Gamblers: An Antibiotic-Induced Evolvable Cell Subpopulation Differentiated by Reactive-Oxygen-Induced General Stress Response

Abstract

Antibiotics can induce mutations that cause antibiotic resistance. Yet, despite their importance, mechanisms of antibiotic-promoted mutagenesis remain elusive. We report that the fluoroquinolone antibiotic ciprofloxacin (cipro) induces mutations by triggering transient differentiation of a mutantgenerating cell subpopulation, using reactive oxygen species (ROS). Cipro-induced DNA breaks activate the Escherichia coli SOS DNA-damage response and error-prone DNA polymerases in all cells. However, mutagenesis is limited to a cell subpopulation in which electron transfer together with SOS induce ROS, which activate the sigma-S (s S ) general-stress response, which allows mutagenic DNA-break repair. When sorted, this small s S -response-''on'' subpopulation produces most antibiotic cross-resistant mutants. A U.S. Food and Drug Administration (FDA)-approved drug prevents s S induction, specifically inhibiting antibiotic-promoted mutagenesis. Further, SOS-inhibited cell division, which causes multi-chromosome cells, promotes mutagenesis. The data support a model in which within-cell chromosome cooperation together with development of a ''gambler'' cell subpopulation promote resistance evolution without risking most cells. RESULTS Cipro-Induced MutagenesisWe developed two assays for cipro-induced mutagenesis without cipro selection of the mutants (Figure 1A). In both assays, strains are grown in liquid, each with cipro at its minimum antibiotic concentration (MAC, final colony-forming units [CFU] are 10% of those of no-drug cultures) (Lorian and De Freitas, 1979). These are ''low-dose'' and ''sub-inhibitory'' relative to MICs (CFU %10 À4 of untreated cells). Table S1 shows MACs and MICs for all strains assayed (wild-type MAC, 8.5 ng/mL). Cells are then removed from cipro and plated selectively for colonies resistant to rifampicin (RifR) or ampicillin (AmpR) antibiotics (Figure 1A), and mutation rates are estimated (STAR Methods). RifR arises by specific base-substitutions in the rpoB gene (Figure S1A), and AmpR arises by ampD null mutations in engineered Escherichia coli (Petrosino et al., 2002) (Figures S1B and S1C; STAR Methods). Strikingly, cipro increased RifR and AmpR mutation rates 26-and 18-fold above no-cipro rates (Figure 1B; Table S2 for all mutation rates). The RifR or AmpR mutants are not selected in sub-inhibitory cipro and are at a slight but significant disadvantage (Figure 1C), implying that mutation, not selection of the mutants, is elevated by MAC cipro. Additional controls show negligible cell death in the lowdose cipro (Figures S1D and S2, other controls). ROS-Dependent Mutagenesis Is s S -Dependent Mutagenic Break RepairThe cipro-induced mutagenesis requires ROS and is inhibited by ROS scavenging or preventing agents thiourea (TU) and 2,2 0 -bipyridine (BP) (Figure 1D; Table S2). The following data indicate that the ROS instigate a s S -licensed mutagenic DNA breakrepair (MBR) mechanism triggered by cipro-induced DSBs.MBR is regulated mutagenesis during repair of DSBs, requiring the SOS and s S respon...

Irregularities in genetic variation and mutation rates with environmental stresses

“…A possibility in E. coli is a potential nutrient specific role of RpoE, another sigma factor impacting on SIM (Gibson et al, 2010). Another possibility is a significant difference in DNA-reactive oxygen species under different nutritional states, as found with some antibioticstressed bacteria (Hoeksema et al, 2018;Pribis et al, 2019). It should be noted that lack of a SIM with some stresses is not restricted to nutritional stresses and there are other examples in the literature where stress does not increase mutation rates (Lilja and Johnson, 2017).…”

“…Beyond the lac system, antibiotic-induced mutagenesis has been investigated and shown to have some similar properties to starvation stress induction with some antibiotics (Gutierrez et al, 2013;Blázquez et al, 2018), involving double-strand breaks and error prone repair (Pribis et al, 2019). The effect of additional environmental variables, including other nutritional stresses, was recently characterized and shown to induce variation in mutational outcomes (Maharjan and Ferenci, 2017b;Shewaramani et al, 2017).…”

Fighting the spread of antibiotic resistance with bacterial competence inhibitors

“…S7) 50 , we constructed two strains in which the entire comCDE operon was under IPTG-inducible control, but in which either the wild-type comD or comE genes were exchanged for two constitutively active mutants. The first mutant, ComD T233I (strain ADP272), constitutively phosphorylates ComE independent of the presence of CSP 30 , while the second, ComE D58E (strain ADP148), mimics the phosphorylated form of ComE, thus inducing competence even in the absence of CSP or ComD 35 . ComD T233I showed a hypercompetent phenotype in presence of IPTG, though TCL was still able to reduce competence activation to nearly half the luminescence production when ComD T233I was induced (Fig.…”

“…Such adjuvants can also go beyond the realm of common drugs to food compounds 28 . Other ideas involve searching for compounds that could reduce the evolution of resistance development, 29,30 or the use of quorum sensing analogous as modulators of pathogenic behaviours and players in biofilm formation 31 . Here, we propose the use of coadjuvants to prevent transformation and the associated spread of resistant genes and virulence factors by inhibiting pneumococcal competence.…”

Exploiting epistasis to perturb the evolution of antibiotic resistance

“…The idea of interfering with resistance evolvability by altering the rate of acquisition of resistance by mutation or horizontal gene transfer is promising and has been investigated in 20 depth (13, [22][23][24][25][26]. Efforts to alter evolvability by other means, while equally promising, have received less attention (16,27).…”