The ArcAB two-component regulatory system promotes resistance to reactive oxygen species and systemic infection by Salmonella Typhimurium

Salmonella enterica Serovar Typhimurium (S. Typhimurium) is an intracellular bacterium that overcomes host immune system barriers for successful infection. The bacterium colonizes the proximal small intestine, penetrates the epithelial layer, and is engulfed by macrophages and neutrophils. Intracellularly, S. Typhimurium encounters highly toxic reactive oxygen species including hydrogen peroxide and hypochlorous acid. The molecular mechanisms of Salmonella resistance to intracellular oxidative stress is not completely understood. The ArcAB two-component system is a global regulatory system that responds to oxygen. In this work, we show that the ArcA response regulator participates in Salmonella adaptation to changing oxygen levels and is also involved in promoting intracellular survival in macrophages and neutrophils, enabling S. Typhimurium to successfully establish a systemic infection.


Introduction
Phagocytic cells play critical roles defending the host against pathogens. However, several microorganisms, such as Salmonella enterica serovar Typhimurium (S. Typhimurium), can survive and multiply in phagocytic cells including monocytic macrophages and neutrophils [1][2][3].
S. Typhimurium survives intracellularly in a membrane-bound compartment known as the Salmonella-containing vacuole (SCV) [2,4]. Salmonella residing inside macrophages can be protected against more lethal neutrophils and inflammatory monocytes [5]. However, despite infection of macrophages and neutrophils appear to be host vulnerable points because pathogens have the capacity to use them to their advantage, they also play a critical role in host protection [6][7][8][9][10]. Indeed, mice lacking neutrophils cannot control S. Typhimurium proliferation [11].
Reactive oxygen species (ROS) are the primary microbicides within macrophages and neutrophils [8,12], particularly under acidic conditions. Two molecules of superoxide (O 2 -) react non-enzymatically to generate hydrogen peroxide (H 2 O 2 ) [13] reaching concentrations between 1-4 μM to 2 mM in the phagosome [8,14]. Further, the slightly alkaline environment in neutrophils triggers the expression and activation of myeloperoxidase [14,15], which catalyzes the production of up to 73 mM hypochlorous acid (HOCl) from H 2 O 2 and chloride [16]. Salmonella modulates gene expression in response to the intracellular lifestyle. Transcriptome analyses of S. Typhimurium collected from infecting macrophages showed that the expression of 919 of 4451 open reading frames becomes significantly altered, and 44% represented coding sequences of unknown function [6]. Similarly, 42% and 31% of Salmonella genes were differentially expressed when infecting macrophages or epithelial cells, respectively [17]. In particular, chlorine-based oxidative stress altered the expression of Salmonella genes encoding proteins implicated in the iron-sulfur cluster assembly, cysteine biosynthesis, stress response, biofilm formation and central metabolism [18].
Several S. Typhimurium porins (OmpD, OmpW, OmpC, OmpS2, and OmpF) mediate H 2 O 2 and HOCl influx. Expression of these porins is also down-regulated in response to H 2 O 2 and HOCl, and their inactivation increases Salmonella ROS resistance [28,29]. Accordingly, the absence of OmpD increases S. Typhimurium invasion and proliferation in RAW 264.7 macrophages, as well as proliferation and systemic dissemination in BALB/c mice [30]. We have previously ruled out a role for SoxS and OxyR in the negative regulation of porin genes; instead, the expression of several of these porins was regulated by ArcA [28,29], a transcription factor in the ArcAB two-component system. ArcAB, which comprises the ArcA response regulator and the sensor histidine kinase ArcB, responds to oxygen [31][32][33]. ArcAB modulates cellular metabolism, biosynthesis and motility of anaerobically and aerobically grown S. Typhimurium [34,35]. It is also required for virulence in Haemophilus influenza, Vibrio cholerae, Actinobacillus pleuropneumoniae, and S. Typhimurium [36][37][38][39].
When S. Typhimurium responds to H 2 O 2 during aerobic growth, ArcA regulates gene expression and contributes to control the levels of glutathione, NADH, and intracellular ROS, as well as the pyruvate dehydrogenase complex and other processes, thus modulating the redox potential of the cell [35]. In this study, we examined in more detail the function of ArcAB two-component system in Salmonella pathogenesis, especially its role in S. Typhimurium infectivity. Our results show that ArcA participates in the bacterial adaptation to changing oxygen levels, and also promotes intracellular survival in macrophages and neutrophils, hence contributing to systemic infection.

Ethics statement
Animals were manipulated following the recommendations in the Guide for the Care and Use of Laboratory Animals of the US National Institutes of Health, and the protocol was approved by the Bioethics Committee of Universidad Andrés Bello, Protocols 001/2012 and 06/2016 in the framework of FONDECYT Grants #1120384 and #1160315.

Bacterial strains and growth conditions
The S. Typhimurium 14028s parental strain, ΔarcA, ΔarcB, ΔarcA/pBR::arcA and ΔarcB/pBR:: arcB, complemented with plasmid pBR322 containing the promoter and coding regions for arcA and arcB, were maintained on LB agar plates in aerobiosis unless otherwise indicated (Table 1).

Cell line cultures
RAW 264.7 (ATCC 1 TIB-71™) murine macrophages and HEp-2 (ATCC 1 CCL-223™) human epithelial cells were cultivated in 25-cm 2 tissue culture flasks (Becton Dickinson Labware) with 5 ml of the specific culture media for each cell type. The Roswell Park Memorial Institute (RPMI) medium and Dulbecco's Modified Eagle Medium (DMEM) were used for epithelial cells and macrophages, respectively. Both cell types were passaged twice a week when they reached 80% confluence. Culture media were supplemented with 10% fetal bovine serum (FBS). Cells were incubated at 37˚C with 5% CO 2 .

Mice bone-marrow-derived neutrophil extraction and infection assays
C57BL/6 female mice (7 to 8 weeks old) were used to obtain bone-marrow-derived neutrophils and for infection assays. Mice were kept in plastic cages in a temperature-controlled environment (22-24˚C). Bacteremia was assessed as described elsewhere [30], Briefly, three groups of six female C57BL/6 mice (7 to 8 weeks old) were infected orally with 1×10 5 colony forming units (CFUs). In parallel, the same number of mice was infected intraperitoneally with 1×10 3 CFUs of parental or mutant strains in 0.1 ml Phosphate Buffered Saline (PBS) (NaCl 137 mM, KCl 2.7 mM, Na 2 HPO 4 10 mM, and K 2 HPO 4 1.4 mM, pH 7.4). A group of six females were used as a noninfected control. The health of all mice was monitored daily. After day 3 (for oral infection) or day 5 (for intraperitoneal infection), mice were euthanized by cervical dislocation. Extracted livers and spleens from infected and noninfected controls were weighed and homogenized with sterile PBS; the homogenates were serially diluted (10-fold increments) in sterile PBS, and CFUs were determined by plating onto LB agar. To prepare mouse neutrophils, bone marrow was extracted as described [40], and bone-marrow-derived neutrophils were obtained using the mouse Neutrophil Isolation Kit (Milenybiotec) according to the manufacturer's instructions. This resulted in the isolation of an average of 800.000 neutrophils/ml with around 85% viability, which were also positive for CD11b and Ly6G, as determined by flow cytometry. In addition, the viability of neutrophils was monitored throughout the experiments with trypan blue staining.

Gentamicin protection assays
Cell infection assays were conducted using S. Typhimurium 14028s and its isogenic derivatives ΔarcA and ΔarcB, as described [41] with minor modifications. Bacteria were grown under microaerophilic conditions by adding an overlay of 500 μl of sterile mineral oil as a barrier to oxygen. Prior to infection assays, bacteria were centrifuged (13,000 rpm, 5 min) and resuspended in 1 ml of cell culture medium (RPMI for HEp-2, neutrophils and DMEM for RAW 264.7) supplemented with 10% FBS; as a result, the concentration of bacteria used to infect was 5x10 8 bacteria/ml. For HEp-2 and RAW 264.7 adherent cells, the assays were performed on 96-well plates. 100 μl of the bacterial cell suspension was added to each well containing cell monolayers (multiplicity of infection of 1:100). After 1 h incubation in 5% CO 2 at 37˚C, technical triplicates of the infected cells were stained with trypan blue to determine cell viability. Also, technical replicates for each strain were lysed with deoxycholate (0.5% w/v in PBS), serially diluted (10-fold) in PBS, and plated onto LB agar plates to obtain the CFU of each strain at 1 h postinfection (hpi). The remaining wells were washed twice with sterile PBS and incubated in 5% CO 2 at 37˚C for 2 h with 100 μl cell medium plus 250 μg ml -1 gentamicin to kill extracellular bacteria. At 3 hpi, the medium was removed, and cells were washed twice with PBS and lysed with sodium deoxycholate (0.5% w/v in PBS). The cell lysates were 10-fold serially diluted in PBS and plated onto LB agar plates to obtained the CFU counts at 3 hpi. The same protocol was used for infection of non-adherent murine neutrophils, except that the cells were kept in 1.5-ml Eppendorf tubes and in each washing step required 5-min centrifugations at 270 g.

Determination of ROS
ROS levels in infected cells were measured using 10 μM of the probe 2',7'-dichlorodihydrofluorescein diacetate (H 2 DCFDA), as described [42]. The fluorescent probe was added just prior to measuring, and fluorescence was determined at 1 and 3 hpi using a TECAN Infinite 200 PRO Nanoquant microplate reader (excitation 480 nm; emission 520 nm). Emission values were first blanked against the background fluorescence of non-infected cells and then normalized to the optical density of bacteria grown as the OD 600 of the co-culture was measured at the same time as the fluorescence was read. Cells in dimethylsulfoxide (DMSO) were used as a blank. Measurements were carried out every 5 min for 100 min. To calculate intracellular ROS, PBS buffer, bacteria, and eukaryotic cells without treatment were used as blanks. The difference in fluorescence was calculated and divided by the elapsed time. This value was normalized by the difference in growth during the respective times. To quantify H 2 O 2 , the Amplex1 Red Hydrogen Peroxide/Peroxidase Assay Kit (ThermoFisher) was used following manufac-turer´s instructions. Additionally, HOCl was determined using GFP bleaching. For this, the pGlo plasmid was introduced into each bacterial strain, remaining episomally and induced overnight with arabinose 50 mM [43] and the loss of fluorescence, as an indirect measure of increasing amounts of HOCl, was determined using a TECAN Infinite 200 PRO Nanoquant microplate reader (excitation 395 nm; emission 509 nm).

Total RNA extraction from infected eukaryotic cells
Infected cells were recovered at 1 and 3 h, washed twice with PBS, and then lysed with sodium deoxycholate (0.5% w/v in PBS). RNA extraction was performed by the acid-phenol method, as described [44]; the pellet was suspended in 30 μl nuclease-free water, and stored at -80˚C until used. The integrity of the RNA was determined by 1.0% agarose gel electrophoresis, quantity was determined spectrophotometrically and quality was verified by OD 260 / 280 ratio. The RNA was treated with 2 U of DNase I (Roche) for 1 h to remove contaminant DNA. To ensure no carry-over DNA in the samples, we routinely performed PCR using primers for bacterial 16s RNA and found no product using the RNA extract as template.

qRT-PCR
cDNA synthesis was performed at 37˚C for 1 h in 25 μl of a mixture containing 2.5 pmol of Random Primers (Invitrogen), 10 μl template RNA (5 mg), 0.2 mM dNTPs, 1 μl sterile water, 4 μl of 5× buffer (250 mM Tris-HCl pH 8.3, 375 mM KCl, 15 mM MgCl2, and 10 mM DTT), and 200 U of reverse transcriptase (Invitrogen). Primers used for qRT-PCR are listed in S1 Table. The relative quantification of each transcript was performed using the Brilliant II SYBR Green QPCR Master Reagent and the Mx3000P detection system (Stratagene). The qRT-PCR mixture (20 μl) contained 1 μl of the cDNA template and 120 nM of each primer. The qRT-PCR was performed under the following conditions: 10 min at 95˚C followed by 40 cycles of 30 s at 95˚C, 45 s at 58˚C, and 30 s at 72˚C. Primer pairs were selected with amplification efficiencies of 3.3 ± 10%. Obtained values were used to calculate fold-change expression of target genes, normalized by the expression of a suitable gene selected in these experimental conditions [45]. We used talB gene expression for normalization, which is stable under the studied conditions, based on whole transcriptomic analyses [46]. Validation of the talB gene expression to normalize transcriptomic experiments under oxidative stress is found in S1, S2 and S3 Figs. The selection was supported by comparing stability and expression of talB gene with other commonly used housekeeping genes in our experimental conditions. Expression levels were compared to those of each gene found in strain 14028s at 1 hpi.

Statistical analyses
Gene expression of each mutant strain was calculated relative to wild type. Gene-by-gene comparisons were performed as individual experiments for each time point using one-way ANOVAs with α = 0.05. Statistical analyses were performed with the Bonferroni correction comparing mutant strains with a wild type strain separately at 1 and 3 hpi using GraphPad 5.01 (Prism1).

The ArcAB two-component system is required for bacterial survival in epithelial cells, macrophages, and neutrophils
Bacteria associated with epithelial cells (HEp-2), macrophages (RAW 264.7), and bone-marrow-derived neutrophils were quantified at 1 and 3 hpi to evaluate the ability of ΔarcA and ΔarcB mutants to adhere to cells and to establish an intracellular niche, respectively. Experiments were also performed with the genetically complemented ΔarcA::aph/pBR::arcA and ΔarcB::caf/pBR::arcB strains, which gave similar results to those obtained using the parental strain 14028s. Salmonella can promote its own entry into host cells by first adhering to the cell surface using specialized fimbriae and adhesins [47,48]. No difference in adhesion to any of the tested host cells was observed between strains 14028s and ΔarcA (Fig 1A), an expected result since ArcA does not regulate genes encoding fimbriae [35], and therefore, its absence should not affect Salmonella adhesion. In contrast, loss of ArcB resulted in 2-fold more bacteria harvested from HEp-2 cells, but not from phagocytic cells (Fig 1A). This suggested that absence of the ArcB sensor histidine kinase altered the entry of Salmonella into epithelial cells, and this effect was independent from the ArcA response regulator. At 3 hpi, the number of intracellular bacteria recovered from infections with ΔarcA in the various cell types was significantly lower compared to S. Typhimurium 14028s (Fig 1B), underscoring the importance of ArcA in establishing an intracellular bacterial niche, especially inside epithelial cells and neutrophils. Similarly, the ΔarcB mutant survived poorly in epithelial cells and neutrophils (Fig 1B). As a control, we carried out the same experiment with a previous incubation of the eukaryotic cells with a phagocytosis inhibitor and found that this treatment reduced drastically the number of recovered CFUs (S4 Fig). The reduced intracellular survival of ΔarcA and ΔarcB mutants (Fig 1) could be due to either increased bacterial susceptibility to intracellular microbicidal effects or a differential induction of ROS upon infection. To distinguish between the two possibilities, we determined the amount of ROS produced by infected cells at the same time points postinfection using H 2 DCFDA. This is a fluorogenic probe that in the presence of ROS is converted to the highly florescent compound 2',7'-dichlorofluorescein (DCF). The probe detected similar amounts of ROS over 3 hpi (Fig 2), which remained relatively constant irrespective of the infecting bacterial strain in the assay. This suggested that both macrophages and neutrophils maintain their total ROS levels throughout the first 3 h of infection irrespective of the presence or absence of arcAB gene expression. Similar results were observed with the quantification of H 2 O 2 production in both cell types and HOCl production in neutrophils (S5 and S6 Figs, respectively). Together, we conclude that an intact ArcAB two-component system contributes to intracellular survival of S. Typhimurium in phagocytes and epithelial cells, and the presence or absence of the system does not influence the levels of ROS production induced in host cells.

ArcA and ArcB regulate detoxification-, membrane permeability-, and toxic resistance-related genes inside macrophages and neutrophils
We evaluated the role of the ArcAB two-component system in modulating oxidative stress resistance during intracellular bacterial survival [17]. RNA was extracted from bacteria after infection (RAW 264.7 and bone-marrow-derived neutrophils) at 1 and 3 hpi. These times were representative of two stages in the eukaryotic cell response, namely the oxidative burst and SCV maturation, respectively [49][50][51][52][53].
It was previously demonstrated that ArcA and Hfq mediate the activation of the hilD promoter, especially during early stationary phase and with vigorous aeration [54]. This can trigger activation of SPI-1 resulting in the secretion of effector proteins required for Salmonella uptake by cells [54]. We therefore examined the expression of hilA and sipC, as these genes are important during the early steps of Salmonella intracellular survival. In neutrophils, ArcA was required for induction of sipC expression (Fig 3). SipC is a protein whose function contributes to the maintenance of the phagosome compartment [54]. Also, hilA expression was upregulated by ArcA at 3 hpi under our experimental conditions (Fig 3). In early stationary phase in vitro culture, mutations in arcA also reduced hilA expression 2-fold [54].
Other important factors required for Salmonella survival inside phagocytes are detoxification enzymes including katG, katE, katN, ahpF, and ahpC, which become highly induced in response to hydrogen peroxide [17]. Inside neutrophils, where the main ROS compound is HOCl, the majority genes encoding catalases, alkyl hydroperoxidase, and superoxide dismutase were up-regulated in the absence of arcA (Fig 4). In the context of biochemical reactions occurring inside this particular phagosome, where superoxide and hydrogen peroxide are consumed rapidly to produce HOCl in the presence of the myeloperoxidase enzyme, these products are arguably scarce and short-lived. Only katE was down-regulated in the absence of arcA at 1 hpi (Fig 4). Because the KatE enzyme is expressed during exponential growth [55], it is likely that this result reflects the reduced intracellular growth of arcA mutant bacteria.
Another bacterial protection mechanism against ROS involves the modulation of the composition of the outer membrane. Under H 2 O 2 and HOCl stress in vitro, S. Typhimurium ArcA downregulates genes encoding OmpC, OmpD, OmpF, and OmpW porins [28,29]. Inside neutrophils, however, bacteria through ArcA upregulate ompD, ompF, and ompW (Fig 5), while in macrophages ArcA downregulates ompD and ompF and upregulates ompW (S7 Fig). These results showing different expression patterns related to the presence or absence of arcA and the type of phagocytic cell suggest that bacteria sense different environments and respond accordingly in each cell type, this adaptation capacity allow Salmonella to survive the ROSrelated conditions that the bacteria encounters during infection.
Gene expression related to central metabolism in response to highly toxic HOCl was also evaluated since these pathways are critical to maintain redox state and prevent oxidative damage, and previous work reported that ArcA, in response to 1 mM H 2 O 2 , regulates key genes required for carbon biosynthesis, energy, and maintenance of the redox state [35] as the main bacterial response to these toxic conditions. We examined the expression of a selection of genes encoding enzymes involved in central metabolism. These included: manZ (encoding a subunit from the mannose transporter) and three genes encoding enzymes involved in glycolysis, namely phosphoglucose isomerase (pgi), 2,3-bisphosphoglycerate-independent phosphoglycerate mutase (pmgI), and fructose-bisphosphate aldolase (fbaB). Our results show that genes related to glycolysis are not part of the ArcA regulon. In contrast, manZ is downregulated by ArcA at 3 hpi (Fig 6). A similar finding was also reported in E. coli, where complex posttranslational regulation of this transporter in response to low glucose and phosphate was noted [56]. We also found a similar response in macrophages (S8 Fig).

arcA and arcB are required for successful systemic infection in C57BL/6 mice
To investigate the relevance of arcA and arcB during host infection, in vivo infection assays were conducted in C57BL/6 mice with each single mutant. In comparison to strain 14028s, both ΔarcA and ΔarcB showed reduced abilities to cause infection when administered orally or intraperitoneally. Upon oral administration of 1×10 5 CFU, the recovery of CFUs corresponding to the ΔarcA strain was reduced more than 90% in both the liver and spleen. Likewise, ΔarcB diminished approximately 60-70% in both organs compared to the wild type strain (Fig  7). A similar result was observed after intraperitoneal infection with a 1×10 3 -CFU inoculum. Recovery of the ΔarcA strain was reduced 85% in the liver and 83% in the spleen. The same The ArcAB two component system and its role in Salmonella infection phenomenon was observed when quantifying viable bacteria from the ΔarcB strain (Fig 7), demonstrating that ArcA and ArcB are essential for S. Typhimurium systemic infection in C57BL/6 mice separately, as seen in the CFU recovered from each singe mutant. As a control, we used a ΔslyA strain, which is highly attenuated in mice infections [57,58]. Inactivating SlyA decreased the recovery of bacteria in a manner similar to that observed ΔarcA and ΔarcB strains. These results, combined with the gene expression patterns observed for each single mutant, underscore the importance of ArcA and ArcB for systemic infection.  detoxification (sodA, sodB, katN, katG, katE, ahpC and ahpF) in S. Typhimurium ΔarcA and ΔarcB strains inside bone-marrow-derived murine neutrophils. Ã P < 0.05; ÃÃ P < 0.01; ÃÃÃ P < 0.001. One-way ANOVA with the Bonferroni correction comparing mutant strains with a wild type strain separately at 1 and 3 h postinfection. Biological replicates n = 5, three technical replicates per experiment. https://doi.org/10.1371/journal.pone.0203497.g004 The ArcAB two component system and its role in Salmonella infection

Discussion
ArcA regulates Salmonella cellular metabolism, flagella biosynthesis, and motility under many in vitro conditions, including anaerobiosis, aerobiosis, and the presence of H 2 O 2 [34,35]. Loss   Fig 5. Relative expression of porin genes (ompC, ompD, ompF and ompW) in S. Typhimurium ΔarcA and ΔarcB strains inside bonemarrow-derived murine neutrophils. ΔarcA and ΔarcB strains inside bone-marrow-derived murine neutrophils. Ã P < 0 05; ÃÃ P < 0 01; ÃÃÃ P < 0 001. One-way ANOVA with the Bonferroni correction comparing mutant strains with a wild type strain separately at 1 and 3 h postinfection. Biological replicates n = 5, three technical replicates per experiment. https://doi.org/10.1371/journal.pone.0203497.g005 The ArcAB two component system and its role in Salmonella infection of arcA also reduces expression of SPI-1 genes invF and hilA in liquid cultures [54]. However, the involvement of ArcA in Salmonella intracellular survival remains unclear. In this study, we investigated the involvement of ArcA in cell invasion and survival in epithelial cells, macrophages, and neutrophils in physiological conditions similar to those found in the host. Our   Fig 6. Relative expression of metabolic genes (manZ, pgi, pmgI and fbaB) in S. Typhimurium ΔarcA and ΔarcB strains inside bone-marrowderived murine neutrophils. ΔarcA and ΔarcB strains inside bone-marrow-derived murine neutrophils. Ã P < 0.05; ÃÃ P < 0.01; ÃÃÃ P < 0.001. One-way ANOVA with the Bonferroni correction comparing mutant strains with a wild type strain separately at 1 and 3 hours postinfection. Biological replicates n = 5, three technical replicates per experiment. results revealed that ArcA plays a role in intracellular bacterial survival, as demonstrated by the significant reduction in the number of viable ΔarcA bacteria recovered from each cell type at 3 hpi. The reduced numbers of bacteria were related to their reduced capacity to replicate within cells and not to defects in their ability for invasion/phagocytosis. Significant differences were not observed in the overall production of ROS by infected macrophages and neutrophils, irrespective of the genetic background of the infecting bacterial strain. This observation suggests that the cues triggering ROS production are independent of the ArcAB regulon. However, the bacteria were in an oxidative environment when survival was measured, which was consistent with the peak in ROS production by macrophages and neutrophils upon infection [59][60][61].
We found that at the transcriptional level, ArcA modulates adaptive bacterial responses to stress conditions inside phagocytes. In neutrophils, which display a harsher environment than macrophages, ArcA was implicated in promoting the expression of genes associated with the SPI-1 T3SS such as hilA and sipC, both of which are associated with invasion, as well as formation and maintenance of the SCV [54,62]. Additionally, inside neutrophils, only some of the typical genes that encode detoxifying enzymes were induced, as previously described for macrophages [17]. Of these enzymes, only katE seems to be relevant in the conditions encountered inside neutrophils during the time range evaluated.
The expression of porin genes was also modulated by ArcA in response to several factors found in the context of infection. For instance, OmpC and OmpF also respond to osmolarity and in this case are regulated by EnvZ and OmpR [63]; also permeability varies with glucose The ArcAB two component system and its role in Salmonella infection levels, including OmpW [64,65], while OmpD is modulated by oxygen levels and changing pH values [66]. Additionally, the glycolytic pathway was independent of the ArcA regulon, although it seems to be required for intracellular replication of S. Typhimurium in macrophages [67]. Only the endpoint of the pathway was induced by ArcA in this case, which could be also related to the interconnection between glycolysis and the pentose phosphate pathway.
In the systemic infection model, significantly reduced numbers of viable bacteria were found in spleen and liver, suggesting ArcAB play an important role in the in vivo infection ability of S. Typhimurium. This observation agrees with our results from ex vivo cell infections and also with our transcriptional analyses. These results differ from the previous report by Evans et al. where no differences in virulence were found for arcA and wild type S. Typhimurium [34]. These could be due to methodological variations in the infection assays. In the report by Evans et al., a lower bacterial inoculum (250 CFU per mice) was used and the mice remained alive for several days. In contrast, a higher bacterial dose, as we have used, would cause a more robust level of inflammation by the wild type bacteria leading to rapid killing of the infected mice, while the arcA and arcB mutants are clearly attenuated. We believe our experimental conditions model the infective dose for Salmonella under natural infections, which is high (10 4 −10 8 ) in comparison to other pathogens [68]. Intestinal invasion of S. Typhimurium depends on the bacterial dose and a linear proportion between the inocula (10 4 −10 10 ) and invasion has been shown [69]. These observations were also replicated in the wax moth larvae infection model [70].
Evans et al. also reported no differences in virulence between arcA mutant and parental strains using a competitive index (CI) assay that included high bacterial doses [34], results we also confirmed in our study (S9 Fig). The absence of an effect of arcA and arcB mutants in competition assays at low bacterial doses can be explained by the dose effect discussed above. Under higher doses, we believe the parental strain in the mixed infection can elicit a robust inflammatory response, which would preclude seeing differences with the mutants. Together, we conclude that the ArcAB regulon, due to its modulatory role in central metabolism, fine tunes the expression of several stress-associated pathways having a more noticeable role under conditions of higher bacterial cell density.
The expression analysis of selected genes shows that the ArcAB system may work in a noncognate manner, since there is a different gene expression pattern in each single mutant strain (Figs 4, 5 and 6). This has been previously observed by others under different experimental conditions [71][72][73][74][75], and also noticed in our global transcriptomic analyses in the presence of H 2 O 2 and HOCl (in preparation). It is possible that ArcAB may be noncanonically activated under certain metabolic conditions. For example, intracellular infection could increase the pool of acetyl phosphates, which would activate ArcA independently of ArcB phosphorylation. Global transcriptomic analyses under multiple stress conditions including intracellular infection may offer clues to understand this behavior of the ArcAB system.  [34]. Animals where infected orally (p.o.) of i.p with a 1:1 mixture of S. Typhimurium 14028 and the arcA mutant. Mice were sacrificed at 4 or 6 days p.i and liver and spleen were collected for processing. CI index was calculated as described [34]. (TIF) S1 Table. Primers used for qPCR. (PDF)