Blocking RpoN reduces virulence of Pseudomonas aeruginosa isolated from cystic fibrosis patients and increases antibiotic sensitivity in a laboratory strain

Multidrug-resistant organisms are increasing in healthcare settings, and there are few antimicrobials available to treat infections from these bacteria. Pseudomonas aeruginosa is an opportunistic pathogen in burn patients and individuals with cystic fibrosis (CF), and a leading cause of nosocomial infections. P. aeruginosa is inherently resistant to many antibiotics and can develop resistance to others, limiting treatment options. P. aeruginosa has multiple sigma factors to regulate transcription. The alternative sigma factor, RpoN (σ54), regulates many virulence genes and is linked to antibiotic resistance. Recently, we described a cis-acting peptide, RpoN*, which is a “molecular roadblock”, binding consensus promoters at the -24 site, blocking transcription. RpoN* reduces virulence of P. aeruginosa laboratory strains, but its effects in clinical isolates was unknown. We investigated the effects of RpoN* on phenotypically varied P. aeruginosa strains isolated from CF patients. RpoN* expression reduced motility, biofilm formation, and pathogenesis in a P. aeruginosa-C. elegans infection model. Furthermore, we investigated RpoN* effects on antibiotic susceptibility in a laboratory strain. RpoN* expression increased susceptibility to several beta-lactam-based antibiotics in strain P. aeruginosa PA19660 Xen5. We show that using a cis-acting peptide to block RpoN consensus promoters has potential clinical implications in reducing virulence and improving antibiotic susceptibility.


Results
Virulence phenotypes were variable in P. aeruginosa isolates from CF patients. P. aeruginosa isolated from different CF patients or within the same CF patient have varied phenotypes and genotypes 36,37 . P. aeruginosa adapts over time, leading to mutations and changes in expression of genes related to motility, quorum sensing, and overall virulence 38,40 . To determine the virulence-related phenotypic profiles of the strains used in this study (Table 1), each P. aeruginosa patient isolate was evaluated for motility and biofilm formation, compared to the virulent positive control strain P. aeruginosa PA19660 Xen5. Several patient isolates were highly motile  forming the most extensive biofilm (Fig. 1C). SCH0254-116, SCH0397-3, and UUH0202 did not form biofilms. The pathogenesis of patient isolates was evaluated in a P. aeruginosa -C. elegans infection model. All patient isolates were compared to E. coli OP50, an avirulent negative control. SCH0057-7 was the most pathogenic in the paralytic killing assay, which is mediated by hydrogen cyanide production 41,42 (Fig. 2A). Other strains were moderately pathogenic, including SCH0256-1, SCH0354-1, SCH0397-3, and UUH0202. SCH0057-7, SCH0338-58, and UUH0202 were highly pathogenic in the slow killing assay, which mimics establishment and proliferation of an infection and is mediated by the lasR, gacA, lemA, and ptsP genes 43 , while UUH0201 was moderately pathogenic (Fig. 2B). As expected, the virulence-associated phenotypes of patient isolates varied widely in vitro and in vivo.

RpoN protein levels varied among patient isolates.
Others reported that the rpoN gene was mutated or lost in approximately 20% of P. aeruginosa isolates from CF patients 38 . Loss or mutation in the rpoN gene can result in phenotypes similar to those observed in the patient isolates evaluated here 20,22,23 . Thus, we evaluated relative protein levels of RpoN in these patient isolates by western blot. RpoN levels were moderately high in the positive control P. aeruginosa PAO1-S, while low or minimal protein levels were detected in the isogenic ΔrpoN mutant negative control (Fig. 3). The faint bands in the ∆rpoN strain and several CF patient isolates are background signals due to non-specific antibody binding to another protein or sigma factor with a similar apparent molecular weight. RpoN levels varied in the CF patient isolates, with high levels in SCH0057-7, SCH0397-3, and UUH0201; intermediate levels in SCH0254-116, SCH0338-58, and UUH0202; and background levels in SCH0254-23, SCH0254-118, SCH0256-1, SCH0354-1, SCH03269, and UUH0101. RpoN* expressed in CF patient isolates reduced virulence-associated phenotypes in vitro. The effect of RpoN* expression on motility and biofilm formation in patient isolates was not known. Unfortunately, some patient isolates could not be transformed, and so only four isolates were evaluated for the effects of RpoN* expressed from a plasmid. SCH0057-7, SCH0256-1, SCH0338-58, and SCH0354-1 were transformed with a plasmid expressing RpoN* or the empty vector and selected with gentamicin. If RpoN* affected transcription of virulence-related genes in different genetic backgrounds as previously reported 32 , we expected attenuation of virulence-related phenotypes in P. aeruginosa CF patient isolates. RpoN* significantly reduced colony diameter in all four patient isolates in the swimming motility assay (Student's t-test, **p ≤ 0.01, ***p ≤ 0.0001) (Fig. 4A, Supplemental Fig. S3 (top two rows)). RpoN* significantly reduced colony diameter in SCH0057-7, Figure 1. Characterization of virulence phenotypes of P. aeruginosa strains isolated from cystic fibrosis patients. P. aeruginosa CF patient isolates were compared to laboratory strain PA19660 Xen5. All assays were conducted at 37 °C for 24 h. (A) Colony diameter of swimming, or flagellar, motility assay conducted on soft (0.3%) agar. (B) Twitching, or pili, motility assay conducted on semi-hard (1.3%) agar. Colony diameter was measured across point of inoculation to the edges of bacteria colony. (C) Biofilm formation assay was conducted in 96-well microtiter plates. Biofilms were stained with crystal violet (0.1%), solubilized in ethanol (95%), and absorbance measured at OD 550 . Bars are the mean ± SD; n = 5 to 6 replicates in motility assays and n = 10 in biofilm assay. Each assay was performed at least three separate times and representative results are shown.

RpoN* increased susceptibility to select antibiotics in vitro.
Antibiotic resistance is a problem in CF patients with P. aeruginosa infections [44][45][46] . We previously reported that RpoN* alters transcription of several genes involved in multidrug efflux pumps that confer natural resistance 32 . Additionally, RpoN is implicated in tolerance to various classes of antibiotics [28][29][30] . We evaluated the effects of RpoN* on antibiotic susceptibility using a MicroScan Neg MIC 43 panel. The test conditions were P. aeruginosa PA19660 Xen5 that was mock-transformed (no vector) or transformed with the empty vector or RpoN* plasmid. We expected that RpoN* would improve antibiotic susceptibility of P. aeruginosa. In PA19660 Xen5 mock-transformed or with the empty vector, antibiotic susceptibility profiles were the same, except for gentamicin, which increased in the empty vector strain due to the GM R selection marker (Table 2). In PA19660 Xen5 expressing RpoN*, susceptibility to five beta-lactam antibiotics was improved at least 2-to 4-fold (Table 2). Results presented in Table 2 are representative of a single experiment, however these results were highly consistent and repeatable over multiple experiments and biological replicates. The antibiotics with improved susceptibility were cefotaxime, cefepime, and ceftazidime (three cephalosporins), piperacillin (a ureidopenicillin), and imipenem (a carbapenem). Susceptibility to some antibiotics was unchanged (data not shown). For piperacillin, there was at least a 2-fold increase in susceptibility. This improvement is clinically relevant as it changed the status from resistant to sensitive (Table 2). For the other drugs, which there was no change in clinical susceptibility status, presence of RpoN* increased the therapeutic potency of the drugs. The results demonstrate that RpoN* expression increased P. aeruginosa susceptibility to several antibiotics.

Discussion
Here, we confirm and expand results of previous studies by showing the ability of RpoN* to abrogate virulence phenotypes in P. aeruginosa isolates from CF patients and to improve susceptibility to several antibiotics. Our working model of the mechanism of action of RpoN* is that it binds the -24 promoter consensus sites, blocking transactivation by RpoN and other sigma factors. By altering the transcriptome, RpoN* reduced virulence in well-characterized laboratory strains 32 . Thus, the motivation for this study was to understand the clinical relevance of RpoN*. We demonstrated that RpoN* expressed in CF patient isolates reduced motility and biofilm formation in vitro, independently of RpoN protein levels. The RpoN* molecular roadblock protected C. elegans from a highly virulent P. aeruginosa patient isolate in an in vivo infection model. RpoN* also improved P. aeruginosa susceptibility to antibiotics.
P. aeruginosa isolated from CF patients are highly variable 36,37 , with the rpoN gene often mutated or lost 38 . The patient isolates evaluated in this study had a broad range of motility, biofilm formation, RpoN protein levels, and virulence in C. elegans. There was no correspondence between most in vitro phenotypes, in vivo pathogenesis, and RpoN levels (Supplemental Fig. S4). The only correlation observed was between twitching, or pili-associated, motility and biofilm formation (Supplemental Fig. S4F, p = 0.0357, R 2 = 0.37050). Other studies suggested that in vitro phenotypes of P. aeruginosa isolates can be related to disease status in CF patients 47 . Patient information and status of P. aeruginosa infections is limited for the isolates described here, so a comparison between phenotypes and patient status is not feasible. Interestingly, two isolates, UUH0201 and UUH0202, were obtained five months apart from the same patient, with UUH0201 collected first. The UUH0202 strain was less motile and RpoN protein levels dropped compared to UUH0201, but virulence increased. This supports the concept that in vitro www.nature.com/scientificreports www.nature.com/scientificreports/ phenotypes reflect P. aeruginosa infection status in CF patients 47 . Further work would be needed to fully elucidate such correlations.
The RpoN* molecular roadblock reduced virulence phenotypes in patient isolates with high or low levels of RpoN. Furthermore, there was no substantial difference between CF isolates transformed with the empty vector and in the absence of manipulation (wild type strains). For instance, RpoN protein levels were higher in SCH0057-7 than PAO1-S, and RpoN* reduced flagellar and pili motility, biofilm formation and pathogenesis. In contrast, relative RpoN protein levels were low in SCH0256-1 and SCH0354-1, and yet RpoN* reduced motility. Thus, the roadblock was effective in the presence or absence of the native sigma factor. This is possible since there is a redundancy among P. aeruginosa sigma factors, with multiple sigma factors having consensus promoter sites for the same gene 31 . Thus, in the absence of RpoN, other sigma factors promote transcription of certain genes. The results here confirm our previous findings, which show that RpoN* reduced virulence in a laboratory strain that was deleted for rpoN 32 . Unfortunately, barriers to transformation precluded evaluating RpoN* in some of the other clinical isolates. However, the strains that were successfully transformed represented much of the diversity across the patient isolates.
The CF patient isolates demonstrated variable pathogenesis in the C. elegans paralytic killing model that spans 6 hours. Only one pathogenic isolate, SCH0057-7, was transformable and thus possible to evaluate the effects of RpoN* in vivo. This strain and several others were also pathogenic in the slow killing assay. This assay was not used to evaluate RpoN* because of difficulty maintaining the plasmid and RpoN* expression. Gentamicin selection and IPTG induction are not durable, we found 32 , because the C. elegans cuticle is impermeable and the compounds are poorly absorbed in the intestine 48 . While it is expected that expressing RpoN* in CF isolates would improve C. elegans survival in the slow killing assay, it is not feasible with the current vector. If issues with maintaining the plasmid and expression of the roadblock were resolved, it would be interesting to evaluate RpoN* in this assay using patient isolates.
The molecular roadblock, RpoN*, binds numerous promoters in bacterial genomes, altering the transcriptome. RpoN* expression in P. aeruginosa greatly reduced transcription of the mex family genes 32 , which are involved in multidrug efflux pumps 49 . Increased expression of mex genes is linked to increased resistance to antibiotics 14 . Therefore, we investigated whether RpoN* alters P. aeruginosa susceptibility to antibiotics. We employed a clinical laboratory assay for testing bacterial susceptibility or resistance to antibiotics and found that RpoN* improved antibiotic susceptibility at least two-fold for five different antibiotics, including imipenem. This agrees with previous studies that showed RpoN is involved in P. aeruginosa tolerance of carbapenems, quinolones, and tobramycin [28][29][30] . Unfortunately, the commercial assay uses pre-determined antibiotic concentrations in a 96-well plate, limiting the scope of the roadblock's effects. Additionally, the P. aeruginosa strain used here is sensitive to quinolones and tobramycin, so RpoN* effects on resistance to these antibiotics was not evaluated. Unfortunately, testing clinical strains was not feasible for this study. However, it will be important to test such strains, particularly those resistant to quinolones, carbapenems, and tobramycin, to determine the effects of RpoN*. Further studies are needed to uncover the full spectrum of RpoN* effects on antibiotic susceptibility.
MRDOs are increasing worldwide, even those with resistance to entire antibiotic classes. Alarmingly, nearly all antibiotics brought to market in the past 30 years are variations on existing drugs 50 . Research into alternative strategies to treat bacterial infections is a priority, including compounds to enhance activity of existing antibiotics or neutralize virulence factors. The molecular roadblock falls into the latter type. RpoN* binds consensus promoters throughout the P. aeruginosa genome, affecting the transcription of numerous virulence factors. Due to  www.nature.com/scientificreports www.nature.com/scientificreports/ the many binding sites for RpoN*, it is unlikely that resistance to it would develop during treatment. The binding sequence of the RpoN consensus promoter is conserved across gram-negative and gram-positive bacteria (36,37). The effects of RpoN* on virulence phenotypes of Pseudomonas putida, Burkholderia cepacia, and Escherichia coli have been explored (unpublished data), suggesting that RpoN* may reduce virulence in multiple organisms. More studies are needed to identify the spectrum of RpoN* activity and its resistance frequency. Currently, RpoN* is a tool for antimicrobial development and is not a usable drug. However, a study was recently published using a stapled RpoN-like peptide to reduce transcription of RpoN-related genes in E. coli 51 . They found that this stapled peptide, which binds to RpoN consensus promoters, could enter P. aeruginosa cells 51 . Finding a small molecule or stapled-peptide that works in the same cis-acting manner as RpoN* would be an effective, clinically relevant strategy to combat P. aeruginosa virulence and antibiotic resistance.

Materials and Methods
Bacteria and nematodes. P. aeruginosa clinical isolates were provided by Seattle Children's Hospital (SCH strains) and Upstate University Hospital (UUH strains). P. aeruginosa PAO1-M was provided by C. Manoil 41 , and P. aeruginosa PAO1-S and ΔrpoN were provided by D. Haas 22 . P. aeruginosa PA19660 Xen5 was purchased from PerkinElmer. E. coli OP50 was provided by D. Pruyne (SUNY Upstate Medical University). All strains are listed in Table 1. For long-term storage, bacteria were grown overnight in LB broth at 37 °C with shaking, and frozen in 10% glycerol at −80 °C. Caenorhabditis elegans N2 was purchased from the Caenorhabditis Genetics Center (University of Minnesota, Minneapolis, MN), and maintained on nematode growth media (NGM) seeded with E. coli OP50 at 20 °C 52 . Populations were synchronized via egg lay and grown to the young adult stage at 20 °C 53 . plasmids. RpoN* and empty vector plasmids were previously described 32 . Plasmids were maintained in E.
transformation. Permissive P. aeruginosa patient isolates and a lab strain were transformed by electroporation prior to all experiments, per standard protocol 54 . Transformed bacteria were selected on LB agar or BHI agar supplemented with gentamicin (30 mg/L). Individual colonies were picked for each assay.

Western blot analysis.
Overnight bacteria cultures were treated with Cell Lytic B Lysis Reagent (Sigma) to generate crude cell lysates. The soluble protein fraction was separated on 10% Mini-PROTEAN TGX Stain-Free protein gels (BioRad), activated for 5 minutes with UV light, imaged and transferred via semi-dry apparatus to a PVDF membrane. Membranes were incubated with primary antibody specific for E. coli RNA σ 54 (1:500, BioLegend) overnight, then with secondary antibody HRP goat anti-mouse (1:10,000, Jackson ImmunoResearch). The chemiluminescent signal was generated with the Pierce SuperSignal West Fempto substrate kit (Thermo Scientific) and detected with ChemiDoc MP Imaging System (Bio-Rad Laboratories). Protein bands and total protein per lane were measured with Image Lab (Version 5.2.1; Bio-Rad Laboratories). RpoN bands were compared to corresponding total detected protein in each lane and the background value in ∆rpoN was subtracted from all samples. phenotyping assays. Assays to measure swimming, twitching 55 , and biofilm formation 56 , were conducted according to standard protocols. Transformed P. aeruginosa clinical isolates were grown in appropriate media supplemented with gentamicin (30 mg/L), and with or without IPTG (1 mM). Motility and biofilm assays were conducted at 37 °C for 24 h. Images of motility assays were obtained with IVIS-50 (Perkin Elmer) and colony diameter was measured with Living Image software (Perkin Elmer). To study biofilm formation, overnight cultures were diluted 1:3 in LB broth with appropriate conditions, grown for 3 h at 37 °C with shaking. Log-phase cultures were diluted 1:50 into M63 minimal media with 0.4% arginine and 1 mM MgSO 4 (30 mg/L gentamicin and 1 mM IPTG included for transformed cultures), and 100 µl added per well to a U-bottom 96-well plate and plates incubated at 37 °C for 24 h. Biofilms were stained with 0.1% crystal violet at RT, extracted in 95% ethanol, and absorbance was measured at 550 nm with a μQuant microplate spectrophotometer (BioTek).

P. aeruginosa -C. elegans infection assays.
For the paralytic killing assay, laboratory strains, clinical isolates, or transformed P. aeruginosa were spread on Brain Heart Infusion (BHI) agar (Difco) with, when applicable, gentamicin (30 mg/L) and with or without IPTG (1 mM). E. coli was spread on BHI agar. All plates were grown overnight at 37 °C. Bacteria colonies were swabbed onto BHI agar, supplemented with gentamicin and/ or IPTG (1 mM) when applicable, and grown at 37 °C for 24 h 41 . Adult C. elegans were added to plates and the assay was conducted at room temperature, per standard protocol 41 . For the slow killing assay, laboratory strains or clinical isolates of P. aeruginosa were grown overnight in LB broth at 37 °C with shaking, and cultures were spread on a modified NGM agar (0.35% bactopeptone, 2% bactoagar) 57 . Plates were incubated at 37 °C for 24 h, then at room temperature for an additional 24 h. The assay was conducted at 20 °C, and worms were scored every 24 h per standard protocol 57 .
Antibiotic sensitivity testing. Transformed or mock transformed bacteria were grown overnight in LB broth with gentamicin (30 mg/L) and IPTG (1 mM) or only IPTG (1 mM), respectively. MicroScan Neg MIC 43 panels (Beckman Coulter Inc., Brea, CA) were used. Panels were set up per manufacturer's protocol (MicroScan Gram Negative Procedure Manual, version 09/2016) using the RENOX system (Beckman Coulter Inc., Brea, CA) with a final well concentration of 3-7 × 10 5 CFU/mL. The following modifications were made to the manufacturer's protocol: LB broth supplemented with IPTG (1 mM) and with or without gentamicin (30 mg/L) was used in place of saline for whole panel. Plates were incubated at 35 °C for 16-20 h and read using a MicroScan