Arch Microbiol (2015) 197:135–145 DOI 10.1007/s00203-014-1030-y

ORIGINAL PAPER

Expression of the lipopolysaccharide biosynthesis gene lpxD affects biofilm formation of Pseudomonas aeruginosa Sahar A. Alshalchi · Gregory G. Anderson 

Received: 9 July 2014 / Revised: 13 August 2014 / Accepted: 15 August 2014 / Published online: 31 August 2014 © Springer-Verlag Berlin Heidelberg 2014

Abstract  Bacterial biofilms are an important cause of nosocomial infections. Microorganisms such as Pseu‑ domonas aeruginosa colonize biotic and abiotic surfaces leading to chronic infections that are difficult to eradicate. To characterize novel genes involved in biofilm formation, we identified the lpxD gene from a transposon-mutant library of P. aeruginosa. This gene encodes a glucosamineN acyltransferase, which is important for lipopolysaccharide biosynthesis. Our results showed that a loss-of-expression mutant of lpxD was defective for biofilm formation on biotic and abiotic surfaces. Additionally, this mutant strain exhibited significantly decreased bacterial attachment to cultured airway epithelial cells, as well as increased bacterial cytotoxicity toward airway cells. However, consistent with a defect in lipid A structure, airway cells incubated with the lpxD mutant or with mutant lipid A extracts exhibited decreased IL-8 production and necrosis, respectively. Overall, our data indicate that manipulating lpxD expression may influence P. aeruginosa’s ability to establish biofilm infections. Keywords  Lipopolysaccharide · Biofilm · Pseudomonas aeruginosa · Cystic fibrosis · Necrosis

Communicated by Erko Stackebrandt. S. A. Alshalchi · G. G. Anderson (*)  Department of Biology, Indiana University Purdue University Indianapolis, 723 West Michigan Street, SL 320, Indianapolis, IN 46202, USA e-mail: [email protected]

Introduction Pseudomonas aeruginosa is a Gram-negative opportunistic pathogen, commonly isolated from infections associated with biofilm formation, including infections of burns and other wounds, otitis media, and complicated lung infections of patients with cystic fibrosis (CF). In fact, it is regarded as one of the most medically relevant biofilm forming bacterial species, contributing significantly to morbidity and mortality in CF patients (Costerton et al. 1999; Gibson et al. 2003; Doring et al. 2011). P. aeruginosa biofilm cells are embedded in a complex matrix of polysaccharides, proteins, and extracellular DNA (eDNA), which protect the bacteria from phagocytes, deleterious compounds released from immune cells, and antimicrobial treatment (Davey and O’Toole 2000; Ryder et al. 2007; Doring et al. 2011). Eventually, persistence of P. aeruginosa in CF patients and the resultant chronic inflammation lead to airway destruction and fibrosis. Bacterial colonization becomes lifelong, contributing to CF patient complications, including respiratory failure and death (Costerton et al. 1999; Doring et al. 2011). The expression of virulence genes in P. aeruginosa has been shown to be reciprocally regulated, with some factors produced during chronic infections, associated with biofilm formation, and others produced during acute planktonic infections. Therefore, biofilm cell gene expression differs profoundly from that in suspended counterparts (Sauer et al. 2002; Ventre et al. 2006). The dramatic changes in gene expression of biofilm cells lead to many phenotypic alterations, including decreased expression of the type III secretion system (T3SS). The toxic effect of T3SS appears to result in host cell death and stimulation of inflammatory responses in mammalian cells (Coburn et al. 2007). The regulation of T3SS expression affects the mode of pathogenesis

13

136

Arch Microbiol (2015) 197:135–145

of P. aeruginosa, which has been demonstrated through cell culture and animal models (Augustin et al. 2007; Coburn et al. 2007; Anderson et al. 2010). Another modification that P. aeruginosa undertakes during biofilm formation is in the lipopolysaccharide (LPS). LPS is a major component of the outer membrane of almost all Gram-negative bacteria. Typically, LPS is comprised of three components: hydrophilic O-polysaccharide, core-oligosaccharide, and hydrophobic lipid A (Coburn et al. 2007). P. aeruginosa produces two different LPS molecules: a conserved polyrhamnose homopolymer known as the A-band and a serotype-specific O antigen known as the B-band. The relative expression of each within the bacterium can lead to variability in binding, virulence, and phenotype. Notable differences in relative amounts of A- and B-band LPS present on the bacterial surface are particularly evident between planktonic and biofilm-state bacterial culture (Beveridge et al. 1997; Valvano 2003; Augustin et al. 2007). The lipid A portion of LPS is primarily responsible for bioactivity by triggering the biosynthesis of diverse mediators of inflammation (Chen et al. 2004; Drewniak et al. 2010; He et al. 2013). The key genes encoding lipid A biosynthesis in P. aeruginosa (lpxA, lpxC, and lpxD) are thought to be crucial for bacterial viability. These genes are highly conserved and are present as a single copy among most Gram-negative bacteria, and so offer potential as therapeutic targets (Gon et al. 2004; Raetz et al. 2007; Bodewits et al. 2010). LpxD protein is a soluble homotrimeric enzyme; it catalyzes the third step in LPS biosynthesis by transferring a 3-hydroxydecanoyl moiety from acyl carrier protein (ACP) to UDP-3-O-(3-hydroxydecanoyl) glucosamine (Buetow et al. 2007; Badger et al. 2011). Crystallographic studies demonstrate that the central domain of LpxD from P. aeruginosa, Escherichia coli, and Chlamydia trachomatis is highly conserved (Badger et al. 2011). On the other hand, structural LPS modifications have been observed involving both the hydrophilic region as well as hydrophobic acyl chain domains of lipid A in some Gramnegative bacteria. In P. aeruginosa, these changes include Table 1  Strains and plasmids

Strains and plasmids

differences in acyl chain length and distribution, which could affect the virulence of the bacterium (Kulshin et al. 1991; Ernst et al. 1999). It has been shown that a specific LPS variant is synthesized in bacteria colonizing the CF airways, in response to environmental changes, and this alternative LPS is associated with chronic infection of P. aeruginosa (Cryz et al. 1984; Ernst et al. 1999). These modifications increase antibiotic resistance and alter LPSmediated immune inflammatory responses (Cryz et al. 1984; Ernst et al. 1999; Raetz and Whitfield 2002). Many effects of lipid A on mammalians cells are secondary to the overproduction of different kinds of cytokines, such as tumor necrosis factor-α (TNF-α) and interleukin1β (IL1β), by macrophages (Beutler and Cerami 1988; Galanos and Freudenberg 1993). These secondary peptide bioactive mediators are capable of inducing many inflammatory responses in host cells by acting independently, in sequence, synergistically, or antagonistically (Galanos and Freudenberg 1993; Rietschel et al. 1994). In our study, we present an analysis of biofilm phenotype alterations induced by mutation of the LPS biosynthesis gene lpxD. This mutant strain was isolated from a library of transposon insertional mutants of P. aeruginosa. To our knowledge, this gene has not previously been identified in global screens of biofilm formation. The alteration in LPS structure in the mutant strain increased the expression of T3SS and induced cytotoxic effects on airway-derived epithelial cells, while decreasing cytokine production and necrosis. We suggest that these lpxD-mediated activities may lead to new therapeutic strategies to control inflammation of biofilm-implicated infections of P. aeruginosa.

Materials and methods Bacterial strains, plasmids, media, and antibiotics All bacterial strains and plasmids used in this study are listed in Table 1. Biofilms were assayed in LB medium

Relevant genotype

Source

P. aeruginosa PA14

Wild type

Anderson et al. (2008)

E. coli S17-1

Laboratory strain for cloning

Anderson et al. (2008)

S. cerevisiae InvSc1 SAA10

In vivo cloning; ura3-52/ura3-52

Invitrogen

ΔlpxD, isogenic lpxD mutant in PA14

This study

Plasmids pMQ30 pMQ70 pS11

Allelic replacement vector; yeast cloning PBAD expression vector

Shanks et al. (2006) Shanks et al. (2006)

Deletion of lpxD; pMQ30 backbone

This study

plpxD

Full-length gene lpxD; backbone pMQ70 (pMQ70::lpxD)

This study

Strains

13

Arch Microbiol (2015) 197:135–145

at 37 °C. For all experiments, bacteria were grown in LB medium overnight with an appropriate antibiotic concentration for plasmid selection pressure: 50 µg/ml gentamicin, 20 µg/ml nalidixic acid, 50 µg/ml carbenicillin, or 250 µg/ ml carbenicillin. Generation of a transposon‑mutant library To generate a library of transposon mutants, we conjugated P. aeruginosa strain PA14 with Escherichia coli S17-1 carrying plasmid pBT20. This plasmid contains the Mariner transposon (Simon et al. 1989). Mutants were selected on LB plates with 50 µg/ml gentamicin and 20 µg/ml nalidixic acid. Biofilm formation assay Biofilm formation was assayed as previously described (O’Toole and Kolter 1998b). Briefly, 100 µl of overnight cultures, diluted 1:100 into LB medium, were dispensed into wells of 96-well polyvinyl chloride (PVC) microtiter plates (BD Falcon, Bedford, MA). To measure biofilm formation, the plates were incubated statically at 37 °C for

137

24 h. At this point, PVC plates were stained with 0.1 % crystal violet (CV) for 10 min. The plates were then rinsed thoroughly with water. The CV was solubilized with 125 µl of 30 % glacial acetic acid for 10 min. Then, the CV was transferred to a flat bottom polystyrene microtiter plate (Greiner bio-one), and absorbance was determined at 550 nm in a SpectraMaxM2 spectrophotometer (Molecular Devices, Sunnyvale, CA). Genetic techniques Arbitrary PCR To determine the DNA sequence flanking the transposon insertion in our mutant, we used arbitrary-primed PCR, as previously described (O’Toole et al. 1999). The primers we used for this procedure are listed in Table 2. Construction of a mutant strain We sought to generate an independent lpxD mutant strain, using a method we have previously used (Anderson et al. 2008). Briefly, 1,000 bp immediately flanking upstream and

Table 2  Primers used in this study Primer

Sequence (5’–3’)

Arbitrary primers ARB1 ABR6 P237 P238 ABR2 P240 P241

GGCCACGCGTCGACTAGTACNNNNNNNNNNGATAT GGCCACGCGTCGACTAGTACNNNNNNNNNNACGCC GGCCACGCGTCGACTAGTACNNNNNNNNNNAGAG TATAATGTGTGGAATTGTGAGCGG GGCCACGCGTCGACTAGTAC ACAGGAAACAGGACTCTAGAGG CACCCAGGTTTCTTGTACAC

Deletion primers S1Ufor S1Urev S1Dfor S1Drev

TCGACTGAGCCTTTCGTTTTATTTGATGCCTGGCAGTTCCACTACTACGCCGGCGGCTTC GGTGACGCAAAAAACGTCAACAACGCAGCTGGTTCATCCG CGGATGAACCAGCTGCGTTGTTGACGTTTTTTGCGTCACC GGAATTGTGAGCGGATAACAATTTCACACAGGAAACAGCTAGGCCGGGACGTCCTTGCCG

Complementation primers CS1for TACTGTTTCTCCATACCCGTTTTTTTGGGCTAGCGAATTCAGGAGGAGCGCGCTATGATGAGTACCTTGTCCTA CS1rev CTGTATCAGGCTGAAAATCTTCTCTCATCCGCCAAA ACAGTTACGCATCAGATGAAGCGT P729 CAGACCGCTTCTGCGTTCTG P730 GCAACTCTCTACTGTTTCTCC qRT-PCR primers PA5110for PA5110rev LPXDfor LPXDrev ExsAfor

CCTACCTGTTGGTCTTCGACCCG GCTGATGTTGTCGTGGGTGAGG CTATGCGGTGATCGAGAGTG CGATCAGCGTATCGGACAG TGCAAGTCTCGTTCGTTGTC

ExsArev

CGGGAGTACTGCTTATCGTC

13

138

downstream lpxD were amplified using the primer pairs S1Ufor/S1Urev and S1Dfor/S1Drev (Table 2). BamHIdigested pMQ30 was mixed with PCR fragments, and this mixture was transformed into Saccharomyces cerevisiae (Shanks et al. 2006). The isolated plasmids were transformed into Escherichia coli strain S17-1 by electroporation, and the plasmid was transferred to P. aeruginosa PA14 by conjugation (Kuchma et al. 2005). The exconjugants were selected on LB plates containing 50 µg/ml gentamicin and 20 µg/ml nalidixic acid. The grown cells were then cultured overnight in LB and plated on 10 % sucrose LB agar. Creation of a complementation plasmid The full length of lpxD was amplified from P. aeruginosa strain PA14 using the primer pair listed in Table 2. This fragment was joined into EcoRI-digested pMQ70, by homologous recombination in S. cerevisiae as described above (Shanks et al. 2006), creating the recombinant plasmid plpxD (Table 1). The plasmid was maintained in E. coli strain S17-1, and the plasmid construct was confirmed by PCR using the primer pair p730/p729 (Table 2). P. aer‑ uginosa strains were transformed with this complementation plasmid by electroporation and selected on LB plates containing 250 µg/ml carbenicillin. DNA sequencing DNA sequencing was performed at the fee-for-service DNA Sequencing Core Facility at the Indiana University School of Medicine (Indianapolis, IN). Attachment and biofilm assays on CF‑derived cells For tissue culture experiments, we used CF-derived airway epithelial cells (CFBE) originally developed from a CF individual homozygous for the ΔF508-CFTR mutation (Anderson et al. 2008). We developed biofilms of wild type and mutant strains on CFBE cells as previously described (Anderson et al. 2008). Briefly, we seeded 2 × 105 cells/ well in 24-well tissue culture plates (Falcon, Franklin, NJ) in minimal essential medium (MEM)(Mediatech, Herndon, VA) supplemented with 10 % fetal bovine serum (Atlanta Biologicals, Lawrenceville, CA), 2 mM l-glutamine, 50U/ ml penicillin, and 50U/ml streptomycin (Lonza, Walkersville, MD). After incubation for 8–10 days at 37 °C and 5 % CO2, the medium was replaced with 0.5 ml MEM/ well (without phenol red, serum, or antibiotics) containing 1.2 × 107 CFU/ml phosphate-buffered saline (PBS)washed bacteria. These wells were incubated for 1 h at 37 °C and 5 % CO2. To assay bacterial attachment to CFBE cells, wells were washed 3–4 times with 0.5 ml PBS and treated with 10 % Triton X-100 in PBS for 10 min to

13

Arch Microbiol (2015) 197:135–145

lyse the CFBE cells. Lysates were harvested, vortexed for 3 min, serially diluted, and plated on LB. Biofilm formation on CFBE cells was assayed by inoculating the cells with P. aeruginosa as above, and after the initial 1-h incubation time, the supernatant was replaced with 0.5 ml of fresh MEM supplemented with 0.4 % arginine. After 4-h incubation at 37 °C and 5 %CO2, the CFU/ml of bacterial cells was enumerated as described above. Cytotoxicity assay Cytotoxicity of tested strains was measured by quantifying lactate dehydrogenase (LDH) released from CFBE cells after 6-h incubation post-inoculation (Anderson et al. 2008), using the CytoTox 96® NonRadioactive Cytotoxicity Assay kit (Promega, Madison, WI) according to the manufacturer’s instructions. A value of 100 % cytotoxicity represents the average value of the supernatant from uninoculated CFBE cells after addition of 1 % Triton X-100, which resulted in a complete lysis of the epithelial cells. Baseline, spontaneous-release LDH values (0 % cytotoxicity) were obtained by quantification of LDH in wells of uninoculated CFBE cells in the absence of Triton X-100. Quantitative RT‑PCR Gene expression of lpxD and exsA was measured by qRTPCR in the WT and lpxD mutant strain as follows. RNA isolation RNA was isolated using the Qiagen RNeasy kit (Qiagen, Valencia, CA) from 1 ml of overnight bacterial cultures. Cultures were centrifuged for 5 min at 4 °C and resuspended in 100 µl of lysozyme-containing TE buffer (pH 8). The manufacturer’s protocol was then followed, including on-column DNase digestion. Isolated RNA was subjected to a second DNaseI treatment (Roche, Diagnostics, Germany) for 1 h at 37 °C, followed by purification by Qiagen RNeasy columns, according the manufacturer’s instructions. The purity of extracted RNA was checked by PCR using the PA5110for/rev primer pair (Table 2). The concentrations of prepared RNAs were measured using a NanoDrop spectrophotometer. To monitor expression of the exsA gene, total RNA of tested strains was isolated under low calcium concentrations to induce T3SS expression, as described previously (Wolfgang et al. 2003). Briefly, overnight cultures of wild type and mutant strains were diluted 1:100 with LB medium in the presence of 5 mM ethylene glycolbis (2-aminoetheylether)-N,N,N′,N′–tetraacetic acid (EGTA) and incubated to mid-exponential growth phase (OD600 = 0.5).

Arch Microbiol (2015) 197:135–145

cDNA synthesis cDNA was synthesized using the SuperScript III FirstStrand Synthesis System for RT-PCR kit (Invitrogen, Carlsbad, CA) with 1 µg of RNA, according to the manufacturer’s instructions.

139

instructions. This staining kit incorporates Hoescht 33342 to detect nuclear morphology (blue), propidium iodide to detect weakened membranes (necrosis, red), and fluoromethyl ketone inhibitor of caspase to detect apoptosis (green). Images were analyzed by an EVOSfl microscope. Swarming motility

qRT‑PCR analysis Quantitative real-time PCR (qRT-PCR) analysis was performed with SYBR Green PCR master mix (Applied Biosystems, CA) in a total volume of 25 µl, using the primer pairs corresponding to the lpxD and exsA genes (Table 2). PCR was carried out in an Applied Biosystems 7300 realtime PCR system as follows: 1 cycle of 50 °C for 2 min, 95 °C for 10 min, followed by 40 cycles of 95 °C for 15 s, 60 °C for 1 min, and 70 °C for 1 min. The dissociation step was 95 °C for 15 s, 60 °C for 1 min, and 95 °C for 15 s. The fbp gene was used as a normalization control (PA5110for/ rev primers).

Swarming plates were prepared as described previously (Xavier et al. 2011). Briefly, freshly prepared swarming plates, composed of M8 medium and 0.4 % agar, were dried for 3 h and spotted gently with 3 µl of overnight culture of wild type and deletion mutant strains. After a brief drying period at room temperature (a few hours), the plates were then incubated at 37 °C for 24 h. Statistical analysis A student’s t test was used to determine statistical significance, with a p value of