suggests a method through which an error-prone polymerase, DnaE2, functions beyond replication forks, impinging on its implications in growth and survival of non-replicating cells. The experimental system in this study provides a novel tool to investigate these mechanisms as well as additional players further and assess impacts of lesion repair and tolerance in replication-independent, but metabolically active conditions, where damage to DNA via molecules including ROS is possible (Gray et al., 2019; Manina and McKinney, 2013) (such as Caulobacter cells in ‘swarmer’ state or other cells outside S phase of cell cycle).

The relevance of the process described here is highlighted by the survival advantage it confers to non-replicating cells. It is possible that NER-coupled DnaE2-mediated gap-filling helps avoid the problems associated with persistent ssDNA gaps (due to extensive NER activity itself) or DNA damage on the chromosome (Jeiranian et al., 2013; Murli et al., 2000; Rudolph et al., 2007). For example, recent study in human cells showed that coordinated action of NER along with Y-family polymerase, Polκ, and exonuclease, Exo1, was crucial for gap-filling and prevention of UV-induced double-stranded breaks in non-S phase cells (Sertic et al., 2018). Such a role for specialized polymerases in gap-filling has also been observed in case of yeast cells (Kozmin and Jinks-Robertson, 2013; Sertic et al., 2011). More generally, our work highlights the possibility of coordinated activity of repair and tolerance pathways canonically studied as functioning independently. The universality of the NER-mediated error-prone polymerase function described here is underscored by its functionality in a diverse range of model systems, from bacteria to yeast and human cells (Janel-Bintz et al., 2017; Kozmin and Jinks-Robertson, 2013; Sertic et al., 2018), independent of the type or family of error-prone polymerase (DnaE2 in Caulobacter vs. PolIV/ PolV in E. coli) employed during gap-filling.

Materials and methods

Key resources table
Reagent type
(species) or resource
DesignationSource or referenceIdentifiersAdditional information
Strain, strain background
Caulobacter crescentus NA1000
Caulobacter crescentus NA1000 strainsPMID:334726
This study
Supplementary file 1
Recombinant DNA reagentPlasmidsThis studySupplementary file 2
Sequence based reagentsOligosThis studySupplementary file 3
AntibodyAnti-Flag (mouse monoclonal)Sigma-AldrichF1804
(RRID:AB_262044)
Western blot
(1:2000)
AntibodyAnti-mouse IgG, HRP-linked antibodyCell Signaling Technology7076S
(RRID:AB_330924)
Western blot
(1:5000)
Commercial assay, kitSuperSignal West Pico Plus Chemiluminescent SubstrateThermo Scientific34577Western blot
Chemical compound, drugMitomycin C (MMC)AG ScientificM-2715DNA damaging agent
Commercial assay, kitSYTOX Green Nucleic Acid StainThermo Fisher ScientificS7020Flow cytometry
Chemical compound, drugPercollGE Healthcare17-0891-01Synchrony
Software, algorithmGraphPad Prism 8GraphPad SoftwareRRID:SCR_002798Analysis
Software, algorithmFiji (ImageJ)Schindelin et al., 2012RRID:SCR_002285Analysis
Software, algorithmMATLAB R2020aMathWorksRRID:SCR_001622Analysis
Software, algorithmOuftiPaintdakhi et al., 2016RRID:SCR_016244Analysis
Software, algorithmMicrobeTrackerSliusarenko et al., 2011RRID:SCR_015939Analysis

Bacterial strains and growth conditions

Request a detailed protocol

Bacterial strains, plasmids, and primers used in the study are listed in Key Resources Table (and Supplementary files 13). Construction of plasmids and strains is also detailed in respective supplementary files. Transductions were performed using ɸCR30 (Ely, 1991). Caulobacter crescentus cultures were grown at 30°C in PYE media (0.2% peptone, 0.1% yeast extract and 0.06% MgSO4) supplemented with appropriate concentrations of antibiotics, as required. While growing strains carrying dnaA under an IPTG-inducible promoter, liquid media were supplemented with 0.5 mM IPTG and solid media with 1 mM IPTG. Microscopy experiments were performed in minimal media containing 1X M2 salts (0.087% Na2HPO4, 0.53% KH2PO4, 0.05% NH4Cl) supplemented with 1% PYE, 0.2% glucose, 0.01 mM FeSO4, and 0.01 mM CaCl2.

Non-replicating swarmer cells were isolated using synchrony protocols described previously (Badrinarayanan et al., 2015; Chimthanawala and Badrinarayanan, 2019). Briefly, cells were grown overnight in minimal media supplemented with IPTG. Cultures in log-phase were depleted for DnaA via washing off IPTG and allowing cells to grow in IPTG (-) conditions for one generation (~130 min). Following this, cultures were synchronized and OD600 of resulting swarmer cells was adjusted to 0.1, prior to treatment with DNA damage. In case of MMC damage, appropriate volume of 0.5 mg/ml MMC (AG Scientific, #M-2715) stock (prepared by resuspending in sterile water) was added into the culture and incubated at 30°C for 30 min. Damage was washed off by pelleting down cells at 8000 rpm for 4 min and resuspending in fresh media. For UV damage, cultures were transferred to a 90 mm petri plate and exposed to specific energy settings in a UV Stratalinker 1800 (STRATAGENE). During recovery (after UV and MMC damage), cells were incubated for 90 min at 30°C and 200 rpm. For strains expressing SSB-YFP, SSB-GFP, or dnaN-YFP under Pxyl, 0.3% xylose was added 1.5 hr prior to imaging. Replication reinitiation after damage recovery was achieved by inducing cultures with 0.5 mM IPTG. DNA damage treatment used was either 0.5 µg/ml MMC (30 min) or 75 J/m2 UV for all experiments, unless otherwise specified.

For flow cytometry analysis, 300 µl of cultures were fixed in 700 µl of 70% chilled ethanol and stored at 4°C until further processing. These samples were treated with 2 µg/ml RNaseA in 50 mM sodium citrate for 4 hr at 50°C. DNA was stained with Sytox green nucleic acid stain (5 mM solution in DMSO from Thermo Fisher Scientific) and analyzed using a BD Accuri flow cytometer.

Fluorescence microscopy and image analysis

Request a detailed protocol

For time course imaging, 1 ml aliquots of cultures were taken at specified time points, pelleted, and resuspended in 100 µl of growth medium. Images were taken without damage treatment (no damage control), after 30 min of damage treatment (+ damage) and again at 0, 30, 60, and 90 min after removal of DNA damage (recovery). Control cells were grown under same treatment regime, but no damaging agent was added to growth media. 2 µl of cell suspension was spotted on 1% agarose pads (prepared in minimal medium) and imaged. For time-lapse imaging, 2 µl cell suspension was spotted on 1.5% GTG agarose (prepared in minimal medium), grown inside an OkoLab incubation chamber maintained at 30°C and imaged at specific intervals for the indicated period of time. For cell division tracking after replication reinitiation, cells were grown on 1.5% GTG agarose prepared in growth medium with 1 mM IPTG.

Microscopy was performed on a wide-field epifluorescence microscope (Eclipse Ti-2E, Nikon) with a 63X oil immersion objective (plan apochromat objective with NA 1.41) and illumination from pE4000 light source (CoolLED). The microscope was equipped with a motorized XY stage and focus was maintained using an infrared-based Perfect Focusing System (Nikon). Image acquisitions were done with Hamamatsu Orca Flash 4.0 camera using NIS-elements software (version 5.1). For excitation at 460 nm, exposure time was set to 300 ms; at 490 nm, exposure time used was 400 ms; and for 550 nm, exposure time of 300 ms was used. Images were analyzed using ImageJ as well as Microbetracker or Oufti in MatLab (Paintdakhi et al., 2016; Sliusarenko et al., 2011). Values for random positions within each cell and relative position of replisome foci were generated using the following custom-written MatLab scripts.

load(''); %mesh_file from oufti
%% extract data for position of spots and cell ids from spots file;
k = 1;
for i = 1:length(cell_List{1,1})
     if isfield(cell_List{1,1}{1,i}, 'spots')==1
          for j = 1:length(cell_List{1,1}{1,i}.spots.l)
               cell_position(k,1)= cell_List{1,1}{1,i}.spots.l(1,j);
               cellids(k,1) = i;
               k=k+1;
          end
     else
          continue
     end
end
%% calculate cell lengths from mesh file
for i = 1:length(cellids)
     var = cellids(i);
     cell_length(i,1)= length(cellList.meshData{1, 1}{1, var}.mesh);
end
%% generate random floating point numbers
for i = 1:length(cell_position)
     if cell_length(i,1)==0
          random_num(i,1)=0;
     else
          random_num(i,1) = rand(1)+ randi(cell_length(i,1)–1); %randomly generated numbers
     end
end
%% distance between random variable to dnaN
for i = 1:length(random_num)
     dist_pix(i,1) = abs(random_num(i,1)-cell_position(i,1)); %in pixels
end
%in microns
dist_micr= dist_pix*0.108;

Graphs were generated in GraphPad Prism 8. Statistical analysis was performed in GraphPad Prism 8. Exact p-values are summarized in Supplementary file 4.

Survival assay

Request a detailed protocol

For calculating viability of an asynchronous steady-state population under DNA damage (Figure 2—figure supplement 1A and D), Caulobacter cultures were grown in PYE with 0.5 mM IPTG to OD600 of 0.3. For assessing survival under MMC, serial dilutions were made in 10-fold increments and 6 µl of each dilution (10−1 to 10−8) was spotted on PYE agar containing 1 mM IPTG and appropriate amounts of MMC. For assessing survival under UV, similar serial dilutions were made, spotted on PYE agar containing 1 mM IPTG and exposed to appropriate doses of UV in a UV Stratalinker 1800 (STRATAGENE). Growth was quantified by multiplying dilution factor of the last visible spot with number of colonies on the last spot. Percentage survival for each strain was calculated by normalizing growth of that specific strain treated with different doses of DNA damage to that in media without DNA damage.

For assessing survival of non-replicating cells under MMC (Figure 5D), non-replicating swarmer cells (10 ml, OD600 0.1) were treated with different concentrations of DNA damage for 30 min. After washing off damage, these replication-blocked cells were taken through either ‘damage recovery’ (90 min recovery) or ‘no recovery’ regime. Cells from both regimes were serially diluted, plated on PYE agar containing 1 mM IPTG, and colony counts were estimated after 48 hr. Fraction survival was calculated by normalizing viable cell count of MMC-treated cells to viable cell count without DNA damage treatment. Refer Figure 5—figure supplement 1B for schematic of the experimental setup.

Rifampicin resistance assay

Request a detailed protocol

Non-replicating swarmer cells (10 ml, OD600 0.1) were grown in ‘no recovery’ or ‘damage recovery’ conditions (as described above for survival experiments; Figure 5—figure supplement 1B). At the end of each experimental treatment, cultures were spun down, resuspended in 10 ml PYE containing 0.5 mM IPTG, and grown at 30°C overnight. These cultures were plated on PYE agar containing 1 mM IPTG and 100 µg/ml Rifampicin. Rifampicin-resistant colonies were counted 48 hr after plating, and mutation frequencies were calculated by normalizing to viable cell count for that specific culture.

Western blotting

Request a detailed protocol

At specific time points of the experiment, 1.5 ml aliquots of 0.1 OD600 cultures were pelleted down at 10000 rpm for 5 min, pellets were snap frozen in liquid nitrogen and stored at −80°C. Pellets were resuspended in SDS sample buffer, and boiled at 95°C for 10 min. Equal amounts of lysates were loaded on 6% SDS-PAGE gel, resolved at 100 V and transferred to polyvinylidene fluoride (PVDF) membrane (BIO-RAD, #1620177) in a wet electroblotting system. Non-specific binding to the membrane was blocked with 5% Blotting-Grade Blocker (BIO-RAD, #170–6404), followed by probing with 1:2000 dilution of monoclonal anti-flag antibody (Sigma, #F1804, RRID:AB_262044) and 1:5000 dilution of HRP-linked anti-mouse secondary antibody (Cell Signaling Technology, #7076S, RRID:AB_330924). The blots were visualized after incubation with SuperSignal West PICO PLUS Chemiluminescent Substrate (Thermo Scientific, #34577) using an iBright FL1000 imager (ThermoFisher Scientific).

Data availability

Data analysed during this study are included in the manuscript. Numerical data files (source data files) have been provided for Figure 1—figure supplement1, Figure 2–5 and corresponding figure supplements.

References

    1. Alves IR
    2. Lima-Noronha MA
    3. Silva LG
    4. Fernández-Silva FS
    5. Freitas ALD
    6. Marques M
    7. Galhardo RS
    (2017)
    Effect of SOS-induced levels of imuABC on spontaneous and damage-induced mutagenesis in Caulobacter crescentus
    DNA Repair 59:20–26.
    1. Lamers MH
    2. Georgescu RE
    3. Lee S-G
    4. O’Donnell M
    5. Kuriyan J
    (2006)
    Crystal Structure of the Catalytic α Subunit of E. coli Replicative DNA Polymerase III
    Cell 126:881–892.
  1. Book
    1. Manina G
    2. McKinney JD
    (2013) A Single-Cell Perspective on Non-Growing but Metabolically Active (NGMA) Bacteria
    In: Pieters J, McKinney J. D, editors. Pathogenesis of Mycobacterium Tuberculosis and Its Interaction with the Host Organism. Springer. pp. 135–161.
    https://doi.org/10.1007/978-3-642-40232-6

Decision letter

  1. Kevin Struhl
    Senior Editor; Harvard Medical School, United States
  2. Maria Spies
    Reviewing Editor; University of Iowa, United States
  3. Andrew Robinson
    Reviewer; University of Wollongong, Australia
  4. Robert Fuchs
    Reviewer

In the interests of transparency, eLife publishes the most substantive revision requests and the accompanying author responses.

Acceptance summary:

In this paper, the authors showed that DnaE2, a bacterial translesion synthesis DNA polymerase participates in long patch excision repair, a sub-pathway of nucleotide excision repair. This paper will be of substantial interest to those in the fields of DNA repair and mutagenesis. The authors employ a novel approach to identify a new function for DnaE2 in bacteria linked to cell survival, thus adding to a growing list of activities for DNA polymerases outside the context of the DNA replication fork. While the study relies on indirect observations, significant novel insight is gained and the conclusions are consistent with the data.

Decision letter after peer review:

[Editors’ note: the authors submitted for reconsideration following the decision after peer review. What follows is the decision letter after the first round of review.]

Thank you for submitting your work entitled "Coordination of NER with replication-independent translesion synthesis is essential for bacterial DNA damage survival" for consideration by eLife. Your article has been reviewed by 3 peer reviewers, and the evaluation has been overseen by a Reviewing Editor and a Senior Editor. The following individuals involved in review of your submission have agreed to reveal their identity: Robert Fuchs (Reviewer #1); Andrew Robinson (Reviewer #2).

Our decision has been reached after consultation between the reviewers. Based on these discussions and the individual reviews below, we regret to inform you that your work will not be considered further for publication in eLife.

This work proposes an intriguing concept that TLS-mediated synthesis takes place on NER-generated gaps in the absence of DNA replication. After discussing the manuscript, the reviewers and the reviewing editor agreed that while the links between NER and DnaE2 are interesting, the work falls short of proving that DnaE2 is participating in TLS and that it more likely has a role in gap resolution that does not depend on lesion bypass. All three reviewers have clear concerns about the level of mechanistic insight provided by the manuscript. The individual reviews provides some suggestions of how such an insight may be achieved, but that would require a significant amount of additional work.

Reviewer #1:

My overall comment is that the data do support neither the title nor the abstract of the paper. In the title "bacterial DNA damage survival" is mentioned while no damage survival data are provided. In the abstract, the authors say they provide evidence for "replication-independent TLS". At the best their data show that a TLS polymerase, namely DnaE2, participate into NER gap-filling which does not mean that DnaE2 performs a lesion bypass. Indeed, a gap generated during NER does in principle not contain a lesion except in the rare situations of closely spaced lesions. If this were indeed the case the authors should observe a quadratic dose response. The involvement of DnaE2 during NER does not mean that DnaE2 is performing Translesion synthesis (TLS). There are more and more situations describing the involvement of TLS polymerases in transactions that do not involve lesion bypass per se. For example, the involvement of Pol kappa has been described many years ago during NER-gap filling in yeast and in human cells without bona fide lesion bypass (see ref below). To assess genuine lesion bypass the authors would need to monitor induced mutagenesis that is the signature of TLS.

Substantive concerns:

1. Proper description of the level of damage induced by MMC and UV irradiation should be provided. At the best I can tell the authors only used one dose of MMC and one dose of UV. There needs to be experiments showing the physiological effect of MMC treatment or of UV irradiation such as survival (as claimed in the title) or mutagenesis.

2. All data are based on fluorescence imaging. Other methodologies (genetics…..) should be implemented to reinforce the study.

Ogi T, Lehmann AR. The Y-family DNA polymerase kappa (pol kappa) functions in mammalian nucleotide-excision repair. Nat Cell Biol. 2006; 8: 640-642. https://doi.org/10.1038/ncb1417 PMID: 16738703

Lehmann AR. New functions for Y family polymerases. Mol Cell. 2006; 24: 493-495. https://doi.org/10.1016/j.molcel.2006.10.021 PMID: 17188030

Reviewer #2:

Joseph et al. report a single-molecule fluorescence microscopy study that reveals interesting new links between nucleotide excision repair and the translesion synthesis DNA polymerase DnaE2 in non-replicating Caulobacter crescentus cells. By analysing non-replicating cells, the authors are able to examine replication/repair activities that occur outside of the context of the replication fork. The data provide strong support for a functional link between NER and TLS occurring under these conditions. The mechanism they propose is reasonable, however I think that an alternative mechanism that involves homologous recombination would fit their data equally well. Decoupling DnaE2 expression from the RecA*-mediated SOS response should allow the authors to distinguish between the two mechanisms.

1. It seems possible to me that the gaps created by the Uvr proteins might be repaired via homologous recombination, as opposed to gap-filling. The recruitment of SSB, HolB, DnaN, and DnaE to the repair intermediates, and the strong dependence on recA would be consistent with this idea. The requirement for DnaE2 could be consistent with a role in recombination, such as D-loop extension. As the authors point out, the footprint of SSB is larger than the typical gaps produced by UvrABC, so SSB foci should not form on these short gaps. SSB foci might instead form if the initial gaps are enlarged in preparation for homologous recombination (perhaps through the actions of RecQ and RecJ, as has been reported in E. coli). The authors also report robust induction of the SOS response. This requires the formation of RecA* nucleoprotein filaments, such as those formed in preparation for homologous recombination. It might be possible for RecA* to form on short gaps produced by UvrABC, but it seems more likely (and consistent with the formation of SSB foci) that RecA* is forming on gaps that have been enlarged for homologous recombination. The observation that recA is not required for the formation of DnaN foci is consistent with the notion that clamps are loaded at gaps created by UvrABC. It remains formally possible, however, that clamps are loaded at recombination intermediates in the presence of RecA, and at SSB-coated gaps in its absence. Both scenarios would lead to focus formation.

2. Decoupling the production of DnaE2 from the formation of RecA* should allow the authors to distinguish between their proposed gap-filling mechanism and a homologous recombination mechanism. They could do this by expressing DnaE2 from a plasmid, or by introducing a lexA null mutation to make the cells SOS constitutive. Each approach has advantages and disadvantages, so ideally both would be tested. If a gap-filling mechanism is at play, the cells should no longer require recA for resolution of gaps introduced by UvrABC as DnaE2 would already be present. If the damage-independent production of DnaE2 (from a plasmid or in a lexA background) removes the requirement for recA this would support the gap-filling mechanism proposed by the authors. If, on the other hand, the gaps are repaired through homologous recombination, the cells would remain dependent on recA for gap resolution. If resolution does turn out to require recA even when DnaE2 is already present, it would be pertinent to test whether recO is also required for resolution. A requirement for both recA and recO would be strong evidence in support of a homologous recombination mechanism.

Note that if the authors chose to express DnaE2 from a plasmid, this construct could be used to complement DnaE2 function in their chromosomal null and catalytic-dead mutations, which would further solidify their further observations.

Reviewer #3:

This manuscript by Joseph et al. demonstrates that the Caulobacter TLS polymerase DnaE2 plays a role in NER in non-dividing cells. Using cell-biological imaging the authors show that replisome components localize on DNA in non-replicating swarmer cells after treatment with DNA damaging agents. Resolution of these foci requires SOS activation and the catalytic activity of DnaE2. These data contribute to a growing view that TLS polymerases contribute to DNA damage tolerance outside the replisome. In all the manuscript is well-written and properly supported by the data. The authors should consider the following issues:

1. I find it odd that ImuB is never discussed given that it is required for ImuC (DnaE2) mutagenesis. While I don't think it is strictly required for publication it would certainly strengthen the paper if the authors tested whether ImuB is required for the resolution of DnaN foci.

2. Figure 2C – Given the very slow kinetics in Figure 2D is the loss of a focus in 2C really representative? How do the authors ensure that the much faster kinetics seen in 2C is not due to photobleaching resulting from a much faster imaging rate compared to 2D?

[Editors’ note: further revisions were suggested prior to acceptance, as described below.]

Thank you for submitting your article "Coordination between NER and specialized polymerase DnaE2 action enables DNA damage survival in non-replicating bacteria" for consideration by eLife. Your article has been reviewed by 3 peer reviewers, and the evaluation has been overseen by a Reviewing Editor and Kevin Struhl as the Senior Editor. The following individuals involved in review of your submission have agreed to reveal their identity: Andrew Robinson (Reviewer #2); Robert Fuchs (Reviewer #3).

The reviewers have discussed their reviews with one another, and the Reviewing Editor has drafted this to help you prepare a revised submission.

Essential Revisions:

After a substantial discussion, the reviewers agreed, that a substantial amount of new data added to the revised manuscript significantly strengthens the authors' conclusions. The Discussion section, however, needs a careful reframing.

1. The discussion between the reviewers and the editor was centered on the importance and relative frequency of the pathway in which DnaE2 participate. The authors use the term NER quite loosely, but do not actually observing all NER events, rather LPER. We agreed that it is important that the authors clarify in the Discussion that LPER is likely a minor pathway, but important when there is a high density of lesions.

2. The authors need to clearly articulate the differences between the regular NER and long patch excision repair, which likely involves DnaE2, and include references for LPER, which seem to be missing.

3. In the methods section, the authors should include the exposure times. The authors should also comment on what they think would be the shortest event that would produce a focus in their images.

Reviewer #1:

The additional experiments and clarifying text have substantially strengthened the manuscript. The authors have addressed my major concerns in my prior review.

I have only one comment that the authors should consider addressing:

What is the abundance of SSB-YFP relative to endogenous SSB? The authors currently remark that gaps must be >30 nt to enable SSB binding. However if there is a substantial pool of unlabeled SSB present this would suggest that the authors are likely undercounting SSB foci (especially if the gap size is on the order of single tetramer binding). Although not required, it would be interesting if the authors described the intensity of these SSB foci. For example do most foci bleach in a single step consistent with a small gap or instead are they quite bright suggesting much longer ssDNA tracks?

Reviewer #2:

Joseph et al., use a novel approach to investigate the roles of the DNA polymerase DnaE2 in the bacterium Caulobacter crescentus. By generating non-replicating cells and using live-cell imaging to observe key DNA replication/repair proteins, they make a compelling case for the existence of a novel DnaE2 activity. The authors' primary conclusion is that under non-replicating conditions, DnaE2 carries out gap-filling within intermediates of the nucleotide excision repair pathway and that this activity supports the survival of cells treated with DNA damaging agents.

Strengths

The use of non-replicating swarmer cells of Caulobacter crescentus enables the authors to experimentally isolate DnaE2 activities that are independent of DNA replication. The use of live-cell fluorescence imaging enables the authors to directly observe DNA replication and repair proteins as they work on the DNA. By combining the two approaches, the authors reveal a novel activity for DnaE2 that is unlikely to have been discovered in any other way. The imaging results are further supported by traditional genetics approaches.

Weaknesses

Due to technical limitations the main players – DNA gaps, UvrA and DnaE2 – are not directly observed in this study. Instead, the molecular model that the authors propose derives from observations of other DNA replication and repair proteins in cells that maintain, or lack, UvrA and DnaE2. It is important to note that the study generates significant new insight and in my opinion this shortcoming does not outweigh the strengths of this study. Additionally, a number of important mechanistic questions are discussed but have not been explored experimentally in the current study. Why would a specialized DNA polymerase required for a simple gap-filling activity? Why is RecA required for gap filling? Why is RecA so important for cell survival (much more so than DnaE2) under non-replicating conditions where no recombination is possible? There is a high likelihood that these questions will be addressed experimentally in follow-up studies.

Overall the authors are successful in identifying new phenomena that are consistent with a role for DnaE2 in gap-filling of repair intermediates generated by nucleotide excision repair. The primary contribution of the study is that it adds to a growing list of activities for DNA polymerases outside the context of the DNA replication fork – an area that has attracted considerable attention within the DNA repair community over the past five years.

There are quite a few spelling and grammar errors throughout the manuscript. The paper would benefit from careful attention to these. Additionally, in some cases the description of the methods is a little ambiguous. A clear example of this is in the description of the survival assays. As it stands it is not clear that survival was measured using non-replicating swarmer cells. It seems likely that survival was measured using swarmers, however this is not clearly stated in the Methods or Figure Captions. Some minor rewriting to improve the clarity and precision of statements throughout the manuscript would be highly beneficial.

Reviewer #3:

By using an interesting experimental system of Caulobacter swarmer cells, the authors address the mechanism of DNA lesion processing in the absence of replication. Lesions generated by MMC and UV light are known to be essentially repaired by NER; in Caulobacter, as well as in E. coli and other model bacteria, NER entails several steps, (i) lesion recognition and excision by the UvrABC complex, (ii) gap-filling and (iii) ligation. The work mostly focuses on the gap-filling step that is thought to be essentially dependent on DNA PolI (line 422). Surprisingly, the authors' work points to the involvement of polymerase DnaE2 (ImuC) in lesion processing under these non-replicative conditions. DnaE2 was previously shown to be the main polymerase involved in induced mutagenesis in Caulobacter and has thus been categorized as a TLS polymerase. Nevertheless, increasing numbers of reports have shown that, polymerases initially discovered as being involved in lesion bypass and mutagenesis, also participate in other metabolic processes and are now more commonly called specialized DNA polymerases.

General comment: the treatments used in the present paper induce thousands of DNA lesions per bacterial genome (i.e. at 75J/m2 of UV light the number of dimers is ≈3000/bacterial genome). Given the huge number of potential repair events, I wonder if their visualization in the form of fluorescent foci (as aimed in this paper) is realistic? The situation is different for ongoing replication forks for which there are only a few per cell.

This paper requires the following major revisions:

1. The main objective of the paper is to study NER following MMC and UV damage under non-replicative conditions. As mentioned above, DNA synthesis during NER is commonly considered to be performed by PolI. While the paper is focused on the potential involvement of DnaE2, it is difficult to understand why the authors did not investigate the localization of PolI and DnaE2 in parallel? In contrast, the authors use fluorescent fusion strains to visualize DnaN, SSB, DnaE and HolD, that appear all to be less central to NER than DnaE2 and PolI.

2. Role of SSB: the authors conclude that SSB binds to NER-generated gaps; classical NER gaps (12-13bp) are way too short to support binding of even a single SSB tetramer (30nt).

However, a pathway, referred to as Long Patch Excision Repair (LPER), has been described in E. coli (Cooper, 1882; Cooper and Hanawalt, 1972). LPER patches are at least 1500 nt long but represent only about 1% of the classical NER events. The mechanism of formation and the biological significance of these long patches has not been elucidated yet. The size of the patches generated during LPER are compatible with SSB filament formation. It is thus possible that these LPER events are the events that specifically require DnaE2 and the ones that are detected in the present paper. This would also explain why there are only few NER-dependent foci/cell (as noted by the authors on line 480). Additional experimentation is necessary to confirm or exclude the LPER hypothesis.

3. Throughout the paper, the authors appear to consider MMC and UV irradiation as being similar DNA damaging agents. This is probably not accurate, as MMC induces ICL's to a much larger extent than UV-light. This point is seen in Figure 2 suppl.1 A and D: compared to wt, dnaE2 is almost equally sensitive to UV light while it is highly sensitive to MMC. This may point to the specific involvement of DnaE2 in ICL repair.

Cooper, P. K. (1982). Characterization of long patch excision repair of DNA in ultraviolet-irradiated Escherichia coli: an inducible function under rec-lex control. Molecular and General Genetics : MGG, 185(2), 189-197.

Cooper, P. K., and Hanawalt, P. C. (1972). Heterogeneity of patch size in repair replicated DNA in Escherichia coli. Journal of Molecular Biology, 67(1), 1-10.

The possibility that DnaE2 is specifically restricted to filling-in the LPER gaps should be considered. In contrast, the 99% of normal NER patches are filled in by PolI.

Comparison between DnaE2 and PolI fusion strains may show a huge difference in foci formation.

https://doi.org/10.7554/eLife.67552.sa1

Author response

[Editors’ note: the authors resubmitted a revised version of the paper for consideration. What follows is the authors’ response to the first round of review.]

Reviewer #1:

My overall comment is that the data do support neither the title nor the abstract of the paper. In the title "bacterial DNA damage survival" is mentioned while no damage survival data are provided. In the abstract, the authors say they provide evidence for "replication-independent TLS". At the best their data show that a TLS polymerase, namely DnaE2, participate in NER gap-filling which does not mean that DnaE2 performs a lesion bypass. Indeed, a gap generated during NER does in principle not contain a lesion except in the rare situations of closely spaced lesions. If this were indeed the case the authors should observe a quadratic dose response. The involvement of DnaE2 during NER does not mean that DnaE2 is performing Translesion synthesis (TLS). There are more and more situations describing the involvement of TLS polymerases in transactions that do not involve lesion bypass per se. For example, the involvement of Pol kappa has been described many years ago during NER-gap filling in yeast and in human cells without bona fide lesion bypass (see ref below). To assess genuine lesion bypass the authors would need to monitor induced mutagenesis that is the signature of TLS.

We are grateful to Dr. Fuchs for thorough and constructive review of our work. Based on the feedback, we have made extensive revisions to our original manuscript. We address all comments individually below. Here we address three key points raised by the reviewer:

A). Role of specialized DNA polymerases in gap-filling on NER-generated substrates

We would like to start by thanking Dr. Fuchs for raising an important criticism about our usage of the term ‘TLS’ and apologise for the issues stemming from lack of clarity in our writing and interpretation of results. We originally used the term TLS to describe DnaE2 action as it is the only known mutagenic polymerase involved in lesion bypass in the limited number of organisms it has been studied in. However, we agree that the usage of this term can be misleading because it invokes the idea of potential mutagenic lesion bypass. Indeed, the focus of our study is to highlight the gap-filling role of this polymerase on NER-generated substrates.

During the course of our revisions, we came across this excellent review article published during the time of our manuscript revisions (Fuji and Fuchs, Nov 2020, MMBR). Here the authors address issues about TLS definitions, and the need to expand this definition to encompass the idea of ‘specialized polymerases’ in bacteria. Fuji and Fuchs make an important argument that error-prone polymerases can function beyond their canonical role of replication-associated lesion bypass. Thus, a more expansive definition of their action must be considered. For this, the term ‘specialized polymerases’ could be employed to consider error-prone polymerases that may have roles other than replication-associated TLS. Thus, in our revised manuscript, we now use the term ‘specialized polymerase’ or ‘low fidelity polymerase’ to refer to DnaE2 activity in gap-filling, irrespective of the possibility of mutagenic lesion bypass. The steps that result in gap-filling mediated by DnaE2 in non-replicating cells is the central focus of this manuscript.

The aspect of resolving the function of DnaE2 in ‘gap-filling alone’ or ‘lesion bypass + gap-filling’ itself is far more complex and this is not a central point in our currently reported investigations. It would be exciting to resolve the specific function of DnaE2 in gap-filling vs mutagenic lesion bypass. However, the complexity of experiments required to do so precludes our ability to conclusively use our current assay design of non-replicating Caulobacter swarmer cells to answer this question (expanded further in point B). We thus feel that to be able to systematically address the specific point of mutagenic potential of DnaE2 activity, an independent comprehensive study is best done. This is indeed part of forthcoming investigations conducted by a PhD student in the lab. Hence, the current manuscript focusses on the gap-filling role of DnaE2. To the best of our knowledge, in case of DnaE2 (a highly conserved but poorly studied error-prone polymerase) (Wu et al., 2012) ours is one of the first in vivo, cell biological studies to probe its activity in a cell cycle independent context, providing direct evidence for DnaE2 function on NER-generated substrates in non-replicating cells.

We have now significantly rewritten the text (and modified the title and abstract) to refer to gap-filling and avoid speculative implications of DnaE2 action on mutagenesis. Based on the feedback and the literature evidence (including the eukaryotic studies highlighted by Dr. Fuchs), we have rewritten all sections of the text to use the term ‘specialized polymerase’ or ‘low fidelity polymerase’ to describe DnaE2 function in gap-filling, avoiding reference to TLS in our context (for example, L73-74 and other sections of the manuscript). We also provide data on dose-dependent effect of DnaE2 action on replisome persistence as well as survival in non-replicating cells (Figure 4—figure supplement 1C, Figure 5D) (expanded further in points B and C).

B). Why is there need for specialized polymerase during replication-independent NER?

Indeed, a current challenge of our study is to understand why DnaE2 action is required for such gap filling. Given persistence of the replicative polymerase (DnaE) in the absence of DnaE2 at ssDNA gaps (suggesting inability of DnaE to complete synthesis across the same) (Figure 4C), and requirement of synthesis activity of DnaE2 (Figure 4D), in our original manuscript, we had speculated on a role of specific NER-generated substrates, such as COLs, in requiring DnaE2-mediated synthesis.

Based on feedback from the reviewer, we performed additional experiments at various doses of MMC damage and observe a dose-dependent effect of DnaE2 action on replisome persistence (Figure 4—figure supplement 1C) as well as cellular survival (measured by estimating viable count of the population) (point C and Figure 5D). We also conducted assays to measure rates of rifampicin resistance generation (as a measure for mutagenesis) in our experimental setup (Figure 4—figure supplement 1G). We find that increase in mutagenesis in response to damage (independent of presence or absence of replication) arises from DnaE2 activity. However, rifampicin resistance assays are technically very challenging in the non-replicating swarmer cell setup (which requires us to conduct several DnaA depletions and density gradient synchronizations across large culture volumes), and it is also difficult to reliably disentangle the confounding effects of replication during the outgrowth period. Our current setup is unique in that it allows us to be 100% sure that there is no ongoing replication in our system and that cells cannot undergo recombination due to absence of a second chromosome (Figure 1A). Thus, given the complexity of the experimental design and challenges associated with conducting the assays to measure mutagenesis with reliability specifically in non-replicating conditions, we prefer to continue to maintain focus on gap-filling by DnaE2 and the survival advantage it confers to non-replicating cells (toning down our reference to COLs). To resolve above raised question in a comprehensive and reliable manner, extensive work on DnaE2 action (which is poorly studied, unlike E. coli error-prone polymerases) would need to be conducted. This is outside the scope of our current study and does not affect the central observations reported here.

We include new data on replisome foci dynamics as well as cellular survival (measured via CFU assays) in the presence and absence of DnaE2 across a range of MMC doses (Figure 4—figure supplement 1C, Figure 5D and L363 to L383). We also include results from rifampicin resistance assays for calculating mutagenesis, and provide explanation about the difficulty to disentangle the impact of the replication-proficient outgrowth period on mutagenesis observed under damage (Figure 4—figure supplement 1G and L300 to L307). We have ensured to tone-down our speculation on the impact of COLs on the pathway we describe in our current study (and removed the speculative model figure from the manuscript) (L427 to L434). We have maintained the focus on the gap-filling role of DnaE2 and its impact on survival.

C). Impact of gap-filling by specialized polymerases on bacterial DNA damage survival:

In our original manuscript, we used live cell imaging-based read-outs to draw direct conclusions on the importance of coordination of NER with replication-independent DnaE2 activity in enabling survival under DNA damage. In Figure 5C and Results section starting L329 we provided evidence for a survival advantage in cells where gap-filling has been completed. For this, we specifically scored survival as the ability of a cell to undergo one or more divisions after release into replication-permissive conditions, with cell division proficiency being a reliable indicator of cells where repair has been completed (Vickridge et al. 2017). We found that cells that engaged in repair had a 4-fold survival advantage over cells that either lacked DnaE2 or were not given the time to repair in non-replicating phase (Figure 5C). We supported this experiment with evidence on cell length elongation, a hallmark of persistent DNA damage and continued SOS response activation (Wehrens et al., 2018). We showed that cell length continued to increase (without intermittent divisions) in cells that either lacked DnaE2 or were not given a window of recovery to complete gap-filling before release into replication-permissive conditions. However, in cells that were able to complete gap-filling, cell length was restored close to a wild type distribution and cell cycle progression as well as division completion was initiated soon after release into replication-permissive conditions (Figure 5—figure supplement 1A and L348 to L362).

Based on feedback from Dr. Fuchs, we have now additionally carried out survival assays via CFU measurement to reinforce the conclusions outlined above. We now conduct survival analysis (via measuring viable cell count) across a range of MMC doses (0.125, 0.25, 0.5 and 0.75 µg/ml of MMC; further explanation of the doses chosen are given in the next point). In the absence of recovery or DnaE2 action, we find that viable cell count is significantly compromised (Figure 5D and L363 to L383) and no survival advantage is observed when non-replicating cells are released into replication-permissive conditions. In line with our cell biological read-outs, this assay supports the conclusion that enhanced survival at higher doses of DNA damage in non-replicating conditions is dependent on DnaE2 action during recovery. We speculate that these effects are likely observed because of incomplete gap-filling as seen by persistence of SSB foci in conditions where DnaE2-mediated synthesis has been compromised (see discussion L504 onwards).

We include results from survival assays conducted via measurement of viable cell count across various doses of MMC treatment (L363 to L383 and Figure 5D). In addition, we explain our cell biological read-outs for survival with clarity (L329 to L362, Figure 5C and Figure 5—figure supplement 1A).

Substantive concerns:

1. Proper description of the level of damage induced by MMC and UV irradiation should be provided. At the best I can tell the authors only used one dose of MMC and one dose of UV. There needs to be experiments showing the physiological effect of MMC treatment or of UV irradiation such as survival (as claimed in the title) or mutagenesis.

We thank Dr. Fuchs for raising this point about damage doses used. We have now conducted key experiments of across several doses of DNA damage.

In our original manuscript, we had used a sub-inhibitory dose of two distinct DNA damaging agents based on literature evidences (MMC and UV; as described previously in Galhardo et al., 2005 and Modell et al., 2011), which elicited similar responses in our assays. In addition, we had tested two doses of MMC and UV in Figure 2—figure supplement 1F, Figures 4—figure supplements 1C and 1D. We had done so based on cell viability assays to ensure that the doses used show essentiality for DnaE2, but non-lethal for wild type. Given the various types of damages (on the guanine residue) caused by MMC, we are uncomfortable to provide estimate of lesion numbers as this may not be accurate. Hence, we rely on estimating the viable cell count for wild type and dnaE2 deleted cells and choose doses of damage where DnaE2 essentiality is observed, but wild type growth is not severely affected (with similar patterns of cell growth for both UV and MMC doses of damage, Figure 2—figure supplements 1A and 1D). Furthermore, unlike E. coli UmuDC that is thought to preferentially act on UV damage, DnaE2 (which is present in all organisms that lack UmuDC, with mostly GC-rich genomes) preferentially acts on MMC-induced damage. Thus, we specifically focus our study on MMC as it creates the damage substrates that DnaE2 specifically acts on (most likely damage on guanine, as proposed by others and supported by our preliminary unpublished data as well), utilizing UV damage only as a corroborating/ secondary source of damage.

In our revised manuscript, we now provide viable cell count data (for steady-state cells with and without DnaE2) across a range of MMC/ UV treatment used to identify the doses of damage treatment used in our study (Figure 2—figure supplement 1A and 1D), where DnaE2 becomes increasingly essential for survival. Using four doses of MMC (0.125, 0.25, 0.5 and 0.75 µg/ml) and two doses of UV damage (75 and 150 J/m2), we carry out cell biological assays to measure replisome focus formation as well as dissociation during recovery in non-replicating conditions. We also assess persistence of foci at all doses of damage in case of dnaE2 deletion (Figure 4—figure supplements 1C and 1D). Finally, we use genetic read-outs to assess survival via viable cell count assay at all doses of MMC damage in the presence and absence of DnaE2. We find that at low damage dose (0.125 µg/ml MMC), ~10% cells show replisome localization in response to damage and this number reduces to ~2% in a DnaE2-independent manner. Similarly, in the viable cell count assay as well, at low dose of damage survival appears to be independent of DnaE2 action. However, at higher doses of damage, both replisome dissociation and survival advantage in non-replicating conditions is dependent on DnaE2 (Figure 5D). Thus, we now report a dose-dependent effect of DnaE2 on replisome persistence as well as cellular survival in non-replicating conditions.

We include characterization of DNA damage doses used in this study in Figure 2—figure supplements 1A and 1D. We also provide a description of damage doses used and literature evidence for preference of DnaE2 action on MMC damage (L166 to L170 and L197 to L199). We further include data assessing effect of DnaE2 action on replisome foci persistence and impact on survival at multiple doses of damage (Figure 4—figure supplements 1C and 1D, Figure 5D).

2. All data are based on fluorescence imaging. Other methodologies (genetics…..) should be implemented to reinforce the study.

We now include other read-outs to support of the key conclusion of our work (see points discussed above, as well as experiments in response to Reviewer 2 and 3).

Ogi T, Lehmann AR. The Y-family DNA polymerase kappa (pol kappa) functions in mammalian nucleotide-excision repair. Nat Cell Biol. 2006; 8: 640-642. https://doi.org/10.1038/ncb1417 PMID: 16738703

Lehmann AR. New functions for Y family polymerases. Mol Cell. 2006; 24: 493-495. https://doi.org/10.1016/j.molcel.2006.10.021 PMID: 17188030

We have now included relevant references in our revised manuscript.

We are grateful to Dr. Fuchs for the insightful review of our work. We would be happy to provide additional clarifications and experiments as advised.

Reviewer #2:

Joseph et al. report a single-molecule fluorescence microscopy study that reveals interesting new links between nucleotide excision repair and the translesion synthesis DNA polymerase DnaE2 in non-replicating Caulobacter crescentus cells. By analysing non-replicating cells, the authors are able to examine replication/repair activities that occur outside of the context of the replication fork. The data provide strong support for a functional link between NER and TLS occurring under these conditions. The mechanism they propose is reasonable, however I think that an alternative mechanism that involves homologous recombination would fit their data equally well. Decoupling DnaE2 expression from the RecA*-mediated SOS response should allow the authors to distinguish between the two mechanisms.

We are grateful to Dr. Robinson for the positive assessment of our manuscript and for the helpful feedback and suggestions. Dr. Robinson has raised an important point with regards to the mechanism of DnaE2-mediated gap-filling (recombination vs direct synthesis on NERgenerated substrates). We have addressed this comment below.

1. It seems possible to me that the gaps created by the Uvr proteins might be repaired via homologous recombination, as opposed to gap-filling. The recruitment of SSB, HolB, DnaN, and DnaE to the repair intermediates, and the strong dependence on recA would be consistent with this idea. The requirement for DnaE2 could be consistent with a role in recombination, such as D-loop extension. As the authors point out, the footprint of SSB is larger than the typical gaps produced by UvrABC, so SSB foci should not form on these short gaps. SSB foci might instead form if the initial gaps are enlarged in preparation for homologous recombination (perhaps through the actions of RecQ and RecJ, as has been reported in E. coli). The authors also report robust induction of the SOS response. This requires the formation of RecA* nucleoprotein filaments, such as those formed in preparation for homologous recombination. It might be possible for RecA* to form on short gaps produced by UvrABC, but it seems more likely (and consistent with the formation of SSB foci) that RecA* is forming on gaps that have been enlarged for homologous recombination. The observation that recA is not required for the formation of DnaN foci is consistent with the notion that clamps are loaded at gaps created by UvrABC. It remains formally possible, however, that clamps are loaded at recombination intermediates in the presence of RecA, and at SSB-coated gaps in its absence. Both scenarios would lead to focus formation.

2. Decoupling the production of DnaE2 from the formation of RecA* should allow the authors to distinguish between their proposed gap-filling mechanism and a homologous recombination mechanism. They could do this by expressing DnaE2 from a plasmid, or by introducing a lexA null mutation to make the cells SOS constitutive. Each approach has advantages and disadvantages, so ideally both would be tested. If a gap-filling mechanism is at play, the cells should no longer require recA for resolution of gaps introduced by UvrABC as DnaE2 would already be present. If the damage-independent production of DnaE2 (from a plasmid or in a lexA background) removes the requirement for recA this would support the gap-filling mechanism proposed by the authors. If, on the other hand, the gaps are repaired through homologous recombination, the cells would remain dependent on recA for gap resolution. If resolution does turn out to require recA even when DnaE2 is already present, it would be pertinent to test whether recO is also required for resolution. A requirement for both recA and recO would be strong evidence in support of a homologous recombination mechanism.

Note that if the authors chose to express DnaE2 from a plasmid, this construct could be used to complement DnaE2 function in their chromosomal null and catalytic-dead mutations, which would further solidify their further observations.

We thank Dr. Robinson for highlighting this possibility. We realize that we should have discussed the same in detail as this was an important consideration we had already taken into account while designing our assay system. In our assays, we ensured that the second copy of the chromosome is unavailable in cells during DNA damage recovery (all cells have only 1n chromosome content; Figure 1A). As stated in results L130, L134, and Figure 1A in our revised manuscript, we isolate Caulobacter swarmer cells that have single chromosomes. Thus, the possibility of recombination is ruled out in this system, suggesting strongly that the TLS polymerase, DnaE2 alone is sufficient to carry out gap-filling on NER-generated substrates. We have now clearly considered this possibility in the main text and provided evidence ruling the same out (L251 onwards).

To further support the idea that DnaE2 directly participates in gap-filling on NER-generated substrates, we also conducted our experiments in cells deleted for recN. This repair protein is essential for RecA-mediated recombination repair in Caulobacter (Badrinarayanan et al., 2015). We find that neither association nor dissociation of the β-clamp is perturbed in the absence of recN (with dynamics as seen in wild type conditions), ruling out a role for recombination in the process described in our manuscript. These data are now included in the manuscript (L257 and Figure 4—figure supplement 1A)

In addition to the above evidences, we also considered the experimental suggestions made by Dr. Robinson. The lexA constitutively active mutant prevents the ability for us to conduct synchronization experiments to isolate non-replicating swarmer cells with a single chromosome. Thus, we prefer not to conduct this experiment as it would make our experimental system noisy. Our current setup is unique in that it allows us to be 100% sure that there is no ongoing replication in our system and that cells cannot undergo recombination due to absence of a second chromosome. Secondly, while the complementation experiment would be a nice method to additionally rule out a role for recombination, complementation of DnaE2 alone would be insufficient. Under the SOS response, three components for Caulobacter TLS are expressed (ImuA, ImuB and ImuC (DnaE2)), all of which are essential for successful synthesis by DnaE2. Indeed, in our revised manuscript we provide evidence for the same (based on feedback from Reviewer 3) (Figure 4D). Thus, we would need to inducibly express all three components at appropriate ratios. This has been extremely challenging till date (from ours and others labs, likely given the large size of the entire operon and presence of one regulatory RNA whose function is currently unknown).

Indeed, for the current revisions as well we attempted to construct such a tool, but faced severe technical challenges in the process. We feel unfortunate, but are confident that our swarmer cell setup (with a single chromosome) as well as recN deletion experiments address the point raised by Dr. Robinson.

We now explain our non-replicating, single chromosome, swarmer cell setup with clarity (Figure 1A and L130 onwards) to rule out role for recombination in the DnaE2-mediated gap-filling process (L251 onwards). We also provide data for cells showing recovery in the absence of the recombination protein RecN (L254 to 258 and Figure 4—figure supplement 1A).

Reviewer #3:

This manuscript by Joseph et al. demonstrates that the Caulobacter TLS polymerase DnaE2 plays a role in NER in non-dividing cells. Using cell-biological imaging the authors show that replisome components localize on DNA in non-replicating swarmer cells after treatment with DNA damaging agents. Resolution of these foci requires SOS activation and the catalytic activity of DnaE2. These data contribute to a growing view that TLS polymerases contribute to DNA damage tolerance outside the replisome. In all the manuscript is well-written and properly supported by the data. The authors should consider the following issues:

1. I find it odd that ImuB is never discussed given that it is required for ImuC (DnaE2) mutagenesis. While I don't think it is strictly required for publication it would certainly strengthen the paper if the authors tested whether ImuB is required for the resolution of DnaN foci.

We thank the reviewer for raising this point. We did not discuss the role of ImuB as we had considered that characterization of the same was outside the scope of our current study. However, based on feedback from the reviewer, we have now conducted our experiments in cells lacking imuB as well. We find that, as in the case of DnaE2, ImuB is also essential for β-clamp dissociation during DNA damage recovery (Figure 4D and L308 to L316). In addition, we highlight the importance of ImuB for DnaE2 function (L66, L309 and L476). We have also confirmed physical interaction between DnaN and ImuB via a bacterial two-hybrid assay and would be happy to provide these data if additionally required.

2. Figure 2C – Given the very slow kinetics in Figure 2D is the loss of a focus in 2C really representative? How do the authors ensure that the much faster kinetics seen in 2C is not due to photobleaching resulting from a much faster imaging rate compared to 2D?

This is an important point raised by the reviewer. Unfortunately, the experimental regimes are not directly comparable between Figure 2C and Figure 2D. Thus, given that these data are not central to our current manuscript, we have removed this figure panel (2C) from the revised manuscript presently as it does not contribute significantly to the conclusions of the study. We maintain consistent focus on the analysis over longer timescales (Figure 2 and elsewhere throughout the manuscript).

[Editors’ note: what follows is the authors’ response to the second round of review.]

Essential Revisions:

After a substantial discussion, the reviewers agreed, that a substantial amount of new data added to the revised manuscript significantly strengthens the authors' conclusions. The Discussion section, however, needs a careful reframing.

We are very grateful for the thoughtful consideration of our work and for guidance with regards to essential revisions we need to make to our manuscript.

1. The discussion between the reviewers and the editor was centered on the importance and relative frequency of the pathway in which DnaE2 participate. The authors use the term NER quite loosely, but do not actually observing all NER events, rather LPER. We agreed that it is important that the authors clarify in the Discussion that LPER is likely a minor pathway, but important when there is a high density of lesions.

We thank the Editor and Reviewers for highlighting this critical point. We agree that we are not observing all NER events and that it is important to consider the possibility that DnaE2 is required only for a subset of NER events that involves gap-filling across long patches (likely at higher doses of damage). Hence, based on this feedback, we have revised our Discussion section to discuss this point with clarity (L433 to L448). We include LPER as a possible scenario where DnaE2 activity may be required, in addition to the alternate model that would involve generation of long ssDNA gaps at problematic intermediates of NER. We state that a limitation of our study is that we are not tracking all NER events (a significant proportion of which could be mediated via gap-filling by PolI across short patches of ssDNA) and that the mechanism by which long ssDNA gaps are generated from a subset of NER events remains unclear (L449 to L453). Our future efforts are aimed at understanding why there is a need for specialized polymerase activity on some NER-generated substrates at higher doses of damage.

2. The authors need to clearly articulate the differences between the regular NER and long patch excision repair, which likely involves DnaE2, and include references for LPER, which seem to be missing.

We have now included reference to LPER. In the introduction (L42) and results (L227) sections we briefly mention the possibility of LPER. In the discussion (L420 onwards), we consider the differences between regular NER and long patch repair and discuss the specific involvement of DnaE2 in long patch repair at higher doses of damage.

3. In the methods section, the authors should include the exposure times. The authors should also comment on what they think would be the shortest event that would produce a focus in their images.

We have revised this section of the manuscript to provide information on exposure times (L571).

We apologise if we have misunderstood the query with regards to focus formation in our experimental regime. We address this comment in two parts and would be happy to provide additional clarifications if required. (1) With regards to stability, we are confident that the foci we observe are likely bound molecules in stable structures, as reported for replisome components in previous studies pertaining to DNA replication and repair in E. coli (Soubry et al., 2019) based on the following observations: In our experiments, we find that varying exposure times (50-400 ms) do not influence the number of foci observed per cell under damage (and no foci are detected in the absence of damage). In support, we also note that once formed, a localization is observed for 6 min on average before dissociation. Since our imaging setup does not have the resolution to allow us to comment on binding kinetics with precision (as we are not detecting single molecule binding events and transient binding events), we prefer not to make conclusions about the same. We would also like to highlight that focus dissociation was repair-dependent (as assessed by lack of dissociation of individual foci in absence of dnaE2). (2). With respect to duration of damage events, we did conduct our experiments at short exposures to UV damage as well. We observed DnaN foci at 30 J/m2 UV exposure (dose used in Henrikus et al., 2018; E. coli study tracking TLS activity under damage via live cell imaging) and 50 J/m2 (dose used in Soubry et al., 2019; E. coli study tracking replisome activity under damage via live cell imaging). As anticipated, the percentage of cells with foci increased with increasing dose of damage. For the experiments reported in our current work, we chose a regime of damage where activity of DnaE2 is required and wild type cells can repair damage (with minimal impact on cell death).

References

Henrikus, S. S., Wood, E. A., McDonald, J. P., Cox, M. M., Woodgate, R., Goodman, M. F., van Oijen, A. M., and Robinson, A. (2018). DNA polymerase IV primarily operates outside of DNA replication forks in Escherichia coli. PLoS Genetics, 14(1), e1007161. https://doi.org/10.1371/journal.pgen.1007161

Soubry, N., Wang, A., and Reyes-Lamothe, R. (2019). Replisome activity slowdown after exposure to ultraviolet light in Escherichia coli. Proceedings of the National Academy of Sciences, 201819297. https://doi.org/10.1073/pnas.1819297116

https://doi.org/10.7554/eLife.67552.sa2

Article and author information

Author details

  1. Asha Mary Joseph

    National Centre for Biological Sciences - Tata Institute of Fundamental Research, Bangalore, India
    Contribution
    Conceptualization, Resources, Data curation, Formal analysis, Funding acquisition, Validation, Investigation, Visualization, Methodology, Writing - original draft, Writing - review and editing
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0002-0465-9799
  2. Saheli Daw

    National Centre for Biological Sciences - Tata Institute of Fundamental Research, Bangalore, India
    Contribution
    Resources, Data curation, Formal analysis, Validation, Investigation, Visualization, Writing - review and editing
    Competing interests
    No competing interests declared
  3. Ismath Sadhir

    1. National Centre for Biological Sciences - Tata Institute of Fundamental Research, Bangalore, India
    2. Max Planck Institute for Terrestrial Microbiology, LOEWE Centre for Synthetic Microbiology (SYNMIKRO), Marburg, Germany
    Contribution
    Resources, Formal analysis, Validation, Writing - review and editing
    Competing interests
    No competing interests declared
  4. Anjana Badrinarayanan

    National Centre for Biological Sciences - Tata Institute of Fundamental Research, Bangalore, India
    Contribution
    Conceptualization, Data curation, Formal analysis, Supervision, Funding acquisition, Validation, Investigation, Visualization, Methodology, Writing - original draft, Project administration, Writing - review and editing
    For correspondence
    anjana@ncbs.res.in
    Competing interests
    No competing interests declared
    ORCID icon "This ORCID iD identifies the author of this article:" 0000-0001-5520-2134

Funding

Human Frontier Science Program (00051/ 2017-C)

  • Anjana Badrinarayanan

Department of Atomic Energy, Government of India (12-R&D-TFR-5.04-0800)

  • Anjana Badrinarayanan

Department of Science and Technology, Ministry of Science and Technology (PDF/2018/001164)

  • Asha Mary Joseph

Department of Biotechnology , Ministry of Science and Technology (IYBA(BT/12/IYBA/2019/10))

  • Anjana Badrinarayanan

CSIR

  • Saheli Daw

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

The authors acknowledge assistance from Prachi Shinde and Dr. Ramya Rajagopalan for preliminary experiments and NCBS Central Imaging and Flow Facility (CIFF) for flow cytometry usage. The authors thank Dr. Rodrigo Reyes-Lamothe, Dr. Sabari Tirupathy, Dr. Stephan Uphoff, Dr. Tung Le, and members of the AB lab for helpful discussions and feedback on the manuscript. The authors are grateful for the constructive feedback and suggestions from the reviewers and editors. This work was supported by fellowships from DBT (DBT-RA), DST-SERB (PDF/2018/001164) (AMJ), and CSIR (SD) as well as grants from DBT-IYBA (BT/12/IYBA/2019/10), HFSP CDA (00051/2017 C), and intramural funding from NCBS-TIFR (Department of Atomic Energy, Government of India, under project no. 12 R and D-TFR-5.04–0800) (AB).

Senior Editor

  1. Kevin Struhl, Harvard Medical School, United States

Reviewing Editor

  1. Maria Spies, University of Iowa, United States

Reviewers

  1. Andrew Robinson, University of Wollongong, Australia
  2. Robert Fuchs

Publication history

  1. Received: February 15, 2021
  2. Accepted: April 14, 2021
  3. Accepted Manuscript published: April 15, 2021 (version 1)
  4. Version of Record published: May 6, 2021 (version 2)

Copyright

© 2021, Joseph et al.

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.

Metrics

  • 528
    Page views
  • 59
    Downloads
  • 0
    Citations

Article citation count generated by polling the highest count across the following sources: Crossref, PubMed Central, Scopus.

Download links

A two-part list of links to download the article, or parts of the article, in various formats.

Downloads (link to download the article as PDF)

Download citations (links to download the citations from this article in formats compatible with various reference manager tools)

Open citations (links to open the citations from this article in various online reference manager services)

Further reading

    1. Cell Biology
    David W Sanders et al.
    Research Article Updated

    Many enveloped viruses induce multinucleated cells (syncytia), reflective of membrane fusion events caused by the same machinery that underlies viral entry. These syncytia are thought to facilitate replication and evasion of the host immune response. Here, we report that co-culture of human cells expressing the receptor ACE2 with cells expressing SARS-CoV-2 spike, results in synapse-like intercellular contacts that initiate cell-cell fusion, producing syncytia resembling those we identify in lungs of COVID-19 patients. To assess the mechanism of spike/ACE2-driven membrane fusion, we developed a microscopy-based, cell-cell fusion assay to screen ~6000 drugs and >30 spike variants. Together with quantitative cell biology approaches, the screen reveals an essential role for biophysical aspects of the membrane, particularly cholesterol-rich regions, in spike-mediated fusion, which extends to replication-competent SARS-CoV-2 isolates. Our findings potentially provide a molecular basis for positive outcomes reported in COVID-19 patients taking statins and suggest new strategies for therapeutics targeting the membrane of SARS-CoV-2 and other fusogenic viruses.

    1. Cell Biology
    2. Structural Biology and Molecular Biophysics
    Agata Szuba et al.
    Research Article Updated

    Septins are conserved cytoskeletal proteins that regulate cell cortex mechanics. The mechanisms of their interactions with the plasma membrane remain poorly understood. Here, we show by cell-free reconstitution that binding to flat lipid membranes requires electrostatic interactions of septins with anionic lipids and promotes the ordered self-assembly of fly septins into filamentous meshworks. Transmission electron microscopy reveals that both fly and mammalian septin hexamers form arrays of single and paired filaments. Atomic force microscopy and quartz crystal microbalance demonstrate that the fly filaments form mechanically rigid, 12- to 18-nm thick, double layers of septins. By contrast, C-terminally truncated septin mutants form 4-nm thin monolayers, indicating that stacking requires the C-terminal coiled coils on DSep2 and Pnut subunits. Our work shows that membrane binding is required for fly septins to form ordered arrays of single and paired filaments and provides new insights into the mechanisms by which septins may regulate cell surface mechanics.