The PAPI-1 pathogenicity island-encoded small RNA PesA influences Pseudomonas aeruginosa virulence and modulates pyocin S3 production

Small non-coding RNAs (sRNAs) are post-transcriptional regulators of gene expression that have been recognized as key contributors to bacterial virulence and pathogenic mechanisms. In this study, we characterized the sRNA PesA of the opportunistic human pathogen Pseudomonas aeruginosa. We show that PesA, which is transcribed within the pathogenicity island PAPI-1 of P. aeruginosa strain PA14, contributes to P. aeruginosa PA14 virulence. In fact, pesA gene deletion resulted in a less pathogenic strain, showing higher survival of cystic fibrosis human bronchial epithelial cells after infection. Moreover, we show that PesA influences positively the expression of pyocin S3 whose genetic locus comprises two structural genes, pyoS3A and pyoS3I, encoding the killing S3A and the immunity S3I proteins, respectively. Interestingly, the deletion of pesA gene results in increased sensitivity to UV irradiation and to the fluoroquinolone antibiotic ciprofloxacin. The degree of UV sensitivity displayed by the PA14 strain lacking PesA is comparable to that of a strain deleted for pyoS3A-I. These results suggest an involvement of pyocin S3 in DNA damage repair and a regulatory role of PesA on this function.


Introduction
Bacterial small RNAs (sRNAs) have been recognized as key contributors to regulatory networks, and have been shown to play critical roles in many intra-and extracellular processes, and in pathogenesis [1][2][3]. Most sRNAs exert their regulatory function post-transcriptionally, acting by base pairing with the mRNA of their target genes ultimately modulating mRNA translation and/or stability. sRNAs can share extended base complementarity when they are cis-encoded on the opposite strand of the target mRNA, or they can interact with the target mRNA via short and imperfect base pairing, as in the case of trans-encoded sRNAs. The a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 expression of most sRNAs is responsive to environmental stress conditions spanning from iron and oxygen limitation, to oxidative, metabolic/nutrient, pH and cell envelope stresses [4].
In the opportunistic human pathogen Pseudomonas aeruginosa the use of transcriptomics approaches has recently led to the identification of numerous new sRNAs, mostly in the attenuated strain PAO1 and some also in the virulent one PA14 [5][6][7]. The bacterium P. aeruginosa is known as a major cause of both acute and chronic lung infections in patients belonging to all age groups who are immunocompromised, or who have defective mucociliary clearance, previous epithelial injury or foreign body placement [8]. Lung infections caused by P. aeruginosa can appear as a spectrum of clinical entities ranging from a rapidly fatal pneumonia in a neutropenic patient to a multi-decade bronchitis in patients with cystic fibrosis (CF). The expression of virulence traits in P. aeruginosa is fine-tuned by a dynamic and intricate regulatory network [9],that leads the expression of P. aeruginosa pathogenic functions with a sharp timing. In this scenario, sRNAs can finely contribute to P. aeruginosa ability to quickly adapt to a new environment and manage to persist.
Previously, bioinformatics and approaches such as the pull-down with the RNA chaperone Hfq have been used for identifying novel P. aeruginosa sRNAs but only a small number of them have been characterized functionally, and even less described as being implicated in the regulation of P. aeruginosa virulence [10][11][12][13][14][15][16][17]. Moreover, most of the studies have been performed using the reference strain PAO1, while very little is known about the biological effects of the sRNA mediated regulation in virulent P. aeruginosa strains, such as PA14. Compared to PAO1, the clinical isolate PA14 is significantly more virulent in a wide range of hosts, including mice, the nematode Caenorhabditis elegans, the insect Galleria mellonella, and the plant Arabidopsis thaliana, and has thus become an important reference strain because of its enhanced virulence [18][19][20]. Generally, PAO1 and PA14 strains share highly conserved genomes, although PA14 possesses a slightly larger one, likely due to horizontal gene transfer resulting in the acquisition of pathogenicity islands (PAIs) [21].
This study focuses on a P. aeruginosa sRNA observed for the first time by a sRNA-sequencing approach [5], in which unique and conserved sRNAs in the P. aeruginosa strains PAO1 and PA14 were revealed. Here, we validated the sRNA with the operative name SPA0021, renamed PesA, as being transcribed from the pathogenicity island PAPI-1, present in PA14 strain, but not in PAO1. In addition, our results display that PesA is expressed in several P. aeruginosa isolates, including environmental and clinical ones isolated from CF patients. Moreover, we show that PesA operates a post-transcriptional regulation of genes involved in S-type pyocin production. Pyocins are narrow-spectrum bacteriocins synthesized by more than 90% of P. aeruginosa strains and presumed to play a role in niche establishment and protection in mixed populations. The pyocin genes are usually located on the P. aeruginosa chromosome and induced by mutagenic agents that cause DNA damage such as mitomycin C and UV irradiations. P. aeruginosa pyocins can be subdivided into insoluble R and F pyocins and soluble S pyocins. S-type pyocins AP41, S1, S2, and S3 are constituted of two components in which the large component carries an endonuclease C-terminal domain responsible for the killing activity that causes cell death by DNA cleavage [22][23][24][25][26][27], while pyocins S4 and S5 have tRNAse and pore-forming activities, respectively [28,29]. More recently, also the new S-type pyocin S6, with rRNase activity, has been described functionally [30]. In our study, we show that PesA deletion leads to decreased expression of pyocin S3, associated to increased sensitivity to UV irradiation and to the fluoroquinolone antibiotic ciprofloxacin. Furthermore, PesA is induced by host-environment stimuli such as low oxygen availability and body temperature, suggesting a key role in P. aeruginosa adaptability to different environmental stresses. Finally, we also found that PesA is involved in the regulation of P. aeruginosa PA14 virulence in CF human bronchial epithelial cells. Our results suggest that a PAPI-1-encoded sRNA can contribute to the modulation of the expression of genes outside PAPI-1, and to different aspects of P. aeruginosa pathogenesis during infection.

Ethics Statement
Study on human P. aeruginosa isolates from Hannover has been approved by the Ethics Commission of Hannover Medical School, Germany [31]. The patients and parents gave oral informed consent before the sample collection. Approval for storing of biological materials was obtained by the Ethics Commission of Hannover Medical School, Germany. The study on human P. aeruginosa isolates from the Regional CF Center of Lombardia (S1 Table)

Bacterial strains and culture conditions
Bacterial strains and plasmids used in this study are listed in S2 Table. E. coli strains were routinely grown in Luria-Bertani broth (LB) at 37˚C. P. aeruginosa strains were grown at 37˚C in Brain Heart Infusion (BHI) rich medium or in LB at 120 rpm unless otherwise indicated. Carbenicillin and gentamicin were added at 300 and 20 μg/ml, respectively, unless otherwise indicated. For P BAD induction in vector plasmid pGM931, arabinose was added to a final concentration of 10 mM. Anaerobic batch cultivations of P. aeruginosa and the shift from aerobic to anaerobic conditions were performed in a 800 ml-Biostat-Q system bioreactor (B-Braun) as described previously [17].
P. aeruginosa mutant in pesA gene was generated by an enhanced method of markeless gene replacement described previously [36] with some modifications to adapt it to P. aeruginosa as described previuosly [17]. PA14 mutant in pesA gene was obtained by allelic exchange with a deletion from-34 to + 140 positions with respect to pesA transcription start site as follows. TS1 region spanning left 533 bp flanking sequence of pesA gene was amplified by PCR with oligos 23/24. TS2 region spanning last 127 nt and right 368 bp flanking sequence of pesA was amplified by PCR with oligos 25/26. PCR amplifications were performed from PA14 genomic DNA. Overlap extension (SOE)-PCR with oligos 23/26 was used to join TS1 and TS2 that carried end complementary regions introduced by 25/26, respectively during their separate PCR amplification [37]. Joined TS1-TS2 DNA fragments were digested with EcoRI-PstI and cloned in CC118 λpir into the poly-linker site of pSEVA612S giving rise to pSEVApa14-ΔpesA.
The TS1-TS2-inserted pSEVApa14-ΔpesA was transferred from E. coli CC118 λpir to PA14, with the assistance of the helper E. coli strain HB101(pRK600) in a conjugative triparental mating. Exconjugant P. aeruginosa clones were selected on M9-citrate with 60 μg ml −1 of gentamicin. Since pSEVA612S derivatives cannot replicate in P. aeruginosa, Gm R exconjugant clones could appear only by co-integration of the construct in the genome of the recipient strain by homologous recombination between joined TS1-TS2 fragments borne by pSE-VA612S and the recipient chromosome. Plasmid pSW-1 was transferred from E. coli DH5α to P. aeruginosa clones bearing genomic co-integrates of pSEVApa14-ΔpesA by triparental mating as above, and pSW-1-recipient P. aeruginosa clones were selected on M9-citrate with 300 μg ml −1 of carbenicillin. Cultures of resulting P. aeruginosa clones carrying pSW-I were grown overnight in LB with 300 μg ml −1 of carbenicillin and then plated on the same medium. Single colonies were screened for loss of gentamicin resistance. Gentamicin-sensitive clones carrying the deleted alleles were then screened by PCR with oligo pairs 27/28 for pesA.
All plasmid constructs and deletion mutant were checked by sequencing with oligos indicated in S3 Table. RNA isolation and analysis Total RNA was prepared as described previously [5] from 2-10 ml of bacterial cell cultures. Quality and concentration of the RNA extracted were assessed by a Biospectrometer (Eppendorf). Northern blot analyses were performed as described previously [5]. DNA oligonucleotide probes were 5'-end labeled with [γ-32 P]ATP (PerkinElmer, NEG502A) and T4 polynucleotide kinase (Promega, M4103) according to manufacturer's instruction. Oligo 1 and 2 were used to probe PesA and 5S RNA, respectively. Treatment with terminator-5'-phosphate-dependent exonuclease was performed in terminator reaction buffer A (Epicentre, TER51020) as described in detail previously [17].
Radioactive bands were acquired after exposure to phosphor screens using a Typhoon™ 8600 variable mode Imager scanner (GE Healthcare BioSciences) and visualized with image-Quant software (Molecular Dynamics).
Quantitative RT-PCR analysis was performed on total RNA extracted from P. aeruginosa PA14 wild-type and ΔpesA at mid-, late-exponential and stationary phase (OD 600 of 0.8, 1.6 and 2.7, respectively). cDNA synthesis was performed from 1 μg of total purified RNA using QuantiTect Reverse Transcription Kit (Qiagen). RT-PCRs were performed using QuantiTect SYBR 1 Green PCR Kit (Qiagen) and oligo pairs 30/31, 32/33, 34/35 for 16S, pyoS3A and pyoS3I amplification, respectively. The reaction procedure involved incubation at 95˚C for 15s and 40 cycles of amplification at 94˚C for 15 s, 60˚C for 30 s and 72˚C for 30s. 16S ribosomal RNA was used as reference.
In vitro and vivo assays of sRNA/mRNA interactions Electrophoretic Mobility Shift Assay (EMSA) to analyze sRNA/mRNA interactions were performed as described previously [17].
Experiments with fluorescent reporters were carried out as described previously [17]. Fluorescence polarization FP 485/535 , fluorescence intensity FI 590/635 and Abs 595 were measured in a Tecan Infinity PRO 200 reader, using Magellan as data analysis software (Tecan). GFP and mCherry activities were expressed in Arbitrary Units (AU) as ratio FP 485/535 /Abs 595 and FI 590/635 /Abs 595 , respectively.

RNA synthesis
RNAs for RNA/RNA interaction assays were prepared by T7 RNA polymerase transcription of gel-purified DNA fragments obtained by PCR as described previously [17]. DNA fragments for PesA RNA and pyoS3A-I mRNA preparations were amplified from P. aeruginosa PA14 genomic DNA with oligo pairs 3/4 and 5/6, respectively. The DNA fragment for RseX RNA was amplified from E. coli C1a genomic DNA with oligos 7/8.

Pyocin S3 spotting assay
Pyocin killing assay was performed using the spotting method as indicated previously [38] with some modifications. 10 μl of filter-sterilized supernatants from cell cultures with OD 600 of 1 were spotted onto LB 1.5% agar plates. A lawn of the pyocin S3 sensitive P. aeruginosa strain ATCC 27853 containing 5 × 10 6 cells ml −1 was plated by inclusion into 0.3% soft agar over the dried spots. Plates were incubated overnight at 37˚C and checked for the formation of the clearing zone on the spotting site, which is indicative of the pyocin S3 activity.

Antibiotic disk diffusion
Susceptibilities of P. aeruginosa strains to antimicrobial agents were analyzed by disk diffusion measurement. Filter disks (Oxoid, CT0425B, CT0013B, CT0207B, CT0052B, CT0058, CT0010B) were placed on a lawn of 10 6 CFU/ml bacterial cells plated by inclusion into 0.3% LB agar. Plates were incubated overnight at 37˚C and the diameters of the clear zones around the disks were measured.

UV sensitivity assay
UV treatment was performed using a Stratalinker 1800 UV Crosslinker (Stratagene). Cell cultures with an OD 600 of 1 (corresponding to 8 x 10 8 CFU/ml) were serially diluted until 10 −7 ; 3 μl of each dilution were spotted in triplicate onto LB-agar plates, dried, and exposed to increasing amounts of UV radiation, from 0 to 100 J/m 2 . Plates were incubated overnight at 37˚C.
Cytotoxicity assays in human CF respiratory cells IB3-1 cells, an adeno-associated virus-transformed human bronchial epithelial cell line derived from a CF patient (ΔF508/W1282X) and obtained from LGC Promochem, were grown as described previously [39]. Cells were infected with P. aeruginosa strains at a multiplicity of infection (MOI) of 100. Cell viability was evaluated using the CellTiter 96 1 Non-Radioactive Cell Proliferation Assay (MTT) kit (Promega, G4000), according to manufacturer's instructions.

Bacterial isolates analysis
Bacterial isolates were plated on 1.5% BHI-agar plates and grown overnight at 37˚C. Culture samples were taken and processed for genomic DNA and total RNA extraction. PAO1 and PA14 strains treated in the same conditions were used as controls. Oligos 9/10 and 37/38 were used for PCR-amplification of the genomic region containing the pesA and 16S (as positive PCR-control) loci, respectively.

Results
The sRNA PesA is encoded in the PAPI-1 and is a processed transcript The SPA0021 sRNA was identified using a comparative sRNA-seq approach in which 52 P. aeruginosa novel sRNAs have been identified either in the attenuated strain PAO1 or in the virulent one PA14. SPA0021 was validated to be % 260 nt in lengths, and one of the 12 sRNAs whose genetic locus is unique to the PA14 strain [5]. Moreover, SPA0021 was found to be encoded within the pathogenicity island PAPI-1, and because of this property, we renamed it as pathogenicity island-encoded sRNA A (PesA). The pesA gene locates downstream and on the same strand of pilM2 gene (belonging to the type IV B pilus operon pil2), and overlaps the 3' of PA14_59370 (gene with unknown function) (Fig 1A and 1B). Since a rho-independent transcription terminator was predicted within the PA14_59370 sequence [40]. The cluster of the sRNA-seq reads that mapped upstream the predicted terminator was considered as the 3' of the pesA gene [5]. The 5'-end was also mapped by the sRNA-seq read-clustering and was in perfect agreement with the size validated by Northern blot [5].
We evaluated PesA expression along the growth-curve in the rich medium BHI (Fig 1C). Northern blot analyses showed a main band of the expected product size of % 260 nt (black arrow) at each analyzed time point, and multiple bands with higher molecular weight (white arrows) especially at early exponential phase (OD 600 = 0.2). The main band of PesA showed to accumulate in late stationary phase. The presence of high molecular weight multiple bands, and the absence of rho-independent transcription terminators downstream pilM2, suggested that the main band of % 260 nt could derive from the processing of a longer transcript. Sensitivity to treatment with terminator 5 0 -phosphate-dependent exonuclease (Fig 1D), which preferentially degrades monophosphate processed transcripts confirmed that the % 260 nt RNA is indeed a processed product.

PesA is widespread expressed in clinical and environmental isolates
To assess the clinical impact of PesA, we validated its presence and expression levels throughout a collection of 29 clinical P. aeruginosa isolates derived from respiratory samples of patients with chronic respiratory diseases, including CF and chronic obstructive pulmonary disease (COPD), and 5 isolates from environmental habitats. In particular, the clinical isolates were recovered both during intermittent infections and at different stages of chronic lung infection. Part of them was previously characterized both in vitro and in vivo [31,41]. We also included in this study the Liverpool epidemic strain LESB58 [42] and PAO1 and PA14 as controls.
In Fig 2, results on pesA-gene amplification and Northern blot analysis show that the pesA gene is present in 17 out of the 27 CF clinical isolates, in the 2 COPD isolates and in 2 out of 5 environmental isolates. Notably, there was no association between the presence/expression of pesA gene and the P. aeruginosa status (chronic vs intermittent). In addition, no differences were observed between clonal isolates recovered from the same patients at different stages of

PesA is induced in anaerobic growth and at 37˚C
We analyzed the transcriptional responsiveness of PesA RNA to environmental or body temperature, and reduced oxygen availability. Temperature sensitivity was tested by probing PesA in early-(OD 600 = 0.8) and mid-exponential phase (OD 600 = 1.8) at both 20 and 37˚C and after 20 min of acclimation following a shift from 20 to 37˚C, as described in detail previously [17]. As shown in Fig 3A, the growth at 37˚C caused an up-regulation of PesA if compared to the growth at 20˚C; no increase of PesA RNA accumulation was observed during the 20 min of acclimation. PesA showed also to be responsive to oxygen availability. PesA was probed at mid-and late-exponential phase (OD 600 of 0.8 and 2, respectively) under anaerobic conditions in BHI with nitrate to sustain anaerobic respiration, as described previously [17]. In addition, bacterial cells were grown in BHI with aeration until mid-exponential phase (OD 600 = 0.8); then, oxygen was excluded from cultures. PesA levels were assessed immediately before oxygen exclusion and 20 (OD 600 = 0.9) and 150 min (OD 600 = 1.3) from the start of anaerobic conditions, as described previously [17]. PesA levels were higher in anaerobic than aerobic conditions both in mid-and late-exponential phase (Fig 3B). In addition, the shift from aerobic to anaerobic conditions caused a progressive increase of PesA levels.

PesA is involved in ciprofloxacin and UV-resistance
To explore the involvement of PesA in the regulation of cellular mechanisms, also linked to P. aeruginosa virulence, we constructed the knock-out mutant strain PA14 ΔpesA and the plasmid vector pGM-pesA carrying the pesA gene under the arabinose inducible P BAD promoter with the aim to measure the effects of perturbing PesA levels on phenotypic traits of the PA14 strain. Deletion of the pesA gene and overexpression from the pGM-pesA vector in the wildtype background were ensured by Northern blot (S1 Fig).   Fig 2. PesA gene dissemination and expression levels among environmental, CF and COPD clinical isolates. Assays on environmental, CF and COPD isolates are shown in three panels, A, B and C, respectively. The strain-collection was plated on BHI-agar plates. After overnight growth at 37˚C, culture samples were taken and processed for total RNA extraction and analysis by Northern blot, and for genomic DNA extraction. Positive or negative PCR-amplification outcomes are indicated as "+" or "-" in the "pesA gene amplification" row, below each Northern Blot. PAO1 and PA14 were used as controls of Northern Blot analysis and for negative or positive gene amplification, respectively. https://doi.org/10.1371/journal.pone.0180386.g002 PesA, a PAPI-1 encoded small RNA of P. aeruginosa We performed different types of phenotype evaluations. The most evident effects of PesA deletion were on UV and ciprofloxacin susceptibility. In particular, PesA deletion resulted in an enhanced sensitivity to the antibiotic ciprofloxacin and to UV irradiation. In fact, the susceptibility of wild-type PA14 and ΔpesA mutant cells to antimicrobials was analyzed by antibiotic disk diffusion on agar plates. The diameters of the inhibitory zones were measured after overnight incubation at 37˚C. As shown in Fig 4, the diameter of the clear zone around ciprofloxacin was higher for mutant ΔpesA (46.33 ±0.58 mm) in comparison to that of wildtype strain (40.33 ± 0.58 mm) thus suggesting a contribution of PesA in the ciprofloxacin resistance mechanism. In addition, we noticed an incremented sensitivity of the ΔpesA mutant strain to UV light, showing a decrease in CFUs with respect to the wild-type starting from treatment with 30 J/m 2 (Fig 5). This suggested the involvement of PesA in improving survival under conditions of genotoxic stress, such as UV irradiation treatment. Intriguingly, the deletion of pyoS3A-I operon, that is positively regulated by PesA (see below), gave rise to same levels of UV sensitivity as pesA deletion (Fig 5).
We did not observe significant differences for other phenotypes analyzed, including growth-curves analysis on rich and minimal media, susceptibility to other antimicrobial agents of different structural families, hemolytic activity, flagellum-mediated motility, pyocyanin and pyoverdine secretion.

PesA is involved in P. aeruginosa pathogenicity in human CF respiratory cells
We evaluated the virulence of P. aeruginosa PA14 strains on the CF bronchial epithelial cell line IB3-1. In particular, we assessed the killing capacity of PA14 ΔpesA mutant compared to the wild-type strain by the MTT assay, which provides a method of determining viable cell number measuring the conversion of 3-(4,5-dimethylthiazol)-2,5-diphenyltetrazolium bromide (MTT) to insoluble formazan by dehydrogenase enzymes of the intact mitochondria of living cells. Our results (Fig 6) showed that cells infected with P. aeruginosa PA14 ΔpesA mutant were more viable with respect to those infected by the wild-type strain, thus indicating that pesA may contribute to P. aeruginosa PA14 acute virulence.
PesA targets pyoS3A-I operon As mentioned previously, PesA is encoded in cis to the 3' of the gene PA14_59370 with unknown function, which makes the study of such putative target difficult to perform. Therefore, we managed to identify direct targets of PesA RNA by the use of the bioinformatics tool TargetRNA [43]. This tool predicted, as a high-scored output, an interaction in the region from -30 to -8 nt upstream the gene pyoS3I of the pyoS3 operon (Fig 7A and S4 Table), predicted to encode pyocin S3 in the PA14 strain. The pyocin S3 genetic locus comprises two structural genes, pyoS3A (PA14_49520) and pyoS3I (PA14_49510), annotated to encode the killing S3A and the immunity S3I proteins, respectively. To confirm this annotation in PA14, we deleted the pyoS3 operon and tested the pyocin S3 production by the killing assay of the sensitive ATCC 27853 P. aeruginosa strain [27]. As shown in Fig 8, deletion of pyoS3A-I completely abolished the production of pyocin S3 by PA14 strain. Remarkably, PesA deletion resulted in strong reduction of pyocin S3 production, thus suggesting a positive role of PesA on the pyocin S3 production.
The predicted region of the pyoS3A-I operon targeted by PesA comprises the Ribosome Binding Site (RBS) of the pyoS3I gene, and locates within the ORF of the pyoS3A gene. To assess this predicted PesA-pyoS3 mRNA interaction, PesA RNA and the pyoS3 mRNA region spanning -116 to +108 from pyoS3I translational start site were produced in vitro, mixed and analyzed on native polyacrylamide gels. As shown in Fig 7B, the two RNAs specifically formed a complex.
We generated distinct types of translational fusions to test the effects of PesA on the pyoS3I gene alone and on the pyoS3A-I mRNA. A first reporter plasmid, named pBBR1-lacZ:: pyoS3A-I::sfGFP, mimics a bi-cistron under the control of the heterologous constitutive promoter P LtetO-1 . It was obtained by cloning a region of 224 nt, comprehensive of the last 117 nt of the pyos3A gene and the first 108 nt of pyoS3I thus generating a first translational fusion of the reporter F'lacZ with the last 39 codons of pyoS3A, and a second translational fusion of the first 36 codons of pyoS3I gene with sfGFP. GFP activity was assayed in PA14 wild-type and PA14 ΔpesA in the absence and presence of PesA overexpression from pGM-pesA. As shown in Fig 9A, there was an approximately 25% reduction in GFP activity in the PA14 ΔpesA background. In the presence of pGM-pesA overexpressing PesA, in wild-type background, GFP activity increased approximately 30%. These results suggested that PesA positively regulates pyoS3I expression. To test whether translation of pyos3A gene was necessary to have this PesA effects on pyos3I, we generated a second mono-cistronic reporter derivative, pBBR1-pyoS3I:: PesA, a PAPI-1 encoded small RNA of P. aeruginosa sfGFP, carrying the same pyoS3A-I region as pBBR1-lacZ::pyoS3A-I::sfGFP in which only the pyoS3I gene was translationally fused with sfGFP. GFP analyses confirmed that PesA influence in a positive manner the regulation of pyoS3I in the mono-cistronic construct, with a~20% increment in GFP activity in the presence of PesA overexpression in the wild-type background, and a~40% decrease in the PA14 ΔpesA background (Fig 9B). This suggested that the translation of the two genes is not merely and solely coupled and that the effect of PesA on pyoS3I do not require the translation of pyoS3A.
To assess the influence of PesA also on the pyoS3A gene, we substituted the sequence coding for the F'LacZ domain in pBBR1-lacZ::pyoS3A-I::sfGFP with the one of the reporter gene mCherry, and monitored simultaneously the activity of both mCherry and sfGFP of the translational fusion mCherry::pyoS3A-S3I::sfGFP. As shown in Fig 9C, the~50% increase in sfGFP activity followed by PesA overexpression, and the~25% decrease in the PA14 ΔpesA background reconfirmed the positive regulation exerted by PesA on the pyoS3I gene. mCherry also showed an increase in activity when PesA was overexpressed, suggesting that PesA also exerts a positive regulation on the upstream gene pyoS3A. Thus, PesA seems to be a positive regulator of the whole pyoS3 operon.
Quantitative RT-PCR on total mRNA of wild-type and ΔpesA strains was also performed to check whether PesA influenced pyoS3 mRNA levels. Samples were taken at mid-, late-exponential and stationary phase (OD 600 of 0.8, 1.6 and 2.7, respectively) and both genes, pyoS3A and pyoS3I, were analyzed for their expression levels. No significant differences were observed either for pyoS3A or pyoS3I in wild-type and ΔpesA backgrounds at any time-point. Notably, the expression levels of the two genes were comparable at every time-point and seemed not to be influenced by the growth phase, being constant along the growth curve (S5 Table). We conclude that PesA exerts positive regulation on pyoS3A-I mRNA by modulating mRNA translatability and without influencing its stability. PesA, a PAPI-1 encoded small RNA of P. aeruginosa By the use of the IntaRNA web tool [44,45], we also detected a putative interacting region between PesA and the leader sequence of the pyoS3 operon, from -76 to -42 from the TTG start codon of the pyoS3A gene. To evaluate whether PesA was also able to influence the S3-operon translation by acting on its leader sequence, we generated the translational fusion leader-pyoS3A::sfGFP, by cloning the whole leader region of the pyoS3 operon and the sequence encoding the first 37 aminoacids of pyos3A gene, in frame with the sfGFP reporter.

Discussion
We studied the novel P. aeruginosa sRNA PesA, which was originally identified as being transcribed from the horizontally acquired pathogenicity island PAPI-1 in the strain PA14. Our analysis revealed that PesA is widespread in clinical isolates from patients affected by chronic respiratory diseases, such as CF, being expressed in 55% of the cases tested. Moreover, PesA expression is responsive to low oxygen conditions, a hallmark of CF and COPD, and impacts P. aeruginosa pathogenicity in CF bronchial cells. These results suggest that PesA could be relevant during P. aeruginosa infection in chronic respiratory diseases. PesA, a PAPI-1 encoded small RNA of P. aeruginosa  We speculated that PesA had the potential to act as a trans-encoded base-pairing sRNA involved also in the post-transcriptional regulation of genes located outside PAPI-1 and thus performed genome-wide bioinformatics screenings for target genes. One predicted target was the pyoS3A-I operon coding for pyocin S3. We then confirmed that both pyoS3A and pyoS3I genes are positively influenced by PesA. The interaction of PesA with the bicistronic pyoS3A-I mRNA is suggested to simultaneously stimulate translation initiation and termination of pyoS3I and pyoS3A, respectively, without influencing stability of the pyoS3A-I mRNA. Interestingly, PesA could impact a putative mechanism of translation coupling between pyoS3I and sfGFP combined with the control vector (pGM931) or the plasmid overexpressing PesA (pGM-pesA), in PA14 wildtype and PA14 ΔpesA. The strains were grown to an OD 600 of 1.8 in LB medium supplemented with gentamicin and carbenicillin, to maintain pBBR1-and pGM-plasmids, respectively, and arabinose, to induce PesA overexpression. Cells were harvested and treated for sfGFP and mCherry activity determination by measuring fluorescence polarization FP 485/535 and fluorescence intensity FI 590/635 , respectively. sfGFP and mCherry activities are expressed as ratio FP 485/535 /Abs 595 and FI 590/635 /Abs 595 , respectively. Data derive from three independent experiments. Values represent the mean ± SD. Statistical significance by Student's t-Test is indicated: *p<0.05; **p< 0.01. https://doi.org/10.1371/journal.pone.0180386.g009 PesA, a PAPI-1 encoded small RNA of P. aeruginosa pyoS3A [22] that remains to be elucidated. However, in this paper we demonstrate that stimulation of pyoS3I by PesA does not require translation of the upstream pyoS3A thus suggesting that the two genes are not strictly translationally coupled. This scenario, as a whole, is compatible with the role of PesA in assuring balanced expression of toxin S3A and antitoxin S3I to prevent deleterious effects on the producing host. Overall, these data obtained with the reporter genes are consistent with the observation that PesA deletion results in strong reduction of pyocin S3 production (Fig 8).
It is conceivable that PesA can be involved in mechanisms of niche establishment via pyocin S3. PesA was in fact also detected in environmental isolates and its expression from PAPI-1 might confer a selective advantage that favours PAPI-1 maintenance in the environment due to its regulation of pyocin S3.
It is likely that PesA has a broad set of target genes whose functions go beyond the niche establishment. The observation that PesA deletion induces less killing in infected CF bronchial epithelial cells suggests that PesA could modulate, directly or indirectly, virulence factors of P. aeruginosa. It was previously shown that pyocin S2 is endowed with cytotoxic activity on human cell lines [46]. It can be argued that pyocin S3 has similar effects and can act as virulence factor whose regulation is under the control of PesA.
Furthermore, PesA could regulate the expression of genes involved in DNA damage repair as suggested by the increased sensitivity of ΔpesA mutant to fluoroquinolone antibiotic ciprofloxacin and to UV irradiation. Intriguingly, the degree of UV sensitivity displayed by the ΔpesA mutant is comparable to that of a strain deleted for pyoS3A-I. These results imply a potential involvement of the DNase activity of pyocin S3 in DNA damage repair, and introduce an intriguing network among sRNAs, pyocins and DNA damage repair that will require additional experiments to be elucidated. To our knowledge this is the first work characterizing a sRNA encoded in a pathogenicity island in P. aeruginosa. In addition, our results indicate that PesA is able to modulate key genes located outside the PAPI-1. In summary, the horizontal acquisition of PAPI-1 could provide the new host with a regulatory function that can switch the expression of genes involved in niche establishment, virulence and stress resistance.

S1 Fig. Validation of PesA deletion and overexpression.
A) PA14 wild-type and PA14 ΔpesA were grown in BHI medium until an OD 600 of 2.7. Culture samples were taken and processed for total RNA extraction and analysis by Northern blot. B) PA14 strains harbouring pGM-pesA or the control empty vector pGM931 were grown in BHI medium with carbenicillin until an OD 600 of 0.8. Cells were split into two flasks, and 10 mM arabinose (ara) was added to one. Culture samples were taken at the indicated OD 600 and processed for total