Pseudomonas aeruginosa Alters Staphylococcus aureus Sensitivity to Vancomycin in a Biofilm Model of Cystic Fibrosis Infection.

The airways of cystic fibrosis (CF) patients have thick mucus, which fosters chronic, polymicrobial infections. Pseudomonas aeruginosa and Staphylococcus aureus are two of the most prevalent respiratory pathogens in CF patients. In this study, we tested whether P. aeruginosa influences the susceptibility of S. aureus to frontline antibiotics used to treat CF lung infections. Using our in vitro coculture model, we observed that addition of P. aeruginosa supernatants to S. aureus biofilms grown either on epithelial cells or on plastic significantly decreased the susceptibility of S. aureus to vancomycin. Mutant analyses showed that 2-n-heptyl-4-hydroxyquinoline N-oxide (HQNO), a component of the P. aeruginosa Pseudomonas quinolone signal (PQS) system, protects S. aureus from the antimicrobial activity of vancomycin. Similarly, the siderophores pyoverdine and pyochelin also contribute to the ability of P. aeruginosa to protect S. aureus from vancomycin, as did growth under anoxia. Under our experimental conditions, HQNO, P. aeruginosa supernatant, and growth under anoxia decreased S. aureus growth, likely explaining why this cell wall-targeting antibiotic is less effective. P. aeruginosa supernatant did not confer additional protection to slow-growing S. aureus small colony variants. Importantly, P. aeruginosa supernatant protects S. aureus from other inhibitors of cell wall synthesis as well as protein synthesis-targeting antibiotics in an HQNO- and siderophore-dependent manner. We propose a model whereby P. aeruginosa causes S. aureus to shift to fermentative growth when these organisms are grown in coculture, leading to reduction in S. aureus growth and decreased susceptibility to antibiotics targeting cell wall and protein synthesis.IMPORTANCE Cystic fibrosis (CF) lung infections are chronic and difficult to eradicate. Pseudomonas aeruginosa and Staphylococcus aureus are two of the most prevalent respiratory pathogens in CF patients and are associated with poor patient outcomes. Both organisms adopt a biofilm mode of growth, which contributes to high tolerance to antibiotic treatment and the recalcitrant nature of these infections. Here, we show that P. aeruginosa exoproducts decrease the sensitivity of S. aureus biofilm and planktonic populations to vancomycin, a frontline antibiotic used to treat methicillin-resistant S. aureus in CF patients. P. aeruginosa also protects S. aureus from other cell wall-active antibiotics as well as various classes of protein synthesis inhibitors. Thus, interspecies interactions can have dramatic and unexpected consequences on antibiotic sensitivity. This study underscores the potential impact of interspecies interactions on antibiotic efficacy in the context of complex, polymicrobial infections.

INTRODUCTION

Cystic fibrosis (CF) is caused by mutations in the cystic fibrosis transmembrane conductance regulator (CFTR) gene. Although CF is a systemic disease, long-term lung infections are primarily responsible for poor patient outcomes (1). Despite routine administration of antibiotics, these infections are often highly resilient and resistant to treatment (2–4). Culture-independent studies have revealed that infections in the airways of CF patients are polymicrobial and complex (2–8); nonetheless, Pseudomonas aeruginosa and Staphylococcus aureus remain two of the most prevalent respiratory pathogens detected in CF patients (9). S. aureus is the most prevalent pathogen in younger patients with CF, while P. aeruginosa is highly prevalent in adult patients (9, 10). The presence of both P. aeruginosa and S. aureus is associated with decreased lung function, as measured by forced expiratory volume in 1 s (FEV1), and poor patient outcomes (11–14).

Interactions between P. aeruginosa and S. aureus have been the focus of several studies. Notably, it has been shown that P. aeruginosa negatively impacts S. aureus by producing 2-n-heptyl-4-hydroxyquinoline N-oxide (HQNO), an inhibitor of the electron transport chain (ETC) of S. aureus. HQNO can cause an increase in S. aureus biofilm formation at least under one condition (15, 16), and prolonged exposure to this compound leads to selection of S. aureus small-colony variants (SCVs) (17). Previously, HQNO has been shown to decrease the sensitivity of S. aureus to various aminoglycoside antibiotics, including streptomycin, dihydrostreptomycin, and tobramycin (17, 18). Additionally, HQNO and P. aeruginosa-produced siderophores have been shown to shift S. aureus to a fermentative mode of growth, eventually leading to reduced S. aureus viability (19–22). Finally, both P. aeruginosa and S. aureus form biofilms, which dramatically alters the expected antibiotic tolerance profiles for these organisms (23). P. aeruginosa forms biofilms within the CF lung, enabling persistence and recalcitrance to treatment (24–27).

In this study, we explored the effects of a dual-species interaction on the antibiotic tolerance of one microbial species in the context of CF lung infections involving bacterial biofilms. Specifically, we tested whether P. aeruginosa influences the susceptibility of S. aureus to vancomycin, a frontline antibiotic used to treat methicillin-resistant S. aureus (MRSA) in CF patients; approximately 25% of CF patients are culture positive for MRSA (9). We discovered that P. aeruginosa decreases the susceptibility of S. aureus biofilms to vancomycin, as well as to other cell wall synthesis inhibitors and protein synthesis inhibitors. We propose a model whereby P. aeruginosa exoproducts cause S. aureus to undergo a metabolic shift, leading to reduced growth and decreased susceptibility to a range of clinically relevant antibiotics.

RESULTS

P. aeruginosa supernatant protects S. aureus from vancomycin on plastic.

Previous work from our lab found that S. aureus 8325-4 downregulates penicillin-binding protein 4 (Pbp4) in the presence of P. aeruginosa when grown on plastic (22). Pbp4 has transpeptidase and carboxypeptidase activities and catalyzes the final step in peptidoglycan synthesis (28). Loss of the pbp4 gene results in increased tolerance to vancomycin (29) and, conversely, decreased tolerance to β-lactam antibiotics (30). Thus, we hypothesized that exposure of S. aureus to P. aeruginosa might alter the susceptibility of S. aureus to vancomycin.

To test this hypothesis, we selected a methicillin-sensitive S. aureus strain (Newman), an MRSA strain (USA300), and P. aeruginosa PA14 for our initial experiments. We previously showed that in our coculture system, S. aureus biofilm cell viability dramatically decreased after 10 to 16 h of coincubation with P. aeruginosa when these microbes are cocultured on CF-derived bronchial epithelial (CFBE) cells or plastic (22). In contrast, exposure of S. aureus to P. aeruginosa culture supernatant for 24 h did not alter S. aureus biofilm cell viability on plastic (Fig. 1A, left 2 bars). Therefore, to avoid P. aeruginosa-mediated, late-stage killing of S. aureus, we used P. aeruginosa supernatant to examine whether S. aureus vancomycin sensitivity is altered by the presence of P. aeruginosa-secreted products. In this assay, we first allowed S. aureus cells to attach for 1 h, and then the planktonic cells were removed and the attached cells washed with fresh medium and incubated for an additional 5 h to allow biofilm formation. Afterwards, the biofilm fraction was exposed to P. aeruginosa supernatant and/or vancomycin for 24 h. We refer to this method throughout as the “biofilm disruption assay,” because we first allow the biofilm to form and then assess the impact of different treatments on disrupting biofilm cell viability. We feel that such an assay more accurately reflects an infection-like condition wherein the microbial community is already established prior to the application of any treatment.

FIG 1 

P. aeruginosa protects S. aureus biofilm and planktonic populations from vancomycin. (A to D) Biofilm disruption assays on plastic were performed with S. aureus (Sa) Newman (A) and USA300 (B), P. aeruginosa PA14 supernatant (Pa sup), and vancomycin at 50 μg/ml. S. aureus biofilm (A and B) and planktonic (C and D) CFU were determined. Data in panels A and C and B and D were from the same experiments. Each column displays the average from at least three biological replicates, each with three technical replicates. Error bars indicate standard deviation (SD). **, P < 0.01, and ***, P < 0.001, by ordinary one-way analysis of variance (ANOVA) and Tukey’s multiple comparisons posttest. bd, below detection.

Incubation of S. aureus Newman with P. aeruginosa supernatant on plastic significantly decreased the sensitivity of S. aureus Newman biofilms to vancomycin, resulting in a 2-log increase in cell viability of S. aureus (Fig. 1A). Furthermore, the wild-type strain of S. aureus Newman and its Δpbp4 mutant derivative exhibited the same sensitivity to vancomycin in the absence and presence of P. aeruginosa supernatant, indicating that this observed decrease in sensitivity was not due to the absence of Pbp4 (Fig. 1A; see Fig. S1 in the supplemental material). Furthermore, P. aeruginosa supernatant protects S. aureus USA300 biofilms from vancomycin to a similar extent as S. aureus Newman (Fig. 1B).

The absence of Pbp4 is not responsible for decreased S. aureus sensitivity to vancomycin. Biofilm disruption assays on plastic were performed with the S. aureus (Sa) Newman pbp4 deletion mutant (Δpbp4), P. aeruginosa PA14 supernatant (Pa sup), and vancomycin at 50 μg/ml. Each column displays the average from at least three biological replicates, each with three technical replicates. Error bars indicate SD. ns, not significant; *, P < 0.05 by ordinary one-way ANOVA and Tukey’s multiple comparisons posttest. Download FIG S1, TIF file, 2.6 MB.

We also observed that P. aeruginosa supernatant protects S. aureus Newman and USA300 planktonic populations from vancomycin to levels of viability not significantly different from those of the untreated control, despite the high concentration of vancomycin used in the experiment (50 μg/ml [Fig. 1C and D]). Treatment of S. aureus Newman or USA300 planktonic populations with 50 μg/ml of vancomycin in the absence of P. aeruginosa supernatant resulted in a reduction of planktonic cell viability to below the detection level of this assay (~200 CFU/ml). Importantly, as was observed for the biofilm-grown bacteria, P. aeruginosa supernatant did not impact the cell viability of planktonic S. aureus Newman or USA300 in the absence of vancomycin (Fig. 1C and D). Furthermore, a comparison of the data from biofilm-grown and planktonic cells highlights the biofilm antibiotic tolerance of both strains versus vancomycin (Fig. 1, compare panels A and C and panels B and D).

We tested whether P. aeruginosa supernatant inactivates vancomycin and thus renders the drug ineffective against S. aureus, thereby explaining the observed increase in biofilm tolerance in the presence of P. aeruginosa supernatant. To test this possibility, we performed a minimum bactericidal concentration (MBC) assay in which we preincubated vancomycin with either minimal essential medium supplemented with 2 mM l-glutamine (MEM+l-Gln) or P. aeruginosa supernatant (prepared in MEM+l-Gln) for 24 h. The MBC of vancomycin was then determined for S. aureus and Streptococcus sanguinis. For S. aureus, the MBC of vancomycin that was preincubated with MEM+l-Gln alone was 3.9 μg/ml, compared to 125 μg/ml when vancomycin was preincubated with P. aeruginosa supernatant and MEM+l-Gln. This result is consistent with our observations in Fig. 1. For S. sanguinis, the MBC of vancomycin that was preincubated with MEM+l-Gln alone was 0.98 μg/ml, compared to 1.95 μg/ml when vancomycin was preincubated with P. aeruginosa supernatant derived in MEM+l-Gln. S. sanguinis remains sensitive to vancomycin in the presence of P. aeruginosa supernatant, indicating that P. aeruginosa supernatant does not inactivate vancomycin under our tested conditions.

We next used the coculture assay to track the kinetics of S. aureus tolerance to vancomycin in the presence of P. aeruginosa supernatant. P. aeruginosa supernatant promotes consistently high cell viability of biofilm-grown S. aureus Newman and USA300 in the presence or absence of vancomycin over the course of 26 h (Fig. 2A and B). In contrast, S. aureus biofilms exposed to vancomycin alone experience a steady reduction in cell viability starting at ~5 h after exposure to the antibiotic (Fig. 2A and B). P. aeruginosa supernatant also maintains high cell viability of S. aureus Newman and USA300 planktonic counterparts in the presence of vancomycin during the same time course (Fig. 2C and D).

FIG 2 

Kinetics of S. aureus biofilm and planktonic populations in the presence of P. aeruginosa supernatant and vancomycin. (A to D) Biofilm disruption assays on plastic were performed with S. aureus (Sa) Newman (A and C) or USA300 (B and D), P. aeruginosa PA14 supernatant (Pa sup), and vancomycin (Vanc) at 50 μg/ml. S. aureus biofilm (A and B) and planktonic (C and D) CFU were determined. Data in panels A and C and B and D were from the same experiments. Each time point displays the average from two biological replicates, each with three technical replicates. Error bars indicate SD.

P. aeruginosa-mediated protection of S. aureus from vancomycin is not specific to P. aeruginosa PA14. Another laboratory strain, P. aeruginosa PAO1, as well as various P. aeruginosa clinical isolates, is able to significantly enhance the protection of S. aureus Newman from vancomycin (see Fig. S2A in the supplemental material). As with P. aeruginosa PA14, these P. aeruginosa supernatants do not impact S. aureus Newman biofilm cell viability in the absence of vancomycin (Fig. S2B).

HQNO and siderophores contribute to the ability of P. aeruginosa to protect S. aureus biofilms. (A) Biofilm disruption assays on plastic were performed with S. aureus (Sa) Newman, supernatant (sup) from the specified P. aeruginosa (Pa) strain or clinical isolate, and vancomycin (Vanc) at 50 μg/ml. (B) Biofilm disruption assays on plastic were performed as described for panel A in the absence of vancomycin. (C) An illustration of the P. aeruginosa PQS pathway, including the protein products encoded by the relevant genes tested in the P. aeruginosa PA14 mutant analyses in panels D and F. (D and E) Shown is an analysis of strains with mutations in genes involved in the P. aeruginosa PQS biosynthetic pathway. Biofilm disruption assays on plastic were performed with S. aureus Newman, wild-type P. aeruginosa PA14, and specified deletion mutant supernatants (Pa sup), in the presence (D) or absence (E) of vancomycin (Vanc) at 50 μg/ml. (F) Biofilm disruption assays on plastic were performed with S. aureus USA300, supernatants from the P. aeruginosa PA14 wild type and the specified deletion mutants (Pa sup), and vancomycin (Vanc) at 50 μg/ml. Each column displays the average from at least two biological replicates, each with three technical replicates. Error bars indicate SD. No significant differences were found when the viability of S. aureus Newman exposed to each P. aeruginosa supernatant was compared to the viability of untreated S. aureus Newman (control) by ordinary one-way ANOVA and Dunnett’s multiple comparisons posttest. *, P < 0.05; ***, P < 0.001. Download FIG S2, TIF file, 19.9 MB.

P. aeruginosa supernatant protects S. aureus from vancomycin on CFBE cells.

Previously, we modified an established epithelial cell-P. aeruginosa coculture system on CFBE cells (31) to create a CFBE dual-bacterial coculture system (22). Here, we verified S. aureus forms biofilms on the epithelial monolayers in this system. First, we used microscopy to image S. aureus microcolony formation on CFBE cells (Fig. 3). We observed that S. aureus forms microcolonies by 6 h and continues to form biofilms for up to 21 h (Fig. 3A and B). S. aureus can also form mixed-species microcolonies with P. aeruginosa (Fig. 3C). Additionally, we performed an MBC90 assay on CFBE cell-grown biofilm cells to test whether the population of S. aureus cells attached to the airway cell monolayer exhibits high tolerance to antibiotics, a well-established feature of biofilms. The vancomycin MBC90 value for S. aureus planktonic cells was 1.95 μg/ml, compared to an MBC90 value of 500 μg/ml for the CFBE cell-grown biofilm fraction. Together, these data indicate that S. aureus forms biofilms on CFBE cells.

FIG 3 

P. aeruginosa protects S. aureus biofilms from vancomycin on CFBE cells. (A and B) Representative images of S. aureus microcolonies on CFBE cells 6 h p.i. (A) and 21 h p.i. (B). White arrows indicate microcolonies. (C) Representative image of mixed-species microcolonies composed of S. aureus (red, DsRed expressing) and P. aeruginosa (green, green fluorescent protein [GFP] expressing) on CFBE cells 6 h p.i. (D) Cytotoxicity of S. aureus (Sa) Newman and/or P. aeruginosa PA14 supernatants (Pa sup) either undiluted (1×) or diluted 1/16× on CFBE cells. Cytotoxicity is normalized to total LDH release. (E) Biofilm disruption assays on CFBE cells were performed with S. aureus Newman, P. aeruginosa PA14 supernatant diluted 1/16×, and vancomycin (Vanc) at 50 μg/ml. Each column displays the average from at least three biological replicates, each with three technical replicates. Error bars indicate SD. ns, not significant.

Next, we performed lactate dehydrogenase (LDH) cytotoxicity assays to measure the health of the CFBE cells upon exposure to S. aureus cells and P. aeruginosa supernatant under our assay conditions. Treatment with S. aureus Newman led to low levels of cytotoxicity similar to the MEM+l-Gln control (Fig. 3D). In contrast, undiluted P. aeruginosa PA14 supernatant led to high levels of cytotoxicity and disrupted the CFBE cell monolayers before the end of the experiment. When diluted 1/16×, P. aeruginosa PA14 supernatant showed low levels of cytotoxicity similar to the MEM+l-Gln control (Fig. 3D).

We then employed this dual-species coculture assay to test whether P. aeruginosa decreases S. aureus susceptibility to vancomycin. In this experiment, we sought to closely mirror the conditions and timing of the biofilm disruption assay on plastic described in the previous section. However, we made several modifications to the protocol used for the biofilm disruption assay on plastic to ensure that the epithelial cell monolayers remained intact throughout the experiment. We shortened the biofilm disruption assay on CFBE cells to 21 h compared to the 30-h duration for the disruption assay on plastic. Additionally, P. aeruginosa PA14 supernatant was diluted to preserve the integrity of the monolayers (Fig. 3D). Despite the lower concentration of supernatant used in these studies (1/16× dilution), we observed significant P. aeruginosa-mediated protection of S. aureus Newman from vancomycin on CFBE cells (Fig. 3E). Furthermore, we showed that the epithelial cytotoxicity remained low (Fig. 3D) for all the treatment conditions shown in Fig. 3E.

HQNO and siderophores contribute to the ability of P. aeruginosa to protect S. aureus from vancomycin.

To determine which P. aeruginosa exoproducts are responsible for decreasing S. aureus sensitivity to vancomycin, we assayed strains with mutations in candidate P. aeruginosa PA14 genes that were previously found to be important for interactions between P. aeruginosa and S. aureus (22). The strains we tested had mutations in genes encoding P. aeruginosa-specific secreted products, including phenazines (phzA-G1/2), elastase (lasB), the master transcriptional regulators of Las and Rhl quorum sensing systems (lasR and rhlR), and several components of the Pseudomonas quinolone signal (PQS) quorum sensing system biosynthetic pathway (pqsA, pqsH, and pqsL) (see Fig. S2C in the supplemental material) and siderophore biosynthesis (pvdA and pchE). The mutant analysis was conducted by performing biofilm disruption assays on plastic using supernatants from these mutants with the addition of vancomycin.

We observed that the addition of P. aeruginosa PA14 deletion mutant supernatants differentially affected S. aureus Newman sensitivity to vancomycin. Most of the mutants tested had no impact on the ability of supernatants to confer tolerance to vancomycin (Fig. S2D). Incubation of S. aureus Newman with supernatant from the P. aeruginosa PA14 ΔpqsA deletion mutant resulted in a significant increase in S. aureus Newman sensitivity to vancomycin compared to incubation with supernatant from the wild-type P. aeruginosa PA14 (Fig. S2C and D). Additionally, supernatants from the P. aeruginosa PA14 strains carrying mutations that blocked both HQNO and siderophore production (ΔpqsA ΔpvdA ΔpchE and ΔpqsL ΔpvdA ΔpchE) resulted in a striking increase of S. aureus Newman sensitivity to vancomycin compared to S. aureus Newman treated with supernatant from the wild-type P. aeruginosa strain, with S. aureus Newman attaining levels of sensitivity similar to exposure to vancomycin alone in the absence of P. aeruginosa supernatant (Fig. S2D). In contrast, supernatant from the P. aeruginosa PA14 ΔpqsH deletion mutant (which eliminates PQS but not HQNO production) did not alter S. aureus Newman sensitivity to vancomycin compared to S. aureus Newman treated with supernatant from the wild-type P. aeruginosa strain (Fig. S2D). Taken together, it appears that the combined effect of HQNO and siderophores confers protection of S. aureus Newman from vancomycin. As we observed for P. aeruginosa PA14 wild-type supernatant, P. aeruginosa PA14 deletion mutant supernatants did not impact the biofilm cell viability of S. aureus Newman in the absence of vancomycin (Fig. S2E).

Next, we tested whether HQNO and siderophores may also play a role in decreasing S. aureus Newman sensitivity to vancomycin on CFBE cells. Exposure of S. aureus Newman to supernatant from the P. aeruginosa PA14 ΔpqsL ΔpvdA ΔpchE mutant led to a significant increase in S. aureus Newman susceptibility to vancomycin to levels similar to those of S. aureus Newman exposed to vancomycin alone (Fig. 3E). This result suggests that P. aeruginosa-produced HQNO and siderophores also play a role in protection of S. aureus Newman from vancomycin on CFBE cells.

In the case of S. aureus USA300, it is less clear which factors are responsible for P. aeruginosa-mediated protection of USA300 from vancomycin. Addition of P. aeruginosa PA14 supernatant from ΔpqsA ΔpvdA ΔpchE and ΔpqsL ΔpvdA ΔpchE mutants resulted in significant, but less dramatic increases in S. aureus USA300 susceptibility to vancomycin compared to the effect of these mutants on the susceptibility of S. aureus Newman (Fig. S2F). Exposure of S. aureus USA300 to supernatant from the P. aeruginosa PA14 ΔpqsA deletion mutant did not significantly impact S. aureus USA300 susceptibility to vancomycin (Fig. S2F). These findings suggest that HQNO and siderophores may be partially responsible for P. aeruginosa-mediated protection of S. aureus USA300 from vancomycin and additional factors may be involved.

To further explore the contribution of HQNO and siderophores, we measured the levels of these exoproducts under our experimental conditions. We quantified HQNO and pyoverdine in supernatants from P. aeruginosa grown either on plastic or on CFBE cells in our coculture assays. To measure the levels of HQNO, we performed a functional assay in which the amount of HQNO corresponds to a P. aeruginosa-mediated decrease in S. aureus cell viability compared to a standard curve using pure, commercially available HQNO (Fig. 4A; see Fig. S3A and B in the supplemental material). A P. aeruginosa ΔpqsL mutant does not produce HQNO and thus is deficient in killing S. aureus; however, killing can be restored by adding exogenous HQNO. A standard curve was used to relate the HQNO concentration to S. aureus CFU (Fig. S3A). Using this approach, we found that by 6 h, higher levels of HQNO are present in supernatants from P. aeruginosa grown on CFBE cells (~15 μg/ml) compared to P. aeruginosa grown on plastic (~8 μg/ml HQNO [Fig. 4A]). We also measured the levels of HQNO in supernatant from P. aeruginosa grown on plastic for 24 h, which is the source of supernatant we use throughout the study. Under this last condition, the amount of HQNO is ~10 μg/ml, which is within the range detected in supernatant from P. aeruginosa grown either on CFBE cells or on plastic for 6 h (Fig. 4A).

FIG 4 

HQNO and pyoverdine quantification. (A and B) The amounts of HQNO and pyoverdine in supernatants from wild-type P. aeruginosa PA14 (Pa) grown either on CFBE cells or on plastic were determined. (A) Levels of HQNO were determined using a standard curve relating the HQNO concentration to S. aureus CFU following coculture with a P. aeruginosa PA14 ΔpqsL ΔpvdA ΔpchE deletion mutant. (B) The levels of pyoverdine were determined by measuring absorbance at 405 nm. Each column displays the average from two biological replicates, each with three technical replicates. Error bars indicate SD.

Pyocyanin is not produced under our experimental conditions. (A) Shown is a standard curve relating HQNO concentration to S. aureus CFU following coculture with a P. aeruginosa (Pa) PA14 ΔpqsL ΔpvdA ΔpchE deletion mutant in the presence of the indicated HQNO concentrations. (B) S. aureus (Sa) Newman was cocultured with a P. aeruginosa PA14 ΔpqsL ΔpvdA ΔpchE deletion mutant in the presence of supernatants from wild-type P. aeruginosa PA14 grown either on CFBE cells or plastic. (C) Pyocyanin levels were quantified in supernatants from P. aeruginosa PA14 grown on either CFBE cells or plastic by measuring absorbance at 520 nm. (D and E) Biofilm disruption assays were performed with S. aureus Newman and the specified concentrations of pyocyanin dissolved in DMSO (D) or DMSO (E). Each column displays the average from at least two biological replicates, each with three technical replicates. Error bars indicate SD. bd, below detection; ns, not significant; ***, P < 0.001 by ordinary one-way ANOVA and Dunnett’s multiple comparisons posttest. Download FIG S3, TIF file, 13.2 MB.

Pyoverdine levels were measured by determining the absorbance at 405 nm, as reported previously (32). Supernatants from P. aeruginosa grown on plastic or on CFBE cells have low levels of pyoverdine by 6 h, similar to supernatant from the ΔpvdA control at 6 h (Fig. 4B). Higher levels of pyoverdine are detected in supernatants from P. aeruginosa grown on CFBE cells by 21 h and plastic by 24 h compared to supernatant from the ΔpvdA control at 24 h (Fig. 4B). These results indicate that P. aeruginosa produces the HQNO and pyoverdine exoproducts in our assay system, but the amounts of production differ between different growth conditions.

We also quantified the levels of the phenazine pyocyanin, another known antistaphylococcal factor produced by P. aeruginosa. It has been shown that pyocyanin can block the S. aureus ETC, leading to growth inhibition and selection for S. aureus SCVs (33, 34). Pyocyanin levels were determined by measuring the absorbance at 520 nm. We found that P. aeruginosa does not produce detectable pyocyanin under our growth conditions (Fig. S3C), perhaps explaining why this factor does not appear to play a role in P. aeruginosa-mediated killing of S. aureus in our model (22) (Fig. S2D). We verified that providing exogenous, commercially available pyocyanin, but not the vehicle control (dimethyl sulfoxide [DMSO]), does indeed decrease S. aureus cell viability under our conditions (Fig. S3D and E).

We have shown above that P. aeruginosa culture supernatant can alter S. aureus tolerance to vancomycin; however, we wanted to test whether coculturing S. aureus with P. aeruginosa cells could produce the same effect. Coculture of S. aureus Newman with wild-type P. aeruginosa PA14 for 21 h on CFBE cells did not lead to increased S. aureus tolerance to vancomycin, but instead decreased S. aureus cell viability (data not shown), consistent with previous findings (22). We have also shown that P. aeruginosa PA14 ΔpqsL ΔpvdA ΔpchE is unable to kill S. aureus (22), and here we show the same factors are required to protect S. aureus from vancomycin-mediated killing. We exploited the lack of killing by the ΔpqsL ΔpvdA ΔpchE mutant to test whether cells of these mutants are unable to protect S. aureus from vancomycin. Consistent with the results from supernatant experiments above, cells of the P. aeruginosa PA14 ΔpqsL ΔpvdA ΔpchE mutant are unable to protect S. aureus from vancomycin (data not shown).

Exogenous HQNO protects S. aureus biofilms from vancomycin.

To confirm the contribution of HQNO in altering the sensitivity of S. aureus to vancomycin, we conducted a biofilm disruption assay on plastic with the addition of pure HQNO (Fig. 5), with and without vancomycin. We used 100, 33, and 11 μg/ml of HQNO in our study, concentrations in the range of those measured above and produced by stationary-phase P. aeruginosa cultures in rich media (35, 36) (Fig. 4). We observed that exogenous HQNO protects S. aureus Newman and USA300 from vancomycin in an HQNO dose-dependent manner (Fig. 5A and B). The concentrations of HQNO and DMSO used in our study did not reduce S. aureus Newman biofilm cell viability (see Fig. S4A and B in the supplemental material). Furthermore, HQNO-mediated protection of S. aureus from vancomycin is not specific to S. aureus Newman and USA300; exogenous HQNO decreased the sensitivity to vancomycin of two out of four S. aureus clinical isolates tested (see Fig. S5A to D in the supplemental material). Finally, we showed that this protection was mediated specifically by HQNO because the addition of high levels of PQS, another end product of this pathway (Fig. S2C), did not protect S. aureus Newman biofilms from vancomycin treatment (Fig. S5E).

FIG 5 

Exogenous HQNO protects S. aureus biofilms from vancomycin. (A and B) Biofilm disruption assays on plastic were performed with S. aureus (Sa) Newman (A) or USA300 (B), vancomycin (Vanc) at 50 μg/ml, and the specified concentrations of HQNO (dissolved in DMSO). Each column displays the average from at least three biological replicates, each with three technical replicates. Error bars indicate SD. ns, not significant; *, P < 0.05, **, P < 0.01, and ***, P < 0.001, by ordinary one-way ANOVA and Tukey’s multiple comparisons posttest.

The concentrations of HQNO and DMSO used in this study do not impact the viability of S. aureus Newman biofilms in the absence of vancomycin, and DMSO does not alter sensitivity of S. aureus Newman biofilms to vancomycin. (A) Biofilm disruption assays on plastic were performed with S. aureus (Sa) Newman and the specified concentrations of 2-n-heptyl-4-hydroxyquinoline N-oxide (HQNO, dissolved in DMSO) and DMSO. (B) Biofilm disruption assays on plastic were performed with S. aureus Newman, the specified concentrations of DMSO, and vancomycin (Vanc) at 50 μg/ml. Each column displays the average from at least three biological replicates, each with three technical replicates. Error bars indicate SD. ns, not significant. Download FIG S4, TIF file, 2.1 MB.

HQNO protects S. aureus clinical isolates from vancomycin to various degrees. (A to D) Biofilm disruption assays on plastic were performed with the specified S. aureus (Sa) clinical isolate, 100 μg/ml of HQNO, and vancomycin at 50 μg/ml. The SMC number indicates the strain number used. (E) Biofilm disruption assays on plastic were performed with S. aureus Newman, 100 μg/ml of PQS, and vancomycin at 50 μg/ml. Each column displays the average from at least three biological replicates, each with three technical replicates. Error bars indicate SD. ns, not significant; *, P < 0.05, by ordinary one-way ANOVA and Tukey’s multiple comparisons posttest. Download FIG S5, TIF file, 14.1 MB.

Anoxia protects S. aureus biofilms from vancomycin.

Our previous studies indicated that in coculture, P. aeruginosa interference with S. aureus ETC function results in S. aureus growing via fermentation (22). That is, growth in the presence of P. aeruginosa shifted S. aureus to anoxic-like growth conditions. Here, we sought to determine whether anoxic conditions impact S. aureus biofilm susceptibility to vancomycin under our specific assay conditions, and if so, whether anoxic conditions recapitulate P. aeruginosa-mediated protection of S. aureus from vancomycin.

Under our assay conditions for biofilms grown on a plastic surface, anoxia decreased S. aureus sensitivity to vancomycin (Fig. 6A and B), which is consistent with a previous finding (37). Additionally, it was reported that while the MICs of vancomycin for S. aureus were similar under normoxia and anoxia, the rate of S. aureus planktonic cell death is greater in the presence of oxygen (38, 39). Addition of supernatant from wild-type P. aeruginosa PA14 under anoxic conditions conferred additional protection to S. aureus Newman and USA300 from vancomycin compared to normoxia alone (Fig. 6A and B). Under anoxic conditions, supernatant from the P. aeruginosa PA14 ΔpqsL ΔpvdA ΔpchE mutant conferred protection to both S. aureus Newman and USA300, illustrating that anoxia recapitulates the activity of P. aeruginosa supernatant in protecting S. aureus biofilms from vancomycin (Fig. 6A and B).

FIG 6 

Anoxia protects S. aureus biofilms from vancomycin. (A and B) Biofilm disruption assays on plastic were performed with S. aureus (Sa) Newman (A) or USA300 (B), P. aeruginosa PA14 (Pa) wild-type (WT) and ΔpqsLΔ pvdA ΔpchE deletion mutant supernatants, and vancomycin at 50 μg/ml under either normoxic or anoxic conditions. Each column displays the average from at least three biological replicates, each with three technical replicates. Error bars indicate SD. ns, not significant; *, P < 0.05, **, P < 0.01, and ***, P < 0.001, by ordinary one-way ANOVA and Tukey’s multiple comparisons posttest.

P. aeruginosa supernatant decreases S. aureus growth.

Our data above indicate that P. aeruginosa supernatant, likely via HQNO and siderophores, protects S. aureus biofilms from vancomycin treatment. Previous studies have observed HQNO-dependent growth inhibition of S. aureus (17, 40, 41). Thus, one likely mechanism for the reduced efficacy of vancomycin observed here, given its cell wall target, is slowed growth of the target bacterium.

We tested whether exposure of S. aureus to exogenous HQNO or supernatants from wild-type P. aeruginosa PA14 or the ΔpqsL ΔpvdA ΔpchE mutant would result in reduced S. aureus growth. In this experiment, we monitored growth of S. aureus over the course of 10 h in shaking flask cultures with the same medium used in the coculture assays (MEM+l-Gln), alone or amended with HQNO or P. aeruginosa supernatant. Exposure of S. aureus to HQNO and P. aeruginosa PA14 wild-type supernatant caused a decrease in S. aureus growth compared to S. aureus grown in MEM+l-Gln alone (control [Fig. 7A]). Conversely, S. aureus treated with P. aeruginosa PA14 ΔpqsL ΔpvdA ΔpchE mutant supernatant resulted in a growth profile similar to that of the control (Fig. 7A); both the control and the culture amended with P. aeruginosa PA14 ΔpqsL ΔpvdA ΔpchE mutant supernatant showed an ~1-log increase in viability between 8 and 14 h. Next, we performed an extended growth curve experiment to monitor growth of S. aureus from 12 to 24 h in the presence or absence of P. aeruginosa supernatants. Starting at 16 h, there is an increase in growth of S. aureus exposed to P. aeruginosa PA14 wild-type supernatant, reaching levels similar to those of the control and the culture treated with P. aeruginosa PA14 ΔpqsL ΔpvdA ΔpchE mutant supernatant by 24 h (Fig. 7B). Thus, while supernatants and HQNO slow the growth of S. aureus in the short term, these treatments do not cause long-term loss of viability, an observation consistent with our conclusion that protection of S. aureus from vancomycin treatment in coculture is likely due to the slower growth of this Gram-positive organism.

FIG 7 

P. aeruginosa supernatant decreases S. aureus growth. (A and B) Growth curve assays of planktonic populations in shaking flasks were performed with S. aureus (Sa) Newman, HQNO at 100 μg/ml, and P. aeruginosa PA14 (Pa) wild-type (WT) and ΔpqsL ΔpvdA ΔpchE deletion mutant supernatants. Samples were collected either every 2 h from 0 to 10 h p.i. (A) or every 2 h from 12 h to 24 h p.i. (B). Each time point displays the average from at least three biological replicates, each with two technical replicates. Error bars indicate SD. (C) Planktonic susceptibility assays in shaking flasks were performed with S. aureus (Sa) Newman, P. aeruginosa PA14 supernatant (Pa sup), and 50 μg/ml of vancomycin. Each column displays the average from at two biological replicates, each with two technical replicates. Error bars indicate SD. *, P < 0.05 by ordinary one-way ANOVA and Tukey’s multiple comparisons posttest.

As a control, we verified that P. aeruginosa supernatant protects S. aureus from vancomycin when grown in shaking flasks (Fig. 7C), the assay conditions used to monitor growth in Fig. 7A and B. Overall, these data indicate that P. aeruginosa-derived HQNO and siderophores can reduce or eliminate S. aureus growth in the minimal medium used in these studies.

Exposure to P. aeruginosa supernatant does not select for SCVs under our experimental conditions.

S. aureus SCVs arise due to mutations in the ETC, resulting in high tolerance to antibiotic treatment (42–45). Previous studies reported that exposure of S. aureus to P. aeruginosa or pure HQNO can select for S. aureus SCVs (15, 17, 34). To determine whether SCVs are playing a role in promoting S. aureus biofilm cell viability in our model, we sought to enumerate any SCVs that may arise under our experimental conditions. We conducted biofilm disruption assays on plastic for up to 5 days in which S. aureus Newman was exposed to either MEM+l-Gln alone (control), or supernatants derived from wild-type P. aeruginosa PA14 or the ΔpqsL ΔpvdA ΔpchE mutant. We did not observe SCVs under any of the conditions tested.

P. aeruginosa supernatant does not further enhance the tolerance of a highly resistant S. aureus small colony variant to vancomycin.

To further probe whether SCVs might contribute to the P. aeruginosa-mediated protection of S. aureus from vancomycin, we used a previously described S. aureus SCV, generated by mutating a key gene for heme biosynthesis, hemB (44). An S. aureus Col hemB SCV variant has a defective ETC and the classical small colony phenotype (44, 46–49). The S. aureus Col hemB mutant is highly tolerant to vancomycin (Fig. 8A), which is consistent with previous findings that SCVs are tolerant to other cell wall-active antibiotics (42, 44). We observed that P. aeruginosa PA14 supernatant did not further enhance the tolerance of the S. aureus Col hemB mutant to vancomycin (Fig. 8A), which is consistent with the mechanism by which ETC inhibition confers tolerance to this antibiotic. In contrast to the S. aureus Col hemB mutant, the Col parental strain is susceptible to vancomycin to levels similar to those of S. aureus Newman, and likewise, P. aeruginosa supernatant protected the S. aureus Col parental strain from vancomycin (Fig. 8B).

FIG 8 

S. aureus small-colony variant biofilms are tolerant to vancomycin independent of P. aeruginosa supernatant. (A and B) Biofilm disruption assays on plastic were performed with S. aureus (Sa) Col, P. aeruginosa PA14 supernatant (Pa sup), and vancomycin. (A) The S. aureus Col hemB mutant was exposed to 250 μg/ml of vancomycin. (B) The S. aureus Col parental strain was exposed to 50 μg/ml of vancomycin. Each column displays the average from at least three biological replicates, each with three technical replicates. Error bars indicate SD. ns, not significant; **, P < 0.01, and ***, P < 0.001, by ordinary one-way ANOVA and Tukey’s multiple comparisons posttest.

P. aeruginosa supernatant reduces S. aureus cell wall thickness.

A commonly observed feature of vancomycin-resistant strains of S. aureus is an increase in cell wall thickness (50, 51). S. aureus cells grown anaerobically also exhibit an increase in cell wall thickness (52). We examined whether exposure of biofilm-grown S. aureus Newman to P. aeruginosa PA14 supernatant causes S. aureus cell wall thickening, which may explain the observed decreased susceptibility to vancomycin of S. aureus in the presence of P. aeruginosa. Surprisingly, we found that S. aureus Newman cells exposed to MEM+l-Gln alone (control) for 24 h had very thick cell walls (average of 70 nm [see Fig. S6A and D in the supplemental material]). In contrast, S. aureus Newman biofilm cells that were exposed to P. aeruginosa supernatant for 24 h had cell walls significantly less thick relative to the control (average of 31 nm [Fig. S6B and D]), values that were comparable to those reported for S. aureus cells grown in rich medium (50, 51). Next, we tested whether HQNO could be mediating the observed decrease in S. aureus cell wall thickness upon exposure to P. aeruginosa supernatant. Indeed, S. aureus Newman biofilm cells treated with exogenous HQNO had less thick cell walls relative to the control (average of 28 nm [Fig. S6C and E]), which recapitulated the effect of P. aeruginosa PA14 supernatant. These results indicate that thickening of the S. aureus Newman cell wall is not the basis for increased vancomycin tolerance when exposed to P. aeruginosa PA14 supernatant.

P. aeruginosa supernatant decreases the cell wall thickness of biofilm-grown S. aureus. (A to C) Representative transmission electron microscopy (TEM) images of S. aureus Newman grown in MEM + l-Gln alone (control [A]) or in the presence of either P. aeruginosa PA14 supernatant (B) or 100 μg/ml of HQNO (C). (D and E) Cell wall thickness was measured for S. aureus (Sa) Newman cells grown as biofilms incubated with or without P. aeruginosa PA14 supernatant (Pa sup) for 24 h (D) or with or without 100 μg/ml of HQNO (dissolved in DMSO) for 24 h (E). Each column displays the average of two biological replicates, each with two technical replicates. Error bars indicate SD. *, P < 0.05, and ***, P < 0.001, by two-tailed unpaired t test. Download FIG S6, TIF file, 3.3 MB.

P. aeruginosa supernatant protects S. aureus from other antibiotics.

The data presented here show that P. aeruginosa supernatant can protect S. aureus biofilms from the antimicrobial effect of vancomycin. We wanted to ask how broadly this protection extended to other antibiotics. We found that this protection phenotype is not specific to vancomycin; P. aeruginosa PA14 supernatant protects S. aureus Newman from another cell wall-active antibiotic—oxacillin (see Fig. S7A in the supplemental material). In contrast, P. aeruginosa supernatant does not protect S. aureus Newman or USA300 biofilms from the antibiotic daptomycin (Fig. S7B and C). Unlike vancomycin and oxacillin, daptomycin does not target the cell wall of S. aureus.

Testing P. aeruginosa-mediated protection of S. aureus biofilms from other antibiotics. (A) Biofilm disruption assays on plastic were performed with S. aureus (Sa) Newman, P. aeruginosa PA14 supernatant (Pa sup), and the specified concentration of oxacillin. (B and C) Biofilm disruption assays on plastic were performed with S. aureus Newman (B) or USA300 (C), P. aeruginosa PA14 supernatant, and the specified concentrations of daptomycin. Each column displays the average from at least three biological replicates, each with three technical replicates. Error bars indicate SD. ns, not significant; *, P < 0.05 by ordinary one-way ANOVA and Tukey’s multiple comparisons posttest. Download FIG S7, TIF file, 1.8 MB.

To more thoroughly test whether P. aeruginosa supernatant can protect S. aureus from other antibiotic classes, we screened Biolog Phenotype MicroArray panels containing 240 antibiotics for a supernatant-mediated protection phenotype. For each panel, S. aureus Newman was exposed to either MEM+l-Gln alone (control) or P. aeruginosa PA14 wild-type supernatant. Interestingly, it appears that P. aeruginosa supernatant can protect S. aureus from a wide range of antibiotics, many of which fall into two broad categories: cell wall synthesis inhibitors and protein synthesis inhibitors (see Table S1 in the supplemental material). Specifically, these antibiotics include multiple representatives from the β-lactam, glycopeptide, aminoglycoside, macrolide, and tetracycline classes. Importantly, we confirmed our previous findings that P. aeruginosa supernatant protects S. aureus from vancomycin and oxacillin (Table S1; Fig. S7). We also observed that P. aeruginosa supernatant reduces S. aureus sensitivity to the nucleic acid inhibitors rifampin and novobiocin, as well as numerous other chemicals (Table S1).

P. aeruginosa supernatant protects S. aureus from several classes of antibiotics. Download TABLE S1, PDF file, 0.1 MB.

Furthermore, we tested whether HQNO and siderophores are required for protection of S. aureus from a subset of antibiotics in the Phenotype MicroArray panels. S. aureus Newman was added to panel 12 and exposed to either MEM+l-Gln alone (control) or P. aeruginosa PA14 wild-type or ΔpqsL ΔpvdA ΔpchE mutant supernatants. We found that HQNO and siderophores contribute to the ability of P. aeruginosa to protect S. aureus from cell wall-targeting antibiotics (β-lactam class), protein synthesis inhibitors (aminoglycoside, macrolide, and tetracycline classes), and nucleic acid inhibitors (rifampin and novobiocin) (Table S1).

DISCUSSION

In this study, we have shown that interspecies bacterial interactions alter antibiotic tolerance in unpredictable ways. Specifically, we found that P. aeruginosa protects biofilm and planktonic populations of S. aureus from vancomycin, a frontline drug to treat MRSA in CF patients and in other infections. Furthermore, we observed that P. aeruginosa-mediated protection of S. aureus from vancomycin occurs with multiple P. aeruginosa and S. aureus strains, as well as clinical isolates.

We have shown that the P. aeruginosa exoproducts HQNO and siderophores contribute to protection of S. aureus from vancomycin. We previously demonstrated that both factors are required to shift S. aureus from respiration to fermentation, thus slowing the growth of the bacterium (22). To test whether inhibition of electron transport resulted in increased tolerance to vancomycin, we exposed S. aureus to anoxic, fermentative conditions to inhibit respiration in a different way. Anoxia recapitulated the effect of P. aeruginosa supernatant. Overall, we propose that the enhanced tolerance of the biofilm is mediated, at least in part, by the reduced growth of S. aureus when this microbe is grown in the presence of P. aeruginosa exoproducts. Our growth assays support this hypothesis. In contrast, our data do not support a role for SCVs in P. aeruginosa-mediated protection of S. aureus from vancomycin in our model system. Consistent with our model, it has been shown that anoxia also slows S. aureus growth in minimal medium (53). Additionally, it has been reported that S. aureus also exhibits slow growth in sputum from CF patients, raising the possibility of increased S. aureus tolerance to vancomycin in vivo (54).

We observed that P. aeruginosa supernatant protects S. aureus from antibiotics with two broad mechanisms of action: cell wall synthesis inhibitors and protein synthesis inhibitors. Thus, the interaction we observe here could have a general effect on antimicrobial therapy in the context of polymicrobial infections. Furthermore, we showed that HQNO and siderophores are required for protection from representatives from both cell wall-targeting antibiotics and protein synthesis inhibitors. We suggest above that slowed growth and/or the shift to fermentative growth of S. aureus by P. aeruginosa exoproducts confers resistance to vancomycin, oxacillin, and other cell wall-active antibiotics. Similarly, we propose a model wherein HQNO- and siderophore-mediated inhibition of the ETC disrupts the electrochemical gradient and thus prevents cell entry of protein synthesis inhibitors that require protein motive force, including aminoglycosides and tetracyclines (55). Furthermore, the mechanism by which P. aeruginosa alters S. aureus antibiotic tolerance here may extend to other polymicrobial interactions. For example, several P. aeruginosa-derived phenazines repress Candida albicans respiratory metabolism and drive this fungus toward fermentative metabolism and the production of ethanol (56–58).

We also tested whether P. aeruginosa supernatant alters the sensitivity of S. aureus to daptomycin, which acts via a different mechanism. Daptomycin inserts into the cell membrane, leading to membrane depolarization, a loss of membrane potential, and subsequent death of the bacterial cell. P. aeruginosa supernatant does not protect S. aureus from daptomycin, further suggesting that P. aeruginosa-mediated protection of S. aureus is dependent on the antibiotic’s mode of action. Moreover, in our Biolog Phenotype MicroArray screen, we observed that P. aeruginosa supernatant did not protect S. aureus from nucleic acid inhibitors apart from rifampin and novobiocin.

P. aeruginosa produces several known antistaphylococcal factors: HQNO, siderophores, pyocyanin, and rhamnolipids (18, 22, 33, 59, 60). Pyocyanin does not seem to be involved in P. aeruginosa-S. aureus interactions here because it is not produced in appreciable amounts by P. aeruginosa under our experimental conditions (Fig. S3C). In our system, HQNO and siderophores are involved in multiple P. aeruginosa-S. aureus interactions: P. aeruginosa-mediated killing of S. aureus (22) and P. aeruginosa-mediated protection of S. aureus from antibiotics. P. aeruginosa-mediated killing of S. aureus occurs after 10 to 16 h of coculture, whereas P. aeruginosa-mediated tolerance of S. aureus to vancomycin can occur after as early as 6 h, indicating the possibility that these interactions could be occurring sequentially. Additionally, S. aureus cells could be experiencing these P. aeruginosa exoproducts at a distance, resulting in a gradient of P. aeruginosa-mediated protection of S. aureus from antibiotics.

Microbes do not exist in isolation, but rather, as members of a polymicrobial community—a fact that is still underappreciated. We have shown that the antibiotic sensitivity of one microbe can change dramatically and unexpectedly when in the presence of another microbial species, underscoring the difficulty of extrapolating from monoculture experiments to polymicrobial settings. In recent years, other groups have shown other examples of microbial interactions influencing antibiotic effectiveness (17, 61–63). Furthermore, the biofilm mode of growth contributes high-level tolerance to antimicrobial agents and must be considered when studying infections involving biofilms. Clearly, neighboring microbes in a mixed infection can impact this antimicrobial tolerance. Elucidation of the molecular mechanisms and consequences of these interspecies interactions may allow us to better anticipate the outcomes of treating a specific patient’s polymicrobial community with a specific antibiotic. We believe our findings are especially relevant to CF—a chronic, polymicrobial disease that requires continuous treatment with numerous antimicrobial agents.

MATERIALS AND METHODS

See Text S1 in the supplemental material for additional details regarding the methods.

Supplemental materials and methods. Download TEXT S1, PDF file, 0.2 MB.

Bacterial strains and culture conditions.

A list of all S. aureus and P. aeruginosa strains used in this study is included in Table S2 in the supplemental material. S. aureus was grown in tryptic soy broth (TSB), and P. aeruginosa was grown in lysogeny broth (LB). All overnight cultures were grown with shaking at 37°C for 12 to 14 h, except for the S. aureus Col hemB mutant, which was grown statically at 37°C for 20 h.

Strains used in this study. Download TABLE S2, PDF file, 0.1 MB.

Biofilm disruption assay on plastic.

Overnight liquid cultures of S. aureus were diluted to an optical density at 600 nm (OD600) of 0.05, washed in phosphate-buffered saline (PBS), and resuspended in minimal essential medium (MEM [Thermo, Fisher Scientific]) supplemented with 2 mM l-glutamine (MEM+l-Gln). Triplicate wells of a plastic 96-well plate were inoculated with 100 μl of the S. aureus suspensions and incubated at 37°C in 5% CO2. Unattached cells were removed 1 h postinoculation (p.i.), and 90 μl of MEM+l-Gln was added to each well. The plate was incubated at 37°C in 5% CO2. Unattached cells were removed 6 h p.i., at which point antibiotic dilutions in MEM+l-Gln, P. aeruginosa supernatant, and MEM+l-Gln were added (total well volume of 90 μl). The plate was incubated at 37°C in 5% CO2. The planktonic cell population was collected 30 h p.i., serially diluted 10-fold in PBS, and plated on mannitol salt agar. After incubation at 37°C for 18 h, planktonic CFU were determined. To collect the remaining biofilm cell population from the 96-well plate, 50 μl of 0.1% Triton X-100 in PBS was added to each well. Next, the plate was gently agitated on an undulating rocker for 30 min, and wells were scraped using a solid multipin replicator. Biofilms were further disrupted by covering the plate with a foil seal and vortexing for 2 min. Biofilm cells were serially diluted, plated, and enumerated as described for the planktonic cells. Viable cell counts for all relevant experiments are reported as log10 transformed CFU per milliliter. For time course assays, samples were collected and analyzed 0, 1, 3, 6, 19, and 26 h after the 6-h p.i. medium change.

MBC assay to test inactivation of vancomycin by P. aeruginosa supernatant.

Vancomycin dilutions were prepared in either MEM+l-Gln or P. aeruginosa supernatant and incubated for 24 h. Minimum bactericidal concentrations (MBCs) of vancomycin (preincubated with either MEM+l-Gln or P. aeruginosa supernatant) for S. aureus and Streptococcus sanguinis were determined. See Text S1 in the supplemental material for additional details regarding the methods.

Biofilm disruption assay on CFBE cells.

Overnight liquid cultures of S. aureus were diluted to an OD600 of 0.05, washed in PBS, and resuspended in MEM+l-Gln. CFBE cells were grown in a plastic 24-well plate until confluent, at which point the CFBE monolayers were washed twice with 500 μl MEM+l-Gln. Next, triplicate wells were inoculated with 500 μl of the S. aureus suspensions and incubated at 37°C in 5% CO2. Unattached cells were removed 1 h p.i., and 450 μl of MEM+l-Gln was added to each well. The plate was incubated at 37°C, 5% CO2. Unattached cells were removed 6 h p.i., at which point 50 μg/ml vancomycin in MEM+l-Gln, P. aeruginosa supernatant, and MEM+l-Gln were added (total well volume of 500 μl). Planktonic cell populations were removed 21 h p.i. Then biofilms were disrupted by adding 250 μl of PBS to each well and scraping thoroughly with a plastic pipette tip. Biofilm cells were serially diluted and plated as previously described for the biofilm disruption assay on plastic.

Bright-field and fluorescence microscopy.

CFBE cells were inoculated with S. aureus and P. aeruginosa. Microcolonies were imaged at 6 and 21 h p.i. by bright-field and fluorescence microscopy. See Text S1 in the supplemental material for additional details regarding the methods.

MBC90 assay on CFBE cells.

MBC90 of vancomycin was determined for planktonic and biofilm populations of S. aureus grown on CFBE cells. See Text S1 in the supplemental material for additional details regarding the methods.

Cytotoxicity assay.

Biofilm disruption assays were performed on CFBE cells as previously described. Supernatants were collected 21 h p.i., and lactate dehydrogenase (LDH) release from the CFBE cells was measured. See Text S1 in the supplemental material for additional details regarding the methods.

Quantification of P. aeruginosa exoproducts.

P. aeruginosa was grown either on plastic or on CFBE cells (in MEM+l-Gln) as described previously, and supernatants were collected 6, 21, or 24 h p.i. HQNO levels in supernatants were measured using a standard curve relating pure HQNO concentration to S. aureus CFU upon coculture with the P. aeruginosa ΔpqsL mutant. Pyoverdine and pyocyanin were quantified as previously described (32, 64). See Text S1 in the supplemental material for additional details regarding the methods.

Other growth assays.

For growth curves in shaking flasks, S. aureus was exposed to either HQNO, P. aeruginosa supernatant, or MEM+l-Gln alone. Samples were collected every 2 h, and planktonic CFU were determined. For the planktonic susceptibility assay in shaking flasks, S. aureus was exposed to either P. aeruginosa supernatant or MEM+l-Gln, with or without vancomycin. For the SCV selection assay, S. aureus was exposed to either P. aeruginosa supernatant or MEM+l-Gln alone. SCV selection was performed as previously described (15). See Text S1 in the supplemental material for additional details regarding the methods.

Transmission electron microscopy.

S. aureus was exposed to P. aeruginosa supernatant, HQNO, or MEM+l-Gln alone. Biofilm cells were fixed, stained, and sectioned. Samples were imaged with a transmission electron microscope, and cell wall thickness was measured. See Text S1 in the supplemental material for additional details regarding the methods.

Biofilm antibiotic susceptibility assay on plastic.

A modified biofilm antibiotic susceptibility assay was performed using Biolog Phenotype MicroArray bacterial chemical sensitivity panels. See Text S1 in the supplemental material for additional details regarding the methods.