Global transcriptome responses including small RNAs during mixed‐species interactions with methicillin‐resistant Staphylococcus aureus and Pseudomonas aeruginosa

Abstract Pseudomonas aeruginosa and Staphylococcus aureus mixed‐species biofilm infections are more resilient to biocide attacks compared to their single‐species counterparts. Therefore, this study used an in vitro model recapitulating bacterial burdens seen in in vivo infections to investigate the interactions of P. aeruginosa and S. aureus in biofilms. RNA sequencing (RNA‐seq) was utilized to identify the entire genomic response, both open reading frames (ORFs) and small RNAs (sRNAs), of each species. Using competitive indexes, transposon mutants validated uncharacterized PA1595 of P. aeruginosa and Panton–Valentine leukocidin ORFs of S. aureus are required for competitive success. Assessing spent media on biofilm development determined that the effects of these ORFs are not solely mediated by mechanisms of secretion. Unlike PA1595, leukocidin (lukS‐PV) mutants of S. aureus lack a competitive advantage through contact‐mediated mechanisms demonstrated by cross‐hatch assays. RNA‐seq results suggested that during planktonic mixed‐species growth there is a robust genomic response or active combat from both pathogens until a state of equilibrium is reached during the maturation of a biofilm. In mixed‐species biofilms, P. aeruginosa differentially expressed only 0.3% of its genome, with most ORFs necessary for growth and biofilm development, whereas S. aureus modulated approximately 5% of its genome, with ORFs suggestive of a phenotype of increased virulence and metabolic quiescence. Specific expression of characterized sRNAs aligned with the genomic response to presumably coordinate the adaptive changes necessary for this homeostatic mixed‐species biofilm and sRNAs may provide viable foci for the design of future therapeutics.

Even though in vivo, S. aureus and P. aeruginosa are found cocolonizing, numerous studies have shown that P. aeruginosa frequently outcompetes or kills S. aureus in vitro (Baldan et al., 2014;Biswas, Biswas, Schlag, Bertram, & Gotz, 2009;Pastar et al., 2013;. Pseudomonas aeruginosa's ability to outcompete S. aureus for the niche have been linked to its ability to sense S. aureus through the shedding of peptidoglycan (Korgaonkar, Trivedi, Rumbaugh, & Whiteley, 2013) and secrete numerous products including phenazines (Cardozo et al., 2013), 4-hydroxy-2-heptylquinoline-N-oxide (Machan, Taylor, Pitt, Cole, & Wilson, 1992), extracellular matrix (Qin, Yang, Qu, Molin, & Tolker-Nielsen, 2009), and proteases (e.g., LasA and LasB) (Park, Lee, Cho, Herzberg, & Lee, 2012). To survive with P. aeruginosa in vivo, S. aureus forms small colony variants (Biswas et al., 2009) and induces expression of many virulence factors and regulatory genes (Park et al., 2012). In wounds, it is known that S. aureus tends to be found near the surface of the wound, while P. aeruginosa is generally found deep in the wound bed (Kirketerp-Moller et al., 2008), though some studies have shown wounds where morphologically distinct bacteria have segregated into microcolonies with P. aeruginosa along the wound margin and other species following behind (Dalton et al., 2011). However, the details of the specific interspecies interactions where they interface are unclear.
Previous studies have addressed gene expression in P. aeruginosa (Folsom et al., 2010;Waite et al., 2006) and S. aureus (Beenken et al., 2004;Resch et al., 2006) biofilms and have selectively assessed factors thought to contribute to the dynamic state of a mixed-species biofilm (Park et al., 2012;Pastar et al., 2013). In this study, we have used an in vitro model to investigate the interactions of P. aeruginosa and S. aureus in biofilm and planktonic cultures and utilized a custom RNA-sequencing method to unbiasedly capture the global transcriptome response including small RNAs of each species. Since noncoding sRNAs can regulate numerous targets and are capable of controlling a wide range of adaptive processes (Gottesman et al., 2006;Storz, Vogel, & Wassarman, 2011;Tomasini et al., 2014), we hypothesized that regulatory sRNAs play a key role in modulating interspecies interactions in the biofilm. Surprisingly, in our model, P. aeruginosa plays a more passive role than previously suspected in the presence of S. aureus during biofilm growth by upregulating only a few genes of which most are important for growth and biofilm development. In contrast, S. aureus adapts robustly by differentially expressing many genes, including those contributing to the production of capsule and leukocidin, which promotes its success in the dual-species biofilms and helps S. aureus coexist in close proximity with P. aeruginosa. Finally, these pathogens express sRNAs that appear to parallel the gene transcription to presumably coordinate the adaptations required for the success of each biofilm species.

| Bacterial strains and growth conditions
Strains used are listed in Table 1. Cultures were grown in either 20% or 100% Brain Heart Infusion medium (Thermo Fisher Scientific Remel Products, Lenexa, KS) supplemented with 2% NaCl and 1% glucose (BHI++). For general cell enumeration experiments, cells were plated on trypticase soy agar plates supplemented with 5% sheep's blood (TSA+). For isolation of either P. aeruginosa or S. aureus, cells were plated on Pseudomonas isolation agar (PIA) (Thermo Fisher) or lipovitellin salt mannitol agar (LSMA) (HiMedia Laboratories, Mumbai, India), respectively. Cells were enumerated using the ProtoCOL automated colony counter (Microbiology International, Frederick, MD). A competitive index (CI) was generated by growing a mixture of P. aeruginosa and S. aureus, seeded together at a 1:1 ratio, and enumerating colony forming units for each species either after reaching mid-logarithmic growth phase or after 24 hr under biofilm growth conditions. The competitive index was calculated as the S. aureus/P. aeruginosa output ratio divided by the input inoculum ratio.

| Biofilm formation
Biofilms were grown on glass disks in a static system or on glass slides/disks in a drip-flow biofilm reactor. To grow biofilms on static glass disks, an overnight culture of P. aeruginosa or S. aureus was diluted 1:20 into fresh 20% BHI++ and grown to an OD 600 of 0.5. For static biofilms, the culture was then diluted to an OD 600 of 0.05 into phosphate-buffered saline (PBS) and incubated with borosilicate glass disks (Ace Glass Inc, Vineland, NJ) in a 24-well plate with one disk per well for 2 hr at room temperature. For mixed-species biofilms, a 1:1 mixture of P. aeruginosa and S. aureus was incubated with the glass disks. After attachment, the disks were rinsed and 1 ml fresh 20% or 100% BHI++ was added to each well. After incubation for 24 hr at 37°C with shaking the disks were rinsed and placed in 1 ml of PBS.
Biofilms were detached by sonication and the cells were plated on TSA+, PIA, or LSMA for enumeration after overnight incubation at 37°C. For drip-flow biofilms, the cultures were prepared as previously mentioned. Following dilution into PBS, 5 ml of cells were injected into each chamber of a drip-flow reactor (Biosurface Technologies Corporation, Bozeman, MT). Each chamber contained one glass microscope slide and three glass disks. Cells were incubated in the reactor for 2 hr at room temperature at which point the nonadherent cells were drained and the drip-flow reaction was allowed to proceed with fresh 20% or 100% BHI++ at a rate of 1 ml per minute per chamber for 24 hr at 37°C.

| RNA extraction
Following 24 hr of growth in the drip-flow reactor, biofilms were removed from the glass slides and preserved in RNA protect (Qiagen).
For planktonic cultures, cells were grown to mid-log phase and harvested by centrifugation. The cells were then preserved in RNA protect. Staphylococcus aureus was incubated with lysozyme (Sigma-Aldrich) for complete cell lysis before RNA extraction. RNA was extracted using the mirVana miRNA Isolation Kit (Life Technologies, Grand Island, NY) according to the manufacturer's instructions for isolation of whole and small RNA fractions. The sample was then treated twice with TURBO DNase (Ambion, Grand Island, NY) according to manufacturer's instructions to remove contaminating genomic DNA. RNA was quantified using a NanoDrop ND-1000 spectrophotometer (Thermo Scientific, Wilmington, DE) and analyzed for quality using an Agilent 2100 Bioanalyzer (Agilent Technologies, Santa Clara, CA).

| Differential expression analyses
For differential expression analysis of genes (Table S2 and S3) and small RNAs (Tables S4-S7)

| Deletion of pa1595
A suicide vector was generated using In-Fusion HD cloning kit (Clontech Laboratories), as described previously (Miller et al., 2016), which allowed for deletion of the pa1595 gene through allelic exchange with an antibiotic cassette, pucGM. Briefly, primers were designed according to the manufacturer's suggestion (

| Swarming assays
Swarming motility assays were performed as described previously (Heurlier et al., 2004). Bacteria were grown in 20% BHI++ medium overnight for 16 hr and 3 μl of culture was plated in four independent replicates. Swarming was evaluated on plates containing 0.5% w/v Bacto agar (Difco), 8 g of nutrient broth (Oxoid) per liter, and 0.5% w/v d-glucose. Swarming was observed after 24 hr of incubation at 37°C and was repeated on three separate occasions.

| Pyocyanin (PYO) assay
PYO quantitation was performed as described previously (Essar, Eberly, Hadero, & Crawford, 1990). Specifically, a 5-ml sample of culture grown overnight in 20% BHI++ was centrifuged at 12,000g for 20 min and the supernatant was extracted with 3 ml of chloroform and then re-extracted into 1 ml of 0.2 N HCl to give a pink to deep red solution. The absorbance of this solution was measured at 520 nm.

| Cross-streak assay
A cross-streak assay was performed as described previously (Michelsen et al., 2014). Briefly, one loopful of an overnight culture of P. aeruginosa was streaked horizontally across TSA+ plates and allowed to dry. One loopful of an overnight culture of S. aureus was then streaked vertically on the same plate. Once the S. aureus streak dried, the plates were incubated overnight at 37°C. The plates were then photographed to allow qualitative interpretation of the P. aeruginosa-S. aureus interaction.

| Statistical analysis
All statistical data were calculated using GraphPad Prism version 5.0 for Macintosh (GraphPad Prism Software, San Diego, CA). Statistical significance was accepted when P values were equal to or less than 0.05.

| RNA-seq data accession number
The RNA-seq reads and the DESeq results have been deposited in NCBI's Gene Expression Omnibus (Edgar, Domrachev, & Lash, 2002) and are accessible through Bioproject accession no PRJNA341717.

| In vitro growth conditions mimic P. aeruginosa and S. aureus burdens seen in dual-species biofilms of the wound environment
To determine which culture conditions would best mimic the wound environment and bacterial burden ratios seen in wound infections, we grew our cultures in two different concentrations of BHI++ (20% and 100%), a chopped meat-based media.
Using 100% BHI++ for the growth of mixed planktonic cultures, we found that when starting with a bacterial ratio of 1:1, there was little difference during log phase, but after 24 hr, P. aeruginosa outcompeted S. aureus by about one log ( Figure 1a). Mixed planktonic cultures grown in 20% BHI++ had similar bacterial ratios as cultures grown in 100% BHI++ during early growth phases, but after 24 hr, S. aureus numbers dropped significantly ( Figure 1a). When growing mixed biofilms, we used direct cell numbers to calculate a competitive index (CI). The CI was calculated as the S. aureus/P. aeruginosa output ratio divided by the input inoculum ratio; therefore, a CI greater than 1 indicates that S. aureus is in greater numbers in the dual-species biofilm. Biofilms grown under static conditions in 100% BHI++ yielded a CI of 87.4 with S. aureus strongly outcompeting P. aeruginosa; whereas growth in 20% BHI++ yielded a CI of 9.9 × 10 −3 (Figure 1b), representing approximately 100 P. aeruginosa for every 1 S. aureus, a similar ratio previously reported in a porcine wound model (Pastar et al., 2013). Therefore, we chose to use 20% BHI++ for further experiments. To obtain sufficient quantities of RNA for sequencing, biofilms were grown under dripflow conditions because the CI of biofilms grown in the drip-flow was not significantly different from the CI of static biofilms grown in 20% BHI++ (CI 1.4 × 10 −2 , p = .14) ( Figure 1b).

| Scanning electron microscope demonstrates P. aeruginosa and S. aureus growth dynamics and burden in in vitro dual-species biofilms
Single-and mixed-species biofilms were grown in 20% BHI++ under static and drip-flow conditions and 100% BHI++ under static conditions. The mixed-species biofilm micrographs depicted trends corresponding to the ratios observed using direct cell counting ( Figure 2).
In both static and drip-flow biofilms grown in 20% BHI++, P. aeruginosa significantly outcompeted S. aureus, while the opposite trend was seen with biofilms grown in 100% BHI++ under static conditions.

| Differential gene expression of P. aeruginosa and S. aureus in mixed biofilm or planktonic cultures when compared to single-species conditions
RNA sequencing was performed on RNA isolated from mixed-and single-species biofilm and planktonic cultures (collected at an OD 600 of 0.5) of P. aeruginosa and S. aureus. Two replicates of each experimental conditional were assessed and gene expression trends were confirmed using quantitative real-time PCR (qRT-PCR) for randomly chosen genes with high levels of differential expression with RNA-seq ( Fig. S1).   (Table S3). Table 3 contains a summary of these results, in which each ORF was assigned to an orthologous group (COG) based on predicted or known function (Angiuoli et al., 2011;Winsor et al., 2011).

| Few genes are differentially expressed in P. aeruginosa when grown in a biofilm with S. aureus
Using our criteria for determining significant differential gene expression (≥2 fold change, p < .05), only 16 ORFs (nine up-and seven downregulated) experienced expression changes when P. aeruginosa was grown in a mixed-versus single-species biofilm ( Table 3) The protein products of exoS and phzB1 are toxic molecules that allow F I G U R E 1 In vitro growth conditions to mimic bacterial burden ratios seen in in vivo mixed-species infections. Staphylococcus aureus grown in 100% BHI++ with Pseudomonas aeruginosa competes better and grows to higher numbers compared to coculture planktonic growth in 20%BHI++, where after 24 hr P. aeruginosa outcompetes S. aureus. Cell enumeration using colony-forming units depicts P. aeruginosa and S. aureus growth dynamics when seeded together at a 1:1 ratio and grown under planktonic growth conditions in either 20% or 100% BHI++ media (a). Pseudomonas aeruginosa and S. aureus subcultures were seeded together at a 1:1 ratio and grown under static or drip-flow biofilm conditions in either 20% or 100% BHI++ media. Unlike 100% BHI++, growth as either a static or drip-flow mixed-species biofilm in 20% BHI++ media yields bacterial ratios previously seen in in vivo infections; specifically, a P. aeruginosa to S. aureus ratio of approximately 100:1. A competitive index (CI) was calculated as the S. aureus CFU/P. aeruginosa CFU output ratio divided by the input inoculum ratio (b). Data represent the mean ± standard deviation from at least four independent experiments for invasion and killing of the host (Cezairliyan et al., 2013;Rangel, Diaz, Knoten, Zhang, & Hauser, 2015). Also, both ExoS and FliC are secreted by the type III secretion system and play important roles in stimulating the host inflammatory response (Ince, Sutterwala, & Yahr, 2015). The upregulated ORFs were categorized as being involved in general cell maintenance functions such as ion transport and utilization (e.g., moeA/PA3914, hasE/, kdpE/PA1632, PA4220, and PA1854) and amino acid metabolism (Table 2). One study demonstrated that amino acid metabolism along with cell motility and stress response is implicated in biofilm development in Salmonella enterica serovar Typhimurium (Hamilton et al., 2009).
Resulting cell lysis and release of cellular components provides material for the structural components of the biofilm (Ma et al., 2009;Rice et al., 2007).

| Pseudomonas aeruginosa grown in mixed planktonic cultures downregulates many ORFs involved in metabolism, general cellular processes, and type VI secretion
In contrast to the limited set of ORFs that we found to be differentially expressed when P. aeruginosa was grown in a mixed-species biofilm, hundreds of ORFs were differentially expressed when P. aeruginosa was grown in a mixed planktonic culture (Tables 3 and S2 ORFs are involved in the type VI secretion system (T6SS) (PA0074 to PA0091). Pseudomonas aeruginosa uses its T6SS to secrete toxins F I G U R E 2 SEM micrograph depicts similar trends in the ratios of Pseudomonas aeruginosa and Staphylococcus aureus in mixed-species cocultures as previously observed using direct cell counting in Figure 1. Single-and mixed-species biofilms were grown in 20% BHI++ under static and drip-flow conditions and 100% BHI++ under static conditions. In mixed cultures grown with 20% BHI++ (static or drip-flow), P. aeruginosa was the dominant species. In static biofilms grown in 100% BHI++, S. aureus significantly outcompetes P. aeruginosa. Images show 10,000× magnification and are representative of three independent experiments (Tse1-3) into competing bacteria that hydrolyze peptidoglycan (Hood et al., 2010;Russell et al., 2011). To protect itself from these toxins, P.
aeruginosa synthesizes immunity proteins Tsi1 and Tsi3 that were also downregulated along with the toxins.
Interestingly, when P. aeruginosa is grown as a planktonic culture with S. aureus, 34% of the most prominent upregulated ORFs have unknown functions which promotes speculation for their role in coordinating events allowing P. aeruginosa to dominate during the later stage of planktonic growth.

| Staphylocccus aureus downregulates amino acid and nucleotide biosynthesis when grown with P. aeruginosa in biofilms
Differential gene expression analysis showed a marked decrease in the expression of amino acid and nucleotide biosynthetic genes in S.
aureus when comparing mixed-and single-species biofilms (

| Various S. aureus virulence factors are differentially expressed in mixed biofilms
RNA-seq analysis showed upregulation of known staphylococcal virulence factors including Panton-Valentine leukocidin (lukS-PV and lukF-PV), protein A (spA), a protease (clp), and a complement inhibitor.
Many of these proteins are secreted factors that facilitate host cell destruction and immune evasion (Lina et al., 1999;Zhang, Jacobsson, Vasi, Lindberg, & Frykberg, 1998). Iron sequestration genes belonging to the Isd family (including isdA, C, E, and G2) were upregulated when S. aureus was grown in a biofilm with P. aeruginosa. This family is involved in acquiring heme for scavenging iron (Tiedemann, Muryoi, Heinrichs, & Stillman, 2008). In addition, three proteins be-

| Staphylocccus aureus downregulates various metabolic processes when grown in planktonic cultures with P. aeruginosa
Similar to the trends observed when S. aureus is grown in a biofilm with P. aeruginosa, under mixed planktonic conditions, there is an overall decrease in expression of genes whose products are necessary for amino acid transport and metabolism. As with mixed biofilm conditions, the two ribonucleotide reductase genes were upregulated while the majority of the rest of the nucleotide transport and metabolism ORFs were downregulated. In addition, there was a downregulation of a number of ORFs involved in translation, including 13 ribosomal proteins and a handful of tRNA synthesis proteins.

| There is differential expression of stressresponse genes and virulence factors when S. aureus is grown in mixed planktonic cultures
Consistent with mixed biofilm conditions, there was a downregulation of fnbB and coa in a mixed planktonic culture. Two ORFs predicted to encode enterotoxins (USA300HOU_0851 and USA300HOU_0852) and staphopain (USA300HOU_1910), a peptidase whose targets include elastin, were also downregulated under these conditions.
Contrary to mixed biofilm conditions, RNA-seq of mixed planktonic cultures revealed an upregulation of two superantigen-like proteins.
The gene encoding staphylokinase (sak) was also upregulated. Sak is known to have a role in allowing invasiveness in the skin and has recently been shown to inhibit biofilm formation (Kwiecinski et al., 2015), however a regulator promoting biofilm formation (rot) (Mootz et al., 2015) was also upregulated.
Two ORFs involved in the oxidative stress response (ahpC1, ahpF) (Poole et al., 2000) were downregulated, while genes respon-  (Sahl & Bierbaum, 1998), their production indicates a defensive strategy by S. aureus when grown in competition. Finally, there was upregulation of two genes implicated in antibiotic resistance, lrgA (which is also part of a holin/ antiholin system) (Groicher et al., 2000) and a tellurite transporter (Taylor, 1999).

| Specific ORFs upregulated under mixedspecies biofilm conditions contribute to competitive success of either S. aureus or P. aeruginosa
To determine the contributions of the ORFs upregulated in either P.
aeruginosa or S. aureus when grown under mixed-species biofilm conditions, transposon mutants deficient of these ORFs were grown in dual-species biofilms with wild-type competitors and colony-forming units of each species after 24 hr growth were assessed to generate a competitive index (CI). Before a CI could be obtained for the mutants, growth curves and biofilm assays were performed for each transposon mutant to ensure that any defect in competition could not be attributed to deficiencies in growth or ability to form biofilms. Any mutant found to be defective in growth or biofilm formation was not considered for future downstream analysis.
All the ORFs upregulated in P. aeruginosa in mixed-species biofilms listed in Table 2 were also important for either growth or biofilm development, and were thus unsuitable for further analysis through competition assays. We chose an upregulated gene, PA1595, which was upregulated fourfold (Table S2) in wild-type P. aeruginosa grown in mixed-species biofilms to characterize in terms of its contribution to competitive ability. PA1595 mutant displayed no growth defects.
PA1595 is a hypothetical protein predicted to localize to the outer membrane, with no identified conserved domains. The CI of a biofilm consisting of the PA1595 transposon mutant (strain PW3851) and wild-type S. aureus was 2.5 × 10 −2 , 2.5-fold higher than the CI of a biofilm consisting of parental strains (TCH1516/MPAO1; CI = 9.9 × 10 −3 , p < .01), while the CI of the clean PA1595 knockout (strain SLRK01) was 2.1 × 10 −2 (p < .05) (Figure 3) suggesting that the lack of PA1595 allows for S. aureus to be more abundant in a mixed-species biofilm and thus important for P. aeruginosa to compete with S. aureus in the biofilm niche.
The S. aureus transposon mutant library was generated in JE2 background (Fey et al., 2013). The CI between JE2 and P. aeruginosa was insignificant from the CI of MPAO1 and TCH1516 (CI = 9.26 × 10 −3 , p = .3). Staphylococcus aureus lukS-PV, a component of the Panton-Valentine leukocidin, were upregulated 2.3-fold when S. aureus was grown in a mixed-species biofilm. Compared to the CI of the parental strains, the transposon mutant lukS-PV (strain NE1848) was significantly impaired in their ability to compete with P. aeruginosa in a mixed biofilm as evidenced by lower CI value (3.4 × 10 −3 , p < .05), which is 2.7-fold lower than the CI of the parental strains ( Figure 3). These data suggest that leukocidin plays a role in the ability of S. aureus to compete with P. aeruginosa in mixedspecies biofilms.

| The inability of mutant S. aureus and P. aeruginosa to compete in a biofilm is not attributed to an inability to combat the competitor's exoproducts
We sought to characterize the mechanisms contributed by these ORFs leading to a decrease in the ability of the aforementioned mu-

| The mutant's lack of competitive success cannot be solely attributed to an inability to secrete exoproducts
We hypothesized that the defect in competition may be due to the mutants' inability to secrete exoproducts. Therefore, we tested the effect of the mutants' spent media on the biofilm viability of the wildtype competitor.

F I G U R E 3
Mutants corresponding to open reading frames upregulated in either Staphylococcus aureus or Pseudomonas aeruginosa during mixed-species biofilms were defective for competition in mixed-species biofilms. An overnight culture of P. aeruginosa or S. aureus was diluted into fresh 20% BHI++ and allowed to reach an OD 600 of 0.5, before seeding the bacteria at a 1:1 ratio for attachment and maturation into a 24-hr static mixed-species biofilm. Biofilms were detached by sonication and the cells were plated on selective media for enumeration. A competitive index (CI) was calculated as the S. aureus CFU/P. aeruginosa CFU output ratio divided by the input inoculum ratio. The PA1595 transposon mutant (strain PW3851) and clean knockout (strain SLRK01) were defective for competition allowing a tip in the CI favoring more S. aureus growth compared to the mixed-species biofilm established with wild-type strains (TCH1516/MPAO1). The lukS-PV (strain NE1848) transposon mutant derived from parental S. aureus strain JE2 was defective for competition with a CI ratio skewed to favor more P. aeruginosa growth compared to the mixed-species biofilm established with wild-type strains (JE2/MPAO1). Data represent the mean ± standard deviation from at least four independent experiments. Significance was determined by unpaired Student's t-test (one-tailed) (*p < .05; **p < .01) We saw strong inhibitory effects of P. aeruginosa spent media on S. aureus biofilm viability, but no significant difference in S. aureus viability when comparing spent media from wild-type P. aeruginosa or PW3851 mutant strain (Figure 4c). No significant difference in P. aeruginosa biofilm viability was found using spent media from either S. aureus mutant NE1848 when compared to spent media from wild-type S. aureus ( Figure 4d). Interestingly, we noted that PW3851 and SLRK01 produce significantly less pyocyanin and have a reduced ability to swarm ( Figure 5).
Collectively, our results suggest that the observed defects in competition cannot be attributed to the mutants' inability to secrete exoproducts.
F I G U R E 4 Defects in competition in mixed-species biofilms cannot be solely attributed to mechanisms of secretion. Contributions of secreted factors from either the mutant or wild-type strains on the formation of static biofilms of the competitor. Spent media supernatants were acquired from bacterial cultures grown overnight in 20% BHI++ at 37°C shaking, harvested by centrifugation, and filter sterilized. Strains were grown 24 hr at 37°C shaking in 96-well plate with BHI++ media supplemented with supernatants at a 1:1 ratio. Biofilm viability was determined using the PrestoBlue Cell Viability Reagent read at an excitation and emission wavelengths of 535 nm and 590 nm, respectively (a-d). Defect in competition of mutants is not attributed to the inability of the mutants to combat exoproducts secreted by the wild-type (either JE2 or MPAO1) competitor (a, b). Defect in competition of mutants is not attributed to the inability of the mutants to secrete exoproducts and inhibit the biofilm viability of competitor in a dual-species biofilm (c, d). Spent media supernatants were boiled as a control to assess if factors in spent media were protein based. Wild-type Staphylococcus aureus (JE2); S. aureus lukS-PV (strain NE1848); wild-type Pseudomonas aeruginosa (strain MPAO1); P. aeruginosa PA1595 transposon mutant (strain PW3851); P. aeruginosa PA1595 clean knockout (strain SLRK01)

| Mechanisms during close interactions contribute to the competitive defects observed with both S. aureus mutants
In order to determine if the lack of competitive success of the mutants could be attributed to mechanisms taking place when the two species are in direct contact, we performed cross-streak assays of mutant strains with wild-type competitors. By comparing the zone of interaction (black arrow) when the wild-type strains were cross hatched (Figure 6a), we noticed that when wild-type P. aeruginosa was grown with NE1848 (Figure 6b), the zones of interaction show more ingrowth of P. aeruginosa and an increased area of clearance, suggesting that contact inhibition may be one mechanism for the observed defects when the S. aureus mutant was grown with P. aeruginosa. When PW3851 and wild-type S. aureus were grown together, the zone of interaction appeared similar to that of the two wild-type strains (Figure 6c).
Using literature and databases of previously identified sRNAs, eliminating duplicate sRNAs (those with sequence overlap <25 nucleotides), and having a cutoff criteria (p > .05; fold change >2) demonstrated approximately 100 and 150 sRNAs were differentially expressed in S. aureus and P. aeruginosa, respectively, during mixedversus single-species biofilm growth. Interestingly, of these sRNAs in both P. aeruginosa and S. aureus approximately 25% were upregulated and 75% were downregulated (Table S4 and S5).

| Few characterized sRNAs, but many previously identified sRNAs were differentially expressed in P. aeruginosa mixed-versus singlespecies biofilms
Three sRNAs differentially expressed in P. aeruginosa during mixedversus single-species biofilms were previously characterized, with two of them having been functionally characterized (Table 4). ErsA (also annotated as SPA0122) was upregulated approximately sixfold in P. aeruginosa in mixed-species biofilms. ErsA is a sRNA whose transcription is strictly dependent on the envelope stress-responsive F I G U R E 5 Pseudomonas aeruginosa PA1595 mutants (PW3851/SLRK01) produce less pyocyanin and have a reduced ability to swarm compared to parental strain. Spent media supernatants of cultures of each strain grown overnight in 20% BHI ++ were extracted with chloroform and HCl and absorbance readings were taken at 520 nm for PYO quantification (a). Swarming motility was assessed for each strain by inoculating 3 μl of a 16 hr overnight culture onto swarm plates (0.5% w/v Bacto agar, 8 g/L of nutrient broth, and 0.5% w/v d-glucose) and allowed to incubate for 24 hr at 37°C (b). Wild-type P. aeruginosa (strain MPAO1); P. aeruginosa PA1595 mutant (strain PW3851); P. aeruginosa PA1595 (strain SLRK01). Data represent the mean ± standard deviation from four independent experiments. Significance was determined by paired Student's t-test (two-tailed) (*p < .05) sigma factor σ 22 (AlgT/U) which is essential for the production of the exopolysaccharide alginate (Ferrara et al., 2015). ErsA has been shown to target the algC transcript, is induced upon envelope stress, increases in temperature, and anaerobic conditions and is thought to influence the dynamics of exopolysaccharide biosynthesis underlying the development of biofilm matrix (Ferrara et al., 2015). RgsA (also known as P16) was downregulated almost fourfold in P. aeruginosa in mixed-species biofilms. RgsA's expression is dependent on RpoS, indirectly influenced by response regulator GacA (Gonzalez et al., 2008;Park et al., 2013), conserved across Pseudomonads (Livny, Brencic, Lory, & Waldor, 2006), upregulated in late stationary and biofilm culture conditions, and thought to promote resistance to oxidative stress (Dotsch et al., 2012).

| Expression of small RNAs in S. aureus involved in capsule formation, amino acid metabolism, and toxin production during biofilm growth with P. aeruginosa
RsaA (also annotated as Sau64 and Teg88) is a sRNA that represses translation of the MgrA regulator which activates biofilm production and inhibits capsule formation (Romilly et al., 2014). SprX (also annotated as Teg15 and RsaOR) is a sRNA that also represses capsule formation but through translational inhibition of the spoVG operon (Eyraud et al., 2014). RsaA and SprX are downregulated approximately two-and threefold, respectively, in S. aureus in mixed-versus singlespecies biofilms (Table 4). RsaE (also annotated as Teg92) is a sRNA that is downregulated 6.3-fold in S. aureus grown in mixed-species biofilms. RsaE regulates metabolic pathways and increases the amino acid pool by upregulating expression of valine, isoleucine, and leucine operons. SprA1 is a cytolytic peptide toxin that is regulated by a cis-antisense RNA, SprA1 AS (also annotated as Teg152) (Fozo et al., 2010;Sayed et al., 2012). Typically, SprA1 AS is generated in excess of SprA1 thus inhibiting the translation of this toxin (Sayed et al., 2012). When S. aureus is grown in a mixed-species biofilm SprA1 is upregulated approximately 1.5-fold, whereas SprA1 AS is downregulated approximately 1.5-fold suggesting an alleviation of translational repression by this antisense RNA and most likely subsequent translation of the SprA1 toxin.

| Pseudomonas aeruginosa the dominant pathogen: model mimics mixed-species infections seen in vivo
While it is interesting to study the planktonic interactions of these two species, the biofilm condition is more relevant medically because chronic biofilm infections significantly impair wound healing.
In contrast to the trend seen in mixed planktonic cultures where  the growth and success of S. aureus. Additionally, we have observed a more rapid doubling time of S. aureus in planktonic culture.
Combined with the abundance of nutrients, the ability to replicate more quickly may explain the ability of S. aureus to outcompete P. aeruginosa in high-energy environments. Overall, biofilms grown in 20% BHI++ and harvested from the drip-flow reactors resulted in bacterial burdens seen in in vivo studies and clinical infections and thus represented an appropriate model to study bacterial interactions in mixed-species biofilms using RNA-seq.

| Pseudomonas' competitive advantage through normal growth and biofilm development processes
Pseudomonas aeruginosa differentially expressed relatively few ORFs (~0.3% of the genome) when grown in a mixed-versus single-species biofilm, in contrast to S. aureus which had a more robust transcriptomic response showing nearly 150 up-and downregulated ORFs (~5% of the genome). A similar trend was observed in a recent study comparing P. aeruginosa strain PA14 and S. aureus strain 8325-4 in an in vitro cystic fibrosis (CF) model (Filkins et al., 2015). Therefore, P. aeruginosa appears to easily maintain itself as a dominant organism in various in vitro systems. Interestingly, most mutants lacking an ORF that was upregulated in P. aeruginosa in mixed-species biofilms were defective for growth or biofilm development and thus their contributions during mixed-species biofilms could not be assessed. Taken together, this suggested to us that P. aeruginosa may play a more passive role in mixed-species biofilms and its success appears to be a result of the ability to carry out normal growth and biofilm developmental processes.
Passive killing of competitors in the niche may be a result of intrinsic events required for biofilm development and maturation.
One of these intrinsic events is bacterial autolysis, a form of programmed cell death that releases lytic enzymes and bacteriocins to which competitors may be sensitive to (Michel-Briand & Baysse, 2002). Holin/antiholin systems are responsible for controlling autolysis which releases lytic enzymes and cellular material that is then utilized to build structural components of the biofilm, foster specific biofilm architecture, and promote dispersion (Ma et al., 2009). One of the few ORFs upregulated in P. aeruginosa during mixed-species biofilm growth is PA3431 and PA3432. PA3432 and PA3431 are lr-gA/cidA and lrgB/cidB homologs, respectively, and are holins or antiholins that can coordinate and control the exact timing of autolysis.
This autolysis may be triggered by nutrient deprivation, stress, or insults from a competitor. In our SEM analyses both monospecies P. aeruginosa and mixed biofilms showed an increased level of extracellular material when grown in 20% BHI++ under static conditions.
In contrast, there is little, if any, of this material present in S. aureus biofilms grown in 20% BHI++ or in P. aeruginosa biofilms grown in 100% BHI++, or in the drip-flow reactor. As 100% BHI++ and growth in the drip-flow reactor provide a rich growth environment, these results suggest that P. aeruginosa is responsible for making this extracellular material under nutrient limiting conditions, and this material may be a byproduct produced by autolytic events required for biofilm development. In addition, we have observed that P. aeruginosa biofilms are significantly more mucoid and will form at the air-liquid interface more readily than those produced by S. aureus, perhaps partly as a result of the extracellular material visualized using SEM (Figure 2). Ultimately, autolysis by P. aeruginosa may provide a competitive advantage through the inadvertent release of lytic enzymes and bacteriocins contributing to subsequent growth inhibition of S. aureus.
Numerous studies have shown that P. aeruginosa can actively inhibit or kill competitors in the niche, including S. aureus. Pseudomonas aeruginosa attacks its competitors indirectly through the production of secreted exoproducts and directly by injecting toxins into competitors by the Type III and Type VI secretion system (T3SS, T6SS) (Russell et al., 2013). Although Park et al. (2012) showed that P. aeruginosa secretes an antibiofilm protease encoded by lasB against S. aureus (Park et al., 2012), we did not see lasB upregulated in P. aeruginosa when grown in the presence of S. aureus. Also, in P. aeruginosa in mixed-species biofilms there was no significant change in expression of T6SS ORFs and exoS, a Type III secreted toxin, was downregulated which further suggested that in our study P. aeruginosa did not mount a defensive response against S. aureus during biofilm growth and competitive advantage might be obtained by carrying out normal growth and biofilm process.

| Contributions of a novel P. aeruginosa ORF during competition with S. aureus
Our studies did reveal a novel ORF that provides a competitive advantage to P. aeruginosa in mixed-species biofilms with S. aureus.
PA1595 is a hypothetical protein predicted to be in an operon with two previously uncharacterized ORFs (PA1593 and PA1594). Cluster of orthologous group (COG) predictions of both PA1593 and PA1594 demonstrated significant similarities to HGG motif-containing thioesterase proteins and possible involvement in aromatic compound catabolism. Conserved domain database (CDD) predictions further suggested that both PA1593 and PA1594 contain domains indicative of phenylacetic acid (PAA) degradation. PAA serves as an antipathogenic factor released at stationary phase and plays a role in inhibiting quorum sensing. Musthafa, Sivamaruthi, Pandian, and Ravi (2012) demonstrated that exogenous addition of PAA leads to poor swimming motility and reductions in QS-dependent pyocyanin, exopolysaccharide, protease, and elastase production in PAO1 (Musthafa et al., 2012). Our results demonstrate that a PA1595 transposon mutant strain (PW3851) and a clean knockout (SLRK01) were both defective during competition with S. aureus. Neither the supernatant of MPAO1, PW3851, or SLRK01 had a significant effect on S. aureus biofilm viability, but interestingly both PA1595 mutant strains, PW3851 and SLRK01, produced significantly less pyocyanin and had a reduced ability to swarm. Although PAA levels were not assessed in our study, the role of PA1595 is most likely dynamic and our work begins to unravel its contributions for success of P. aeruginosa in mixed-species biofilms.

| Staphylococcus aureus has a robust genomic response to growth with P. aeruginosa suggestive of increased virulence and capsulation, with a concomitant decrease in biofilm formation
Despite S. aureus being significantly outcompeted by P. aeruginosa, it is well known that S. aureus is a particularly virulent bacterium, able to cause severe disease, impair wound healing, and has numerous virulence factors (Pastar et al., 2013;. Similar to previous reports, S. aureus upregulates a few of these virulence genes in response to being grown with P. aeruginosa, including both ORFs encoding leukocidin (lukS-PV and lukF-PV) and α-hemolysin (hla) (Pastar et al., 2013).
Contrary to a previously published report that saw a downregulation of spA in a mixed-species biofilm in vivo (Pastar et al., 2013), we saw significant upregulation of spA in our in vitro conditions and these differences in expression may be a result of disparities in experimental conditions. However, our global study allowed insight into myriad of other factors at play and interestingly, several S. aureus genes known to be important for biofilm formation including fibronectin-binding proteins A and B (FnbAB), clumping factor B (ClfB), and coagulase (Coa) (Abraham & Jefferson, 2012;Friedrich et al., 2003;McCourt et al., 2014;Mulcahy et al., 2012;O'Brien et al., 2002;Zapotoczna et al., 2015) were downregulated when S. aureus was grown in a biofilm with P. aeruginosa.
Also, when S. aureus was grown in a biofilm with P. aeruginosa, 13% and 45% of the downregulated ORFs were those involved in purine and pyrimidine biosynthesis and amino acid transport and metabolism, respectively, while 15% of the upregulated ORFs were involved in capsular biosynthesis, though it has been shown that these strains of S. aureus do not actually form a capsule due to mutations in various capsule ORFs (Boyle-Vavra et al., 2015). As shown by Filkins et al., 2015; the presence of P. aeruginosa exoproducts can drive S. aureus from aerobic to lactic acid fermentation (Filkins et al., 2015). We noted similar upregulation of fermentation related ORFs, including pflAB, ldh, and adh. Taken together, it is plausible that S. aureus increases production of specific virulence factors, downregulates biofilm formation, relies on alternative respiration in order to coexist in this competitive niche with the dominant pathogen, P. aeruginosa.

| Contributions of specific S. aureus ORFs during competition with P. aeruginosa
As previously mentioned, S. aureus is able to successfully colonize biotic and abiotic surfaces and persist in chronic wound environments by avoiding killing by the host immune system. Many S. aureus strains produce a polysaccharide capsule that promotes abscess formation and prevents phagocytosis by polymorphonuclear leukocytes (PMNs) (Herbert et al., 1997;Karakawa, Sutton, Schneerson, Karpas, & Vann, 1988;Kiser, Cantey-Kiser, & Lee, 1999). Of the ORFs upregulated in S. aureus in mixed-versus single-species biofilms, 15% corresponded to type 5 capsule (cap5EF, cap5H-P). Although this strain is known to not form a capsule due to specific mutations in various capsule ORFs (Boyle-Vavra et al., 2015), others strains do, and it is possible that upregulation of capsule could serve as a protective mechanism for those strains when they are faced with adverse conditions such as host immune responses and competition with other microbial species.
Panton-Valentine leukocidin (PVL) is an important virulence factor implicated in severe pneumonia and necrotizing fasciitis (Boyle-Vavra & Daum, 2007) and is composed of two subunits (lukF-PV and lukS-PV) that combine to form pores, particularly in the membranes of PMNs.
While ORFs encoding PVL are upregulated in mixed infections (Pastar et al., 2013), it is unknown what role PVL plays in S. aureus' competitive success during mixed-species growth. We saw that a lukS-PV transposon mutant strain (NE1848) was defective for competition with P. aeruginosa and sought to identify if PVL was harmful to not only host PMN cells, but also bacterial cells. Subjecting P. aeruginosa to spent media isolated from NE1848 showed no significant change in biofilm viability, though growing NE1848 in cross-hatch assays with P. aeruginosa suggesting that PVL may play a role in S. aureus' competitive success when S. aureus is in close proximity to a competitor such as P. aeruginosa.
RNA-seq and analysis of ORFs makes it clear that S. aureus responds strongly when grown in the presence of P. aeruginosa. While CI numbers show that P. aeruginosa outcompeted S. aureus, S. aureus mounts a robust genetic response, upregulating various virulence factors and converting to lactic acid fermentation while downregulating ORFs involved in basic metabolism and biofilm formation. Increased virulence and capsulation (for capsulated strains) are likely strategies that S. aureus adopts in order to prevent its total elimination from a mixed-species biofilm.

| sRNAs parallel genomic response in mixedspecies biofilms
Because sRNAs can target mRNAs and regulate their translation without influencing the transcriptional output, we speculate that this phenomenon may be occurring in P. aeruginosa to dictate a phenotype providing a competitive advantage in mixed-species biofilms. ErsA was upregulated in P. aeruginosa during mixed-species biofilm growth and is a sRNA that is induced upon envelope stress. It is interesting to speculate how ErsA may help counteract any agents S. aureus is producing. On the other hand, ErsA is involved in regulating the dynamic state of the exopolymeric substance and may be expressed to coordinate efforts required for typical biofilm maturation. We suspect from the ORF RNA-seq analyses that P. aeruginosa is minimally affected by the presence of S. aureus after the establishment of a biofilm. In support of this, RgsA, a sRNA that has been shown to help resist against oxidative stress, was downregulated in P. aeruginosa in mixed-species biofilms. Nevertheless, numerous uncharacterized sRNAs were differentially expressed and our study begins to help define their roles in the pathobiology of P. aeruginosa.
RsaA and SprX are two small RNAs known to repress capsule formation and were downregulated in S. aureus in mixed-versus singlespecies biofilms. This is consistent with our RNA-seq analyses of ORFs because approximately 15% of upregulated ORFs in S. aureus in mixed-versus single-species biofilms were those involved in capsule formation. 80% of S. aureus isolates recovered from infected patients are encapsulated (Cocchiaro et al., 2006), therefore capsules must also be an important tool against or protected from bacterial competitors such as P. aeruginosa. Furthermore, RNA-seq assessing small RNAs demonstrated that S. aureus induces expression of a type I toxinantitoxin (TA) module (SprA1/SprA1 AS ) when grown as a biofilm in the presence of P. aeruginosa. SprA1 toxin's translation is repressed by the cis-antisense RNA, SprA1 AS . SprA1 was expressed approximately 10 times higher than SprA1 AS in S. aureus in mixed-species biofilms suggesting that this toxin is being expressed. SprA1 is a cytolytic peptide that is autolytic, kills eukaryotic cells, as well as Gram-positive and -negative bacteria (Fozo et al., 2010;Sayed et al., 2012). Autolytic TA systems are involved in enriching the population of persister cells because metabolically dormant cells are refractory to the toxins whereas rapidly growing siblings are sensitive. This suggests that S. aureus may undergo this altruistic behavior in effort to coexist and combat P. aeruginosa in mixed-species biofilms. Finally, we saw a downregulation of RsaE, a sRNA regulator of metabolism which activates amino acid biosynthesis, and is most likely contributing to the decrease in amino acid biosynthesis pathways indicated by our RNA-seq analyses of ORFs.
Collectively, our results suggested that the sRNA profile closely mimics the trends seen when the ORFs were analyzed, and expression of small RNAs appears to steer S. aureus' phenotype toward capsule formation, decreased amino acid metabolism, and production of toxins that promote persistence during biofilm growth with P. aeruginosa.

| CONCLUSION
The advent of deep sequencing (RNA-seq) has allowed researchers to identify the physiological contributions of regulatory small RNAs and thus our work utilized RNA-seq to identify the entire genomic response, both ORFs and small RNAs, required for the success of each pathogen in a mixed-species biofilm. Our methods aimed to recapitulate the dynamic nature of a mixed-species biofilm seen in in vivo infections.
Based on our results we speculate that during an acute (planktonic) mixed-species infection, there is a robust response or active combat from both pathogens until a state of equilibrium is reached during the maturation of a chronic infection (biofilm). Ultimately, one organism dominates in the environment (P. aeruginosa) and proceeds by carrying out the typical operations necessary for establishing a biofilm community crowding out, but not completely eliminating its competitor (S. aureus). It is likely that combating these infections may necessitate a reactivation or tip in this equilibrium. Also, our work suggests that sRNAs are coordinating with the genomic response to carry out the adaptive changes necessary to attain a homeostatic mixed-species biofilm. Grant # NIH P30 DK089507 supported the production of the P. aeruginosa transposon mutant library (Jacobs et al., 2003) used in this study. RNA sequencing was performed by SeqWright Genomic Services, Houston, Texas.