Interbacterial Antagonism Mediated by a Released Polysaccharide

ABSTRACT

Pseudomonas aeruginosa and Staphylococcus aureus are two common pathogens causing chronic infections in the lungs of people with cystic fibrosis (CF) and in wounds, suggesting that these two organisms coexist in vivo. However, P. aeruginosa utilizes various mechanisms to antagonize S. aureus when these organisms are grown together in vitro. Here, we suggest a novel role for Psl in antagonizing S. aureus growth. Psl is an exopolysaccharide that exists in both cell-associated and cell-free forms and is important for biofilm formation in P. aeruginosa. When grown in planktonic coculture with a P. aeruginosa psl mutant, S. aureus had increased survival compared to when it was grown with wild-type P. aeruginosa. We found that cell-free Psl was critical for the killing, as purified cell-free Psl was sufficient to kill S. aureus. Transmission electron microscopy of S. aureus treated with Psl revealed disrupted cell envelopes, suggesting that Psl causes S. aureus cell lysis. This was independent of known mechanisms used by P. aeruginosa to antagonize S. aureus. Cell-free Psl could also promote S. aureus killing during growth in in vivo-like conditions. We also found that Psl production in P. aeruginosa CF clinical isolates positively correlated with the ability to kill S. aureus. This could be a result of P. aeruginosa coevolution with S. aureus in CF lungs. In conclusion, this study defines a novel role for P. aeruginosa Psl in killing S. aureus, potentially impacting the coexistence of these two opportunistic pathogens in vivo.

IMPORTANCE Pseudomonas aeruginosa and Staphylococcus aureus are two important opportunistic human pathogens commonly coisolated from clinical samples. However, P. aeruginosa can utilize various mechanisms to antagonize S. aureus in vitro. Here, we investigated the interactions between these two organisms and report a novel role for P. aeruginosa exopolysaccharide Psl in killing S. aureus. We found that cell-free Psl could kill S. aureus in vitro, possibly by inducing cell lysis. This was also observed in conditions reflective of in vivo scenarios. In accord with this, Psl production in P. aeruginosa clinical isolates positively correlated with their ability to kill S. aureus. Together, our data suggest a role for Psl in affecting the coexistence of P. aeruginosa and S. aureus in vivo.

INTRODUCTION

Cystic fibrosis (CF) is a genetic disease that results in an abnormally thick mucus lining in the lung, which causes chronic pulmonary infections and reduced lung function, leading to high morbidity and mortality in this cohort (1, 2). Pseudomonas aeruginosa and Staphylococcus aureus are two dominant microorganisms colonizing CF airways (3, 4). These organisms are commonly coisolated from CF sputum samples, with one-third of people with CF being coinfected with both pathogens (5). P. aeruginosa and S. aureus are also coisolated from chronic wound environments (6). Coinfection with both pathogens correlates with increased disease severity (7–9).

Although they coexist in vivo, P. aeruginosa antagonizes S. aureus growth in vitro. The P. aeruginosa quorum-sensing system PQS (Pseudomonas quinolone signal) (10), protease LasA (11), iron sequestering pyoverdine (12), and rhamnolipid (13) all inhibit the planktonic growth of S. aureus in vitro.

Psl (polysaccharide synthesis locus) is a P. aeruginosa exopolysaccharide consisting of a pentasaccharide repeat including d-glucose₁, d-mannose₃, and l-rhamnose₁ (14). There are two forms of Psl, a cell-associated form that is associated with the outer membrane and a cell-free form that is released into the extracellular environment. The molecular determinants for Psl localization are, however, unknown. The amount of Psl produced varies between P. aeruginosa strains and clinical isolates (15–17). Psl functions in early biofilm formation and maintaining biofilm structure (18). There have been several reports that Psl or other P. aeruginosa polysaccharides may be involved in interactions with different bacterial species, specifically, Staphylococcus spp. Staphylococcus protein A (SpA) can bind to P. aeruginosa clinical isolates via Psl and type IV pili (19). P. aeruginosa exopolysaccharides, including Psl and Pel, can disrupt established Staphylococcus epidermidis biofilms (20). It has also been reported that Psl does not affect the relative abundance of P. aeruginosa and S. aureus in a mature dual-species biofilm but does reduce S. aureus aggregate formation in the early stages of biofilm formation (21).

Given the above-cited reports that implicate Psl in P. aeruginosa and Staphylococcus sp. interactions, we decided to investigate the potential roles of Psl in mediating interactions with S. aureus. During the course of our study, we found that the cell-free form of Psl could induce S. aureus lysis. This is the first study to investigate the role of Psl in killing S. aureus.

RESULTS

P. aeruginosa Psl production antagonizes the growth of S. aureus.

To investigate the impact of Psl on S. aureus growth, we first performed a cross-streak assay (22). S. aureus strain USA300 was inoculated horizontally on solidified medium, and the P. aeruginosa strain was streaked vertically, perpendicularly intersecting the S. aureus streak (Fig. 1A). When cocultured with wild-type P. aeruginosa PAO1 in this assay, an inhibition zone of S. aureus growth was observed in the region neighboring the intersection. However, when cocultured with an isogenic psl promoter deletion (ΔP_psl) mutant that does not produce Psl, this inhibition zone was reduced, and S. aureus was able to grow almost to the point of intersection with P. aeruginosa (Fig. 1A). This suggests that S. aureus survival is negatively affected by P. aeruginosa Psl production.

FIG 1.

P. aeruginosa Psl production antagonizes growth of S. aureus. (A) S. aureus USA300 was grown in coculture with either P. aeruginosa PAO1 or ΔP_psl on solidified medium using a cross-streak assay. Arrows point to the inhibition zones of S. aureus growth that were observed after overnight incubation. (B) S. aureus survival over the course of 8 h when grown alone or cocultured with PAO1 or ΔP_psl. Data are means and standard deviations (SD); individual points indicate the means for biological replicates (N = 5; n = 3). Significance was determined using Student’s t test. *, P < 0.05 compared to PAO1. (C) S. aureus survival when cocultured for 7 h with PAO1, ΔpslD, or the complemented ΔpslD/pslD⁺ strain. N = 4; n = 3. Data are means and SD; individual points indicate biological replicates. S. aureus survival (B and C) is presented as CFU normalized to the starting CFU at 0 h. Significance was determined using a one-way ANOVA. *, P < 0.05 compared to PAO1.

We therefore hypothesized that Psl may have a role in antagonizing S. aureus. To further investigate this, S. aureus was grown in planktonic coculture with either P. aeruginosa PAO1 or the ΔP_psl mutant (23). S. aureus survival was quantified by enumerating CFU. After 5 h, S. aureus levels decreased when cocultured with both PAO1 and the ΔP_psl mutant, compared to when grown alone. However, when cocultured with the ΔP_psl mutant, S. aureus had increased survival (10- to 20-fold) compared to coculture with PAO1 (Fig. 1B). Similarly, S. aureus had increased survival when cocultured with a P. aeruginosa ΔpslD mutant, compared to the coculture with PAO1 (Fig. 1C). Introducing the wild-type pslD allele in trans (ΔpslD/pslD⁺) restored the activity of antagonizing S. aureus to that of PAO1 in coculture (Fig. 1C). Together, these data indicate that Psl production antagonizes S. aureus. The ΔP_psl mutant retained some antagonism toward S. aureus, likely due to other known killing P. aeruginosa mechanisms (10–13).

Cell-free Psl antagonizes the growth of multiple S. aureus isolates.

P. aeruginosa produces both cell-associated and cell-free Psl (24). The mechanism underlying Psl localization are, however, unknown. Prior studies with three classes of monoclonal antibodies directed against Psl suggested that one of the most functionally active anti-Psl antibodies (Cam-003) failed to bind to synthetic Psl oligosaccharides (25, 26). The authors suggested that Cam-003 bound to a Psl isoform that was found in the native polymer but not in the synthetic forms. Furthermore, mild alkaline treatment during carbohydrate purification eliminated Cam-003 binding, suggesting that Psl may contain lipid moieties that were not identified in the reported Psl structure (24, 25). To investigate the potential for lipid modification of Psl, we performed a limited genetic screen of transposon mutants (27) in nonessential genes annotated to be involved in lipid A biogenesis, acyl-transferases, or acyl carrier proteins. The goal was to identify mutants with altered Psl localization, aiming to uncover mechanisms involved in Psl association with the cell. Using an immunoblot assay, we found that wild-type PAO1 had an even distribution of Psl between the two forms, while many of the mutants screened had increased ratios of cell-free Psl to cell-associated Psl (Fig. 2A; also, see Table S2 in the supplemental material). Interestingly, several mutants that impacted Psl distribution were related to lipid A synthesis and modification, warranting future investigations of lipid A in mediating the association of Psl with the cell outer membrane.

FIG 2.

Cell-free Psl derived from P. aeruginosa spent medium antagonizes S. aureus growth. (A) Cell-associated (associated) and cell-free (free) Psl from P. aeruginosa was quantified by immunoblotting. The cell pellets and spent media of centrifuged overnight cultures were used as the source of cell-associated and cell-free Psl, respectively. Psl was quantified by densitometry of the blot and normalized to PAO1 cell-associated Psl. Data are means and SD; individual points indicate biological replicates (N = 3; n = 3). Significance was determined with one-way ANOVA. *, P < 0.05 compared to PAO1. (B) S. aureus survival when cocultured for 7 h with P. aeruginosa PAO1, ΔP_psl, pagP::Tn, or pagP::TnΔpsl strains. S. aureus survival is presented as CFU normalized to the starting CFU at 0 h. Data are means and SD; individual points indicate biological replicates (N = 3; n = 3). Significance was determined using Student’s t test compared to PAO1. *, P < 0.05; ns, not significant. (C) S. aureus survival when grown for 5 h in spent medium of PAO1 or the ΔP_psl or pagP::Tn strain that was diluted 1:1 in fresh medium. Data are means and SD; individual points indicate means for biological replicates (N = 4; n = 3). Significance was determined using Student’s t test. *, P < 0.05 compared to PAO1. (D) S. aureus was grown for 4 h in spent medium from PAO1, ΔP_psl, or pagP::Tn strains that had been pretreated with (+) or without (−) PslG. Data are means and SD; individual points indicate biological replicates (N = 4; n = 3). S. aureus survival (B and C) is presented as CFU normalized to the starting CFU at 0 h. Significance was determined using Student’s t test, compared to the no-PslG pretreated group. *, P < 0.05; ns, not significant.

To determine if a specific form of Psl was responsible for mediating the antagonism toward S. aureus, we used one of the mutants identified from the screen, a pagP transposon mutant (pagP::Tn PA1343::ISlacZ/hah), as it produces predominately cell-free Psl (Fig. 2A; Table S2). PagP is a palmitoyltransferase which transfers palmitate from outer membrane phospholipids to lipid A (28). We grew S. aureus in planktonic coculture with either P. aeruginosa PAO1 or the ΔP_psl or the pagP::Tn strain (Fig. 2B). As previously observed (Fig. 1B), coculture with the ΔP_psl mutant led to a significant increase in S. aureus survival, compared to coculture with PAO1. In contrast, when grown in coculture with the pagP::Tn mutant, S. aureus survival was significantly reduced (more than 10-fold) compared to that in coculture with PAO1 (Fig. 2B). To determine if this increased antagonism is dependent on Psl, a pslBCD deletion was created in the pagP::Tn background (pagP::TnΔpsl). We found that S. aureus survival when cocultured with the pagP::TnΔpsl strain was comparable to that seen with the ΔP_psl mutant (Fig. 2B). This suggests that the enhanced antagonism toward S. aureus from the pagP::Tn strain is dependent on Psl production. We also observed the same trends when other S. aureus strains, including LAC and a methicillin-sensitive S. aureus (MSSA) strain, were cocultured with either P. aeruginosa PAO1, ΔP_psl, or pagP::Tn (Fig. S1). The above data suggest an important role for cell-free Psl in antagonizing the growth of S. aureus during planktonic coculture.

Since cell-free Psl is secreted and appears to be the form that mediates S. aureus antagonism, we tested the ability of P. aeruginosa spent medium to antagonize S. aureus growth. Similar to planktonic coculture (Fig. 2B), S. aureus had reduced survival when grown in spent medium from PAO1 and the pagP::Tn strain, compared to growth in ΔP_psl spent medium, the latter of which did not show any changes in S. aureus survival when quantified by both CFU (Fig. 2C) and optical density at 600 nm (OD₆₀₀) (Fig. S2). Furthermore, S. aureus survival was significantly reduced when grown in pagP::Tn spent medium, compared to PAO1 spent medium (Fig. 2C; Fig. S2).

PslG is a hydrolase that cleaves the polymeric chain of Psl into monomers (29). To further support the role of cell-free Psl in antagonizing S. aureus growth, P. aeruginosa spent medium was pretreated with PslG to degrade Psl (Fig. S3). S. aureus was supplemented with PslG treated spent media from PAO1 or the ΔP_psl or pagP::Tn strain and grown for a further 4 h. S. aureus survival significantly increased by 3- and 15-fold when grown with PAO1 and pagP::Tn PslG-treated spent medium, respectively, compared to spent medium without PslG (Fig. 2D). As expected, PslG treatment of ΔP_psl spent medium resulted in no change to S. aureus survival, indicating that PslG does not affect S. aureus growth (Fig. 2D).

Psl-mediated antagonism toward S. aureus is independent of other P. aeruginosa factors that impact S. aureus growth.

P. aeruginosa utilizes various mechanisms to antagonize the growth of S. aureus, including PQS (10), pyoverdine (12), LasA (11), and rhamnolipid (13). Alginate, another exopolysaccharide produced by P. aeruginosa, has been found to promote P. aeruginosa and S. aureus coexistence by downregulating some of the above mechanisms in P. aeruginosa (23, 30). We therefore wanted to examine if Psl-mediated antagonism toward S. aureus was dependent on any of these mechanisms. There were no significant differences found in rhamnolipid and LasA activity in PAO1 and the ΔP_psl mutant (Fig. S4A to C). A deficiency in PQS or pyoverdine did not affect the ability of P. aeruginosa cell-free Psl to antagonize S. aureus growth during coculture (Fig. S4D). As controls, corresponding mutants for each mechanism showed expected loss of these antagonistic factors (see the supplemental methods and Fig. S4A to D). Collectively, these data indicate that cell-free Psl antagonizes the growth of S. aureus independently of rhamnolipid, LasA, PQS, and pyoverdine.

Cell-free Psl kills S. aureus by disrupting the cell envelope.

Exopolysaccharides isolated from bacteria and fungi can disrupt bacterial cell envelopes (31–33). We therefore hypothesized that Psl similarly caused S. aureus cell lysis. To test this, S. aureus was incubated with diluted spent medium from P. aeruginosa PAO1, ΔP_psl, or pagP::Tn for 2 h and processed for transmission electron microscopy (TEM). This time point was selected as S. aureus was still viable after 2 h of incubation with P. aeruginosa spent medium (Fig. 2C). TEM imaging revealed that S. aureus incubated with either PAO1 or pagP::Tn spent medium had disrupted cell membranes, while S. aureus incubated with ΔP_psl spent medium showed little evidence of cellular damage, similar to the no-treatment control (Fig. 3A; Fig. S5). Image analysis, which quantified the percentage of cells with membrane damage relative to the untreated control, revealed that approximately 25% of S. aureus cells incubated with PAO1 spent medium had disrupted cell envelopes (Fig. 3B). However, when incubated with spent medium from the pagP::Tn strain, the number of S. aureus cells with membrane damage increased to approximately 50% (Fig. 3B). This is consistent with our previous results which showed decreases in S. aureus cell numbers (Fig. 2C) and OD₆₀₀ (Fig. S2) when the organism was grown in medium supplemented with P. aeruginosa spent medium, suggestive of cell lysis. The data indicate that cell-free Psl likely kills S. aureus by disrupting the cell envelope.

FIG 3.

Cell-free Psl kills S. aureus by disrupting the cell envelope. (A) S. aureus USA300 was incubated with spent medium from P. aeruginosa PAO1, ΔP_psl, or pagP::Tn for 2 h. Changes in cell morphology were visualized by TEM. Arrows indicate cells with disrupted cell envelopes. (B) Total and disrupted cell counts for each group were enumerated, and the percentage of disrupted cells within each group was calculated. Data are means and SD; individual points indicate biological replicates (N = 4; n = 3). Significance was determined using a one-way ANOVA compared to PAO1. *, P < 0.05.

Psl promotes S. aureus killing in conditions that mimic in vivo growth.

The above-described experiments were performed using planktonic cultures grown in a rich medium. We wanted to further determine if cell-free Psl could mediate S. aureus killing under in vivo-like conditions. Biofilm formation is one of the key factors contributing to P. aeruginosa and S. aureus chronic infections (34, 35). As such, we next examined if cell-free Psl could kill S. aureus grown in biofilms (Fig. 4A). Pre-established S. aureus biofilms were incubated for 5 h with spent medium from P. aeruginosa. Biofilm biomass was then quantified by crystal violet staining (36). Compared to the control group with no P. aeruginosa spent medium added, S. aureus biofilm biomass was significantly decreased when treated with both PAO1 and pagP::Tn spent media. No significant difference was found between the control and ΔP_psl spent media. This suggests that cell-free Psl also promotes killing of S. aureus in biofilms.

FIG 4.

Cell-free Psl promotes S. aureus killing under in vivo-like conditions. (A) Twenty-four-hour S. aureus USA300 biofilms were incubated with or without P. aeruginosa PAO1, ΔP_psl, or pagP::Tn spent medium for 5 h. Biofilm biomass was quantified by crystal violet staining. Data are means and SD; individual points indicate biological replicates (N = 5; n = 3). Significance was determined using one-way ANOVA. *, P < 0.05, and ns, not significant, compared to the medium-only control. (B) S. aureus survival when cocultured with P. aeruginosa PAO1, ΔP_psl, or pagP::Tn in SCFM2 for 7 h. Data are means and SD; individual points indicate biological replicates (N = 3; n = 3). Significance was determined using Student’s t test. *, P < 0.05 compared to PAO1.

Synthetic CF sputum medium (SCFM2) mimics the CF sputum composition and has been used to culture both P. aeruginosa and S. aureus (23, 37). We quantified S. aureus survival when cocultured with P. aeruginosa PAO1, ΔP_psl, or pagP::Tn in SCFM2 for 8 h (Fig. 4B). Like our previous results in rich media (Fig. 2B), S. aureus survival significantly increased by 10-fold when cocultured with the ΔP_psl mutant, compared to coculture with PAO1 (Fig. 4B). Furthermore, S. aureus survival significantly decreased by 5-fold in coculture with the pagP::Tn mutant compared to coculture with PAO1 (Fig. 4B). This suggests that cell-free Psl may contribute to S. aureus killing in nutritional environments like a CF lung. Together, these data suggest that Psl-mediated killing of S. aureus occurs in conditions that are reflective of in vivo scenarios.

Increased cell-free Psl production in P. aeruginosa clinical isolates promotes S. aureus killing.

Some P. aeruginosa clinical isolates lack or have reduced Psl production (15–17). Given that cell-free Psl appears to kill S. aureus, it is possible that reduced Psl production in P. aeruginosa occurs during coevolution with S. aureus, facilitating polymicrobial infections. To test if there is a correlation between Psl production and S. aureus killing among P. aeruginosa clinical isolates, 16 P. aeruginosa isolates derived from CF sputum samples were collected from the Cure CF Columbus Translational Core at Nationwide Children’s Hospital (Fig. 5). Since mucoid P. aeruginosa can coexist with S. aureus, due to the overproduction of alginate (23), only nonmucoid P. aeruginosa isolates were evaluated. Psl production of each nonmucoid P. aeruginosa isolate was quantified by immunoblot assay. S. aureus USA300 was grown in planktonic coculture with each CF isolate, as well as PAO1 or the ΔP_psl or pagP::Tn mutant, and S. aureus survival was quantified. A correlation (R² = 0.3578) was found between the cell-free Psl production of the P. aeruginosa strains and the cocultured S. aureus survival (Fig. 5A). Furthermore, 6 of the 16 P. aeruginosa CF isolates produced no Psl, and each had reduced S. aureus killing activity compared to PAO1 (Fig. 5B). This suggests that low Psl production in P. aeruginosa clinical isolates can result in enhanced survival when in coculture with S. aureus and that Psl production may be one factor involved in their coexistence during infections.

FIG 5.

Psl production in P. aeruginosa clinical isolates positively correlates with the ability to kill S. aureus. S. aureus USA300 was cocultured with each of the 16 P. aeruginosa CF isolates, PAO1, and the ΔP_psl and pagP::Tn strains. The number of CFU at 7 h was divided by that at 0 h to quantify survival. (A) Linear regression analysis was performed to determine any correlation between the two parameters indicated. Cell-free Psl production by the designated strains was measured by immunoblot assay using Psl antibody and then normalized to PAO1. Individual points indicate biological replicates (N = 3; n = 3). (B) P. aeruginosa clinical isolates with low Psl production showed reduced S. aureus killing activity. Six of the CF isolates produced no cell-free Psl (−), while the other 10 produced variable amounts of cell-free Psl (+; the arrow indicates the increasing production of Psl), compared with that produced by PAO1. The number of CFU of S. aureus USA300 cocultured with each of the isolates, PAO1, and the ΔP_psl and pagP::Tn strains for 7 h was divided by that at 0 h to quantify S. aureus survival. Data are means and SD (N = 3; n = 3).

Purified cell-free Psl is sufficient to kill S. aureus.

Exopolysaccharides produced by some bacteria and fungi have antimicrobial activities (31–33). Our results demonstrate that cell-free Psl mediates the killing of S. aureus. However, this was determined from planktonic coculture of P. aeruginosa and S. aureus (Fig. 1 and 2A and B) or growth of S. aureus in P. aeruginosa spent medium (Fig. 2C and D), both conditions which could contain Psl-independent antagonists of S. aureus growth. We therefore tested if cell-free Psl purified from PAO1 and the pagP::Tn mutant could directly inhibit S. aureus growth. S. aureus survival was quantified by measuring the OD₆₀₀ after 16 h of growth in medium supplemented with increasing concentrations of purified cell-free Psl (0 to 100 μg/mL). S. aureus grown alone, without the addition of cell-free Psl (untreated), was used as the control. S. aureus survival decreased as the concentration of purified Psl increased (Fig. 6A). The ΔP_psl strain also underwent the identical purification process, and the resulted extraction had no effect on S. aureus survival. In addition, we found that purified Psl originating from ΔpqsA and ΔpvdA mutants showed activity similar to that of Psl from the parent strains in killing S. aureus (Fig. S4E). This demonstrates that purified cell-free Psl, in a concentration-dependent manner, is sufficient to mediate the killing of S. aureus.

FIG 6.

Purified cell-free Psl is sufficient to kill S. aureus. Survival of S. aureus grown in media supplemented with (A) increasing concentrations of purified cell-free Psl from PAO1 or the pagP::Tn mutant or cell-free products from the ΔP_psl mutant or (B) 60 μg/mL of d-mannose (Man), 20 μg/mL of d-glucose (Glu), 20 μg/mL of l-rhamnose (Rha), or a combination of the three monosaccharides (Combined). S. aureus grown without cell-free Psl or monosaccharides (untreated) was used as a control. OD₆₀₀ was measured after 16 h. S. aureus survival is presented as OD₆₀₀ normalized to that of the control. (A) The x axis is a log scale. Data are means and SD; individual points indicate the mean of biological replicates (N = 3; n = 3). Significance was determined using Student’s t test compared to the ΔP_psl mutant. *, P < 0.05.

Cell-free Psl consists of d-mannose, d-glucose, and l-rhamnose, at a ratio of 3:1:1, respectively (24). The influence of these individual monosaccharides on S. aureus survival was also evaluated. There was no change in survival of S. aureus grown in medium supplemented with either each monosaccharide separately or a combination (Fig. 6B). This indicates that an integral structure of Psl is critical for the S. aureus killing property.

DISCUSSION

P. aeruginosa and S. aureus often cause chronic coinfections that can lead to more severe disease outcomes (7–9). Understanding the dynamic between the two pathogens is key to successfully treating and eliminating such coinfections. Here, we demonstrate a novel role for P. aeruginosa exopolysaccharide Psl in killing S. aureus in vitro (Fig. 1). Specifically, we found that purified cell-free Psl mediated S. aureus killing (Fig. 6), likely by disrupting the cell envelope and causing cell lysis (Fig. 3). This antagonistic relationship would likely impact the coexistence of these two opportunistic pathogens in vivo.

Furthermore, we demonstrate that cell-free Psl can kill S. aureus in mature biofilms (Fig. 4A), suggesting that the antimicrobial properties of Psl are not limited to highly metabolically active, planktonic S. aureus. In support of this, P. aeruginosa extracellular Psl from spent medium can disrupt established S. epidermidis biofilms. Interestingly, this was independent of any bactericidal effect, as spent media of wild-type P. aeruginosa and a psl mutant were able to comparably inhibit the growth of planktonic S. epidermidis (20). This suggests that cell-free Psl may have different killing mechanisms for S. aureus and S. epidermidis. It is also possible that Psl-mediated killing of planktonic S. aureus and disruption of biofilms occur through distinct mechanisms. Conversely, Staphylococcus protein A (SpA) secreted by S. aureus can inhibit P. aeruginosa biofilm formation, and this can be prevented by Psl binding to SpA (19). However, we found that a spa transposon mutant (spa::Tn) was as susceptible as wild-type S. aureus to killing by PAO1 and pagP::Tn when grown in planktonic coculture (Fig. S6). This suggests that SpA does not play a role in the cell-free Psl-mediated killing of S. aureus described here.

In addition to Psl, P. aeruginosa produces two other exopolysaccharides, alginate and Pel (38, 39). In contrast to the antimicrobial properties of Psl toward S. aureus observed here, alginate mediates P. aeruginosa and S. aureus coexistence (23, 30). Alginate overproduction, resulting in a mucoid phenotype, downregulates known antagonistic mechanisms in P. aeruginosa, including rhamnolipid, pyoverdine, and HQNO (2-heptyl-4-hydroxyquinoline N-oxide) (23). Interestingly, alginate and Psl share the same precursor, mannose-1-phosphate (40), leading to reduced Psl production when alginate is overproduced (41, 42). This is in accord with our finding that Psl plays the opposite role of alginate, by promoting S. aureus killing. This study focused exclusively on S. aureus antagonism mediated by nonmucoid P. aeruginosa. Prior studies showed that some nonmucoid clinical isolates harbor secondary mutations in algTU (43, 44), and we did not determine if isolates used in this study indeed harbored these mutations. Jones et al. (42), on the other hand, found that functional Psl, albeit at reduced levels, is still produced in some mucoid P. aeruginosa isolates, warranting further investigation of S. aureus killing by P. aeruginosa producing both Psl and alginate. Pel also contributes to P. aeruginosa and S. aureus polymicrobial interactions. It can reduce cross-linking within the biofilm matrix, enabling the expansion of P. aeruginosa in a dual-species biofilm with S. aureus (21). Secreted Pel can also remove established S. epidermidis biofilms (20). However, our data suggest that Pel is not directly involved in killing S. aureus (Fig. S7).

Our data also suggest that the structure of Psl may affect its antimicrobial activity. PslG-treated PAO1 and pagP::Tn spent media had no detectable Psl, similar to the ΔP_psl mutant (Fig. S3). However, while they both showed reduced S. aureus killing compared to untreated spent medium (Fig. 2D), S. aureus survival was still significantly reduced compared to that with ΔP_psl spent medium (Fig. 2D). This suggests that PslG-treated PAO1 and pagP::Tn spent media demonstrate intermediate S. aureus antimicrobial properties. It is possible that a higher-molecular-weight (MW) structure is needed for the antimicrobial property of cell-free Psl. One interpretation of these data is that PslG may partially disrupt Psl polymers such that it cannot be recognized by Psl antibodies, but these forms can have reduced, although some, S. aureus killing activity. Consistent with this, Maalej et al. found that a high-molecular-weight polysaccharide produced by Pseudomonas stutzeri showed antimicrobial activity toward S. aureus and other bacteria (31). This polysaccharide consists of a trisaccharide backbone chain containing glucose, mannose, and lactyl rhamnose, similar to P. aeruginosa Psl. It is likely that these two polysaccharides, both isolated from Pseudomonas species, share similar properties in killing S. aureus. Maalej et al. also compared antimicrobial abilities of the P. stutzeri polysaccharide in different MWs and found that higher-MW polysaccharides had increased Gram-positive bacterium-killing activity. In addition, we found that the monosaccharides making up cell-free Psl showed no killing of S. aureus (Fig. 6B). Together, these observations suggest that the structure of cell-free Psl may contribute to its antimicrobial activity.

From our TEM analysis, we propose that cell-free Psl damages S. aureus cell envelopes, causing cell lysis (Fig. 5). Our observations contribute to the growing evidence that exopolysaccharides produced by bacteria and fungi can have antimicrobial properties by disrupting cell envelopes (31–33). A polysaccharide isolated from Streptomyces virginiae, consisting of mannose, glucose, and galactose, also inhibits S. aureus growth (33). Interestingly, this polysaccharide has a sugar composition similar to that of Psl and a bactericidal concentration comparable to what we determined for cell-free Psl (Fig. 6). As mentioned above, the polysaccharide produced by P. stutzeri is antimicrobial toward S. aureus (31). It was speculated that the amphiphilicity of the polymer may provide structural affinity between the S. aureus cell wall and polysaccharide, presumably leading to lysis. Metal chelation was suggested as another potential mechanism of P. stutzeri polysaccharide permeabilization of the S. aureus cell envelope (31). This might also apply to the Psl-mediated killing of S. aureus, as Psl can sequester and store iron (45). Despite determining that cell-free Psl causes S. aureus cell lysis, we have yet to elucidate the mechanism. However, we observed no difference in PAO1 and ΔP_psl spent media in lysing heat-killed S. aureus (Fig. S4C). This suggests that killing requires active S. aureus growth/metabolism. Future investigations will focus on the mechanism(s) involved in Psl-mediated killing.

It is also possible that Psl can antagonize S. aureus by other mechanisms, in addition to cell lysis. The loss of cell-associated Psl leads to reduced aggregation (46). Cell aggregation promotes P. aeruginosa pyoverdine (47) and quorum sensing signal production, including PQS (48), both of which are involved in killing S. aureus. In addition, purified cell-associated Psl can act as a signal to stimulate cyclic di-GMP production in P. aeruginosa (49). It is therefore possible that Psl could affect the production of known S. aureus-killing factors by promoting cell aggregation and/or acting as a signal molecule in P. aeruginosa. However, this is not supported by our finding that purified Psl is sufficient to promote S. aureus killing and that pqs and pvdA mutants still showed Psl-dependent killing.

Psl is important for P. aeruginosa biofilm formation and contributes to chronic infections (18). However, several P. aeruginosa clinical isolates produce little or no Psl (15–17). Consistent with this, we found that 6 of the 16 examined CF isolates produced no Psl. A positive correlation was observed between Psl production and S. aureus killing activity among the isolates (Fig. 5), suggesting a role for Psl in affecting the coexistence of the two pathogens in CF lungs. Coexistence with S. aureus can be beneficial to P. aeruginosa. For example, S. aureus is predominant in early dual-species biofilms and can promote P. aeruginosa attachment and biofilm formation (50). We speculate that P. aeruginosa may undergo adaptation to reduce its antagonistic weapons, like Psl, to better coexist with S. aureus in vivo.

Taken together, our findings demonstrate that cell-free Psl produced by P. aeruginosa can disrupt the S. aureus cell envelope, leading to S. aureus cell death. Psl-mediated killing also occurs under in vivo-like conditions, including in SCFM2 and S. aureus biofilms. Consistent with this, the amount of Psl produced in P. aeruginosa clinical isolates positively correlates with their antistaphylococcal activity. This study demonstrates a novel role for exopolysaccharides in polymicrobial infections and highlights the need for future investigations in polymicrobial coevolution during chronic infections.

MATERIALS AND METHODS

Bacterial strains, primers, and growth conditions.

All bacterial strains and plasmids are listed in Table 1. All primers used in this study are listed in Table S1. Gene deletion constructs were incorporated into the P. aeruginosa genome using homologous recombination as previously described (51). S. aureus was grown in lysogeny broth (LB; 10 g/L tryptone, 5 g/L yeast extract, 10 g/L NaCl). P. aeruginosa was grown in LB with no salt (LBNS). Overnight cultures were grown at 37°C with 200 rpm shaking for 16 h. Antibiotic concentrations were 10 μg/mL gentamicin and 100 μg/mL ampicillin for selection of Escherichia coli and 300 μg/mL carbenicillin for P. aeruginosa.

TABLE 1.

Strains and plasmids

Strain or plasmid	Description	Source or reference
E. coli
NEB5α		New England Biolabs
S17
P. aeruginosa
PAO1	WT P. aeruginosa	55
ΔPpsl mutant	psl production deficient; psl operon promoter deletion mutant	55
ΔpslD/V mutant	psl production deficient; ΔpslD with an empty vector pUCP18	24
ΔpslD/pslD+ mutant	pUCP18::pslD was used to complement ΔpslD	24
PAO1/V	WT P. aeruginosa with an empty vector pUCP18	24
pagP::Tn mutant	pagP transposon mutant (UWGC:PW3442, PA1343::ISlacZ/hah)	27
htrB1::Tn mutant	htrB1 transposon mutant (UWGC:PW1009, PA0011::ISlacZ/hah)	27
htrB2::Tn mutant	htrB2 transposon mutant (UWGC:PW6435, PA3242::ISlacZ/hah)	27
lpxO1::Tn mutant	lpxO1 transposon mutant (UWGC:PW8596, PA4512::ISphoA/hah)	27
lpxO2::Tn mutant	lpxO2 transposon mutant (UWGC:PW2702, PA0936::ISphoA/hah)	27
pagL::Tn mutant	pagL transposon mutant (UWGC:PW8859, PA4661::ISphoA/hah)	27
phoP::Tn mutant	phoP transposon mutant (UWGC:PW3128, PA1179::ISlacZ/hah)	27
phoQ::Tn mutant	phoQ transposon mutant (UWGC:PW3131, PA1180::ISlacZ/hah)	27
pmrB::Tn mutant	pmrB transposon mutant (UWGC:PW9023, PA4777::ISlacZ/hah)	27
pagP::TnΔpsl mutant	pagP transposon mutant with pslBCD deletion	This study
ΔpvdA mutant	pvdA deletion mutant	This study
ΔpqsA mutant	pqsA deletion mutant	This study
ΔpvdA ΔPpsl mutant	psl promoter and pvdA double-deletion strain	This study
ΔpqsA ΔPpsl mutant	psl promoter and pqsA double-deletion strain	This study
rhlA::Tn mutant	rhlA transposon mutant (UWGC:PW6886, PA3479::ISphoA/hah)	27
lasA::Tn mutant	lasA transposon mutant (UWGC:PW4282, PA1871::ISlacZ/hah)	27
ΔpelA mutant	pelA deletion mutant	56
0CM2DB	CF clinical isolate	This study
0CMJ92	CF clinical isolate	This study
0CPQ3U	CF clinical isolate	This study
0CO5ON	CF clinical isolate	This study
0CMJ8Y	CF clinical isolate	This study
0CM9TX	CF clinical isolate	This study
0CHBKC	CF clinical isolate	This study
0CHBKD	CF clinical isolate	This study
0CH5M4	CF clinical isolate	This study
0CLLNJ	CF clinical isolate, coisolated with S. aureus	This study
0CLEJ5	CF clinical isolate, coisolated with S. aureus	This study
0CH7HJ	CF clinical isolate, coisolated with S. aureus	This study
0CHGWG	CF clinical isolate, coisolated with S. aureus	This study
0CGU6K	CF clinical isolate, coisolated with S. aureus	This study
0CI6XZ	CF clinical isolate, coisolated with S. aureus	This study
0CHOI3	CF clinical isolate, coisolated with S. aureus	This study
S. aureus
USA300	WT	57
MSSA	ATCC 29213, methicillin sensitive	ATCC
LAC	MRSA	58
spa::Tn mutant	spa transposon mutant (NE286, NARSA)	57
Plasmids
pEX18Gm	For allelic exchange in P. aeruginosa
pΔpqsA	For pqsA deletion	This study
pΔpvdA	For pvdA deletion	This study
pMpsl-KO1	For pslBCD deletion	59

P. aeruginosa and S. aureus coculture.

Cross-streak and planktonic cocultures were performed as previously described with modifications (23). For the cross-streak coculture, overnight cultures of P. aeruginosa strains and S. aureus USA300 were normalized to an OD₆₀₀ of 1.5. A loopful of S. aureus culture was streaked horizontally on lysogeny agar with no salt (LANS) followed by a vertical streak of P. aeruginosa. The cross-streaked plate was incubated overnight at 37°C. For the planktonic coculture, overnight cultures of P. aeruginosa and S. aureus were diluted to an OD₆₀₀ of 0.05 and combined at a ratio of 1:1 in 2 mL of either fresh LBNS or SCFM2 (52). The coculture was incubated at 37°C with shaking at 200 rpm for up to 8 h. Aliquots were taken at 0, 3, 5, 6, 7, and 8 h, serially diluted to plate on Difco Pseudomonas isolation agar (PIA) and BBL mannitol salt agar (MSA) to enumerate CFU of P. aeruginosa and S. aureus, respectively. S. aureus survival at each time point was normalized to the number of CFU at 0 h.

Crude extraction of cell-free and cell-associated exopolysaccharide.

P. aeruginosa strains were grown overnight at 37°C and 200 rpm in 5 mL of LBNS and normalized to an OD₆₀₀ of 2. Crude exopolysaccharide was isolated using a modified protocol (53). Briefly, the bacteria were centrifuged at 8,000 × g for 10 min at room temperature to harvest the spent medium and pellet. For cell-free exopolysaccharide, the spent medium after centrifugation was removed and passed through a 0.22-μm filter to remove any residual bacteria. For cell-associated exopolysaccharide, the bacterial pellet was suspended in 2 mL of 1.5 M NaCl, vortexed vigorously, and placed on a platform rocker at room temperature for 15 min. After extraction, the sample was centrifuged at 5,000 × g for 10 min at room temperature. The supernatant was removed and passed through a 0.22-μm filter to remove any residual bacteria.

Psl immunoblot assay.

Two microliters of the crude exopolysaccharide extracts and Psl standards (500 μg/mL, 250 μg/mL, 100 μg/mL, 50 μg/mL, and 10 μg/mL) were spotted onto a nitrocellulose membrane (0.2 μm; Bio-Rad) and incubated at room temperature with 5% skim milk for 1 h, Psl antibody (AstraZeneca; 1:3,000 dilution) for 1 h, and secondary antibody (goat anti-human IgG; 1:5,000 dilution; Abcam) for 1 h sequentially. Western blotting detection reagent (ECL) was added to the membrane for visualization via the ChemiDoc imaging system (Bio-Rad). Densitometry analysis was performed with ImageJ software, and results were compared to a standard curve of pure Psl.

S. aureus killing assay using P. aeruginosa spent medium.

Overnight cultures of P. aeruginosa and S. aureus were diluted to an OD₆₀₀ of 0.05 and incubated at 37°C with shaking at 200 rpm for 7 h and 4 h, respectively. Psl-mediated antagonism was observed when S. aureus was grown in coculture with P. aeruginosa for 7 h, so we selected this time point to harvest P. aeruginosa spent medium. A P. aeruginosa culture was filter sterilized using a 0.22-μm filter (Fisherbrand) and diluted 1:1 in fresh LBNS. Four milliliters of stationary-phase S. aureus (OD₆₀₀ = 1.5) was centrifuged, resuspended in 4 mL of diluted P. aeruginosa spent medium, and incubated at 37°C with shaking at 200 rpm for up to 4 h. Aliquots were taken every hour, and either absorbance (OD₆₀₀) was measured using a spectrophotometer (Thermo Fisher), or the sample was serially diluted and CFUs were enumerated. For experiments with PslG-treated spent media, filter-sterilized P. aeruginosa spent medium was treated with 100 μM PslG for 1 h at 37°C to degrade cell-free Psl, prior to the addition of S. aureus. S. aureus survival at each time point was normalized to the CFU or OD₆₀₀ at 0 h.

Cell-free Psl purification.

Psl was isolated by following an established protocol (24, 49) with modifications. Briefly, concentrated spent medium of P. aeruginosa was subjected to 3 rounds of ethanol precipitation, treatment with 0.1 mg/mL of DNase I, RNase A, and proteinase K, and 3 rounds of dialysis (Slide-A-Lyzer dialysis cassette; molecular weight cutoff [MWCO] of 3,500). Concentration and carbohydrate content of the resulting purified cell-free Psl were determined by immunoblotting and phenol-sulfuric acid assay (54), respectively.

S. aureus killing assay by purified Psl.

Overnight S. aureus culture was diluted to an OD₆₀₀ of 0.05 and incubated at 37°C with shaking at 200 rpm for 4 h to reach an OD₆₀₀ of 1.5. The culture was resuspended in fresh LBNS containing increasing concentrations (0, 2.5, 5, 10, 20, 40, 80, and 100 μg/mL) of purified cell-free Psl, and 200 μL was added into each well of a 96-well plate. To test if the monosaccharides of Psl contributed to S. aureus killing, 20 μg/mL l-rhamnose, 20 μg/mL d-glucose, 60 μg/mL d-mannose, or a combination of the three monosaccharides was added to S. aureus. Plates were incubated in a plate reader (SpectraMax i3x; Molecular Devices) at 37°C for 16 h. OD₆₀₀ was measured every 30 min as an indication of survival and compared to that of the control (S. aureus grown alone). Relative S. aureus survival was measured by OD₆₀₀ normalized to the control.

TEM imaging.

S. aureus cultures were treated with P. aeruginosa spent medium for 2 h, centrifuged, and washed with phosphate-buffered saline (PBS). Samples were fixed in 2.5% glutaraldehyde overnight and then embedded in 2% agarose for TEM processing. Samples were postfixed with 1% osmium tetroxide, stained en bloc with 1% aqueous uranyl acetate, dehydrated in a graded series of ethanol, and embedded in Eponate 12 epoxy resin (Ted Pella Inc.; 18012). Ultrathin sections were cut with a Leica EM UC6 ultramicrotome (Leica Microsystems; EM FC7) and collected on copper grids. Images were acquired with an FEI Technai G2 Spirit transmission electron microscope (FEI), a Macrofire (Optronics) digital camera, and AMT image capture software. Total and disrupted cell counts were manually enumerated. All intact and disrupted cells are individually labeled in Fig. S5.

Biofilm killing assay.

An overnight culture of S. aureus was diluted in fresh LB to an OD₆₀₀ of 0.01. Two hundred microliters was aliquoted into each well of a 96-well plate and incubated at 37°C statically for 24 h. Biomass was washed with PBS 3 times, and the remaining attached biofilm was treated with diluted P. aeruginosa spent medium for 5 h. After the removal of planktonic cells and a PBS wash, the remaining biofilm was stained with 200 μL of 0.1% crystal violet for 15 min at room temperature. Biofilm was then washed with PBS 3 times, and crystal violet was extracted with 200 μL of 100% ethanol for 15 min at room temperature. OD₅₉₀ was measured by a plate reader as an indication of biofilm mass.

Statistical analysis.

Statistical significance was determined using either a one-way analysis of variance (ANOVA) with Tukey’s post hoc test or Student’s t test. Analyses were performed using GraphPad Prism v.7 (GraphPad Software). Statistical significance was determined using a P value of <0.05.