DNA supercoiling differences in bacteria result from disparate DNA gyrase activation by polyamines

Abstract

DNA supercoiling is essential for all living cells because it controls all processes involving DNA. In bacteria, global DNA supercoiling results from the opposing activities of topoisomerase I, which relaxes DNA, and DNA gyrase, which compacts DNA. These enzymes are widely conserved, sharing >91% amino acid identity between the closely related species Escherichia coli and Salmonella enterica serovar Typhimurium. Why, then, do E. coli and Salmonella exhibit different DNA supercoiling when experiencing the same conditions? We now report that this surprising difference reflects disparate activation of their DNA gyrases by the polyamine spermidine and its precursor putrescine. In vitro, Salmonella DNA gyrase activity was sensitive to changes in putrescine concentration within the physiological range, whereas activity of the E. coli enzyme was not. In vivo, putrescine activated the Salmonella DNA gyrase and spermidine the E. coli enzyme. High extracellular Mg2+ decreased DNA supercoiling exclusively in Salmonella by reducing the putrescine concentration. Our results establish the basis for the differences in global DNA supercoiling between E. coli and Salmonella, define a signal transduction pathway regulating DNA supercoiling, and identify potential targets for antibacterial agents.

About the Authors:

Alexandre Duprey

Roles Conceptualization, Formal analysis, Investigation, Visualization, Writing – original draft

Affiliation: Department of Microbial Pathogenesis, Yale School of Medicine, New Haven, CT, United States of America

ORCID logo https://orcid.org/0000-0001-8310-8508

Eduardo A. Groisman

Roles Conceptualization, Funding acquisition, Supervision, Writing – review & editing

* E-mail: [email protected]

Affiliations Department of Microbial Pathogenesis, Yale School of Medicine, New Haven, CT, United States of America, Yale Microbial Sciences Institute, West Haven, CT, United States of America

ORCID logo https://orcid.org/0000-0001-6860-7691

Introduction

All living cells supercoil their DNA. The resulting changes in DNA structure enable compaction of the DNA so that it fits within the limited space of a bacterial cell or eukaryotic organelle [1]. DNA supercoiling is necessary for transcription in prokaryotes [2,3] and eukaryotes [4], recombination in bacteriophages [5,6] and eukaryotic viruses, including HIV [7], and chromosome segregation [8]. Because some of the enzymes governing bacterial DNA supercoiling are essential, their pharmacological inhibition is an efficient antibacterial strategy [9,10].

Global DNA supercoiling in bacteria results from the opposing activities of topoisomerase I, which relaxes DNA, and DNA gyrase, which compacts DNA. (Although DNA topoisomerase IV also changes DNA supercoiling, its contribution is negligible compared to that of DNA gyrase under physiological conditions [11].) The global DNA supercoiling effects of these enzymes are in addition to the primarily local effects of gene transcription and DNA replication [12] and nucleoid structuring proteins [13].

Formation of negative supercoiled DNA by the purified DNA gyrase from Escherichia coli requires Mg2+, adenosine triphosphate (ATP) and spermidine [14], the most abundant polyamine in bacteria [15]. Mg2+ and ATP are required for strand cleavage and strand passage, respectively [16,17], but the specific role that spermidine plays has remained unknown. By contrast, neither ATP nor polyamines are required for activity of the investigated bacterial topoisomerase I [18], and its eukaryotic equivalent does not even require Mg2+ [19]. Therefore, topoisomerase I activity is currently considered to be independent of metabolic co-factors [9]. Synthesis of topoisomerase I is controlled by a homeostatic mechanism whereby an increase in negatively supercoiled DNA promotes transcription of the topoisomerase I-encoding topA gene [20]. It is currently unknown whether specific metabolites control the activities of DNA gyrase and topoisomerase I in living cells.

Polyamines are a family of positively charged molecules consisting of aliphatic hydrocarbon chains harboring several amine groups [21]. Polyamines are found in all living organisms except some Archaea (i.e., Methanobacteriales and Halobacteriales [22]). The three most common polyamines are spermidine, putrescine and spermine. The abundance of individual polyamines often differs across organisms. For example, E. coli has high amounts of spermidine and its precursor putrescine but lacks the spermidine-derived spermine [15]. Polyamines have RNA- and DNA-binding capacities [23]. In bacteria, they have been implicated in translation [24], protection against acidic [25] and oxidative [26] stresses, iron scavenging [27], and virulence [28,29].

Salmonella enterica serovar Typhimurium (S. Typhimurium) supercoils its DNA less than E. coli under basal growth conditions [30]. This is surprising because these two bacterial species are closely related [31] and also because their TopA (topoisomerase I) proteins share 96% amino acid identity, and their GyrA and GyrB proteins, which constitute the two subunits of DNA gyrase, share 91% and 97% identity, respectively. Differences in the latter proteins are responsible, in part, for the distinct global DNA supercoiling behaviors of the two species because: (i) identical mutations in gyrA [32] or gyrB [33] have different effects on the two species; (ii) the E. coli DNA gyrase is faster than the S. Typhimurium enzyme in vitro; and (iii) DNA is less negatively supercoiled in an E. coli strain expressing a chimeric GyrA protein consisting of the E. coli GyrA with its 35 C-terminal residues replaced by the corresponding region from the S. Typhimurium GyrA than in an E. coli strain expressing the wild-type E. coli GyrA [30].

Here, we report that in vivo, putrescine and spermidine specifically activate the E. coli and S. Typhimurium DNA gyrases, respectively. We identify an environmental condition that controls DNA supercoiling in S. Typhimurium by decreasing putrescine abundance and establish that this condition does not alter DNA supercoiling in E. coli. Our findings suggest new targets to disrupt DNA supercoiling specifically in bacterial pathogens without harming commensal bacterial species or humans.

Results

Spermidine and putrescine stimulate DNA supercoiling in E. coli

Because activation of the E. coli DNA gyrase in vitro requires spermidine [14], we wondered whether spermidine plays such a role in vivo. Thus, we examined the DNA supercoiling status of a reporter plasmid in isogenic wild-type and speE E. coli using the classical agarose/chloroquine gel method [34]. The speE gene specifies a protein that converts putrescine into spermidine, and spermidine into spermine [35] (Fig 1). Therefore, a speE null mutant is unable to make spermidine and spermine (Fig 1).

[Figure omitted. See PDF.]

Fig 1. The putrescine biosynthetic and degradative pathways in S. Typhimurium and E. coli.

Black and gray arrows represent biochemical reactions carried out by the protein(s) indicated next to them. Black letters indicate proteins investigated in this work, gray letters indicate other proteins related to putrescine abundance. Some intermediate metabolites are omitted for clarity. Blue arrows denote changes in the expression of genes mediating putrescine import, export, biosynthesis or degradation resulting from growth of S. Typhimurium in different Mg2+ concentrations as measured by the RNA-seq experiment discussed below (Table 1). Bold blue arrows denote pathways upregulated during growth in 10 mM Mg2+ compared to 0.8 mM Mg2+. Slim light blue arrows denote pathways downregulated during growth in 10 mM Mg2+ compared to 0.8 mM Mg2+. *: the PuuABCD pathway is present in E. coli but not in S. Typhimurium.

https://doi.org/10.1371/journal.pgen.1009085.g001

[Figure omitted. See PDF.]

Table 1. Extracellular Mg2+ controls the mRNA amounts of genes governing putrescine abundance in S. enterica.

https://doi.org/10.1371/journal.pgen.1009085.t001

DNA was more negatively supercoiled in wild-type E. coli than in the isogenic speE mutant (Fig 2A), suggesting that spermidine and/or spermine increases negative DNA supercoiling. We ascribe this activity to spermidine because spermine was not detected under conditions in which spermidine and putrescine were (Fig 3), in agreement with published results [15].

[Figure omitted. See PDF.]

Fig 2. Putrescine increases negative DNA supercoiling in vivo.

DNA supercoiling in the indicated E. coli (A) and S. Typhimurium (B) strains, which differ in putrescine abundance. E. coli strains used were: MG1655 (wild-type), AAD85 (speE), AAD96 (speE oat puuA), AAD95 (speB speC speF). S. Typhimurium strains used were: 14028s (wild-type), AAD46 (speE), JY979 (oat), AAD58 (speE oat), EG6501 (sapABCDF), AAD212 (speB argE). Coloured circles qualitatively represent expected concentrations of putrescine (blue) or spermidine (green). Different lowercase letters indicate pairwise statistical significance (Tukey’s HSD, p<0.05, n = 7 for S. Typhimurium wild type, n = 4 for S. Typhimurium speE, n = 2 otherwise). The growth curves corresponding to the cultures used in these experiments are presented in S1 Fig.

https://doi.org/10.1371/journal.pgen.1009085.g002

[Figure omitted. See PDF.]

Fig 3. Intracellular concentrations of polyamines determined.

A,B: Concentrations of putrescine (A), spermidine (B) and spermine were assessed by HPLC in exponentially growing wild-type S Typhimurium strain 14028s or the isogenic speE oat mutant (AAD58) in N-minimal medium with the indicated Mg2+ concentrations. Stars denote statistical significance compared to wild-type in 800 μM Mg2+. ***: p<0.001; *: p<0.05; ns: not significant (Student’s t-test, n = 3). C,D: Same as above but for wild-type E. coli strain MG1655. None of the tested polyamines was detectable in the speB speC speF mutant (AAD95). Spermine was not detected under the investigated conditions (detection limit = 0.2 mM).

https://doi.org/10.1371/journal.pgen.1009085.g003

The oat and puuA genes specify proteins that convert putrescine into 4-aminobutanoate (Fig 1). Therefore, the oat puuA speE triple mutant is predicted to accumulate putrescine but lack spermidine and spermine. DNA was more negatively supercoiled in the oat puuA speE triple mutant than in the speE single mutant or in wild-type E. coli (Fig 2A), suggesting that putrescine can increase global DNA supercoiling in vivo.

Putrescine is produced from ornithine by either the SpeF or SpeC proteins, depending on the environmental condition, notably extracellular pH [36], or from agmatine by the SpeB protein (Fig 1). The speB speC speF triple mutant displayed the lowest negative DNA supercoiling of the investigated E. coli strains (Fig 2A) as it is unable to make putrescine, spermidine and spermine (Fig 1). Cumulatively, the results in this section indicate that both putrescine and spermidine promote DNA supercoiling in E. coli, and that spermidine plays a dominant role in this process.

Putrescine stimulates DNA supercoiling in S. Typhimurium, but spermidine has a negligible effect

The SpeE proteins of S. Typhimurium and E. coli share 94% amino acid identity, which is higher than the 90% median amino identity between proteins present in the two species [37]. Therefore, inactivation of the speE gene was anticipated to result in analogous phenotypes in S. Typhimurium and E. coli: decreased negative DNA supercoiling. Unexpectedly, DNA was actually more negatively supercoiled in the S. Typhimurium speE mutant than in the wild-type parent (Fig 2B). This is the exact opposite phenotype observed in E. coli upon speE inactivation (Fig 2A). These results suggest that putrescine is more important for DNA supercoiling in S. Typhimurium than in E. coli.

That putrescine promotes DNA supercoiling in S. Typhimurium is further supported by the phenotypes of mutants in pathways predicted to alter putrescine abundance. For instance, DNA was more negatively supercoiled in the oat speE double mutant than in oat or speE single mutants. In turn, DNA was more negatively supercoiled in the latter two mutants than in wild-type S. Typhimurium (Fig 2B). Deletion of the putrescine exporter-encoding sapABCDF operon [38] increased negative DNA supercoiling more than deletion of the oat and speE genes (Fig 2A). By contrast, DNA was less negatively supercoiled in a speB argE double mutant than in wild-type S. Typhimurium (Fig 2B), likely due to the inability of the mutant to synthesize putrescine and spermidine (Fig 1).

Because the S. Typhimurium speB speC speF triple mutant had a severe growth defect in minimal media, it was not possible to determine its DNA supercoiling status. This is in contrast to the E. coli speB speC speF triple mutant, which grew well enough to measure DNA supercoiling, albeit not as well as wild-type E. coli. Therefore, mutations in homologous genes that compromise putrescine abundance result in different phenotypes in S. Typhimurium and E. coli.

The results in this section indicate that putrescine strongly stimulates DNA supercoiling in S. Typhimurium and that the contribution of spermidine to this process is minimal. Moreover, they raise the possibility that the different basal global DNA supercoiling status of S. Typhimurium and E. coli [39] result from disparate abundances of specific polyamines in the two species, and/or from the response of their respective topoisomerases or DNA gyrases to these polyamines. As stated above, differences in DNA gyrase activity contribute to the distinct basal DNA supercoiling of S. Typhimurium and E. coli [30].

The concentrations of putrescine, spermidine, and spermine are similar in S. Typhimurium and E. coli

We determined that wild-type S. Typhimurium strain 14028s harbors 20 mM putrescine, 4.5 mM spermidine, and no detectable spermine when grown in N-minimal media with glycerol as carbon source and 0.8 mM Mg2+ (Fig 3), a concentration similar to that of free cytoplasmic Mg2+ [40]. The abundance of these polyamines is in contrast to those reported for a different S. Typhimurium strain grown in complex LB medium in which putrescine and spermine were detected but spermidine was not [41].

Wild-type E. coli strain MG1655 has polyamine concentrations similar to those of S. Typhimurium (15 mM putrescine, 2.5 mM spermidine, no detectable spermine) when grown in the same medium (Fig 3). The latter concentrations are lower than those reported for the E. coli strain W3110 grown in nutrient broth, albeit in the same order of magnitude [23,42].

The speE oat double mutant had more total putrescine (38 mM) than wild-type S. Typhimurium (20 mM) and no detectable spermidine (Fig 3). This result makes physiological sense given that the speE oat mutant is defective in both putrescine degradation and spermidine synthesis (Fig 1).

Cumulatively, the data in this section argue against the possibility of the disparate DNA supercoiling phenotypes exhibited by S. Typhimurium and E. coli upon speE inactivation resulting from differences in polyamine abundance between the two species (Fig 2). Moreover, they suggest that the differences in basal negative DNA supercoiling between S. Typhimurium and E. coli result from the response of their respective DNA gyrases and/or topoisomerases to particular polyamines.

The S. Typhimurium and E. coli DNA gyrases are differentially activated by individual polyamines

The genetic data presented above argue that negative DNA supercoiling is stimulated by putrescine in S. Typhimurium and by spermidine in E. coli (Fig 2). Thus, we examined the ability of these polyamines to activate the purified DNA gyrases from the two species. To guide the design of an in vitro DNA gyrase activation assay using polyamine concentrations in the physiological range, we considered the affinity constants between polyamines and various cellular components and determined that only ~40% of putrescine and ~5% of spermidine are free in the cell [23,42] (i.e., not bound to DNA, RNA, ATP or phospholipids). If the proportion of free polyamines is the same in S. Typhimurium and E. coli, then, based on our measurements of total polyamines (Fig 3), the concentrations of free putrescine and spermidine are estimated to be ~8 mM and ~0.2 mM, respectively.

The in vitro activities of both the S. Typhimurium and E. coli DNA gyrases were stimulated by low concentrations of both putrescine and spermidine (Fig 4). However, the two enzymes responded differently to an increase in the concentrations of these polyamines. For instance, 11 mM putrescine activated the Salmonella DNA gyrase as much as 0.4 mM spermidine (Fig 4). Both putrescine and spermidine activated the E. coli enzyme more strongly than the S. Typhimurium DNA gyrase, the effect being greater for putrescine than for spermidine (Fig 4). Activation of the E. coli DNA gyrase plateaued at a putrescine concentration between 3.4 mM and 11 mM (from +122% to +152%, i.e. a 1.2-fold increase). This is in contrast to the S. Typhimurium enzyme, the activity of which increased 2.1-fold between these two concentrations (from +26% to +53%).

[Figure omitted. See PDF.]

Fig 4. Polyamines regulate the in vitro activity of the purified DNA gyrases from S. Typhimurium and E. coli.

Topoisomerase-treated plasmid DNA was incubated with purified DNA gyrase alone (reference for 0% variation in DNA gyrase activity) and with the indicated amounts of each polyamine. The experiment was performed with the DNA gyrases from E. coli and (separately) from S. Typhimurium. DNA gyrase activity was calculated as ΔLk/hour from agarose/chloroquine gels and normalized to the activity of the purified DNA gyrase without polyamines. Coloured circles quantitatively represent added concentrations of putrescine (blue) or spermidine (green). Lowercase letters indicate pairwise statistical significance (Tukey’s HSD, p<0.05, n = 2).

https://doi.org/10.1371/journal.pgen.1009085.g004

Putrescine, spermidine, and spermine modified DNA supercoiling in a DNA gyrase-dependent manner because no changes in DNA supercoiling were observed when DNA gyrase was left out of the reaction (S2 Fig), and also because neither putrescine nor spermidine modified the activity of the purified topoisomerase I from E. coli (S2 Fig).

Putrescine inhibited both DNA gyrases at 40 mM (Fig 4). This in vitro result is unlikely to have physiological significance because inactivation of the putrescine-degrading genes speE and oat increased putrescine concentration only two-fold in vivo (Fig 3), and the putrescine concentration would need to increase five-fold to inhibit DNA gyrase activity (i.e., from 8 to 40 mM free putrescine). Likewise, spermine, which is not detected in Salmonella (this work) or E. coli (this work and [15]), inhibited both DNA gyrases at 0.86 mM, but strongly activated them at 0.29 mM (Fig 4). By contrast, spermidine had no inhibitory effect at the highest tested concentration (i.e., 2 mM) (Fig 4).

Cumulatively, the results in this section indicate that both putrescine and spermidine activate the S. Typhimurium and E. coli DNA gyrases at physiological concentrations. However, these enzymes differ in that the S. Typhimurium DNA gyrase is further activated (2.1 fold) by increases in putrescine concentrations that have negligible (1.2 fold) effects on the activity of the E. coli DNA gyrase.

An increase in the concentration of extracellular Mg2+ decreases global DNA supercoiling in S. Typhimurium but not in E. coli

Mg2+ is the most abundant divalent cation in living cells [43]. It is necessary for DNA gyrase activity in vitro [14] and competes with polyamines for binding to negatively charged molecules such as DNA [44]. These properties raise the possibility of a regulatory connection existing among the Mg2+ concentration, polyamine content, and DNA supercoiling. To test this possibility, we explored whether changes in the Mg2+ concentration in the growth media modify global DNA supercoiling in S. Typhimurium and E. coli.

DNA supercoiling was maximal in wild-type S. Typhimurium experiencing ~1 mM Mg2+ (Fig 5A), which corresponds to the concentration of free Mg2+ in living cells [40]. DNA became less supercoiled both when the Mg2+ concentration in the media decreased to 10 μM Mg2+ and when it increased above 2 mM Mg2+ (Fig 5A). By contrast, DNA supercoiling in E. coli was insensitive to changes in the Mg2+ concentration in the media between 10 μM and 10 mM (Fig 5A). These results raise the question: How does an increase in Mg2+ promote a decrease in DNA supercoiling in S. Typhimurium?

[Figure omitted. See PDF.]

Fig 5. Mg2+ controls S. Typhimurium DNA supercoiling in a putrescine-dependent manner.

A: In vivo DNA supercoiling of wild-type S. Typhimurium (14028s) (black) and wild-type E. coli (MG1655) (grey) grown in N-minimal medium with the indicated Mg2+ concentrations. B: In vivo DNA supercoiling of wild-type S. Typhimurium (14028s) grown in N-minimal medium with the indicated Mg2+ concentrations in the absence or presence of putrescine (1 mM). ***: p<0.001; ns: not significant (Student’s t-test, n = 2).

https://doi.org/10.1371/journal.pgen.1009085.g005

Growth in high Mg2+ reduces putrescine abundance in S. Typhimurium

Both putrescine (Fig 4) and ATP [45] are required for in vitro activity of the S. Typhimurium DNA gyrase. Thus, we considered the possibility that growth in high Mg2+ decreases DNA supercoiling in Salmonella by reducing the concentration of putrescine and/or ATP. Such a reduction would not take place in E. coli because growth in high Mg2+ does not alter DNA supercoiling in this species (Fig 5A).

We determined that the putrescine concentration was 25% higher in S. Typhimurium during growth in 0.8 mM than in 10 mM Mg2+ (Fig 3). By contrast, the putrescine concentration did not change in E. coli experiencing these Mg2+ concentrations (Fig 3). Putrescine supplementation to the growth medium overcame the decrease in DNA supercoiling provoked by high extracytoplasmic Mg2+ in S. Typhimurium (Fig 5B). Growth at the high Mg2+ concentration resulted in a 1.5-fold decrease in ATP concentration (S3 Fig), which is too small to account for the observed changes in DNA supercoiling [46] (i.e., the decrease in ATP concentration would have to be >5 fold). These results indicate that a growth condition that reduces putrescine abundance also reduces global negative DNA supercoiling in S. Typhimurium. How, then, does growth in high Mg2+ decrease putrescine abundance?

Growth in high Mg2+ reduces the mRNA amounts of genes that increase putrescine abundance

We carried out an RNA-seq experiment to identify genes differentially expressed at Mg2+ concentrations resulting in high versus low global negative DNA supercoiling (0.8 mM and 10 mM, respectively) in wild-type S. Typhimurium. Growth in 10 mM Mg2+ changed the mRNA abundance of multiple genes that encode proteins predicted to lower putrescine amounts (Fig 1; values for each gene presented in Table 1). For example, the mRNA abundance of the speA, speB, and speC genes was lower in 10 mM Mg2+ than in 0.8 mM Mg2+ (Table 1). speA and speB specify enzymes that participate in the synthesis of putrescine from arginine (Fig 1); and speC mediates putrescine synthesis from ornithine (Fig 1).

The mRNA abundance of the putrescine-degrading oat gene and of the arg genes mediating the conversion of ornithine into arginine (Fig 1) was higher in 10 mM Mg2+ than in 0.8 mM Mg2+. The most strongly upregulated gene in 10 mM relative to 0.8 mM Mg2+ was argI, which specifies the rate-limiting enzyme in the conversion of ornithine into arginine (Fig 1). Thus, the transcriptional response to high Mg2+ favors expression of genes that promote putrescine degradation and deplete the putrescine precursor ornithine. However, similar agmatinase activity mediated by the SpeB protein was observed in crude extracts prepared from S. Typhimurium grown at 10 mM and 0.8 mM Mg2+ (S4 Fig), and ornithine decarboxylase activity mediated by the SpeC and SpeF proteins was undetectable(S4 Fig). Therefore, it is presently unclear which of the genes whose transcription changes in 10 mM vs. 0.8 mM Mg2+ (Table 1) is responsible for the resulting change in putrescine abundance (Fig 3).

The mRNA abundance of the putrescine importer yeeF gene [47] was lower in 10 mM relative to 0.8 mM Mg2+. Taken together with the data presented above, this result indicates that high Mg2+ reduces the putrescine concentration in S. Typhimurium by altering the expression of genes predicted to decrease putrescine import and synthesis, and to increase putrescine degradation.

Finally, expression of gyrA (but not gyrB), which is also significantly regulated by Mg2+ in the growth media though not to the extent of argI, speE or oat, may contribute to the observed changes in global negative DNA supercoiling. Therefore, the coordinated response to Mg2+ (rather than individual enzymatic activities), appears to be responsible for the observed variations in polyamine amounts (Fig 3) and the resulting changes in global negative DNA supercoiling. The effects of Mg2+ point to specific gene regulation by Mg2+ rather than effects on DNA supercoiling-sensitive promoters because none of the genes differentially expressed at 10 mM vs. 0.8 mM Mg2+ displayed sensitivity to DNA supercoiling when 11 different DNA supercoiling-altering conditions were used to examine their expression behavior (see S5 Fig and GEO entry GSE137586).

Discussion

We have now established that different polyamines activate the DNA gyrases from S. Typhimurium and E. coli. The different response of these DNA gyrases to the polyamines putrescine and spermidine dictates the distinct basal global DNA supercoiling set point of the two species. We determined that environmental Mg2+ regulates DNA supercoiling in S. Typhimurium, but not in E. coli, by altering putrescine abundance. Our findings reveal a novel physiological role for polyamines, which play critical functions in all domains of life [48]. Moreover, they help explain why DNA is less negatively supercoiled in S. Typhimurium than in E. coli when these bacteria experience the same growth condition. And they suggest potential targets for antibacterial agents to inhibit the essential process of DNA supercoiling.

The molecular basis for the DNA supercoiling differences between S. Typhimurium and E. coli

The activity of the purified DNA gyrases from S. Typhimurium and E. coli is enhanced or inhibited by individual polyamines depending on their concentrations (Fig 4). This result explains the surprising finding that the S. Typhimurium speE mutant, which is unable to make spermidine but accumulates its precursor putrescine (Fig 1), is more supercoiled than wild-type S. Typhimurium (Fig 2B), but that the equivalent speE mutation in E. coli decreases DNA supercoiling below the values of wild-type E. coli (Fig 2A). Although putrescine activates the E. coli DNA gyrase at lower concentrations than the S. Typhimurium DNA gyrase, further increases in putrescine concentration stimulate S. Typhimurium DNA gyrase activity but not that of E. coli (Fig 4). Therefore, the increased putrescine amounts resulting from speE inactivation in E. coli have little effect on its DNA gyrase because the enzyme is already maximally activated (Fig 6).

[Figure omitted. See PDF.]

Fig 6. Differential sensitivity of DNA gyrases to distinct polyamines dictates DNA supercoiling differences between E. coli and S. Typhimurium.

E. coli DNA gyrase activity requires minimal amounts of putrescine and is therefore insensitive to variations in putrescine amounts (top left). As a result, the decreased DNA supercoiling resulting from loss of spermidine in a E. coli speE mutant is not compensated by increased amounts of putrescine, resulting in relaxed DNA (bottom left). S. Typhimurium DNA gyrase activity requires high amounts of putrescine to be activated and is highly sensitive to variations in putrescine amounts (top right). Putrescine accumulation in a S. Typhimurium speE mutant hyperactivates its DNA gyrase resulting in more negatively supercoiled DNA (bottom right).

https://doi.org/10.1371/journal.pgen.1009085.g006

By contrast, the putrescine accumulation resulting from speE inactivation in S. Typhimurium furthers DNA supercoiling because activation of its DNA gyrase requires higher putrescine amounts than does activation of the E. coli DNA gyrase (Fig 3). That conditions resulting in putrescine accumulation hyperactivate DNA gyrase in S. Typhimurium is further supported by the increased DNA supercoiling exhibited by the sapABCDF mutant (Fig 2), which is defective in putrescine export [38]. The different DNA supercoiling phenotypes resulting from inactivation of the speE gene in S. Typhimurium and E. coli are reminiscent of the different behaviors exhibited by these species harboring equivalent mutations in the gyrB, seqA, and mukB genes [33]

DNA gyrase is a tetramer consisting of two GyrA and two GyrB proteins [9]. The behavior of organisms with chimeric DNA gyrases originating from S. Typhimurium and E. coli, and the clustering of amino acid differences between the two GyrA proteins in the C-terminus, implicate this region in the species-specific DNA supercoiling behavior. Although the E. coli DNA gyrase is faster in vitro than the S. Typhimurium DNA gyrase in the presence of spermidine, a S. Typhimurium strain bearing the E. coli GyrA protein exhibited decreased DNA supercoiling compared to wild-type S. Typhimurium [30]. Our findings raise the possibility of putrescine and spermidine interacting with the C-terminal region of GyrA to activate the S. Typhimurium and E. coli DNA gyrases, respectively. Under this hypothesis, loss of supercoiling in the S. Typhimurium strain with the E. coli GyrA may reflect disrupted activation by polyamines.

Growth in >2 mM Mg2+ reduces DNA supercoiling in S. Typhimurium (Fig 5A) by promoting expression of genes specifying products that reduce putrescine accumulation (Table 1). Because the Mg2+ effect on DNA supercoiling is independent of the ATP control of DNA gyrase activity [46], the uncovered mechanism allows S. Typhimurium to modify DNA supercoiling while avoiding disruption of energy-dependent cellular processes. Furthermore, it enables S. Typhimurium to alter DNA supercoiling in response to different, and a priori independent, metabolic signals–ATP and putrescine.

S. Typhimurium experiences a decrease in global negative DNA supercoiling inside mammalian cells [49]. A mildly acidic pH was previously proposed to play a role based on an in vivo synergistic effect of the DNA gyrase inhibitor novobiocin and acidic pH on DNA relaxation [50]. However, this observation can be explained by a mildly acidic pH increasing permeability to novobiocin [51]. Changes in the abundance of putrescine and spermidine, or in a metabolite yet to be identified, may contribute to the reported decrease in global DNA supercoiling.

E. coli differs from S. Typhimurium in that it does not alter its basal DNA supercoiling in response to an increase in Mg2+ concentration in its surroundings (Fig 5A). Under the investigated conditions, the levels of negative DNA supercoiling in E. coli are similar to those in S. Typhimurium (0.5 RSU compared to 0–1 RSU, respectively). This argues against the C-terminus of GyrA being responsible for the reported lower DNA supercoiling in S. Typhimurium vs. E. coli [30]. However, the latter experiments were carried out using both a different S. Typhimurium strain (LT2-derived versus 14028s in the current study) and growth conditions (LB at 30°C versus defined media at 37°C in the current study), which raises the possibility of the differences in DNA supercoiling between S. Typhimurium and E. coli resulting from the signals they respond to and the magnitude of those responses. The DNA gyrase subunits of S. Typhimurium and E. coli are >90% identical, and this is also the case for the SpeE protein. Therefore, these findings alongside others [30,52,53] demonstrate how phenotypic differences between closely related bacterial species can result from allelic differences in conserved genes and in their regulation, rather than from the presence of a gene(s) in one species and its absence from another.

Inhibiting DNA supercoiling by manipulating polyamine abundance

The vast majority of antibacterial agents that inhibit DNA supercoiling do so by targeting DNA gyrase [54]. Unfortunately, many of these agents are toxic to humans [55], and resistance to these agents is spreading quickly [56]. Furthermore, despite continued efforts to discover compounds that alter DNA gyrase activity [57], there has been no clinical success since the year 2000.

The identification of putrescine and spermidine as key regulators of DNA gyrase in vivo and in vitro raises the possibility of altering DNA supercoiling by targeting the pathways governing polyamine abundance (as opposed to DNA gyrase). That polyamine biosynthesis is essential in several pathogens (reviewed in [48]) supports this possibility. In S. Typhimurium, there is a direct correlation between DNA supercoiling and expression of the invasion gene invA [58], and DNA supercoiling decreases when S. Typhimurium is inside macrophages [59], promoting expression of genes required for intramacrophage survival [59]. Moreover, mutants defective in putrescine synthesis or export exhibit aberrant DNA supercoiling (Fig 2B) and are attenuated for virulence [28,29].

Spermine inhibited DNA gyrase at submicromolar concentrations (Fig 4). This raises the possibility of organisms using spermine as an endogenous DNA gyrase inhibitor. It also suggests using a spermine scaffold to design novel DNA gyrase inhibitors.

Finally, humans lack some of the putrescine biosynthetic enzymes utilized by bacteria (S6 Fig) as well as putrescine transporters, making it less likely that an agent targeting putrescine abundance in bacteria would inhibit human proteins carrying out analogous functions. That pathogenic S. Typhimurium and commensal E. coli differ in their control of DNA supercoiling suggests that an antimicrobial agent may target pathogen-specific pathways while sparing those operating in commensal organisms.

Materials and methods

Bacterial strains, plasmids and growth conditions

Strains, plasmids and primers are described in S1 Table.

Mutations were created by λred recombination [60]. Plasmid pSIM6 was used to supply λred in all cases. After PCR verification of the strains, mutations were moved into a clean genetic background using P22-mediated transduction in the case of S. Typhimurium [61] and P1-mediated transduction in the case of E. coli.

Strains AAD35, AAD46 and AAD85 were built by λred recombination using pKD4, pKD3 and pKD4 as templates, respectively, and primer pairs 16608/16609, 16651/16652, 16834/16835, respectively.

Strain AAD58 was built by P22 transduction using a lysate prepared in strain AAD46 to infect strain JY979.

Strain AAD181 was built by P22 transduction using a lysate prepared in strain AAD35 to infect strain AAD58.

For strain AAD212, single argE::kan and speB::cat mutations were built in separate strains by λred recombination using pKD4 and pKD3 as templates, respectively, and primer pairs 16698/16699 and 16702/16703, respectively. The speB::cat strain (AAD62) was then transformed by plasmid pCP20 to remove the cat cassette and yield a ΔspeB strain (AAD65). Finally, the ΔspeB strain was transduced with a P22 lysate prepared in strain AAD61 (argE::kan) to yield strain AAD212.

For strain AAD96, single oat::cat and puuA::cat mutations were built separately by λred recombination using pKD3 and pKD4 as templates, respectively, and primer pairs 16840/16841 and 16842/16843, respectively. The oat::cat (AAD86) strain was then transformed with plasmid pCP20 to remove the cat cassette and yield Δoat strain (AAD88). Finally, the Δoat strain was sequentially transduced with P1 lysate from AAD85 (speE::kan) and from AAD91 (puuA::cat) to yield strain AAD96.

For strain AAD95, single speB::cat, speC::kan and speF::cat mutations were built in separate strains by λred recombination, using pKD3, pKD4 and pKD3 as templates, respectively, and primer pairs 16832/16833, 16838/16839, and 16836/16837 respectively. The speB::cat strain (AAD89) was then transformed with plasmid pCP20 to remove the cat cassette and yield a ΔspeB strain (AAD92). Finally, the ΔspeB strain was sequentially transduced with a P1 lysate prepared in the speC::kan strain (AAD87) and from the speF::cat strain (AAD90) to yield strain AAD95.

Strains AAD85, AAD95 and AAD96 were sequenced alongside the parental MG1655, and Snippy (https://github.com/tseemann/snippy) was used to detect mutations. AAD96 has a mutation that specifies a WaaO protein with the Asp36Glu substitution. The four strains were otherwise isogenic except for the intended mutations. Data are available in the Sequence Read Archive (SRA) under accession numbers SRR12282603, SRR12282604, SRR12282605, SRR12282606.

All strains were grown in HH800 minimal medium at 37°C except otherwise indicated. HH800 is N-minimal medium[62] supplemented with 0.8 mM MgCl2, 0.1% casamino acids, 0.27% glycerol, and adjusted to pH 7.7. HH was the same medium, but with 10 mM MgCl2, and was used as a high magnesium concentration to investigate regulation of polyamine abundance by Mg2+. Antibiotics were used at the following concentrations: ampicillin, 50μg/mL; kanamycin, 50μg/mL; chloramphenicol, 25μg/mL. E. coli strains were further supplemented with 1μg/mL biotin and 1μg/mL thiamine when grown in minimal medium.

Measurement of DNA supercoiling on agarose/chloroquine gels

Overnight precultures of strains bearing plasmid pJV were washed once in water, then diluted into fresh HH800 medium to a starting OD600 of 0.05. Cells were grown until mid-exponential phase (OD600 = 0.7 ± 0.1 for wild-type, and adjusted accordingly for mutants with altered growth), then plasmids were immediately extracted using a QIAGen Plasmid Mini kit. 800 ng of purified plasmid for each sample was then loaded on a Tris-Borate-EDTA, 0.8% agarose, 2.5 μg/mL chloroquine gel. Gels were run overnight at 1.3 V/cm, washed in water for at least 4 h, then stained using EZ-vision (VWR) and imaged with an ImageQuant LAS400 (GE healthcare). The intensity of each band was quantified with ImageJ. The linking number (Lk) of the top band was arbitrarily set to 0, then the band immediately below had Lk = 1, the next one Lk = 2… The intensity-weighted average Lk was calculated for each lane. The measured DNA supercoiling for S. Typhimurium was normalized across experiments to the supercoiling exhibited by the wild-type strain 14028s following growth in HH (= HH800 + 10 mM MgCl2), defined as 0 Relative Supercoiling Units (RSU), and the supercoiling in WT in HH800, defined as 1 RSU.

Considerations about quantification of DNA supercoiling

This work concerns in vivo DNA supercoiling, which is overall negative. Therefore, relaxed DNA (i.e. supercoiling closer to 0) is referred to as “low supercoiling” and has low RSU values. By contrast, supercoiled DNA (i.e., strongly negative DNA supercoiling) is referred to “high supercoiling” and has high values in RSU units. Because apparent linking numbers on agarose/chloroquine gels often vary from gel to gel, we use a relative scale instead (RSUs), which allows easy comparison of different gels thanks to a common set of controls. Under the investigated conditions, one linking number difference out of 7 or 8 visible bands typically converts into 2 RSUs. As we focused on quantitative variations in DNA supercoiling rather than absolute values, we did not attempt to distinguish positive and negative DNA supercoiling on gels.

HPLC measurement of total polyamines

Strains were cultured as described above for the determination of DNA supercoiling. Five OD600 units were pelleted and washed three times with PBS. Cells were then resuspended in 200 μL lysis buffer (100 mM MOPS, 50 mM NaCl, 20 mM MgCl2, pH 8.0) and plated to obtain Colony Forming Units (CFUs).

Cells were frozen in liquid nitrogen, then thawed at 37°C three times. 60 μL trichloroacetic acid 40% was added and the mixture was incubated on ice for 5 min. The supernatant was submitted for derivatization.

Derivatization was performed by adding 250 μL NaOH 2M, 10 μL benzoyl chloride and incubating at room temperature for 20 min. The reaction was stopped by adding 400 μL NaCl, then polyamines were extracted using 300 μL diethyl ether. The ether was evaporated at 37°C and polyamines were dissolved in methanol/water 45%/55% (v/v).

For the standard curve, a mixture of 20 mM putrescine, 10 mM spermidine and 10 mM spermine was made. This mixture was serially diluted in 4-fold increments. 50 μL of the polyamine mix was added to 200 μL lysis buffer, then derivatized as before.

HPLC was performed on a Waters Acquity UPLC H Class with PDA detector (Waters, Milford, MA). The column was a Waters Acquity UPLC BEH C18, 1.7 μm (Waters 186005604), kept at 60°C throughout the experiment. Flow rate was 0.3 mL/min. Solvent was 45% methanol/55% water (v/v). Detection was performed in UV at 254 nm.

Cell volume was measured in a separate experiment by phase microscopy, and estimated to be 4 μm3 for the tested strains and conditions.

Intracellular polyamine concentrations were then calculated from the measured concentrations by HPLC, the CFUs and the cell volume.

RNA-seq analysis

Overnight precultures of strains bearing plasmid pJV were washed once in water, then diluted into fresh HH800 medium supplemented or not with 10 mM Mg2+ to a starting OD600 of 0.05. Cells were grown until OD600 = 0.8 ± 0.1 (similarly to supercoiling measurements) and RNA was extracted using a QIAGen RNeasy Mini kit. In addition to the included DNase treatment in the kit, DNA was further eliminated by treatment by turbo DNase (Ambion). RNA was finally re-purified using a QIAGen RNeasy Mini kit.

Library construction and sequencing was performed by the Yale Center for Genome Analysis. rRNA was depleted using a RiboZero kit (Illumina). cDNA synthesis was performed by adding A bases to the 3’ end of fragments, followed by oligodT priming. The 4 samples were barcoded and multiplexed into a single flow cell. Sequencing was performed by on an HS4000 (Illumina), Hiseq 75x2 paired-end, unstranded.

Reads were mapped to 14028s genome (Genbank CP001363.1) using bowtie 2.2.9 [63] and differential expression analysis was performed using cuffdiff from the cufflinks 2.2.1 [64] package. Data was deposited to GEO (GSE150757)

Assay of ornithine decarboxylase and agmatinase activities

Cells were grown until mid-exponential phase as for DNA supercoiling measurements. Then, cells were washed twice in PBS, and sonicated on a Bioruptor sonicator (Diagenode, NJ, USA) with the following settings: High intensity, 30s ON/30s OFF, 15 min. The supernatant was used to determine activities and protein concentrations.

Agmatinase activity was determined by assaying the release of urea by alpha-isonitroso propiophenone [65] and ornithine decarboxylase by solubilization of 2,4,6 trinitrobenzenesulfonic acid adducts into DMSO [66].

Protein concentrations were determined using the Bradford colorimetric assay.

Examining the activities of topoisomerase I and DNA gyrase

Wild-type S. Typhimurium strain14028s transformed with pJV was grown in HH800 + 10 mM MgCl2 + 25 μg/mL novobiocin until OD600 = 1, then pJV was extracted from 400 mL of culture using a QIAGen Plasmid Maxi kit. This yielded a moderately relaxed pJV.

For topoisomerase I assays, 800 ng of purified pJV was incubated with 0.5 units of DNA topoisomerase I (NEB M0301) in Cutsmart buffer (NEB B7204) at 37°C for 2 h, then inactivated at 65°C for 20 min.

For gyrase assays pJV was first further relaxed by topoisomerase I as for the topoisomerase I assay. Then, for E. coli gyrase assays, 800 ng of relaxed pJV was incubated with 2 units of DNA gyrase (TopoGen TG-2000G-3) at 37°C for 5 h, then inactivated at 65°C for 20 min. For S. Typhimurium gyrase assays, 800 ng of relaxed pJV was incubated with 0.62 pmol GyrA and 0.62 pmol GyrB (both purified from SL1344, which has 100% identity with 14028s gyrase subunits, gifted by Lesley Mitchenall and Anthony Maxwell, John Innes Centre, Norwich, UK, [32]) at 37°C for 90 min, then inactivated at 65°C for 20 min. The reaction buffer was as follows: Tris pH 7.5, 35 mM; KCl, 24 mM; MgCl2, 4 mM; DTT, 2 mM; ATP, 1 mM; glycerol, 6.5% (v/v); Bovine Serum Albumin, 0.1 mg/mL.

In all cases, the whole reaction was loaded onto a Tris-Borate-EDTA, 0.8% agarose gel, 0.4 μg/mL chloroquine gel. Gels were migrated, stained, imaged and quantified as for in vivo DNA supercoiling gels. DNA gyrase activity was defined as the increase in average Lk, per hour, compared to the initial topoisomer distribution. DNA gyrase activity in the absence of any polyamine was used as a baseline (i.e. 0% relative variation).

Supporting information

S1 Fig. Growth curves for the various strains presented in Fig 2.

A: E. coli strains B: S. Typhimurium strains. Each line represents a single biological replicate. The large red square represents the point where samples were taken for DNA supercoiling.

https://doi.org/10.1371/journal.pgen.1009085.s001

(PDF)

S2 Fig. In vitro effect of putrescine, spermidine and ornithine on topoisomerase I.

Assays were conducted as described in Methods using the indicated amounts of the listed metabolites.

https://doi.org/10.1371/journal.pgen.1009085.s002

(PDF)

S3 Fig. ATP amounts vary little in response to excess Mg2+.

Intracellular ATP concentrations of wild-type S. Typhimurium (14028s) were measured using the Bac-Titer Glo assay. *: p<0.05 (Student’s t-test, n = 3)

https://doi.org/10.1371/journal.pgen.1009085.s003

(PDF)

S4 Fig. Activity of putrescine biosynthetic enzymes in crude S. Typhimurium extracts.

Agmatinase and ornithine decarboxylase activities were assayed in crude extracts from wild-type S. Typhimurium (14028s). ns: not significant (Student’s t-test, n = 3)

https://doi.org/10.1371/journal.pgen.1009085.s004

(PDF)

S5 Fig. Expression of key putrescine-related genes is not dependent on DNA supercoiling.

Expression of the presented S. Typhimurium genes (in FPKM) was plotted as a function of DNA supercoiling (data taken from GEO entry GSE137586). Open circles represent independent data points, and lines represent linear regressions.

https://doi.org/10.1371/journal.pgen.1009085.s005

(PDF)

S6 Fig. The putrescine biosynthetic and degradative pathways in H. sapiens.

Note that OAT is unrelated to the bacterial Oat and catalyzes a different reaction.

https://doi.org/10.1371/journal.pgen.1009085.s006

(PDF)

S1 Table. Strains, plasmids and primers used in this study.

https://doi.org/10.1371/journal.pgen.1009085.s007

(PDF)

Acknowledgments

HPLC analysis was performed by the Yale Keck proteomics platform. We thank Drs. Lesley Mitchenall and Anthony Maxwell (John Innes Centre, Norwich, UK) for the purified Salmonella gyrase and for comments on the manuscript. We thank Jennifer Aronson for comments of the manuscript and help with the figures.

Citation: Duprey A, Groisman EA (2020) DNA supercoiling differences in bacteria result from disparate DNA gyrase activation by polyamines. PLoS Genet 16(10): e1009085. https://doi.org/10.1371/journal.pgen.1009085

References

1. Stuger R, Woldringh CL, Weijden CC van der, Vischer NOE, Bakker BM, Spanning RJM van, et al. DNA Supercoiling by Gyrase is Linked to Nucleoid Compaction. Mol Biol Rep. 2002;29:79–82. pmid:12241080

2. Pruss GJ, Drlica K. DNA supercoiling and prokaryotic transcription. Cell. 1989;56:521–523. pmid:2645054

3. Dorman CJ. DNA supercoiling and transcription in bacteria: a two-way street. BMC Mol Cell Biol. 2019;20:26. pmid:31319794

4. Dunaway M, Ostrander EA. Local domains of supercoiling activate a eukaryotic promoter in vivo. Nature. 1993;361:746–748. pmid:8441472

5. Higgins NP. Measuring In Vivo Supercoil Dynamics and Transcription Elongation Rates in Bacterial Chromosomes. The Bacterial Nucleoid. Humana Press, New York, NY; 2017. pp. 17–27. https://doi.org/10.1007/978-1-4939-7098-8_2 pmid:28842872

6. Benjamin KR, Abola PA, Kanaar R, Cozzarelli NR. Contributions of Supercoiling to Tn3 Resolvase and Phage Mu Gin Site-specific Recombination. J Mol Biol. 1996;256:50–65. pmid:8609613

7. Carteau S, Mouscadet JF, Goulaouic H, Subra F, Auclair C. Effect of Topoisomerase Inhibitors on the in Vitro HIV DNA Integration Reaction. Biochem Biophys Res Commun. 1993;192:1409–1414. pmid:8389550

8. Holmes VF, Cozzarelli NR. Closing the ring: Links between SMC proteins and chromosome partitioning, condensation, and supercoiling. Proc Natl Acad Sci. 2000;97:1322–1324. pmid:10677457

9. Champoux JJ. DNA topoisomerases: structure, function, and mechanism. Annu Rev Biochem. 2001;70:369–413. pmid:11395412

10. Pedro-Botet ML, Yu VL. Treatment strategies for Legionella infection. Expert Opin Pharmacother. 2009;10:1109–1121. pmid:19405787

11. Zechiedrich EL, Khodursky AB, Bachellier S, Schneider R, Chen D, Lilley DMJ, et al. Roles of Topoisomerases in Maintaining Steady-state DNA Supercoiling in Escherichia coli. J Biol Chem. 2000;275:8103–8113. pmid:10713132

12. Gilbert N, Allan J. Supercoiling in DNA and chromatin. Curr Opin Genet Dev. 2014;25:15–21. pmid:24584092

13. Verma SC, Qian Z, Adhya SL. Architecture of the Escherichia coli nucleoid. PLoS Genet. 2019;15. pmid:31830036

14. Gellert M, Mizuuchi K, O’Dea MH, Nash HA. DNA gyrase: an enzyme that introduces superhelical turns into DNA. Proc Natl Acad Sci. 1976;73:3872–3876. pmid:186775

15. Shah P, Swiatlo E. A multifaceted role for polyamines in bacterial pathogens. Mol Microbiol. 2008;68:4–16. pmid:18405343

16. Noble CG, Maxwell A. The Role of GyrB in the DNA Cleavage-religation Reaction of DNA Gyrase: A Proposed Two Metal-ion Mechanism. J Mol Biol. 2002;318:361–371. pmid:12051843

17. Basu A, Parente AC, Bryant Z. Structural Dynamics and Mechanochemical Coupling in DNA Gyrase. J Mol Biol. 2016;428:1833–1845. pmid:27016205

18. Wang JC. Interaction between DNA and an Escherichia coli protein omega. J Mol Biol. 1971;55:523–533. pmid:4927945

19. Nitiss JL, Soans E, Rogojina A, Seth A, Mishina M. Topoisomerase Assays. Curr Protoc Pharmacol. 2012;CHAPTER: Unit3.3. pmid:22684721

20. Tse-Dinh Y-C, Beran RK. Multiple promoters for transcription of the Escherichia coli DNA topoisomerase I gene and their regulation by DNA supercoiling. J Mol Biol. 1988;202:735–742. pmid:2845101

21. Michael AJ. Polyamines in Eukaryotes, Bacteria, and Archaea. J Biol Chem. 2016;291:14896–14903. pmid:27268252

22. Gevrekci AÖ. The roles of polyamines in microorganisms. World J Microbiol Biotechnol. 2017;33:204. pmid:29080149

23. Igarashi K, Kashiwagi K. Polyamine Modulon in Escherichia coli: Genes Involved in the Stimulation of Cell Growth by Polyamines. J Biochem (Tokyo). 2006;139:11–16. pmid:16428314

24. Tabor CW, Tabor H. Polyamines in microorganisms. Microbiol Mol Biol Rev. 1985;49:81–99.

25. Jung IL, Kim IG. Polyamines and Glutamate Decarboxylase-based Acid Resistance in Escherichia coli. J Biol Chem. 2003;278:22846–22852. pmid:12670930

26. Tkachenko AG, Akhova AV, Shumkov MS, Nesterova LY. Polyamines reduce oxidative stress in Escherichia coli cells exposed to bactericidal antibiotics. Res Microbiol. 2012;163:83–91. pmid:22138596

27. Brickman TJ, Armstrong SK. The ornithine decarboxylase gene odc is required for alcaligin siderophore biosynthesis in Bordetella spp.: putrescine is a precursor of alcaligin. J Bacteriol. 1996;178:54–60. pmid:8550442

28. Jelsbak L, Thomsen LE, Wallrodt I, Jensen PR, Olsen JE. Polyamines Are Required for Virulence in Salmonella enterica Serovar Typhimurium. PLOS ONE. 2012;7:e36149. pmid:22558361

29. Parra-Lopez C, Baer MT, Groisman EA. Molecular genetic analysis of a locus required for resistance to antimicrobial peptides in Salmonella typhimurium. EMBO J. 1993;12:4053–4062. pmid:8223423

30. Rovinskiy NS, Agbleke AA, Chesnokova ON, Higgins NP. Supercoil Levels in E. coli and Salmonella Chromosomes Are Regulated by the C-Terminal 35–38 Amino Acids of GyrA. Microorganisms. 2019;7:81. pmid:30875939

31. Adeolu M, Alnajar S, Naushad S, Gupta RS. Genome based phylogeny and taxonomy of the “Enterobacteriales”: proposal for Enterobacterales ord. nov. divided into the families Enterobacteriaceae, Erwiniaceae fam. nov., Pectobacteriaceae fam. nov., Yersiniaceae fam. nov., Hafniaceae fam. nov., Morganellaceae fam. nov., and Budviciaceae fam. nov. Int J Syst Evol Microbiol. 2016 [cited 22 Sep 2016]. Available: http://ijs.microbiologyresearch.org/content/journal/ijsem/10.1099/ijsem.0.001485?crawler=true

32. Webber MA, Buckner MMC, Redgrave LS, Ifill G, Mitchenall LA, Webb C, et al. Quinolone-resistant gyrase mutants demonstrate decreased susceptibility to triclosan. J Antimicrob Chemother. 2017;72:2755–2763. pmid:29091182

33. Champion K, Higgins NP. Growth Rate Toxicity Phenotypes and Homeostatic Supercoil Control Differentiate Escherichia coli from Salmonella enterica Serovar Typhimurium. J Bacteriol. 2007;189:5839–5849. pmid:17400739

34. Vetcher AA, McEwen AE, Abujarour R, Hanke A, Levene SD. Gel mobilities of linking-number topoisomers and their dependence on DNA helical repeat and elasticity. Biophys Chem. 2010;148:104–111. pmid:20346570

35. Tabor CW, Tabor H, Xie QW. Spermidine synthase of Escherichia coli: localization of the speE gene. Proc Natl Acad Sci. 1986;83:6040–6044. pmid:3526348

36. Charlier D, Bervoets I. Regulation of arginine biosynthesis, catabolism and transport in Escherichia coli. Amino Acids. 2019;51:1103–1127. pmid:31267155

37. McClelland M, Sanderson KE, Spieth J, Clifton SW, Latreille P, Courtney L, et al. Complete genome sequence of Salmonella enterica serovar Typhimurium LT2. Nature. 2001;413:852–856. pmid:11677609

38. Sugiyama Y, Nakamura A, Matsumoto M, Kanbe A, Sakanaka M, Higashi K, et al. A Novel Putrescine Exporter SapBCDF of Escherichia coli. J Biol Chem. 2016;291:26343–26351. pmid:27803167

39. Husna AU, Wang N, Cobbold SA, Newton HJ, Hocking DM, Wilksch JJ, et al. Methionine biosynthesis and transport are functionally redundant for the growth and virulence of Salmonella Typhimurium. J Biol Chem. 2018; jbc.RA118.002592. pmid:29720401

40. Alatossava T, Jütte H, Kuhn A, Kellenberger E. Manipulation of intracellular magnesium content in polymyxin B nonapeptide-sensitized Escherichia coli by ionophore A23187. J Bacteriol. 1985;162:413–419. pmid:2984182

41. Fang S-B, Huang C-J, Huang C-H, Wang K-C, Chang N-W, Pan H-Y, et al. speG Is Required for Intracellular Replication of Salmonella in Various Human Cells and Affects Its Polyamine Metabolism and Global Transcriptomes. Front Microbiol. 2017;8. pmid:29187844

42. Miyamoto S, Kashiwagi K, Ito K, Watanabe S, Igarashi K. Estimation of Polyamine Distribution and Polyamine Stimulation of Protein Synthesis in Escherichia coli. Arch Biochem Biophys. 1993;300:63–68. pmid:7678729

43. Romani AM, Scarpa A. Regulation of cellular magnesium. Front Biosci J Virtual Libr. 2000;5:D720–734. pmid:10922296

44. Rowatt E, Williams RJP. The binding of polyamines and magnesium to DNA. J Inorg Biochem. 1992;46:87–97. pmid:1522415

45. Kongsoi S, Yokoyama K, Suprasert A, Utrarachkij F, Nakajima C, Suthienkul O, et al. Characterization of Salmonella Typhimurium DNA gyrase as a target of quinolones. Drug Test Anal. 2015;7:714–720. pmid:25381884

46. Workum M, Dooren SJ, Oldenburg N, Molenaar D, Jensen PR, Snoep JL, et al. DNA supercoiling depends on the phosphorylation potential in Escherichia coli. Mol Microbiol. 1996;20:351–360. pmid:8733233

47. Kurihara S, Suzuki H, Oshida M, Benno Y. A Novel Putrescine Importer Required for Type 1 Pili-driven Surface Motility Induced by Extracellular Putrescine in Escherichia coli K-12. J Biol Chem. 2011;286:10185–10192. pmid:21266585

48. Michael AJ. Polyamine function in archaea and bacteria. J Biol Chem. 2018; jbc.TM118.005670. pmid:30254075

49. Duprey A, Reverchon S, Nasser W. Bacterial virulence and Fis: adapting regulatory networks to the host environment. Trends Microbiol. 2014;22:92–99. pmid:24370464

50. Colgan AM, Quinn HJ, Kary SC, Mitchenall LA, Maxwell A, Cameron ADS, et al. Negative supercoiling of DNA by gyrase is inhibited in Salmonella enterica serovar Typhimurium during adaptation to acid stress. Mol Microbiol. 2018;107:734–746. pmid:29352745

51. Pomares MF, Corbalán NS, Adler C, de Cristóbal R, Farías RN, Delgado MA, et al. Macrophage environment turns otherwise MccJ25-resistant Salmonella into sensitive. BMC Microbiol. 2013;13:95. pmid:23634875

52. Duprey A, Nasser W, Léonard S, Brochier-Armanet C, Reverchon S. Transcriptional start site turnover in the evolution of bacterial paralogous genes–the pelE-pelD virulence genes in Dickeya. FEBS J. 2016;283:4192–4207. pmid:27727510

53. Chen HD, Jewett MW, Groisman EA. Ancestral Genes Can Control the Ability of Horizontally Acquired Loci to Confer New Traits. PLOS Genet. 2011;7:e1002184. pmid:21811415

54. Maxwell A. DNA gyrase as a drug target. Trends Microbiol. 1997;5:102–109. pmid:9080608

55. Rubinstein E. History of Quinolones and Their Side Effects. Chemotherapy. 2001;47:3–8. pmid:11549783

56. WHO Advisory Group on Integrated Surveillance of Antimicrobial Resistance, World Health Organization. Critically important antimicrobials for human medicine: ranking of antimicrobial agents for risk management of antimicrobial resistance due to non-human use. 2017. Available: http://apps.who.int/iris/bitstream/10665/255027/1/9789241512220-eng.pdf

57. Shapiro A, Jahic H, Prasad S, Ehmann D, Thresher J, Gao N, et al. A Homogeneous, High-Throughput Fluorescence Anisotropy-Based DNA Supercoiling Assay. J Biomol Screen. 2010;15:1088–1098. pmid:20930214

58. Galán JE, Curtiss R. Expression of Salmonella typhimurium genes required for invasion is regulated by changes in DNA supercoiling. Infect Immun. 1990;58:1879–1885. pmid:2160435

59. Ó Cróinín T, Carroll RK, Kelly A, Dorman CJ. Roles for DNA supercoiling and the Fis protein in modulating expression of virulence genes during intracellular growth of Salmonella enterica serovar Typhimurium. Mol Microbiol. 2006;62:869–882. pmid:16999831

60. Datta S, Costantino N, Court DL. A set of recombineering plasmids for gram-negative bacteria. Gene. 2006;379:109–115. pmid:16750601

61. Davis RW, Botstein D, Roth JR. Advanced Bacterial Genetics: A Manual for Genetic Engineering. Cold Spring Harbor, N.Y: Cold Spring Harbor Laboratory Pr; 1980.

62. Snavely MD, Miller CG, Maguire ME. The mgtB Mg2+ transport locus of Salmonella typhimurium encodes a P-type ATPase. J Biol Chem. 1991;266:815–823. pmid:1824701

63. Langmead B, Salzberg SL. Fast gapped-read alignment with Bowtie 2. Nat Methods. 2012;9:357–359. pmid:22388286

64. Trapnell C, Roberts A, Goff L, Pertea G, Kim D, Kelley DR, et al. Differential gene and transcript expression analysis of RNA-seq experiments with TopHat and Cufflinks. Nat Protoc. 2012;7:562–578. pmid:22383036

65. Carvajal N, López V, Salas M, Uribe E, Herrera P, Cerpa J. Manganese Is Essential for Catalytic Activity ofEscherichia coliAgmatinase. Biochem Biophys Res Commun. 1999;258:808–811. pmid:10329468

66. Ngo TT, Brillhart KL, Davis RH, Wong RC, Bovaird JH, Digangi JJ, et al. Spectrophotometric assay for ornithine decarboxylase. Anal Biochem. 1987;160:290–293. pmid:3578755