Single-Nucleotide Polymorphisms Found in the migA and wbpX Glycosyltransferase Genes Account for the Intrinsic Lipopolysaccharide Defects Exhibited by Pseudomonas aeruginosa PA14

ABSTRACT

Pseudomonas aeruginosa PA14 is widely used by researchers in many laboratories because of its enhanced virulence over strain PAO1 in a wide range of hosts. Although lipopolysaccharide (LPS) is an important virulence factor of all P. aeruginosa strains, the LPS of PA14 has not been characterized fully. A recent study showed that the structure of its O-specific antigen (OSA) belongs to serotype O19. We found that the OSA gene cluster of PA14 shares ∼99% identity with those of the O10/O19 group. These two serotypes share the same O-unit structure, except for an O-acetyl substitution in one of the sugars in O10. Here we showed that both PA14 and O19 LPS cross-reacted with the O10-specific monoclonal antibody MF76-2 in Western blots. Analysis by SDS-PAGE and silver staining showed that PA14 LPS exhibited modal chain lengths that were different from those of O19 LPS, in that only “very long” and “short” chain lengths were observed, while “medium” and “long” chain lengths were not detected. Two other novel observations included the lack of the uncapped core oligosaccharide epitope and of common polysaccharide antigen (CPA) LPS. The lack of the uncapped core oligosaccharide was caused by point mutations in the glycosyltransferase gene migA, while the CPA-negative phenotype was correlated with a single amino acid substitution, G20R, in the glycosyltransferase WbpX. Additionally, we showed that restoring CPA biosynthesis in PA14 significantly stimulated mature biofilm formation after 72 h, while outer membrane vesicle production was not affected.

IMPORTANCE P. aeruginosa PA14 is a clinical isolate that has become an important reference strain used by many researchers worldwide. LPS of PA14 has not been characterized fully, and hence, confusion about its phenotype exists in the literature. In the present study, we set out to characterize the O-specific antigen (OSA), the common polysaccharide antigen (CPA), and the core oligosaccharide produced by PA14. We present evidence that PA14 produces an LPS consisting of “very-long-chain” and some “short-chain” OSA belonging to the O19 serotype but is devoid of CPA and the uncapped core oligosaccharide epitope. These intrinsic defects in PA14 LPS were due to single-nucleotide polymorphisms (SNPs) in the genes that encode glycosyltransferases in the corresponding biosynthesis pathways. Since sugars in CPA and the uncapped core are receptors for different bacteriocins and pyocins, the lack of CPA and an intact core may contribute to the increased virulence of PA14. Restoring CPA production in PA14 was found to stimulate mature biofilm formation.

INTRODUCTION

Pseudomonas aeruginosa is an important opportunistic pathogen that infects a wide range of organisms, including humans. It can result in fatal consequences in compromised individuals with cancer, AIDS, burn wounds, or cystic fibrosis (CF). It infects the lungs of ∼80% of CF patients, and the infection usually becomes chronic, resulting in death in a majority of these patients (1). Two P. aeruginosa strains, PAO1 and PA14, have been used widely in research relating to either biofilms or virulence. Their genomes have been sequenced, and transposon libraries that cover the majority of the genome of each have been constructed (2–5). Compared to strain PAO1, strain 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 (6–8). Strain PAO1 and strain PA14 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) (4). However, no evidence has been presented to correlate the presence of these PAIs with the increased virulence of PA14 compared to PAO1 (4).

Of the many virulence factors of P. aeruginosa, the surface lipopolysaccharide (LPS) is a major one, as defects in LPS biosynthesis have been shown to increase the 50% lethal dose (LD₅₀) >1,000-fold in a burned-mouse infection model (9). LPS is the predominant component of the outer leaflet of the outer membrane, mediating direct interactions with the environment and host cells. Structurally, LPS can be divided into three domains: lipid A (the endotoxin which anchors the molecules into the outer membrane), a short core oligosaccharide (OS), and the distal portion of long-chain polysaccharide, called the O antigen (O-Ag) (Fig. 1A). Most P. aeruginosa strains simultaneously produce two distinct forms of O antigen, namely, the O-specific antigen (OSA; historically called the B band), which is a heteropolymer composed of repeating units of two to five distinct sugars, and the common polysaccharide antigen (CPA; historically called the A band), which is produced by the majority of P. aeruginosa strains and contains a homopolymer of the rare sugar d-rhamnose (d-Rha) (Fig. 1A).

FIG 1.

Illustrations of the various glycoforms of LPS in P. aeruginosa and the organization of gene clusters involved in the respective biosynthesis of OSA and CPA. (A) Diversity of the various LPS glycoforms on the cell surface of a typical strain of P. aeruginosa. *, Glc^IV is not present in all serotypes; †, the average number (n) of CPA repeat units is around 20; §, the OSA polymer from serotype O19 is displayed as a representative polymer. (B) OSA and CPA gene clusters in strain PA14. Diagrams are not drawn to scale.

The sugar composition and the linkages between O-Ag units of the OSA form the basis for the International Antigen Typing Scheme (IATS) and classify P. aeruginosa into 20 serotypes (10–12). The chemical structures of the OSA of all 20 IATS serotypes have been elucidated (13). The first characterized OSA biosynthesis cluster was that for strain PAO1 (O5 serotype) (14, 15), followed by those of serotypes O6 (8) and O11 (16). The OSA clusters are located in the same relative position in P. aeruginosa genomes and are flanked by himD/ihfB (pa3161) on the 5′ end and wbpM (pa3141) on the 3′ end. These findings eventually led to the cloning and sequencing of the OSA loci of all 20 IATS serotypes by Raymond et al. (17), who showed that the OSA clusters could be divided into 11 distinct groups based on the protein families encoded by the genes and the presence of insertion sequences (IS) and deletions. The genome of PA14 has been sequenced by Lee et al. (4), and the locus of the PA14 OSA biosynthesis cluster was identified. Using NCBI BLASTN analysis (http://blast.ncbi.nlm.nih.gov/Blast.cgi?PROGRAM=tblastn&PAGE_TYPE=BlastSearch&LINK_LOC=blasthome), we found that the PA14 OSA cluster belongs to the O10/O19 serotypes (Fig. 1B). The OSA repeat unit of PA14 contains the structure(-4)-α-l-GalNAcA-(1-3)-α-d-QuiNAc-(1-3)-α-l-Rha-(1-) (18), which corresponds to the O-Ag structure of IATS O19 (13). However, no biochemical or immunological properties of the PA14 LPS have been established to date.

CPA is composed of a trisaccharide repeat unit of α1→3,α1→2-linked d-Rha (Fig. 1A). The monoclonal antibody (MAb) N1F10 raised against CPA could detect the expression of this particular form of O-Ag in 14 of the 20 P. aeruginosa IATS serotype strains. The strains that did not react to the MAb were of serotypes O7, O12, O13, O14, O15, and O16 (19, 20). It is worth noting that the majority of clinical P. aeruginosa isolates from chronically infected cystic fibrosis patients lack OSA but produce CPA as the predominant surface polysaccharide antigen (19, 21). CPA biosynthesis genes are organized into two contiguous clusters, an eight-gene operon consisting of rmd, gmd, wbpW, wzm, wzt, wbpX, wbpY, and wpbZ (pa5454-pa5449) (22–24) and a newly identified, five-gene operon, pa5455-pa5459, in PAO1 (25) (Fig. 1B). In the eight-gene operon, three genes, rmd, gmd, and wbpW, are involved in the biosynthesis of the nucleotide sugar precursor GDP–d-rhamnose, another three genes, wbpX, wbpY, and wbpZ, encode distinct glycosyltransferases thought to be involved in the assembly of the d-Rha repeat units, and two other genes, wzm and wzt, encode the ABC transporter system involved in the transfer of the d-rhamnan precursor from the cytoplasmic to the periplasmic side of the inner membrane (26). In the five-gene operon, pa5455 and pa5456 encode glycosyltransferases, pa5457 and pa5459 encode methyltransferases, and pa5458 codes for a putative acetyltransferase (25). The importance of the latter operon to CPA biosynthesis has been established based on evidence collected from mutagenesis and overexpression experiments (25). CPA was found to play important biological roles as a receptor for bacteriophage A7 (27) and lectin-like bacteriocins (28) and as a surface antigen for P. aeruginosa that is important for adherence to human bronchial epithelial cells (29). In a recent study, Murphy et al. (30) showed that CPA is required for strain PAO1 to develop mature and robust biofilms. When we performed a BLASTN analysis of the genome sequence of PA14, we found that PA14 possesses the eight-gene and five-gene CPA biosynthesis clusters. The organization of these two operons is depicted in Fig. 1B. The presence of both clusters might have led researchers to assume that PA14 is capable of producing CPA (31). However, the presence of CPA in PA14 has not been demonstrated experimentally, and in this study, we found that PA14 possesses intrinsic LPS deficiencies, one of which is the inability to produce CPA on the bacterial cell surface.

The core oligosaccharide (OS) in P. aeruginosa LPS can be divided into the inner and outer cores. The inner core structure is highly conserved among Gram-negative bacteria, whereas the outer core is species specific. In P. aeruginosa, the outer core is composed of galactosamine (GalN), Rha, and three glucose residues (Glc^I to Glc^III) (32) (Fig. 1A and 2). Most P. aeruginosa strains simultaneously synthesize two distinct forms of core OS glycoforms, called the capped and uncapped cores. The “capped core” is covalently attached to different lengths of O antigen repeat units, while the “uncapped core” is not linked to an O antigen and is the shortest form of LPS (Fig. 1A). The capped and uncapped cores have distinct chemical structures. In the capped core, l-Rha is bonded to Glc^I through an α1-3 linkage by the action of the rhamnosyltransferase WapR, while in the uncapped core, l-Rha is bonded to Glc^II through an α1-6 linkage by the enzyme MigA (33, 34) (Fig. 2). The expression of wapR is negatively regulated, while migA is positively regulated by the quorum sensing system RhlR/RhlI (34). In some strains, the uncapped core also contains a fourth Glc residue (Glc^IV) (35, 36) linked to l-Rha (Fig. 2). Probing of Western blots with the MAb 5c-101, specific for recognizing the Glc^IV residue as part of the uncapped core epitope (37), could separate the 20 standard IATS serotype strains of P. aeruginosa into two groups: a group that reacted (Glc^IV-positive strains, including PAO1 and strains of IATS serotypes O2, O7, O8, O10, O16, O18, O19, and O20) and a group without reaction (Glc^IV-negative strains, including LESB58 and strains of serotypes O1, O3, O4, O6, O9, O11, O12, O13, O14, O15, and O17) (32, 36). The gene responsible for adding the Glc^IV residue to the uncapped core is wapB (32). All serotypes contain a functional wapB gene, but a deletion in the wapB promoter region in some strains of P. aeruginosa accounts for the lack of the Glc^IV residue in the uncapped core of those serotypes (37). Despite being a short oligosaccharide, the LPS uncapped core is exposed, interacts with the environment, and mediates various interactions (38). For example, an intact outer core with an exposed terminal glucose residue was found to be necessary for maximum association and ingestion of P. aeruginosa by corneal epithelial cells, indicating that the uncapped core is the ligand that interacts with eukaryotic cells (39). Moreover, R-pyocins, which are phage tail-like bacteriocins produced by some P. aeruginosa strains that are able to kill susceptible cells, bind to susceptible P. aeruginosa cells through sugar residues of LPS core (37, 40). To date, there is no information about the presence and/or structure of the uncapped core LPS of PA14.

FIG 2.

Intact uncapped core oligosaccharide structure of P. aeruginosa and R-pyocin targets. The enzyme MigA is responsible for the transfer of an l-Rha residue to the Glc^II residue. The Glc^IV residue is transferred to l-Rha by the enzyme WapB. The presence of Glc^IV is required for recognition by MAb 5c-101. R1 pyocin targets the l-Rha sugar residue of the uncapped core, R2 pyocin targets the Glc^II residue, R3 pyocin targets the Glc^IV residue, and R5 pyocin targets the Glc^I and Glc^III residues. The dashed line means that the involvement of the β-Glc residue (dashed arrow) is less favored than the involvement of the terminal α-Glc residue (solid arrow) in R5 pyocin recognition.

Compared to the LPS of standard laboratory P. aeruginosa strain PAO1 and other strains from the standard IATS serotype sets, LPS of PA14 appeared to exhibit deficiencies. Therefore, the focus of this study was to determine the causes of the LPS deficiencies in PA14. The LPS of PA14 was characterized by using SDS-PAGE, silver staining, and Western immunoblotting. The immunoprobes used included the outer core-specific MAb 5c-101 and the CPA-specific MAb N1F10. Results from these experiments conclusively demonstrated the intrinsic defect in LPS banding patterns for PA14 LPS samples. A series of complementation assays and site-directed mutagenesis were carried out. Our results showed that the LPS phenotype defects in PA14 were due to single-nucleotide polymorphisms (SNPs) in two genes: wbpX and migA. These constitute point mutations in PA14 and explain its lack of an intact core and CPA on the cell surface. We also examined the impact of expressing CPA in PA14 and its role in biofilm formation and outer membrane vesicle (OMV) biogenesis. Expressing CPA in PA14 allowed it to produce more robust biofilms by 72 h of growth; however, the presence of CPA had no impact on OMV biogenesis.

MATERIALS AND METHODS

Bacterial strains, plasmids, and growth conditions.

Plasmids and bacterial strains used in this study are listed in Table 1. To construct the complementary plasmid pHERDwbpX, primers wbpX-F-SacI (TGCGAGCTCCATGACGCGTCTGTTGGTCG) and wbpX-R-XbaI (TGCTCTAGATTAACCGCAATCCGCATGTACTG) were used to amplify the wild-type wbpX gene from PAO1, and the purified PCR product was then digested with SacI and XbaI and inserted into the pHERD20T vector (between the SacI and XbaI sites). The same procedure was used to construct pHERDwbpZ, using the primer pair wbpZ-F-SacI (TGCGAGCTCAATGCGGGTACTGCACTTC) and wbpZ-R-XbaI (TGCTCTAGACTTACCGAGCGGCCTTCAC). A standard electroporation protocol (41) was used to transform various plasmids into P. aeruginosa PA14, P. aeruginosa PAO1(wbpX::lacZ), or P. aeruginosa PAO1(migA::Gm^r). P. aeruginosa strains were grown in lysogeny broth (LB; also commonly known as Luria-Bertani medium) at 37°C unless otherwise stated. When necessary, antibiotics were added at the following concentrations: 300 μg ml⁻¹ carbenicillin, 90 μg ml⁻¹ tetracycline, and 150 μg ml⁻¹ gentamicin. Escherichia coli strains were grown in LB medium at 37°C with the following antibiotic concentrations: 100 μg ml⁻¹ ampicillin and 15 μg ml⁻¹ tetracycline.

TABLE 1.

Bacterial strains and plasmids used in this study

Strain or plasmid	Relevant feature(s)	Source or reference(s)
Strains
E. coli DH5α	F− ϕ80dlacΔM15 Δ(lacZYA-argF)U169 deoR recA1 endA1 hsdR17(rK− mK−) supE44 λ− thi-1 gyrA96 relA	Invitrogen
P. aeruginosa strains
PAO1	Wild type; serotype O5	55
PAO1(migA::Gmr)	PAO1 derivative; Gmr-marked migA mutant	33
PAO1(wbpX::lacZ)	Mutant with transposon insertion mutation of wbpX gene obtained from the transposon insertion library constructed by the University of Washington; mutant ID 5532	3
PA14	Wild type; serotype O19	4, 56
PA14(rmd)	Mutant with transposon insertion mutation of rmd gene of PA14	5
Plasmids
pHERD20T	Broad-host-range vector containing pBAD promoter; Apr Cbr	57
pYHJL9	pHERD20T containing the whole pa5455-pa5459 operon from PAO1 between the XbaI and HindIII sites	25
pYHJL18	pHERD20T containing wzm and wzt from PAO1 between SacI and XbaI sites	25
pHERDwbpX-Z	pHERD20T containing wbpX, wbpY, and wbpZ from PAO1 between SacI and XbaI sites	25
pHERDwbpX	pHERD20T containing wbpX from PAO1	This study
pHERDwbpZ	pHERD20T containing wbpZ from PAO1	This study
pHERDwbpXG20R	pHERD20T containing G20R site-directed mutant construct of PAO1 wbpX	This study
PHERDwbpXA303V	pHERD20T containing A303V site-directed mutant construct of PAO1 wbpX	This study
PHERDwbpXA450T	pHERD20T containing A450T site-directed mutant construct of PAO1 wbpX	This study
miniCTX-migA	Chromosome integration vector miniCTX2 containing the PAO1 migA gene	34
miniCTX-migAR92Q	R92Q site-directed mutant of PAO1 migA	This study
miniCTX-migAV170M	V170M site-directed mutant of PAO1 migA	This study
miniCTX-migAR92Q-V170M	PAO1 migA containing two point mutations, R92Q and V170M; the amino acid sequence is the same as that encoded by PA14 migA	This study
pHERDrmd	pHERD20T vector containing the PA14 rmd gene	This study

Site-directed mutagenesis of wbpX and migA.

Three point mutants of wbpX_PAO1 (G20R, A303V, and A450T) and two point mutants of migA_PAO1 (R92Q and V170M) were constructed by following the QuikChange mutagenesis protocol (Agilent), using the primers listed in Table S1 in the supplemental material and with pHERDwbpX and miniCTX-migA, respectively, used as templates. The mutant constructs were transformed into PA14, PAO1(wbpX::lacZ), or PAO1(migA::Gm^r) by electroporation to test their in vivo activity.

LPS preparation, SDS-PAGE, silver staining, and Western immunoblotting.

LPS were prepared from P. aeruginosa strains by using the method of Hitchcock and Brown (42). Glycine SDS-PAGE gels at 12.5% were cast according to the manufacturer's instructions (Bio-Rad) and used for analysis of OSA and CPA LPS. Silver staining of SDS-PAGE-resolved LPS was performed by the superfast method described by Fomsgaard et al. (43), and the Western immunoblotting protocols used were as described previously (25, 43). To characterize the core OS of PA14, Tricine SDS-PAGE gels, which provide a higher resolution for low-molecular-weight molecules, were used following a previously described protocol (34).

Biofilm development assay and SEM.

Biofilm development was assayed according to previously described methods (30, 44). Briefly, each strain was grown in tryptic soy broth (TSB) to an optical density at 600 nm (OD₆₀₀) of 0.5. Two-hundred-microliter aliquots of a 1:40 dilution of each culture were then added in triplicate to a 96-well culture plate. Plates were incubated for 24 h, 48 h, and 72 h at 37°C, after which the cultures were removed and the wells were washed three times with phosphate-buffered saline (PBS) and left to dry for 1.5 h. Next, 2% crystal violet was added to each well and left for 15 min, and then the wells were washed at room temperature with distilled H₂O to remove excess stain and left to dry for 1.5 h. Finally, glacial acetic acid (33% [vol/vol]) was added to the wells to release the stain, and the absorbance at 600 nm (A₆₀₀) of each well was measured. For scanning electron microscopy (SEM), each strain was grown for 24 h, 48 h, and 72 h, as described above, in 96-well plates containing plastic coverslips. After incubation, the coverslips were fixed, dehydrated, critical point dried, coated with 15 nm of gold by using an Emitech K550 sputter coater as described previously (30), and were then viewed using a Hitachi S-570 scanning electron microscope.

Isolation and qualitative and quantitative assessments of OMV.

OMV were isolated mainly following the procedure described previously (30). Briefly, different P. aeruginosa strains were grown in 800 ml TSB (Becton, Dickinson, Mississauga, Ontario, Canada) for 16 h at 37°C in a rotary shaker at 200 rpm. For strains containing plasmids [PA14(pHERD20T), PA14(pHERDwbpX-Z), PA14(rmd)(pHERD20T), and PA14(rmd)(pHERDrmd)], carbenicillin was added to a concentration of 200 μg ml⁻¹. Whole cells were removed by centrifugation at 12,000 × g for 10 min at 4°C, and the resulting supernatant, containing OMV, was subjected to ultracentrifugation at 150,000 × g for 1.5 h at 4°C. The pellet was resuspended in 25 mM HEPES buffer (pH 7.4), and the remaining cellular debris was removed via syringe filtration through a 0.4-μm-pore-size cellulose acetate membrane filter. The filtrate was centrifuged at 21,000 × g for 30 min at 4°C, and the pellet, containing OMV, was resuspended in 100 μl of 25 mM HEPES. To quantify OMV production, 10 μl of purified OMV was assessed in triplicate by using the fluorescent lipophilic dye FM4-64 (Molecular Probes, Burlington, Ontario, Canada) as described previously (30). Transmission electron microscopy was used to examine the OMV samples of each strain, as described previously (30). The protein content of the OMV produced by each strain was analyzed using SDS-PAGE via silver staining following a previously described procedure (30).

Statistical analysis.

Prism 5 (GraphPad Software, Inc., San Diego, CA) was used to complete the statistical analysis of the data. One-way analysis of variance (ANOVA) with Tukey-Kramer multiple-comparison posttests was used to analyze OMV production and cell lysis. Information describing the distribution of OMV sizing data was calculated using Microsoft Excel (Microsoft, Redman, WA). Since the data for biofilm experiments did not satisfy the homogeneity-of-variance assumption required by a parametric ANOVA, the Kruskal-Wallis test was completed. Dunn's multiple-comparison tests were used to evaluate specific differences within and between the 24-h, 48-h, and 72-h time points. The level of significance was set at an α value of 0.05 for all tests.

RESULTS

PA14 produces OSA LPS reactive to O10 antiserum.

The O-specific antigen (OSA) gene clusters of the 20 IATS serotyping strains are divided into 11 highly divergent groups. The O10 and O19 OSA gene clusters, of ca. 16 kb, are identical to each other (17). We examined the 16-kb OSA gene cluster sequence of PA14 (containing 12 genes, from wzz to wbpM) against the 20 sequenced OSA gene clusters by BLASTN analysis, and we determined that it is almost identical to those of the O10 and 019 OSA gene clusters, with only two nucleotide differences among 15,607 nucleotides. Note that the O10 and O19 O-repeat units are also identical to each other, with the exception that the O10 O antigen has an O-acetyl substitution in C-2 of the l-Rha residue, i.e., O19 has a trisaccharide repeat unit of (→3)-l-Rha-(α1→4)-l-GalNAcA-(α1→3)-d-QuiNAc-(α1-, while O10 has a trisaccharide repeat unit of →3)-l-Rha2Ac-(α1→4)-l-GalNAcA-(α1→3)-d-QuiNAc-(α1-) (13). Since the gene encoding the enzyme for the acetylation of l-Rha is not located within the 16-kb OSA gene cluster, one cannot specifically assign PA14 to O10 or O19 based on in silico serotyping alone. The LPS of PA14 and of O10 and O19 strains were prepared and analyzed by SDS-PAGE, silver staining, and Western immunoblotting. From the silver-stained gel, it is clear that the modal chain lengths of the LPS produced by the three strains are different. All three strains are able to produce very-long-chain LPS molecules (Fig. 3); however, O19 also exhibits a proportion of long-chain LPS molecules, while O10 produces both long- and medium-chain-length LPS molecules. When the gel was immunostained with MAb MF76-2, which is specific against O10 OSA LPS (45), reactivity of the antibody to LPS bands of PA14 and O19 was observed, including intense bands in the high-molecular-weight (HMW) or very-long-chain LPS area of the sample lanes. As expected, LPS from the homologous O10 strain also reacted with MF76-2, but the intensity of the signal was stronger. Compared to PAO1 and the O10 and O19 strains, PA14 appeared to produce proportionally more short-chain LPS (core + 2, core + 3, etc.) than HMW LPS (Fig. 3 and 4). The inner core-specific MAb 5c-7-4 (35) was included with the antibody probes to standardize sample loading, because the inner core OS is conserved among all 20 serotypes, and as anticipated, reactivities were apparent for the core, core + 1, and core + 2 bands (Fig. 4).

FIG 3.

Characterization of the phenotype of OSA-containing LPS of PA14. (A) Silver-stained SDS-PAGE gel. (B) Western blot probed with MAb MF76-2 (specific for IATS O10). O19 and PA14 LPS show blotting signals of similar intensities, both of which are weaker than that of O10 LPS. The negative control, PAO1 LPS, showed no blotting signal.

FIG 4.

PA14 lacks an intact uncapped core structure due to point mutations in migA. (A) Silver-stained Tricine SDS-PAGE gel. The black arrow indicates the intact uncapped core band of PAO1, the green arrow indicates the truncated uncapped core band of PA14, and the red arrow indicates the core + 1 band of PA14. (B) Western blot obtained using MAb 5c-101 (outer core specific). (C) Western blot obtained using MAb 5c-7-4 (inner core specific).

PA14 lacks an intact uncapped core structure due to point mutations in the migA gene.

When P. aeruginosa LPS is analyzed by SDS-PAGE and silver staining, LPS with an uncapped core, which has the lowest MW, migrates the fastest and forms the lowest band on an SDS-PAGE gel. The next band in the gel, above this fast-migrating band, is the core + 1 O-repeat band, i.e., core OS with one O repeat of OSA (35). When the PA14 LPS were examined on Tricine SDS-PAGE gels, the lowest band observed (Fig. 4A, green arrow) migrated faster than the uncapped core band of strain PAO1 (indicated by a black arrow in Fig. 4A). This indicates that the uncapped core of PA14 has a lower MW than that of PAO1. Directly above this lowest band is a darkly stained band (Fig. 4A, red arrow) that migrated slightly slower than the uncapped core band of PAO1 (Fig. 4A, black arrow), indicating that it is most likely the core + 1 O-repeat band of PA14. Both of the aforementioned bands reacted with the inner core-specific MAb 5c-7-4 in a Western blot (Fig. 4C), indicating that the inner core OS was present in the uncapped core and core + 1 LPS of PA14. However, when the LPS of PA14 was probed with the outer core-specific MAb 5c-101 in a Western immunoblot, no reactivity could be discerned (Fig. 4B). The outer core-specific MAb 5c-101 recognizes the Glc^IV residue of uncapped core (37). Members of our lab previously showed that the glycosyltransferase WapB is responsible for the addition of the terminal Glc^IV residue to the uncapped core (Fig. 2) (32), so we examined the genome sequence of PA14 (accession no. GCA_000404265.1) for the presence of a wapB homolog gene. Surprisingly, we found not only that PA14 contains a wapB homolog gene (PA14_51220) and the upstream promoter region but also that both wapB_PA14 and its upstream DNA region show 100% sequence identity to those of PAO1 (data not shown). This means that the lack of the Glc^IV residue in the uncapped core of PA14 is not due to the lack of a functional wapB gene but rather to the lack of other functional glycosyltransferases involved in uncapped core assembly.

The activities of the PAO1 rhamnosyltransferases MigA and WapR are critical for determining the ratio of synthesized uncapped/capped core LPS (Fig. 2). The migA knockout mutant produces a truncated core that lacks the α1-6-linked l-Rha moiety in its core, as well as the terminal Glc^IV residue that requires the former as a receptor sugar; hence, the LPS sample of such a mutant exhibits a low-MW band that migrates faster than the uncapped core OS band on an SDS-PAGE gel (33, 34, 46), and such a band is not recognized by MAb 5c-101. Comparing the sequences of migA_PA14 and migA_PAO1 revealed that the former contains SNPs resulting in two amino acid substitutions: R92Q and V170M. To verify if these amino acid substitutions are responsible for abrogating the function of MigA, leading to blockage of the synthesis of the intact uncapped core in PA14, we decided to complement PA14 with migA_PAO1. We prepared a migA_PAO1 construct in the integration plasmid miniCTX2, yielding miniCTX-migA_PAO1, and then transformed this construct into PA14 to facilitate expression of a single copy of migA_PAO1 to avoid a gene dosage effect. The complemented strain harboring miniCTX-migA_PAO1 showed restored synthesis of the intact uncapped core OS structure, as indicated by reaction with the outer core-specific MAb 5c-101 (Fig. 4B). When the membrane was blotted with the inner core-specific MAb 5c-7-4, the uncapped core band (lower band) of PA14 was weakly stained (Fig. 4C), while the core + 1 band (upper band) was intensely stained (Fig. 4C), indicating that the PA14 strain produces more core + 1 LPS species than uncapped core LPS, while for the complemented strain PA14(miniCTX-migA_PAO1), the uncapped core band (lower band) was intensely stained, and the intensity of the core + 1 band (upper band) was reduced. These results indicated that the level of synthesis of uncapped core LPS was increased, while the level of synthesis of core + 1 LPS was reduced. This could be explained by the competition between MigA and WapR for the same substrate, as observed previously (34). In wild-type PA14, without a functional MigA protein, more substrate is available for WapR to synthesize core + 1 LPS. However, when PA14 was complemented with a functional migA_PAO1 gene, less substrate was available for WapR. These results unequivocally show that PA14 produces a truncated uncapped core LPS due to point mutations in the migA_PA14 gene.

To elucidate which of the single amino acid substitutions, i.e., R92Q, V170M, or both, might be responsible for the dysfunction of MigA_PA14, two single site-directed mutants (SDMs) of PAO1 migA (migA^R92Q and migA^V170M) and a double point mutant (migA^R92Q-V170M) were prepared. All three constructs were then used to complement the migA knockout strain PAO1(migA::Gm^r). The PAO1(migA::Gm^r) knockout mutant produced a truncated uncapped core OS that ran faster than the intact uncapped core on SDS-PAGE, and this fast-migrating band did not react with the outer core-specific MAb 5c-101 (33, 34) (Fig. 4B). Complementing the PAO1(migA::Gm^r) knockout mutant with miniCTX-migA_PAO1 restored the biosynthesis of the intact uncapped core, which showed reactivity with MAb 5c-101 (Fig. 4B). The use of the miniCTX-migA^R92Q construct also successfully complemented the PAO1(migA::Gm^r) knockout to restore the biosynthesis of the intact uncapped core structure (Fig. 4B). This shows that the R92Q amino acid substitution has very little effect on the function of MigA. When the miniCTX-migA^V170M construct was used in the complementation experiment, a band that corresponded to the truncated core OS migrating faster than the intact core was observed in the silver-stained SDS-PAGE gel (Fig. 4A). However, on Western blots with MAb 5c-101, a very weak band appeared, indicating that a small amount of intact core had been synthesized (Fig. 4B). These results suggested that the amino acid substitution V170M affected the glycosyltransferase activity of the migA_PA14 gene product, though not completely abrogating its function. The miniCTX-migA^R92Q-V170M construct could not complement the PAO1(migA::Gm^r) strain, and no uncapped core OS band was detected (Fig. 4B). Based on these results, we concluded that both the R92Q and V170M amino acid substitutions are required for the complete abrogation of MigA function in PA14.

PA14 lacks CPA LPS due to a point mutation in the wbpX gene.

When PA14 LPS was probed with MAb N1F10 (specific for CPA LPS) (19) in Western blots, no signal could be detected, indicating that strain PA14 lacks CPA LPS (Fig. 5B). Comparing the CPA clusters between PAO1 and PA14 showed that they are highly conserved (∼99% sequence identity). Hence, it was not immediately apparent which CPA biosynthesis gene is nonfunctional and would lead to abrogation of CPA biosynthesis in PA14. To determine which CPA biosynthesis gene is not functional in PA14, we first transformed different sets of CPA genes from PAO1 into PA14 by using the broad-host-range vector pHERD20T and examined the restoration of CPA biosynthesis. Our results showed that the plasmid construct containing the three glycosyltransferase genes, wbpX_PAO1 to -Z_PAO1 (pHERDwbpX-Z_PAO1), was able to restore CPA biosynthesis in PA14 (Fig. 5B). Compared to the proteins in PAO1, WbpX_PA14 has three amino acid substitutions (G20R, A303V, and A450T), WbpZ_PA14 has two substitutions (T181A and E338D), and WbpY_PA14 is identical to WbpY_PAO1. We therefore hypothesized that the changes in wbpX, wbpZ, or both account for the abrogation of CPA biosynthesis in PA14. To test this, we transformed the wbpX_PAO1 and wbpZ_PAO1 genes individually into PA14, and we found that wbpX_PAO1 alone was able to restore the biosynthesis of CPA in PA14 (Fig. 5C), suggesting that the point mutations in wbpX must render it nonfunctional and account for the lack of CPA LPS in PA14. It should be noted that when strain PA14 was complemented with either wbpXYZ or wbpX alone, the Western blot signal was much weaker than that of the PAO1 wild-type strain. This was likely due to overexpression of the wbpX gene. In fact, we previously found that when wbpXYZ were overexpressed in PAO1, the total amount of CPA synthesized was reduced (25). Another possible reason is that the cognate inactive wbpX_PA14 gene that was still present competed with the introduced PAO1 wbpX gene from the complementation plasmid for binding of the substrates.

FIG 5.

PA14 does not synthesize CPA LPS due to the lack of a functional wbpX gene. (A) Silver-stained SDS-PAGE gel. (B) Western blot probed with CPA-specific MAb N1F10. Providing the three contiguous glycosyltransferase genes, wbpXYZ, in the broad-host-range vector pHERD20T restored the biosynthesis of CPA. (C) Western blot probed with the CPA-specific MAb N1F10. Providing wbpX alone was sufficient to restore CPA biosynthesis in PA14.

To determine which particular amino acid substitution observed in WbpX_PA14 is responsible for abrogating its function, we constructed the three corresponding SDMs of WbpX_PAO1 (G20R, A303V, and A450T) and used these to complement a wbpX transposon insertion knockout mutant strain [PAO1(wbpX::lacZ)]. We found that two of the three SDMs (wbpX^A303V and wbpX^A450T) were able to complement the PAO1(wbpX::lacZ) mutant and restored CPA production to the same level as that when wild-type wbpX_PAO1 was used for complementation. In contrast, the wbpX^G20R mutant was not able to complement the PAO1(wbpX::lacZ) mutant (Fig. 6A). Similar results were obtained when the plasmids containing these SDM wbpX constructs were transformed into strain PA14 (Fig. 6B): only the G20R mutant was unable to restore CPA biosynthesis in PA14. These results convincingly show that the single amino acid substitution G20R is responsible for the loss of function of WbpX_PA14 and the lack of CPA biosynthesis in PA14.

FIG 6.

The point mutation G20R in wbpX_PA14 is responsible for the loss of function of this gene. (A) Complementation of the PAO1(wbpX::lacZ) knockout mutant by using wild-type and SDM constructs of wbpX. CPA biosynthesis could be restored by wild-type wbpX, wbpX^A303V, and wbpX^A450T but not by wbpX^G20R. (B) Complementation of PA14 by using wild-type wpbX from PAO1 and the three SDM constructs. Only wbpX^G20R failed to restore CPA production in PA14.

Expression of CPA LPS in PA14 stimulates mature biofilm formation.

To determine if restoring CPA biosynthesis in PA14 alters the ability of cells to form biofilms, we tested the biofilm-forming ability of three strains, PA14 (wild type; CPA⁻), PA14(pHERD20T) (empty vector control; CPA⁻), and PA14(pHERDwbpX-Z) (CPA⁺), by using a standard 96-well biofilm assay. For strains PA14(pHERD20T) and PA14(pHERDwbpX-Z), carbenicillin was added to the growth medium at 200 μg ml⁻¹. All three strains were able to grow at similar growth rates under the tested conditions (see Fig. S1 in the supplemental material). At 24 h, all three strains demonstrated similar levels of substantial biofilm growth, while at 48 h, all three strains showed a significant drop in total biofilms (Fig. 7). Compared to PA14 and PA14(pHERD20T), strain PA14(pHERDwbpX-Z) produced slightly more biofilm at 48 h, although the difference was not statistically significant. However, at 72 h, the total biofilm growth for the CPA⁺ strain, PA14(pHERDwbpX-Z), was significantly increased compared to that at 48 h (P < 0.05), while the total amounts of biofilm produced by the two CPA⁻ strains, PA14 and PA14(pHERD20T), showed no significant change during that time (Fig. 7). Hence, restoring CPA expression in PA14 significantly stimulated mature biofilm formation at 72 h (P < 0.05). The two CPA⁻ strains, PA14 and PA14(pHERD20T), behaved similarly in terms of biofilm formation, indicating that adding carbenicillin to the growth medium at 200 μg ml⁻¹ to maintain the pHERD20T plasmid in the latter strain had no effect on P. aeruginosa PA14 biofilm growth.

FIG 7.

Examination of biofilm formation by PA14, PA14(pHERD20T), and PA14(pHERDwbpX-Z). Quantification of crystal violet staining associated with biofilm formation at 24, 48, and 72 h was performed for the different P. aeruginosa O-polysaccharide-producing strains. Values represent the solubilization of crystal violet and were quantified by determining the A₆₀₀ (n = 9 independent experiments/strain). *, P = 0.05 for PA14, PA14(pHERD20T), and PA14(pHERDwbpX-Z) at 48 and 72 h versus 24 h; **, P = 0.05 for PA14(pHERDwbpX-Z; within-strain difference) at 72 h versus 24 h. Error bars represent standard deviations.

We then used SEM for high-resolution visualization of biofilm formation by P. aeruginosa PA14 (CPA⁻) and PA14(pHERDwbpX-Z) (CPA⁺) at 24 h, 48 h, and 72 h (Fig. 8). At 24 h, both strains formed robust biofilms that displayed evidence of biofilm maturation, based on the observable presence of amorphous materials identified in many biofilm studies to represent exopolysaccharide and other biofilm matrix materials. At 48 h, both strains showed a significant decrease in the total number of attached biofilm cells. At 72 h, there was a continued reduction in attached biofilm cells for the PA14 strain (CPA⁻), as only individual surface-attached cells and small colonies were observed at this point, but the PA14(pHERDwbpX-Z) strain (CPA⁺) displayed a dramatic increase in biofilm growth and formed a thick, compact, mature biofilm at this point. These observations substantiated the quantitative results obtained for each strain (Fig. 7), suggesting that expressing CPA in PA14 by using the plasmid pHERDwbpX-Z significantly increased mature biofilm formation after 72 h.

FIG 8.

Visualization of PA14 and PA14(pHERDwbpX-Z) biofilm production by SEM. Micrographs of PA14 and PA14(pHERDwbpX-Z) biofilms at 24, 48, and 72 h. Insets show higher-magnification views of the central region of each micrograph. Bars = 10 μm (main panels) and 3 μm (insets).

Expression of CPA LPS in PA14 showed no significant effect on OMV.

Outer membrane vesicles (OMV) are lipid bilayer vesicles derived from the outer membranes of Gram-negative bacteria. They have diameters ranging from 20 to 250 nm (47) and play important biological roles, including transfer of genetic information, secretion of enzymes, and release of virulence factors. A recent study by Macdonald and Kuehn (31) used an rmd transposon insertion mutant of PA14 (PA14-rmd::Tn) to study the effect of a CPA⁻ LPS phenotype on OMV biogenesis. As the evidence provided in the present study revealed that strain PA14 is already CPA⁻, we expected that the PA14-rmd::Tn mutant would have the same LPS phenotype as strain PA14. To reconcile these findings, we obtained the same rmd mutant used by Macdonald and Kuehn (31) from the transposon insertion library produced by the Ausubel lab (5) and characterized the LPS phenotype. As expected, all three strains, i.e., PA14, the PA14-rmd::Tn mutant, and the complemented strain PA14(rmd::Tn)(pHERDrmd), showed exactly the same LPS phenotype, i.e., CPA⁻ OSA⁺ (see Fig. S2 in the supplemental material). To determine if restoring CPA biosynthesis in PA14 would have an effect on OMV biogenesis, PA14, PA14(pHERD20T), and PA14(pHERDwbpX-Z) were examined for the ability to produce OMV. Data on the average size and protein content of the OMV produced by each strain were also recorded. First, to ensure that OMV production was not caused by instability in the outer membranes of the mutant cells, we assessed cell lysis by performing dot immunoblotting using an anti-RNA polymerase antibody. All three strains showed negligible levels of cell lysis after 16 h of growth (see Fig. S3). We then quantified the OMV produced by each strain. The results showed that there was no statistical difference in OMV production among the strains tested (Fig. 9A), indicating that expression of CPA in PA14 did not significantly change the quantity of OMV produced in any of the three strains. To measure the average size of OMV produced by each strain, we obtained negatively stained transmission electron micrographs of purified OMV (Fig. 9B) and measured the mean, median, range, and skew of the diameters of the OMV for each strain. No significant difference among the three strains examined could be detected (Table 2). This suggests that CPA has no significant effect on the size of OMV produced by PA14. By means of SDS-PAGE and silver staining to compare the overall protein content of the OMV produced by each strain, no significant differences in protein banding patterns could be discerned among the three strains (Fig. 9C). From these results, we can conclude that CPA does not have a significant effect on OMV biogenesis and properties in P. aeruginosa PA14.

FIG 9.

Measurements of OMV production, sizes, and protein contents. (A) Quantitative assessment of OMV production from 24-h cultures of P. aeruginosa PA14, PA14(pHERD20T), and PA14(pHERDwbpX-Z) by using the lipophilic dye FM4-64. (B) Negatively stained transmission electron micrographs of OMV purified from 24-h cultures of PA14, PA14(pHERD20T), and PA14(pHERDwbpX-Z) cells. Bars = 1 μm. (C) Purified OMV from each P. aeruginosa strain cultured for 24 h and analyzed by silver-stained SDS-PAGE. The purified OMV demonstrated for these strains show no major differences in composition.

TABLE 2.

OMV size distribution^a

Strain	OMV diam (nm)
	Mean ± SE	Median	Range	SD
PAO1	88.2 ± 1.0	87.2	21.3–231.9	31.2
PA14	82.7 ± 1.2	82.1	22.6–292.3	45.0
PA14(pHERD20T)	89.3 ± 1.1	86.8	29.4–297.6	42.8
PA14(pHERDwbpX-Z)	91.8 ± 1.3	88.6	30.7–315.2	43.9

^aValues are for OMV diameter measurements taken from three independent experiments comprising multiple micrographs for each strain. A total of 1,000 OMV were measured for each strain.

DISCUSSION

PA14 has been gaining popularity as a more virulent reference strain than the universal standard laboratory strain P. aeruginosa PAO1, and it is used by many researchers in the fields of Pseudomonas virulence and biofilm research. However, LPS, one of the important virulence factors of this organism, has not been characterized fully for this strain. To date, there is only one study, by Coulon et al. (18), who reported the structure of the O-repeat units of the OSA LPS of PA14 as a trisaccharide repeat of -4)-α-l-GalNAcA-(1-3)-α-d-QuiNAc-(1-3)-α-l-Rha-(1-) (18). This structure corresponds to that reported earlier, by Knirel's group (13, 48), as the representative structure of the O-Ag of IATS O19. The genome sequence of PA14 revealed that it contains an OSA gene cluster identical to the O10/O19 OSA gene clusters. Note that the OSA gene clusters of O10 and O19 are identical to each other (17) and that the difference between O10 and O19 O-repeat units is that in O10, the l-Rha residue is O-acetyl substituted at the C-2 position (48). The acetyltransferase gene responsible for the 2-acetyl group was not localized within the OSA cluster in the O10/O19 group; hence, this presumed acetyltransferase gene in O10 must reside elsewhere in the chromosome of the bacterium. Both the O-Ag structure and the sequences of the OSA cluster of O10 and O19 suggested that the two belong to a cross-reactive group among the 20 IATS serotypes. Hence, it was not surprising that the MAb MF76-2, produced against serotype O10 (45), readily cross-reacted with O19, and also with PA14. Western immunoblot results showed that MAb MF76-2 readily interacted with PA14 and O19 LPS but reacted weakly compared to the interaction with O10 LPS (Fig. 1). The results from the Western immunoblots suggested that PA14 LPS more closely resembles O19 LPS. However, on examining the silver-stained SDS-PAGE gels of the LPS from these strains, PA14 OSA exhibited modal chain lengths (reflected by the LPS banding pattern) that were different from those of both the O10 and O19 serotype strains. PA14 lacks the medium-chain and long-chain LPS and produces predominantly very-long-chain LPS and short-chain LPS. In PAO1, the production of long-chain and very-long-chain LPS is regulated by Wzz₁ (PA3160) and Wzz₂ (PA0938), respectively. When wzz₁ is mutated, PAO1(wzz₁) produces very-long-chain LPS and short-chain LPS but lacks the long-chain LPS; when wzz₂ is mutated, PAO1(wzz₂) lacks the very-long-chain LPS; and when both wzz₁ and wzz₂ are mutated, the double mutant produces O-Ag with a nonmodal (random) chain length (49). The wzz₂ gene is located outside the OSA gene cluster, while wzz₁ is usually located at the 5′ end of the OSA gene cluster. By performing ClustalW sequence alignment comparisons, we observed that wzz₂ of PAO1 had a very high sequence identity (99% amino acid sequence identity) to the corresponding gene in PA14 (PA14_52130), while homology comparison between wzz₁ of PAO1 and its homolog in PA14 (PA14-23360) showed that the identity remained relatively low, with an amino acid sequence identity of 56%. It is plausible that the wzz₁ homolog, PA14-23360, is not functional; hence, the modal chain lengths displayed by PA14 LPS in SDS-PAGE analysis represent a banding pattern that is a result of chain length regulation by the wzz₂ homolog, PA14_52130. Mutational analysis of wzz₁ and wzz₂ homologs of PA14 is under way in our lab.

As mentioned earlier, the uncapped core structure plays important biological roles. An intact uncapped core with an exposed terminal glucose is the ligand that interacts with eukaryotic cells (39). Without the intact uncapped core OS structure, the possibility that P. aeruginosa PA14 can be associated with and ingested by host epithelial cells is significantly reduced. The sugars in the uncapped core are also the recognition sites of different R-type pyocins (37, 40). Pyocins, the bacteriocins produced by P. aeruginosa, are able to kill cells of members of the same or closely related species. R-pyocins make up one of the three different families of pyocins (including the soluble S-pyocins, the flexible F-pyocins, and the rod-shaped R-pyocins) produced by P. aeruginosa (50, 51). They are genetically and morphologically related to P2 bacteriophages (50), but unlike bacteriophages, they lack a phage head structure, do not contain DNA, and are not replicative. R-pyocin is believed to kill susceptible cells by first binding to the bacterial cell surface through receptors and then contracting their sheath and inserting their core structure into the cell envelope, which finally results in host cell lysis by depolarization of the cell membrane (52). The killing efficiency of R-pyocin is very high. One particle can cause the death of a bacterium in 20 min (52). Based on the tail fiber sequences and their killing spectra, R-pyocins can be divided into five different types (R1 to R5) (53). Since the majority of clinically isolated P. aeruginosa strains produce pyocins, the pyocin typing technique became one of the standard methods for classifying P. aeruginosa strains in the microbial community related to cystic fibrosis specimens (54). Recently, it was found that different sugars in the uncapped core of P. aeruginosa are the receptors for different R-type pyocins (37, 40) (Fig. 2). For example, R1 pyocin targets the α1-6-linked l-Rha residue (transferred by MigA) in the uncapped core (40), R2 targets the Glc^II residue (40), and R3 targets the Glc^IV residue (37) (Fig. 2). Because of the mutation in migA, the truncated core produced by the PA14 strain would lack the α1-6-linked l-Rha and Glc^IV residues. This will make PA14 resistant to killing by R1 and R3, which gives it an advantage in competition with other P. aeruginosa strains or other closely related bacterial species in the environment. In this regard, point mutations in the migA gene have the potential to contribute to the virulence of P. aeruginosa PA14.

CPA was found to be a receptor for bacteriophage A7 (27) and for lectin-like bacteriocins (28). Lectin-like bacteriocins, such as pyocin L1 and putidacin L1, display a genus-specific killing activity. McCaughey et al. (28) reported that recognition of CPA occurs through binding of d-Rha at the conserved QxDxNxVxY sugar binding motifs of the C-terminal lectin domain and that this interaction is a prerequisite for bactericidal activity. The lack of CPA in PA14 should make it resistant to killing by these bacteriocins and provide an advantage during competition against other P. aeruginosa bacteria in a mixed bacterial community environment.

Since the CPA biosynthesis cluster can be localized in the whole-genome sequence of PA14, it is not surprising that some researchers would assume that the bacterium should have a CPA⁺ phenotype. This turned out to be the case in a recent study where the authors designed their experiments to include an rmd::Tn mutant of PA14 as a CPA⁻ strain and used it for comparison to wild-type PA14 in order to investigate the effect of having CPA LPS on OMV biogenesis (31). Our results unequivocally demonstrate that strain PA14 cannot produce CPA due to a G20R amino acid substitution in the glycosyltransferase WbpX. To substantiate our finding about the CPA⁻ phenotype of PA14, we characterized the LPS of the same rmd::Tn mutant and showed that the LPS phenotype of this mutant is the same as that of wild-type PA14, i.e., both are CPA⁻. By transforming a plasmid containing the wbpXYZ genes from PAO1, we observed CPA production in the PA14 transconjugant bacteria. We then investigated the effect of expressing CPA in PA14 on OMV production, and the results showed that CPA had no effect on OMV biogenesis and properties. This is in good agreement with the findings from similar experiments conducted in the strain PAO1 background by Murphy et al. (30). In their study, it was also observed that a PAO1 CPA⁻ mutant strain could not develop into robust biofilms and exhibit changes in cell morphology and biofilm matrix production (30), indicating that CPA plays an important role in the development of a mature biofilm. In this study, we found that in accordance with the findings for PAO1, restoring the expression of CPA in PA14 significantly stimulated mature biofilm formation after 72 h.

In conclusion, the importance of PA14 as a standard P. aeruginosa strain used by many laboratories worldwide warrants a better understanding of the LPS phenotype of this strain. In this study, we performed a systematic characterization of the LPS of P. aeruginosa PA14. Based on genome sequence data and immunochemical analysis, we confirmed that the OSA of PA14 belongs to O19. LPS produced by PA14 and the O19 strain cross-reacted with MAb MF76-2, which is specific to O10 LPS. PA14 produces LPS of either a short chain length or very long chain length and lacks glycolipid moieties with a long chain length; this LPS characteristic is consistent with that exhibited by the wzz₁ mutant of strain PAO1. Although a CPA biosynthesis cluster is localized in the genome of PA14, this strain is defective in CPA production due to a SNP resulting in a G20R substitution in the rhamnosyltransferase WbpX. PA14 also harbors mutations in migA, which explains why this strain cannot produce an intact uncapped core OS. Since CPA and an intact core could both serve as receptors for bacteriocins, the lack of CPA and an intact uncapped core likely provides an advantage to PA14 during competition against other P. aeruginosa strains, contributing to its virulence. Finally, the role of CPA in PA14 biofilm formation and OMV biogenesis is similar to that observed in PAO1.