Pseudomonas aeruginosa DesB Promotes Staphylococcus aureus Growth Inhibition in Coculture by Controlling the Synthesis of HAQs

Pseudomonas aeruginosa is a pathogen that can cause serious infections and usually coexists with other pathogens, such as Staphylococcus aureus. Virulence factors are important for maintaining a presence of the organisms in these multispecies environments, and DesB plays an important role in P. aeruginosa virulence. Therefore, we investigated the effect of DesB on S. aureus reduction under competitive situation. Liquid cultures of P. aeruginosa wild type (WT) and its desB mutant were spotted on agar plates containing S. aureus, and the size of the clear zones was compared. In addition, interbacterial competition between P. aeruginosa and S. aureus was observed over time during planktonic coculture. The transcriptional profiles of the WT and desB mutant were compared by qRT-PCR and microarray to determine the role of DesB in S. aureus reduction at the molecular level. As a result, the clear zone was smaller for the desB mutant than for P. aeruginosa PAO1 (WT), and in planktonic coculture, the number of S. aureus cells was reduced in the desB mutant. qRT-PCR and microarray revealed that the expression of MvfR-controlled pqsA-E and phnAB operons was significantly decreased, but the mexEF-oprN operon was highly expressed. The results indicate that intracellular levels of 4-hydroxy-2-heptylquinoline (HHQ), a ligand of MvfR, are reduced due to MexEF-OprN-mediated efflux in desB mutant, resulting in the decrease of MvfR binding to pqsA-E promoter and the reduction of 4-hydroxy-2-alkylquinolines (HAQs) synthesis. Overexpression of mexEF-oprN operon in desB mutant was phenotypically confirmed by observing significantly increased resistance to chloramphenicol. In conclusion, these results suggest that DesB plays a role in the inhibition of S. aureus growth by controlling HAQ synthesis.


Introduction
The gram-negative opportunistic pathogen Pseudomonas aeruginosa is a causative agent of nosocomial and life-threatening infections in injured, burned, and immunocompromised patients [1]. This human pathogen produces multiple extracellular factors, such as elastases,

Bacterial strains and culture conditions
The bacterial strains used in this study are listed in Table 1. All bacterial strains were kind gift. Among these strains, desA, desT, and fabA mutants are P. aeruginosa PAOl harboring each truncated gene, and desB mutant harbors an insertion of an ISlacZ/hah transposon in desB gene. The strains were routinely maintained on Luria-Bertani medium (LB; 10 g/L tryptone, 5 g/L yeast extract, 10 g/L NaCl; Difco) and grown at 37°C.

Lysis of S. aureus
S. aureus was incubated in LB broth at 37°C for 18h. This overnight culture was then diluted to 1:25 with fresh LB broth and incubated to mid-log growth phase. An aliquot of this subculture then was mixed with 2 mL of LB medium containing 0.8% agar to an optical density at 600 nm (OD 600 ) of 0.2. After allowing the plate to solidify, 3 μL of an overnight culture of P. aeruginosa was spotted onto the plate, and it was incubated at 37°C for 24 h. After incubation, the plate was imaged using a universal digital camera.

Interspecies growth competition assay using planktonic cultures
The interbacterial competition assay was performed as described previously with minor modifications [22]. S. aureus and P. aeruginosa strains were streaked on LB agar plates and incubated at 37°C for 24 h. The next day, colonies of approximately the same size were selected from the plates, inoculated in 5 mL of LB broth, and incubated for 18 h. The overnight cultures were washed with 1 mL of PBS and resuspended to an OD 600 of 1.0 and 2.5 for P. aeruginosa and S. aureus, respectively. P. aeruginosa and S. aureus were mixed at 1:1 (vol/vol). A 10-μL aliquot of the mixture was spotted on a cellulose acetate filter disc and placed on the LB agar plate, which was incubated at 37°C. The growth of individual bacterial species was analyzed by resuspending the filter disc in 0.5 mL of PBS and plating the suspension on Cetrimide agar (CA; Sigma, St. Louis, MO, USA) plates for P. aeruginosa and on Mannitol salt agar (MSA; Difco) plates for S. aureus.

Minimum inhibitory concentration (MIC) determination
P. aeruginosa strains were grown overnight (18 h) and subcultured in 5 mL of sterile LB broth to log phase (OD 600 0.7-1.0). The culture was diluted with LB broth to OD 600 0.1. Then, 100 μL aliquots of serial two-fold dilutions of chloramphenicol in LB broth were prepared in a 96-well plate at final concentrations of 0-512 μg/mL, and an equal volume of bacterial culture (at OD 600 0.1) was added. After a 24-h incubation, inhibition of P. aeruginosa growth was assessed by measuring the OD 600 using a microplate reader (Bio Tek Instruments, Inc., Winooski, VT, USA). The MIC was defined as the concentration at which was no growth was observed.
Serial dilution spotting assay P. aeruginosa strains were cultured in LB broth at 37°C for 18 h, and then the overnight culture (approximately density: 10 8 CFU/mL) was serially diluted (10-fold, 10 0 −10 6 CFU/mL) in PBS. 2 μL of the diluents was vertically spotted on LB agar containing 0, 8, 16, 32, 64, or 128 μg/mL chloramphenicol. After 24 h of incubation at 37°C, spot formation was assessed, and the chloramphenicol resistance of the WT and desB mutant strain was compared.
Quantitative RT-PCR (qRT-PCR) P. aeruginosa colonies freshly grown on LB agar plates were inoculated into 5 mL of LB broth and incubated at 37°C for 18 h. These overnight cultures were diluted 1:25 in fresh LB broth, and grown to an OD 600 of 0.4-0.5. Total RNA was extracted from 1 mL cultures of P. aeruginosa strains using the Qiagen RNase mini kit (Qiagen, Hilden, Germany) according to the manufacturer's instructions. Then, one U of RNase-free DNase I (Invitrogen, CA, USA) was added to one μg of extracted total RNA, and the reaction mixture was incubated for 15 min at room temperature. Next, the enzyme was inactivated by adding 1 μL of 25 mM EDTA, and heating the mixture at 65°C for 10 min. cDNA was synthesized from the RNA using the Superscript III first-strand synthesis system (Invitrogen). The PCR reaction mixture (20 μL) contained 10 μL of VeriQuest SYBR Green qPCR master mix (USB, Affymetrix), 1 μL of 10 μM forward primer, 1 μL of 10 μM reverse primer, 2 μL of cDNA, and 6 μL of sterile distilled water. The reaction mixtures were denatured by incubation at 95°C for 10 min, which was followed by 35 cycles of 95°C for 15 s and at 60°C for 1 min. The target genes mvfR, pqsA, mexE, mexF, mesT, and oprN were amplified using the designed primers listed in Table 2. The constitutively expressed housekeeping gene rpoD was used to normalize gene expression. The relative expression of the genes was obtained from the calculated Ct values.

Microarray analysis
Total RNA was prepared from the P. aeruginosa strains using the same as procedure as described for qRT-PCR, except that a greater volume of culture was used. The quality of the purified RNA was confirmed using an Agilent 2100 Bioanalyzer System. cDNA was generated and labeled using the Bioprime labeling kit (Invitrogen), and the microarray hybridization was performed using Hybridization solution (MYcroarray.com). The microarray data were normalized and analyzed using Genowiz 4.0 (Ocimum Biosolutions, India). Total 5544 genes were analyzed, and among these, the comparative transcriptional profile of HAQ-related genes between WT and desB mutant was used for data interpretation.

Statistical analysis
Interspecies growth competition assay was repeated with two samples in each repeat, and MIC assay was repeated with six samples in each repeat. The bacterial cell counts (log CFU mL -1 ) and OD values at several timepoints or chloramphenicol concentrations were analysed by the mixed model procedure (SAS Institute, Cary, NC, USA). Pairwise t-tests were performed for comparisons of least squares means among the interactions at alpha = 0.05. All experiments except for microarray analysis were performed in replicate, but microarray data were confirmed by qRT-PCR.

Results and Discussion
DesB is involved in S. aureus reduction Other microbes are usually detected in P. aeruginosa infections. It was previously shown that P. aeruginosa lyses S. aureus during coculture [23]. P. aeruginosa and S. aureus affect each other, and interspecies communication is required for the expression of various virulence factors [24]. Certain virulence factors obstruct the proliferation of S. aureus by lysis or growth inhibition. Schweizer and Choi [18] found that DesB, an aerobic desaturase expressed by P. aeruginosa, is closely associated with the production of virulence factors, particularly elastase, rhamnolipids, and pyocyanin. Therefore, we hypothesized that DesB also plays a role in S. aureus reduction under mixed culture conditions. Accordingly, we conducted comparative analyses of the virulence of WT and desB mutant strains during coculture with S. aureus. In addition to desB, genes involved in unsaturated fatty acids (UFAs) synthesis, such as desA, desT, and fabA was also evaluated in order to investigate if UFA synthesis is involved in S. aureus reduction. In the spot assay, the desB mutant showed clear zones with smaller diameters than those surrounding the WT strain (Fig 1), whereas the desA (phospholipid acyl desaturase) and desT (TetR family transcriptional regulator) mutants showed clear zones of similar diameter to those surrounding the WT strain. In addition, the fabA mutant showed slightly lower activity of cell number decrease than the WT strain, and the fabA desB double mutant showed no S. aureus reduction (Fig 1). The P. aeruginosa fabA gene encodes β-hydroxydecanoyl-ACP dehydrase, which is involved in fatty acid synthesis under aerobic and anaerobic conditions [25]. In our previous study, we found that the fabA mutant used in the present study has a very low growth rate compared to other mutants, including desA, desB, and desT mutants [21]. Therefore, the decreased S. aureus reduction ability of the fabA mutant, and the zero-activity of the fabAdesB double mutant are not due to a deficiency of FabA activity but result from the slow-growth phenotype of the fabA mutant. This result indicates that DesB is an important factor for S. aureus reduction. Furthermore, a time-course growth competition assay between P. aeruginosa and S. aureus was conducted to confirm the deficiency phenotype of S. aureus reduction in the desB mutant in planktonic culture. Mutation of the desB gene did not influence the growth of P. aeruginosa; however, it did affect the extent of S. aureus reduction in coculture (Fig 2). The total S. aureus cell count decreased after 10 h of coculture with the WT strain, whereas a decrease in the number of S. aureus cells was observed after 12 h of coculture with the desB mutant strain, suggesting that S. aureus reduction by the desB mutant was retarded due to the lack of functional DesB. In addition, a greater number of S. aureus cells were maintained in a coculture with the desB mutant than in a coculture with WT over the same incubation period. However, comparatively smaller difference between two strains was observed at the 24 h timepoint than at the 10 h and 12 h timepoints. It can be inferred that the phenotype of desB mutant at this timepoint was attributed to delayed production of lysis-or growth inhibition-associated virulence factors. Also, it may be possible that S. aureus responds to WT or desB mutant in terms of SCV formation in S. aureus under coculture condition. Pyocyanin stimulates SCV selection in S. aureus and the SCVs appeared after 24h of cocultivation [11]. Since desB mutant displayed reduced pyocyanin production [18], SCVs may be formed less in desB mutant than in WT. However, SCVs were not found in this study because coculture experiment was conducted until 24 h of coculture.
This indicated that P. aeruginosa DesB is involved in S. aureus reduction. P. aeruginosa produces various extracellular antimicrobial substances associated with a decrease in S. aureus cell number that are mostly regulated by the pqsA-E operon [26][27][28]. In addition, P. aeruginosa secretes LasA protease, which lyses S. aureus, and transcription of the lasA gene is controlled by LasR [23,29]. Thus, we assumed that the reduction of S. aureus cells may result from correlation between DesB and production of these factors. However, it should be pointed out that the desB mutant retains some reduction ability. This finding could be explained by the presence of DesB-independent factors that participate in S. aureus reduction even in the absence of DesB activity [14,27,30]. It was discussed more detailedly later in this paper.  DesB positively regulates the transcription of the MvfR-regulated genes pqs and phn qRT-PCR and microarray analyses were conducted to determine the possible molecular mechanism underlying DesB-involved S. aureus reduction. Mashburn et al. [13] reported that a P. aeruginosa pqsA mutant exhibited reduced S. aureus lysis during coculture, indicating that PqsA or anthranilate-coenzyme A ligase [31] is essential for complete S. aureus lytic activity. Thus, we investigated the possibility of a correlation between desB and two other genes, pqsA and its regulator mvfR. The results showed that in the desB mutant, the transcription of mvfR was slightly reduced, whereas pqsA expression was reduced approximately 50-fold compared to the levels in WT (Fig 3). This result suggested that DesB plays a significant role in P. aeruginosa-catalyzed S. aureus reduction in mixed culture by controlling pqsA gene expression. PqsA is required for HAQ synthesis, which regulates the production of the cell-to-cell communication factors and virulence factors of P. aeruginosa, including elastase and pyocyanin [32,33]. However, from this result, we could not predict how DesB regulates pqsA expression at the molecular level. Consequently, a microarray analysis was performed to elucidate the molecular mechanism underlying DesB-related S. aureus reduction. Based on the fact that DesB controls pqsA, transcriptional levels of HAQ-related genes including QS genes in WT and desB mutant were compared.
The expression patterns of pqsA and mvfR in the qRT-PCR analysis were also consistent with microarray data ( Table 3). The levels of all the pqs genes, including pqsA, were significantly reduced in the desB mutant, whereas the expression levels of two other S. aureus lysisrelated genes, lasA (encoding LasA protease) and lasR (required for lasA transcription) in the desB mutant and WT strains were similar.
Recently, Beaume et al [34] reported that lasA is involved in S. aureus lysis, whereas pqs is responsible for S. aureus growth inhibition rather than lysis. Therefore, this indicates that desB mutant have reduced ability in S. aureus growth inhibition by a considerable decrease in pqs expression. Also, microarray analysis revealed that LasA plays a role in S. aureus lysis as a DesB-independent factor in the absence of functional DesB. In addition, Beaume et al. [34] demonstrated that a carB gene, involved in pyrimidine biosynthesis, is required for S. aureus growth inhibition without any influence on PQS synthesis. Thus, CarB also contributes to S. aureus growth inhibition in a DesB-independent manner [34].
P. aeruginosa produces S. aureus growth inhibition-associated materials, such as 4-hydroxy-2-alkylquinolines (HAQs), which have antimicrobial activity. These HAQs include 4-hydroxy-2-heptylquinoline (HHQ), 4-hydroxy-2-nonylquinoline (HNQ), pseudomonas quinolone signal (PQS), and 4-hydroxy-2-heptylquinoline N-oxide (HQNO) [35]. Synthesis of these HAQs is mainly controlled by the P. aeruginosa pqs system, which comprises pqsA-E and pqsH. PqsA-D catalyzes the synthesis of HHQ molecules, which are converted to PQS by PqsH. The pqsA-E genes are expressed under the control of the transcriptional regulator MvfR, whereas pqsH is regulated by LasR, but not by MvfR. Thus, the transcriptional pattern of pqsA-E is distinct from that of pqsH in the desB mutant. The PQS system is interlinked with two quorum-sensing systems, las and rhl [36]. MvfR, a regulator of the pqs system, is positively controlled by LasR and negatively regulated by RhlR [37]. Comparative transcriptional analysis showed that in the desB mutant, expression of the pqs operon was reduced, whereas the rhl and las QS genes were normally expressed, indicating that rhlR and lasR are not involved in the reduced pqs expression observed in the desB mutant. However, rhlI expression was reduced by desB mutation compared to the levels in WT. Since, according to McKnight et al. [38], rhlI expression is positively regulated by PQS, we could assume that the decreased rhlI transcription observed in the desB mutant may be attributed to reduced levels of PQS.
In addition, expression of phnA and phnB in the desB mutant was significantly lower than that in WT. In addition to the pqs operon, transcription of the phnAB genes is also under the control of MvfR. The phnAB-encoded proteins are responsible for conversion of shikimic acid to anthranilate, which is continually used in HAQs synthesis catalyzed by PqsA-D. TrpE and TrpG also catalyze anthranilate synthesis from chorismic acid, and the anthranilate produced by this pathway is known to be utilized for either tryptophan or HAQ synthesis [39]. In addition to the PhnAB pathway, there is an alternative anthranilate synthesis pathway, called the kynurenine pathway, which consists of KynA, KynB, and KynU, and catalyzes the conversion of tryptophan to anthranilate. However, dissimilar to pqs expression, transcriptional expression of kynA, kynB, kynU, and trpEG was not reduced in the desB mutant. Consequently, the above results indicate that anthranilate synthesis is controlled by DesB, and that only MvfR-regulated phnAB expression is associated with this regulation (Table 3). Although pqsA-E and phnAB are co-regulated by MvfR [40], only a small difference in mvfR expression between the desB mutant and WT was observed in the microarray analysis. This phenomenon could be explained by MvfR activation. HHQ functions as a ligand for the LasR-type transcriptional regulator MvfR and enhances MvfR binding to the pqsA-E promoter [41], thereby activating the pqsABCDE and phnAB operons, which is followed by increased production of HAQs. Therefore, in the desB mutant, HHQ was not sufficient to stimulate MvfR binding to the promoter, even if mvfR was expressed at a similar level to that in the WT strain; thus, resulting in decreased MvfR-governed activation of the downstream pqs and phn genes in the desB mutant and reduced production of HAQs.
In addition, the amounts of crude PQS produced by WT and desB mutant were indirectly measured by spot assay for S. aureus growth inhibition in order to compare PQS levels between two strains. As a result, extract of desB mutant exhibited the considerable decrease of S. aureus growth inhibition compared to the one of WT, meaning that desB mutant produced much less PQS than WT (data not shown).
The mexEF-oprN operon is overexpressed in the desB mutant We hypothesized that the low levels of HHQ and concomitant decrease of PQS might lead to the reduced S. aureus growth inhibition via a yet unknown regulatory mechanism. Interestingly, we showed, both by qRT-PCR and by microarray analyses, that the mexEF-oprN operon was highly overexpressed in the desB mutant compared to the levels in WT (Table 3, Fig 3). Although the mexF expression levels in the microarray analysis appear to differ from the levels in the qRT-PCR analysis due to the relatively lower accuracy of microarray data compared to qRT-PCR, the same tendency for transcriptional expression of mexEF-oprN was shown in the results of both methods. The mexEF-oprN operon in P. aeruginosa encodes a resistance-nodulation-cell division (RND)-type efflux pump, MexEF-OprN. The MexEF-OprN system is not induced in most P. aeruginosa strains, thereby allowing the expression of QS-regulated virulence determinants [42]. The mexEF-oprN multidrug efflux operon is highly expressed in the presence of antibiotics, nitrosative stress, or disulfide stress [43,44]. Expression of the mexEF-oprN operon is known to be positively regulated by MexT, which is encoded by a gene located immediately upstream of the mexEF-oprN operon, and the mexT gene is negatively regulated by MexS, an oxidoreductase [45,46]. This pump is an important factor for antibiotic resistance, and it transports various molecules, such as chloramphenicol, fluoroquinolones, triclosan, and trimethoprim [47]. Fukuda et al. [48] reported a norfloxacin-resistant mutant of P. aeruginosa PAO1, called an nfxC-type mutant, and showed that an nfxC-type mutant overexpresses the MexEF-OprN efflux pump. Kohler et al. [49] reported that the nfxC-type mutant shows decreased rhlI expression, and the resulting overexpression of the efflux system negatively affects cell-to-cell signaling in P. aeruginosa. The transcriptional profile of the desB mutant is similar to that of the nfxC-type mutant in terms of the levels of pqsA, phnAB, and type III secretion system-related gene expression [50], but different in terms of lasB and rhlAB expression. The wild-type PAO1 strain used in this study contains a 8-bp insertion in mexT, which encodes an inactive and uninducible protein, whereas its isogenic nfxC mutant harbors a deletion of the 8-bp insert and produces functional MexT [42]. However, our transcriptional analyses revealed that the mexEF-oprN operon was greatly overexpressed in desB mutant, even if mexT expression was not altered due to suppression via increased mexS expression (Table 3, Fig 3). According to Kohler et al. [45], nfxC-type mutant produces the effector of MexT, thus causing MexT activation at the posttranslational level and consequently constant overexpression of the mexEF-oprN operon. In addition, MexT also positively regulates several other genes such as PA1744, PA1970, PA2759, PA3229, PA4623, and PA4881 [51], and our transcriptional analysis revealed that these genes were highly expressed in desB mutant (data not shown). Hence, we could assume that posttranslationally activated MexT contributes to the overexpression of mexEF-oprN and other additional genes. Alternatively, the results suggest that mexEF-oprN overexpression in the desB mutant is MexT independent. Kumar and Schweizer [52] observed that large colony variants lacking several efflux pumps exhibited mexEX-oprN overexpression, even though they harbor nonfunctional MexT [52]. This finding suggested that metabolic stress due to the fitness impairment of the variants caused overexpression of the MexEF-OprN efflux pump via a yet uncharacterized regulatory mechanism in the absence of MexT activation [52]. Likewise, our findings could be explained by a disruption of normal metabolism due to the desB mutation, which affects cell fitness and facilitates MexEF-OprN overexpression in the absence of MexT-controlled regulation.
In order to confirm the phenotype of mexEF-oprN efflux pump overexpression in the desB mutant, we determined the MIC of chloramphenicol, a representative substrate for MexE-F-OprN and performed serial dilution-spotting assays. The results showed that the desB mutant was considerably more resistant to chloramphenicol that the WT, and the MIC of the desB mutant was greater than 512 μg/mL, compared to 128 μg/mL for WT (Fig 4A). In the serial dilution-spotting assay, all ten-fold dilutions of the desB mutant grew normally up to 128 μg/mL of chloramphenicol, whereas growth of WT was reduced starting at 16 μg/mL ( Fig  4B). This result demonstrated that the desB mutant is highly resistant to antibiotics due to MexEF-OprN overexpression. Lamarche and Deziel [53] demonstrated that besides antibiotics, the MexEF-OprN efflux pump also exports HHQ, which results in low levels of HHQ and PQS inside the bacterial cells. Likewise, overexpressed MexEF-OprN in the desB mutant also led to reduced production of HAQs. In addition, Olivares et al. [50] reported that PQS in P. aeruginosa overexpressing mexEF-oprN was not detected in early stationary phase but PQS was produced in late stationary phase. Thus, it can be explained that reduced S. aureus growth inhibition in desB mutant comes from delayed PQS production (Fig 2).

Conclusion
This study demonstrated that the desB mutation results in overexpression of MexEF-OprN, which subsequently contributes to 1) decreased HAQs levels inside the cells, 2) reduced MvfR binding to the pqsA-E promoter, and 3) suppression of HAQ synthesis. Ultimately, these events lead to impaired production of the virulence factors involved in S. aureus growth inhibition (Fig 5). In other words, P. aeruginosa DesB is very involved in S. aureus growth inhibition in mixed microbial communities. However, further studies are needed to determine how desB mutation is linked to MexT-independent mexEF-OprN expression at the molecular level. WT and desB were exposed to chloramphenicol (0-512 μg/mL). After 24-h incubation, the optical density at 600 nm (OD 600 ) was measured to determine bacterial growth in the presence of chloramphenicol. (B) Overnight cultures of WT and desB were 10-fold serially diluted and then spotted on chloramphenicol-containing agar plates (0, 8, 16, 32, 64, and 128 μg/mL). After incubation, growth was observed. Means with (*, **) are significantly different (P < 0.05) in WT or desB mutant. Supporting Information S1 Appendix. Microarray data comparing gene expression of P. aeruginosa PAO1 (WT) and its desB mutant (desB mutant). (PDF)