Arginine Metabolism Powers Salmonella Resistance to Oxidative Stress

ABSTRACT

Salmonella invades host cells and replicates inside acidified, remodeled vacuoles that are exposed to reactive oxygen species (ROS) generated by the innate immune response. Oxidative products of the phagocyte NADPH oxidase mediate antimicrobial activity, in part, by collapsing the ΔpH of intracellular Salmonella. Given the role of arginine in bacterial resistance to acidic pH, we screened a library of 54 single-gene mutants in Salmonella that are each involved in, but do not entirely block, arginine metabolism. We identified several mutants that affected Salmonella virulence in mice. The triple mutant ΔargCBH, which is deficient in arginine biosynthesis, was attenuated in immunocompetent mice, but recovered virulence in phagocyte NADPH oxidase deficient Cybb^−/− mice. Furthermore, ΔargCBH Salmonella was profoundly susceptible to the bacteriostatic and bactericidal effects of hydrogen peroxide. Peroxide stress led to a larger collapse of the ΔpH in ΔargCBH mutants than occurred in wild-type Salmonella. The addition of exogenous arginine rescued ΔargCBH Salmonella from peroxide-induced ΔpH collapse and killing. Combined, these observations suggest that arginine metabolism is a hitherto unknown determinant of virulence that contributes to the antioxidant defenses of Salmonella by preserving pH homeostasis. In the absence of phagocyte NADPH oxidase-produced ROS, host cell-derived l-arginine appears to satisfy the needs of intracellular Salmonella. However, under oxidative stress, Salmonella must additionally rely on de novo biosynthesis to maintain full virulence.

INTRODUCTION

Despite significant progress by global health initiatives to ensure access to potable water, adequate sanitation, and food safety education, diarrheal disease caused by pathogens like Salmonella remains one of the top 10 leading causes of mortality worldwide (1). Non-typhoidal Salmonella are responsible for roughly 1.3 billion cases of intestinal disease per year (2). Given the broad host range, high prevalence in the food chain (3), and rising rates of antimicrobial resistance (4), Salmonella infections are a major public health concern.

After crossing the intestinal epithelium, Salmonella establishes an intracellular niche inside host epithelial cells and macrophages, replicating inside a modified endosome termed the Salmonella-containing vacuole (SCV). Within the SCV of professional phagocytes, Salmonella are exposed to reactive oxygen species (ROS) produced by the phagocyte NADPH oxidase (NOX2), one of the most potent arms of the innate host defense (5). NOX2-generated ROS can damage DNA, iron-sulfur cluster-containing proteins, and protein cysteine and methionine residues (6, 7). To survive the initial response to infection that is dominated by ROS produced by phagocyte NADPH oxidase, Salmonella employs a multitude of antioxidant enzymes (8, 9). For instance, superoxide dismutases degrade superoxide anions (O₂-) into oxygen (O₂) and hydrogen peroxide (H₂O₂) (10), catalases decompose H₂O₂ into O₂ and H₂O, peroxidases catalyze the reduction of H₂O₂ and other hydroperoxides into H₂O using a variety of reductants (11), and the low-molecular-weight thiol glutathione can scavenge H₂O₂ (12). Yet, despite decades of research on the topic, the mechanisms by which Salmonella protects itself against the ROS produced during the innate host response are incompletely understood (13, 14).

During periods of high electron flow through the electron transport chain (ETC), damage caused by NOX2 to the NADH dehydrogenase encoded by the nuo operon results in a collapse in the bacterial ΔpH, disrupting membrane energetics, and precipitating increased protein misfolding events (15). In response, Salmonella undergoes overflow metabolism and utilizes fermentation-produced acetate to stabilize the alkaline pH of the bacterial cytoplasm (15, 16). Here, we investigated whether there are additional mechanisms Salmonella could employ to maintain ΔpH during oxidative stress.

Salmonella utilizes a number of different inducible and constitutive strategies to maintain pH homeostasis (17). Constitutive acid resistance systems, which enable survival at the typical gastric pH of less than 2.5, depend on the presence of l-arginine (18, 19), lysine (20), or ornithine (21). Inducible acid tolerance systems in Salmonella require a pre-adaption exposure to a mildly acidic pH prior to an acid challenge at pH ≥ 3 (22). The inducible acid tolerance systems include acid shock proteins that are controlled by RpoS, OmpR, Fur, and PhoPQ (23), regulation of F₀F₁-ATPase (24–26), and induction of amino acid decarboxylase (27) or deiminase (28) systems.

l-arginine is the only amino acid identified as one of the seven key nutrients that support full Salmonella virulence in mice (29). While both l-arginine catabolism (28) and l-arginine transport (30) are known to be critical for Salmonella virulence in mice, it is not yet known whether this necessity for l-arginine during host infection is related to resistance to the innate host response. Interestingly, research in other organisms implicates l-arginine catabolism in resistance to oxidative stress (31, 32), and Salmonella deficient in both cytoplasmic superoxide dismutases displays increased levels of l-arginine (33). Given these observations, the following experiments sought to explore how Salmonella utilizes l-arginine metabolism during host infection to promote pathogenesis during periods dominated by the antimicrobial activity of phagocyte NADPH oxidase.

RESULTS

l-arginine metabolism is a critical determinant of Salmonella pathogenesis.

In the following investigations, we examined the role of l-arginine metabolism in Salmonella pathogenesis. Towards this end, a targeted, single-gene knockout library of 54 genes involved in l-arginine metabolism and several neutral control mutants was screened in a C57BL/6 murine model of acute S. Typhimurium infection. This screen identified several critical “pinch points” in l-arginine biosynthesis, transport, and catabolism that impact the ability of Salmonella to establish an infection in the liver tissue of C57BL/6 mice (Fig. 1; Table S1). Deletions in the argA or argG genes, which are both required for de novo l-arginine biosynthesis, substantially decreased Salmonella fitness in vivo. In addition, analysis of the combined effect of deletions in argCBH arginine biosynthesis genes significantly (P = 0.0077) attenuated Salmonella in C57BL/6 mice. These data suggest that de novo arginine biosynthesis from α-ketoglutarate is necessary for Salmonella to maintain full virulence in the host. Our screen also identified l-arginine transport as an important determinant of bacterial survival in the host. Salmonella harboring a deletion in either artP or artQ, subunits of the l-arginine uptake system encoded by the artPIQM-artJ gene cluster, or in argT, the periplasmic binding protein of the argT-hisJQMP gene cluster (34), suffered impaired growth during infection. Additionally, deletion of a putative l-arginine/ornithine antiporter, ydgI, decreased bacterial survival in the host. Combined, the decreased fitness of this facultative intracellular pathogen in the absence of either l-arginine transport or biosynthesis suggests that l-arginine is a critical nutrient that promotes Salmonella survival in the host. Not only do our results complement the results of a previous study that showed that l-arginine transport-deficient Salmonella is attenuated in BALB/c mice (30), but our data also suggests that Salmonella’s l-arginine requirements during infection cannot be maintained exclusively by transport from the host, but must also involve de novo synthesis.

FIG 1.

Significant hits from an arginine metabolism-specific single-gene KO library screen in mice. Five C57BL/6 mice were inoculated intraperitoneally with 3,000 CFU of a library containing barcoded deletion mutants of 54 arginine metabolism genes, and then livers and spleens were harvested 3 days after infection. Log₂ fold changes in barcode abundance were determined between the input and output libraries. Fitnesses with an adjusted P-value of less than 0.05 in both organs are marked. Color key: fitness defect (red), no difference (blue), not tested (no color). Figure created with BioRender.com.

Our screen revealed that l-arginine deiminase pathway genes, STM14_5361 (arcA) and arcC, were also important for Salmonella virulence in the C57BL/6 mouse model. The attenuation of the ΔSTM14_5361 and ΔarcC mutants in C57BL/6 mice is consistent with a prior study that showed that the l-arginine deiminase pathway contributes to the virulence of Salmonella in C3H/HeN mice (28). While the results of our in vivo screen indicate that both host-derived and de novo-synthesized l-arginine appear to be important for Salmonella fitness in the host, the fitness advantage conferred by the l-arginine deiminase pathway may reflect the eventual fate of these l-arginine pools. In light of these observations, the following investigations sought to characterize the mechanism by which l-arginine biosynthesis contributes to Salmonella virulence.

l-arginine metabolism enhances Salmonella oxidative stress resistance.

To further analyze the role of l-arginine biosynthesis in Salmonella pathogenesis, we generated a deletion mutant deficient in the argCBH operon, which was identified in our screen as a critical component of Salmonella fitness in vivo. As expected from previous studies in related organisms (35), ΔargCBH Salmonella exhibited an auxotrophy for l-arginine as demonstrated by the inability of this mutant to grow in E salts minimal medium containing glucose (EG) as the only carbon source (Fig. 2a). The growth defect of ΔargCBH Salmonella in EG minimal medium was reversed upon complementation with the low copy plasmid pWSK29 expressing the argCBH operon from its native promoter or upon the addition of exogenous l-arginine (Fig. 2a).

FIG 2.

l-arginine biosynthesis is important during oxidative stress. (a) Aerobic growth of wild-type (WT) and ΔargCBH mutant Salmonella in EG minimal media in the presence or absence of 80 μg/mL supplemental l-arginine as measured by OD₆₀₀ with a Biotek Synergy H1 plate reader. Data are shown as mean ± SEM (n = 2 to 6). (b) Aerobic growth of the indicated Salmonella mutants in the presence of 150 μM H₂O₂ in EG minimal media supplemented with 80 μg/mL l-arginine as measured by OD₆₀₀ with a Biotek Synergy H1 plate reader. Data are shown as means ± SEM (n = 3 to 9). (c) Competitive index (CI) of the ΔargCBH mutant and WT Salmonella inoculated in competition into C57BL/6 and Cybb^−/− mice. 250 CFU of each strain was inoculated intraperitoneally and then harvested 4.5 days after infection for C57BL/6 mice or 2.5 days after infection for Cybb^−/− mice. Data were analyzed by unpaired, two-tailed t test with Welch’s correction where ****, P < 0.0001. Box and whisker plots represent minimums to maximums, 25th and 75th percentiles, and medians. n = 5 for C57BL/6 and n = 8 for Cybb^−/− mice.

To investigate whether l-arginine metabolism contributes to Salmonella’s resistance to oxidative stress, wild-type and ΔargCBH Salmonella were challenged with bacteriostatic concentrations (150 μM) of hydrogen peroxide in EG minimal media supplemented with 80 μg/mL l-arginine (Fig. 2b). Despite sufficient l-arginine availability to fully restore the growth of the mutant to wild-type levels, ΔargCBH Salmonella took significantly longer than wild-type controls to recover from the H₂O₂ challenge. We also tested the virulence of ΔargCBH Salmonella by inoculating C57BL/6 mice with equal numbers of mutant and wild-type Salmonella. ΔargCBH Salmonella were at a competitive disadvantage for growth against wild-type controls in C57BL/6 mice. However, the l-arginine auxotroph was able to recover virulence in Cybb^−/− mice lacking the membrane-bound gp91phox subunit of the phagocyte NADPH oxidase (Fig. 2c). Together, these data demonstrate that l-arginine biosynthesis plays a previously unappreciated role in Salmonella’s defense against the ROS generated by phagocyte NADPH oxidase.

l-arginine metabolism mediates peroxide resistance in the presence of oxygen.

To understand how l-arginine levels in Salmonella could impact resistance to ROS, an l-arginine re-feed protocol was utilized to probe the consequence of a low l-arginine supply in ΔargCBH Salmonella (Fig. S1). Briefly, mutant and wild-type Salmonella grown overnight in a limiting amount of exogenous l-arginine (20 μg/mL) were resuspended in fresh EG minimal medium that contained 80 μg/mL l-arginine, a nonlimiting concentration that promotes growth of the auxotroph ΔargCBH Salmonella (Fig. S2). Following a 30-min incubation period in l-arginine-containing media, bacterial cultures were then treated for an additional 30 min with 400 μM H₂O₂ in a shaking incubator. Using this approach, we uncovered that the l-arginine auxotroph is remarkably susceptible to the bactericidal effects of H₂O₂. Without functional l-arginine biosynthesis or sufficient intracellular l-arginine stores, the susceptibility of Salmonella to peroxide stress was so severe that within 30 min of peroxide exposure there was a 3-log reduction in the number of surviving ΔargCBH Salmonella (Fig. 3a). In comparison, wild-type Salmonella completely resisted H₂O₂ killing under the experimental conditions used in these investigations. Building off prior research in Salmonella that showed that electron flux through the electron transport chain can dictate susceptibility to H₂O₂ (15), we employed the re-feeding assay to microaerobic and anaerobic cultures. When grown overnight in fully aerated conditions and then incubated under anaerobic conditions in fresh EG minimal medium for 30 min prior to treatment with H₂O₂, ΔargCBH Salmonella was partially resistant to the drastic killing seen in fully oxygenated media (Fig. 3b). In contrast, when both grown overnight and treated with H₂O₂ under anaerobic conditions, the hypersusceptibility of ΔargCBH Salmonella to peroxide stress was completely abolished (Fig. 3c). Combined, these observations suggest that l-arginine metabolism protects Salmonella against the toxic effect of ROS during periods of high electron flow through the electron transport chain.

FIG 3.

l-arginine biosynthesis-dependent peroxide killing is exacerbated in the presence of oxygen. Survival of wild-type (WT) and ΔargCBH mutant Salmonella grown overnight in EG minimal media supplemented with 20 μg/mL l-arginine after 30 min of treatment with 400 μM H₂O₂ in EG supplemented with 80 μg/mL l-arginine. (a) Salmonella in aerobic experiments were grown overnight and treated with H₂O₂ in the presence of oxygen. (b) Salmonella in aerobic to anaerobic shift experiments were grown overnight in the presence of oxygen and then treated with H₂O₂ under anaerobic conditions. (c) Salmonella in anaerobic experiments were grown overnight and treated with H₂O₂ under anaerobic conditions. Data are shown as means ± SEM (n = 6) and was analyzed by two-way ANOVA with Tukey’s correction where *, P < 0.05; ****, P < 0.0001.

l-arginine metabolism buffers Salmonella’s cytoplasm during oxidative stress.

Expanding on these observations, we began to further dissect the molecular mechanisms that drive Salmonella’s reliance on l-arginine metabolism to resist oxidative stress. We uncovered that ΔargCBH Salmonella grown overnight in a limiting amount of l-arginine was hypersusceptible to the bactericidal effects of H₂O_2, regardless of the presence of l-arginine in the resuspension media (Fig. 4). In contrast, ΔargCBH Salmonella grown overnight in EG minimal medium that contained 80 μg/mL l-arginine was resistant to the bactericidal effects of H₂O₂, regardless of whether there was l-arginine in the resuspension media or not.

FIG 4.

l-arginine levels help to maintain resistance to oxidative stress. Salmonella grown overnight in EG minimal media supplemented with either 20 μg/mL l-arginine (Low Arg.) or 80 μg/mL l-arginine (High Arg.) was resuspended in either EG media (No Arg.) or EG media with 80 μg/mL l-arginine (High Arg.) for 30 min prior to treatment. Susceptibility of wild-type (WT) and ΔargCBH mutant Salmonella to 30 min treatment of 400 μM H₂O₂ in EG media +/− l-arginine. Data are shown as means ± SEM (n = 6) and was analyzed by two-way ANOVA with Tukey’s correction where **, P < 0.01; ***, P < 0.001.

Increased respiratory activity has previously been associated with hypersusceptibility to oxidative stress (15). To understand if the concentration of l-arginine available to the auxotroph affected respiration levels, thus contributing to the increased susceptibility of ΔargCBH Salmonella to H₂O₂, we measured the oxygen consumption rate under each individual condition (Fig. 5a). Compared to wild-type controls, subculturing of ΔargCBH Salmonella grown in EG minimal medium containing 20 μg/mL or 80 μg/mL l-arginine into fresh EG minimal medium without l-arginine stopped respiratory activity. In contrast, moving ΔargCBH Salmonella from either EG minimal media containing 20 μg/mL or 80 μg/mL l-arginine into fresh EG minimal medium containing l-arginine did not cause any measurable arrests in aerobic respiration. Neither decreased nor increased respiratory capacity appeared to be predictive of susceptibility to peroxide stress. These results were surprising given that the susceptibility of ΔargCBH Salmonella to peroxide killing was greatly diminished under anaerobic conditions and that the sensitivity of Salmonella to oxidative stress is positively correlated with aerobic respiratory activity (15).

FIG 5.

l-arginine levels help to maintain pH homeostasis during oxidative stress. All assays were performed using Salmonella grown overnight in EG minimal media supplemented with either 20 μg/mL l-arginine (Low Arg.) or 80 μg/mL l-arginine (High Arg.) and then resuspended in either EG media (No Arg.) or EG media with 80 μg/mL l-arginine (High Arg.) for 30 min prior to measurements. (a) Effect of supplemental l-arginine on respiration of wild-type (WT) and ΔargCBH mutant Salmonella in EG media +/− l-arginine. Aerobic respiration was measured polarographically. Data are shown as means ± SEM (n = 5). (b) Intracellular pH of wild-type (WT) and ΔargCBH mutant Salmonella measured ratiometrically with pHluorin before and after treatment with 400 μM H₂O₂ in EG media +/− l-arginine. Data are shown as means ± SEM (n = 6) and was analyzed by unpaired, two-tailed t test where *, P < 0.05; ****, P < 0.0001.

Given recent research that the killing activity of phagocyte NADPH oxidase collapses the cytoplasmic ΔpH of Salmonella (15) and the known role of l-arginine catabolism in maintaining pH homeostasis during acid stress (18, 19), we sought to measure the changes in intracellular pH during peroxide challenge in Salmonella grown in varying l-arginine concentrations (Fig. 5b). Compared to wild-type controls, the intracellular pH of ΔargCBH Salmonella grown overnight in 20 μg/mL of l-arginine was collapsed to media pH levels after incubation in fresh EG minimal medium without l-arginine. However, when moving from 20 μg/mL to 80 μg/mL l-arginine media, ΔargCBH Salmonella and wild-type controls had identical intracytoplasmic pH values. Interestingly upon H₂O₂ treatment, ΔargCBH Salmonella grown in 20 μg/mL l-arginine suffered a rapid and complete collapse of the intracellular pH below the media pH, regardless of the presence of l-arginine in the fresh media or not. Moreover, ΔargCBH Salmonella grown overnight in EG minimal medium with 80 μg/mL l-arginine had a similar starting pH as wild-type controls when incubated in fresh media containing l-arginine, but a lower starting pH when incubated in fresh media without l-arginine. However, in striking contrast to the results of the 20 μg/mL l-arginine cultures, 80 μg/mL l-arginine cultures of ΔargCBH Salmonella were able to maintain their intracellular pH at or above media pH levels when challenged with H₂O₂. Combined, these data indicate that a collapse in intracellular ΔpH is more predictive of the susceptibility of ΔargCBH Salmonella to oxidative stress than changes in respiratory flux through the ETC.

DISCUSSION

l-arginine metabolism is an essential determinant of virulence in Salmonella (28–30). However, we are still lacking a mechanistic understanding of why l-arginine metabolism is important for Salmonella during infection. Herein, we demonstrate that Salmonella relies on l-arginine metabolism both in vivo and in vitro to resist the antimicrobial actions of ROS produced by the phagocyte NADPH oxidase. We show that increased l-arginine concentrations protect the pathogen from peroxide-induced cytoplasmic ΔpH collapse and subsequent killing. Thus, l-arginine metabolism contributes to the resistance of Salmonella to oxidative killing by preserving pH homeostasis.

Defects in l-arginine biosynthesis impair the virulence of diverse pathogenic bacteria in the host, despite the possession of functional l-arginine transport systems (36–39). Our study shows that ΔargCBH S. Typhimurium auxotrophic for l-arginine are attenuated for virulence in the host. The restoration of virulence of ΔargCBH Salmonella to wild-type levels in Cybb^−/− mice that lack ROS generated by the phagocyte NADPH oxidase suggests that l-arginine biosynthesis plays a role in bacterial resistance to oxidative stress. Consistent with this observation, ΔargCBH Salmonella are hypersusceptible to both the bacteriostatic and bactericidal activity of exogenous H₂O₂. The dispensability of l-arginine biosynthesis for Salmonella virulence in the absence of oxidative stress demonstrates either a specific necessity for biosynthesis during oxidative stress or an overall higher l-arginine requirement during oxidative stress that cannot be met by transport alone. The hypersusceptibility of ΔargCBH S. Typhimurium to peroxide stress can be rescued if the bacterial cells are grown in higher concentrations of l-arginine (80 μg/mL) prior to challenge, which lends credence to the idea that Salmonella experiences high l-arginine requirements during the oxidative stress engendered in the innate response in C57BL/6 mice.

During periods of high electron flow through the ETC, the killing activity of the phagocyte NADPH oxidase acts through a collapse in the intracellular ΔpH (15). In response to ROS impairing the proton-pumping NDH-I isoform of NADH dehydrogenase, Salmonella shifts redox balance and ATP production from the ETC and oxidative phosphorylation to fermentation and substrate-level phosphorylation. Synthesis of acetate from acetyl-CoA helps Salmonella to maintain a ΔpH by consuming intracellular protons that accumulate in the absence of sufficient NDH-I activity. The work herein demonstrates that l-arginine metabolism must be considered as an additional component by which Salmonella maintain ΔpH during periods of oxidative stress. Salmonella possesses three different catabolic pathways that could be functioning to preserve pH homeostasis through either direct consumption of intracellular protons or production of ammonia to act as a proton sponge during oxidative stress: the l-arginine decarboxylase (ADC) pathway, the l-arginine deiminase (ADI) pathway, and the l-arginine succinyltransferase (AST) pathway. Salmonella bearing deletions in the inducible, degradative form of l-arginine decarboxylase (adiA) are known to be impaired in their ability to tolerate acidic pH (18, 21). Interestingly, research in E. coli has already demonstrated a role for adiA in resistance to both acid and oxidative stress in vitro (31). Although the role of the ADI pathway in acid stress has not yet been evaluated in Salmonella, the ADI pathway is both active and critical for full virulence in mice (28). Additionally, work in Staphylococcus epidermidis has identified the importance of the ADI pathway in ameliorating endogenous oxidative stress conferred during acid stress (32). The second step in the AST pathway catalyzed by succinylarginine dihydrolase degrades the guanidino group of succinylarginine yielding succinylornithine and two ammonia. Based on the current literature, the likelihood of the AST pathway acting as a major pH homeostasis mechanism seems low as the pathway appears to mainly function as a mechanism to regenerate glutamate during nitrogen-limiting conditions (40). While either the ADI or ADC pathway could be the primary contributor to pH homeostasis during oxidative stress, the identification of the proton-consuming enzymes in the ADI pathway in our fitness screen may suggest that under in vivo conditions, the ADI pathway is more relevant than the ADC pathway.

In addition to considering that catabolism of l-arginine may be acting to directly preserve pH homeostasis during oxidative stress, other alternative or parallel mechanisms could explain the role arginine biosynthesis plays in Salmonella pathogenesis and resistance to oxidative stress. l-arginine may also serve as a source of nitrogen during infection, catabolized through the AST pathway to generate glutamate for diverse anabolic reactions such as the production of the antioxidant glutathione. However, this particular explanation seems unlikely given the absence of any fitness defect in AST pathway genes seen in our in vivo screen. While our results indicate a specific role for l-arginine biosynthesis during oxidative stress, the identification of l-arginine transporters in our fitness screen may suggest that l-arginine metabolism is also important for dampening the effects of reactive nitrogen species generated by inducible nitric oxide (NO) synthase (iNOS). Many pathogens have evolved strategies not only to detoxify NO, but also to divert and deplete host l-arginine to limit NO production (41). Salmonella upregulates both host arginase II (42) and host cationic amino acid transporters (30) during infection, facilitating a shift in host l-arginine catabolism away from iNOS and transport of host-derived l-arginine into the SCV. Stealing l-arginine from the host not only dampens NO production, but may also serve as a means to regulate T cell proliferation (43). Interestingly, amino acid transport mutants have been previously shown to exacerbate virulence defects tied to amino acid biosynthesis (44). Lastly, the role of l-arginine metabolism in Salmonella pathogenesis may be explained by the downstream formation of polyamines through the ADC pathway. Polyamines have been implicated in resistance to oxidative stress in diverse organisms (45); however, in S. Typhimurium, it appears that polyamines play only a minor role in oxidative stress resistance and are more important for defense against nitrosative stress (46). These alternative explanations are not incompatible with the arginine catabolism theory and could be functioning in parallel.

Respiratory flux may be functionally linked to resistance of Salmonella to oxidative stress, as strongly suggested by the fact that aerobically and anaerobically respiring Salmonella are more sensitive to peroxide killing than anaerobically fermenting cultures (15). Salmonella undergoing oxidative stress increases glucose consumption to fuel energetic and redox needs through a combination of substrate-level phosphorylation and aerobic respiration (16). Adding to these observations, we demonstrate that anaerobic fermentation can rescue hypersusceptibility to oxidative stress; however, our results also indicate that respiration alone is not fully predictive of increased sensitivity to oxidative stress. Salmonella that has stopped respiring after removal of l-arginine is not less sensitive to oxidative stress than actively respiring cells. What appears to be more predictive of hypersusceptibility to oxidative stress is the collapse of ΔpH associated with low levels of l-arginine. This observation is in line with the recently proposed model of oxidative killing by which a failure to maintain pH homeostasis and membrane energetics affects the redox status of proteins that rely on the ETC to catalyze the folding of periplasmic proteins (15). Inability to maintain pH homeostasis may also have profound effects on fundamental aspects of bacterial physiology, shifting the cytoplasmic pH into a range that is not compatible with enzymatic activity or transcription in general (13).

In summary, Salmonella l-arginine metabolism is an important determinant of pathogenesis during the selective pressure imposed by the phagocyte NADPH oxidase. However, in the absence of oxidative stress, l-arginine biosynthesis appears to be dispensable for Salmonella pathogenesis.

MATERIALS AND METHODS

Bacterial strains, plasmids, and growth conditions.

Wild-type Salmonella enterica subsp. enterica serovar Typhimurium strain 14028s and its mutant derivatives were used for all assays. The deletion mutant was constructed using the λ-Red homologous recombination system as previously described (47). The chromosomal deletion was confirmed by genotyping PCR with external check primers followed by Sanger sequencing to confirm the correct orientation of the antibiotic cassette. The complementation plasmid was generated by amplification of the gene of interest from the wild-type chromosome followed by ligation of the gene fragment into the low-copy number plasmid pWSK29 (48). The complementation plasmid was confirmed by Sanger sequencing with internal and external check primers. Strains were stored at −80°C in LB + 7% DMSO broth. See Tables S2 to S4 for the details of all strains and plasmids used in this study.

Cultures were routinely started from either a small amount of −80°C DMSO stock or a single colony, which was inoculated into the appropriate liquid growth medium and grown overnight (16 h to 20 h) aerobically at 37°C with shaking. The liquid growth media used for all assays was either Lysogeny broth (LB) or essential salts minimal media with glucose (EG) (1.7 mM MgSO₄, 9.5 mM citric acid, 57.4 mM K₂HPO₄, 16.7 mM H₅NNaO₄P, 0.4% d-Glucose, pH 7.0 unless otherwise noted). As appropriate, penicillin or chloramphenicol was added at final concentrations of 250 or 20 μg/mL, respectively. To determine CFU, bacterial cultures were serially diluted 10-fold in phosphate-buffered saline (PBS) and spread onto LB agar plates followed by growth at 37°C for 16 h.

Growth assays.

Salmonella grown overnight in LB were diluted to 2.5 × 10⁶ CFU/mL into fresh EG minimal media in the presence or absence of 80 μg/mL supplemental l-arginine. For growth assays in the presence of hydrogen peroxide, 150 μM H₂O₂ was added directly to the cultures after resuspension into fresh media. The optical densities at 600 nm (OD₆₀₀) of 200 μL aliquots of bacterial cultures were recorded with a Biotek Synergy H1 plate reader (Agilent, Santa Clara, CA) every 15 min for 40 h. The time (h) at which each culture reached an OD₆₀₀ of 0.2 was calculated to determine statistical significance.

Ethics statement and murine model.

All mice were used according to protocols approved by the Institutional Animal Care and Use Committee (IACUC) at the University of Colorado Anschutz Medical Campus School of Medicine. C57BL/6J and congenic Cybb^−/− mice deficient in the gp91phox membrane subunit of the phagocyte NADPH oxidase (49) originated from breeders obtained from The Jackson Laboratory. Cybb^−/− mice were regularly verified by genotyping PCR against the C57BL/6J background controls. All mice were housed in static cages in groups of three to five mice per cage under specific pathogen-free conditions. During experiments, mice were housed in static cages in groups of three to five mice per cage under BSL2 and specific pathogen-free conditions. All experiments used 6- to 8-week-old male and female mice, and animals were assigned randomly to experimental groups. Animal subjects were not involved in any previous procedures prior to experimentation.

Library construction, sequencing, and analysis.

The targeted ARG gene deletion library consisted of 54 specific single gene deletion mutants representing genes participating in Salmonella’s arginine metabolism. Barcodes were introduced into the respective mutants of a previously published collection of Kanamycin resistant mutants (50) using Lambda-red recombination of a Tetracycline resistance cassette amplified from the S. enterica sv Typhimurium LT2 strain TT25401 (https://rothlab.ucdavis.edu/drugs/tetra.shtml). Recombination relied on extensive homology between the original mutants and the barcode-containing primers PCR_L1 and PCR_R1 in Table S3. The ARG pool also contained seven differently barcoded malXY mutants as neutral controls with no expected phenotypes.

Five mice were given an intraperitoneal injection of 3,000 CFU of the input pool. The input pools and output pools from spleens and livers were prepared for sequencing essentially as previously described (51). Briefly, bacteria were harvested and grown in LB + Tet, pelleted, washed with water, and lysed. The lysate was subsequently used as template for PCR using indexed Illumina sequencing primers that annealed directly, flanking both sides of the barcodes in primer PCR _L1. The frequency of each barcode was then determined by Illumina sequencing. Raw outputs were normalized to the median of the barcoded malXY neutral controls. Significance was determined using a two-tailed t test. The ratio of the medians of the organs and inputs were reported as log₂ fold changes.

Murine competition assays.

Mice were inoculated intraperitoneally (i.p.) with ~500 CFU of a mixture containing equal numbers (250 CFU strain 1 + 250 CFU strain 2) of two strains of Salmonella that were each grown separately overnight in LB. Mice were humanely euthanized 3 days after inoculation by CO₂ inhalation followed by cervical dislocation. Spleens and livers were harvested from each mouse and emulsified in 5 or 3 mL PBS, respectively. CFU were quantified by serial dilution onto LB agar plates containing the appropriate antibiotics. Plates were incubated for 16 h at 37°C prior to determining colony counts. Competitive index was calculated as follows: ([strain 1/strain 2]_output/[strain 1/strain 2]_input).

Susceptibility to H₂O₂.

Salmonella grown overnight in EG media with either 20 μg/mL or 80 μg/mL l-arginine were resuspended to an OD₆₀₀ of 0.4 in 3 mL of either fresh EG media or EG media + 80 μg/mL l-arginine. Cultures were then allowed to equilibrate in the fresh media for 30 min at 37°C with shaking prior to taking a “time zero” sample for serial dilution onto LB agar. The cells were challenged by adding 400 μM H₂O₂ directly to the cultures and incubating for another 30 min at 37°C with shaking. The number of surviving bacteria at “time 2” were quantified after serial dilution onto LB agar and following incubation at 37°C for 16 h. Percent survival was calculated as follows: ([CFU from “time 2”/CFU from “time zero”] × 100).

Polarographic O₂ measurements.

Consumption of O₂ was measured using an ISO-OXY-2 O₂ sensor attached to an APOLLO 4000 free radical analyzer (World Precision Instruments, Inc., Sarasota, FL) as described in Husain et al. (52). Briefly, Salmonella cultures grown overnight in EG media with either 20 μg/mL or 80 μg/mL l-arginine were resuspended to an OD₆₀₀ of 0.4/mL in either fresh EG media or EG media + 80 μg/mL l-arginine. Cultures were then allowed to equilibrate in the fresh media for 30 min at 37°C with shaking. Then, 3 mL of bacterial cultures were rapidly withdrawn, vortexed for 1 min, and immediately recorded for O₂ consumption. A two-point calibration for 0% and 21% O₂ was done as per manufacturer’s instructions.

Intracellular pH determination.

To estimate the intracytoplasmic pH of Salmonella, the pH-sensitive pHluorin plasmid driven from the PBAD promoter was used (53). Salmonella grown overnight in EG media with 0.2% L-arabinose and either 20 μg/mL or 80 μg/mL l-arginine were resuspended to an OD₆₀₀ of 0.2 in 3 mL of either fresh EG media or EG media + 80 μg/mL l-arginine. Cultures were then allowed to equilibrate in the fresh media for 30 min at 37°C with shaking. The cultures were scored for fluorescence using a Shimadzu R5300C spectrofluorometer with an excitation wavelength scan from 300 to 490 nm and emission at 510 nm for 5 min. The cells were then treated with 400 μM H₂O₂ and the excitation spectra was measured for an additional 15 min. To determine the value of internal pH from the obtained fluorescence ratio (390/480 nm), pHluorin-expressing Salmonella were incubated in EG media + 80 μg/mL l-arginine pH 5.0 to 8.5 in the presence of protonophore potassium benzoate. A Boltzmann sigmoid best-fit curve was applied to the plot of fluorescence ratio and the known pH values obtained using the standards (Fig. S3).

Statistical analysis.

All statistical analyses were performed using GraphPad Prism Version 9.0 for macOS (GraphPad Software, La Jolla California USA, www.graphpad.com). P values of < 0.05 (two-tailed) were considered statistically significant. *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001. For analyzing in vivo data, the unpaired, parametric, two-tailed t test was used. Each dot within the scatterplot represents data from one animal (biological replicate). Box and whisker plots represent minimums to maximums, 25th and 75th percentiles, and medians. For analyzing in vitro data, the two-way ANOVA with Tukey’s correction was used. Each dot within the interleaved scatterplot represents one biological replicate (average of two technical replicates per biological replicate). Bars represent mean ± SEM.