OUP user menu

Gene expression characteristics of a cystic fibrosis epidemic strain of Pseudomonas aeruginosa during biofilm and planktonic growth

Jim Manos, Jonathan Arthur, Barbara Rose, Scott Bell, Pholawat Tingpej, Honghua Hu, Jeremy Webb, Staffan Kjelleberg, Mark Douglas Gorrell, Peter Bye, Colin Harbour
DOI: http://dx.doi.org/10.1111/j.1574-6968.2008.01472.x 107-114 First published online: 1 March 2009

Abstract

Epidemic Pseudomonas aeruginosa have been identified in cystic fibrosis (CF) patients worldwide. The Australian Epidemic Strain-2 (AES-2) infects up to 40% of patients in three eastern Australian CF clinics. To investigate whether AES-2 isolates from chronically infected CF adults differentially express well-conserved genes potentially associated with transmissibility, we compared the transcriptomes of planktonic and biofilm-grown AES-2, infrequent P. aeruginosa clones and the reference P. aeruginosa PAO1 using the Affymetrix PAO1 array. The most interesting findings emerged from comparisons of planktonic and biofilm AES-2. AES-2 biofilms upregulated Type III secretion system genes, but downregulated quorum-sensing (QS)-regulatory genes, except lasR, QS-regulated, oxidative-stress and iron-storage genes. QS-regulated and iron-storage genes were downregulated to a greater extent in AES-2 biofilms compared with infrequent clone and PAO1 biofilms, suggesting enhanced anaerobic respiration in AES-2. Chitinase and chitin-binding protein maintained high expression in AES-2 biofilms compared with infrequent clone and PAO1 biofilms. Planktonic AES-2 upregulated QS regulators and QS-regulated genes, iron acquisition and aerobic respiration genes, and had high expression of Group III Type IV pilA compared with low expression of Group I Type IV pilA in infrequent clones. Together, these properties may enhance long-term survival of AES-2 in CF lung and contribute to its transmissibility.

Keywords
  • Pseudomonas aeruginosa
  • epidemic clone
  • quorum sensing
  • cystic fibrosis
  • microarray
  • infection

Introduction

Pseudomonas aeruginosa infection is a major determinant of morbidity and mortality in patients with cystic fibrosis (CF). Chronic infection and inflammation promote a progressive decline in lung function, leading to premature death. Over the last decade, pulsed-field gel electrophoresis (PFGE) has identified dominant P. aeruginosa clones or clonal complexes in paediatric and adult CF clinics in Australia (Anthony et al., 2002), the United Kingdom (Scott & Pitt, 2004; Salunkhe et al., 2005), Brazil (Pellegrino et al., 2006) and Denmark (Jelsbak et al., 2007). The Australian Epidemic Strain-2 (AES-2) (otherwise known as PII/P2) (O'Carroll et al., 2004) infects up to 40% of patients in three CF clinics in eastern Australia. The overrepresentation of a particular clone in a CF population can be due to its abundance in the environment (Jones et al., 2003) and/or due to epidemic spread (McCallum et al., 2001; Armstrong et al., 2003). AES-2 has not been found in the environment, suggesting person-to-person transmission. Many epidemic clones, including AES-2, also exhibit signs of increased virulence (Al-Aloul et al., 2004; O'Carroll et al., 2004).

The determinants of transmissibility and virulence in epidemic P. aeruginosa clones remain undefined, and there are no clear indications to date of conserved processes. The Liverpool Epidemic Strain (LES) has shown differential expression of quorum-sensing (QS)-regulated genes compared with reference P. aeruginosa strain PAO1 (Salunkhe et al., 2005), and the Manchester clone (MA) (Lewis et al., 2005) contains a putative transmission-linked gene island containing a bacteriophage cluster designated Pf4 (Webb et al., 2004), highly homologous to the Pf1 bacteriophage of P. aeruginosa. While acquisition of novel genes may play a role in mucosal infectivity, the differences in the expression of genes common across all CF strains should not be overlooked. Recently, Mathee et al. (2008) compared the genomes of a chronically infecting infrequent clone (strain PA2192) and the MA abundant clone and found that each carries a comparatively modest number of unique ORFs, representing c. 9% and 1.3% of all ORFs, respectively. The authors infer that with over 90% of the genome being conserved, there is ample opportunity for expression of well-conserved genes to play a role in the increased infectivity of the MA clone.

An insight into the properties characterizing epidemic clones is needed to improve understanding of their increased infectivity and to promote the development of strategies for limiting their spread. This study provides a comparative transcriptomic analysis of planktonic and biofilm-grown AES-2 and infrequent P. aeruginosa CF clones and the reference P. aeruginosa strain PAO1 (Stover et al., 2000), aimed at identifying well-conserved genes with potential roles in the increased infectivity of AES-2.

Materials and methods

Bacterial strain selection

Microarray analyses were conducted on eight P. aeruginosa isolates grown planktonically and as a biofilm, comprising two AES-2 isolates from patients attending the adult CF Centre at The Prince Charles Hospital, Brisbane, Australia, and two AES-2 and four infrequent P. aeruginosa clones from the Adult CF Centre, Royal Prince Alfred Hospital, Sydney, Australia. A dominant clone is defined as one infecting at least three patients and showing less than a three-band difference on SpeI macrorestriction analysis according to the criteria of Tenover et al. (1995). An infrequent clone is defined as one with more than a three-band difference from other clones and infecting less than three patients. Planktonic and biofilm-grown samples of P. aeruginosa strain PAO1 (ATCC 16592) were microarrayed for comparison. Patients were pair-matched for forced expiratory volume (FEV1), age and gender, with all patients having been infected with P. aeruginosa for at least 4 years, and with the same strain (as defined by SpeI macrorestriction digest, followed by PFGE) for at least 1 year. (Age and gender: two female AES-2: 26.3±1.1 years vs. two female infrequent clonal: 25.8±3.3 years; two male AES-2: 24.2±2.7 years vs. two male infrequent clonal 22.2±1.7 years.)

Bacterial growth methods

Luria–Bertani broth (LB) was used for planktonic growth because it has been widely used as a nonspecialized growth medium for planktonic P. aeruginosa transcriptomics in both CF and non-CF studies (Schuster et al., 2003; Salunkhe et al., 2005; Waite et al., 2005; Alvarez-Ortega & Harwood, 2007). In accordance with the strategies adopted for transcriptome expression in other studies (Wagner et al., 2003; Hentzer et al., 2005), cells for planktonic growth were harvested at the mid log phase (OD600 nm=0.5±0.05) while biofilm-grown cells were harvested at 72 h. Growth curves were used to determine mid-log growth in planktonic culture. For biofilm growth, mid-log phase cells were inoculated (1 : 100) into 800 mL of LB in a Centers for Disease Control and Protection bioreactor. To obtain sufficient RNA for transcriptomic analysis, glass slides were attached to the rods to increase biofilm coverage. The assembly was incubated (37 °C water bath) and stirred (100 r.p.m.). At 72 h, biofilms were washed off with ice-cold 1 × phosphate-buffered saline (PBS) and pelleted (3 min/5000 g/4 °C). The pellet was washed in 1 × PBS and resuspended in RNAprotect (Qiagen). In follow-up studies, twitching motility was measured after 48-h growth following a stab inoculation of 0.8% (w/v) LB agar plates (Beatson et al., 2002).

RNA extraction and cDNA synthesis

Cells were treated with RNAprotect, and total RNA was extracted using the RNeasy Mini kit (Qiagen). The RNA concentration was determined by absorbance at 260 nm, with a minimum 500 ng μL−1 required to proceed to cDNA synthesis. RNA quality and the presence of residual DNA were checked by formaldehyde agarose gel electrophoresis.

cDNA was synthesized, fragmented and labelled as per the Affymetrix technical protocol. cDNA was purified using the MinElute Purification kit (Qiagen), fragmented with DNaseI, and 3′-end-labelled using the GeneChip® DNA Labelling Reagent (Affymetrix). DNA fragmentation was quality-checked using a Bioanalyser, and cDNA suitability was assessed with a ‘test3’ array (100 housekeeping genes), before hybridization to the P. aeruginosa PAO1 array (Affymetrix). Hybridization conditions were induced at 50 °C for 16 h at 60 r.p.m., followed by washing and scanning of the chip (Affymetrix GeneChip Scanner 3000 at 532 nm for excitation and 570 nm for emission). CEL and CHP files were generated using the scanner program gcos (Affymetrix).

Data analysis

Microarray data were analysed using bioconductor (Gentleman et al., 2004). Data normalization used the robust multiarray average method incorporating probe-level background-correction, quantile normalization and linear extraction of a final expression measure for each gene per array, which was used to determine differential expression by the empirical Bayes method (Smyth, 2004). The false discovery rate method was controlled to reduce false positives. A positive B-statistic was used as a guide for statistically significant differential expression. Data were combined with the latest information from the P. aeruginosa annotation project at http://v2.pseudomonas.com.

Microarray replicates

The AES-2 and infrequent clonal clones were treated as biological replicates in a two-group differential-expression analysis. Two isolates were arrayed as biological replicates to assess biological variability at the culture level (same isolate, but different culture, RNA extraction and microarrays): one planktonic AES-2 (isolate 9 as per GEO submissions below) was arrayed in duplicate, and one planktonic infrequent clone (isolate 8) was arrayed in triplicate. Substitution of different biological (culture) replicates had little effect on the prediction of differentially expressed genes. Pair-wise comparisons of the planktonic AES-2 replicate data gave R2=0.74, while the three infrequent clone replicates gave R2 values of 0.59, 0.62 and 0.65.

Two technical replicates (same isolate, culture and RNA extraction, different microarrays) of two biofilm AES-2 (isolates 9 and 10) and two planktonic infrequent clones (isolates 6 and 7) were arrayed for technical variability. Fluorescence value comparison for all genes showed no significant difference. The microarray data are available on the Gene Expression Omnibus (GEO) website: http://www.ncbi.nlm.nih.gov/projects/geo (accession numbers: GSE10304 and GSE6122).

Follow-up work

Microarray validation

Array data for eight genes (algP, azu, exsA, lasI, oprD, oprF, pscJ and rbsA), chosen for high differential expression under all conditions, and association with virulence and/or known function, were validated by quantitative SYBR-green-RT-PCR on cDNA synthesized from microarray RNA. Reverse transcription (RT), using 50 U SuperScriptII RT (Invitrogen) and 1 μg total RNA, was carried out as per the manufacturer's protocol (Invitrogen).

Gene detection-PCR

PCR was used for the detection of PA0979: 979F (5′-CCGAGTTCAAACGAGAG-3′)-979R (5′-CGAGTTCGTCCGACATC-3′); 979F2 (5′-GAACAGCAGAAGATCCAGGAA-3′) and 980R (5′-CAATCTCATCCGATCCTCC-3′) and the pilA pilin gene (Spangenberg et al., 1995).

Results and discussion

Transcript levels averaged 79% for planktonic and 84% for biofilm isolates, comparable to other studies (Ochsner et al., 2002; Wagner et al., 2003). Pseudomonas aeruginosa PAO1 was used as the primary reference, as with other P. aeruginosa pathobiology studies. However, because strain PAO1 was wound infection-derived, comparisons were also made with infrequent clones from chronically infected CF patients representing the same ecological niche and similar infection stage. The microarray expression ratios for the eight selected genes, carried out on both planktonic and biofilm-grown cells (Materials and methods), showed a good correlation (Fig. 1) with the quantitative RT-PCR ratios (correlation coefficient: R2=0.6044).

Figure 1

Correlation plot of microarray and quantitative RT-PCR (qRT-PCR) fold values. Correlation plot for validation of microarray data, comparing the fold values obtained for eight genes (algP, azu, exsA, lasI, oprD, oprF, pscJ and rbsA), by microarray (MA) and quantitative SYBR-green-RT-PCR (qRT-PCR). All genes were chosen for high differential expression under all conditions, association with virulence and/or known function, and were tested in duplicate under both planktonic and biofilm-growth conditions.

In analysing infrequent clones from different chronically infected CF patients, we were aware of the inherent diversity and the likelihood that variations in the selective pressures of individual lungs may affect the mechanisms by which individual isolates adapt to long-term survival. Array analysis cannot distinguish between genes ‘switched off’ and genes ‘absent’ due to strain diversity or mutation. In defining differentially expressed genes here as those differentially expressed across all isolates in each group, we have identified a core set of genes that may hold clues to the increased infectivity of AES-2.

The most interesting findings emerged from comparisons of planktonic and biofilm AES-2. A large number of genes were downregulated in AES-2 biofilms compared with planktonic AES-2 (Table 1 and Supporting Information, Table S1). Adaptation of P. aeruginosa to long-term survival in the CF lung typically involves downregulation of genes involved in acute infection such as proteases and the Type III secretion system (T3SS), and upregulation of genes sustaining biofilm formation, such as regulators of oxidative stress response and antibiotic efflux (Waite et al., 2005). In contrast to this expectation, exsA(PA1713), the master activator of exoenzyme S (ExoS) and the central regulator of T3SS in P. aeruginosa (Hogardt et al., 2004), and prcG(PA1705), another regulator of the T3SS, were both more highly expressed in AES-2 biofilms than in planktonic AES-2 (7.5 × and 4.2 ×, respectively) (Table 1) and the average expression of all T3SS genes was 2.3 × higher in AES-2 biofilms than in planktonic AES-2 (see GEO microarray data). The higher expression of T3SS-related genes in planktonic growth may contribute to the pathogenesis of AES-2. However, the possibility that these findings reflect technical factors related to the model systems used in this study cannot be excluded.

View this table:
Table 1

Gene Expression in Pseudomonas aeruginosa AES-2 – Genes of known function

GeneBiofilm AES-2 vs. planktonic AES-2Biofilm AES-2 vs. Biofilm ICPlanktonic AES-2 vs. planktonic ICPlanktonic AES-2 vs. PAO1Biofilm AES-2 vs. PAO1Description
Upregulated genes of known function
PA0852 cbpD6.8Chitin-binding protein CbpD precursor
PA1172 napC5.0Cytochrome c protein
PA1175 napD3.9Periplasmic nitrate reductase protein D
PA1245 aprX4.5Membrane protein
PA1246 aprD2.7Alkaline protease secretion protein D
PA1247 aprE2.3Alkaline protease Secretion protein E
PA1248 aprF2.7Alkaline protease Secretion protein F
PA1249 aprA6.0Alkaline metallo-proteinase
PA1250 aprI2.7Alkaline protease inhibitor
PA1705 pcrG4.2Type III secretion regulator
PA1713 exsA7.5Transcriptional regulator
PA2300 chiC14.3Chitinase
PA2386 pvdA22.9Pyoverdine
PA2396 pvdF5.4Pyoverdine synthetase F
PA2620 clpA2.4ATP-binding protease
Downregulated genes of known function
PA0432 sahH−8.5S-adenosyl-l-homocysteine hydrolase
PA0447 gcdH−9.8Glutaryl CoA dehydrogenase
PA0792 prpD−4.1Propionate catabolic protein
PA0852 cbpD−18.2Chitin-binding protein CbpD precursor
PA0958 oprD−5.9−11.5Outer membrane porin protein OprD
PA1432 lasI−6.0Autoinducer synthesis protein LasI
PA1581 sdhC−3.0Succinate dehydrogenase subunit C
PA1582 sdhD−2.7−1.8Succinate dehydrogenase subunit D
PA1777 oprF−7.0Outer membrane protein OprF precursor
PA1793 ppiB−5.4Peptidyl-prolyl-cis-trans isomerase
PA1871 lasA−7.0−3.6LasA protease precursor
PA1947 rbsA−4.2Ribose transport protein
PA1985 pqqA−4.1Pyroloquinoline quinone biosynthesis protein A
PA2001 atoB−4.1−4.1Acetyl-CoA acetyltransferase
PA2025 gor−3.2Glutathione reductase
PA2300 chiC−42.0Chitinase
PA2386 pvdA−6.3Pyoverdine
PA2396 pvdF−8.8Pyoverdine synthetase F
PA2532 tpx−3.2Thiol peroxidase
PA2622 cspD−4.6Cold-shock protein CspD
PA2623 icd−10.5Isocitrate dehydrogenase
PA3049 rmf−53.1Ribosome modulation factor
PA3477 rhlR−5.2Transcriptional regulator RhlR
PA3478 rhlB−14.2Ramnosyltransferase chain B
PA3479 rhlA−15.3Ramnosyltransferase chain A
PA3531 bfrB−8.3Bacterioferritin
PA3622 rpoS−7.4Sigma factor RpoS
PA3724 lasB−16.1Elastase
PA4481 mreB−2.5Rod-shape determining protein
PA4922 azu−7.5−1.9Azurin precursor
PA5018 msrA−2.9Methionine sulfoxide reductase
PA5253 algP−7.5Alginate regulatory protein AlgP
PA5255 algQ−2.3Alginate regulatory protein AlgQ
PA5288 glnK−5.7Nitrogen regulatory protein P-II 2
PA5429 aspA−4.7Aspartate ammonia lyase
Non-P. aeruginosa PAO1 genes
Pae L37109cds110.611.2Type IV pilin (pilA)
Pae L81176cds315.67.2Flagellar cap protein (fliD)
Pae L81176cds53.97.7Flagellar protein
PaeAF241171cds302.2Probable transcriptional activator (M. leprae)
  • Genes of known function that showed statistically significant differential-expression (B-statistic >0), between strain PAO1, AES-2 and infrequent clonal isolates grown planktonically and as a biofilm. The P-value has been corrected using the false discovery rate and only genes with a B-statistic >0 and having above-threshold fluorescence after normalization are shown.

  • IC, Infrequent clones.

AES-2 biofilms, infrequent clone and PAO1 biofilms all downregulated QS regulator rhlR and effector genes rhlA, rhlB and lasB. Interestingly, expression of QS regulator lasR remained unchanged in AES-2 biofilms. Some QS-regulated genes were selectively downregulated in AES-2 biofilms including PA0432 sahH (S-adenosyl-l-homocysteine hydrolase), used by P. aeruginosa to convert SAH to the homocysteine required for synthesis of the QS signal autoinducer-2 (Walters et al., 2006), and PA1777 oprF (outer membrane protein F). Thus, the effects of different expressions of the QS regulators remain unclear. The QS-regulated chitinase chiC(PA2300) (Folders et al., 2001) and the associated chitin-binding protein cbpD(PA0852), found in most clinical CF isolates, were significantly upregulated in planktonic AES-2 compared with infrequent clones, but subsequently significantly downregulated in AES-2 biofilms. cpbD was also upregulated in planktonic cultures of the chronic Liverpool epidemic strain LES431 compared with the non-CF isolate LES400 (Salunkhe et al., 2005). CpbD mediates attachment to chitin-containing substrates and protects P. aeruginosa against the proteolytic activity of elastase and also putatively aids biofilm formation (Folders et al., 2000). The downregulation of these genes in AES-2 biofilms indicates that they are no longer essential once the biofilm is established.

Eight QS-regulated genes were significantly upregulated in planktonic AES-2 compared with infrequent clones, notably the alkaline protease system genes (PA1245-50). The ability of AES-2 isolates from chronic infection to express proteases is consistent with findings from our phenotypic studies reported previously (Tingpej et al., 2007). LES431 also showed significant upregulation of alkaline protease genes compared with LES400 and strain PAO1 (Salunkhe et al., 2005). These genes were downregulated in AES-2 biofilms, but to a lesser extent than in the infrequent clones (see GEO microarray data).

Also contrary to expectation was the finding that AES-2 biofilms downregulated several enzymes essential in protecting bacteria against oxidative damage compared with infrequent clones, including (PA2025-glutathione reductase, PA2532-thiol peroxidase, PA3529-a probable peroxidase and PA5018-methionine sulphoxide reductase) (Table 1 and Table S1). Bacteria typically require protection of oxidative stress enzymes to prevent potentially lethal DNA damage; however, downregulation of oxidative stress genes has been shown to be a putative indicator of hypermutability in CF strains, allowing them to better survive adverse conditions, and PA3529 was previously downregulated in biofilm cultures of CF isolates (Driffield et al., 2008). Also significantly downregulated in AES-2 biofilms, but not the infrequent clone or PAO1 biofilms, were the proteins algP and algQ, essential for alginate production and mucoidy. This suggests that alginate production is not essential for AES-2 once the biofilm structure has been established.

Genes upregulated in AES-2 biofilms compared with P. aeruginosa strain PAO1 included napC(PA1172) and napD(PA1175). napC encodes the tetrahaeme membrane protein NapC, the electron donor for periplasmic nitrate reductase NapD (Hasegawa et al., 2003). This increased transcription of napC clusters may reflect anaerobically increased nitrate turnover in AES-2 biofilms compared with strain PAO1. AES-2 napC and napD were also slightly upregulated (1.3 ×), in comparison with infrequent clones.

Interestingly, both planktonic and biofilmAES-2 expressed pilA while infrequent clones did not (Table 1: non-PAO1 genes). In fact, the average AES-2 pilA expression was within the top 25% of upregulated genes of the AES-2 genome. PCR analysis showed that AES-2 isolates had a Group III Type IV pilA (L37109) (Table 1: non-PAO1 genes) rather than the Group I pilA found in strain PAO1 (PA4525) and the majority of CF P. aeruginosa strains (Kus et al., 2004) including those in this study. In subsequent PCRs using Group III-specific pilA primers, AES-2 yielded PCR products of size consistent with Group III pilA, while the infrequent clones yielded PCR products consistent with Group I pilA. Sequencing of the PCR product from two AES-2 isolates (GenBank accession number: 858318) showed 97% and 96% amino acid identity, respectively, with the Group III pilA of P. aeruginosa G7/G9 (ID L37109) and P. aeruginosa UCBPP-PA14, The MA clone and the virulent clinical strain PA14 also have Group III Type IV pili (Askiyan et al., 2008). Strains with Group III or Group V, but not Group I or Group II Type IV pili, have been shown to express large amounts of surface pili (Askiyan et al., 2008). Loss of twitching motility has been reported in CF isolates from chronic infection (Head & Yu, 2004); however, in our subsequent studies, twitching motility was evidenced in all AES-2 isolates, but not in any of the infrequent clones, which corresponds with our pilA expression data. Further studies are needed to establish whether the high expression of Group III Type IV pili confers an advantage for AES-2 during initial attachment to CF lung epithelium or sputum.

Planktonic AES-2 showed enhanced expression of iron acquisition genes pyoverdine pvdA(PA2386) and pyoverdine synthetase pvdF(PA2396) (22.9 × and 5.4 ×, respectively, Table 1) compared with infrequent clones; however, expression was significantly downregulated (−6.3 × and −8.8 ×) in AES-2 biofilms, which showed expression similar to infrequent clone biofilms. Alkaline protease expression facilitates siderophore-mediated iron uptake by P. aeruginosa (Kim et al., 2006); thus, enhanced expression of pyoverdine in planktonic culture may be due to the increased alkaline protease activity. It is possible that the infrequent clones utilize a different sequestration pathway to AES-2, because they had significantly higher expression (8.3 × ) than AES-2 of the B subunit of bacterioferritin, bfrB(PA3531), another major iron-storage protein in bacteria. Interestingly, two genes with functions in aerobic respiration, sdhC(PA1581) and sdhD(PA1582), were downregulated in planktonic AES-2 compared both with strain PAO1 and the infrequent clones. Succinate dehydrogenase D (sdhD) appears to be downregulated under iron-limited conditions, as shown in PAO1 (Wilderman et al., 2004), however, no such conditions were used in this study.

Our data showed that while AES-2 isolates are resistant to the glycylcycline-type antibiotic tigecycline and β-lactam antibiotic carbapenem (not shown), this resistance must have occurred along different pathways to the infrequent clones. The transposase-related gene PA0979, part of the MexCD-OprD efflux pump regulatory gene nfxB involved in glycylcycline resistance (Dean et al., 2003; Dumas et al., 2006), was not expressed in any AES-2 isolate, and all tested negative for PA0979 by PCR. In contrast, all infrequent clones contained the expected 267 bp band (results not shown). This may well be a feature of AES-2 isolates in general. Similarly, the outer membrane porin protein oprD (PA0958), associated with resistance to carbapenem in CF isolates, was downregulated in AES-2, but not in the infrequent clones.

In summary, our findings suggest that differential expression of well-conserved P. aeruginosa genes plays a role in the enhanced infectivity of AES-2. Some of the properties associated with AES-2, such as high expression of alkaline protease genes and cpbD and the possession of the Group III Type IV pilin, are shared with other CF clonal strains. Analysis of isolates from acute infections in younger CF patients will provide further insight into the properties of AES-2. Because P. aeruginosa is capable of considerable genetic exchange, it is likely the genes not present on the PAO1 array also contribute to the transmissibility of AES-2.

Statement

Microarray data access: Gene Expression Omnibus (GEO) accession numbers: GSE10304 and GSE6122.

Supporting Information

Additional Supporting Information may be found in the online version of this article:

Table S1. Microarray analysis of AES-2, showing genes of probable, putative or unknown function that showed statistically significant differential-expression (B-statistic >0), between strain PAO1, AES-2 and infrequent clones grown planktonically and as a biofilm.

Acknowledgements

We thank Drs Claire Wainwright, Melanie Syrmis and Mark O'Carroll of the Adult Cystic Fibrosis Centre, The Prince Charles Hospital, and the Department of Respiratory Medicine, Royal Children's Hospital, Brisbane, Australia, for two AES-2 specimens and clinical test results, and Dr Mark Elkins and Carmel Moriarty at the Cystic Fibrosis Centre, Royal Prince Alfred Hospital, Sydney, Australia, for AES-2, infrequent clone specimens and clinical test results.

Footnotes

  • Editor: Craig Winstanley

References

View Abstract