RAW246.7 murine macrophages (ATCC, TIB71) were cultivated in Dulbecco’s modified Eagle’s medium (DMEM; Corning, United States) supplemented with 10% fetal bovine serum, 1 mM glutamine, 100 U/μL penicillin and 100 μg/mL streptomycin (all from EuroClone, Italy). The cells were grown in tissue culture flasks or multiwell plates, at 37°C and 5% CO₂. Human monocyte-derived macrophages (HMDMs) were differentiated in vitro from monocytes isolated from the buffy coats of healthy donors, as previously described (Del Porto et al., 2011). Briefly, peripheral blood mononuclear cells were isolated by density gradient centrifugation (Lympholyte; Cedarlane, Hornby, CA, United States), and were selected with an anti-CD14 monoclonal antibody coupled to magnetic beads (Miltenyi Biotec, Bergisch Gladbach, Germany). The CD14⁺ cells were differentiated for 7 days in Roswell Park Memorial Institute (RPMI) 1640 (Gibco-BRL, Invitrogen Corporation, Carlsbad, CA, United States) supplemented with 20% fetal bovine serum and 100 ng/mL recombinant macrophage colony stimulating factor (PeproTech Inch, Rocky Hill, NY, United States).

The bacterial strains and plasmid used in this study are listed in Table 1. The P. aeruginosa mutant strains were recovered from frozen stocks and analyzed for the gene deletions. Genomic DNA extraction and gene amplification were performed as previously reported (Di Domenico et al., 2015), using primers listed in Supplementary Table S1. As expected, sodB and sodM amplification bands were detected in the P. aeruginosa WT (PAO1). In contrast, sodB and sodM amplification bands were absent in the sodB and sodM mutants, respectively (Supplementary Figure S1). All of these strains were grown from single colonies in Luria-Bertani (LB) medium (Sigma, United States) at 30°C, with liquid cultures grown with shaking at 180 rpm. The sodB and sodM mutant strains were routinely grown in LB medium containing 50 μg/mL tetracycline and gentamycin, respectively. As expected from previous studies (Hassett et al., 1995; Iiyama et al., 2007), the sodB mutant grew more slowly than the parental PAO1 strain (Supplementary Figure S2). The P. aeruginosa strains that expressed green fluorescent protein (GFP) were obtained by electroporation of pUC30T-gfpmut3 (Barbier and Damron, 2016) and selection in 15 μg/mL gentamycin on LB agar plates.

The day before infection, the macrophages were seeded into 48-well plates (10⁵ cells/well) in culture medium without antibiotics, and incubated at 37°C in 5% CO₂. When applied, the macrophages were pretreated (i.e., before infection) with the NADPH oxidase inhibitor, diphenyleneiodonium (DPI; Sigma, United States), at 10 μM for 30 min. Exponentially growing P. aeruginosa cells were prepared by refreshing the overnight cultures in LB broth at 30°C (Supplementary Material). After two washes in phosphate-buffered saline (PBS), the P. aeruginosa were resuspended in antibiotic-free cell-culture medium and added to the macrophages at a multiplicity of infection (MOI) of 10. The infection was synchronized by centrifugation of the multiwell plates (550 × g for 5 min), which were then incubated at 37°C in 5% CO₂ for 30 min to 60 min. At the end of the infection, the cells were washed with PBS and incubated in DMEM containing 1 mg/mL amikacin and 1 mg/mL ceftazidime for 15 min. Afterward, the macrophages in selected wells (defined as t₀) were lysed in 1% Triton X-100 for 10 min at room temperature, and finally diluted to 1 mL PBS. The cell viability was determined according to the colony-forming unit (CFU) assay. In the remaining wells, the medium was replaced with culture medium supplemented with a sub-inhibitory concentration of the antibiotics (0.1 mg/mL amikacin; 0.1 mg/mL ceftazidime) and incubated for a further 60 min (t₆₀) or 180 min (t₁₈₀). At the end of the incubations, the live P. aeruginosa were recovered as described above. The bacteria survival was calculated according to Eq. 1:

The macrophages (i.e., RAW 264.7 cells, HMDMs) were seeded in 24-well plates (2 × 10⁵ cells/well) in antibiotics-free medium the day before infection, and then infected with GFP-expressing PAO1 and PAO1 sodB P. aeruginosa strains, at a MOI of 25. Phagocytosis was carried out by incubation of the infected macrophages for 30 min or 60 min at 37°C in 5% CO₂. Afterward, the cells were gently washed two or three times with PBS, enzymatically detached, and analyzed by flow cytometry (BD FACSCalibur, France). Phagocytosis was evaluated as the fraction of GFP⁺ cells in the bulk population. The data were analyzed using the CellQuest software, and the images were processed with FlowJo.

Intracellular O2- levels were measured using luminol (Sigma, United States). Briefly, the macrophages were resuspended in Hank’s balanced salt solution (HBSS) without phenol red (Sigma, United States), supplemented with 25 μg/mL luminol, and seeded in white 96-well plates (Sarstedt, Germany). The RAW264.7 macrophages were seeded at 3 × 10⁵ cells/well, with the HMDMs at 10⁵ cells/well. The macrophages were challenged with the P. aeruginosa strains at a MOI of 10, and the chemiluminescence was measured at given time using a multilabel counter (Wallac 1420 Victor2). The data were corrected based on the controls without macrophages. Quantitative analysis was performed by determination of the areas under the curve (AUC) using the GraphPad Prism software. Furthermore, O2- levels were measured by the nitroblue tetrazolium (NBT, Sigma) reduction assay, which was carried out according to Choi et al. (2006), with minor modifications. Briefly, the macrophages seeded in 24-well plates were supplemented with 1 mg/mL NBT, and infected with the P. aeruginosa strains (MOI = 10). After 60 min of infection, intracellular NBT was solubilized and the optical densities were determined spectrophotometrically (Supplementary Material).

The extracellular H₂O₂ released from the infected macrophages was measured by the production of Resofurin, using Amplex Red assays (Invitrogen). Briefly, HMDMs were seeded in 96-well plates (10⁵ cells/well) in antibiotic-free culture medium the day before infection, and then washed with PBS and incubated with 50 μM Amplex Red and 0.1 U/mL horse-radish peroxidase in KRPG buffer (145 mM NaCl, 5.7 mM sodium phosphate, 4.86 mM KCl, 0.54 mM CaCl₂, 1.22 mM MgSO₄, 5.5 mM glucose, pH 7.35). Macrophages were challenged with P. aeruginosa strains at a MOI of 10 and fluorescence was measured at 30 min intervals with a fluorescence scanner (Amersham Typhoon 9600). Values were corrected for controls without macrophages in KRPG supplemented with 50 μM Amplex Red and 0.1 U/mL horse-radish peroxidase. The concentration of H₂O₂ in the samples was calculated using standard curves obtained with defined H₂O₂ concentrations (0–20 μM).

10⁶ RAW264.7 macrophages were infected with P. aeruginosa strains for 1 h, as described above. After infection, macrophages were gently washed and lysed in Hepes 50 mM pH 7.4, NaCl 150 mM, EDTA 20 mM, NaF 100 mM, Na₃VO₄ 10 mM, 1% Triton X-100, protease inhibitor cocktail. 25–30 μg protein samples were separated by SDS-PAGE and transferred to nitrocellulose blotting membrane (GE Healthcare, Italy). After blocking in PBS, Tween 0.1%, skim milk 5% (SIGMA, United States), membranes were incubated with the primary antibody for 16–18 h at 4°C. Next day, membranes were washed, incubated at room temperature with the horse-radish-peroxidase–conjugated secondary antibody (GE Healthcare, Italy) and visualized with a Chemi Doc XRS system (Bio-Rad Laboratories Ltd., Hemel Hempstead, United Kingdom). Quantitative Western blotting was performed using the ImageJ software. The primary antibodies used were: LC3 (Cell Signaling, Italy; #2775), diluted at 1:1000; and GAPDH (Santa Cruz, Italy; sc-47724), diluted at 1:400.

All of the data are reported as a means ± standard deviation. The Figures and statistical analyses were constructed using the GraphPad Prism software (GraphPad Software Inc.). The statistical tests used are indicated in the corresponding Figures. Differences were significant for a P-value cut-off of 0.05.

The periplasmic P. aeruginosa SOD might contribute directly to the scavenging of the superoxide anion (O2-) produced by the macrophage NOX2, with a possible impact on the bacterial survival. To test this hypothesis, we analyzed the survival of the P. aeruginosa sod mutants in the macrophages. For this, RAW264.7 macrophages were infected with WT PAO1 or the sod mutants, as either PAO1 sodB or sodM, and the live intracellular bacteria were determined using the CFU assay. These data showed greater intracellular survival of the PAO1 sodB mutant, with respect to the PAO1 WT and PAO1 sodM, with these last two showing similar survivals (Figure 1A). As expected, inhibition of NADPH oxidase by the DPI pretreatment resulted in significant increases in the live P. aeruginosa recovered from the macrophages infected with PAO1 WT or PAO1 sodM, which confirmed the oxidative burst as the primary killing mechanism in the macrophages early after infection (Figure 1B). However, no differences were detected in the PAO1 sodB survival in DPI-treated macrophages, with respect to the untreated cells (Figures 1A,B). These data suggested that SODB contributes to the NADPH-dependent killing of P. aeruginosa by the macrophages. Similarly, HMDMs showed significant increases in the intracellular survival of PAO1 sodB, but not PAO1 sodM, with respect to PAO1, in both the untreated and DPI-treated cells (Figures 1C,D). This thus extended the role of P. aeruginosa SODB to primary human macrophages.

To further support these data, we analyzed phagocytosis by determining the fraction of macrophages that engulfed the PAO1 or PAO1 sodB strains. RAW264.7 macrophages and HMDMs were infected with GFP-expressing strains for 30 and 60 min. Subsequently, the non-internalized P. aeruginosa were removed by several washes, and the sub-population of infected macrophages (GFP⁺) was evaluated by flow cytometry. Figures 2A,D show no differences in the GFP⁺ sub-population between the cells infected with PAO1 WT and the PAO1 sodB mutant, in both the untreated and DPI-treated macrophages. Furthermore, quantitative analysis confirmed that the GFP⁺ sub-populations were similar, irrespective of whether the P. aeruginosa infecting strain was PAO1 WT or the PAO1 sodB mutant, and whether macrophages were untreated (Figures 2B,E) or DPI-treated (Figures 2C,F). Collectively, these data strongly suggest that the phagocytosis of PAO1 and PAO1 sodB was substantially similar at 30 and 60 min after infection.

Having observed greater survival of PAO1 sodB with respect to PAO1 WT and no differences in bacterial phagocytosis, we hypothesized that the bacterial SOD, and in particular SODB, contributes to the killing of the intracellular P. aeruginosa through the modulation of NADPH-dependent ROS production.

Taking into consideration the enzymatic activities of NADPH oxidase and SOD, the most likely hypothesis to explain these data was that the superoxide radicals produced by the macrophage NOX2 were promptly converted into H₂O₂ by the bacterial SOD, which in turn contributes to the killing of the intracellular P. aeruginosa. If this is the case, the macrophages infected with the sod mutants should be characterized by higher O2- and lower H₂O₂ levels than those infected with PAO1 WT. Therefore, we analyzed the O2- and H₂O₂ levels in macrophages infected with PAO1 WT and the sodB mutant. To differentiate between O2- and H₂O₂, we took advantage of their different mobilities through the membrane. Indeed, while the mobility of O2- is highly restricted by its negative charge, as H₂O₂ is neutral and long-lived, it can leak from the cells (Fang, 2011). Thus, O2- was measured intracellularly by luminol oxidation (Bedouhène et al., 2017), and H₂O₂ extracellularly using Amplex Red assays (Schürmann et al., 2017).

Macrophages challenged with PAO1 WT and PAO1 sodB were loaded with luminol, and ROS production was determined by chemiluminescence measurements. Kinetics analysis revealed a rapid increase in luminol chemiluminescence, which is a readout of O2- production, with peaks at 15 to 25 min (Figure 3). Additionally, the macrophages infected with PAO1 sodB showed higher chemiluminescence signals than those infected with PAO1 WT, both for the RAW264.7 cells (Figure 3A) and the HMDMs (Figure 3C). As expected, inhibition of NADPH oxidase with DPI greatly reduced the luminol oxidation (Figures 3A,C). Quantitative analysis of the luminol chemiluminescence AUC confirmed the significantly higher O2- production in the cells infected with PAO1 sodB, with respect to PAO1 WT (Figures 3B,D). Similar data were obtained with nitroblue tetrazolium reduction assays, which detected higher levels of the superoxide anion in macrophages infected with the sodB mutant, with respect to PAO1 WT (Supplementary Figure S3). Collectively, these data suggest that P. aeruginosa SODB contributes to O2- dismutation, and that the extent of its contribution depends on the total amount of O2- produced by the host cells, which, in our experimental models is higher in human than murine macrophages.

According to our hypothesis, the absence of bacterial SODB activity increases O2- levels and decreases H₂O₂ production in macrophages infected with the sodB mutant. To confirm this, we evaluated extracellular H₂O₂ leakage using the Amplex Red assay. As expected, infection of HMDMs with the P. aeruginosa sodB mutant was associated with lower H₂O₂ release, with respect to cells infected with PAO1 WT (Figure 4). No signal above the background (i.e., uninfected macrophages) was detected for the RAW264.7 macrophages, possibly due to the low levels of ROS produced upon infection (data not shown).

Overall, the macrophages infected with the PAO1 sodB mutants were characterized by higher O2- levels and lower H₂O₂ levels, with respect to the cells infected with the WT bacteria, This supported the concept that P. aeruginosa SODB converts the phagocytic O2- to H₂O₂.

It has been well established that macrophages can kill engulfed bacteria using different mechanisms that can be activated sequentially, starting with the oxidative burst, which in turn activates other mechanisms, including autophagy (Huang et al., 2009; Lam et al., 2010). Thus, to evaluate the possible role of the bacterial SODB in microbicidal mechanisms other than the oxidative burst, we analyzed P. aeruginosa survival at later time points after infection. After infection and treatment with antibiotics, the RAW264.7 macrophages were incubated in bacteria-free medium for 1 h and 3 h, and the live intracellular bacteria were counted using the CFU assay (Figures 5A–C). As previously observed, at the end of the infection (t₀), more viable PAO1 sodB were recovered from the infected macrophages than the PAO1 WT (Figures 5A,B, time 0). However, the CFU recovered from the infected macrophages at 1 h and 3 h from the infection were similar, independent of the bacterial strains (Figures 5A,B). Similar data were obtained in the HMDMs (Figures 5D,E). Consequently, the long-term survival of PAO1 WT appeared to be greater than that of PAO1 sodB in both of these macrophage models (Figures 5C,F). This was also described by the slopes of the killing curves, which represent the rates of bacterial killing, and which were higher for PAO1 sodB, with respect to PAO1 WT (Supplementary Table S2).

Collectively, these data demonstrate that the long-term survival of the PAO1 sodB mutant in both murine and human macrophages is less than for PAO1 WT, which suggests that SODB ultimately promotes bacteria survival.

It has been shown that autophagy activation by Toll-like receptor signaling or receptor-mediated phagocytosis depends on NOX2 activity and ROS production, with the latter required to recruit LC3 to phagosomes and to target intracellular Salmonella typhimurium to (auto)phagosomes (Huang et al., 2009). Additionally, autophagy activation by P. aeruginosa has been demonstrated in different cell types, including macrophages, where it contributes to restrict intracellular survival (Yuan et al., 2012; Junkins et al., 2013; Jabir et al., 2014). Having observed that P. aeruginosa SODB contributed to the long-term survival of P. aeruginosa within macrophages, we speculated that it might inhibit autophagy through reduction of the levels of intracellular O2-. To test this hypothesis, RAW264.7 cells and HMDM macrophages were infected with PAO1 WT and PAO1 sodB, and autophagy activation was monitored by analysis of changes in the levels of lipidated LC3 (LC3-II) with Western blotting (Kabeya et al., 2000). In whole-cell lysates from these non-infected macrophages, LC3-I and LC3-II were detected at similar levels in the RAW264.7 cells, while in HMDMs the intensity of the LC3-II band was slightly higher than that of LC3-I. In contrast, 1 h after infection, LC3-II increased specifically in the infected macrophages (Figures 6A,C). Accordingly, quantitative Western blotting showed higher levels of LC3-II in cells infected with the PAO1 sodB mutant, with respect to those infected with PAO1 WT (Figures 6B,D). This suggests that the SODB activity of the intracellular P. aeruginosa reduces autophagy activation, which thus provides a link between the oxidative and non-oxidative roles of SODB in modulation of P. aeruginosa survival in macrophages.