P. aeruginosa strain PGN5 produces a high amount of alginate

PAO1, the parent strain of PGN5, is a wild‐type P. aeruginosa strain that produces relatively small amounts of alginate and exhibits a non‐mucoid phenotype; thus, PGN5 is also non‐mucoid when cultured (Fig. 3A). In PAO1, the alginate biosynthetic operon, which contains genes required for alginate production, is negatively regulated. Activation of this operon leads to alginate production and a mucoid phenotype. For example, the P. aeruginosa strain VE2 (i.e. PAO1+mucE) displays high levels of alginate production and a mucoid phenotype (Fig. 3B) (Qiu et al., 2007). This strain was generated through a mariner‐based transposition mutagenesis on PAO1 and carries a gentamicin resistance cassette upstream of the mucE gene, which drives its overexpression. The mucoid phenotype in this strain is stable, with no observed reversion to non‐mucoidy after over 12 passages on non‐selective media. Thus, overexpression of MucE, an activator of the alginate biosynthetic pathway, induces a strong and stable mucoid phenotype in the PAO1 strain. The plasmid pUCP20‐pGm‐mucE, which constitutively overexpresses MucE, was used to test whether the genetically modified PGN5 strain could produce alginate (Qiu et al., 2007). Indeed, the presence of this plasmid in PGN5 (PGN5+mucE) induced a mucoid phenotype (Fig. 3B). Through over 12 passages on non‐selective media, the predominant phenotype of PGN5+mucE colonies is mucoid; however, some reversion to non‐mucoidy occurs. Thus, the mucoid phenotype in this strain is relatively less stable than VE2, presumably because overexpression of MucE occurs from a plasmid, rather than a chromosomal locus. To measure the amount of alginate produced by PGN5+mucE on a cellular level, we performed a standard carbazole assay, which showed that the PGN5+mucE and VE2 strains produce comparable amounts of alginate (Fig. 3C). Indeed, VE2 and PGN5+mucE produced averages of 24.49 ± 1.64 g l⁻¹ and 26.41 ± 3.39 g l⁻¹ of alginate (dry weight) respectively.

To examine whether the alginate produced by PGN5+mucE was similar in composition to alginate produced by VE2, we performed HPLC to compare the M and G content of alginate produced by each strain. The HPLC chromatograms obtained from alginate prepared from VE2 and PGN5+mucE were similar to each other and to a seaweed alginate control (FMC 8133) (Fig. 3D). The M:G ratios of VE2 and PGN5+mucE were consistently determined around 80:20, while seaweed alginate was consistently around 70:30. To confirm that the physical properties of VE2 and PGN5+mucE alginates were also similar, we prepared 2% sodium alginate gels from alginate produced by each strain, as well as two different commercial seaweed controls (Sigma 2158 and Sigma 180947) and measured the viscosity and yield stress. The average viscosities of VE2 and PGN5+mucE alginate gels were consistent and comparable at 73.58 and 72.12 mPa respectively (Fig. 3E). While the viscosities of seaweed alginate gels varied between samples, viscosities of VE2 and PGN5+mucE alginate gels were within the range of gels prepared from seaweed alginate (Fig. 3E). Likewise, the average yield stress of VE2 and PGN5+mucE alginate gels was similar at 47.34 and 47.16 Pa respectively (Fig. 3F). The average yield stress of sodium alginate gels prepared from Sigma 2158 and Sigma 180947 was 33.64 and 32.15 Pa respectively; thus, they have a lower tensile strength than gels prepared from VE2 and PGN5+mucE alginate (Fig. 3F).

P. aeruginosa strain PGN5 did not cause mortality in mice

To test whether the pathogenesis of PGN5 was attenuated, C57BL/6 mice were challenged with intraperitoneal injection of 5 × 10⁸ cells of the PCR‐ and phenotype‐validated strains VE2, PGN5+mucE or E. coli BL21, or PBS as a negative control. Injection with the VE2 strain was fatal in 95% of mice within 48 h (Fig. 4). In contrast, injection with BL21 cells resulted in 20% mortality within 48 h, while no mortality was observed from injection with either the PGN5+mucE strain or PBS (Fig. 4). The mice were monitored for 4 weeks post‐injection, and no change in mortality was observed.

Open in a separate window

Figure 4

Per cent survival of male and female C57BL/6 mice after injections with P. aeruginosa strains VE2, PGN5+mucE or E. coli BL21.

A. Per cent survival of male mice n = 10.

B. Per cent survival of female mice, n = 10.

C. Per cent survival of combined survival of both male and female, n = 20.

Gene deletions

Retrieval and analysis of P. aeruginosa gene sequences were performed at the Pseudomonas Genome Database website: http://www.pseudomonas.com (Winsor et al., 2016). Five genes, toxA, plcH, phzM, wapR and aroA, were sequentially deleted from the chromosome of the wild‐type P. aeruginosa strain PAO1. The pEX100T‐NotI plasmid was used to mediate the in‐frame marker‐less deletion of each gene (Schweizer and Hoang, 1995; Damron et al., 2009). This plasmid carries the genes ampR, which confers resistance to carbenicillin, and sacB (B. subtilis), which provides sucrose sensitivity. Plasmid inserts used to delete toxA, plcH, phzM and wapR were generated by PCR amplification of 500–1000 bp of sequence directly upstream and 500‐1000 bp of sequence directly downstream of each target gene, followed by fusion of these DNA fragments via cross‐over PCR. The resultant PCR product was digested and ligated into pEX100T‐NotI. For in‐frame deletion of aroA, about 800 bp of upstream sequence adjacent to about 900 bp of downstream sequence of the target gene was synthesized, digested and ligated into pEX100T‐NotI by the company GenScript (Piscataway, NJ, USA). Each of the final plasmids was transformed into OneShot^TM TOP10 Electrocomp E. coli (Invitrogen, Carlsbad, CA, USA).

For each deletion, the pEX100T‐NotI plasmid carrying its specific insert was introduced into P. aeruginosa via triparental conjugation with the helper plasmid pRK2013. The target gene was deleted with a two‐step allelic exchange procedure. Briefly, homologous recombination between one site on the plasmid and its target site on the chromosome integrated the plasmid into the P. aeruginosa chromosome (i.e. a single‐cross‐over event). Single cross‐overs were selected by plasmid‐conferred resistance to carbenicillin and sensitivity to 10% (w/v) sucrose supplemented in PIA. Single cross‐overs were grown overnight in LB broth to allow for homologous recombination between the second site on the plasmid and its target site on the chromosome (i.e. a double‐cross‐over event), which removes the entire plasmid sequence along with the target gene sequence. Double cross‐overs were selected by sensitivity to carbenicillin and resistance to sucrose, indicating that the plasmid sequence had been removed. Sucrose‐resistant, carbenicillin‐sensitive colonies were sequenced to verify the in‐frame gene deletion.

Gene sequencing

Sanger sequencing to confirm in‐frame deletions was performed by the West Virginia University (WVU) Genomics Core Facility (Morgantown, WV, USA). Whole‐genome re‐sequencing on these strains was performed by CD Genomics (Shirley, NY, USA). Genomic assembly of the data utilized the SPAdes microbial isolate assembler (Bankevich et al., 2012) followed by transfer annotation using PROKKA (Seemann, 2014). Homology‐based taxonomy assignment utilized PanGIA (Li et al., 2018), along with in‐house derived confidence level settings based on positive control DNA mixtures. Reference‐based analysis employed bowtie2 (Langmead and Salzberg, 2012) compared to the parental strain genome (GenBank accession number AE004091.2) with variant calling. bowtie2 settings include a 1000‐nt window and a 200‐nt step size. Whole‐genome sequences are available for PGN4 (GenBank accession number CP032540) and PGN5 (GenBank accession number CP032541). Pairwise genome comparison utilized the Artemis Comparison Tool for visualization (Carver et al., 2005).

Western blot for Exotoxin A

A Western blot was used to verify the deletion of toxA in PGN5 by the absence of the Exotoxin A product. Cells from 24‐h broth cultures of P. aeruginosa strains PAO1, Exotoxin A‐positive PA103 and PGN5 were removed by centrifugation. Supernatant was filter‐sterilized with a 0.2 μm syringe filter and concentrated with an Amicon Ultra‐15 Centrifugal Filter Unit with a 10‐kDa cut‐off (Sigma, St. Louis, MO, USA) according to the manufacturer's instructions. Total protein concentration of the supernatant (extracellular fraction) was quantified with the bicinchoninic acid assay (BCA) using a bovine serum albumin (BSA) standard (Pierce, Rockford, IL, USA). For Coomassie staining, equal amounts of protein (60 mg) were run with SDS‐PAGE on a TGX 4‐20% gel (Bio‐Rad, Hercules, CA, USA), stained with Coomassie R‐250 (Protea Biosciences, Morgantown, WV, USA) and imaged on a ProteinSimple Fluorochem M (San Jose, CA, USA). For Exotoxin A western blot, samples with equal total protein concentrations (0.05 μg/μL) were run on the ProteinSimple Wes automated Western blot system using the 12–230 kDa separation module and the 8 × 25 capillary cartridge (San Jose, CA, USA). Exotoxin A was detected with a polyclonal rabbit anti‐Exotoxin A antibody (Sigma, St. Louis, MO, USA) diluted 1:200 and the ProteinSimple anti‐rabbit detection module (San Jose, CA, USA).

Analysis of haemolytic activity

To verify deletion of plcH, haemolytic activity of E. coli strain BL21 and P. aeruginosa strains PAO1, VE2 and PGN5 was assessed by culturing bacterial strains for 48 h on blood agar (trypticase soy agar with 5% sheep blood; Remel, Lenexa, KS, USA). Before inoculation, plates were treated with a 1 mg ml⁻¹ solution of aromatic amino acids and dried. Plates were imaged for the presence or absence of clear plaques indicating haemolysis surrounding bacterial growth.

Extraction and quantification of pyocyanin

To verify deletion of phzM in PGN5, a two‐step extraction procedure with chloroform and HCl was used to recover and quantify the secreted pyocyanin pigment (Essar et al., 1990). Briefly, cells from 24‐h broth cultures of P. aeruginosa strains PAO1 and PGN5 were removed by centrifugation and the supernatant was filter‐sterilized with a 0.2‐μm syringe filter. Pyocyanin was extracted from 7.5 ml samples with 4.5 ml of chloroform. Pyocyanin was re‐extracted from 3 ml of the chloroform layer with 1.5 ml of 0.2 M HCl to give a pink colour. The absorption at 520 nm (A₅₂₀) of the aqueous HCl layer was measured with a SmartSpec™ 3000 spectrophotometer (Bio‐Rad, Hercules, CA, USA), and pyocyanin concentration was calculated as follows:

LPS extraction, silver staining and Western blot for O‐antigen

To verify deletion of wapR, lipopolysaccharides (LPS) were analysed for the presence of O‐antigen. LPS was extracted using the Hitchcock and Brown method (Hitchcock and Brown, 1983), resolved by electrophoresis on 12% SDS‐PAGE and stained for visualization using an ultrafast silver staining method (Fomsgaard et al., 1990). For Western immunoblotting, LPS was transferred onto BioTraceNT nitrocellulose membranes (Pall) and detected with a 1:1 mixture of monoclonal antibodies MF15‐4 (OSA specific) and 5c‐7‐4 (inner core specific) overnight at room temperature, followed by incubation with an alkaline phosphatase‐conjugated goat anti‐mouse Fab₂ secondary antibody (Jackson ImmunoResearch, West Grove, PA, USA). The blots were developed using nitroblue tetrazolium (NBT) and 5‐bromo‐4‐chloro‐3‐indolylphosphate (BCIP) as described previously (Blake et al., 1984; de Kievit et al., 1995).

ELISA for AroA

An enzyme‐linked immunosorbent assay (ELISA) was used to verify deletion of aroA by determining the absence of the 3‐phosphoshikimate 1‐carboxyvinyltransferase product. A 96‐well untreated plate was coated with equal amounts of carbonate buffer and of broth culture of the P. aeruginosa strains PAO1 and PGN5. The plate was then incubated for 2 h at 37°C and blocked in skim milk overnight. A polyclonal rabbit anti‐AroA primary antibody (LifeSpan BioSciences, Inc., Seattle, WA, USA) and a polyclonal goat anti‐rabbit IgG secondary antibody (Sigma, St. Louis, MO, USA) were used to detect 3‐phosphoshikimate 1‐carboxyvinyltransferase. Samples were incubated in 1‐Step™ Ultra TMB ELISA Substrate Solution (Thermo Scientific, Rockford, IL, USA), followed by addition of acid stop solution (R&D Systems, Minneapolis, MN, USA). The absorbance was immediately measured at 450 nm (A₄₅₀) on a Spectramax^® i3x microplate reader (Molecular Devices, Sunnyvale, CA, USA). Absorbance reported was the relative absorbance detected between PAO1 and PGN5. Reported values represent the mean ± SD of three independent experiments. Two‐tailed Student's t‐test was used to determine statistical significance.

Growth assays

To quantify rescue of growth of the PGN5 strain by aromatic amino acid supplementation, PIB supplemented with gentamicin was inoculated with a single colony of the P. aeruginosa strains PGN5+mucE or VE2 and grown at 37°C until reaching an A₆₀₀ of 1, as measured on a SmartSpec™ 3000 spectrophotometer. The seed cultures were then used to inoculate PIB supplemented with either 0, 0.5 or 1 mg ml⁻¹ of aromatic amino acids (Y, F, W). Cultures were grown for 72 h at 37°C, and A₆₀₀ measurements were taken every 24 h to estimate the total number of cells in each culture. The results presented represent three independent experiments.

Carbazole assay

The carbazole assay is a standard uronic acid detection method that was used to quantify the amount of alginate produced in culture. The P. aeruginosa strains VE2 and PGN5+mucE were grown on PIA for 48 h at 37°C. Alginate and cells were collected with 0.85% NaCl. Cell suspensions were incubated in sulfuric acid/borate solution and carbazole at 55°C. As carbazole reacts with uronic acid, the solution turns a pink to purple colour. Absorbance at 530 nm (A₅₃₀) was measured on a SmartSpec™ 3000 spectrophotometer. The concentration of alginate (μg/ml OD⁻¹ ₁₆₀₀) was calculated using a standard curve generated with purified mannuronic alginic acid (Sigma, St. Louis, MO, USA).

Collection of alginate

Pseudomonas aeruginosa strains VE2 and PGN5+mucE were cultured for 72 h in 1 l of PIB or PIB supplemented with 1 mg ml⁻¹ of aromatic amino acids. Alginic acid was precipitated with three volumes of ethanol, vacuum‐filtered through a Twill Dutch weave wire cloth (5 μm; Dorstener Wire Tech, Spring, TX, USA) and dried in a vacuum oven. The dried weight of alginate collected from VE2 and PGN5+mucE strains was presented as the mean ± SD of five independent experiments.

Compositional analysis of alginate via HPLC

Determination of the M:G sugar ratios was performed by HPLC analysis using an adapted pre‐column derivatization of hydrolysed alginate with 1‐phenyl‐3‐methyl‐5‐pyrazolone (PMP) (Honda et al., 1989). Seven mg of dried alginic acid was suspended in 962 μl of water and 38 μl of trifluoroacetic acid (TFA) and hydrolysed in a sealed vial at 100°C. After 4 h, the pH of the samples was adjusted to 9.5 and 0.5M PMP in methanol (0.15 ml) was added. The samples were incubated at 70°C for 1.5 h with shaking at 1000 r.p.m. The pH was adjusted to 6.5, and excess PMP reagent was removed by repeated extraction with chloroform prior to analysis. The M/G ratios were determined by HPLC using an Agilent Infinity II HPLC system equipped with a 1260 quaternary pump (G7111B), vial sampler (G7129A), multicolumn thermostat (G7116A), diode array detector (G7115A) and Openlab CDS ChemStation Edition (v. C.01.08 [210]) software. Chromatographic separation was performed using an Agilent Eclipse Plus C18 column (4.5 × 150, 3.5 μM) and a mobile phase consisting of 0.1 M phosphate buffer (pH 6.7) with acetonitrile at a ratio of 82:19 (v/v, %). The column temperature was maintained at 25°C, and elution of the derivatized M/G monomer at 1 ml min⁻¹ was detected at 245 nm. As a control, commercially available alginate from seaweed (FMC Protanal CR 8133, Wilmington, DE, USA) was prepared and run along with P. aeruginosa alginate samples.

Physical property testing of alginate

To test the physical properties of bacterial alginates, sodium alginate gels were prepared from dried alginic acid. Briefly, alginic acid was dissolved in NaOH pH 8.5–9 to form sodium alginate, which was then precipitated with ethanol, filtered and dried in a vacuum oven. Sodium alginate gels (2% w/v) were prepared by dissolving sodium alginate in distilled water. Sodium alginate gels were prepared for comparison from seaweed alginate samples purchased from Sigma: alginic acid sodium salt, from brown algae (Sigma 2158; catalogue number A2158) and alginic acid sodium salt (Sigma 180947; catalogue number 180947). Viscosity and yield stress were measured on a HAAKE MARS I Rheometer (Thermo Scientific, Karlsruhe, Germany) using preset parameters. Results are presented as the mean ± SD of 3–5 independent experiments. Analysis and calculations were performed with RheoWin Software (version 4.82, Thermo Scientific).

Bacterial pathogenicity in mice

Pseudomonas aeruginosa strains PGN5 and VE2, and E. coli strain BL21 were grown in LB broth to generate frozen stocks for injection into C57BL/6 mice. Cell cultures were concentrated to 2.5 × 10⁹ cells per ml, flash‐frozen in liquid nitrogen and stored at −80°C. The cultures used were strain‐verified by PCR and phenotype, and cell concentrations were verified by viable plate counts before freezing and before mouse injection. A total of eighty 10‐ to 12‐week‐old mice were divided into four groups of 10 males and 10 females. Each group received 200 μl intraperitoneal injections of one of the following: PBS, 5 × 10⁸ cells of the P. aeruginosa strains VE2 or PGN5+mucE, or 5 × 10⁸ cells of the E. coli strain BL21. After injection, mice were monitored for mortality every 3 h for 72 h and then every 12 h for 7 days.

Localization of bioluminescent bacteria in mice

PAO1 and PGN5 were marked with bioluminescence using the method and plasmids developed in the Schweizer laboratory (Choi and Schweizer, 2006). The pUC18 mini‐Tn7T‐Gm‐lux plasmid was used, which allows for insertion of the luxCDABE operon, along with an FRT‐flanked gentamicin resistance cassette into a neutral site downstream of the glmS gene. Briefly, electrocompetent PAO1 and PGN5 cells were prepared with 300 mM sucrose and electroporated with the pUC18 mini‐Tn7T‐Gm‐lux and pTNS2 plasmids. Pure stocks were generated from resultant gentamicin‐resistant and bioluminescent colonies. The pFLP2 plasmid was used to remove the gentamicin resistance cassette. Final stocks used for mouse injection were PCR‐verified, bioluminescent, plasmid‐cured, and gentamicin and carbenicillin‐sensitive. For mouse injection, stocks were prepared and injected as described above. Mice were imaged on an IVIS Lumina XRMS (PerkinElmer, Waltham, MA, USA) every 6 h for 18 h and monitored for 4 weeks. By 18 h post‐injection, bioluminescence was only detected at the injection site of all mice.

This work was supported by the National Institutes of Health (NIH) grants R44GM113545 and P20GM103434. We would like to acknowledge the WVU Genomics Core Facility, Morgantown WV, for support provided to help make this publication possible. We would also like to thank Dr. Beverly Delidow's 2018 Summer Writing Group, which helped with critical review during the writing of the manuscript.