Antibiotic resistance alters through iron-regulating Sigma factors during the interaction of Staphylococcus aureus and Pseudomonas aeruginosa

Iron is a limiting factor in such a condition that usually is sequestered by the host during polymicrobial infections of Pseudomonas aeruginosa and Staphylococcus aureus. This study aimed to investigate the interaction of S. aureus and P. aeruginosa, which alters iron-related sigma factors regulation and antibiotic resistance. The antibiotic resistance of P. aeruginosa and S. aureus was investigated in a L929 cell culture model. The expression level of pvdS, hasI (P. aeruginosa sigma factors), and sigS (S. aureus sigma factor) genes was determined using Quantitative Real-Time PCR. pvdS and hasI were downregulated during co-culture with S. aureus, while the susceptibility to carbapenems increased (p-value < 0.0001). Also, there was a direct significant relationship between resistance to vancomycin with sigS. Regarding the findings of the current study, iron-related sigma factors of P. aeruginosa and S. aureus play a role in induction susceptibility to various antibiotics, including carbapenems and vancomycin.

(KPC)-acquired through transferable genes, are classified in class A of Ambler classification system. Moreover, overproduction of efflux pumps such as MexAB-OprM and reduced expression of porins including OprD lead to resistance to carbapenems [15][16][17] .
Vancomycin is an essential treatment for staphylococcal infections. Resistance to this drug becomes a critical issue in recent years 18 . Vancomycin-resistant and vancomycin-intermediate S. aureus (VRSA and VISA) are controlled through the vanA gene transferred on plasmid and missense mutations in walk/R genes, respectively 19,20 . In addition, fluoroquinolones are a suitable choice as treatment of staphylococcal infections. S. aureus resists against this class of antibiotics through over-expression of efflux pumps (norA gene) and mutation in topoisomerase IV encoded by gyrA/B and grlA/B 21,22 .
It has been reported that iron metabolism plays a central role in antibiotic resistance in Escherichia coli 23 ; however, the relationship between iron concentration and antibiotic resistance in polymicrobial infections is unclear. Moreover, the role that ECF sigma factors play in antibiotic resistance is vague.
Therefore, this study aimed to determine how the iron-related ECF sigma factors would alter during S. aureus and P. aeruginosa interaction and how this alteration influenced the antibiotic resistance of persisters and wild-type isolates.

Results:
Coexistence of S. aureus and P. aeruginosa: inhibitory or stimulatory effect. A cell line-based model was developed to investigate the viability of S. aureus (SA-1) in co-culture with four clinical strains of P. aeruginosa (PA-1, PA-2, PA-3, and PAO1) in wound infections ( Table 3). As illustrated in Fig. 1, P. aeruginosa caused a significant decrease in the viable colony counts of S. aureus recovered from the planktonic and biofilm www.nature.com/scientificreports/ states compared to monoculture. During the first hour of co-culture, a reduction in viability was detected for S. aureus, which was more remarkable in SA-1/PA-1 and SA-1/PAO1 co-cultures. The S. aureus viability in SA-1/ PA-1 and SA-1/PAO1 in the planktonic and biofilm states decreased to approximately half of the monoculture (Fig. 1a,b). PA-1 belongs to sequence type 111(ST-111), a high-risk clone producing different toxins and phenazines. Compared to PA-2 and PA-3 (belong to ST-235), a more remarkable killing effect was observed in the combination of SA-1/PA-1. In contrast, a stimulatory effect was observed in the co-culture of PA-2 and S. aureus. To clarify, the viable colony counts of S. aureus in SA-1/PA-2 co-culture almost reached the monoculture state (Fig. 1a,b). Interestingly, no significant decrease in viability was detected in S. aureus when co-cultured with PA-3 in both planktonic and biofilm conditions. In comparison to the planktonic state, the viability of S. aureus more notably decreased in the biofilm condition.
As depicted in Fig. 2, the viability of PA-1 and PAO1 did not change compared to the monoculture. The viable colony counts in both biofilm and planktonic forms indicated a negligible effect of S. aureus on P. aeruginosa viability. Contrary to PA-1/SA-1 and PAO1/SA-1, the viability of PA-2 and PA-3 reduced in comparison to monoculture. Remarkably, following co-culture with S. aureus, the recovered P. aeruginosa strains demonstrated a highly mucoid phenotype in which the phenazine-producing strains converted to a non-phenazine producer.
S. aureus strain converted to a slow-growing phenotype during co-culture. Notably, following the co-culture with P. aeruginosa, the growth rate of S. aureus colonies reduced significantly. Although P. aeruginosa strains grew naturally at 37 °C and in ambient air after overnight incubation, S. aureus indicated a slow growth. In other words, slow-growing, tiny, non-hemolytic colonies of S. aureus were recovered five days postplating. The diluted samples were plated on Columbia agar supplemented with 5% sheep blood and incubated at 37 °C and 5% CO 2 , while the monoculture of S. aureus was recovered typically at 37 °C and ambient air after 18-24 h. Also, mannitol consumption was restrained in these phenotypes during the growth on MSA. Also, auxotrophy to menadione, thymine, and hemin was investigated on the persister strains, but no auxotrophy was detected.
The activity of iron regulating ECF sigma factors altered during co-culture. Iron regulating ECF sigma factors of P. aeruginosa. Iron is tightly regulated in co-culture conditions due to its critical role in the adaptation of S. aureus and P. aeruginosa. Therefore, the ECF sigma factors were investigated in a co-culture model. As illustrated in Fig. 3, the expression level of hasI and pvdS decreased significantly in PA-1/SA-1 and PAO1/SA-1 combinations compared to PA-1 and PAO1 monocultures. Meanwhile, the expression level of hasR (a receptor gene to sigma factor hasI) and pyoverdine production reduced remarkably. Although hasI downregulated to two-fold in the planktonic state, a more than ten-fold decrease was observed in the biofilm state in PA-1 and PAO1. A similar downregulation (a two-and five-fold decrease in the planktonic and the biofilm states, respectively) was observed in pvdS, which lead to a reduction in pyoverdine production in the two mentioned strains. In contrast, pvdS and hasI expression levels increased to approximately ninety percent in the biofilm state, whereas a half increase was detected in the planktonic conditions of PA-2 and PA-3.
The siderophore production and ECFs downregulation occurred in consistence to increased killing of S. aureus by PA-1 and PAO1 (Figs. 3 and 5a). During the co-culture of PA-1 and PAO1 with S. aureus, pyoverdine production decreased compared to monoculture. Also, hasR downregulated due to iron boost in the co-culture media (Fig. 3a,b). Conversely, the pyoverdine production indicated a non-significant increase in PA-2 and PA-3 in comparison to monoculture conditions and the wild-type strain (Fig. 5a). Compared to the monoculture findings and gene expression level in P. aeruginosa strains in iron-rich and -starved media, the remains of S. aureus were used as iron sources by P. aeruginosa.
Iron regulating ECF sigma factor of S. aureus. sigS is the only known ECF sigma factor in S. aureus. Contrary to P. aeruginosa, the role of sigS in iron regulation has not been investigated; however, it may play a role in Fenton's reaction through iron regulation. Therefore, the expression level of sigS was observed in the co-culture of S. aureus with different strains of P. aeruginosa.
As depicted in Figs. 4, sigS upregulated in the co-culture with PA-1 and PAO1. The expression level of sigS increased two-and three-fold in the planktonic and biofilm forms compared to monoculture. Interestingly, the siderophore production decreased in the co-culture with PA-1 and PAO1 in both planktonic and biofilm states. Regarding the considerable killing effect of the two mentioned strains on S. aureus and slow-growing isolates recovered from co-culture, it seems that the iron starvation leads to sigS upregulation. Therefore, the expression level of sigS was investigated in iron-rich (> 1 µM) and -starved (< 0.5 µM) media, and the findings mentioned above were confirmed (Fig. 4).
Despite the upregulation of sigS in the co-culture with PA-1 and PAO1, a slight increase was observed in SA-1/ PA-2 and SA-1/PA-3. Moreover, the siderophore production indicated a non-remarkable decrease in the SA-1/ PA-2 and SA-1/PA-3 combinations compared to monocultures (Fig. 5b). Although a coexistence relationship was observed between SA-1 and PA-2 and PA-3, the expression level of sigS increased and recovered isolates of S. aureus slowly grew. Consider the significant increase of siderophores in co-culture with PA-2 and PA-3 in the iron-rich medium; it is suggested that despite coexistence, SA-1 competes for nutrients with the biofilm-forming and susceptible strains of P. aeruginosa.
Antibiotic resistance had a relationship with ECF sigma factors. Iron  www.nature.com/scientificreports/ hypothesis of whether the antibiotic resistance is influenced by co-culture or not. Changes of different antibiotics categories in P. aeruginosa were listed in Table1. Carbapenems, including imipenem, meropenem, and doripenem, showed various changes after co-culture with S. aureus. To illustrate, the MIC of meropenem and doripenem increased in the PA-1 and PAO1, while the imipenem resistance level decreased remarkably in PA-1 (MIC: 2 µg/mL) and increased in PAO1. The expression level of genes encoding resistance to carbapenems, including kpc, efflux pumps (mexA-mexB-oprM), and oprD was investigated to indicate the reason for this contradiction. The expression level of kpc decreased more than 100-fold in PA-1, whereas mexA-mexB-oprM upregulated (Fig. 6a,c). As well, oprD downregulated in PAO1 (Fig. 6b). Unlike PAO1, oprD upregulated in PA-2, while the expression level of kpc upsurged in PA-3 (Fig. 6). The MIC of carbapenems decreased in PA-2, whereas a significant increase was detected in PA-3.
The correlation between ECFs and antibiotic resistance was investigated, and as demonstrated in Table 1, there is a strain-dependent relationship between the expression level of ECFs and kpc, mexA-mexB-oprM, and oprD. In other words, a significant direct relationship was observed between kpc downregulation, oprD upregulation, and ECFs' decreased expression level. No association was identified between ECFs and resistance to meropenem and ciprofloxacin regarding the high expression level of mexA-mexB-oprM and the MIC amounts in all strains.
In addition, the resistance level was examined in an iron-rich media. As the medium was supplemented with iron, a similar reaction was observed in SA-1/PA-2 and SA-1/PA-3 combinations, too. Likewise, the apparent relationship between ECF and kpc downregulation was detected in monocultures of P. aeruginosa in the ironrich medium. Interestingly, the viability of S. aureus augmented two-fold in combination with PA-1 and PAO1 in the iron-rich medium.
As mentioned above, S. aureus strain converted to a slow-growing phenotype. The antibiotic susceptibility was investigated in this phenotype. As shown in Table 2, the MIC of ciprofloxacin increased from 0.5 to 64 µg/ mL after co-culture with all four strains of P. aeruginosa. The expression level of norA as an efflux pump involves resistance to ciprofloxacin was checked. According to Fig. 6, norA upregulated in S. aureus after co-culture. However, no changes were detected in the expression level of gyrA/B and grlA/B in comparison to monoculture.
NorA overproduction and resistance to ciprofloxacin were observed in both iron-rich and -starved media; however, to a lesser extent in the former. Unlike the co-culture conditions, norA was overexpressed in the monocultures in the iron-starved medium, while such reaction was not detected in monocultures of the ironrich medium.
Moreover, a notable resistance to vancomycin was detected in S. aureus when co-culture with P. aeruginosa strains. The MIC of vancomycin upsurged from 0.5 to 256 µg/mL after co-culture. The mechanisms for vancomycin resistance, including vanA, vanS, and walk/R, were investigated to determine the reason for the MIC escalation. SA-1 did not possess the vanA gene and van operon as examined by the PCR method. As depicted in Fig. 6, the expression level for walk/R increased significantly during co-culture compared to monoculture.
Resistance to vancomycin in SA-1 was studied in the iron supplemented medium, and a heterogeneous population of S. aureus was recovered from co-culture. Some vancomycin-sensitive isolates grew normally after 18-24 h, while some isolates slowly recovered after five days and were completely resistant to vancomycin. Although a slight increase in sigS expression level was detected in the first group, a remarkable upsurge in the expression of sigS and walk/R was observed in the second group in comparison to monoculture.
The correlation between sigS and resistance to ciprofloxacin and vancomycin was explored. According to Fig. 6 and Table 2, a significant direct relationship was noticed between sigS and vancomycin resistance due to walk/R overexpression. Also, norA overexpression and sigS expression levels were increased consistently and indicated a significant statistical association.

Discussion
The essential content of iron in bacteria is 10 -7 -10 -5 M; however, the host maintains its extracellular iron content as 10 -8 M. Therefore, it is necessitated for the pathogens to acquire iron using different mechanisms, including siderophores, heme assimilation systems, and ferric iron uptake systems 24 . During coinfection, the competition for micronutrients, including iron, determines the type of relationship between S. aureus and P. aeruginosa. Moreover, S. aureus aggregates and attaches to the cell substratum and initiates the biofilm formation process. The early biofilm formed by S. aureus contributes to P. aeruginosa to co-aggregation and development of dual-species biofilms 25 . As depicted in Figs. 1 and 2, the viability of S. aureus varied in co-culture with different strains of P. The viable colony counts of PA-2 recovered from the planktonic and biofilm conditions in monoculture and co-culture with SA-1. (c) The viable colony counts of PA-3 recovered from the planktonic and biofilm states in monoculture and co-culture with SA-1. (d) The viable colony counts of PAO1 recovered from the planktonic and biofilm conditions in monoculture and co-culture with SA-1. (e) The phenazine production was inhibited during co-culture with SA-1. PA-1 produced green pigment before co-culture with SA-1, while the pigment production was inhibited during the co-culture with SA-1.  www.nature.com/scientificreports/ aeruginosa. As SA-1 co-cultured with PA-1 (ST111) and PAO1, the viability reduced to half of the monoculture. During co-culture, P. aeruginosa secrets LasA protease and HQNO (2-heptyl-4-hydroxyquinoline-N-oxide) to kill S. aureus and obtain iron using siderophores 26,27 . LasA protease production increased in P. aeruginosa strains recovered from co-culture compared to the monocultures (data not shown). Moreover, the pyoverdine production dramatically decreased in PA-1 and PAO1 as co-cultured with SA-1 (Fig. 5). P. aeruginosa and S. aureus co-infection provide an iron-rich environment for the former, leading to decreased pyoverdine production 28 .
Regarding the decreased expression level of pvdS and hasI, it is suggested that PA-1 and PAO1 used SA-1 as an iron source. Mashburn and et al. reported that during co-culture with S. aureus, P. aeruginosa downregulated the iron uptake genes due to the presence of a sustainable source of iron 28 . The ECF sigma factors play different roles in the survival and fitness costs of the bacteria. As iron sources decreased in the environment, the anti-sigma factor protein degrades, therefore the expression levels of hasI and pvdS increase. Consequently, the production of HasR (a corresponding receptor of hasI) and pyoverdine would be augmented 29,30 . The ECF downregulation, HasR and pyoverdine decrease in the co-culture of S. aureus with P. aeruginosa were observed in the iron-rich medium. Figure 5. Siderophore production (a) P. aeruginosa and (b) S. aureus in the monoculture, the planktonic and biofilm states of co-culture. Each data set was analyzed using the student's t-test, and the Holm-Sidak method for multiple comparison. The data were presented as Mean + SEM. *p-value < 0.05; **p-value < 0.01; ***p-value < 0.001; ****p-value < 0.0001. Table 1. Antibiotic susceptibility of recovered P. aeruginosa strains. *P-value < 0.01; **P-value < 0.001; ***P-value < 0.0001.  www.nature.com/scientificreports/ Contrary to the high killing ability of PA-1 and PAO1, no significant decrease in viability was observed in SA-1 co-cultured with PA-2 and PA-3. PA-1 strain belonged to ST-111, while PA-2 and PA-3 belonged to ST-235. Although both STs are hypervirulent, high-risk clones, it is suggested that the characteristics of the strains, including toxin production, played an essential role in the relationship of SA-1 with PA-2 and PA-3 31,32 . As demonstrated in the results section, PA-2 and PA-3 indicated a negligible effect on SA-1 viability, and also ECFs downregulated. The pyoverdine and HasR production increased slightly. The viability of SA-1 in the co-culture with PA-2 and PA-3 did not differ significantly in the iron-rich and -starved media. Therefore, it seems a straindependent killing behavior caused the coexistence of SA-1 with PA-2 and PA-3.

Antibiotics
ECFs play various roles in antibiotic resistance, virulence, and metabolism. The relationship between antibiotic resistance and P. aeruginosa ECFs was investigated during co-culture with S. aureus. Moreover, genes encoding resistance to carbapenems were investigated. Resistance to carbapenems often occurs due to carbapenemases (encoding by kpc) preferably against imipenem, porins (encoding by oprD) against imipenem and doripenem, and efflux pumps (maxA-mexB-oprM) against meropenem. As depicted in Fig. 3, the expression level of hasI and pvdS decreased in staphylolytic strains of P. aeruginosa. Iron boost in the environment due to the lysis of S. aurues led to a decrease in kpc expression level, while maxA-mexB-oprM and oprD were overexpressed. Imipenem resistance preferably occurs due to the loss of oprD and carbapenemases (enzymes encoding by kpc), whereas maxA-mexB-oprM efflux pumps cause resistance to meropenem. PA-1 and PA-3 strains possess the kpc gene, and after co-culture with S. aureus, a significant decrease in kpc expression level was observed in PA-1 while PA-3 showed upregulation. The MIC of imipenem in PA-1 and PA-3 confirmed that iron limitation leads to kpc upregulation. At the same time, maxA-mexB-oprM upregulated to tenfold in PA-1. Also, the expression level of maxA-mexB-oprM increased in PAO1. MIC of meropenem increased from 32 and 0.5 to 64 for PA-1 and PAO1. The norA and walk/R changes in S. aureus strains in the planktonic and biofilm forms of the co-and monoculture. Each data set was analyzed using the student's t-test, and the Holm-Sidak method for multiple comparison. The data were presented as Mean + SEM. * p-value < 0.05; **p-value < 0.01; ***p-value < 0.001; ****p-value < 0.0001. Table 2. Antibiotic susceptibility of recovered S. aureus strains. * P-value < 0.01; **P-value < 0.001; ***P-value < 0.0001.  www.nature.com/scientificreports/ PA-2 does not possess kpc gene, but a significant decrease in the carbapenems' MIC was detected in this strain.
The expression level of oprD as an import gate for carbapenems increased remarkably due to iron limitation. Unlike previous reports that described a direct relationship between the iron limitation and maxA-mexB-oprM overexpression, our findings indicated that S. aureus beyond the role as an iron source might affect the regulation of antibiotic resistance 33 . Moreover, the MIC of ciprofloxacin elevated from 16 to 512 in consistency with maxA-mexB-oprM overexpression. As reported by Ankley, the iron chelation lead to a decrease in resistance to antimicrobial compounds 34 . Moreover, iron plays an important role in the early stages of biofilm formation, which indirectly may influence antibiotic resistance 35 . Overexpression of ferric reductase leads to antibiotic mediated cell death through the Fenton reaction in P. aeruginosa. Such an effect was induced abundantly in exposure to gentamycin, norfloxacin, tetracycline, and ampicillin 13 . sigS as the only ECF identified in S. aureus plays different roles in survival during starvation, stresses caused by DNA and cell wall damages, and oxidative stresses 36 . The expression level of sigS increased during the co-culture with P. aeruginosa strains. Although the survival of SA-1 in combinations with PA-2 and PA-3 were not influenced extensively, the metabolism and growth were affected. The recovered isolates of SA-1 after co-culture slowly grew and lost the ability for hemolysis, mannitol consumption, and pigment production. Different studies reported that P. aeruginosa lyses S. aureus during co-culture to provide iron and nutrient to grow. Moreover, P. aeruginosa acts as a stress agent for S. aureus and leads it to persistence 25,37 . According to Fig. 4, the siderophore production decreased in co-culture with P. aeruginosa strains, particularly PA-1 and PAO1. Moreover, sigS expression level reduced in SA-1 co-cultured with P. aeruginosa in an iron-rich medium.
Based on Table 2, the MIC of ciprofloxacin increased from 0.5 to 8 µg/mL in SA-1 and norA as an encoding gene for NorA efflux pump, overexpressed in both co-culture states. Iron limitation leads to norA overexpression and increased resistance to ciprofloxacin in S. aureus [38][39][40] . The overexpression of norA occurs due to the mgrA regulatory effect, which causes resistance to ciprofloxacin and resistance to vancomycin through an unknown mechanism 41 . sigS expression contributes to S. aureus to persist in starvation condition and confer resistance against DNA damages 42 .
Interestingly, the slow-growing isolates of S. aureus remarkably became resistant to vancomycin. The MIC of vancomycin upsurged from 0.5 to 512 µg/mL after co-culture. Although SA-1 did not possess van operon, the expression level of walk/R increased significantly after co-culture with P. aeruginosa. Reduced vancomycin susceptibility occurs due to inactivation of walk/R two-component system, responsible for cell wall synthesis in S. aureus 43,44 . As Miller and et al. reported, sigS mutants were more susceptible to cell-wall targeting antibiotics and DNA damaging agents, including ciprofloxacin; it is suggested that sigS overexpression in co-culture with P. aeruginosa is related to walk/R inactivation and decreased sensitivity to vancomycin 42 . Also, sigS upregulation concomitant with H 2 O 2 production leads to a reaction with intracellular iron (Fenton reaction), DNA damage, and cell death 10,11 . The Fenton reaction contributes to cell death induced by antibiotics in bacteria 12,13 . Oxidative stress might play a role in resistance to ciprofloxacin in ΔkatA (mutation in genes encoding catalase) mutants; and therefore, the iron assimilation genes would be silenced to defend bacteria against oxidative stress 45 .
Iron as a critical factor regulates resistance to different antibiotics. During co-culture, iron metabolism altered because of competition between P. aeruginosa and S. aureus, resulting in changes in antibiotic resistance. The ECF sigma factors play a role in regulating iron and, consequently, influence antibiotic resistance of the infection strains. Although a direct relationship between sigS and antibiotic resistance was observed in the current study, it is necessary to conduct more studies on the exact association between resistance and ECF sigma factor in S. aureus during co-culture.

Materials and methods
Study design. One clinical strain of S. aureus and three clinical strains of P. aeruginosa were selected based on their characteristics, mentioned in Table 3. The strains were obtained from the microbial bank of the Microbiology Laboratory of Hamadan University of Medical Sciences. The clinical strains were chosen regarding their characteristics, including biofilm formation, toxin production (T1SS, T2SS, and T3SS secretion systems for P. aeruginosa, Panton-Valentine Leukocidin (PVL) and alpha toxins for S. aureus), iron-related sigma factors, and antibiotic susceptibility. Also, the clinical strains of P. aeruginosa were chosen regarding their phenotypic and genotypic characteristics (based on molecular detection of virulence and resistance genes, quantitative measure- Co-culture condition. Mono-and co-culture assessment of the bacteria were done on L929 monolayer based on Dehbashi et al. 8 . Briefly, the monolayer of L929 was prepared in 24-wells cell culture plates. Then, the DMEM medium was removed, and the bacterial cultures (in exponential phase) were washed in PBS. 100 µL of bacterial suspension with OD 600 :0.1 in 1 mL MEM containing l-Glutamine was added to the wells as co-culture (each S. aureus and P. aeruginosa strain in each well). The plates were incubated at 37 °C and 5% CO2. In 1, 6, 12, and 24 h intervals, the media were aspirated, diluted, and plated on MSA and CA. Then the fresh medium was added. After 24 h incubation, to recover the S. aureus and P. aeruginosa strains, the planktonic bacteria were diluted in fresh PBS and cultured on MSA, CA, BHI, and Columbia agar. Then, following twice washing with PBS, 200 µL of 0.1% Triton X-100 was added to each well, and gently agitated for 30 min. The cells were scraped by a cell scraper to disrupt the biofilm. Then, the bacteria were diluted and plated as described for the planktonic co-culture. Each experiment was done in triplicate.
Antimicrobial susceptibility. The antimicrobial susceptibility of recovered strains of S. aureus and P. aeruginosa were tested based on CLSI 2019 for different categories of antibiotics, including beta-lactams, aminoglycosides, fluoroquinolones, and carbapenems using E-test strips (Liofilchem, Italy). The categories of antibiotics were selected based on the clinical guidelines of infections' treatments. Moreover, the antimicrobial susceptibility of S. aureus and P. aeruginosa in the co-culture was investigated. Five concentrations of each antibiotic were added to the infected cell monolayer, and in 1, 6, 12, and 24 h intervals, the medium was replaced with fresh MEM + antibiotics, and then the samples were plated as described in the past section. All the tests were done in triplicate.
Siderophore production. The spectrophotometry method based on El-fouly et al. was used to investigate siderophore production 46,47 . Briefly, P. aeruginosa strains were inoculated to RPMI1640 (Invitrogen, USA) and incubated at 37 °C by shaking 100 rpm overnight. The OD 600nm of the cultures was measured. After centrifugation at 200 g for 30 min, the supernatants were collected and filtered by 0.22 µm Millipore filters (Merck, Germany). The OD 405nm of supernatants was measured spectrophotometrically, and then Relative Pyoverdine Production (RPP) was calculated by the following formula: RPP: OD 405 /OD 600 . The siderophore amount in S. aureus was measured using succinic acid broth, containing K 2 HPO 4 , KH 2 PO 4 , (NH 4 ) 2 , MgSO 4 , succinic acid. The suspected strains were inoculated into the media and incubated at 30 °C for 24 h, shaking at 200 rpm. Then, the media was centrifuged at 10,000×g for 10 min. The supernatant's absorbance was measured at OD 400 using a spectrophotometer 48 .
RNA isolation and RT-PCR. The total RNAs isolation and cDNA synthesis were performed on bacteria during the co-cultures using the GeneAll extraction and cDNA synthesis kit (GeneAll, Korea) based on the manufacturer's instructions.
The gene expressions of hasI, pvdS, sigS, kpc, oprD, mexA-mexB-oprM, norA, and walk/R were analyzed using real-time PCR based on the primers listed in Table 4. Based on previous studies, nucA and rpoD were selected as the most suitable reference genes among other housekeeping genes of S. aureus and P. aeruginosa. 10 µL of 2X Syber Green PCR Master Mix (Amplicon, Denmark), 1 µL of each primer (20 pmol), and 2 µL cDNA, and DEPC-treated water was added to each tube to a final volume of 20 µL. The amplification was done based on the following program: 95 °C for 15 min, 40 cycles of 95 °C for 20 s, and 56 °C for 30 s. All the tests were performed in triplicate and in three days.

Statistical analysis.
For all the collected data, a Student t-test was performed by GraphPad Prism 6.0 (Graph Pad Software, USA). The p-values were corrected for multiple testing errors, with a 5% false discovery rate (FDR 5%). The student's t-test was applied to compare the gene expression ratios determined by qRT-PCR between the co-cultures and monocultures. The tests were performed using Holm-Sidak test for multiple average comparisons, considering a p-value of 0.05 or less as significant. All data were presented as mean ± SEM. The