Contribution of the nos-pdt Operon to Virulence Phenotypes in Methicillin-Sensitive Staphylococcus aureus

Nitric oxide (NO) is emerging as an important regulator of bacterial stress resistance, biofilm development, and virulence. One potential source of endogenous NO production in the pathogen Staphylococcus aureus is its NO-synthase (saNOS) enzyme, encoded by the nos gene. Although a role for saNOS in oxidative stress resistance, antibiotic resistance, and virulence has been recently-described, insights into the regulation of nos expression and saNOS enzyme activity remain elusive. To this end, transcriptional analysis of the nos gene in S. aureus strain UAMS-1 was performed, which revealed that nos expression increases during low-oxygen growth and is growth-phase dependent. Furthermore, nos is co-transcribed with a downstream gene, designated pdt, which encodes a prephenate dehydratase (PDT) enzyme involved in phenylalanine biosynthesis. Deletion of pdt significantly impaired the ability of UAMS-1 to grow in chemically-defined media lacking phenylalanine, confirming the function of this enzyme. Bioinformatics analysis revealed that the operon organization of nos-pdt appears to be unique to the staphylococci. As described for other S. aureus nos mutants, inactivation of nos in UAMS-1 conferred sensitivity to oxidative stress, while deletion of pdt did not affect this phenotype. The nos mutant also displayed reduced virulence in a murine sepsis infection model, and increased carotenoid pigmentation when cultured on agar plates, both previously-undescribed nos mutant phenotypes. Utilizing the fluorescent stain 4-Amino-5-Methylamino-2',7'-Difluorofluorescein (DAF-FM) diacetate, decreased levels of intracellular NO/reactive nitrogen species (RNS) were detected in the nos mutant on agar plates. These results reinforce the important role of saNOS in S. aureus physiology and virulence, and have identified an in vitro growth condition under which saNOS activity appears to be upregulated. However, the significance of the operon organization of nos-pdt and potential relationship between these two enzymes remains to be elucidated.


Introduction
Nitric Oxide (NO) is a highly reactive free radical gas that easily and rapidly diffuses across biological membranes [1,2].One cellular source of NO production is via the NO synthase enzymes (NOS), whose three isoforms are well-characterized in mammals and are involved in a wide variety of cellular processes, including regulation of blood pressure, nervous system signaling, and protection against pathogens [3][4][5][6][7][8][9].All three mammalian NOS enzymes consist of both an oxygenase and reductase domain, which together catalyze a five-electron heme-based oxidation of the guanidine nitrogen of L-arginine to L-citrulline [7].The power and versatility of NO as a signaling molecule is exemplified by its ability to diffuse several cell diameters away from its site of synthesis in eukaryotes [1,10].NO can react directly or indirectly via formation of reactive nitrogen species (RNS) with a variety of cellular targets, including DNA, lipids, and proteins containing transition metal centers and thiol groups (reviewed in [11]).
A number of studies in recent years have revealed that some bacterial species also contain a NOS homologue in their genomes, equivalent in most cases to the oxygenase domain of eukaryotic NOS (reviewed in [3]).The role of bacterial NOS can be quite varied, participating in diverse physiological functions such as biofilm development, biosynthesis of plant phytotoxins, stress resistance and virulence [12][13][14][15][16][17][18].Although this versatility suggests a specificity of function to individual bacterial species, an increased understanding of the regulation and exact cellular targets of bacterial NOS may help divulge an evolutionary link between prokaryotic and eukaryotic organisms.Previous analysis of bacterial genomes has shown that bacterial NOS enzymes are primarily found in Gram-positive species such as Streptomyces, Bacillus, and the staphylococci [3].A number of biochemical studies have focused on determining the crystal structure and characterizing the enzymatic activity of these bacterial NOS enzymes (reviewed in [4]).By comparison, much less is understood about the regulation of bacterial NOS enzyme expression, activity, and participation in cell signaling.
Staphylococcus aureus is a versatile pathogen that is highly transmissible in healthcare settings as well as in the community atlarge.Of increasing concern is the fact that many methicillinresistant S. aureus (MRSA) strains are genetically-resistant to nearly all antibiotics currently used in the clinical setting.In fact, a recent report by the CDC estimates that 80,000 MRSA infections and 11,000 MRSA-associated deaths occur in the US per year [19].These alarming statistics underscore the importance of deciphering new drug targets to expand our arsenal against this devastating pathogen.Bacterial NOS enzymes are an attractive target for drug development, since they have been shown to confer tolerance to oxidative stress [14,15,20] and antibiotics [18,21], promote bacterial survival inside neutrophils [20] and macrophages [14,21], and contribute to virulence in vivo [14,21].A search for small molecule inhibitors of Bacillus subtilis NOS has also recently been undertaken and described [22].With respect to S. aureus, exogenous addition of NO to cultures protects this bacterium from challenge with H 2 O 2 and antimicrobials [15,18].More recent studies of S. aureus nos mutants have revealed that this gene contributes to a number of virulence properties, including protection against oxidative stress [20,21], antibiotics [21], antimicrobial peptides [21], and leucocyte-killing [20,21].A role for saNOS in S. aureus virulence was also recently confirmed in a murine abscess model of infection [21].Although multiple models infer the importance of bacterial NOS during infection, the exact molecular mechanisms by which NOS elicits its control on bacterial stress resistance and virulence have not been elucidated.
Given the established relevance of saNOS in virulence and as a potential drug development target [22], it is surprising that relatively little attention has been given to the genetics, regulation, and the exact cellular targets of this enzyme.Inspection of the S. aureus genome has revealed that the nos open reading frame (ORF; SAR2007 of the MRSA252 genome) is separated by only 19 nucleotides from the downstream SAR2008 ORF.No identifiable transcription termination signals appear to be present between nos and SAR2008, suggesting that they may comprise an operon.SAR2008 (henceforth referred to as the pdt gene) encodes a prephenate dehydratase (saPDT) enzyme, which catalyzes the formation of phenylpyruvate from prephenate during phenylalanine biosynthesis [23].Therefore, this current study was undertaken to probe the potential genetic and functional relationships between saNOS and saPDT in a clinical methicillin-sensitive S. aureus strain.These experiments demonstrate that the nos and pdt genes are co-transcribed, and that the expression of this operon is upregulated during low-oxygen growth.Furthermore, a nos::erm mutant displayed reduced virulence in a murine sepsis model and increased carotenoid pigmentation when cultured on agar plates, both previously-undescribed nos mutant phenotypes.An assay to detect intracellular NO/RNS in S. aureus was also developed using the fluorescent stain DAF-FM diacetate, and demonstrated decreased levels of intracellular NO/RNS in the nos mutant when cultured on agar plates.

Bacterial strains and growth conditions
Escherichia coli and Staphylococcus aureus strains and plasmids used in this study are listed in Table 1.Planktonic S. aureus cultures were grown either aerobically (1:10 volume to flask ratio, 250 RPM) or under low-oxygen conditions (7:10 volume to flask ratio, 0 RPM), as indicated for each experiment.S. aureus was freshly streaked from a frozen stock onto tryptic soy agar (TSA) for 24 hours, followed by overnight aerobic growth in tryptic soy broth (TSB) for each experiment, unless otherwise indicated.As indicated, antibiotics were included in agar plates and broth cultures at the following concentrations: 5 mg/ml or 10 mg/ml chloramphenicol (Cm), 2 mg/ml or 10 mg/ml erythromycin (Erm).E. coli was routinely grown under aerobic conditions (1:10 volume to flask ratio, 250 RPM) in Luria-Bertani (LB) broth with 50 mg/ ml ampicillin (Amp) or 50 mg/ml kanamycin (Km), as indicated in Table 1.Glycerol stock cultures of all strains were maintained at 280uC and were prepared by mixing equal volume of overnight culture with sterile 50% (vol/vol) glycerol in cryogenic tubes.

Creation of nos::erm and Dpdt mutant strains
To create a nos::erm insertion mutant, plasmid pTR27 (Table 1) was created as follows: The nos gene was amplified from S. aureus genomic DNA by PCR using the nos1-F and nos1-R primers specified in Table 2, followed by Topo-cloning into pCR2.1 (Invitrogen).An internal BglII site was introduced into this cloned sequence 232 bp downstream from the saNOS start codon in the sequence 5'-gttaaatgtcattgatgcaagAGATGTtactgacgaagcatcgttcttatc-3', where the capital letters are located, changing the sequence from AGATGT to AGATCT, a BglII restriction enzyme cut site.An EcoRI fragment harboring this modified allele was moved into the EcoRI site of pBluescript SK (pBSK), a vector backbone that does not contain a BglII site.A 1.1 kb BamHIdigested fragment harboring the erm resistance cassette from Tn1545 [24] was then cloned into the BglII site of the nos fragment cloned in pBSK.The resulting nos::erm allele was subsequently ligated into the EcoRI site of the temperature sensitive shuttle vector pBT2 [25] to generate pTR27, which was first transformed into strain RN4220 by electroporation, followed by phage transduction into UAMS-1, with growth at 30uC for all steps [26][27][28].To create strain KR1010 (UAMS-1 nos::erm mutant), integration of pTR27 into the nos gene on the UAMS-1 chromosome was initiated by growth at 43uC on TSA +10 mg/ml Erm (non-permissive temperature for plasmid replication), to promote integration of the plasmid into the chromosome via homologous recombination at the nos gene.To induce a second recombination event, a single isolated colony was used to inoculate TSB (no antibiotic) and grown at 30uC for 5 days.Every 24 hours, an aliquot of the culture was diluted 1000-fold into fresh TSB (no antibiotic).On days 3-5, the culture was serially-diluted and spread on TSA +10 mg/ml Erm, and isolated colonies were then screened for Erm-resistant and Cm-sensitive phenotypes by picking and patching onto TSA +2 mg/ml Erm and TSA + 5 mg/ml Cm.PCR using the nos1-F and nos1-R primer pair, as well as Southern blotting, were used to screen candidate mutants to confirm that the chromosomal nos gene had been correctly replaced by the nos::erm allele.
To create a Dpdt deletion mutant, plasmid pKR13 (Table 1) was created as follows: PCR was performed with primer pairs pdt1-F/pdt1-R (amplifying the 1.2-kb region upstream of nucleotide 151 of the pdt open reading frame), and pdt2-F/pdt2-R (amplifying the 1.0-kb region downstream of nucleotide 672 of the pdt ORF) using AccuPrime TM Pfx SuperMix (Invitrogen Life Technologies) and UAMS-1 genomic DNA as template.Each PCR product was cloned into pCRBlunt (Invitrogen Life Technologies) following the manufacturer's protocols for ligation and E. coli transformation, followed by Sanger DNA sequencing (ICBR, University of Florida) to verify that no mutations had been introduced.The pdt1 and pdt2 inserts were each excised from pCRBlunt by restriction digestion with HindIII-ClaI and ClaI-XbaI, respectively, followed by gel-purification of each insert (DNA Gel Extraction Kit, Zymo Research).Plasmid pBT2 was digested with HindIII-XbaI, and a ''triple'' ligation reaction using T4 DNA ligase (New England Biolabs) containing this linearized plasmid, HindIII-ClaI pdt1 and ClaI-XbaI pdt2 was incubated for several hours at room temperature, followed by heat-shock transformation into E. coli DH5a.The resulting pKR13 was moved into S. aureus RN4220 and UAMS-1 by electroporation and phage transduction, respectively, as described above.Creation of strain KR1013 (UAMS-1 Dpdt mutant) was performed as described above, except that isolated colonies recovered after induction of the second recombination event were screened for chloramphenicol sensitivity by picking and patching onto TSA (no antibiotic) and TSA +5 mg/ml Cm.PCR with primers nos2-F and pdt2-R was used to screen candidate mutants to confirm that 521bp of the pdt gene had been deleted.

Creation of complementation plasmids
To create pMKnos, primer pair nos2-F/nos2-R and Thermalace enzyme (Invitrogen Life Sciences) was used to PCR-amplify a 1.6 kb product (spanning 564 nucleotides upstream of the nos ATG start codon through the entire nos ORF), using UAMS-1 genomic DNA as template.This PCR product was cloned into pCRBlunt (Invitrogen Life Technologies) according to the manufacturer's protocols for ligation and E. coli transformation, followed by Sanger DNA sequencing to verify that no mutations had been introduced.This DNA fragment was excised from pCRBlunt by restriction digestion with BamHI and EcoRI, gelpurified, and ligated to BamHI-EcoRI digested pMK4 [29].This ligation was transformed into E. coli, followed by movement into S. aureus RN4220 and KR1010 by electroporation and phage transduction, respectively, as described above.To create pMKpdt, primer pair nos3-F/nos3-R was used to PCR-amplify a product spanning 564 nucleotides upstream of the nos ATG start codon, and primer pair pdt3-F/pdt3-R was used to amplify a 0.97 kb product spanning the pdt ORF.Each PCR product was cloned into pCRBlunt (Invitrogen Life Technologies) following the manufacturer's protocols for ligation and E. coli transformation, followed by Sanger DNA sequencing.The nos promoter fragment was excised from pCRBlunt by restriction digestion with BamHI and SmaI, gel-purified, and ligated to BamHI-SmaI digested pMK4 and transformed into E. coli.The resulting plasmid was then digested with SmaI and EcoRI, and ligated to the 0.97-kb pdt insert that was excised from pCRBlunt by digestion with these same two enzymes.This ligation was then transformed into E. coli, followed by movement into S. aureus RN4220, KR1010, and KR1013 by electroporation and phage transduction, respectively, as described above.S. aureus UAMS-1, KR1010, and KR1013 were also transduced with pMK4 to create vector-only control strains.

Growth experiments and RNA isolation
To measure growth-phase dependent expression of nos and pdt, overnight cultures of UAMS-1 were diluted to an OD 600 = 0.05 in TSB, and were grown in both aerobic and low-oxygen conditions as described above (n = 3 biological replicates).Culture samples were harvested from each flask at 2, 6, and 12 hours growth.To compare nos and pdt expression between UAMS-1 (pMK4), nos::erm mutant (pMK4) and nos::erm complement (pMKnos) strains, cultures of each (n = 3 biological replicates) were grown for 6 hours under low-oxygen conditions.To compare expression of nos, pdt, and various carotenoid pigment-related genes (Table 2) between UAMS-1 (pMK4), nos::erm mutant (pMK4), nos complement (nos::erm mutant containing pMKnos) and pdt complement (nos::erm mutant containing pMKpdt) strains, cultures of each (n = 3 biological replicates) were grown for 28 hours on TSA +5 mg/ml Cm.For all experiments, RNA was isolated with the FASTPREP system (Lysing Matrix B, Q-Biogene, Inc.) and Qiagen RNeasy mini kit, using previously-published methods [30].

Quantitative real-time PCR (qRT-PCR) analysis of gene expression
To remove residual genomic contamination, isolated RNA was treated with Turbo DNase using the TURBO DNA-free TM kit (Applied Biosystems).A mock qRT-PCR was then performed with primer pair sigA-F/sigA-R and DNAse-treated RNA as template to check for amplification of contaminating genomic DNA.Each RNA sample (0.750 mg) was then converted to cDNA using the iScript Reverse Transcriptase kit (BioRad), which generates total cDNA using a mixture of oligo(dT) and random hexamer primers.Expression of genes of interest (Table 2) was measured in the cDNA from each sample by qRT-PCR using iQ SYBR green supermix (BioRad) and the Eco Real-Time PCR system (Illumina).The Livak method (2 -DDCt ) [31] was used to calculate the relative fold change between the calibrator (indicated in each figure legend) and test samples.Primers specific to the housekeeping gene sigA (sigA-F/sigA-R) were used to normalize the data.The amplification efficiency of sigA-F/sigA-R was within 5% of all of the other primer pairs used in qRT-PCR.All qRT-PCR reactions were performed on triplicate technical replicates of each RNA isolated from n = 3 biological samples.

Co-transcription PCR
To determine whether the nos and pdt genes are co-transcribed, 0.750 mg of UAMS-1 RNA isolated from 6 hour low-oxygen cultures (n = 3 biological replicates) was converted to cDNA using the iScript select cDNA kit (BioRad) and primer pdt5-R, which anneals to the stop codon region of the pdt gene.As a control, these reactions were also performed in the absence of reversetranscriptase (RT) enzyme, to check for genomic DNA contamination.PCR reactions (30 cycles) were then performed on 0.075 mg of each RT+ and RT-cDNA sample, as well as a UAMS-1 genomic DNA template control and no template control.Primer pair nos5-F/pdt5-R was used to amplify a 1.9kb product that spans the region between the nos start codon and pdt stop codon.PCR products were visualized by agarose gel electrophoresis and ethidium bromide staining, and imaged on a Molecular Imager Gel Doc XR+ (BioRad).

Carotenoid pigment assays
The pigment differences in UAMS-1 (pMK4), nos::erm mutant (pMK4), nos complement (nos::erm mutant containing pMKnos), and pdt complement (nos::erm mutant containing pMKpdt) strains were quantitatively assessed using a modified method from Morikawa et al. [32].In brief, each strain was grown at 37uC on TSA +5 mg/ml Cm.After 48 hours growth, cells were scraped from each plate, suspended in 1 ml dH 2 O, and centrifuged for 3 minutes at 13,000 RPM.The supernatants were removed and cells were washed and centrifuged a second time as described above.Cell pellets were then resuspended in 420 ml methanol, vortexed for 10 seconds, and 20 ml was withdrawn to measure the OD 600 of each suspension.The remaining 400 ml was incubated at 55uC for 5 minutes, and then centrifuged at 13,000 RPM for 2 minutes.Finally, 350 ml supernatant (containing extracted pigments) was withdrawn from each tube, transferred to a cuvette containing 650 ml methanol, and the absorbance of each sample at 465 nm.All samples were standardized by dividing the A 465 reading by the sample's corresponding relative OD 600 value [calculated by dividing the OD 600 of each sample by the average OD 600 of UAMS-1 (pMK4)], to account for slight variations in the amount of cells harvested from the agar plates.This experiment was performed on n = 3 independent experiments per strain.

Assessment of phenylalanine biosynthesis
To determine if saPDT is required for phenylalanine biosynthesis, UAMS-1 (pMK4), nos::erm mutant (pMK4), Dpdt mutant (pMK4), and corresponding complement strains were grown in chemically-defined media (CDM) [33] in the presence and absence of phenylalanine.In brief, each strain was cultured overnight in TSB containing 5 mg/ml Cm, followed by centrifu-gation to harvest cells.Cell pellets were each washed once in CDM lacking phenylalanine, then resuspended in the same media.These washed cell suspensions were then used to inoculate 1 ml of CDM with phenylalanine or CDM lacking phenylalanine to an OD 600 = 0.02.Aliquots (200 ml) of each diluted culture were transferred in triplicate to wells of a 96-well tissue culture plate (Costar 3596), and grown at 37uC.OD 600 readings were measured in a Synergy HT microplate reader (Biotek) every 2 hours over a 24 hour growth period.This experiment was performed on n = 4 biological replicates (acquired in 2 independent experiments) per strain.

Fluorescent detection of endogenous NO/RNS
To detect intracellular accumulation of NO and RNS in S. aureus, a fluorescent assay using the cell-permeable stain 4-Amino-5-Methylamino-2',7'-Difluorofluorescein (DAF-FM) Diacetate (Invitrogen Life Technologies) was developed.To validate the use of this stain as a measure of intracellular NO/RNS in S. aureus, strain UAMS-1 was grown overnight in TSB supplemented with 3% (wt/vol) NaCl and 0.5% (wt/vol) glucose (TSB-NaGlc).Overnight cultures were then diluted to an OD 600 = 0.05 in TSB-NaGlc, and 1 ml aliquots were transferred to wells of a 24well tissue culture plate (Costar 3524; previously coated overnight at 4uC with 20% human plasma) and grown for 7 hours at 37uC.Following growth, the total well (biofilm + supernatant) was harvested, centrifuged and cell pellets were resuspended in 16 Hank's Buffered Salt Solution (HBSS) buffer containing 5 mM DAF-FM diacetate.Cell suspensions were incubated for 1 hour at 37uC, then cells were collected by centrifugation, washed once in HBSS buffer, and resuspended in HBSS alone (''untreated'') or HBSS supplemented with 100 mM Diethylamine (DEA; Sigma) or 100 mM DEA/NO (Cayman Chemicals).Aliquots (200 ml) of each untreated or treated cell suspension were immediately transferred in triplicate to wells of a Costar 3904 96-well plate, which was incubated at 37uC in a Biotek Synergy HT fluorescent plate reader.Fluorescence (EX/EM 485610/516610) and OD 600 measurements were recorded after 30 minutes, and data were reported as relative fluorescent units (RFU) per OD 600 of each well.Where indicated, 2-(4-carboxyphenyl)-4,5-dihydro-4,4,5,5tetramethyl-1H-imidazolyl-1-oxy-3-oxide, monopotassium salt (cPTIO), an NO scavenger, was added to cell suspensions to a concentration of 150 mM during the 1 hour DAF-FM staining step.
To determine the relative levels of endogenous NO/RNS of S. aureus when grown on TSA plates, UAMS-1 (pMK4), nos::erm mutant (pMK4), nos complement (nos::erm mutant containing pMKnos) and pdt complement (nos::erm mutant containing pMKpdt) strains were cultured for 26 hours on TSA +5 mg/ml Cm.Cells were harvested from each plate by scraping with a sterile toothpick, and resuspended in 1 ml HBSS.Cell pellets were harvested by centrifugation, washed once in HBSS buffer, and then resuspended in HBSS containing 5 mM DAF-FM.Aliquots (200 ml) of cell suspension were immediately transferred in quadruplicate to wells of a Costar 3904 96-well plate, which was incubated at 37uC in a Synergy HT fluorescent plate reader.Fluorescence and OD 600 measurements were recorded after 90 minutes, and data were reported as RFU per OD 600 of each well.

Hydrogen peroxide sensitivity assay
The sensitivity of UAMS-1 (pMK4), nos::erm mutant (pMK4), Dpdt mutant (pMK4), and corresponding complement strains were assessed for sensitivity to H 2 O 2 based on the method reported by Nudler et al. [15].In brief, each strain was inoculated in LB broth (+ selective antibiotic, as needed), and grown overnight at 37uC and 250 RPM.These cultures were diluted to an OD 600 = 0.05 in 20 ml LB (no antibiotic) in a 250 ml Erlenmeyer flask.After growth for 4.75-5 hours at 37uC and 250 RPM, 1 ml was removed for serial dilution, track plating [34] and CFU/ml determination, and the remaining culture was incubated in the presence of 250 mM H 2 O 2 for 2 hours.After this 2 hour treatment, 1 ml of culture was removed from each flask for serial dilution and track plating to determine CFU/ml.

Murine model of sepsis and dissemination
These experiments were conducted as described previously [35,36].Briefly, 6 week old, female CD-1 Swiss mice were purchased from Charles River Laboratories, and housed at the vivarium in the College of Medicine, University of South Florida.Animals were housed in standard metal cages (2 per cage), with ad libitum food and water access.For infection purposes, previouslystocked cultures of UAMS-1 and KR1010 were thawed, washed twice in PBS, and diluted in PBS to 1610 9 CFU/ml.Mice were randomly assorted into two groups (n = 9 mice per strain) and subsequently inoculated by tail vein injection with 100 ml bacterial suspension, giving a final inocula of 1610 8 CFU/ml.The infection was allowed to proceed for 7 days, or until mice reached a pre-moribund state (used as a measure of mortality).Mice were monitored every 3 hours for the first 24 hours after inoculation, then every 8 hours (up to 72 hours post-inoculation), and twice daily for the remaining infection period.The criteria used for determining the pre-moribund state included development of a hunched posture, labored breathing, decreased activity, immobility, inability to eat or drink, and ruffled fur.Mice were then euthanized with carbon dioxide, and death was verified by the cessation of cardiovascular and respiratory activity.The liver, kidneys, heart, and lungs were collected and stored at 280uC.Any mouse sacrificed before day 7 was recorded for mortality, and their organs were stored for enumerating bacterial burden.Each organ was subsequently homogenized in 3 ml sterile PBS, and the CFU/ organ determined via serial dilution and viable count enumeration.This study was performed in strict accordance with the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health.The protocol was approved by the Institutional Animal Care and Use Committee of the University of South Florida (Permit Number: A-4100-01).

Statistical analyses
All statistical analyses used for the in vivo data were performed using SAS software (version 9.2, SAS Institute, Cary, NC).The distribution of data was determined in SAS through tests for normality (SAS proc univariate) and equality of variance (SAS proc ttest).The statistical significance of bacterial recovery from the murine model of sepsis was evaluated using a Mann-Whitney Test.Mortality was measured using a Log Rank and Chi-square test with 1-degree of freedom.For all in vivo statistical analyses the significance level was set at a = 0.05.In vitro assays were subjected to statistical analysis using Sigmaplot version 12.5 software (Build 12.5.0.38,Systat Software, Inc.).Data were tested for normality and equal variance, then analyzed with a parametric (if normality and/or equal variance tests passed) or non-parametric (if normality and/or equal variance tests failed) one-way ANOVA, followed by an appropriate multiple-comparison test to detect differences between individual groups.

Results
The S. aureus nos and pdt genes comprise a co-transcribed operon Inspection of the S. aureus nos genomic region revealed the presence of an ORF located only 19-bp downstream of the nos gene.This ORF (SAR2008 of the MRSA252 genome, herein referred to as pdt) encodes a predicted prephenate dehydratase (saPDT) enzyme, which catalyzes the formation of phenylpyruvate from prephenate during phenylalanine biosynthesis [23].A conserved domain search [37] of the S. aureus pdt ORF revealed the presence of signature catalytic (PDT; IPR001086) and regulatory (ACT; IPR002912) domains present in mono-functional PDT enzymes that are common to Gram-positive bacteria [23].A BLAST search [38] using the S. aureus MRSA252 nucleotide sequence spanning the nos and pdt ORFs revealed that this organization is present in all sequenced S. aureus genomes, as well as all sequenced coagulase-negative staphylococcal genomes currently available in the National Center for Biotechnology Information (NCBI) genome database (Fig. 1).As reported previously [3] and verified by our own BLAST search of all sequenced microbial genomes in the NCBI database, numerous other bacterial genomes contain a gene encoding a predicted NOsynthase that is highly homologous to the S. aureus nos ORF (. 50% amino acid identity).However, inspection of the genomic context of nos in these other bacteria revealed that the putative nos-pdt operon structure appears to be unique to the staphylococci (Fig. 1).
To gain insight into the growth conditions under which nos and pdt are expressed, quantitative real-time PCR (qPCR) was performed on RNA isolated from aerobic and low-oxygen cultures at early-exponential (2 hours growth), late exponential (6 hours growth) and stationary (12 hours growth) growth phase (Fig. 2).The expression patterns of nos and pdt were similar, in that RNA levels of each gene were greatly increased under low-oxygen growth compared to aerobic growth.Furthermore, expression of both of these genes was growth-phase dependent, peaking at late exponential growth phase, and declining in stationary phase.Cotranscription of nos and pdt was verified by synthesizing cDNA from UAMS-1 RNA using a primer that anneals to the 3' end of the pdt gene, followed by PCR using primers designed to amplify a 1.9 kb product that spans the start codon of nos and stop codon of pdt.As shown in Fig. 3A, a distinct 1.9 kb PCR product was amplified using cDNA as template, confirming that nos and pdt are indeed co-transcribed.This amplification only occurred when reverse-transcriptase (RT) enzyme was present in the cDNA synthesis reaction, indicating that this product was not amplified from contaminating genomic DNA.Because the nos mutant used in this study was created by inserting an erythromycin resistance gene cassette into the nos coding sequence, the effect on pdt transcription in this mutant was also assessed using quantitative real-time PCR (Fig. 3B).Comparison of pdt expression in the nos mutant by this method revealed that pdt RNA levels were about 3fold less than the wild-type strain in low-oxygen 6 hour cultures, and pdt expression was not restored when the nos mutant was complemented with nos expression from a plasmid.These results collectively demonstrate that the nos and pdt genes represent a cotranscribed operon, and that the insertion mutagenesis used to create the nos mutant has a partial polar effect on pdt expression when cultured in TSB.

S. aureus PDT is required for growth in the absence of phenylalanine but does not affect oxidative stress resistance and carotenoid pigmentation
The crystal structure of saPDT has been previously solved and the predicted enzymatic function of this enzyme (conversion of phenylpyruvate to prephenate) has been confirmed in vitro [23].However, the operon structure of nos and pdt suggest that the functions of each encoded enzyme may be inter-related.Therefore, the ability of wild-type, nos mutant, pdt mutant, and corresponding complement strains to grow in the presence and absence of phenylalanine was compared using a chemicallydefined media (CDM; Fig. 4).In complete CDM (containing phenylalanine), all wild-type, mutant, and complement strains displayed similar growth patterns over a 24-hour period, and each achieved a similar OD 600 after 24 hours growth (Fig. 4A).However, in CDM lacking phenylalanine, the pdt mutant displayed greatly reduced growth relative to the wild-type strain, and reached an OD 600 that was 83% less than the wild-type strain after 24 hours (Fig. 4B).This growth defect was fully complemented by expressing the pdt gene in trans on a plasmid in the pdt mutant strain, confirming a requirement for this enzyme in S. aureus phenylalanine biosynthesis.The pattern of nos mutant growth was nearly identical to the wild-type strain in the absence of phenylalanine (Fig. 4C), even though this mutation had a partial polar effect on pdt expression when grown in TSB (Fig. 3B).Data retrieved from a previously-published RNA-seq experiment performed on S. aureus strain USA300 HOU [39] suggests that there may be other potential promoter sites within the nos-pdt operon capable of driving pdt expression (Fig. S1 in File S1).At least two of these regions are located downstream of the inserted erm cassette in the nos mutant, and may represent a potential reason why growth in the absence of phenylalanine was not impaired in the nos mutant.Overall these results suggest that (1) saNOS is not required for growth in the absence of phenylalanine, and (2) there is adequate expression of pdt in the nos mutant to support phenylalanine biosynthesis when grown in chemically-defined media.A role for saNOS in resistance to oxidative stress has been established [18,20,21], but a similar role for saPDT has not been previously explored.Given that nos [40] and pdt [40,41] expression were shown by others to be induced in response to H 2 O 2 , it is possible that saPDT may also contribute to oxidative stress resistance, possibly by generating the a-keto acid anion phenylpyruvate that has been proposed to act as an antioxidant by direct chemical interaction with H 2 O 2 [42].To explore this possibility, wild-type, nos mutant, pdt mutant, and corresponding complement cultures were grown to late-exponential phase, and challenged with 250 mM H 2 O 2 (Fig. 5).As expected, the nos mutant displayed a significant (.3-log) reduction in cell viability after H 2 O 2 challenge compared to the wild-type strain, and this phenotype was fully-restored by supplying the nos gene on a plasmid (Fig. 5A).However, the sensitivity of the pdt mutant to oxidative stress was unaffected in this assay (Fig. 5B), suggesting that, at least in this in vitro growth condition, saPDT does not appear to function in an antioxidant capacity.
During routine culturing of the wild-type, nos mutant, pdt mutant, and complement strains used in these studies, the nos mutant consistently appeared to display increased carotenoid (golden-yellow) pigmentation when cultured on TSA agar plates.This observation was verified by measuring the A 465 nm of methanol-extracted carotenoid pigments from 48-hour TSA plate cultures of wild-type, nos mutant, and complement strains (Fig. 6A).Consistent with qualitative observations, the nos mutant displayed a statistically-significant (p,0.05,Student-Newman-Keuls Test) increase in pigmentation compared to the wild-type strain.Furthermore, pigment was restored to wild-type levels by complementation with a nos-expressing plasmid but not with a pdt-expressing plasmid, verifying that increased pigmentation of the nos mutant is solely due to loss of saNOS function.To determine if increased pigmentation in the nos mutant was a result of altered expression of pigment-related genes, qPCR was employed to measure changes in expression of crtN (a member of the carotenoid biosynthesis operon [43]), asp23 (alkaline shock protein 23, highly-regulated by alternative sigma factor B [44,45]), and purH (purine biosynthesis gene H), as mutation in these genes and/or regulatory circuits have each been shown to greatly alter pigment production [43,45,46].Surprisingly, asp23, crtN, and purH RNA levels were not significantly different between wildtype, nos mutant, and nos complement strains (Fig. 6B), suggesting that increased pigmentation of the nos mutant is not due to altered expression of these genes.

Detection of intracellular NO/RNS in S. aureus by DAF-FM diacetate
To detect and compare relative levels of intracellular NO in S. aureus, an assay using the cell-permeable fluorescent stain DAF-FM diacetate was developed.This compound diffuses into cells and once inside, is cleaved by intracellular esterases to liberate weakly-fluorescent DAF-FM, which in turn becomes highly fluorescent upon reaction with NO or other specific RNS [47].To validate the use of this stain as an indirect indicator of intracellular NO levels in S. aureus, wild-type biofilms grown for 7 hours were subjected to DAF-FM diacetate staining, followed by treatment with HBSS buffer alone, DEA/NO (a short half-life NO donor), DEA alone, cPTIO (an NO scavenger), and DEA/NO + cPTIO (Figure 7A).As expected, treatment with DEA/NO resulted in a marked increase in DAF-FM fluorescence in S. aureus cells, whereas treatment with DEA alone did not affect fluorescence relative to the untreated condition.Interestingly, treatment of cells with cPTIO alone decreased intracellular DAF-FM fluorescence relative to untreated cells (buffer alone), suggesting that this NO scavenger was able to reduce the level of endogenous NO detectable by intracellular DAF-FM.Furthermore, when cPTIO-treated cells were subsequently exposed to NO donor, the fluorescent signal did not greatly increase compared to the untreated sample.Collectively, these results validate the use of DAF-FM diacetate as a fluorescent tool to indirectly measure intracellular NO levels in S. aureus.
DAF-FM diaceate was initially used to measure relative endogenous NO levels in late-exponential phase low-oxygen TSB cultures of the wild-type, nos mutant, and complement strains, as nos expression was increased at this phase of low-oxygen growth (Fig. 2).However, no detectable differences in DAF-FM fluorescence were observed (Fig. S2 in File S1), suggesting that saNOS activity may be minimal or very transient at this phase of growth.Alternatively, its contribution to endogenous NO in this growth condition may be masked relative to other cellular sources of NO/RNS.Since phenotypic changes in pigmentation were observed when these strains were cultured on TSA plates (Fig. 6A), 28-hour TSA plate cultures of wild-type, nos mutant, and complement strains were also assessed for endogenous NO levels by staining with DAF-FM diacetate (Fig. 7B).Intriguingly, DAF-FM fluorescence was significantly (p,0.05,Tukey test) decreased relative to the wild-type and complement strains, suggesting that saNOS enzyme activity and/or expression is increased in S. aureus when cultured on TSA plates.Given that the level of DAF-FM fluorescence only decreased by about 35% in the nos mutant, it is likely that other cellular sources of endogenous NO and/or RNS are also present that are able to react with DAF-FM in this assay.

Contribution of saNOS to survival in a murine sepsis model
Although the contribution of saNOS to MRSA virulence has been previously-demonstrated in a murine abscess model of infection [21], a role for this enzyme in virulence of MSSA has not been reported.Therefore, the virulence of MSSA strain UAMS-1 and its isogenic nos mutant was compared using a previouslypublished murine sepsis model of infection [35,36].As demonstrated in Figs.8A-8D, the bacterial load (as measured by CFU/ organ) in the lungs, kidneys, and liver were significantly (p,0.05,Mann-Whitney Test) lower in the nos mutant-infected mice compared to wild-type infected mice.Although not a statistically significant difference, bacterial counts from the heart were also decreased in nos mutant-infected mice relative to the wild-type infected mice.As well, the nos mutant-infected mice displayed significantly (p,0.05,Log Rank and Chi Squared Test with 1degree of freedom) less mortality than the wild-type infected mice in this model of infection, with 8/9 nos mutant-infected mice and 6/9 wild-type-infected mice surviving throughout the 7-day infection (Fig. 8E).Given that a partial polar effect on pdt expression was observed in the nos mutant when grown in lowoxygen TSB cultures (Fig. 3B), it is possible that the in vivo effects of the nos mutant observed in this study are due, in part, to partial loss of pdt expression.These results, combined with previouslypublished data showing that a MRSA nos deletion mutant displayed decreased virulence in an abscess infection model [21], suggest that saNOS and/or saPDT contribute to S. aureus dissemination and survival in vivo.

Discussion
An emerging body of evidence has demonstrated that bacterial NOS enzymes act as important regulators of cellular stress resistance and physiology.For example, NO-derived from the NOS of plant-pathogenic Streptomyces species is required for the nitration of thaxtomin phytotoxins, a modification that is essential for its phytotoxicity [12].The NOS enzymes of B. subtilis, B. anthracis, and S. aureus have also been shown to promote resistance to oxidative stress imposed by hydrogen peroxide [14,15,21] and by certain antibiotics [18,21].This oxidative stress resistance has been proposed to be mediated by several mechanisms, including transient suppression of the enzymatic reduction of intracellular cysteines by NO (and subsequent inhibition of the Fenton reaction) [14,15], NO activation of catalase activity [14,15], and NO-induced expression and/or increased activity of superoxide dismutase (SOD) [18,21].Our current results have corroborated the contribution of saNOS to hydrogen peroxide resistance in the MSSA strain UAMS-1, and furthermore, have also illustrated that carotenoid pigmentation is increased in a nos mutant when grown on TSA plates (Fig. 6A), a phenotype that has not been previously described for S. aureus nos mutants.Staphyloxanthin, the major carotenoid pigment synthesized by a series of enzymes encoded by the crt operon [43], has been shown to be a potent antioxidant defense against hydrogen peroxide and the immune cell respiratory burst [48,49], as well as conferring tolerance to host antimicrobial peptides [50] and fatty acids [51] by altering membrane fluidity.Therefore, at first glance, it appears paradoxical that a S. aureus nos mutant would be more sensitive to oxidative stress while displaying increased carotenoid pigmentation.However, the hydrogen peroxide challenge assay utilized in this study was performed during mid-exponential growth phase in planktonic cultures, a growth condition under which differences in staphyloxanthin production is unlikely to be observed, since maximal expression of crtM has been shown to occur in stationary phase in planktonic culture [46].Although loss of saNOS does not appear to alter expression of the crt operon and regulators of crt expression when grown on TSA plates (Fig. 6B), it is possible that NO derived from saNOS may regulate the activity of one or more of the enzymes involved in the synthesis of staphyloxanthin or one of its precursors.Another possibility stems from the fact that the degree of conjugation (increased number of double bonds) in carotenoids tends to correlate to the observed intensity of orange-yellow pigmentation [52].Since NO is known to have an affinity for biological membranes [53] and can attack the double bonds of b-carotene [54], it is possible that increased pigmentation observed in the S. aureus nos mutant is simply a result of loss of NO accumulation in the cell membrane.
In these presented studies, we have also optimized the use of DAF-FM diacetate to monitor relative intracellular NO levels in S. aureus.Using this technique, a decrease in intracellular NO/RNS was observed in the S. aureus nos mutant when cultured on TSA plates, a growth condition where increased pigmentation of the nos mutant was also observed.Although DAF-FM can also react with certain RNS such as nitrosonium ions, the majority of the fluorescent signal has been shown by others to be due to NO [55,56].Since the NO radical itself is relatively unstable and may quickly yield other RNS upon exposure to cellular components, intracellular DAF-FM fluorescence should be considered an indirect measurement of NO levels.In previous studies of S. aureus [21], B. anthracis [14], and B. subtilis [57] NOS activity, a NO-specific copper fluorescein (CuFL) stain [58,59] was utilized to detect intracellular NO production in wild-type and nos mutants.When used in S. aureus, the cells were incubated in a buffer containing arginine prior to CuFL staining [21] and it was not specified in this study whether excess arginine was necessary to detect NO production.Arginine addition may artificially induce nos expression and/or enzyme activity in cells not normally producing nos, and may overestimate the actual number of cells that contain high-level saNOS enzyme activity when growing in vivo.Furthermore, CuFL specifically interacts with NO to yield high-level fluorescence, but it may be more difficult to detect endogenous NO if it is produced in a short biological window or if it reacts quickly with other cellular targets to yield other RNS.In these cases the use of DAF-FM diacetate may provide a viable alternative method of intracellular fluorescent NO detection.
The protective effects of endogenous NO observed in bacteria in vitro likely translate to an important contribution of NOS to the virulence of these Gram-positive pathogens during human infection.For example, a B. anthracis nos mutant displayed decreased survival when engulfed by macrophages, and an increased lethal dose (LD 50 ) of spores compared to the parental strain in a murine model of systemic infection [14].Furthermore, a nos mutant generated in a highly virulent community-acquired MRSA (USA300) strain was shown to be more sensitive to killing by neutrophils by both intracellular and extracellular mechanisms, as well as decreased survival in macrophages [21].This MRSA nos mutant was also less virulent in a murine subcutaneous model of infection, displaying decreased abscess sizes and decreased recovered CFUs compared to the wild-type strain [21].The UAMS-1 nos::erm mutant strain used in our current study was also less virulent than its wild-type counterpart in a murine sepsis model of infection (Fig. 8).A partial polar effect on pdt expression was observed in this mutant when grown in low-oxygen TSB cultures (Fig. 3B), therefore it is possible that the in vivo effects observed in the sepsis model were due, in part, to partial loss of pdt expression.Given the interest in both NOS and PDT as potential drug development targets [22,23], our future studies will focus on assessing the individual contributions of nos and pdt to S. aureus virulence in vivo using non-polar deletion mutants.
As discussed above, S. aureus NOS clearly contributes to the oxidative stress response and virulence of this deadly pathogen.Considering the interest in this enzyme as a potential drug target, studies aimed at understanding the upstream regulators of both nos gene expression and saNOS enzyme activity are surprisingly lacking [22].In the study presented here, we have confirmed that S. aureus nos is co-transcribed with the downstream pdt gene, encoding a prephenate dehydratase which catalyzes the formation of phenylpyruvate from prephenate in the phenylalanine biosynthesis pathway.Although saNOS contributes to oxidative stress resistance in S. aureus, we have demonstrated that expression of nos-pdt is upregulated in response to low-oxygen growth and is maximally expressed at late-exponential growth phase.These observations suggest that one or both of these enzymes are required for an as-yet unknown function in the normal physiology of this organism during low-oxygen growth.Although the genetic regulators of nos-pdt expression have not yet been identified, a search of the S. aureus microarray meta-database (SAMMD) [60] has revealed that nos expression was upregulated in purH [46], lytS [61], and clpP [62] mutants, as well as by challenge with H 2 O 2 [40], and downregulated when challenged with chlorination [63], mupirocin [64], and various antimicrobial peptides [65].
The S. aureus PDT enzyme has also been postulated to be a potential drug target since it is involved in phenylalanine synthesis and is highly conserved among the staphylococci [23].As such, the enzymatic activity of purified recombinant S. aureus PDT has been demonstrated and the crystal structure of this protein has been solved [23].Although study of this gene in S. aureus has not been previously described, the data presented in this manuscript begin to fulfill this knowledge gap by confirming a role for saPDT as being required for growth in the absence of phenylalanine, as well as identifying low-oxygen growth as a signal that stimulates expression of pdt in vitro.RNA microarray analysis of S. aureus cultures previously revealed that nos and pdt were upregulated 2.5and 2.2-fold, respectively, in H 2 O 2 -treated cultures [40].Given these collective results, we hypothesized that saPDT, like saNOS, may also contribute to oxidative stress resistance, possibly by generating excess phenylpyruvate under these conditions.The aketo acid anions, such as b-hydroxypyruvate and b-phenylpyruvate, can chemically react with H 2 O 2 , and have been proposed to act as antioxidants by nucleophilic attack of the mono-deprotonated peroxide species at the C-2 carbonyl group carbon center [42].Furthermore, pyruvate, another a-keto acid, is commonly added as an H 2 0 2 -scavenger in certain cell culture media [66], and exogenous pyruvate is able to protect these cultured cells from H 2 O 2 -induced lysis [67].However, our data suggests that saNOS is not required for activity of saPDT when participating in phenylalanine biosynthesis, and saPDT itself does not appear to contribute to hydrogen peroxide resistance under the in vitro conditions tested in this study (Figs. 4 and 5).Therefore the potential functional and/or biological relationship between these two enzymes remains a mystery and subject of our on-going investigations.

Supporting Information
File S1 Contains the following files: Figure S1.Expression of the nos-pdt operon in S. aureus USA300 HOU as assessed by RNA-seq analysis.Data depicted in this figure was generated as part of a previously-published RNA-seq study [39], whereby RNA from 3 hour TSB cultures of strain USA300 HOU was subjected to Ion Torrent sequencing and resulting reads were aligned to the USA300 FPR genome.Arrows represent potential internal promoters and/or RNA processing sites.Figure S2.Detection of intracellular NO/RNS in low-oxygen cultures.UAMS-1 (pMK4), KR1010 (nos::erm mutant; pMK4), and KR1010 (complement; pMKnos) were grown as low-oxygen TSB cultures for 6.25 hours, followed by harvesting and resuspension in HBSS +5 mM DAF-FM dicaetate as described in Materials and Methods.Fluorescence (RFU) and OD 600 readings were recorded after 90 minutes incubation in a microplate reader.Data represents the average of n = 3 independent experiments, error bars = SEM.(PDF)

Figure 1 .
Figure 1.Genetic arrangement of the S. aureus nos and pdt genes.The representative bacterial genera and/or species that contain a nos gene are listed below the S. aureus nos-pdt operon diagram.The ''+'' designation under nos indicates its presence in each respective genome, whereas the ''+/2'' designation under pdt indicates if this gene is located immediately downstream of nos in each respective bacterial genera/species.The predicted molecular weight (kDa) of the nos and pdt ORFs are indicated above each gene, and the size (bp) of each gene is shown below.Gene designations refer to the MRSA252 genome.The site where the erm resistance cassette was inserted to create the nos::erm mutant (KR1010) is indicated by the black block arrow.*denotes that all sequenced strains in NCBI contain the indicated genes.doi:10.1371/journal.pone.0108868.g001

Figure 2 .Figure 3 .
Figure 2. Growth-phase dependent nos (A) and pdt (B) expression.Total RNA was isolated at the indicated time points from n = 3 biological replicates of UAMS-1 grown under highly aerated (light grey bars; 250 RPM, 1:10 volume-to-flask ratio) or low-oxygen (dark grey bars; 0 RPM, 7:10 volume-to-flask ratio) conditions.Quantitative realtime PCR was performed on reverse-transcribed cDNA from each sample using nos (A) or pdt (B) specific primers.The Livak (2 -DDCt ) method was used to determine relative fold change of nos and pdt expression, using measured sigA expression as the reference gene and the 2 hour aerobic sample as the calibrator.Error bars = standard error of the mean (SEM).*denotes statistical significance between aerobic and low-oxygen fold-change values at each indicated time point (Two-Tailed T-Test, p,0.01);Ndenotes statistical significance relative to t = 2 hour low-oxygen time point (Dunnett's Test, p,0.05).doi:10.1371/journal.pone.0108868.g002

Figure 7 .
Figure 7. Detection of intracellular NO/RNS with DAF-FM diacetate.A: Cells harvested from replicate UAMS-1 7 hour static biofilms were resuspended in HBSS containing 5 mM DAF-FM diacetate.After incubation for 1 hour at 37uC, cells were collected by centrifugation, washed, and resuspended in HBSS alone (''untreated'') or HBSS supplemented with 100 mM DEA or 100 mM DEA/NO.Where indicated, 150 mM cPTIO (NO scavenger) was added to cell suspensions during the 1 hour DAF-FM diacetate staining step.Aliquots (200 ml) of each cell suspension were immediately transferred to a 96-well plate, and incubated at 37uC in a Synergy HT fluorescent plate reader.Fluorescence and OD 600 measurements were recorded after 30 minutes, and data were reported as relative fluorescent units (RFU) per OD 600 of each well.Data represents the average of n = 3 independent experiments, error bars = SEM.*statistical significance compared to untreated UAMS-1 (p,0.05,Tukey Test).B: UAMS-1 (pMK4; wild-type), KR1010 (pMK4; nos mutant), and KR1010 (pMKnos; nos complement) were grown on agar plates for 26 hours, followed by harvesting and resuspension in HBSS +5 mM DAF-FM dicaetate.Fluorescence (RFU) and OD 600 readings were recorded after 90 minutes incubation in a microplate reader as described above.Data represents the average of n = 3 independent experiments, error bars = SEM.*statistical significance compared to UAMS-1 (pMK4) (p,0.05,Tukey Test).doi:10.1371/journal.pone.0108868.g007

Figure 8 .
Figure 8.Effect of nos on S. aureus virulence.1610 8 CFU/ml of UAMS-1 (wild-type) and KR1010 (nos mutant) cultures were each injected into the tail veins of female CD-1 Swiss mice (n = 9 mice per group).The infection proceeded for 7 days, or until mice reached a pre-moribund state (used as a measure of mortality).Mice were then euthanized and organs were collected and stored at 280uC prior to homogenization and determination of recovered CFU/organ.Graphs represent the calculated bacterial burdens for the kidneys (A), heart (B), lungs (C) and liver (D), and the percent survival (E).For A-D, data are graphed as box-whisker plots, indicating the 25 th , median, and 75 th percentiles (box lines), the 90 th and 10 th percentiles (error bars), and mean (+) of each data-set.*denotes statistically significant difference of mortality (p,0.05,Log Rank and Chi Squared Test with 1-degree of freedom), **denotes statistical significance of bacterial recovery (p,0.05,Mann-Whitney Test).doi:10.1371/journal.pone.0108868.g008

Table 1 .
Bacterial strains and plasmid constructs used in this study.

Table 2 .
PCR primers used in this study.