Bacterial metallothionein, PmtA, a novel stress protein found on the bacterial surface of Pseudomonas aeruginosa and involved in management of oxidative stress and phagocytosis

Overexpression of a wild-type gene can be a powerful tool to identify mutant phenotypes that may be missed by observing phenotypes associated with a clean deletion (25). We constructed a pmtA-overexpressing mutant (P_BADpmtA) using a broad-host range Tn7 transposon with a P_BAD arabinose-inducible promoter, pTJ19 (26) (Fig. S1). This P_BAD Tn7 contains the promoter of the arabinose operon and its regulatory gene, araC, allowing for tight control of expression (27). We inserted pmtA downstream of this promoter using EcoRI and HindIII restriction sites and the Gibson Assembly protocol. We verified the gene was in the correct orientation by sequencing and then conjugated this suicide delivery vector (pBADpmtA) into PAO1 (WT) generating P_BADpmtA (Fig. S1A). The insertion of the Tn7 into the chromosome was verified by PCR amplification of the glmS region, allowing for less concern for vector plasmid maintenance and burden within bacterial cells (Fig. S1B) (20, 26, 28, 29). No growth rate changes were observed in P_BADpmtA when grown in tryptic soy broth (TSB), Luria–Bertani (LB) (supplemented various levels of arabinose) or M9-A (arabinose as a sole carbon source) media when compared to either the WT or to other PAO1 isogenic strains (Table 1; Fig. S2 to S4). Using quantitative PCR (qPCR), we tested the effects of different media (i.e., TSB plus 1% arabinose, TSB-A; LB plus 1% arabinose, LB-A; M9-A; and mammalian cell infection media plus 1% arabinose [MCIM-A]) on the expression of pmtA in WT PAO1 alone and found that the amount of pmtA expressed did not change with nutrient availability (Fig. 1A). We next tested promotor activation in P_BADpmtA grown in M9-A, LB-A, TSB-A, or MCIM-A with arabinose and found that in the presence of arabinose, P_BADpmtA showed a 15- to 30-fold increase in pmtA expression when compared to wild type (Fig. 1B through E). The MCIM-A had the highest fold change difference in expression of pmtA when compared to WT: this could be due to the presence of fetal bovine serum (FBS) present in this media. Arabinose is the second most abundant pentose in plants and a component of cattle feed as such, which may represent an additional source of arabinose in the MCIM-A (30). Alternatively, the pBAD promoter has been shown to be regulated by virulence factor regulator (Vfr) which could be activated in the presences of serum (31). Since we have previously shown that the deletion of pmtA in PAO1 results in diminished pyocyanin production and biofilm formation, we investigated these same phenotypes in P_BADpmtA (Fig. 2). Inducing overexpression of pmtA in TSB 1% arabinose leads to more pyocyanin production (Fig. 2A) and greater biofilm formation (Fig. 2B and C) when compared to similarly treated WT and other isogenic PAO1 strains (Table 1). Interestingly, the overproduction of pyocyanin in P_BADpmtA does not lead to a decrease in the growth rate when compared to WT as has been seen in other studies where pyocyanin was induced (Fig. S2), suggesting that PmtA can also provide endogenous protection from pyocyanin-induced oxidative stress (32).

A variety of ROS are produced by the host immune system as a defense mechanism against invading pathogens. The damage to cells by the oxidation of proteins, lipids, or nucleic acids can be extensive, causing disruptions of metabolic processes, membrane destabilization, and other effects (36). P. aeruginosa can employ several mechanisms to respond to ROS stress (36). Given its intrinsic thiol content, PmtA may contribute to antioxidant mechanisms that protect against immune-derived oxidants. Pyocyanin secretion has also been shown to have toxic effects on P. aeruginosa secreting these molecules. P. aeruginosa offsets these effects by biofilm formation, energy decreases, and by the production of catalase, which can diminish oxidative stress (36). Since pyocyanin production has been shown to be nutrient dependent, we compared pyocyanin levels produced in cultures grown in M9-A to cultures grown in complex media (TSB). Pyocyanin is not secreted at detectable levels in M9-A media compared to TSB (Fig. 3A and B), and so. M9-A was used to minimize the presence of pyocyanin in experiments designed to determine the effects of PmtA on ROS exposure (treating cells with 40 mM H₂O₂ or 0.3% NaOCl for 90 mins and calculating the percent survival) (Fig. 3C and D). When we compare P_BADpmtA to the other PAO1 strains, overexpression of pmtA resulted in a higher percent survival when exposed to these two different oxidants, indicating a role for PmtA in relieving oxidative stress. The deletion mutant (ΔpmtA) was not significantly different from WT in oxidant sensitivity, which we have also seen in previous studies (20), indicating that when PmtA is not present, P. aeruginosa can use other mechanisms to address oxidative stress.

Since evidence suggests that PmtA plays a role in the management of oxidative stress and in a model of innate virulence (20), we explored the effect of PmtA on phagocytosis in co-culture with a human macrophage cell line (THP-1, ATCC TIB-202). For these experiments, THP-1 cells were treated to differentiate them to an adherent macrophage-like phenotype (37), then were seeded and cultured for 3 days prior to infection (38). Cultures were then infected with ∆pmtA, P_BADpmtA, or WT PAO1 strains (grown in M9-A, washed in phosphate-buffered saline [PBS], and resuspended in MCIM-A) at a multiplicity of infection (MOI) of 10. Live, intracellular bacteria in these infected cultures were determined by plating and counting CFU/mL obtained from cultures that had been antibiotic treated and washed. Unexpectedly, at both time points after infection (30 and 60 min), significantly more live intracellular ∆pmtA were observed when compared to both WT and P_BADpmtA PAO1 strains (Fig. 4A and B). We also observed significantly fewer live intracellular P_BADpmtA when compared to both ∆pmtA and wild-type PAO1 strains at these time points (Fig. 4A and B). These data indicate that the loss of pmtA expression results in increased phagocytosis, while overexpressing pmtA reduces the frequency of phagocytic engulfment.

We visualized phagocytosis of P. aeruginosa by tagging each of the PAO1 strains with a fluorescent marker encoded on Tn7 with a P_BAD arabinose-inducible promoter controlling ORFs. The fluorescent proteins dTomato or superfolder GFP (sfGFP) are regulated in Tn7 by constitutively active P_Tac promoters (Table 1; Fig. S5). We conjugated the dTomato fluorescent/Tn7 into P. aeruginosa ∆pmtA and into WT PAO1 strains and verified with PCR (Fig. S5C). We also verified that the pmtA mutation was still present in ∆pmtA:dtomato after selection of the fluorescent marker transformants and observed no changes in growth profiles (Fig. S5D and E). P_BADpmtA was labeled with sfGFP fluorescent protein using the same approach. We performed the phagocytosis assay using tagged ∆pmtA:dtomato, P_BADpmtA:sfgfp, and WTPAO1:dtomato strains, taking fluorescent images at times 0, 30, and 60 min (Fig. 5). At time 0, extracellular bacteria are visible. Gentamicin was added to coincubations of bacteria with THP-1 at 30 and 60 min, after which the cultures were washed and imaged. Most of the extracellular bacteria were removed by this treatment except for a few that remained attached to the glass bottom of the dish. More ∆pmtA bacterial cells were closely associated with THP-1 cells when compared to wild-type and P_BADpmtA PAO1 strains, indicating a role of pmtA in attachment and phagocytosis.

We repeated the phagocytosis assay at the 30-min timepoint and visualized the cultures using confocal microscopy, capturing images across the focal planes that span the THP-1 cell volume (Fig. 6). At 30 min, we again observed more ∆pmtA bacterial cells both inside the THP-1 cells and at the cell surface when compared to the wild-type PAO1 strain.

To identify pathogens for phagocytosis, macrophages use pattern recognition receptors (PRR) to bind microbial-associated molecular patterns (MAMPs) (39). Since we observe more phagocytosis and close association of ∆pmtA bacterial cells to THP-1 macrophages, we next examined if PmtA can be found on the surface of WT PAO1. Using a commercially available anti-PmtA monoclonal antibody and fluorescein isothiocyanate (FITC)-labeled goat anti-mouse IgG secondary antibody, we evaluated the binding of anti-PmtA to PAO1 WT strain using fluorescence microscopy (Fig. S6). We observed that PAO1 cells, first incubated with monoclonal mouse IgG1 anti-PmtA followed by a secondary FITC-labeled goat anti-mouse IgG antibody, stained the bacterial cells at a higher level than PAO1 cells exposed to the secondary antibody alone (Fig. S6). We quantified this binding using both flow cytometry and ELISA (40) (Fig. S7; Fig. 7A and B). Anti-PmtA was incubated with each of the three isogenic strains, and then, binding was detected with Alexa Fluor 647-labeled goat anti-mouse IgG. Flow cytometry revealed that monoclonal anti-PmtA antibody binds to PmtA that is available on the surface of WT PAO1 and other pmtA-expressing PAO1 strains (Fig. 7A) while having significantly lower binding to ∆pmtA. Surface binding of the anti-PmtA mAb can also be detected by ELISA using whole bacteria from PAO1 strains that have been immobilized on the ELISA microwell surface (Fig. 7B). We observed significantly more binding of anti-PmtA on PAO1-expressing strains than on ∆pmtA. While monoclonal antibodies are designed to be specific, binding of antibodies to intact, live bacteria has been shown to occur at a lower level owing to nonspecific interactions of antibodies with components of the cell wall, extracellular polysaccharides, and/or extracellular DNA (41, 42).

We also evaluated the role of PmtA as a contributor to intracellular survival after phagocytic engulfment. To do this, we repeated the phagocytosis assay as already described, but at 60 min post-infection, we added gentamicin and then continued the THP-1/P. aeruginosa co-incubation for another 60 min (Fig. 8). The THP-1 cells were then lysed, and the lysate plated on Tryptic Soy Agar (TSA). We calculated the percent persistence by comparing CFU at 60 min and CFU at 2 h. We found that significantly fewer intracellular ∆pmtA bacterial cells were recovered after this 2-h incubation when compared to similarly treated P_BADpmtA and WT PAO1. These data indicate a role for PmtA in intracellular bacterial persistence that could relate to instances of prolonged or chronic P. aeruginosa infection in vivo.

All chemicals and reagents were purchased from Fisher Scientific (Pittsburgh, PA) unless otherwise stated.

THP-1 cells were cultivated in RPMI-1640 medium (ATCC 30–2001) supplemented with 10% fetal bovine serum, 100 U/mL of penicillin, and 100 mg/mL of streptomycin. The cells were grown in tissue culture flasks, multiwell plates, or glass bottom dishes (WillCo-dish, GWSt-3533) at 37°C and 5% CO₂. Differentiation into macrophages was induced by treatment with 40 ng/mL of phorbol 12-myristate 13-acetate (PMA) (ThermoFisher Scientific J63916.M resuspended in dimethyl sulfoxide) for 72 h.

P. aeruginosa strain PAO1 (53) was used as the wild-type (WT) strain, and all P. aeruginosa strains were grown at 37°C in LB containing 10 g of Bacto Tryptone, 5 g of yeast extract, and 10 g of NaCl per liter or TSB or M9 media (arabinose as sole carbon) (20) or on agar-solidified plates or MCIM (RPMI-1640 medium 236.25 mL (ATCC 30–2001), 3.5% FBS, 20 mM HEPES). When bacterial media was supplemented with 1% arabinose, the media was designated as LB-A, TSB-A, M9-A, or MCIM-A). Where a reduced amount of arabinose was added, the concentration was noted. For agar plates, 15 g/L of Bacto Agar (BA) was added. BA contained 44 g of Columbia BA base per liter. For each assay, overnight cultures were established from a single colony. All relevant bacterial strains and plasmids are described in Table 1. Antibiotic selection for conjugations was performed with 1.5 mg/mL of trimethoprim and 25 mg/mL of spectinomycin as previously described (26). Plasmids were maintained in Escherichia coli using 100 mg/mL of trimethoprim, 100 mg/mL of ampicillin, and 100 mg/mL of kanamycin.

The full-length of the pmtA gene (258 bp) was PCR amplified from P. aeruginosa strain PAO1 genomic DNA (gDNA) using the primer pairs pmtAFEcoR1 (TACCCATGGGATCTGATAAGAATTCATGAACAGCGAAACCTGTGC) and pmtARHindIII (GTACCGGGCCCGCGGCCGCAAGCTTGCTCCTCAGGGCGAGATC). The PCR mixture contained 2× Phusion High-Fidelity PCR Master Mix with HF Buffer (New England Biolabs [NEB]; Cat.#: M0531S/L), 10 mM of pmtAF, 10 mM of pmtAR, and 25 ng/mL of gDNA in a final volume of 50 mL. The amplification conditions were as follows: (i) 4 min at 98°C and (ii) 30 cycles of 25 s at 98°C, 30 s at 63°C; and 60 s at 72°C; and (iii) 10 min at 72°C. Simultaneously, pTJ19 was linearized using EcoRI-HR (NEB R3101T) and HindIII-HF (NEB R3104T) following manufacturer’s suggested conditions. Amplification fragments were assembled using Gibson assembly (NEB E2611S) and transformed into NEB5-alpha (NEB C2987H) competent E. coli cells following manufacturer conditions yielding pBADPmtA. The suicide delivery vector (pBADPmtA) was verified for correct fragment orientation by sequencing using Plasmidsaurus.com ( Fig. S1). A quadriparental mating was performed to conjugally transfer the Tn7 with a P_BAD arabinose-inducible promoter in front of the pmtA gene from pBADPmtA into the glmS intergenic region of PAO1, yielding strain P_BADpmtA. The insertion of the Tn7 downstream of glmS was verified as previously described (20, 26, 29, 54) (Fig. S1).

PAO1 strains were also fluorescently tagged by generated a Tn7 with trimethoprim plus ORFs encoding the fluorescent proteins dTomato or sfGFP. Expression scaffolds and Tac promoters were PCR amplified from pTW41552 and pTW416 (35) using the following primer sets, respectively, and as described above: GFPpmtAF (GCCTTCGCGAGGTACCAAGGGCAGATTGTGTCGACC) and GFPpmtAR (TTGCGGCCGCGGGCCCGCTAATTCGATCATGCATGAGCTCAC) dtomatoR (ACCGGGCCCGCGGCCGCAAGCTTGCATGAGCTCACTAAATTCTTGACAATTAAT) and dtomatoF (TCGACCTGCAGGCATGCAAGCTTGACCATGTGGTCACGCTTTTCG). For the over-expressing PAO1 strain using a GFP tag, pBADPmtA was linearized using restriction enzymes ApaI (NEB R0114S) and KpnI (R3142S) following manufacturer conditions. For both WT PAO1 and DpmtA using the dTomato tag, pTJ19 was linearized using HindIII-HF (NEB R3104T) following manufacturer’s recommended conditions. Amplification fragments were assembled using Gibson assembly (NEB E2611S) and transformed into NEB5-alpha (NEB C2987H) competent E. coli cells following manufacturer’s conditions yielding the following suicide delivery vectors, pPBADpmtAGFP, and pdTomatoPBADTpTn7 (Fig. 5S). These suicide delivery vectors were verified for correct fragment orientation by sequencing using Plasmidsaurus.com (Fig. S5). A quadriparental mating was performed as described above in WT PAO1 and DpmtA yielding; PAO1dtomato, DpmtAdtomato, and P_BADpmtAGFP. The insertion of the Tn7 downstream of glmS was verified as previously described (20, 26, 28, 29, 33, 54) (Fig. S5C). To reverify that DpmtAdtomato maintained the clean deletion, pmtA was PCR amplified as previously described from all new strains (Fig. S5D lane 3) (20).

P. aeruginosa strains were grown for 30 h in triplicate, and 1 mL of cell-free supernatants were harvested. Pyocyanin was measured as previously described (20).

Biofilm formation was quantified using a previously described microtiter plate assay (20).

Colony morphology assays were performed as previously described (55–57), with the following modifications: solid media was composed of 1% tryptone, 1% agar, 40 mg/mL of Congo red (Fisher Scientific, S70401-1), and 20 mg/mL of Brilliant Blue Coomassie (Fisher Scientific, BP101-25), which were added into 500 mL of deionized water and autoclaved for 20 min before pouring plates. Overnight cultures of DpmtA, P_BADpmtA, and WT PAO1 strains were grown for 24 h in M9 media. Five microliters of overnight cultures was spotted on colony morphology assay media and incubated at 25°C, >95% humidity for 5 days. Images of the pink centers from the biofilm were analyzed using ImageJ software (58).

Overnight cultures of DpmtA, P_BADpmtA, and WT PAO1 strains were grown for 24 h in M9 media. From each strain, 10⁸ cells were added to 1 mL of fresh M9 media, and starting the CFU/mL was verified by serial dilutions and plating. Cells were exposed to either 40 mM of H₂O₂ or 0.3% NAOCl for 90 min at 37°C. Then, cultures were serial diluted and plated for end point CFU/mL determination. Percent survival was calculated using end point CFU/mL/starting CFU/mL.

Overnight cultures of DpmtA, P_BADpmtA, and WT PAO1 strains were grown for 24 h in M9-A, TSB-A, LB-A, or MCIM-A. Total RNA was extracted using manufacturer’s RNA extraction protocol from MasterPure DNA and RNA purification kit (Biosearch technologies). RNA was converted to cDNA with an iScript cDNA synthesis kit (Bio-Rad, 1708890). Gene-specific primers to pmtA and the housekeeping ribosomal gene (RNA polymerase, subunit alpha) were used as previously described for qPCR with the following modifications (12). All reactions were setup in 20-mL volume including the following: iTaq Universal SYBR Green Supermix (50% of reaction) (Bio-rad 1725120) 10 mM forward and reverse primers (15% of reaction), nuclease-free H₂O (10% of reaction), and 100 ng of cDNA. Negative controls with no template were prepared and tested with each set of reactions. Reactions were amplified in triplicate using a CFX96 Real-Time Thermocycler (Bio-Rad) following the thermal cycle protocol as recommended by the manufacturer. The relative gene expression for pmtA was calculated using the cycle threshold (Ct) value compared to RNA polymerase, with the alpha subunit (RpoA) gene as the internal reference standard. Fold changes were expressed as 2^−ΔΔCt values.

THP-1 cells were differentiated using PMA (38) and seeded in 24-well plates (1 × 10⁵ cells/well), 3 days before infection with PAO1 strains at an MOI of 10. Phagocytosis assays were carried out by growing PAO1 strains in M9-A media for 24 h. A total of 10⁸ cells were washed 2 times with PBS and serial diluted in infection media containing 1% arabinose. Cells were then incubated for 30 or 60 min at 37°C in 5% CO₂ in air. Gentamicin (300 mg/mL) was then added to each well, and an additional 30 min of incubation was performed. Afterward, the cells were gently washed three times with PBS, and macrophages were detached with 2 mM of EDTA and lysed with 0.1% Triton-X 100. The lysed cells were serially diluted and plated on TSA, and CFU/mL was calculated.

For imaging, the differentiated THP-1 cells (1 × 10⁶) were plated 3 days before infection on Bioptechs 30-mm glass bottom culture dishes. Cells were exposed to PAO1 strains at an MOI of 10 and gentamicin treated/washed as described above. Cells were imaged on a Zeiss Axiovert 100 fluorescence microscope with a ×63 oil immersion objective or on Nikon AXR confocal with ×66 oil immersion objective. For confocal images, Celltracker Green (Invitrogen, catalog no. C7025) was used to visualize THP-1 cells. Images were acquired using an Imaging QIClick camera and processed using ImageJ (58).

PAO1 was grown overnight in M9 media, and 10⁸ cells were washed twice with PBS. The bacteria were resuspended in 20 mL of PBS. Bacteria were then heat fixed onto the slide, and 20 mM of mouse monoclonal IgG₁ anti-PmtA (StressMarq Biosciences Inc., SMC-553, clone 8D8) in 50 mL of PBS was incubated on the fixed cells for 1 h at RT. The slide was washed 3× with 1000 mL of PBS. FITC-labeled goat anti-mouse IgG secondary antibody (SouthernBiotech, 1030-02) was diluted 1:400 in PBS, added to the slide and incubated 1 h at room temperature (RT) in the dark. The slide was washed 3× with 1,000 mL of PBS, and then, cells were imaged on a Zeiss Axiovert 100 fluorescence microscope with a ×63 oil immersion objective or on Nikon AXR confocal with a ×66 oil immersion objective. Images were acquired using an Imaging QIClick camera and processed using ImageJ (58).

Flow cytometry binding assay was performed as previously described with the following modifications (40). P. aeruginosa strains were grown overnight in M9 media, and 10⁷ bacterial cells were washed twice with 1 mL of PBS. Cells were then resuspended in 100 mL with 1.67 × 10⁻⁵ M of mouse anti-PmtA. Bacteria–antibody mixtures were incubated for 1 h at 4°C then spun down for 5 min at 15,000 × g. Bacterial cells were then washed with 1 mL of PBS. Bacteria were then resuspended, pelleted, and then resuspended in 1:400 Alexa Fluor 647-labeled goat anti-mouse IgG for 1 h at 4°C in the dark. Bacteria were then spun down for 5 min at 15,000 × g and washed with 1 mL of PBS. Cells were then fixed for 1 h at 4°C with 4% (vol/vol) paraformaldehyde. After fixation, cells were centrifuged at 15,000 × g for 5 min and resuspended in 300 mL of PBS. Samples were analyzed using flow cytometry on a BD Fortessa. Bacteria were gated using their forward and side scatter profile and analyzed for Alexa-Fluor 647 intensity.

PmtA accessibility on the surface of intact bacterial cells was tested with monoclonal anti-PmtA binding using ELISA in Immulon 2HB 96-well plates as previously described with the following modifications (40). P. aeruginosa strains were grown overnight in M9 media, and 10⁸ bacterial cells were washed twice with 1 mL of PBS. Microtiter plates then were coated with 100 mL of 5 × 10⁷ CFU prewashed cells resuspended in PBS and incubated overnight at 4°C. Plates were then washed and blocked with 2% bovine serum albumin overnight at 4°C. Following another wash step, the plate was then incubated with 5 mg/mL of anti-PmtA for 1.5 h at RT, subsequently washed, and incubated with goat anti-mouse Ig-AP (SouthernBiotech, 1010-04) for 1.5 h at RT. After the final wash, 1 mg/mL of p-nitrophenyl phosphate substrate was added to the plates, and optical density values at 405 nm were measured every 30 s for 10 min using a Spectramax (Molecular Devices) plate reader.

THP-1 cells were differentiated using PMA and seeded in 24-well plates (1 × 10⁵ cells/well) 3 days before infection with PAO1 strains at an MOI of 10. Phagocytosis assay was performed as described above, and infected cells were incubated for 60 min at 37°C in 5% CO₂. Then, 300 mg/mL of gentamicin was added to each well. For selected wells, an additional 30 min of incubation was performed. Cells were gently washed three times with PBS, and macrophages were detached with 2 mM of EDTA and lysed with 0.1% Triton-X 100 at time zero (CFUt₀). In the remaining wells, gentamicin (300 mg/mL) was added and then incubated for an addition 60 min (CFUt₆₀). There was no difference in gentamicin sensitivity between the different P. aeruginosa strains. At the end of the incubation, the cells were gently washed three times with PBS, and macrophages were detached with 2 mM of EDTA and lysed with 0.1% Triton-X 100. The lysed cells were serially diluted and plated on TSA, and CFU/mL and percent persistence (P = CFU_t60/CFU_t0 × 100) was calculated.